HANDBOOK OF STABLE ISOTOPE ANALYTICAL TECHNIQUES
Front Cover: Cover photograph shows a multiple collector inductively coupled plasma mass spectrometer (ICP-MS). Copyright of photo by IRMM, Retieseweg, 2440 Geel, Belgium. Publication kindly permitted by IRMM and acknowledged by the editor.
HANDBOOK OF STABLE ISOTOPE ANALYTICAL TECHNIQUES VOLUME 1
Pier A. de Groot editor Economic Geology Research Institute, School of Geosciences, University of the Witwatersrand, Johannesburg, South Africa and Delta Isotopes Consultancy, Pastoor Moorkensstraat 16, 2400 Mol-Achterbos, Belgium (present address)
2004
ELSEVIER Amsterdam - Boston - Heidelberg - L o n d o n - New York - Oxford Paris - San Diego - San Francisco
- Singapore
- Sydney
- Tokyo
ELSEVIER B.V. Sara Burgerhartstraat 25 P.O. Box 211,1000 AE Amsterdam, The Netherlands
ELSEVIER Inc. 525 B Street, Suite 1900 San Diego, CA 92101-4495 USA
ELSEVIER Ltd The Boulevard, Langford Lane Kidlington, Oxford OX5 1GB UK
ELSEVIER Ltd 84 Theobalds Road London WC1X 8RR UK
9 2004 Elsevier B.V All rights reserved. This work is protected under copyright by Elsevier B.V, and the following terms and conditions apply to its use: Photocopying Single photocopies of single chapters may be made for personal use as allowed by national copyright laws. Permission of the Publisher and payment of a fee is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. Special rates are available for educational institutions that wish to make photocopies for non-profit educational classroom use. Permissions may be sought directly from Elsevier's Rights Department in Oxford, UK: phone (+44) 1865 843830, fax (+44) 1865 853333, e-mail:
[email protected]. Requests may also be completed on-line via the Elsevier homepage (http://www.elsevier.com/locate/ permissions). In the USA, users may clear permissions and make payments through the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; phone: (+ 1) (978) 7508400, fax: (+ 1) (978) 7504744, and in the UK through the Copyright Licensing Agency Rapid Clearance Service (CLARCS), 90 Tottenham Court Road, London W1P 0LP, UK; phone: (+44) 20 7631 5555; fax: (+44) 20 7631 5500. Other countries may have a local reprographic rights agency for payments. Derivative Works Tables of contents may be reproduced for internal circulation, but permission of the Publisher is required for external resale or distribution of such material. Permission of the Publisher is required for all other derivative works, including compilations and translations. Electronic Storage or Usage Permission of the Publisher is required to store or use electronically any material contained in this work, including any chapter or part of a chapter. Except as outlined above, no part of this work may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without prior written permission of the Publisher. Address permissions requests to: Elsevier's Rights Department, at the fax and e-mail addresses noted above. Notice No responsibility is assumed by the Publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein. Because of rapid advances in the medical sciences, in particular, independent verification of diagnoses and drug dosages should be made. First edition 2004 Reprinted 2005
Library of Congress Cataloging in Publication Data A catalog record is available from the Library of Congress. British Library Cataloguing in Publication Data A catalogue record is available from the British Library. ISBN: 0 444 51114 8
Working together to grow libraries in developing countries www.elsevier.com I www.bookaid.org www.sabre.org : L S E V I ER
~ooK 1...... "..... iAID ...... 1
~abrc" l - ~"O t l l ~ d ~ t
"
101
@ The paper used in this publication meets the requirements of ANSI/NISO Z39.48-1992 (Permanence of Paper). Printed in The Netherlands.
Those incredible deltas
They measured day and night On their super isotope machine And ended that analytical fight With zero errors, as you imagine Then ... they had actually begun Without switching the bloody thing on. C. Brenninkmeijer
This Page Intentionally Left Blank
Dedication to S.M.F. Sheppard
Simon M. F. Sheppard
viii
Dedication to S.M.F. Sheppard
Dedication The two volumes on Stable Isotope Techniques are dedicated to my colleague and former tutor Simon M.F. Sheppard. He was the person who introduced me, during my study period in the Centre de Recherches P6trographiques et G6ochimiques, at Vandoeuvre-l~s-Nancy, France, to the wonderful world of stable isotopes applied to earth science, especially in the field of geochemistry and mineral deposits. It was, on reflection a critical period in developing my career, motivating me to understand the full ramifications of stable isotope chemistry. Since that time my interest in stable isotope chemistry has constantly progressed. If Simon had not stimulated me to work in this field, I would not have organised this publication presenting the essential guidelines of analytical procedures and techniques for the measurement of stable isotope ratios in samples for a large number of scientific disciplines. Simon Sheppard, born April 16, 1938, at Salisbury, England, was educated at the University of Cambridge, England, where he gained a BA Tripos in Natural Sciences in 1962. His first studies in isotope geochemistry began at McMaster University in Canada, under the guidance of Henri Schwarcz gaining his PhD in 1966 on "Stable isotope (C, O) geochemistry of metamorphic rocks". From 1966 to 1968 Simon was Research Fellow at the California Institute of Technolog~ Pasadena, U.S.A., where he first worked with Samuel Epstein and Hugh Taylor Jr. Here he completed two important publications on porphyry copper mineralization linked to fluid-rock interactions (Sheppard et al., 1969, 1971). Simon was then appointed to an Assitent Professorship from 1968 until 1970, at the University of Texas, Austin, U.S.A. He then returned to the Uinted Kingdom to develop stable isotope research facilities at the Scottish Universities Research and Reactor Centre (SURRC), East Kilbride, Scotland. In 1976 Simon was invited to the Centre de Recherches P6trographiques et G6ochimiques (CRPG), Vandoeuvre-16s-Nancy, in France, by Bernard Poty with the intention of coupling fluid inclusion research with stable isotope systematics. This scientific challenge could not be resisted and so Simon left SURRC to join CRPG sponsored by CNRS, the French national research organization, ultimately becoming the 'Directeur' of CRPG in 1980 for ten years. In 1991 Simon was appointed to the newly founded l~cole Normale Sup6rieure (ENS) in Lyon, France, where he continued his stable isotope research work until his recent retirement in September 2003. During his scientific career, Simon Sheppard has become highly regarded as a stable isotope geochemist in Europe and worldwide. His main area of interest is in the origin and transfer of aqueous, carbonic, sulfurous and silicate fluids in the mantle, crust and hydrosphere. Related to this theme is the use of fluid inclusion analytical methods pioneered by Alain Weisbrod and Bernard Poty at CRPG. Simon always has been fascinated by the application of stable isotope studies to economic geolog~ particularly material transfer by fluids and stable isotope exchange processes (fluid -
Dedication to S.M.F. Sheppard
ix
mineral interactions). His extensive experience in different scientific areas is also shown by his contribution of two chapters in Volume L Part 1, Review and Discussion on Developments in Stable Isotope Analytical Technologies, where Simon is the co-author of Chapter 2, on "Analysis of Fluids from Clays and Sediments", and author of Chapter 4:6, on "The Experimental Determination of Isotopic Fractionations'. This is the first compilation on experimental methods on this subject to be published. Besides possessing a substantial list of frequently cited publications Simon has been Associate Editor for Precambrian Research (1977- 1989), Geochimica Cosmochimica Acta (1979 - 1985), and is still active as Associate Editor for Lithos (from 1984) and Chemical Geology (from 1991). During my contact with Simon over the years, I learned that Simon is an 'amateur expert' in architecture. I well remember his request to see the 'Rietveld - house' in Utrecht, the Netherlands (now protected by UNESCO as world cutural heritage site). Living in France Simon enjoys fine wine, and has developed a curiosity for goat cheese. Currently he is writing a book on French cheeses which should be published in the near future.
References Sheppard S. M. F., Nielsen R. L. & Taylor H. P. Jr. (1969) Oxygen and hydrogen isotope ratio of clay minerals from porphyry copper deposits. Econ. Geol., 64: 755-777.
Sheppard S. M. F., Nielsen R. L. & Taylor H. P. Jr. (1971) Hydrogen and oxygen isotope ratios in minerals from porphyry copper deposits. Econ. Geol., 66: 515-542.
This Page Intentionally Left Blank
Introduction
INTRODUCTION
The honour for the creation of this two volume book series on stable isotope analytical techniques must be given to James O'Neil. At the very start of this project I did not realize that he, together with Zachary Sharp, had begun writing a book on a very similar subject. Originally James O'Neil and Zachary Sharp were among the first authors I invited to write a chapter for my book series. It was in this way that I discovered about their own book project, which already had been in progress for some time. It was Zachary Sharp in particular who convinced and motivated me, supported by James O'Neil, to continue my project, after he recognized that the books I had in mind had a different approach and aimed at a far larger range of elements (isotopes) than their own. The prospective publication by James O'Neil and Zachary Sharp will have the form of a textbook and is concentrated more on a purely geochemical field. I am grateful for their "open-minded attitude". At the start, the aim of these books was to concentrate on stable isotope analytical methods of purely geochemical interest, but very soon it became clear it was not easy to draw a line between purely geochemical methods with geological applications and methods used in other fields of science. Considering the interest isotope chemists working in other disciplines could have in a handbook on analytical techniques made me decide to extent the contents of these books, and to include a wide range of other disciplines, where stable isotope analysis is used for different purposes. Disciplines of interest, besides geochemistry, are: anthropology, archaeology, agronom~ atmospheric science, biology, bio(geo)chemistr~ climatology, drug detection methodology, ecology~ environmentology, food science or alimentology (e.g. detection of adulterat-ion), forensic science, hydrolog}~ marine science, medical science, metallurgy~ meteoritic science, metrolog~ nutrition studies, palaeontolog~ petrochemistry, pharmacology, planetary science, and toxicology. The range of elements chosen is purely arbitrary but dependent on the choice of the invited authors. Since new methods were actually developed for Br isotope analysis, by analogy with C1 isotope methods, I decided the element of Br to be the upper limit for these books. This also considerably extended the range of basic techniques. For instance, stable isotope analysis of elements such as Li, B, Mg, Si, K, Ca, Ti, Cr, Fe, Ni, Cu, Zn, Ga, Ge, and to some extent Se, concentrated historically more on solid source mass spectrometry rather then gas source mass spectrometry. Other tools for isotope measurement cross this classical boundary between "solid source mass spectrometry (MS) - gas source MS", such as secondary ionization mass spectrometry (SIMS), inductively coupled plasma mass spectrometry (ICP-MS), laser related mass spectrometry systems, fast atom bombardment mass spectrometry (FAB-MS), or nuclear activation analysis (NAA) techniques. Far less common techniques for stable isotope analysis, such as glow discharge mass spectrometry (GDMS), accelerator mass
xii
Introduction
spectrometry (AMS), and some forms of optical spectrometry, are developed and some already are commonly used, while others are at different stages of development. For example, optical analytical methods were developed for gas samples as a faster but less precise method, parallel with the mass spectrometry technique. Stable isotope analytical methods were developed soon after the discovery of the existence of isotopes. For example, in 1934 there was the discovery of deuterium by Harold Urey for which he received the Nobel prize. Early techniques were based on determination of isotopic ratios by densimetric, gravity, electric resistivity, pycnometry type of methods. The development of a usable mass spectrometer, by Nier, and improved by McKinney and coworkers in the late 1940's - early 1950's, gave an important impuls for the use of stable isotope techniques in scientific studies. These early methods were generally complicated, time consuming procedures, and had relatively low precision and accuracy. First improvements were on precision of methods, and soon attempts were made to simplify preparation procedures. In early methods sample size was comparatively large, in the order of a few to 10's or 100's of mgs. Decrease of sample size was another aim, while improving analytical methods. Accuracy of methods is controlled by certified standard or reference materials (see Volume L Part 2, Chapter 40). Standardization of techniques and correct calibration methods are an important concern and need continuous attention to avoid comparison of isotopic values based on different or badly calibrated scales. Reduction of sample size was made possible first by use of so-called "static mass spectrometers" with a single inlet, thus avoiding the continuous pumping of samples while not being actively measured and basically reducing the needed gas volume by half. This technique is still in use at present, for example in stepped heating procedures (Volume L Part 1, Chapter 13) or in fluid inclusion analysis techniques in geological, geochemical, meteoritic or planetary studies. The development of secondary ionization mass spectrometry (SIMS) and laser techniques introduced the possibility of analysing in situ for specific type of samples. The first generation SIMS were limited in the elements on which stable isotopes could be measured caused by low mass resolution characteristics and precision was comparatively low. High resolution SIMS has recently been developed (e.g SHRIMP-II and Cameca ims-1270), increasing the number of isotopes which can be measured, and with improved precision compares with the first generation machines (see Volume L Part 1, Chapter 30). The introduction of laser technology, after the pioneering work by Ian Franchi, Douglas Rumble or Zachary Sharp for stable isotope measurement decreased the sample size considerably. Techniques for analyzing powdered or grain size samples or analyzing on a microscopic scale in an in situ mode are available with their own specific characteristics and limitations. Single grain or spot analysis inside grains is made possible with the laser and the SIMS techniques. First laser types used were infra-red (IR) Nd:YAG and CO2 lasers. Newer developments are with ultra-violet (UV) lasers such
Introduction
xiii
as excimer lasers, quadrupoled Nd-YAG lasers, Ar-F or Kr-F2-Ne or Xe-C1 gas mixture lasers, and doubled frequency Cu-vapour lasers (Volume L Part 1, Chapter 20 or Volume II, Chapter 6-1.5.1). The development for measuring organic, fluid or solid samples is moving a different wa N into on-line systems (originally a converted elemental analyser (EA) was used) with oxidizing, reducing or pyrolyses reactors, eventually in a combined order depending on the sample material and the gas of interest for isotopic measurement, and with application of a carrier gas (generally He, seldomly H2 or N2 are considered; Ar may introduce problems in the ion source of a MS by sputtering effects) in a socalled continuous flow (CF) system to transport reaction gases through the system. The carrier gas may contain 02 for oxidation purpose in combustion (oxidizing) systems. Addition of gas chromatographs (GC's) for separation or purification of sample materials a n d / o r for purification of effluent gases after reaction in the EA section of these on-line systems became a common feature (see alsoVolume I, Part 1, Chapter 8). The advantage of such techniques is the very small sample size needed, the high number of samples that can be analyzed in short time periods compared with classical methods, the possibility of automation of these systems, reducing labour intensity (and thus costs) for analyzing, and the option to combine the measurement of different effluent gases for different isotopic ratios in consecutive way. Moreover, with application of CF-IRMS techniques there is no need for vacuum conditions, as was the case in precursory techniques. Increasingly, special designed EA's are used in on-line systems for analysis of organic materials and fluids. New developments also include inorganic materials such as" phosphates, sulfates, and nitrates for oxygen, sulfur and/ or nitrogen isotopes. Automation, as mentioned earlier, is another trend in analytical techniques. A large group of materials are suitable for such automated systems (e.g. organic materials, carbonates, water or fluid samples) while other materials are not suitable (e.g. rock or mineral samples for O- and H-isotope analysis, including fluorination systems). Another trend is to combine measurement of several isotopes in a sample in an online system. Organic matter is suitable for such an approach, and combinations of some of H-, 0-, C-, S-, and N-isotopes can be applied. This places some constraints on the MS-side in such systems, where an MS must be able to jump from one m/z ratio to another quickly and to handle the measured peaks for samples and references, including background and other corrections by advanced computer automation. In the fast development of laser technology, tunable diode lasers must be mentioned for possible application on a number of materials by optical spectroscopic methods as a highly probable technique in the future (see also Volume L Part 1, Chapters 33 and 34 for optical systematics). Other analytical techniques and tools have been developed but details of all these developments are not given here in this introduction. As presented in the foregoing
xiv
Introduction
sections, the modern tendency is to analyse on smaller samples (including in situ analysis), preferably for a major part automated, and faster analysis per sample. Analysis of a number of different isotopic ratios on the same sample, generally in a sequential way, are applied either on separated phases in effluent gases produced from samples in reactors, or by sequential analysis of separated sample compounds, or by subsequent handling of the same sample with different treatments (e.g. gas equilibration methods). This handbook consist of two volumes: The first, edited volume contains two parts. Part I includes contributions presenting 'subjective' reviews on analytical techniques for specific stable isotopes or materials, reviews on stable isotope analysis by selected machines, descriptions of specialized and novel methods in stable isotope analytical techniques. Readers are guided to modern analytical techniques and are advised which techniques are the best to use for specific materials or conditions. Part 2 includes matters that are not strictly confined to analytical techniques themselves but related to analysis of stable isotopes, such as" views on the development of mass spectrometers and ion source stability, matters concerning isotopic scales, standards and reference materials, calibration and correction matters, a review on experimental isotopic fractionation determination and directives for setting up a laboratory. Appendices present the internet-based stable isotope discussion list named: 'isogeochem list', the also internet-based stable isotope fractionation calculator, and information on suppliers of stable isotope reference materials The second volume aims to present an encyclopedic overview of stable isotope analytical techniques in an 'objective' way. The chapters in this volume are intended to be complementary to the chapters in the first volume. In the second volume analytical techniques from historical times up to the most recent developments, are presented as a classical order of elements. Short descriptions of methods and diagrams of analytical devices are presented. Many classical techniques, of which several were never used in an operational form or became obsolete or forgotten, are included. Many of these older techniques formed the basis for present-day techniques. They also may improve the understanding of the development of analytical techniques which are used in preference today. Much of the experience from the old technology can be useful in applying to, or in constructing modern analytical systems.
Acknowledgement
I like to express my gratitude to all who were helpful to me during the preparation of these two book volumes. The list is far too long to mention everyone, but I hope that I will be forgiven if their names are not specifically highlighted. All help was important to me, unrelated to the size of this help. Finally, I like to thank all the reviewers, whose comments and suggestions considerably improved the quality of this publication.
xv LIST OF C O N T R I B U T O R S J.K. Aggarwal E.A. Atekwana G. Beaudoin G.E. Bebout M. Berglund S. Borella S.R. Boyd M.E. Bi~ttcher W.A. Brand T.D. Bullen K.L. Casciotti L.-H. Chan
C.C.Y. Chang M. Coleman
L. Dallai P. De Bi6vre J. Diemer T. Ding T. Durakiewicz H.G.M. Eggenkamp G.D. Farquhar
D.F. Ferretti M.L. Fogel K.S. Gan
Institute of Mineralogy, Mtinster University, Corrensstrasse 24, D-48149 M/~nster, Germany Department of Geology, Indiana University Purdue University, 723 W. Michigan Street, SL 122, Indianapolis, IN 46202-5132, USA D6partement de G6ologie et de G6nie G6ologique, Universit6 Laval, Qu6bec, GIK 7P4, Canada Department of Earth & Environmental Sciences, Lehigh University, Bethlehem, Pennsylvania 18015, USA European Commission- Joint Research Centre, Institute for Reference Materials and Measurements (IRMM), Retieseweg, 2440 Geel, Belgium Climate and Environmental Physics, Physics Institute, University of Bern, Sidlerstrasse 5, 3012 Bern, Switzerland CRPG-CNRS, 15 rue Notre Dames des Pauvres, B.P. 20, 54501 Vandoeuvreles-Nancy Cedex, France Max-Planck-Institute for Marine Microbiology, Department of Biogeochemistry, Celsiusstr.1, D-28359 Bremen, Germany Max-Planck-Institute for Biogeochemistry, PO Box 100164, 07701 Jena, Germany Water Resources Division, MS-420, U.S. Geological Survey, 345 Middlefield Road, Menlo Park, CA 94025, USA Water Resources Division, U. S. Geological Survey, 12201 Sunrise Valley Drive, Reston, VA 20192, USA Department of Geology & Geophysics, Louisiana State University, Baton Rouge, LA 70803-4101, USA Water Resources Division, U.S. Geological Survey, 345 Middlefield Rd, MS 434, Menlo Park, CA 94025, USA Postgraduate Research Institute for Sedimentology, University of Reading, UK, current address Center for Life Detection, Jet Propulsion Laboratory, California Institute of Technology, 4800 Oak Grove Drive, MS 183-301, Pasadena, CA 91109-8099, USA CNR-Instituto di Geologia Ambientale e Geoingegneria, Sez. Roma "La Sapienza", P.le Aldo Moro 5, 00185 Rome, Italy Institute for Reference Materials and Reference Measurements, JRC-European Commission, B-2440 Geel, Belgium European Commission, Joint Research Center, Institute for Reference Materials and Measurements, Retieseweg, B-2440 Geel, Belgium Institute of Mineral Resources, Chinese Academy of Geological Sciences, Beijing 100037, P. R. China Los Alamos National Laboratories, Condensed Matter & Thermal Physics Group, Mailstop K764, Los Alamos, NM 87545, USA Department of Geochemistry, Faculty of Earth Sciences, Utrecht University, P.O.Box 80021, 3508 TA, Utrecht, The Netherlands Environmental Biology Group, Research School of Biological Science, Institute of Advanced Studies, Australian National University, GPO Box 475, Canberra, ACT 2601, Australia National Institute of Water and Atmospheric Research, Wellington, New Zealand Carnegie Institution of Washington, Geophysical Laboratory, 5251 Broad Branch Rd., NW, Washington, DC 20015, USA Environmental Biology Group, Research School of Biological Science, Institute of Advanced Studies, Australian National University, GPO Box 475, Canberra, ACT 2601, Australia
xvi S. Ghelli H.A. Gilg I. Gilmour J.-P. Girard
M. GrSning C. Guillou
S. Halas K. Hashizume
V..M. Holland
J. Horita T.R. Ireland T.M. Johnson J.C. Johnston H.R. Karlsson C. Kendall E. Kerstel H. Kipphardt R.V. Krishnamurthy H.R. Krouse K. Leckrone
C. Ldcuyer
M. Leuenberger R. Lucchini I,
Lyon
SPIN, via Tamagno, 3, 42048 Rubiera (RE), Italy Lehrstuhl ffir Ingenieurgeologie, Technische Universit~it M~inchen, Arcisstr. 21, 80290 Mfinchen, Germany Planetary and Space Sciences Research Institute, Open University, Milton Keynes, Buckinghamshire, MK7 6AA, UK BRGM, Department Analysis and Mineral Characterization, BP6009, 45060 Orldans cedex 2, France International Atomic Energy Agency, Agency's Laboratories Seibersdorf, Isotope Hydrology Laboratory, A-1400 Vienna, Austria European Commission, Joint Research Centre, Institute for Health and Consumer Protection, Physical and Chemical Exposure Unit, 1-21020 Ispra (VA), Italy Uniwersytet Marii Curie-Sklodowskiej, Instytut Fizyki, Pracownia Spektrometrii Mas, P1. M. Curie-Sklodowskiej 1, 20-031 Lublin, Poland Centre de Recherches Pdtrographiques et Gdochimiques, 15 Rue NotreDame des Pauvres, B.P. 20, 54501 Vandoeuvre-16s-Nancy Cedex, France, and Department of Earth & Space Sciences, Osaka University, Toyonaka, Osaka 560-0043, Japan European Commission, Joint Research Centre, Institute for Health and Consumer Protection, Physical and Chemical Exposure Unit, 1-21020 Ispra (VA), Italy Chemical Sciences Division, Oak Ridge National Laboratory, P.O.Box 2008, MS 6110, Oak Ridge, TN 37831-6110, USA Research School of Earth Sciences, The Australian National University, Canberra ACT 0200, Australia Geology Department, 245 Natural History Bldg., MC-102, University of Illinois, Urbana-Champaign, Urbana, IL 61820, USA Iterations, P.O. Box 590805, San Francisco, California 94159, USA Department of Geosciences, and Department of Chemistry and Biochemistry, Texas Tech University, Box 1053, Lubbock, TX 79409, USA Water Resources Division, U.S. Geological Survey, 345 Middlefield Rd, MS 434, Menlo Park, CA 94025, USA Center for Isotope Research, Department of Physics, University of Groningen, The Netherlands Bundesanstalt ffir Materialforschung und -pr~fung (BAM), D-12200 Berlin, Germany Department of Geosciences, Western Michigan University, 1187 Rood Hall, Kalamazoo, MI, 49008, USA Department of Geology and Geophysics, University of Calgary, 2500 University Drive NW, Calgary, Alberta, T2N 1N4, Canada Department of Chemistry, Roosevelt University, 430 S. Michigan Ave., Chicago, IL 60605, USA Laboratoire CNRS UMER 5125, "Paldoenvironnements & Paldobiosph6re", Batiment ,~ Geode ,, Campus de la Doua, Universit6 Claude Bernard Lyon 1, 27-43 Boulevard du 11 Novembre 1918, 69622 Villeurbanne cedex, France Climate and Environmental Physics, Physics Institute, University of Bern, Sidlerstrasse 5, 3012 Bern, Switzerland Institut de Mindralogie et Pdtrographie, BFSH-2, CH-1015 Lausanne, Switzerland Department of Earth Sciences, The University of Manchester, Manchester, M13 9PL, UK
xvii B. Marty
B. Mayer W. Meier-Augenstein G. M~not
G. Michalski J. Miller C.R. Qu~tel M. Rehk/imper
F. Reniero
S. Rezzi
M. Ricci S.J. Sadofsky M. Saurer B. Schnetger
M.A. Sephton Z.D. Sharp S.M.E Sheppard R. Siegwolf S.R. Silva D. Smith
P. Spanel
H. Le Q. Stuart-Williams
B.E. Taylor P.D.P. Taylor
Centre de Recherches P~trographiques et G~ochimiques, 15 Rue NotreDame des Pauvres, B.P. 20, 54501 Vandoeuvre-l~s-Nancy Cedex, France, and Ecole Nationale Sup6rieure de G6ologie, Avenue du Doyen Roubault, 54501 Vandoeuvre-l~s-Nancy Cedex, France Department of Geology and Geophysics, University of Calgary, 2500 University Drive NW, Calgary, Alberta, T2N 1N4, Canada Queen's University Belfast, Environmental Engineering Research Centre, David Keir Building, Belfast, BT9 5AG, UK Institute of Geology, University of Bern, Baltzerstrasse 1, 3012 Bern, Switzerland, present address Woods Hole Oceanographic Institution, Department of Geology and Geophyics, Massachusetts, USA University of California, Department of Chemistry and Biochemistry, San Diego, CA 92039, USA NOAA Climate Monitoring and Diagnostics Laboratory, Boulder, CO, USA, and CIRES, University of Colorado, Boulder, CO, USA European Commission, Joint Research Center, Institute for Reference Materials and Measurements, Retieseweg, B-2440 Geel, Belgium Institute of Isotope Geology and Mineral Resources, ETH Ztirich, NO C61, CH-8092 Ztirich, Switzerland, and Institute of Mineralogy, Mtinster University, Corrensstrasse 24, D-48149 Mtinster, Germany European Commission, Joint Research Centre, Institute for Health and Consumer Protection, Physical and Chemical Exposure Unit, 1-21020 Ispra (VA), Italy European Commission, Joint Research Centre, Institute for Health and Consumer Protection, Physical and Chemical Exposure Unit, 1-21020 Ispra (VA), Italy Department of Biological Sciences, University of Idaho, Moscow, ID 83844, USA Department of Earth & Environmental Sciences, Lehigh University, Bethlehem, Pennsylvania 18015, USA Paul Scherrer Institute, CH-5232 Villigen PSI, Switzerland Carl-von-Ossietzky University, Institute for Chemistry and Biology of the Marine Environment (ICBM), P.O. Box 2503, D-26111 Oldenburg, Germany Planetary and Space Sciences Research Institute, Open Universit3r Milton Keynes, Buckinghamshire, MK7 6AA, UK Department of Earth and Planetary Sciences Northrop Hall, Albuquerque, NM, 87131-1116, USA Laboratoire de Science de la Terre and CNRS-UMR 5570, Ecole Normale Sup~rieure de Lyon, 46 All6e d'Italie, 69364 Lyon, France Paul Scherrer Institute, CH-5232 Villigen PSI, Switzerland Water Resources Division, U.S. Geological Survey, 345 Middlefield Rd, MS 434, Menlo Park, CA 94025, USA Centre for Science and Technology in Medicine, School of Postgraduate Medicine, Keele University, Thornburrow Drive, Hartshill, Stoke-on-Trent, ST4 7QB, UK V. Cerm~ik Laboratory, J. Heyrovsky Institute of Physical Chemistry, Academy of Sciences of the Czech Republic, Dolejskova 3, 182 23, Prague 8, Czech Republic Research School of Biological Sciences, Australian National University, GPO Box 475, Canberra City, ACT 2601, Australia Geological Survey of Canada, Ottawa, Ontario K1A OES, Canada Institute for Reference Materials and Reference Measurements, JRC-European Commission, B-2440 Geel, Belgium
xviii M.A. Teece P. Therrien M.H. Thiemens
S. Toyoda
S. Valkiers B.H. Vaughn R.M. Verkouteren
S. Wankel D.M. Wayne J.W.C. White E Wombacher S.C. Wong
N. Yoshida
C.-E You
State University of New York - College of Environmental Science and Forestry, 1 Forestry Drive, Syracuse NY 13210, USA D6partement de G6ologie et de G6nie G6ologique, Universit6 Laval, Qu6bec, GIK 7P4, Canada Department of Chemistry and Biochemistry, University of California, San Diego, La Jolla, California 92093-0356, USA Department of Environmental Chemistry and Engineering, Frontier Collaborative Research Center, and SORST, Japan Science and Technology Corporation (JST), Kawaguchi, Saitama, Japan Institute for Reference Materials and Reference Measurements, JRC-European Commission, B-2440 Geel, Belgium INSTAAR, University of Colorado, Boulder, CO, USA Surface and Microanalysis Science Division, Chemical Science and Technology Laboratory, National Institute of Standards and Technology, Gaithersburg, MD 20899, USA Water Resources Division, U.S. Geological Survey, 345 Middlefield Rd, MS 434, Menlo Park, CA 94025, USA NMT-15, Pit Disassembly and Nuclear Fuels Technologies, MS E 530, Los Alamos National Laboratory, Los Alamos, NM 87545, USA INSTAAR, University of Colorado, Boulder, CO, USA Institute of Mineralogy, M~inster University, Corrensstrasse 24, D-48149 M~inster, Germany Environmental Biology Group, Research School of Biological Science, Institute of Advanced Studies, Australian National University, GPO Box 475, Canberra, ACT 2601, Australia Department of Environmental Chemistry and Engineering, Frontier Collaborative Research Center, and Department of Environmental Science and Technology, Tokyo Institute of Technology, 4259 Nagatsuta, Midori-ku, Yokohama 226-8502, Japan, and SORST, Japan Science and Technology Corporation (JST), Kawaguchi, Saitama, Japan Department of Earth Sciences, National Cheng Kung University, Tainan, Taiwan, ROC
CONTENTS VOLUME I Dedication Introduction List of Contributors PART 1 Chapter I
Chapter 2
Chapter 3 Chapter 4
Chapter 5
Chapter 6
Chapter 7
Chapter 8
Chapter 9
Chapter 10
Chapter 11
Chapter 12
Chapter 13
Chapter 14
Stable Isotope Analysis of Water and Aqueous Solutions by Conventional Dual-Inlet Mass Spectrometry -J. Horita & C. Kendall Conventional and Less Conventional Techniques for Hydrogen and Oxygen Isotope Analysis of Clays, Associated Minerals and Pore Waters in Sediments and Soils - H.A. Gilg, J-P. Girard & S.M.F. Sheppard Techniques for Stable Isotope Analysis of Fluid and Gaseous Inclusions - L. Dallai, R. Lucchini & Z.D. Sharp Selected Ion Flow Tube Mass Spectrometry (SIFT-MS) and Flowing Afterglow Mass Spectrometry (FA-MS) for the Determination of the Deuterium Abundance in Water Vapour - P. Spanel & D. Smith Natural Abundance 2H-NMR Spectroscopy. Application to Food Analysis - S. Rezzi, C. Guillou, E Reniero, V.M. Holland & S. Ghelli Mass Spectrometric Techniques for the Determination of Lithium Isotopic Composition in Geological Material -Chan, L.-H. Thermal Ionization Mass Spectrometry Techniques for Boron Isotopic Analysis: A Review -You, C.-F. GC and IRMS Technology for 13C and 15N Analysis on Organic Compounds and Related Gases - W. Meier-Augenstein Preparation of Ecological and Biochemical Samples for Isotope Analysis - M.A. Teece & M.L. Fogel Extraction of Dissolved Inorganic Carbon (DIC) in Natural Waters for Isotopic Analyses - E.A. Atekwana & R.V. Krishnamurthy Compound Specific Isotope Analysis of the Organic Constituents in the Murchison Meteorite - M.A. Sephton & I. Gilmour A New Method for the Isotopic Examination of Sub-Milligram Carbonate Samples, Using Sulphamic Acid (NH2.SO3H) at Elevated Temperatures - H. Le Q. Stuart-Williams Determination of the Abundance and Stable Isotopic Composition of Trace Quantities of C and N in Geological Samples: The Practice and Principles of Stepped-Heating at High Temperature Resolution - S.R. Boyd Stable isotope measurements of atmospheric CO2 and CH4 - B.H. Vaughn, J. Miller, D.F. Ferretti & J.W.C. White
Page vii - ix xi - xiv xv - xviii
1 - 37
38 - 61 62 - 87
88 - 102
103 - 121
122- 141
142- 152
153- 176
177 - 202
203 - 228
229 - 236
237- 255
256 - 271 272- 304
xx Chapter 15 Preparation and Analysis of Nitrogen-bearing Compounds in Water for Stable Isotope Ratio Measurement - C.C.Y. Chang, S.R. Silva, C. Kendall, G. Michalski, K.L. Casciotti & S. Wankel Chapter 16 615N Analyses of Ammonium-Rich Silicate Minerals by Sealed-Tube Extractions and Dual Inlet, Viscous-Flow Mass Spectrometry - G.E. Bebout & S.J. Sadofsky Chapter 17 Nitrogen Isotopic Analyses at the Sub-Picomole Level Using an Ultra-low Blank Laser Extraction Technique - K. Hashizume & B. Marty Chapter 18 Mass Independently Fractionated Ozone in the Earth's Atmosphere and in the Laboratory -J.C. Johnston & M.H. Thiemens Chapter 19 Site-specific Nitrogen Isotope Analysis in N20 by Mass Spectrometry - S. Toyoda & N. Yoshida Chapter 20 Fluorination Methods in Stable Isotope Analysis -Bruce E. Taylor Chapter 21 Oxygen Isotope Analysis of Plant Water Without Extraction Procedure - K.S. Gan, S.C. Wong & G.D. Farquhar Chapter 22 Oxygen Isotope Analysis of Phosphate - C. L6cuyer Chapter 23 Pyrolysis Techniques for Oxygen Isotope Analysis of Cellulose M. Saurer & R. Siegwolf Chapter 24 Sample Homogeneity and Cellulose Extraction from Plant Tissue for Stable Isotope Analyses S. Borella, G. M6not & M. Leuenberger Chapter 25 Analytical Methods for Silicon Isotope Determinations T. Ding Chapter 26 Procedures for Sulfur Isotope Abundance Studies - B. Mayer & H.R. Krouse Chapter 27 Direct Measurement of the Content and Isotopic Composition of Sulfur in Black Shales by Means of Combustion-Isotope-Ratio-Monitoring Mass Spectrometry (C-irmMS) M.E. BOttcher & B. Schnetger Chapter 28 Summary of Methods for Determining the Stable Isotope Composition of Chlorine and Bromine in Natural Materials - H.G.M. Eggenkamp Chapter 29 Selenium, Iron and Chromium Stable Isotope Ratio Measurements by the Double Isotope Spike TIMS Method T.M. Johnson & T.D. Bullen Chapter 30 SIMS Measurement of Stable Isotopes T.R. Ireland Chapter 31 Stable Isotope Analysis by Multiple Collector ICP-MS - M. Rehk~imper, F. Wombacher & J.K. Aggarwal Chapter 32 Different Isotope Ratio Measurement Applications for Different Types of ICP-MS: Comparative Study of the Performance Capabilities and Limitations C.R. Qu6tel & J. Diemer Chapter 33 Isotope Ratio Analysis Techniques using Photoionization as a Source of Ions - I. Lyon Chapter 34 Isotope Ratio Infrared Spectrometry E. Kerstel -
-
-
-
-
-
-
-
305 - 347
348 - 360
361 - 374
375 - 389 390 - 399 400 - 472 473 - 481 482 - 496 497- 506
507- 522 523 - 537 538- 596
597 - 603
604 - 622
623-651 652-691 692 - 725
726- 745
746- 758 759- 787
xxi Chapter 35 Glow Discharge Mass Spectrometry: F u n d a m e n t a l s and Potential Applications in Stable Isotope Geochemistry - D.M. Wayne Chapter 36 The Use of Molecular Sieves in Stable Isotope Analysis - H.R. Karlsson Chapter 37 Introduction to Isotope Dilution Mass Spectrometry (IDMS) - M. Berglund
PART 2 Chapter 38 Mass Spectrometer H a r d w a r e for Analyzing Stable Isotope Ratios - W.A. Brand Chapter 39 Techniques of Ion Current Stabilization in Isotope Ratio Mass Spectro-
Chapter 40 Chapter 41
Chapter 42 Chapter 43
Chapter 44 Chapter 45 Chapter 46
Chapter 47
metry - S. Halas & T. Durakiewicz International Stable Isotope Reference Materials - M. GrOning The Nature and Role of Primary Certified Isotopic Reference Materials: A Tool to U n d e r p i n Isotopic M e a s u r e m e n t s on a Global Scale - P.D.P. Taylor, P. De Bi6vre & S. Valkiers Traceability in Isotopic M e a s u r e m e n t s - H. K i p p h a r d t Strategies and Practicalities in the Production and Use of Gas Isotope Standard Materials - R.M. Verkouteren Data Corrections for Mass-Spectrometer Analysis of SO2 - M. Coleman O x y g e n Isotope Corrections in Continuous-Flow M e a s u r e m e n t s of SO2 - K. Leckrone & M. Ricci Experimental M e a s u r e m e n t of Isotopic Fractionation Factors and Rates and M e c h a n i s m s of Reaction - S.M.F. Sheppard Laboratory Set-Up for GC-MS and Continuous-Flow IRMS - W. Meier-Augenstein
788 - 804 805 - 819 820 - 834
835 - 856
857 - 873 874- 906
907- 927 928 - 943
944- 956 957- 970 971 - 991
992- 1037 1038 - 1042
A p p e n d i x A Isogeochem list A p p e n d i x B The Web Stable Isotope Fractionation Calculator - G . Beaudoin & P. Therrien A p p e n d i x C Suppliers of Reference Materials
1043 - 1044
References
1053- 1181
Subject Index
1183- 1234
1045 - 1047 1048 - 1052
This Page Intentionally Left Blank
PART 1 Review and Discussion on Developments in Stable Isotope Analytical Technologies A Guidance to Modern Analytical Methods and Related Matters
This Page Intentionally Left Blank
Handbook of Stable Isotope Analytical Techniques, Volume 1 P.A. de Groot (Editor) 9 2004 Elsevier B.V. All fights reserved.
CHAPTER 1 Stable Isotope Analysis of Water and Aqueous Solutions by Conventional Dual-Inlet Mass Spectrometry Juske Horital & Carol Kendall2 1 Chemical Sciences Division, Oak Ridge National Laboratory, P.O.Box 2008, MS 6110, Oak Ridge, TN 37831-6110, USA 2 Water Resources Division, U.S. Geological Survey, 345 Middlefield Rd, MS 434, Menlo Park, CA 94025, USA e-mail:
[email protected] ;
[email protected]
1.1 Introduction The foundation of various analytical methods for the stable isotope composition of water and other aqueous samples (natural abundance, 1H. 2H (D) - 99.985" 0.015 atom%, and 1 6 0 " 170 9180 -- 99.762 90.038 90.200 atom%) was established during the Manhattan Project in the U.S.A., when large amounts of heavy water were produced for nuclear reactors (see Kirshenbaum, 1951, for a detailed account). From early on, there was great interest in the oxygen and hydrogen isotopic compositions of water, because they are the ideal tracers of water sources and reactions. The increased analytical precisions made possible by the subsequent development of modern gas-source isotope-ratio mass spectrometers with dual-inlets and multi-collectors, have caused the proliferation of new analytical methods and applications for the oxygen and hydrogen isotopic compositions of water. These stable isotopes have found wide applications in basic as well as applied sciences (chemistry, geology, hydrology, biology, medical sciences, and food sciences). This is because water is ubiquitous, is an essential and predominant ingredient of living organisms, and is perhaps the most reactive compound in the Earth. In this article, we review recent developments and refinements of analytical methods for preparing waters and other aqueous samples of different origins for the measurement of the oxygen and hydrogen isotopes by conventional dual-inlet, dynamic gas-source isotope-ratio mass spectrometry. Earlier review articles include those by Gonfiantini (1981), Wong & Klein (1986), Platzner (1997), and Coplen (2001). During the past decade, other in-situ (laser-ablation and secondary ion), static, quadrupole, and continuous-flow mass-spectrometric techniques have been developing rapidly. We briefly discussed emerging techniques of continuous-flow mass-spectrometry because the same preparation methods are employed in both dual-inlet and continuous-flow mass-spectrometry.
2
C h a p t e r 1 - J. Horita & C. K e n d a l l
The size, nature (free water or physically/chemically bound water within a matrix), chemical composition (pH, salts, and other dissolved constituents), and isotopic abundance of aqueous samples vary widely depending on their type, origin, and history. Examples are hydrologic (water vapor, precipitation, surface waters, soil waters, groundwater, geothermal water), geologic (fluid inclusions, hydrous minerals, dissolved water in minerals), biological (plasma, urine, saliva, human milk, breath water, plant water), agricultural (juices, wine, milk) and laboratory (synthetic, experimental) fluids. Waters enriched (and sometimes depleted) in deuterium and/or 180 are used for hydrological and biological tracer studies, where isotopic effects and fractionation are not of concern. Some meteorites have extremely wide ranges of oxygen and hydrogen isotopic compositions, reflecting nucleosynthesis and planet-forming processes in the early stages of the universe. Most isotopic studies, however, deal with aqueous samples whose isotopic compositions are within the normal range of terrestrial isotopic compositions (-400%0 < 6D < +50%0 and -50%0 < 6180 < +10%o on the VSMOW-SLAP scale). Investigations are made of natural (biologic, geochemical, and physical) and laboratory processes that cause a variety of kinetic and equilibrium isotope effects (isotope fractionations). For such studies dealing with relatively small isotopic variations, precisions of < 1 - 2%0 (6D, 62H) and < 0.1 - 0.2%0 (6180) are required. In addition to the aspects regarding water/aqueous samples discussed above, the choice of an analytical method depends on many other factors and requirements (expertise/equipment, time/labor, and precision/accuracy). While the conventional analytical techniques developed for natural waters in the 1950's and 1960's are still among the most precise methods, recent developments are shifting toward simplified procedures or automated analyses of small-size samples. In particular, automated preparation systems (available from commercial mass spectrometer companies) connected to the mass spectrometer (i.e., "on-line" systems) not only reduce the time and labor required for a large sample throughput, but also eliminate systematic errors of human origin (Barrie & Prosser, 1996; Brand et al., 1996). Automation, of course, requires substantial capital investments other than the mass spectrometer to keep up with state-of-the-art techniques. With the increasing popularity of stable isotope techniques in many disciplines and the vast amount of isotopic data produced, data-correction procedures, quality control, and inter-laboratory standardization of isotopic data are becoming crucial issues.
1.2 Collection, transportation, extraction, and storage of water samples Great care must be exercised in the collection, transportation, extraction, and storage of water samples prior to their isotopic analysis with techniques described in the following sections. First, it is critical that the sample collected actually be representative of the feature of interest. For example, groundwater wells need to be pumped for a time sufficient to remove the stagnant water in the well before sampling, and samples from deep rivers should be depth-integrated or sampled at weirs or other convergence points. Samples do not need to be filtered or chilled to preserve the isotopic compositions, but samples may need to be poisoned by HgC12 to suppress microbial activity. The most serious problem is evaporation, which can be avoided by quickly transferring the sample into an appropriate container. For large-volume samples of
Stable IsotopeAnalysisof Water and Aqueous Solutions ...
3
hydrological, biological and agricultural fluids, the use of glass or high-density polyethylene bottles with tight caps is recommended. The bottle size should closely match the size of the sample (i.e., small dead volume); small amounts of water in big bottles can readily fractionate by isotope exchange with vapor. From our experience, caps with conical inserts or ones with Teflon liners are the most reliable. Glass bottles should not be filled entirely to the top if there is any chance of freezing or of large pressure changes (such as might be expected in under-pressured airplane luggage compartments) during transportation. One convenient and reliable way to transport large numbers of bottles is to put them back into the original cardboard trays, wrap the trays in bubble paper, and then put the trays in insulated ice chests or coolers (at room temperature), packed securely with plastic peanuts. Long-term storage in leaky containers or low-density polyethylene bottles results in the alteration of isotopic composition due to evaporation and diffusion (Stewart, 1981; Rozanski & Rzepka, 1991). It is reported that the isotopic composition of water stored in some high-density polyethylene bottles with a thin (2mm) wall changed over a period of 2 years, due apparently to diffusion through the wall (Stuiver et al., 1995). For long-term storage, waxing the caps (dip bottle tops in melted paraffin and store upside-down) or torchsealing the waters in glass ampoules should be considered. Samples of waters with known isotopic compositions should be stored along with the samples as a check on possible fractionation. Samples containing large amounts of organic matter (e.g., oil) may need to be stored in cool and dark locations to prevent degradation. Atmospheric water vapor in air can be sampled by opening a valve of a pre-evacuated flask or by circulating air through a flask for a few minutes. Alternatively) water vapor can be collected by slowly pumping air through a glass or metal trap cooled at -60~ or below (Craig & Gordon, 1965; Ehhalt, 1974; White & Gedzelman, 1984; Helliker et al., 2002). A filter may be installed upstream to prevent contamination by rain, snow, or ice. The flow rate, coolant temperature, and other aspects (geometry of a trap, glass bead filler to increase surface area, etc.) must be carefully considered to ensure quantitative trapping of water vapor, depending on the temperature and humidity of air. Various desiccants (silica-gel, molecular sieves, Mg(C104)2, etc.) may be used to collect water vapor without a coolant, but they invariably contribute background water, when heated for the recovery of water. Soil and plant samples should also be kept in the same types of tight containers and bottles with small dead-volumes until extraction and analysis. Water contained in soil and plants can be extracted by various methods (vacuum and azeotropic distillation, squeezing/pressure extraction, and centrifugation) prior to isotopic analysis. Comparison of several studies obtained by these different extraction methods highlight potential problems associated with each method, especially for samples with low water contents (Revesz & Woods, 1990; Ingraham & Shadel, 1992; Walker et al., 1994; Aragu~is-Aragu~is et al., 1995; Landon et al., 1999). Direct CO2 and H2 equilibration methods recently developed for soil and plants samples offer an alternative, promising approach as discussed below (Scrimgeour, 1995; Hsieh et al., 1998; McConville et al., 1999; Koehler et al., 2000).
4
C h a p t e r 1 - J. Horita & C. K e n d a l l
Brines, and agricultural and biological fluids contain high concentrations of dissolved salts and other compounds (sugar, alcohol, protein, etc.). In the decomposition methods for the isotopic analysis of waters discussed below, water must be extracted from these samples prior to isotopic analysis or as part of a single-step extraction/ reduction procedure. It is often very difficult to extract all water from these samples by means of vacuum or azeotropic distillation without potential isotopic exchange and fractionation, even at elevated temperatures. Centrifugation helps remove suspended materials from samples of blood, juice, and other substances. Some organic compounds (light hydrocarbons, oil, and tar) must be removed from natural waters because they can poison metal/chemical reagents used in the decomposition methods and can interfere with mass spectrometric measurements. The separation can be achieved by extracting them with water-insoluble solvents or adsorption on warm paraffin wax. CO2 and H2 equilibration methods are preferred methods because these methods require little pretreatment of aqueous samples. In addition to these analytical problems, the presence of certain kinds of dissolved constituents can change the measured isotopic composition of water, due to "the isotope salt effect", as discussed below. Geologic aqueous samples (fluid inclusions in minerals, water dissolved in minerals and glass, hydrous minerals) require the extraction of water prior to isotopic analysis by means of crushing, thermal decrepitation, and thermal heating (Godfrey, 1962; Roedder et al., 1963; Vennemann & O'Neil, 1993). Because these methods are bulk extraction techniques, the presence of different types of water (inclusion water and dissolved water in minerals) and different generations of fluid inclusions pose serious analytical problems (e.g., Kazahaya & Matsuo, 1985). Complete extraction of water is often a difficult task (e.g., Ihinger et al., 1994). Reactions occurring during extractions (e.g., adsorption of water to mineral surface, thermal reactions among C-O-H gases) could also alter the isotopic composition of inclusion water. The amount of water in these samples is often a limiting factor for isotopic analysis because conventional dual-inlet mass spectrometry requires water samples of a 10 tlmol (0.2 tlL). With decreasing sample size, contamination and memory effects from extraction and preparation systems become increasingly problematic. An extraction step, if necessary, is an important part of overall analytical procedure, and errors associated with these processes have to be critically evaluated. 1.3 Review of analytical methods
1.3.1 Hydrogen Isotopes Hydrogen gas (H2) is the preferred gaseous species for D / H ratio measurements in gas-source isotope ratio mass spectrometry, although other forms of gases (water, methane, acetylene, ethane, propanol) have been used with limited success. Many metals (U, Zn, Zn-CaO, W, Mg, Mn, Cr) and carbon have been tested for quantitative conversion (reduction) of water to H2 gas at elevated temperatures (400 - 1000~ the water-oxygen is converted to metal oxides or CO/CO2 (see reviews by Kirshenbaum, 1951; Wong & Klein, 1986). M + xH20 ~ xH2 + MOx
M: metal or carbon
[1.1]
Stable Isotope Analysis of Water and Aqueous Solutions ...
5
Quantitative conversion of the water to H2 is very important to avoid or minimize any potential isotopic fractionation. High concentrations of dissolved salts and compounds in brines, and in biological and agricultural fluids often interfere with the quantitative conversion reactions as discussed below. Conversion methods can be divided into (1) dynamic methods where multiple aliquots of water are reacted oneby-one with the metal reagent in some kind of reactor, and (2) static batch methods where water samples are reacted separately with a metal reagent in a closed vessel. A conceptually different method of analysis is to equilibrate the hydrogen isotopes between water samples and H2 gas. One main difference between conversion and equilibration methods is that the 6D values produced by conversion methods are composition values whereas the 6D (and 6180 values) produced by equilibration methods are activity values. This important distinction will be discussed further in section 1-4.
1.3.1.1 Dynamic conversion~decomposition methods Conventional U and Zn reactor methods
Among the metals, uranium (U) and zinc (Zn) have been most successfully used for D / H ratio measurements of natural waters in a dynamic conversion system (U: Bigeleisen et al., 1952; Stewart & James, 1981; Sajjad &Tasneem, 1983; Wong et al., 1984; Vaughn et al., 1998; Zn: Graft & Rittenberg, 1952; Friedman, 1953; Horibe & Kobayakawa, 1960; Schiegl & Vogel, 1970; Lyon & Cox, 1980; Dubois, 1985; Morse et al., 1993). These techniques, especially the uranium reduction method, can provide very high precision (lo = 0.2%o, Craig & Gordon, 1965), but require intensive labor and care because of the nature of largely manual procedure. Although dynamic reduction methods with uranium and zinc have been a standard method in the past, their use has significantly declined during the last decade. Recentl~ Cr (Gehre et al., 1996a, b), platinized Mg (Halas & Jasinska, 1996), and Mn (Tanweer & H a n , 1996; Shouakar-Stash et al., 2000) have also been reported as suitable reducing reagents for conversion. The basic procedure for using a uranium furnace is as follows (Bigeleisen et al., 1952)" uranium turnings (depleted in fissile 235U) are broken into small pieces, cleaned with HNO3 and distilled water, and packed tightly into a U-shaped or double-walled quartz tube, with quartz and/or copper wool loosely packed at each end (and perhaps in the middle) to prevent dispersal of the uranium oxide dust that is formed during the reaction (see Bigeleisen et al., 1952; Gonfiantini, 1981). The furnace is heated to 600 - 800~ in a vacuum line and the introduced water is passed once or more through the furnace by moving a liquid nitrogen dewar back and forth between small U-traps on either side of the furnace, and heating the glass line (by heat tape, heat gun, or torch). To ensure that the H2 gas introduced into the mass spectrometer is homogeneous, the gas may be mixed by use of a Toepler pump or a magnetically-operated rotor pump (Schiegl & Vogel, 1970). The U in a vessel should be replaced every few hundred analyses to avoid memory effects caused by the uranium oxides.
6
Chapter 1 - J. Horita & C. Kendall
The resulting H2 gas can be transferred to a reservoir or mass spectrometer inlet by expansion, automated Toepler pump, adsorption to charcoal with liquid nitrogen, or reaction with uranium metal at 80~ to form uranium hydride, which can be decomposed by heating to 700~ (Friedman & Hardcastle, 1970). For analyzing ~2 ~L water samples using the former method, several grams of charcoal (made from coconut shells) is put in glass sample vessels, and the charcoal is completely degassed at 350~ prior to use (J.R. O'Neil, pers. commun., 1999). H2 gas is transferred to the sample vessels with liquid N2, and then admitted into a mass spectrometer without heating the charcoal. There is about a 2%o fractionation in 6D values for H2 gas of normal sample size (a 100 ~mol). However, this does not pose a problem as long as H2 gas prepared from standard and sample waters are processed in the same way. It was observed that for small samples (ca. < 20~mol), 6D values of H2 gas transferred with charcoal became progressively lower (J.R. O'Neil, pers. commun., 1999).
On-line preparation methods Large sample throughput for D / H ratio measurements can be accomplished by (1) off-line, multi-sample preparation systems (section 1-3.1.2) and (2) automated, online preparation systems (i.e., connected to a mass-spectrometer). One early waterreduction unit using uranium metal was installed in the inlet system of a mass spectrometer so that the H2 produced was directly introduced to it (Nief & Botter, 1959). This design was further improved for multi-sample measurements (Thurston, 1970, 1971; Hartley, 1980; Thurston & James, 1984, Gehre et al., 1996b) and for simultaneous measurements of D / H and 1 8 0 / 1 6 0 with twin mass spectrometers (Hagemann & Lohez, 1978; Wong et al., 1984; Barrie & Coward, 1985). This method can be fully automated by the use of an autosampler (Cr: Brand et al., 1996; U: Vaughn et al., 1998). The Aqua SIRA (VG Isogas Ltd, UK) U-reduction system described by Wong et al. (1984) was designed for deuterium-enriched saline fluids derived from clinical and nutritional tracer studies. To reduce memory effects, each sample is injected and analyzed 6 - 20 times, with an average of 7 injections per sample. Samples and standards were injected sequentially (i.e., there is no dual inlet) with precisions of about 0.6 ppm (4%o) for deuterium-enriched samples, which is satisfactory for most biomedical studies. By analyzing a water reference between every 5 samples, precisions as good as 0.2%o for oxygen and hydrogen isotopes are obtainable (Hagemann & Lohez, 1978); however, Wong et al. (1984) report optimum precisions for natural abundance samples of 1.1%o and 0.4%o for hD and ~180, respectively, with their system. Although use of U reduction methods seemed to be on the wane, the successful development of a high-precision, automated, high-throughput U reduction system (Vaughn et al., 1998) may generate renewed interest in the method. Unlike most former methods, only small quantities of uranium are required: 4000 injections (350 400 samples) per 0.5g. During analysis, a 10 - 50 ~L aliquot of water (depending on the size of the sample loop used) is drawn from each septa-topped vial using a Gilson HPLC autosampler (Gilson, Middleton, WI, USA). The autosampler dispenses the water into a 6-port HPLC injection valve. When each sample is injected 4 times, and each vial in a set is reanalyzed twice (called two "tours" of the sample set), a reproduc-
Stable Isotope Analysis of Water and Aqueous Solutions ...
7
ibility of about 0.3%o can be achieved (Vaughn et al., 1998). The authors attribute the high precision of their method to maintaining a highly repeatable sample size (+ 1%) by using an HPLC injection valve (instead of using a syringe), keeping the fractionation during sample evaporation constant, devising methods to reduce and correct for memory effects, and monitoring changes in "machine slope" (%o-scale shrinkage) over time. A comparison of the precision and accuracy of this method with the H2-water equilibration method (described below) for 165 Antarctic water samples analyzed by both methods (Hopple et al., 1998), showed that the 6D values of 94% of the samples analyzed by the two methods differed by less than 3.5%o, and 6% of the values differed by as much as 5%0. The "cutoff" value of 3.5%0 was chosen by the authors because it represents the 3~J standard deviation of most 6D analysis methods. Chromium (Cr) has also been used as a reducing agent (Gehre et al., 1996a, b; Nelson & Dettman, 2001) with good analytical precisions (1%o), and Brand et al. (1996) successfully automated the method using an autosampler. Chromium can be used both for off-line preparation where the sample is collected in glass ampoules, or as an on-line method where 1 ~L water samples are injected directly into the Cr furnace, and the gas flows from there into the expansion bellows of the mass spectrometer system (Gehre et al., 1996b). The Cr reactor for on-line reduction consists of about 50 g of chromium powder (Patinal < 0.3 mm, Merck) in a quartz furnace tube that is heated to 900-1000~ depending on sample type (Gehre et al., 1996b). The method is suitable for a variety of aqueous and organic compounds (gaseous or liquid); however, the temperature of the furnace and reaction time must be adjusted for each type. For example, water samples react very rapidly (80 sec from injection to starting the measurement) whereas methane requires 45 minutes for complete reduction (Gehre et al., 1996b). A disadvantage of the Cr method is that the reactors need to be refilled with Cr after 100 - 200 samples (Gehre et al., 1996b; Brand et al., 1996). Recently, Morrison et al. (2001) tested a Cr reactor method with continuous-flow isotope-ratio mass spectrometry. Water samples as small as 50nL (typically 3gL) was injected with an autosampler into a Cr reactor heated at 1050~ in a quartz tube, and H2 produced was carried with He gas into a mass spectrometer equipped with an electrostatic energy filter. They reported precision of _~0.5%0 with small to negligible memory effects. Recently, Shouakar-Stash et al. (2000) developed a Mn reaction unit installed on an inlet of an isotope-ratio mass spectrometer. Water, brines, natural gas, and organic solvents were reacted with Mn at 900~ for 20 sec (water) to 9 min (chlorinated solvents), and H2 produced was introduced directly to a mass spectrometer. Precision of ~ 1.02.0%0 was obtained. Using 5 gL of water per sample, 200 reductions can be obtained from 50 g of Mn. Advantages of Mn include availability~ low cost, lack of any pretreatment, and lack of memory effect for samples differing by less than about 200%0. Memory effects can be avoided by flushing with 5 gL of sample prior to an analysis. The most recent and very promising technique of water analysis is a high-temperature reduction/pyrolysis method in the presence of carbon, nickel (Ni) and other metals. In the presence of carbon, water reacts to form CO, (CO2), and H2. Ni and other metals appear to catalyze this reaction, and the production of CO2 is limited at
8
Chapter 1 - J. Horita & C. Kendall
high temperatures. Tobias et al. (1995) described a Ni-metal furnace to reduce water to H2 at 850~ for a continuous-flow mass spectrometer; a heated (330~ palladium (Pd) foil "filter" was used to separate the resulting H2 from the carrier gas and other impurities. This Ni reduction at + 1000~ is finding application in off-line preparation with Ni pyrolysis bombs, where H2 diffuses out to a quartz tube around the Ni tube (Gray et al., 1984; Motz et al., 1997). On-line measurements of D / H and/or 1 8 0 / 1 6 0 in waters, organic matter, or inorganic matter in continuous-flow mode, especially when automated by coupling a high temperature elemental analyzer (sometimes called a "pyrolysis" unit) or gas chromatograph (Begley & Scrimgeour, 1996, 1997; Werner et al., 1996; Koziet, 1997; Farquhar et al., 1997; Kelly et al., 1998; Br6as et al., 1998; Loader & Buhay, 1999; Kornexl et al., 1999a, b; Hilkert et al., 1999), are rapidly becoming popular methods in many laboratories for the analysis of small water samples. Recently, Sharp et al. (2001), and Eiler & Kitchen (2001) applied the high-temperature carbon pyrolysis/Hecarrier gas continuous-flow mass spectrometry to the determination of D / H and/or 180 / 160 ratios of nano- to pico-liters of water samples. Coupled with laser ablation or heating, it is becoming reality to analyze both D / H and 180/160 ratios of minute amounts of water in geologic samples (fluid inclusions and hydrous minerals) with unprecedented spatial resolution. There is an active ongoing debate about the best type of pyrolysis reactors and the optimum pyrolysis temperature for different types of samples. This method is described further in Chapter 3.
Memory effects
A major drawback of all dynamic systems for the conversion of water to H2 is memory effects caused by (1) incomplete removal of samples from a preparation system between analyses, (2) memory from a metal reactor itself, and (3) adsorption of water onto the walls of the system, followed by isotopic exchange between the following water sample and this reservoir of adsorbed water. Memory effects of up to 4% difference in 6D values between two consecutive samples are reported, depending on the type and amount of metal and water (Bigeleisen et al., 1952, Graft & Rittenberg, 1952; Hartley, 1980; Lyon & Cox, 1980; Wong et al., 1984). A detailed study of the memory effect was carried out by Morse et al. (1993), using a dynamic zinc reduction method with BDH zinc (BDH Chemicals Ltd., Poole, Dorset, UK)(Figure 1.1). Their study concluded that the furnace blank (0.11 ~mol H2) was insignificant compared to the amount of H2 contributed by adsorption of water from the previous sample (1.2 gmol); the adsorbed water contributed as much as 3 - 10% of the total H2 yield, depending on sample size, and that the amount of contamination was independent of original sample size. If the memory effect is caused by a constant amount of the previous sample (Morse et al., 1993), then the measured 6D of any sample can be satisfactorily corrected by applying a "blank" correction of this amount, having the gD value of the previous sample. The correct 6D value of the sample can be calculated by the mass-balance equation: 6Dsam = [~Dm (Hsam + Hblk)- hDprev Hblk] / Hsam
[1.2]
Stable Isotope Analysis of Water and Aqueous Solutions ...
9
Figure 1.1 - Memory effects of 6D values of SLAP after VSMOW from a Zn reduction unit. The memory effect increases with decreasing sample size, showing the contamination of 1.2 ~mol H2 from the previous samples regardless of sample size. After Morse et al. (1993)..
where (~Dmis the measured value of the sample, ~)Dprevis the measured value of the previous sample, and Hsam and Hblk are the amounts of hydrogen from the sample and blank (memory effect), respectively. Todd (1955) showed that there are two types of water in glass systems, surfacebound adsorbed water and water that diffuses out of the glass at temperatures > 400~ Surface-bound water is strongly bound and does not exchange with hydrogen gas (Graft & Rittenberg, 1952). Heating to 400~ was required to completely degas surface-bonded water, and heating to 100~ only removed about 25%; the amount of water bonded to glass increased with age of the glass and exposure to humid conditions (Todd, 1955). Other causes of memory effects include traces of water left in syringes from previous samples (syringes should be flushed several times with sample to avoid this) or from adsorbed atmospheric moisture, gas trapped in dead volumes of valves, and incomplete conversion in the reduction furnace. Memory effects can be reduced by heating the glass walls, reducing the internal surface area of the system, and by successive injections of the same sample. It appears that the continuous-flow isotope-ratio mass spectrometry with a He-carrier gas is less susceptible to the memory effects. 1.3.1.2 Static batch reduction methods Zn During the last decade, batch reduction methods using Zn in closed glass reservoirs gained popularity, and some methods were automated. Recently, similar methods with platinized Mg (Halas & Jasinska, 1996) and Mn (Tanweer &Han, 1996) were developed and are discussed below. The "grandfather" of all the many versions of zinc batch reduction methods in use today was the method of Coleman et al. (1982) who first identified a type of zinc shot (Analar zinc, BDH Chemicals Ltd., UK) that could
10
Chapter 1 - J. Horita & C. Kendall
be reliably used in a batch method. The basic method is as follows" zinc shot is sieved to retain the-30 to +60 mesh size grains, rinsed in 30% nitric acid, washed in distilled water, dried, and outgassed under vacuum at about 300~ Some users insist that the cleaned zinc be kept under vacuum (maybe even heated) prior to use, and others leave it exposed to the air for years with no problems. About 0.25 g of zinc is put in the bottom of a Pyrex vessel that has a stopcock with a Teflon plug (Figure 1.2); the vessel is connected to a vacuum line, evacuated, and outgassed at about 100~ The closed vessels are then removed from the line, the plugs of the stopcocks are removed, the vessels are filled with dry nitrogen (to avoid contamination with atmospheric moisture), and 1-10 iLL of water is injected onto the zinc using a syringe. The plugs are then inserted quickly, the closed vessels returned to the vacuum line, the water frozen with liquid N2, and the vessels thoroughly evacuated. The closed vessels are then put in a heating block at 450~ (m.p. - 419.5~ for about 30 minutes for quantitative reduction of the water to H2. The vessels are then attached to the inlet system of the mass spectrometer and analyzed. Modifications of the Coleman method abound. Most groups use modifications of the original Coleman vessel type (Stanley et al., 1984; Florkowski, 1985; Wong et al., 1987b; Penman & Wright, 1987; Tanweer et al., 1988), including a very large version designed for a 60-port automated gas handling system (Kendall & Coplen, 1985), and a vessel with a sidearm for soil samples (Turner & Gailitis, 1988). However, other groups report on the successful use of Pyrex or quartz/Vycor tubes (Kendall & Coplen, 1985; Sudzuki, 1987; Vennemann & O'Neil, 1993; Schimmelmann & DeNiro, 1993: Dem6ny, 1995; Yang et Figure 1.2 - Variations of the Coleman reaction vessel for batch Zn reduction method. Wall bubble for samples of biological fluids (urine, plasma, saliva, human milk, etc.)(Wong et al., 1987) and sidearm for porous samples (sand, soils, etc. ) (Turner & Gailitis, 1988). The original vessel by Coleman et al. (1982) had neither bubble nor side arm.
Stable Isotope Analysis of Water and Aqueous Solutions ...
11
al., 1996b; Karhu, 1997). Tanweer et al. (1988) suggested modifying the method by increasing the reaction temperature to 460~ and using 10 times the stoichiometric amount of zinc. Other workers use higher temperatures for glassware outgassing (up to 500~ zinc outgassing (up to 350~ and zinc oxidation (up to 500~ Water introduction methods also differ. Instead of removing the vessels to inject the water, water can also be introduced into the vessels on-line, either by direct injection through a septum, or by freezing water derived from decomposition of minerals or organic matter, or by breaking open microcapillaries containing water. Alternatively) the water can be syringed into hairpin-shaped microcapillaries which are loaded into the vessels, along with glass-enclosed iron bars, at the same time as the zinc; a magnet is used to raise and then drop the bar to break open the capillary just before putting the vessels into the block heater (Kendall & Coplen, 1985). Most workers have either used BDH zinc or a zinc available from Biogeochemical Laboratory at Indiana University (variably referred to as "Bloomington," "Indiana," or "Hayes" zinc)l. A comparison of the analytical precisions for water and human-fluid samples prepared using the "original" BDH zinc and Indiana zinc (Wong et al., 1992a) showed no significant difference; Vennemann & O'Neil (1993) also reported no significant differences in yields or 6D values with the two zinc types. The method can be used for many saline and hypersaline waters, including biomedical fluids (Wong et al., 1987b; Tanweer, 1993a), without prior distillation, but with some limitations (see section 1-4). It was observed that low-pH waters (e.g., acid lake water, water extracted from some hydrous minerals) react incompletely with Zn resulting in poor precisions.
Reactivity of Zn The choice of zinc reagent is critical to the successful use of the method. A number of groups have reported difficulties with a variety of zinc used for batch reduction (Coleman et al., 1982; Stanley et al., 1984; Kendall & Coplen, 1985; Florkowski, 1985; Wong et al., 1992a). In fact, many groups have problems with recent batches of BDH zinc (e.g., Wong et al., 1992a). Tanweer (1993a) reported that a new batch of BDH zinc with 0.3 - 1.5 mm grain size required larger amounts of zinc (+ 2.5 g) for 8 ~L water, compared to that of the original batch (0.25 g), and that reaction temperature must be increased from 460 ~ to 480~ One early comparison of the metal compositions of a number of types of analytical grade zinc found that the only significant difference among the brands was that BDH zinc contained more lead (Kendall & Coplen, 1985). Subsequent analyses of batches of BDH zinc have revealed other possibly important trace contaminants. A recent comparison of the trace contaminants in BDH and SHG zinc (Outokumpu Zinc Co., Kokkola, Finland) showed that BDH zinc contained much higher quantities of sodium (Karhu, 1997). The emission spectrographic analyses performed by Kendall & Coplen (1985) showed no difference in sodium contents among zinc types, perhaps because the detection limit for sodium was so high (0.1%) whereas it was at 0.1 to 100 ppm (depending on element) for the other metals.
Contact Dr. A. Schimmelmann, Dept. of Geol. Sci., Biogeochemical Lab., Indiana University, Bloomington, IN 47405-1405, U.S.A, e-mail:
[email protected].
12
Chapter 1 - J. Horita & C. Kendall
Several groups have tried to enhance the reductive properties of zinc by the addition of trace contaminants. Addition of I - 1.5 g lead granules to l g aliquots of otherwise unusable brands of cleaned zinc, or to already-used but recleaned BDH zinc (remelting not necessary), produced moderately acceptable results (Kendall & Coplen, 1985); replacements of the lead with tin or cadmium granules resulted in incomplete reduction. The so-called Indiana zinc has been produced since about 1985 by melting zinc with the addition of undisclosed contaminants, and then reforming the zinc mixture (Hayes & Baker, 1986; Hayes & Johnson, 1988), and is commercially available from Biogeochemical Laboratory of Indiana University (see the above address). This modest commercial effort was initiated in about 1986 to remedy the problem many laboratories were experiencing getting reliable results using different batches of BDH zinc. Karhu (1997) determined that the addition of trace quantities (~ 200 ppm) of sodium can also improve the reductive capability of an otherwise unusable zinc. Zinc-sodium mixtures were made by melting weighed aliquots of SHG zinc and various amounts of sodium (Merck) in a borosilicate tube under a constant flow of Ar (the Ar is necessary to prevent oxidation and eventual ignition of Na during heating); about 200 ppm sodium produced optimum results (reproducibility of + 0.7%o). After cooling, the cylinders were lathed and the thick lathings (1 - 1.5 mm thick) were broken into 15 mg chunks which could be used without further preparation. Samples were then prepared using a sealed-tube method (Kendall & Coplen, 1985; Vennemann & O'Neil, 1993), where about 30 mg of zinc per I mL of water was added to each 6 mm OD quartz tube (Karhu, 1997). Heating the tubes to 250~ prior to opening them into the mass spectrometer was found sufficient in reducing the small systematic bias resulting from adsorption of H2 by zinc (Karhu, 1997), in contrast to the temperature of 480~ recommended by Dem6ny (1995). Many groups now routinely make reliable zinc by adding Na to commercially available zinc (J.R. O'Neil, person, comm., 2000). Remelting zinc to improve the purity or to alter surface features has also been attempted by a few groups. SEM photographs of thoroughly cleaned zinc grains showed that BDH zinc had a much smoother and homogeneous surface (Kendall & Coplen, 1985; Tanweer, 1990) than other types examined. Several types of zinc were melted in long Vycor tubes in an attempt to produce zinc with identical surface configurations (Kendall & Coplen, 1985). When cooled, the impurities were concentrated at one end and removed. The solid, purified, zinc cylinders were lathed to produce fine turnings which were recleaned and tested; about I - 1.5 g of reformed zinc from two other sources behaved just as well as BDH zinc (Kendall & Coplen, 1985). Sublimation of an otherwise unusable zinc on the walls of a vessel or tube prior to reaction with water is another way to produce a usable zinc (Kendall & Coplen, 1985; Sudzuki, 1987). For example, melting of pieces of zinc wire (Nihon Denkyu Kogyo, Japan) under vacuum to decompose any zinc hydroxide, followed by introduction of the water and sublimation of the zinc onto the walls to enhance reaction rates, produces results comparable to the Coleman et al. (1982) method (Sudzuki, 1987). Noto & Kusakabe (1995) reported that Zn powder (no grain size information available) can also be used successfully after washing with nitric acid and degassing in vacuum at 250~ Hence, there is ample evidence that many other types of zinc are potentially
Stable Isotope Analysis of Water and Aqueous Solutions ...
13
usable but the reforming or sublimation of the zinc is time consuming. These experiments suggest that successful water reduction and zinc oxidation is highly dependent on surficial casting-produced features and not on just chemical composition.
Hydrogen reservoir in Zn There is some evidence for a source of exchangeable H in zinc. In one of the few published studies that made a direct comparison of 6D values produced by the conventional U furnace method and the zinc batch method (Kendall & Coplen, 1985), it was noted that one batch of BDH zinc produced 6D values for VSMOW and SLAP that were within 1 - 2%0 of the values obtained by uranium reduction, whereas three other BDH batches showed considerable scatter and produced 6D values that were 2 10%o lower than those determined with U reduction. Similar differences among different batches of Zn have been observed in many other laboratories. Schimmelmann & DeNiro (1993) also reported that gD values of water prepared with the Indiana zinc in Pyrex tubes at 500~ were up to 14%o lower than those of U-reduction method. The source of exchangeable H in the zinc could apparently be removed by melting the zinc under vacuum (Kendall & Coplen, 1985). Florkowski (1985) showed that there was a strong "amount effect " with some types of zinc (including one batch of BDH zinc), causing large (> 20%0) fractionations for 0.8 g of zinc; some zinc types and batches also showed significant differences among various fractions < 1.5 mm. Subsequent work by others at the IAEA lab (Tanweer et al., 1988) has confirmed the existence of small differences in 6D values for waters prepared with various weights of BDH zinc, but suggested that the large depletions in D seen in earlier studies might be an artifact of incomplete cleaning/degassing of the zinc. Furthermore, they concluded that the amount of BDH zinc (10 times the stoichiometric amount) and a conversion temperature of 460~ not the grain size, provides the optimum conditions for precise and accurate results (Tanweer et al., 1988). Lower stoichiometric ratios (0.1 to 0.2 mg zinc for 8 - 12 ~L water) produced low 6D values because of incomplete reaction, and high values also produced low gD values (Tanweer et al., 1988). The theory that the low 6D values associated with excess zinc was caused by absorption of hydrogen in the zinc (Florkowski, 1985) is supported by the results of some subsequent experiments (Tanweer et al., 1988; Dem6ny~ 1995). Schimmelmann & DeNiro (1993) demonstrated that "Bloomington zinc" leached with nitric acid released 44 gmol H2/g with 6D values of -163 to -229%0 upon heating to 450~ even after drying in vacuum at 100~ It has been suggested that one of the main reasons why many laboratories experienced poor reproducibility with BDH zinc is because of failure to follow the cleaning procedures outlined by Coleman et al. (1982) and explained in more detail in Tanweer et al. (1988), resulting in incomplete removal of adsorbed water and oxides on zinc surfaces (Tanweer, 1990). Ultrasonic cleaning before the acid wash has been suggested as a means to remove the fine particles (Turner & Gailitis, 1988) that can be a minor factor in producing lower 6D values than obtained using coarser fractions (Tanweer et al., 1988). Although they only report data from a single batch of BDH zinc, Tanweer et al. (1988) suggest that proper cleaning of the zinc eliminates the variations in performance that had been observed with different batches of BDH zinc.
14
C h a p t e r 1 - J. Horita & C. K e n d a l l
Hydrogen reservoir in the glasses The presence of a source of H on and/or within Pyrex glass has been demonstrated by several experiments. Pyrex vessels, preheated to 100 ~ - 200~ under vacuum, and then filled with H2 to a pressure of 33 KPa (25~ with a 6D of-122%o required 8 days of continuous heating at 430~ to reach a steady-state composition of about-175%o; pretreated vessels filled with H2 with a 6D of-690%o showed an increase to -665%o in 24 h, but had not reached a steady-state composition in 8 days (Kendall & Coplen, 1985). These results strongly suggest a source of hydrogen within the Pyrex with a 6D value intermediate between the two isotopic compositions of H2. These degrees of isotopic exchange were found regardless of whether the vessels had previously been baked for several hours with hydrogen of-122 or -689%o, suggesting that the reservoir was large and the exchange rate was slow. Turner & Gailitis (1988) also measured the extent of gaseous exchange with the Pyrex vessels by degassing vessels at 100~ adding H2 with a known 6D at 19 KPa (25~ and baking the vessels at 450~ for 2 hours. The heating resulted in a shift of 0.4 - 2%0, with no significant difference in results for vessels that had previously been exposed to H2 with 6D values showing a range of about 250%0. Use of Vycor (or quartz) vessels or sealed-tubes was found to cause significantly less isotopic exchange than Pyrex vessels or tubes by several thorough studies (Kendall & Coplen, 1985; Vennemann & O'Neil, 1993; Karhu, 1997), supporting the existence of a sizable reservoir of exchangeable hydrogen in borosilicate glass. However, it should be noted that because in many of the experiments described in the paragraphs above the glass was not heated to 400~ (the temperature apparently required to completely degas surface-bonded water, according to Todd, 1955) between experiments, it is possible that some of the results attributed to exchangeable hydrogen in the glass might be due to surface-bonded water. The degree of exchange is correlated with the surface to volume ratio and time; H2 in 6 or 9 mm sealed-tubes is more exchanged than H2 in large vessels at a constant pressure (Kendall & Coplen, 1985). Vycor and quartz behaved identically and showed much less exchange with the H2 of samples than Pyrex (Kendall & Coplen, 1985; Sudzuki, 1987; Vennemann & O'Neil, 1993; Karhu, 1997). Hence, quartz or Vycor are an obvious, but expensive, choice for a relatively trouble-free sealed-tube technique.
Considerations for routine-analysis The effect of hydrogen isotope exchange with Zn reagents and/or glass during the reaction and storage is to decrease the difference between the 6D of waters (i.e., "shrink the %o-scale")(Kendall & Coplen, 1985; Turner & Gailitis, 1988). Storage of reacted (and unreacted) samples in Vycor tubes for 18 months did not affect the 6D of the gas (the reacted samples were NOT reheated before analysis), making Vycor tubes ideal for researchers who stockpile samples (Kendall & Coplen, 1985). This experiment is in sharp contrast to the findings of Dem6ny (1995) who reported that heating quartz or Pyrex tubes to 480~ immediately prior to introduction of the H2 into the mass spectrometer was necessary to eliminate a systematic bias in the 6D that he attributed to adsorption of H2 into the zinc (Hayes zinc), and the somewhat similar findings of Karhu (1997) who reported that heating of the tubes to 250~ was needed to remove a systematic bias with the sodium-spiked SHG zinc. These results suggest
Stable Isotope Analysis of Water and Aqueous Solutions ...
15
that the degree of apparent absorption into zinc may depend on zinc type. BDH zinc may be less subject to absorption; Kendall & Coplen (1985) and Vennemann & O'Neil (1993) reported no significant amount effect with BDH zinc. Nevertheless, if the same sample volumes and water/zinc ratios are maintained, a linear normalization correction will remove the bias caused by the use of borosilicates and any possible zinc absorption effects. In fact, many users of zinc batch methods may be completely unaware that their version of the method is routinely producing results that may be biased by as much as 10 - 20%o. They might only discover this fact if they compare their zinc-batch results with results obtained using other reduction methods in their own laboratory (which is, unfortunately, rarely done), send their breakseals to another laboratory for analysis, or wonder about apparent No-scale shrinkage. Since the degree of bias (fractionation) depends on the details of the preparation system used, when such gas samples prepared in one laboratory are analyzed in a different laboratory where the mass spectrometers are calibrated by using gas samples prepared using a slightly different version of the zinc method, the values must be normalized using water reference standards prepared along with the water samples (in the first laboratory). One of the keys to successful use of the zinc method is to maintain consistent analytical conditions. It is a very unforgiving method where seemingly small methodological changes can produce large problems with reproducibility. Users should follow procedures of a published method exactly~ or else make whatever modifications seem necessary~ test the modified method thoroughly, and once the method is working fine, change nothing. Even under ideal conditions in a laboratory where there are experienced workers with the zinc method, the fussy method occasionally produces ~D values that are 5 - 50%0 fractionated. Hence, it is recommended that whenever possible, samples be prepared in duplicate and/or any meteoric-derived samples that deviate significantly from the GMWL (Global Meteoric Water Line) be considered for reanalysis. The standard zinc batch method has been modified for the analysis of pore water in soils by addition of a sidearm positioned below the stockcock (Figure 1.2), which can be loaded with soil samples contained in small reusuable glass tubes (Turner & Gailitis, 1988). The side arm remains above the top of the heating block, and the soil water evaporates to react with the zinc. For samples up to 150 gL and water contents < 20%, the modified technique requires a reaction time of 2.5 h. Insufficiently long reaction times, combined with water amounts > 30 gL, result in significant decreases in the 6D values due to incomplete reduction of water. There is a consistent bias in the 6D values of water extracted from kaolinite, apparently from an additional source of exchangeable H from the clays, that can be corrected for by use of standards of known composition (Turner & Gailitis, 1988). Other metals Problems with the reliability of batches of BDH zinc have prompted searches for other reagents that might be suitable for use in a batch reaction mode. Chromium can
16
C h a p t e r 1 - J. Horita & C. K e n d a l l
be used for the successful analysis of small (1 gL) water samples, either in an off-line batch mode or in an automated dynamic mode (Gehre et al., 1996b)(section 1-3.1.1). The use of magnesium (Mg) powder activated by a thin film of platinum (Pt) for analyzing water samples was first reported by Deqiu & Zhengxin (1985); this method was described in more detail by Halas & Jasinska (1996). The basics of the method are as follows" granular Mg (0.5 - 2 mm) is sieved, vacuum roasted, coated with dissolved platinum by a rather complicated procedure, the treated granules are dried and outgassed, and then about 4 gL of water is reacted with 120 mg of Mg-Pt reagent at 400~ for I h. The Pt coating step involves the use of several hazardous chemicals, including an acetone-ether solution that has to be distilled prior to use, and aqua regia. Previous attempts to use magnesium for either batch or dynamic reduction had been stymied by the production of Mg(OH)2 on the surface; the Mg(OH)2 can be decomposed at about 550~ but this would preclude the use of Pyrex reaction vessels (Halas & Jasinska, 1996). Coating the Mg with platinum black (Mg-Pt) may result in total decomposition of water to H2 at 400~ because the Mg-Pt is so reactive that the formation of hydroxides does not occur (Halas & Jasinska, 1996). The preliminary experimental data reported by Halas & Jasinska (1996) show some promise, but more work is needed to demonstrate optimum conditions for long-term precision and accuracy. They report that the method is rather sensitive to the amount of reagent used (perhaps because of diffusion of hydrogen into the metal), and that they hope that it might be suitable for analyzing untreated brines because of the high reactivity of the reagent. Manganese (Mn) shows much more promise in that it requires almost no pre-treatment and a preliminary report shows analytical precision of 0.4-0.8%o (Tanweer & Han, 1996). Although two types of Mn were tested, the 0.1 mm size Mn available from Fluka Chemicals Ltd. (Buchs, Switzerland) was shown to be a better choice than the finer-grained Mn from Merck Chemicals (Darmstadt, Germany). Optimum conditions using Coleman-type vessels (Coleman et al., 1982) appear to be 8 gL of water, 0.8 g of Fluka Mn, and a reaction time of 40 minutes at a temperature of 520~ (Tanweer & Han, 1996). Major advantages of the method are that the Mn requires no pretreatment, does not seem to show an amount effect (for 0.1 mm Fluka zinc). Furthermore, it is possible that other brands of Mn could be used, if they were sieved and rinsed to remove the fine powders that apparently are responsible for the amount effects. Shouakar-Stash et al. (2000) developed a Mn reactor unit installed on an inlet of an isotope-ratio mass spectrometer (see section 1-3.1.1). Recently, Ward et al. (2000) showed that lithium aluminum hydride can be used to reduce water for D / H ratio measurements by the reaction: LiA1H4 + 4H20 --* LiAI(OH)4 + 4H2
[1.3]
9 - 11 /~L of water was reacted with 0.04 to 0.05g of LiA1H4 in a vacutainer, and the reaction occurred instantly apparently at room temperature. Because LiA1H4 contributes to hydrogen in H2 produced, D / H measurements must be standardized with VSMOW-SLAP or other standards prepared by the same procedure. They reported precision of + 4 - 8%0 for waters with natural D / H ratios.
Stable IsotopeAnalysisof Water and Aqueous Solutions ...
17
Van Kreel et al. (2000) and Previs et al. (2000) report D / H ratio measurements of biological fluids enriched in D20 after converting water to acetylene with calcium carbide CaC2 + H20 --~ CaO + C2H2
[1.4]
The reaction was complete within a few second at room temperature, carried out in a closed container (e.g., Vacutainer). The mass ratio of m / e - 27 (C2HD) to 26 (C2H2) was determined with a mass spectrometer. The current precision (few tens of %0), however, limits its application.
1.3.1.3 H2-water equilibration H2-water equilibration methods were originally developed during the 1930's1940's for determining deuterium contents of heavy water by equilibrating it with H2 gas in the presence of Pt catalysts (Kirshenbaum, 1951). A major problem was a sharp decline in the catalytic activity of Pt catalysts used due to coverage of the surface by water. During the 1970's, new types of Pt catalysts were developed, which retain a hydrophobic surface because of the matrix materials used (Teflon, styrene divinyl benzene, etc.). These Pt catalysts significantly enhanced D / H exchange between liquid water and H2 (Rolston et al., 1976). Ohsumi & Fujuno (1986), and Horita (1988) reported precise (lo = 0.8%o) D / H ratio measurements of natural waters, including s. Subsequently, Horita et al. (1989) demonstrated that the H2-water equilibration method with the hydrophobic catalysts can be readily automated using a gas equilibrator designed for the automated CO2-water equilibration method (Figure 1.3). Several investigators tested and improved this method for natural waters (Coplen et al., 1991; Ohsumi, 1991; Brand et al., 1996; Thielecke et al., 1998; Bourg et al., 2001) and for biological and agricultural fluids (Scrimgeour et al., 1993; Coplen & Harper, 1994; Br6as et al., 1996; Thielecke & Noack, 1997). The major advantages of the H2-water equilibration method are: (1) no pretreatment of water samples is required regardless of the type of aqueous solution (e.g., fresh water, brine, biological and agricultural fluids), except for removal of dissolved H2S; (2) very precise data (lo < 0.5%o, in external precision) can readily be obtained; and, (3) the entire procedure can be readily automated with an on-line gas equilibrator, thus reducing required labor and time, and reducing human errors. Both D / H and 180/16 0 ratios of water can be determined successively with the same gas equilibrator. Gas equilibrators and multi-preparation systems with septum-sealed glass vials (e.g., Vacutainers) for gas equilibration methods are currently available from commercial mass spectrometer firms (Analytical Precision, Finnigan MAT, Micromass, PDZ Europa). The gas equilibrator is specially designed for fully automated high-precision gas equilibration methods, and the multi-preparation system is versatile for many other sample preparations. Recently, the H2-water equilibration method has also been adapted to a continuous-flow mass spectrometer with a unique geometry (Prosser & Scrimgeour, 1995). The addition of a small energy filter on conventional continuous-flow mass spectrometers to better separate the He and H2 peaks has allowed the modification of the equilibration systems to operate under continuous-
18
Chapter 1 - J. Horita & C. Kendall
Figure 1.3 - Schematic of an automated, on-line gas equilibrator for H2- and CO2-water equilibration methods. Air in the equilibration vessels is p u m p e d out through a capillary tube to minimize a loss of water. Temperature control of + 0.05 ~ and + 0.1~ or better is required for H2- and CO2-water equilibration methods, respectively.
flow mode using He as a carrier gas. Precisions of 1%o for hydrogen and 0.1%o for oxygen or better are presently being reported by the commercial vendors.
Exchange reactions, catalysts and normal procedures Hydrogen isotope exchange between gaseous H2 and liquid water proceeds via two consecutive reactions; HD(g) +H20(v) ~=~H2(g) + HDO(v) and
[1.5]
19
Stable Isotope Analysis of Water and Aqueous Solutions ... H D O ( v ) + H20(1) e,, H 2 0 ( v ) + HDO(1)
[1.6]
The overall reaction is HD(g) + H20(1 ) k H2(g ) + HDO(1)
[1.7]
where k is the rate constant. Pt-catalysts catalyze the isotope exchange reaction [1-5], which is otherwise extremely sluggish. The fractionation factor, a - (D/H)water/(D/ H)H2, of the overall reaction [1-7], is very large (3.81 at 25~ Rolston et al., 1976), resulting in extremely low D / H ratios (ca. 0.004%) of H2 gas equilibrated with natural waters (0.015%). The D / H fractionation factor is also very sensitive to temperature, and precise temperature control (___0.05~ is required. The original hydrophobic Pt catalyst used by Horita (1988) and Horita et al. (1989) is Hokko Beads (125 - 250 mm diameter styrene divinyl benzene porous resin beads coated with 3 wt% Pt), manufactured by Shoko Ltd. (Minato-ku, Japan). Another type (1 mm, I wt% Pt) is also available. Finnigan MAT (Bremen, Germany) provides Pt catalysts manufactured by MS-Analysentechnik (Berlin, Germany). The Atomic Energy of Canada Limited (AECL) of Chalk River National Laboratory (Ontario, Canada) also produces similar hydrophobic Pt catalysts. The hydrophobic Pt catalysts are generally carefree, and can be used many times after washing with distilled water and drying in air. It is reported that heating the I mm Hokko Beads at 60~ resulted in the loss of catalytic activity, whereas heating at 60 ~ - 100~ has no effect on the powdered Hokko Beads and other hydrophobic catalysts (ISOGEOCHEM1 discussion). Dissolved HzS is known to poison the catalysts; hence, water samples suspected to contain H2S should be treated by reacting the water with Cu metal or AgNO3 overnight (Coplen et al., 1991). Although the human nose is very sensitive to low concentrations of H2S, prolonged exposure is harmful; hence, water samples should not be routinely tested by sniffing. Heating the beads that were exposed to H2S at 230~ for 2 hr in a H2 stream can restore the activity (Ohba & Hirabayashi, 1996). Among biological and agricultural fluids, alcohol in wine is reported to interfere with mass spectrometric measurements (Br6as et al., 1996). However, it is not clear whether alcohol poisoned the catalysts or interfered with mass spectrometric measurements. The treatment of urine samples with activated charcoal is also recommended. The catalysts are electrostatic, and can be better handled by gluing to a stick (Coplen et al., 1991) or containing them in a Teflon bag (Ohba & Hirabayashi, 1996). The suppliers of the hydrophobic Pt catalysts are: 9 Shoko Co. Ltd. (attn: Wataru Maruyama)" 3-8-3 Nishi-Shimbashi, Minato-ku, Tokyo 105, Japan. Tel. +81 3-3459-5106; Fax (81) 3-3459-5081
Information on the ISOGEOcHEM list can be found in Appendix A of Part 2 in this Volume of the Handbook.
20
Chapter 1 - J. Horita & C. Kendall U.S. agent of Shoko" ICON Services Inc., Stable isotopes and labeled compounds (attn: John Kilby), Ox Bow Lane, Summit, NJ 07901, U.S.A. Tel. +1 (800) 322-4266; Fax +1 (908) 273-0449; www.iconisot.ios.com
9 MS Analysentechnik: c/o Bogen-Elektronik, Potsdamer Str 12-13, 14163 Berlin, Germany. Fax +49 30 8226072 9 AECL: Marketing and Sales/Nuclear Products and Services (attn: Chris Knight), Chalk River National Laboratory; Chalk River, Ontario, Canada K0J 1J0. Tel +1 (613) 584-8811 (ext 6029); Fax +1 (613) 584-1438. For normal operation, 1 to 5 mL of water or aqueous sample is loaded into a reaction vessel, together with a few to tens of mg of the Pt catalyst: the amount of the catalyst required probably depends on the type and manufacture. Large amounts of water are preferred because then (1) loss of water during the evacuation of air in a reaction vessel through a capillary tube is negligible when a gas equilibrator is used, (2) the evaporation of water within a reaction vessel (typically 10 - 20 mL) is negligible, and (3) there is little change in the D / H ratio of water samples during isotopic exchange between H2 and water. However, water samples as small as 0.1 - 0.25 mL have been successfully analyzed (Coplen et al., 1991; Thielecke et al., 1998). The vessels are then evacuated either after freezing the water or through a capillary tube without freezing. H2 gas is admitted into the vessels to a given pressure, usually one atmosphere. It is recommended that the same amounts of water and H2 gas be used for the entire set of standard and sample waters, so that isotopic changes caused by the above three processes should be constant and accounted for by standardization (normalization). Valves to the reaction vessels are closed and the vessels are immersed into a water or air bath. The bath temperature should be near room temperature (generally at 18 40~ and controlled to within + 0.05~ because the water-H2 D / H fractionation factor is very sensitive to temperature (6.3%o per ~ The recommended maximum + 0.05~ deviation in temperature corresponds to ~ 0.3%o in 6D (H2) values. The use of hollow plastic balls on the surface of the water bath can drastically reduce temperature fluctuations. A bath temperature below room temperature may be preferred in order to avoid the condensation of water inside the vessels. However, the above three factors (water loss, evaporation, and isotopic exchange) should be evaluated carefully for any experimental procedure, and suitable corrections made to the data, if necessary. Shaking the vessels is not advisable, because unlike CO2-water equilibration discussed below, it does not facilitate the isotope exchange reaction, and because it could spread the catalyst powders, plugging the capillary tube. After isotopic equilibration, H2 gas is expanded into a sample-side reservoir of a mass spectrometer through a Utrap cooled with dry ice or liquid N2 that removes water moisture. When the sample of water is small (< 1 mL), water loss during the evacuation of the vessels could alter the isotopic composition of water. This effect must be critically evaluated depending on the exact procedure used (i.e., vessel volume, evacuation time, geometry of capillary tube, etc.). To minimize problems with small-size samples, the analyst should: (1) use small-volume reaction vessels to minimize the loss of water
Stable Isotope Analysis of Water and Aqueous Solutions ...
21
to the vapor phase, and (2) use the same amount of water for standard and sample waters, so that the effect of water loss will largely cancel out. In either case, the accuracy of ~SD values may not be as high as that of large-volume samples. On the other hand, the effect of isotopic exchange on 6D values of waters can be readily corrected by the following equation, similar to that used in the CO2-water equilibration method (Craig, 1957). [1.81
5corr -- (l+p / ot)6meas - (p / Ot)Sinit
where i~corr, 6meas, and Dinit a re 6D values of corrected, measured, and initial H2 gases against a working standard H2. The p is the molecular H2/water ratio, and c~= (D/ H)water / (D / H)H2 at equilibrium (3.81 at 25~ The overall D / H isotope exchange rate between H2 and water follows a first-order reaction, and the following equation is applied; f n o n - e q u i l - ( ~t - ~ ) / ( 6 0
- 6~) - exp(-kt)
ln(fnon-equil) - ln(6t - ~ ) / (~0 - ~ )
- -kt
[1.9a] [1.9b]
where fnon-equil is the fraction of non-equilibrated H2 gas (1.0 at t = 0, and 0 at t - ~), and 60, 6t, and ~ are the 6 values of H2 at the initial time, time t, and at the equilibrium time, respectively. In the presence of the hydrophobic Pt catalysts, the half-reaction time ( t l / 2 - 0.693.k-1, the time needed to reach a fnon-equil value of 0.5) is only 2.5 min (Horita et al., 1989). This means that after 25 min. (i.e., ten times the tl/2), the extent of isotopic equilibrium attained between H2 gas and liquid water is 99.9% (i.e., fnon-equil - 0.1%). Putting the beads in a Teflon bag slows down the isotope exchange rate (tl/2 - 40 min, Ohba & Hirabayashi, 1996). In contrast, isotopic equilibration takes 3 days using non-hydrophobic Pt-A1203 catalysts (Scrimgeour et al. 1993). The activities of different hydrophohic Pt catalysts have not yet been compared.
Special considerations There are several aspects to be considered for D / H ratio measurements with the H2-water equilibration method. The H2 gas used as a working standard in dual-inlet mass spectrometers should have a 6D value close to that of H2 equilibrated with the samples and standards. For example, the 6D value of H2 gas equilibrated with VSMOW and SLAP at 25~ is-740 and -850%o on the VSMOW-SLAP scale, respectively. Commercial H2 gas cylinders have a wide range of 6D values (-100 to -800%o), depending on manufacturing processes: electrolysis of water generally produces high purity gases with 6D values in the range of-100 to -200%o, while H2 from petrochemical processes (e.g., thermal cracking) has more negative 6D values (< -600%o). Because the D / H ratios of the H2 gas equilibrated with natural waters are very low (D/H = 0.004% at 25~ the relative contribution of H3 + (formed in the source) becomes a very large part of the total M / e - 3 (HD + and H3 +) signal in the mass spec-
22
C h a p t e r 1 - J. Horita & C. K e n d a l l
trometer. The production of H3 + is proportional to the square of H2 pressure in an ionization chamber of a mass spectrometer, and is in the range of 5 to 10ppm per 10-9 A of M / e - 2 with many commercial gas-source isotope-ratio mass spectrometers. Thus, when D / H ratio measurements were conducted at 5 x 10-9 A of M / e = 2, there is 25 to 50 ppm contribution of H3 + to the approximate 80 ppm of HD +. Thus, attaining a small and, more importantly, very stable H3 + contribution during the course of a mass-spectrometry session is crucial to precise measurements of D / H ratios with the H2-water equilibration method. Possible gradual drifts of the H3 + contribution or other ion source conditions with time can be easily monitored and corrected by measuring the D / H ratios of H2 gas equilibrated with a standard water several times during a session. With precise temperature control (< + 0.05~ of the water/air bath and a linear time-correction of small drifts in the H3 + contribution, an external precision of + 0.5 - 1.0%o (lo) or better can be readily achieved (Brand et al., 1996). This precision is nearly as good as that of conventional uranium-reduction methods, and is better than that of a batch Zn-reduction method in many laboratories. A set of two or more laboratory standard waters, which are calibrated on the VSMOW-SLAP scale (as discussed in the section on calibration) should always be analyzed together with the samples. Since the H2-water equilibration method does not require any pretreatment of samples (except for removal of H2S) prior to analysis, this technique is a preferred method for various aqueous samples with high concentrations of dissolved salts and compounds. Scrimgeour (1995) used the H2-water equilibration method to determine D / H ratio of water in soils and plants by directly equilibrating with H2 gas without any extraction of water. A soil sample yielded results consistent with that of an azeotropic method, but twig samples from shrubs and trees did not equilibrate completely with H2. Incomplete isotopic exchange is due partly to the low catalytic activity of the PtA1203 used, and this technique needs further investigation. Koehler et al. (2000) determined 6D values of pore water in clay-rich core samples directly using the H2-water equilibration method. With the Hokko beads, isotopic equilibrium between H2 and pore water was attained within 4 h, and a high precision (< + 1%o) was obtained. The analytical methods for D / H ratios are summarized in Table 1.1.
1.3.2 Oxygen isotopes For oxygen isotope analysis of water and other aqueous samples with dual-inlet, gas-source isotope-ratio mass spectrometry, C02 is a preferred gas species. Oxygen gas is preferred for the measurement of all three isotopes (160, 170, and 180). Water can be directly injected to a mass spectrometer, but this causes a large memory effect and, in the case of saline waters, rapid corrosion of the filament (Wong et al., 1984). CO is used as an analyte in continuous-flow isotope-ratio mass spectrometry using methods where oxygen in water or organic materials is reduced ("pyrolyzed") with C to form CO.
Table 1.1 - S u m m a r y of analytical methods for D / H ratios of water and aqueous samples, which are currently in use or u n d e r development.
Method
Reagent
Reduction
U
Reduction
Zn
Reduction
Zn
Reduction
Cr
Reduction
Mg/Pt
Reduction
Mn
Reduction
C
Reduction
LiA1H4
Reduction
CaC2
Electrolysis Equilibration H2 / Pt
Reaction
T (~
Reaction time (min)
Sample size (~L)
Precision (%o, lc~)
U + 2H20--* UO2 + 2H2 in reactor Zn + H 2 0 ~ ZnO + H2 in reactor ( w / q t z sand) Zn + H 2 0 ~ ZnO + H2 in closed tubes
400-800
15-30
0.5-10
0.2-1.0
380-450
15-30
0.5-10
1.0-2.0
450-550
10-60
1-10
0.5-2.0
2Cr + 3 H 2 0 ~ Cr203 + 3H2 in reactor
850-950
1-2
1-4
0.5-2.0
Mg + H 2 0 ~ MgO + H2 in closed tubes M n + H20--* MnO + H2 in reactor / closed tubes C + H 2 0 ~ C O / C O 2 + H2 in Ni bomb or pyrolysis furnace LiA1H4 + 4H20 --* LiAI(OH)4 + 4H2 in closed tubes CaC2+H20 --* CaO+C2H2 in closed tubes H 2 0 --* H2 + 1/202 H20(liq) ~=~ H2
Comment
9 o r~
> Use of Toepler / charcoal / expansion 0.5- 1% m e m o r y effect Use of Toepler / charcoal / expansion 1 - 2% m e m o r y effect Zn selection, degassing, Zn / water ratio critical Dissolved salts / organic interfere Commercial a u t o m a t e d unit on-lined to MS 1% m e m o r y effect Dissolved salts / organic slightly interfere Need to prepare platinized Mg Mg / water ratio i m p o r t a n t Grain size critical (100-200 ~m optimum)
400
60
4-10
2 (?)
520-900
0.5-40
5-8
0.7-1.5
950
20
5
2.0
ambient
<1
9-11
2-4
Reagent contribute H to H2
ambient
<1
3-5
3-4
M / e ratios of 27 / 26 determined
? 25-40
10 1-3 hr
20 >100
0.1(?) 0.3-1.0
~,,io
o
> O
O ~. 9
Commercial a u t o m a t e d unit on-lined to MS
Complete electrolysis Commercial a u t o m a t e d unit on-lined to MS No sample pretreatment (except for H2S) t,J
24
Chapter 1 - J. Horita & C. Kendall
1.3.2.1. Conversion methods
Many chemicals have been used to convert water for oxygen isotope analysis: guanidine hydrochloride (Boyer et al., 1961; Dugan et al., 1985; Dugan & Borthwick, 1986; Wong et al., 1987a), BrF5/C1F3/CoF3 (O'Neil & Epstein, 1966a; Bottinga & Craig, 1969; Blattner, 1973; Gulens et al., 1985; Dugan & Borthwick, 1986; Suvorova & Dubinina, 1996; Baker et al., 2002), and carbon with metal catalysts (Majzoub, 1966; Bariac et al., 1982; Brenninkmeijer & Mook, 1981; Ferhi et al., 1983; Gray et al., 1984; Edwards et al., 1994). BrF5 and C1F3 decompose water to 02, which is usually converted to CO2 with graphite. Other chemicals (Br2, HgC12-Hg(CN)2, and Na2S208) were also tested in the past. Recently, Meijer & Li (1998) improved an electrolysis method for precise 6170 (+ 0.07%o) and ~180 (+ 0.10%o) analysis of water. All these decomposition methods are designed to convert small amounts (< 10 ~L) of water quantitatively to 02, CO and CO2. Among these decomposition methods, the guanidine hydrochloride method is the only one routinely used in some isotope laboratories for the oxygen isotope analysis of water, due to the relative ease of its procedures and the high precision (+ 0.10%o) attainable. The carbon pyrolysis technique, which converts water exclusively to CO at high temperatures (1000 - 1300~ has recently been adapted for isotopic analysis of water, other fluids, organic matter, and minerals with a continuous-flow mass spectrometer (Brand et al., 1994; Werner et al., 1996; Begley & Scrimgeour, 1996, 1997; Koziet, 1997; Farquhar et al., 1997; Br6as et al., 1998; Loader & Buhay, 1999; Houerou et al., 1999; Kornexl et al., 1999a,b; Sharp et al., 2001); this new method is described elsewhere in this book (Volume II, Part 3, Chapter 6-2.3.6). In the guanidine hydrochloride method, water is directly converted to CO2 (Dugan et al., 1985; Wong et al., 1987a). About 100 mg of very pure guanidine hydrochloride is placed in a Pyrex tube, which is then evacuated and heated to melt the guanidine; the tube is then cooled, about 10 gL of water is frozen into the tube, any contaminants are pumped away, and the tubes are sealed with a torch. The sealed tube is then heated in a furnace at 260~ for 8 - 16 h (or longer), for the water and guanidine to react together to produce ammonia and CO2. As the tubes later cool below 70~ the gases further react to form solid ammonium carbamate (NH4NH2CO2). The cooled, pre-scored, reaction tube is then either placed inside an evacuated, ball-jointed, tube-cracking assembly (Dugan et al. 1985) or is mounted in an apparatus containing a flexible tube-cracker (Wong et al., 1987a). Both of these units contain a separate reservoir containing about 0.5 mL of 100% H3PO4. After the tube is cracked open, the unit is removed from the vacuum line and the lower part of the unit placed in an oven at 80~ for about an hour for the ammonium carbamate to react with the phosphoric acid, and decompose to produce CO2, ammonium chloride, and ammonium phosphate. The CO2 is then purified on a vacuum line and analyzed. The method can be fussy and special precautions to observe include purifying the guanidine by melting it twice, and never letting any of the other reagents directly contact the H3PO4. Erratic results, including anomalously high ~13C values of the CO2, can be caused by the accidental production of C12 gas. With BrF5 or C1F3, small amounts (5 - 10 ~L) of water are converted to 02 in the same manner as samples of silicates and oxides (O'Neil & Epstein, 1966a; Bottinga &
Stable IsotopeAnalysisof Water and Aqueous Solutions ...
25
Craig, 1969; Blattner, 1973; Gulens et al., 1985; Dugan& Borthwick, 1986; Suvorova & Dubinina, 1996). Water samples have to be transferred quantitatively to an Ni bomb and the contamination of atmospheric vapor has to be avoided. Three to 10 times the stoichiometric amounts of the reagents are used, and the reaction can proceed readily at 80 - 350~ within a few minutes. The 02 is usually converted to CO2 with heated graphite. Alternatively, 02 can be directly introduced into an isotope-ratio mass spectrometer as an analyte, and the both 170/160 and 180/160 ratios of water can be determined. This method is useful for the study of non mass-dependent isotopic fractionation and extraterrestrial waters.
1.3.2.2 Direct introduction of water Majzoub & Nief (1968) first described an isotope ratio mass spectrometer for direct 180/160 measurement of water where the H2180 + and H2160 + ion currents were monitored. Although high precision results (+ 0.08%o) were achieved, the entire inlet and source units had to be heated to minimize memory effects. Hagemann & Lohez (1978) built a twin mass spectrometer system for simultaneous oxygen and hydrogen isotope measurement of water, using the direct injection method for 180/160 and online uranium reduction method for D / H (discussed above). A commercial version of the twin mass spectrometer (Aqua SIRA, VG, UK) was built and tested (Wong et al., 1984; Barrie & Coward, 1985). However, the direct introduction method has never become a routine technique. 1.3.2.3 C02-water equilibration The CO2-water equilibration method has been most widely used for the oxygen isotope analysis of large-volume water samples (hydrologic, agricultural, and biological fluids). This method, based on the classic work by Harold Urey and his colleagues (Cohn & Urey, 1938, Mills & Urey, 1939; 1940), was first applied to natural waters in the early 1950's (Baertschi, 1953; Dansgaard, 1953; Epstein & Mayeda, 1953). This method is often referred to as the "Epstein-Mayeda" method. The CO2-water equilibration method is a simple and very precise method for obtaining 6180 values of water. A few mL of water are placed in a vessel equipped with a stopcock. The air is removed by successive freezing, evacuating, and then thawing the water sample. A few 100 ~moles of CO2 are then frozen into the vessel, the stopcock closed, and the vessel placed in a water or air bath where the temperature (usually 25~ is controlled to about + 0.1~ The vessels are periodically shaken to speed up the equilibration rate. After several hours or days, the vessels are returned to a vacuum line, the CO2 is separated from the water by freezing the water first in liquid nitrogen and then a dry ice-alcohol slush, and then analyzed. Roether (1970) eliminated the time-consuming freezing and thawing steps that were required to remove air from the vessels without fractionating the water, by evacuating air from the reaction vessel through a capillary tube. This innovation minimized oxygen isotope fractionation of the water to about 0.1%o. With this technique, Fairbanks (1982) achieved a high precision (_+0.03%0) for seawater samples. The Roether technique laid the foundation for the automated oxygen and hydrogen isotope analysis of water samples by on-line gas equilibration (Buzek, 1983; Chiba et al., 1985;
26
Chapter 1 - J. Horita & C. Kendall
Brenninkmeijer & Morrison, 1987; Wong et al., 1987b; Horita et al., 1989). The CO2water equilibration technique was also modified for rapid manual preparation of multiple water samples, using inexpensive labwares such as plastic syringes and preevacuated glass vials (Yoshida & Mizutani, 1986; Graber & Aharon, 1991; Socki et al., 1992).
Exchange reactions The overall oxygen isotope exchange between gaseous CO2 and water proceeds via two reactions. ke
Gas exchange:
CO2(g ) ~ CO2(aq ) kh
[1.10] kl
m
Hydration / dissociation: CO2(aq ) + H20(1 ) ~ H 2 C O 3 r H + + H C O 3
[1.11]
The overall reaction is, ktot C1602(g) + H2180(1) r C18016 O(g) + H2160(1)
[1.12]
The hydration reaction (kh) is orders of magnitude slower than the dissociation reaction (M), and the enzyme carbonic anhydrase is known to catalyze the hydration reaction. When a reaction vessel is not shaken, the gas exchange (ke) is rate-limiting in the overall reaction (Roether, 1970). .~ However, with increasing fie~ quency of shaking (i.e., an ~ increase in the effective sur~ face area and mixing of the ~ solution), the overall exchange ~ rate increases rapidly, and the ~ hydration reaction becomes ~ rate-limiting (Figure 1.4:). ~• Roether (1970) described the ~ overall rate constant (ktot) for ~ each case; o ~
Figure 1.4 - Effect of shaking on overall rate constant (Mot, min-1) of 180/160 exchange in CO2-water equilibration method. With increasing shaking cycles, the reaction changes from gas exchange-limited to hydration-limited processes. After Roether (1970).
O
27
Stable Isotope Analysis of Water and Aqueous Solutions ... ktot - ke'CO2(diss) / CO2(g)
(gas exchange limited)
[1.13]
ktot - 1 / 3"kh'CO2(diss) / [CO2(g) +CO2(diss)]
(hydration limited)
[1.14]
where CO2(diss) is total dissolved CO2 (CO2(aq) + H2CO3 + HCO3-). The characteristic time (ktot -1) decreased an order of magnitude with increasing shaking frequency, from that of the gas exchange reaction to that of hydration (18 min.) at 18~ (Roether, 1970) (Figure 1.4). The ktot also depends on the amount of water. Under normal conditions, vigorous shaking decreases significantly the half-life time of the isotope exchange (tl/2 - 0.693"ktota1-1) from 100 -400 min. to 10 - 20 min. Thus, the oxygen isotope equilibration between gaseous CO2 and water can be complete after 2 - 3 hr (99.9% equilibrium after ten times h/2) rather than I - 2 days, when not shaken vigorously (Figure 1.5). For detailed discussion on the kinetics of oxygen isotope exchange between gaseous CO2 and water, see Roether (1970) and Taylor (1973). The required equilibration time must be established in each laboratory for the conditions and types of samples anticipated, rather than relying on published times. The oxygen isotope exchange between CO2 and alkaline waters (pH = 9), where carbonate ion (CO32-) iS the dominant inorganic carbon species, is very slow (Mills & Urey, 1940). A large fraction of CO2 gas introduced into the vessel is absorbed into the water. For these reasons, the pH of alkaline waters needs to be adjusted to below 6 - 7 with acid (anhydrous H3PO4 or other acids). High salt contents also slow down the isotope exchange rate, and vigorous shaking of the vessel becomes necessary (see the section below).
Figure 1.5 - Examples of 180/160 exchange in CO2-water equilibration method. Shaking decreases a half-life time (tl/2=ln2.ktot -1) by a several factor. Salinity and pH affect significantly overall isotope exchange time. The exact half-life time also depends on many factors (geometry of a vessel, shaking amplitude and frequency, amount of water and CO2, and chemistry of water). Data of pure water from Gonfiantini (1981). A Dead Sea brine is of Na-Ca-K-C1 type with 320 g/L salinity (NaC1 - 1.95 molal, KC1 - 0.15 molal, MgC12- 2.00 molal, and CaC12- 0.49 molal).
28
Chapter 1 - J. Horita & C. Kendall
The equilibrium oxygen isotope fractionation factor ~ = (180 / 160)CO2(g) / (180 / 160)H20(1) is 1.0412 at 25~ (Hut, 1987), and its temperature dependence is 0.2%0/~ Thus, temperature control of + 0.1~ is adequate for precise measurement. The oxygen isotope composition of the dissolved CO2 species (CO2(aq), H2CO3, HCO3-) is different from that of gaseous CO2, and degassing of dissolved CO2 should be minimized during the extraction from the vessel. Depending on the ratio of CO2 to water and their initial isotopic compositions, the isotopic composition of the water generally changes during the isotopic exchange with CO2. This effect should be corrected, if necessary, as with the H2-water equilibration method (equation [1.8])(Craig, 1957). The isotopic composition of water samples should be measured along with laboratory standard waters, which are in turn calibrated on the VSMOW-SLAP scale (see discussion below).
Small samples For small amounts of water (< 0.1 mL), water loss during evacuation through a capillary tube of an automated gas equilibrator a n d / o r evaporation of water inside a reaction vessel (typical volume 10 - 20 mL) becomes significant. Isotopic exchange with CO2 gas also causes a significant change in the isotopic composition of water. To circumvent these problems, Kishima & Sakai (1980) developed a "micro-scale" CO2water equilibration method (a.k.a the MCE method) for small amounts (0.5 - 10 ~L) of water using small-volume (0.5 mL) reaction vessels and applying the corrections required to account for isotopic exchange and unavoidable evaporation. Rozanski et al. (1987) tested this method. Ohba (1987) and Socki et al. (1999) further improved this technique using a sealed Pyrex glass tube as a reaction vessel. The correction equation for evaporation and isotope exchange is; ~corr -- C~. p(~meas - ~init) + 103"(1+10-3"6meas)[(1 + (a'-l)~,] - 103
[1.15]
where p is the ratio of oxygen atoms in CO2 to that in H20 in the vessel, and ~, is the ratio of vapor to liquid water in the vessel. The a - (180/160)CO2(g)/(180/160)H20(1) = 1.0412 (Hut, 1987) and a' - (180/160)H20(v) / (18O/160)H20(1 ) - 0.99074 at 25~ (Horita & Wesolowski, 1994). The 5cor, ~meas, and ~init are 6180 values of corrected, measured, and initial CO2 gases, respectively, relative to working standard CO2. When the evaporation of water is negligible (~, ~ 0), this equation is the same as equation [1.8]. The tl/2 ranges from 2 to 6 hr for 0.5 to 10 ~L of water (Kishima & Sakai, 1980), which is comparable to those of the original CO2-water equilibration method. Contamination of water from moisture and other sources has to be avoided while loading the sample into a reaction vessel. For accurate and precise data, the transfer and separation of water and CO2 have to be complete. Furthermore, the amount of water and CO2 introduced to a reaction vessel must be measured precisely to make the needed corrections. With these precautions and the above corrections, the 5180 values of water samples as small as 0.5 ~L can be determined precisely (lo - 0.05 0.1%o). Using CO2 with a 6180 close to that expected after equilibration can significantly shorten reaction times and improve analytical precisions. In the absence of liquid water, the isotope exchange between CO2 and water vapor is prohibitively slow,
Stable Isotope Analysis of Water and Aqueous Solutions ...
29
and the reaction has to be catalyzed by Pt at high temperatures (Dostrovsky & Klein, 1952). One important advantage of the micro-scale CO2-water equilibration method is that the 6D value of the same water sample can be determined using a reduction (Zn, U, etc.) method after the removal of the CO2 gas. In fact, if water samples prepared for the micro-scale CO2-water equilibration method cannot be weighed (or otherwise measured) to sufficient precision, the yield of H2 gas must be measured to correct the data. Most of the decomposition methods discussed above for the ~5180 analysis of small amounts of water cannot be used for 6D analysis of the same water sample. An exception is the high-temperature (1450~ carbon pyrolysis method, by which 6D and 6180 analysis can be done on a single sample (Motz et al., 1997). Water is converted to H2 and CO within an elemental analyzer, which are then separated by GC and introduced into a mass spectrometer with He carrier gas. A peak-jump between H2 and CO enables a sequential measurement of 6D and 6180 values. Splitting small water samples into two or more aliquots for separate analyses is not recommended because of probable fractionation associated with the splitting. Scrimgeour (1995), Hsieh et al. (1998), McConville et al. (1999), and Koehler et al. (2000) applied CO2 equilibration methods to water in soil and plants. Oxygen isotope equilibrium between CO2 and soil water was attained within I - 2 days. The precision ranged from + 0.12%o (McConville et al., 1999) to + 0.3 - 0.4%o (Hsieh et al., 1998; Koehler et al. (2000). The CO2-water equilibration method was also successfully adopted for continuous-flow mass spectrometry for as small as 40 ~L of water, using septum-sealed glass vials (Analytical Precision, Finnigan MAT, Micromass, PDZ Europe; Fessenden et al., 2002). Leuenberger & Huber (2002) developed a novel on-line continuous-flow mass spectrometric method for 180/160 analysis of ice cores. CO2 was injected into a stream of liquid water generated by layer-by-layer melting of an ice core. After dissolved CO2 reaches an (near) isotopic equilibrium with water at 50~ dissolved CO2 was separated through a gas-permeable membrane and carried with a He to a mass spectrometer. Leuenberger & Huber (2002) attained a precision of better than 0.1%o with a resolution of I - 3 cm of an ice core. The analytical methods for 180/160 ratios are summarized in Table 1.2.
1.4 Effects of dissolved compounds High concentrations of dissolved compounds in water (brine, biological and agricultural fluids) can pose serious analytical problems. In the decomposition methods for 6D and 6180 analyses of water and other aqueous samples discussed above, water must first be extracted from the samples. If the water is injected directly into a reaction unit, serious contamination and memory effects would result. The most common method for the extraction of water is vacuum distillation. However, it is not always easy to extract water quantitatively from high-salinity brines, juices, and blood even with prolonged heating at elevated temperatures. Incomplete extraction of water from
Table 1.2 - S u m m a r y of analytical m e t h o d s for 1 8 0 / 1 6 0 ratios of w a t e r a n d a q u e o u s samples, w h i c h are currently in use or u n d e r d e v e l o p m e n t . Method
Reagent
Reaction
Equilibration
CO2
H20(liq) <=~ CO2
2 0 - 30
3 - 10 hr
>100
0.05- 0.1
C o m m e r c i a l a u t o m a t e d unit on-lined to MS S h a k i n g facilitates the reaction Slow reactions for brines
Equilibration
CO2
H20(liq) ~=~ CO2
2 0 - 30
3 - 10 hr
0.5-10
0.05-0.1
For small s a m p l e s Corrections for v a p o r i z a t i o n a n d CO2 e x c h a n g e
260 80
8 - 10hr I hr
10
0.1 - 0.15
5 - 10
0.1 - 0.2
Conversion
guanidine
Conversion
BrF5(C1F3)
Conversion
Electrolysis
graphite / diamond
g u a n i d i n e + 2 H 2 0 --* N H 4 N H 2 C O 2 + NH4C1 N H 4 N H 2 C O 2 + NH4C1 + H3PO4 ~ CO2 + NH4C1 + (NH4)3PO4 H 2 0 + BrF5 --* 1 / 2 0 2 + BrF5 + 2HF 0 2 + C ~ CO2
T (~
80 - 350
Few to 30
Sample size (~L)
Precision (%o, l o )
Comment
R e a g e n t s explosive r
C + H 2 0 --~ C O / C O 2 + H2 950 - 1200 (in Ni b o m b or pyrolysis tube) 350 2 C O ~ CO2 + C H 2 0 ~ H2 + 1 / 2 0 2
Reaction time (min)
?
20 15
2 - 10
40
1000
0.2 - 0.3
0.1
C o m m e r c i a l a u t o m a t e d unit on-lined to MS
r~
Partial (<1%) electrolysis Calibration r e q u i r e d for fractionation
o
!
2.
(3
Stable Isotope Analysis of Water and Aqueous Solutions ...
31
brines enriched in MgC12 and CaC12 results in lower-than-expected (< 10%o for 6D and < 5%0 for 6180) and scattered isotopic compositions (Grabczak et al., 1986; Fritz et al., 1986; Horita, 1989a; Koehler et al., 1991; Yang et al., 1996b). These alkaline-earth elements can be removed as carbonate precipitates, and water recovered by azeotropic distillation (Horita & Gat, 1988). However, this is a tedious procedure, and is not suited for routine analysis. Brines were directly introduced to a reaction vessel for D/ H ratio measurements with the batch Zn-reduction method (Tanweer, 1993b). Although this technique was successful for many salt solutions, the optimum amount of Zn had to be established for each salt solution. The same problem is encountered during the extraction and isotopic analysis of brine inclusions in minerals (Horita & Matsuo, 1986). 6180 values of biological samples (saliva, urine, plasma, and milk) determined by the guanidine hydrochloride method were systematically different from those with the CO2-equilibration method (Wong et al., 1987a). In addition to incomplete recovery of water, it is also possible that isotopic exchange between water and dissolved organic matter occurs in decomposition methods at elevated temperatures. The H2- and CO2-water equilibration techniques are the methods of choice for aqueous samples with high concentrations of dissolved compounds because these methods do not require the extraction of water or any sample treatment (except removal of H2S). With equilibration methods, the time required to reach isotopic equilibrium between H 2 / C O 2 gases and water increases with salinity. For example, it takes about 2 hr for gaseous H2 to reach D / H equilibrium with a Dead Sea brine (salinity, 320 g/L) instead of I hr for distilled water (Horita, 1988). Isotopic exchange between CO2 gas and brines slows down significantly with salinity, particularly with MgC12 and CaC12 solutions (Grabczak et al., 1986; Fritz et al., 1986; Horita, 1989a; Fortier, 1994). This is due largely to a sharp decrease in the solubility of CO2 in salt solutions. Vigorous shaking of reaction vessels and acidifying of brines, if alkaline, becomes necessary (Figure 1.5)(Horita, 1989a; Fortier, 1994). A fundamental characteristic of water in brines and other aqueous solutions is that the isotope activity ratios of water are not the same as the isotope composition (concentration) ratios. The difference between the isotope activity and composition (concentration) ratios ("the isotope salt effect"), which is caused by interactions between solutes and water (hydration, structure-making and structure-breaking processes, etc.), can be defined as:
F -
a(HDOorH2180)/a(H2160)= ~,(HDOorH 2180 )
X(HDOorH2180)/X(H2160)
~, ( H 2 1 6 0 )
[1.16]
Where a, X, and ~, denote activity, mole fraction, and activity coefficient of isotopic water molecules, respectively (Horita et al., 1993b). Then, the F can be determined as: 103 In F = ~activity - 6composition - Y(ai" mi)
[1.17]
32 where ai and mi are the experimentally determined coefficient and molality (mol/kg H20) of the dissolved salt i, respectively. Examples of the isotope salt effects of geochemical interest are shown in Table 1.3 and Figure 1.6.
Chapter 1 - J. Horita & C. Kendall Table 1.3 - The isotope salt effects at 25~ (see equations [1.16] and [1.17]). Consult Horita et al. (1993b) for data at other temperatures.
Salt NaC1
1031nF (ai, %o per molal) 6D 6180 2.15
0
KC1 2.54 0.16 Gas (H2 and C O 2 ) equilibration CaC12 5.4* -0.45* methods yield the isotope activity MgC12 5.0* -1.11 ratios, and all other decomposition MgSO4 5.7 -1.04 methods discussed above yield the isoNa2SO4 4.8 tope composition ratios. The isotope HC1 0.80 salt effects in hypersaline brines are far H2SO4 -8.2# larger than the errors associated with the isotopic analysis, and need to be * not well determined corrected for. Many experimental stud- # at 20~ ies show that the effect is proportional to the molality (mol/kg H20) of single salt solutions. For complex natural brines, the effect is a sum of the effects caused by each single salt. If gas equilibration methods are carried out at temperatures different from 25~ the temperature dependence of the isotope salt effect has to be taken into account (see Horita et al., 1993b). Not all salts of geochemical interest are well investigated, but any dissolved salts with concentrations higher than 0.1 molal need serious consideration for the isotope salt effect; this includes waters with very high concentrations of sulfuric acid (e.g., mine-drainage ponds). It is also known that some non-electrolytes (urea, pyridine, dioxane) change the isotope activity ratio, but the effects of major dissolved compounds in biological and agricultural fluids (plasma, protein, sugar, alcohol, etc.) are not determined yet.
The geochemical implications of the isotope salt effect are discussed in detail by Horita et al. (1993a). Briefly, isotopic water molecules mix according to their 6 composition values (i.e., they mix conservatively), but they react according to their 6 activity values (e.g., evaporation of surface waters, degassing/boiling from hydrothermal waters, precipitation of minerals from water), just like other chemical constituents. Horita & Gat (1989) reported very large D / H salt effects (up to 18%o) for Dead Sea brines (salinity ~ 320 g/L) by direct measurements of both D / H activity and composition ratios. They showed that the evaporation process of the Dead Sea is properly interpreted with isotope activity ratios, rather than composition ratios. The D / H salt effects are detectable even for seawater and mid-ocean ridge vent fluids (1.2%o, Shanks et al., 1995), whereas the magnitude of the 180/160 effect in seawater is negligible. 1.5 Mass-spectrometric measurements and standardization Operating procedures for mass spectrometric measurements are discussed in detail by Gonfiantini (1981) and Coplen (2001). Several corrections must be applied to raw g values measured against a working standard gas: (1) zero-enrichment factors,
Stable Isotope Analysis of Water and Aqueous Solutions ...
33
Figure 1.6 - The isotope salt effects of dissolved salts and Dead Sea brine at 25~ The addition of salts to a fresh water (O) increases i~Dactivity and decreases h18Oactivity values, respectively as indicated by the two arrows: exception is 618Oactivity of KC1, which increases. Gas (H2 and CO2) equilibration methods yield 6activity values, and conversion methods (U, Zn, guanidine hydrochloride, etc.) yield 6composition values. Notice that the most commonly used methods (U or Zn reduction for 6D and CO2-equilibraiton for 6180) yield the two isotopic ratios on different scales, 6D(c) - 6180(a) (2) instrumental corrections (valve mixing, residual, and tail contributions), (3) the 170-correction for CO2, a.k.a, the "Craig correction" (Craig, 1957), (4) the H3 + effect for H2, and (5) the effect of isotope exchange in H2- and CO2-equilibration methods. The zero-enrichment cancels out w h e n gases prepared from samples and reference standards u n d e r the same conditions are m e a s u r e d against a c o m m o n w o r k i n g standard gas. The instrumental corrections are very small with m o d e r n mass spectrometers. Some uncertainty still exists about the appropriate 170-correction (Santrock et al., 1985; Verkouteren et al., 1995). A mass spectrometer should be tuned for a small H3 + contribution, but its stability is the primary concern, especially for the H2-water equilibration method. The effect of isotopic exchange on the isotopic composition of water can be easily corrected with equation [1.8]. Many of these corrections are included in the data-reduction software of m o d e r n mass spectrometers, but the users should be familiar with the exact procedure of these corrections, since there m a y be optional choices.
34
Chapter 1 - J. Horita & C. Kendall
A fundamental principle in the calibration of stable isotope data of samples is the parallel preparation and mass-spectrometric measurements of samples and standards under the same conditions. This is because isotopic fractionation may occur during the preparation of samples to produce the appropriate analyte gases (H2, CO2, CO, 02, etc.). This is obvious for the gas (H2 and CO2) equilibration methods, where large equilibrium isotope fractionations, c~(CO2-H20) and c~(H20-H2), are involved. These fractionation factors should not be used for back-calculating the isotopic composition of water samples because the "actual" values of the fractionation factors very likely vary depending on sample preparation procedures and the mass spectrometer tuning. Even with nominally quantitative decomposition methods, it is likely that samples are slightly fractionated, depending on the method, procedure, preparation equipment, and operators. For example, the 6D value of water prepared by metal (U, Zn, and Cr) reduction methods in various laboratories varies over more than 10%o (Brand, ISOGEOCHEM). By preparing samples and standards at the same conditions and conducting mass-spectrometric measurements in the same session, all potential isotopic fractionations involved during the entire course of preparation and measurements can be collectively cancelled out. Isotope effects and fractionation caused by slow, systematic changes in preparation systems (e.g., temperature) and a mass spectrometer (e.g., source conditions such as H3 + effect) can be identified and corrected by measuring separate aliquots of the same laboratory and/or international standard waters several times during the analysis of a set of samples. Two international standard waters (VSMOW and SLAP) are distributed by IAEA and NIST for inter-laboratory calibration of the stable isotopic compositions of waters. The amount of these standards available for each isotope laboratory is limited (one 20 mL bottle each, every 3 years). Thus, laboratory standard waters need to be prepared and calibrated against the VSMOW-SLAP scale for routine measurements of a large number of water samples. Local tap water can serve as a laboratory standard. However, it is rather difficult to obtain or prepare a laboratory standard water with low 6D and ~180 values. The reason for using two standard waters with widely different isotopic compositions for the calibration of sample water is to "normalize" the %o-scale the necessary process of stretching or shrinking the isotope ratio scale associated with preparation procedures and mass spectrometry in each laboratory, so that isotopic data of water from different laboratories can be directly compared. The isotopic compositions of VSMOW and SLAP are defined as 6D - 0, ~180 - 0%0 (VSMOW) and ~D -428, 6180 =-55.5%o (SLAP)(Gonfiantini, 1978). When all samples and two laboratory standards (A and B) are prepared at the same conditions and their isotopic ratios are determined in the same mass spectrometry session against the same working standard gas (WS), the isotopic ratio of samples on the VSMOW-SLAP scale can be calculated as; 6Sample/VSMOW-SLAP- (a-~Sample/WS) + 6 a = (~A/VSMOW-SLAP - 6B/VSMOW-SLAP) / (~A/WS - 6B/WS) 6 - [(6B/VSMOW-SLAP" ~A/WS) - (~A/VSMOW-SLAP" ~B/WS)] /(6A/WS - ~B/WS)
[1.18a1
Stable Isotope Analysis of Water and Aqueous Solutions ...
35
where 6Sample/VSMOW-SLAP is the 6 value of sample relative to the VSMOW-SLAP scale, for example (Coplen, 1988; 2001). The VSMOW-SLAP factor (fVSMOW-SLAP) is the overall correction for stretchingshrinking an isotopic scale, which is characteristic for the sample preparation procedures and mass-spectrometric conditions in each laboratory;
f VSMOW- SLAP =
(~)B/VSMOW- SLAP - ~)A/VSMOW- SLAP)[ 1 + ( 10-3 " ~)A/WS)]
[1.19al
(6B/WS- 8A/WS)[ 1+ (10 -3 " ~A/VSMOW - SLAP) ]
When VSMOW and SLAP are two standards, the above two equations are reduced to; 6Sample / VSMOW-SLAP - 6SLAP / VSMOW'(6Sample/WS - 6VSMOW/WS) / (6SLAP/WS - 6VSMOW/WS)
[1.19b]
and fVSMOW-SLAP = 6SLAP/VSMOW'[1 + (10 -3" 6VSMOW/WS)] / (6SLAP/WS - ~VSMOW/WS)
[1.19b] For isotopic studies using water enriched in heavy isotopes, several isotopic standard waters with 6D and 6180 values greater than VSMOW are available from IAEA (Wong et al., 1993). An inter-laboratory comparison of two juices was also conducted (Koziet et al., 1995). The use of commercially available standard gases (H2, CO2) for the actual standardization of samples is not recommended because there is no guarantee that the sample preparation and calibration methods used for such standard gases are consistent with those in other laboratories. However, such gases make good working standards if their 6 values are close to those of analytes. Another choice for a working standard (or for other reference samples) is to seal isotopically identical aliquots of gas into glass tubes or ampoules, and break open a new tube either daily or when needed; such ampoules can be prepared according to the method described in Coplen & Kendall (1982). 1-6 Comments and future directions
It has been often noticed in the literature that only 6180, not 6D, values are reported for many hydrologic water samples. This is due mainly to the fact that 6D measurements of water by conventional dynamic and batch reduction methods discussed before have been more labor-intensive and time-consuming (i.e., more expensive) than 6180 measurements by the classic CO2-water equilibration method. Recent developments in analytical techniques for the 6D value of water (H2-water equilibration and automated reduction systems) are expected to fill this serious gap. Automated gas (H2 and CO2) equilibration methods are very suitable for this purpose, and are becoming standard techniques in many isotope hydrology laboratories.
36
Chapter 1 - J. Horita & C. Kendall
For natural hydrologic samples, it is clearly desirable to determine both 5D and 5180 values, because they can provide two, rather than one, dimensional information. For example, the so-called "deuterium excess" value (d = 5D - 8.5180) is a very useful parameter for identifying the source region and climate conditions for precipitation and old groundwaters. Furthermore, plotting 5D versus 5180 values of samples (before reporting or using the data), and comparing the values to the GMWL, is one of the best means for identifying samples that perhaps should be reanalyzed (i.e., outliers). However, it must be noted that 5180 and 5D values of natural waters, especially young meteoric waters, are usually highly correlated to each other, indicating that little additional information could be gained by measuring the second isotope ratio, which might double the cost of the analysis. Many researchers have, thus, concluded that for broad surveys of waters (and especially hydrograph separations where a hundred stream samples might be collected over a single rain storm), 5180 alone can be used as an investigatory tool. If there are no significant differences in 5180 values among a set of related samples (e.g., a glacier core), it is unlikely that there will be significant differences in 5D. Hence, considerable money can be saved by examining the 5180 results before considering what to analyze for 5D. It is recommended that some 10% of the samples analyzed for 5180 in such surveys should also be analyzed for 5D, in case there is some unexpected information to be added by the second isotope. For a good, but rare example of the usefulness of analyzing both 5180 and 5D of waters for hydrograph separations, see page 276 of Kendall et al. (1995b). For small-size (< 0.1 mL) water samples, metal (U, Zn, Cr, Mn, etc.) reduction is still the method of choice. Batch reduction methods and automated dynamic system by means of an autosampler can equally well serve this need. However, the quality of isotopic data obtained by these reduction methods, especially by batch methods, varies, depending on many known and unknown factors (reagent, sample, vessel, etc.) as discussed above. Cross-analysis with other methods and/or duplicate measurements of samples are recommended, especially for waters with unusual composition and matrix (salinity, pH, etc.). Recent new developments of continuous-flow mass spectrometry are making it possible to determine both 5D and 5180 values of water samples simultaneously (high-temperature carbon pyrolysis method). Such techniques are particularly useful for small-size samples (e.g., water vapor, single fluid inclusions). With decreasing sample size, possible isotopic fractionations associated with extraction from the matrix, contamination of the water, and adsorption/desorption of water onto and from the entire system will become major issues (Morse et al., 1993). It is interesting to notice that the past decade witnessed the most significant progress in analytical techniques for stable isotope analysis of water since the original developments almost half a century ago. As a result, the number of stable isotope laboratories in the world increased rapidly, and an exploding number of isotopic analyses of water are being reported in the literature. Overall quality control of isotopic data (e.g., sample documentation and tracking, information management for sample preparation, mass-spectrometric measurements, and data reduction) is becoming an
Stable Isotope Analysis of Water and Aqueous Solutions ...
37
important issue.
Acknowledgement
We thank W.A. Brand, T.B. Coplen, J.R. Gat, R. Gonfiantini, J.R. O'Neil, A. Schimmelmann, A. Tanweer, and W.W. Wong for their comments and reviews. The work of JH was sponsored by the Division of Chemical Sciences, Geosciences, and Biosciences, Office of Basic Energy Science, U.S. Department of Energy under contract number DEAC05-00OR22725, Oak Ridge National Laboratory, managed and operated by UT-Battelle. LLC. Use of firm, brand, and trade names in this manuscript is for identification purposes only and does not constitute endorsement by the U.S. Geological Survey.
Handbook of Stable Isotope Analytical Techniques, Volume 1 P.A. de Groot (Editor) 9 2004 Elsevier B.V. All fights reserved.
CHAPTER 2 Conventional and Less Conventional Techniques for Hydrogen and Oxygen Isotope Analysis of Clays, Associated Minerals and Pore Waters in Sediments and Soils H. Albert Gilgl, Jean-Pierre Girard2 & Simon M. F. Sheppard3 Lehrstuhl ffir Ingenieurgeologie, Technische Universit~it M~inchen, Arcisstr. 21, 80290 M(inchen, Germany 2 BRGM, Department Analysis and Mineral Characterization, BP6009, 45060 Orl6ans cedex 2, France 3 Laboratoire de Science de la Terre and CNRS-UMR 5570, Ecole Normale Sup6rieure de Lyon, 46 All6e d'Italie, 69364 Lyon, France e-mail: 1
[email protected] 1
2.1 Introduction
Clay minerals form in a wide range of environments near the earth's surface during weathering, diagenetic processes or hydrothermal alteration. Hydrogen (D/H) and oxygen (180/160) isotope studies of clays are particularly useful in understanding their genesis, low-temperature fluid-rock interaction, sedimentary processes and paleoclimates and have a wide range of applications (e.g., Girard & Fouillac, 1995; Mizota, 1996; Sheppard & Gilg, 1996; Savin & Hsieh, 1998). Due to the very fine particle size and the commonly complex mineral composition of clays, special pretreatment and isolation techniques are required prior to isotope analysis. This paper reviews the most commonly applied and some unconventional techniques related to the isotope analysis of clays. Additionally we include sections on isotope analysis of pore waters and associated non-clay minerals in clay-rich rocks.
2.2 Techniques of clay separation
A major difficulty in isotope studies of natural clays is to extract pure, monomineralic clay fractions from rocks. This is because clays often occur as intimate physical mixtures of intergrown clay and/or non-clay minerals. Sample purity is critical to the precise determination of isotope ratios in clays. Isotope data collected on imperfectly segregated clay samples may be acceptable for studies in which accuracy is not required. However, it should not be encouraged. The presence of O-free and/or Hfree mineral contaminants mixed with clays usually has no bearing on the accuracy of isotope ratio determinations, at least to the extent that contaminants do not react during the process of O and/or H extraction either to form new minerals containing O and/or H, or products leading to mass interferences in the mass spectrometer. In contrast, the presence of O and/or H-bearing contaminants (organic and/or inorganic) in clay separates may shift measured 8 values significantly away from the clay value,
Conventional and Less ConventionalTechniquesfor Hydrogenand OxygenIsotope ...
39
eventually leading to erroneous interpretations. Particular care must be given to the separation and purification of clay samples, as well as to the mineralogical characterization of purified clay segregates. This section is a summary of clay separation techniques most commonly used for sediments and soils. Additional information can be found in Girard & Fouillac (1995). A flow chart of recommended laboratory procedures is given in Figure 2.1.
2.2.1 Physical separation techniques Sample selection Clays occurring as fracture a n d / o r vug filling can often be sampled from slabs or thick sections by use of high resolution micro-drilling techniques. In samples where the clay of interest is spread throughout the rock, the clay fraction must be extracted from the bulk rock using physical a n d / o r chemical treatments. Such treatments can be very time-consuming and should be undertaken on selected samples. That is, samples for which chances of success are greatest based on careful and educated characterization. Standard techniques such as optical microscopy, scanning electron microscopy (SEM), electron microprobe analysis (EMPA) and X-ray diffraction (XRD) are routinely used to determine the nature and the abundance of the mineral phases present in a bulk sample (Tucker, 1988). Relating XRD data obtained on bulk or clay-size fractions with thin section observations may be difficult. Beaufort et al. (1983) developed a powerful approach combining EMPA with XRD on minute amounts of clays microsampled from thin section to precisely identify clays from different micro-sites in a rock sample (Girard et al., 1989). The first criteria for sample selection is the abundance of the clay mineral of interest. However, other criteria that greatly influence the ease with which the clay can be segregated out should also be considered. For instance, a sandstone containing abundant diagenetic kaolinite systematically mixed with illite will be less amenable to isolating the kaolinite fraction than a sample in which kaolinite is far less abundant but free of illite (illite and kaolinite are particularly difficult to separate from one another). The spread in grain size distribution and the freshness of the mineralogical constituents of the rock are also important criteria for selection. Because physical segregation of clays largely relies on grain size separation techniques (see below), generally the finer the constituents and the lesser the spread in grain size, the more difficult the clay separation. The mode of occurrence of the clay in the rock is also important to consider. Clays replacing other silicates are often difficult to segregate due to the presence of fine-grained relicts of the replaced mineral and other reaction products. Notable exceptions can be the argillic alteration of porphyritic igneous rocks and granites (Sheppard et al., 1969). Finally, major difficulties are to be expected with samples containing two or more populations-generations of a same clay mineral. It is notoriously difficult to separate a pure fraction of diagenetic illite from shales because they are rich in detrital illite (Girard & Fouillac, 1995). Similarly, it would be virtually impossible to isolate weathering kaolinite from a soil developed on a kaolinite-rich bed rock. As a rule of thumb, ideal samples for successful clay segregation usually are samples
40
Chapter 2 - H.A. Gilg, J.-P. Girard & S.M.F. Sheppard
Figure 2.1 - Flow sheet for clay separation and purification process for stable isotope analysis (modified after Girard & Fouillac, 1995).
Conventional and Less Conventional Techniques for Hydrogen and Oxygen Isotope ...
41
in which the clay of interest is abundant, occurs in a grain size range different from other mineral constituents of the rock, and is not physically mixed with other clay minerals.
Bulk rock disaggregation In the particular situation of rocks containing no other silicates besides the clay of interest, the clay fraction can be obtained by appropriate chemical techniques (see below) that will digest the non-silicate fraction leaving a pure fraction of the clay behind. In almost all other situations, the bulk rock must be disaggregated in order to physically separate the clay fraction. The most critical point in the disaggregation process is to preserve the original, natural grain size of the different mineral constituents. The use of a mechanical crusher (ball mill crusher or similar type) is therefore not recommended. Mechanically crushing or powdering a dry rock sample artificially homogenizes the grain size distribution of the various minerals, and unavoidably complicates segregation of clay particles. Gentle disaggregation techniques performed on wet samples yield better results. Poorly indurated samples (sediments, soils...) can be disaggregated by use of mortar-and-pestle or sonification (bath or probe) techniques. When using mortar-and-pestle, hitting and sheering the sample should be avoided or minimized. When using ultrasonic probes at high power for long duration, one should be aware of potential contamination by particles released from the probe and/or from the glassware (Rendigs & Commeau, 1987). The heatingfreezing cycling technique developed by Liewig et al. (1987) is best for preserving original grain-size. However, it is time consuming and requires samples with relatively high permeability. Other methods, more specifically adapted to soil samples and based on the use of Na resins (Amberlite) and sonification, have been described by Bartoli et al. (1991). Grain size separation Although grain size separation is often performed on the untreated disaggregated bulk fraction, it is best to perform appropriate chemical treatments prior to grain size segregation because a significant proportion of the clay may be enclosed in polymineralic aggregates cemented by organics, carbonates, sulfates, etc. Chemical treatments (see next section) will break these aggregates apart, freeing the clays. In addition, chemical treatments also help increase the stability (dispersion) of a clay suspension in water, which in turn helps performing high quality grain size separation. Grain size separation is traditionally done by settling in water columns or by centrifugation, and application of Stokes law. Purest clay fractions are usually found in the <10 ~m size range. If the amount of material permits, it is useful to separate several consecutive size fractions (i.e., <0.1, 0.1-0.2, 0.2-0.5, 0.5-1.0, 1.0-2.0, 2.0-5.0 and 5.010.0 ~m). Ultrafine size fractions (<0.01, 0.01-0.05 and 0.05-0.1 ~m) can be separated out by continuous flow ultracentrifugation techniques. In general, illite and illitesmectite tend to concentrate in the finest size range (< 2 ~m) while kaolinites and chlorites may be more abundant in coarser size ranges (> 2 ~m).
42
Chapter 2 - H.A. Gilg, J.-P. Girard & S.M.F. Sheppard
It may be useful to characterize the grain size distribution of the bulk clay-size fraction prior to grain size separation by use of standard particle sizers (COULTER COUNTER@, laser diffraction particle sizers, etc.). Identification of different size modes, possibly directly corresponding to different minerals, will guide the strategy for grain size separation in order to isolate each of the different constituents. It may also help separating different generations of a clay mineral present in a sample (GiralKacmarc~ et al., 1998).
Na-polytungstate densimetry Densimetry separation techniques based on the use of heavy liquids (bromoform, TBE ...) are known to perform poorly on clays. A method specifically adapted to clays and using Na-polytungstate solutions as the separating medium was recently developed (Cassagnab6re, 1998). Na-polytungstate is highly soluble in water and solutions with densities ranging from 2.00 to 3.00 g/cm3 can be made with a precision of 0.005 g/cm3. The separation process is accelerated by use of centrifugation. Because the density exhibited by clay minerals is variable and because significant overlap exist between the different clay groups, success of the Na-polytungstate densimetry method is not guaranteed for any given sample. Trials must be performed prior to applying the technique to large series. Applications to segregate diagenetic kaolinitedickite from reservoir sandstones (Cassagnab6re, 1998) and weathering kaolinite from lateritic soils (Girard et al., 2000) have been successful. This technique, howeveI; is not recommended for samples rich in exchangeable Ca (formation of Ca-polytungstate would modify solution density). In spite of this, it holds great promise for future investigations.
High gradient magnetic separation
High gradient magnetic separation techniques can be used to concentrate and segregate clay minerals from one another (Russel et al., 1984; Tellier et al., 1988). The clay suspension is circulated in a continuous flow loop through a strong magnetic field (2 Tesla or more) applied to a plexiglass box filled with steel wool which acts as a trap for magnetic particles. The efficiency of this technique requires to find the right compromise between the strength of the magnetic field, the concentration of the clay suspension and the flow rate. It involves a fair amount of trial and error depending on the nature of the clay samples. Good success at segregating Fe-rich clays (chlorite group or berthierine) was reported in many applications (Righi & Jadault, 1988; Hillier et al., 1996; Hornibrook & Longstaffe, 1996; Ryan et al., 1998). Obtaining pure separates of non-magnetic clays is often more difficult.
2.2.2 Chemical purification
In many studies, physical segregation techniques alone will not yield a pure fraction of the studied clay and purification must be achieved by use of specific chemical treatments. Most of the selective dissolution treatments described below have been developed by soil scientists in the fifties and sixties (Jackson, 1979). Many have been modified to variable degrees and improved as they were applied to other types of rocks. In this section, we present the treatments that are most commonly used today.
Conventional and Less Conventional Techniques for Hydrogen and Oxygen Isotope ...
43
Selective dissolution of mineral contaminants Carbonates and sulfates" Ca-Mg carbonates and sulfates (gypsum) are readily removed by treating the sample with buffered (pH=5) 1N acetic acid for a few hours at 25-50~ 1N HC1 can equally be used to remove carbonates, but may affect any chlorite present in the sample (see below). The dilute acid treatment will also remove exchangeable cations and salts (halite, sylvite...). Longer duration, higher temperature and greater acid concentrations are necessary to get rid of Fe carbonates (siderite). Iron oxides and hydroxides: Two techniques are commonly used to remove Fe oxides and hydroxides: the Na-dithionite-citrate-bicarbonate method (DCB) of Mehra & Jackson (1960) and the photolytic ammonium oxalate method of de Endredy (1963). The DCB method is extremely efficient and rapid (15-30 min. at 70~ 12 hours at 25~ The photolytic ammonium oxalate method is rapid (30 min.) but may affect hydrogen isotopes (see below). Multiple treatments are often necessary for samples extremely rich in iron (tropical soils, banded iron formations...) with either method, and may induce alteration of the clays (Giral, 1994). Ryan & Gschwend (1991) developed a technique, referred to as the Ti(III)-citrate-EDTA method, less aggressive to clay minerals. However, its potential effect on the isotopic composition of clays is yet to be explored. The EDTA method of Borggaard (1982) is even more gentle to clays, but it only removes amorphous or poorly crystalline Fe oxide-hydroxides and not crystalline hematite or goethite. Amorphous silica and aluminosilicates: Amorphous silica, aluminum hydroxides and poorly ordered aluminosilicates (allophane, imogolite, etc.) are dissolved by treatment with hot 0.5 M NaOH for 2.5 minutes (Hashimoto & Jackson, 1960; Smith, 1994). Chlorites and montmorillonites are generally not attacked by this method, but crystalline gibbsite can be dissolved (Smith, 1994). Less aggressive, but also less effective (Smith, 1994) are treatments with hot or cold 3-5% Na2CO3 solutions (Longstaffe, 1986; Bird et al., 1992). Chlorite and gibbsite" Chlorite can be selectively dissolved from illite-chlorite clay mixtures by treating the sample with 1N HC1 at 80~ for 1-2 hours (Eslinger, 1971; Longstaffe, 1986). This method is based on the fact that chlorite dissolves faster than illite in HC1 (Ross, 1969). The treatment should be followed by dissolution of amorphous aluminosilicates produced by the decomposition of chlorite using 5% Na2CO3 solution (Longstaffe, 1986; Bird et al., 1992) or 0.5 M NaOH (Hashimoto & Jackson, 1960). In a similar way, gibbsite can be removed from kaolinite-gibbsite mixtures derived from lateritic soils by treating the sample with boiling 6N HC1 for 1-3 hours (Bird et al., 1992, Giral, 1994). It is important to keep in mind when using these methods that part of the illite or the kaolinite fraction (the smaller and/or more poorly crystallized crystals) will be lost. This may affect the representativeness of the clay sample (Longstaffe, 1986; Giral, 1994). Removal of organic matter Most clay-bearing sedimentary rocks, including soils, shales and mudstones, contain organic matter. Although usually present as a minor or trace constituent, organic matter is an additional reservoirs for hydrogen and, to a much lesser extent, oxygen in the rock (H/C and O / C atomic ratios in organics are usually between 0.5 to 4, and 0.1
44
Chapter 2 - H.A. Gilg, J.-P. Girard & S.M.F. Sheppard
to 0.5 respectively). Separation of organic matter from clays raises a number of special problems because (1) it may be present not only as a separate phase but also adsorbed on the surfaces or edges of clays or accommodated in interlayer spaces, (2) it is not detected by most of the widely available mineral identification techniques, and (3) it can resist conventional cleaning treatments (e.g. Farmer & Mitchell, 1963; Jackson, 1979). In certain ways, the problems of removing organic matter from clays is comparable to the removal of organic matter from carbonates. In both cases, a given technique may be applied without carefully checking the effects of the method on the isotopic properties of the mineral, and comparative studies have very rarely been published in detail. Another major problem common to both is the difficulty of having standard samples which are really comparable to natural samples in order to check rigorously the effects of the pretreatment processes. Leaching techniques using organic solvents such as methanol, ethanol, benzene, toluene or hot (60~ chloroform and a Soxhlet-type apparatus are useful to remove liquid hydrocarbons from samples such as petroleum reservoirs or source rocks. However, leaching techniques do not remove organics entirely and further cleaning must be done (Monin et al., 1978). The most efficient way to clean samples from organic matter relies on techniques permitting complete oxidation of organic matter. Several wet and dry pretreatment techniques for removing organic matter have been developed (see Table 1 in Sheppard & Gilg, 1996; Menegatti et al., 1999). They are considered under the following headings: hydrogen peroxide, sodium hypochlorite, sodium peroxodisulphate, plasma ashing, step-heating. Other techniques, such as the use of bromine water at 40~ (Mitchell & Smith, 1974) are not discussed. The present situation suggests that the hydrogen peroxide, sodium peroxodisulphate and plasma ashing methods are the most satisfactory. Hydrogen peroxide: Pretreatment of soil samples with 30% H202 solution is the recommended technique given by Jackson (1979) before carrying out chemical analyses. The same technique has been applied for isotopic analyses (e.g. Yeh, 1980). Reaction with H 2 0 2 is rapid, usually being complete within 1 /2 hour. Some organicrich shale sample, however, have to be repeatedly treated with H 2 0 2 for more than a day to effectively remove organic matter (Hyeong & Capuano, 2000). The efficiency of organic matter removal has to be checked by total organic carbon (TOC) analysis. Bleaching of the sample and absence of bubbling upon addition of H202 are not indicators of complete organic matter removal (Hyeong & Capuano, 2000). The reaction is so strongly exothermic that the sample-H202 mixture in a beaker can erupt violently like a volcano and splutter over the top of the beaker, with loss of sample. Placing the beaker in a cold water bath can prevent this problem (and perhaps retard isotopic exchange with the clays). Eslinger (1971) and Yeh (1974, 1980) have stated that neither the hydrogen nor oxygen isotope compositions of their minerals and rocks (organic contents not given) were modified. In the absence of good standards and further verification with, in particular, organic-rich shales, it should not be automatically assumed that the hydrogen peroxide treatment cannot modify the isotopic composition of the clays. In a recent study, Hyeong & Capuano (2000) found that H202 treatment results in a shift in the 6D value of H26 smectite
Conventional and Less Conventional Techniques for Hydrogenand Oxygen Isotope ...
45
standard by up to -8 %o and a illite/smectite-rich Gulf Coast shale sample by-13 %o. The shift was accompanied by a reduction in the reproducibility of the analyses with standard deviations increasing from +2 %o before the H202 treatment to +12 %o after treatment. The authors suggest that deuterium is no longer evenly distributed in the samples and invoke the retention or removal of a hydrogen bearing phase as the cause of the decreased reproducibility. Sodium hypochlorite: The pretreatment of biogenic carbonates with a sodium hypochlorite (NaC10) solution is widely practiced and has been applied to laterites and shales (Taieb, 1990; Bird et al., 1992). From a series of tests on the same samples, Taieb (1990) concluded that the yields were generally lower and the isotopic values were less reproducible compared to the hydrogen peroxide method. The hypochlorite technique is not discussed further and is not recommended. Sodium peroxodisulphate: A new method of organic matter removal from clay-bearing sediments using sodium peroxodisulphate (Na2S208) combined with a NaHCO3 buffer (pH of 7.8 to 8.5) has been described recently by Meier & Menegatti (1997) and Menegatti et al. (1999). Best yields were achieved with a oxidant/sample mass ratio of 40 and a buffer/oxidant mass ratio of 1.1 at 98~ The method is very efficient: more than 98 wt. % of the total organic matter from a black shale containing 10.9 wt.% Corg was removed, whereas less than 87 wt.% of the total organic matter was oxidized from the same sample using the sodium hypochlorite and hydrogen peroxide methods (Meier & Menegatti, 1997). The application of the sodium peroxodisulphate method to some standard clay minerals (kaolinite, illite and montmorillonite) have not altered their oxygen and hydrogen isotope compositions (Menegatti et al., 1999). Plasma ashing: Plasma ashing refers to the decomposition of organic substances by reaction with gas excited in a radio frequency (RF) discharge at low pressure (e.g. Gleit & Holland, 1962; Brenna & Morrison, 1984; Taieb, 1990; Hogg et al., 1993). The technique is also termed "low temperature ashing or LTA". A number of laboratory scale ashers or plasma furnaces are commercially available (e.g. Emitech, USA; Harrick Scientific Corporation, USA; VG Microtech, UK) and they have similar specifications. The powdered sample is spread out on a non-metallic support such as a glass thin section, watch glass or silica glass sample holder and placed in the reaction chamber that can be evacuated to a pressure of about 0.1-0.2 torr by a rotary vacuum pump. The sample is most efficiently loaded in as dry a form as possible to reduce pump down time. RF power is applied around the reaction chamber (13.56 MHz; up to 100-150 watts), for example a cylinder 150 mm long by 100 mm diameter, and the energy is transferred to the carrier gas that is being bled into the chamber, now at a pressure of about I torr. The carrier gas is excited by the RF field leading to partial ionization and dissociation of the gas molecules. This more reactive product attacks the surface of the sample. A visible glow or plasma around the sample can be seen. The working gas can be air or, more efficiently, oxygen, because it is principally oxygen which reacts with the organic material to give volatile constituents that are carried away in the gas stream. Non-volatile minerals remain on the sample support that becomes only warm during the oxidation reactions (probably <60~ as the support can be handled immediately after reaction is stopped and the chamber is let up
46
Chapter 2 - H.A. Gilg, J.-P. Girard & S.M.F. Sheppard
to atmospheric pressure); hence the designation LTA. Note that "low temperature" of LTA may be a little misleading. At the site of interaction between the plasma and the organic matter, the temperature could be substantially higher than the overall temperatur (<60~ of the sample, because the oxidation reaction liberates energy. The temperature may be high enough to locally modify the associated mineral. For example, although no evidence for the modification of clay minerals has been reported, biogenic aragonite may be isotopically altered during LTA, even though hydrothermal aragonite remains unaffected (SMFS, unpublished data). Further tests are warranted. If oxygen is used as the carrier gas, the ordinary rotary pump oil must be changed to Fomblin| oil to avoid any chance of production of an explosive mixture between hydrocarbon oil and oxygen. Note that the fomblinisation of a rotary pump that has already worked with conventional oils involves more than just changing the oil; the pump needs to be thoroughly cleaned and seals replaced, if necessary. As indicated above, the oxidation reactions takes place at the surface of the sample. For this reason the clay sample needs to be spread out as thinly as possible on its support to expose the maximum surface; it is turned over several times during the ashing process. There are no hard and fast rules concerning the best working conditions. The variables are nature, abundance and occurrence of organic matter, mass and granulometry of the sample, exposure time, plasma power, nature of carrier gas and its flow rate, position of sample in chamber, and pressure in the chamber. To give an idea, tests on a simulated carbonate-free black shale, produced by mixing natural quartz, kaolinite, illite, smectite and hard coal (13 wt. %; H / C - 0.7) that had passed through a 63 mm sieve indicate that after 2 hours of exposure in a plasma furnace operating at 150 watts, >90% of the coal carbon had been removed. The hydrogen isotope analysis of this pretreated whole rock shale is identical within analytical precision to that calculated from the clay mineral constituents (Taieb, 1990). In the author's laboratory (SMFS), samples are run in a 100 watt plasma furnace for at least 3 hours unless specific test indicate otherwise. Step-heating: For H-isotope analysis, preheating of the sample, under vacuum, at temperatures below 350~ and thus below the decomposition temperatures of the associated clay minerals has been tried (Sheppard & Gilg, 1996). Although certain hydrocarbons, etc. will decompose, complete removal of hydrogen from the organics is not evident and exchange between the liberated organic hydrogen and hydroxyl hydrogen is possible. The technique may work satisfactorily for samples in which organic compounds are thermally unstable and clay minerals do not dehydroxylate below 350~ But, it is unlikely to be a generally applicable technique.
Cleaning and purity checks Following chemical treatments, the purified clay should be thoroughly washed and cleaned of all chemicals prior to drying by repeated centrifugation-agitation cycles in distilled water. If time permits, a more complete cleaning is obtained by dialyzing clay suspensions for a few days in distilled water (periodically renewed). The mineralogy and the purity level of the different size fractions segregated can then be determined by XRD and transmission electron microscopy (TEM). If the amount of clay available is sufficient (>lg), chemical analysis by X-ray fluorescence, atomic
Conventional and Less Conventional Techniques for Hydrogenand OxygenIsotope ...
47
absorption spectrometry and/or inductive coupled plasma techniques can be performed. The latter techniques usually can detect smaller amounts of contamination than XRD or TEM. DTA-TGA analysis and FTIR spectrometry may also be very useful to identify and quantify minerals present in mixtures of clays, in particular clay minerals closely related chemically or polytypes (i.e., kaolinite, dickite, nacrite). Yields of organic carbon removal can be analyzed using conventional techniques, e.g., coulometric titration. Purity checks are most critical as very small amounts of contamination may result in significant shifts of measured isotopic compositions from those of the clays and erroneous interpretations (Girard & Fouillac, 1995).
Effects of separation treatments on the isotopic composition of clays Most of the treatments presented above have been verified not to affect significantly (i.e. beyond analytical uncertainty) the isotopic composition of clays. Many of these tests were done in a variety of studies, over the years in a patchy fashion, and using different experimental conditions and different clay reference materials depending on availability for each particular study. Many were conducted at a time when refined mineralogical techniques for the characterization of clay minerals were not routinely used. A comprehensive validation of each of the different chemical treatment is yet to be carried out on well characterized international clay standards with the ambition to investigate specimen of variable chemistry and/or crystallinity from all groups of clay minerals. The current state of knowledge with regards to the absence or existence of significant isotopic effect for most common clay minerals is summarized in Table 2.1. References of relevant studies in which demonstrative validation tests were reported are indicated or shown in the section Appendix. The reader is referred to these publications for the details of the experimental conditions. Investigators with an interest in modifying experimental conditions of any particular chemical treatment or in applying a chemical treatment to clay samples significantly different (chemically- or crystallinity-wise) from those used in the reference study are advised to conduct their own validation tests.
2.3 Isotopic analysis of clay minerals 2.3.1 Hydrogen isotopes Clay minerals have various reservoirs of hydrogen. These include hydrogen in hydroxyl groups, adsorbed water and in some cases interlayer water. Adsorbed and interlayer water of clays exchange isotopically rapidly with atmospheric vapor at ambient temperatures (Moum & Rosenqvist, 1958; Savin & Epstein, 1970). Consequently, they must be removed from OH hydrogen prior to analysis. For most clay minerals, pre-heating in vacuum at temperatures of about 200~ for two or more hours removes most of these waters without isotopic exchange with the structural OH hydrogen (Savin & Epstein, 1970). However, evidence for significant hydrogen isotope exchange between interlayer water and OH-hydrogen during degassing of halloysite were detected by Lawrence (1970). Additionally, recent experimental studies (Longstaffe & Mizota, 1999; Hsieh & Yapp, 1999) document rapid hydrogen isotope exchange between water and halloysite, especially the hydrated 10/~ form, even at
Table 2.1 - Effect of selective dissolution and other treatments on the isotopic composition of clays and associated minerals Mineral
Treatment
Effect on 6180
Effect on 6D
Recommended o u t g a s s i n g T~
U s e f u l references
Illite
Ac-Ac, H202, DCB, P1-Ash
none
none
200-250~
Illite/smectite Smectite
1N-HC1 Ac-Ac, H202, DCB Ac-Ac, H202, DCB, P1-Ash
none none none*
? none none-strong
200-250~ 180-250~
none none* none none none none small
none none none-slight small ? none ?
200-250~ 200-250~
Halloysite
Ac-Ac, H202, DCB Ac-Ac, H202, P1-Ash DCB Am-Ox 1N-HC1, 6N-HC1 Na-polytungstate densimetry H202, DCB
100~
Eslinger (1971), Yeh (1974, 1980) Hogg et al. (1993), Taieb (1990) Eslinger (1971) Eslinger (1971), Yeh (1974, 1980) Yeh (1980), Taieb (1990), Stern et al. (1997), Hyeong & Capuano (2000) Eslinger (1971), Yeh (1980) Yeh (1980), Ta~'eb (1990) Bird et al. (1992) Giral (1994), Appendix Giral (1994), Appendix Eslinger (1971), Bird et al. (1992), Giral (1994) Appendix, Girard et al. (2000) Hsieh (1997), Hsieh & Yapp (1999)
Goethite Gibbsite
5M N a O H DCB
none none
strong none
100-120~ 110-130~
Yapp (1991), Chazot (1997) Bird et al. (1989)
Chlorite
Kaolinite
Ac-Ac = buffered acetic acid 1N; H202 = h y d r o g e n peroxide 30%; DCB = Na-dithionite-citrate-bicarbonate; Am-Ox = a m m o n i u m oxalate; P1-Ash = plasma asher; * 9presumably because of the absence of effect on 6D value.
4~ oo
r
to !
4~ r
t
Conventional and Less Conventional Techniques for Hydrogenand OxygenIsotope ...
49
room temperature. Thus, D-H ratios of halloysite obtained by conventional techniques are probably only of limited value. Minor D-H exchange problems during degassing have also been reported by Lawrence (1970) and Fagan & Longstaffe (1997) for some smectites. Marumo et al. (1995) showed that the conventional degassing temperature of 200~ may not be sufficient to remove completely interlayer water from some smectites. Especially, poorly crystallized smectites seem to retain considerable amounts of interlayer water at temperatures at which dehydroxylation starts. Residual interlayer water after incomplete degassing is strongly enriched in deuterium with 6DSMOW values up to + 29%o (Marumo et al., 1995). Thus, significant contamination from residual D-enriched interlayer water has to be expected in the case of incomplete interlayer water removal. Thermogravimetric or differential thermogravimetric analyses in vacuum, stepwise heating experiments and precise calculation of hydrogen yields are very useful to constrain appropriate temperature cuts for separating interlayer and hydroxyl water of such samples. Dehydroxylation of clay minerals under vacuum generally starts at temperatures above 350 to 400~ (Brindley & Lemaitre, 1987). The previously degassed samples are further heated without exposure to the atmosphere either with a resistance furnace or an induction oven to temperatures above 900~ The expelled water is collected with a liquid nitrogen trap. Clay minerals with high Fe2+ concentrations, e.g. some chlorites and smectites, can produce significant amounts of free hydrogen during dehydroxylation which has either to be oxidized to water using for example a CuO furnace or collected with a Toepler pump. The dehydroxylation water is subsequently reduced to hydrogen using, for example, hot uranium, zinc or other metals (e.g., Bigeleisen et al., 1952; Vennemann & O'Neil, 1993; see Volume IL Part 3, Chapter 1-2.3/12 of this book) and its isotope composition is measured. Analytical reproducibility is typically better than +2 %o for kaolinites, but often worse than that for smectites and some illites.
2.3.2 Oxygen isotopes The conventional method to extract oxygen from clays is the fluorine oxidation, or fluorination, technique developed by Taylor & Epstein (1962) and Clayton & Mayeda (1963). Typically ~ 10 mg of clay is reacted with excess F2 or BrF5 in a nickel tube at 500-600~ overnight. C1F3 has also been used as a fluorinating agent with some success (Borthwick & Harmon, 1982). The liberated oxygen is cleaned of extraneous reaction gases by standard cryogenic techniques and converted to CO2 over hot graphite prior to analysis on the mass spectrometer. Because clays are not refractory the fluorination reaction under such conditions is complete. Analytical uncertainty, based on reproducibility of measured 6~80 values, typically is +0.2-0.3 %o. As for hydrogen analysis, one practical difficulty of oxygen analysis of clays is to remove adsorbed water, and interlayer water when dealing with expandable clays, prior to fluorination reaction. Most of the non-structural water can be removed by placing the samples in a dried atmosphere (drybox with P205) for 12 hours to 24 hours (Savin, 1967, Savin & Epstein, 1970). This step can be extended to a few days for minerals such as smectites, smectite-bearing mixed-layers, vermiculite and halloysite
50
Chapter 2 - H.A. Gilg, J.-P. Girard & S.M.F. Sheppard
in order to reduce the amount of residual interlayer water as much as possible. Additional removal of non-structural water is done by heating clay samples in the nickel tubes under vacuum to a temperature of 150-250~ for I to 2 hours (Savin, 1967; Savin & Epstein, 1970). Recommended outgassing temperatures for common clays and associated minerals are given in Table 2.1. Alternatively, a prefluorination treatment of a few minutes at room temperature with a small amount of reagent may be applied (Clayton & Mayeda, 1963). Prefluorination at higher temperatures and/or for longer periods of time is not advisable because it might cause partial reaction of the clay and loss of structural oxygen (Hamza & Epstein, 1980; Hogg et al., 1993; Girard & Savin, 1996). On the other hand, if the prefluorination treatment is too short, water adsorbed on clays and on the internal wall of the Ni tubes may not be completely removed. Best suited conditions (time, amount of reagent, temperature) should be determined by the investigator. The possibility of using the laser fluorination techniques recently developed to extract oxygen from small amounts (< lmg) of silicates and oxides (Sharp, 1990) has not yet been extensively investigated on clays. The laser technique consists of loading many samples in a single reaction chamber, and reacting each of them in a sequence by heating with a laser in the presence of reagent (F2 or BrF5). Because clays are susceptible to passive fluorination (partial reaction at room temperature) the laser technique does not appear well suited. Cross-contamination of oxygen from the different samples present in the chamber may occur. We know of a few attempts to laser fluorinate clays with variable degrees of success. Unpublished preliminary experiments by one of the authors (JPG) using the 1 - 10 ~m fraction of Macon kaolinite (Girard & Savin, 1996) gave encouraging results. Average 6180 of thirteen laser determinations performed in three different loads was within 0.1%o of the conventional 6180 value and the associated standard deviation was 0.4%o. Even better results with an analytical precision similar to the conventional fluorination technique (0.2 to 0.3 %o) were reported from non-swelling clay minerals, such as kaolinite, dickite or illite analyzed at the Scottish Universities Research and Reactor Centre, East Kilbride (Maliva et al., 1999; A.E. Fallick, 2000, pers. comm.). Addition of an airlock module to a standard laser fluorination system allows samples to be transferred and reacted in the reaction chamber individually (Spicuzza et al., 1998b). This approach yielded good result for whole rock powders containing clays and holds promise for laser analysis of clay minerals (J.W. Valley, 1998, pers. comm.). The potentials of in-situ oxygen isotope measurements of clay minerals using secondary ion mass spectrometry have been explored by Williams & Hervig (1997) and Williams et al. (2001a). Their preliminary study on two standard clay minerals, montmorillonite (SWy-1) and kaolinite (KGa-1), however, revealed large analytical errors (~1 to 3 %o, lo) and mineral-dependend instrumental mass fractionations.
2.3.3 Intracrystalline fractionation of oxygen isotopes Two types of structural oxygen can be distinguished in clay minerals: hydroxyl oxygen (OH oxygen) and non-hydroxyl oxygen (non-OH oxygen). Hydroxyl oxygen mainly occurs in A1-OH, Fe-OH and Mg-OH bonds located in the octahedral sheets of
Conventional and Less Conventional Techniques for Hydrogenand OxygenIsotope ...
51
clays and in the brucite-like layer of chlorites. It represents 45% of the stoichiometric oxygen in kaolinite and chlorite, and 17% in illite. The non-OH oxygen occurs in M-OM bonds, where M is a cation other than H (commonly Si, A1, Mg, Fe...), and is found in the tetrahedral sheets and in the bonds bridging tetrahedral and octahedral sheets. Because the chemical bonds in which these two types of oxygen are involved are different, their isotopic composition (180/16 0 ratios) must be different (Taylor & Epstein, 1962). The fractionation between OH and non-OH oxygen is referred to as intracrystalline oxygen isotope fractionation and should be expected to be temperature dependent. In principle it ought to be possible to determine the temperature and the 6180 value of the water from which a clay mineral formed, directly from the 6180 values of OH and non-OH oxygen. Intracrystalline fractionation of oxygen isotopes in clays therefore constitutes a potential single mineral geothermometer. Two analytical approaches, partial fluorination and thermal dehydroxylation, have been used in an attempt to measure the intracrystalline fractionation of natural clays (Savin, 1967; Hamza & Epstein, 1980; Bechtel & Hoernes, 1990; Girard & Savin, 1996). In the partial fluorination approach, the clay sample is only partly reacted by use of an insufficient amount of reagent, an insufficient reaction temperature or an insufficient reaction time. This approach is based on the assumption that OH groups react faster than the remainder of the structure, and that OH oxygen is liberated prior to non-OH oxygen (Savin, 1967). In the thermal dehydroxylation approach, the clay is dehydroxylated by heating in vacuum and the liberated water is collected (by freezing) and analyzed isotopically (Bechtel & Hoernes, 1990; Girard & Savin, 1996). To be successful, the dehydroxylation process must occur with no kinetic fractionation during the liberation of water (only 50% of the OH oxygen is driven off as water) and no exchange between OH and non-OH sites during diffusion of the water through the particles. The results of analytical studies (Savin, 1967; Hamza & Epstein, 1980; Bechtel & Hoernes, 1990; Girard & Savin, 1996) agree with theoretical calculations (Zheng, 1993) from a qualitative standpoint. All indicate that OH oxygen is significantly depleted (up to a few tens of per mil) in 180 relative to non-OH oxygen in a same clay. They also suggest that intracrystalline fractionation decreases with increasing temperature of formation (Hamza & Epstein, 1980; Bechtel & Hoernes, 1990). However, the accuracy of intracrystalline fractionation measurements conducted with each of the two approaches remains questionable. Intracrystalline fractionations measured for similar low temperature (supergene) kaolinites in four different studies range from 17 to 27 %o, and are much smaller than values of ca. 40 %0 predicted by theoretical calculations (Zheng, 1993). In an extensive and detailed study of kaolinite and dickite, Girard & Savin (1996) showed that both partial fluorination and thermal dehydroxylation were associated with potential flaws previously overlooked and related to intrinsic properties of the clay samples, including particle size distribution and crystallinity. Partial fluorination experiments conducted under variable conditions (temperature, reagent amount, reaction duration) suggested that liberation of OH and non-OH oxygen was simultaneous and lead to unreliable determinations of the intracrystalline fractionation. Thermal dehydroxylation experiments yielded consistent results and was con-
52
Chapter 2 - H.A. Gilg, J.-P. Girard & S.M.F. Sheppard
cluded to be a more promising approach than partial fluorination, provided the less than I gm size fraction of the clay was used. A similar comparative study conducted on serpentine by Plas & Frfih-Green (1997) lead to the same conclusions. Although application of the approach to natural systems (Bechtel & Hoernes, 1993; Plas & Frfih-Green, 1997; Harris et al., 1999) and experimental studies (Fortier et al., 1994) seems to give reasonable results, additional developmental work is necessary before the method can be routinely used and applied as a single-mineral geothermometer.
2.3.4 Controlled isotope exchange technique The controlled isotope exchange technique (CIE) was developed by Labeyrie & Juillet (1982) for oxygen isotope analysis of biogenic opal (diatom silica) and was later applied to a Mg-rich smectite (stevensite) by Escande et al. (1984) . The method is based on the concept that such minerals contain two reservoirs of oxygen. One that isotopically exchanges readily with the ambient water (or water vapor) even during preparation and is related to interlayer water and silanol groups (Escande et al. 1984). The second reservoir does not show such an easy isotope exchange with the environment and its isotope composition reflects the conditions of mineral formation. In the CIE method, the weakly bound oxygen is exchanged with water vapors of known isotopic composition under controlled conditions (200~ and 24 h for stevensite) using a special setup (Figure 2.2). Following the controlled isotope exchange experiment, the sample is rapidly calcinated at 1000~ (and thus dehydroxylated) to prevent re-equilibration with water vapor during loading in a conventional silicate oxygen isotope extraction line. The precision thus obtained is +0.3 %o. From the experiments using isotopically different water vapors, it is possible to calculate the fractionation factor between easy exchangeable oxygen and water vapor at 200~ and additionally the
Conventional and Less Conventional Techniques for Hydrogenand OxygenIsotope ...
53
percentage of exchangeable oxygen in the clay. The latter varied between 40% and 2.5% for stevensites synthesized at temperatures between 25 and 400~ (Escande, 1983). Because the oxygen isotope values of the dehydroxylated clay mineral is measured, the results of this technique cannot be easily compared to samples analyzed by the conventional degassing method (section 2.3.2). In addition, sample dependent isotope exchange between OH oxygen and non-OH oxygen during calcination cannot be excluded (Girard & Savin, 1996). A similar controlled isotope exchange approach was developed for D / H measurements of goethites (Yapp & Poths, 1995) and halloysites (Hsieh & Yapp, 1999). The CIE method is very time consuming and cannot be regarded today as a standard technique for clay minerals, but it has potentials for clay minerals with considerable amounts of easy exchangeable oxygen or hydrogen, such as halloysites or some smectites.
2.4 Isotope analysis of pore water in clay-rich rocks The isotope analysis of pore waters in clay-rich rocks is very important in understanding the movements and chemical evolution of interstitial waters in low-permeability rocks or soils, but also their effects on the alteration of rocks. Liquid water occurs in very different "pools" or "compartments" within such rocks: a) free, mobile, interstitial or pore water, b) surface-adsorbed water and c) structural water in hydrous minerals and organics. Some pools of loosely bound water in clay-rich rocks are physically not separable, such as surface-adsorbed water and structural interlayer water in swelling clays or channel water in zeolites. These are here referred to as "sorbed" water. It has been demonstrated that sorbed water fractionates hydrogen and oxygen isotopes with pore waters in a rock (e.g. France-Lanord & Sheppard, 1992; AraguasAraguas et al., 1995). A technique of estimating the isotope composition of "sorbed" water in clay-rich rocks by mass balance calculations is given in France-Lanord & Sheppard (1992) and will not be discussed further here. Various techniques have been developed to analyze the isotopic composition of pore waters from low-permeability rocks (Anonymous, 2000). Each one has its potentials, but also its limitations. The techniques and some of their problems are discussed in the following paragraphs.
2.4.1 Physical techniques of water extraction High-pressure squeezing techniques to extract pore waters from clay-rich sediments are documented for example in Manheim & Sayles (1974), Entwistle & Reeder (1993), Cuevas et al. (1997), B6ttcher et al. (1997a) and here (Figure 2.3). Maximum stresses obtained using a hydraulic press are in the order of 100 to 200 MPa. The efficiencies of pore-water extraction increase with the initial moisture content (measured as the weight percentage of water loss during oven-drying at 105~ and vary between 0 and 70%. The minimum moisture content necessary to extract any pore water from an unconsolidated sediment using high-pressure squeezers is about 7 to 15% (Entwistle & Reeder, 1993; B6ttcher et al. 1997a), but may be higher for consolidated shales. Ultra-centrifugation (e.g. Litaor, 1988; Walker et al. 1994) is an alternative to squeezing for relatively water-rich and unconsolidated samples. Although these physical extraction techniques may not allow a complete extraction of the pore waters, they are generally less affected by fractionation processes compared to distillation processes. Ultrafiltration effects by a clay membrane (Coplen & Hanshaw,
54
Chapter 2 - H.A. Gilg, J.-P. Girard & S.M.F. Sheppard
1973), however, may result in some minor isotope fractionation (< 3%o for 6D and < %0 for 6180) during extraction from clay-rich samples. 2.4.2 Distillation methods In the vacuum distillation method (e.g., Jusserand, 1980), the sample is heated to temperatures ranging from 35 to 200~ under vacuum and the released water collected in a vessel cooled with liquid nitrogen. A variant of this technique has been used by Moreau-Le Golvan et al. (1997) for indurated shales with low water content (< 5 wt. %). The samples are crushed to variable grain sizes, mostly less than 5 mm. Contact time with the atmosphere was varied (15-60 min.). The chosen vacuum distillation temperature was 60~ and the duration 14 hours. A significant effect of grain size and atmospheric exposure time was detected mainly for 6180, but not for 6D values. The mean absolute standard deviation obtained for 6D values of 44 pore water extracts from a shale was + 1.3 %0. In the azeotropic distillation method (e.g., Revesz & Wood, 1990; Leaney et al., 1993) an immiscible hydrocarbon (e.g. toluene, hexane or kerosene) is added to the soil sample forming an azeotrope mixture. Heating of these mixtures leads to boiling at temperatures of less than 100~ The resulting vapor mixture is condensed and the extracted soil water forms an immiscible liquid in the condensation tube. Both methods require complete extraction of the water as the distillation process involves major isotopic fractionation if the amount of water remaining in the samples is greater than 2% of the original present (Araguas-Araguas et al., 1995). Both temperature and extraction yield affect the isotope composition of the distillate. Various comparative tests of the different distillation techniques have been published (e.g., Ingraham & Shadel, 1992; Walker et al., 1994; Araguas-Araguas Figure 2.3 - Hydraulic press for the extraction of pore waters from clay-rich rocks.
Conventional and Less ConventionalTechniques for Hydrogenand OxygenIsotope ...
55
et al., 1995). The results show that with increasing clay mineral content, decreasing grain size and decreasing water content the inter- and intra-laboratory discrepancies of isotope measurements increase. High-temperature distillation yields larger deviations than low-temperature distillation techniques. The main problem with these techniques is that in addition to mobile or free pore waters also some of the adsorbed and interlayer water with its distinct isotopic composition (France-Lanord & Sheppard, 1992) is extracted. 2.4.3 Direct equilibration method The direct equilibration method for the analysis of 6180 values of soil waters is described in detail by Hsieh et al. (1998). It is a variant of the classic CO2-water equilibration technique (e.g. Epstein & Mayeda, 1953). A soil sample of 6 to 10 g, which is sterilized by exposure to gamma rays, is placed in a glass vessel on a vacuum line. After freezing the soil sample with ethanol-dry ice slurry, the vessel is opened to vacuum. Several thawing-pumping-freezing cycles are needed to release non-condensable soil gases. A defined volume of tank CO2 is added to the frozen soil sample and is allowed to equilibrate for 12 to 48 hours with the sample at a thermostated water bath temperature of 25~ An aliquot of the CO2 is taken with a gas-tight syringe through a rubber septum and injected into the mass spectrometer. The calculation of the soil water 6180 value using mass balance consideration and appropriate fractionation factors are documented in the Appendix A of Hsieh et al. (1998). The reproducibility of the method is between 0.3 and 0.4%o. McConville et al. (1999) and Koehler et al. (2000) developed automated techniques for measuring both 6180 and 6D values of porewaters by direct CO2 and H2 equilibration. The D / H composition of pore waters from clays-rich samples is determined by equilibration of the sample with H2 during 4 hours using a Pt catalyst at 25~ The 6D results are with 1%o of those derived from waters collected from piezometers (Koehler et al., 2000). 2.4.4 Radial diffusion cell method This relatively new method was developed by Van der Kamp et al. (1996) to determine the isotopic composition, chemistry and effective porosities for groundwaters in aquitards. The method is based on diffusive exchange between interstitial water in a cylindrical samples and water in a reservoir placed along the axis of the sample. The sealed diffusion cell has an outer diameter of 7.2 cm and a length of ca. 12 cm. The drilled out central reservoir (diameter of 2.5 cm, about 25 mL) is stabilized with a porous or perforated plastic liner and filled with isotopically distinct distilled water (test fluid). Equilibrium between test fluid and pore water is reached after about 60 days in the investigated samples. The equilibrated reservoir water is removed with a syringe and measured for its isotope composition. The central reservoir is refilled again with the initial test fluid. The equilibration, extraction and refilling is repeated several (about 4 to 5) times. Mass balance calculations and plots of net mass added versus isotope composition of the repeated experiments allow the determination of the initial isotope composition of the pore water. Experiments at room temperature (25~ and sampling temperatures (4~ showed no discernible effects. The radial diffusion method was tested by Van der Kamp et al. (1996) for D / H of pore water in a clay-rich glacial till. The cell results are within 2 to 3 %o for 6D of accumulated pore
56
Chapter 2 - H.A. Gilg, J.-P. Girard & S.M.F. Sheppard
waters collected in situ from piezometers.
2.5 Isolation and isotope analysis of minerals intimately associated with clays 2.5.1 Silica minerals Quartz, cristobalite, tridymite or opal can be significant constituents of finegrained sediments, such as clays, soils, or shales, and argillically altered hydrothermal systems. The isotopic compositions of these silica minerals can be highly variable and yield information about their genesis (e.g., Sheppard et al., 1971; Blatt, 1987; Mizota 1996), but are also important for isotope mass balance calculations of non-separable mixtures containing silica minerals. The isolation of quartz generally follows the flow sheet presented in Sridhar et al. (1975). After a treatment of the sample with 6 N HC1 at 100~ (removal of carbonates, hydrous iron oxides, some amorphous material and clay minerals) and gravimetric size fractionation (settling, centrifugation), the sample is fused with powdered sodium pyrosulfate. The melt is dissolved with 3 N HC1 (removal of phyllosilicates) and washed. Residual amorphous silica and feldspars are dissolved by shaking in H2SiF6 that has been cleaned from HF by reaction with sized 1-100 ~m commercial ground quartz (Jackson et al., 1976). After dissolution of fluorates using saturated H3BO3 and the purity checks by XRD and SEM, the quartz isolates are weighted for quantitative determination of the quartz content. This isolation procedure does not affect the 5180 of the quartz (Syers et al., 1968; Sridhar et al., 1975). Cristobalite or tridymite is separated from quartz in size fractions > 2/~m before pyrosulfate fusion using a fluid with a density of ca. 2.4 g /cm3. The lighter cristobalitebearing fraction is given an acid-base-fluosilicic (ABFS) treatment (Henderson et al., 1972): following several treatments with 6N HC1 at 80~ and later with 0.5 N NaOH at 100~ residual minerals like feldspars are removed with 30% H2SiF6 at 18~ These treatments can also be used for isolation of amorphous biogenic silica, such as diatom or phytolith opal, or diagenetic opal-cristobalite (Henderson et al., 1972). These waterand OH-bearing forms of amorphous or semi-amorphous silica, however, contain significant amounts of easily exchangeable oxygen. Therefore, they need a special analytical procedure, such as rapid high-temperature dehydration (e.g., Labeyrie, 1974; Wang & Yeh, 1985; Brandriss et al., 1998), controlled isotope exchange (e.g., Labeyrie & Juillet, 1982; Juillet-Leclerc & Labeyrie, 1987) or step-wise fluorination (Matheney & Knauth, 1989).
2.5.2 Hydroxides and oxyhydroxides of iron and aluminum The purification of iron (III) oxides and oxyhydroxides, specially goethite or hematite, from clay-rich rocks for oxygen isotope analysis can be achieved by selective dissolution of silicates using hot 5 M NaOH (Yapp, 1991). Chazot (1997) conducted tests on the effect of the 5M NaOH treatment on 5180 values of one natural and one synthetic goethite. Tests were performed on the goethites alone as well as on mixtures of goethite and kaolinite, and using a solution markedly depleted in 180 (Figure 2.4). The results indicate that the 5M NaOH treatment has no effect on oxygen isotopes when done at 80~ for 4.5 hours. In contrast, goethite 5180 value may be significantly modified, presumably due to isotopic exchange with the NaOH solution, when treatment is done at 100~
Conventional and Less ConventionalTechniquesfor Hydrogenand OxygenIsotope ...
57
(~ 18OSMOW ( ~ o ) -16
~ ~ I ~ l
-14
"7-'~
-12 I
-10
-8
I
~
NaOH solution estimate
untreated goethite
H
-6 I
-4 I-
H
goethite + 5M NaOH at 80~ for 4.5 hours goethite + kaolinite + 5M NaOH at 80~ for 4.5 hours
!
goethite + 5M N aOH at 100~ for 4.5 hours synthetic goethite
natural goethite
Figure 2.4 - Measured 6180 values of untreated and treated goethites and goethite-kaolinite mixtures using the 5M NaOH method modified after Yapp (1991). Data from Chazot (1997). Heavy-liquid and high-field magnetic separation are also useful for the enrichment of these minerals. However, a complete isolation may not be possible in all cases, especially for aluminum oxides and oxyhydroxides (gibbsite, boehmite etc.) and require material balance calculations using chemical analyses and other selective dissolution methods (e.g., Bird et al., 1989, 1992; Yapp, 1991, 1998). Special techniques are necessary for hydrogen isotope analysis of goethite including removal of organics with H202 and degassing in vacuum up to 3 hours at 100~ (Yapp & Pedley, 1985) and controlled isotope exchange of "high temperature" nonstoichiometric hydrogen (Yapp & Poths, 1995). For oxygen isotope analysis, the goethite is prefluorinated with BrF5 at room temperature to remove nonstoichiometric water (Yapp, 1987). Goethite may also contain small amounts of a Fe(CO3)OH component. Its carbon isotopic composition can be determined by stepwise decarbonation(-dehydration) in vacuum at ca. 230~ (Yapp & Poths, 1991; 1993). 2.5.3 Titanium oxides
Anatase and rutile can be isolated from kaolinite-rich rocks by boiling in 6N HC1 and consequently dissolving the remaining impurities using H2TiF6 at 45~ (Sayin & Jackson, 1975; Bird et al., 1992, 1993). Residual quartz or mica can be removed from the titanium oxides using a magnetic separator, as minor Fe is substituted in TiO2.
58
Chapter 2 - H.A. Gilg, J.-P. Girard & S.M.F.
Sheppard
2.5.4 Phosphates Mizota et al. (1992) determined the oxygen isotope composition of both Ca- and Al-forms of soil phosphates by selective extraction. Aluminum-bound phosphate is extracted with 0.5 M NH4F solution at pH 8.2 and room temperature for 24 hours. A subsequent extraction with 0.5 M HC1 dissolves completely the Ca-forms of phosphate, mainly apatite (Williams et al., 1967). The two extracts were purified, the phosphate precipitated as BiPO4.0.5H20 and finally fluorinated following the procedures of Tudge (1960). 2.5.5 Carbonates, organic matter and isotopic analysis of whole rock shales The H- and O-isotope compositions of whole rock shales cannot usually be directly analyzed by conventional whole rock techniques unless the ubiquitously associated carbonates and organic matter are only present as trace constituents (< few tenths of wt. %). During H-isotope analysis using induction furnace heating of the crucible, hydrocarbons can distill in the reaction tube, leading to incomplete hydrogen yields. Although heating the reaction tube with an external electric furnace may eliminate this problem, the maximum working temperature of the furnace may be insufficient to liberate all the hydrogen from the minerals, again leading to incomplete yields. During the extraction of oxygen gas using a fluorinating agent, carbonate carbon can react with oxygen and fluorine to give COF2 and COBr2 compounds. The yields of molecular oxygen gas are thus too low. For these reasons a multi-technique approach is necessary. Carbonates The C- and O-isotope compositions of associated carbonates can be determined using the classic H3PO4 acid attack on the whole rock powder. The resulting CO2 gas, however, may contain contaminants arising from reactions of the acid with organic matter and sulfur minerals. Because of mass interference in the mass spectrometer, the measured 645 value will not necessary be equal to the ~13C value. These contaminants can be removed by reacting the sample gas with silver phosphate (Smith & Croxford, 1975; Charef & Sheppard, 1984). Organics and whole rock analysis The total organic matter of a rock cannot usually be separated in a form suitable for isotopic analysis. The techniques developed to separate kerogen from a rock by destroying the minerals are so drastic (e.g. HC1 and HF 6N attack; Durand & Nicaise, 1980) that some fraction of the organic matter is probably also destroyed. An alternative approach is to determine the H- and C-isotope composition of the total organic matter by difference: analysis of the whole rock sample after removal of carbonates (clays plus organics) plus analysis of the whole rock sample after removal of both carbonates and organics by, for example, plasma ashing. The whole rock minus carbonates sample can be analyzed in the following way. Two reaction tubes in silica glass are interconnected (Figure 2.5). In one of the tubes the "whole rock" sample is mixed with copper oxide (CuO) and in the other only copper oxide is present. In the tube with the sample, a layer of copper oxide is placed above the sample. Before introducing the copper oxide to the tubes it was degassed at 800~ to reduce the carbon
Conventional and Less Conventional Techniques for Hydrogen and Oxygen Isotope ...
59
blanks. The tube with the sample is degassed at 120~ to remove adsorbed water. The extraction is started by heating the tube with copper oxide by itself to 850-900~ to generate a partial pressure of oxygen in the two reaction tubes. Then the sample tube is heated progressively to 1000~ so that all the volatile constituents are oxidized and gases such as CH4, H2S and other hydrocarbons cannot form. The sample tube is kept at 1000~ for about one hour. The gases are then passed through two liquid nitrogen Figure 2.5 - Apparatus for the extraction of total hydrogen and carcooled traps to com- bon from organic-rich whole rocks by oxidation, using copper oxide pletely separate the (CuO). condensable gases (H20, CO2, SO2) from the excess oxygen and other non-condensable gases. Finally the water and carbon dioxide are separated from each other and purified, and then analyzed isotopically in the usual way. The above technique where the sample is always in a significant oxygen atmosphere during the heating stage gives more satisfactory results than the sealed tube method. In the sealed tube method the sample and copper oxide mixture are sealed in a silica tube and then heated in a muffle furnace at 1000~ After cooling down the tube is opened under vacuum with a tube cracker and the water and carbon dioxide are separated and purified in the usual way. Hydrogen yields, usually presented as H20 +, are similar to or often lower than the two reaction tube method. This may be due to back reactions and hydrated mineral formation in the sealed tube during cooling. The H- and C-isotope compositions of the organic matter are calculated from the yield and isotopic data derived from the analyses of the whole rock sample and whole rock minus organic matter sample. Combined with the C- and O- isotope analyses of the carbonates and mass balance considerations, the H-, C- and O-isotope compositions of both the whole rock and principal constituents can be derived. This multi-
60
Chapter 2 - H.A. Gilg, J.-P. Girard & S.M.F. Sheppard
analysis approach may not be essential if the organic matter content of the whole rock is less than about 0.6 wt. % and carbonate is only a trace mineral. 2.6 Conclusions
Over the past three decades, significant progress has been made in the development of physical and chemical extraction and purification techniques of clays for subsequent stable isotope analysis. The flow sheet (Figure 2.1) combines these techniques into a standard procedure. Most of these treatments do not to alter the isotopic composition of the clay minerals in a significant way (see Table 2.1). However, for certain clay minerals and some treatment procedures, rigorous tests are missing. There are also difficulties in separating mixtures of some clay minerals, such as kaolinite and illite, or of different generations of the same mineral in a sample. Specific techniques for hydrogen and oxygen isotope analysis of clay minerals have been established and yield satisfactory results for most minerals with a precision of ca. + 0.2 %o for 6180 and + 2 to 3 %o for 6D. However, the precision of isotope analysis is often lower for clay minerals containing interlayer water, such as halloysites or some smectites, because either they show significant isotope exchange between OH groups and interlayer water during degassing or have important amounts of easy exchangeable oxygen and/or hydrogen. For these minerals, alternative techniques, for example a modification of the controlled isotope exchange technique of Labeyrie & Juillet (1982), should be developed and improved. The isotopic analysis of pore waters in clay-rich rocks is still in its infancy. We suggest that if the pore water content is sufficiently high, physical extraction techniques, such as high-pressure squeezing, are preferable to distillation techniques, as the latter techniques are not able to separate of free pore water and the isotopically distinct adsorbed and interlayer water of clay minerals. Promising alternatives are the newly developed direct equilibration and the radial diffusion method.
61
Conventional and Less Conventional Techniques for Hydrogen and Oxygen Isotope ...
Appendix Effect of Ammonium Oxalate (AM-OX) and Na-dithionite-citrate-bicarbonate (DCB) treatments for removal of iron oxide and hydroxide on the isotopic composition of kaolinite. The kaolinite used for the test is the Macon reference kaolinite described in Girard & Savin (1996). Treatment
untreated untreated DCB DCB DCB AM-OX AM-OX
N u m b e r of treatments
Temperature (~
Duration (h)
0 2 yield (%)
6180 (%0)
H 2 0 yield (%)
0 0 1 1 1 3 10
25 55 70 -
50 1 3
97 97 104 100 99 103 96
21.0 21.2 21.2 20.9 21.1 21.4 21.6
15.2 14.7 14.3 14.7 14.7 12.2 14.7
6D (%o) -54 -54 -56 -57 -57 -60 -60
Effect of Na-polytungstate densimetry (NaPT) on the isotopic composition of kaolinite Treatment
untreated NaPT NaPT
Temperature (~ Room Room
Duration (days)
0 2 yield (%)
6180 (%o)
H 2 0 yield (%)
6D (%o)
5 5
99 98 104
21.7 21.6 21.6
15.0 14.4 12.4
-59 -54 -60
Handbook of Stable Isotope Analytical Techniques, Volume 1 P.A. de Groot (Editor) 9 2004 Elsevier B.V. All rights reserved.
CHAPTER 3 Techniques for Stable Isotope Analysis of Fluid and Gaseous Inclusions Luigi Dallail, Raffaele Lucchini2 & Zachary D. Sharp3 1 CNR-Instituto di Geologia Ambientale e Geoingegneria, Sez. Roma "La Sapienza", P.le Aldo Moro 5,
00185 Rome, Italy 2 Institut de Min6ralogie et P6trographie, BFSH-2, CH-1015 Lausanne, Switzerland 3 Department of Earth and Planetary Sciences Northrop Hall, Albuquerque, NM, 87131-1116, USA e-mail:
[email protected]
3.1 Introduction
Combined stable isotope and fluid inclusions studies of ore deposits and vein systems have greatly increased our understanding of the fluid phase involved in mineral deposition (e.g. Kerrich et al., 1978; Selby & Nesbitt, 1996). The isotope composition of a fluid phase can be calculated on the basis of mineral-fluid equilibrium fractionation at a given temperature, from the 6-values of minerals that formed in equilibrium with the fluid itself. This procedure is not valid when the mineral experienced open system behaviour and original isotopic equilibrium is no longer preserved (e.g. Frezzotti et al., 2000). A more direct approach is to measure the isotopic composition of the fluid remaining as fluid inclusions in a mineral" this fluid generally represents a sample of the fluid trapped within a mineral or a sediment at the time of their formation. The significance of the isotope data of the inclusion fluid is based on the assumption that its isotope composition has been preserved over geological time; that is neglegible isotopic exchange between inclusion fluid and host minerals has occurred. The assumption of preserved fluid composition has been tested in several studies. The stable isotope composition of waters extracted from different minerals within a given lithology are the same (e.g. Vityk et al., 1993; Genty et al., 2002; Naden et al., 2003), and when accurately measured, the isotopic composition of inclusion fluids reflect calculated equilibrium composition of the host mineral phase. For these reasons in the last fifteen years the number of the stable isotope laboratories performing stable isotope analyses of fluid inclusions has noticeably increased and the amount of new data is rapidly growing (see Figures 3.1-3). The aim of this paper is to discuss the various methods of fluid inclusion extraction, their drawbacks and problems, and the effort that have been made to improve their accuracy.
Techniques for Stable Isotope Analysis of Fluid and Gaseous Inclusions
3.2 Methods of extraction Stable isotope analysis of fluid inclusions involves two separate steps: a) quantitative extraction of fluid phase from the host mineral; b) mass spectrometric analyses of the extracted phase. The least ambiguous analyses can be performed on macroscopic fluid inclusions. Piperov & Penchev (1982) described a procedure of withdrawing the inclusion water by puncturing the inclusion itself: they introduced a capillary into the inclusion and drew the water out. Similarly Lazar & Holland (1988) and Genty et al., Figure 3.1a,b,c - Compilation of measured 6D values of fluid inclusions extracted from a) plutonic and b) & c) metamorphic rocks and veins. The reference number on the X-axes corresponds to the reference number in the Appendix. Temperatures refer to homogenization temperatures. The 6D values of the fluid inclusions from magmatic rocks (a) fall in the compositional field of the magmatic waters (- 40 < 6D < - 80). The isotopic composition of fluids extracted from inclusions in contact metamorphic rocks (b) also fall in the magmatic range. The 6D values of inclusions with higher TH (closer to the pluton) are those typically found in primary fluid inclusions of plutonic rocks. The hydrogen isotope composition of metamorphic rocks (b) shows the open-system vs. closed-system behaviour of the investigated rocks. In regional metamorphic rocks, hydrogen isotope data are more scattered in samples with low homogenization temperatures, refleting meteoric or seawater sources. At higher Th values, 6D values converge on those of magmatic origin.
# reference
63
64 3.2a,b,c - Compilation of measured 613C values of fluid inclusions extracted from magmatic (a) and metamorphic (b, c) rocks and veins. The reference numbers on the X-axis correspond to the reference numbers in the Appendix. Fluid inclusions of magmatic rocks (a) have measured ~13C values in the range of the magmatic CO2 assumed on the basis of carbonatite 613C values (- 3 t o - 8%o). The very low values were interpreted in terms of mixing in the source regions of juvenile carbon and CO2 derived from subducted sediments (Mattey et al., 1984). The 613C values of the CO2 produced by metamorphic reactions (b) vary over a broad range. Generally 613C values < - 8%o are related to oxidation a n d / o r partial exchange reactions with organic matter. Fluid inclusions from carbonate-bearing rocks (marbles, calciteschists) have the highest ~13C values (from- 4.9%o to + 5%o). The 613C values of fluid inclusions from schistose and gneissic rocks are scattered, likely reflecting mineralogical heterogeneities and different carbon sources. The ~)13C values of inclusions extracted from meta-igneous rocks are slightly depleted compared to the magmatic range, and may indicate the involvement of external CO2 or partial exchange with organic matter during metamorphism. The isotopic compositions reported for contact metamorphic rocks (c) are influenced both by the mineralogical and the external CO2 imput as mentioned above. Figure
Chapter 3 - L. Dallai, R. Lucchini & Z.D. Sharp
Techniques for Stable Isotope Analysis of Fluid and Gaseous Inclusions
65
Figure 3.3- Correlation of the measured 6180 vs. 6D values of aquaeous fluid inclusions ex-tracted from halites and speleothems. Also shown is the global meteoric water line (GMWL). Data plotting to the right of the GMWL are related to evaporation. The reference data are reported in the Appendix: Horita & Matsuo, 1986 (A); Ohba & Matsuo, 1988 (El) Horita, 1990 (11); Yang et al., 1995 (~); Yang et al., 1996b ( , ) ; Dennis et al., 2001 (A); Naden et al., 2003 (O); Vityk et al., 1993 (O).
(2002) were able to drill a hole into the inclusions walls (halite crystals and spelothem calcite, respectively), and extract the inclusion fluid with a micropipette. In these cases (Madan galenas, Red Sea halites, and Villars c speleothems, respectively), the inclusions were millimetric in size, a situation which is rarely encountered. More common fluid inclusions are at the micron scale, so that an entire population of inclusions must be released to generate enough material for analysis. The methods used for bulk fluid inclusion extractions are thermal (vacuum heating to decrepitation, pyrolysis) a n d / o r mechanical (crushing or grinding in vacuum). Both methods are destructive. As of now, only a few attempts have been made on non-destructive analytical methods, such as Raman spectroscopy (Rosasco et al., 1975; Marshall et al., 1994), but instrumental detection limits and analytical reproducibility lead to results that are not satisfactory for geological investigation. The crushing method is based on the principle of breaking the minerals hosting the inclusions in order to break and open the inclusions themselves (Roedder et al., 1963). Volatiles trapped within the minerals are released in a sealed, evacuated vessel (e.g. stainless-steel tube, steel or pyrex ball-mill) and then cryogenically transferred to a vacuum line for isotope extraction. Thermal decrepitation method involves heating the samples in a vacuum line to sufficently high temperatures for the inclusions to crack open due to high internal pressure (Roedder et al., 1963). Specific advantages and drawbacks of the mechanical
66
Chapter 3 - L. Dallai, R. Lucchini & Z.D. Sharp
and the thermal methods of extraction make these methods complementary. Once the inclusion fluids have been extracted, the analytical methods generally used for the stable isotope analysis of gaseous and aqueous phases are applied to convert H-C-N-O-S compounds to gases suitable for analysis in the mass spectreometer, namely CO2, H2, N2, and SO2. Oxygen isotope analysis of fluid inclusions water have been performed according to different methods: micro C O 2 - water equilibration (Kishima & Sakai, 1980); fluorination (O'Neil & Epstein, 1966b); guanidine hydrochloride (Dugan et al., 1985); continuous-flow water-reduction (Sharp et al., 2001). Hydrogen isotope analyses have been made either by Zn reduction (Coleman et al. 1982), uranium reduction (Bigeleisen et al., 1952; Friedman & Smith, 1958), and most recently, water reduction in continuous flow systems (Prosser & Scrimgeour, 1995; Burgoyne & Hayes, 1998; Sharp et al., 2001). The details concerning the methods mentioned above are reported in specific chapters of these books and will not be treated hereafter. 3.3 Thermal decrepitation The procedure of heating the host mineral and decrepitating inclusions has been successfully employed to extract inclusion fluids from numerous minerals, including quartz, fluorite, barite, sphene, galena, halite and calcite. With regards to the latter two phases further considerations will be added in the drawbacks paragraphs. Thermal decrepitation is a simple and inexpensive method: samples are placed in a quartz tube and evacuated. The tube is then heated using a thermocouple-monitored furnace. The degree of heating necessary to cause decrepitation is determined on the basis of the pressure changes in the vacuum line. Generally, heating continues until a stable pressure is reached. The schematic of a simple and efficent extraction line is shown in Figure 3.4 (after Lucchini, 1997, unpublished). The quartz tube is heated to high temperature prior to sample loading to remove any water that could desorb from the walls. Samples are loaded and thoroughly degassed under vacuum and low temperature heating (100-150 ~ C). Further heating is applied to extract the fluid inclusions. In some cases, decrepitation allows for selective extractions of distinct fluid inclusion populations (Hattori & Sakai, 1979) by heating the sample to different temperatures (stepwise heating, see also Chapter 13). The extraction of fluids belonging to different stages of the rock evolution is possible by step-heating when the temperatures of decrepitation for each generation are distinct. For instance if late stage fluid inclusions trails are observed in a sample, decrepitation temperature may be as low as 100-120~ The gas released by this generation can be either measured or discarded. The furnace temperature can be raised progressively to reach the estimated temperature of decrepitation. However, decrepitation experiments at high temperatures (T > 750 ~ C) have shown that additional release of water or other volatiles may occur once the decrepitation temperature range is significantly overstepped. Piperov & Penchev (1973) carefully monitored the reactions taking place after they decrepitated H20 - CO2 inclusions in an allanite crystal at temperatures between 350 ~ to 600~ Almost no additional gas was released until 900~ where a sharp increase of H20 and CO2 production was observed. They interpreted the newly
Techniques for Stable Isotope Analysis of Fluid and Gaseous Inclusions
Figure 3.4 - Schematic of the extraction lines for fluid inclusions. Upper figure illustrates the extraction line for analysis of hydrogen isotope ratios from water or methane, and carbon isotopes from CO2 or methane. Lower figure is for determining the 6180 values of water. H20 is equilibrated with CO2 derived from the reservoir. The configuration of a single line can be changed to accommodate either type of analysis.
67
68
Chapter 3 - L. Dallai, R. Lucchini & Z.D. Sharp
released gas in terms of decrepitation of sub-micron inclusions and release of hydrogen gas dissolved in the crystal structure. If the sample was then reheated, a considerable amount of hydrogen was released at temperature above 500~ with a sharp hydrogen release at 800 to 900~ Variable amounts of the other gas species, such as CH4 and CO were also present; thus gas species other than the expected C O 2 and H20 were derived from chemical reaction during heating. The amount of CO was greater at higher temperature of decrepitation. Water can be trapped both as fluid inclusions and structurally bond water (bubble water) during mineral growth, the isotopic composition of water released from high temperature decrepitation resulting in mixed 6D values. Simon, (2001) grinded and thermally decrepitated at temperature of 1200~ aliquots of hydrothermal quartz with different grain size, in order to evaluate the contribution of bubble water to the 6D values of the released water. Water yields obtained by thermal decrepitation decrease as a function of grain size because water from large fluid inclusions can be lost while grinding the sample. Because the structurally bonded water is D-depleted, the final hydrogen isotope composition is influenced by the volumetric ratio between fluid inclusion water and bubble water. The gD values of water extracted by thermal decrepiatation decrease with decreasing grain size fraction because the bubble water fraction becomes dominant in fine grain material. Most accurate 6D estimates of the fluid inclusion water are obtained when fluid inclusion/bubble water ratio is high (minerals formed at medium to low temperature), and when coarse-grained material is decrepitated. Release of CO2-rich fluid inclusion from quartz at temperatures of 1100~ has been successfully performed (see Nesbitt & Muehlenbachs, 1995); at such high temperature CuO was used to oxidize possible carbon species; this technique is described in the oxidative pyrolysis paragraph (see below). Decrepitation accomplished by melting of the host mineral fails to exctract the inclusion fluids: a quartz xenolith bearing high density pure CO2 inclusions within a basalt was heated and melted at temperatures in excess of 1200~ (Dallai et al, unpublished). No inclusion decrepitation and no CO2 was released was observed until T > 1460~ was reached. At this temperature the quartz (and the quartz-glass tube) melted, and several newly formed bubbles were observed into the molten sample (Figure 3.5). These bubbles were interpreted in terms of decrepitated gas unable to escape from the viscous mineral but no CO2 was detected by Raman spectroscopy of these newly formed bubbles, because the CO2 density was too low. Only an H20 vapour phase was produced during heating above 600~ Because inclusions contained only high-density CO2, H20 desorption from mineral and/or sample holder, or dissolved water in basalt were inferred as possible sources. Very high temperature vacuum volatilization has been proven reliable to extract fluid inclusions from different mineral phases such as halite (Knauth & Beeunas, 1986), plagioclase (+ quartz + apatite), and clinopyroxene (+ olivine; Kelley & FrtihGreen, 2001). Care shall be taken to quarry all the impurities (i.e. organic matter,
Techniques for Stable Isotope Analysis of Fluid and Gaseous Inclusions
69
Figure 3.5 - Photograph of the melted quartz sample (now glass) with newly formed vescicles (scale 400:1).
hydrous phases). This latter procedure is often difficult and time consuming. Vacuum crushing method (see below) is generally advised for fluid inclusion extraction from halites. Recently, Sharp et al., (2001) have obtained accurate and reproducible data by vaporizing extremely small halite samples in a helium stream, and processing the evolved water in a CuO furnace to covert any H2 to H20. The water is then cryofocused and released into a high T reduction furnace that allowed for quantitative conversion of H20 to H2 and CO gas. Both 6D and 6180 values could be determined from the same sample by this technique.
Significant drawbacks of decrepitation are: 1) the method is generally not applicable to fluid inclusions hosted by hydrous minerals (Kazahaya & Matsuo, 1985); 2) the possibility exists for chemical reactions to occur between released gases during heating repetitions; 3) low density inclusions that homogenize to a vapor phase will not decrepitate because of the flat isochores; 4) problems are encountered when extracting calcite and/or dolomite-hosted inclusions (calcite starts decomposing at around 450~ CO2 gas may be produced from carbonate breakdown and add to CO2 from fluid inclusions. L6cuyer & O'Neil, (1994) observed small amounts of CO2 being produced during decrepitation of H20-bearing fluid inclusions in carbonate shells. Above 500~ the amount of CO2 evolved was extensive and the amount of CO2 was correlated with the temperature of decrepitation. However, no correlation between the 6180 values of the extracted H20 and the amount of CO2 (or the temperature) was found. In general it is advised not to heat carbonate-hosted inclusions above 150~ (Dennis et al, 2001). With regards to D / H determination in calcite-hosted inclusion water (and calcitebound water), it has been noted that high-T decrepitation leads to ~SD values that are significantly lower than the actual inclusion values (Yonge, 1982). Matthews et al., (2000) reported an isototopic fractionation of- 30.0 _+1.4 %0 between the 6D values of
70
Chapter 3 - L. Dallai, R. Lucchini & Z.D. Sharp
inclusions in cave deposits and the values of the waters from which they grew. A slightly lower value (- 22.1 + 3.9 %0), was reported by Yonge, (1982); this difference may be related to the different temperature of heating (Yonge used 700~ whereas Matthews et al. used 900~ and/or to the different behaviour of the analyzed cave deposits (Matthews et al., 2000).
Reproducibility for
hydrogen isotope measurements of fluid inclusions Hattori & Sakai, (1979) and Jenkin et al., (1992) reported a precision better than 2%0 (2 o). For carbon isotope composition of fluid inclusions Kreulen (1980) reported a precision between 0.5 and 1.0 %0 (2 ~J). The reproducibility of some of the analyses perfomed on the extraction line of Figure 3.4 is reported in Table 3.1.
3.4 Oxidative Pyrolysis The stepped heating procedure at very high temperatures can be applied in order to disrupt the crystal structure of the mineral and extract the gas dissolved in the crystal. The oxidative pyrolysis method is mainly applied to the recovery of the reduced carbon species in minerals formed at high-temperature (e.g. olivine, pyroxene), typically in volcanic glasses. The h-values of the gas species extracted with this method represent the isotopic compositions of gas mixtures, namely gas released from the fluid inclusions and gas phase dissolved into the crystal. Therefore measured carbon isotope compositions are correlated to the abundance of these two components, and possible effects of isotopic fractionation need to be evaluated (Nadeau et al., 1990). The method involves loading the sample in a pre-baked quartz glass combustion vessel, which is then evacuated. The sample is heated to 400 ~ - 500~ in order to remove any surfacial organic contamination (Mattey et al., 1984). The temperature is raised further (800 ~ to 1200~ in the presence of a low partial pressure of 02, (10 - 20 mbars) evolved from heating CuO. Oxygen converts the solid and reduced carbon to CO2. Separation of CO2 from other gas species (SO2, H 2 0 ) is achieved by standard cryogenic methods (Nadeau et al., 1990; Kelley & Friih-Green, 2001). Mattey et al., (1984; 1989) reported that all the carbon is released from basalt glasses after 1 hour pyrolysis at 1200~ repeated extractions at 1300~ of pyrolyzed samples showed almost no further carbon release. Watanabe et al., (1984) reacted the samples for 24 hours in the presence of 0 2 to extract CO2-rich inclusions from olivine crystals (oxygen was used in order to convert any graphitic carbon left in the crystal after inclusion decrepitation). The partial pressure of oxygen in the reaction furnace is used to convert the reduced forms of released carbon to CO2 for isotopic analyses (Watanabe et al., 1983; Mattey et al., 1984:; 1989; Nadeau et al., 1990). The reported reproducibility of the oxidation pyrolysis method is 1.2%o for the ~13C of fluid inclusions in olivine crystals analyses (Watanabe et al., 1984). Mattey et al. (1984:; 1989) performed their isotopic analyses on a static mass spectrometer in order to measure very small amount of gas (Carr et a1.1986). The overall precision they obtained on the measured 613C of carbon dissolved in basalt glasses was + 5.0%0.
Sample Wt (g)
Mineral Tdecr.('C ) Gas phase extracted from fluid inclusions c02 H20 N.C. mbar
A5.1 A5.1 A5.1 83.2 B3.2 B3.3 B3.3 MR19 MR19 MR17 MR17 MR25 MR25 r141 1-141 r104 r104
%, vol mol. (MS) C02 H20 NC
?, vol mol. (other methods) c02
H20
NC
21.3
77.4
0.2
51.O
40
9
49
48
4
55
41
4
~ ~ ~ C Volt P DMAT251 B ~ DSM O Wmbar
2.6 2.5 2.7 4.1 2.7 3.6 4.1 1.1 1.3 1.0 1.1 0.8 1.0
Qtz Qtz Qtz Qtz Qtz Qtz Qtz Qtz Qtz Qtz Qtz Qtz Qtz
540 550 600 600 600 600 600 520 520 520 520 520 520
20.2 19.2 19.9 220 120 98.4 120 8.9 9.5 8.2 11.0 16.0 17.0
-3.0 -3.2 -3.3 5.4 5.2 4.3 4.6 2.4 2.5 -3.1 -3.0 0.0 0.1
6.0 4.8 8.0 6.4
Qtz Qtz Qtz Qtz
550 550 550 550
0.2 3.3 7.1 4.8
-5.4 -5.6 -3.4 -3.7
0.972 1009
-35.2 -36.6
80 50
20.7 19.5
76.6 78.8
2.7 1.7
6741 3313 5312 7056 0.099 0.102 0.110 0.176 0.105 0.125 Volt DeltaE 2620 1548 840 593
-73.6 -75.8 -19.8 -22.8 2.7 -2.6 -29.3 -32.4 -96.8 -99.7
380 400 450 440 100 150 50 70 110 100
17.9 16.8 27.8 29.7 47.2 46.2 45.0 40.9 60.1 58.4
80.8 82.2 71.1 68.4 43.7 41.3 50.3 54.6 32.8 35.7
1.2 1.0 1.0 1.9 9.1 12.5 4.7 4.5 7.1 5.9
31.5 28.8 37.0 33.4
0.01 0.01 0.27 0.13
4.3 4.9 14.0 15.8
95.6 94.9 80.9 79.9
0.1 0.1 5.0 4.2
Techniques for Stable Isotope Analysis of Fluid and Gaseous Inclusions
Table 3.1 - Example of data reproducibility using the extraction line as in Figure 34.2 for fluid inclusions extracted from quartz veins. VMAT251 and VDeltaE are referred to the hydrogen gas voltage (mV) in the Faraday cup 2 with bellow loo?, open.
71
72
Chapter 3 - L. Dallai, R. Lucchini & Z.D. Sharp
3.5 Crushing or grinding in vacuum
Several detailed papers exist which describe the method of crushing minerals in vacuum and inclusion fluids release for stable isotope analyses (i.e. Andrawes & Gibson, 1979; Kita, 1981; Kazahaya & Matsuo, 1985; Horita & Matsuo, 1986; Norman & Sawkins, 1987; Ohba & Matsuo, 1988; Dennis et al., 1998). Mineral crushing was initially performed in aluminum tubing. The samples were loaded in the tube; the tube was then evacuated and squeezed until the content had been finely ground and the volatiles released (Roedder et al, 1963). By using this method low gas yields were generally obtained, and the tube could not be re-used. Ball-mill apparata and/or stainless steel crushing cells have been successively adopted. The ball-mill devices made by Kita (1981) and Ohba & Matsuo, (1988), are shown in Figure 3.6a,b, respectively. They consist of pyrex cylinders filled with aluminum, pyrex or stainless steel ball which moves up and down by hand shaking, magnetic breaker or any motorized device. More recently a crushing cell has been developed (Dennis et al., 2001)" such a device consists of a stainless steel flanged tube with a hardened base, containing a piston that is electromagnetically raised and lowered (Figure 3.6c). Halite and calcite minerals have been successfully analyzed using the ball mill and the crushing cell, whereas the crushing efficency on fluorite, quartz or harder mineral phases is generally unsatisfactory. Using the apparatus of Kita, (1981) the uncrushed quartz fractions after 1 hour milling was 60%. The fraction left ucrushed by Ohba & Matsuo's ball-mill, (1988) was 44%. Further attempts are needed to constrain the efficency of the crushing cell. The main drawbacks of crushing are selective crushing, gas generation during crushing, and adsorption. Crushing very small inclusions is problematic and selective release of volatiles from bigger inclusions may occur (L4cuyer & O'Neil, 1994). A correlation between the size of inclusions and the 8180 of the extracted waters has been observed, the smaller inclusions having lower oxygen isotope values due to exchange with the host mineral and/or diffusional 180 loss from inclusion to the host quartz (Ohba et al., 1995). Therefore measured isotopic compositions of inclusions from incompletely crushed samples may result in slightly enriched 8180 values. The major drawback of crushing is adsorption, which takes places mainly when H20-vapour is released from fluid inclusions and interacts with the newly formed surfaces of the powdered mineral. These surfaces can be chemically active; oxygen and hydrogen isotope fractionation occurs due to the H20 vapor adsorption, particularly on quartz and fluorite (Barker & Torkelson, 1974). It is worthwhile to note that no adsorption has been observed in absence of crushed sample, thus "adsorption occurs only when new surfaces are being created" by fracturing the mineral (Ware & Pirooz, 1967). The results obtained by grinding the sample in a "pepper-mill"-like device showed that "the smaller the final grain size, the more enriched in deuterium the extracted water", and that "the larger the final grain size, the smaller the water yield" (Simon, 2001). Crushing quartz grains above 0.6 mm in size resulted in constant water yields, and in 8D values similar to the ones obtained by thermal decrepitation. The positive correlation between the 8D values of water extracted and the amount of surfaces formed by grinding was interpreted in terms of adsorbed water being D-depleted (Simon, 2001). Similar conclusions were inferred from 8180 values of desorbed water (post-crushing heating) lower than 8180 values of water collected upon crushing (Matsuo, 1991). Interestingly, the latter experiment yielded contrasting results for hydrogen isotope
Techniques for Stable Isotope Analysis of Fluid and Gaseous Inclusions
Figure 3.6a, b, c - Sectional views of the Kita's (1981), Ohba & Matsuo's (1988) and Dennis et al., (2001) crushing devices.
73
74
Chapter 3 - L. Dallai, R. Lucchini & Z.D. Sharp
fractionation (6D values of desorbed water higher than 6D of collected water; Matsuo, 1991). In order to reduce adsorption, the crushing device should be on-line with a vacuum extraction line (adding a cold finger at liquid nitrogen temperature to the crushing apparatus to collect the released gas results in significantly higher gas yields (Ohba & Matsuo, 1988)), and heated to moderate temperatures (150 ~ - 200~ during the whole procedure (Dennis et al., 2001; Simon, 2001). The adsorbed gas can be released by heating the powdered crystal after completing the crushing procedure; as adsorption is temperature-dependent and very high temperatures (up to 600~ are necessary to completely desorb water from silica surfaces (Hockey & Pethica, 1961; Moore & Rose, 1973). However, there is no consensus about the possibility to obtain a quantitative recovery of the adsorbed water on crushed material was from heating the powdered material; Kazahaya & Matsuo, (1985), Horita & Matsuo, (1986), and Matsuo, (1991) reported successful experiments from heating halite samples to 180 ~ to 220~ Conversely, water desorption from crushed quartz was considered unsatisfactory for accurate isotope analyses by Ohba & Matsuo, (1988). These authors addressed the problem of measuring the original oxygen isotope composition of inclusion waters, due to the incorporation of oxygen atoms into the newly formed quartz surfaces during adsorption. Mathematical corrections to the measured 6D and ~180 values were adopted, assuming isotopic fractionation of the adsorbed water could be modelled as a Rayleigh fractionation process. Even so, results were less than ideal. Continous flow techniques can be employed to transport the released gas in an inert gas stream to the cryogenic trap (e.g. Li & Shi, 1983), or directly into a gas chromatograph (Andrawes & Gibson, 1979) or a mass spectrometer (e.g. Sharp et al., 2001). Andrawes & Gibson (1979) reported that quantitative water extraction was possible by using a helium flow-assisted crushing device. However chromatographic analysis of other gases simultaneously with water was unsuccessful and water needed to be measured separately. Still, different gases are adsorbed to different extent. Experiments by Barker & Torkelson, (1974) and Matsuo, (1991) indicate that CO2, N2 and CH4 adsorption on powdered grains is negligible. Crushing has been proven accurate and reproducible: reported precision for ~)13C analysis of CO2 from fluid inclusions hosted in quartz is + 0.5%o or better (Kreulen, 1980; Matsuo, 1991). Measured 6180 and 8D values of water rich inclusions hosted in halites and carbonates have reproducibilities within + 1.6%o and + 1.7%o, respectively (Matsuo, 1991) and + 0.3, + 2.0%o (Yang et al., 1996b). Horita & Matsuo, (1986) reported 6180 and 6D values of water hosted in halites differing by - 0.2 %o and - 3.0 %o from the values of the host brine. Duplicate 6D analysis of water inclusions in speleothems agree to + 3.0%o (Harmon et al., 1979; Dennis et al., 2001).
34.6 Cryogenic separation and molar volumes calculation C O 2 , H20, CH4 and N2 gas species released during decrepitation and/or crushing need to be separated prior to analysis. Different methods are used for the various gas species. Most involve a simple vacuum extraction line illustrated in Figure 3.4: water and CO2 are released and frozen into the trap closest to the original sample (Trap A)
75
Techniques for Stable Isotope Analysis of Fluid and Gaseous Inclusions
using liquid nitrogen. They can be cryogenically separated using liquid nitrogen and ethanol-dry ice slush, respectively. Methane does not freeze at liquid nitrogen temperature and is free to expand throughout the line. CH4 is oxidized at T - 800~ in the cuprous oxide furnace. Cu20 is used instead of CuO because it has much lower 02 vapor pressure at 800~ During oxidation, methane is converted to CO2 and H20 and frozen into the trap adjacent to the furnace (Trap B). A small piece of platinum in the furnace catalyzes the reaction, increasing the oxidation rates. Again CO2 can be separated from the water cryogenically and frozen into a sample container for 813C and 8180 analyses (the 8180 value of the oxidized methane has no significance). The water
Samples containing hydrocarbons CO2, H20, N2, hydrocarbons Thermal Decripitation ~ Production in some cases of CO, 02, H2S, SO2 / Crushing / Cracking of Hydrocarbons
I Condensible
Separation with liquid Nitrogen
[ Non-Condensible ] N2, CH4, CO ] Oxidation with CuO at 800~ CH4+202~CO2+2H20 2CO+O2-+2CO2
CO2 H20
Separation with liquid Nitrogen Condensible Non-Condensible
Separation with dry ice + ethanol
CO2
I
I
H20
I
I
~ Introduction of CO2
I
I
N2
I
I
Adsorb on zeolite trap at LN2 temp.
Equilibration at 25~ 24hrs CO2, H20
I CO2, H20 I
Separation with dry ice + ethanol
Non-Condensible
I
I C~
Separation with dry ice + ethanol
CO2 ~r
I
I
I
8180(H20)
I
H20
H20 Reaction of H20+Zn 500~ 30min
I Reaction of H20+Zn 500~ 30min
813C(CO2) 8180(CO2)
Condensible
8H(H20)
815N
813C(CH4 - CO)
Figure 3.7- Schematic diagram for procedure of gas treatment and separation.
8H(CH4)
76
Chapter 3 - L. Dallai, R. Lucchini & Z.D. Sharp
F i g u r e 3.8a, b - Schematic of the laser-based fluid inclusion extraction line. The fluid inclusion is opened during laser 'drilling'. Helium gas passes through the heated sample chamber, incoroporating water released during exposure of the fluid inclusion. Water and CO2 are collected on the liquid nitrogen trap in a 6-way valve assembly. After 3 minutes, all water is transferred. The 6-way valve is switched and the trap is heated, releasing the water as a coherent 'pulse'. Water passes through a 1450~ microfurnace filled with glassy carbon causing complete reduction of water to H2 and CO, through a gas chromatographic column (5A molecular sieve) to separate the reaction products, and finally into the mass spectrometer for isotopic analysis. (b) Schematic of reaction chamber. Helium enters and exits the chamber through two holes in the bottom. Laser radiation is admitted through a UV-grade silica window. The seal is maintained with a Viton O-ring, affixed firmly on the top of the sample chamber.
can be frozen into a tube with zinc for reduction to ZnO and H2 for hydrogen analysis (Coleman et al., 1982). Alternatively, very small amounts of water can be equilibrated with CO2 for oxygen isotope determinations (Kishima & Sakai, 1980; Yu, 1991), equilibrated with H2 gas for hydrogen isotope determinations (Horita, 1988; Horita et al., 1989) or exchanged with guanidine for oxygen isotope determinations (Yang et al., 1996b). The general techniques described above allow for 6D and ~13C values from methane, 613C and 6180 values from CO2 and 6D and 6180 values from water to be deter-
Techniques for Stable Isotope Analysis of Fluid and Gaseous Inclusions
77
mined. Remaining non condensable N2 gas can be adsorbed onto 5/k molecular sieve cooled to liquid N2 temperatures for N2 analysis. A schematic step-diagram for gas separation is reported in Figure 3.7. Yields can be determined manometrically.
34.7 Further Developments The choice of the method of extraction depends on the mineral hosting the inclusions and the fluid characteristics. Crushing and decrepitation are complementary methods (e.g. Hattori & Sakai, 1979; Alderton & Harmon, 1991). However these two conventional methods bear obvious limitations, such as the generally inability to separate different generations of inclusions. Analysis of individual inclusions has only recently been made successfully using a laser GC system (see below). Several attempts have been made using Raman microprobe, based on the relationship between the different isotopic peaks for CO2 and the isotopic ratios of the carbon species (Rosasco et al., 1975; Dhamelincourt et al., 1979; Marshall et al., 1994). It is our experience that the measured isotopic ratios for a single fluid inclusion using Raman spectroscopy are quite precise, but inaccurate due to varying sample geometry, depth to inclusion and thickness. Thus, the results of this method are still unsatisfactory. Recently, Sharp et al., (2001) developed a method for microliter water analysis based on carbon reduction, merging the advantages that gas chromatography and continuous flow technology provide for mass spectrometric analysis techniques. Water samples are entrained in a helium stream and carried through a furnace filled with glassy carbon and heated to 1450~ The high temperature heating causes reduction of the water to H2 and CO, which are then separated in a gas chromatographic column and directly measured for D / H and 180/160 ratios in a mass spectrometer. Heating fluid inclusion-bearing samples in the He-stream is easily incorporated into the system. The system has recently been modified to incorporate a UV laser to open individual large (~ 100 mm diameter) inclusions (Figure 34.8a,b). In this way different generations of inclusions can be analyzed.
Aknowledgements
The authors are indebt to Paul Dennis, who provided insights to the crushing procedures. Reviews by J. Horita, A. Matthews, J. O'Neil, W. Yang, and an anonymous referee improved the original version of the manuscript. Financial support by the Swiss National Science Foundation (21-049302.96) is acknowleged.
Appendix C o m p i l a t i o n of p u b l i s h e d d a t a o n fluid inclusions 6D, 613C a n d 6180 analyses of minerals f r o m rocks in different geological settings. # (1) Location
Method(2) Sample
Contact metamorphism
Mineral
H o s t rock
Met. g r a d e
T~
6DSMOW%o max min
150 150 200-440 250 250 250 250 350 350
-105 -83 -43 -78 -69 -51 -47 -61 -59
-41
Ap Dol Qtz Cal Qtz Sulfate Py Qtz Qtz
Limestones Calc-silicate Flysch Limestones Limestones Limestones Hornfels Limestones Calc-silicate
Qtz-Cal(def.)
Bulk
Metasediment
Green.
90-150
-158
-61
D
Qtz-Cal(undef.)
Bulk
Metasediment
Green.
90-150
-90
-30
Lake Louise, Canada
D
Qtz-Ank-CC
Bulk
Metasediment
Prehnite
90-185
-150
-120
4
Lake Louise, Canada
D
Qtz-Ank-Py
Bulk
Metasediment
Prehnite
90-185
-91
-72
5
Rocky Mountains, $52 ~ Canada
C
Qtz
Limestones
Amph.
100-186
-150
-100
6 7 8
South Cornawall, UK Connemara, Ireland Purcell Mountains, canada
D D D
Low grade
110-150 118-250 149-333
-80 -32 -170
-49 -27 -61
C C C C C C C C C
1 2 3 4 5 6 7 8 9
Mines Gasp6, Quebec, Canada Weolag Tu-Mo deposit, Korea South Cornawall, UK Mines GaspG Quebec, Canada Mines Gasp6, Quebec, Canada Mines Gasp6, Quebec, Canada Weolag Tu-Mo deposit, Korea Mines Gasp6, Quebec, Canada Weolag Tu-Mo deposit, Korea
1
Jasper, Canada
D
2
Jasper, Canada
3
Skarn Qtz-vein (Tu-Mo) Qtz-vein(W-Sn) Skarn Skarn Skarn Qtz-vein (Tu-Mo) Skarn Qtz-vein (Tu-Mo)
Regional metamorphism
Dae Hwa W-Mo Mine, South Korea 10 Dae Hwa W-Mo Mine, South Korea 11 South Cornawall, UK
9
Qtz-vein
Flysch Qtz Qtz-vein(Pb-Zn) Qtz, Cal, Brt Metagabbro Qtz-vein Metasediment Qtz Qtz-vein
Amph.
-23
-43
Reference
Shelton, 1983 So et al., 1983 Wilkinson et al., 1995 Shelton, 1983 Shelton, 1983 Shelton, 1983 So et al., 1983 Shelton, 1983 So et al., 1983 Nesbitt & Muehlenbachs, 1997 Nesbitt & Muehlenbachs, 1997 Nesbitt & Muehlenbachs, 1997 Nesbitt & Muehlenbachs, 1997 Nesbitt & Muehlenbachs, 1995 Wilkinson et al., 1995 O'Reilly et al., 1997 Nesbitt & Muehlenbachs, 1997 Shelton et al., 1987
n.s. Qtz-vein(W-Mo)
Cal
Orthogneiss
150-225
-101
n.s. Qtz-vein(W-Mo)
Dol
Orthogneiss
150-225
-75
-71
Shelton et al., 1987
Qtz-vein(Sn-Cu)
Qtz
Flysch
150-250
-25
-25
Wilkinson & al 1995
D
Low grade
Appendix continued >
(3
!
c~ c~ ~,,,i ~ ~,,,i o
N
> Appendix continued
C~
12 Rocky Mountains Trench, N52 ~ Canada 13 Rocky Mountains, N52 ~ Canada 14 Proterozoic West R. M. T., $52 ~ Canada 15 Proterozoic West R. M. T., N52 ~ Canada 16 Mount Robson, Canada
C
Qtz-vein
Qtz
Metasediment
Green-Amph
150-250
-150
-120
C
Qtz-vein
Qtz
Sandstone
Green-Amph
150-300
-130
-40
C
Qtz-vein
Qtz
Metasediment
Amph.
150-300
-60
-150
C
Qtz-vein
Qtz
Metasediment
Green-Amph
150-320
-150
-90
D
Qtz-vein
Qtz
Metasediment
Prehnite
160-260
-155
-62
17 Cariboo Mountains, Canada
D
Qtz-vein
Qtz
Metasediment
Green.
165-330
-142
-109
18 Malton Gneiss, N52 ~ Canada
C
Qtz-vein
Qtz
Gneiss
Green-Amph
170-210
-80
-40
19 Slide Mountain Terrane, N52 ~ Canada 20 West Onimeca Ext. Compl, $52 ~ Canada 21 Kootenay Arc, $52 ~ Canada
C
Qtz-vein
Qtz
Metabasalts
Green-Amph 200-280
-150
-120
C
Qtz-vein
Qtz
Amph.
200-280
-90
-150
C
Qtz-vein
Qtz
Amph.
200-300
-100
-130 -87
Massive sulfide Massive sulfide Qtz-vein Qtz-vein Qtz-vein
Wooklawn, Canada Wooklawn, Canada Soutern Alps, New Zeland Soutern Alps, New Zeland Rocky Mountains Trench, $52 ~ Canada 27 Barkerville Terrane, N52 ~ Canada 28 Central Alps, Switzerland 29 Tete jaune cache, Canada
C C D D C
30 Dae H w a W-Mo Mine, South Korea 31 Onimeca Ext. Complex, $52 ~ Canada
n.s. Qtz-vein(W-Mo)
22 23 24 25 26
Cal Qtz Cal Qtz Qtz
Metavolcanic Metavolcanic Qtz-Feld schists Qtz-Feld schists Metasediment
Green. Green. Amph. Amph. Amph.
200-320 200-320 200-320 200-320 205-295
-127 -89 -68 -59 -140
-53 -44 -150
C
Qtz-vein
Qtz
Metasediment
Green-Amph 220-280
-150
-120
D D
Qtz-vein Qtz-vein
Qtz Qtz
Syenite Metasediment
Green. Green.
240-280 250-270
-76 -74
-33
Chal
Granitic Gneiss
250-300
-75
250-320
-100
C
Qtz-vein
Qtz
Amph.
-140
Nesbitt & Muehlenbachs, 1995 Nesbitt & Muehlenbachs, 1995 Nesbitt & Muehlenbachs, 1995 Nesbitt & Muehlenbachs, 1995 Nesbitt & Muehlenbachs, 1997 Nesbitt & Muehlenbachs, 1997 Nesbitt & Muehlenbachs, 1995 Nesbitt & Muehlenbachs, 1995 Nesbitt & Muehlenbachs, 1995 Nesbitt & Muehlenbachs, 1995 Lusk & Krouse, 1997 Lusk & Krouse, 1997 Jenkin et al., 1994a Jenkin et al., 1994a Nesbitt & Muehlenbachs, 1995 Nesbitt & Muehlenbachs, 1995 Mullis et al., 1994 Nesbitt & Muehlenbachs, 1997 Shelton et al., 1987
9 9
9
r o C~
~,,,i ~
9
Nesbitt & Muehlenbachs, 1995 Appendix continued > ~D
> Appendix continued 32 Onimeca Ext. Complex, $52 ~ Canada 33 Canadian Cordillera, Canada 34 Canadian Cordillera, Canada 35 Canadian Cordillera, Canada 36 Dae Hwa W-Mo Mine, South Korea 37 Okanagan Valley, British Columbia 38 Brunswick, Canada 39 Brunswick, Canada 40 South Cornawall, UK 41 Naxos, Greece 42 Central Alps, Switzerland 43 Western Carpathians 44 Western Carpathians 45 Western Carpathians 46 Central Alps, Switzerland 47 Central Alps, Switzerland 48 Wilson Terrane, NVL, Antarctica 49 Dae Hwa W-Mo Mine, South Korea 50 Dae Hwa W-Mo Mine, South Korea 51 Central Alps, Switzerland 52 Naxos, Greece 53 Central Alps, Switzerland 54 Central Alps, Switzerland 55 Central Alps, Switzerland 56 Central Alps, Switzerland 57 Soutern Alps, New Zeland 58 Western Carpathians 59 Naxos, Greece 60 Naxos, Greece 61 Naxos, Greece
O C D D D n.s.
Qtz-vein
Gneisses
Amph.
Qtz Qtz Qtz Sche
Oceanic Terranes "Suspect Terranes Island Arc Terranes Granitic Gneiss
Qtz
Graphitic schists
C Massive sulfide C Massive sulfide D Qtz-seg C Qtz-seg D Qtz-seg D Qtz-seg D Shear zone D Qtz-seg D Qtz-seg D Qtz-seg D Qtz-vein n.s. Qtz-vein(W-Mo)
Cal Qtz Qtz Qtz Qtz Qtz Qtz Qtz Qtz Qtz Qtz Wo
Metavolcanic Metavolcanic Flysch Schists(chl) Marble Granitoides Granitoides Metasediment Phillite Schists Pelite Granitic Gneiss
n.s. Qtz-vein(W-Mo)
Qtz
Granitic Gneiss
D C D D D D D D C C C
Qtz Qtz Qtz Qtz Qtz Qtz Qtz Qtz Qtz Qtz Qtz
Orthogneiss Marble Syenite Gneiss Phillite Schists Qz-feld schists Granitoides Bio schist Marble Marble
D
Qtz-carb-vein(Au) Qtz-carb-vein(Au) Qtz-carb-vein(Au) Qtz-vein(W-Mo)
Qtz
Qtz-vein(Au)
Qtz-vein Qtz-seg Qtz-vein Qtz-vein Qtz-vein Qtz-vein Qtz-vein Qtz-vein Qtz-seg Qtz-seg Qtz-seg
268-286 Green. Green. Low grade Amph. Green. Green. Green. Green. Amph. Amph. Green.
Green. Amph. Green. Amph. Amph. Amph. Amph. Green. Amph. Amph. Amph.
250-330
-120
-150
250-350 "250-350 250-350 260-300
-158 -141 -123 -74
-125 -130 -93 -60
-148
-121
270-370 270-370 290-330 350 300-400 300-400 300-400 300-400 320-340 320-340 350-420 350
-169 -169 -18 -70 -59 -57 -39 -35 -90 -42 -84 -65
-88 -125 -10
375-390
-78
-73
Shelton et al., 1987
380-430 385 400-420 400-450 400-450 400-450 400-500 450-500 530 565 620
-65 -88 -61 -68 -54 -51 -42 -64 -59 -66 -69
-39
Mullis et al., 1994 Rye et al., 1976 Mullis et al., 1994 Mullis et al., 1994 Mullis et al., 1994 Mullis et al., 1994 Jenkin et al., 1994a Hurai et al., 1997 Rye et al., 1976 Rye et al., 1976 Rye et al., 1976
Nesbitt & Muehlenbachs, 1995 Nesbitt et al., 1989 Nesbitt et al., 1989 Nesbitt et al., 1989 Shelton et al., 1987 Zhang et al., 1989
-34 -27 -76
-98
-47 -41 -38
Lusk & Krouse, 1997 Lusk & Krouse, 1997 Wilkinson et al., 1995 Rye et al., 1976 Mullis et al., 1994 Hurai et al., 1997 Hurai et al., 1997 Hurai et al., 1997 Mullis et al., 1994 Mullis et al., 1994 Frezzotti et al., 2000 Shelton et al., 1987 r
!
,o
c~ ~..L. ~..L.
N Appendix continued >
> Appendix continued
=r
62 Naxos, Greece C 63 Bushveld complex, South Africa C 64 North Thompson fault, Canada D
Qtz-seg Quarzite Qtz-vein
Qtz Qtz Qtz
Gneiss Metasediment Metasediment
Amph. Green. Amph.
680
-78 -34 -158
-146
Rye et al., 1976 Schiffries & Rye, 1990 Nesbitt & Muehlenbachs, 1997
r
Retrograde alteration South Cornawall, UK Connemara, Ireland Connemara, Ireland Connemara, Ireland
r~
D D D D
Qtz-vein Rock Rock Rock
Qtz Qtz Qtz Cal
Flysch Metagabbro Granite Metagabbro
Low grade Green. Green. Green.
265-315 230-400 230-310 275
-28 -27 -26 -22
-13 -29 -18
Wilkinson et al., 1995 Jenkin et al., 1992 Jenkin et al., 1992 Jenkin et al., 1992
-55 -44 -43 -42 -41 -27 -23 -9 -34 -58 -81 -73 -163 -76 -65 -142 -62 -39 -38 -71 -65 -79 -69 -78 -62
-10 -3 -15
Knauth & Beeunas, 1986 Knauth & Beeunas, 1986 Knauth & Beeunas, 1986 Knauth & Beeunas, 1986 Yang et al., 1996b Knauth & Beeunas, 1986 Knauth & Beeunas, 1986 Knauth & Beeunas, 1986 Shemesh et al., 1992b Shemesh et al., 1992b Yang et al., 1995 Yang et al., 1995 Yang et al., 1995 Yang et al., 1995 Yang et al., 1995 Madu et al., 1990 Schwarcz et al., 1976 Schwarcz et al., 1976 Schwarcz et al., 1976 Schwarcz et al., 1976 Schwarcz et al., 1976 Harmon et al, 1979 Harmon et al, 1979 Harmon et al, 1979 Harmon et al, 1979
o o
Sedimentary rocks Palo Duro Basin, Texas USA Palo Duro Basin, Texas USA Palo Duro Basin, Texas USA Death Valley, USA Qaidam Basin, China Verdi Valley, Arizona USA Lyons, Kansas USA Salado Fm, New Mexico USA Mount Hemon, Israel Mount Hemon, Israel McArthur Pass, Canada Vermillion Pass, Canada Nahanni Butte, Canada Berry, Canada Kotaneelee, Canada Snowbird, British Colunbia M a m m o t h Cave NP, Kentuky Crystal Cave, Bermuda San Luis Potosi, Mexico N o r m a n Bone Cave, W Virginia Grapevine Cave, W Virginia Coldwater cave, Iowa N o r m a n Bone Cave, W Virginia Tumbling Creek Cave, Missouri M a m m o t h Cave NP, Kentuky
D Cpl C D D D D D D D C C C C C D C C C C C C C C C
Cal-vein Cal-vein Dol cement Dol cement Dol cement Dol cement Dol cement Qtz-vein(Au, Sb) Cal speleothem Cal speleothem Cal speleothem Cal speleothem Cal speleothem Cal speleothem Cal speleothem Cal speleothem Cal speleothem
Halite Halite Halite Halite Halite Halite Halite Halite Cal Cal Dol Dol Qtz, Dol Dol Dol Qtz Cal Cal Cal Cal Cal Cal Cal Cal Cal
Evaporite Evaporite Evaporite Evaporite Evaporite Evaporite Evaporite Evaporite Carbonates Carbonates Dolomite Dolomite Dolomite Dolomite Dolomite Listwanites
35 64 142 145 163 166-175 170 240
-45
-3 -12 -40 -74 -72 -42 -139 -39 -8 -15 -52 -51 -67 -52 -34 -39
Appendix continued >
o ~,,io
C3 r~
9
o
oo
> Appendix continued Mt. Seldom, Israel England Yunan Province, China Yunan Province, China Yunan Province, China Hubei Province, China Hubei Province, China Qinghai Province, China Qinghai Province, China Searles Lake, California Great Lakes North Atlantic Pacific California St Marteen, Caribbean Sea Florida Keys Soreq cave, Israel Soreq cave, Israel Soreq cave, Israel Soreq cave, Israel Villars cave, France Clamouse cave, France
to
C Cal speleothem C C M D C M C M C Biogenic Crb. D Biogenic Crb. D Biogenic Crb. D Biogenic Crb. D Biogenic Crb. D Cal speleothem D Cal speleothem D Cal speleothem D Cal speleothem D Cpl Cal speleothem Cpl Cal speleothem
Plutonic rocks 1 2 3
Connemara, Ireland D Connemara, Ireland D San Diego & Riverside C Counties, US 4 San Diego & Riverside C Counties, US 5 San Diego & Riverside C Counties, US 6 Connemara, Ireland D 7 SW granites, England D/C 8 Bushveld complex, South Africa C 9 SW-Indian Ridge D 10 SW-Indian Ridge D
Halite Cal Halite Halite Halite Halite Halite Halite Halite Halite Cal Cal Cal Cal Cal Cal Cal Cal Cal Cal Cal
-53 -51 -68 -91 -86 -38 -82 -29 -45 -80 -80 -44 -50 -35 -75 -24 -28 -20 -38 -41 -35
Evaporite Evaporite Evaporite Evaporite Evaporite Evaporite Evaporite Evaporite
-34 -39
-78
74 -60 -26 -50 -20 -10 -21 -23 -16 -34 -39 -33
Ohba & Matsuo, 1988 Dennis et al, 2001 Horita & Matsuo, 1986 Horita & Matsuo, 1986 Horita & Matsuo, 1986 Horita & Matsuo, 1986 Horita & Matsuo, 1986 Horita & Matsuo, 1986 Horita & Matsuo, 1986 Horita, 1990 Lecuyer & O'Neil, 1994 Lecuyer & O'Neil, 1994 Lecuyer & O'Neil, 1994 Lecuyer & O'Neil, 1994 Lecuyer & O'Neil, 1994 Matthews et al., 2002 Matthews et al., 2002 Matthews et al., 2002 Matthews et al., 2002 Genty et al., 2002 Genty et al., 2002 r
~r
Granite Granite Gabbro
125-205 270-340 520-565
-45 -24 -58
-17 -17 -51
O'Reilly et al., 1997 O'Reilly et al., 1997 Taylor et al., 1979
Aplite-peg
Gabbro
565-700
-78
-57
Taylor et al., 1979
Graphic-peg
Gabbro
565-700
-52
-50
Taylor et al., 1979
600
-24 -55 -47 -52.0 -50.0
-14 -48
O'Reilly et al., 1997 Alderton & Harmon, 1991 Schiffries & Rye, 1990 Kelley & Friih-Green, 2001 Kelley & Friih-Green, 2001
Qtz-vein Qtz-vein Pocket peg
Qtz-vein(Mo) Rock Pegmatoid Rock Rock
Qtz, F1, Brt Qtz, Cal
Qtz Qtz Qtz Plg Plg
Granite Granite Gabbro O1-Gabbro Gabbro
243-252 259-292
-27.0 -35.0
Appendix continued >
!
~,,i~ ~,,i~
N
> Appendix continued 11 SW-Indian Ridge
D
Rock
Plg/rock
12 SW-Indian Ridge 13 SW-Indian Ridge
D D
Rock Rock
Plg/rock Pig
~r
O1-Gabbro Norite Pegm-Gabbro Felsic rock
234-268 -54.0
-33.0
Kelley & Fr~ih-Green, 2001
333-404 -71.0 190-318 -44.0
-7.0 -28.0
Kelley & Fr~ih-Green, 2001 Kelley & Fr~ih-Green, 2001
Hydrothermal alteration SW granites, England
D/C
SW granites, England
D/C
Amba-Dongar, India
D
SW granites, England Tui mine, New Zeland Tui mine, New Zeland SW granites, England
D/C D D D/C
Bushveld complex, South Africa Bushveld complex, South Africa Tui mine, New Zeland Tui mine, New Zeland Babine Lake, British Columbia Babine Lake, British Columbia Babine Lake, British Columbia Bushveld complex, South Africa Beregovo deposit, Ukraine
C C D C n.s. D D C C
Milos Island, Greece
C
# (1) Location
Qtz-vein (Pb-Zn-Ti) Qtz-vein (Pb-Zn-Ti) Fl-vein
Granite
110-135
-45
Qtz
Granite
110-135
-36
2
Alderton & Harmon, 1991
115-150
-55
-46
230-400 250 250 250-300
-39 -24 -17 -41
-16
Palmer & William-Jones, 1996 Alderton & Harmon, 1991 Robinson, 1974 Robinson, 1974 Alderton & Harmon, 1991
300-600 300-600 350 350 400
-63 -38
-101 -100 -94 -29 -52
Schiffries & Rye, 1990 Schiffries & Rye, 1990 Robinson, 1974 Robinson, 1974 Zaluski et al., 1994 Sheets et al.. 1996 Sheets et al.. 1996 Schiffries & Rye, 1990 Vityk et al., 1993
-0.3
Naden et al., 2003
F1
Qtz-vein (Sn-W) Qtz-vein (Au) Qtz-vein (Au) Qtz-vein (Cu-Pb-Zn) Hydro-vein Qtz-vein Qtz-vein (Au) Qtz-vein (Au) Qtz-vein(CU) Qtz/Cc-vein Qtz/Cc-vein Qtz-plug Mineralized Qtz-vein Vein
Method(2) Sample
F1
Qtz Qtz Sph Qtz
Carbonatite Granite Andesite Andesite Granite
Qtz Qtz Gal Qtz Qtz Qtz Qtz Qtz Qtz
Gabbro Green. Green. Gabbro Andesite Andesite Rhyodacite Porphyry Cu-dep. Porphyry Cu-dep. Gabbro Rhyolite tufts
200-250
-71 -53 -70 -35 -151 -153 -150 -48 -94
Qtz
Tuffs-Ignimbrites
150-284
-7.8
Mineral
Host rock
Met. g r a d e
T~
Alderton & Harmon, 1991
-9
~)13CPDB%o min max
Or~
9 9
~,,d.
9
oo
o o~ C~
o
Reference
Contact metamorphism 1 2 3 4
Dome de l'Agout, France Dome de l'Agout, France Weolag Tu-Mo deposit, Korea Dome de l'Agout, France
C C C C
Rock Rock Qtz-vein (Tu-Mo) Qtz-and-seg
Bulk Bulk Qtz Bulk
Calcsilicate Calcsilicate Calcsilicate Schists(bio)
350 Horn.
-5.3 -4.9 -3.4 -14.7
1.2 0.8
Kreulen & Schuiling, 1982 Kreulen & Schuiling, 1982 So et al., 1983 Kreulen & Schuiling, 1982 Appendix continued >
OO
> Appendix continued 5 6 7 8 9
South Cornawall, UK Soutern India Dome de l'Agout, France Dome de l'Agout, France Weolag Tu-Mo deposit, Korea
4~ Qtz-vein(W-Sn) Qtz Rock Qtz-Grt Qtz-seg Qtz Qtz-seg Qtz Qtz-vein (Tu-Mo) Py
D C C C D C D D D D D D D P D C D D C P D C D C C D C C D
Qtz CalcSchistss Qtz-vein Qtz Marble Qtz-seg Qtz Marble Qtz-seg Qtz Marble Qtz-seg Qtz Marble Qtz-vein Qtz Dolomites Rock Qtz Flysch Qtz-vein(Pb-Zn) Qtz Flysch Qtz-vein(Sn-Cu) Qtz Flysch Qtz-vein Qtz Anphibolites Qtz-vein Qtz, Cal, Ba Metagabbro Qtz-vein Qtz Granitoides Shear zone Qtz Granitoides Qtz-vein Qtz Pegmatite Rock Qtz Orthogneiss Qtz-vein Qtz Orthogneiss Qtz-vein Qtz Orthogneiss Qtz-vein Qtz Pelite Qtz-vein Qtz Gneiss Qtz-seg Qtz Gneiss Rock Qtz Paragneiss Qtz-vein Qtz Paragneiss Qtz-vein Qtz Metasediment Qtz-vein Qtz Schists(Bt) Qtz-seg Qtz Schists(Chl) Qtz-seg Qtz Schists Qtz-vein Qtz Schists Qtz-seg Qtz Schists(graphitic) Qtz-seg Qtz Schists(graphitic) Qtz-vein(Au)
Regional metamorphism 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29
Tauern Window, Austria Naxos, Greece Naxos, Greece Naxos, Greece Wilson Terrane, NVL, Antarctica Naxos, Greece South Cornawall, UK South Cornawall, UK South Cornawall, UK Tauern Window, Austria Connemara, Ireland Western Carpathians Western Carpathians southern india Tauern Window, Austria Central Alps, Switzerland Central Alps, Switzerland Wilson Terrane, NVL, Antarctica Naxos, Greece southern india Central Alps, Switzerland Central Alps, Switzerland Western Carpathians Naxos, Greece Dome de l'Agout, France Central Alps, Switzerland Naxos, Greece Naxos, Greece Okanagan Valley, British Columbia
Flysch Gneiss Gneiss(mu) Gneiss(sill) Hornfels
D P C C C
200-440 Gran. 250
Amph. Amph. Amph. Amph. Amph. Amph. Low grade Low grade Low grade Amph. Green. Green 9 Gran. Amph. Green. Green. Green. Amph. Gran. Green. Green. Green. Amph. Green. Green. Amph. Amph.
400-600 385 565 620 110-150 150-250 290-330 400-600 118-250 300-400 450-500 400-600
680
300-400 530
380-690 380-690 268-286
-11.9 -13.3 -8.5 -10.3 -15.2
-9.0 -6.3 -12.0 -9.4
Wilkinson et al., 1995 Jackson et al., 1988 Kreulen & Schuiling, 1982 Kreulen & Schuiling, 1982 So et al., 1983
-3.8 -2.9 1.7 3.6 -0.1 -7.2 -13.9 -3.9 -14.1 -7.0 -16.8 -10.5 -9.2 -9.5 -5.3 -10.5 -7.0 -3.4 -3.5 -15.8 -6.5 -4.7 -11.8 -8.0 -16.6 -9.4 -5.0 -13.0 -8.9
-1.5
Hoefs & Morteani, 1979 Rye et al., 1976 Rye et al., 1976 Rye et al., 1976 Frezzotti et al., 2000 Kreulen, 1980 Wilkinson et al., 1995 Wilkinson et al., 1995 Wilkinson et al., 1995 Hoefs & Morteani, 1979 O'Reilly et al., 1997 Hurai et al., 1997 Hurai et al., 1997 Harris et al., 1993 Hoefs & Morteani, 1979 Hoefs & Stalder, 1977 Hoefs & Stalder, 1977 Frezzotti et al., 2000 Rye et al., 1976 Harris et al., 1993 Hoefs & Stalder, 1977 Hoefs & Stalder, 1977 Hurai et al., 1997 Rye et al., 1976 Kreulen & Schuiling, 1982 Hoefs & Stalder, 1977 Kreulen, 1980 Kreulen, 1980 Zhang et al., 1989
2.5 5.0
-7.7 -1.8 -3.4 -4.0 -3.1 -6.2 -4.4 -9.0 -2.0 -2.0 -1.0 -5.0 -8.2
Appendix continued >
r
!
C~
N
> A p p e n d i x continued 30 Soutern Alps, N e w Zeland
t~ t~
D
Qtz-vein
Qtz
Schists(Qtz-Feld)
Amph.
200-320
-11.7
-8.2
Jenkin et al., 1994a t~
Volcanic rocks 1 2 3 4 5 7 6 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27
Loihi S e a m o u n t Scotia sea Mariana t r o u g h Loihi S e a m o u n t Indian ridge Mid Atlantic Indian ridge Mid Atlantic Indian ridge Juan de Fuca Explorer S e a m o u n t h Hawai Juan de Fuca Ridge Sea Rise east pacific rise Kilauea East rift C a y m a n Rise Famous zone Mid Alantic Ridge Galapagos Galapagos Ridge Loihi S e a m o u n t C a y m a n Rise Famous zone Mid Alantic Ridge Kilauea East rift Famous zone Mid Alantic Ridge Rita zone East pacfic rise Loihi S e a m o u n t N o r t h w e s t e r margin, N o r t h America 28 Marianna arc
P P P P P P P P C P P P P C P P C P P P P P P P C P P
Rock Rock Rock Rock Rock Rock Rock Rock Rock Rock Rock Rock Rock Rock Rock Rock Rock Rock Rock Rock Rock Rock Rock Rock Rock Rock Rock
Glass Glass Glass Glass Glass Glass Glass Glass Glass Glass Glass Glass Glass Glass Glass Glass Glass Glass Glass Glass Glass Glass Glass Glass Glass Glass Bulk
P
Rock
Glass
C
Pocket peg
P
Rock
Alkali basalt BABB BABB Basanite MORB MORB MORB MORB MORB MORB OIB OIB Tholeiite Tholeiite Tholeiite Tholeiite Tholeiite Tholeiite Tholeiite Tholeiite Tholeiite Tholeiite Tholeiite Tholeiite Tholeiite Trans basalt Xenolites
-8.3 -12.0 -11.8 -5.8 -11.7 -5.3 -10.5 -7.8 -5.8 -7.4 -5.4 -6.7 -24.7 -9.1 -8.8 -7.1 -7.0 -6.4 -6.4 -6.3 -6.3 -6.2 -6.1 -6.0 -5.6 -5.9 -9.9
-6.5 -9.0
-7.3 -6.7 -5.3
-3.5 -22.3 -4.2 -4.9 -6.0 -5.5 -5.6 -5.8 -5.1
-5.8 -0.7 -4.0
Exley et al., 1986 Mattey et al., 1984 Mattey et al., 1984 Exley et al., 1986 Mattey et al., 1989 Mattey et al., 1984 Mattey et al., 1989 Exley et al., 1986 Mattey et al., 1989 Mattey et al., 1984 Mattey et al., 1984 Mattey et al., 1984 Sakai et al., 1984 Pineau & Javoy, 1983 Des Marais & Moore, Des Marais & Moore, Pineau & Javoy, 1983 Des Marais & Moore, Sakai et al., 1984 Exley et al., 1986 Sakai et al., 1984 Sakai et al., 1984 Sakai et al., 1984 Des Marais & Moore, Pineau & Javoy, 1983 Exley et al., 1986 N a d e a u et al., 1990
~r r~
o 9 r~
9 w,,~.
1984 1984
o 9
1984 C~
o
1984
-29.7
-24.7
Mattey et al., 1984
-15.5
-3.0
Taylor et al., 1979
-17.1
-8.3
Miller & Pillinger, 1997
Plutonic rocks San Diego & Riverside Counties, US Cornubian Batholith, England
Gabbro Qtz
Granite
520-565
A p p e n d i x continued >
OO
> Appendix continued Carrock Fell, England Xihuashan Cornubian Batholith, England Dome de l'Agout, France Dome de l'Agout, France Connemara, Ireland Southern india SW-Indian Ridge SW-Indian Ridge SW-Indian Ridge
P P C C C D D D D D
Rock Rock Rock Qtz-seg Peg Qtz-vein Rock Rock Rock Rock
Qtz Qtz Qtz Qtz Bulk Qtz, F1, Ba Qtz Plg/ rock Plg/ rock Pig / rock
SW-Indian Ridge SW-Indian Ridge Soutern India
D D P
Rock Rock Rock
Plg/ rock Plg/ rock Qtz-Grt
D D
Qtz pebbles Qtz pebbles
Sedimentary rocks Stanleigh Mine, Canada Venterrdorp Contact Reel South Africa Snowbird, British Colunbia, Canada Providencia, Mexico Providencia, Mexico
D
Qtz-vein(Au, Sb)
C C
Granite Granite Granite Granite Granite Granite Charnockite O1-Gabbro Gabbro O1-Gabbro Norite Pegm-Gabbro Felsic Vein Charnockite
125-205
Conglomerates Conglomerates
-16.0 -9.7 -8.3 -11.8 -2.7 -18.7 -12.4 -11.7 -10.2 -10.4
-16.0 -3.0 -7.4 -8.9 -9.5 -8.1 -1.9 -2.4 -4.6
Miller & Pillinger, 1997 Miller & Pillinger, 1997 Miller & Pillinger, 1997 Kreulen & Schuiling, 1982 Kreulen & Schuiling, 1982 O'Reilly et al., 1997 Harris et al., 1993 Kelley & Friih-Green, 2001 Kelley & Fr/.ih-Green, 2001 Kelley & Fr/ih-Green, 2001
-23.9 -24.9 -11.9
-1.7 -8.4 -6.3
Kelley & Fr/ih-Green, 2001 Kelley & Fr/ih-Green, 2001 Jackson et al., 1988
-8.0
1.5
1.1 -8.6
Qtz
Listwanites
240
-10.0
Sph Cal
Limestones Limestones
290-365 330-350
-7.0 -11.0
T~
~}180%o
Vennemann et al. 1992 Vennemann et al., 1992 Madu et al., 1990 Rye & O'Neil, 1968 Rye & O'Neil, 1968
r =r r~
# (1) Location
McArthur Pass, Canada Vermillion Pass, Canada Nahanni Butte, Canada Berry, Canada Kotaneelee, Canada Dabsun Lake, China Mt. Seldom, Israel Inghilterra ? Yunan Province Yunan Province
Method(2) Sample
C C C C C M C C C M
Dol Dol Dol Dol Dol
cement cement cement cement cement
Cal speleothem
Mineral
H o s t rock
Met. g r a d e
min Dol Dol Qtz, Dol Dol Dol Halite Halite Cal Halite Halite
Dolomite Dolomite Dolomite Dolomite Dolomite Evaporite Evaporite Evaporite Evaporite Evaporite
142 145 163 166-175 170
-13.7 -11.4 -18.3 -5.7 -3.7 0.93 -1.9 -8.44 -2.2 -21.9
Reference !
max
-11.1 -11.2 -1.7 1.26 2.6 -5.2
Yang et al., 1995 Yang et al., 1995 Yang et al., 1995 Yang et al., 1995 Yang et al., 1995 Yang et al., 1996b Ohba & Matsuo, 1988 Dennis et al., 2001 Horita & Matsuo, 1986 Horita & Matsuo, 1986 Appendix continued >
=r
~,,ao
N
> Appendix
continued
Hubei Province Hubei Province Qinghai Province Qinghai Province Searles Lake, California Providencia, Mexico Providencia, Mexico Providencia, Mexico Kaneuchi deposit, Japan Beregovo deposit, Ukraine
C M C M C C C C C/D C
Milos Island, Greece
C
Qtz-vein Mineralized Qtz-vein Vein
Halite Halite Halite Halite Halite Sph Qtz Cal Qtz Qtz Qtz
Evaporite Evaporite Evaporite Evaporite Evaporite Limestones Limestones Limestones Rhyolite tufts
290-365 330-315 330-350 150-350 200-250
3.3 -14.2 6 -16.8 -9.8 5.8 -4.5 -0.2 -1.2 -7.1
14.1 6.2 -3.7 0.8 -0.3 -0.9
Horita & Matsuo, 1986 Horita & Matsuo, 1986 Horita & Matsuo, 1996 Horita & Matsuo, 1996 Horita, 1990 Rye & O'Neil, 1968 Rye & O'Neil, 1968 Rye & O'Neil, 1968 Ohba et al., 1995 Vityk et al., 1993
Tuffs-Ignimbrites
150-284
-7.8
-0.3
Naden et al., 2003
1) Reference n u m b e r in Figures 3.1abc and3.2abc; 2) Extraction method, C: crushing; D" decrepitation; D / C: both crushing and decrepitation; P: pyrolysis; M: melting; CpI: capillary; n.s." not specified. 3) TH~ h o m o g e n i z a t i o n temperature for the fluid inclusions or calculated temperature of trapping. Mineral abbreviations according to Kretz, (1983).
o o r~
>
9
r~ o
9
Handbook of Stable Isotope Analytical Techniques, Volume 1 P.A. de Groot (Editor) 9 2004 Elsevier B.V. All fights reserved.
CHAPTER 4 Selected Ion Flow Tube Mass Spectrometry (SIFT-MS) and Flowing Afterglow Mass Spectrometry (FA-MS) for the Determination of the Deuterium Abundance in Water Vapour Patrik Spanell & David Smith2 I V. Cerm~ik Laboratory, J. Heyrovsky Institute of Physical Chemistry, Academy of Sciences of the Czech Republic, Dolejskova 3, 182 23, Prague 8, Czech Republic 2 Centre for Science and Technology in Medicine, School of Postgraduate Medicine, Keele University, Thornburrow Drive, Hartshill, Stoke-on-Trent, ST4 7QB, U.K. e-mail: I
[email protected]; 2
[email protected]
4.1 Introduction In this chapter, we describe novel methods by which on-line, real-time water vapour deuterium abundance can be measured in single breath exhalations and above aqueous liquids, including urine and blood. The need for these measurements became apparent as our trace gas analytical techniques began to be used for the determination of trace gas metabolites in breath for clinical diagnosis and therapeutic monitoring (Smith & Spanel, 1996a, b; Spanel et al., 1998). In particular, it became clear that a rapid, non-invasive method to determine total body water in patients suffering from end-stage renal failure would have great value if it could be used directly in the clinical environment. Deuterium isotope dilution within the body coupled with a new analytical approach offers a route to this objective. Deuterium abundance is conventionally determined in liquid water, urine, saliva and blood samples by first equilibrating a sample of these media with gaseous hydrogen in the presence of a catalyst, thus producing a H2/HD mixture above the liquid. Then conventional mass spectrometry, sometimes coupled with gas chromatography (GCMS) is used to quantify H2 + and HD + ions derived from this mixture (Begley & Scrimgeour, 1997, Part 1, Chapter 1). However, this approach requires long-term sample preparation and relatively laborious analysis. Thus, several days pass from sample acquisition to analysis. The value of a method that would provide an immediate result is obvious. In response to this need, we have developed the selected ion flow tube mass spectrometric method (SIFT-MS) and the flowing afterglow mass spectrometric method (FA-MS) for the on-line, real time determination of the deuterium abundance in water vapour. Both methods involve the flow of thermalised H30+(H20)3 ions along a fast flow tube. These ions react in multiple collisions with molecules of water vapour
Selected Ion Flow Tube Mass Spectrometry(SIFT-MS)and FlowingAfterglow ...
89
introduced into the flow tube and their isotopic composition reaches equilibrium and is analysed by a mass spectrometer located downstream. A typical mass spectrum obtained when mass selected hydronium ions, H3160 + (m/z = 19), are used as precursor ions for the analysis of humid air and breath by the SIFT-MS method (Smith & Spanel, 1996a, b) is shown in Figure 4-1a. Additional ions appear at m / z of 20 and 21 in the product ion mass spectrum (see the spectra in Figure 4-1). These product ions are the isotopomers of H30 + containing respectively D, 170, and 180, which are formed in the helium carrier gas by isotope exchange reactions between the injected, thermalised H3160 + ions and the isotopomers of water, i.e. H2160, HD160, H2170 and H2180. Sequences of ion-molecule reactions also occur that form the hydrated hydronium (water cluster) ions H30+~ H30+.(H20)2 and H30 +~(H20)3 at m / z of 37, 55, and 73 respectively, which appear on the mass spectra together with their corresponding D, 170 and 180 isotopomers at m / z 38 and 39, 56 and 57 and 74 and 75 (see Figure 4-1a). Clearlyi there is information in these mass spectra on the isotopic composition of the water present in the carrier gas. Hence, the deuterium content of a water sample introduced into the helium carrier gas can be determined from such spectra. To properly understand that which follows, we need to distinguish between the isotopic composition of the following three "phases"" the liquid water sample (designated by the subscript liq), the water vapour transferred from an aqueous sample headspace into the helium carrier gas (designated by the subscript yap) and the H30+(H20)o,1,2,3 ions that comprise the ion swarm created in the carrier gas (designated by the subscript ion).
4.2 Background science 4.2.1 Partition of HDO between liquid water and its vapour In water containing a low abundance of deuterium almost all the deuterium is contained in HDO molecules. Therefore, in order to determine the deuterium isotope abundance ratio in a liquid water sample, Rlliq = D / ( H + D), by analysing its vapour, the partition of HDO between the liquid and vapour phases needs to be addressed. A difference arises because HDO has a lower saturated vapour pressure than H20 at sub-boiling temperatures. Thus, the deuterium abundance in the headspace vapour, Rlvap, is lower than that in the liquid, Rlliq. The ratio of these parameters, i.e. the temperature dependent dimensionless partition coefficient K 1 - Rlvap/Rlliq, can be calculated from the data and the equations given by Van Hook (1972) and Jancso & Van Hook (1974) using Raoult's law (Atkins, 1990). In Figure 4-2, K1 is plotted as a function of temperature. Similarly, the values of the isotope abundance ratio of 170 in the water vapour, R2vap - 170/(160 + 170 + 180) and that of 180, R3vap - 180/(160 + 170 + 180) are proportional to their corresponding abundance in the liquid R2liq and R3liq. The partition coefficients K 2 - R2vap/R2liq and K3 - R3vap/R3liq derived from the data given by Jancso & Van Hook (1974) are also given in Figure 4-2. Note that K2 is very close to unity and that there is a small but discernible effect for 180 that can be accounted for by K3.
90
Chapter 4- P. Spanel & D. Smith
Figure 4.1 - Spectra obtained by SIFT-MS from water vapour plotted on a semi-logarithmic scale as counts per second (c/s) versus mass to charge ratio (m/z). a) A spectral scan over the m / z range 10 to 100 for tap water when H30 + is injected ( m / z = 19). Note the production of the three hydrates H30+(H20)1,2,3 at m / z values 37, 55 and 73. Note also the appearance of their D, 170, and 180 isotopomers, b) A spectral scan from m / z 72 to 76 for tap water. Note the clear separation of the mass peaks. Mean count rates were obtained for each m / z value by recording and averaging the count rates at four mass settings around each peak (see text), c) A spectral scan from m / z 72 to 76 for a 1% mixture of D20 in tap water. Note the much larger count rate at m / z =74 compared to that in b) for tap water alone
Selected Ion Flow Tube Mass Spectrometry (SIFT-MS) and Flowing Afterglow ...
91
Figure 4.2 -Dimensionless water vapour/liquid water partition coefficients K1, K2 and K3 plotted as a function of temperature for the isotopomers HDO, H2170 and H2180 respectively, calculated with respect to H20 from the data and equations given in Van Hook, 1972, and Jancso & Van Hook, 1974, using Raoult's law (Atkins, 1990). The vertical lines at 30~ and 37~ (body temperature) indicate typical temperatures used for measurements.
4.2.2 Ion molecule reactions; equilibrium between ions and water vapour molecules in the gas phase To determine the deuterium abundance in water vapour using our novel methods, it is essential to understand the ion chemistry that generates the isotopomers of the ions. The isotopic composition of the ion swarm is primarily determined by the kinetics of the ion-molecule reactions. The initial reactions that occur (Spanel & Smith, 2000) are: H30 + + HDO <--,H2DO + + H20
[4.1]
Studies of the analogous isotope exchange reactions of H30 + with D20 indicate that the rate coefficients are close to their collisional rate coefficients (Adams et al., 1982; Smith et al., 1980) and that statistical mixing of the H and D isotopes amongst the ionic and neutral products occurs. Thus, equilibrium in reaction system [4.1] will be approached rapidly. The equilibrium constant, K, for this reactive system is influenced by the enthalpy change, AH - 0.4 kJ/mol and by the entropy change, AS - 3.4 J / m o l / K (Spanel & Smith, 2000). Thus, the equilibrium constant obtained using the relations In K = - AG/RT and AG = AH - TAS is K = 1.11. This value is significantly lower than the statistical value 3/2 expected on the basis of the number of H and D atoms alone.
92
Chapter 4- P. Spanel & D. Smith
In addition, the following three-body association reactions occur in parallel with reaction [4.1]: H2DO + + H20 + M --* (H4DO2+) * + M H30 + + HDO + M ~ (H4DO2+) * + M
[4.2a] [4.2b]
Because of these, the reactive system [4.1] above can never reach true equilibrium as the number density of H2DO + and H30 + ions decrease with time (distance along the flow tube). Further sequential association reactions also occur producing the higher order hydrates. Hence, these complications prohibit the use of the H30+(H20)0,1,2 ions for deuterium analyses (Spanel & Smith, 2000). Fortunatel~ H30+(H20)3 ions become the dominant ionic species in the carrier gas at sufficiently large water molecule concentrations. For these triply-hydrated H30 + ions further association with water molecules is much slower. Thus, H30+(H20)3 cluster ions are considered to be in equilibrium with water vapour molecules. The following isotope exchange and ligand switching reactions establish the equilibrium: H30+(H20)3 + HDO ~ H8DO4 + + H20 H30+(H20)3 + HDO --* H30+(H20)2HDO + H20
[4.3] [4.4]
Efficient mixing of H and D atoms occurs within the intermediate reaction complex ion (H10DO5+)*, as was demonstrated for the analogous reactions of H30+(H20)3 with D20 by Adams et al. (1982) and Henchman et a1.(1982). Calculations of the enthalpy and entropy changes for reaction [4.3] are not trivial, but it is reasonable to assume that the AH will be smaller than that for reaction [4.1]. In effect, AH is zero at 300 K. We can also postulate that in reaction [4.3] the translational and rotational entropy changes are also relatively small and that AS is entirely described by statistical factors (Spanel & Smith, 2000). Thus, when equilibrium is established by reactions [4.3] and [4.4], the deuterium abundance ratio in the H30+(H20)3 cluster ion swarm, Rlion, will be equal to that in the water vapour, Rlvap. This has been experimentally validated under both SIFT-MS (Spanel & Smith, 2000) and FA-MS (Spanel & Smith, 2001) conditions using standard deuterium enriched water. The speed of approach to equilibrium in these reactive systems depends on the number density of water molecules in the carrier gas and the rate coefficients for the isotope exchange reaction [4.3] and the equivalent ligand switching reaction [4.4]. These rate coefficients are close to 2 x 10-9 cm3s-1 at 300 K (Ikezoe et al., 1987). Thus for SIFT-MS and FA-MS deuterium analyses, the number density of water molecules in the carrier gas should be more than 1013 cm -3, in order to establish equilibrium in about I ms (Spanel & Smith, 2000), which is short compared to the reaction time.
4.2.3 Ion signals and isotopomer overlap The final important issue to be addressed is the relationship of Rlion to the observed ion signals. The equilibrium distribution of the various isotopomer ions
93
Selected Ion Flow Tube Mass Spectrometry (SIFT-MS) and Flowing Afterglow ...
(and hence their count rates at the detector; see Figure 4.3) can be expressed according to a binomial distribution (Karasek & Clement, 1988). For the case of d e u t e r i u m isotopes only, the relative intensities of the individual isotopic variants of ions at mass M ( m / z - 73), M + 1, M + m, etc., can be expressed in terms of Rlion as: [4.5a] [4.5b]
[M] = ( 1 - Rlion) n [M + 1] = nRlion (1 - Rlion) n-1
n! m [M + m] - m ! ( n - m)! R l i ~
)n-m - Rli~
[4.5c]
Figure 4.3 - Schematic diagrams of the SIFT-MS and FA-MS instrumental configurations. In SIFT-MS, ions are selected by a quadrupole mass filter and injected into the helium carrier gas. In FA-MS, the ions are created in the carrier gas by a microwave discharge thus producing a weak flowing afterglow along the flow tube. Direct samplings of breath and liquid headspace are achieved as shown in the insets.
94
Chapter 4- P. Spanel & D. Smith
Here, n is the number of hydrogen atoms in the molecular ion (n = 9 for H30+(H20)3) and m is the number of deuterium atoms in the isotopic variants of the ions (m = 0, 1, 2,...). Thus, the ratio of the concentrations of singly deuterated (m - 1) and the nondeuterated (m = 0) variants of H30+(H20)3, is: [H8DO4+]/[H904 +] - [M + 1 ] / [ M ] - 9
Rlion 1 -Rlion
[4.6]
So the deuterium abundance is amplified by a factor of ~ 9 relative to water vapour. The presence of the oxygen isotopes in the ions can be accounted for in a similar way. The 170 containing ions contribute to the measured [M + 1]/[M] ion signal ratio. Due to the relatively small mass differences of the oxygen isotopes, kinetic isotope effects and enthalpy and entropy differences are not significant in the following oxygen isotope exchange reactions at 300 K (Spanel & Smith, 2000): H3160 + + H3170 (or H3180)3 ~ H3170 + (or H3180+)3 + H2160 + [4.7] Therefore, R2ion = R2vap and thus the 170 isotopomers contribute an amount 4R2ion/(1 - R2ion) to the [M + 1]/[M] signal ratio, which can be expressed as: N(74)/N(73)
- [H8D1604+]/[H 9160 4 +] + [H 91704+] / [ H 91604+]
Rlion
--
Rlion
- (91 ---1-~'-l;on+ 41 -R-l-ion)
[4.81
N(73) and N(74) are the signal intensities for all ions at m / z = 73 and m / z = 74. Since mass discrimination between these adjacent mass ions is insignificant and the sampling efficiencies are the same, we assume that the ion signals intensities (count rates) are proportional to the number densities of the various ions in the carrier gas, that is N(74) ~[H8D1604 +] + [H91704 +] a n d N(73)~[H91604+].
Similarly, the 180 isotopomer of the H904 + ion at m / z - 75 will have a signal level of 4R3ion/(1 - R3ion) relatively to the 160 isotopomer, with a potential additional contribution from the doubly deuterated ion H7D204 + for rich H D O / H 2 0 mixtures. However, it can be shown using equations [4.5a] and [4.5c] that when R1 < 10-3 the contribution of doubly deuterated ions is negligible (< 3 x 10-5 representing < 0.15% of a typical R3 value of 0.02). Thus, from an accurate measurement of the ion count rate ratio, Q - N(74)/N(75), the D abundance, Rlvap, in water vapour is determined as" - 4 Q R 3 v a p - 4R2vap Rlvap 9
[4.9]
The Q value obtained for normal water is typically 0.35 (see Figure 4.1b). Thus, the actual measurements involve ion count rates at m / z of 74 and 75 that are not very dif-
Selected Ion Flow Tube Mass Spectrometry (SIFT-MS) and Flowing Afterglow ...
95
ferent, which is inherently more accurate than attempting to measure count rates (and ion currents) that are vastly different (e.g. Begley & Scrimgeour, 1997). Now we show how these principles relating Rlvap and the isotopomer ion mass spectra are technically implemented in the SIFT-MS and FA-MS techniques.
4.3 The SIFT-MS and FA-MS experimental techniques 4.3.1 SIFT-MS The SIFT-MS technique has been described in detail in several recent reviews (Smith & Spanel, 1996a,b). It is based on the selected ion flow tube, SIFT, technique which has been used for more than two decades to determine the rate coefficients and ion products of ion-molecule reactions at thermal energies (Smith & Adams, 1988). The principle of the SIFT technique is illustrated in Figure 4.3. Ions are created in an ion source and a current of mass selected (precursor) ions of a given m / z is obtained using a quadrupole mass filter. These precursor ions (typically H30 +, NO + and 02+; Smith & Spanel, 1996a, b) are injected into a fast flowing (~ 60 m/s) inert carrier gas (helium at a pressure of ~ 1 Tort (= 100 Pa)) through a Venturi-type orifice (diameter 1 to 2 ram). The ions are convected along a flow tube (typically 30 to 100 cm long) by the fast flowing carrier gas as a thermalised ion swarm possessing a Maxwellian velocity distribution appropriate to the carrier gas temperature (usually 300 K). The sample of air/gas mixture to be analyzed is introduced at a known flow rate into the carrier gas and its component molecules react with the precursor ions. The precursor ions and the product ions of the reactions are sampled from the flowing swarm via a pinhole orifice (~ 0.3 mm diameter) located at the downstream end of the flow tube and analyzed by a differentially pumped quadrupole mass spectrometer (p < 10-4 Tort) with a single channel multiplier ion counting detector. Concentrations of the trace gases in the carrier gas can be calculated from the ion signal levels and the rate coefficients for the reactions of the precursor ion with the individual trace gases (Spanel & Smith, 1996). In this way, the trace gases in complex mixtures such as human breath can be quantified (Smith & Spanel, 1996a, b). For the deuterium isotopic analyses, H30 + precursor ions are used exclusively. Injected H30 + ions convert to H30+(H20)1,2,3 ions and their D, 170 and 180 isotopomers when a water vapour sample is introduced into the carrier gas, as is discussed above. Thus, ions are produced at m / z values of 73, 74 and 75 (see Figure 4.1). If the count rates of all three ions are to be determined accurately the dynamic range of the ion counter must be large. This is because the count rate at m / z of 73 (H904 + ions) is typically more than 100 times those for the ions at m / z of 74 and 75 for tap water vapour (see Figure 4.1b). Since it is desirable to maximise the count rates of the minor product ions (i.e. the (H8DO4 + + H917OO3 +) and H9180 + ions) in order to minimise random counting errors, this corresponds to relatively large count rates for the m / z 73 ions. A practical upper limit for the count rate of ions at m / z of 73 that can be obtained in SIFT-MS (i.e. mass filtered ions) is about 105/s. Then the count rates at m / z values of 74 and 75 are typically 103 /s. To record count rates as large as 105/s accurately, the counting system must be capable of counting at about 107 c/s and even then the dead time of the detection system has to be accounted for (Fahey 1998). To avoid potential problems relating to such high count rates, the deuterium analysis can
96
Chapter 4- P. Spanel & D. Smith
be done by recording only the N(74)/N(75) ratio, and then exploiting equation [4.9]. This is the basis of the SIFT-MS (and the FA-MS) technique for D / H analysis. Because the count rates of ions at m / z - 74 and 75 ions are comparable, accuracy of measurement is optimised. But poorer precision is obtained using SIFT-MS for deuterium analyses, due to the relatively low count rates at m / z of 74 and 75, which are at a maximum of about 103/s (see section 4-5). However, greater precision is obtained by the much higher count rates obtained using FA-MS, as we indicate later. The abundance sensitivity of the analytical mass spectrometer must be better than 10-5, so that less then 10-5 of the major intensity peak ( m / z - 73) overlaps with the adjacent minor peak (m/z - 74:). Mass discrimination between the m / z 74 and 75 ions must be minimised (to less than 0.1%) and therefore the mass spectrometer must be operated at the lowest practical resolution consistent with the required abundance sensitivity. Then, the results that we obtain using standard samples indicate that mass discrimination between m / z = 74 and m / z = 75 is negligible (Spanel & Smith, 2001). Accurate definition of the ion peaks is essential to ensure that the ion counts can be integrated over m / z intervals having identical width (0.3 mass units is chosen in this work) and spaced exactly by I m / z unit (see Figure 4.1). In practice, this means that a 16 bit D / A converter must be used to control the mass setting of the mass spectrometer. The actual measurement of the count rates also requires some discussion. We have found that operating the downstream mass spectrometer in the multiple-ion-monitoring mode (Spanel & Smith, 1996) improves precision of measurement in comparison to scanning over the two ion peaks. Thus, for example, to measure the mean count rates of ions at m / z of 74 and 75, the count rates at m / z values of 73.8, 73.9, 74.0, 74.1 and 74.8, 74.9, 75.0, 75.1 are recorded by fast stepping between these masses, using appropriate dwell times, (typically 20 ms)(Spanel & Smith, 1996). This minimizes the possible effects of any fine structure on the peak shapes. Limited mass scans around m / z of 74 and 75 are shown in Figures 4.1b and 4.1c. It can be seen that the chosen m / z settings straddle the required peaks. For analysis, the ion signals N(74) and N(75) were then taken as the mean values of the four count rates,/, obtained for each of the two ions. The ratio Q is then calculated as: N(74) I(73.8) + I(73.9)+ I(74.0)+ I(74.1) Q - N(75) = I(74.8) + I(74.9)+ I(75.0)+ I(75.1)
[4.10]
For the commercially available instrumentation (Trans Spectra Limited, U.K.) all numerical analysis using equations [4.9] and [4.10] and the K1 partition coefficient are performed on-line by an embedded computer, thus providing an instantaneous readout of the Q, Rlvap and Rlliq values.
4.3.2. FA-MS We introduced the flowing afterglow mass spectrometric method, FA-MS, to improve measurement precision and to allow deuterium analyses to be made in the water vapour contained in single breath exhalations that last about 5 s. This obviously requires a considerable increase in the H30 + precursor ion count rate to values that
Selected Ion Flow Tube Mass Spectrometry(SIFT-MS)and Flowing Afterglow ...
97
cannot be achieved with the upstream mass filter used for SIFT-MS (see Figure 4.3). So, for FA-MS, a weak microwave discharge is created in a narrow glass tube that carries a fraction of the helium carrier gas into the stainless steel flow tube, thus forming a flowing afterglow in the flow tube (see Figure 4.3). The gas phase ion chemistry involving the trace amount of water present in He results in the formation of H30 + ions in the flow tube at much higher number densities than can be produced using SIFT-MS. In every other way FA-MS is identical to SIFT-MS. The H30 + ions are very rapidly converted to the H30+(H20)3 + precursor ions on addition of the water vapour sample into the helium carrier gas (see Figure 4.1). Using this arrangement, very large count rates at m / z of 73, typically several millions per second, are seen at the mass spectrometer/ion detection system. Such large count rates cannot be counted sufficiently accurately by a conventional counting system. However, the count rates of the ions at m / z - 74 and 75 are also much greater than are achievable in SIFT-MS. For optimum accuracy and precision, it is best to arrange that the count rates of the m / z 74 and 75 ions are within the range 10,000 to 30,000 c/s. These can be adjusted by varying the helium flow through the microwave discharge. Now the count rates of the ions at m / z = 75 and 74 are used to determine the deuterium content of the water vapour sample (e.g. in breath) according to equation [4.9]. The price paid for bypassing the upstream mass filter and creating large ion count rates is the presence of other minority ions (such as 02 + and NO +) in the carrier gas. This precludes the use of FAMS for traces gas analysis of complex mixes, but does not pose problems for deuterium isotope analysis. Again, four mass settings are used for each m / z peak and then the analysis is effectively identical to that described for SIFT-MS. Commercial FA-MS instrumentation (Trans Spectra Limited, U.K) are capable of real time calculations of the deuterium abundance exploiting equation [4.9] and [4.10] using the embedded computer.
4.4 Sampling methods 4.4.1 Direct breath sampling Air or breath is sampled into the flow tube via a heated, calibrated capillary tube (typically stainless steel of internal diameter 0.3 mm, length 8 cm) that is coupled to the flow tube via heated stainless steel tubing (internal diameter 5 mm, length about 30 cm). The tip of the capillary extends into a stainless steel coupling positioned perpendicularly to the capillary axis (Figure 4.3). Exhaled breath is introduced into the coupling via a standard disposable cardboard mouthpiece (about 15-mm diameter). This arrangement is patient friendly, offering a suitable resistance to the flow of breath such that a steady exhalation can be made over a few seconds. The exhaled breath totally displaces the ambient air from the entrance to the sampling capillary and so a sample of breath enters the capillary and immediately expands into the coupling tubing (pressure of about I Torr; Smith & Spanel, 1996a, b). The entrance to the capillary is again exposed to the ambient air upon oral inhalation. By suitable choice of the dimensions of the calibrated capillary, the flow rate of the sampled air/breath is arranged to be about 3% of the helium flow rate. This sample flow rate is sufficient to convert the majority of the precursor H30 + ions to the desired H30+(H20)3 ions. In the current SIFT-MS and FA-MS instruments, the helium flow rate is typically 60 Torr.litres/sec and that of the air/breath sample is typically 2 Torr.litres/sec. The flow
98
Chapter 4- P. Spanel & D. Smith
time of air/breath through the capillary and the coupling tubes is about 10 ms, comparable to the flow time of the sampled gases along the flow tube. Thus, the net response time of the instrument is about 20 ms. Since the duration of a typical breath exhalation is about 5 s, time profiles of the individual ion signals can be defined. Typical time profiles of the m / z = 74 and 75 ions are shown in Figure 4.4a for three breath exhalation/inhalation cycles. It is imperative to heat the capillary and the coupling lines to minimise the condensation of water, and thus to eliminate "memory" effects. An example of the application of this technique to determine deuterium abundance in breath water vapour has been described by Davies et al. (2001b). An individual was dosed with a small amount of D20 (typically 0.3 g per kg of body weight) and then the dispersal of the deuterium throughout the body was monitored by FA-MS breath analysis. A typical time variation of deuterium in the breath water vapour is shown in Figure 4.4b. The interpretation of such data and its use to derive total body water is discussed by Davies et al. (2001b). This breath sampling technique has been used for SIFT-MS quantification of a variety of breath metabolites in clinical investigations (Spanel et al., 1998; Davies et al., 1997, 2001a; Smith et al., 2002b).
4.4.2 Headspace sampling For medical diagnostic studies, we are using SIFT-MS to analyse the volatile metabolites in the headspace of blood, urine and the dialysate fluid used for peritoneal dialysis. To date, our urine headspace analyses have been most productive in clinical diagnosis (Spanel et al., 1999). FA-MS can be used very effectively to accurately determine deuterium (HDO) in the headspace above the surface of these media using the following sampling method. About 10 ml of fluid is placed in a septumsealed 200-ml glass bottle. The bottle is then placed in a temperature controlled water bath and the headspace is allowed to develop for about 20 min. A hypodermic needle connected directly to the input line of the SIFT-MS or FA-MS instrument (Figure 4.3) is then used to pierce the septum and sample the headspace vapour. As before, the sampling lines are held at about 100~ to inhibit the condensation of water vapour and other condensables. The headspace above the liquid is drawn into the carrier gas by the pressure differential (from atmosphere to flow tube pressure). The analytical mass spectrometer can be operated in either the full scan mode to obtain mass spectra over a predetermined m / z range or in the multi-ion mode to track the concentrations of volatiles or to determine the deuterium abundance in the headspace water vapour. Since the sample bottle is sealed, its internal pressure, and hence the sample flow rate, decreases with time (Spanel & Smith, 2001). This is illustrated by the FA-MS data in Figure 4.5a, which show the decrease in the m / z = 74 and 75 ions as the headspace above a sample of dialysate fluid flows into the helium carrier gas. The deuterium abundance does not change during the sampling time, only the actual count rates of the analytical ions (at m / z = 73, 74 and 75) reduce as the flow of the H 2 0 / H D O vapour diminishes. Integration of the signal counts over the sampling time period (typically 30 seconds) improves the precision of the deuterium abundance determination. Figure 4.5b shows the increase with time of the deuterium abundance in dialy-
Selected Ion Flow Tube Mass Spectrometry (SIFT-MS) and Flowing Afterglow ...
F i g u r e 4 . 4 - a) Mean values of the FA-MS ion count rates (c/s obtained at the four mass settings as described in the text and indicated in Figure 4.1) for m / z values of 74 and 75, observed as three breath exhalations are directly sampled. The mean values of the ion count rate ratios, Q, over the alveolar breath intervals indicated are used to calculate the values of Rlvap (and then Rlliq) given for each exhalation, b) An example of a typical long time variation of Rlliq determined for breath water following oral ingestion of 18.7 g of D20 by a volunteer (filled squares). The open squares are simultaneous measurements for a control who has not ingested D20.
99
100
Chapter 4- P. Spanel & D. Smith
Figure 4.5 - a) Mean values of FA-MS ion count rates at m / z = 74 and 75, accumulated as the headspace from a sealed bottle containing dialysate fluid at 37~ flowed into the helium carrier gas. The decrease with time of the count rates for each ion is due to the decrease in the water vapour/air pressure in the sealed bottle during sampling. The mean ion count rate ratio, Q, is used to calculate Rlvap (and then Rlliq). b) Time profile of Rlliq for dialysate samples taken during a CAPD (continuous ambulatory peritoneal dialysis) session following a dialysate exchange at time 0. The patient ingested 17g of D20 four hours prior to t=0.
Selected Ion Flow Tube Mass Spectrometry(SIFT-MS)and FlowingAfterglow...
101
sate fluid in the peritoneal cavity following a regular dialysate exchange by a patient that had orally ingested a small quantity of D20 four hours previously (Ashgar et al., 2003). Such data are being used to study the flow of water across the peritoneal membrane (Ashgar et al., 2003). They exemplify the precision that can be obtained using this FA-MS analytical method. SIFT-MS has been also used to study isotopic composition of ethanol and acetaldehyde produced by yeast cells fermenting deuterium labeled glucose (Smith et al., 2002a).
4.5 Accuracy and precision The accuracy of the measured Rlvap also depends on the accuracy of the adopted values of R2vap and R3vap. The values we use are derived from the known values of R2liq and R3liq, viz. 0.000379 and 0.002006 respectively (Li et al., 1988; Baertschi, 1976). The liquid/vapour phase partition coefficients K2 and K3 are slightly temperature dependent (Figure 4.2). When breath is analysed for deuterium, values corresponding to the alveolar interface temperatures must be used, which can range from 34 ~ to 37~ (Wilson et al 2001). This spread in temperature results in only a 0.3% variation of K1. Also, the accuracy of the derived value for the deuterium abundance in the liquid is influenced by the accuracy of the K1 value. Fortunately, K1 is well characterised (Figure 4.2). The accuracy may be compromised in FA-MS measurements by the production of unwanted ions via the reactions of the ions formed in the microwave discharge with trace organic impurities, possibly causing isobaric interference at m / z 74 and 75. A check for any such interference can be made by the repeated analyses of water/ breath samples with a known standard deuterium abundance, e.g. normal breath (Davies et al 2001b). Then the presence of any constant background interference can be easily accounted for in the analysis, if required. The precision of the Q value obtained using equation [4.10] is described by the standard error in Q, which can be calculated by considering a Poisson distribution of the total numbers of ions as: AQ _ /~,~N,74,+dN(75) ~ N(74) N(75)
[4.11]
N(74) and N(75) represent the total numbers of ions counted at m / z - 74 and 75. The standard error in R1 calculated using equation [4.9] is then predicted as: tiN(75)) N(74) + N(75)
AR 1 - ~R3(~/N(74)
[4.12]
Validation of the SIFT-MS and FA-MS methods for deuterium analyses, carried out using standard mixtures, demonstrate that both accuracy and precision are typically 1% for headspace sampling when using the procedure described above (Spanel & Smith, 2000, 2001).
102
Chapter 4- P. Spanel & D. Smith
4.6 Concluding remarks The novel SIFT-MS and FA-MS techniques can be used for the on-line, real time analysis of the deuterium abundance in breath water vapour and in the headspace of aqueous liquids, including blood, urine and dialysate fluid. Of the two methods, we emphasise that FA-MS results in greater precision because of the larger ion count rates involved, although the microwave discharge that is exploited may introduce small fractions of impurity ions. Our experience shows that this is not a serious problem. In this regard, SIFT-MS can provide "cleaner" analyses, because the H30 + precursor ions are selectively introduced into the helium carrier gas. However, longer ion counting integration times need to be used to obtain an acceptable precision for the D / H analyses in SIFT-MS.
Handbook of Stable Isotope Analytical Techniques, Volume 1 P.A. de Groot (Editor) 9 2004 Elsevier B.V. All fights reserved.
CHAPTER 5 Natural Abundance 2H-NMR Spectroscopy. Application to Food Analysis S. Rezzil, C. Guilloul*, F. Renierol, V. M. Holland1 and S. Ghelli2 1 European Commission, Joint Research Centre, Institute for Health and Consumer Protection, Physical and Chemical Exposure Unit, 1-21020Ispra (VA),Italy 2 SPIN,via Tamagno, 3, 42048Rubiera (RE), Italy e-mail: *claude.guillouC~rc.it
5.1 Introduction
The determination of stable isotope ratios is of great interest in various fields of scientific research such as geochemistry, hydrology, medicine, biochemistry and food science. Two analytical techniques are mainly used for the measurement of the stable isotope content. These are Isotope Ratio Mass Spectrometry (IRMS) for 2H/1H, 180/ 160, 13C / 12C, 15N / 14N and 34S / 32S global ratios and deuterium Nuclear Magnetic Resonance (2H-NMR). 2H-NMR has been demonstrated to be very useful for site-specific characterisation of organic molecules, i.e. for the intramolecular distribution of deuterium. Despite its lack of sensitivity, which implies the use of relatively large sample sizes, 2H-NMR provides a "fingerprint" of the deuterium content that can be correlated with its natural or synthetic a n d / o r geographic origin. This Chapter is intended as a brief, introductory survey of the fundamentals and the use of 2H-NMR at natural abundance. It is not intended to be an exhaustive review of all applications of NMR used as an isotopic technique, including studies with labelled compounds. We briefly present the basic principles of NMR spectroscopy with specific considerations about the observation of deuterium at natural abundance from both theoretical and practical aspects. Some examples of applications in food science are also discussed. 5.2 General considerations on N M R spectroscopy 5.2.1 Basic principles
NMR spectroscopy is based on the properties of some nuclei that have a non-zero spin angular momentum L, i.e. a rotational motion on itself (Abragam, 1961). These nuclei can be regarded as small spinning bar magnets when the sample is placed in a magnetic field (Bo). Since nuclei possess an electric charge, the existence of a spin gives rise to a non-zero dipolar magnetic moment ~ collinear with L. Due to quantum rules, only few orientations of L, and consequently also of ~, are possible as expressed by the following equations:
104 +
Chapter 5 - S. Rezzi, C. Guillou, F. Reniero, V.M. Holland & S. Ghelli -)
, - 7L
[5.1]
Lz - mi(h/2~)
[5.2]
m I = I, I - 1, I - 2 , . . . , - I [5.31 where Lz is the projection of L on z axis parallel to the direction of the magnetic field B0. The values of both 7 (magnetogyric ratio) and I (spin quantum number) depend on the nucleus (see Table 5.1). Different atoms have therefore different values of ~. The term mI is the magnetic quantum number which characterises the Eigen-states or energy levels of the nucleus. The number of these energy levels (i.e. the number of possible orientations of L and ~) are equal to 2I + 1. The spin quantum number I is an integer or half an integer. In the simplest case, I is equal to 1 /2, as for 1H, 13C and 15N. In those cases, only two levels are possible in agreement with the quantum rules, which are commonly indicated as R (or 1 / 2 ) and 13(or _ 1 / 2 ) spins. R and 13spins correspond to clockwise and counter clockwise rotation of the nucleus (Figure 5.1). When a stable isotope, having ~ ~ 0 (i.e. I ~ 0) is placed in a magnetic field B0, magnetic interactions between B0 and g take place. The different possible orientations of L thus take different values of energy i.e., the interaction with B0 induces a non-equivalence of the orientations of L and the existence of a ground state and an excited state for the spin (Figure 5.1). This results in the ground spin state being more populated with respect to the excited spin state. Transitions between these two spin states can be achieved by a net absorption of energy AE given by Larmor equation [5.4] where m is called the Larmor frequency or resonance frequency: AE - hv - 7(h/27t)B o ~ m - 7B o
[5.4]
Larmor frequency depends thus on ~, and B0 which are relative to the kind of nucleus and the type of spectrometer, respectively. Since the Table 5.1 - Nuclear characteristics of some nuclei values of ~, are different for different isotopes, each of them Nucleus I a (%) 7 (rad T-1 s-l) S has a well-defined resonance frequency. Consequently, it is 1H 1/2 99.98 26.7510 x 107 1 possible to collect separated 2H 1 0.015 4.1064 x 107 9.65 x 10-3 NMR spectra for each isotope having I ~ 0. The value of m lies inside radio wave range, i.e. low energy if compared with other spectroscopies, such as infrared, visible and ultraviolet,
13C
1/2 1
15N
1/2 5/2
14N 170
99.63
1.108
6.7263 x 107 1.9331 x 107
1.59 x 10-2 1.01 x 10-3
0.37
-2.7116 x 107
1.04 x 10-3 2.91 x 10-2
0.037
-3.6264x 107
I 9nuclear spin number; a: natural abundance; 7: nuclear magnetogyric ratio; S: relative sensitivity.
Natural Abundance2H-NMRSpectroscopy.Applicationto Food Analysis
105
Figure 5.1 - Energy levels of the spin system with I = 1/2, in the absence and in the presence of a magnetic field Bo. The NMR signal intensity is proportional to the probability of transition from the ground state a to the excited state 13. This probability is proportional to the excess of spin a with respect to the exited state 13. The number of nuclei in the ground (Na) and excited states (N~) are governed by the Boltzmann relation [5.5], where k and T are the Boltzmann constant and the absolute temperature, respectively. Nc~ AE N~ = exPkT
[5.5]
Since AE is small, NR is only slightly larger than Nf~ and the resulting NMR signal is very low. The difference in spin populations between these two energy levels can be modified by applying a magnetic field B1, rotating at the resonance frequency of the observed nucleus in the plane perpendicular to Bo. After application of B1, the resulting excess of spin in the excited state returns to the ground state through relaxation mechanisms until reaching the initial Boltzmann equilibrium. During these relaxation mechanisms different effects are occurring and are registered under a time domain signal, the Free Induction Decay (FID). Following this principle of excitation/relaxation, the pulse-NMR technique allows the obtention of spectra acquired by summing many single spectra, i.e. by acquiring many scans, in order to increase the signal-tonoise ratio (S/N) which is proportional to the square root of the number of scans. Finally, a Fourier Transformation converts the FID into the final NMR spectrum in the frequency domain. Thus, all the effects coming from the interaction between B0 and g occurring during the relaxation mechanisms, longitudinal and transversal relaxation (also called spin-lattice and spin-spin relaxations, respectively), are observable in an NMR spectrum. The more important of these are the chemical shift, the scalar and dipolar couplings.
106
Chapter 5 - S. Rezzi, C. Guillou, F. Reniero, V.M. Holland & S. Ghelli
5.2.1.1 The chemical shift Following Lenz's law, the electronic environment near the nucleus induces a small magnetic field opposed to B0. This phenomenon gives rise to a magnetic shielding which makes the local magnetic field at the nucleus smaller than the applied field B0 by an amount equal to oB0, where o is known as the shielding or screening constant. Consequently, the effective magnetic field at the nucleus is equal to Beff- B0 - oB0 instead of B0 and thus, the Larmor equation becomes: AE - hv - ~,hB0(1 - o) ~ m - ~,B0(1 - o)
[5.6]
The frequency corresponding to the spin transition also depends on o whose magnitude is closely related to the chemical environment at the nucleus. The value of ~, is much greater than o. That means that ~, and o correspond to two effects very different in magnitude which can be seen by splitting the frequency m into two terms 011 -- ~'B0 and m2 - -~,Boo. This component 0)2, called chemical shift, is directly linked and is very sensitive to the chemical environment of the nuclei. The chemical shift (hi) of a nucleus measures the position of its resonance signal in the NMR spectrum. Its value is expressed in part per million (ppm) as the order of magnitude of m2 is about a million times smaller than ml. In practice, it is calculated by the following formula, where the frequency of the signal of reference is calibrated to zero: ~i --
V i - Vref v0
X
10
6
[5.7]
It is clear that nuclei possessing the same atomic environment, i.e., magnetically equivalent, resonate at the same chemical shift.
5.2.1.2 Scalar coupling Scalar coupling is due to the magnetic interactions that occur between the magnetic moments /a of nuclei connected by chemical bonds. This effect produces the splitting of the signal relative to a single nucleus into 2nI+1 lines, where n is the number of coupled nuclei and I the corresponding spin number. The frequency difference between two successive lines of the splitting gives the coupling constant nJxy, where n is the number of bonds which separate the coupled nuclei X and Y. Its magnitude varies from a few hundred Hz for 1Jxy to a few Hertz for nJxy with n greater than 3. In practice, broad-band decoupling techniques are often used to remove the coupling with protons in order to simplify the spectra.
5.2.1.3 Dipolar coupling Dipolar coupling arises from the magnetic interactions between g through space, i.e. without the necessity of the presence of a chemical bond. This coupling results in a transfer of magnetisation called Nuclear Overhauser Enhancement (NOE). This effect is typically observed in 13C NMR spectra acquired in proton decoupling mode.
Natural Abundance 2H-NMR Spectroscopy. Application to Food Analysis
107
5.2.1.4 The signal intensity It is finally fundamental to underline that, under appropriate conditions of acquisition, the area of the signals observed in a NMR spectrum is directly proportional to the number of nuclei in resonance. 5.2.2 The NMR instrumentation The NMR spectrometer is composed of five main parts: the magnet, the shim system, the probe, the radiofrequency unit and the computer (Figure 5.2).
The magnet has a cylindrical shape with a bore in the centre which receives the probe where the sample tube is placed. A strong permanent electric current circulating in a superconducting coil placed in liquid helium produces the magnetic field B0. The shim system is made of 16-28 room temperature coils, positioned inside the bore between the magnet and the probe. The electric currents allowed to circulate in these room temperature coils, usually known as shims, produce the compensation magnetic fields which are used to improve the homogeneity of the magnetic field B0. In practice, the operators adjust up to 6 to 8 of the more sensitive parameters of these shims, generally using the lock signal intensity, to optimise the resolution of the NMR spectrum. The probe contains a radiofrequency circuit behaving as a radio antenna. Its main component is the coil where the sample is placed for measurement. The coil is used to transfer radiofrequency to the sample to excite the spin systems and also to detect the NMR signal. Capacitors and inductances complete the circuits and make the antenna tuneable on different frequencies. The radiofrequency unit contains all the devices for radiofrequency generation (synthesisers), amplification (linear amplifiers), detection (receiver) and digitalisation (ADC). It also contains the circuit for the "lock system" which is necessary to ensure the stability of the magnetic field during the experiments. This circuit performs a continuous monitoring of a reference frequency of a nucleus different from the observed one. This is achieved through the monitoring of the resonance frequency of either a deuterated or fluorinated substance. Solvents highly enriched Figure 5.2- The NMR spectrometer.
108
Chapter 5 - S. Rezzi, C. Guillou, F. Reniero, V.M. Holland & S. Ghelli
in deuterium or containing fluorine are usually used for this purpose. The spectrometer can be equipped with an automatic sample changer and automation programs enabling the execution of field frequency locking and field homogeneity adjustment (automatic shimming) in order to measure a series of samples. Finally, a computer system manages all the devices and a host computer operates as an interface for the user to set-up the instrumental conditions and for data processing. 5.3 Deuterium NMR spectroscopy (2H-NMR)
5.3.1 The intramolecular distribution of deuterium The total deuterium content of natural compounds depends on several physical, chemical and/or physiological effects leading to isotope fractionation. These isotope effects can be associated with pedoclimatic parameters, the habitat of the producing plant, the geographic latitude of origin, the deuterium content of local rainwater, the amount of rainfall, the evapotranspiration and also with biosynthetic or technological processes. The deuterium content of a product is therefore an interesting probe for origin and/or authenticity proof. The joint structural and quantitative dimensions of 2H-NMR allows one to observe large variations in deuterium content between the different sites of a given molecule with respect to the expected statistical distribution (Martin & Martin, 1981). Hence, 2H-NMR provides a quantitative tool for establishing the deuterium fingerprint which can be used as a probe of the chemical, biochemical and technological history of the product (Martin et al., 1982a; Martin & Martin, 1990). Due to the low natural abundance of deuterium, only monodeuterated molecules are occuring. In order to remove the line splittings arising from the scalar couplings between deuterium and proton, the deuterium spectra are always acquired with 1H decoupling mode. As a consequence, each signal, assigned to each magnetically nonequivalent site, corresponds to one monodeuterated species called an isotopomer. The 2H-NMR spectrum of the three monodeuterated isotopomers of ethanol (the corresponding sites I, II and III) are presented Figure 5.3. Obviously, the assignment of the deuterium NMR signals to the monodeuterated isotopomers is a prerequisite to the interpretation of the isotopic fingerprints. 1HNMR and if necessary, multidimensional NMR experiments are carried out to ensure the correct attribution of proton and thus deuterium signals.
5.3.2 Spectroscopic particularities of the deuterium nucleus 5.3.2.1 The sensitivity The observation of 2H resonances at natural abundance suffers from sensitivity limitations due to both the intrinsic low receptivity of the nucleus and its low natural abundance. The resulting low sensitivity (S) is also affected by the strength of the instrumental magnetic field (B0), the nuclear magnetogyric ratio (~,)which is a characteristic of the nucleus, the isotopic abundance (a) and on the concentration of the sam-
Natural Abundance 2H-NMR Spectroscopy. Application to Food Analysis
109
Figure 5.3 - Natural abundance 2H-NMR spectrum of ethanol.
ple. Table 5.1 gives the characteristics of some nuclei and their relative sensitivity to NMR. As indicated in this Table 5.1, the sensitivity is given relative to that of proton, which is the most sensitive nucleus in NMR spectroscopy. The sensitivity of deuterium is around 9.65 x 10-3 (Table 5.1). So, the observation of deuterium is approximately 1.5 x 106 times more difficult than that of proton. This sensitivity limitation is partly overcome by the use of high field spectrometers and of pulsed NMR techniques, which allow the accumulation of a large number of scans. In addition, specific deuterium probes are designed to allow the measurement of a relatively large sample size. Historically NMR probes of 15mm were even used for the acquisition of deuterium spectra at natural abundance. However, the use of large diameter probes could lead to several technical and practical limitations (e.g. difficulties in ensuring field homogeneity, efficiency of decoupling power...). Nowadays, 10 mm probes are actually the better compromise between the gain in sensitivity and these technical limitations. As a result the actual sample to be measured in 2H-NMR must fill at least the active volume of the measuring cell (typically about 2 ml). That means that when analysing a sample available only in small quantities (e.g. extracted flavour compounds) one has to consider the use of an appropriate solvent for the preparation of the sample. 5.3.2.2 Resolution
It must be emphasized that, expressed in frequency units, the chemical shift discrimination observed in the deuterium spectrum is 6.5 times lower than that of the proton NMR spectrum acquired in the same magnetic field. At a proton nominal frequency of 400 MHz, for example, deuterium NMR has therefore a resolution power somewhat analogous to that of 1H NMR at 60 MHz. The analyst must consider that the resolution achievable for a 2H-NMR spectrum is strictly correlated to intensity and
110
Chapter 5 - S. Rezzi, C. Guillou, F. Reniero, V.M. Holland & S. Ghelli
homogeneity of the magnetic field of the NMR spectrometer. In practice, for a given instrument, i.e. a given magnetic field, one can improve the resolution by optimising the magnetic field homogeneity (shimming) and by decreasing the viscosity of the sample (temperature, concentration, choice of the solvent).
5.3.2.3 Relaxation Since its spin number is different from 1 / 2 (I=1), deuterium has an electric quadrupolar moment, which arises from the non-spherical repartition of nuclear charges. Thus in the case of deuterium, the quadrupolar mechanism of relaxation is widely predominant. A consequence of this is the limitation of spin-spin dipolar interactions and therefore the loss of Nuclear Overhauser Enhancement. Hence the enhancement of the signals occurring in proton decoupling mode is not observed for deuterium. This property is highly beneficial for quantitative analysis of 2H NMR spectra as timeconsuming gated decoupling sequences are not necessary for the acquisition of quantitative spectra. 5.3.3 Quantitative 2H-NMR spectroscopy In the 2H-NMR spectrum, the signal intensities are directly proportional to the number of moles of deuterium nuclei in resonance. Consequently, the quantitative measurement on each site, i.e. of each isotopomer, can be carried out in order to determine the site-specific isotope ratios, which we shall now define.
5.3.3.1 Definition of the site-specific isotope ratios
Isotope ratios are commonly defined as the ratio of the number of heavy to light isotopes in a chemical species, 2H/1H or D / H in the present case. The NMR method allows the measurement of absolute values of the site-specific isotope ratios (D/H)i defined a s : ( D / H ) i - Di/(PiN H)
[5.8]
where Di is the number of deuterium atoms in site i, NH the number of fully protonated molecules, and Pi the number of equivalent positions of i. The stoichiometric number of hydrogen (P) and the number of atoms of deuterium (D) in a molecule are given respectively by: P - ~ Pi
[5.91
t=l n
D -
~ Di t=l
[5.10]
111
Natural Abundance 2H-NMR Spectroscopy. Application to Food Analysis
The real fi and statistical Fi molar fractions are thus defined as: [5.11] [5.12]
f i - Di/D F i = Pi/P Then, the site-specific isotope ratios (D/H)i can also be expressed as follows: (D/P)i-
(fi/Fi)(D/H)global
[5.13]
where (D/H)global represents the overall deuterium isotope ratio of the molecule. The relative intramolecular repartition Ri/j of deuterium between two sites i and j of the molecule has been previously defined (Martin & Martin, 1981)" Ri/j - PjDi/D j - P i ( D / H ) i / ( D / H ) j
[5.14]
Absolute values of site specific isotope ratios are measured by adding a precisely known amount (mws) of a working standard (ws) to the NMR tube containing a mass (ma) of sample A. The deuterium site specific isotope ratio of isotopomer i (D/H)i, expressed in parts per million (ppm), is calculated from: (D/H)i-
(D/H)ws(PwsmwsMASi)/(PAmAMwsSws)
[5.15]
where P, M and S are the number of hydrogen atoms, the molecular mass and the intensity of the deuterium NMR signal of the working standard (ws) and of the sample (A), respectively. The working standard generally used is tetramethylurea (TMU) calibrated against the V-SMOW (Vienna-Standard Mean Ocean Water) whose 2H/1H ratio has been determined (155.76 x 10-6) (IAEA Tecdoc-825 and references herein). Certified TMU (Guillou & Martin, 1993) is available from the Institute of Reference Materials and Measurements (IRMM) in Geel (Belgium). As discussed later on, for some applications other working standards are sometimes used. In such cases careful calibration against certified TMU must be performed. The results can also be expressed in the relative scale 6D%o usually adopted by mass spectroscopists" AD~ %o - 1000((D/H) i - 155.76)/155.76
[5.16]
5.3.3.2 Precision and accuracy
NMR is a very attractive quantitative method because of the direct proportionality of signal intensity to the number of nuclei in resonance. However, as discussed elsewhere in this paper, the precision of the method depends strongly on the signal-tonoise ratio, and other parameters such as phase adjustment, base line artefacts and resolution.
112
Chapter 5 - S. Rezzi, C. Guillou, F. Reniero, V.M. Holland & S. Ghelli
In 1988, the results of a collaborative study establishing the repeatability and reproducibility of the NMR measurement of the deuterium isotope ratios of ethanol were published (Guillou et al., 1988). This work, organized by the Community Bureau of References (BCR) in Brussels, involved sixteen different spectrometers (4.7-9.4T) installed in fifteen European laboratories. Three ethanol samples of different botanical origin (sugar-beet, grape and maize) were selected in order to check a relatively wide range of natural isotope contents. Quantitative determinations were carried out by measuring the intensity ratio (Ti) between signals associated with the methyl, methylene and hydroxyl sites of ethanol and the signal of the tetramethylurea used as internal reference. The internal repartition parameter R ( R - 2(D/H)II/(D/H)I, where II and I stand for methylene and methyl, repectively) was also measured. The results of this collaborative study demonstrated that a good precision can be obtained using quantitative 2H-NMR and that it is possible to measure isotope ratios with confidence intervals of 0.25%, a repeatability around 1% and an acceptable reproducibility (2-3%). From a practical point of view, the precision reached in routine analysis of individual samples in a given laboratory is high enough to detect, for example, the addition of less than 5% v / v of exogenous ethanol in a beverage. However, for comparing results from different laboratories and for the standardization of the results, the use of reference materials (ethanols in the present case) of a certified site-specific ratio is necessary. For that purpose, sealed NMR tubes containing ethanols of three different botanical origins mixed with TMU were established as Certified Reference Materials (CRM) (Martin et al., 1994). These CRM are also available from the IRMM in Geel. Recently, a number of new CRM suitable for isotopic analysis of food products have been prepared within the frame of the European project SMT3-CT96-2086 "Establishing field reference materials for the authentication of food and beverages by isotopic analyses" funded by DG Research of the European Commission. Two of these new CRM (BCR656 ethanol from wine 96% vol. and BCR660 hydro alcoholic solution 12% vol.) are specifically intended for use in 2H-NMR and have also been available since the beginning of 2002 from the IRMM. The 2H-NMR of ethanol is a very favourable case and it is also interesting to study the precision and accuracy of the measurement for other compounds. This was done by Martin & Naulet (1988) using several kinds of molecular species of different natural origin (water, ethanol, anethole and vanillin). They compared IRMS and 2H-NMR measurements and found a good agreement for the results of these two techniques for water. The results for the organic compounds were also relatively consistent for these two techniques although systematic deviations of a few ppm had been detected for vanillin. 5.4 Practical aspects of 2H-NMR
5.4.1 Preparation of the sample Obtaining accurate values for isotope ratio measurements involves precise experimental procedures. As discussed above, 2H-NMR spectroscopy requires a large quantity of sample. The first question that one should address is the quantity of material available for analysis. In practice, the compound of interest often needs to be extracted from a complex matrix. Usual physico-chemical treatment of the starting
Natural Abundance 2H-NMR Spectroscopy. Application to Food Analysis
113
materials (e.g. solvent extraction, distillation, precipitation...) and also sometimes chromatographic techniques are required to obtain, and purify, a sufficient amount of sample. These extraction and purification steps must not cause any significant isotope fractionation effects (Moussa et al., 1990). The test sample measured by NMR is prepared by mixing precise amounts of the analyte and the working standard. The precision of the weighing in the preparation of the NMR tube is critical as is illustrated below in the case of the analysis of ethanol. Sometimes, due to problems of solubility or to avoid contamination of the analyte, the working standard can also be introduced in a coaxial tube. In equation [5.15], the site specific parameters (D/H)i depend on the ratio of the number of moles of the analyte to that of the internal reference. The purity of both the compounds, which must be determined by an appropriate method, has therefore to be taken into account. Alternatively the molar ratios of analyte vs. internal reference can also be obtained by quantitative 1H-NMR of the same mixture prepared for the 2H-NMR. As illustrated in the case of ethanol, quantitative 1H-NMR spectra, acquired through the proton decoupling coil of the selective deuterium probe, gave results completely comparable to those derived by the conventional procedure (KarlFischer titration for determination of the alcoholic grade) (Fauhl & Wittkowski, 1996). One could think of using this 1H-NMR approach as a general tool for the determination of molar ratios required in equation [5.15]. However, conditions for the acquisition (pulse length, relaxation delay) of 1H-NMR spectra must be optimised in order to ensure quantitative measurements. In particular, this includes knowledge about the spin-lattice relaxation times (T1) of the proton signals, which is fundamental for determining the relaxation delay. On the other hand, the parameters describing the internal distribution of deuterium (i.e. the Ri/j parameters and the molar fractions fi) do not depend on the purity of the product, at least where no signals arising from the impurities, overlap with the peaks of the molecule under study. This remark becomes particularly interesting in the case of products containing residual impurities which are difficult to remove or quantify. In 2H-NMR, the use of a 19F locking device is desirable in order to avoid field drift which can result in the broadening of the signals. Indeed, fluorine nuclei are used to ensure the field-frequency locking. For that purpose, a lock substance such as hexafluorobenzene is generally added in the NMR tube. The use of a solvent is often required for the preparation of the NMR tube. In general the solvent cannot be used as a working standard since its peak intensity is often not comparable to those of the analyte. The choice of the solvent should take into account its solubilization properties for all components of the mixture (i.e. analyte, working standard, lock substance) and its deuterium signals, which must not interfere with the peaks of interest of the analyte. In principle, a higher concentration of the sample leads to an increase in the signal to noise ratio. However, a high concentration of the sample can also increase the viscosity and thus be detrimental to the actual resolution of the spectrum. The choice of the solvent can also be guided by the possible effect on the chemical shift and the shape of the NMR signals. The best compromise
114
Chapter 5 - S. Rezzi, C. Guillou, F. Reniero, V.M. Holland & S. Ghelli
has thus to be experimentally established as illustrated later in the case of raspberry ketone. Finally, in order to obtain the best possible resolution and reproducibility of the measurements, it is necessary to use high quality NMR tubes and avoid the presence of insoluble particles in the solution placed in the NMR tube. 5.4.2 2H-NMR data acquisition The NMR spectrometer must be equipped with a probe tuned to the characteristic resonance frequency of deuterium for the corresponding field. The specific deuterium probes, which are generally used, have a proton decoupling channel and also a fieldfrequency stabilization channel tuned at the fluorine frequency (lock). Broad-band 1H decoupling is applied to simplify the 2H spectrum by removing the splitting of the signals due to the coupling between deuterium and proton. This is achieved by the strong irradiation of the protons at the appropriate radio frequency, which also leads to the heating of the sample. Nowadays, by exploiting the pulse techniques, low power decoupling techniques, such as the WALTZ pulse sequence, permits the acquisition of decoupled spectra with less generation of heat in the sample. This facilitates the control and the stability of the temperature of the sample and therefore contributes to an improvement in the reproducibility of the results. The stability of temperature of the sample during the whole experiment is critical to obtain satisfactory resolution. Thus, the temperature of the sample is generally adjusted slightly above room temperature. However, heating of the sample to a higher temperature may sometimes be desired in order to improve the resolution of overlapping NMR signals (Hermann, 1999). Due to the small frequency range of the deuterium spectra, a frequency window of 1200Hz at 61.4MHz and a 16K memory size, corresponding to a 6.8s acquisition time, are sufficient to ensure the good resolution of all signals. In the case of ethanol for example, the resolution measured on the spectrum, expressed as the half-width of the methyl and methylene signals of ethanol and the methyl signal of TMU, must be less than 0.5Hz (Commission regulation (EEC), No. 2676/90, 1990). The resolution of the spectrum can be checked by transforming the FID without exponential multiplication (LB - 0) and by measuring the line width at half-height. The length of the acquisition time must be equal to 3 or 5 times the longest deuterium longitudinal relaxation time, to ensure complete relaxation, i.e. a complete return of the exited nuclei to the Boltzmann equilibrium between successive pulses. To obtain more precise information on the deuterium relaxation times, several techniques can be used for their measurement (Guillou, 1986). One such technique is the classical Inverse Recovery Fourier Transformation pulse sequence. In quantitative 2HNMR, a pulse length (typically 10 to 20 x 10-6s) corresponding to a 90-degrees flip angle is conventionally applied in order to ensure a maximum steady-state-magnetization. The number of scans is then determined to achieve a good signal-to-noise ratio (S/N). For example, for a sample of 95% ethanol, investigated at 61.4MHz, in a 10mm tube, a S/N ratio of around 140 can be reached after 240 scans. Since the experiment must be repeated up to 10 times if very precise results are desired, 4.5 hours experimental time is necessary with the above conditions. The total experimental time
Natural Abundance 2H-NMR Spectroscopy. Application to Food Analysis
115
depends obviously on the amount of product in the tube and on the required degree of precision, i.e. the number of experiments. So, by increasing the concentration of the product, the total experimental time can be drastically reduced. The analyst should find the optimum compromise between the concentration of the product, which beyond a certain level diminishes the resolution of the signal, and the number of scans necessary to obtain a S/N ratio which produces a quantitative measurement. 5.4.3 Data processing Modern NMR spectrometry is based on radio frequency pulsed techniques and on the sampling of the signal of free induction decay (FID) by digital conversion. This time domain data is submitted to Fourier transformation to obtain the digital NMR spectrum in the frequency domain. Several mathematical window functions can be applied to the FID prior to the Fourier transformation in order to enhance either the signal to noise ratio or the resolution. In the case of deuterium, an exponential multiplication of the free induction decay associated with a line broadening (LB) of 1-2 Hz is usually applied. The resulting increase in the signal-to-noise ratio is beneficial for the determination of peak heights and areas. On the other hand, the line broadening resulting from this window function, is detrimental to the resolution and increases the overlap between signals with similar chemical shifts. The signal intensities and heights are determined, using the calculation routines of the NMR instrument, after baseline and phase corrections of the spectrum. Automatic baseline and phase corrections are available in the NMR software but they may not be fully reliable depending on the quality and complexity of the spectrum. It must be emphasized that the correction of the phase may have dramatic effects on the integration results and this is known as one of the major problems for quantitative NMR. Hence manual corrections performed by the NMR spectroscopist are still often preferred although it remains somehow 'subjective'. Sharp and well separated peaks such as those of the ethanol deuterium spectrum, allow the adjustment of the first and second order phase using visual peak symmetry criteria. Indeed, in that situation good reproducibility of results is achieved for peak intensities and peak heights. Manual phase correction becomes more difficult when considering broader lineshapes and overlapping peaks. Moreover, in most cases the linewidths are different from one peak to another and it is therefore necessary to measure the peak areas to compute the (D/H) ratios. The use of algorithms which perform the optimal definition of the NMR signals is therefore desirable. Several iterative algorithms based on the theory of the NMR signal have been proposed for computing the peak intensity, chemical shift, linewidth and the phase. Approaches optimised for the particular case of natural abundance deuterium spectra have been investigated (Martin, 1994; Cremonini et al., 1998).
5.5 Examples of application The precise control of food products is becoming increasingly important since the difference in price between labelled food products (i.e. with a defined geographical origin and/or botanical source) and non-labelled ones is substantial. 2H-NMR has proven to be reliable in this respect as well as in the implementation of legislation regarding food products throughout the food industry.
116
Chapter 5 - S. Rezzi, C. Guillou, F. Reniero, V.M. Holland & S. Ghelli
5.5.1 Ethanol
2H-NMR spectroscopy has been applied to the discrimination between the various botanical and synthetic origins of ethanol (Martin et al., 1982b, 1983a). Sugars of different botanical origin are characterized by their specific deuterium content. The deuterium content of the non-exchangeable sites of a sugar is correlated to the plant species which is affected by different photosynthetic pathways and by the geoclimatic conditions during biosynthesis. For example, beet root sugar shows a lower deuterium content than cane sugar. It has been demonstrated that the deuterium content of a sugar is partly transferred to the ethanol produced by its fermentation (Martin et al., 1986a, 1991a, b). During the fermentation process two hydrogen atoms of sugar are transferred to the methyl group of the ethanol. The other carbon-bound hydrogen atoms of ethanol are introduced from the water of the fermentation medium. The hydroxyl group exchanges with the water and has no important meaning for practical use.
Due to its molecular properties ethanol is much more accessible to 2H-NMR than sugar and, moreover, it can be easily extracted by distillation from wine or other alcoholic beverages like spirit drinks (Martin et al., 1983a), beers (Martin et al., 1985a; Franconi et al., 1989), etc. The determination of the addition of sugar to sugar-containing products like musts (Commission regulation (EEC) No 2676/90, 1990; Monetti et al., 1996), fruit juices (Martin et al., 1991c, 1996a,b; Pupin et al., 1998), honeys (Linder et al., 1996; Giraudon et al., 2000), maple syrup (Martin et al., 1996c), etc.., can also be performed on the ethanol obtained after controlled fermentation. All these applications require several preparation steps before 2H-NMR (e.g. fermentation, distillation...). These need to be carefully monitored to prevent any deuterium fractionation artefacts, which could affect the final result. In this respect the end of fermentation is determined by testing for the presence of residual sugars. The yield of the distillation must be established by the weighing of the starting material and of the distillation products and by the determination of their alcoholic grade. Also, the absence of evaporation of alcohol during the preparation of the NMR tube, before the measurement, needs to be checked (Guillou, 1991). In the wine sector, isotopic techniques have attracted much attention for their capacity to check the compliance of wine products with national and European regulations, for instance, as regards enrichment with exogenous sugars and for the detection of fraud and illegal practices (false declaration of origin and watering). In 1990, the E.C. Regulation N ~ 2676/90 officially adopted the Nuclear Magnetic Resonance (NMR) method for the control of enrichment of wine with exogenous sugars (chaptalisation) which is carried out to increase the alcoholic grade of wine (Commission regulation (EEC) No 2676/90, 1990; Martin et al., 1986b). For example, an addition of 17g/L of sugars is equivalent to a 1% increase in the alcoholic grade. Several publications reported the correlation between the isotopic ratio of ethanol in wines and their geographical origin (Martin et al., 1988; Monetti et al., 1994). Correlations have been found between geoclimatic parameters (latitude, longitude, rainfall and temperature) and the isotope ratios. The application of 2H-NMR as a method for the control of the wine market needs to take into account this natural variability. To this aim databases
Natural Abundance2H-NMRSpectroscopy. Applicationto Food Analysis
117
of the isotopic ratios of authentic wines are established, on a yearly basis, in order to provide reference data for the different vintages of the wine-producing regions (Commission regulation (EEC) No 2729/2000, 2000; Gim6nez-Miralles et al., 1999; Kosir et al., 2001). 5.5.2
Raspberry ketone
Raspberry ketone is the compound mainly responsible for the raspberry aroma which is widely used in the food industry. Due to the fact that the amount of this compound found in the raspberry fruit itself is very low, an alternative source was necessary to meet demand. 2H-NMR provides a means of distinguishing between <
Figure 5.4 - Methods of biogeneration and of chemical synthesis of raspberry ketone.
118
Chapter 5 - S. Rezzi, C. Guillou, F. Reniero, V.M. Holland & S. Ghelli
Figure 5.5 - 2H NMR spectrum of raspberry ketone in different solvent systems (above: dioxane, below: benzene / dioxane).
ing it very suitable for use as a working standard. The results presented in Table 5.2 show the clear discrimination between the natural products obtained using biological methods with natural precursors, and those obtained using synthetic precursors. The deuterium pattern of (D/H)2 and (D/H)3 on the aromatic ring is particularly informative. Because of the higher dispersion of (D/ H)4 and (D/H)5 values for samples of both natural and synthetic origin, the deuterium pattern of the aliphatic chain is less informative whereas the deuterium content of the methyl site (D/H)6 is very similar for all origins.
5.5.3 Phenylacetic acid/benzoic acid
The flavour compounds originating from L-phenylalanine are of particular interest in the food industry. Transamination or oxidative deamination and decarboxylation of phenylalanine provides phenylacetaldehyde, leading, in turn, by reduction, to phenylethanol and, by oxidation, to phenylacetic acid (Figure 5.6). These products, as well as the esters formed by phenylethanol and phenylacetic acid with a great variety of acids and alcohols, substantially contribute to the aroma of many foods and are used in large quantities in aromatic formulations. Phenylethanol is directly accessible by extraction, i.e. from rose oil and certain fusel oils, but no convenient extractive sources of phenylacetic acid seem available. In the case of phenylacetic acid the NMR resonances of the nuclei of the benzene ring are unresolved and only the total (D/H) ratio of all aromatic positions is there-
119
Natural Abundance 2H-NMR Spectroscopy. Application to Food Analysis Table 5.2 - Mean Values of (D/H)i and fi of raspberry ketonea (D / H)2
(D / H)3
(D / H)4
(D / H)5
mean SD
135.4 (11.1)
170.0 (14.4)
121.3 (41.2)
mean SD
128.3 (11.8)
119.9 (8.7)
155.5 (29.1)
119.1 (6.5)
f2
f3
f4
f5
mean SD
0.192 (0.029)
0.243 (0.048)
0.170 (0.049)
mean SD
0.190 (0.019)
0.185 (0.030)
0.210 (0.042)
(D / H)6
(D / H)3 / (D / H)2
(D / H)5 / (D / H)4
Natural samples (six samples) 114.3 (31.0)
125.5 (2.4)
1.26 (0.09)
0.96 (0.13)
124.8 (5.0)
0.94 (0.04)
0.79 (0.14)
f6
f3 / f2
f5 / f4
1.26 (0.08)
0.94 (0.14)
0.97 (0.09)
0.82 (0.09)
Synthetic samples (nine samples)
Natural samples (six samples) 0.156 (0.030)
0.240 (0.014)
Synthetic samples (nine samples) 0.167 (0.017)
0.248 (0.014)
a The numeration i of the sites of raspberry ketone is presented in Figure 5.1 and corresponds to the decreasing chemical shift of the observed peaks in the 2H-NMR spectrum, f2-6 are the real molar fractions as define in equation [5.11].
fore accessible. Moreover this isotope ratio and that relative to the benzylic methylene group adjacent to the carboxyl group show a great variability for both natural and synthetic sources and this information therefore does not help in the determination of the origin of phenylacetic acid. However, when phenylacetic acid is converted into
Figure 5.6 - Biogeneration of phenylacetic acid and phenylethanol from L-phenylalanine.
120
Chapter 5 - S. Rezzi, C. Guillou, F. Reniero, V.M. Holland & S. Ghelli
Figure 5.7 - Conversion of phenylacetic acid into benzoic acid.
benzoic acid (Figure 5.7), a great dispersion of the signals of the aromatic deuterium nuclei is achieved (Figure 5.8) (Aleu et al., 2002). The synthetic materials showed a uniform distribution of deuterium in the three different positions (o, m and p for the ortho, meta and para positions, respectively) of the aromatic ring, with ratios (D/H) fo/(fm+fp) of 0.672, close to the statistical value ( 2 / 3 = 0.666). On the contrary, samples from natural origin showed a ratio with higher or lower values with respect to the statistical distribution. Interestingly~ the data of the benzoic acids from natural precursors, for which the complete quantitative analysis of the deuterium NMR spectrum was obtained, show for the three spectroscopically different positions the following (D/H) enrichment order: para > ortho > meta.
Table 5.3 - Various applications of 2H-NMR. Chemical
Reference
Acetic acid
Vallet et al. (1988), Remaud et al. (1992), Martin et al. (1993a), Hermann (2001) Gonzalez et al. (1998) Barbeni et al. (1997) Fronza et al. (1993) Fronza et al. (1995) Fronza et al. (1995) Hagedorn (1992), Martin et al. (1993a), Remaud et al. (1997a) Martin et al. (1983b, 1993a) Martin et al. (1983b, 1985b, 1993a) Toulemonde et al (1983), Maubert et al. (1988), Martin et al. (1993a), Remaud et al. (1997b) Remaud et al. (1997b) Remaud et al. (1997c) Martin et al. (1986c) Hanneguelle et al. (1992) Hanneguelle et al. (1992), Martin et al. (1993b) Carle et al. (1992) Deiana et al. (2001) Zhang et al. (1998a) Lai et al. (1995), Quemerais et al. (1995), Aursand et al. (1997, 2000) Vallet et al. (1991), Martin et al. (1992) Martin et al. (1992), Williams et al. (1993), Zhang et al. (1994a,b) Danho et al. (1992) Jamin et al. (1997) Hays et al. (2000)
Citric acid 3-Z-Hexenol 6-Decanolide 2-Phenylethanol 2-Phenylacetate Benzaldehyde Estragoles Anetholes Vanillin p-Hydroxybenzaldehyde Allyl isothiocyanate R-pinenes Linalyl acetate Linalool Bisabolol Squalene Glycerol Fatty acids, triacylglycerols Amino acids Carbohydrates Caffeine Nicotine Cocaine, heroin
Natural Abundance 2H-NMR Spectroscopy. Application to Food Analysis
5.5.4 2H-NMR applications to various molecules 2H-NMR spectroscopy has been successfully used to provide an isotopic fingerprint characterizing the origin (botanical, semisynthetic or synthetic ) of a wide range of compounds belonging to various molecular families. Table 5.3 reports various examples, apart from these reported above, most of which coming from natural sources.
121
Figure 5.8 - 2H NMR spectrum of natural benzoic acid.
5.6 General conclusions
The existence of the non random intramolecular distribution of deuterium has allowed many research possibilities for the study of this isotope in food science. The necessary requirements of this technique remain that of a relatively important and costly apparatus, as well as a computerized environment to facilitate data treatment as well as the automation of the analytical procedures. Despite the need of a large amount of sample in comparison with other isotopic techniques, 2H-NMR yields, under optimised experimental conditions, an amount of unique information, such as the direct quantitative determination of the intramolecular content of deuterium in each site of a given molecule. The resulting information has proven to be a very interesting data set for quality assessment of food products with respect to their natural, botanical and/or geographical origin. Further developements in 2H-NMR at natural abundance are anticipated particularly in the study of mixtures (Lai et al., 1995; Quemerais et al., 1995; Aursand et al., 1997, 2000) without the isolation of individual compounds.
Handbook of Stable Isotope Analytical Techniques, Volume 1 P.A. de Groot (Editor) 9 2004 Elsevier B.V. All fights reserved.
CHAPTER 6 Mass Spectrometric Techniques for the Determination of Lithium Isotopic Composition in Geological Material Lui-Heung Chan Department of Geology & Geophysics, Louisiana State University, Baton Rouge, LA 70803-4101, U.S.A. e-mail:
[email protected]
6.1 Introduction The element lithium has two stable isotopes: 6Li and 7Li. Their atomic abundances in natural sources are approximately 7.59 and 92.41%, respectively (Qi et al, 1997). The cosmic abundance of lithium isotopes reflects primordial nucleosynthesis, galactic cosmic-ray spallation, and destruction processes (Olive & Schramm, 1992). On Earth, the two isotopes are susceptible to separation in geological processes, due to their relatively large difference in mass. Lithium isotopic compositions of natural material therefore have important geochemical and cosmochemical implications. Lithium isotopes are also of interest to nuclear and biomedical sciences. One nuclear application is the use of 6Li as shielding material in nuclear reactors. In medical science, lithium is a therapeutic drug for manic depression. Despite considerable interest in lithium isotopes, applications in earth science and other disciplines have been seriously hampered by the intrinsic difficulty of precise measurement of low-mass isotopes. The relative mass difference between the two isotopes of lithium is about 16%, which is among the highest of thermally ionized elements. Isotopic fractionation during mass spectrometric and other instrumental analyses is severe and, in the absence of a third isotope, the effect of mass discrimination cannot be internally corrected. For several decades, many investigators have studied the isotopic composition of lithium in an effort to determine the natural variation of its isotopic abundance. The results of research prior to the early 1980's display a wide range of isotopic ratios in geological samples and meteorites (see reviews of Heier & Billings, 1972 and Chan, 1987). In addition to natural fractionation, this large measured variation is also attributable to isotopic separation during chemical preparation and instrumental analysis. In the last two decades considerable progress has been made to minimize mass fractionation in mass spectrometric analysis. As a result, the isotopic compositions of the major lithium reservoirs are becoming known. This chapter is devoted to the
Mass Spectrometric Techniques for the Determination of Lithium Isotopic ...
123
review of recent advances in mass spectrometric techniques for lithium isotope measurements. The intent is not to provide an exhaustive coverage of all recent developments, but to describe chemical and instrumental procedures that have proven useful for lithium isotope analysis in geological material. The emphasis will be on thermal ionization mass spectrometry (TIMS) and the relatively new method of inductively coupled plasma mass spectrometry (ICP-MS). The merits and shortcomings of the various methods for geological applications will be evaluated. Finally, the state of knowledge of the terrestrial variation of lithium isotope composition and examples of geological applications will be presented. Ion microprobe technique will be briefly mentioned here but will be discussed in detail in Part 1, Chapter 30. 6.2 Historical aspects Traditionally the isotopic analyses of lithium are performed on the monovalent ions (6Li+ and 7Li+) in a mass spectrometer. The ions are produced by thermal ionization of an ion source consisting of a lithium compound such as chloride, iodide, sulfate, and phosphate (Shima & Honda, 1963; Svec & Anderson, 1965; Dews, 1966; Michiels & De Bi6vre, 1983). Alternatively the ions can be derived from a sputtering ion source (Gradsztajn et al., 1967). Pronounced isotopic fractionation effect in evaporation and ionization processes is present in all of these methods (Brewer, 1936; Hutchinson, 1954). In addition, instrumental factors, such as accelerating voltage and optical properties, can also affect ion current ratios of light mass ions (Hutchinson, 1954; Gradsztajn & Guez, 1969). As a result, the isotopic ratio may change as a function of the fraction of lithium evaporated during the course of a run or may exhibit constant but irreproducible values. In an effort to combat these problems, investigators have used synthetic isotope standards to calibrate the magnitude of fractionation and have performed analysis under strictly controlled instrument conditions (Flesch et al., 1973; Michiels & De Bi6vre, 1983). Heavier molecular ions were also used to reduce the percentile difference in mass and hence mass discrimination effect. For example, Cameron (1955) made measurements on 6Li7LiI+ and 7Li7LiI+ at masses 140 and 141 produced by electron bombardment of vapor sublimed from lithium iodide. The mass spectrometric determinations in these earlier studies yielded 7Li/6Li ratios varying from 10 to 12.7 for terrestrial and meteoritic samples. The large variation may be mostly due to analytical artifacts.
Ion microprobe employs a finely focused primary beam of energetic ions to perform in situ analysis of solid samples by secondary ion mass spectrometry (Shimizu & Hart, 1982a). This technique has been employed to study lithium isotopic composition of meteorites (Poschenrieder et al., 1965; Klossa et al., 1981; Chaussidon & Robert, 1998) and serpentinized peridotites (Decitre et al., 2002). Atomic absorption spectrometry has been considered a rapid and economic means for the isotopic determination of lithium (Wheat, 1971; Divis &Clark, 1978). The doublet of the 670.8 nm resonance line of lithium undergoes an isotopic shift resulting in a triplet for the natural mixture of lithium isotopes. The use of monoisotopic 6Li and 7Li lamps permits separation of the specific adsorption components of the two isotopes and hence the determination of the isotopic ratio of lithium. The precision of the
124
Chapter 6- L.-H. Chan
method is limited by the precision of the absorbance measurement with a hollow cathode lamp and instrument drift (Meier, 1982; Divis & Clark, 1978). As a result, an uncertainty of several weight percent has been observed. This method therefore does not provide a high degree of precision necessary for geological studies. Nuclear reactions provide a means of lithium isotope determination without the difficulty of isotope fractionation (Brown et al., 1978). This technique makes use of the reactions 6Li(d,c~)4He, 7Li(d,c~)5He(n)4He and 7Li(d,n)aBe(c~)4He. Lithium isotope ratio is directly proportional to the a particles produced by the specific reactions. Errors inherent in this method include the dependence of reaction cross section on beam energy and incident angle, interference reactions, and target thickness. By calibration against isotopic standards the overall accuracy of this method is estimated to be about 1%. This method has been applied to establish the consistency of the lithium isotope ratios of meteorites at the 1% level (Rajan et al., 1980). Because of its relatively low precision, the nuclear technique has not been widely used for geological studies. 6.3 M o d e r n t e c h n i q u e s
Among all the techniques, mass spectrometry is inherently the most precise method for lithium isotope determination. Recent advancement includes the development of appropriate ion source compounds for the conventional thermal ionization method and the adaptation of inductively coupled plasma mass spectrometry to this extremely light element. This section includes a description of chemical preparation procedures followed by the modern mass spectrometric methods in the chronological order of their development. 6.3.1 Chemical preparation The chemical procedure is designed to achieve quantitative and clean extraction of lithium from aqueous and solid samples. Total recovery is crucial because loss of lithium in the course of chemical separation will lead to isotopic fractionation. Impure samples will require higher filament current during the ionization process, causing isotope fractionation and rapid sample decay. Extraneous ions can also affect mass bias during analysis in an inductively coupled plasma mass spectrometer and consequently isotope ratio measurement. 6.3.1.1 Rock and sediment dissolution For lithium isotope analysis, silicate rocks are best ground in an agate mortar and decomposed by digestion in a mixture of 48% HF and 72% HC104 (4:1 by volume). Care must be taken because lithium forms insoluble fluoride and is easily lost in the residue. In the dissolution procedure, cations in the silicates form fluorides upon reaction with HF. As evaporation proceeds, the fluorides are converted to perchlorates because of the low volatility of perchloric acid. The reactions with lithium in the silicates are shown below.
Li-silicate + HF ~ LiF + SiF4 LiF + HC104 ~ LiC104 + HF
[6.1a]
[6.1b]
Mass Spectrometric Techniques for the Determination of Lithium Isotopic ...
125
The perchlorates of cations are easily dissolved in water, resulting in total dissolution in an aqueous solution. Lithium in carbonate rocks is extracted by dissolution in dilute acetic acid. The use of weak acid prevents reaction with silicate detritus that is rich in lithium. Lithium concentrations of natural waters and rock solutions are determined prior to isotope analysis. Lithium may be determined at the gg/1 level by using flame emission and inductively coupled plasma optical emission spectrometry. Standard additions are carried out to remove the matrix effect. For low lithium abundance or small samples, lithium concentration may be measured by isotope dilution mass spectrometry using lithium phosphate as an ion source material (You &Chan, 1996).
6.3.1.2 Chemical separation Lithium is separated from other ions in the sample solution by the conventional ion exchange chromatography technique. It is well known that significant fractionation of lithium isotopes occurs during ion exchange chromatography. As lithium passes through a column of inorganic (e.g. sodium zeolite) or organic (e.g. sulfonated polystyrene) cation exchanger, the leading eluant is rich in 7Li and the trailing sample is rich in 6Li (Taylor & Urey, 1938; Oi et al., 1991; Moriguti & Nakamura, 1998a). It is therefore essential to achieve total recovery of lithium from the ion exchange column. The resin used for cation exchange is a styrene-divinylbenzene cross-linked copolymer with sulfonic acid functional groups (Bio-Rad AG50Wx8 or AG50Wx12, 200-400 mesh). Lithium is the smallest of alkali ions but has the largest hydrated radius. As a result, lithium is least strongly held on the resin, and the ion selectivity of the alkali ions follows the order of decreasing size of the hydrated ions: Li+
126
Chapter 6- L.-H. Chan
F i g u r e 6 . 1 - Elution curves of lithium in seawater and various rock solutions from cation exchange columns: (a) 100 ml column for the borate method (Chan, 1987), and (b) 15 ml column for the phosphate method (You &Chan, 1996). Lithium peak shifts to earlier eluant volumes with increasing Mg/Li. In practice, an appropriate cut of the lithium fraction was taken before the Na peak to accommodate rock compositions ranging from basalt to peridotite. In the example shown in Figure 6.1b, the 38-64 ml fraction (shaded area) was collected for quantitative recovery of lithium from the rocks. (Figure from Chan et al., 2002c).
Mass SpectrometricTechniquesfor the Determinationof LithiumIsotopic ...
127
To analyze lithium in the form of monovalent ion (Li+), only 100 ng of lithium or less is required. In this technique, You & Chan (1996) employed a smaller cation exchange column with an inner diameter of 1 cm filled to a height of 15 cm. Clean separation of lithium from other cations can be achieved with a single elution of 0.5N HC1. Lithium phosphate is prepared by adding excess phosphoric acid to LiC1 to form the ion source material (Moriguti & Nakamura, 1993). Moriguti & Nakamura (1998a) adopted multi-stage ion exchange chromatography to purify lithium from major ions and organic matter. The separation procedure employed a succession of four cation exchange columns with small resin volumes (0.1-1 ml) and HC1 of various normalities and mixed ethanol-HC1 as eluants. This work clearly demonstrated the shift of the lithium elution peak through the series of basalt, andesite and rhyolite. James & Palmer (2000) achieved good separation of lithium and sodium using a 2.7-ml resin volume and two passes of sample through the column. Although the 4-step procedure of Moriguti & Nakamura (1998a) is tedious, the advantage of small resin volumes is low procedural and reagent blanks (11 pg and 0.8 pg). In comparison, the total blanks from the 15-ml columns used by You & Chan (1996) are 100-190 pg and those from the procedure of James & Palmer (2000) are 80150 pg. Nitric acid mixed with methanol is an effective agent for separation of lithium from sodium because of their different partition coefficients between this solution and the ion exchanger (Strelow et al., 1974). Tomascak et al. (1999a) used I M HNO3 in 80% methanol to elute Li from an AG50Wx8 column (8 mm inner diameter, 30 cm high) for isotope analyses by ICP-MS. 6.3.2 Thermal ionization mass spectrometry Two approaches have been undertaken to overcome the problem of mass fractionation in thermal ionization mass spectrometry: (1) the synthesis of high-mass molecular ions for isotopic measurement, and (2) the selection of a loading form that produces Li + ions with minimum temperature-induced fractionation. In the first method, the isotopic effect is reduced because the relative mass difference between complex isotopic species is smaller. In the second approach, the goal is to increase the volatilization of a molecular species relative to that of the atomic species, followed by congruent ionization. Chan (1987) and Green et al. (1988) have developed methods based on the measurement of molecular ions Li2BO2+ and Li2F+. Among the lithium compounds, Li2B407 and Li3PO4 are considered to be the best loading forms for the measurement of Li + (Xiao & Beary, 1989; Xiao et al., 1992a; Moriguti & Nakamura, 1993; You & Chan, 1996). Of these mass spectrometric techniques, the diborate molecular ion method (Chan, 1987) and the phosphate method (Moriguti & Nakamura, 1993, 1998a; You & Chan, 1996) have been used most extensively for high-precision studies of geological material. 6.3.2.1 Isotopic determination based on the measurement of molecular ion (Li2B02 +) In this method, lithium is converted to lithium tetraborate (Li2B407) by reaction with boric acid. The tetraborate serves as an ion source from which Li2BO2§ is pro-
Chapter 6 - L.-H. Chan
128
duced by thermal ionization in the mass spectrometer. Isotopic ratios are measured on the molecular ion at masses 56 and 57, thus reducing the isotope effect compared to the normal measurement of Li § at masses 6 and 7. Chan (1987) first developed the chemical separation and mass spectrometric procedures for this method and applied it to the study of geological samples. Datta et al. (1992) studied in detail the factors that influence the ion beam behavior and precision. There are 36 possible isotopic configurations of the Li2BO2+ molecular ion having mass numbers from 54 to 61. The most abundant Li2BO2+ ions are of the mass numbers 55, 56, and 57 when all the constituent elements are of natural origin (Table 6.1). 6Li/7Li is determined from the most intense ion currents at m / z - 56 and 57. If we ignore the relatively insignificant contributions of 170 and 180, the ion intensity ratio at these two masses may be written as:
I56
26Li7LillB1 60 2 + 7Li 2 10B 160 2 I5.--7 = 7Li211B160 2 = 2L + B
[6.2]
where I56 and I57 are ion current intensities at m / z = 56 and 57, L is 6Li/7Li in the sample, and B is lOB/11B in the boron reagent. In actual isotopic analysis, the diborate ion is synthesized with liB-enriched boric acid so that I56/I57 is more sensitive to the change in L (Chan, 1987). The error introduced in the 6Li/7Li ratio by using the approximation in equation [6.2] is less than 0.2%o for B - 0.05, the boron isotopic ratio of the boric acid used in this technique (Chan, 1987). A weaker peak is observable at mass 55. If oxygen is assumed to be monoisotopic, its intensity relative to that of the mass 57 peak is I55
/57
=
26Li7Li1~
6
+ Li21
1B1602
= L(L + 2B)
[6.3]
7Li211B1602
In the absence of isotopic fractionation, L calculated from I55/I57 and I56/I57 should agree within the limit of uncertainty. However,/55 is about 15 times lower than /56 in natural samples when B - 0.05 and therefore carries a larger measurement error. For this reason,/56/I57, rather than I55/I57 is used for ratio determination. In the presence of traces of sodium impurity following the ion exchange procedure, LiNaB407 is produced along with Li2B407. Abundance ratio may be measured on LiNaBO2 + at masses 72 and 73 according to
/72
6Li23Na11B1602 + 7L i23Na 10B 160 2
173
7L i23Na 11B 160 2
= L+B
[6.4]
129
Mass Spectrometric Techniques for the Determination of Lithium Isotopic ...
Lithium isotope analysis of tetraborate is Table 6.1 - Possible isotopic configurations of performed on a magnetic sector thermal Li2BO2+ molecular ion ionization mass spectrometer equipped Mass number Species with multiple collectors. The operational conditions and procedures have been 54 6Li210B1602 described by Chan (1987) and Datta et al. (1992). The following discussions are based 6Li7Li10B1602 55 on these studies. 6Li211B1602 6Li210B160170
The prepared Li2B407 is loaded on 56 degassed tantalum filament. The filament current is gradually increased until the Li2BO2+ signal reaches a suitable intensity. The abundance ratio at m / z 56 and 57 is then measured using simultaneous collection with Faraday cups. Datta et al. (1992) observed that the use of high signal 57 strength (>400 mV at mass 57) leads to enhancement in the rates of signal decay and fractionation. Therefore, the molecular ion ratio is measured at low signal intensity between 30 mV and 100 mV (0.3 x 10-12 to 1 x 10-12 A). Under these conditions, the intensity ratio remains largely constant for more than 10 blocks (2Om = 0.4- 1%o for 100 ratios), indicating small in-run isotopic fractionation.
7Li210B1602 6Li7LillB1602 6Li211B160170 6Li7Li10B160170 6Li210B160180 6Li210B1702 7Li211B1602 7Li210B160170 6Li211B160180 6Li7Li11B160170 6Li7Li10B160180 6Li7Li10B1702 6Li210B170180 6Li211B1702
In some samples LiNaBO2 + ions ( m / z = 72 and 73) are detected along with the Li2BO2+ ions (Chan et al., 1992). Although the lithium peak is cleanly separated from the sodium peak during cation exchange chromatograph3r traces of sodium may be introduced from the anion exchange resin as it is converted to the hydroxide form with 1N NaOH. Thus during mass spectrometric runs, both LiNaBO2 + and Li2BO2+ ions are monitored and the more dominant species is used to determine the abundance ratio. 6Li/7Li ratio can be calculated from the ion intensities at masses 72 and 73, according to equation [6.4]. When both ions are prominent in a sample, measurements at m / z - 56/57 and 72/73 yield identical 6Li/7Li ratio within error, indicating absence of isotopic fractionation between the coexisting dilithium and lithiumsodium compounds. The mean measured 6Li/7Li ratio of a lithium carbonate standard NIST L-SVEC is 0.08282 based on 10 separate analyses, with a standard deviation (lo) of 0.00010 or 1.3%o. Later the ratio determined with the same method was revised to be 0.083062 + 0.000054 (lo) (Chan et al., 1992) probably due to different instrument conditions. Analyses of seawater samples yielded a mean 67Li value of 33.4 + 0.5%o (lo) (Chan& Edmond, 1988). With the development of this method, it became possible to deter-
130
Chapter 6 - L.-H. Chan
mine lithium isotopic composition of geological samples to a precision of about 1%o. However, because of low analytical sensitivity, applications have been restricted to high lithium samples including oceanic rocks and hydrothermal fluids (Chan et al. 1992, 1993, 1994) and saline brines (Bottomley et al., 1999, 2003; Chan et al., 2002d). It is theoretically possible to solve for lithium and boron isotopic composition and fractionation from analyses of more than two masses of Li2BO2+ (Sahoo & Masuda, 1995; Bickle et al., 2000). Sahoo & Masuda (1995) employed a Daly ion counter to measure Li2BO2+ ions at masses 54, 55, 56, and 57 and determined lithium and boron isotopic ratios by solving simultaneous equations [6.2] and [6.3]. The method yielded internally consistent results with a relative standard deviation (RSD) of 0.5%o for lithium and 0.1%o for boron. The isotopic ratio obtained for L-SVEC was 0.082289. Datta et al. (1992) pointed out that/56/I57 and I55//57 a r e affected to different degrees for a given change in the 6Li/7Li ratio, as apparent from equations [6.2] and [6.3]. This means that unless true intensity ratios are obtained, simultaneous measurements of I56//57 and 155//57 a r e not expected to yield the same value for L. Bickle et al. (2000) evaluated the possibility of fractionation correction by measuring multiple species of Li2BO2+. They conclude that it is not possible to correct for fractionation because of the loss of the much larger Li + beam from lithium tetraborate. Thus the usefulness of the simultaneous isotopic determination of lithium and boron is limited. 6.3.2.2 Isotopic analysis based on the measurement of atomic ion (Li +) While the ionization of Li2B407 to Li2BO2+ is typically weak, lithium having a low ionization potential readily ionizes to Li § in a thermal ion source. Consequently, measurement of monovalent ions results in two to three orders of magnitude increase in sensitivity compared to molecular ions. However, the light isotope tends to vaporize more readily causing the measured 7Li/6Li ratio to increase in the course of a sample run. Two lithium compounds have been developed as filament loading forms that produce Li + ions with minimal isotopic fractionation.
6.3.2.2.1 Tetraborate as the loading form Xiao & Beary (1989) and Xiao et al. (1992a) examined the temperature effect on the isotopic ratio of Li + emitted from several lithium compounds (LiNO3, LiF, LiC1, LiI, LiOH, Li3PO4, and Li2B407) using a double filament technique. The 7Li/6Li ratio increases with ionization filament temperature below 1200~ for all compounds (Figure 6.2). At temperatures above 1200~ the ratio becomes constant with temperature. Among the lithium compounds investigated, Li2B407 produced a strong Li § ion beam with the most stable isotopic ratio. This technique is therefore suitable for determination of isotopic composition of samples containing 10 ng to I gg lithium. The protocol for the preparation of Li2B407 is similar to that for the diborate molecular ion (Xiao & Beary, 1989). Mass spectrometric analysis is carried out using double rhenium filaments. The ionizing filament current is raised until the temperature reaches 1500~ This procedure yields a precision of 0.23%o for L-SVEC (RSD). This technique has been used to measure urine (Xiao & Beary, 1989) and foraminifera
Mass Spectrometric Techniques for the Determination of Lithium Isotopic ...
131
(Hoefs & Sywall, 1997). Qi et al. (1997) used this method to calibrate isotope reference material with a l~J precision of 1.4%o. The mass fractionation correction for this method is small (K - Rtrue / Rmeasured - 0.9979, where R - 6Li / 7Li).
6.3.2.2.2 Phosphate as the loadingform Moriguti & Nakamura (1993) considered lithium phosphate to be superior to Li2B407 as an ion source material. Due to its stability at high temperatures, it is more efficiently vaporized as a molecular species without decomposition to the monovalent ion. No fractionation was observed in Li3PO4 at temperatures above 750~ contrary to the observation of Xiao et al. (1992a). You & Chan (1996) confirmed that lithium phosphate gives rise to an ion beam with stable isotopic ratio between 1000 and 1400~ (Figure 6.2). Because of the high ionization efficiency of Li3PO4, this technique may be applied to samples containing tens of ng to 100 ng lithium. Lithium phosphate is analyzed using a double rhenium filament assembly, Moriguti & Nakamura (1993) recommended an ionization filament temperature of 1150~ (equivalent to a current of 1.55-1.65 A), whereas You & Chan (1996) chose 1350~ (2.1 A). James & Palmer (2000) also observed excellent stability of 6Li/7Li between 850 and 1200~ Isotope ratio measurements are performed with a Li + ion current of about I x 10-11 A (evaporation filament current 0.6-0.8 A) (Moriguti & Nakamura, 1993, 1998a; You &Chan, 1996; James & Palmer, 2000). It is important to note
Figure 6.2 - Variation of the 7Li / 6Li ratio with ionization filament temperature for different compounds. Data for Li3PO4 are from Moriguti & Nakamura (1993) and You & Chan (1996). All other data are from Xiao & Beary (1989). (Figure from You & Chan, 1996).
132
Chapter 6 - L.-H. Chan
that the isotopic ratio of light elements is sensitive to beam focus; hence ion beam focusing is performed at the beginning of each block (10 ratios) during ratio measurement. The procedure of You & Chan (1996) yielded a relative standard deviation (RSD) of 0.4%o for analyses of L-SVEC, 1.3%o for seawater and better than 0.5%o for standard rocks (Chan et al., 2002c). Intercalibration with the molecular diborate ion method showed agreement better than 3%o for a wide range of geological material, with an average difference of 1%o between the two methods (You & Chan, 1996). Moriguti & Namura (1998a) achieved a RSD of 0.4%o for standard, rocks, and seawater samples. James & Palmer (2000) achieved a precision of better than 1%o for seawater and rock analyses. Thus, with proper column calibration, the phosphate method can achieve a precision of +1%o (lo) or better for geological samples. Because of its high sensitivity, this technique extends the analytical capability to low abundance lithium samples such as river waters (Huh et al., 1998), sediment pore waters (Zhang et al., 1998b; James et al., 1999; Chan & Kastner, 2000), marine carbonates (You & Chan, 1996), and highly depleted volcanic rocks (Chan et al., 2002a).
6.3.3 Inductively coupled plasma mass spectrometry The first generation of ICP mass spectrometer is equipped with a quadrupole mass analyzer and a single multiplier detector. This necessitates sequential measurement of isotope abundance, and the precision of isotope ratio determination is limited by the stability of ion signal generated by the ICP source. The most advanced ICP mass spectrometer combines the ICP ion source with a double focusing magnetic sector mass analyzer and nine Faraday cups. This multi-collector ICP mass spectrometer (MCICP-MS) allows simultaneous isotopic measurement and opens the door to high precision determination of isotopic composition of refractory and high-ionization potential elements. Although lithium is readily ionized in a thermal source, ICP-MS is a potentially favorable alternative for isotopic determination because of its speed, low detection limit, and tolerance of sample impurities. Gr6goire et al. (1996) first applied the technique to geological material using a quadrupole ICP-MS. Measured isotopic ratios varied with ion lens voltages, aerosol carrier gas flow rates, and lithium concentration in the solution. These mass discrimination effects were corrected by analyzing the isotopic standard IRM-016 (Michiels & De Bi~vre, 1983) before and after the unknown sample. The procedure resulted in a mean RSD of 0.8%0. Kosler et al. (2001) demonstrated that quadrupole ICPMS can be successfully used for precise measurement of low-Li marine carbonates. Lithium isotopic composition was measured on foraminiferal shells containing 5-10 ng of lithium with a precision of 1%o (la). Tomascak et al. (1999a) were the first to employ a multi-collector sector ICP-MS (VG Plasma 54) to determine the isotopic composition of lithium. Desolvating nebulizer is used to remove most of the water in the sample solution, thereby increasing the amount of sample ionized in the plasma. The conventional nine Faraday cup array cannot accommodate the 16% mass dispersion between 6Li+ and 7Li+. On VG Plasma
Mass SpectrometricTechniques for the Determination of Lithium Isotopic ...
133
54, an additional Faraday cup on the high-mass side of the flight tube together with the axial Faraday cup permits simultaneous measurement of the two isotopes. Mass fractionation effects are severe in MC-ICP-MS, especially for low-mass isotopes. An important cause of mass bias is believed to be the space charge effect in the skimmer cone as ions enter the mass spectrometer (Gilson et al., 1988; Tanner, 1992). Heavy ions experience less mutual repulsion and consequently are transmitted more efficiently than lighter ions. In practice, the measured 7Li/6Li ratio varies with time with a drift up to 6% over 12 hours (Tomascak et al., 1999a). The isotopic ratio may increase or decrease over the course of a sample run and the standard exhibits varied values that may be higher or lower than the reported values (Figure 6.3). These observations suggest that the mass discrimination processes of lithium are complex and that the fractionation factor is very large. To correct for mass fractionation, an isotopic standard (L-SVEC) is run before and after each unknown sample. 67Li value of the unknown is calculated by normalizing to the mean isotopic value of the adjacent standards. In spite of the severe mass bias, the average precision of this technique, based on replicate analyses of standards and geological samples, is estimated to be +0.6%o (lo). The results for seawater and a basalt standard (JB-2) agree with the published TIMS values within analytical uncertainty (see Table 6.3). This method is suitable for analysis of small samples (~40 ng Li). In addition, lithium isotope determination appears to be insensitive to the presence of trace sodium and magnesium contaminants in the sample solution (Tomascak et al., 1999a; Nishio and Nakai, 2002). However, higher concentrations of matrix elements can significantly affect mass bias and hence the measured isotope ratio (Nishio and Nakai, 2002). Applications of the MC-ICP-MS technique include the investigations of volcanic rocks and natural waters (Tomascak et al., 1999c, 2000, 2002, 2003).
6.4 References and standards for lithium isotope composition 6.4.1 L-SVEC (NIST Reference Material 8545) The isotopic standard is L-SVEC, a Li2CO3 standard distributed by U.S. National Bureau of Standards (NBS), presently National Institute of Standards and Technology (NIST). The carbonate was prepared from spodumene ore and standardized against primary standards that were prepared by blending highly enriched lithium isotopic materials (Flesch et al., 1973). Its absolute 6Li/7Li abundance ratio has been determined by mass spectrometry to be 0.0832 + 0.0002 (Flesch et al., 1973). Qi et al. (1997) calibrated the same standard and obtained an absolute 6Li/7Li ratio of 0.08215 _+ 0.00012 (lo)(Table 6.2). Because of mass discrimination effects, isotopic ratio measured on TIMS is rarely the true value of the sample. The isotopic ratios for L-SVEC obtained by various TIMS techniques at different laboratories are summarized in Table 6.2. All TIMS measurements were performed on magnetic sector mass spectrometers, except for the work of Hoefs & Sywall (1997) which was carried out on a thermal ionization mass spectrometer with a quadrupole mass analyzer. The values range between 0.0821 and 0.0844. The differences between laboratories could be due to instrument factors and operating conditions specific to the method employed. The methods based on the measurements of complex molecular ions have a greater RSD (0.6 to 1.2%o for diborate) than
134
Chapter 6 - L.-H. Chan
F i g u r e 6 . 3 - Lithium isotope analyses of geological samples (circles) using multi-collector, magnetic sector ICP-MS: (a) seawater; (b) JB-2 basalt standard; (c) saline groundwater. Mass fractionation is corrected by alternate runs of an external standard, L-SVEC (diamonds) (Figure from Tomascak et al., 1999a).
135
Mass Spectrometric Techniques for the Determination of Lithium Isotopic ...
Table 6.2 - Lithium isotope ratio of reference standards L-SVEC and IRM-016 measured by thermal ionization mass spectrometric methods Loading form
Measured ion
6Li/7Li
lc~
RSD ( % o )
L-SVEC LiI LiI Li2B407 LiF Li2B407 Li2B407 Li2B407 Li2B407 Li2B407 Li2B407 Li3PO4 Li3PO4 Li3PO4 Li3PO4 Li3PO4
Li + Li + Li + Li2F + Li2BO2 + Li2BO2 + LiNaBO2 + Li2BO2 + Li + Li + Li + Li + Li + Li + Li +
0.0832* 0.08214 0.08215* 0.08201 0.08282 0.083062 0.083013 0.082289 0.082212 0.0844 0.082543 0.082612 0.082533 0.082740 0.082757
0.0002 0.00008 0.00012 0.00013 0.00010 0.000054 0.00009 0.000043 0.000019 0.0004 0.000022 0.000029 0.000033 0.000030 0.000028
2.4 0.9 1.4 1.6 1.2 0.65 0.90 0.5 0.23 4.7 0.26 0.36 0.39 0.36 0.34
IRM-O 16 LiI LiI Li2B407
Li + Li + Li +
0.08137* 0.08181* 0.08212*
0.00017 0.00031 0.00014
2.1 3.8 1.7
Reference
Flesch et al., 1973 Lamberty et al., 1987 Qi et al., 1997 Green et al., 1988 Chan, 1987 Chan et al., 1992 Chan et al., 1993 Sahoo & Masuda, 1995 Xiao & Beary, 1989 Hoefs & Sywall, 1997 Moriguti & N a k a m u r a , 1993 You & Chan, 1996 Moriguti & N a k a m u r a , 1998a Huh et al., 1998 James et al. 1999
Michiels & De Bi6vre, 1983 Lamberty in Qi et al., 1997 Qi et al., 1997
*absolute values calibrated with synthetic standards
those from measurements of monovalent ions from tetraborate and phosphate (0.2 to 0.4%0). The greater uncertainty in the analysis of the complex ions may be weak ion current due to inefficient thermal production, as well as in-run fractionation. However, analysis of Li + on a quadrupole instrument carries a greater error of 5%0 (Hoefs & Sywall, 1997), indicating that a magnetic sector instrument is necessary for highprecision lithium isotope analysis. It is noted that the best measured 6Li/7Li ratios from the phosphate ion source are all higher than the absolute ratio obtained for LSVEC by Qi et al. (1997) but lower than that of Flesch et al. (1973) (Table 6.2). In thermal ionization mass spectrometry; the mass discrimination effect is mainly due to differential vaporization of the isotopes, hence the measured 6Li/7Li ratio should be higher than the true ratio. The result of Qi et al. (1997) is in keeping with the expected effect of mass dependent fractionation, suggesting that the higher value of Flesch et al. (1973) may be inaccurate. This isotopic standard is currently used by many laboratories as the reference to compute the ~ value for natural and other samples. Because 6Li is the rare isotope, the isotopic composition of Li in a sample has been expressed as 66Li.
136
Chapter6 - L.-H.Chan -( 6Li]CLL/} sample
56Li -
6Li]
-1
x 1000
[6.5]
%)std
By definition, a higher 86Li value indicates a relatively light isotopic composition. To conform to the 6 notation of other stable isotopes, the isotopic composition currently is expressed as c57Li:
(7Li]
samp,e
57Li -
-1
x 1000
[6.6]
6Li) std
The conversion relationship of the two notations is &7Li -
-&6Li 1 + D6Li~ 1000/
[6.7]
6.4.2 I R M - 0 1 6 ( I R M M reference material) There exists another lithium isotope standard (IRM-016) that is used in some laboratories (Oi et al., 1997; Gr6goire et al., 1996). This standard is distributed by the Institute for Reference Materials and Measurements (IRMM), Belgium. The absolute 6Li/ 7Li ratio of this lithium carbonate standard was originally determined by Michiels & De Bi6vre (1983) to be 0.08137 + 0.00017 (lo). Recently Qi et al. (1997) recalibrated this standard to have 6Li/7Li - 0.08212 + 0.00014 (lo), in agreement with the absolute value they obtained for L-SVEC, 0.08215 + 0.00012 (lo). According to these authors, the absolute isotopic abundance of IRMM-016 is identical to that of L-SVEC. 6.4.3 Seawater as an isotopic standard
Lithium is a conservative element in the ocean and has an oceanic residence time that is much longer than the ocean mixing time (Stoffyn-Egli & McKenzie, 1984). The isotopic composition of lithium in seawater is therefore homogeneous everywhere in the ocean ( C h a n & Edmond, 1988). Seawater analysis provides a means to test the suitability of the entire chemical and mass spectrometric procedure. The published data of seawater are summarized in Table 6.3. Both 66Li and 57Li relative to L-SVEC are given to accommodate the two notations used in literature. All errors are quoted as lo. Chan & Edmond (1988) showed that seawaters from the
137
Mass Spectrometric Techniques for the Determination of Lithium Isotopic ... Table 6.3 - Lithium isotopic composition of seawater and standard rocks relative to L-SVEC Standard Seawater
66Li (%o) 67Li(%0) lo (%o)
n
Method
Reference Chan & Edmond, 1988 You &Chan, 1996 James & Palmer, 2000 Moriguti & Nakamura, 1998a Tomascak et al., 1999a Nishio & Nakai, 2002
-32.3 -31.4 -31.5 -29.1
33.4 32.4 32.5 30.0
0.5 1.3 0.5 0.4
5 6 7 5
TIMS-Li2B407 TIMS-Li3PO4 TIMS-Li3PO4 TIMS-Li3PO4
-30.8 -28.5
31.8 29.3
1.0 0.5
15 3
MC-ICP-MS MC-ICP-MS
-4.9
4.9
0.4
5
TIMS-Li3PO4
-6.8 -5.1 -5.1 -4.3
6.8 5.1 5.1 4.3
0.1 0.2 0.3 0.3
3 7 4 5
TIMS-Li3PO4 TIMS-Li3PO4 MC-ICP-MS MC-ICP-MS
-3.9
3.9
1
TIMS-Li3PO4
-3.9
3.9
0.15
3
MC-ICP-MS
Moriguti & Nakamura, 1998b Nishio & Nakai, 2002
BHVO-1 Hawaii basalt
-5.8 -5.2 -5.0
5.8 5.2 5.0
0.8 0.2 0.8
3 3 8
TIMS-Li3PO4 TIMS-Li3PO4 MC-ICP-MS
James & Palmer, 2000 Chan & Frey, 2003 Bouman et al., 2002
JA-1 Andesite
-5.7
5.7
1
TIMS-Li3PO4
-5.8
5.8
0.3
5
MC-ICP-MS
Moriguti & Nakamura, 1998b Nishio & Nakai, 2002
-3.8 -3.9
3.8 3.9
0.4 0.2
3 3
TIMS-Li3PO4 TIMS-Li3PO4
James & Palmer, 2000 Chan & Frey, 2003
JB-2 Japan basalt
JB-3 Japan basalt
JR-2 Rhyolite
Moriguti & Nakamura, 1998a James & Palmer, 2000 Chan et al., 2002c Tomascak et al., 1999a Nishio & Nakai, 2002
Pacific a n d Atlantic h a v e the s a m e 67Li v a l u e s w i t h i n the limit of uncertainty. Five ocean w a t e r s a m p l e s g a v e a m e a n v a l u e of 33.4 + 0.5%o b y the Li2BO2 + technique. U s i n g p h o s p h a t e as the ion source of Li +, You & C h a n (1996) o b t a i n e d a m e a n 67Li v a l u e of 32.4 + 1.3%o for Atlantic s e a w a t e r a n d J a m e s & P a l m e r (2000) o b t a i n e d a similar v a l u e (32.5+ 0.7%0). U s i n g the s a m e ion source c o m p o u n d , M o r i g u t i & N a k a m u r a (1998a) r e p o r t e d a l o w e r 157Li value, 30.0 + 0.4%0 for w a t e r f r o m the M a r i a n a Trough. Based on replicate a n a l y s e s b y MC-ICP-MS, T o m a s c a k et al. (1999a) d e t e r m i n e d 67Li of a Pacific O c e a n w a t e r s a m p l e to be 31.8 + 1.0%o. U s i n g the M C - I C P - M S t e c h n i q u e , N i s h i o & N a k a i (2002) o b t a i n e d a relatively l o w v a l u e of 29.3 + 0.5%o for the n o r t h Pacific. The results of different m e t h o d s therefore s p a n a r a n g e of 4%o, g i v i n g a m e a n of 31.6 + 1.6%o for global o c e a n water. L i t h i u m i s o t o p e c o m p o s i t i o n of this n a t u r a l reference s o l u t i o n m u s t be f u r t h e r e v a l u a t e d .
138
Chapter 6 - L.-H. Chan
6.4.4 Standard rocks
A number of international standard rocks have been analyzed and can serve as inter-laboratory comparison (Table 6.3). All TIMS measurements are based on Li + emission from phosphate. With the exception of JB-2 by James & Palmer (2000), the results demonstrate good agreement among laboratories for TIMS-phosphate and MC-ICP-MS techniques. We can conclude that these two methods can yield precise and accurate results to +1%o (lo) for isotope analysis of lithium in geological material. 6.5 Evaluation of mass spectrometric techniques In this section, I summarize the merits and drawbacks of the methods discussed. TIMS measurement based on the high-mass molecular ions reduces mass fractionation effects. However, complex lithium compounds such as Li2B407 do not ionize to molecular ions efficiently, and consequently a relatively large sample size (3-4/~g) is required. Such sample size necessitates the use of a large ion exchange column and great quantities of reagents. It also restricts geological application to Li-rich samples. Moreover, because the ion beam intensity of Li2BO2+ is orders of magnitude weaker than Li +, the in-run precision is inferior. In addition, error in the boron isotope determination and presence of oxygen isotopes all contribute to the uncertainty of the method. This pioneering technique is therefore giving way to other more sensitive mass spectrometric techniques.
Thermal ionization of lithium phosphate and tetraborate in a double filament source produces a strong and stable Li + ion beam. The method is therefore applicable to the analysis of small lithium samples (optimally 50 to 250 ng). This is a most important advantage, which allows measurement of low-lithium abundance or rare samples. The method is however sensitive to sample purity and beam focusing. Extraneous ions or organic matter in the sample source can cause isotopic fractionation. Small samples (< 5 ng Li) are especially susceptible to the interference of impurities. With proper sample preparation, this method can achieve a precision better than +1%o (lo) (Tables 6.2 and 6.3). Because of its intrinsic precision and accuracy, thermal ionization mass spectrometry has been the method of choice for isotope ratio determination. This technique however requires labor-intensive preparation of high-purity ion source material, tedious sample loading and conditioning routine, and long measurement time. The advantages of ICPMS over TIMS are high-ionization efficiency of the plasma source, tolerance of sample impurity, relatively simple sample preparation, and speed and ease of analysis. MC-ICP-MS is now a increasingly popular tool for precise and accurate isotope analysis of lithium. Ratio determination can be performed with 40 ng of lithium. Sample is introduced by solution nebulization, eliminating the need of preparing suitable loading compound and sample conditioning protocol. A major disadvantage is the pronounced isotopic effect as shown by large variation of the measured isotopic ratio with time and its large deviation from the true value. The mass fractionation effect in ICP-MS is corrected by normalizing to an external standard that brackets the unknown sample. This method of correction could potentially introduce error due to dissimilar behavior of the samples and the standard and matrix effects. A mag-
Mass SpectrometricTechniques for the Determination of Lithium Isotopic ...
139
netic sector MC-ICP-MS is also costly and therefore not readily available. A quadrupole ICPMS does not yield high precision but offers a more affordable means for plasma analysis. Development of the use of single collector magnetic sector ICPMS for high-precision lithium isotope determination may provide a compromise. In summary, existing methods permit determination of lithium isotope ratio in natural material to a precision of about +1%o (lo) or better. The method based on the measurement of the analysis of diborate ions is becoming obsolete because of inferior analytical sensitivity and precision, and laborious sample preparation. Two methods have proven most useful for high precision measurement of geological samples: the direct analysis of Li + from a phosphate source using TIMS and Li + from a plasma source using MC-ICP-MS. These methods are capable of measuring nanogram quantities of lithium thus providing the analytical capability for a wide range of geological materials.
6.6 Terrestrial variation and geological applications Much of the knowledge of terrestrial variation of lithium isotopes was generated in the last decade following the development of reliable mass spectrometric techniques and suitable chemical preparation procedures. There is now a growing database of the isotopic compositions of the major reservoirs of the Earth. Figure 6.4 presents the current knowledge of the isotopic ranges of lithium in ocean water, oceanic crust, marine sediments and interstitial waters, volcanic arcs, continental rocks, river waters and subsurface brines. Lithium isotopic composition is expressed as 67Li relative to L-SVEC. Ocean water is uniform in lithium isotopic composition with 67Li = 31.6 + 1.6%o (Table 6.3 and references therein), which is distinctly heavier than MORB (+1.5 to +4.7%o; Chan et al., 1992; Moriguti & Nakamura, 1998b; Tomascak & Langmuir, 1999). Alteration of seafloor basalts at low temperatures results in the addition of seawater lithium to the rocks and consequently heavier isotopic compositions in weathered basalts (6 to +21%o) (Chan et al., 1992, 2002a). Lower oceanic crust that has been altered by hydrothermal fluids is low in 7Li (67Li to -2%o) (Chan et al., 2002c). Hightemperature vent fluids from the mid-ocean ridge systems have a 67Li range of 6 to 11%o, indicating extraction of basaltic lithium with preferential retention of 6Li in greenschist facies alteration minerals (Chan et al., 1993; Bray, 2001). These studies show that lithium isotopic composition is an excellent indicator of seawater-ocean crust interaction. Marine sediments were reported to have 67Li values between-1 and 15%o (Chan et al., 1994, 1999; You et al., 1995a; James et al., 1999; Chan & Kastner, 2000). New determinations and revision of published data show relatively light isotopic ratios in pelagic and hemipelagic sediments (2-4 %o; Chan et al., 2001, 2002b; Chan, unpublished). Consequently, the isotopic range for clay-rich sediments is revised t o - 1 to 6%o. Marine carbonates have a wide range of isotopic compositions reflecting diagenesis and possibly variation of seawater composition with time (You &Chan, 1996; Hoefs & Sywall, 1997; Kosler et al., 2001). Lithium isotopic composition of sediment
140
Chapter 6 - L.-H. Chan I
'
I
"
I
=
Seawater MORB OIB Weathered MORB Hydrothermally altered MORB Mid-ocean ridge hot springs Marine sediments Sediment pore fluids Marine carbonates Arc lavas Continental crust River waters
t
-10
I
0
i
I
10
i
I
t
20
I
30
i
I
|
40
50
I
60
67Li%o Figure 6.4 - The ranges of isotopic compositions of lithium in various reservoirs. sources are given in the text.
Data
interstitial waters is sensitive to sediment-water reactions. These include volcanic ash alteration, ion exchange, sediment diagenesis, hydrothermal reactions and fluid expulsion from accretionary prisms (You et al., 1995a; You & Chan, 1996; Zhang et al., 1998b; James et al., 1999; Chan & Kastner, 2000). The isotopic composition of the mantle is not well known. Hawaiian shield volcanoes exhibit a narrow range of 67Li (2.5 to 5.7%0) ((Tomascak et al., 1999c; C h a n & Frey, 2003). Lithium isotope ratio is a novel tracer of arc magma genesis and subduction zone processes (Moriguti & Nakamura, 1998b; Chan et al., 1999, 2002c; Tomascak et al., 2000, 2002). Lithium concentration and g7Li typically increases in arc lavas relative to mantle-derived rocks due to contribution of subducted components. Continental rocks, including granitic rocks, rhyolites, schist, gneiss, and shale vary between-3.6 and 11%o (Sturchio & Chan, 2003; Bottomley et al., 2003). River waters display a wide range of isotopic compositions, 6 to 33%0 (Huh et al., 1998). River suspended material is relatively light (-2 to 6%o;Huh et al., 2001), suggesting large fractionation of lithium isotopes in the weathering regime. Lithium isotopes have also been used to decipher the origin of solutes in salt lakes (Tomascak et al., 2003) and subsurface brines in crystalline rocks and sedimentary basins (Bottomley et al., 1999,
Mass Spectrometric Techniques for the Determination of Lithium Isotopic ...
141
2003; Chan et al., 2002d). 6.7 Conclusion
In the last two decades much effort has been devoted to method development for high precision lithium isotope measurement. It is now possible to analyze lithium in geological samples with a standard deviation of +1%o or better. This capability has produced a significant database for the reservoirs of the Earth, which show a span of 60-70%o in lithium isotopic composition. The isotopic systematics has proven valuable for understanding a wide range of geological processes. Important examples are oceanic crust alteration, hydrothermal activities, sediment-water interaction, arc magma genesis, and subsurface brine evolution. There are inquiries that have just begun, such as lithium isotopic composition of the mantle, fractionation in the weathering regime, and isotopic variation in seawater with time. In particular we are ignorant of the isotopic fractionation factors between minerals and water. It is fair to conclude that the problem of instrumental fractionation of lithium isotopes has been greatly reduced but not completely overcome. There remains great challenge in analyses of extremely small samples such as foraminiferal shells. With further research in instrumental techniques and refinement of analytical methods, we can continue to expand the applications of lithium isotopes and the fundamental knowledge of their geochemistry.
Acknowledgement
I gratefully acknowledge the collaboration and help of many colleagues in my lithium isotope research. National Science Foundation provided continuous support for the work from method development to geological studies. Louisiana Board of Regents provided funds for a Finnigan MAT 262 thermal ionization mass spectrometer. This chapter benefited from the reading of Dr. Y. Huh, and the comments of Drs. M. Berglund, T. Elliott, T. Zack and an anonymous reviewer.
Handbook of Stable Isotope Analytical Techniques, Volume 1 P.A. de Groot (Editor) 9 2004 Elsevier B.V. All fights reserved.
CHAPTER 7 Thermal Ionization Mass Spectrometry Techniques for Boron Isotopic Analysis: A Review Chen-Feng You Department of Earth Sciences, National Cheng Kung University, Tainan, Taiwan, ROC e-mail:
[email protected]
7.1 Introduction
Boron is a quintessential crustal element that is widely distributed in surface rocks and aqueous fluids on Earth (Leeman & Sisson, 1996; Anovitz & Grew, 1996). It has two natural stable isotopes, 10B and nB, with an average abundance of approximately 19.9 and 80.1% respectively. The unique geochemical characteristics of B, which include high solubility in aqueous fluids, high magmatic incompatibility and large relative mass difference between two isotopes, make B and 6nB useful tracers of deep earth fluids and the recycling of subducted materials in convergent margins (Kotaka, 1973; Kakihana et al., 1973; Palmer, 1991; Ishikawa & Nakamura, 1994; You et al., 1996; Palmer & Swihart, 1996). For instance, the boron isotopic compositions in natural waters (such as seawater, hydrothermal fluids, groundwater and spring water) vary approximately 80 %o (You et al., 1994; Barth, 1993) and show distinctive isotopic compositions in each hydrological reservoir (Figure 6.1). The 5riB distribution in aqueous solutions, therefore, serves as a diagnostic proxy for tracing water sources in hydrological processes or possible water/rock interactions (Barth, 1997, 1998; Heumann et al., 1995). Throughout this chapter the boron isotopic composition (511B) is expressed as per mil (%0) deviation from NIST SRM 951. 611B - {[(11B / 10B)sample / (11B / 10B)SRM 951]-1}X103
[7.11
where SRM 951 is the NIST boric acid standard, prepared from a Searles Lake borax and has a certified 11B/10B ratio of 4.0437 + 0.0033 (Cantanzaro et al., 1970). Boron is an underused tracer in earth sciences, in particular, when compared to the well-established H, O, and C stable isotopes. This mainly is a result of lacking proper analytical technique for B isotopic analysis. The first boron isotopic measurement, in fact, was made more than 30 years ago (Thode et al., 1948; McMullen et al., 1961; Shima, 1962; Agyei & McMuller, 1968; Schwarcz et al., 1969). Since then the progress of this field has been extremely slow because of problems associated with mass spectrometric analysis. Not until the last decade did a rapid increase in the number of
Thermal Ionization Mass Spectrometry Techniques for Boron Isotopic Analysis: A Review
143
Groundwater (Australia) Fresh and salty water (Australia) Lake Brine (Australia) Hot and cold spring water (China) Lake brine (China) Hot spring water (Yellowstone) Fumarolic condensates (Japan) Pore water Hydrothermal fluids Seawater I
-20
I
I
0
I
I
20
I
I
40
I
I
60
Figure 7.1 - Variation of boron isotopic compositions in aqueous fluids on Earth's surface environments (after Barth, 1993).
studies of the stable boron isotope composition in natural samples start to be seen in the related literature (Spivack & Edmond, 1986; Klotzli, 1992; Barth, 1993; Palmer & Swihart, 1996). During this period, a variety of instrumental analytical methods with varying degrees of accuracy and precision were proposed and subsequently have been evaluated by researchers in different fields (Heumann et al., 1995; Aggarwal & Palmer, 1995). These techniques include inductively coupled plasma mass spectrometry, ion microprobe, glow discharge mass spectrometry, and thermal ionization mass spectrometry (TIMS) (see Swihart, 1996). Below I will focus attention to discuss two techniques, namely alkali-borates positive ion (e.g., Cs2BO2+) TIMS and BO2-negative ion TIMS. Other more detailed information on the analytical procedures, including chemical separation, other instrumental techniques, and relevant geological implications are referred to in recent review articles by Swihart (1996), Palmer & Swihart (1996), Aggarwal & Palmer (1995), Heumann et al. (1995) and Barth (1993). 7.2 Thermal ionization mass spectrometry
Among all available techniques for isotopic analysis, TIMS produces the most reliable results and has become the method of choice for precise boron isotopic determination. According to the charge condition of ionized species inside mass spectrometer, the TIMS methods are sub-divided into two categories, namely the positive-TIMS and the negative-TIMS. The first reliable B isotopic determination was made feasible using
144
Chapter 7 - C.-F. You
Na2B407 mass spectrometric method by McMullen et al. (1961) and Nomura et al. (1973), where Na2BO2 + molecular ions were measured with precision up to 2 ~ 3%o. Subsequently Cs2B407 m a s s spectrometric was developed by Ramakumar et a1.(1985) and Spivack & Edmond (1986) for a more accurate and precise isotopic measurement, which is capable of producing data with an uncertainty of better than 0.3%o. About the same time, a negative-TIMS method utilizing BO2- was developed for B concentration and isotopic measurements, with a precision approximately 2%o (Heumann & Zeininger, 1985; Vengosh et al., 1989). The alkali-borates mass spectrometric techniques measure molecular ions at much higher masses than the BO2- TIMS. The potential artifact of temperature-induced mass fractionation associated with alkaliborates TIMS, therefore, is reduced and better in-run precision is achievable. However, the low ionization sensitivity and efficiency of alkali-borates has limited these techniques to be applied in low B abundance samples (e.g., foraminiferal tests or river waters). On the other hand, the BO2-negative-TIMS enhances the ionization efficiency more than 1000 times, but has suffered seriously from uncontrollable mass fractionation artifacts, as a result of the relative low masses of borate ions, which caused a rather poor analytical precision of approximately 2%o. Fortunately, there is room for improvement in terms of analytical precision of the BO2- TIMS technique, if precise measurement of other minor oxide peaks can be performed simultaneously. Sahoo & Masuda (1995) apply the Li-borate TIMS for precise measurement of Li2BO2+ molecular ions by employing a Daly ion-counting system. By solving simultaneous quadratic equations of intensity ratios at masses 54, 55, 56 and 57, they have obtained precise and accurate measurement of Li and B isotopes in one single analysis. Applying a similar concept, I will discuss possibility of using the 1 6 0 / 1 8 0 ratio for 11B/10B mass fractionation correction during the BO2-negative-TIMS in the later section. Nevertheless it is a fair comment to say that no one single technique is appropriate for all studies; instead, the choice of analytical method depends on the type of sample analyzed, the level of precision required, and the analytical equipment available as quoted by Aggarwal & Palmer (1995). For instance, the BO2- method achieves excellent ionization efficiency and high reproducibility for porewaters with little or no chemical procedure required (Zuleger et al., 1996; You et al., 1993a; Brumsack & Zuleger, 1992). The analytical precision of 2 to 3%o for BO2- TIMS is sufficient to resolve the porewater variations, as large as 30%0 in ODP Legs 127 and 131 (Brumsack & Zuleger, 1992; You et al., 1993b). Below I will firstly compare the advantages/disadvantages in terms of chemical efficiency, sensitivity, and analytical precision achieved by available positive-TIMS and negative-TIMS techniques for boron isotopic analyses. Subsequently a new experimental procedure applying the 180/160 normalization procedure for possible mass fractionation correction will be outlined. 7.3 Positive-TIMS
Several alkali borate compounds, including Li-borate, Na-borate, K-borate, Rbborate and Cs-borate, have been investigated as ionization sources for boron positiveTIMS analyses over the past four decades (Table 7.1).
Thermal Ionization Mass Spectrometry Techniques for Boron Isotopic Analysis: A Review
145
Table 7.1. Summary of different TIMS techniques for boron isotopic ratio determination Technique
Measured ion
Relative standard deviation (%o)
Advantages/disadvantages
References
P-TIMS
Na2BO2 +
0.8-3.0
Low sensitivity Isobaric interference by
Aggarwal & Palmer (1995)
Low to medium sensitivity Less in-run fractionation The three K isotopes interfered with B mass spectrum
McMullen et al. (1961)
K2BO2+
N-TIMS
Rb2BO2 +
0.94
Low to medium sensitivity Less in-run fractionation Large Rb + ion beam
Gensho & Honda (1971) Xiao et al. (1991 )
Li2BO2 +
0.1-1.0
Low to medium sensitivity Large in-run fractionation Simultaneous isotopic measurement of both Li and B
Gensho & Honda (1971 ) Chan (1987) Sahoo& Masuda (1995)
LiNaBO2 +
1.0
Low to medium sensitivity Large in-run fractionation Simultaneous isotopic measurement of both Li and B
Chan (1987)
Cs2BO2 +
0.11-0.45
Low to medium sensitivity Negligible in-run fractionation Large Cs + ion beam
Spivack & Edmond (1986) Ishikawa & Nakamura (1990)
BO2-
1.6
Large in-run fractionation No chemistry High sensitivity No isobaric interference
Zeininger & Heumann (1983)
1.4-3.4
Same as above
Duchateau et al. (1986)
Same as above
Vengosh et al. (1989)
Large in-run fractionation No chemistry High sensitivity CNO- isobaric interference
Hemming & Hanson (1992)
0.66
(Table 7.1 continued)
146
Chapter 7 - C.-F. You
(Table 7.1 continued) Technique Measured ion N-TIMS
BO2-
Relative standard deviation (%o) 1.0
Advantages/disadvantages
References
Large in-run fractionation No chemistry High sensitivity No isobaric interference Integrate all ion beam
You et al. (1993a)
1.0
Large in-run fractionation Minimal chemical preparation High sensitivity No isobaric interference
Aggarwal & Palmer (1995)
0.58
In-run fractionation correction
You et al. (1999)
The Li-borate was used first in a preliminary investigation of boron isotopic ratios (Gensho & Honda, 1971). This compound was then applied for lithium isotope determination at masses 54 and 55 using a known boron spike (Chan, 1987). The Li-borate method has two major disadvantages because of the relative large mass difference between the two ions and the interference by natural occurring Li or B isotopes. Sahoo & Masuda (1995) utilized this compound for the simultaneous determination of Li and B isotopes; however, the potential mass fractionation artifact was assumed to be of no importance in the study. The Na-borate method using Na2B407 salt and analyzed Na2BO2 + species at masses of 88 and 89 was first proposed by Palmer (1958) and, then, was found to be susceptible to impurities in the early development. The Na-borate has a rather low ionization efficiency, which normally needs at least a few tens of micrograms of boron for each analysis. Additionally there is a large in-run isotopic fractionation during the course of an individual analysis, with the 11B/10B ratio becoming progressively heavier as the fraction of evaporated material increased. The temperature induced mass fractionation is commonly on the order of 2 to 3%0 at least (You et al., 1999). As only two stable isotopes of boron occur in nature, it is not possible to carry out an online correction. This limits the Na-borate method analytical precision to a level approximately 3%o (2o). There is a further potential isobaric interference by 88Sr§ and 89y+ if the chemical separation is not sufficiently pure or if the mass spectrometer is used also for Sr isotope analyses. The uncontrollable mass fractionation associated with positive-TIMS method can be reduced by increasing the molecular mass of the alkali-borate species analyzed (Figure 7.2). Use of the K-borate compound was first reported by McMullen et al. (1961); however, it was not a suitable compound because the presence of three K isotopes (39K, 40K and 41K) leads to overlapping K2BO2+ peaks in the mass spectrum that are not easily isolated.
Thermal Ionization Mass Spectrometry Techniques for Boron Isotopic Analysis: A Review
147
Figure 7.2 - A comparison of the isotopic fractionation curve for BO2-, Na2BO2 + and Cs2BO2+ during the analysis of SRM 951 standard boric acid assuming the thermal ionization processes on hot metal surface follow a simple Rayleigh model (after Klotzli, 1992). The masses for each molecular ion measured by various techniques are 308 and 309 for Csborate, 88 and 89 for Na-borate, 42 and 43 for BO2-. The fractionation factor used for Csborate, Na-borate, and BO2- curve calculation is 1.0016, 1.0057 and 1.0118, respectively.
The Rb-borate method was first suggested in 1971 for precise boron isotopic measurement (Gensho & Honda, 1971). As the Rb2BO2+ masses are sufficiently high at 212 to 217 amu, the fractionation problem is greatly reduced. Xiao et al. (1991) used a Rb:B ratio of approximately 2:1 and have successfully found strong Rb2BO2+ beams and obtained a relative standard deviation of 0.94%o (2o). However, the reported significant memory effect of 87Rb+ caused potential isobaric interference on 87Sr in later Sr isotopic analysis. The Cs-borate method was suggested nearly thirty years ago by Rein & Abernathey (1972); however, the first study of the method did not appear until the papers by Ramakumar et al. (1985) and Spivack & Edmond (1986). The Cs-borate analyzes the molecular ion of Cs2BO2+ at masses of 308 and 309, where the mass difference between the two is sufficiently small and thus avoiding most of the in-run mass fractionation. It is the most precise technique for boron isotopic determination currently, an average in-run precision of better than 0.2%o is normally attainable with 150 ratios (Ishikawa & Nakamura, 1993, 1994; Gaillardet & Allegre, 1995). Its low ionization efficiency however, has limited its applications in low abundance natural materials. The Cs-borate technique has been through some stages of modification after the original proposal by Ramakumar et al. (1985) and Spivack & Edmond (1986). These modifications include the use of graphite slurry (Xiao et al., 1988) or La(NO3)3 to enhance the Cs2BO2+ ionization (Klotzli, 1991), the employment of mannitol to prevent the volatil-
148
Chapter 7 - C.-F. You
ization of boron in HF digestion (Ishikawa & Nakamura, 1990; Nakamura et al., 1992), the optimization of the Cs:B ratio, as well as the optimization of boron loaded and the selection of proper filament material (Xiao et al., 1991; Aggarwal & Palmer, 1995). Due to the potential Cs deposition on the lens stacks which causes charging the lenses and drift the focus position, the Cs:B ratio was suggested to be reduced from the earlier suggestion in the range of 0.5 to 2 down to 1:7 (Aggarwal & Palmer, 1995).
7.4 Negative-TIMS The BO2- ion is one of the oxides with an electron affinity of greater than 2 eV, which form sensitively as negative ions at the surface of a hot metal filament (Heumann et al., 1995). Other similar compounds including ReO4-, OsO3- and TcO4- have also been applied as target ions for precise Re, Os and Tc negative-TIMS isotopic determination respectively. The first BO2- negative-TIMS analysis was reported by Papic et al (1979). The technique, in principle, can offer a method with high ionization efficiency, minimal chemical preparation, free of isobaric interference and can achieves a comparable level of precision as that of the positive-TIMS under strictly controlled conditions. The major disadvantage of the BO2- negative-TIMS, however, is the large mass difference between the masses analyzed at 42 and 43 amu. This causes severe mass dependent isotope fractionation during the course of the isotopic analysis, leading to the attainable precision of only 0.8%o (2o) under strictly controlled mass spectrometric running conditions (Figure 7.2). Even worse, the determined 11B/10B ratio for NBS 951 standard boric acid by BO2-negative-TIMS varies from laboratory to laboratory around the world and shows large discrimination compared with the NBS certified ratio or ratios obtained by the Cs-borate positive-TIMS method (Table 6.2). The most likely cause of the inconsistent NBS 951 results is the degree of mass fractionation, which must vary from laboratory to laboratory and from sample to sample. Table 7.2 - Comparison of boron isotopic composition of NBS 951 boric acid measured by various TIMS techniques Technique
Measured ion
11B/ 10B
Reference
P-TIMS
Cs2BO2+
4.0456 4.0503 4.0512 4.0529 4.0533 4.0506 4.0357
Spivack and Edmond(1986) Xiao et al. (1988) Nakamura et al. (1992) Aggarwal and Palmer(1995) Zhai et al. (1996) Tonarini et al. (1997) Sahoo and Masuda (1995)
4.018 4.010 3.996 4.001 4.001 3.987 4.043
Zeininger and Heumann (1983) Vengosh et al. (1989) Vengosh et al. (1991) Hemming and Hanson (1992) Barth (1998) Palmer et aI. (1998) You et al. (1999)
Li2BO2+ N-TIMS
BO2-
Thermal Ionization Mass SpectrometryTechniques for Boron Isotopic Analysis: A Review
149
You et al. (1993a) have examined a method by integrating the ion beam intensity until the entire sample is exhausted to improve the precision; however, the technique is not very practical in terms of machine time consumed. Detailed discussion on a potential method to normalize the in-run mass fractionation will be followed below. Various reagents, including CaC12, BaC12, NaNO3, Ca(NO3)2, and La(NO3)3, have been used as activator to load together with sample during the BO2- negative-TIMS (Heumann & Zeininger, 1985; Xiao et al., 1988; Klotzli, 1992; Hemming & Hanson, 1992). It is believed that these activators lower the electron work function of the filament and thus enhance the ionization efficiency by two to three order of magnitudes. A similar observation established that boron-free seawater matrix serves also an excellent activator for aqueous solutions with seawater-like matrix, such as porewaters or hydrothermal vent fluids on seafloor. Hemming & Hanson (1994) reported potential CNO- interference at 42 amu, which was possibly generated by organic matter decomposition and can be removed by peroxide treatment or ultrafiltration before samples loaded onto filament. It is advisable to monitor carefully of CN- peak at 26 amu to avoid possible isobaric contamination at 42 amu. Hemming & Hanson (1994) advocate rejecting any analysis with signal at mass 26 amu. However, recent studies have demonstrated that most natural fluids can be analyzed directly using the BO2negative-TIMS technique without any chemical separation, except for those of complex samples with high organic content (Spivack & You, 1997). 7.5 Mass fractionation correction for BO2- negative-TIMS The BO2-negative-TIMS technique is characterized by a considerably higher analytical sensitivity compared to the positive-TIMS, thus allowing the methodology to be applied for low abundance environmental samples. However, the associated large analytical uncertainty has prevented it from use in any high-resolution geochemical application.
The uncontrollable instrumental mass fractionation is one of the most important factors affecting the analytical precision of the isotopic ratio in mass spectrometry, in particular for those of low mass isotopes (e.g. B and Li). Normally for elements with more than three isotopes, one can normalize one ratio between one pair of isotopes to a standard value and correct the mass fractionation effect on other ratios (e.g. Sr and Nd isotopic analyses). It is, however, not possible to apply the internal normalization procedure for two-isotope elements as there is only one ratio available. Fortunately, there are many two-isotope elements which form oxide ions in TIMS (e.g. BO2-, LaO + and CeO+), with ionization efficiencies much higher than metal ions (Shen et al., 1992). For BO2-negative-TIMS, theoretically there must be at least six oxide ions appearing at different masses with four peaks that can be measured precisely. This opens the possibility of normalizing the ratio between oxide ions to that between the oxygen isotopes. For instance, I45/I43 (= 11B160180/11B160160) can be normalized to 180/160.
150
Chapter 7 - C.-F. You
The traditional BO2- negative-TIMS measured only signals at masses 42 and 43 amu and hence no mass fractionation correction could be made. Several potential analytical problems may have affected literature data that applied the same technique. Firstly the measured 11B/10B ratio of NBS 951 varies from laboratory to laboratory and shows large deviations from the certified ratio of 4.0436 or from Cs-borate positive-TIMS results (Table 6.2). Secondly a strict control in TIMS running parameters is required to avoid changes in the degree of fractionation among samples. Thirdly there is no evaluation of at what stage data acquisition needs to be started or abandoned during analysis. Fourthly there occurs a random rejection of unwanted blocks during data reduction to achieve high in-run precision. Lastly the potential artifacts due to intrinsic oxygen isotopic variation or oxygen fractionation are neglected. All the difficulties will be removed if the fractionation artifacts can be normalized. There are six possible nuclide combinations of the BO2- molecular ions, in a mass range between 42 and 47, and the first four peaks can be measured with sufficient precision. From statistical consideration, the measured intensity ratios for I43/I42, I44/I42, I45/I42 can be expressed as the following formula: I43 / I42 = B+2R17 I44/I42 = 2BR17+2R18+R17R17 I45 / I42 = 2BR18+BR17R17+2R17R18
[7.2] [7.3] [7.41
where B is the 11B/10B ratio; R17 and R18 represents 1 7 0 / 1 6 0 and 1 8 0 / 1 6 0 ratio respectively. There are three independent equations and in principle, precise values of B, R17, and R18 can be calculated by solving simultaneous quadratic equation described. Unfortunately, the BO2- fractionation does not follow strictly any known mass fractionation laws and thus it became more complicated to carry out the correction procedures. Mathematically one can derive two more equations of I45/I42 and I45/I43.
I45 /I43 = (2BR18+BR17R17+2R17R18)/ (B+2R17) ~ 2R18 I45/I42 = 2BR18+BR17R17+2R17R18 ~ 2BR18
[7.5] [7.6]
It is clear from the equation [6.5] and [6.6] that I45/I43 and I45/I42 represents actually 2R18 and 2BR18 respectively. It will become extremely useful for normalization correction if any systematic variation between the two variables can be observed. In other words, the internal fractionation correction will be possible if the slope in the I45/I43 versus I45/I42 plot does not change with time. Indeed we have observed very promising preliminary results (You et al., 1999). There are a few exciting observations in the I45/I42 versus I45/I43 plot presented in You et al. (1999). The two intensity ratios, indeed, show large but well-correlated variations. There are even larger variations from run to run, but they follow a similar correlation for the same sample. This implies that one is able to use an empirical 180/160 ratio derived from individual machine for normalization. You et al. (1999) report a
Thermal Ionization Mass Spectrometry Techniques for Boron Isotopic Analysis: A Review
Figure 7.3 - (a) Variation of I43 / I42 intensity ratios against I45 / I43 (~-2.180 / 160) for SRM 951 in duplicated analyses using BO2-negative TIMS. Note a large variation of I43/I42 ratios before the 180/160 normalization correction was made. These results show good correlation between the I43 / I42 and 180 / 160. (b) Variation of I45 / I42 intensity ratios against I45 / I43 (~2" 180 / 160) for SRM 951 in duplicated analyses using the BO2- negative TIMS technique. Twenty-nine duplicated analyses show excellent correlation between the I45/I42 and 180/ 160 ratios. Using a common 180/160 of 0.002143 in our MAT 262 (You et al., 1999), we obtained an average 11B/ 10B ratio of 4.0429 with excellent precision 0.58%o (2a) for SRM 951 after applying the 180/160 normalization procedures.
151
152
Chapter 7 - C.-F. You
value of 4.0429 +__0.0003 (2or - 0.58%o) based on 29 duplicated analyses of NBS 951 boric acid. The measured ratios of I43/I42 and I45/I42 are plotted along with the 180/160 in Figure 7.3. Before any correction, the NBS 951 has a m e a n l l B / I ~ value of 4.015 + 3.5%o with a variation range between 4.01007 and 4.02355, which differs significantly from the certified ratio. After the mass fractionation correction based on 180/160, the mean value for NBS 951 is 4.0429 + 0.0003, which agrees excellently with the certified or positive-TIMS results. The analytical uncertainties involved in this procedure mainly stem from the 180/160 measurement due to the low abundance at 44 amu. This technique maintains the high ionization efficiency (1000 times better than Cs2BO~- ), but also improves the analytical precision a few times to a similar level as Cs2BO~- positive-TIMS. The ability to resolve a small change in boron isotopic compositions is critical for further evaluation of oceanic pH variation in the past based on foraminiferal shells (Palmer et al., 1998; Sanyal et al., 1995; Spivack et al., 1993). mean llB/I~
7.6 Conclusions
Over the last ten years there has been a considerable increase in the application of the boron isotopes as a geochemical tracer in earth sciences. A variety of instrumental techniques for boron isotope analyses have been developed and are now available for selection to fit various research goals or types of samples. The Cs-borate positive-TIMS and the BO2 negative-TIMS are the most commonly used techniques today. The Cs-borate technique achieves a better precision, but with low ionization efficiency and lengthy chemical preparation. The modified BO2 negative-TIMS applying the oxygen normalization may attain a similar precision level as Cs-borate, but can be applied to low abundance natural samples with simple chemistry. By determining the BO2 peak abundance at mass 42, 43, 44, and 45 simultaneously, the effect of instrumental fractionation can be corrected based on the 180/160 ratios. The new BO2 negative-TIMS technique is crucial for re-evaluation of small change of 11B/1~ in foraminferal shells and to check possible oceanic pH variation in the past.
Acknowledgements
The author thanks constructive reviews by Dr. T. Ishikawa and Dr. G. H. Swihart, which improved the content of this paper significantly. I thank Dr. Joris Gieskes to help for improving English in the earlier draft. Drs. J. Shen and T. Lee provide wonderful cooperation in conducting the B02 negative-TIMS experiments at the Institute of Earth Sciences, Academia Sinica. This work is supported by MOE and NSC Taiwan, Republic of China.
Handbook of Stable Isotope Analytical Techniques, Volume 1 P.A. de Groot (Editor) 9 2004 Elsevier B.V. All fights reserved.
CHAPTER 8 GC and I R M S T e c h n o l o g y for 13C and 15N A n a l y s i s on Organic C o m p o u n d s and Related G a s e s Wolfram Meier-Augenstein Queen's University Belfast, Environmental Engineering Research Centre, David Keir Building, Belfast, BT9 5AG, UK e-mail:
[email protected]
General preface One of the aims of this book and, hence, this chapter is to provide both novices and the well seasoned "isotopist" with a reference work presenting the status quo in stable isotope work as well as practical guidelines. To accommodate all that, a general prefAbbreviation/ Meaning Acronym APE
Atom %Excess
CSIA GC GC / C-IRMS
HVOC IRMS
Compound Specific Isotope Analysis Gas Chromatography Gas Chromatography - Combustion Isotope Ratio Mass Spectrometry Gas Chromatography - Mass Spectrometry High Performance (Pressure) Liquid Chromatography High Resolution Chromatography High Resolution capillary Gas Chromatography High Temperature capillary Gas Chromatography High-Volatile Organic Compound Isotope Ratio Mass Spectrometry
LC MDGC MS PLOT SIA SIM tBDMS TLC TMS VOC
Liquid Chromatography Multi-Dimensional Gas Chromatography Mass Spectrometry Porous Layer Open Tubular (column) Stable Isotope Analysis Selected Ion Monitoring tert-B u tyl dimeth ylsil yl Thin Layer Chromatography Trimethylsilyl Volatile Organic Compound
GC-MS HPLC HRC HRcGC HTcGC
Other acronyms used in the literature
GC-C-IRMS; irmGC-MS
GIRMS; GIMS for Gas Isotope Ratio Mass Spectrometry
MSD for Mass Selective Detection
154
Chapter 8 - W. Meier-Augenstein
Figure 8.1 - Set-up of an isotope ratio mass spectrometer coupled to a gas chromatograph via a combustion interface to measure 13C/12C (Carbon mode) or 15N/14N ratios (Nitrogen mode). This schematic shows the reference gas set-up used for automated internal isotopic calibration, ref. = reference.
ace and a table of the most commonly used abbreviations precede this chapter. Gas isotope ratio mass spectrometry (GIRMS or simply IRMS) is probably the oldest type of mass spectrometry used in analytical chemistry. IRMS has been a standard tool in areas such as geochemistry, quaternary sciences and environmental sciences. However, only since the commercial availability of IRMS instruments coupled to a gas-chromatograph (GC) via a combustion interface (C) (Figure 8.1) in 1990, has IRMS received the attention of other areas of applied analytical chemistry; areas where GCMS and LC-MS are commonly used for analytes which are part of a complex matrix or are present in variable concentrations. In contrast to organic mass spectrometers (MS) that yield structural information by scanning a mass range over several hundred Dalton for characteristic fragment ions, IRMS instruments achieve highly accurate and precise measurement of isotopic abundance at the expense of the flexibility of scanning MS. Since GC-MS can be used to measure stable isotope enrichment, the question arises why one should embrace GC/ C-IRMS. Scanning mass spectrometers use a single detector and therefore cannot simultaneously detect particular isotope pairs for isotope ratio measurement. For isotope ratio measurement, the MS is best operated in selected ion monitoring mode (SIM) to optimise sensitivity to selected masses. Even in SIM mode, limited accuracy and precision of such isotope ratio measurements impose a minimum working enrichment for 13C and 15N of at least 0.5 atom% excess (APE) (Preston & Slater, 1994; Rennie et al., 1996). In other words, organic MS cannot provide reliable quantitative information in cases where low turnover or low rate of incorporation results in isotopic enrichment of less than 0.5 APE.
GC and IRMS Technology for 13C and 15N Analysis on Organic Compounds and Related Gases
155
In contrast, GC/C-IRMS can measure isotopic composition at low enrichment and natural abundance level. This means that minute variations in very small amounts of the heavier isotope are detected in the presence of large amounts of the lighter isotope. Since the small variations of the heavier isotope habitually measured by IRMS are of the order of-0.07 to +1.09 APE, the g-notation in units of per mil [%0] has been adopted to report changes in isotopic abundance as a per mil deviation compared to a designated isotopic standard: a s - ([Rs-Rstd] / Rstd) x 1000 [%o]
[8.1]
where Rs is the measured isotope ratio for the sample and Rstd is the measured isotope ratio for the standard. To give a convenient rule-of-thumb approximation, in the 5-notation, a 13C enrichment in the range of-0.033 to +0.0549 APE corresponds to 613C value range of-30%o to +50%o. A change of +1%o is approximately equivalent to a change of 0.001 APE and 0.0003 APE for 13C and 15N, respectively. The sensitivity of GC/C-IRMS is such that tracer/tracee (mol/mol) ratios down to 10-5 can reliably be detected (Brenna et al., 1997); in the same review, Brenna et al. also provide an in-depth discussion of notations and elementary calculations such as mass balance and pool mixing equations. Due to its high sensitivity, GC/C-IRMS depends on careful sample preparation and high-resolution capillary GC (HRcGC) (Meier-Augenstein, 1999). Demands on sample size, sample derivatization, quality of GC separation, interface design and isotopic calibration have been discussed in a number of reviews (Brenna, 1994; Brand, 1996; Meier-Augenstein, 1997a; Ellis & Fincannon, 1998; Metges & Petzke, 1999). 8.1 Introduction
That increased sensitivity automatically leads to improved results in analytical chemistry is often asserted, hardly ever proven, and, in fact, false. What initially appears to be a slap in the face of instrument manufacturers and analytical chemists alike, under closer scrutiny proves to be nothing but the sum of day to day experience. Pushing the detection limits down further and further, it becomes increasingly likely that minority components introduced during sample collection and sample preparation will interfere with the analysis, i.e. detection and quantification, of the actual target compounds. These minority components (otherwise referred to as dirt, rubbish or muck) may originate from solvent impurities, stationary phases used in TLC or ion exchange resins to name but a few. In our case, the analysis of isotopic abundance in a given organic compound, the introduction of compounds alien to the original sample can have a detrimental effect on accuracy as well as precision. Accuracy will almost always be impaired, as the presence of a compound, chemically identical to the target compound, yet from a different source, will adulterate the true isotopic composition of the target compound. Accuracy and precision will both deteriorate when the target compound partially reacts with another compound, e.g. in the injector of the GC. Unresolved adjacent GC
156
Chapter 8 - W. Meier-Augenstein
peaks will also compromise accuracy and precision because even a small peak overlap will contribute part of the isotopic signature of the adjacent compound peak to that of the target compound. Overlapping peaks that lead to isobaric interference or gas reactions in the ion source can further compromise precise measurement of isotope ratios. In addition to the aforementioned, there is another source of error that is unique to compound specific isotope analysis (CSIA), namely mass discrimination or isotopic fractionation. In principle, two different types of isotope effects can cause isotopic fractionation, kinetic isotope effects and thermodynamic isotope effects. In general, isotope effects are caused by differences in vibration energy levels of bonds involving heavier isotopes as compared to bonds involving lighter isotopes. This difference in bond strength can lead to different reaction rates for a bond when different isotopes of the same element are involved (Melander & Saunders, 1980). The most significant kinetic isotope effect is the primary isotope effect, whereby a bond containing the atom or its isotope in consideration is broken or formed in the rate-determining step of the reaction. Rieley has presented an excellent in-depth discussion of kinetic isotope effects and associated theoretical considerations in 1994 (Riele)~ 1994). The second kind of isotope effect is associated with differences in physico-chemical properties such as infrared absorption, molar volume, vapour pressure, boiling point and melting point. Of course, these properties are all linked to the same parameters as those mentioned for the kinetic isotope effect, i.e. bond strength, reduced mass and, hence, vibration energy levels. However, to set it apart from the kinetic isotope effect, this effect is referred to as thermodynamic isotope effect (Meier-Augenstein, 1997a) because it manifests itself in processes where chemical bonds are neither broken nor formed. Typical examples for such processes in which the results of thermodynamic isotope effects can be observed are infrared spectroscopy, distillation and any kind of two-phase partitioning. This thermodynamic isotope effect, or physicochemical isotope effect, is the reason for the higher infrared absorption of 13CO2 as compared to 12CO2, the enrichment of ocean surface water with H2180 and the isotopic fractionation observed during chromatographic separations. Hence, every step of sample manipulation from sample collection over sample preparation, sample injection to sample separation can result in isotopic fractionation of the target compounds. It is therefore the aim of this chapter to point out potential pitfalls in the on-line isotopic analysis of organic compounds and related gases. At the same time, its intention is to provide general guidelines to aid with sample preparation, GC analysis and troubleshooting for on-line IRMS. Since this book is aimed at a wide audience ranging from the novice to the well-seasoned "isotopist", this chapter endeavours to meet the anticipation of both.
8.2 Sample preparation It may sound trivial but it must be clearly understood that the quality of results of hyphenated analytical techniques in general and CSIA in particular cannot be better than the quality of the sample analysed. In other words, sample preparation is a crucial part of the whole analysis because it constitutes more often than not the perfor-
GC and IRMS Technologyfor 13Cand 15NAnalysison OrganicCompoundsand RelatedGases
157
mance-limiting step. For high-precision CSIA by GC/C-IRMS close attention must be paid to the following points: 9 Every step of the sample preparation protocol (collection, work up, derivatization) must be scrutinised for potential mass discriminatory effects to avoid isotopic fractionation of the target compounds. ~ If the potential of isotopic fractionation cannot be ruled out conclusively, an internal standard, of a similar chemical nature (but not requiring derivatization) and of known isotopic composition, should be added to the sample prior to the sample preparation procedure. 9Signal size and isotopic composition of the standard(s) must match those of the analyte(s) (Brenna et al., 1997). ~ The potential of all GC parameters (polarity of stationary phase, carrier gas management, temperature programme) and techniques should be exploited to their fullest. 9 The isotopic signature of the derivatization agents used must be homogenous throughout the duration of a project involving GC/C-IRMS. This is best achieved by acquiring a large stock from the same batch and by appropriate storage. The latter may include storage over drying agents, at low temperatures, under an inert gas and not exposed to light.
8.2.1. Sampling and sample preparation Gases and volatile organic compounds (VOCs) One area of stable isotope analysis (SIA) to benefit from the introduction of GC/CIRMS with respect to sampling and sample preparation is the analysis of gases. With off-line techniques, CSIA of gases was all but confined to permanent gases such as CO2, N2, 02 and noble gases. CISA of natural gases, for instance, was only possible by a laborious and time-consuming procedure in which samples of natural gases were separated and individual sample components were cryogenically collected using prep-GC on packed columns. After combustion and separation of CO2 from H20 by fractionated distillation, the resulting CO2 samples were analyzed off-line by dualinlet IRMS. This all changed with the commercial availability of high sensitivity GC/C-IRMS systems that coincided with the advent of porous layer open tubular (PLOT) fused silica capillary columns. Stationary phases for gas analysis such as Chromosorb 106, HayeSep N or Carbosphere that hitherto could only be used on packed columns became available for capillary GC. The combination of PLOT capillary columns and GC/C-IRMS not only reduced analysis time but also sample size and simplified sample preparation (Table 8.1). Whereas previously large sample volumes had to be collected and concentrated (for trace gases at ppb level up to 800 L of air are needed), the new technology meant that sample sizes ranging from 175 mL to as little as 50 ~tL were sufficient (Brooks & Atkins, 1992). Samples can either be directly injected on to the PLOT column (Baylis et al., 1994; Merritt et al., 1995b) or purged on to the column from an on-line cryogenic pre-concentration set-up (Brand, 1995a; Rudolph et al., 1997; Ehleringer & Cook, 1998).
158
Chapter 8 - W. Meier-Augenstein
The success story of this technique however, should not blind us for the pitfalls encountered with gas SIA. The most serious threat to accurate and precise isotope abundance measurement in gaseous compounds is isotopic fractionation. In this case, it is almost exclusively the thermodynamic isotope effect that causes isotopic fractionation. Small molecules or, in other words, compounds of low molecular weight are much more susceptible to mass discrimination caused by differences in molar volume, vapour pressure and boiling point. It is for this reason that gas samples are exclusively injected in splitless mode (Table 8.1) to avoid mass discrimination in the injector. Mass discrimination when dealing with gases and VOCs can occur during sampling by adsorption, during sample preparation, e.g. water and/or CO2 trapping by adsorption and even during GC analysis. What these procedures have in common is that they are all systems of twophase partitioning. Slight differences in association and dissociation constants for compound/matrix interaction will lead to mass discrimination. Baylis et al. (1994) observed isotopic fractionation for CO2 and highly volatile hydrocarbons on an Alumina-KC1 PLOT column which they attributed to adsorption phenomena. Sohns et al. (1994) reported mass discrimination for N2 on cooled molecular sieve 5~ suggesting a weaker interaction between the adsorbent and 15N14N as compared to 14N14N. They also noted that the Molsieve 5~ packed column, used in their GC, retained CO2 that subsequently bled off the column causing mass interference in the IRMS. Trapping water by means of adsorption has also been reported to cause isotopic fractionation of CO2. A minor shift towards higher 613C-values of +1.43%o vs. baseline was observed using MgC104 as desiccant. Using molecular sieves of 4~ and 5~ however resulted in shifts in 613C-values vs. baseline of +13.43%o and +21.44%o, respectively (MeierAugenstein et al., 1994a). Not surprisingly, the isotopic fractionation was even more pronounced for the 6180-values for CO2. Here, the shifts in 6180-values were +24.87%o and +7.97%o for molecular sieve 4~ and 5~,, respectively. The only drying agent that, like cryogenic water trapping, did not cause any isotopic fractionation was molecular sieve 3~,. Trapping trace gases in ambient air or breath is fraught with the risk of loosing precious material on the adsorbent (Kohlmuller & Kochen, 1993) and the same is true for sample preparation of VOCs by dynamic headspace or extraction with diethyl ether (Schumacher et al., 1995).
Organiccompounds
One could be forgiven for assuming organic compounds with a molecular weight above 150 amu are less prone to isotopic fractionation than gases and VOCs. However, there is evidence to the contrary. Hofmann et al. observed a shift of 615N-values to more negative values of approx. 2%0 for amino acids that were isolated by ion exchange chromatography (Hofmann et al., 1995). Caimi and Brenna (1997) reported that the beginning of an HPLC peak had an isotope ratio sharply enriched relative to the parent material, while the end of the peak is mildly depleted. These observations show that quantitative peak collection of the entire LC peak is important for accurate isotope ratio analysis. Hence, LC techniques must be applied with caution when used
C3 r
Table 8.1 - Methods used for GC analysis of permanent gases and volatile organic compounds (VOCs). Target Compound(s)
Column
Typical Set-Upa)
Remarksa)
C1 to C6 n- and iso-hydrocarbons from natural gas samples (e.g. headspace of oil wells).
PORAPLOT Q
carrier gas: Helium; flow rate: 2.5 ml / min (0.32 mm I.D.) 80 ~ (5 min) to 200 ~ at 5~
C1 to C6 alcohols, aldehydes, ketones and esters as trace constituents (e.g. in breath).
PORAPLOT U
C2 to C5 aliphatic and olefinic
PORAPLOT Q
splitless injection of 50 250 ~1; injected sample is cryo-focused on column (-40 ~ C, 2.5 min). splitless injection; target compounds arepre-concentrated on-line from 100 ml of sample prior to injection by cryo-trapping (liquid N2). splitless injection; target com pounds are pre-concentrated on-line from 175 ml of sample prior to injection by cryo-traping (liquid N2). splitless injection; target compounds are pre-concentrated on-line from as little as 300 ~1 of ambient air for CO2 analysis. Molsieve 5~ retains CO2 and can thus lead to memory effects resulting in massinterference in the IRMS.
hydrocarbons in ambient air.
CO, CO2, N2, NO, N 2 0 in ambient air.
02, N2, CH4, CO
PORAPLOT Q
Molsieve 5A
carrier gas: Helium; flow rate: 10 ml / min (0.53 mm I.D.) 50 ~ (5 min) to 150 ~ at 20~ carrier gas" Helium; flow rate: i ml / min (0.32 mm I.D.) 35 ~ (10 min) to 200 ~ at 3~ carrier gas: Helium; flow rate: 1.5 ml / min (0.32 mm I.D.) 25 ~ isothermal carrier gas: Helium; flow rate: 10 ml / min (0.53 mm I.D.) 100 ~ isothermal
a) The information given in these columns should be read as generalized guidelines providing a starting point for the reader to resolve individual analytical tasks.
~q C~
=r o 9 0xa o
r
La,a
r L*l
> ~,,,i.
9
b,,,i 9
r9 9
r~
C3 r~
L,l
160
Chapter 8 - W. Meier-Augenstein
for sample preparation or sample clean up of complex mixtures for isotope ratio analysis. Again, we are dealing with a form of two-phase partitioning where slight differences in solute/stationary phase interaction leads to mass discrimination. Very recently, Filer presented a comprehensive overview of isotopic fractionation during chromatography (Filer, 1999).
8.2.2. Sample derivatization It is quite remarkable that publications of studies dedicated to the influence of sample derivatization on high-precision CSIA by GC/C-IRMS are few and far between (Meier-Augenstein, 1997a; Meier-Augenstein, 1999). Observations to this effect are usually passed on by word of mouth, be it at conferences and, more recently, as postings to discussion groups on the internet. Since one of this book's intentions is to serve as a reference book, I would like to take this opportunity to bring some of these none the less valuable observations into the public domain. When derivatizating a sample prior to CSIA, one should bear in mind the various ways this can influence the analysis and its results. (I) Isotopic fractionation due to kinetic isotope effects during derivatization. (II) Change of isotopic composition in the target compound(s) due to derivatization agent(s), e.g. tracer dilution. (III) Gas chromatographic consequences. (IV) Potential adulteration of isotope ratio measurement caused by non-quantitative sample conversion.
Isotopicfractionation
During derivatization reactions, kinetic isotope effects come only to bear if one of the reagents reacts non-quantitatively. Any such effect associated with derivatization reactions and theoretical considerations for calculating true ~513C-values from derivatized compounds have been discussed in a recent article (Rieley, 1994). However, in most cases the possibility of kinetic isotope effects during derivatization can be safely ignored. The most widely used derivatization reactions can be divided in three classes, silylation, esterification and acetylation. Silylation, which replaces the hydrogen in a hydroxyl or an amino group with an alkylated silyl group, will not cause a carbon isotopic fractionation since no carbon-containing bond is involved. Since silylation is carried out using a huge excess of the silylation agent it is not likely to cause isotopic fractionation of oxygen or nitrogen either. In esterification and trans-esterification reactions, a carbon centre contained in the final product is involved and, hence, a carbon kinetic isotope effect is possible. These reactions usually employ an excess of derivatization agent, e.g. an alcohol, with a catalyst. Since the carbon of interest is in the compound and these reactions are rapid and usually quantitative, no carbon kinetic isotope effect will be expressed. In acetylation reactions, however, the situation is reversed. The carbon involved in the rate-determining step is in the derivatization agent rather than in the compound of interest. Since the derivatization agent is used in excess and, hence, will not react quantitatively, a carbon kinetic isotope effect at the acetyl carboxy-carbon is possible although the compound of interest may react quantitatively.
GC and IRMS Technology for
13Cand 15N Analysis on Organic Compounds and Related Gases
161
Change of isotopic composition Changes of the 13C/12C isotope ratio due to derivatization are usually a minor consideration unless the true isotopic composition of the target compound(s) must be known. Khalfallah et al. (1993) reported a method to correct for 13C-tracer dilution through carbon added during derivatization of plasma lactate. Demmelmair and Schmidt (1993) described a carbon balance equation to calculate ~3C-values of free amino acids from ~13C-values of their N-acetyl, O-propylate (NAP) derivatives at natural abundance level. However, in most biochemical applications such as in-vivo or in-vitro tracer studies the interest is focused on changes in isotopic enrichment rather than the absolute values. This is easily achieved since measurements are compared against baseline or background samples that have undergone the same sample preparation procedure.
Effects on gas chromatographic separation Next, the effects of derivatization on sample integrity and the quality of GC separation have to be considered. For a variety of reasons, silylation agents have enjoyed increasing popularity over the last decade. Derivatization introducing trimethylsilyl (TMS) or tert-butyl-dimethylsilyl groups (tBDMS) is usually quantitative, reaction conditions are relatively mild, the reactions as such are easily performed and samples can be injected directly after the derivatization has been successfully completed. However, there are other reasons as to why silylation agents should be used with caution when it comes to derivatizing samples for GC/C-IRMS. Derivatization of amino acids and a-ketocarboxylic acids with TMS leads to formation of several derivatives for one compound (i.e. mono-, di-, tri-TMS and cis/trans isomers). Sample stability is another concern. Recent observations with tBDMS-derivatives of amino acids obtained from protein hydrolysates showed that these derivatives can deteriorate in a matter of days despite being stored in dark glass vials at subzero temperature (Hofmann et al., 1995). Derivatization by silylation might also hamper GC separation as the extremely apolar nature of trimethylsilyl (TMS) and tertbutyldimethylsilyl (tBDMS) derivatives can obscure compound characteristics that could otherwise be gas chromatographically exploited. For example, despite the entirely different chemical nature of the pure amino acids, the tBDMS-derivatives of L-aspartic acid and L-4-hydroxyproline exhibit similar gas chromatographic properties. A temperature gradient of 5 ~ starting at 80~ after a 5-minute isothermal step leads to a 100% peak overlap of the two compounds. Only by employing three different temperature gradients separated by two isothermal steps, can a baseline separation be achieved. Furthermore, derivatization with TMS or tBDMS limits one's choice of GC stationary phases to apolar phases such SE30 and SE52/SE54 (for stationary phases and commercial equivalents, please, refer to Table 8.2). Lastly, sample derivatization can interfere with sample conversion into CO2 and N2 thus potentially compromising both accuracy and precision of isotope ratio measurements. For instance, an excessive carbon load introduced by derivatization might result in incomplete combustion thus compromising accurate isotopic analysis. The tBDMS-derivative of L-leucine contains 18 carbons (6 from leucine plus 6 per tBDMS
162
Chapter 8 - W. Meier-Augenstein
group) and one nitrogen atom per molecule. That means in order to generate I moleequivalent of N2, 36 mole-equivalents of CO2 will be generated. Segschneider et al. (1995) reported that accurate 615N from tBDMS were obtained only when more than 7 nmol of N2 per individual amino acid were introduced into the ion source of the IRMS. For leucine this means that in this case 252 nmol of CO2 were generated; this amount is roughly equivalent to a volume of 5.6 mL. One practical implication of this is that on average the CO2 trap will clog up after three runs only. This excessive carbon load can also lead to a temporary overload of the oxidation catalyst, thus resulting in the additional formation of carbon monoxide. Carbon monoxide cannot be trapped cryogenically and, hence, will enter the ion source together with N2 thus causing isobaric interference with the measurement of N2 isotope ratios (Rennie et al., 1996). Potential interference of silyl groups with the oxidation catalyst should also be taken into consideration. As yet, a potentially negative influence of SiO2, formed during combustion of silylated compounds, on the performance of the oxidation catalyst has not been ruled out conclusively. The formation of siliceous deposits on the oxidation catalyst is one possibility as this would not only reduce its reactive surface area but would also impede gas flow through the combustion reactor. For the same reason, i.e. non-quantitative sample conversion due to interference with the oxygen donor / oxygen sink system, the use of trifluoroacetates (TFA) or heptafluorobutyrates (HFB) should be avoided (Hofmann et al., 1995). Fluorine forms extremely stable fluorides with Cu (CuF2) and Ni (NiF2) thus irreversibly reducing combustion efficiency of the CuO/NiO system. In addition, fluorine irreversibly poisons the platinum that most on-line combustion-IRMS systems use in their combustion reactors. If CuO/Pt is used as combustion catalyst, this effect will show up immediately. Co-injecting equimolar amounts of N-TFA, O-propyl L-leucine and methyldecanoate (both compounds contain 11 carbons) on to a GC/C-IRMS system fitted with a fully oxidized CuO/Pt combustion chamber yielded a peak area for the amino acid that was only half the size of that obtained from the fatty acid (MeierAugenstein, 1997a). In other words, for the TFA-derivatized amino acid only 50% of the expected CO2 yield was produced. Combustion systems based on C u O / N i O / P t seem not immediately effected. This is probably due to the large oxygen sink capacity of the NiO but these systems will deteriorate with time (depending on sample throughput but usually after 100 - 150 sample injections), too.
Overview of derivatization methods Naturally, there are circumstances in which the risks mentioned above can be deemed acceptable. Yet, this decision lies solely with the analyst. Of course, one way of avoiding the problems with derivatization is not to derivatize the sample at all. For free fatty acids, amines and alkanols, this approach involves the use of moderately polar to polar stationary phases and high-temperature GC (Table 8.2). However, not every polar compound is amenable to these techniques, e.g. amino acids, and hightemperature capability of polar stationary phases is limited even when oxygen free helium is used as carrier gas. In the following, amongst others, derivatization
GC and IRMS Technology for 13C and 15N Analysis on Organic Compounds and Related Gases
163
Table 8.2- Stationary phases, their commercial equivalents and typical applications. Stationary Phasea)
Commercial Equivalents
Applications
SE 30 apolar; 100% polymethylsiloxane, 0% functionalities
DB-1, CP-Sil 5 CB, BP1, Rtx-1, HP Ultra-l,
Aliphatic, olefinic and aromatic hydrocarbons; long chain fatty acids and alcohols; waxes; steroids; TMS andtBDMS derivatized compounds.
SE 52 (SE 54)b) apolar; 5% phenyl
DB-5, CP-Sil 8 CB, BP5, Rtx-5, HP Ultra-2
Aliphatic, olefinic and aromatic hydrocarbons; long chain fatty acids and alcohols; waxes; steroids; TMS andtBDMS derivatized compounds; NPP derivatized amino acids.
OV-1701 moderately polar; 7% cyanopropyl, 7% phenyl
DB-1701, CP-Sil 19 CB, BP10, Rtx-1701, HP-1701
Fatty acid methyl esters (FAME); amino acids as NAP derivatives, TFA methylates / propylates, methylor ethyl-chloroformates; alcohols.
OV-17 moderately polar; 50% phenyl
DB-17, CP-Si124 CB, Rtx-50. HP-17
Methyl- and ethylcloroformates of amino acids; ethylchloroformates of organic acids.
OV-225 moderately polar; 25% cyanopropyl, 25% phenyl
DB-225, CP-Si143 CB, BP225, HP-225, Rtx-225
FAME, phenols, ethyl-chloroformates of organic acids; NAP derivatives of amino acids.
Carbowax polar; PEGc)
DB-Wax, CP-Si152 CB, BP20, HP-Wax, HP-20M, Stabilwax
Aldehydes; ketones; esters; alcohols; amines, flavours; FAME; amines.
OV-351 polar; PEG nitroterephthalic acid ester
DB-FFAP, CP-Wax 58 CB, BP21, HP-FFAP
Free fatty acids; flavours; FAME.
Silar 7CP highly polar; 75% cyanopropyl, 25% phenyl
BPX70
FAME; NAP derivatives of amino acids; TFA methylates/ propylates of amino acids.
Silar 10C highly polar; 100% cyanopropyl
CP-Sil 88 CB
FAME.
a) Stationary phases are listed in order of increasing polarity from top to bottom. b) Strictly speaking, SE 54 is not the same as SE 52. However, since there are no commercially available columns coated with 5% phenyl, 1% vinyl polymethylsiloxane, it is usually listed together with SE 52 equivalents (5% phenyl polymethylsiloxane). c) PEG = polyethylene glycol
164
Chapter 8 - W. Meier-Augenstein
methods for amino acids are mentioned, the products of which will not interfere with one-line sample combustion.
Trimethylsilylates and tert-butyldimethylsilylates (TMS and tBDMS) Silylation agents still enjoy considerable popularity for derivatization of polar functionalities such as hydroxyl- (OH), amino- (NH2) and thiol- (SH) groups. Derivatization introducing TMS or tBDMS groups is usually quantitative, reaction conditions are relatively mild and the reagents can be used off the shelf. In cases where the ready-to-use silylation reagents do not achieve complete derivatization, e.g. OHgroups in sterols, the addition of a reaction promoting base such as pyridine or triethylamine helps to achieve a quantitative reaction. However, as pointed out earlier, silylation should be regarded as a last resort when it comes to sample preparation for GC / C-IRMS. Apart from the potential formation of several derivatives for one compound (i.e. mono-, di-, tri-TMS and cis/trans isomers), derivatisation with TMS or tBDMS adds a large number of additional carbons to the molecule of interest thus, in the case of C isotope ratio measurement, decreasing its 613C-value by 20%0 and more.
Methylation
Methylation is the standard derivatization method for carboxylic acids in general and for fatty acids in particular. With the commercial availability of BF3/methanol solution, there is no need to use the cumbersome HC1- and HBr-methanol solutions. BF3 is a Lewis acid that catalyses the esterification with methanol just as well as HC1 or HBr. Its advantage lies in its high volatility and the virtual absence of corrosive properties.
Ethyl-chloroformates (ECF):
Actuall~ the reaction of amino acids with ethyl chloroformate (ECF) yields their N-carboxy-ethoxy, O-ethylates (Husek, 1995). This is a very fast, easy to use reaction and almost as convenient to perform as silylation reactions. This method can also be used for other organic acids such as hydroxy-, amino- and keto-carboxylic acids (Husek & Matucha, 1997; Husek, 1998). However, in dealing with amino acids this method tends to yield non-quantitative derivatization for Glu, Gln, Tyr, and Hyp. Arg and Asp cannot be recovered at all.
N-pivaloyl, O-isopropylates (NPP): This method, developed by Metges et al., is very reliable and produces good derivatization yields even for problem amino acids such as Glu, Asp, Lys, His and Tyr (Metges et al., 1996; Metges & Petzke, 1997).
N-acetyl, O-proplylates (NAP):
The classic method, however, there are two aspects of the original protocol (Wolfe, 1992) that might cause some inconvenience. First, quantitative removal of HBr is difficult and time consuming. Secondly, esterification using alcohol and inorganic acid requires the presence of a water scavenger and the original protocol suggests
Table 8.3- Overview of widely used derivatization methods together with typical conditions for their GC separation and GC / C-IRMS caveats. C o m p o u n d Class
Derivatization
Typical GC Conditionsa)
Caveats
Long Chain Alcohols
TMS
SE 30:50 ~ (2 min) to 200 ~ at 40~ 200 ~ to 320 ~ at 3 ~/ min
none
r (3
=r 9 o 0-Q
Fatty Acids
Methylation using Methanol/BF3
SE 52 / SE 54:120 ~ (4 min) to 280 ~ at 4 ~
none
H y d r o x y / A m i n o / KetoCarboxylic Acids
TMS or tBDMS
SE 52 / SE 54:80 ~ (5min) to 150 C at 2~ 150 ~ C to 220 o C at 3.5~ 20~ to 300 ~ C
TMS: multiple derivatives for same c o m p o u n d tBDMS: excessive carbon load
Risk of non-quantitative derivatization
r
t,o
>
H y d r o x y / Amino / KetoCarboxylic Acids
Ethylchloroformates
OV-17:60 ~ (3 min) to 300 ~ at 6 ~/ min
H y d r o x y / A m i n o / KetoCarboxylic Acids
Acetyl, methylates
OV-1701:70 ~ (5 min) to 220 ~ at 5 ~/ min
none
9
Amino Acids
TMS or tBDMS
SE 52 / SE 54:50 ~ (5 min) to 150 ~ at 6~ 150 ~ to 300 ~ at 1 2 ~
TMS: multiple derivatives for same c o m p o u n d tBDMS: excessive carbon load
r
o
~..L.
Amino Acids
TFA, methylates TFA, iso-propylates
OV-1701:70 ~ (5 min) to 180 ~ at 3.5~ 180 ~ to 230 ~ at 5~ 10~ to 270 ~
Poisoning of combustion catalyst
Amino Acids
Ethylchloroformates
OV-1701" 60 ~ (3 min) to 100 ~ at 5 ~ 100 ~ to 300 ~ at 10 ~/ min
Risk of non-quantitative derivatization
Table 8.3 continued >
o o
> Table 8.3 continued Amino Acids
N-Pivaloyl, iso-propylates
SE 52 / SE 54:70 ~ (1 min) to 220 ~ at 3 ~ 220 ~ to 300 ~ at 10 ~/ m i n
none
Amino Acids
N-Acetyl, propylates
OV-1701:70 ~ (5 min) to 200 ~ at 4~ 200 ~ to 300 ~ at 6 ~/ min
none
a) The information given in these columns should be read as generalized guidelines providing a starting point for the reader to resolve individual analytical tasks.
r
!
r
~,,~o
GC and IRMS Technology for 13C and 15N Analysis on Organic Compounds and Related Gases
167
2,2-dimethoxy-propane. Unfortunately, this compound leads to side reactions, which turn the initially clear solution crimson (this even happens when heating a blank solution of the reagents without amino acids being present). The products of these side reactions and, hence, the colour are almost impossible to remove. For GC/C-IRMS analysis of amino acids as their NAP derivatives, the following protocol is suggested which avoids the pitfalls of the classic method. The first step is a trans-esterification reaction that eliminates the need for a water scavenger as no water is being formed during the condensation step. The amino acid(s) [up to a total amount of 20 ~mol] are dissolved in 400 /~L of propylacetate and 200 ~L of propanol/BF3 (14%). The mixture is heated in a crimped autosampler vial at 110~ for 30 min. From the cooled down solution, solvent and excess reagent are removed in a gentle stream of N2 at 60~ The residue is taken up in 500 ~L of propanol/BF3 and blow dried again. Repeat the procedure with 300 ~L of propanol. This reaction ideally results in a solid, crystalline residue of almost colourless appearance. The amino acid propylates are dissolved in a mixture of 200 ~L of acetonitrile and 100 ~L of 1,4 dioxane. To this solution, 150 ~L of triethylamine (mix well) and 90 ~L of dry acetic anhydride (either freshly distilled or stored over molecular sieve 3 ~) are added. The resulting mixture is heat in a crimped autosampler vial at 60~ for 15 minutes (the reaction time can be increased if necessary, but the temperature must not exceed 60~ After cooling down, 300 ~L of CHC13 are added and the mixture is neutralized with 600 ~L of 0.001 M NaHCO3 in two steps (2 times 300 ~L). The aqueous phase is discarded and the chloroform phase is dried over molecular sieves (3 ~, beads, pearl form). Compound classes of biochemical interest, various derivatization methods and corresponding typical GC conditions are summarized in Table 8.3.
8.3 GC/C-IRMS For isotope ratio measurement the analyte must either be a simple gas such as CO2 or N2, or must be converted into a simple gas, isotopically representative of the original sample, before entering the ion source of an IRMS. It is also important to remember that IRMS, in fact, determines the difference in isotope ratio with great precision and accuracy rather than the absolute isotope ratio. IRMS measurements yield the information of isotopic abundance of the analyte gas relative to the measured isotope ratio of a standard or reference gas. This is done to compensate for mass discriminating effects that may fluctuate with time and from instrument to instrument. From the above it is patently obvious that a GC cannot be directly coupled to an IRMS. The need for sample conversion into simple gases has prompted the design of a combustion interface where the GC effluent is fed into a combustion reactor. This reactor, either a quartz glass or ceramic tube, is typically filled with CuO/Pt or CuO/ NiO/Pt and maintained at a temperature of approximately 820~ or 960~ respectively (Merritt & Hayes, 1994; Merritt et al., 1995a). The influence of combustion tube packing on analytical performance of GC/C-IRMS has been reported by Eakin et al. (1992) and, more recently, by Ellis & Fincannon (1998). To remove water vapour generated during combustion, a water trap is required. Most instrument manufacturers
168
Chapter 8 - W. Meier-Augenstein
employ a Nation [Perma Pure Inc.] tube for this purpose. Nation is a fluorinated polymer that acts as a semi-permeable membrane through which H20 passes freely while all the other combustion products are retained in the carrier gas stream. Quantitative water removal prior to admitting the combustion gases into the ion source is essential because any water residue would lead to protonation of CO2 to produce HCO2 +, which interferes with analysis of 13CO2(isobaric interference). Very recently, Leckrone & Hayes (1998) published a detailed study of this effect. In contrast to dual-inlet IRMS machines, GC/C-IRMS systems produce almost Gaussian shaped signals. In addition, due to the "chromatographic isotope effect" (Matucha et al., 1991; Matucha, 1995; Filer, 1999) the m / z 45 signal (13CO2) precedes the m / z 44 signal (12CO2) by 150 ms on average (Rautenschlein et al., 1990; Brand, 1996), an effect not observed in ordinary Continuous Flow-IRMS (CF-IRMS) systems. This time displacement depends on the nature of the compound and on chromatographic parameters such as polarity of the stationary phase, column temperature, and carrier gas flow (Meier-Augenstein et al., 1996). Therefore, loss of peak data due to unsuitably set time windows for peak detection and, hence, partial peak integration will severely compromise the quality of the isotope ratio measurement by GC/CIRMS. The same is true for traces of peak data from another sample compound due to close proximity resulting in peak overlap with the sample peak to be analysed. Due to the fact that isotope ratios cannot be determined accurately from the partial examination of a GC peak, high resolution capillary gas chromatography (HRcGC) resulting in true baseline separation for adjacent peaks is of paramount importance for high-precision CSIA. It should be noted that the chromatographic isotope effect is not caused by a vapour pressure effect. It is rather the result of different solute/stationary phase interactions that are dominated by Van der Waals dispersion forces leading to an earlier elution of the heavier isotopomer (Matucha et al., 1991). This difference in chromatographic solute/stationary phase interaction is caused by lower molar volumes of the labelled, and thus heavier, compounds. The reason for the decrease in molar volume is the increased bond strength and thus shortened bond length between 13C-H and, to a lesser degree, 12C-13C,and 12C-H and 12C-12C,respectively. A problem not easily overcome is the partial peak overlap of compounds showing similar characteristics in both chemical and chromatographic respect. There is many an application where closely related compounds simply cannot be resolved resulting in peaks overlap. To make matters worse, CO2 and N2 disperse more freely within the carrier gas stream than their parent organic compounds which can result in overlapping CO2 peaks even for baseline resolved GC peaks. To extract valuable information obscured by such peak overlaps, Goodman & Brenna (1994, 1995) suggested software algorithms for improved data processing. These algorithms were based on curve fitting rather than the summation (the industrial standard) using combinations of exponentially modified Gaussian (E) and Haarhoff/Van-der-Linde (H) functions and were tested on up to 70% valley peak overlap. When the adjacent peaks were of equal abundance (leading peak" trailing peak 1"1) combinations of HE and HH appeared to
GC and IRMS Technology for 13C and 15N Analysis on Organic Compounds and Related Gases
169
provide the best recovery of isotope ratios. In the case of unequal abundance in favour of the leading peak (10:1), the HH combination gave the best accuracy. When the abundance was reversed (1"10), the EH combination provided the best accuracy but only for peak overlap up to 40% valley. Despite these encouraging results, curve-fitting algorithms for restoring lost accuracy have not been incorporated into any commercial IRMS data reduction software by the manufacturers. However, any progress in this direction aided by full evaluation of new algorithms for routine use (under 'real life conditions') requires user access to such algorithms and hence their incorporation into IRMS manufacturers' software.
8.3.1 High resolution capillary gas chromatography (HRcGC) As pointed out earlier, baseline separated gas chromatographic peaks are the basis for high-precision CSIA. To achieve this goal, in the first instance, basic gas chromatographic rules must be observed: (1) The polarity of the stationary phase should meet the polarity of the sample constituents (Table 8.2). (2) Column head pressure and, hence, carrier gas velocity should be set to suit column diameter (Table 8.4). (3) Temperature gradients should be chosen that exploit the maximum of the column length (i.e. the longer the column, the slower the temperature rise per minute; Table 8.4). (4) Samples should be injected in splitless mode. However, to avoid peak broadening and peak tailing, the split should be opened 10 seconds after injection. (5) Use of a retention gap (also known as pre-column or guard-column) will prolong the lifetime of the GC column and improve chromatographic performance. The retention gap (a deactivated piece of fused silica capillary tubing, typically 1 to 3 m long) helps focusing the injected sample thus leading to sharper peaks. Table 8.4 - Recommended values for carrier gas velocity and temperature gradient according to column length when using helium as carrier gasa). Column length [m] 10 15 20 25 30 40 50
Elution of methane [s]b) 35 53 70 88 105 140 175
Temperature gradient [~ 2.5 1.65 1.25 1.05 0.84 0.63 0.5
a) Based on working directions given by K. Grob (Grob, 1986). b) Set GC oven temperature to 30 ~ Set split ratio to about 1:30, inject a few ~L of natural gas (or lighter gas) and measure elution time of the first peak (FID signal). Adjust column head pressure to match recommended elution time.
170
Chapter 8 - W. Meier-Augenstein
Further to these principles, HRcGC techniques such as multi-dimensional capillary GC (MDcGC), enantio-selective GC, porous layer open tubular (PLOT) column GC for analysis of VOCs and high-temperature capillary GC (HTcGC) are powerful tools for high-precision CSIA when used in combination with GC/C-IRMS. Nitz and co-workers were the first to report the advantages of using MDcGC in GC/C-IRMS (Nitz et al., 1992). MDcGC is now, often in combination with enantioselective GC, almost exclusively used in authenticity control of flavours and fragrances by CSIA (Casabianca et al., 1995; Juchelka et al., 1998). Similarly, HTcGC is strongly associated with CSIA of steroids and long-chain fatty acids, e.g. (Woodbury et al., 1998a). It is often inferred that high temperature capillary GC is limited to the use of crossbonded apolar stationary phase such as SE 30 and SE 52 and therefore only of limited applicability. This assumption is mainly based on the maximum allowable operating temperature (MAOT) given by GC column manufacturers. Of course, strictly speaking high temperature capable GC columns should be purpose-made. However, most stationary phases except for polyethylene glycol (PEG) based ones, can be safely used up to 360~ (apolar phases up to 450~ if certain precautions are observed. (I) To avoid damage to the stationary phase at high temperatures, the carrier gas must be free of oxygen and moisture traces. Considering the cost for replacing a GC column, the investment in a high capacity gas purifier will soon have paid off. (II) It is a remarkably little known fact that polar stationary phases are light sensitive (Grob, 1986). Even exposure to indirect daylight or light from fluorescent tubes will lead to column deterioration. The first sign of this happening is usually a dramatic increase in column bleed. (III) Conditioning the column properly prior to high temperature usage will minimize normal column bleed (e.g. at 4 ~ to 300~ maintain at 300~ for 5 hours; repeat but program to 360~ and maintain at 360~ for 2 hours). (IV) If a polar phase is to be used for HTcGC, chose a polysiloxane based phase with a high cyanopropyl content. In contrast to common belief, cyanopropyl substituted polysiloxane phases are more stable and inert than phenyl substituted phases. Kurt Grob suggested that this is caused by high surface tension of phenyl substituted phases which in turn causes problems with film stabilization during column coating (Grob, 1986). Regrettably, the achievements of HRcGC in terms of optimum peak-shape and baseline separation are likely to be impaired during combustion and the subsequent passage through the interface. Changes in tubing diameter and frequent use of unions to connect the various parts of tubing lead to a loss in peak definition (peak broadening; peak distortion) and even to partial peak overlap. All of the aforementioned have a detrimental effect on accuracy and precision of isotope ratio measurement (MeierAugenstein et al., 1996). Very recently, Keith Goodman reported a single-capillary interface design (SCID) which he developed to overcome these problems. As the name suggests, a single capillary was used to connect the GC column to the open-split in front of the IRMS (Goodman, 1998). This capillary was threaded through a furnace and accommodated 2 CuO wires positioned thus as to coincide with the furnace dimensions. So far, this design has been tested for 13C isotopic abundance analysis of n-alkanes.
GC and IRMS Technology for
13Cand 15N Analysis on Organic Compounds and Related Gases
171
8.3.2 Isotopic fractionation and isotopic calibration during GC/C-IRMS analysis Owing to its unique design, it is not possible in GC/C-IRMS to calibrate target compounds against a standard of known isotopic composition, introducing the standard in exactly the same way as the analyte. There are only three feasible means of introducing a standard; (a) addition of reference compounds to the sample, (b) introduction of reference gas pulses to the carrier gas stream or (c) introduction of reference gas pulses directly into the ion source. The results of an extensive study into methods of isotopic calibration by Merritt et al. (1994) emphasised these demands. Comparing the use of internal reference compounds with the introduction of reference gas pulses directly in the ion source of the IRMS, Merritt et al. (1994) found an offset of > 2%0 between the two methods in the case of incomplete combustion and other systematic errors affecting only the analytes. These systematic errors affected both the analytes and the co-injected reference compounds but were not reflected by the external reference gas pulses. Other groups reported similar observations (Caimi et al., 1994; Caimi & Brenna, 1996). In the absence of such systematic errors, Merritt et al. (1994) found that both methods of isotopic calibration gave consistent results as long as multiple reference peaks were used to permit drift correction. Only one reference peak for isotopic calibration, albeit from an internal reference compound, is not enough to compensate for the influence of GC parameters such as analyte/stationary phase interaction or column temperature on measured isotope ratios (Meier-Augenstein et al., 1996). There are several stages during GC/C-IRMS analysis where mass discrimination and, hence, isotopic fractionation can occur. Closer inspection identifies seven potential sources. (1) Isotopic fractionation during sample injection (which can be overcome by oncolumn or time programmed splitless injection). (2) Chromatographic isotope effect. (3) Chromatographic peak distortion (leading and trailing peak tail). (4) Combustion process. (5) Peak distortion of N2/CO2 gas peak during passage of the combustion interface. (6) Changing flow conditions at the open split prior to the IRMS. (7) The IRMS itself. Obviously, the external reference gas pulses only compensate for item (7), whereas internal reference-compounds reflect all of the aforementioned. Recently, a method for isotopic calibration was reported that, provided a combustible gas was used, could reflect the systematic errors caused by items (4) to (7). This method combines the convenience and practicability of external reference gas calibration with the advantage of reflecting the majority of physical influences to which analytes are subjected in a GC/ C-IRMS system (Meier-Augenstein, 1997b).
172
Chapter 8 - W. Meier-Augenstein
8.4 Examples of high-precision CSIA using GC/C-IRMS High-precision CSIA of 13C isotopic abundance at both natural abundance and low enrichment levels yields measurements of 613C-values with a precision of 0.3%0 or better. Thanks to this high precision, even minute changes in the 13C isotopic composition can be reliably detected. Not surprisingly therefore, GC/C-IRMS has become the method of choice to determine the origin of a given organic compound by measuring its characteristic isotope 'finger print'. One area where HRcGC techniques in conjunction with GC/C-IRMS are unsurpassed is authenticity control of flavours, fragrances and essential oils. The potential of GC/C-IRMS for this kind of application has been very successfully exploited and developed further. The powerful combination of GC/C-IRMS with enantioselective capillary GC and MDcGC for authenticity control of flavours, wines and spirits have been reviewed recently by A. Mosandl (Mosandl, 1995; Mosandl et al., 1995; Mosandl, 1997). Another aspect of authenticity control is food quality control to prove and prevent fraudulent adulteration of a high quality product such as maize germ oil or virgin olive oil with oils of minor quality (Woodbury et al., 1995; Woodbury et al., 1998b; Woodbury et al., 1998a; Angerosa et al., 1997) High-precision CSIA of fatty acids in general is one application of GC/C-IRMS at both natural abundance and enrichment level that has yielded a wealth of information (Meier-Augenstein, 2002). Ecological Sciences have greatly benefited from this technique using CSIA of fatty acids to study terrestrial and marine ecosystems, food webs, even migration patterns of birds. Similarly, Bio-medical Sciences have exploited this technique using 13C-enriched fatty acids to study precursor/product relationships of poly-unsaturated fatty acids in-vivo. Furthermore, GC/C-IRMS was successfully used to trace the origin of narcotic drugs. Variation in 13C isotopic abundance of heroin was found to be dependent on its geographical site of production (Desage et al., 1991). Variations in both ~13C and 615N values provided valuable information to discriminate the different origin of several batches of confiscated 3,4-(methyldioxy)-methylamphetamin (MDMA, Ecstasy) tablets (Mas et al., 1995). Coca leaves from South America were found to vary in their ~13C (-32.4 %o to-25.3 %o) and 615N (0.1-13.0%o) values (Ehleringer at al., 2000). Humidity levels and the length of the rainy season and differences in soils were thought to affect the fixation processes and cause the observed subtle variations in 13C and 15N contents, respectively. In conjunction with the variations of trace alkaloids (truxilline and trimethoxycocaine) contents found in cocaine, researchers were able to correctly identify 96 % of 200 cocaine samples originated from the regions studied. High-precision CSIA might also provide the means for an unambiguous doping test in athletes. Independent work by two groups showed a drop in 613C values of endogenous 5a- and 5b-androstanediol from -26.52%o to -30.21%o, in some cases even
GC and IRMS Technology for 13C and 15N Analysis on Organic Compounds and Related Gases
173
down to -32.44%o, after synthetic testosterone was administered (Becchi et al., 1994; Aguilera et al., 1996; Shackleton et al., 1997). This drop in 613C value lasted for up to 10 days before g13C values of androstanediol returned to their former baseline value. Stable isotope analysis of natural gases, long chain alkanes and other fossil biomarkers by GC/C-IRMS is used to determine origin and maturity of natural gas and crude oil sources (Bakel et al., 1994; Bjoroy et al., 1994; Hughes et al., 1995; Berner & Faber, 1996), to identify sources of oil spills and oil pollution (Uzaki et al., 1993; Mansuy et al., 1997) as well as ocean transported bitumen (Dowling et al., 1995) and to characterise refractory wastes at heavy-oil contaminated sites (Whittaker et al., 1995; Whittaker et al., 1996). Applying high temperature capillary GC/C-IRMS, Evershed and co-workers gleaned new insights in paleo-dietary habits and prehistoric lifestyle from organic residues preserved in archaeological artefacts (Evershed et al., 1994; Evershed et al., 1995; Evershed et al., 1997; Raven et al., 1997). In a multidisciplinary approach, variations in 615N values from soil amino acids were used to determine differences in land use during Bronze Age, medieval and early modern times. One of the observations made were the consistently low levels of 15N abundance (615N < 0%o) in the amino acids Thr and Phe from soils of unmanured sites of cereal production. All the other amino acids showed positive 615N values (Simpson et al., 1997). Interestingly enough, a similar pattern in 615N values of free plasma amino acids was found in fasting human subjects (Metges & Petzke, 1997). Of course, this is only a limited selection out of the wide spectrum of GC/C-IRMS applications. However, over the last few years several review articles have collated publications in the field of GC/C-IRMS, to which the interested reader is referred (Preston, 1992; Whittaker et al., 1995; Brenna, 1994; Brand, 1996; Brenna et al., 1997; Meier-Augenstein, 1999).
8.5 Hyphenated GC/C-IRMS Techniques In recent years, the research efforts of different groups working in the field of GCIRMS have focused on extending the scope of on-line CSIA towards the measurement of organic 180/16 0 and organic 2H/1H isotope ratios. Research has also been aiming at high-precision measurement of two different elemental isotope ratios such as 2H/ 180, 13C/180 and 13C/15N, from the same compound source in one analytical run. In addition, hybrid systems have been developed that enable CSIA while at the same time recording a conventional mass spectrum of the target compound to aid its identification (Meier-Augenstein et al., 1994a; Meier-Augenstein, 1995; Hall et al., 1999). Placing a PORAPLOT Q capillary column between the combustion reactor and the IRMS, thus achieving baseline separation of N2 from CO2 by 100 s, 15N/14N and 13C/ 12C isotope ratios from alanine, leucine and phenylalanine could be measured in one single gas chromatographic analysis (Rennie et al., 1996). The excellent separation of N2 from CO2 provided ample time to switch IRMS ion source parameters from N2- to
174
Chapter 8 - W. Meier-Augenstein
Figure 8.2- Dual isotope measurement of 615N and 613C from alanine (as N-acetyl, O-propyl derivative) during the same analysis. The arrow indicates when the ion source parameters were switched from nitrogen mode (m/z 28) to carbon dioxide mode (m/z 44). A 100 s baseline to baseline separation of N2 from CO2 was achieved by passing the combustion products past-reduction through a PORAPLOT Q column, held at 35 ~ The results obtained from dual-isotope analyses (615N:-7.78 + 0.10%o vs. air; 613C:-40.22 + 0.14%o vs. PDB) were in good agreement with those obtained from separate analyses (615N: -7.86 + 0.38%o vs. air; 613C:-40.20 + 0.21%o vs. PDB).
CO2-mode (Figure 8.2). In 1994, using a GC-based IRMS system, Brand et al. showed that CSIA of 1 8 0 / 1 6 0 ratios was possible by converting oxygen-containing organic compounds on-line to CO by means of a pyrolytic reaction (Brand et al., 1994). The on-line coupling of GC and IRMS via a pyrolysis interface (GC/Py-IRMS) was used for the simultaneous determination of ~513C- and 6180-values for vanilla from different origins (Hener et al., 1998). Farquhar et al. converted bulk plant matter into N2 and CO by an automated on-line pyrolysis-based reaction using nickelized carbon at about 1100~ and separating N2 from CO post-pyrolysis in a GC fitted with a 5 ~ molecular sieve PLOT column (Farquhar et al., 1997). Independently, Begley & Scrimgeour (1997) reported on high-precision 62H and 6180 measurement for water and VOCs by using 20% nickelized carbon to generate both H2 and CO at temperatures of between 1050~ and 1100~ Standard deviation was + 2%o and + 0.3%o for 62H and 6180, respectivel~ with samples ranging from urine, water and VOCs. Their pyrolysis system was based on earlier work that was aimed at simultaneous 62H and 6180 determination from small water and urine samples (0.5 mL) (Begley & Scrimgeour, 1996). Common to both studies was the use of a
GC and IRMSTechnologyfor 13Cand 15NAnalysison OrganicCompoundsand RelatedGases
175
novel IRMS with the high dispersion necessary for separation of the 2HIH+ and 4He+ ion beams (Prosser & Scrimgeour, 1995). Very recently, this highomaSs dispersion IRMS was coupled to a GC via a pyrolysis interface including a 5 A molecular sieve PLOT column to achieve CSIA for 2H of fatty acids. In a technical brochure, 62H-values for 16:0 and 18"1 fatty acids (as methyl esters) from tuna oil were reported as148.5 +_4:.1%o and-155.3 _+1.0%o (vs. VSMOV) respectively (Scrimgeour et al., 1999). This GC/Py-IRMS system has also been applied to the measurement of lipid synthesis in humans, using deuterium oxide (D20) incorporation into fatty acids. Administering safe deuterium enrichments (< 0.5 atom%), the labelled fatty acids did not to contain more than one deuterium per molecule of fatty acid, and showed similar chromatographic behaviour to natural abundance samples. Following overnight incorporation of D20, plateau palmitate enrichments were measured by GC/Py-IRMS with a relative standard deviation of 0.5% (Scrimgeour et al., 1999). A different approach to CSIA for H/D ratios of organic compounds such a ethyl benzene and cyclohexanone was published by Tobias & Brenna (1996). Initially using a two stage reactor interface (CuO at 850~ followed by Ni held at 950~ (Tobias & Brenna, 1996), they found that better precision for 62H was achieved by employing an empty alumina tube held at about 1150~ (Tobias & Brenna, 1997). Since their IRMS was not capable of fully resolving sample derived 2HIH from excess 4He carrier gas, they used a heated Pd filter in conjunction with a make-up pressure unit to prevent He from entering the IRMS. This set-up selectively admitted only hydrogen through the Pd foil membrane into the ion source. The measurement of intra-molecular variations in isotopic abundance due to kinetic isotope effects during biosynthesis is another recent development to extend the scope of GC-IRMS. The group around H.L. Schmidt employed 13C isotope pattern analysis for distinction of natural compounds from corresponding synthetic products (Weber et al., 1997; Weilacher et al., 1996). In 1997, an on-line pyrolysis system for position specific isotope analysis (PSIA) of selected compounds from a complex mixture was described in detail for the first time (Corso & Brenna, 1997). They coupled a GC (GC-1) for sample separation prior to pyrolysis to the GC (GC-2) separating pyrolytic products of the selected sample compound. Furthermore, they installed a valve into GC-2 to permit separated pyrolysis fragments to be admitted to an organic MS for structure analysis of these fragments. 8.6 Conclusion
High-precision CSIA by GC/C-IRMS of gases and organic compounds at natural abundance and low enrichment levels is a powerful tool that provides quantitative information such as bio-availability, assimilation, turnover, incorporation and metabolism in biological and ecological systems. If great care is taken to avoid the pitfalls associated with sampling, sample preparation and sample separation, GC/C-IRMS will yield insights into complex systems offering answers to biochemical, physiological, and environmental questions that cannot be obtained with any other analytical instrumentation.
176
Acknowledgements
Chapter 8 - W. M e i e r - A u g e n s t e i n
The author gratefully acknowledges financial support through the Dietmar-HoppFoundation for Medical Research and Studies (Walldorf, Germany). Many thanks are due to Dr Helen F Kemp for critical reading of the manuscript.
Handbook of Stable Isotope Analytical Techniques, Volume 1 P.A. de Groot (Editor) 9 2004 Elsevier B.V. All fights reserved.
CHAPTER 9 Preparation of Ecological and Biochemical Samples for Isotope Analysis Mark A. Teecel, 2. & Marilyn L. Fogel2** State University of New York - College of Environmental Science and Forestry, 1 Forestry Drive, Syracuse NY 13210, USA 2 Carnegie Institution of Washington, Geophysical Laboratory, 5251 Broad Branch Rd., NW, Washington, DC 20015, USA e-mail: *
[email protected], **
[email protected] 1
9.1 Introduction Stable isotope analyses have proven to be a critical source of information for delineating processes in ecological and ecosystem studies (Rundel et al., 1989; Lajtha & Michener, 1994), which canencompass the study of habitats and the interactions of organisms with their environment. The main power of isotopic analyses resides in the ability to study both specific processes and also to trace sources of materials and flows of energy through complex ecological webs. Biological processes, such as photosynthesis and metabolism, fractionate materials and change isotope compositions. For example, plants fractionate carbon during CO2 uptake and therefore carbon isotope studies are a useful tool to study the photosynthetic process. Conversely, at the whole organism level, the fractionation associated with animal feeding is much smaller. As biological or inorganic matter that transfers between components in a particular ecosystem may have unique isotopic compositions (e.g., Fry & Scherr, 1984; Michener & Schell, 1994; Dittel et al., 1997; Chamberlain et al., 1997; Kelly, 1999), isotopes may be useful as indicators of food web structure and trophic level status. For example, the living components of all terrestrial communities, including humans, as well as many marine communities rely on the primary production of organic matter by photosynthesis to provide energy for consumers. Organic matter and energy are then passed along food chains, with energy being dissipated at each trophic level. Stable isotopes can be one of the most effective tools for tying together these ecological processes that affect, in turn, affect the biosphere, the atmosphere, and the geosphere at both local and global scales (e.g. Gearing, 1991; Quay et al., 1989; Paerl & Fogel, 1993). This chapter is intended as an overview of the application of stable isotope techniques to addressing ecological questions. We offer suggestions for ecologists who might be considering including isotope analyses to their work, as well as for more experienced isotope geochemists who are interested in expanding their research into a new field. The chapter expands on previous work including the several books that
178
Chapter 9 - M.A. Teece & M.L. Fogel
have been published on isotope application to ecological studies (Rundel et al., 1989; Lajtha & Michener, 1994; Griffiths, 1998), and specific texts on methodology for stable isotope use in environmental studies (Coleman & Fry, 1991; Knowles & Blackburn, 1993; Boutton & Yamasaki, 1996). Although techniques in measuring stable isotopes have changed to more automated devices, the information included in specific chapters of these books (i.e., Ehleringer, 1991; Boutton, 1991) is still appropriate. The focus of this chapter is on the proper sampling protocols for accurately determining the natural abundance and distribution of stable isotopes, particularly carbon and nitrogen, in ecological systems and does not address methods that include the addition of isotopically labeled compounds at enriched levels. Furthermore, methods for the analysis of H isotopes in different sample matrices are presented elsewhere in this book and BOX
9.1 - C a s e s t u d i e s o f a n a q u a t i c e c o s y s t e m
A simplified representation of the complex interactions and pathways of energy transfer in an aquatic ecosystem. Examples of external energy and substrate sources are also represented.
F i g u r e 9.1 -
C a s e s t u d y 9.1 -
Role of terrestrial carbon in aquatic food webs
Terrestrial carbon enters aquatic ecosystems via land runoff and may be a significant source of carbon to the consumers of the food web. The size and importance of this external source of carbon to the aquatic food web can be assessed by measuring the stable carbon isotope ratios of different elements of the food web. The isotopic composition of phytoplankton will reflect an aquatic signal (e.g. lacustrine or oceanic signal), whereas isotope analysis of dissolved and particulate organic matter will indicate whether such materials are terrestrial in origin. Measuring the isotopic composition of larval fish, whose diets include phytoplankton and particulate organic matter, will indicate the relative importance of terrestrial carbon to the aquatic food web. Furthermore, analysis of planktivorous and canivorous fish and generalist top feeders (e.g. trout, bass, and birds such as gulls) will confirm the role of terrestrial carbon in the aquatic ecosystem. Such approaches were employed in studies of the aquatic food web in Loch Ness in Scotland (Jones et al., 1998), and in an Australian tropical stream (Bunn et al., 1997).
Preparation of Ecological and Biochemical Samples for Isotope Analysis
Case study 9.2 -
179
Nutrient loading in aquatic ecosystems
Management decisions on issues of water pollution frequently revolve around high concentrations of nutrients, their sources and how they affect water quality. A research study can consider a study of phytoplankton blooms in a lake, and ask the question, "What is the source of nutrients supporting a phytoplankton bloom in a lake?" A study of this sort requires measuring the concentration and distribution of nutrients such as phosphate, nitrate and ammonium, in conjunction with measuring levels of bacterial activity and phytoplankton productivity. Nitrogen isotope measurement of phytoplankton and different sources of nutrients will aid in determining the major source of nutrients which supports the phytoplankton bloom. Dissolved inorganic nitrogen has at least four major sources: terrestrial runoff, regeneration from benthic processes, atmospheric deposition, and regeneration in the water column by bacteria. To best characterize each source, their concentrations and 615N values should be monitored periodically (weekly to monthly) over the course of the year as temporal variations occur. In conjunction with these measurements, bacterial activity should be measured to determine the relative importance of these processes over the year. In addition, it is important to determine the dominant species of phytoplankton which comprises the bloom, and also the change in algal population over the year. Isotopic compositions of the different algal species collected will provide insights into the temporal utilization of different nutrient sources with respect to time. The overall result of this type of combined study will indicate the relative importance of particular nutrient sources to phytoplankton productivity in the lake. Examples of these types of studies include those by McClelland et al. (1997) and Cifuentes et al. (1988, 1989).
the reader is referred to the chapters by Horita & Kendall (Chapter 1, this volume) and Volume IL Part 3, Chapter 1-2/3 on hydrogen isotopes. Our chapter is presented in three sections: Section I deals with the development of experimental approaches and the potential uses of stable isotopes, while Section II describes different methods of storage and preparation of samples for isotopic analysis. In Section III we outline chemical methods for bulk and compound specific isotope analyses of biological materials pertinent to ecological studies.
9.2 Section I: Stable isotope biogeochemistry in ecological research Before undertaking a study in this field, a clear understanding of the ecological questions that are of relevance is of the utmost importance. In any particular ecosystem, the number of plant, animal, water, soil, and air samples available for analysis is almost infinite, and although methods for analyzing bulk isotopic ratios are becoming more and more automated (Barrie & Prosser, 1996; Fry et al., 1992; Wong et al., 1992b), resources can easily be wasted when trying to investigate a large-scale ecosystem. Conversely, nothing is more vexing to the scientist, and to the community, as isotopic results that are incomplete in the sense that sources and sinks, and interconnections between matter pools have not been adequately characterized either chemically, biologically, or isotopically. Once hypotheses have been established, the site needs to be assessed biologically, chemically, and physically. Literature surveys are a good source of biological inventories, but the most useful information usually arises from the studies and observations of ecologists who have often worked for many years at a particular site. For example, the U. S. National Science Foundation funds a series of Long Term Ecological Research Sites (LTER) and in recent years, some of these sites have been the focus of
180
Chapter 9 - M.A. Teece & M.L. Fogel
isotopic investigations (e.g. Fry, 1991). As in any new scientific endeavor, the first iteration of a large-scale isotopic project will very likely involve some unforeseen problems associated with the analytical design. Many factors need to be considered in an attempt to reduce these potential problems, thus we suggest an approach that can be tailored to the specific questions being addressed. Illustrative examples of how this general approach might be applied to investigations of aquatic (Figure 9.1) or terrestrial (Figure 9.2) ecosystems are also described in Box 9.1 and Box 9.2 respectively. In general, the first step to approaching the isotopic study of an ecosystem requires the consideration of issues that are specific to natural abundance stable isotope analysis of ecosystems and their components, and takes place before actual sampling of B O X 9.2 - C a s e s t u d i e s o f a t e r r e s t r i a l e c o s y s t e m
Schematic of the major components of a typical terrestrial ecosystem. Representative examples of different trophic levels and particular pathways of energy and matter transfer are shown. A well defined stable isotope study can aid in understanding the complex interactions between these components. F i g u r e 9.2 -
C a s e s t u d y 9.3 - The impact of atmospheric carbon dioxide on a terrestrial ecosystem A question relevant to the present day surficial environment is whether a signal of increasing CO2 concentrations in the atmosphere can be found in terrestrial plants, animals, or soils? This question provides an opportunity to study changes in CO2 concentrations over several different timescales by designing a protocol which involves the analysis of carbon from a variety of "pools". For example, the 613C value of plant tissue is related to both the isotopic composition and concentration of atmospheric CO2. This isotopic ratio of plants is set during a single growing season, whereas the 613C value of tree rings provide excellent yearly signals. At longer timescales, the isotopic composition of leaves is passed up the food chain to herbivores, such as deer. Collagen in their bones, or apatite in their teeth take several years to form and will therefore contain an integrative isotopic signal. As a result, a longer-term record of 613C values more in line with the tree ring study can be determined. Lastly, the carbon isotope composition of soil carbonates which are formed over thousands of years, provide a record of long-term changes in climate and CO2 concentration.
Preparation of Ecological and Biochemical Samples for Isotope Analysis
181
m a t e r i a l begins. It is i m p o r t a n t to c o n s i d e r the different factors t h a t affect the stable isotope c o m p o s i t i o n of material. Biological a n d b i o c h e m i c a l p r o c e s s e s fractionate isot o p e s a n d therefore h a v e a s u b s t a n t i a l effect on the isotope c o m p o s i t i o n of o r g a n i s m s . O n the other h a n d , the f r a c t i o n a t i o n associated w i t h the c o n s u m p t i o n of organic m a t ter b y h i g h e r trophic level o r g a n i s m s is small, facilitating the use of i s o t o p e s in source identification. Therefore, the isotopic c o m p o s i t i o n of a p a r t i c u l a r m a t e r i a l in an ecos y s t e m will d e p e n d on b o t h its source isotopic c o m p o s i t i o n a n d p r o c e s s - r e l a t e d frat i o n a t i o n s associated w i t h its f o r m a t i o n .
C a s e s t u d y 9.4 - Assessment of the effect of fertilization nitrogen on agricultural crops The primary question is the following: Are nutrients completely available for crop growth? There is also a secondary concern whether fertilizer pollutes ground water and affects other plant and aquatic ecosystems. For this study, characterizing the isotopic and chemical concentrations of separated nitrogen-containing nutrients (i.e. nitrate and ammonium) is critical. A sampling regime should be implemented which includes the collection of plants, crops, soil, and groundwater samples. In conjunction with concentration data, nitrogen isotopes can be used to trace the transport of nutrients and fertilizer nitrogen species into plants and groundwater. Initially, baseline values of concentration and isotopic composition of the nitrogenous nutrients in potential sources, such as groundwater, should be sampled. After application of the fertilizer, soil and groundwater nutrient concentrations should be monitored over time, preferably over the entire course of the growing season, and also subsequent seasons. If there is evidence of elevated nutrient concentrations, then more complicated analysis of the 615N of dissolved nitrogen species should be added to the study (see Box 8.3). Assessing plant uptake of nutrients involves sampling and analyzing the bulk nitrogen isotope composition of plants, both crop and indigenous species, in the agricultural fields on a monthly or weekly basis. In addition, different parts of the plants should be sampled in order to determine the degree of isotopic heterogeneity in a large plant sample. Lastly, at least 5-10 samples of the crop plants from various locations in the field should be collected in the event that the fertilizer application was not uniform. C a s e s t u d y 9.5 - Determination of dietary sources Stable isotopes are increasingly being used to investigate the feeding ecology and behavior of animals and birds. The question may be "What is the relative contribution of freshwater fish to the diet of a particular bird," or "Does the diet of an animal change over the seasons". In addressing these questions, it is imperative that the potential diet sources under investigation exhibit distinct isotopic compositions (i.e. d values of sources differ by more than 2%o). For example, marine species of fish (613C = -17%o + 4) have isotope compositions which are substantially different to freshwater species (613C = -24%o + 4). In such diet studies, potential dietary sources are collected and analyzed and compared with particular tissue samples from the consumer. Similar sample types should be measured (e.g. feathers, muscle tissue, teeth), as the isotopic composition of individual tissue types can vary significantly within a single organism (Gearing, 1991; Tieszen et al., 1983). The carbon isotope composition of an organism should reflect its diet and the nitrogen isotope composition may indicate the trophic level of the organism in the ecosystem. It is important to understand the limitations of these methods, as there can be considerable sources of isotopic variability in the diet and the particular tissue analyzed. Gearing (1991) provides a general review of the study of trophic relationships with stable isotopes and addresses many of the issues of isotopic heterogeneity in sample types. There is a wealth of studies employing stable isotopes in elucidating feeding strategies of a large variety of organisms including birds (Bearhop et al., 1999), fish (Doucett et al., 1999), whales (Hobson and Schell, 1998), crabs (Fantle et al., 1999) and primates (Schoeninger et al., 1998).
182
XQ
0
Chapter 9 - M.A. Teece & M.L. Fogel
Differentiation of sources using isotopic ratios: Because stable isotopes are most often used as tracers of different sources in an ecosystem (Case study 9.1), the first consideration should be to establish whether sources truly have unique or even distinct isotopic ratios (e.g. Farquhar et al., 1989a; Quay et al., 1992). As a rule of thumb, if the 6 values of sources differ by less than 1%o, a major isotopic study designed to differentiate their movement in an ecosystem should probably be reconsidered (e.g. Benner et al., 1987). While current instrumentation and methods are capable of distinguishing much smaller differences in isotopic composition (0.2%o for 613C or 0.3%o for 615N), in complex systems involving multiple interactions, the reality of distinguishing sources of materials with such small differences is a concern. Overall, spatial and temporal heterogeneity in 6 for a single material, and isotopic variation between specific compound classes in organisms, makes distinguishing and tracing particular sources of organic matter virtually impossible at the level of less than 1%o. Therefore, it is important to study ecosystems where the ratio of signal (distinct isotope ratios) to noise (e.g. isotopic heterogeneity) is favorable. Isotopic heterogeneity in individual samples" The most important variable that can blur isotopic signals is heterogeneity within a single sample. Individual samples may be isotopically heterogeneous on spatial scales from meters to nanometers, and therefore sampling of material requires careful consideration of these factors. For example, within a single animal, individual tissues (e.g. muscle, blood, hair) have different isotopic compositions (Tieszen et al., 1983), and in a study on isotopic heterogeneity of leaf tissue, we observed that the ~13C value of plant mesophyll is consistent over a leaf, however the veins and petioles show significant deviation. In addition, individual materials are isotopically heterogeneous on a molecular scale with different biochemical compound classes (e.g. proteins, lipids and carbohydrates) having a range of isotopic compositions. This molecular heterogeneity results from different process-related fractionations that occur during the biosynthesis of specific compound classes. Since the development of continuous flow gas chromatography-combustion-isotope ratio mass spectrometry (GC-C-IRMS) for the isotopic analysis of individual compounds (Hayes et al., 1990), the literature is filled with publications defining the breadth and scope of isotopic ratios on the compound specific level. Not only are different biochemical fractions quite distinct in their isotopic compositions (range from 3-5%o), but the individual compounds that comprise the pools also have almost an order of magnitude greater variation (20-30%o) (e.g., Blair et al., 1985; Rieley et al., 1991; Abrajano et al., 1994; Fantle et al, 1999). These isotopic variations between individual compounds generally result from different isotopic fractionations during their biosynthesis. Furthermore, the heterogeneity in such biochemical isotopic fractionations can be overprinted by modification of physiological processes, such as the variations in the ~513Cvalue of phytoplankton as a function of dissolved CO2 concentrations (Bidigare et al., 1997; 1999). These physiological variations can result in shifts of C or N isotope ratios of the bulk material by at least 5%0 (e.g., Ambrose, 1991; Bird et al., 1995; Buchman et al., 1996; Cifuentes et al., 1988). Therefore, unless a very specific biochemical pool if sampled from a physiologically invariant source, then the likelihood of source vari-
Preparation of Ecological and Biochemical Samples for Isotope Analysis
183
ations is very real.
3. Spatial variation: Sample heterogeneity governed by physiological or biochemical changes, can be exploited for delineating ecosystem dynamics. For example, the stable carbon isotope compositions of vegetation and air in forest ecosystems are influenced by height. With increasing height, the canopy air becomes progressively enriched in 13C, as the influence of soil-respired CO2 and tropospheric CO2 changes. Therefore, as the carbon source for photosynthetic fixation (i.e. canopy CO2) changes, the isotope values of leaves of individual trees vary as a function of height (Buchmann et al., 1997). Similarly, the isotopic composition of leaves may change with overall elevation (Sparks & Ehleringer, 1997). 4. Temporal variation" The stable isotope composition of components of an ecosystem can vary on a multitude of different timescales from seconds to thousands of years. The temporal variations of an ecosystem and how it responds on a diel, seasonal, or longer time spans rely specifically on isotopic fluctuations (Case study 9.2). In such studies, researchers hypothesize or predict that isotopic variations will be linked to specific biological or chemical processes (Cifuentes et al., 1989; Fogel et al., 1999). Such studies have also led to discoveries of the importance of particular pools of matter or organisms that were previously considered unimportant in the functioning of an ecosystem (e.g. Stapp et al., 1999). In addition, the isotopic composition of individual materials may reflect a specific time period and this parameter is an important one to consider (e.g. Ben-David et al., 1997; Johnson et al., 1998). For example, carbon isotope ratios in leaf tissue reflect a single growing season (e.g. Brooks et al., 1997), whereas the carbonate in apatite of teeth of animal feeding on leaves will have an isotopic signal reflecting several years of an animal's life (0-5 years) (Lee-Thorp & van der Merwe, 1991). Trees with significantly sized trunks can live for several centuries; thus long-term trends are recorded in the 613C values in growth rings (Stuiver & Braziunas, 1987; Bert et al., 1997). The 613C values of soil carbonates, which form over thousands of years, reflect long-term climate change that occurs on continental scales (e.g., Cerling, 1992). If one study included all of these carbon pools, the leaves, teeth, wood, and soil carbonates, multiple ecological processes occurring over very different time signals could be discovered (Case study 9.3). Considering all the issues discussed above and incorporating the experiences of several researchers, we summarize below an approach to a general ecological investigation that involves stable isotope techniques:
Step 1" After defining the research questions or hypotheses, decide whether there is a good possibility that stable isotopes will be able to provide answers to the questions. Is there a good "isotope signal to noise ratio"? Which isotopes will be the most useful? Are there methods already established to analyze the stable isotopes in the ecological materials that are part of the study design?
184
Chapter 9 - M.A. Teece & M.L. Fogel
Step 2" Design a survey of the ecosystem to measure baseline isotopic values in the pools that are of interest. Is the study site accessible year round or seasonally? Does the ecosystem vary seasonally, and most importantly, does the scientific question involve studying the ecosystem over a period of time? Is special sampling equipment required? Are permits needed for sampling or importation(e.g. CITES) across international borders? Step 3" Begin the planning stage for collecting, processing, and analyzing samples. Assemble the equipment needed for the collection keeping in mind the storage needed until samples are analyzed (Section II). A list of all reagents, bottles, and chemicals that are needed in the field can be extremely helpful. Ancillary analyses, for example nutrient concentrations, are important considerations that are often overlooked. Step 4" Preliminary sampling trip and isotopic analyses: Take the first set of samples, if the opportunity exists to travel easily to the ecosystem to be studied. Collection of 2-3 times more samples than you intend to analyze will allow you to choose from this array of samples at the time of isotope analysis. If the sampling site is remote, check the availability of equipment at the site, and confirm that all necessary items of equipment and chemicals are on hand. Attempt the first set of isotopic analyses. The first set of analyses should determine whether the analytical procedures are robust, if sources have distinct isotopic compositions and whether the stable isotope techniques used will address the research objectives. Step 5: Rethinking the experiment and the sampling protocols" Often, the first field collection is an adventure that is problematic and incomplete. Analyze the initial data and reformulate the questions and experimental design to reflect the reality of the ecosystem. Consider adding additional parameters, sample types, or time points, at this stage, that now are more important, and eliminate those that will not contribute to the final results. In many instances, the material that is collected consists of a complex mixture (e.g. soil, sediment, filtered material) and therefore a more refined approach to isotope analysis may be required. In such cases, measurement of the isotopic composition of individual organic compounds can provide unique information for studying complex interactions where physical separation and even biological separation are difficult (for example the transfer of dissolved organic carbon from phytoplankton to bacterioplankton). The methods and techniques for such analyses are discussed in section 9.3.3. 9.3 Section II: Methods of sample collection and storage Different types of samples require different methods of collection and storage, and the choice of methods will be influenced by the location of the ecosystem, the availability of cold storage, and the kinds of analyses to be performed. Whatever methods are chosen, it is imperative that they do not affect the isotopic integrity of the material being sampled. Isotopic ratios of organic material are susceptible to alteration if degradative processes take place during storage. Samples that are poorly preserved are at
Preparation of Ecological and Biochemical Samples for Isotope Analysis
185
risk of possible bacterial decay, which may result in mineralization or solubilization of organic carbon. Alternatively, contamination can alter isotopic ratios, especially of samples destined for compound specific isotope analysis. Contamination while sampling is minimized through the use of gloves and minimal handling of materials. Geochemical materials with ancient organic matter are sampled and stored at room temperature (e.g. light hydrocarbons in petroleum samples), whereas other materials require more refined techniques and, in some cases, specialized collection equipment is required.
9.3.1 Biological materials Biological sampling is based on capturing and preserving the isotopic and biochemical composition of an organism at a particular point in time. In order to accomplish this, living processes must be effectively inhibited, because at death, organisms release enzymes (e.g. proteases, nucleases, and lipases) that specifically target all of the biochemical classes of compounds. Cold storage is the best method for slowing down living reactions, while also inhibiting degradation pathways. If liquid nitrogen is available on the site, quick-freezing and storage in an enclosed vessel will preserve all of the high molecular weight biochemical information, as well as any isotopic information. Dry ice or freezing in a conventional-20~ freezer are also acceptable methods of cold storage, and are often available on oceanographic research vessels or nearby laboratories. Prior to freezing biological material it is important to rinse any adhering sediment or salt off of the organism with distilled or deionized water. The use of fixatives (e.g formalin or ethanol) to preserve samples is not recommended, as we have found that tissue preserved by the addition of these reagents has an altered carbon isotope ratio, most likely because of binding between the solvent and the tissue. In remote areas, there are several ways to overcome the inconvenient access to cold storage. First, long-life liquid nitrogen dewars can be purchased that maintain liquid nitrogen temperatures of-196~ for at least 30 days. The dewars are initially expensive, but are reusable and should provide several years worth of service. For this method to be useful, there must be a source of liquid nitrogen available. Most universities have reasonably priced liquid nitrogen reservoirs, and dewars such as these require filling only one time. Transporting liquid nitrogen to the field requires careful planning and prior to using such methods the reader is encouraged to understand the potential hazards associated with working with cryogenic liquids. Safety information is available from gas supply companies and also the National Research Council (1981). An alternative to cold storage is drying in the field. Plants can be effectively dried in plant presses on herbarium paper. We have tested this method with samples taken in the Australian desert, which is characterized by low humidity and high temperatures. Drying occurred in 2-3 days, with no change in plant appearance. Care should be taken to use clean blotter paper rather than newspaper, as the ink from the newsprint could contaminate the plant sample. Last, some plant presses are connected to a source of heat to promote drying; a simple light bulb should be chosen over heated charcoal.
186
Chapter 9 - M.A. Teece & M.L. Fogel
An isotopic comparison of samples preserved by drying, with those preserved using liquid nitrogen revealed no difference in 615N o r ~13C. Keep in mind that one aspect of drying that may cause complications is the use of heat to accelerate drying, which could act to increase the activity of degradative enzymes. Another complication often encountered in tropical regions is high humidity, which will increase drying times. A third and simple technique to dry and preserve samples in the field is to use desiccants (e.g. silica gel). The sample can be wrapped in foil and placed in a sealed jar or sample bag that contains a desiccant. With frequent replenishment of the desiccant, this method can remove a significant amount of water. Storage of animal tissue is a separate problem, because drying tissue can attract insects that might consume it. As a result ample consideration should be given to the type of tissue to be studied. For example, collagenous or keratinous material from bone, horn, nails, or feathers can be easily rinsed with distilled water and dried without any chemical or isotopic alteration. Teeth are also very resistant tissues that are especially desired because they contain well preserved organic and inorganic isotopic reservoirs. However, if tissue samples are needed, they should be frozen immediately, preferably in liquid nitrogen or put on ice as soon as possible.
9.3.2 Aquatic particulate material One of the most frequently used methods for collecting aquatic material for isotopic analysis is filtration. There are, however, many things that should be considered prior to venturing into the field. First, how much material needs to be collected for the desired analysis? Second, will the filter interfere with the analysis, or contaminate it, in any way? Third, what is the particle size of interest? For on-line isotopic methods for determining bulk C, N, H, and O isotope ratios of organic material, less than a milligram of material is needed, even including replicate analyses. In contrast, for compound specific work, especially if a target molecule (e.g. cholesterol) is sought, 5-10 mg of organic matter is a reasonable amount to collect to allow for replicate analyses. For example, if the environment that is to be studied is rich in biological material, typically a liter of water is filtered for bulk analysis. For sampling larger amounts of material or in regions where there is a very low amount of particulate biological material, large filters are required and often each scientist constructs filtration devices that fit individual needs. Large volume filtration apparatus have been designed to operate in situ, to be connected with complex pumping system, or to be contained in pressurized vessels (Wakeham & Volkman, 1991). The size of the filter can be determined by the user, and for most materials sheets of glass fiber or polycarbonate can be cut to fit the design of the filter holder. At this stage the type of filter material is an important consideration. For a filter to be compatible with an analytical technique it should not contaminate the sample. Most researchers use glass fiber filters, which can be cleaned by heating (450~ for 4 hours), and conventional filter holders (47 mm in diameter), which can be readily obtained from scientific supply companies. Filters containing particles (e.g. GF/F filters) are wrapped in clean aluminum foil, stored frozen, until they are dried or pro-
Preparation of Ecological and Biochemical Samples for Isotope Analysis
187
cessed in the laboratory prior to isotopic analysis. If the method requires a specific amount of weighed material, the filter should be chosen that will allow the sample to be recovered quantitatively and with ease (e.g., a Nuclepore filter is the perfect choice, because material can be readily rinsed off the surface of the filter). The particle size is the last consideration for filtration. By the simplicity of their design, filters are only crude ways of separating material by size. The most common filter is the GF/F filter (glass fiber) with a pore size of 1 mm, however, as filtration proceeds, material is trapped in the fibers of glass, so that the effective or nominal pore size is reduced to 0.7 mm. A GF/F filter can effectively trap most phytoplankton, many bacteria, and all zooplankton. Typically, a much larger size filter or mesh is used to separate zooplankton from phytoplankton and bacteria. Mesh sizes > 100 mm will separate the majority of zooplankton, and filter sizes > 20 mm will remove a majority of the microzooplankton. The sample trapped between I mm and 20 mm is the material typically captured on a GF/F filter. It will usually include phytoplankton of all types, detrital material, some attached bacteria, and microheterotrophs including flagellates and rotifers. Unfortunately, many studies in aquatic ecology are aimed at defining trophic relationships among these different groups of organisms, and thus simple filtration fails to provide adequate separation. In such cases, a compound specific approach may be useful, especially if the organisms under investigation produce source-specific biological marker compounds, as outlined in section 9.4.3. However, if the concentration of these source-specific compounds is low, then substantially larger volumes of material would be required, collection of which may not be logistically feasible in the field.
9.3.3 Collection of water samples Water samples collected for the isotopic analysis of ammonium or nitrate should be frozen immediately. Typically one liter of water is collected, and if the sampling location is very turbid or biologically-active (e.g. a river or lake), it is best to filter the sample through a pre-combusted GF/F filter. Seawater samples from the open ocean that contain very little particulate matter can be collected and frozen directly. In addition, a subsample (approximately 25 mls) should be frozen or analyzed for nutrient concentrations immediately (Solarzano, 1969; Strickland & Parsons, 1972). For isotopic analysis, the frozen sample will be returned to a laboratory for subsequent analysis, however long-term storage should be avoided, because concentrations of ammonium, for example, are often lower after storage. If a freezer is not available in the field, water samples should be kept cold on ice. If no method of refrigeration is available, adding I ml of 2 N HC1 to I liter of water will generally preserve the sample by halting bacterial growth and ensuring that the pH of the solution favors the ionic state of ammonium.
9.3.4 Storage of samples
In the laboratory, samples are typically dried and stored in a freezer prior to isotopic analyses. The preferred method of drying is freeze-drying, but alternatives such as oven drying may be employed. Freeze-drying employs a vacuum to remove water from tissues that are kept frozen throughout the procedure, thereby minimizing the
188
Chapter 9 - M.A. Teece & M.L. Fogel
possible degradative processes caused by heat drying. However, freeze-dryers typically are not resistant to acid, so that samples can not be pre-treated with acid (e.g. samples containing carbonate carbon) and subsequently dried. For such acid-treated samples, a similar type of drying, which was developed for biological tissues and is termed centrifugal vacuum drying, can be employed. In these instruments, samples are placed into a centrifuge, and as vacuum drying proceeds, the sample is concentrated in the bottoms of the centrifuge tubes. These drying units are generally built with chemically resistant housing and traps so that samples treated with acid or solvents can be dried readily. The most common alternative to freeze-drying is oven drying. Most ecological laboratories house large drying ovens maintained at 80~ with or without flowing dry gas. Drying biological tissue, that is clearly dead and no longer metabolically active, can be accomplished at temperatures between 50 ~ and 80 ~ At lower temperatures, flowing an inert gas such as nitrogen, argon, or helium over the sample both minimizes oxidation, and also decreases drying time, as water is swept efficiently from the oven. Vacuum ovens offer another alternative means of drying, and often desiccants (e.g. P205 or Silica Gel) are added to increase the efficiency of water removal.
9.3.5 Sample preparation for isotope analysis
The initial preparation of samples for isotope analysis is dependent on the type of analysis to be performed. If the isotope composition of the whole sample is sought, then after drying the material, little further preparation is required. Samples that may contain carbonate carbon, however, should be treated with dilute HC1 (0.1N) and thoroughly rinsed with distilled water prior to drying. Carbonate carbon is typically more enriched in 13C than organic carbon and may confuse analyses of organic material if not removed prior to isotope analysis. The carbonate present in dissolved materials or on filters can be removed by incubating the sample for 24 hours in a glass dessicator over fuming HC1. Carbonate can be removed from sediments, which can contain up to 90 wt% CaCO3, by acidifying samples with either concentrated HC1 in silver sample cups (Nieuwenhuize et al., 1994) or with a modified HC1 vapor phase method (Yamamuro & Kayanne, 1995). Complex sample matrices (e.g. soils, leaf litter, stomach contents) should be ground to a fine powder using a mortar and pestle, as the necessity for sample homogeneity cannot be overemphasized. Material frozen with liquid nitrogen in a mortar, or crucible, can be easily ground to a fine powder using a glass pestle. Alternatively, commercially available grinders can be used which pound samples to fine powders under cryogenic conditions (e.g., Spex Mill, Wig-L-Bug). Such equipment is very useful when working with materials such as wood, bone, soils and sediments. If a compound specific approach is to be undertaken then substantial preparation, including specific chemical and biochemical techniques may be performed prior to isotope analysis. These techniques are discussed with relation to specific analyses later in this chapter.
Preparation of Ecological and Biochemical Samples for Isotope Analysis
189
9.4 SECTION III: Stable isotope analysis of ecological samples The present chapter concentrates on techniques to measure the isotopic composition of important nutrients, bulk organic matter and the three major classes of organic compounds (lipids, amino acids and carbohydrates). The reader is also directed to Chapters 8 and 23 for additional methods on such subjects. In addition, it may be important to measure the isotope composition of inorganic material in the ecological setting under investigation. The specific techniques required to measure these parameters are discussed in other chapters on isotopes in hydrology (Chapter 1), dissolved inorganic carbon (Chapter 10) and atmospheric gases (Chapter 14).
9.4.1 Isotopic analysis of ammonium Measurement of the two major N-containing nutrients, ammonium and nitrate, is important for nitrogen isotope studies. The analysis of dissolved nitrate is described in a Chapter 15 (applications in Liu et al., 1996; Brandes & Devol, 1997; Chang et al., 1999). There are several methods for ammonium analysis, and in all of them, NH4 + is separated from solutions or samples by either gas phase transfer (i.e., distillation or diffusion) or adsorption and separation by a substrate (e.g., molecular sieve Zeolite BOX 9.3 - Nitrogen isotope analysis of ammonia
Two distinct methods of measuring 15N-NH4 + are available. One method is a modification of the ammonia diffusion method and is described in detail in Holmes et al. (1998). The second method utilizes steam distillation and zeolite adsorption to measure the nitrogen isotopic composition of ammonia (Velinsky et al., 1989). The steam distillation unit requires a hardy refrigeration unit, a source of ammonia-free water for steam generation, a condenser, and a receiving flask. All of the glassware used in this method should be cleaned and rinsed with 1% HC1 and then rinsed thoroughly with distilled water. After the glassware for distillation and receiving is assembled, and prior to preparing the first sample, 200 mls of pure ethanol is distilled into the receiving vessel. This rinse is followed by a subsequent wash with 200 mls of distilled water. Both solutions are discarded after distillation. This washing and rinsing procedure is repeated between each sample. The recovery flask is then prepared as follows: a pre-combusted Pyrex Pasteur pipette is attached to the output tube from the condenser unit. The flask contains 25 mls of dilute HC1 (0.00I N), and care is made to have the tip of the pipette underneath the surface of the acid solution. A filtered water sample (250 mls) is added to the distillation flask with 1 ml of 10N NaOH. The amount of NaOH added will depend on the salinity and the buffering capacity of the environmental water. Once the NaOH is added to the solution, the NH4 + is converted to NH3, so distillation should begin immediately. It is essential that >99% of the ammonia is distilled, and during all phases of this procedure, individual steps should be checked for yields by conventional ammonium analysis (Strickland and Parsons, 1972). It is important to ensure that the pH of the acid trap remains below 6. After distillation is completed, the receiving flask is removed from the unit and the ammonium is adsorbed onto a zeolite molecular sieve. A bed of zeolite is made on a GF/F filter (cleaned by preheated to 550~ for 2 hours) and the ammonium solution passed through the filter. Gravity filtration takes up to 60 minutes but results in approximately 100% recovery. The filtered zeolite is placed in a drying oven at 50~ for no less than 48 hours. If a dry gas can be passed through the oven, water vapor is more efficiently removed. After drying, the zeolite can be scraped off the filters and analyzed by combustion in either a sealed tube or by EA-IRMS. The precision and accuracy of this method is about +0.5%o, with the error probably originating from slight isotopic fractionation of the NH4 + onto the zeolite. For EA analysis, the zeolite should be dried in the presence of either He or Ar, because N2 from air can be trapped within the pores of the zeolite, which interferes with the analysis.
190
Chapter 9 - M.A. Teece & M.L. Fogel
W-85) (Box 9.3). All of the procedures that have been developed are relatively simple, however, attention must be paid to almost all of the details, because large isotopic fractionations occur in the transfer of ammonium [NH4 +] or ammonia [NH3] between different solutions. For example, there is a 30%o isotope effect between dissolved ammonium and gaseous ammonia. Therefore, to minimize this potential fractionation during distillation or diffusion it is essential that all (>99%) of the a m m o n i u m / a m m o nia in solution is transferred to the receiving solution. Smaller isotope effects are associated with adsorption. In many of the employed methods distillation and adsorption are often coupled together (e.g. Velinsky et al., 1989), and sloppy techniques with either process can result in decreased analytical precision and accuracy. Moreover, with adsorption methods substantial losses in recovery are possible. Usually, the amount of the sample that is necessary for isotopic analysis requires the methods to be as efficient as possible, with 20 mg of N being the lower limit for routine analysis. Surface water NH4 + concentrations in most aquatic ecosystems are below 5-10 mM, such that 200 to 350 mls of water need to be processed for acceptable precision. Therefore, it is essential to prevent contamination of NH4 +, which can arise from many sources, as ammonia is a gas. In particular, investigators should be keenly aware of any enriched tracer work that is being done in the same laboratory or ship, and if so, separate areas and glassware are warranted. The other sources of possible contamination are from cleaning solutions, smoke, and distilled water. The isotope composition of low levels of ammonium in marine waters can be measured using an adaptation of the ammonia diffusion method. Large volumes of water (up to 4L) are used and during the process, ammonium is converted to ammonia, which is trapped on glass fiber filters (Holmes et al., 1998). To determine the appropriate volume of water required for analysis, it is necessary to know how much ammonium is in the sample (Solarzano, 1969; Strickland & Parsons, 1972; Holmes et al., 1999). This method may be initiated in the field or on-board ship, although diffusions generally last two weeks. Two distinct advantages of this method are that multiple samples can be run concurrently and that the labor involved per sample is minimal.
9.4.2 Isotope analysis of bulk organic matter The isotope composition of bulk tissue samples can be measured using one of two methods. In the first, the organic material is placed in a quartz tube with reduced copper and copper oxide (CuO), sealed under vacuum and combusted in a muffle furnace at 900~ The resulting gases are purified on a vacuum line and are then analyzed by conventional IRMS (for a thorough review see Boutton, 1991). The second more common approach uses a continuous flow system, which comprises an elemental analyzer (EA) directly attached to an IRMS (Fry et al., 1992). Samples are weighed into tin or aluminum boats and loaded into a multiple sample carousel on the EA. The whole sample including the metal boat is combusted and the resulting gases (CO2, N2 and H20) are separated on a gas chromatography column. These gases are then introduced into the IRMS, and both 613C and 615N values can be
Preparation of Ecologicaland BiochemicalSamplesfor IsotopeAnalysis
191
obtained from the same sample. Typical sample sizes range from 100 - 200 mg for animal tissue to 1- 2 mg for sediments. It is good laboratory practice to analyze a standard after every 5 samples to check precision, and ideally the standard should be similar in elemental composition to the samples of interest. Examples of standards include mesquite leaves for the analysis of plant material and collagen for muscle tissue analysis. The EA-IRMS system has many advantages over the traditional sealed tube method, including a rapid throughput of samples (up to 80 samples per day) and sample sizes of less than 100 mg (e.g. organic rich material such as muscle tissue). It should be noted that results from EA-IRMS analysis differ from those of off-line techniques, as a result of an isotopic fractionation which occurs during the introduction of the gases into the mass spectrometer. Therefore, a correction should be applied to the EA-IRMS results which should be calculated on a daily basis through the analysis of known, preferably NBS, standards. In some instances, such as the analysis of muscle, the isotope composition of fatfree tissue may be required to simplify trophic web analysis (e.g. Hesslein et al., 1993). Removal of fats or lipids from whole tissue will affect the 613C values as lipids are typically depleted in 13C relative to bulk material. The techniques employed to remove lipids range from vigorous Soxhlet extraction (e.g. Focken & Becker, 1998) to mild solvent extraction (Hesslein et al., 1993). If the goal of the study is to measure fatfree tissue, then the extraction method employed should aim to remove fats and little else. Generally, this can be easily achieved by a mild extraction technique, such as immersion of the sample in hexane, with gentle agitation and decanting of the solvent. More vigorous techniques may remove other organic compounds, such as the more polar compounds (e.g. cholesterol), which will in turn affect 613C values of the "fat-free" tissue. In addition, vigorous extraction of blood samples may affect 615N values, as compounds such as uric acid and urea may be preferentially extracted (Bearhop et al., 2000). Stable isotope analysis of bulk organic matter has been widely used in ecological research (reviews include Griffiths, 1998; Lajtha & Michener, 1994; Ehleringer et al., 1993; Rundel et al., 1989; Peterson & Fry, 1987). For example, the isotopic composition of plant organic matter can be related to changes in physiology, associated with differences in water use efficiency (Bert et al., 1997; Johnson & Tieszen 1994; Read et al., 1992; Farquhar et al., 1989a). Environmental parameters such as light, water, salinity and air pollution (Buchmann et al., 1996; D u e t al., 1998; Farquhar et al., 1989a) mitigate ~)13C values as well. Stable isotopes of animal tissue are also used to investigate trophic structure, migration, and metabolism (Hobson & Clark, 1992; Koch et al., 1995; Ben-David et al., 1997; Anderson & Polis, 1998; Witt et al., 1998; Schmutz & Hobson, 1998; Wainwright et al., 1998; Jones et al., 1998; Schoeninger et al., 1998; Alisauskas et al., 1998; Marra et al., 1998; Hansson et al., 1997; Nelson et al., 1998; Ostrom et al., 1997).
9.4.3 Compound specific isotope analysis of major biochemicals The isotopic composition of individual organic compounds in ecological and biological samples can be extremely useful in deconvoluting processes and interactions
192
Chapter 9 - M.A. Teece & M.L. Fogel
between organisms in complex systems. In studies where bulk isotope analysis may provide an overview of the whole ecosystem, targeting specific molecules will provide information on the role of individual organisms. This approach relies on different organisms synthesizing different molecules. Through numerous chemical and biochemical procedures, these molecules can be separated and their isotopic compositions analyzed individually. The approach is particularly powerful in studies where physical separation of material is difficult such as in the determination of sources to water column particulates in lakes. Filtering particulate matter from the water column provides an overview of the multiple sources to the lake and the isotopic composition of the complete filter reflects an average signal of input. The isotope composition of specific molecules will provide more detailed information pertaining to the importance of different sources. The 613C of individual n-alkanes indicate different types of higher plant sources (Spooner et al., 1994; Rieley et al., 1991), whereas the isotopic composition of a particular suite of sterols may provide information on phytoplankton inputs (Canuel et al., 1997). A more recent application of compound specific isotope analysis is aimed at elucidating particular metabolic pathways in organisms. The isotope composition of individual compounds is dependent on the enzymatic reactions which occur during their biosynthesis, and therefore retains a signal of their pathway of synthesis. Several studies have exploited these techniques to elucidate the pathways of lipid synthesis in organisms such as bacteria (Teece et al., 1999; Summons et al., 1998; van der Meer et al., 1998) and pigs (Stott et al., 1997a). The compound specific approach requires a significantly greater effort to prepare a sample for analysis, and in some cases this additional work may not be worthwhile. For example, the preparation of a single sample for the analysis of the 613C values of individual amino acids by gas chromatography combustion isotope ratio mass spectrometry (GC-C-IRMS) may take up to three days in the laboratory. Typically, the isotope composition of samples are measured in triplicate with each analytical run being up to 60 minutes in length. As a result of these protracted periods of preparation and analysis, it is difficult to acquire large data sets. Secondly, GC-C-IRMS systems are not abundant and are not simple off-line attachments which are free of problems. As with any analytical instrument, these systems require continual monitoring and maintenance, however the results obtained from such analyses are often worth the hard work involved. Therefore, in determining whether such a compound specific isotope approach is warranted, several questions should be addressed: 9Can bulk isotope analysis provide the answer? ~ Is the determination of sources or biochemical pathways important? 9 Are there multiple sources (pathways) that cannot be resolved by bulk techniques? 9Do the source organisms (pathways) synthesize different molecules? 9Are these compounds specific to the source organism (pathway)?
Preparation of Ecological and BiochemicalSamplesfor Isotope Analysis
193
Once the specific compounds of interest have been determined, the question of analysis techniques need to be addressed. The primary technique to analyze the isotope composition of individual compounds is gas chromatography combustion isotope ratio mass spectrometry (GC-C-IRMS). The sample is injected into a gas chromatograph and the compounds of interest are separated on a well-chosen chromatography column. The column eluent passes through a micro-volume furnace where the material is combusted and the resulting CO2 (or N2) is introduced into the mass spectrometer (Hayes et al., 1990; Eakin et al., 1992). The limiting factor in compound specific analysis using GC-C-IRMS is that only volatile compounds can be analyzed as the inlet system is a gas chromatograph. There has been limited success in designing systems where the inlet system is a liquid chromatograph, though commercial experimental systems have been developed and marketed. In many cases, chemical derivatization of a compound is necessary to synthesize a component that is amenable to gas chromatographic analysis, and therefore GC-C-IRMS analysis. Volatile compounds, such as certain classes of lipids, can be directly analyzed, while amino acids and carbohydrates require extensive chemical derivatization. The analytical procedures to separate particular classes of compounds are numerous, accordingly we will present an overview of suggested approaches. It should be noted that in order to obtain reproducible isotope values, individual compounds must be sufficiently separated on the gas chromatography (GC) column prior to mass spectrometric analysis. Isotopic precision depends on adequate, preferably baseline, separation of components. Therefore, it is important to verify both the methods of GC separation as well as IRMS techniques for the isotopic analysis of specific compounds. It is essential to check the analytical procedures used by the repeated analysis of compounds of identical chemical composition. Checking the methodology in this manner should confirm that the specific compounds of interest are quantitatively extracted from the sample, and also indicate whether any isotope fractionation occurs during analysis. Furthermore, it is good practice to add one or more "internal" standard compounds of known chemical and isotopic composition to the sample prior to extraction/analysis. These internal standards should have similar chemical properties to the compounds of interest, however should not be present in the natural sample or coelute with other compounds. The isotopic composition of these compounds can be measured in the final analyses as a check to indicate the reproducibility of the methods used. For example, nC36 alkane is typically employed as an internal standard in hydrocarbon analyses, and nC19:0 fatty acid used in fatty acid analyses. The approach to measuring the isotope composition of individual compounds requires numerous steps and these should be addressed prior to sample collection (Figure 9.3).
Step 1" Collection and storage of samples for analysis should minimize contamination. GC-C-IRMS is a very sensitive technique, with nanogram quantities of material detected (cf. milligram amounts for bulk isotope analysis). Human hands
Chapter 9 - M.A. Teece & M.L. Fogel
194
Step I
Sample Collection Freeze-dry I
Dry Sample
]
Bulk Isotopic Analysis
Step 2
Extraction
Complex Mixture of Compounds
Step 3
Chemical Separation
I I
I
CompoundCroup
I I
I
CompoundCroup
Derivatization Scheme
Derivatization Scheme
I Step 4
GC-amenable Compounds [ [ GC-amenable Compounds [ Step 5
Identification
Identification Step 6
I Analysis by GC-C-IRMS
I Analysis by GC-C-IRMS
Step 7
9 . 3 - A generalized analytical scheme for the stable isotope analysis of individual organic compounds in biological samples. The numbered steps refer to analytical procedures that are specific to the compounds of interest, and are discussed in more detail in the text. Figure
Preparationof Ecologicaland BiochemicalSamplesfor IsotopeAnalysis
195
contain many of the lipids and proteins that will be subsequently analyzed, so handling should be minimized and gloves used. The use of chemical preservatives, such as ethanol or formaldehyde, is discouraged and samples should ideally be stored in clean glassware or sterile plasticware. For most analyses, material should be freeze-dried as discussed above.
Step 2" Extraction of the compounds of interest requires the correct extraction method. Hydrophobic lipids and fats are extracted using a series of organic solvents, whereas carbohydrates and amino acids are hydrophilic and sulfuric and hydrochloric acids are employed in their extraction. The chemicals, materials and conditions differ for all compounds and selecting the correct method is paramount in the successful extraction of material. Prior to extraction, (internal) standard compounds should be added to check the methodology. Step 3" The mixture of extracted compounds may be complex and require additional separation techniques to isolate the compounds of interest in greater purity. Several methods of separation can be used including column chromatography, thin layer chromatography, ion exchange chromatography, and electrophoresis. The particular techniques required differ with each compound class and extensive literature can be found relating to these analytical challenges. Step 4: The mixture of compounds may be chemically derivatized to produce compounds, which can be analyzed on a gas chromatograph. Volatile lipid compounds (e.g. hydrocarbons) can be analyzed directly, such as those present in epicuticular waxes of higher plants (Rieley et al. 1991; Collister et al. 1994; Lockheart et al. 1997). However, fatty acids and alcohols require derivatization prior to analysis, in order to attain sufficient chromatographic separation required for reproducible isotopic analysis. Numerous techniques for derivatization are published and depend on the chemical functionality of the particular compounds: Fatty acids are methylated whereas alcohols are typically analyzed as the trimethylsilyl ethers. Step 5: To acquire reproducible isotope numbers, the components of complex mixtures must be adequately separated on the gas chromatography (GC) column to reduce the influence of contaminating or co-eluting peaks during analysis. Separation of compounds using GC requires the correct choice of chromatographic column. There are a multitude of chromatographic columns available from several vendors, varying in polarity and size, and the choice of column is dependent on the chemical properties of the components being analyzed. Nonpolar columns, such as DB-1, HP-1, Ultra-l, are used for hydrocarbon analyses and a column containing a 5% phenyl substituted methylpolysiloxane phase (HP-5, CP-Sil 8C) is useful for analysis of derivatized alcohols and fatty acids. Step 6" The identity of the specific compounds isolated should be confirmed by GCMS prior to isotope analysis. This step will ascertain whether the separation on the chosen GC column is adequate and will also indicate whether any contam-
196
Chapter 9 - M.A. Teece & M.L. Fogel
ination was introduced during the analytical scheme.
Step 7: Stable isotope analysis of the individual compounds by GC-C-IRMS. The analyses should be performed in triplicate to determine variations on instrumental analysis. Analysis of standards of similar chemical structure will aid in determining the reproducibility of results. Step 8: If a compound was derivatized prior to analysis, then the isotopic composition of the original compound must be calculated. Derivatization usually involves the addition of a carbon-containing component in order to produce the volatile compound. Calculation of the isotopic composition of the original compound requires consideration of the contribution of the added carbon to be assessed if measuring ~)13C values. Therefore, it is essential to analyze compounds of known chemical and isotopic composition to calculate the effect of the derivatizing agent. For example, fatty acids are methylated prior to analysis and the isotopic contribution of the added methyl carbon is assessed by isotopic analysis of underivatized and subsequently derivatized fatty acid standards (see Abrajano et al., 1994:). Simple mass balance calculations allow the contribution of the 613C-derivatizing agent carbon to be assessed, which can then be applied to the unknown compounds to calculate the original isotope compositions. These steps are general guidelines and what follows are more specific examples of approaches to analyzing specific molecules. The approaches below are suggestions and each method can be altered or customized to the particular question being addressed.
9.4.3.1. Isotope composition of individual lipids
Lipids are a class of hydrophobic compounds that include fatty acids, phospholipids, triacylglycerides, alcohols, hydrocarbons and sterols (e.g. cholesterol). Lipids perform numerous functions including acting as a source of energy (e.g. triacylglycerides) to a protection mechanism against predators (e.g. plant wax esters). The carbon isotopic composition of individual lipids reflects both the isotopic composition of the carbon source utilized by the organism and isotopic fractionations accompanying biosynthesis. Isotopic fractionations which result from the enzyme-mediated reactions during biosynthesis of lipids are ultimately dependent on environmental conditions. Therefore, the isotope composition of individual lipids can provide valuable information on both the carbon source and potential changes in the organisms environment. Such an approach has been used in several studies to determine changes in past vegetation patterns by isotopic analysis of higher plant-derived hydrocarbons (France-Lanord & Derry, 1994; Bird et al., 1995; Lichtfouse et al., 1995; 1997) and lignin-phenols (Goni & Eglinton, 1996) in lake and marine sediments. Extraction of lipid compounds is a relatively simple procedure, however some caution should be applied to the methods employed. Ideally, the samples should be dry, having been freeze-dried as described above. There are two main methods of lipid extraction, both of which use similar organic solvents. The choice of organic sol-
Preparation of Ecological and Biochemical Samples for Isotope Analysis
197
vents in the extraction of lipids is important and for most procedures mixtures of dichloromethane and methanol are applicable [dichloromethane should be used rather than chloroform - dichloromethane has similar chemical properties to chloroform but is substantially less harmful]. Glass distilled or Optima grade solvents should be used whenever possible. All glassware should be thoroughly cleaned and rinsed with solvents prior to use to minimize contamination. One method for lipid extraction uses a Soxhlet system, in which clean boiling solvents (mixtures of dichloromethane and methanol) are refluxed through the sample for periods ranging from 30 minutes to 24 hours. This method is best employed when extracting lipids from a complex matrix such as sediments, soils and some woods, where the compounds of interest may be physically trapped or weakly bound in the matrix. For biological material, particularly living tissue (e.g. plants, muscle, and blood), a simpler and less harsh extraction scheme should be used. This second extraction method (Box 9.4) uses a mixture of dichloromethane and methanol at room temperature, with repeated mixing using an ultrasonic tank or vortex mixer (modified from Bligh & Dyer, 1959). Once the lipids have been extracted from the sample, the task of separation and derivatization of the particular compounds of interest needs to be addressed. The total organic extract of a sediment may contain more than 80 different compounds, so that one or more clean-up steps may be required before analysis by GC-C-IRMS. The goal of these procedures is to obtain a sample fraction, which contains the molecules of interest in a mixture that can be adequately separated on the gas chromatograph. In the case of hydrocarbon analysis, a single purification step allows adequate separation of more than 20 compounds (e.g. Freeman et al., 1990; Bjoroy et al., 1991) whereas low concentrations of cholesterol may require multiple clean up steps to isolate the compound in sufficient purity (e.g. Canuel et al., 1997). Similarly, analysis of ligninphenols, a class of compounds that are uniquely synthesized by vascular plants, requires a complex procedure of chemical oxidation and derivatization procedures prior to analysis by GC-C-IRMS (Goni & Eglinton, 1996).
BOX 9.4 - Lipid extraction of plant tissue
Freeze-dried leaves are ground to a fine powder using a pestle and mortar. A mixture of dichloromethane:methanol (1:1) (Optima grade) is added to the plant samples (15-20mg) in a precombusted glass test tube fitted with a Teflon-lined cap. The mixture is thoroughly mixed for up to 10 minutes, using a combination of ultra-sonication and vortex mixing. If required, the sample can be centrifuged (1,000g; 5 mins) to pellet the plant matter, and the solvent removed with a clean glass pipette. This procedure should be repeated 5 times to ensure complete extraction with 5ml of clean solvent mixture being used for each cycle. The extract is transferred to a round-bottomed flask and the solvent removed using a rotary evaporator. The total organic extract (TOE) is redissolved in the minimum dichloromethane required to quantitatively transfer it to a clean precombusted vial fitted with a Teflon-lined cap. The solvent is removed under a stream of nitrogen at room temperature to produce the dried TOE, which is stored in a freezer prior to further treatment.
198
Chapter 9 - M.A. Teece & M.L. Fogel
BOX 9.5 - Isotope composition of individual fatty acids from muscle tissue
After extraction of muscle tissue using dichloromethane and methanol, the total organic extract (TOE) is subjected to mild alkaline hydrolysis to release the esterified fatty acids. A 6% KOH solution in methanol (5 ml) is added to the TOE, the tube tightly sealed and allowed to react overnight at 50~ After cooling, 3 ml of double-distilled deionized water is added and the neutral lipids extracted three times (each 5 ml) into a mixture of hexane:diethyl ether (9:1). Neutral lipids are transferred, and stored dry, in glass vials with Teflon-lined caps. The remaining aqueous methanol phase is acidified to pH 2 with 6N HC1 (Pierce Chemicals, constant boiling HC1). The fatty acids present in this fraction are extracted three times (each 5 ml) into a mixture of hexane:diethyl ether (9:1) and transferred to a clean precombusted glass test tube. Solvent is then removed under a stream of nitrogen at room temperature. Fatty acids are subsequently converted to their corresponding fatty acid methyl esters (FAMES). 200ml of a solution of 14% BF3 in methanol (Pierce Chemicals) is added to the fatty acid fraction, the tube tightly sealed and allowed to react for 15 minutes at 60~ After cooling, 3ml of doubledistilled deionized water and 3ml of hexane is added to the mixture. FAMES are extracted into the hexane fraction (3 x 3 ml), transferred to a glass vial and the solvent removed under a stream of nitrogen. Samples should typically be analyzed within two days of preparation using a nonpolar GC column (Ultra-i). Measurement of the ~)13C values of individual fatty acids (see Box 9.5) requires a single clean-up step and a subsequent derivatization step. After organic extraction of the sample, the fatty acid fraction is isolated using ion exchange chromatography and converted to the corresponding fatty acid methyl esters (FAMES). The 613C values of fatty acids have been reported in studies of plants (Collister et al., 1994; Rieley et al., 1991; Vogler & Hayes, 1980), changes in diets (Rhee et al. 1997; Trust-Hammer et al., 1998; Gilmour et al., 1995a), and in ecological studies of mussels and shrimp in hydrothermal vents (Pond et al. 1998; Rieley et al., 1999). The techniques to analyze particular lipid c o m p o u n d classes are reported in the literature (e.g. Blau & Halket, 1993) and require various chemical separation and derivatization techniques. Isotopic analysis of hydrocarbons has been used in studies ranging from palaeoenvironmental reconstruction (Rieley et al., 1991; Freeman et al., 1994; Bird et al., 1995) and elucidation of input sources of vascular plants to coastal sediments (Canuel et al., 1997), to the feeding habits of insectivorous bats (Des Marais et al., 1980). The isotopic analysis of PCBs (polychlorinated biphenyls) and PAHs (polyaromatic hyrdrocarbons) may also be employed to determine sources of toxic chemicals in the environment (O'Malley et al., 1994; Ballentine et al., 1996; Jarman et al., 1998). Alcohols can be analyzed as trimethylsilyl (TMS) derivatives (Jones et al. 1991) and such techniques have been applied to elucidation of dietary sources through the m e a s u r e m e n t of 613C of cholesterol in pigs (Stott et al. 1997a) and fossil whalebones (Stott et al. 1997b).
9.4.3.2. Isotope composition of individual amino acids The amino acids are the building blocks of proteins, and the isotopic composition of these c o m p o u n d s reflect their pathways of biosynthesis. Amino acids are ubiquitous c o m p o u n d s that give very little clue as to their origin or source by their relative abundance alone. The isotopic compositions of these molecules, however, are diverse and can be quite powerful in delineating pathways of biosynthesis or processes of
Preparation of Ecological and Biochemical Samples for Isotope Analysis
199
diagenesis (Macko et al. 1987; Engel et al., 1990; Engel & Macko, 1997; Fogel et al., 1997; Fogel & Tuross, 1999; Fantle et al., 1999). Predictable and fundamental isotopic patterns have been studied in a diverse group of organisms from differing ecological and geological contexts. Typically in plants, glycine and aspartic acid are the most enriched in 13C with valine being the most depleted (Fogel et al., 1997; Fogel & Tuross, 1999). The distribution of isotope signatures of amino acids in higher trophic level organisms is dependant on, and indicative of, the food source. The 813C and 815N values of individual amino acids in such organisms may be predicted on the basis of their pathway of biosynthesis (non-essential amino acids) or their direct incorporation from the diet (essential amino acids) (Fogel et al., 1997). Isotopic compositions of amino acids have been analyzed by off-line combustion after the compounds had been separated by high performance liquid chromatography (Abelson & Hoering, 1961; Macko et al., 1987; Hare et al., 1991; Balzar et al., 1997). As stated, liquid chromatography has yet to be routinely interfaced with an isotope ratio mass spectrometer, therefore a majority of the analyses are now performed on derivatized amino acids separated by gas chromatography. Tissues are hydrolyzed with HC1, dried down, and then derivatized by a number of different methods (see Chapter 8). The method detailed in Box 9.6 is straight-forward, robust, and a majority of the major biological amino acids can be separated for isotopic analyses. In addition, the method is suitable for chiral detection of D- and L- amino acids (Silfer et al., 1991; Metges et al., 1996). A disadvantage of this method is the addition of fluorine molecules that provide volatility can contaminate oxidation furnaces and reduce the working lifetime of these furnaces. The isotope composition of amino acids is determined by isotope fractionations associated with the specific enzyme-mediated reactions which occur during their synthesis. Glycine is commonly the amino acid with the most 13C-enriched isotopic comBOX 9.6- Amino acids in bone collagen
Samples of isolated collagen (Ambrose, 1990; Koch et al., 1994) are weighed (1-3 mg), and loaded into pre-combusted hydrolysis tubes, and hydrolyzed under an atmosphere of N2 in lml of 6N constant-boiling HC1 (Pierce Chemical) (20 hours at 110~ The resulting hydrolysates are dried under a stream of N2 at 100~ and the released amino acids esterified by I ml of anhydrous acidified iso-propanol (1 hour at 110~ and subsequently acylated by 0.5 ml trifluoroacetic anhydride (TFAA) in 0.5 ml dichloromethane (10 min. at 110~ Samples are analyzed on a GC-CIRMS system using a non-polar GC column (e.g. Ultra-I). A suite of amino acid standards are derivatized and analyzed with each batch of samples. For carbon isotope analyses, about 2 mg of hydrolyzed protein is injected onto the column. After various splits in the injector (1:10) and the open split (1:2), about 10 ng of carbon per peak enters the ion source of the mass spectrometer. For nitrogen isotope analysis, about 3 mg of sample is injected in the splitless mode, such ca. 100 ng of amino acid is converted to N2. Each sample should be analyzed in triplicate with corresponding standard amino acids. Reproducibility for carbon isotopes ranges from _+0.2%o to _+1.0%, depending on peak size, with a typical error of _+0.4 %o. For nitrogen isotopes, because the derivative contains no added nitrogen, values are determined directly. Standard deviations on the major peaks (e.g. proline, aspartate, glutamate, lysine, and arginine) are about _+0.5%. Well-resolved but very small peaks (e.g. valine and leucine) have higher errors _+1-2%.
200
Chapter 9 - M.A. Teece & M.L. Fogel
position. It has a simple and central role in metabolism. In terrestrial plants, glycine is formed from serine during photorespiratory processes. Isotopically heavier carbon is shunted into this pathway with 13C depleted carbon going into subsequent amino acids synthesized in the Krebs cycle. Glycine is a key intermediate in the formation of porphyrins and purines, thus an understanding of its isotopic composition may be critical in understanding the isotopic composition of molecules, such as chlorophyll. Many amino acids are essential for the growth of animals and the isotopic signatures imparted to these particular essential amino acids can be retained as proteins from primary producers are digested. For example, in omnivorous animals, meat or other animal protein is the primary contributor of lysine to the diet. Therefore, the isotopic composition of lysine should indicate the source of dietary protein, whereas the b u l k ~)13C should reflect the primary carbohydrate or energy source (Fogel et al., 1997). Furthermore, isoleucine can be 6-10%o more positive in 613C than leucine and valine in bacterial and algal proteins. The isotopic difference between leucine and isoleucine transfers up the food chain into fish, birds, and mammals and therefore the 613C of these components can be used to trace trophic interactions (Fogel et al., 1997). T h e ~)13C of
alanine may be an indicator of the energy status of an organism. Alanine is one of the most variable amino acids in terms of isotopic composition, and derives its carbon skeleton directly from that of pyruvate. Pyruvate, and alanine, are central metabolites between the TCA cycle and carbohydrate metabolic pathways (e.g., glycolysis in case of heterotrophs or the Calvin cycle of photosynthesis). Lastly, ranges in gl3C of amino acids are also useful indicators of source and processes. For example, in higher plants, a range of 25%0 in 613C is common, and in phytoplankton, the spread in 613C varies from 15 to 20%o, whereas in sharp contrast, cultured bacterial amino acids have a very narrow range of 6-8%o (Macko et al., 1987; Fantle et al., 1999; Fogel et al., 1997; Teece & Fogel, unpublished results). The different ranges of 613C values is governed by isotope fractionations occurring during amino acid biosynthesis and such distributions may be used in future to determine sources of these compounds to natural ecosystems. 9.4.3.3.
Isotope composition of carbohydrates
The largest reservoir of carbon in the biosphere resides in the carbohydrates, and can comprise up to 40% of the dry weight of bacteria and up to 70% of that of vascular plants. Carbohydrates serve as metabolic storage products, as carbon and energy sources in non-photosynthetic metabolism (e.g. starch, inulin, glycogen, mannitol), and are associated with cell walls and membranes that provide protection, stability and strength (e.g. cellulose, hemicellulose, pectins, chitin, agar, peptidoglycans, lipopolysaccharides and glycolipids). Carbohydrates represent the major form of photosynthetically fixed carbon and are critical energy components of food webs. In the oceans, most of the organic material produced by phytoplankton is consumed by herbivorous zooplankton and protozoans in the photic zone, which are in turn consumed by higher trophic level organisms. Carbohydrates, in particular glucose, represent the major organic carbon
Preparation of Ecological and Biochemical Samples for Isotope Analysis
201
BOX 9.7- Isotope composition of individual monosaccharides in zooplankton Samples of isolated zooplankton (20rag) are added to 72% H2SO4 (lml) at room temperature for 90 minutes. After dilution to 1.2M H2SO4, the sample is further hydrolyzed for 3 hours at 100~ An internal standard (myo-inositol) is added to the hydrolysate, and both are subsequently neutralized to pH 6.5 with BaCO3. The resulting BaSO4 precipitate is washed and removed by centrifugation, and the neutralized hydrolysate reduced to 2ml by rotary evaporation. The released monosaccharides are reduced to alditols by reaction with 0.5ml of a freshly prepared solution of NaBH4 (500rag in 10ml H20) for at least 12 hours at room temperature. The residual NaBH4 is decomposed by the addition of glacial acetic acid, and after effervescence ceases (pH 5.5), a second internal standard (erythritol) is added, and the solution evaporated to dryness under reduced pressure. Boric acid is removed by repeated additions of methanol, followed by evaporation to dryness. Acetylation is performed in sealed tubes in a solution of pyridine:acetic anhydride (1:1) for 2 hours at 100~ After the addition of water (2ml), the alditol acetates are extracted into dichloromethane (3 x 2ml) and evaporated to dryness under a stream of nitrogen. The derivatives are analyzed in triplicate by GC-C-IRMS using a polar column (e.g. BPX-70) and compared to a suite of standard compounds treated in an identical manner.
component transferred between trophic levels. The isotopic composition of glucose provides an important indicator of the processes and pathways of energy utilization in ecosystems (e.g. Moers et al., 1993). Measurement of the isotope composition of carbohydrates is an analytical challenge, and therefore there are few published reports of analyses. Macko et al. (1990; 1991) measured the 613C of individual monosaccharides by off-line combustion of components which had been previously isolated by liquid chromatography. Carbohydrates are highly hydrophilic and the challenge to produce derivatives, which are amenable to GC analysis, requires several chemical reactions and derivatization steps (Moers et al., 1993). An approach to the analysis of individual monosaccharides is presented in Box 8.7 (modified from Cowie & Hedges, 1984), however as many steps are involved it is paramount to test these procedures with standard compounds to determine whether the reaction scheme results in isotope fractionation during derivative preparation. In the majority of biological material, carbohydrates are enriched in 13C relative to amino acids and lipids, and also bulk tissue. Individual monosaccharides can be up to 10%o enriched in 13C relative to whole tissue, such as in cyanobacterial mats (Moers et al, 1993). In herbaceous and woody pant tissues, cellulose and hemicellulose are typically enriched in 13C by 1-2%o relative to whole plant material (Benner et al, 1987). During the diagenesis of organic matter, the isotopic composition of carbohydrates provides a means to trace sources of these compounds, either resulting from new bacterial production or degradation (Macko et al., 1991). In the future, the isotopic composition of individual monosaccharides may provide additional vital information in tracing sources of organic matter and the transfer of energy in ecosystem studies. The methods for the isotopic analysis of these compounds has been established, and now provides an additional tool for investigating ecological questions on a molecular scale.
'
202
Chapter9 - M.A.Teece & M.L.Fogel
9.5 Conclusions
Because of new technological advancements in the elemental analyzers linked to isotope ratio mass spectrometers (e.g., Kornexl et al., 1999a), stable isotope analysis is quickly becoming a standard measurement for interpretation of various biological and environmental parameters. We easily imagine that in the next ten years these instruments will take their place along side of nutrient autoanalyzers, fluorescence instruments, and UV/V is spectrophotometers. The challenge then arises for ecologists to use stable isotope tools as wisely and as carefully as those who developed the early methods, perfected the techniques, and pioneered their use. Graduate students in ecology are currently learning the "ins and outs" of stable isotope fractionation and systematics, as part of their core education. The compound specific techniques are deeply rooted in chemistry and will probably remain in the domain of geochemists and biochemists. Chemistry departments, on the other hand, are finding that creating links to environmental and biological colleagues through interdisciplinary collaborations open up new ways of examining complex ecosystems. Students in ecology will naturally gravitate to laboratories where these more complex isotopic methods are being practiced and applied to biological problems. The ability to couple the power and speed of elemental analyzer online methods to the specificity of individual compound work will be key for the full integration of stable isotopes as essential parameters for understanding and delineating ecosystems and ecosystem processes.
Acknowledgments The authors would like to thank Henry Fricke, Susan Ziegler, Timothy Filley, Joachim Bebie, Matthew Wooller, and Richard Ash for helpful advice and discussions during the preparation of this manuscript. Helpful comments were provided by the reviewers Brian Fry and Page Chamberlain.
Handbook of Stable Isotope Analytical Techniques, Volume 1 P.A. de Groot (Editor) 9 2004 Elsevier B.V. All fights reserved.
C H A P T E R 10
Extraction of Dissolved Inorganic Carbon (DIC) in Natural Waters for Isotopic Analyses E. A. Atekwanal & R. V. Krishnamurthy2 Department of Geology, Indiana University Purdue University, 723 W. Michigan Street, SL 122, Indianapolis, IN 46202-5132, USA 2 Department of Geosciences, Western Michigan University, 1187 Rood Hall, Kalamazoo, MI, 49008, USA email: I
[email protected], 2
[email protected] 1
10.1 Introduction 10.1.1 Dissolved inorganic carbon (DIC) Dissolved inorganic carbon (DIC) consists of CO2(aq), H2CO3, HCO3-, and CO32-. Its concentration in natural waters is governed by the interaction of various carbon species and can be represented by the following reactions (Garrels & Christ, 1965). CO2(g) + H 2 0 *--, H 2 C O 3 H2CO3 ~ H + + HCO3H + + C a C O 3 ,--, H C O 3 - + C a 2+
[10.1a] [10.1b] [10.1c]
The source of CO2(g) initiating the above reaction [10.1a] in natural waters is either atmospheric or biogenic. The biogenic component is predominantly from root respiration or oxidation of organic matter. The carbonate component [10.1c] in ground and surface waters is derived from the dissolution of soil and/or bedrock carbonates. In studies of DIC in natural waters, determining both the concentration and carbon isotopic ratio of DIC (~113CDIC) are useful. The concentrations of DIC and ~)13CDIC in natural waters are controlled by carbon flux, as well as, biogeochemical cycling in the system. Any process that contributes to or removes carbon from natural waters influences the DIC pool. Changes in DIC concentration can occur due to dissolution/precipitation reactions within the aqueous carbonate system, biogenic uptake and release of CO2 and from mixing of carbon from different sources. Changes in the ~)13CDIC would result from fractionation accompanying carbon transformations during the above processes.
10.1.2 Utility of DIC measurements
In hydrologic systems, the concentrations of various species of DIC are indicative of the reaction pathways of waters. Using thermodynamic principles, the concentra-
204
Chapter 10- E.A. Atekwana & R.V. Krishnamurthy
tion and ~)13CDIC can be used to define the DIC and hydrologic evolution of waters (e.g. Deines et al., 1974; Wigley et al., 1978). Examples of a few studies in which DIC has also been utilized include, tracing groundwater recharge (e.g. Landmeyer & Stone, 1995; Cane & Clark, 1998), studies of carbon evolution of groundwater (e.g. Reardon & Fritz, 1978; Chapelle & Knobel, 1985; Rose & Long, 1989; White & Chuma, 1987; McMahon & Chapelle, 1991; Cai et al., 2003), carbon cycling in streams (e.g. Pawellek & Veizer, 1994; Flintrop et al., 1996; Yang et al., 1996a; Atekwana & Krishnamurthy, 1998; Helle et al., 2002; Finlay, 2003), carbon cycling in lakes (e.g., Rau, 1978; Quay et al., 1986; Herczeg, 1987; Weiler & Nriagu, 1987; Wachniew & Rozanski, 1997; Nakamura et al. 1998; Vreca, 2003), determining the extent of surface water-groundwater interaction (e.g. Taylor & Fox, 1996) and assessing microbial influence of groundwater isotopic composition in natural and polluted groundwater (e.g. Chapelle et al., 1988; Chapelle & McMahon, 1991; Nascimento et al., 1997; Hunkeler et al., 1999; Fang et al., 2000; Pombo et al., 2002).
10.1.3 Principles of DIC extraction DIC in natural waters is extracted by converting carbon components into CO2 gas. Quantitative measurements of the evolved CO2 permit determination of DIC concentration; the CO2 is introduced into an isotope ratio mass spectrometer (IRMS) for 613C determination. Isotope ratios are reported in the 6-notation where: ~) (%0)- ((Rsample / Rstandard)-1) x 10 3
[10.2]
R is 13C/12C. Values are reported relative to VPDB (Vienna Pee Dee Belemnite). In studies where the carbon-14 activity of DIC is of interest, the extracted CO2 is prepared for C-14 analysis and measured using counting techniques or Accelerator Mass Spectrometry (AMS). 10.2 DIC extraction
The information presented in this chapter is essentially a guide for selecting an appropriate technique for DIC extraction for isotopic analysis. The focus is on extraction techniques routinely used for DIC isotopic analyses. Mass spectrometric analyses are not discussed because of differences in mass spectrometric types, individual machine configuration and analytical setup. The techniques presented assume that mass spectrometric CO2 analysis is off line, although the procedures described can be adapted to online systems with slight procedural modifications. For micro-extraction of DIC and ~)13CDIC determination, the techniques of Graber & Aharon (1991) and Salata et al. (2000) can be consulted. For extraction of individual carbon components for isotopic analysis from water, the technique of Games & Hayes (1976; 1977) can be used. St-Jean (2003) describes a method for automated 613C analysis of DIC on continuous-flow isotope ratio mass spectrometer (CF-IRMS). The techniques described here do not consider water sampling methods or protocol. It is assumed that sampling of groundwater, surface water (grab or depth specific samples) or pore water in sediments or rock will be done in a manner that preserves sample integrity, such as preventing CO2 degassing from the sample or atmospheric
205
Extraction of Dissolved Inorganic Carbon (DIC) in Natural Waters for Isotopic Analyses
CO2 contamination. Techniques currently used for DIC extraction for isotopic analy-
ses can be categorized as variants of the Gas Evolution, Precipitation, Precipitation and Gas Evolution, and Vapor Phase Equilibration techniques (Figure 10.1). 10.3 Gas evolution technique In the gas evolution technique, DIC in a water sample is converted to CO2 under vacuum conditions by acidification (Mook, 1968; Tan et al., 1973; Games & Hayes, 1976; 1977; Reardon et al., 1979; Hassan, 1982; Graber & Aharon, 1991; Holt et al., 1995; Atekwana & Krishnamurthy; 1998; Cane & Clark, 1998). The yield of CO2 extracted is converted to DIC concentration and the CO2 is introduced into a mass spectrometer for 613CDIC determinations. The main advantage of the gas evolution technique is the simultaneous determination of DIC concentration and the 813CDIC of the same water sample. Although the gas evolution technique is simple, it can be plagued with problems that potentially cause erroneous concentration and 613CDIC determinations. Hassan (1982) has shown that the gas evolution technique seems to suffer from CO2 gain/loss from water in sample containers during shipping and laboratory storage prior to DIC extraction, and incomplete transfer of evolved gas phase in sample containers to the vacuum line
~
$
D][C
1"Coi c
8~sCo,c
$
l~ol c
s~n]~
,.~C03
(
l'~oiC
I-~ a c l ~ e
DIG
HeMsp~e
H~O ~po~
81RaOlC
Figure 10.1 - Flow diagram showing common techniques and major steps involved in DIC extraction and possible measurements associated with CO2 extracted by each method.
206
Chapter 10- E.A. Atekwana & R.V. Krishnamurthy
during extraction. Additionally, water sample storage problems can be compounded by bacterial activity (including photosynthesis and respiration) which could alter DIC concentrations and ~513CDICin samples not stored in opaque glass bottles, cooled to 4~ and/or preserved by the addition of a bactericide such as HgC12 (e.g. Mook, 1968; Yang et al., 1996a; Cane & Clark, 1998). One technique that overcomes these problems, is simple, and allows for accurate determination of both DIC concentrations and 613CD~C (Atekwana & Krishnamurthy, 1998) is detailed below.
10.3.1 The Vacutainer Gas Evolution Technique (Atekwana & Krishnamurthy, 1998) This procedure uses vacutainer septum tubes (blood serum vials) to collect, react, and extract DIC from natural water samples. We call it the vacutainer technique to distinguish it from other gas evolution techniques. The vacutainer tubes are preloaded with 85% phosphoric acid and magnetic stir bars. In the lab, air is evacuated from the septum tubes after loading with phosphoric acid and magnetic stir bars. In the field, water samples are introduced into the septum tubes using a syringe. Sample collection and injection takes less than 30 seconds. The acid-water reaction begins immediately upon injection. The evolved CO2 is extracted in the laboratory, where yields are measured and isotope ratios determined.
10.3.1.1 Procedure Sample container preparation List of supplies~equipment needed 9 16 x 100 or 16 x 165 mm glass septum tubes (VACUTAINER| Serum Tubes, Becton Dickson & Company, Franklin Lakes, NJ 07417) 9 85 % phosphoric acid 9 13 x 8 mm magnetic stir bars (Fisher brand| cat. No. 14-511-62)1 9 Extraction needle (consisting of an inner luer ground joint joined to a piece of 9mm diameter tube by glass blowing. A 26 gauge needle (small diameter to prevent septum coring) is affixed to the luer end 9 Laboratory vacuum extraction system To prepare the septum tubes for DIC extraction, the septum of each tube is removed and a magnetic stir bar inserted into each of the tubes. If the CO2 is removed from the septum tubes using ultrasonic water bath, the magnetic stir bar is omitted. Approximately 0.5 ml of 85% phosphoric acid is dispensed into each tube and their septa replaced. Each septum tube is evacuated of air introduced into the tubes while loading with phosphoric acid and magnetic stir bars using an extraction needle attached to the laboratory vacuum line by way of a Cajon union (Figure 10.2, insert A). Evacuation of air takes a few minutes and is complete when the vacuum gauge on the system has recovered to proper vacuum conditions. During this step, most of the moisture and any CO2 in the phosphoric acid is removed.
1. A modification of the DIC extraction technique where CO2 is removed from the septum tubes using an ultrasonic water bath does not require use of magnetic stir bars (see DIC extraction section).
Extraction of Dissolved Inorganic Carbon (DIC) in Natural Waters for Isotopic Analyses
207
Figure 10.2- Simplified schematic showing a laboratory v a c u u m extraction system. Insert (A) shows an extraction needle used to evacuate septum tubes pre-loaded with phosphoric acid and a magnetic stir bar. Insert (B) shows a DIC extraction set-up used to transfer CO2 from water-acid reaction in septum tubes into the laboratory vacu u m system. This set-up consists of a heater~stirrer, a water bath (or an ultrasonic water bath), an extraction needle and an external U-tube moisture / ice trap.
Water sample collection
List of supplies~equipment needed 10-ml plastic syringes with a luer lock tip (Becton Dickson & Company, Franklin Lakes, NJ 07417) 9 26-gauge needles (Becton Dickson & Company, Franklin Lakes, NJ 07417) 9 Prepared septum tubes (Pre-evacuated septum tubes loaded with phosphoric acid and magnetic stir bars)
9
In the field, a known volume of water (usually 5 or 10 ml) is introduced into the prepared septum tube using a plastic syringe fitted with the 26-guage needle. Before collecting a water sample, the needle is removed from the syringe in order to facilitate rapid uptake of water into the syringe and to prevent possible vacuum degassing of the sample due to the small needle size. After filling, the needle is quickly reattached to the syringe. Air bubbles, if present, are removed by tapping the syringe gently and excess water subsequently expelled to obtain the desired volume. Filtration of the water before injection in the septum tube can be accomplished using syringe filter with a female luer-lok inlet and male luer slip outlet. The sample is then injected into the prepared septum tube. The tube is shaken for a few seconds to mix the acid and water. The reaction between the water and the phosphoric acid starts instantaneously.
208
Chapter 10 - E.A. Atekwana& R.V. Krishnamurthy
Reacted samples are transported as such to the laboratory for CO2 extraction. No special storage conditions are needed except care to prevent breakage of the glass sample tubes.
Laboratory extraction of C02 from the water~acid reaction mixture List of supplies~equipment needed 9 Septum tubes with reacted samples 9 50~ water bath (water in a beaker on a hot plate/stirrer is sufficient) or water in an ultrasonic agitator (e.g. Fisher Scientific Catalog 15-335-6) 9 -80~ ice slush (mixture of dry ice and organic solvent) ~ Liquid nitrogen 9 External U-tube trap (9 mm Pyrex glass tube blown to fit vacuum system; see Figure 10.2, insert B) 9 Extraction needle (consisting of a 23-gauge needle attached to an inner luer ground joint, joined to a piece of 9 mm tube by glass blowing) 9 Laboratory vacuum extraction system Evolved CO2 in septum tubes is extracted using an external modification to the vacuum system (Figure 10.2, insert B). During extraction, the septum tubes with reacted samples are placed in a water bath at about 50~ (or in an ultrasonic water bath). A removable external U-tube trap is utilized to ensure that the bulk of moisture is trapped prior to gas entry into the extraction line. A Cajon union is used to connect the external U-tube trap to the vacuum system. The extraction needle is attached to the external U-tube trap by way of a threaded vacuum seal. Prior to CO2 extraction, the needle attached to the external U-tube trap (Figure 10.2, insert B) is partially inserted into the septum of the tube containing the reacted sample, thereby covering the needle opening. This step is required to evacuate the dead volume between the stopcock on the vacuum line and the needle-tip (Figure 10.2). After evacuation, the needle is pushed through the septum and the CO2 is extracted dynamically while the reaction mixture is continuously stirred. This step is carried out for 5 to 10 minutes. It is noteworthy to mention that stirring was found necessary in this procedure. Without stirring the extraction efficiency for standard solutions varied between 85 and 93 %. If the ultrasonic water bath is used, sonic agitation is sufficient to effectively transfer all CO2 from the sample/acid mixture in the septum tube to the vacuum line. The extracted CO2 is trapped in the vacuum line with liquid nitrogen (-180~ Residual non-condensable gases are pumped away and the trapped CO2 is purified cryogenically by substituting the liquid nitrogen in the trap with-80~ ice slush. The CO2 released is then re-trapped in a calibrated cold finger using liquid nitrogen and its yield measured manometrically. The CO2 is collected from the cold finger and introduced into the mass spectrometer for carbon isotopic measurements.
Extraction efficiency and r
precision of the Vacutainer Gas Evolution Technique
Several experiments were conducted to verify the accuracy and precision of the vacutainer gas evolution technique. Initially, C02 evolved from "solid-form" samples,
Extraction of Dissolved Inorganic Carbon (DIC) in Natural Waters for Isotopic Analyses
209
each representing different amounts of solid Na2CO3, were reacted with 100% phosphoric acid and extracted using the method of Krishnamurthy et al. (1997). The amount added, the experimental yield and percentage yield as CO2 and the corresponding 613C of the extracted CO2 are presented in Table 10.1. The data shows that the extraction efficiency (% yield as CO2) of the solid standard is better than 99% with a ~13C of 7.5 + 0.1%o. Next, a solution was prepared using the solid Na2CO3 for DIC extraction. Evolved CO2 from 5-ml aliquots of the standard and 5 ml of natural waters were extracted using the above DIC extraction procedure. Natural waters included laboratory tap water, stream water, runoff from a wetland and spring water downgradient of a landfill, and water obtained from a landfill leachate collection system. The purpose was to determine the optimum time required for CO2 extraction from a reaction mixture and the extraction efficiency of the procedure. Individual extractions were carried out for 5, 10, 15, and 20 minutes. The results of this time series extraction of the Na2CO3 solution and natural water samples are shown in Table 10.2. It is observed that 100 % of the theoretical CO2 yield from the Na2CO3 solution is obtained from the reaction mixture within five minutes. Additionally, the ~13C of the Na2CO3 solution (-7.5 + 0.1%o) was identical to that of the Na2CO3 solid (-7.5 + 0.1%o, Table 10.1). The natural waters extracted by this technique also yield highly reproducible isotopic results. With a standard deviation of CO2 yields of less than 1 ~mole, at least 98% of the DIC was extracted from the natural samples within five minutes. Except for the stream water with a 613CDIC range of 0.5%0, the rest of the natural waters do not vary by more than 0.2%0 for different extraction time intervals. The lower precision of the 613CDIC for the stream water could be due to natural variability of DIC concentrations in flowing systems. Precision and accuracy of the technique was further verified by extracting a batch of 10 samples of Na2CO3 solution at 50~ for 10 minutes. The results are also presented in Table 10.2. The results show that the extraction process is 99 + 1% efficient, and the ~13C is measurable to better than 0.1%o. Therefore, an extraction time of 5 to 10 minutes at 50~ with constant stirring provides highly reproducible DIC concentrations and 613C with this technique. Modification of the vacutainer technique where water samples are collected in septum tubes without magnetic stir bars and CO2 is removed from the septum tubes and transferred to the vacuum line by ultrasonic agitation also produced precise and accurate results, a total of 10 samples of Na2CO3 Table 10.1 - CO2 yield and h13C from solid Na2CO3 standard. Run # 1 2 3 4 5
Amount Added (m moles CO2)
Experimental Yield% (m moles CO2)
Yield as CO2
613C
167.7 121.7 109.1 86.9 75.7
167.1 119.9 107.4 86.2 76.4
99.8 98.5 98.4 99.2 100.9 99.4 + 1.0
-7.6 -7.6 -7.5 -7.5 -7.4 -7.5 + 0.1
210
Chapter 10 - E.A. Atekwana & R.V. Krishnamurthy
Table 10.2 - Time series extraction of Na2CO3 solution and selected water samples at 50~ Sample
Extraction Time (minutes)
CO2 Yield (/~ moles)
% Yield as CO2 (~ moles) 100.0 99.8 98.5 99.7 99.5 + 0.7
~)13CDIC (%o)
Na2CO3 1 2 3 4
5 10 15 20
101.3 101.1 99.8 101.0 100.8 + 0.7
Tap Water 1 2 3 4
5 10 15 20
25.4 25.4 25.6 25.9 25.6 + 0.2
-10.6 -10.7 -10.5 -10.9 -10.7 + 0.2
Stream Water 1 2 3 4
5 10 15 20
22.3 22.0 21.8 21.9 22.0 + 0.2
-10.8 -11.1 -11.3 -11.0 -11.1 + 0.2
Wetland Runoff 1 2 3 4
5 10 15 20
96.2 95.6 95.2 94.3 95.3 + 0.8
4.8 5.0 5.0 5.0 5.0 + 0.1
Spring 1 2 3 4
5 10 15 20
101.0 101.5 101.3 101.5 101.3 + 0.2
7.7 7.9 7.9 8.0 7.9 + 0.1
Landfill Leachate 1 2 3 4
5 10 15 20
131.7 131.0 131.3 130.6 131.2 + 0.5
3.6 3.8 3.7 3.7 3.7 + 0.1
Na2CO3 (n = 10)
10
98.9 + 1.0
-7.4 -7.5 -7.5 -7.5 -7.5 + 0.1
-7.5 + 0.1
s o l u t i o n e x t r a c t e d u s i n g t h i s m o d i f i c a t i o n w a s 99 + 1% efficient, a n d t h e (~13C w a s m e a s u r a b l e t o b e t t e r t h a n 0.1%o ( D a t a n o t s h o w n ) .
10.3.1.2 Advantages of the vacutainer gas evolution technique compared to other gas evolution technique T h e v a c u t a i n e r g a s e v o l u t i o n t e c h n i q u e of A t e k w a n a & K r i s h n a m u r t h y (1998) d e s c r i b e d a b o v e has a d v a n t a g e s in that the p r o c e d u r e o v e r c o m e s m a j o r p r o b l e m s a s s o c i a t e d w i t h t h e o t h e r g a s e v o l u t i o n t e c h n i q u e s . A l l e v i a t e d p r o b l e m s are d i s c u s s e d below.
Extraction of Dissolved Inorganic Carbon (DIC) in Natural Waters for Isotopic Analyses
211
Sample reaction in C02-free atmosphere Prior to sample collection, the septum tubes are preloaded with magnetic stir bars and phosphoric acid and then evacuated. This not only removes the ambient air introduced during loading but also removes any CO2 from the acid, consequently, preventing sample contamination.
Transfer of aliquot of water sample with representative DIC to the vacuum system during extraction The problem of selecting a suitable container for water collection that allows for complete transfer of evolved CO2 from the sample container (Hassan, 1982) is also solved by this technique. For example, one difficulty with water samples collected via other techniques is in assessing the CO2 partitioned in the headspace relative to the water phase. This is especially true for water with high CO2 content. The septum tubes do not suffer from headspace CO2 partitioning because all the CO2 evolved in the septum tube via the acid reaction is from the parent-sample and is effectively transferred to the laboratory vacuum line in this procedure.
Post sampling alteration of DIC Upon water collection and injection into the prepared septum tube, DIC is immediately converted into CO2 and confined therein. The highly acidic condition of the reaction mixture (pH <2.5) prevents secondary production and/or consumption of DIC components by microbial respiration or photosynthesis thus eliminating the need for bactericide.
Sample storage in septum tubes at room temperature Tap water along with natural water samples were collected, reacted and extracted after storage at room temperature for varying lengths of time. These samples were extracted at 50~ for 25 to 30 minutes without stirring. The storage time ranged from a few hours to 43 days. Results of the storage experiment are presented in Table 10.3. The results show near consistency for both the yield of CO2 and 613CDIC. Variations of up to 6% in CO2 yields could result from the fact that CO2 was extracted from the sample tubes at 50~ without stirring. In addition, natural variability in DIC concentration could contribute to this error. These experiments were carried out with the sole purpose of checking the integrity of the septum tubes over long periods of storage. The data from Table 10.3 shows that septum tubes are suitable sampling devices in which water samples can be reacted and stored for a reasonable length of time prior to DIC extraction. In addition, data from the time series extraction (Table 10.2) and storage experiments (Table 10.3) demonstrate good agreements between field replicates.
10.4 Precipitation Technique The conventional precipitation technique and subsequent modifications involves precipitation of DIC as carbonate by adding a SrC12-NH4OH, BaC12-NaOH or Ba(OH)2 reagent to water samples in the field. The precipitate is subsequently filtered and dried. A portion of the precipitate is reacted with phosphoric acid under closed system conditions, the evolved CO2 purified and its isotopic ratio measured by mass spectrometry (e.g. Gleason et al., 1969; Friedman, 1970; Deines et al., 1974; Barnes et
212
Chapter 10 - E.A. Atekwana & R.V. Krishnamurthy
Table 10.3 CO2 yields -
and 613Cof natural waters after storage at room temperature and different time
intervals. TAP WATER 613C Time CO2 Yield (days) (~moles) (%o) 0.5 3.5 4.2 10.8 43.5
24.53 -10.9 22.44 -11.0 24.86 -11.0 22.88 -10.9 23.32 -10.8 23.6+1.0 -10.9+0.1
STREAM WATER CO2 613C Yield (~moles) (%o) 20.68 21.89 23.21 20.90 19.36 21.1+1.4
WETLAND RUNOFF CO2 613C Yield (/~moles) (%o)
SPRING WATER CO2 613C Yield (tlmoles) (%o)
94.93 7.6 80.19 8.3 -9.4 90.75 7.8 83.82 8.1 -9.3 96.91 7.5 -9.1 84.37 8.3 94.49 7.5 84.26 8.0 -9.3 88.44 7.9 82.17 8.3 -9.3 -9.3+0.1 83.0+1.8 8.2+0.1 93.1+3.4 7.7+0.2
LANDFILL LEACHATE c02 613c Yield (~moles) (%o) 140.58 137.94 145.97 132.55 136.95 138.8+4.9
5.3 5.6 5.5 5.3 5.5 5.4+0.1
al., 1978; Pearson et al., 1978; Hassan, 1982; Heathcote, 1985; Bishop, 1990; Aravena et al., 1992; Kusakabe, 2001). The precipitation procedure based on SrC12-NH4OH is most commonly used for this technique. Two procedures, one based on the techniques of Gleason et al. (1969); Barnes et al. (1978), Heathcote (1985), Hassan, (1982) and Bishop (1990) and another based on a modification by Kusakabe, (2001) are described below. The modification of the precipitation technique by Kusakabe (2001) uses saturated Ba(OH)2 as the precipitation reagent, and water sampling, carbonate precipitation, and washing of the precipitate is done in plastic syringes. To distinguish the Kusakabe (2001) technique from the traditional precipitation technique, we call it the syringe precipitation technique.
10.4.1 Pocedure 10.4.1.1 Preparation of precipitation reagents Traditional precipitation technique: SrCI2-NH40H (ammoniacal strontium chloride or amm. SrCI2) reagent
List of supplies~equipment needed 9 9 9 9
SrC12 NH4OH Glass or polyethylene bottles PVC (polyvinylchloride) tape
The amm. SrC12 is prepared by reacting SrC12 (s) with NH4OH (1). Two preparation methods reported in the literature are as follows" 1) mix 2,366 ml of NH4OH (approx. 30% NH3) and 453.6 g of SrC12.6H20 (Hassan, 1982) 2) mix 200 ml of NH4OH (approx. 25% NH3) and 180 g of SrC12.6H20 (Barnes et al., 1978; Heathcote, 1985; Bishop, 1990) The amm. SrC12 solution is stored in glass or polyethylene bottles, sealed with PVC tape to prevent any subsequent CO2 absorption. Precaution" Preparation of amm. SrCI2 requires a C02-free atmosphere.
Extraction of Dissolved Inorganic Carbon (DIC) in Natural Waters for Isotopic Analyses
213
Syringe precipitation technique: Ba(OH)2 reagent List of supplies~equipment needed 9 9 9 9 9
Anhydrous Ba(OH)2 2L glass bottles 50 ml disposable syringes and syringe caps (details of the cap not provided) Ascarite II (CO2 absorber) Aluminum bags (details of the type and source of bags not provided)
A large batch of saturated Ba(OH)2 solution is prepared using CO2-free dionised water and stored in 2 L glass bottles. If more than one 2 L bottle us used, there are connected in a train, and the ascarite II is attached to the air intake of the first bottle. A 50 ml syringe is attached via a plastic tube to the train of 2 L bottles containing the saturated Ba(OH)2 solution. A 20 ml aliquot of the solution is slowly withdrawn into the syringe, which is immediately capped tightly and stored in a sealed bag made of aluminum foil. The saturated Ba(OH)2 solution store in this manner for 1 year did not show any visible change in the solution, suggesting that atmospheric CO2 did not permeate the syringe (Kusakabe, 2001). Precaution: Although not stated in the original publication, preparation of the saturated
Ba(OH)2 solution requires a C02-free atmosphere.
10.4.1.2 Water sampling and carbonate precipitation Traditional precipitation technique List of supplies~equipment needed 9 atom. SrC12 reagent 9 Glass or polyethylene bottles 9 PVC (polyvinylchloride) tape The water to be sampled is introduced into the sample bottle along with amm. SrC12. The ratio of water to amm. SrC12 varies. A ratio of 1:1, 1:19 and 1:9 have been used by Gleason et al., (1969), Barnes et al., (1978) and Bishop (1990) respectively. Generally, the sample container is filled with the sampled water and the appropriate amount of amm. SrC12 is added to the bottom. The sample container is immediately capped, sealed with PVC tape, well shaken to ensure complete mixing and allowed to react for 48 hours for precipitation of DIC as SrCO3.
Precaution:This step requires minimal exposure of the amm. SrCI2 reagent and reagent water sample mixture to atmospheric C02.
Syringe precipitation technique List of supplies~equipment needed 9 50 ml plastic syringes containing 20 ml of saturated Ba(OH)2 9 PVC (polyvinylchloride) tape During water sampling, the cap on the syringe is removed under the water to be sampled. Between 20-30 ml of water is withdrawn into the syringe, and the syringe is re-capped (presumably under water). The cap of the syringe is taped and syringe with Ba(OH)2-water mixture is well shaken and stored in the aluminum bag until precipi-
214
Chapter 10- E.A. Atekwana& R.V. Krishnamurthy
tate recovery in the laboratory.
10.4.1.3 Precipitate recovery List of supplies~equipment needed 9 Filtration system (vacuum filtration system) 9 Glass fiber filter paper 9 Drying system (oven or glove box) 9 Sample storage containers (20-ml glass vials) 9 Deionised water 9 50 ml plastic syringe1 9 3-way plastic stopcock1 Filtration In the traditional precipitation technique, a filtration system is assembled using glass fiber filter paper. Precipitated samples are vigorously shaken and then poured through the filter system and filtered for a period of 10 minutes. The collected precipitate is next rinsed with deionised water and filtered for an additional 5 minutes. Precaution: Filtration of reacted samples in air for which all the amm. SrCI2 reagent was not consumed during precipitation could result in additional C02 adsorption. Filtering can be done in a C02-free atmosphere in a glove box. In the syringe precipitation technique, the syringe with the BaCO3 precipitate is held upright. The cap is removed and the syringe connected to a 3-way plastic stopcock (Figure 10.3). Another syringe with CO2-free deionised water is connected to the syringe with the precipitate by way of the stopcock. Most of the mixture of water and excess Ba(OH)2 solution is pushed out of the syringe, ensuring that the precipitate is not lost in this process. Deionised water is pushed into the syringe containing the precipitate. The syringe from which the deionised water was dispensed is disconnected. The syringe with the precipitate and deionised water is immediately capped and shaken vigorously to wash the BaCO3 precipitate. The syringe is left standing upright to allow the precipitate to settle to the bottom of the syringe. After settling, the most of water used for washing the precipitate is removed as previously described. The process of washing the precipitate is repeated 5-6 times until the washing solution is no more alkaline. This procedure ensures that the precipitate 1. Used only in the syringe precipitation technique
F i g u r e 1 0 . 3 - Schematic illustration of the procedure to wash alkaline BaCO3 precipitated in a syringe without exposure to air (Modified from Kusakabe, 2001).
Extraction of Dissolved Inorganic Carbon (DIC) in Natural Waters for Isotopic Analyses
215
is washed clean of the Ba(OH)2 with making contact without atmospheric CO2. The precipitate is recovered by filtering using a 0.45 millipore filter.
Drying Collected precipitate is dried on the filter paper in an oven at 110~ for 4 hr. Alternatively, one may dry the precipitate in a glove box under a nitrogen atmosphere if additional CO2 adsorption is a concern.
Precipitate storage The precipitate is scraped off the filter paper and ground to homogenize. The powdered precipitate is then stored in airtight glass vials and kept in a desiccator until analysis. Care should be taken to ensure that all the precipitate is scraped off the filter paper for storage. Precaution: Store precipitate under airtight conditions to prevent subsequent C02 exchange.
10.4.1.4 C02 evolution from SrC03 precipitate CO2 is evolved from the SrCO3 or BaCO3 precipitate by reacting with acid. Two
methods (McCrea, 1950 and Krishnamurthy et al., 1997) can be used. Another method not presented here that can be used to evolve CO2 from carbonate precipitates is described by Kusakabe (2001). Although these methods were designed to react carbonates for simultaneous determination of the (~13Cand 6180 of the evolved CO2, they can be used for evolving CO2 from SrCO3 for carbon isotopic analysis. In the above procedures, the carbonates are reacted under vacuum and at constant temperature using 100% phosphoric acid. Since only 613C determination of the carbonate is required, the reaction need not be conducted at constant temperature or with 100% phosphoric acid.
List of supplies~equipment needed 9 100 % phosphoric acid (prepared by removal water from 85% phosphoric acid using P205 or by vacuum evaporation). Note that 100% phosphoric acid is not essential if the 6180 is not to be determined on the CO2. 9 Constant temperature water bath 9 Liquid nitrogen 9 -80~ ice slush 9 Reaction vessels (varies depending on the reaction technique used) For the method of McCrea, (1950) 9 Reaction vessels with a side arm are used For the method of Krishnamurthy et al. (1997) 9 16 x 100-mm glass septum tubes (VACUTAINER| Serum Tubes, Becton Dickson & Company, Franklin Lakes, NJ 07417). 9 Glass boats (prepared by flame sealing one end of 5-mm lengths of 9-mm diameter Pyrex tubes. 9 Extraction needle (constructed by attaching a 26 gauge needle (~12 mm long) affixed to an inner Luer ground joint and joined to a piece of 9 mm Pyrex tube
216
Chapter 10 - E.A. Atekwana & R.V. Krishnamurthy
9 Glue (Elmer's Wonderbond plus Super glue, Borden Inc. HHPG Columbus OH. 43215) 9 Septum holding rack 9 Laboratory vacuum extraction system The principle for carbonate reaction and CO2 purification are similar in both the McCrea (1950) and the Krishnamurthy et al. (1997) procedures. The main difference lies in the reaction vessels.
The McCrea (1950) technique: The vessel is loaded with a known amount of precipitate at the bottom of the vessel and phosphoric acid is dispensed into the side arm (Figure 10.4, insert A). The vessel is evacuated of air by attaching it to the vacuum system by way of a coupling. Complete evacuation of air is monitored by vacuum gauge. After evacuation, the vessel is placed in a constant temperature water bath set at 25 ~ or 50~ for about 20 minutes to achieve isothermal equilibrium. The vessel is tilted under water so as to allow the acid to flow out of the side arm and react with the carbonate. The reaction is allowed to continue while in the bath for an additional 4 or 2 hours for reactions at 25 ~ or 50~ respectively, before CO2 extraction and purification in the vacuum system.
The Krishnamurthy et al. (1997) technique: For this technique, glass boats and septurn tubes are used (Figure 10.4, insert B). The septa of the tubes are removed and a known amount precipitate is carefully loaded at the bottom of the tube using wax paper funnels. The glass boats are carefully glued inside the vacutainer tubes about 5 mm below the septum. One ml of phosphoric acid is poured into the boat and the septa replaced. Air is removed from the septum tubes using an extraction needle attached to the vacuum system by way of a Cajon union (Figure 10.4, insert B). After evacuation, the septum tubes are placed in a holder, and placed in a constant temperature water bath set at 25 ~ or 50~ for 15 minutes to achieve isothermal equilibrium. The tubes are tilted under water to allow the acid to flow out of the boat into the bottom of the tube and react with the carbonate. The reaction is allowed to continue for an additional 4 or 2 hours for reactions at 25 ~ or 50~ respectively, before CO2 extraction and purification in the vacuum system.
CO2 extraction and purification: For carbonate reaction using the reaction vessel with a side arm, the vessel is attached to the vacuum system by a Cajon union (Figure 10.4). The portion between the vacuum system stopcock and that of the reaction vessel is evacuated. The CO2 is released into the vacuum system and trapped using liquid nitrogen, and the residual non-condensable gases are pumped away. The CO2 is purified cryogenically by replacing the liquid nitrogen with a -80~ ice slush. The CO2 is transferred to a calibrated cold finger where the yield is measured manometrically. The CO2 gas is then collected and introduced into the mass spectrometer for isotopic ratio measurement.
Extraction of Dissolved Inorganic Carbon (DIC) in Natural Waters for Isotopic Analyses
217
For carbonate reactions using septum tubes, the extraction needle is attached to the vacuum system by way of a Cajon union (Figure 10.4). The septum part of the tube is inserted into the needle deep enough to cover the needle opening in order to pump away the air between the extraction needle and the vacuum system stopcock. After evacuation, the needle is pushed deep enough to puncture the septum releasing the CO2 into the vacuum system. The CO2 is purified and its yield measured in the same way as described above.
10.4.1.5 Extraction Efficiency and Precision of Precipitation Technique The precipitation technique produces highly reproducible 613CDIC. The accuracy of the technique is heavily dependent on the care taken to prevent the precipitating reagent and the precipitate from reacting with atmospheric CO2. Experiments to determine the extraction efficiency and precision of the traditional precipitation technique have been conducted by Hassan (1982) and Bishop (1990) and the syringe precipitation technique by Kasukabe (2001). Hassan (1982) prepared SrC12-NH4OH
Figure 10.4 - Simplified schematic of reaction vessels used for acid reaction of carbonate precipitates for (insert A) the McCrea (1950) and (insert B) the Krishnamurthy et al. (1997) techniques. A simplified schematic of a laboratory vacuum extraction line is shown in Figure 10.2.
218
Chapter 10 - E.A. Atekwana & R.V. Krishnamurthy Table 10.4 - Accuracy and precision of 613CDIc for 0.01 molar NaHCO3 solution (613C =-15.33 + 0.07%o) extracted by the precipitation technique (modified from Hassan, 1982). Sulfate Moles /liter 0.0 0.001 0.005 0.01 0.1
Effinciency of extraction (%) 95 + 3 95 + 3 90 + 5 90 + 2 70 + 3
~13CDIC (%o) -15.36 -15.38 -15.26 -15.25 -14.91
+ 0.05 + 0.04 + 0.02 + 0.02 + 0.03
*A13CDIC (%o) 0.0 -0.02 +0.10 +0.21 +0.45
+ 0.06 + 0.05 + 0.05 + 0.06
* A13CDIC = 613Cmeasured (MSO4) - ~13Ctrue (zero SO4). (Where MSO4 is the molarity of $04).
reagent and precipitated carbonates in a 0.1-molar NaHCO3 solution of known isotopic composition (613C - -15.33 + 0.07%o). In these experiments, the sulfate concentration was varied. The precipitate was filtered and dried in a nitrogen atmosphere in a glove box. A representative sample of the precipitate was reacted with 100% phosphoric acid to evolve CO2. The results of ~13C analysis are presented in Table 10.4. The efficiency of extraction of DIC was 95% except for solutions with greater than 0.005 molar sulfate content, which was 90% or less. The ~13CDIC w a s different from the solid by up to 0.1%o for solutions containing 0.001 molar sulfate or less. Bishop (1990) used a 0.01-molar solution of BDK| KHCO3 (established ~13C -- -31.6 + 0.2%o from 11 analyses of the solid) in deionised water to precipitate carbonates. The precipitate was filtered in air, and oven dried at 110~ for 4 hours or filtered and dried in a CO2-free atmosphere. For isotopic analysis, 20 mg of the precipitate was reacted with 100 % phosphoric acid to evolve CO2. Results for 6 analyses of precipitates filtered and dried in air and 3 analyses of precipitates filtered and dried in a CO2-free atmosphere in a glove box are shown in Table 10.5. The results of these experiments show that the precipitates filtered in air and oven dried had on average 613C values 1%o more positive than that of the solid standard. Results similar to the expected isotopic values of the standard were obtained when the precipitates were filtered and dried in a CO2-free atmosphere of nitrogen. The ~13CDIC of t h e S r C O 3 precipitates show that the technique produces Table 10.5 - Accuracy and precision of 613CDIC for 0.01 molar KHCO3 solution precise results. (613C =-31.6 + 0.2%o) extracted by the precipitation technique (Modified from However, the Bishop, 1990). accuracy of Precipitates prepared in CO2-free media 613CDIC is afPrecipitates prepared in air Sample reference NO. i~13CDIC(%o) Sample reference NO. 613CDIC(%o) fected by the 1 -30.7 7 -31.0 treatment of 2 -30.6 8 -32.2 the precipitate 3 -30.6 9 -31.7 (Gleason et al., 4 -30.6 Mean -31.6 + 0.6 5 -30.2 1969; Hassan, 6 -30.2 1982; Bishop, Mean -30.6 + 0.2 1990).
219
Extraction of Dissolved Inorganic Carbon (DIC) in Natural Waters for Isotopic Analyses
Experiments to determine the reproducibility and accuracy of the syringe precipitation technique were conducted by Kusakabe (2001). In one set of experiments, analytical grade NaHCO3 was used to prepared standard solutions with concentrations that varied from 2.5 to 20 mmole/kg in CO2-free deionised water (Solutions A-l, A-2, A-3, A-4, A-5; Table 10.6). In addition, to test the effect of variable sulfate content in natural waters on the accuracy of the ~13C of the DIC precipitated as carbonate, another set of N20 mmole/kg NaHCO3 solutions were prepared in which varying amounts of analytical grade Na2SO4 were added to make solutions with different H C O 3 - / S O 4 2 - ratios (Solution B, C, D, E, F, and G; Table 10.6). The ~13CDIC of the solutions with varying concentration of NaHCO3 (Solutions A-1 through A-5) gave an averaged value of-11.65 + 0.03%o (n=18). This suggests that the technique gave highly reproducible results and that varying DIC concentrations did not affect the reproducibility of the ~13CDIC. Since the ~13C of the solid NaHCO3 was not measured or reported, it is not possible to determine the accuracy of the technique. 613CDICof the test NaHCO3 solutions where the HCO3-mole fraction (HCO3-/(HCO3- + SO42-)) was greater than 0.34 (Solution B, C and D; Table 10.6) gave similar results to the pure NaHCO3 standard solutions (Solutions A-1 through A-5; Table 10.6). Below HCO3mole fraction of 0.34, the yield of CO2 from the precipitate decrease, the ~13CDIC increased and the standard deviation of the 613C of the precipitate increased as the HCO3-mole fraction decreased (Solutions B through H; Table 10.6)(Kusakabe, 2001). Table 10.6 - Reproducibility of hl3CDIc for carbonates precipitated from NaHCO3 solutions of different concentrations and the effect of variable sulfate content on the h13CDIC of carbonate precipitates (Modified from Kusakabe, 2001). Sample
NaHCO3 (mmol/kg)
Na2SO4 (mmol/kg)
HCO3Fraction*
~13C (%o)
Std Dev. lo
Number of samples (n)
Solution Solution Solution Solution Solution Average
A-1 A-2 A-3 A-4 A-5 of solution A
20.49 15.12 10.09 5.13 2.51
0.0 0.0 0.0 0.0 0.0
1.00 1.00 1.00 1.00 1.00
-11.67 -11.64 -11.64 -11.67 -11.64 -11.65
0.03 0.03 0.02 0.04 0.02 0.03
6 3 3 3 3
Solution Solution Solution Solution Solution Solution Solution
B C D E F G H
19.94 20.55 21.44 20.52 20.18 21.17 21.00
1.89 20.84 40.71 80.41 118.35 159.85 204.45
0.91 0.51 0.34 0.20 0.15 0.12 0.09
-11.71 -11.70 -11.66 -11.28 -10.39 -9.32 -9.78
0.03 0.03 0.01 0.13 0.18 0.19 0.25
4 4 3 5 3 3 4
*HCO3- / (HCO3- + SO42-)
220
Chapter 10 - E.A. Atekwana& R.V. Krishnamurthy
10.4.1.6 Effect on the accuracy of ~313CDICdue to atmospheric C02 contamination of reagent and precipitate Reagent storage: Bishop (1990) exposed SrC12-NH4OH reagent by passing air into the solution using a peristaltic pump and periodically exposing another bottle half-full with the reagent to air. The 613C of the precipitate generated from the two experiments were -26.7 and -31.7%o respectively. Another bottle filled with the reagent and well sealed did not generate any carbonate. These experiments show that the accuracy of the ~13C of the precipitate can be affected by exposure of the reagent solution to atmospheric or other extraneous CO2 sources.
Storage of precipitate prior to analysis: Carbonate samples may exchange carbon with atmospheric CO2 if stored for long periods (Mook, 1968; Hassan, 1982). This is especially true if the carbonate is moist. The g13C of carbonate precipitates stored in a desiccator and a single sample stored open to the atmosphere for 1 year showed shifts of + 0.5 and + 1.5%o respectively (Bishop, 1990).
10.5 Precipitation and gas evolution technique In the conventional precipitation technique described above, the DIC concentration is not determined. However, a modification of the conventional precipitation technique, which we call the precipitation and gas evolution technique, allows for the simultaneous determinations of DIC concentration and ~)13CDICfrom the same water sample. In the modified procedure, the DIC is precipitated as carbonate in the field and subsequently reacted with acid to evolve CO2 without recovering the precipitate, thus avoiding the steps of filtration, drying and storage of the precipitate in airtight containers as in the conventional precipitation technique. Two procedural modifications (Taylor & Fox, 1996; Nakamura et al., 1998) of the conventional precipitation technique can be used if the DIC concentration is to be determined.
10.5.1 The Taylor & Fox (1996) technique In this technique, DIC in water samples are precipitated in the field as in the conventional precipitation technique. In the laboratory, the precipitated carbonate is reacted with phosphoric acid and the evolved CO2 gas is collected to determine the DIC concentration and ~)13CDIC. Estimate of DIC concentrations by this technique require the measurement of reagent, sample and gas weights and is detailed below. 10.5.1.1 Procedure Preparation of SrCl2 - NaOH reagent List of supplies~equipment needed 9 SrC12 (Analar grade) 9 NaOH (Analar grade) 9 Pyrex bottles (50-ml) 9 Analytical balance 9 pH meter 9 Deionised water
Extraction of DissolvedInorganicCarbon (DIC)in NaturalWaters for IsotopicAnalyses
221
The SrC12-NaOH reagent is prepared by reacting SrC12 (s) with NaOH (1) in deionised water. Enough NaOH is added to the mixture to adjust the pH to > 10. The reaction is carried out in a CO2-free atmosphere. The mixture is stored in 50-ml Pyrex bottles, sealed tight and weighed.
Water sampling
List of supplies~equipment needed 9Pre-weighed 50-ml pyrex bottles containing the SrC12-NaOH reagent ~ Clean pre-weighed pyrex bottles (1 L) In the field, the pre-weighed 1-L sample bottle is partially filled with water. The SrC12-NaOH reagent is added from the 50-ml bottle avoiding any spillage and leaving the last few ml at the bottom containing specks of precipitate. The 1-L sample bottle is quickly topped with sample water and both the sample and the solution bottles are tightly closed and re-weighed upon arrival at the laboratory. This allows for precise determination of the weight of the water which is used to estimate the DIC content.
Conversion of precipitate to C02
List of supplies~equipment needed 9 9 9 9 9 9 9 9 9
Orthophosphoric acid (85% phosphoric acid) Large vacuum jar CO2-free air Magnetic stir bar Magnetic stirrer -70~ dry ice/alcohol trap Analytical balance Small stainless steel bottles Laboratory vacuum extraction system
To release the carbonate precipitate as CO2, the bottle containing the precipitated sample is inserted snugly into a large vacuum jar attached to the vacuum line and flushed with CO2-free air. To ensure that the precipitation process is complete, an aliquot of the reacted solution (water sample + SrC12-NaOH) is withdrawn and reacted with acid to determine if any CO2 evolves. Next, the vacuum jar containing the sample bottle is opened to the vacuum extraction system and evacuated of air. Excess phosphoric acid is introduced into the sample bottle by a syringe through a rubber septum, presumably on the large vacuum jar. The solution is stirred continuously with a magnetic stir bar or agitated by ultra sound and the released CO2 is purified by passage through a double dry ice/alcohol trap at or below -70~ and trapped in liquid nitrogen. Extraction of the CO2 takes about 2 hours for complete recovery of the initial vacuum conditions. The CO2 in this procedure is transferred from the vacuum line and frozen in small pre-weighed stainless steel bottles, which are later weighed on an analytical balance to determine the released quantity of gas. The CO2 weights are converted to DIC content and reported as mmol/kg of water. Aliquots of the CO2 are introduced into a mass spectrometer for isotope ratio measurements.
222
Chapter 10 - E.A. Atekwana & R.V. Krishnamurthy
10.5.1.2 Extraction efficiency and precision of technique Duplicate water samples collected in the field was used to demonstrate the accuracy and precision of this technique (Taylor & Fox, 1996). Results of these analyses are shown in Table 10.7. The precision of this technique to determine the DIC concentration is + 0.013 mmol/kg, taking into account the variance due to weighing errors of the empty pyrex bottles and stainless steel containers. The precision of the 613CDIC for this technique was 0.1%o based on two duplicate analyses of natural water samples. Due to the fact that a solution of known DIC concentration and 613C was not prepared using this technique, the accuracy in DIC concentration and ~)13CDIC is not known. Since replicates of natural water samples show an overall difference in DIC concentration and (~13CDIC of + 0.013 m m o l / k g and <0.14%o respectively; DIC extraction by this procedure is reproducible and probably accurate. 10.5.2 The Nakamura et al. (1998) technique In this technique, DIC in water samples are precipitated in the field as in the conventional precipitation technique. In the laboratory, the precipitated carbonate is reacted with phosphoric acid and the evolved CO2 gas is collected to determine the DIC concentration. The procedure as discussed by Nakamura et al. (1998) is detailed below. 10.5.2.1 Procedure In this procedure as described by Nakamura et al. (1998), the method of preparation of amm. SrC12 reagent is not presented but is presumably similar to those described in the conventional precipitation technique. The discussion thus focuses on sample collection, DIC precipitation, CO2 release from the precipitate and DIC concentration determination.
Water sampling and carbonate precipitation List of supplies~equipment needed 9 ~ 9 9
amm. SrC12 reagent 300 ml Erlenmeyer flasks Apiezon grease Adhesive tape
In the field, 300 ml of Table 10.7- Reproducibility of DIC concentrations and ~)13CDICof water is transferred into duplicate field analyses (From Taylor and Fox, 1996). Erlenmeyer flasks to which 613CDICpair (%o) DIC 10 ml of amm. SrC12 rea- Sample (mmol / kg) (Values mean) / o gent is added to precipitate the DIC as SrCO3. The CCB12 0.950 0.900 1.9 flasks are sealed tightly CCB13 0.606 0.625 0.7 with their covers by using CCB14 0.392 0.396 0.3 0.6 0.566 0.581 CCB15 Apiezon grease and adhe0.3 0.605 0.612 CCB16 sive tape. The reacted -12.54 -12.54 0.5 0.584 0.572 CCB17 water samples are trans-6.59 -6.45 1.0 CCB18 0.449 0.423 ported to the laboratory
Extraction of Dissolved Inorganic Carbon (DIC) in Natural Waters for Isotopic Analyses
223
where the precipitate is converted into CO2 for DIC determination and mass spectrometric analysis.
Conversion of precipitate to C02 List of supplies~equipment needed 9 Orthophosphoric acid (85% phosphoric acid) 9 Glove box 9 Ascarite-II, sodium hydroxide on a nonfibrous silicate carrier (Thomas Scientific, USA) 9 Magnetic stirrer 9 -100~ liquid nitrogen/methanol mixture 9 Laboratory vacuum extraction system In the laboratory, the DIC-free portion of the sample is decanted in a CO2-free atmosphere in a glove box. The CO2-free atmosphere is prepared by exposing the air in the glove box to alkali materials (ca. 300 cc of Ascarite-II, sodium hydroxide on a nonfibrous silicate carrier) for several hours. To release the carbonate precipitate as CO2, a 10-ml container with 5 ml of phosphoric acid is placed at the flat bottom of the flask. The flask is fitted with a vacuum stopcock, sealed and evacuated. The SrCO3 is decomposed to CO2 by toppling over the 10-ml container with phosphoric acid, bring the acid in contact with the precipitate and left overnight to ensure complete decomposition of the SrCO3. On the next day, the flask is connected to the vacuum extraction system and the released CO2 is purified by passage through a liquid nitrogen/methanol mixture trap at-100~ and trapped in liquid nitrogen (-180~ Residual non-condensable gases are pumped away and the trapped CO2 is then released in calibrated volume and its yield measured manometrically. DIC concentrations are obtained from the CO2 yield and the initial amount of water collected for DIC extraction. Aliquots of the CO2 are collected and introduced into a mass spectrometer for isotope ratio measurements.
10.5.2.2 Extraction efficiency and precision of technique Duplicate water samples collected in the field were used to demonstrate the accuracy and precision of this technique (Nakamura et al. 1998). The DIC concentration and ~13CD~C for the duplicate runs were + 0.01 mmol/kg and 0.1%o respectively. Although a standard solution of known concentration and ~13CDIC w a s not extracted by this technique (or reported in this study), the field replicates suggest that the reproducibility of the DIC concentration is good and the accuracy of g13CDIC is not affected by the procedure. 10.6 The vapor phase equilibration technique
In this technique, a known volume of water is outgassed in a pre-evacuated known volume within a sample container. The outgassed CO2 can be collected and purified in a vacuum line or introduced into a gas chromatograph to separate the CO2. The CO2 c o n c e n t r a t i o n (CO2(aq) + H2CO3 0) is determined from the collected gas and its isotopic ratio is measured. Because only an aliquot of the evolved gas is col-
224
Chapter 10 - E.A. Atekwana & R.V. Krishnamurthy
lected for mass spectrometer measurement, the stable carbon ratio and DIC concentration in the water sample is obtained by mass balance. The technique is based on equilibrium thermodynamics and requires that the outgassed-CO2 only comes from the dissolved CO2 and that aH2CO3 - MH2CO3 (Hassan, 1982). Although vapor phase equilibration is routinely used in estimates of DIC concentration in natural waters (e.g. Pearson et al., 1978), measurement of the isotopic ratio was proposed by Hassan (1982) and measured by gas chromatograph/combustion furnace/isotope ratio mass spectrometer (GC/C/IRMS) by Miyajima et al. (1995). The vapor phase equilibration technique described below is based on that of Miyajima et al. (1995). In this procedure, water samples are collected in airtight containers of known inner volume. Subsequently, a headspace of known volume is created in each bottle and the sample acidified. The liberated CO2 is allowed to equilibrate with the headspace gas at a known temperature. A sub-sample of the headspace is introduced into the GC/C/IRMS and the isotopic composition and concentration of the CO2 determined. The isotopic ratio of the remaining liquid phase is calculated using the temperature dependent isotope discrimination between the gas and aqueous phase. The X C O 2 (DIC) of the water sample is obtained by mass balance.
10.6.1 Procedure 10.6.1.1 Water sample collection
List of supplies~equipment needed
9 Glass serum bottles (Nichiden-Rika Glass Co., type V-50; inner volume, 68.6 + 0.5 ml) fitted with a butyl rubber septum 9 Aluminum seals 9 Crimping tool 9 HgC12 9 Paraffin wax 9 Calibrated gas tight syringes
In the field, water samples are collected in 68.6-ml glass serum bottles. The bottles are immediately closed with the butyl rubber septum and sealed with aluminum seals using a crimping tool. Precautions are taken to ensure that there are no gas bubbles in the sample containers. For lengthy storage of samples prior to analysis, 0.20 ml of water from each sample container is withdrawn and replaced with a similar volume of saturated HgC12 using gas tight syringes. As a precaution to prevent atmospheric CO2 invasion following HgC12 introduction, the butyl rubber septum's surface is covered with paraffin wax. The collected water samples are transported to the laboratory for further analysis.
10.6.1.2 Vapor phase equilibration List of supplies~equipment needed 9 Glass serum bottles containing water samples ~ Ultrapure He gas (nominally >99.9999%) 9 10 ml gas tight syringes (Dynatech Precision Sampling Corp.) 9 CO2-free 6N HC1 (prepared by bubbling with ultrapure He gas for at least 3 h)
Extraction of Dissolved Inorganic
Carbon
(DIC) in Natural Waters for Isotopic Analyses
225
The vapor phase equilibration is carried out at constant room temperature (23~ A headspace of known volume (5 ml) is created in each sample container by injecting ultra-pure He using a gas tight syringe. A similar volume of water is withdrawn with another syringe (Figure 10.5). Another plastic syringe is used to acidify the water sample by injecting 0.5 ml of CO2-free HC1 solution (6N) through the septum. The samples are shaken vigorously by hand and left upside down and in the dark for at least 40 hours to equilibrate. After equilibration a known amount of headspace gas is withdrawn with a 1-ml gas tight syringe and injected into the G C / C / I R M S for CO2 concentration and isotopic ratio determination. 10.6.1.3 Determination of DIC concentration and c513CDIC Before outlining the procedure for determining DIC concentrations, a simplified description of the GC/C~ IRMS would be appropriate. The G C / C / I R M S is a gas chromatograph interfaced with an isotope ratio mass spectrometer by a combustion furnace. The GC is used to separate the constituents in an injected gas mixture in a packed column using a carrier gas (usually He). The separated constituents are passed through the combustion furnace. The combustion furnace, which operates at temperatures of 800 ~ - 900~ is used for combusting organic carbon constituents. However, in DIC analysis, it is maintained at the proper operating temperature to stabilize analytical conditions even though no combustion is needed. The separated CO2 is then passed into the IRMS where isotopic ratios are determined by comparison to a reference CO2.
Equilibrating CO2 between the headspace and the liquid phase in glass serum bottles containing acidified water sample with headspace (A). Withdrawing a portion of the headspace gas with a gas-tight syringe (B). (Modified from Miyajima et al., 1995). F i g u r e 10.5 -
226
Chapter 10 - E.A. Atekwana & R.V. Krishnamurthy
The output from the GC/C/IRMS consists of the calculated (~13C, along with some chromatographic parameters such as retention times, peak width, and peak area. The ion current signal for M/e (molecular weight per electronic valence) of 44 (12C1602+) of the mass spectrometer is used to estimate the CO2 concentration. As the fraction of 12C1602+ exceeds 98% in natural CO2 samples, the peak area for M/e is closely related to the amount of CO2 injected into the GC/C/IRMS (Miyajima et al., 1995). This allows the concentration of aliquots of CO2 analyzed from the headspace gas to be determined from the peak area for M/e - 44. The ~)13Cvalue of XCO2 (DIC) for the sample is calculated from aliquots of the headspace gas injected into the GC/C/IRMS using the following equation (Miyajima et al., 1995)" a
~)13C(YCO2) - [V h x n + (V b - V h ) x ~ x (n + ~g)] + [V h + (V b - V h ) • ~]
[10.3]
w h e r e Vh is the volume of the headspace, Vb is the inner volume of the serum bottle, n is the ~)13Cvalue of the headspace CO2, ~ is the Ostwalt solubility coefficient and ~ga is the isotopic difference between the gas and aqueous CO2. Both 13and ega are tempera-
ture dependent.
10.6.1.4 Precision and accuracy of the technique The precision and accuracy of determining both the concentration and ~)13CDIC using the vapor phase equilibration method depends on the assumptions used in the headspace equilibration and the analytical precision of isotopic analysis. Equilibration between the CO2 in the headspace and aqueous phases is assumed which can differ, however, depending on the nature of water sample. Factors that affect this equilibrium in water, such as salinity, ionic strength, and concentration of humic substances, are usually not measured or are difficult to determine. This could potentially lead to error in DIC analyses. No experiments have been conducted to determine how these parameters affect the precision or accuracy of the DIC analysis possibly restricting this technique to dilute waters where these effects are considered minimal (Miyajima et al., 1995). Optimization of analytical conditions and correction for isotopic discrimination of headspace CO2 injected into the GC/C/IRMS, as well as, the CO2 sample size affect precision and accuracy of DIC concentration and g13CD~C determination using this technique. A detailed discussion related to these and other parameters in this technique is found in Miyajima et al., (1995). 10. 7 General considerations and conclusions
Widespread use of DIC in research has been hampered by the technical difficulties associated with the DIC extraction and mass spectrometric procedures. In terms of the extraction procedure, ease of sampling and reliability/reproducibility of the DIC extraction procedure is paramount. The salient points of the main DIC extraction procedures are summarized in Table 10.8. The vacutainer gas evolution technique of Atekwana & Krishnamurthy (1998) is simple, versatile, requires less technical sophistication and provides accurate DIC concentration and 613CDIC. Compared to the pre-
Table 10.7 -Summary comparison of the methods of DIC extraction from natural waters Gas Evolution
Sample volume collected
Precipitation
5 or 10 mL
Precipitation and Gas Evolution
20 m L - 1 L
Ca
Vapor Phase Equilibration
300 ml - 1 L
o 9
68 mL
9 f0r...r0utine...D!C...~y.se.s .< Water sample preservaNot required for the method Not required Not required Saturated HgC12 for tion of Atekwana and long term storage ....................................................................................K r ! s ~ a m u r t h y ( ! 9 9 8 ) O Simultaneous determination Routinely determined Not routinely determined Routinely determined Routinely determined of DIC and 6 1 3 C D I C Ca ........................................................................................................................................................................................................................... Extraction of DIC from small Possible, limited by amount Difficult to process on a Difficult to process on a Posslbiewi~:hmodificafions) ....... r sample volume <1 ml - 5 mL of CO2 required for isotopic routine basis routine basis but difficult to process on a o routineb., asis (e.-g- from S.g.d.imentcorgs)...............ana!ysis .. Processing water with very Collection of water in several Precipitation of large volume Precipitation of large volume Equilibrate larger volume of low DIC content tubes and mixing CO2 in the of water of water water r laboratory Effect of sulfate concentration Not determined Lower DIC yields and posiNot determined Not determined on the DIC yield and & 1 3 C D I C ........................................................................................................................................................................................................................... tiye...shi.ft..in...613C.!)!.c........................................................................................................................................................................................................................................... . Ease of water sample collecEasy due to sample container Could be difficult due to Could be difficult due to Easy due to sample tion from remote locations type and sample volume sample volume collected and sample volume collected and containers type and sample collected. Sampling requires field procedures. Requires field procedures. Requires volume collected. Requires limited training competence in procedure to competence in procedure to competence in procedure to prevent reagent and sample prevent reagent and sample eliminate trapped air in contamination by CO2 contamination by CO2 sample container Shipping of sampling equipEasy p~{en{~a~y`~..d{f`~cu~{~.`due...{~`[hei~-~-`~p~{en{ia~y.``.di~f~cu~{~--due---{~...{he ....................... Easy o ment to and shipping of colamount of equipment and amount of equipment and 9 lected samples from remote weight considerations weight considerations Ca locations by commercial car> riers Shipping of extracted CO2 to Possible Possible Possible Not routine. Possible with other laboratories for analysis significant procedural modifications Time required for routine sam- Hours Two or more days Two or more days Two or more days pie collection to isotopic measurement of extractedCO2 Allows for 14C analysis Possible Possible Possible Not routine. Possible with of DIC significant procedural tO modifications .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
..................................................................
.............................................................................................................................
tO --.a
228
Chapter 10 - E.A. Atekwana & R.V. Krishnamurthy
cipitation, precipitation and gas evolution or the vapor phase equilibration techniques, it is experimentally less involving. Using the vacutainer gas evolution technique, water samples can be obtained from many regions of the world with little difficulty and without the need for the researcher to travel to remote sites if cost of travel is a problem. For example, prepared sample tubes have been shipped to scientists in Central America and southern Africa where water samples were collected and shipped to the USA for DIC extraction. Although possible with the other techniques described here, both the logistics and the level of expertise required to conduct similar remote sampling are daunting. Using the vacutainer technique, water samples can be collected and extracted in a matter of a few hours. In a single vacuum set-up, more than 30 samples can be extracted in an 8hour day. Commercial laboratories can easily adapt this technique for routine DIC analysis. These laboratories can prepare the sample tubes, ship them with appropriate instructions to field sites for water sampling. The tubes with water samples can be shipped back to the laboratories for later extraction and isotopic determinations since no special storage conditions are necessary and the length of storage is not problematic. Although the vapor phase equilibration technique offers similar advantages for remote sampling, the DIC extraction and analysis procedure is also technically demanding relative to the vacutainer gas evolution technique. In addition, sample preservation is recommended for long periods of storage requiring additional chemical expense. The method is less amenable to water samples with low DIC concentration since only a small portion of the gas phase is sampled for isotopic analysis. In addition, it does not allow for use of the extracted DIC for possible 14-C analysis if needed. The precipitation and gas evolution techniques provide accurate determinations of t h e 13C of DIC. Accurate determinations of DIC concentration are possible with the precipitation and gas evolution technique. These techniques are less attractive due to the elaborate procedure and precautions needed to obtain accurate results. In addition, the large volumes of water sample (approximately 20 ml to 1L) used in current techniques make it less attractive for analyses involving smaller volumes of water extracted from sediment cores, for example. The main advantage of the techniques lies in the fact that sufficient CO2 can be produced for both conventional and AMS 14C determinations.
Handbook of Stable Isotope Analytical Techniques, Volume 1 P.A. de Groot (Editor) 9 2004 Elsevier B.V. All fights reserved.
CHAPTER 11 Compound Specific Isotope Analysis of the Organic Constituents in the Murchison Meteorite M. A. Sephtonl & I. Gilmour Planetary and Space Sciences Research Institute, Open University, Milton Keynes, Buckinghamshire, MK7 6AA, United Kingdom e-mail: 1
[email protected]
11.1 Introduction Most meteorites are pieces of asteroids that have been propelled towards the Earth by collisions in the asteroid belt (Wetherill & Chapman, 1988). The parent asteroids have escaped the extensive geological recycling prevalent on the planets and, consequently, are composed of extremely ancient materials. As a result, meteorites constitute a record of the early solar system and to some extent the interstellar environment from which it formed. Of all the meteorites, the carbonaceous chondrites are the most primitive and attract particular attention from the organic cosmochemist because they contain up to several percent carbon, the majority of which is present as organic matter. Less than 25% of the organic matter in carbonaceous chondrites is present as low molecular weight free compounds which can be extracted with common organic solvents, while the remaining 75% or so is present as high molecular weight macromolecular materials which are insoluble in common organic solvents (Hayatsu & Anders, 1981). Meteoritic organic matter may have been created in a number of extraterrestrial environments associated with the transformation of an interstellar cloud, through a solar nebula, into a star and planetary system (Cronin et al., 1988). Subsequently, once on the asteroidal parent body, the organic matter may have undergone modification by secondary thermal and aqueous alteration. Scrutinising the record contained within meteoritic organic matter can lead to an improved understanding of the inputs and processes contributing to the early solar system environment. Inexorably, the study of meteoritic organic matter also provides valuable information on prebiotic molecular evolution and the origin of life.
1. Correspondence should be adressed to this author.
230
Chapter 11 - M.A. Sephton & I. Gilmour
Much of our current understanding of meteoritic organic matter has come from investigations of the Murchison meteorite, approximately 100 kg of which fell in Australia in 1969. The indigenous nature of the organic matter within Murchison was quickly established (Kvenvolden et al. 1970) and over the last three decades it has been the focus of intensive research using the most modern techniques. The Murchison organic inventory has become a valuable reference to which all other extraterrestrial organic matter may be compared.
11.2 Stable isotopes The abundances of stable isotopes are generally expressed as 6 values. These indicate the difference, in per mil (%o), between the relevant ratio in the sample and the same ratio in an international standard as follows: %o - ( Rsample- Rstandard / x 1000 Rstandard
[11.1]
The stable isotopic composition of extraterrestrial materials may be determined by nucleosynthesis, nuclear processing and chemical fractionation processes (Penzias, 1980). By establishing the isotopic composition of meteoritic organic matter the significance of each extraterrestrial source region in its production and modification may be constrained. Early studies of this kind obtained isotopic measurements on broad assemblages of meteoritic compounds in the form of either the bulk macromolecule, acid residue or evaporated unfractionated solvent extracts (Smith & Kaplan, 1970; Krouse & Modzeleski, 1970; Kvenvolden et al., 1970; Belsky & Kaplan, 1971; Becker & Epstein, 1982; Robert & Epstein, 1982; Yang & Epstein, 1983). Later work achieved higher resolution and measured isolated fractions containing compounds with broadly similar chemical structures. Such crude fractions were isolated by standard preparative chromatography and then subjected to direct combustion-mass spectrometry (MS). Fractions analysed in this way have included hydrocarbons (Krishnamurthy et al., 1992), amino acids (Epstein et al., 1987; Pizzarello et al., 1991), carboxylic acids (Krishnamurthy et al., 1992; Epstein et al., 1987; Cronin et al., 1993) and polar hydrocarbons (Pizzarello et al., 1994). These studies represent significant advances in confirming the indigenous nature of organic matter by detecting noticeable enrichments in the heavy isotopes of C, N and H (D) when compared to terrestrial values. These measurements also provide valuable clues as to the extraterrestrial environments that contributed to the various classes of meteoritic organic matter, with perhaps the most notable of these being Denrichments characteristic of interstellar space. However, these isotopic measurements can only be averages of a wide variety of organic compounds. Isotopic analysis at the molecular level, known as compound specific isotope analysis (CSIA) represents a further stage in the deconvolution of the structural and isotopic constitution of meteoritic organic matter.
231
Compound Specific Isotope Analysis of the Organic Constituents in the Murchison Meteorite
Table 11.1 - Results from carbon isotopic measurements performed on individual molecules from the Murchison meteorite.
Compound Amino acids glycine D-alanine L-alanine glycine + alanine c~-aminoisobutyric acid c~-aminoisobutyric acid + c~-aminobutyric acid aspartic acid L-glutamic acid isovaline isovaline + valine glycine + alanine
613C %o
Ref.
+22 +30 +27 +41 +5 -1
(1) (1) (1) (2) (1) (2)
+4 +6 +17 +30 +41
(2) (1) (1) (2) (2)
Carboxylic acids, CO and C02 CO -32.0 +2.0 CO2 +29.1 +0.2 acetic acid +22.7 +0.2 propionic acid +17.4 +0.2 isobutyric acid +16.9 +0.2 butyric acid +11.0 +0.3 isovaleric acid +8.0 +0.2 valeric acid +4.5 +0.2
(3) (3) (3) (3) (3) (3) (3) (3)
Sulphonic acids methyl suphonic acid ethyl suphonic acid propyl suphonic acid isopropyl sulphonic acid
(4) (4) (4) (4)
Light hydrocarbons methane ethane propane isobutane butane ethene Normal alkanes n-C12 n-C13 n-C14 n-C15 n-C16 n-C17 n-C18
+29.8 +9.1 -0.4 -0.9
Compound Aromatic compounds benzene toluene naphthalene 2-methylnaphthalene 1-methylnaphthalene phenanthrene fluoranthene pyrene chrysene benzo[ghi]fluoranthene benzo(e)pyrene benzo(j)fluorenthene
(3) (3) (3) (3) (3) (3)
-35.5 -28.5 -33.2 -28.2 -25.3 -31.0 -30.6
(5) (5) (5) (5) (5) (5) (5)
Ref.
-28.7 +0.2 -28.8 +1.1 -12.6 +2.3 -5.8 -11.1 -7.5 +1.2 -5.9 +1.1 -13.1 +1.3 -14.5 +2.2 -14.2 +2.2 -22.3 +4.1 -15.4 +3.3
(3) (6) (6) (6) (6) (7) (7) (7) (7) (7) (7) (7)
Fragments of macromolecular material toluene -24.6 +0.2 -1.3 +2.0 ethylbenzene m-xylene p-xylene methylethylbenzene benzaldehyde phenol 2-methyl phenol 3-methylphenol naphthalene
+9.2 +1.0 +3.7 +0.1 +1.2 +0.1 +4.4 +0.1 +2.4+0.1 +0.1 +0.4
~)13C%o
benzothiophene 2-methylnaphthalene 1-methylnaphthalene acenaphthene
-5.4 -21.9 +1.4 -19.6 -21.7 +0.4 -20.3 -17.8 -18.1 -23.7 +0.6 -24.0 +0.3 -24.1 -10.3 -10.4 -13.9 +0.9 -18.5 +1.9 -6.5 +2.5 -5.5 +0.5 -6.4 +0.8 -15.8 -5.6 +2.1 -6.4 +0.8 -7.2 +2.0 -7.1 -5.9 +1.7
(6) (8) (8) (6) (6) (8) (8) (6) (6) (8) (8) (6) (6) (6) (8) (8) (6) (8) (8) (6) (6) (8) (6) (8) (6)
References: (1) Engel et al., 1990; (2) Pizzarello et al., 1991; (3) Yuen et al., 1984; (4) Cooper et al., 1997; (5) Sephton et al., 2001; (6) Sephton et al., 1998; (7) Gilmour & Pillinger, 1994; (8) Sephton & Gilmour, 1999.
232
Chapter 11 - M.A. Sephton & I. Gilmour
Table 11.2 - Results of nitrogen isotopic measurements performed on individual amino acids from the Murchison meteorite (Engel & Macko, 1997).
Compound R-aminoisobutyric acid sarcosine isovaline glycine f~-alanine D-alanine L-alanine L-leucine D,L-proline D,L-aspartic acid D-glutamic acid L-glutamic acid
615N %o + 184 +129 +66 +37 +61 +60 +57 +60 +50 +61 +60 +58
Table 11.4 - Results of hydrogen isotopic measurements performed on individual amino and sulphonic acids from the Murchison meteorite. Parenthical values are corrected for hydrogen exchange during the water extraction procedure. Compound
8D %o
aspartic acid glutamic acid c~-aminoisobutyric acid isovaline + valine glycine + alanine methyl sulphonic acid ethyl suphonic acid n-propyl sulphonic acid i-propyl sulphonic acid
Ref
+214 (+630) +523 (+1013) +67 (+149) +713 (1014) +1072 (+2448) +483 +787 +536 +852
(1) (1) (1) (1) (1) (2) (2) (2) (2)
References: (1) Pizzarello et al., 1991; (2) Cooper et al., 1997.
Table 11.3 - Results of sulphur isotopic measurements performed on individual sulphonic acids from the Murchison meteorite (Cooper et al., 1997). Compound methyl sulphonic acid ethyl suphonic acid n-propyl sulphonic acid i-propyl sulphonic acid
633S %0
(~34S %0
636S %0
A33
+7.63 +0.33 +0.20 +0.32
+11.27 +1.13 +1.20 +0.68
+22.5 +0.8 +2.1 +2.9
+2 -0.24 -0.40 -0.02
11.3 Compound Specific Isotope Analysis (CSIA) Determining the isotopic compositions of individual molecules in complex organic mixtures has been a long-held goal of stable isotope mass spectrometry. In recent years this type of analysis has produced a significant data set for meteoritic organic components (Tables 11.1-4). Such measurements are particularly useful as they can expose the reaction mechanisms and possibly source environments from which the organic constituents have been formed. Some attempts at CSIA involve the use of preparative chromatography to isolate fractions containing simple mixtures or individual molecules. Subsequently, in a separate procedure, the fractions are then combusted to CO2 or N2 for carbon and nitrogen isotopic measurements respectively or transformed to H2 or SF6 for determining the hydrogen or sulphur isotopic compositions. These gases are then introduced into an isotope ratio mass spectrometer and the isotope ratios measured. Measurements obtained by this direct combustion-MS method include the carbon and hydrogen isotopic compositions of C2 to C5 amino acids (Pizzarello et al., 1991) and the carbon, hydrogen and sulphur isotopic compositions of C1 to C3 sulphonic acids (Cooper et al., 1997).
Compound SpecificIsotope Analysisof the Organic Constituentsin the MurchisonMeteorite
233
Recent analytical advances have led to the combination of separation, combustion and isotopic measurement as a single procedure. The resulting gas chromatographyisotope ratio mass spectrometry (GC-IRMS) system allows the relatively rapid determination of the isotopic compositions of individual molecules within complex organic mixtures (Mathews & Hayes, 1978). As GC-IRMS relies on the GC-amenability of the analytes, volatile meteoritic compounds with hydrocarbon skeletons and few functional groups have been the focus of several studies. Using GC-IRMS, carbon isotopic compositions have been determined for CO, CO2, C2 to C5 aliphatic hydrocarbons (Yuen et al., 1984), C1 to C5 carboxylic acids (Yuen et al., 1984), C6 to C20 aromatic compounds (Yuen et al., 1984; Gilmour & Pillinger, 1994; Sephton et al., 1998) and C12 to C18 n-alkanes (Sephton et al., 2001). GC-IRMS analyses of the less volatile polar compounds are more problematical. These compounds need to be derivatised to increase their volatility and transform them into compounds amenable to analysis by GC (see Chapter 8 for a detailed description of derivatisation and GC methods). The 613C value of the derivative can then be determined by GC-IRMS and the true 613C value can be obtained following a correction for the carbon added during derivatisation. Using this procedure, Engel et al. (1990) and Engel & Macko (1997) have successfully measured the carbon and nitrogen isotopic composition of C2 to C6 amino acids from the Murchison meteorite as trifluoroacetic acid-isopropyl ester derivatives. Compound specific isotope analysis of the macromolecular material represents a formidable challenge. Although by far the most abundant organic component in carbonaceous chondrites and, therefore, arguably the most important meteoritic organic fraction, its study by CSIA has only been attempted recently. This neglect is presumably due to the analytical difficulties associated with a solvent-insoluble high molecular weight organic network. In its natural state the macromolecular material will not yield combined structural and isotopic information at the molecular level by either combustion-MS or GC-IRMS methods. The intractable nature of the macromolecular material has been overcome by the use of off-line hydrous pyrolysis to liberate C7 to Cll aromatic and alkyl aromatic fragments of the macromolecular material for carbon isotopic analysis by GC-IRMS (Sephton et al., 1998). However, although the small-scale nature of these experiments reduced the conventional hydrous pyrolysis sample size requirements from several hundred grams to one or two grams (Sephton et al., 1999), this is still too demanding for many of the meteorite samples available for analysis. Recently, in an attempt to accommodate this restriction, on-line pyrolysis-GC-IRMS has been used effectively on milligram-sized samples of the Murchison meteorite, providing carbon isotopic measurements for C7 to Cll aromatic and alkyl aromatic fragments of the macromolecular material (Sephton & Gilmour, 2001).
234
Chapter 11 - M.A. Sephton & I. Gilmour
11.5 CSIA of organic compounds from the Murchison meteorite 11.5.1 Carbon
Figure 11.1 shows that the C1 to C5 aliphatic hydrocarbons, amino acids, carboxylic acids and sulphonic acids from the Murchison meteorite appear to follow a common trend when their 613C values are plotted against carbon number. 613C values generally decrease as the amounts of carbon in the molecules increase. This trend has been interpreted as the result of a kinetic isotope effect during the stepwise formation of higher molecular weight compounds from simpler precursors (Yuen et al., 1984). The more reactive 12C is preferentially added during the synthesis of the carbon skeleton of these compounds. T h e C6 to C20 free and macromolecular aromatic compounds from the Murchison
meteorite also appear to follow a definite trend when their ~13C values are plotted
40 30
J
20 10
A ,,']:,
m I
r-1 El i r-1
A
0
@
,.~ I
0 A A
-10 -20
$8e
o
0 0
C'
/'5, I~
-30
00
V
X X
-40
o
0
' 5
X X
' , 10 15 carbon rtumber
9
O (3,
X X
X
, 20
Figure 11.1 - Carbon isotopic compositions plotted against carbon number for individual compounds from the Murchison meteorite. Standard deviations and references are listed in Table 1. Assignments are as follows: (A) free aliphatic hydrocarbon, (11) free amino acid, (~) free sulphonic acid, (rl) free carboxylic acid, (O) free aromatic compound, (O) macromolecular aromatic compound, (A) macromolecular oxygen-containing compound, (V) free CO, (V) macromolecular benzothiophene, (x) free normal alkane.
25
Compound Specific Isotope Analysis of the Organic Constituents in the Murchison Meteorite
235
against carbon number. This trend is expressed as an arch where C6 to Cll compounds have ~13C values that increase with additional carbon atoms while Cll to C20 compounds have ~i13Cvalues that decrease with increasing carbon number. The Cll to C20 trend is consistent with the C1 though C5 compounds from Murchison and indicates an origin by a synthetic process (Gilmour & Pillinger, 1994). The C6 to Cll trend is the opposite of that seen for the C1 to C5 compounds and, if this trend is also the result of a kinetic isotope effect, suggests that bond breaking or 'cracking' has produced these compounds (Sephton et al., 1998). During cracking the more reactive 12C bonds are preferentially broken leading to a concentration of this isotope in the products. Therefore the aromatic entities in Murchison seem to have been produced by a process in which both bond formation and destruction was significant. A further noticeable feature of the C6 to C20 aromatic compounds in Murchison is the similarity in 613C values for both free and macromolecular moieties suggesting a genetic relationship. Close inspection of the cases where 813C values are available for both free and macromolecular compounds reveals a consistent enrichment in 12C in the free compounds. This indicates that free aromatic compounds have been released from the macromolecular material in a pre-terrestrial generation event (Sephton et al., 1998). The ~13C values for the C12 to C18 n-alkanes from Murchison form a distinct group
in which none of the compounds exhibit the 13C-enrichment characteristic of indigenous extraterrestrial organic matter. In fact the ~13C values are identical to those of petroleum products or other terrestrial fossil hydrocarbons (Sephton et al., 2001). The Murchison n-alkanes also lack the distinctive increase or decrease in 813C value with carbon number common to the other compound classes. These features confirm the long-held suspicion that these molecules are contaminants from the terrestrial environment added to the meteorite following its fall to Earth (Cronin & Pizzarello, 1990; Sephton et al., 2001). The possibility that terrestrial contamination has produced another meteoritic organic feature, an apparent excess of the L-forms in amino acids, was investigated by Engel et al. (1990) who obtained carbon isotopic compositions of individual enantiomers. The similar 13C-enrichment of both D- and L-forms of alanine from Murchison implies that the L-excess is an indigenous extraterrestrial property.
11.5.2 Nitrogen The 615N values for individual amino acid enantiomers in the Murchison meteorite have been used to determine the origin of these compounds. Engel & Macko (1997) established that both L- and D-enantiomers of alanine and glutamic acid have significantly heavier ~15N values (ca. +60 %0) than their terrestrial counterparts (ca.-10 to +20 %o). On this basis, they argued that the excess of L-forms over D-forms is an extraterrestrial feature and not the result of terrestrial contamination. Furthermore, the 15N-enrichments in the C2 to C6 amino acids in the Murchison meteorite suggested an interstellar source for these compounds or their precursors (Engel & Macko, 1997). It
236
Chapter 11 - M.A. Sephton & I. Gilmour
should be noted, however, that Pizzarello & Cronin (1998) questioned this work and believed that other meteoritic amino acids were coeluting with L-alanine and contributing to the L-excess and 815N determinations.
11.5.3 Sulphur As sulphur contains four stable isotopes it is possible to construct a three isotope plot on which normal mass-dependent fractionations lie on a line of 633S = 0.5 634S. Deviations from this line indicate mass-independent fractionations (expressed a s A33S = 633S - 0.5 ~)34S).Cooper et al. (1997) discovered that the sulphur isotopic composition of methyl sulphonic acid from Murchison implied a mass-independent enrichment in 33S. These isotopic features were attributed to the interstellar gas-phase UV-irradiation of symmetrical CS2 molecule leading to the production of the methyl sulphonic acid precursor.
11.5.4 Hydrogen Compound specific hydrogen isotopic measurement for amino and sulphonic acids display a substantial enrichment in D when compared to terrestrial ratios (Pizzarello et al., 1991; Cooper et al., 1997). These deuterium enrichments indicate the formation of the hydrocarbon skeleton of the amino and sulphonic acids in a low temperature environment such as is found in interstellar clouds. Furthermore, it has been proposed that the relatively constant D / H ratios of the different sulphonic acids imply that the hydrogenation of their unsaturated precursors occurred within a pool of nearly uniform D-enrichment.
11.6 History of meteoritic organic matter It is becoming widely considered that significant amounts of meteoritic organic matter, or its precursor materials, are synthesised in an interstellar environment (Cronin & Chang, 1993). It is also becoming clear that, although the feedstock for meteoritic organic matter may predate the solar system, its final molecular architecture is strongly determined by the effects of aqueous alteration on the meteorite parent body. Organic precursors synthesised in interstellar space are hydrolytically transformed to water-soluble polar organic compounds (Bunch & Chang, 1980). Portions of the macromolecular material are liberated to become soluble compounds (Sephton et al., 1998) and the chemical composition and degree of condensation of the residual macromolecular material is directly controlled by extent of aqueous processing (Sephton et al., 1999). The influence that the interstellar and asteroidal environments appear to exert on the final constitution of meteoritic organic matter has led to this model being called "the interstellar-parent body hypothesis" (Cronin & Chang, 1993). Finally, following the arrival of a meteorite at the Earth's surface, its organic assemblage may be compromised by the addition of terrestrial contaminants.
Acknowledgements The authors are grateful for the constructive reviews of Dr J.R. Cronin and Dr F. Robert.
Handbook of Stable Isotope Analytical Techniques, Volume 1 P.A. de Groot (Editor) 9 2004 Elsevier B.V. All fights reserved.
CHAPTER 12 A New Method for the Isotopic Examination of Sub-Milligram Carbonate Samples, Using Sulphamic Acid (NH2.SO3H) at Elevated Temperatures H. Le Q. Stuart-Williams Research School of Biological Sciences, Australian National University, GPO Box 475, Canberra City, ACT 2601, Australia e-mail :
[email protected]
Abstract High-concentration polyphosphoric acids used for the reaction of carbonates for isotopic analysis have equilibria which are complex and difficult to quantify, resulting in reactions with carbonates having rather uncertain isotopic values. This method of reacting sulphamic acid with carbonate was developed in an effort to find an alternative technique for producing isotopically representative CO2 from small carbonate samples without using an automated device. The method is also interesting because it is entirely anhydrous and takes place at a relatively high temperature. The fixed stoichiometry of sulphamic acid potentially results in excellent reproducibility, producing gas with an isotopic composition dependent only on the make-up of the carbonate and the reaction temperature. Finely ground carbonate material is mixed with sulphamic acid (NH2.SO3H) in evacuated borosilicate glass tubes and heated to 220~ to 235~ in an isothermal furnace. Dry CO2 is produced which can be passed directly into a mass spectrometer. Typical analytical precision for the method is 0.04%0 (813C) and 0.06%o (8180), based on a test run of 14 samples. Recommended, tested, processing parameters are 200 mm by 6 mm borosilicate tubes, sample sizes of 100-350/~g matched to 3.5 to 4.5 times as much dry sulphamic acid, reacted for 30 minutes at 220~
12.1 Introduction This paper is presented in three main parts. This, the first section, is an introduction to the aspects of carbonate reaction addressed by this study and is a brief review of the chemistry of phosphoric acid, examining why the reaction is difficult to control in small scale experiments. The second part (section 12.2) describes the reaction of sulphamic acid with carbonates and quantifies preferred reaction conditions while the third part (section 12.3) provides the experimental background for the parameter selection, based on approximately 200 analyses. Different subsets of the data are used in different figures. These subsets were selected so that as many variables as possible were well controlled to demonstrate the effect of varying one particular parameter.
238
C h a p t e r 12 - H. L e Q. S t u a r t - W i l l i a m s
T a b l e 12.1 - D a t a sets u s e d in t h e F i g u r e s 12.1-14. Figure number
Material
Mass range (~g)
R a n g e of sulphamic
Reaction temperature
L e n g t h off reaction
acid: calcite
(~
(minutes)
masses 1
BangC
52 - 4187
3.7 - 4.4
220
30
2
BangC
88 - 4187
3.7 - 4.4
220
30
3
BangC
200 - 246
3.82 - 4.28
220 - 450
30 - 36
4 5a 5b
BangC BangC BangC
160 - 356 52 - 4187 4 4 - 1281
3.18 - 4.29 3.47 - 4.36 3.56 - 5.0
220 - 450 220 235
20-60 20-60 30
6
BangC
118 - 410
2.23 - 3.94
220
5 - 60
7 8 9
BangC BangC BangC
179 - 646 179 - 646 52 - 4187
1.51 - 4.36 1.51 - 4.36 3.47 - 4.36
220 220 220
20 - 60 20 -60 20 - 60
10 11
BangC BangC
52 - 4187 52 - 4187
1.51 - 21.05 3.47 - 3.46
220 220
5 - 60 20 - 60
12a & 12b
BangC
52 - 4187
3.47 - 3.46
220
20 - 60
13
BangC
161 - 235
3.83 - 4.24
235
30
14
(3 s i e v i n g s ) Synth-0% n=l
210
4.15
235
30
Synth-25% n=3
227 - 235
3.85 - 4.17
235
30
Synth-50% n=3
232 - 236
3.92 - 4.16
235
30
Synth-75% n=2
204 - 239
4.02 - 4.19
235
30
Synth-100% n=2 BangC n=18 NBS-18 n=4
191 - 246 170 - 242 2 0 7 - 249
3.99 - 4.29 3 . 8 0 - 4.33 3.89 - 4 . 1 9
235 235 235
30 30 30
NBS-19 n=4
217 - 248
3.99 - 4.17
235
30
These data sets are summarised in Table 12.1. Attempts to analyse sub-milligram samples of carbonate manually using phosphoric acid (McCrea, 1950) often produce isotopic results which vary considerably from analyses of larger samples of the same material. Automated devices of either the common- or separate- reaction-vessel type have good analytical precision for samples as small as 50 gg or less, but are not available in all laboratories. Due to the complexity of the phosphoric acid reaction and the variability of the methods used to formulate the acid, it may be desirable to have a reaction based on a reagent supplied in the condition in which it will be used, for inter-laboratory comparisons. To limit the impact of variables other than reaction temperature, the reagent should be non-aqueous, preferably reacting only above 100~ to reduce the effects of water produced in the reaction, and inert at room temperature. If an acid reactant is used then the production of H20 is inevitable (making opportunities for isotopic exchange), by the reaction CaO + H2X CaX+ H20. Consequently, in these experiments emphasis was placed on materials that contain a radical such as SO3 which might combine rapidly with produced water SO3 + HaO ---, 2H + + SO4a-. Sulphamic acid (NHz.SO3H) was selected as it has a melting point of about 210~ and is widely available. -
A New Methodfor the Isotopic Examinationof Sub-MilligramCarbonate Samples ...
239
Off-line analyses of carbonates by reaction with phosphoric acid are typically made using versions of the method published by McCrea (1950) in which the carbonate is reacted with concentrated phosphoric acid under vacuum. Usually this process takes place at room temperature or, nowadays, somewhat warmer to make temperature regulation easier. After an extended reaction at a constant temperature the reaction vessels are then opened and the CO2/H20 mixture is passed through a cold-trap to remove water vapour before the sample is admitted to the mass spectrometer. Automated methods are generally similar in principle except that the reaction temperature is normally close to 90~ and the CO2 is commonly frozen into a liquid nitrogen cooled cold finger as it is evolved. At the high concentration of phosphoric acid used for this reaction the major component is pyrophosphoric acid (Jameson, 1959). As the first dissociation constants of pyrophosphoric acid are greater than those of orthophosphoric acid, the reaction of carbonate with these mixtures may be approximated by: CaCO3 + H4P207 ~ CaH2P207 + H20 + CO2
[12.1]
Acid for the off-line process is typically made up by mixing 85% phosphoric acid with P205 with dehydration by heating until either no more P205 can be dissolved in the heated mixture or the relative density has reached 1.9 to 1.92 (Coplen et al., 1983). The product is commonly termed 102% phosphoric acid or similar. The recipe used in the laboratory where this research was conducted (University of East Anglia (UEA), Environmental Sciences, RF. Dennis, pers. com.) is rather different from the method used by Coplen et al. (1983) and consists before heating of 1 litre of 85% H3PO4 and 1.25 kg of P205. The UEA method achieves the same density as stated above but with less boiling. The resulting acids, once they have been cooled and stood, probably have similar compositions, except for minor kinetic effects relating to polymerization of the phosphate. This is not the place for a detailed discussion of the chemistry of phosphate but it is important to emphasize that phosphoric acid mixtures are to all practical intents and purposes "black boxes". For example it is clearly not possible to have 102% orthophosphoric acid (H3PO4) and the product produced is in reality more than 70% pyrophosphoric acid (H4P207)(Jameson, 1959) plus orthophosphoric acid and minor amounts of more highly polymerized phosphates, such as H5P3010. Even '100%' H3PO4 would contain approximately 10% of pyrophosphoric acid (Jameson, 1959). The first dissociation constants for all these acids are different; for example pyrophosphoric acid is a much stronger acid than orthophosphoric acid. Re-equilibration rates subsequent to deprotonation are also probably different as a result of different rates of hydrolysis and polymerization. In short, it is difficult to be certain which acid is actually reacting with the sample in what proportions. Perhaps more importantly it is also difficult to know how long water is available for equilibration with the produced CO2, although McCrea (1950) did not consider this to be a significant problem. Acid variability is not so much of a concern in automated systems where great reproducibility of conditions is possible but it could greatly reduce precision in offiine preparation, especially when
240
Chapter 12 - H. Le Q. Stuart-Williams
very small samples are being processed. In addition to poorly constrained acid chemistry, variables such as grain size must also influence the reaction rates. Despite these apparent difficulties, some off-line methods for small samples have been published (e.g. Ball et al., 1996; Wada, 1988). Common acid-bath automated systems use elevated temperatures to speed the reequilibration of the acid between samples, to reduce the viscosity of the acid and to speed the release of trapped gas. Additionall~ freezing the evolved CO2 as it is produced reduces problems associated with solution of the CO2 and probably removes the water from the system before the acid equilibrium is substantially perturbed" H4P207 + H20 --~ 2H3PO4 is slow at room temperature. This sulphamic acid method attempts to avoid these difficulties by using fast reactions in a non-aqueous environment. 12.2 Sulphamic acid reaction of calcium carbonate
12.2.1 Equipment A tube furnace with accurate temperature regulation is required. For these experiments a furnace from the author's phosphate reaction line (Stuart-Williams & Schwarcz, 1995) was used. The furnace consists of a 5 cm diameter nickel plated copper rod about 12 cm long. The rod (which serves as a large thermal mass with a very even temperature) has a 13 mm hole bored lengthwise through it, lined with a 12 mm O.D. quartz tube into which the 6 mm borosilicate sample tubes are inserted. For phosphate reactions the furnace is heated by three windings but for the low temperatures required by the sulphamic acid reaction only the central Nichrome wire winding is used. The temperature is sensed by a "J" type thermocouple inserted into a small boring in the side of the rod. The plated copper rod assembly is mounted in a stainless steel casing and insulated (electrically) with split mica sheets and (thermally) by vermiculite. The temperature of the furnace is regulated by a Eurotherm 2116 PID controller, using a solid-state relay to switch a 15 volt current from a variable transformer. The temperature varies less than 1~ from the target 220~ or 235~ except for a brief period when a cold sample tube is first inserted. The furnace was not built for these experiments and is not ideal. For example, an aluminium block with multiple cartridge-heater inserts and borings to hold several 6 mm sample tubes would permit the processing of a number of samples simultaneously, so that standards and samples could be reacted together. Samples tubes are about 200 mm long and are made in the laboratory from 6 mm thin-wall borosilicate glass tube. A carrier tool was made for loading the powdered samples into the tips of the glass tubes without smearing powder on the sides, which would result in low yields and incomplete reactions. The tool consists of a 400 mm long, 3 mm diameter stainless steel rod with a small trough milled into it close to the end. The trough is 1.3 mm wide, about 1.5 mm deep and 10 mm long.
A New Method for the Isotopic Examination of Sub-Milligram Carbonate Samples ...
241
A small, glass tube cracker was used to break the tubes and release the gas into the VG SIRA II IRMS.
12.2.2 Chemicals Most testing was p e r f o r m e d using a UEA internal standard marble, BangC (short for Bangor Carrara). X-ray diffraction of the marble detects only calcite. The isotopic composition of the marble has been d e t e r m i n e d to be: ~13C 1.995 + 0.006%0 VPDB and 6180 -1.312 + 0.020%0 VPDB. This is largely based on approximately 25 comparisons with another UEA internal standard, UEACMST. UEACMST has been extensively c o m p a r e d (hundreds of analyses) with reference material NBS-19 and has an isotopic composition very close to 613C 1.988%o and 6180-2.044%0 VPDB. The majority of the BangC used was sieved to the size fraction 63-125 ~m but is u n w a s h e d and contains some finer powder. Other materials used are noted in Table 12.2. NBS-18 (carbonatite) and NBS-19 (limestone) were also analysed, as they are well characterised and have widely separated isotopic values. To fill in the gaps and extend the tested range, additional synthetic carbonates were prepared. Distilled water was evaporated on a hotplate to 15% of its original v o l u m e and then mixed in varying proportions with u n e v a p o r a t e d distilled water. Carbonates were p r e p a r e d by leaving dissolved s o d i u m carbonate in the p r e p a r e d waters for two days and then a d d i n g calcium chloride to precipitate calcium carbonate rapidly. The materials are n a m e d Synth-X%, where X% is the proportion of 180 enriched water p r e p a r e d by evaporation. These p r e p a r e d carbonates are extremely fine-grained and could therefore not be w a s h e d or sieved. The fineness of the material was not regarded as a problem as there w a s insufficient time for isotopic exchange before the carbonate was reacted. X-ray diffraction d e m o n strated only the presence of pure calcium carbonate. The sulphamic acid (NH2.SO3H) used is from 2 batches of reagent grade chemical m a n u f a c t u r e d by BDH Chemicals Ltd., Poole, England. The reagent is p r o v i d e d as dry, coarse crystals. Table 12. 2 - Standard and reference materials used in this study. Material
NBS-18 Synth-0% Synth-25% Synth-50% Synth-75% Synth-100% NBS-19 BangC
Sulphamic analysis
Sulphamic analysis
Phosphoric analysis
Phosphoric analysis
Number of 618OPhossamples 618Osulp (sulphamic / phosphoric)
613C
8180
~13C
6180
n
A180
- 5.22 - 6.19 - 6.61 - 6.37 - 6.39 - 6.29 1.81 1.91
- 17.03 - 4.00 - 0.57 3.10 6.92 10.56 4.09 5.12
- 5.08 - 6.34 - 6.38 - 6.35 - 6.39 - 6.27 1.91 2.00
- 13.07 - 0.80 2.88 6.58 10.29 14.15 8.04 8.92
4/ 3 1/ 3 3/2 3/ 3 2/ 2 2/ 3 4/ 3 24/5
3.96 3.20 3.45 3.48 3.37 3.59 3.95 3.80
242
Chapter 12 - H. Le Q. Stuart-Williams
12.2.3 Method
12.2.3.1 Sample Preparation The 6 mm borosilicate sample tubes are placed in batches in a tube furnace and heated to 500~ for at least 30 minutes before use. They are then covered with polythene sheeting to prevent contamination by dust and stored in racks until required. Samples of carbonate (mainly BangC, see above) are weighed out on aluminium foil boats on a Sartorius Supermicro $4 balance. Optimum masses for 200 x 6 mm tubes range from 100- 350/~g. The weighed powders are carefully tipped into the carrier described above and placed into the tips of the glass tubes, striking the carrier gently to remove adhering powder. The carrier is cleaned with de-ionized water and dried with compressed air between samples. Once loaded the tubes are kept vertical until reacted to prevent spreading of the sample. Sulphamic acid is prepared by grinding the rather coarse crystals in an agate mortar. The mortar with the ground sulphamic acid is placed in the glass drying oven at 80~ when not in use. This helps to keep the chemical dry and to prevent clumping, making it easier to weigh out small amounts. The sulphamic acid is weighed out in a similar manner to that used for the samples and placed in the tubes on top of the samples using the stainless steel carrier. The masses of the sulphamic acid 9sample powder need to be matched at a ratio of 3.5 to 4.5 91. Once the acid is added to the powder the tubes should be evacuated fairly promptly: typically groups of 4 are loaded and then put on a vacuum line and pumped down to less than 6 x 10-6 mbar. Pumping down usually takes 5 to 10 minutes, so that each group completes pumping while the next group is weighed. Preheating the samples in the tubes to 100~ to drive off water had no apparent effect on the analyses and is probably unnecessary. Once the tubes are evacuated they are cutoff and sealed using a gas torch. The tubes can then be stored but must be kept upright.
12.2.3.2 Sample reaction and analysis The furnace is preheated to 220~ Before reaction the sample and reactant are mixed by holding the tube at 45 ~ from the vertical and gently rolling and tapping it. The end of sample tube containing the carbonate and sulphamic acid is then inserted into the furnace, leaving the other half of the tube projecting. The sample is reacted for approximately 30 minutes, after which it is placed in a rack at room temperature to cool. Once cool, the sample tubes are scored and inserted into the cracker on the mass spectrometer. No removal of water from the gas is required. The tubes are then run manually on the VG SIRA II mass spectrometer or as single-sample auto-runs, using a 6 minute cold finger freeze-down. Very approximate relative yields are measured by recording the beam strength and valve configuration when the bellows were not in use. The gas was not measured manometrically for this study to avoid possible contamination but more accurate yield assessment would be desirable for routine analyses.
A New Method for the Isotopic Examination of Sub-Milligram Carbonate Samples ...
243
12.2.4 Results
The results presented in this section apply only to data that were produced precisely by the method stated above. The majority of the remaining data (> 180 points), associated with testing and refinement of the method, were obtained with somewhat different parameters and are presented in the discussion section that follows. The model data here include 18 samples of BangC ranging in mass from 52 ~g to 259 ~g and 2.17 mg to 4.19 mg, plus three samples of NBS-18 carbonatite ranging from 177 ~g to 210 ~g. The 2 smallest samples (52 ~g and 88 ~g) had to be run at ion-beam strengths significantly less than the optimum 5E-9 Amps (mass 44). The mean CO2 analyses and standard deviations for the 18 BangC samples are 613C 1.941 + 0.058%o and 6180 5.407 + 0.095%o, compared with the accepted values from phosphoric acid reaction of 613C 1.995 + 0.006%o VPDB and 6180 - 1.312 + 0.020%o VPDB. One sample (the most isotopically enriched) has greater than 99% probability of being an outlier (American Society for Testing Materials Tn test) and was rejected. Without this sample and the other > 1000 ~g samples the standard deviations become 0.042%o and 0.056%o for carbon and oxygen respectively. The means and standard deviations for the 3 NBS-18 samples a r e ~13C - 5.121 + 0.012%o and 6180-16.417 + 0.030%o. The isotopic composition of CO2 from NBS-18 was determined (using phosphoric acid, see below) as 613C - 5.081 + 0.016%o and 6180 - 13.074 + 0.06%o. Using these numbers, a laboratory-standardised calcite/CO2 fraction factor of 10.229%o and the previously determined values for BangC, A18OSulphamic-Phosphoric c a n be determined for NBS-18 and BangC using phosphoric and sulphamic analyses. A18OSulphamic-Phos phoric(NBS-18) - 3.596%o and A18OSulphamic-Phosphoric(BangC) - 3.510%o. From this, the approximate fractionation factor for CO2-calcite using sulphamic acid reaction at 220~ is determined as c~ - 1.00669. The analytical precision is excellent, especially for the smaller samples. One other strength of this method should be emphasised: there is every evidence that the composition of the CO2 produced by the reaction is highly reproducible. Different batches of sulphamic acid off-the-shelf produce results which are statistically identical (see below): there is no individual variability such as may occur in the mixing of phosphoric acid. The only significant variable is reaction temperature which can be reproduced very reliably. This means that not only can repeatable values be obtained from reference materials but that both phosphoric and sulphamic methods can be compared for the same material. Despite the somewhat limited variation in the parameters used for these runs, there is some spread in the data which should be examined. Two particular types of analytical error are shown in Figures 12.1 and 12.2. Figure 12.1 shows what may be a kinetic isotope effect as each of the three sub-sets of data (large, normal and tiny) shows a very approximate fit to a slope of A180 - 2A13C. The origin of this effect is uncertain but may be related to yield (see below). The samples (with the exception of the outlier: 613C 2.105%o) have ~)13C that clusters around the probable true value for BangC" 1.995%o VPDB. This is to be expected as all carbon in the sample should be
244
Chapter 12 - H. Le Q. Stuart-Williams Figure 1 2 . 1 - The correlation between 813C and 6180. Very small samples lie slightly off the main trend, while large samples show increased kinetic fractionation with this combination of parameters. Regression of samples > 100 ~tg shows the relationship : 8180 = 1.56 813C + 2.39, with R2 = 0.88. Data parameters in Tablel2.1.
released in the reaction and converted to CO2. The 2 smallest samples may be kinetically fractionated from a gas with a slightly different initial composition. It is, unfortunately, difficult to separate mass spectrometer and reaction effects from each. Figure 12.2 shows the relationship between ~180 and relative yield calculated as beam I strength divided by sample mass (Amps-mass 44 initial beam)/(gg of sample) for the sub-milligram samples. The outlier (noted above) was removed from this diagram to produce a better graphical spread of the other data. Measurements using a variable capacitance manometer show that samples of 200 ~g produce about 70% of the theoretical maximum yield of CO2. Higher relative yields tend to produce isotopically heavier gas over a small range of variation, although larger variation is shown Figure 12.2- The relationship between relative yield and 6180. Data set as for Figure 12.1 but the smallest sample (52 /2g), which had a very low yield, was removed to improve graphic spread. Relative yield is calculated as the mass 44 beam strength in amps divided by the sample mass in /2g. Samples with higher yields are typically less isotopically fractionated. Data parameters in Table 12.1.
A New Method for the Isotopic Examination of Sub-Milligram Carbonate Samples ...
245
below to produce the opposite effect, perhaps due to the introduction of extraneous oxygen.
12.3 Experimentation: The effects of changing reaction variables 12.3.1 Introduction The experimental section is based on the entire data set. The experimental variables are highly correlated with each other, so a summary of the relationships in this paragraph is a good introduction to the more detailed discussion that follows. Much of the variation in isotopic analyses by sulphamic reaction can be attributed to changes in yield, but yield changes with a number of other parameters such as sample size, reaction temperature, reaction time, ratio of sulphamic acid:carbonate and grain size. Most of the isotopic variation probably results from the reaction of the carbonate with the sulphamic acid but non-systematic errors include fractionation during the introduction of the samples into the mass spectrometer. The introduction of organic dust may cause outliers, such as seen in samples where substantial errors in ~13C are not mirrored in the 6180. The relatively high reaction temperature may cause problems with gas being produced from carbohydrate and hydrocarbon contaminants. At lower temperatures, increasing yield is correlated with increasing reaction temperature, rising to a peak at 235 ~ - 250~ and then decreasing at higher temperatures. If these yield effects at different temperatures are removed by attempting an oxygen isotopic correction based on the 613C of the carbonate obtained by phosphoric acid reaction and the 613C obtained by the reaction with sulphamic acid, then an approximate 6180/temperature function of-0.013%o ~ is obtained. The precision of analyses gets worse at temperatures higher than 220~ for that reason 220~ is the preferred reaction temperature despite the slightly lower yield. The reaction is relatively insensitive to increases of reaction period over 30 minutes but periods from 10 to 20 minutes show increasing yield and increasing 613C and 6180. Yield is also sensitive to the sulphamic acid 9calcite ratio - increasing amounts of sulphamic acid give higher yields but the analytical precision is worse. Larger grains (for example the coarse NBS-18 grains) result in decreased yields and depleted isotopic values. Sample size is not correlated with isotopic ratio over small ranges but very large and very small samples show changed yields and isotopic ratios. This may be related to the partial pressure of the CO2 in the reaction tube. It is possible to construct functions to correct for the yield effects but this is unnecessary if the range of reaction conditions is restricted.
12.3.2 Temperature effects Substantial temperature variation produces two particular effects" changing yield and changing isotopic fractionation. The variation of yield (calculated from mass 44 beam strength, as discussed above) with reaction temperature is shown in Figure 12.3. Temperatures below 220~ produce a negligible yield as the sulphamic acid does not fuse completely. The highest yields are found at 235 ~ - 250~ Above those temperatures the yields decrease. The reasons for this decrease are unknown but may relate to either the way in which the molten acid spreads up the tube or to the production of different chemical products. Once the acid is molten it starts to diffuse up the walls of the tube and the reaction ceases either when the carbonate has been completely reacted or when it is no longer in contact with the molten acid. At 220~ some acid
246
Chapter 12 - H. Le Q. Stuart-Williams Figure 12.3 - Relative yield (see Figure 12.2) versus reaction temperature. M a x i m u m yields w e r e obtained at 235 ~ to 250~ b u t better precision w a s achieved at 220~ The grey line is not a statistical fit a n d represents the a u t h o r ' s interpretation. Data par a m e t e r s in Table12.1.
always stays at the bottom of the tube but this is not the case at higher temperatures. The exact reactions involved are unknown: the chemical products are currently unidentified: x-ray diffraction of the glassy, water soluble residues shows only the remains of the two compounds initially i n t r o d u c e d - sulphamic acid and calcite. Blanks containing just sulphamic acid produce no gases when fused at 220~ The temperature dependent fractionation is especially interesting as this is a truly anhydrous process. The results of tests made at 220~ to 450~ are shown in Figure 12.4. Three sets of points are plotted: 613C, ~180 and corrected 5180. The carbon isotopic composition should remain constant as the single carbon atom per molecule should not be temperature fractionated: any fractionation that does occur should therefore be as a result of other process associated with partial yields. The "uncorrected" 5180 includes all normal corrections to relate the gas composition to VPDB but does not include partial yield correction. The corrected points putatively have partial yield effects removed, based on the assumption that a) the carbon composition should be constant, b) that points resulting from a temperature dependent fractionation should plot close to a straight line over the temperature range tested and c) that the carbon and oxygen fractionation should be proportional. An average carbon composition for samples up to 275~ was calculated as 513C 1.93%o. The oxygen was then "corrected" according to the relationship: ~18Ocorrected = ~ 1 8 0 +
((1.93 -
~13C ) x
2)
[12.2]
assuming a mass dependent fractionation with a slope of A13C - 2 A 1 8 0 . Slopes close to 2 are indicated by other results in this study (see below). Gradients of changing fractionation factors with varying temperature are presented in Table 12.3.
247
A New Method for the Isotopic Examination of Sub-Milligram Carbonate Samples ...
IPtqlI 4b
2
IA~mL
>
o0
9
cX~
L)
d ~A
Ak
o all
-2
~)13C E3
cr)
O
~)180
-4
E3
9
6180 (Corrected)
-6
I 200
250
300
1-350
400
450
Reaction temperature ~ Figure 12.4 - The temperature dependence of 613C and 6180. Reaction periods were mostly 20 or 30 minutes but some samples were reacted for up to 60 minutes. The filled circles represent 6180 plots which have been approximately corrected to account for fractionation due to kinetic isotope effects: 618Ocorrected = 618Ouncorrected + 2 (1.93 - 613C). Data parameters in Table 12.1.
Table 12.3 - The relationship of temperature variation and changing fractionation factors. The data are subsets of the data presented in Figure 12.4. The data m a r k e d as 'Corrected' have been modified for mass d e p e n d e n t fractionation, based on variations in 613C as described in equation [12.2] in the text. Full temperature range (Corrected) Slope (%o / ~ 25~ intercept Fit (R2)
220 ~ to 300~ (Corrected)
220 ~ to 300~ (Uncorrected)
- 0.0104 + 0.001%o
- 0.0125 + 0.0005%0
- 0.0129 + 0.0005%0
7.34%o
7.9%o
7.93%o
0.92
0.95
0.96
248
Chapter 12 - H. Le Q. Stuart-Williams
The agreement of the corrected and uncorrected results for the temperature range from 220~ to 300~ is good and not substantially different from the corrected results for the full temperature range. The 25~ intercept is calculated only for comparison with phosphoric acid data: the sulphamic reaction does not occur below 200~ It is noteworthy that the projected intercept to 25~ is less than 2.5%o different from the carbonate/carbon dioxide fractionation for phosphoric acid at that temperature. The results suggest that a furnace designed to stay within 3~ of the target temperature would have a maximum thermally induced error of less than 0.04%0. The variation of sample gas ~180 with temperature is probably one of the main reasons that reaction at the lower temperature (220~ gives better precision than higher temperatures (235~ to 250~ even though the yields are greater for the higher temperatures. At 220~ the reaction can only occur between the melting point of the sulphamic acid (N 210~ and the target tem5.0 p e r a t u r e - a range of only A r d ~ 10~ whereas the range of L NL possible reaction tempera2.5 . L . I tures is much greater when a T:~U L L the furnace target temperar,/3 ~ L ture is higher. 0 . 0 immmiRiam.,aimm ,m.mm. aa--,,. ~.-_ m L
n
D n
~m ,.~ m[m O
C, D n~mu=u
I0
12.5- A) Reproducibility of analyses at 220~ For description of reaction time length see Figure 12.4. Sample masses are indicated in the top histogram. Two 6180 series are shown. The top series (180) shows the raw data. The bottom series (18Ob)is partially corrected for kinetically induced fractionation: ~)18Ocorrected = ~)18Ouncorrected + 2 (1.93 - ~ ) 1 3 C ) - 1. The subtraction of 1%o is to prevent superimposition of the data in the display. The data were acquired during two periods separated by several months. B) Reproducibility of analyses at 235~ from a single batch of analyses. Sample masses are indicated in the top histogram. The two series of 6180 analyses are as described in Figure 12.5A. Data parameters in Table 12.1.
Figure
b,~>
4
3 ~
<
~
,,2n ~ ~ m
~176 ~
0
~~ ~
n=uo
~
'~~
~176176176
D.ao,,
Batch I - Batch 2
,I,
~ 180
,,---
q
/
1 ~~
~"~" 0
oo~176
2
B
n=n n=n
~176176176
~
m
m
13C
Samples (in order created)
1.5
|i iI
EL~ 1.0 a ~:~u
Ull
~ ~ o.5
,,o mmlm o. o .--., .m mmlIm mmlmmnil Iln minin n Iml m... m.
oo c~D
5
....
~"" ' ~"~
"
"=~
'~-'" o ." o
180 m
"~ 0
b~>
9 m.m
._
~
,oooo, o~176
9 mm-m
m.m
m
,
m-,O
m
~
mI:m
.
'~
18Ob
o
CO ~
o
2
m.m. n m . m
ram-ran
luuEuh.mm I
-
mk.m.Eu
ml..mmdl I
mD411Vdl m ,.m.4ndllm
Samples (in order created)
13C
A New Method for the Isotopic Examination of Sub-Milligram Carbonate Samples ...
249
Figure 12.5A shows both ~)13C and 5180 for 33 samples of BangC run at 220~ over a period of about 5 months. A third trace shows "kinetically corrected" 5180 based on equation [12.2], which tends to reduce the g180 of the isotopically most enriched results. One permil was subtracted from the resulting values to separate the plots. The resulting plot is remarkably uniform for most samples, suggesting that almost all apparent isotopic variation results from kinetically related effects. There is little variation related to sample size, except that non-systematic errors increase for samples less than 100 g g mass. Figure 12.5B illustrates the isotopic data for 34 samples of BangC run at 235~ over a period of 2 weeks. The traces represent the same parameters as in Figure 12.5A. All analyses show conspicuously more variation than was found for reactions at 220~ Samples of less than 100 gg in particular show much greater variance with one conspicuous outlier in the centre of the diagram.
12.3.3 Reaction time and proportion of material reacted The effects of varying the length of the reaction at a temperature of 220~ are shown in Figure 12.6. Insufficient gas was obtained for an analysis for reaction times < 5 minutes. Reaction times of less than 20 minutes show low yields and correspondingly more depleted isotopic compositions. Essentially similar analyses are obtained for reaction times e 20 minutes although there is some scatter in the data resulting from a lack of a standardised sulphamic acid:carbonate ratio which was adopted later in the testing. The limiting factor for the length of reaction would appear to be the rate at which the molten sulphamic acid migrates up the walls of the tube away from the sample, resulting in little increased yield for longer reactions. Less than 100% yields can be produced in several ways, including incomplete reaction, incomplete recovery of "free" gas from a complete reaction and loss of gas by back reactions. In the case of grains of carbonate being reacted at their surfaces by an acid environment it seems probable that there can be little isotopic fractionation as
t )K C~
5
~
4
O
3
~
8
~
N
6180 > 613C
,/
2
i! 2 r
,t
Yield 0
o
':j
0
10
20
a ~CU
i
30
40
50
length of reaction (minutes)
60
0
Ox~ Z
Figure 12.6 - The effect of the length of reaction on 613C, 6180 and yield. The sulphamic acid: calcite mass ratio is lower and more variable than was employed in later tests. This has led to some relative yield variation as seen in the 30 and 45 minutes reactions. Yields are shown as e m p t y triangles, related to the scale on the right of the graph. The isotopic analyses are barely time dependent for periods > 20 minutes. Data parameters in Table 12.1.
250
Chapter 12 - H. Le Q. Stuart-Williams
there is no possibility of re-equilibration of the gas with the carbonate. Partially reacted material should therefore produce a gas of representative composition, but surprisingly it has been shown that this is not the case: the isotopic composition of CO2 evolved by the reaction of phosphoric acid with carbonate does change during the reaction (Fritz & Fontes, 1966; Walters et al., 1972). The identification of traces of carbonate by XRD in the residue of sulphamic reactions indicates that partial reaction of the carbonate may also produce fractionated gases in this case. If the cause of the low yields were incomplete recovery of an unfractionated gas then purely mass dependent fractionation could be expected, with A180 - 2A13C but in most cases the gas composition corresponds to A180 < 2A13C. Despite this, the general approximation to &180 - 2A13C probably indicates partial gas recovery as a major factor. Gas may be dissolved or lost as bubbles which are frozen into the melt as it cools to a rigid, impermeable glass. To test for back-reactions a group of sample tubes was treated in three different ways following heating: some tubes were held at 110~ for 30 minutes before cooling, some were allowed to cool by standing at room temperature and some were quenched rapidly in cold water. There was no obvious difference between the isotopic results. In summary, partial reaction certainly occurs and may influence the results, partial recovery of a fractionated gas may also be important but it is improbable that the composition is influenced by back-reactions. There is no significant exchange of oxygen between the sulphamic acid and the evolved CO2 or longer reactions would probably show changing 6180, as the 6180 of all the sample gases and the sulphamic acid cannot have been in equilibrium. As a further test of the possibility of exchange between the evolved CO2 and the sulphamic acid melt, some BangC samples were processed using another, very much older, bottle of the chemical from a different manufacturer. The analyses were essentially identical (6180 5.18%o at 235~ apart from a single outlier that may represent contamination of the reagent. Figure 12.7- The effect on the yield of increasing the sulphamic acid : calcite ratio in the reaction. Yields increase with increasing ratios, although other tests (Figure 12.10) show that isotopic analyses are less good at higher ratios of sulphamic acid : calcite. Data parameters in Table 12.1.
3.5 ~, • :Va
~A
3.0 2.5
N
"~
2.0
9 6
1.5
~
<
1.0
J~ I
d~ ql
ll~
diaL__
Tlr |
|0,
I
9
0.5 i
1.5:1
2.0:1
2.5:1
3.0:1
3.5:1
calcium carbonate
4.0:1
4.5:1
A New Method for the Isotopic Examination of Sub-Milligram Carbonate Samples ...
251
12.3.4 Sulphamic acid: car3.5 bonate ratio Ideally the isotopic com~a 3.0 position of the evolved CO2 • should not be dependent on ~ 2.5 :::L the molar ratio of the sulphamic acid reactant to the car- r 2.0 Q i @ bonate being digested. Within | @ 9 9 I 6 small ranges of variability 1.5 this is true but two effects I ! become apparent if the ratios 1.o < vary too widely. The first effect is that at sulphamic:car0.5 5.1 5.2 5.3 5.4 5.5 5.6 bonate mass ratios of 1.5 : I to 6" 1 there is increasing yield, 6180 (%o VPDB) with a slight tendency for Figure 12.8 - There is no significant relationship between relaisotopically more enriched tive yield and 6180. Data as for Figure 12.7. Data parameters CO2 to be associated with the in Table 12.1. higher yields. Relative yields (mass 44 beam strength//~g sample) are shown in Figure 12.7. Figure 12.8 shows that the 6180/yield relationship is rather weak, although a subset of the data with less range in the reactant ratios, plotted as 613C/h180 (Figure 12.9) shows that the gas is strongly mass dependently fractionated. The second isotopic effect is associated with sulphamic 9 carbonate ratios greater than 6 9 1" the evolved CO2 is isotopically depleted, at least in the case of gas produced from BangC. The effect is relatively small: at ratios of 20 9 1 the CO2 is only < 1%o depleted 5.8 (Figure 12.10). This may be 5.7 due to an admixture of gas or water from the reagent during 5.6 the reaction but blanks of sul4b o phamic acid alone as large as ~> 5.5 2 mg failed to produce any ODa measurable amount of CO2. 5.4 0 Due to these slightly ~ 5.3 unpredictable effects it is 6 advisable to control the sul5.2 I e phamic" carbonate ratio fairly 5.1 closely. Between mass ratios 1.80 1.90 2.00 2.10 of approximately 3.5" 1 and ~13C (%o V P D B ) 4.5" 1 there is no appreciable correlation with the isotopic Figure 12.9 - Correlation of i~13C t o 6 1 8 0 . Regression : 6 1 8 0 = composition of the CO2 pro1.59 613C + 2.34. R2 = 0.64. There appears to be some fractionduced, either with or without ation of the gas which is not entirely mass controlled, resulta correction for the apparent ing in a slope of < 2. Data parameters in Table12.1. .m. w
I
.m. .m.
252
Chapter 12 - H. Le Q. Stuart-Williams
kinetic isotope effect (Figure 12.11).
6.0 i
5.5
I
i_JIi
12.3.5 Sample size > There is very little rela5.0 tionship between sample size and analytical error for O 4.5 oo either 6180 (Figure 12.12A) oo or 613C (Figure 12.12B) 4.0 within the range of masses 0:1 5:1 11):1 15:1 20:1 25:1 recommended (100 - 350 Ratio of mass sulphamic acid: ~g) for use in 200 x 6 mm calcium carbonate glass tubes. Larger samples are enriched in both 13C Figure 12.10 - The effect on the 6180 of increasing the sulphamic and 180 by the KIE noted. acid:calcite ratio. Figure 12.7 shows that increasing the Sulphamic acid : calcite ratio over a narrow range improves the yield Very small samples show a (and isotopic analytical accuracy) but larger increases in the more substantial enrich- amount of sulphamic acid produce poor results. This effect may ment in 13C than 180. This result from fractionation of produced CO2 (by entrapment of may imply minor contami- gas in the acid) or from the introduction of extraneous oxygen. nation by an oxygen-free Data parameters in Table12.1. source of fractionated carbon, either in the processing or in the mass spectrometer. ,II
5.8 9 ~)180
5.7
j
O 5180
(corrected)
"~
5.6
>
5.5
0
~b
5.4
O
oo oO
C: 9
o_~~
~
5.3
OI
@
5.2 5.1
3.4
3.6
3.8
4.0
0
0
I
0
4.2
4.4
Ratio of mass sulphamic acid: calcium carbonate Figure 12.11 - The influence of minor variation in the sulphamic acid:calcite ratio on 6180. When the range of ratios is restricted to 3.47 to 4.36 there is no significant relationship between the variables: R2 = 0.00 for uncorrected 6180. Data parameters in Table 12.1.
12.3.6 Grainsize effects X-ray diffraction, as noted above, has failed to identify any of the reaction products in the resulting glassy mass. Compounds produced must therefore either be soluble in the melt or essentially amorphous, which increases the probability that the reaction is not hampered by coating of the grains with reaction products. Approximately twice as much sulphamic acid as is required for a stoichiometric relationship with the carbonate is present in the tube, so incomplete reaction of the grains is probably a consequence of the thermally induced migration of sulphamic acid away from the reac-
A New Method for the Isotopic Examination of Sub-MilligramCarbonate Samples ... The effect of sample size on 613C and 6180. Very large (> 1000 ~g) and very small (< 100 ~g) samples are often relatively enriched in 13C and 180. Data parameters in Table 12.1.
Figure 12.12.A,B -
253
:a~ 5000 4000 .m.,
8000 tion site. This effect is likely r,,o r,,rj to be more apparent with larger grains, which will take ,~ 2000 longer to react. Figure 12.13 shows analyses of three different grain sizes of BangC, 1000 coarse (125 - 180 ~m), medium ( 6 3 - 125 ~m) and @ ~~Jm edr 9 0 fine (< 63/am). The data sug1,80 1 ,,.,,., '~ 1 ,~1:1 1 ,95 '2,Ll(I 2,05 2,10 2,15 gest lower relative yields for (~13C (%0 V P D B ) the coarser sizes with slightly lower isotopic values as a consequence. The same effect B 50130 was noted with NBS-18 which has a grain size of 4000 about 250 ~m: before grinding relative yields averaged 2 mA/~g-1, rising to 2.3 mA Hg8000 1 after grinding in a mortar, o'J o'J despite the coarser material having been reacted at 235~ 2000 which usually produces a higher yield. Very high relative yields of about 2.5 mA 1000 Hg -1 were obtained from the o 9 9 9 9 ed~ q ~ j d ee'e'i 9 extremely fine grained synO 4 a thetic carbonates produced 5,1 F, 9 5 ,P, 5,4 5 ,F, 5 ,~, ~ 7 5,8 by direct precipitation 6180 (~o V P D B ) (Synth-0% to Synth-100%, Table 12.2), with close to 100% true yield. This results in a smaller ~)18OPhos - 618OSulph. Grain-sizes of 100 ~m or less are desirable to obtain good isotopic analyses and high yields, which is compatible with results from studies using phosphoric acid (Swart et al., 1991). ..
lie
.
.
.
.
.
.
".11
12.3.7 Experimentation - the overall validity of the results It has been shown that high precision analyses can be obtained using this technique by controlling variables such as the sulphamic 9carbonate ratio, but for the method to be useful the relationship between the analyses by sulphamic acid and the
254
Chapter 12 - H. Le Q. Stuart-Williams Figure 12.13 - More finely ground calcite tends to have a higher yield of gas and produces a more accurate analytical result. Coarsely ground (125 to 180/am) BangC marble has a lower yield and produces a more depleted isotopic analysis than medium (63 to 125/am) or fine (< 63 /am) sievings of the same material. Data parameters in Table 12.1.
5.4 Finb
5.3
Fin
w
q~
Fihe Coarse Mediumq}
C~ 5.2
0
O
5.1
Medium
9 5.0
O
Coarse
4.9 4.8
O
Coarse
oo
O
1.6
1.8
2.0
2.2
2.4
2.6
Yield Amp (mass 44) / ~g x 1E-n
true isotopic values must be predictable. To test this eight isotopically varied calcium carbonates, ranging from NBS-18 to an isotopically enriched synthetic carbonate prepared for this study (Table 12.2), were analysed by both conventional phosphoric acid analysis and sulphamic acid reaction. The sulphamic acid analyses for these comparisons were performed at 235~ so the oxygen isotopic values are about 0.2%o depleted relative to analyses by the preferred method at 220~ and the precision is less good than at the lower temperature. Despite the reduced precision, regression analysis of the results shows the oxygen isotopic relationship to have a slope of 1.007 and R2 = 0.999. Sulphamic analyses are on average 3.63%o depleted relative to phosphoric analyses.
15 ~
I
75%~'"
10
o
"l,
9
0
~ ~
-5
50%~" NBS-19 \9 25%4" '~/'"
Bar gC /D & ,- NBS-19
0%,,.o"/" /-,'
~
~ /BangC
8180
I-i
~~
J"'d"~
Figure 12.14- The relationship of calcite analyses using sulphamic and phosphoric acids. Carbon isotopic analyses between the two methods are very similar, with a crossplot of the two methods having a slope of unity and a zero/zero intercept. The relationship of 6180 analyses also has a slope of unity but sulphamic acid analyses are depleted by approximately 3.7 1 relative to phosphoric acid analyses at 235~ Data parameters in Table 12.1.
D 100%
,,,,."
/ / .,~ _~'NBS-18 ' II Synthetic
--813 C
-10
d"" 1
-15 -20
9 NBS-18 I
-15
-10 -5 0 813C & 8 1 8 0 (%o
5
VPDB) Sulphamic acid reaction
10
15
A New Methodfor the IsotopicExaminationof Sub-MilligramCarbonateSamples ...
255
The final data column in Table 12.2 shows that Asulphamic-phosphoricis larger for the natural carbonates than for the synthetic ones. This is almost certainly, as noted above, the result of the extremely fine grain-size of the Synth- samples and their very high yields. 12.4 Conclusions
These tests have shown that if the method described in section 12.2 is followed that isotopic analyses of 100 to 350 ~g calcite samples can be made off-line with precisions of 0.04%o (813C) and 0.06%o (6180) respectively. With good control of the reaction furnace very similar replicate analyses should be achievable in any laboratory. While loading clean tubes with weighed ingredients is somewhat time consuming, the preparation of phosphoric acid is avoided and there is no need to pass the gas through a transfer line to remove water vapour. The method could be improved by the construction of a purpose built furnace which would facilitate the reaction of numerous samples and standards simultaneously. Sulphamic analysis of carbonates is also interesting because it offers the opportunity to compare isotopic effects encountered in conventional phosphoric acid analyses with those in a high temperature anhydrous environment.
Acknowledgements
This work was conducted in the Stable Isotope Laboratory of the School of Environmental Sciences at the University of East Anglia and was funded by the laboratory. I would like to thank Paul Dennis (Head of the Laboratory), for his support and discussion, and Alex Etchells (Research Technician) for his assistance with equipment. Karen (my wife) encouraged me when things did not work! Paul Kennedy of The University of Wales, Bangor, kindly supplied the Carrara marble used for BangC. I would like to thank H. Wada and an anonymous reviewer for their help in making this paper more concise and correct.
Handbook of Stable Isotope Analytical Techniques, Volume 1 P.A. de Groot (Editor) 9 2004 Elsevier B.V. All fights reserved.
CHAPTER 13 Determination of the Abundance and Stable Isotopic Composition of Trace Quantities of C and N in Geological Samples: The Practice and Principles of Stepped-Heating at High Temperature Resolution Stuart R. Boydl CRPG-CNRS, 15 rue Notre Dames des Pauvres, B.P. 20, 54501 Vandoeuvre-les-Nancy Cedex, France
13.1 Introduction Carbon and nitrogen can both be present within geological samples in a variety of unrelated components that may have markedly different isotopic compositions. A simple example would be a metasedimentary feldspar crystal, containing 10 ppm of biologically fixed ammoniacal nitrogen, whose surfaces were 'contaminated' by gaseous N2 and air-borne organic nitrogen. Other samples can contain several components that are all indigenous to the sample. For example, carbonaceous meteorites contain trace quantities of nano-diamonds (ppm level) co-existing with high concentrations of organic matter (% level). Clearly, bulk isotope analyses of these samples would lead to meaningless results, and for carbonaceous meteorites, the isotope signal from the nano-diamonds would be totally masked by the organic matter. In such cases, stepped heating techniques (+ various pre-treatments) can be used to release C or N from the various components in a sequential manner, allowing the content and isotope composition of each element to be determined separately. Samples are heated to progressively higher temperatures in a step-wise manner (e.g., 100~ 150 ~ 200~ etc.), either under vacuum ('pyrolysis') or in an oxygen atmosphere ('combustion'). The duration of each temperature step is usually between 30 and 40 min. At the end of each step, the resulting gases are collected and purified, prior to quantification of the element of interest, and the determination of its isotopic composition. For more details on historical aspects of the technique, see Wright & Pillinger (1989) and Robert & Halbout (1990) and references therein. Components which decompose or oxidise at different temperatures will release their gases over different temperature intervals, allowing for their separation. Components occurring in trace (ppm) quantities can be concentrated by chemical/physical pre-treatments prior to the determination of the abundance and isotopic composition the of elements present. For example, Russell et al. (1991) and Hough et al. (1995) 1. Deceased in November 2001.
Determination of the Abundanceand Stable Isotopic Compositionof Trace Quantities ...
257
extracted fine-grained diamonds from whole-rock samples by a series of acid treatments. During these procedures, the components risk becoming contaminated by organic matter, either related to the chemical reagents themselves, or present within the atmosphere. Such contamination, usually rich in elements such as C and N, would interfere with any attempt at a bulk analysis. However, since the contaminants usually oxidise at low temperatures (< 500~ they can be removed during the early stages of a stepped-heating experiment (or by a precombustion; Ash et al., 1990), to leave the component of interest in a pristine state, from which the gases can be liberated at higher temperatures. For most studies, stepped-heating has only been used to isolate different components and measure their isotopic compositions. However, in the last few years, it has become apparent that stepped-heating techniques have other applications. For example, they can be used to characterise the thermal decomposition of materials, to aid in the development of new techniques (Boyd et al., 1993a), although what is perhaps of more general interest is that, by increasing the number of temperature steps (i.e. decreasing the size of each temperature increment), the shape of the release profile can aid in the identification of the component in which the element is present. This feature is advantageous when only a small amount (< 1 mg) of material is available for study, as is often the case for extra-terrestrial samples. However, by increasing the number of temperature steps, the length of the experiment may become prohibitively long for manually operated systems, hence the possibility of automation of stepped-combustion is being explored (Grady et al., 1996; Verchovsky et al., 1997). Automation will also allow for strict control of operating conditions which is imperative for subse-
Figure 13.1 - Apparatus used for high resolution stepped-combustion mass spectrometry of carbon-bearing samples: see text. E, extraction; P, purification; Q, quantification; M, mass spectrometry; HV, high vacuum; CM, capacitance manometer; MS, mass spectrometer; A-D, valves; SF, sample furnace; LNT, liquid nitrogen trap; CuO-Cu20, copper oxide furnace; VCT, variable temperature cryogenic trap.
25 8
Chapter 13 - S.R. Boyd
quent kinetic analysis of the data. 13.2 Technical aspects Figure 13.1 shows the basic apparatus for a system used for the stepped-combustion of carbonaceous samples (Boyd et al., 1998). A system for nitrogen has been described by Boyd et al. (1988). For both carbon and nitrogen, the systems that are used have four aspects in common: extraction (E), purification (P), quantification (Q) and finally mass spectrometry (M), to determine the isotopic composition of the element.
13.2.1 Extraction (E) Extraction involves the heating of the samples in a step-wise manner in regular temperature intervals (e.g. 50~ resulting in steps of 250~ 300~ 350~ etc.) for a fixed duration of 30 minutes, for example. It is important to adhere strictly to the chosen duration, otherwise a subsequent kinetic analysis of the data would be impossible. Samples are either pyrolysed under vacuum, or combusted in an oxygen atmosphere, the latter being provided by an on-line CuO-Cu20 furnace (Figure 13.1), in which wire-form copper oxide is wrapped in platinum foil and partially depleted in oxygen by pumping on the furnace with the temperature at around 900~ (see Boyd et al., 1988, 1995). Such a furnace, when operated at 850 to 920~ can provide oxygen pressures of between 10 and 90 torr, although the oxygen pressure needs to be verified periodically, since the oxygen content becomes depleted with use. At the end of the period of combustion, excess oxygen can be resorbed by changing the copper oxide temperature to 450~ A liquid nitrogen trap (LNT#1;-196~ is also present within the extraction section to remove condensible species such as CO2, which would otherwise build up in the gas phase and inhibit further reaction between the sample and the oxygen. Such a trap is also necessary, in the case of the pyrolysis of carbonates, to minimise 'backreaction' between evolved CO2 and the residual metal oxide. If it is necessary to separate trace quantities of NOx from a large amount of CO2, then a liquid nitrogen trap can be replaced by a CaO-Cu furnace (Boyd et al., 1994; 1995). When nitrogen is being extracted from samples, a CaO trap can be used to retain all condensible gases in the extraction section of the line, since there is little point in transferring them to the purification section.
13.2.2 Purification (P) For carbon, this involves the purification of CO2 for which there are numerous papers (e.g. Sakai et al., 1976; DesMarais, 1978a) and only the simplest example is covered here. After the combustion/pyrolysis period, valve A (Figure 15.1) is opened and C O 2 , SO2 and H20 are condensed onto a variable temperature cryogenic trap (VCT; Figure 13.1), held at-196~ The VCT is contained within the purification section and can
Determination of the Abundanceand StableIsotopicCompositionof Trace Quantities ...
259
operate between -196 ~ and about +150~ the latter being used for overnight degassing. The VCT is essentially a tube of glass surrounded by a heating element, the whole unit being immersed in liquid nitrogen. With no current flowing through the element, the trap is at -196~ and will condense CO2, SO2 and H20. After condensation of the gases, valve A can be closed, the next temperature step started, and valve D (Figure 15.1) opened to pump non-condensible gases such as N2. After the non-condensible gases have been removed, valve D can be closed and the temperature of the VCT raised to -130~ allowing CO2 to expand into the purification section, whilst retaining SO2 and H20; see also Miller & Pillinger (1997). A system for nitrogen purification has been described recently by Boyd et al. (1995).
13.2.3 Quantification (Q) and Mass Spectrometry (M) After extraction of the gas and its purification, valve B (Figure 13.1) can be opened and the gaseous species of interest transferred to the quantification section of the line. A gauge, such as a capacitance manometer (see Boyd et al., 1995; Boyd, 1997 for their operation and performance), is commonly used for both carbon and nitrogen although, for very small samples of nitrogen, the intensity.of the major ion beam (as recorded by the mass spectrometer) has to be used (Boyd et al., 1993b). Ideally, one would like a range of international standards that contained ppm amounts of C and N, which were homogeneous to better than + 5%, e.g. a carbon content of 250 + 10 ppm. Also it would be desirable that this result could be guaranteed for aliquots weighing about 1 mg. At present this is not the case. For example, consider the standard steel SRM - 368 which has a quoted N content of 100 + 10 ppm. In order to achieve this level of accuracy and precision, it is recommended that at least 1 g of material be used, yet the stepped-heating lines are designed to study samples having a maximum weight of typically < 50 mg. Although international standards are not yet available, Boyd & Pillinger (1991) developed a technique for producing small amounts of nitrogen standards which, for milligram amounts of material, gave errors of<+5%. Since these 'ideal' standards are not yet available, one is forced to use pure crystalline solids, such as diamond or ammonium sulphate, in which the element of interest is an essential part of the crystal structure. Simple methods for using such standard materials (very rich in the element of interest) for the determination of the abundance of trace quantities of C and N have been discussed by Boyd & Pillinger (1990) and Boyd et al. (1993b, 1995). After quantification, valve C (Figure 13.1) can be opened and the gas admitted to the mass spectrometer (MS) for the determination of ~13C (or 615N). The isotopic analysis is performed either using a conventional dual-inlet mass spectrometer for gmol(and occasionally nmol-) sized samples, or by more specialised high-sensitivity static vacuum mass spectrometers for nmol-and pmol-sized samples (Frick & Pepin, 1981; Carr et al., 1986; Wright et al., 1988; Hashizume & Sugiura, 1990; Prosser et al., 1990; Marry et al., 1995.)
260
Chapter 13 - S.R. Boyd
Figure 13.2 - Stepped-combustion of an acid-resistant residue from the Cold Bokkeveld carbonaceous meteorite. Filled symbols represent the isotopic composition of carbon released during each step (right-hand axis), the block graph represents the yield at each step (left-hand axis). The results indicate the presence of two components both highly enriched in 13C.
13.2.4 An example Figure 13.2 shows the results obtained from an acid resistant residue, extracted from the Cold Bokkeveld carbonaceous meteorite, that was believed to contain SiC (Russell, 1992). The sample was precombusted in an earlier stepped-combustion experiment (up to 550~ to remove any nano-diamonds that may have been present. The residue (~ 100/~g) was then subjected to stepped-combustion where, apart from the 550 ~ and 1250~ steps ( 50~ increments), the temperature increments were 25~ and the resulting CO2 was analysed by static vacuum mass spectrometry. Figure 13.2 shows that the residue contained at least two components, both highly enriched in 13C (~13C up to +1400%o), the first being released between 700 ~ and 800~ the second between 1000-1100~ Whilst the nature of the first component remains unknown, the release temperature of the second was comparable to that of synthetic SiC. Of interest here is the small amount of sample that was required for the analysis (~ 100~g), and
Determination of the Abundanceand Stable Isotopic Compositionof Trace Quantities ...
261
that the meteoritic SiC occurred at the ppb level, and its presence would not have been detected by a bulk analysis, which would have been dominated by the isotopic composition of the carbonaceous matter present within the matrix of the whole-rock meteorite. 13.3 Modelling the results obtained from pure samples and reference materials 13.3.1 General
In the previous sections, the principles behind stepped-heating have been covered briefly and the basic techniques employed have been outlined. Figure 13.2 showed how the technique, when combined with physical and chemical pre-treatments, can be used to isolate and analyse trace components occurring within whole-rock samples. However, in recent years, it has become apparent that the nature of the release of gas from a component (i.e. the shape of release profile), together with the temperature of maximum release, can be used to help identify the component under investigation, or at least eliminate a range of possibilities. Ideally, one would have a 'library' of results from standard materials, which could then be compared to the results obtained from samples. In anticipation of this, and to aid in the interpretation of existing data, several purely theoretical and combined experimental/theoretical studies have considered the types of result that may be expected from a range of 'simple' single-component systems. The following section considers these various 'end-member' situations, however, it must be born in mind that there may be several processes occurring simultaneously within complex samples. The simplest reactions are listed in [13.1] to [13.5] below, where A is the component of interest, and P is the gaseous product that is released (usually CO2 or N2). They are written in a simplified form to aid in the discussion that follows. A
+
A ?
--* ~
02 02
~ --*
P d i s s o l v e d in A
A A
+ +
~
P P P P P
+ + +
residue residue residue
+
oxide residue
[13.1] [13.2] [13.3] [13.4] [13.5]
Reaction [13.1] is a simple break-down scheme, such as the decomposition of carbonates, where the component breaks down to release the gas, and leave a solid residue. Reaction [13.2] is where the gas of interest is released indirectly, due to some other reaction within the component. For example, the release of ammoniacal nitrogen from micas is believed to be due to the release of water (Boyd, 1997). Reaction [13.3] is where the product is dissolved throughout the component and is released by volume diffusion. Although this has yet to be investigated fully by experiment, a mathematical treatment is provided by Robert & Halbout (1990). Reaction [13.4] is the simplest type of combustion experiment, where the component oxidises totally to give the gas, with there being no residue. Reaction [13.5] is where the component combusts to release the gas, but an oxide residue remains, a good example being the
262
Chapter 13 - S.R. Boyd
oxidation of SiC (Figure 13.2). Up to present, reactions [13.1], [13.2] and [13.4] have been studied in detail by experiment and simple models have been developed. 13.3.2 R e a c t i o n [13.1]: A ~ P + residue
A pertinent example of this is the decomposition of carbonates of the Group II metals, to give CO2 gas and a metal oxide residue. Boyd et al. (1997) studied CaCO3, SrCO3 and BaCO3. The basic reaction under consideration is,
k6
MCO 3 ~ MO
+ CO 2
[13.6]
where k6 is the rate constant for the forward reaction. The dependence of this rate constant on temperature will determine the shape of the carbon release profiles, and the temperature of maximum release. The three carbonates were subjected to stepped-heating ; in each case the samples were powders of greater than 99.9% purity, having a grain size of < 5~m. The powders were pyrolysed using 50~ temperature increments and the duration of each was 30 minutes. CO2 was quantiffed using a piezoresistive gauge and the 613C values determined by dynamic vacuum mass spectrometry. The results are shown in Figure 13.3.
Figure 13.3 - Stepped-pyrolysis of carbonates. The layout is the same as for Figure 13.2.
The reduction of the carbonate data was performed in the following way. The samples were considered to be 'perfect' powders, comprised of 'particles' of MC03, each containing only one carbon atom (either 12C o r 13C), completely independent of their neighbours. For a given particle of MC03, and a given temperature, there will be a
263
Determination of the Abundance and Stable Isotopic Composition of Trace Quantities ...
probability of this particle undergoing spontaneous decomposition. For a large number of particles this probability will manifest itself as a rate constant. At constant temperature, this model gives the same law that governs radioactive decay. So one can write, [ C I T - [Cloe-k6.t
[13.7]
where '[C]o' is the initial amount of carbon present (as carbonate) and '[C]T' is the amount of carbon remaining after time 't' at temperature 'T' and '1%'is the rate constant. Because the rate constant is dependent upon temperature, and the experiments are performed in a stepwise manner, the Arrhenius expression, k6 - A. e-Ea/RT
[13.8]
has to be incorporated into [13.7] and summed over the relevant temperature steps, 'T=I' being the temperature of the first step and 'T=n' being the temperature of the 'nth' step. T = n/ _
[C]n
-
[C]oe
Yk 6
9
~,T = lJ
[13.9]
'[Cn]' is now the amount of carbon remaining after the 'n th' step. From the stepped heating data, values of 'k6' can be obtained for each temperature, which can then be Cl
900'G '.,i
%\% '~ ~1~
m'l
5130~C
?00'G
".,,.,i
i
I
"%'. "%.,... i.
"'.~, -4
(k6],
|'~
",... %
~, ~'~. ~.\
~3 Figure 13.4 - Arrhenius diagram for the carbonate results shown in Figure 15.3. The straight lines obtained suggested good agreement between the model and the experimental data.
%,.., "X%
It., % ~1~,.
i
~.,
"1%.
" ',,
Ba%
,
I
10
"1
",.% 9
%
'%, s~, k
-10 -1'.;'
Ca
',,,.
~,.,...,
".|. "~.,
•]]l,m| '% "l
k\.
"X. I
1'.;'
104/T (.K).
14
m
"",
16
264
Chapter 13 - S.R. Boyd
plotted as a function of temperature on an Arrhe-nius diagram. Li-near relations will indicate good agreement be-tween the model and the experimental data. The results are shown in Figure 13.4. All three carbonates gave linear relationships suggesting good agreement between the model and results. However, when one looks at the isotope data (Figure 13.3), then it is apparent that the reality is more complex. If the particles of MCO3 (idealised to only contain 1 carbon atom) were completely independent of their neighbours, then one might expect that those containing 12C would break down first, since these 'proto-12CO2' molecules would be vibrating faster. Thus, during the experiments, one may have expected that the residues would have become progressively enriched in 13C. This is clearly not the case for CaCO3 and SrCO3 where the residues became mildly enriched in 12C (i.e. note the decreasing 613C with temperature in Figure 13.3). It is possible that during diffusion out of the sample as a whole, the CO2 exchanged with residual carbonate, causing a depletion in 13C in the latter, since CO2 in equilibrium with CaCO3 is enriched in 13C at these temperatures (Bottinga, 1969). The reason for the marked fractionation (11%o) associated with Figure 13.5 - Stepped-pyrolysis of ammonium bearing mica and BaCO3 is unclear. It could feldspar separated from the same granite sample. The layout is possibly be a kinetic effect the same as for Figure 13.2. The dashed lines are (a), the temperalinked to the far greater ture at which magmatic micas begin to sinter and (b), the temperature at which potassium feldspars melt incongruously. The bond strengths involved. shaded area (a) is the region over which micas lose their water, which coincides with the release of nitrogen.
Determination of the Abundance and Stable Isotopic Composition of Trace Quantities ...
265
13.3.3 R e a c t i o n [15.2]: A + ? ~ P + residue
Ammonium (NH4 +) commonly substitutes for potassium (K+) in minerals such as micas and feldspars. In the case of micas, ammonium is thus lo-cated between sheets rich in hydroxyl ions, whereas feldspars are structurally anhydrous. In the study of Boyd et al. (1993a), mica and feldspar samples were heated in 100~ temperature increments from 500 ~ to a maximum of 1300~ Apart from the 500~ step, which was performed in an oxygen atmosphere, the samples were pyrolysed. The duration of the steps was 50 minutes for the 500~ step and 36.5 minutes for the higher temperature steps. The nitrogen was quantified by static vacuum mass spectrometry. Figure 13.5 shows the results obtained by Boyd et al. (1993a) from mica and feldspar separated from the same granite (Cornubian batholith of southwest England). The dashed lines show the temperature at which magmatic micas begin to sinter (Figure 13.5a), and the temperature at which feldspar melts incongruously to leucite plus a viscous melt (Figure 13.5b). Mica loses most of its nitrogen during the 900 ~ and 1000~ temperature steps (Figure 13.5a), well before the sintering temperature of 1100~ whereas for feldspars, which are structurally anhydrous, there was insignificant release of nitrogen below the melting temperature (Figure 13.5b). For mica, the release of nitrogen coincides with the breakdown of the hydroxyl groups and the release of water which occurs between 850 ~ and 940~ (Smykatz-Kloss, 1974). Thus for micas, the release of nitrogen is related to the release of water. Boyd (1997) suggested that the following may be the rate-determining step controlling the release profile of nitrogen from mica kl0 N H 4 + OH---~ N H 3 + H20 where
d[NH4] dt - kl~ [NH4]" [OH]
[13.10] [13.11]
where kl0 is the rate of the forward reaction (of [13.10]), which is a second order reaction dependent on the concentration of two species. Micas are thus a good example of where the release of gas is related to other processes occurring within the mineral. The relevance of the study of Boyd et al. (1993a) was that, once the thermal release of ammonium had been established by stepped-heating techniques and static vacuum mass spectrometry, the results could be used to develop a bulk extraction technique (Boyd, 1997; Bebout & Sadofsky~ Chapter 16). 13.3.4 R e a c t i o n [15.4]: A + 0 2 ~ P
The simplest example of this is the oxidation of fine-grained diamond (Boyd et al., 1998). Because this is a surface reaction, the grain size has an important effect. For a single grain of diamond, at a constant temperature, the mass of carbon (mt) remaining
266
Chapter 13 - S.R. Boyd
after a given time (t) will be given by, 4
mt - P ' 3 " II" ( r o - k - t
)3
[13.12]
where r is the density of diamond (3.515 mg mm-3), ro is the initial radius of the grain (mm), k (mm min-1) is the rate constant, which in this case in the rate of decrease of the radius of the grain (Boyd et al., 1998). Since k is related to temperature by the Arrhenius equation [13.8], it has to be summed over all of the temperature steps, T = 1, being the temperature of the first step and T= n the temperature of the nth step. Equation [13.12] is for a single grain, but in reality there will be numerous grains present, so the mass of each grain has to be multiplied by the total number of grains present (N) to give the total amount of diamond remaining after the nth step (mn),
mn -
4 /3ro IT-nil
p.~.II-
-
Yk
T
-t
.N
[13.13]
1
The example chosen to test the model comes from Russell (1992) where stepped combustion experiments were undertaken on nano-diamonds which had been retrieved from a meteorite by a series of acid treatments. The stepped-heating experiment was performed at very high resolution (10~ steps for the most part), the duration of each step being 30 minutes and the C concentration was measured by capacitance manometry. In total, 34.535 gg of carbon were released during the experiment. It is possible (using equation [13.13]) to create a spread-sheet which allows the comparison of experimental results to the model (Boyd et al., 1998). The analyst can vary the initial grain size (ro) and this necessarily gives the initial mass of each grain from which the total number of grains (N) can be calculated, k values for each temperature can be calculated from the experimental data of Boyd et al. (1998). Using the spread-sheet, a theoretical release profile (block-graph) can be brought into coincidence with the experimental data. Figure 13.6 shows the results of such a comparison. There is excellent agreement between the model and meteorite residue suggesting that equation [13.13] is a valid representation. The data also suggest that there was a limited range of grain-sizes present in the natural sample. Where there is a range of grain-sizes present a very broad release can result. To highlight this point, the results for a single grain of diamond (total weight equal to the sum of the nano-diamonds) have been included in Figure 13.6, where the maximum release is close to 800~ A sample with a continuum of grain sizes, between say 10 nm and 200 gm, may exhibit a broad release between 500 ~ and 800~ In the current example, the chosen grain-size (8.8 nm) is quite close to the value (3 nm) for meteoritic diamond suggested by Lewis et al. (1987). Boyd et al. (1998) point out that the present model is simply a first approximation that will be improved with time.
Determination of the Abundance and Stable Isotopic Composition of Trace Quantities ...
267
13.4 Examples from nature In the preceding sections, it has been outlined that stepped-heating techniques are very simple in principle. There is a strong physical-chemical background and, most importantly, pure singlecomponent samples give uncomplicated results, that can be modelled quite easily. When one changes from simple reference materials, to multicomponent samples, then the results can become more complicated, this being especially true for extraterrestrial samples such as lunar soils (e.g. Brilliant et al., 1994). Not Figure 13.6 - Stepped combustion of fine-grained diamond: a comonly is this due to there parison between theoretical / experimental results and a natural being incomplete separa- sample. The open block graphs are for the theoretical / experimention of components, but tal data and were derived by computer. The block graph to the also because there might right assumes a single grain size of 0.27 mm, that to the left be several release mecha- assumes numerous grains with a diameter of 8.8 nm. The filled block graph is for a natural sample believed to be of the order of 3 nisms operating simultanm grain size. Note the excellent agreement in the shapes for the neously, that may inter- natural and theoretical / experimental (8.8 nm) profiles. fere with one another. For example, the oxidation of carbon may locally raise the temperature above that of steptemperature, causing a second component to be released at slightly lower step-temperatures than expected from studies of the pure material.
However I would stress that if complications arise in the results, they will be due to the nature of the samples, and not be artefacts of the technique. In the following discussion, three examples are considered. The examples chosen reveal aspects of the samples which would have been difficult or impossible to find using other techniques (e.g. bulk extraction). 13.4.1 Nitrides occurring as a trace c o m p o n e n t in an iron meteorite Figure 13.7 is from Franchi et al. (1993b) and depicts the results obtained from the iron meteorite ALH 77250. It was suspected that a nitride was present as a trace component so an acid attack on the bulk sample was performed. The resulting residue was subjected to high-resolution stepped-combustion mass spectrometry. A static vac-
268
Chapter 13 - S.R. Boyd
uum mass spectrometer was used and the sample was oxidised in 50~ temperature increments from 200 ~ to 1000~ the final two increments being 100~ The duration of each increment was 30 minutes. From Figure 13.7 it can be seen that there were two distinct nitrogenbearing components. The first was released over 250-550~ and had a ~15N value characteristic of terrestrial organic matter and is thus likely to be contamination introduced during the preparation of the residue. The second release Figure 13.7- Stepped combustion of an acid-resistant reaches a maximum in the 650- residue from an iron meteorite. Layout the same as for 850~ range and, with a 615N value Figure 13.2. of close to -65%o, is clearly indigenous to the sample and has cosmochemical significance (Franchi et al., 1993b). 13.4.2 Archean chert
Cherts are valuable samples in that they contain some of the best evidence for life in the Precambrian. Evidence is preserved directly as fossils or indirectly as biogenic isotope ratios in trapped organic matter (see Schopf, 1983; Schopf & Klein, 1992). Up to present the most complete study of nitrogen in organic matter from cherts is that of Beaumont & Robert (1999). However 'organic nitrogen' is not the whole story for nitrogen in Precambrian cherts. Figure 13.8 shows stepped-heating results from the Barberton greenstone belt (South Africa), which has an age of between 3.3 and 3.5 Ga, and is believed to have remained closed to argon loss for at least 3.0 Ga (de Wit et al., 1982). This was a whole rock (single chip) sample which was pyrolysed in 100~ temperature steps. The step lengths were 36.5 minutes and the nitrogen was quantified by static vacuum mass spectrometry. In addition to measuring the ~15N value of the nitrogen, the intensity of the 40Ar ion beam was determined. In crustal rocks it is common to relate nitrogen to argon since they are intimately associated. The most common form of nitrogen in crystalline crustal rocks is ammonium which substitutes for potassium (see section 13.3.3). During metamorphism, nitrogen is lost from the rocks as N2 together with water and any 40Ar from the decay of potassium, hence the close association. Nitrogen was released in a single broad peak which is correlated with the release of 40Ar. However, the variations in both 615N and 40Ar/28N2 suggested that at least three components were present. At the lowest temperatures (300~ there was a component rich in 15N with low 40Ar/28N2. This probably results from something quite
Determination of the Abundance and Stable Isotopic Composition of Trace Quantities ...
269
volatile on the surface of the sample which is unlikely to have been of Archean age. The main release of nitrogen occurred between 600 ~ and 1000~ and is characterised by a ~15N value quite close to the modern atmospheric value (0%o), with 40Ar/ 28N2 about 10 times lower (for modern air samples the measured 40Ar/28N 2 was 0.02 in the system used for the analysis). This component may correspond to water-rich fluid inclusions containing potassium (brines), since the measured 40Ar/ 36Ar ratios of between 2,000 and 100,000 (de Wit et al., 1982) show that the 40Ar was produced by the Figure 13.8 - Stepped combustion of an Archean chert. The filled in situ decay of 40K and symbols are the nitrogen isotopic composition, the open symbols was not derived from the are for the argon / nitrogen ratio. The filled block graph is for the ancient atmosphere. release of nitrogen, the open block graph is for the release of These inclusions may cor- argon. respond to metamorphic fluids derived from the breakdown of ammonium-bearing metasediments which would have been capable of providing the water, nitrogen and potassium. The most likely source of the nitrogen released at the highest temperatures was trapped organic matter, since the 615N value (close to +7%o) is similar to values obtained from organic matter which has been isolated from cherts of similar age (Beaumont & Robert, 1999). This example demonstrates that a lot of information can be gained through a simple study, aiding in the design of future experiments and shows clearly that a wholerock analysis, using a bulk-extraction procedure, would have not detected the more interesting features.
270 13.4.3 Reduced carbon in a
Chapter 13 - S.R. Boyd
lamproite
In this section, the author describes an attempt to assess the concentration of micro-diamonds in a rock-sample that was known to be rich in commercial-sized stones. With a mean grade of 680 carats per 100T (1.36 ppm by weight), the Argyle lamproite, Western Australia is the most diamondiferous diatreme discovered so far. It was with a sample from this location that an attempt was made to use stepped-combustion to find fine-grained diamonds in a rock being exploited commercially. A 6.78 g mass was powdered and subjected to acid attack (HF/HC1); note that nearly half of the rock was sediment which had been incorporated into the magma during its passage through the crust, thus the term 'lamproite' is used rather loosely. The acid-resistant residue (4.074 mg) was then analysed using stepped-combustion mass spectrometry. The step lengths were 30 minutes, carbon was quantified by capacitance manometry, and the 613C values determined using a dynamic vacuum mass spectrometer. Figure 13.9 shows the results. There were two carbonaceous components present in the residue and they are interpreted as lowtemperature organic contamination (~450~ and reduced carbon (600-800~ indigenous to the sample. The reduced carbon was assumed to be diamond and, in a manner similar to Figure 15.6, a theoretical peak brought into coincidence with the experimental data (Figure 13.9). The results are consistent with the presence of 20 ppm diamond with a mean diameter of 20 gm. However, when one considers the isotopic composition of the carbon (Figure 13.9), it is apparent that, even if diamond had been detected, the result would have been of little use in the assaying of the grade of the rock. The m e a n ~113C value of the commercial-sized stones is close to-11%o (Jaques et al., 1989). However, the peak which occurs between 600 and 800~ has a ~)13C value between -26 and -24%o. So, if it was
Figure 13.9- Stepped combustion of an acid-resistant residue from a lamproite. The layout of the lower diagram is the same for Figure 13.2. Upper diagram is a block graph, derived by computer, which assumes the presence of diamond of 20/~m diameter.
Determination of the Abundance and Stable Isotopic Composition of Trace Quantities ...
271
diamond, it was unrelated to the commercial stones. The observed &13C of the carbon in the residue is characteristic of graphite. The graphite is most likely from sediments that became incorporated into the magma, during its ascent to the surface. Due to the high temperatures, organic matter in the sediments was graphitised and now occurs as finely disseminated grains, that were concentrated by the simple HF/HC1 acid attack. Any micro-diamond that may have been present would have been masked by this graphite. Even if the above attempt was 'unsuccessful', the results demonstrate that this technique could be used to find fine-grained, dispersed graphite in ancient high-grade metasediments. In the study of Boyd & Philippot (1998), the content and isotopic composition of ammoniacal nitrogen in minerals and whole rock samples from the Moine metasediments, Scotland were determined. Whole rock ammonium contents were high (140-422 ppm NH4+), consistent with the presence of organic matter in the sediments prior to diagenesis and metamorphism. However, no traces of graphite were observed in the thin sections. The absence of graphite may be used to infer that nitrogen is (far) more stable than carbon during high-grade metamorphism. Alternatively, the graphite may have been present within the samples, but was too fine-grained and dispersed to be discerned in thin section. One way to test this would be to perform an acid attack (similar to that performed on the lamproite) and subject the residue to high-resolution stepped-combustion.
13.5 Concluding remarks Stepped-heating techniques have evolved from fairly simple experiments (3-5 steps of 200~ duration), to the situation today where many steps are performed, with the operating conditions being strictly controlled. The techniques are best suited for (1) pilot studies, where an idea of the number of components present, and their respective isotopic compositions, can be obtained from a simple whole-rock analysis, (2) experiments where only a very small amount of material is available for stud34 (3) specialised applications, such as the characterisation of a new reference material. For manually operated systems, the experiments may be prohibitively long, so fully automated systems are being developed. The latter would lead to increase in the quality of the data, permit more pilot studies to be undertaken, and allow the analysts to do more fruitful things, such as reading in the library.
Acknowledgements I.P. Wright and H.R. Karlsson are thanked for constructive criticism of the manuscript.
Handbook of Stable Isotope AnalyticalTechniques, Volume 1 P.A. de Groot (Editor) 9 2004 Elsevier B.V. All fights reserved.
CHAPTER 14 Stable isotope measurements of atmospheric CO2 and CH4 B. H. Vaughnl*, J. Miller2,4, D. F. Ferretti3 & J. W. C. White1 INSTAAR,University of Colorado, Boulder, CO, USA 2 NOAAClimate Monitoring and Diagnostics Laboratory, Boulder, CO, USA 3 National Institute of Water and Atmospheric Research, Wellington, New Zealand 4 CIRES,University of Colorado, Boulder, CO, USA e-mail: *
[email protected] 1
14.1 Introduction Atmospheric carbon dioxide and methane are important greenhouse gases with increasing concentrations in the atmosphere, and have significant chemical affects on the atmosphere and radiative forcing impacts on climate (IPCC, 1990,1995, 2001). Isotopic analyses of CO2 and CH4 provide key tools for better understanding global budgets of these trace gases. In this chapter we describe two different methods for measuring carbon and oxygen isotopes in atmospheric CO2 and a method for measuring carbon isotopes in atmospheric CH4. The first method for CO2 uses traditional dual inlet Isotope Ratio Mass Spectrometry (IRMS) with cryogenic extraction on fairly large (~450 mL) air samples. The second method for CO2 uses Gas ChromatographyIsotope Ratio Mass Spectrometry (GC-IRMS), which utilizes an order of magnitude smaller (45 mL) air samples. The GC-IRMS method is less precise than the dual inlet, but has other attractions such as the elimination of liquid nitrogen use for CO2 extraction, and no requirement for the N20 correction (which contributes to masses 44, 45 and 46 on a dual inlet system). And finally we describe a GC-IRMS system for analyzing carbon isotopes in atmospheric CH4 that uses cryogenic focusing and small samples (40 mL) of air.
14.1.1 What do isotopes in atmospheric CO2 tells us? Carbon and oxygen stable isotope ratios of atmospheric carbon dioxide, ~)13CO2 and 6C18OO, provide important, independent information about carbon sources and sinks. Combined with CO2 mole fraction measurements, the 613CO2 measurements can be used to quantitatively separate fluxes between the atmosphere and the terrestrial biosphere from fluxes between the atmosphere and the ocean (e.g. Keeling et al., 1989, 1995; Tans et al., 1990; Ciais et al., 1995; Francey et al., 1995). This is because C3 plants discriminate against 13CO2during photosynthesis while little isotopic discrimination occurs during carbon uptake by the ocean. The 613C measurements have mainly been used to indicate the one-way carbon fluxes (in the case of fossil fuel release) or net carbon fluxes (e.g. the resultant of photosynthetic uptake and respira-
Stable isotope measurements of atmospheric C O 2 and CH 4
273
tory releases), whereas 6C18OO measurements more reflect the large (gross) natural cycling of CO2 between the atmosphere and surface reservoirs, and are only just beginning to be usefully exploited. The two main mechanisms for controlling 6C18OO on annual to decadal time scales are oxygen isotopic exchange with soil water and oxygen isotopic exchange with leaf water. This isotope can potentially be used to separate photosynthetic and respiratory fluxes for land plants (e.g. Francey & Tans, 1987; Farquhar et al., 1993; Ciais et al., 1997). Note, the gross fluxes become a complicating factor in 613C interpretation when isotopic disequilibria between the atmosphere and surface reservoirs develop (e.g. due to 13C depleted fossil fuel release) and result in second-order 613C changes not reflecting net CO2 exchange, and require careful consideration of response times of exchange (Tans et al., 1993). As the atmosphere integrates surface processes over space and time, CO2 concentration measurements, combined with isotopic measurements, provide constraints for regional scale sources and sinks of atmospheric CO2 on time scales of months and longer. Indeed, these measurements provide our primary constraints on surface fluxes and thus the processes and factors, climatic and otherwise, controlling these fluxes. In the future, atmospheric monitoring may play a central role in verifying any international carbon emission agreements in much the same way that seismic monitoring was used to monitor compliance with nuclear test ban treaties. The degree to which isotopic measurements made on atmospheric samples are useful is seriously constrained by the precision of the mass spectrometer used. For example, a change of just 0.02 %o in 613C measured at one site could translate to an equivalent of 1.0 x 109 metric tons of carbon in models of surface fluxes. Such precision is challenging enough on a short-term basis, but this precision is needed over decades if we are to use the data to study trends over longer periods of time. A high precision instrument is required along with diligent, frequent intercalibrations between laboratories. This section describes the dual inlet isotopic measurements made at the Stable Isotope Lab at the Institute for Arctic and Alpine Research (INSTAAR) at the University of Colorado. Similar, but slightly different methods of measurement are done at other laboratories, including Scripps Institution of Oceanography (SIO) at UCSD, Lajolla, California; The Center for Atmospheric and Oceanic Studies, Commonwealth Scientific and Industrial Research Organization- Atmospheric Research (CSIRO) in Aspendale, Australia; and at Tohoku University (TU), Japan (Keeling et al., 1989; Nakazawa et al., 1997). 14.2 Dual inlet mass spectrometry measurements of atmospheric CO2 Since 1991, the Stable Isotope Laboratory at INSTAAR, University of Colorado has been measuring the stable isotopic composition of atmospheric CO2 from weekly flask samples of air obtained from the network of sites operated by the NOAA Carbon Cycle Group, at the Climate Monitoring and Diagnostics Laboratory (CMDL) in Boulder, Colorado. This operation began with a selection of six sites and two ships in 1990, and in subsequent years, the measurement effort has grown to include all 55 sites in the CMDL program. During calendar year 1999 over 11,000 isotopic analyses of 613C
274
Chapter 14 - B.H. Vaughn, J. Miller, D.F. Ferretti & J.W.C. White
and 5180 of C O 2 w e r e made, including 7,800 sample flasks and 3,200 measurements of air standards used for calibration. In 1990 the program began making measurements with a VG SIRA Series II dual inlet mass spectrometer. This instrument and extraction system routinely achieved an overall reproducibility of 0.03 %o for 613C and 0.05 %o for 5180 (lo). This includes errors in the extraction process and is determined from replicate measurements of air standards over many years. A Micromass Optima dual inlet IRMS was used beginning in 1996. This machine is fitted with a custom manifold and extraction system, and is used exclusively for making measurements on atmospheric gases. The overall reproducibility for the Optima system is +0.01%o 513C and +0.03 %o for 6180 (at lo, for n= 30).
14.2.2 Experimental 14.2.2.1 Sample and standard manifold The automated flask measurement system is composed of four parts: a 40-port manifold; the CO2 purification system; the mass spectrometer, and a computer than controls the system. The system performs a number of operations simultaneously (multi-tasking), as one sample is analyzed in the mass spectrometer, CO2 from the next sample is being purified from the air. For a measurement run, 20 pairs of 2.5 L glass flasks are attached to a 40-port manifold using Cajon UltraTorr fittings. The sample flasks used in the NOAA/CMDL network are Pyrex, taped for safety purposes, and have two stopcocks that allow flushing of the flask when the samples are obtained. Samples are always taken in pairs and filled simultaneously. Results are retained if there is acceptable pair agreement in the isotopic ratios determined (discussed below). The sample flasks are filled to 1.0 to 1.3 x 105 Pa, depending on the altitude of filling. After other gas measurement systems at NOAA/CMDL have extracted aliquots (for CO2, N20, CO, SF6, H2), the pressure in the flasks is 0.8 to 1.1 x 105 Pa when they arrive for analysis of ~)13C and 5180 of C O 2 . Air standards used for this analysis system are obtained from the NOAA Climate Monitoring and Diagnostics Laboratory (CMDL) Carbon Cycle Gases Group (CCGG) and are stored at 2000 psi in aluminum tanks (LUXFER size ALl50, bare aluminum with brass pack-less valves, Scott Marrin, Riverside, CA, USA.). The cylinders are filled from a clean air site at Niwot Ridge, Colorado, USA and details of the preparation of these standards can be found at: http: / / www.cmdl.noaa.gov / ccgg / refgases / airstandard.html. Two such atmospheric air standards are attached to the manifold, using stainless steel tubing (0.0625 inch OD x 0.05 inch ID, with Swagelok-to-VCO fittings), and regulators (High-Purity, Single-Stage, Stainless Steel, model E11-C444A, Air Products and Chemicals, Inc, USA) set to a delivery pressure of 6 psi. There may be other regulators that can work as well, however, it should be noted that considerable testing led to choosing this regulator over others. For example, a variety of problems have been discovered with regulators that employ Viton seals. Manufacturer's tests for analytical contamination are of limited use, and successful performance of a regulator in isotopically sensitive systems can only be determined by long term testing. This includes extended periods of non-use, to allow for any effects of de-gassing of sealant materials
Stable isotope measurements of atmospheric CO 2 and CH 4
275
into the body of the regulator to be seen. A single stage regulator was chosen over a two-stage regulator because 1) they have fewer wetted parts, 2) they are cheaper, and 3) precise outlet pressure is not needed, as the downstream mass flow controller sets the flow rate. Any changes in outlet pressure as the tank drains over time are easily adjusted. Air samples from sample flasks or standard tanks enter the manifold through air actuated, low dead volume bellows valves (stainless steel, model SS-6LV-BNBW4, Nupro Company, Willoughby, Ohio, USA). The 14/35 ground glass joints on the flasks seal well with 0.5 inch Cajon UltraTorr fittings with thick wall (#2-111, 0.104 inch thick) Viton-7 O-rings used in place of the standard wall (#2-014, 0.070 inch thick) O-rings. Electro-polished stainless steel tubing (0.25 inch O.D. thick wall) is used to connect the flask ports to the Cajon VCO fittings on the manifold. 14.2.2.2 The C02 extraction system Air samples are pulled from the manifold through the extraction system using a rotary vane vacuum pump (model RV-3, BOC-Edwards, Wilmington, MA, USA) at a flow rate of 40 standard cubic centimeters per minute (scc/m). The flow is maintained by a mass flow controller (BOC-Edwards, model 825, series B, 0-100 scc/m)(Figure 14.1). During the first minute of air extraction, 40 scc are allowed to waste directly to the vacuum pump, flushing the lines. Following the flush, the sample is then diverted first through a glass water trap, held in an ethanol bath chilled to-85~ by a refrigerated probe (Cryocool, model cc-100, Thermo NESLAB, Portsmouth, NH, USA), and then through a CO2 trap, cooled to liquid nitrogen temperature. This is basically the Triple Trap extraction system provided by Micromass with two modifications: the addition of a flow controller, and the replacement of the standard water trap with a more efficient one. At the end of the sample extraction time, the flask port is closed, and when the flow rate drops to 10 scc/m, a valve is opened that allows the remaining sample air to bypass the mass flow controller (Figure 14.1).
As complete removal of water vapor is critical to high precision measurements, and as the efficiency of the water trap determines in part the speed with which air flows through the gas purification system, we use a water trap that employs a number of features to maximize the trapping efficiency. The water trap is made of Pyrex with an initial 0.75 inch O.D. vertical section that is filled with glass beads to provide a large, cold surface area (Figure 14.2a). A 40-60 micron glass frit at the bottom of this section keeps the beads in place and helps prevent ice crystals from escaping this section of the trap. This is followed by two loops of 0.25 inch OD glass that are 80% submerged in the ethanol. This section traps and re-traps any water vapor or ice crystals that might have escaped the beads and frit. The glass trap is removable, and is held in place using two 0.25 inch Cajon UltraTorr fittings. This allows wet traps to be removed and dried offiine after each daily run. Traps are dried in a 120~ oven with forced air circulation to speed the drying time. The CO2 trap is modeled after the Micromass Triple Trap, employing a liquid nitrogen bath, and an open-bottomed 2.0 inch O.D. tube surrounding the 'U' shaped
276
Chapter 14 - B.H. Vaughn, J. Miller, D.F. Ferretti & J.W.C. White
Figure 14.1 - A general schematic showing the flask/reference tank manifold with the extraction system, including mass flow controller, water trap, CO2 trap, and sample/reference bellows leading to the dual inlet mass spectrometer. The connection from the extraction system to the reference bellows allows large extractions of CO2 from tank air to be used as reference gas. Tank air CO2 has isotopic concentrations very similar to flask samples which minimize any gas memory issues in the mass spectrometer source.
sample trap (Figure 14.2b). The trap is electrically insulated with glass tape, and wrapped with resistive heating wire (~350 cm of 18 gauge, 80% NiChrome wire). With a vent valve on top of the tube closed, the activated heating element boils the LN2, building N2 gas pressure in the open bottom tube. This displaces the LN2, allowing the resistive heating element to further warm the sample to release the previously trapped CO2. Once the C O 2 is extracted from the air and frozen in the trap, the system is pumped to high vacuum (5 E-8 mBar) for 60 seconds. Typically, each 400 cm3 air sample introduced to the system yields ~6.5 mmol CO2 for isotopic analysis. The isolated sample is then warmed to -20~ and expanded for 60 seconds into the sample bellows. The mass 44 sample beam is then balanced to match the reference target beam to less than +2% at 5 x 10-9 a. 14.2.2.3 The mass spectrometer
The mass spectrometer currently used at SIL is a Micromass Optima dual inlet IRMS machine with an electro-magnet. The mass spectrometry is standard. Sample and reference gas flow through matched capillary tubes into the source where they are ionized, and focused into an ion beam. After exiting the source, the ion beam is deflected by the magnetic field, directed irtto Faraday cups, where the beam currents
Stable isotope measurements of atmospheric CO 2 and CH4
277
Figure 14.2- a)The CO2 trapping system, based on a Micromass design, allows the trap to be alternately controlled to set points of-196~ -20~ and +25~ Liquid nitrogen provides the low point. The warmer temperatures are reached by controlling a NiChrome heating wire (24VAC) on the trap and closing the top vent on the (2 in. dia.) displacement tube. The resulting pressure of gaseous N2 displaces the liquid N2 out the open bottom of the displacement tube, allowing further heating, b) The pyrex beaded water trap incorporates an 40-50 micron glass frit at the bottom of a 0.75 in. dia. tube filled with 5 mm diameter beads. There are two open loops that extend out of the chilled bath, to ensure complete removal of water from the air stream by warming and re-trapping of any mobilized ice crystals. are m e a s u r e d for masses 44, 45, and 46. Beam currents are amplified, converted to frequencies, and transmitted to the microprocessor by fiber optic cables. The beams are allowed to stabilize for 60 seconds before measurement. The ratios of beams are calculated from integrations taken during 12 r e f e r e n c e / s a m p l e switches, with outlier rejection b e y o n d l c~. The n u m b e r of r e f / s a m switches is determined by two factors. The lower limit is set by the desired internal precision of less than 0.005 %0 for 645. This typically requires 8 switches. As the system multi-tasks, extracting one sample while the previous one is being analyzed, additional time is available for switches. The u p p e r limit on switches thus is set by the time required to extract the next sample. The raw data are corrected for experimental artifacts, including two corrections: a correction for the presence of small quantities (~10-3 relative to CO2) of N 2 0 (Mook & Jongsma, 1987), which is t r a p p e d along with CO2, and an "ion correction", accounting for the contribution of isotopic species containing 170 (Gonfiantini, 1981). The carbon isotope data are reported as ~ values relative to VPDB-CO2, in units of per mille (%o). The 6 notation is given as"
278
Chapter 14 - B.H. Vaughn, J. Miller, D.F. Ferretti & J.W.C. White
~)13C= [(13C/12C)sam / (13C/12C)ref- 1] x 1000
[14.1]
and applies to ~)180 as well using the VSMOW scale, where 0 %o VPDB - CO2 - 41.47 per mil VSMOW. 14.2.3 Results and discussion
14.2.3.1 Technique performance Data quality is checked in three ways: flask pair agreement, outlier rejection, and internal standards run as samples. All data are retained and reported so that users can assess data quality. Flags are provided with the data to signify known problems, however, so that data with known problems can be easily removed before the data is used. Air samples are taken in paired flasks that are filled simultaneously. For all species, data are retained only from those flasks that demonstrate good agreement between each sample pair. The criteria for acceptable pair agreement used at SIL is agreement with three standard deviations, or I A~)13C I <0.03%o and I A~)180 1 <0.15%o. Between trips to the sampling sites, the flasks are filled with dry air of known nonatmospheric concentrations and isotopic ratios of CO2 and other gases. Flasks that are improperly flushed with air can thus be identified and the data flagged. In addition, the data are examined as time series and are routinely filtered for outliers relative to the general trends observed. Such outliers may be real and represent unusual climatic or atmospheric circulation features, or they may represent non-baseline conditions when air was coming from known sources of contamination. These sources may include nearby industries or, in the case of shipboard sampling, air blown back from the engine exhaust stack. Flagging for each type of outlier is handled differently so data that may represent real atmospheric events are not lost. As an internal quality check, SIL began in late 1996 to measure three aliquots of air from a cylinder of air in the middle of each daily run on the mass spectrometer. This additional cylinder, called the "quality control tank" or "trap", is handled in exactly the same manner as the samples, and provides an independent check on the performance of the entire system. The variability of this tank over time is used to determine our overall reproducibility, and problems in the analysis system are frequently first seen and subsequently diagnosed using this trap tank. The recent behavior of the trap tank is shown in Figure 14.3 and illustrates the value of this quality check. Shown are 10 point running means of the standard deviation (lo) for ~)13C and ~)180 values of the trap tank. The average standard deviation during the last 160 analysis runs is 0.010 %o for ~)13C and 0.042 %o for 6180. Known factors that can cause this number to vary include: source tuning and cleanliness, contamination of the inlet/extraction system by water or organic compounds from gas regulators, and overheating in the CO2 trap. This latter problem occurs when the autorun sequence stops during heating of the trap to release the CO2 and the trap temperature exceeds 20~ This isotopic effect on subsequent samples is not clearly understood, but may result from changes to the stainless steel trapping surfaces inside the tubing that occurred when essentially 'baked' at abnormally high temperatures.
Stable isotope measurements of atmospheric CO 2 and CH 4
279
Figure 14.3 - The 10 point running means of the standard deviation of the quality control tank 613C and 6180 as a function of run number.
14.2.3.2 Standards and calibrations
The results are calibrated to external standards, and to internal whole-air references. The internal set of tank-references consists of a hierarchy of 18 aluminum tanks filled with dry atmospheric air obtained from a clean air site located at 4,000 meters elevation on Niwot Ridge in the Rocky Mountains, Colorado, USA. Eight of these tanks are cycled through as the 'daily' reference measurement tank, after first performing as the trap tank. Five more tanks are measured on a monthly basis, with one tank purposely 'spiked' with 13CO2 t o be 2%o different from the others in 813C. Five other tanks are analyzed every 6 months, with two of those being spiked by 1%o and 2%0. The tanks are stored horizontally to help minimize any gravitational fractionation, and each tank is assigned its own regulator for the life of the tank air. Multiple tanks are used to guard against any systematic bias resulting from one drifting tank. It is extremely unlikely that any long-term changes in the isotopic concentration experienced by one tank would be identical for all. It is essential for providing long-term continuity and precision, that a suite of standards be used, where some are measured daily, and some less frequently. The less frequently measured tanks span years of analyses, there by linking the daily tanks.
280
Chapter 14 - B.H. Vaughn,J. Miller, D.F. Ferretti & J.W.C. White
The standard tanks are tied to the IAEA V-PDB CO2 scale via measurements of carbonates and waters provided by the IAEA and measured by established procedures (Coplen, 1983). As these measurements are made on different extraction systems than the one used for CO2 in air, we have found that the precision with which air standards can be calibrated to the V-PDB CO2 scale is less than the precision with which air standards can be intercalibrated. Consequently, four labs around the world have been working with the IAEA to resolve internal scale differences and establish air standards for stable isotopes in atmospheric CO2. So far, air standards can be measured to a much greater precision than carbonates. This remains, to date, a fundamental problem with tying the carbon isotopes of atmospheric CO2 to the carbonate scale. Two types of comparisons between laboratories are maintained as an essential element in the globalization of atmospheric isotopic measurements. The first is an exchange of 5 CLASSIC cylinders circulated by Commonwealth Scientific and Industrial Research Organization- Atmospheric Research, (CSIRO) among four different labs, including Scripps Institution of Oceanography, (SIO), The Center for Atmospheric and Oceanic Studies, Tohoku University (TU), INSTAAR-CMDL, and CSIRO. Each lab group analyzes the cylinders for CO2, N20, 813C, and 8180, and can compare results. While individual labs make their own internal tie to primary carbonate and water (IAEA) standards, the tanks allow very small relative differences between labs to be examined. The second comparison method between labs is the flask inter-comparison program (ICP). ICP takes an entirely different approach by allowing two or more labs to make measurements on the same flask samples obtained from the same site at the same time, on a regular basis though with lower precision (Francey et al., 1994; Masarie, et al., 2001). The ICP compliments the cylinder measurements by providing 1) ongoing intercomparisons that allow weekly feedback on sampling and measurement methods, giving laboratories an opportunity to pinpoint problems, and 2) the ICP flask samples are sampled and measured in the same way as the flask measurements themselves, thereby eliminating any potential differences in measurement that may arise from pressurized tanks vs. sample flasks. For example, the ICP program with CSIRO at Cape Grim, Tasmania has identified calibration offsets between the two labs that have led to minor modifications of analysis procedures at both labs. ICP calibrations, combined with rotating tanks such as the CLASSIC tanks, are essential to global integration of greenhouse gas measurements. 14.2.3.3 Problems with humid air Drying of air during sampling is important to the quality of the 8180 measurements, made using the NOAA network flasks. For samples collected at humid, tropical locations without drying, the 180/160 measurements are highly variable and consistently more depleted in 180 due to the exchange of oxygen atoms between CO2 and H20 molecules on the walls of the flasks. Systematic tests at INSTAAR (Gemery et al., 1993) showed that the exchange takes place during storage in the flasks, and that this exchange can occur at humidities less than saturation values. There is also clear dependence on the physical characteristics of each flask. This problem can be
281
Stable isotope measurements of atmospheric CO 2 and CH 4
seen in the percentage of retained flask pairs, which is much lower for ~180 (68%) than for 613C (94%). The reason for the lower success rate in 6180 pair agreement is closely tied to the moisture problem, as seen in the trend in low pair agreements at low latitude (generally humid) sites (Figure 14.4). Consequently, data from 30~ to 30% are all flagged as suspect. While the seriously contaminated ~180 of CO2 data are obvious (offsets of 4%o), it is difficult to unambiguously determine water exchange problems in the more subtle cases. However, preliminary comparisons of "wet" and "dry" air at several locations indicate that sites outside of this range are likely not affected. Field testing of a new prototype air sampling apparatus began at American Samoa (SMO) in September 1994 and Cape Kumukahi, Hawaii (KUM) in May 1995. The new AIRKIT (Air Kitzis sampler) differs from the previously used MAKS (Martin and Kitzis Sampler) in two important ways. First, it uses a thermoelectrically cooled condenser to remove water vapor from the air stream, and secondly it has a microprocessor to control the sampling process so that collecting the sample is more automated and less subject to operator error. The effect of drying the air sample is most dramatic for the measurement of 180/160 in CO2. Figure 14.5 shows the dramatic improvement in the 6180 values of CO2 from Seychelles, (Mahe Island, 4 ~ 40' S, 55010, E) when the switch to using dried air was made in September 1998. It is also not clear if the 613C values may be affected by the moisture as well. For example, in autoruns with larger numbers of "wet" flasks, the trap tank value appears to be very slightly shifted in 613C (less than 0.003%0). As the precision of isotopic measurements improves, this suspected problem may become an issue. []
80
.
13c
-~
~
n cP~
I~ 3
~6 [] []
[]
%o~e []
~176
[]
[]
[]
9
[]
0
9
n[]
[] 6 1 8 0
40
[]
I' [] 9
oo
oi
oo O0 9
[]
Figure 14.4 - Percentage of flasks from the NOAA network retained for both 613C and 6180 at 50 different network sites, plotted against latitude. Moisture is a key factor in low retention rates at low latitudes for 6180.
l
9
-40 [] []
[]
9
-80 i
0
,
I
I
20
40
I
I
I
I
r-i
h,I
l
60
Percent Retained
i -r
8O
100
282
Chapter 14 - B.H. Vaughn, J. Miller, D.F. Ferretti & J.W.C. White
Figure 14.5 - Measured ~)180 of atmospheric CO2 in flask samples from Seychelles site (4~ 55~ showing data from analyses on wet flasks prior to 1999 and dried flasks there after. 14.2.4 Dual inlet CO2 summary The dual inlet method of measuring the carbon and oxygen isotopes in atmospheric CO2 provides a robust, high precision technique for measuring long term variations in concentrations. The overall reproducibility of standards used in this system is +0.01%o ~)13C and +0.03 %o for ~)180. However, a relatively large sample size (>400 mL) is required, and significant quantities of liquid nitrogen are consumed in the process of analysis. Barring improvements or breakthroughs in the precision of GC-IRMS methods or optical absorption techniques, this method will remain a benchmark method for measuring the carbon and oxygen isotopes in atmospheric CO2. 14.3 Gas chromatography-isotope ratio mass spectrometry (GC-IRMS) measurements of atmospheric CO2 14.3.1 GC-IRMS introduction Three problems are experienced with traditional dual-inlet IRMS methods used to perform isotopic measurements of atmospheric CO2" 1) co-extraction of contaminating N20 such that post-measurement correction is required 2) the requirement for relatively large samples of CO2 (a few tenths to several retool), and 3) the requirement for coolant use.
To avoid the problems and sources of error that arise due to the inclusion of N20 in the CO2 sample during dual-inlet analysis, either a standardized correction procedure must be used by all laboratories, as recommended by Allison et al. (1995), or N20-free analysis of samples must be achieved.
Stable isotope measurements of atmospheric CO 2 and CH 4
283
To address these shortcomings in the dual inlet technique, this paper describes a GC-IRMS technique that enables N20-free, isotopic analysis of atmospheric CO2 that has been recently developed at National Institute of Water and Atmospheric Research (NIWA) in New Zealand (Ferretti et al., 2000). 14.3.2 Experimental 14.3.2.1 GC-IRMS inlet system The layout of the GC-IRMS system is shown in Figure 14.6. Aliquots of air samples are introduction into an evacuated sample loop and pressure-equilibration-volume (PEV) and are described in section 14.3.2.3. A 10-port valve (Valco Instruments Co.
F i g u r e 1 4 . 6 - Gas chromatographmisotope ratio mass spectrometry (GC-IRMS) schematic. The 10-port valve and adjacent plumbing are enclosed in a temperature controlled box (TCB) and automatically switch between load/back flush (solid loops) and inject (dashed loops). Pressure equilibration of sample gas before injection to the GC column occurs in the pressure-equilibration-volumn (PEV). After GC separation, gas effluent flows through a Nation drier and enters the open split. The Nation drier is cooled by a thermoelectric cooler (TEC) and is purged by a countercurrent flow of clean dry "zero"air. To enable gas stream switching, the sample and purge helium capillaries alternately move into the immediate proximity of the tip of the transfer capillary, which is permanently positioned at the bottom of the open split. The open split has two positions: (1) "open split in," (sample capillary extended and helium capillary retracted, as shown in Figure 14.1), where the Finnigan MAT 252 IRMS receives undiluted GC sample effluent and, (2) "open split out" (sample capillary retracted and helium capillary extended), where the Finnigan MAT 252 IRMS, receives pure helium. The shutoff valve (SOV) is closed to enable operation of the IRMS in dual-inlet mode.
284
Chapter 14 - B.H. Vaughn, J. Miller, D.F. Ferretti & J.W.C. White
Inc., Houston, Texas) is used to direct sample air into the GC, an HP5890 Series II (Hewlett Packard, Avondale, Pennsylvania, USA) where gas separation is performed with a packed column described in section 14.3.3.1. Switching between sample/reference gases and vacuum on/off is performed with pneumatic valves (Nupro) controlled by LabView (National Instruments, Austin, Texas, USA). After GC separation, sample gas enters the open split interface via a Nation MD Series Gas Drier (Perma Pure Inc., Toms River, New Jersey,) that is discussed further in section 14.3.3.2.1. Two independent air-actuated pistons switch between the sample capillary (deactivated quartz glass, 0.32 mm ID flow rate 3.0 mL/min) and the helium purge gas capillary (deactivated quartz glass, 0.32 mm ID flow rate 5 mL/min). Transfer of sample gas or purge helium from the open split to the IRMS is through a deactivated glass capillary, 0.11 mm ID, 1 m long, (SGE, International Pty Ltd., Ringwood, Victoria, Australia) at a flow rate of 0.3 mL/min. Obtained precision was found to be critically dependent on the open split design, which is further discussed in section 14.3.3.2.2. Isotopic analysis of CO2 is performed by a Finnigan MAT 252 IRMS with an accelerating potential of 10 kV. This IRMS utilizes a M u l t i e l e m e n t - Multicollector (MEMCO) system with Faraday cups to simultaneously measure m / e 44, 45, and 46 ion currents, respectively. The ion currents for the m / e 44, 45, and 46 beams are integrated at 4 Hz as the Gaussian peaks elute from gas chromatographic column and the Finnigan software calculates ~)13C and 6180 values. The IRMS was tuned for a compromise between maximum sensitivity (<1200 molecules/ion) and minimum nonlinearity (<0.10 %o/V for 613C and <0.15 %o/V for 6180, respectively) between ~0.2 and 1.5 V (~0.7 to 5 nA) m / e 44. In the course of the development, new Finnigan source slits were installed to the IRMS, the advantages of which are discussed in section 14.3.3.1.
14.3.2.3 Analysis procedure
Routine analysis is provided by injection and analysis of (1) a reference gas aliquot, (2) three aliquots of a sample, and (3) a second aliquot of the same reference gas. This bracketing of the sample aliquots with reference gas aliquots further reduces the effect of any instrumental drift (e.g. as a result of laboratory temperature change) that may occur during the period of analysis. The first and last step of the analysis procedure purges the sample loop and adjacent plumbing (see Figure 14.6) with reference gas for 30 s at a flow rate of 60 mL/ min. After the reference gas flow stops, the pressure is allowed to equilibrate for 5 s to ambient pressure after which injection through both the precolumn and main column occurs for 230 s. Following injection, the precolumn is back flushed for 270 s. Meanwhile, the open split interface switches from directing pure helium to the IRMS to sample effluent, ~60 s before the eluting CO2 peak. Integration of the CO2 peak follows. Approximately 45 s after the final integration point of the CO2 peak, the open split redirects the flow to the IRMS to pure helium.
Stable isotope measurements of atmospheric CO 2 and CH 4
285
The middle step of the procedure introduces 3 sample gas aliquots. The sample loop and adjacent plumbing is effectively flushed by evacuation prior to sample introduction so that sample wastage is minimal. To ensure peak reproducibility, the same amount of gas must be introduced to the IRMS for each sample aliquot. Thus each sample aliquot must be at the same pressure, volume, and temperature (PVT). To facilitate this, a vacuum is applied to the sample flask valve, through the 0.5 mL sample loop and into a 5 mL PEV, 1.36m x 3.175 mm OD (1.016 mm ID.). After introduction of the sample aliquot to the sample loop, the PEV valve is opened to enable sample loop equilibration to ambient pressure. For sample flasks with above ambient pressures, outflow from the PEV to the laboratory atmosphere occurs, and for samples at or below ambient pressure, inflow from the laboratory atmosphere into the PEV occurs. This equilibration is necessary to ensure that sample/reference gas injections are at the same PVT, assuming minimal variations in lab pressure. During this equilibration for samples collected at or below ambient pressure, the PEV is necessary to avoid contamination by stopping "back flow" of laboratory air into the sample loop. Furthermore, during sample loop filling, up to 5 mL of sample gas purges through the sample loop to ensure that no residual from a previous sample or reference gas is injected as a contaminant. Initially, for each new sample, the PEV valve is closed, and all plumbing to the flask valve is evacuated by a turbo molecular pump to a pressure of ~0.01 Pa. This volume is filled, reevacuated, and refilled before an analysis commences. The vacuum pump connection then switches from turbo to roughing pump for the automated run. After the reference gas injection and 450 s prior to each subsequent sample injection, the PEV valve is closed, and the PEV, sample loop, and adjacent plumbing are evacuated. Typically, a vacuum of --0.02 Pa is obtained in this time, after which the vacuum inlet is closed, and the sample valve opens, filling the sample loop, PEV, and adjacent plumbing. To ensure equilibration between the sample flask and sample loop, the sample valve remains open for 30 s. Pressure equilibration of each sample aliquot to ambient pressure then occurs by opening the PEV valve for 5 s. To avoid sample contamination during pressure equilibration of each sample aliquot, it is critical that inflow of laboratory air into the inlet volume must not pass beyond the PEV. For 0.5 litre air samples at ambient pressure this condition was met so long as less than 6 aliquots of air sample are extracted. For routine analysis, total sample usage, including all dead volume, for three aliquots of an ambient pressure air sample is 45 mL. This usage is high compared with the total amount actually injected onto the column (three 0.5 mL aliquots) and can be lowered by reducing the currently limiting 15 mL inlet volume (sample loop, PEV, and adjacent tubing and valve volumes). 14.3.3 Results and discussion
14.3.3.1 Separation of C02 and N20 A Porapak-Q column provided separation between C O 2 and N20 peaks. After the installation of the new proprietary Finnigan source slits, the "memory effect" of the Finigan MAT-252 mass spectrometer was significantly reduced. As a result the "tail-
286
Chapter 14 - B.H. Vaughn, J. Miller, D.F. Ferretti & J.W.C. White
ing" of the CO2 peak was reduced, and the CO2- N20 separation was further improved. This enabled complete separation of N20 from the integrated CO2 peak (Figure 14.7) giving N20-free determinations of 513C and 5180 of atmospheric CO2. The column used is a Porapak-Q, 3.66 m x 1.59 m m OD (1.016 m m ID), stainless steel, 80/100 mesh with a 1.83 m pre-column of the same material (Alltech). The column is maintained at a temperature of 41~ (+ 0.05~ with a helium carrier gas flow rate of 3.0 mL / min.
14.3.3.2 Maximization of signal-to-noise ratio 14.3.3.2.1 Contamination and carrier gas purity High carrier gas purity (99.995%) is required to minimize background contamination and noise. This is achieved with the use of an in-line gas purifier (ALL-Pure Helium Purifier, Alltech) to purify helium of purity greater than 99.995% so that contaminants (CO, CO2, 02, H20, and NMHCs) were reduced to ppb levels.
Figure 14.7- (a) 45/44 isotope ratio and (b) mass 44 chromatogram. Results of separation using the memory effect reducing source slits: The tick marks shown are the positions at which the software starts and ends the peak integration. The 1000 times smaller N20 peak, which is visible as a small blip on the large CO2 tail, is completely separated from the integrated CO2 peak, allowing for a completely N20-free calculation of 613Cand 5180 in CO2.
Stable isotope measurements of atmospheric CO 2 and CH 4
287
Entrainment of laboratory air into the inlet system results in increased background levels and noise and loss of precision. This is mainly because laboratory air is relatively "wet," and the formation of HCO2 + molecules in the ion source contributes to a loss of precision and accuracy (Leckrone & Hayes, 1998). Thus it is imperative for high-precision results to maintain the inlet system completely leak free and dry. IRMS background water levels are minimized and maintained at a constant level by passing all GC effluent through a Nation drier consisting of a 610 mm long, 0.762 mm ID Nation tubular membrane in a 1.59 mm OD stainless steel sheath. Clean, dry "zero" air (dew point < -80~ purges the Nation drier at a flow rate of --100 mL/min. No detectable effect on the CO2 mixing ratio was observed when using zero air compared to air containing CO2 by this technique. The Nation purge flow (~100 mL/min) is high compared to the sample flow (--3 mL/min) to ensure high drier performance (Leckrone & Hayes, 1997). A thermoelectric cooler (TEC)(Tropicool, Christchurch, New Zealand) cooled the second half of the Nation to 0~ By cooling the Nation the vapor pressure in equilibrium with the membrane decreases, and the effectiveness of the drying is enhanced (Leckrone & Hayes, 1997). The addition of the TEC decreases the dew point of the emerging dried gas from ~ -45~ to < -80~ corresponding to a drop in background water (m/e 18), as measured on the most sensitive detector in the Finnigan MAT 252 IRMS, from 2.2 to 1.030 V (2.2 to 1.03 pA).
14.3.3.2.2 Open split design The open split interface is an integral part of the system as it forms the critical link between the GC and the IRMS and allows for the continuous flow of either sample gas or pure helium into the IRMS. We designed the open split to minimize entrainment, contamination, and sample dilution and maximize reproducibility and precision. The background noise is further magnified when operating in GC-IRMS mode, owing to variations in open split entrainment. This occurs because the relative pressure difference between the slightly above ambient pressure open split (due to the purge helium) and the laboratory is affected by pressure variations. Our design minimised entrainment as a result of normal pressure fluctuations in the laboratory so that no detectable contamination occurred. Our open split design consists of a Pyrex glass test tube, open at the top, with the following dimensions: 60 mm long, 1.9 mm OD, and 1.3 mm ID. Gas stream switching, previously described in Figure 14.6, enables a high signal-to-noise ratio. The split ratio is ~1:9; that is, of the sample effluent that enters the open split at 3.0 mL/min, only ~10 % actually enters the ion source at 0.3 m L / m i n (this factor is currently limited by chromatography and not the open split design). Thus, in the technique described here and for samples at current atmospheric levels, N8 nmol CO2 is injected onto the GC column from each 0.5 mL aliquot of air sample, and 0.8 nmol of this enters the IRMS source.
14.3.3.3 Reproducibility: Temperature effects Reproducibility and precision of the measured CO2 mixing ratio are strongly affected by temperature and pressure variations that occur for each eluting peak dur-
288
Chapter 14 - B.H. Vaughn, J. Miller, D.F. Ferretti & J.W.C. White
ing a GC-IRMS analysis. Because of the IRMS nonlinearity, ments are also affected.
6180 and
613C measure-
Temperature variations are generally non-linear and non-monotonic over the analysis period and are therefore important for reproducibility. In addition, temperature variations change the value of the high-precision resistor that regulates the magnet current in the Finnigan MAT 252 IRMS, (W. Brand, personal communication, 1999). The detected signals are therefore affected because magnetic field strength variations cause the ion beam to be shifted. These affects are specific to GC-IRMS, and do not affect the dual inlet measurements made on the instrument. The effects of laboratory temperature variations should be minimized by stabilizing the temperature over the entire inlet system to <0.2~ Insulation of the inlet system, from the reference gas regulators through to the open split, can achieve this if laboratory temperature stability is not to this level. In addition, temperatures must be stabilized in the transfer capillary and SOV to the IRMS to <0.1~ once again, by insulation if required. Furthermore, the sample loop, PEV, and adjacent plumbing are maintained at a stable temperature of 35.00 + 0.02~ by enclosing this section within an insulated, temperature-controlled box (TCB). Temperature control to this level within the TCB is achieved with the use of an external GC temperature sensor and a 40 W cartridge heater. A 2 W electric fan circulated air within the box, and two thermal masses of 150 mL water and 1.6 kg brass were positioned in the box. These measures reduced the uncontrollable temperature effects due to laboratory air temperature variations and improved technique precision.
13.3.3.4 Analysis time
The analysis time is determined by the requirements for sufficient N 2 0 - CO2 separation, (greater than 20 seconds between the end of the integrated CO2 peak and start of the N20 peak), high signal-to-noise ratio, and a sufficient number of reference and sample aliquots for good analysis statistics. The analysis time includes the time required for the separation of CO2 from the air sample. Increasing the carrier flow rate in the column decreases the analysis time. However, reduced separation and increased sample dilution occurs, resulting in lowered signal strength, reduced signalto-noise ratio, and lowered precision. For three aliquots of an air sample with two reference gas aliquots bracketing the sample (one at the beginning, one at the end) the balance between analysis time, adequate separation, dilution, and precision is achieved with an analysis time of 40 min, comparable to that required for dual-iMet analysis. Because a longer GC column or a slower injection flow rate would need to be used for the technique described here, mass spectrometers that have a large memory effect may not be used (without modification) to obtain high-precision, N20-free isotopic analyses of atmospheric CO2 in short analysis times.
14.3.3.5 Technique performance
Maximum performance can be evaluated by considering the "shot noise limit." This limit is based on ion collection statistics and refers to the precision that would be obtained if the ion beam were the only significant noise source (Petterson & Hayes,
Stable isotope measurementsof atmospheric C O 2 and C H 4
289
1978). A simplified expression given by Merritt & Hayes (1994) for the shot noise limited precision (cJ~) expressed as a function of the integrated m / e 44 signal area (44A, V.s), for the reference and sample peaks respectively) is ~62 - 0.00446(1 / 44Areference + 1 / 44Asample)
[14.2]
For the GC-IRMS technique described here in this paper, Ferretti et al. (2000) gave the precision of the technique for 613C, 6180, and CO2 mixing ratio, determined by both analysis of atmospheric air samples and analyses of air standards, to be 0.02 %o, 0.04%0 and 0.4 ppm, respectively. Comparing the observed precision to the theoretical shot noise limited 613C precision given by the above equation for our system, it was seen that the technique performs on average at the shot noise limit of 0.02%0. Evaluating technique performance over a range of sample sizes can be performed by introducing differing amounts of CO2 to the ion source of the IRMS by varying the open split dilution. The split ratio of 1:9 was increased by introducing reference and sample CO2 with the helium capillary also extended and increasing the open split helium purge flow from the normal 5 m L / m i n to ~20 mL/min. Replicate analyses of an air standard by this method are shown in Figure 14.8. Over the observed range of 150 pmol to 1 nmol CO2 in the source, the technique performs within a factor of 1.3 of the shot noise limit. At the lower limit of 150 pmol CO2 in the source, the technique performs within a factor of 1.5 from the shot noise limit, indicating that the effects of noise or systematic error become more significant at these lower sample sizes.
Figure14.8 - Observed and theoretical (shot noise limited) standard deviations for varying amounts of CO2 in the IRMS source.
290
Chapter 14 - B.H. Vaughn, J. Miller, D.F. Ferretti & J.W.C. White
It can be seen from the theoretical prediction (Figure 14.8) that by increasing the amount of gas reaching the IRMS source (while maintaining matched reference and sample sizes), higher performance (i.e., lower standard deviation) is possible. However, this performance is ultimately limited by IRMS detector overload, which for eluting CO2 peaks from the experimental technique presented here in a Finnigan MAT 252 IRMS, would occur at ~4 nmol CO2 in the source. A sustained shot noise limited precision of 0.01%o is theoretically possible with 3 nmol CO2 in the source. Without overloading the GC column, this would be possible by this technique if the split ratio could be reduced to ~1:2.2. This precision of the GC-IRMS would then be comparable to that of the dual inlet technique. 14.3.4 GC-IRMS method summary While the theoretical precision of 0.01%o for ~13C has yet to be sustained on a routine basis, GC-IRMS measurements, even at a coarser precision, opens many new doors for analysis of isotopes in atmospheric CO2. These new opportunities for trace gas research are now possible because of the smaller sample requirement of GC-IRMS. The logistics of new sampling methods (e.g. sampling from pilot less aircraft and international commercial flights) will be eased as will sampling and analysing trace gases from ice cores. The GC-IRMS technique developed in this work can be used as a "front end" in a modular approach to GC-IRMS for other trace gases (e.g. CH4, CO and N20) where on-line preparation of these gases can be performed by miniaturized versions of existing methods and injected directly into the CO2 GC-IRMS system presented here for isotopic and mixing ratio analysis. In addition, for special applications, the sample usage of the technique described here could be minimised to less than 10 mL (with the inclusion of low dead-volume vales and tubing).
As high-precision measurements of isotopes in atmospheric C O 2 a r e key for separating terrestrial biospheric and oceanic exchanges of carbon with the atmosphere and for potentially separating regional scale respiratory and photosynthetic fluxes in terrestrial ecosystems. The primary challenge is to achieve and maintain sufficiently high precision over decades to observe variability of the carbon cycle. Typically, GC-IRMS systems have less precision than dual inlet techniques, and thus, for atmospheric monitoring, the primary challenge for GC-IRMS techniques is to achieve acceptable precision in the isotopic measurement. Thus we are pushing our analytical capabilities to their limit in this application of stable isotopes, and small artefacts that may be unobservable or unimportant in sample collection and mass spectrometry may become important in the future. The keys to success in this approach are diligent intercalibration and intercomparisons of laboratories from around the world, as well as the use of multiple techniques such as dual inlet and GC-IRMS mass spectrometry. To this effect, we have commenced intercomparison studies between the laboratories discussed throughout this paper (INSTAAR, NOAA/CMDL, NIWA, and CSIRO). These programmes will undoubtedly be invaluable over the years to come both as quality control for each laboratory and for the decadal consistency from data from different laboratories as it is integrated into global carbon cycle models.
Stable isotope measurements of atmospheric CO 2 and CH4
291
14.4 Measurements of 13C/12C in atmospheric CH4 14.4.1 Introduction
Atmospheric CH4 is an important chemical component of both the stratosphere and troposphere and is a major contributor to the enhanced greenhouse effect. In the stratosphere, methane is a major source of water vapor (Jones & Pyle, 1984) and is the primary sink for chlorine radicals (Cicerone & Oremland, 1988), and thus plays an important role in the regulation of stratospheric ozone levels. In the troposphere, CH4 consumes about 25% of all hydroxyl radicals, and as a result is an in situ source of CO and 03 (Thompson, 1992). Models indicate that the contribution of methane emissions to greenhouse warming is twenty times that of CO2 on a per molecule basis (Lashof & Ahuja, 1990). It is estimated that methane accounts for approximately 20% of the increase in radiative forcing by trace gases since the onset of the industrial era (Myhre et al., 1998). The atmospheric burden of methane has more than doubled in the last 150 years (Etheridge et al., 1992; 1998) and over that time is highly correlated with human population (Blunier et al., 1993). Over the last 40 years the growth rate of methane in the atmosphere has averaged nearly 1% per year (Cicerone & Oremland, 1988; Etheridge et al., 1998) but has been steadily decreasing over the last 15 years (Steele et al., 1992; Dlugokencky et al., 1998). Neither the rapid increase nor the recent slowdown is fully understood, and this is directly related to the large uncertainties in the magnitudes and spatial distribution of identified methane sources. Estimates of the emission rates of various sources are typically based upon small-scale field measurements (Cicerone & Oremland, 1988, and references within) that are extrapolated to large spatial scales. A few studies have used forward (Fung et al., 1991) and inverse (Brown, 1993; Hein et al., 1997; Houweling et al., 1999) modeling approaches to estimate source distributions based on atmospheric measurements. Nonetheless, considerable uncertainties remain in the estimates of source strengths. The measurement of the stable carbon isotope ratio in atmospheric methane (e.g., Lowe et al., 1994; Quay et al., 1999) and in methane sources (e.g., Tyler, 1986; Conny & Currie, 1996) may allow for a significant reduction in the uncertainties of the magnitudes of various methane sources. If we can m e a s u r e 13C / 12C of atmospheric methane with sufficient precision, and the kinetic fractionation associated with its consumption by the hydroxyl radical (Cantrell et al., 1990) and soil microbes (King et al., 1989), then we can determine the mass-weighted isotopic average of all methane sources at steady-state. When the mole fraction or ~13C of CH4 are not at steady-state, we also need to know their growth rates. If an isotopic "signature" can characterize different methane sources, then the mass-weighted average will be a constraint on the magnitudes of various methane sources. From a 13C point of view, the sources of methane may be divided into three categories: bacterially produced methane, like that from wetlands or ruminants; fossil methane, like that associated with coal and natural gas deposits; and methane produced from biomass burning. Each of these three classes has a fairly distinct isotopic signature, with bacterial methane ~13C ~-60%0, thermogenic methane ~13C ~--40%0, and
292
Chapter 14 - B.H. Vaughn, J. Miller, D.F. Ferretti & J.W.C. White
biomass burning methane ~13C ~ - 2 5 % o (e.g., Quay et al., 1999). Individual methane sources may differ significantly from their source type's characteristic signature, but the values above are averages that are probably valid on large spatial scales. In principle, we should be able to constrain the emissions from these three source types from global atmospheric measurements. A few studies have reported globally and temporally distributed values of ~13C in CH4 (Quay et al., 1991, 1999; Stevens, 1995). Quay et al. (1999) reported more than 600 measurements between 1988 and 1995 from biweekly sampling at Barrow, Alaska, Olympic Peninsula, Washington, Mauna Loa, Hawaii, and American Samoa in addition to less frequent sampling at Cape Grim, Tasmania, and from Pacific Ocean ship transects. Stevens (1995) reported 201 measurements, mostly from the continental United States, between 1978 and 1989. ~)13C of methane in the Southern Hemisphere has also been regularly monitored at Baring Head, New Zealand since 1990 (Lowe et al., 1994). We have established high-precision measurements of ~13C of methane on a global basis, using a subset of sites in the NOAA/CMDL Cooperative Air Sampling Network (e.g., Conway et al., 1994). Since January 1998, we have measured ~13C of methane from six sites (Table 14.1) ranging in latitude from 90~ to 71~ from pairs of flasks collected on a weekly basis. The NOAA network gives us the potential to measure 613C of methane from more than 60 land and ship-based sites. In order to take advantage of the high temporal and spatial density offered by the network, we have designed an automated gas chromatography- isotope ratio mass spectrometry (GC-IRMS) system that analyzes samples using 200 mL of air in less than fifteen minutes. Traditional analysis methods (e.g., Stevens & Rust, 1982), on the other hand, are severely constrained by the 15 - 60 L of air typically used and the labor intensive sample extraction and analysis. This section describes the analysis system and presents some preliminary measurement data. Table 14.1 - NOAA/CMDL Air Sampling Sites Used in this Study Site Code BRW CGO MLO NWR SMO SPO
Site Pt. Barrow, AK Cape Grim,Tasmania Mauna Loa, HI Niwot Ridge, CO American Samoa South Pole
Country
Latitude
Longitude
Elevation (masl)
USA Australia USA USA USA Antarctica
71 ~ 40~ 19~ 40~ 14~ 89~
156o36' W 144~ E 155~ W 105~ W 170o34' W '24~ W
11 94 3397 3749 42 2810
N S N N S
Stable isotope measurements of atmospheric CO 2 and CH 4
293
14.4.2 M e t h o d s
Sample analysis can be separated into six steps: sample introduction, methane preconcentration, cryo-focusing, chromatographic separation, combustion, and mass spectrometric analysis. The details of the reference air used, batch analysis and quality control will also be discussed below.
14.4.2.1 Sample collection Ambient air is pumped through a pair of serially connected 2.5 L glass flasks fitted with two glass-piston stopcocks sealed with Teflon O-rings. Conway et al. (1994) have described the collection method in detail. Whole air reference gas is collected in aluminum high-pressure cylinders at the NOAA/CMDL cooperative site at Niwot Ridge, Colorado, USA (40~ 105~ 3040m). On average, flasks contain less than 1.5 standard liters of air by the time they are analyzed for 1 3 C / 1 2 C ratio of methane, which was a major constraint in the design of the analysis system. Air pressure in the flasks is also about 2 x 104 Pa below ambient pressure (8.5 x 104 Pa) when extracted for measurement.
14.4.2.2 Sample introduction Flasks are attached to a manifold described in detail by Lang et al. (1990) in preparation for analysis. The circular manifold (Figure 14.9) is evacuated up to the flask stopcocks by a rotary pump (Edwards E2M5) to a pressure less than 3 x 10-2 mbar. The stopcocks on the flasks are then opened allowing the air inside to expand through tubing to an eight-port stream selection valve (Valco SD8, Valcon M rotor) fitted to a sixteen position electric actuator. These extra actuation positions allow the manifold to be in a "blanked off" position between the analyses of samples. A diaphragm pump (KNF) then pulls air out of the flask at rate of 100 m L / m i n (STP), controlled by an electronic mass flow controller (Edwards 1605). The air then flows through an Ascarite II (NaOH on a silica substrate) and Mg(C104)2 trap to remove CO2 and water vapor from the sample. The CO2/water trap is a 15 cm x 6 mm I.D. glass trap consisting of a six cm layer of Ascarite II sandwiched between two, 2 cm layers of Mg(C104)2, with small plugs of glass wool at each end. The Cajon UltraTorr fitting holding the trap on the downstream side also has a 10 mm stainless steel frit to prevent particles from entering the rest of the system. After leaving the trap, the air flows to a 40 mL sample loop positioned on a six-port, two-position injection valve (Valco 6-UW, Valcon E rotor). After flushing the sample loop and trap for 120 seconds, the injection valve is switched so that a flow of He (99.999 % purity, further purified by Alltech "All-Pure" He purifier) flushes the contents of the sample loop to another six-port, two-position valve containing the pre-concentrator (Figure 14.9). Note that the flow rate of the He stream is only pressure regulated resulting in changing flow rates with temperature and flow path. The flow rates through the pre-concentrator are 22 mL/ min (STP) at room temperature and 30 m L / m i n (STP) at-120~ The introduction of air from a reference tank has been designed to be as similar as possible to the introduction of flask air, so as to minimize any potential offset between analysis of reference air and sample air. The only difference is that reference air flows
294
Chapter 14 - B.H. Vaughn, J. Miller, D.F. Ferretti & J.W.C. White
Figure 14.9 - Plumbing diagram for methane separation and combustion apparatus.
through the diaphragm pump while it is off. A downstream regulator pressure of at least 2 x 104 Pa above ambient pressure on the reference air tank is needed to overcome the resistance of the water/CO2 trap and maintain a flow of 100 m L / m i n (STP). A total of approximately 200 standard mL is used in each sample analysis. This volume is more than four times the volume of tubing that is flushed but decreases the chances that the trap contains any "memory" of the previous sample from run to run.
14.4.2.3 Sample pre-concentration Pre-concentration of the CH4 within the air sample is necessary to ensure that N2, 02, and Ar do not co-elute with methane from the analytical column. N2 entering the combustion furnace can be oxidized to N20, which interferes with the m / z - 44 and 45 signals that result from CH4-derived CO2. In general, we want only CH4-derived CO2 (and He) in the mass spectrometer during its analysis. The pre-concentration step is to isolate methane on a substrate while N2, 02, and Ar are vented. Our pre-concentrator is based on the design of Merritt et al. (1995b) and modified to ease automation. The pre-concentrator is a linear 1/8" O.D. (0.085" I.D.) x 20 cm stainless steel column packed with 4 cm of 80/100 mesh Haysep-D surrounded by 5 cm of 60/80 mesh glass beads and I cm of glass wool on either side. The column is encased in a 12 cm x 6 mm I.D. glass tube, fitted with two 1/4" O.D. side-arms, as shown in Figure 14.10. A 1 cm thick insulating layer of open-cell foam covers the glass tube. The column is centered within the glass tube by a pair of 1/2" to 1/4" Cajon UltraTorr reducing unions
Stable isotope measurements of atmospheric CO2 and CH4
295
Figure 14.10- Methane pre-concentration device. CH4 is trapped at-120~ and bulk "air" is vented, after which CH4 is released by heating to 0~ Cooling by liquid nitrogen and heating by NiCr wire are controlled by a temperature controller.
through which the column extends. The central 10 cm of the column is wrapped with fiberglass insulated NiCr heating wire (0.23 mm diameter, Omega). The wire is wrapped over a narrow gauge K-type (alumel/chromel, Omega) thermocouple positioned about 2 cm from the center of the column, just beside the liquid N2 outlet (Figure 14.10). The column is fitted to the six-port, two-position valve with 1 / 16" stainless steel tubing and 1 / 16" to 1 /8" reducing unions fitted with 10 mm screens (Valco) and sealed with Teflon ferrules. The column is maintained at-120~ by opening and closing a solenoid valve on a pressurized liquid N2 tank that is plumbed to the inlet of the jacket surrounding the pre-concentration column. The valve is controlled by the central computer, which monitors the thermocouple at a frequency of about 5 Hz. Cold N2, mainly in the vapor phase, enters through one of the side-arms on the glass outer jacket and exhausts through the other side-arm and the gaps between the 1/8" O.D. column and the 1/4" ends of the Cajon UltraTorr fittings. Tests demonstrated that allowing liquid nitrogen to exhaust through the exit side-arm and both ends of the glass jackets provided the most uniform temperatures. The pre-concentrator is kept at-120 + 3~ for 3 minutes prior to the sample injection to ensure that the entire diameter of the column has cooled. Once the sample air has been injected onto the pre-concentrator, it is held at-120~ for 2 minutes allowing the bulk of the "air" to vent. Immediately after the cooling is stopped, the NiCr wire (total resistance - 19.7 f~) is heated to 0~ by applying a 12 V potential across the NiCr wire. This temperature is controlled via a software feedback loop from the central computer. As soon as the heating begins, the six-port valve is switched so that the ~ 30 m L / m i n (STP) of He through the pre-concentrator is replaced by a 2.0 m L / m i n (STP)
296
Chapter 14 - B.H. Vaughn, J. Miller, D.F. Ferretti & J.W.C. White
electronically controlled flow (Tylan FC-260). The low flow rate is required by the analytical column and ensures a reasonable split ratio prior to entering the mass spectrometer. We chose 0~ to minimize the amount of water vapor released by the preconcentrator on to the cryo-focus stage. After the elution of CH4, the high He flow is returned to the pre-concentrator and it is heated to 110~ for 5 minutes to purge the column of H20 and any other remaining condensables. The temperatures and timings for the pre-concentrator were determined by analyzing both the venting flow and the slow eluting flow by Flame Ionization Detection (FID). At the measured temperature of-120~ methane was retained indefinitely on the pre-column. Although the FID is not directly sensitive to air, the flow disturbance caused by its elution is evident at about 15 seconds. The additional 105 seconds was used to let the tail elute. A column heating rate of about 40~ corresponding to an application of 12 V resulted in the elution of methane at 45 seconds after the valve switch and the start of heating, with a peak width (FWHM) of about 30 seconds. Tests using an NDIR analyzer (Li-Cor 6251) indicated that CO2 co-elutes with methane in the absence of the pre-sample loop CO2/H20 trap.
14.4.2.4 Sample cryo-focusing and separation The methane eluting from the pre-concentrator is transferred to the GC through a 0.32 mm I.D. deactivated fused silica transfer capillary (SGE). There it is cryo-focused at the head of the analytical column (Molecular Sieve 5A, 0.32 mm x 25 m, Chrompack) so that its peak width can be reduced. The cryo-focusing is achieved by cooling the first 10 cm of the column to about-150~ The head of the column is encased in a section of 1/4" O.D. stainless steel tubing with a tee at one end, and a cross at the other (Swagelok). The column is held in place by custom-drilled 1 / 4" - 0.5 mm graphitized-vespel reducing ferrules. The tee is used as the inlet for liquid N2 while the cross is used as an outlet and as a port for a K-type thermocouple. The central computer controls the temperature in the identical manner as the pre-concentrator. The head of the column is cooled one minute prior to the heating of the preconcentrator to ensure that all eluting methane is trapped. It is held at-150~ for an additional 2 minutes, which corresponds to the FID - determined elution of methane from the pre-column plus one additional minute of "safety" time. The head of the column is heated by stopping the flow of liquid N2 and simply allowing the cryo-focus device to warm to the GC temperature of 80~ The column warms to 0~ within about 3 minutes, although design tests indicate methane begins to desorb from the column at about-100~ Methane and residual air from the pre-concentration step, along with air from leaks and carrier gas impurities are cryo-focused on the head of the analytical column. Some of this air passes through at-150~ but the portion that is retained must be fully separated prior to combustion and analysis in the mass spectrometer. Although the dominant choice of analytical column in similar systems has been 0.32 mm x 25 m Poraplot Q (Zeng et al., 1994; Merritt et al., 1995b; Sansone et al., 1997), we have found that the separation of CH4 from air is enhanced on Molecular Sieve 5A. At a GC oven temperature of 80~ 02 elutes at 100, N2 at 150, and CH4 at 190 seconds after the
Stable isotope measurements of atmospheric CO 2 and CH 4
297
warming of the cryo-focus region. Furthermore, the strong retention of CH4 on Molecular Sieve 5A allows for a much smaller length of column to be used in cryo-focusing. The GC effluent prior to the elution of CH4 is diverted from the source of the mass spectrometer through a change-over valve located downstream of the open-split (Figure 14.9). The wide separation ensures that when CH4 is present in the combustion furnace and the analyzer section of the mass spectrometer, no other species (other than He carrier gas) are present. The width (FWHM) of the methane peak after conversion to CO2 is five seconds as measured by the mass spectrometer. The peak height is typically about 9 nA (Figure 14.11) but can vary depending upon both the sensitivity of the mass spectrometer and the temperature of the cryo-focus unit. CO elutes at 350 seconds, but the ratio of its peak area to that of methane indicates that only a portion of the initial CO in the sample is trapped during methane pre-concentration. Although the Molecular Sieve column has excellent separating characteristics, it irreversibly adsorbs water and CO2 at room temperature. The presence of the trap
Figure 14.11 - Typical peaks of m / z = 44 (thick line) and m / z = 45 (thin line, x 100) from a standard or sample air run showing CH4-derrived CO2 chromatographic peak and the reference CO2 peak admitted from the bellows of the mass spectrometer. Time is relative to the injection of the pre-concentrated sample onto the cryo-focus region of the analytical column.
298
Chapter 14 - B.H. Vaughn, J. Miller, D.F. Ferretti & J.W.C. White
upstream of the sample loop prevents the majority of water and C O 2 from reaching the column, but the column must be baked out after every ~500 samples at greater than 200~ to remove adsorbed water and CO2.
14.4.2.5 Sample combustion After eluting from the capillary column the methane peak is transferred to the combustion furnace via a 20 cm section of 0.32 mm I.D. fused silica capillary. The combustion furnace is composed of a 3 mm O.D. x 0.5 mm I.D. x 300 mm high-density alumina tube (Alsint, Bolt Technical Ceramics) mounted co-axially within a 400 W cylindrical heater. The combustion tube is attached to transfer capillaries on either end by 1/8" - 1/16" reducing unions (Valco), and the seal is made with 1/8" graphitized-vespel ferrules and 1/16" gold-plated stainless steel ferrules (Valco). The output of the heater is controlled by an electronic temperature controller (Omega 9000A) using an R-type (Platinum and Rhodium/Platinum, Omega) thermocouple. The ceramic tube extends 6 cm beyond the edges of the heater to ensure that the fittings remain cool. Glass wool is used to plug both ends of the annulus between the combustion tube and the heater to minimize the temperature gradient within the heated zone. The combustion tube is filled with Ni and Pt wires that run the length of the furnace. The Ni wire is used as a substrate for oxygen required in combustion, and the Pt wire serves as a catalyst. In order to maximize the amount of oxygen available for combustion and the surface area available for catalysis, six 0.05 mm Ni (99.994% purity) and two 0.05 mm Pt wires (99.95 % purity) are used (Alfa Aesar, Ward Hill, Massachusetts). All wires were braided together to facilitate insertion. The furnace is maintained at 1150~ lower temperatures allow some methane to remain uncombusted. The Ni inside the furnace was initially oxidized by passing pure 02 (99.999% purity) through the furnace at 5 m L / m i n (STP) at 500~ for 4-6 hours, and then at 1150~ for 10-12 hours (Merritt et al., 1995b). However, repeated oxidation is not necessary. This is, most likel?4 because of the small amount of oxygen eluting through the column and passing into the furnace every time a sample is analyzed. The increased surface area of Ni wire, compared to that of Merritt et al. (1995b), may also provide a larger reservoir of oxygen available for combustion. This design yields a consistent amount of CO2, no CH4 and no CO, as measured by the mass spectrometer, FID, and reduction gas analyzer, respectively. Based on these tests we infer a combustion efficiency of 100%. Although water is produced in the combustion of methane, it is not removed from the He stream prior to admittance to the mass spectrometer. Normally, transient amounts of water are removed to limit the extent of the gas-phase ion-molecule reaction between CO2 and H + in the source of the mass spectrometer. In this reaction, a proton bonds to the CO2 resulting in a species of m / z - 45 that does not correspond to CO2 containing 13C. This reaction occurs in all IRMS's, but is "invisible" when its contribution is the same for both running gas and sample gas. In our case the rate of this reaction is substantially higher when our CH4-derived CO2 peak enters the source than when our pure CO2 running gas does, resulting in a systematic error to our measurements. Such systematic errors can be accounted for by calibration. However, ran-
Stable isotope measurements of atmospheric C O 2 and C H 4
299
dom variations in the H20 peak and drifts in the background concentration of H20 in the source over time do contribute to imprecision in our measurements. Fortunately, as shown below, these random errors are small.
14.4.2.6 Mass spectrometric analysis After the CH4-derived CO2 peak leaves the combustion furnace it is transferred to an open split. The split consists of a 0.11 mm I.D. capillary placed 4 cm within a 0.32 mm I.D. capillary that is bathed in He. A 1 m section of the 0.11 mm capillary leads through the change-over valve to the source region of the mass spectrometer (Micromass Optima or Micromass Isoprime), resulting in a pressure of 5 - 6 x 10-6 mbar. The split ratio is approximately 1:6. Although a larger split ratio would allow more CH4derived CO2 to be analyzed, the mass spectrometer cannot operate at pressures greater than I x 10-5 mbar. In the mass spectrometer ion source, the CH4-derived CO2 is ionized and the signals for m / z = 44, 45, and 46 are simultaneously measured. After the tail of that peak has disappeared, after about one minute, a pulse of pure CO2 "running gas" ("bonedry" quality, 99.8% CO2, < 10 ppm H20) from the bellows of the dual-inlet portion of the mass spectrometer is mixed into the He stream and admitted to the source region (Figure 14.11). The purpose of the pure CO2 running gas is to track and correct for changes in the mass spectrometer ion source that occur over periods of half an hour to hours. This square peak of CO2 is thirty seconds wide with a height of about 6 nA. The CO-derived CO2 peak elutes about 20 seconds after the end of the running gas CO2 peak. Once the baseline has returned to normal after another 60 s, the signal collection is stopped. Each aliquot of air, from either a sample flask or reference tank is measured relative to running gas, so that drifts in the source or analyzer regions of the mass spectrometer at time scales of greater than a few minutes are taken into account. Specifically, the m / z - 44, 45, and 46 peaks are integrated for both the sample and running gas, and ratios of the areas are calculated. The data analysis software measures the current at the beginning and end of the data collection period, linearly interpolates between those points, and subtracts these "zero" lines from the raw signals. The m / z = 44, 45, and 46 peaks have slightly different elution times, requiring each peak to have unique integration limits. The software makes an "isotope-shift" correction to the m / z - 45 and 46 peaks that are typically -40 ms and +20 ms, respectively. In order to correct for the contribution of 12C160170 to the m / z - 45 signal, a "Craig Correction" is made (Craig, 1957) based on the area of the m/z=46 peak. Finally, the 613C value of the sample peak is calculated relative to that of the running gas, and then converted to the V-PDB scale using the user-entered V-PDB value of the running gas.
The ~)13Cvalue of our running gas relative to V-PDB is-36.9%o as determined on a dual inlet instrument (Micromass- Optima) in our lab. However, we cannot be certain that this is the 613C value that is admitted to the source. The running gas is probably fractionated in the stainless steel capillaries between the bellows and the mass
300
Chapter 14 - B.H. Vaughn, J. Miller, D.F. Ferretti & J.W.C. White
spectrometer, and the degree of fractionation can vary with the pressure in the bellows. Other day-to-day variability may result from changing baseline conditions and their effect on zero - subtraction. The consequence of these errors is that at this point the raw delta values of both our samples and references differ from their true values by +1.0 + 0.2 %o, on average.
14.4.2.7 Reference gases and calibration In order to know the "true" value of our samples and references, our references have been externally calibrated using traditional, dual-inlet, off-line techniques. Four reference gas cylinders filled with ambient air collected at Niwot Ridge, Colorado have been calibrated by Dr. Stanley Tyler at the University of California, Irvine using a technique based on that of Stevens & Rust (1982; Tyler, 1986; Lowe et al., 1991). The 813CH4 of the reference air was measured relative to pure CO2 reference gas that had been calibrated against IAEA-NZCH (see e.g. Lowe et al., 1999). The isotopic compositions of our samples and one additional reference air tank have been determined relative to these calibrated references. Our reference air is whole air that has been dried by Mg(C104)2 and pumped into aluminum cylinders to about 150 bar at Niwot Ridge, CO. In the future, at least one of our original reference air tanks will be re-measured by the Tyler group to check for drift in the 813C value. All measurements are reported relative to V-PDB (Coplen, 1996a). 14.4.2.8 Analysis sequence Each sample flask is measured as part of a batch of eight. The analysis sequence starts with the analysis of five consecutive aliquots of reference air, of which the first is typically an outlier (greater than 2o from the mean), and always rejected. The measurement of the flask samples then begins, and each sample analysis is alternated with a reference analysis until all eight samples have been measured. The analysis sequence ends with the measurement of four consecutive aliquots of reference gas. Once the first reference measurement has been excluded, the reference measurements are averaged in three groups of five, i.e. run #'s 2, 3, 4, 5 and 7, 9, 11, 13, 15 and 17; and 19, 21, 22, 23, 24. In this way, the drift of the total system over times of about two hours is tracked. Reference gas and sample gas are alternately introduced to the system to reduce the chances of "memory" of a previous sample affecting future samples. Standard gas 813C values are linearly interpolated between the averages of groups 1, 2, and 3. Flask sample 813C values are then re-calculated relative to the interpolated standard gas values to correct for drift. Drifts of about 0.1%o are typically observed between the beginning and end of a run (about 6 hours), with the ending standard gas 813C values heavier than those at the start. One possible explanation for this drift is the accumulation of water vapor in the source region of the mass spectrometer over the course of the run. Water produced as a result of methane combustion and admitted through leaks may not be pumped away from the tubing downstream of the furnace, and the source, as fast as it is produced. From one sample/standard analysis to the next, this effect would be difficult to observe, but over the six hour period of the run, we would expect to observe some accumulation. Regardless of the cause of the drift, our frequent use of reference gas gives us confidence in the accuracy of our measurements relative to that of the externally calibrated reference air.
Stable isotope measurements of atmospheric CO 2 and CH 4
301
14.4.2.9 Quality Control 14.4.2.9.1 Flask tests In order to quantify systematic biases in the measurement of air from under-pressure flasks versus that from over-pressure tanks, we conducted systematic flask tests. Eight flasks were filled from a tank of standard gas to about 0.5 bar, which is the typical pressure in flasks when they are analyzed. The 613C values of these flasks were measured, in the manner stated above, and compared to the 613C values of the standard aliquots of the same batch analysis. Analysis was repeated twice more on these flasks to simulate three total measurements. No systematic bias was detected within the noise (lo = 0.05 %o) to which all samples and standards were subject. Additionally, the 613C values of the flasks from the first and third runs were not distinguishable, implying that we can analyze a flask at least three times without error.
14.4.2.9.2 Flask pair differences One measure of the precision of flask analyses is the difference between the ~13C values of a single flask and its mate. The mean pair difference is-0.018%o (first flask measured minus the second), and the mean of the absolute values of pair differences is 0.118%o (n=630). The distribution of pair differences is well approximated by a normal distribution centered on zero (Figure 14.12), indicating that there is no systematic bias in the order in which a pair of flasks is measured. A m o n g good pairs, defined as those pairs with a difference less than 0.2%o, the mean pair difference is-0.009%o and the mean absolute difference is 0.071%o (n=554).
14.4.2.9.3 Precision of standards We can also use the standard deviation of the standards in a batch analysis as a proxy for the precision of flask measurements. The mean standard deviation of aliquots from standards in any given run is 0.08%0 + 0.02%o (lo, n= 172) (Figure 14.13). Since all measurements are corrected for the drift of standards during a run, we also calculate the absolute difference between the measured ~13C value and the ~13C value Figure 14.12- Histogram showing the distribution of differences between a flask and its mate (1st flask2nd flask). The super-imposed gaussian has a width of sigma = 0.08%0.
302
Chapter 14 - B.H. Vaughn, J. Miller, D.F. Ferretti & J.W.C. White
F i g u r e 1 4 . 1 3 - Standard deviation of standard air aliquots during batch analyses over time. Squares represent rejected runs and circles are retained. Solid line is the long-term mean of retained runs, 0.08 %o.
of the linearly interpolated drift line, at the same point in time. The standard deviation of these differences is 0.07%o + 0.02%0. Using a 40 mL air sample, the shot-noise limited precision of our measurement is ~ 0.02%o, so we are within a factor of four of this limit.
14.4.2.9.4 Sample size v.
~13C
relationship--'linearity"
The relationship between sample size and 813C value was checked by making repeated measurements from a single standard tank using 40 mL and 25 mL sample loops. Although peak area as measured by the mass spectrometer varied in proportion to sample loop size, the 813C value was constant to within typical experimental uncertainty of ~0.05%o. Given that the mole fraction of methane in sample flasks varies by a m a x i m u m of +_15%, we are confident that "non-linear" effects in the chromatographic/combustion system or ion source do not compromise our measurements.
14.4.2.9.5 Internal comparison of reference tanks We have measured the 813CH4 values of our standard tanks relative to one another and compared the measured differences to the differences between tanks as originally measured at UCI. Since the ~13C values encompass a range from -47.17 to -47.27%0, we measured only the two tanks at the ends of the scale. These two tanks are also the tanks that have provided the standard gas for close to 90% of our sample measure-
Stable isotope measurements of atmospheric CO 2 and CH 4
303
ments. Treating the tank "Harpo" as the standard and the tank "Lucy" as an unknown, the (~13Cvalue of Lucy was determined to be -47.14+0.01%o (standard error of the mean, n=16), whereas the assigned 613C value of "Lucy" as determined at UCI is -47.17+0.03%o (standard error of the mean, n=2). The difference in the means is not significant (P>0.2). 14.4.2.9.6 Contamination levels We intermittently assess the level of contamination in our analysis system by injecting a sample loop filled with He instead of air. Such blank runs never yield CH4derived CO2 peak areas of greater than 0.1% of the sample peak area. As an alternative test, we inject a He filled sample loop into the system but bypass the pre-concentration device. These tests yield peak areas only 0.03% of sample peak area. Thus, the contamination that is present is mostly due to condensation of leaks and carrier gas impurities during sample pre-concentration. 14.4.2.10 Future measurements of D/H
The system described above is well suited for adaptation to make measurements of 6D in atmospheric methane. The oxidation furnace currently in line could be replaced by a furnace that would directly convert CH4 to H2 (Burgoyne & Hayes, 1998; Hilkert et al., 1999). The hydrogen isotopic ratio could then be analyzed by an isotope ratio mass spectrometer appropriately tuned. The other change that would have to be made would be to increase the size of the sample loop to account for the lower relative abundance of D compared to 13C and the lower ionization efficiency of H2 relative to CO2. Assuming a CH4 to H2 conversion efficiency of near 100% and using a 100 mL sample loop, precision close to 1%o should be attainable. 14.4.4 Measurements of 13C/12C in atmospheric CH4 summary A method for high-precision, low volume, automated and relatively fast measurements has allowed us to analyze air samples for 613C of methane on a weekly basis from six sites around the globe. This was made possible only by employing the GCIRMS method and the NOAA global flask sampling network. Note that all 613C of methane data, including those flagged for sampling and analytical problems are available at: ftp: / / ftp.cmdl.noaa.gov /ccg / ch4c13 / and that these data have been analyzed by Miller et al. (2002).
Our biggest remaining experimental challenge is the comparison of our reference scale with those of other laboratories measuring 613C of methane. A program of intercomparison of both standards and simultaneously collected samples will best achieve this aim. We have just begun such an effort with the University of California, Irvine (S. Tyler) and with NIWA, (D. Lowe). Ongoing analyses of samples and standards from these and other labs are vital for the maximum scientific application of any one lab's data.
304
Acknowledgments
Chapter 14 - B.H. Vaughn, J. Miller, D.F. Ferretti & J.W.C. White
The authors wish to acknowledge the affiliation and support of the NOAA Climate Monitoring and Diagnostics Laboratory, Carbon Cycle Group in Boulder Colorado, without whom, this development would not have occurred. We also wish to acknowledge the painstaking efforts of our anonymous reviewers, whose attention to detail added greatly to the manuscript.
Handbook of Stable Isotope Analytical Techniques, Volume 1 P.A. de Groot (Editor) 9 2004 Elsevier B.V. All fights reserved.
C H A P T E R 15 Preparation and A n a l y s i s of N i t r o g e n - b e a r i n g C o m p o u n d s in Water for Stable I s o t o p e Ratio M e a s u r e m e n t Cecily C.Y. Chang1*, Steven R. Silva1, Carol Kendall1, Greg Michalski2, Karen L. Casciotti3 & Scott Wankell 1 Water Resources Division, U.S. Geological Survey, 345 Middlefield Rd, MS 434, Menlo Park, CA 94025, USA 2 University of California, Department of Chemistry and Biochemistry, San Diego, CA 92039, USA 3 Water Resources Division, U. S. Geological Survey, 12201 Sunrise Valley Drive, Reston, VA 20192, USA
e-mail: *
[email protected]
15.1 Introduction Recent concern about the potential danger to water supplies posed by agricultural chemicals has focused attention on the mobility of various solutes. In aquatic systems, nitrate and pesticides, are of particular interest (Vistousek et al., 1977; Caraco & Cole, 1999 and references therein). Nitrate concentrations in public water supplies have risen above acceptable levels in many areas of the world, largely caused by overuse of fertilizers and contamination by human and animal waste. The World Health Organization and the United States Environmental Protection Agency set a limit of 10 mg/L nitrate (as N) (Dourson et al., 1991) for drinking water because high-nitrate water poses a health risk (Ward et al., 1996), especially for children, due to methemoglobinemia (blue-baby disease) (Johnson et al., 1987). High concentrations of nitrate in rivers and lakes can cause eutrophication, often followed by fish-kills due to oxygen depletion (Rablais et al., 1999; Peterson et al., 2001). Increased atmospheric loads of anthropogenic nitric and sulfuric acids have caused many sensitive, low-alkalinity streams Table of acronyms as are used in Chapter 15. EA-IRMS: DIN: DOC: DON: IAEA: MEQ: MIFL: NIST: PTFE: SLAP: VSMOW:
Elemental Analyzer Isotope Ratio Mass Spectrometer Dissolved inorganic nitrogen Dissolved organic carbon Dissolved organic nitrogen International Atomic Energy Association Millequivalents Mass Independent Fractionation US National Institute of Standards and Technology Polytetrafluoroethylene Standard Light Antarctic Precipitate Vienna Standard Mean Ocean Water
306
Chapter 15 - C.C.Y. Chang, S.R. Silva, C. Kendall, G. Michalski, K.L. Casciotti & S. Wankel
in North America and Europe to become acidified (Vilagura et al., 2001 and references therein). Streams that are not yet chronically acidic may undergo acidic episodes in response to large rainstorms and/or spring snowmelt. These acidic "events" can seriously damage sensitive local ecosystems. Future climate changes may exacerbate the situation by affecting biogeochemical controls on the transport of water, nutrients, and other materials from land to freshwater ecosystems. The development of effective management practices to preserve water quality and remediation plans for sites that are already polluted, requires identification of nitrogen sources and an understanding of the processes affecting concentrations. In particular, a better understanding of solute sources is required to determine the potential impact of contaminants on water supplies. Determination of the relation between nitrate concentrations in ground water and surface water and the quantity of nitrate introduced from a particular source is complicated by (the occurrence of multiple possible nitrate point and non-point sources in many areas, and the co-existence of several biogeochemical processes that alter nitrogen and other chemical concentrations. Under ideal circumstances, isotopes offer a direct means of source identification because nitrates from different sources often have isotopically distinct nitrogen and oxygen isotopic compositions. In addition, biological cycling of nitrogen often changes isotopic ratios in predictable and recognizable directions that can be reconstructed from the isotopic compositions. This chapter reviews methods for analyzing "natural abundance" nitrogen-bearing compounds in water for 615N, 6180, and 6170. Among the techniques to be reviewed are Kjeldahl distillation and digestion, diffusion, micro diffusion, ion exchange, and microbial denitrifier methods. More emphasis is given to modern (since 1990) methods, often automated, used by geochemists and hydrologists for analysis of precipitation, surface water, and groundwater samples, than to the voluminous literature of Kjeldahl-type methods developed by soil scientists, initially for use in N-uptake studies. The method of choice will depend on the species of nitrogen and concentration, which in turn affects sample volume, as well as potential interfering compounds. Cost and the convenience are also factors. In short, there is no universal method for isolating ammonium (NH4+), nitrate (NO3-) or organic-N. For methods that can be used to recover more than one form of nitrogen, the order of presentation will be ammonia, nitrate and organic-N. Because there are many steps and caveats for each method, the reader should refer to the original papers before proceeding. This is not a chapter on 15N-tracer methods; for a good review of tracer methods, see Mulvaney (1993) and Knowles & Blackburn (1993). This chapter partially overlaps the contents of some sections in Chapter 9, which covers the preparation of ecological and biogeochemical samples for isotope analysis by Teece & Fogel. Chapter 9 also covers sections on particulate organic nitrogen and biological samples. Note that there are several other materials that can also be analyzed for isotopic composition to provide useful information about N sources and cycling, including: dissolved N2, N20 (Vogel et al., 1981; B6hlke & Denver, 1995), dissolved 02 (Aggar-
Preparation and Analysis of Nitrogen-bearing Compounds in Water for Stable Isotope Ratio ...
307
wal & Dillon, 1998; R6v6sz et al., 1999a, b; Wassenaar & Koehler, 1999), and solid organic samples. A multi-isotope approach using the C, O, and S isotopic compositions of associated organic and inorganic components will also aid interpretation of h15N and its variations in the ecosystem. The reader should keep in mind that isotopic compositions generally cannot be interpreted successfully in the absence of other chemical and hydrologic data. In particular, since redox reactions have such a profound impact on isotope fractionations, information on redox chemistry and dissolved oxygen concentrations are especially useful. 15.2 Fundamentals
Nitrogen isotopes Nitrogen comprises 78% of the atmosphere, oxygen 21% and argon 1% (Schlesinger, 1997). Estimates of nitrogen abundance in the earth's crust are more variable and often more difficult to compare than estimates of atmospheric abundance, in part because different methods, units, and portions of the crust have been used to calculate abundance (Holloway & Dahgren, 2002). For instance, of the 97.88% partitioned in to the terrestrial pool (2.007% in atmosphere), 99.775% has been estimated to be in rocks (S6derlund & Svensson, 1976; Winteringham, 1980). In contrast, Schlesinger (1997) estimates that 20% of the global nitrogen is in rocks. Bulk rock nitrogen concentration has been estimated to be 1.27 + 1 ppm (All6gre et al., 2001), 60 ppm in continental crust, 83 in upper crust, and 34 in the lower crust (Wedepohl, 1995). There are two stable isotopes of N: 14N and 15N and a wide range of oxidation numbers exhibited by nitrogen compounds, ranging from +5 (NO3-) to -3 (NH4+). The average abundance of 15N in air is very constant (Junk & Svec, 1958), with 15N/14N - 1/272. Nitrogen isotope ratios are generally reported in permil (%0) relative to N2 in atmospheric Air (see also Part 2, Chapter 40), using the standard definition of 6 (which is pronounced "delta", not "del')" ~15NAir = {[(15N/14N)x / (15N/14N)AIR]-1}. 1000
[15.1]
where x - sample. Analytical precisions of 0.1%o or better are common. To improve interlaboratory comparisons, 615N values should be normalized to the compositions of reference materials with widely different 615N values (B6hlke & Coplen, 1995). For example, 615N values can be normalized to the values of the IAEA ammonium sulfate reference materials N-1 and N-2, which we found to have compositions of +0.45%0 and +20.35%0, respectively (Kendall & Grim, 1990). Some additional reference materials are available for interlaboratory comparisons (B6hlke et al., 1993). Different sources of nitrate span a range of 615N of values, and as such, can be used to differentiate between sources (Figure 15.1) ( Komor & Magner, 1996; Harrington et al., 1998 and many others). However, due to the overlap in values between some sources, using 6180 values in conjunction with 615N provides better source separation (Arevena et al., 1993; Wassenaar, 1995; Mayer et al., 2002 and many others). In some
308
Chapter 15 - C.C.Y. Chang, S.R. Silva, C. Kendall, G. Michalski, K.L. Casciotti & S. Wankel
cases, it is possible to discern nitrate that has undergone denitrification. This can be seen when the remaining 615N and 6180 of nitrate becomes heavier or "enriched" in a ratio of I : 2 (oxygen to nitrogen) (Amberger & Schmidt, 1987; B6ttcher et al., 1990; Voerkelius, 1990), as the lighter ~15N and 6180 is preferentially denitrified by microbes, leaving the enriched nitrate behind.
Oxygen isotopes Oxygen has three stable isotopes:
160, 170, and 180; the relative abun-
dances are 99.76%, 0.04%, and 0.20%, respectively. Stable oxygen isotopic c o m p o s i t i o n s (6180 and ~170) are given in terms of 1 8 0 / 1 6 0 and 1 7 0 / 160 ratios, respectively, using the 6 definition given above. Precisions of 0.1%o and better are common. The 6180 values of nitrate are reported in %0 relative to the standard V-SMOW (Vienna Standard Mean Ocean Figure 15.1 - Ranges of 615N and 6180 values of difWater). There is almost an 80%0 range ferent sources of NO3 (from Kendall, 1998). in 6180 values, corresponding to a 30%0 range in 615N values (Kendall, 1998) for nitrate. Most of the spread in 6180 values is caused by the high values of rain and snow samples. The 6180 of nitrate is a promising new tool for determining sources and cycling of nitrate. Although several techniques have been developed since the 1980s for the analysis of nitrate for 6180 (Amberger & Schmidt, 1987; R6v6sz et al., 1997; Silva et al., 2000), there have been fewer applications of 61sO of nitrate, probably because the methods are labor-intensive may and the involve hazardous materials. Analysis of nitrate for ~1170 is in its infancy; the first published values were reported in Michalski & Thiemens (2000). Until a few years ago, there was thought to be little reason to analyze any terrestrial oxygen-bearing materials for ~170 because there was a defined physical chemical relation between the 8180 and ~170 as a result of mass-dependent fractionations. The relation produces twice the change in the 180/ 160 ratio than the associated 1 7 0 / 1 6 0 ratio because of the difference in relative mass difference of two (18 vs 16). Since the discovery two decades ago that certain molecular systems undergo mass independent fractionation (MIF) (Thiemens & Heidenreich, 1983), investigations into these chemical processes and their applications have dramatically increased and in fact, produced a new generation of research projects. In
Preparation and Analysis of Nitrogen-bearing Compounds in Water for Stable Isotope Ratio ...
309
classical isotope fractionations described by Urey (1947), partitioning of the minor isotope between products and reactants is determined by their isotopic reduced partition functions and their reliance upon the atomic mass differences (i.e., 170 - 0.5,180). In contrast, MIF results from molecular isotopic symmetry and hence the alteration of isotope ratio has no relation to the mass differences (see also Chapter 18 for a discussion on MIF). The mass independent 8170 values of atmospheric ozone (see Chapter 18) are believed to be the result of the allowed couplings in molecular vibrations, where quantum mechanical selection rules allow asymmetric molecules to couple more readily than symmetric molecule (Hathorn & Marcus, 1999). In this instance the ultimate isotopic composition of a stable product derives from symmetry factors, not mass. This symmetry effect should not be confused with the symmetry number, which is a statistical normalization of multiple reaction pathways and which does not alter isotope ratios. In mass dependent fractionations (thermodynamic, kinetic, equilibrium), the dominating factor is the difference in isotopically substituted molecular vibrational frequency. For elements with three stable isotopes, these frequency differences are related to their mass by a proportionality constant z
z :
[15.2]
where M1, M2, and M3 are the major and two minor isotopes respectively. For oxygen isotopes, this relationship can be converted into 6 notation to give 8 1 7 0 = 0.52 8 1 8 0 . In a 3-isotope plot, this N 0.5 slope line defines the terrestrial fractionation line that has been experimentally verified in numerous biogeochemical systems (Figure 15.2). In mass independent fractionations 8 1 7 0 ~ 0.52 8 1 8 0 and are quantified by A 1 7 0 o r "cap delta 17"" A 1 7 0 - 8 1 7 0 - (0.52 8 1 8 0 ) . Experimental studies of pure MIF typically result in a line with a slope of 1 connecting products and reactants in 3-isotope space. In nature, MIF species typically plot on a ~ 0.5 slope line offset from the terrestrial fractionation line, which is the result of some MIF followed by various mass dependant fractionations (Thiemens et al., 2001). The most studied MIF system is the A 1 7 0 signal generated during the formation of ozone. The discovery of a large A170 in atmospheric nitrate should not have been surprising, for several reasons: The intimate cycling between NOx and ozone in the atmosphere, the end product of this cycle is HNO3, and because the observed 8180 values for ozone and atmospheric nitrate are high. Only recently has an experimental method been developed to measure 8170 (A170) in nitrate (Michalski et al., 2002). A NaNO3 standard (USGS-35) was standardized and found to have a A 1 7 0 = +21.56 + 0.11%o and a 8 1 8 0 = +54.1 + 1.5 %o (Michalski et al., 2002; B6hlke et al., 2003).
310
Chapter 15 - C.C.Y. Chang, S.R. Silva, C. Kendall, G. Michalski, K.L. Casciotti & S. Wankel
Figure 15.2 - Schematic of relationship between 5180 and 8170 values. Modified from Michalski et al. (2002).
To facilitate comparison of data between labs and methods, B6hlke et al. (2003) prepared large quantities of two nitrate salts with contrasting 6160, ~170, and 5180 compositions for distribution as reference material (Table 15.1). USGS34 (KNO3) has low 5180 and USGS35 (NaNO3) with high 5180 and high mass independent 5170. These are reproducible within + 0.2 - 0.3 %o with respect to VSMOW (Vienna Standard Mean Ocean Water) and SLAP (Standard Light Antarctic Precipitate) for the two reference materials as well as IAEA-N3 (NO3-). The reference materials are available in small aliquots (0.5 - 1.0 g for USGS34 and 1.0 - 1.5 g for USGS35) from the International Atomic Energy Agency, (IAEA)I and US National Institute of Standards and Technology (NIST)2.
Mass-spectrometric measurements and standardization Operating procedures for mass spectrometric measurements are discussed in detail by Gonfiantini (1981) and Coplen (2000). The special concerns for analyzing organic samples are discussed in detail in Barrie & Prosser (1996). Several corrections must be applied to raw 5 values measured against a working standard gas" 1) zero1. IAEA, Isotope Hydrology Laboratory, Wagramerstrasse 5, PO BOX 100, A-1400 Vienna, Austria. 2. NIST Standard Reference Materials Prgram, Room 204, Building 202, Gaithersburg, MA 20899-6689, USA.
Preparation and Analysis of Nitrogen-bearing Compounds in Water for Stable Isotope Ratio ...
enrichment factors, 2) instrumental corrections (valve mixing, residual, and tail contributions), and 3) the 170-correction for ~180 of CO or CO2, a.k.a, the "Craig correction" (Craig, 1957). The zero-enrichment cancels out. Gases prepared from samples and reference standards under the same conditions are measured against a common working standard gas. The instrumental corrections are very small with modern mass spectrometers. Some uncertainty still exists about the appropriate 170-correction (Santrock et al., 1985; Verkouteren et al., 1995). Many of these corrections are included in the data-reduction software of modern mass spectrometers, but the users should be familiar with the exact procedure of these corrections, since there may be optional choices.
311
Table 15.1 - Delta values for reference materials reported by B6hlke et al. (2003). Values are reproducible + 0 . 2 - 0.3 %o, lo, with respect to VSMOW and SLAP (Standard Light Antarctic Precipitate). Values for 6170 are mass independent. Standard ID
60
Value in %0
IAEA-N3 (NO3-)
6180 6170
+25.6 +13.2
USGS32 (NO3-)
6180
+25.7
USGS34 (KNO3)
6180 6170
-27.9 -14.8
USGS35 (NaNO3)
6180 6170
+57.5 +51.5
A fundamental principle in the calibration of stable isotope data of samples is the parallel preparation and mass-spectrometric measurements of samples and standards under the same conditions. This is because isotopic fractionation may occur during the preparation of samples to produce the appropriate analyte gases (N2, CO2, CO, 02, etc.). Even with nominally quantitative decomposition methods (like EA-IRMS methods), it is likely that samples are slightly fractionated, depending on the method, procedure, preparation equipment, and operators. By preparing samples and standards under the same conditions, and conducting mass-spectrometric measurements in the same session, all potential isotopic fractionations involved during the entire course of preparation and measurements can be collectively cancelled out. Isotope effects and fractionation caused by slow, systematic changes in preparation systems (e.g., temperature) and a mass spectrometer can be identified and corrected by measuring separate aliquots of the same laboratory and/or international standard several times during the analysis of a set of samples. Several solid materials are distributed by IAEA and NIST for inter-laboratory calibration of the stable isotopic compositions of N- and O-bearing solid materials. The amount of these standards available for each isotope laboratory is limited. Thus, laboratory standard material should be prepared and calibrated against these standards for routine measurements of a large number of samples. The reason for using two standards with widely different isotopic compositions for the calibration of sample is to "normalize" the %o-scale - the necessary process of stretching or shrinking the isotope ratio scale associated with preparation procedures and mass spectrometry in each laboratory, so that isotopic data from different laboratories can be directly compared.
312
Chapter 15 - C.C.Y. Chang, S.R. Silva, C. Kendall, G. Michalski, K.L. Casciotti & S. Wankel
For natural abundance determination using an EA-IRMS system where the ion source is nonlinear, it is essential to ensure that all samples and standards within a run contain the same amount of nitrogen, otherwise the precision of the results will be unacceptable. For this reason, some analysts perform a preliminary run using small aliquots of each sample and standards of different sizes. The peak sizes of the samples are compared against the standards to determine the amount of each sample needed.
15.3 Collection, transportation, and storage of samples Sample collection and storage is equally important as laboratory analytical procedure. Adequate assessment of the temporal and spatial variability in potential endmembers is essential. For instance, sampling during fertilizer or manure application should be avoided because of rapid reactions and consequent fractionations in the soil soon after application (i.e., ammonia volatilization and nitrification). Also, the nitrate615N values of materials applied at the surface (e.g., fertilizer, manure, treated waste) are best sampled beneath the application sites, after the N-bearing materials have been nitrified during downward transport through the soil zone. Great care must be exercised in the collection, transportation, extraction, and storage of water samples prior to their isotopic analysis. First, it is critical that the samples represent the feature of interest and that the study objectives are met. For example, groundwater wells need to be pumped for a time sufficient to remove the stagnant water in the well before sampling, and samples from deep rivers should be depthintegrated or sampled at weirs or other convergence points. When collecting stream samples, it is necessary to evaluate whether depth integrated samples are required. In some instances, lagrangian samples may be needed. There are several methods available for collecting dissolved inorganic and organic nitrogen (DIN and DON) species from natural waters and preparing them for 615N analysis. Since these species are biologically labile, samples should be filtered immediately after collection, using 0.45~m or finer filters, (Patton 1995) and chilled at 4~ If more than a few days elapse before samples are analyzed, samples should be frozen. This is especially recommended for NH4 + samples and filtering with 0.2~m filters is common. Use of silver filters for DON samples can aid preservation efforts. Water samples are commonly preserved with sulfuric acid, but not if they are to be analyzed for DON. This is because the humic substances will precipitate in acidic solutions. Fewer labs are using mercuric chloride because it is a hazardous material, and because it precludes the use of microbial methods (Sigman et al., 2001; Cassiotti et al., 2002). Mercuric chloride also poses problems during steam distillation of ammonium (Heaton & Collett, 1985). In any case, it is advisable to follow the lab's preferred method of preservation. Several designs are available for rainfall and throughfall collectors. Typically, these consist of a funnel suspended above a sample reservoir. Although bird droppings on foliage can be expected to contribute to the N present in throughfall samples, a bigger problem is birds perching on the rim of rainfall collectors and contaminating the contents directly. This results in the 'background' nitrogen of interest being effectively
Preparation and Analysis of Nitrogen-bearing Compounds in Water for Stable Isotope Ratio ...
313
masked, as well as drastically increasing the bacterial load in the sample and hence the danger of microbial transformations taking place. Birds can be deterred from perching by adding soft plastic spikes to the rim of the collector; however, it may be advisable to check for bird contamination by colorimetrically testing the samples for phosphorus content and discarding any heavily contaminated samples. Garten (1992) used ion exchange resins to remove the ions of interest from rainwater during the collection period in the field, thus by-passing the problems associated with chemical and biological alteration of samples left in the field for considerable periods of time. This also avoids the need to carry bulky and heavy samples between remote collection sites and the laboratory. It should be noted however, that this approach cannot be used at temperatures below freezing. For large-volume samples of hydrological, biological and agricultural samples, the use of glass or high-density polyethylene bottles with tight caps is recommended. The bottle size should closely match the size of the sample (i.e., small dead volume). Caps with conical inserts or Teflon liners are the most reliable. Glass bottles should not be filled entirely to the top if there is a chance of freezing or of large pressure changes (such as might be expected in under-pressured airplane luggage compartments) during transportation. A convenient and reliable way to transport large numbers of bottles is to put them back into the original cardboard trays, wrap the trays in bubble paper, put the trays in insulated ice chests or coolers (with ice), and pack the bottles securely with plastic peanuts.
15.4 Methods for concentrating and/or removing nitrogen-bearing compounds from waters for 615N analysis Introduction
There are several nitrogen-bearing compounds that occur in natural waters. Almost all methods for analyzing N-bearing materials for 615N involve conversion of the material to N2 gas. This gas is preferred over alternatives (i.e., NO, N20, NO2, or NH3) because it is chemically inert, has a low molecular weight, can quickly be pumped from the mass spectrometer (i.e., has minimal memory effect), can easily be generated from a variety of organic and inorganic compounds, and interpretation of the analyses is simpler because interference from other elements is minimal (Hauck & Bremner, 1976; Mulvaney, 1993). These methods usually consist of three steps: removal of the N-bearing species of interest from the sample, concentration of the removed species, and conversion of the species to N2. This latter step is often automated so that the resulting gases are introduced into the mass spectrometer under computer control. Regardless of the species of interest, a recovery of 100% is necessary to avoid fractionation. For instance, in distillation, the light 615N will be transferred first and when using exchange columns, the heavy material will exit first. Incomplete recovery will produce a sample that is too light or too heavy for distillation and exchange columns respectively. Several approaches were considered when organizing this chapter. To describe methods based on: 1) the species of nitrogen (ammonia, nitrate, or organic-N), 2) type
314
Chapter 15 - C.C.Y. Chang, S.R. Silva, C. Kendall,G. Michalski, K.L. Casciotti& S. Wankel
of method, i.e. distillation, ion exchange, 3) the concentration of N, and 4) historical development. None of these approaches would avoid repetition and overlap since some methods can be used for more than one species. This chapter is organized by species/method/historical development. That is, we show how the methods and refinements evolved. Typically methods were first developed for solutions with highN concentrations and later refined for low concentration samples. We present the ammonium methods first, followed by nitrate and dissolved organic nitrogen. This order is followed when a method has multiple applications as well. We hope that this structure will be useful for individuals with little experience in stable isotope methodology, as well as those possessing more some background. We describe here various well-established techniques: evaporation or freeze-drying, Kjeldahl distillation, micro-diffusion, and ion exchange, and several newer methods for isolating, concentrating, and analyzing the ~15N of dissolved organic matter, ammonium, and nitrate from natural waters. Analysis of 6180 and ~170 of nitrate is also described for techniques that allow it. Finally, an appendix section provides details of some of the procedures described only briefly in the text.
15.4.1 Background information about ammonium-815N In the late 1990's several papers describing methods to recover and analyze water samples for 615N-NH4 + were published. Each method was developed to accommodate a particular type of sample (e.g. freshwater samples with low or high inorganic nitrogen concentrations). These methods were refinements of existing methods developed for use with soil extracts, in which NH4 + concentrations typically exceeded 50 M (Bremner & Keeney, 1965; Adamsen & Reader, 1983; Brooks et al., 1989; Kelley et al., 1991; Sorensen & Jensen, 1991; Liu & Mulvaney, 1992; Lory & Russelle, 1994). Some of the methods described for recovery of 615N-NH4 + can be used to recover 615N-NO3- and organic-N. There are three basic approaches to isolating 615N-NH4+: distillation, diffusion, and ion exchange. In distillation and diffusion, ammonium (NH4 § is converted to gaseous ammonia (NH3) by raising the pH of the solution to ~ 9.5 by the addition of NaOH or MgO. If g15N-NO3- is desired, Devarda's alloyl is used to reduce the nitrate to NH4 +. The sample is distilled or diffused and the ammonium gas is collected on an acidic trap, i.e. a small filter, boric acid indicator solution (see section on Kjeldahl distillation below for details), or zeolite. To sequester the nitrogen in organic-N, the sample must first be digested to convert the organic nitrogen to NH4 +. This is accomplished by the addition of sulfuric acid (H2SO4) and heating of the sample to convert organic nitrogen to ammonium (NH4+). For details describing sample digestion protocol, the reader should refer to the section following the description of inorganic sequestration. Though sequential sequestration of NH4 + followed by NO3- for 615N analysis is often done for tracer studies in which 15N is added, this is not recommended for 615N at the natural abun1. Devarda's alloy is an alloy of aluminum (45%), zinc (5%), and copper (50%).
Preparation and Analysis of Nitrogen-bearing Compounds in Water for Stable Isotope Ratio ...
315
dance levels, particularly if DON concentrations are high (Khan et a1.1998). Selection of the appropriate method depends on several factors, including the number of samples to analyze, the time required for each of the methods, (hours to days), and the concentrations of NH4 + and DON in the sample (the latter of which determine sample volume).
15.4.2 Kjeldahl distillation and digestion Kjeldahl distillation is a multi-step process involving specialized apparatus that provides a means of determining the N-content of inorganic and organic compounds (Table 15.2). Kjeldahl techniques were developed mainly for soil extracts that contain relatively high N concentrations compared to natural abundance water samples that require larger sample volumes. It has been used by soil scientists to determine the concentration of inorganic nitrogen in soil extracts in which the NH4 + concentrations typically exceed 50 ~M. This approach was necessary because the colorimetric methods for inorganic nitrogen are subject to interference by colored, turbid, and high DON extracts. For instance, organic interferences can hamper color development for the Berthelot reaction, used for NH4 + analysis (Rhine et al., 1998). Because DIN concentrations in soils are typically > 100 ~M, much smaller volumes (10 to 20 mLs) are required than for stream samples with lower DIN concentrations samples. However, the high DON concentrations can be problematic. This is because heating and extended distillation times may convert some DON to ammonium through hydrolysis, thus contaminating the sample (Mulvaney & Khan, 1999). Conversely, distillation methods are often unsuitable for natural waters because the low Table 15.2 - Steam distillation methods for determination of inorganic forms of N in soil extracts (from Mulvaney, 1996). Form of N
Method*
NH4 +
Steam distillation with MgO
NO3-
Steam distillation with MgO and Devarda's alloy after destruction of NO2- with sulfamic acid and removal of NH4 + by steam distillation with MgO**
NH4 + + NO3-
Steam distillation with MgO and Devarda's alloy after destruction of NO2- with sulfamic acid**
NH4 + + NO2-
Steam distillation with MgO and Devarda's alloy after destruction of NO2- with sulfamic acid and removal of NH4+ by steam distillation with MgO
NH4 ++ NO3- + NO2-
Steam distillation with MgO and Devarda's alloy
*
With each method, NH3 liberated by steam distillation is collected in H3BO3- indicator solution and determined by titration with 0.0025 M H2SO4. ** If NO2- is absent, the sulfamic acid treatment is omitted.
316
Chapter 15 - C.C.Y. Chang, S.R. Silva, C. Kendall, G. Michalski, K.L. Casciotti & S. Wankel
DIN concentrations require larger volumes of water that cannot be easily distilled at one time (~ 500 mL). Distillation was adapted to sequester 15N at the natural abundance level (Bremner & Keeny, 1965) and later for the isolation of ammonium in tracer studies (Miyazaki et al., 1973; Harrison, 1978; Glibert et al., 1982; Lipschultz et al., 1986). The predecessor of all natural abundance methods for inorganic 615N at oceanic F i g u r e 1 5 . 3 - Steam distillation apparatus. Available from O'Brien's Sciconcentrations is the entific Glass Blowing, 725A West Bridge Street, Monticello, IL 61856, distillation procedure USA (from Mulvaney, 1996; Fig 38-1). of Cline & Kaplan (1975). To overcome the problem of low nitrate concentrations in seawater, these authors roto-evaporated the samples prior to reduction with Devarda's alloy and distillation. Later, Velinsky et al. (1989) used 1.5 L custom made distillation flasks to perform serial distillations on successive 500 mL aliquots for samples containing < 5 gM NH4 +, typical of seawater and estuarine waters. There are many configurations of Kjeldahl digestion and distillation equipment, and commercial units are now readily available. Distillation units designed to fit Kjeldahl digestion flasks allow the digested samples to be immediately distilled (Figure 15.3). Macro Kjeldahl digestion uses 300 to 800 mL flasks while semi-micro or micro units use 30 to 50 mL flasks (Bremner, 1996). In cases where more N is required for analysis than can be distilled at once, samples may either be concentrated by evaporation or multiple sample aliquots can be combined. Although the digestion step is only necessary for organic N, the term Kjeldahl is often applied to the isolation and preparation of inorganic N which uses Kjeldahl apparatus.
15.4.2.1 Kjeldahl distillation of ammonium (NH4+) Methods suited for samples with high NH4 + concentrations Ammonium (NH4 +) can be quantitatively determined by the addition of MgO and steam distillation for 3 to 4 minutes to liberate gaseous ammonia (NH3). The NH3 is
Preparation and Analysis of Nitrogen-beating Compounds in Water for Stable Isotope Ratio ...
317
collected on an acidic trap usually boric acid (H3BO3) o r sulfuric acid (H2SO4). Hydrochloric acid (HC1) is unsuitable because NH4C1 is volatile at relatively low temperatures (Hauck 1982). To quantitively determine NH4 +, an indicator solution1 is added to the boric acid allowing the quantity of ammonia distilled to be easily determined by titration with a standard sulfuric acid solution (Bremner & Keeney, 1965, Bremner, 1996). After titration, an additional amount of sulfuric acid is added to ensure that all of the ammonium is retained as ammonium sulfate (NH4)2SO4 upon drying. However, any traces of residual boric acid may be eliminated by the addition and re-evaporation of methanol (Mulvaney et al., 1997). Sulfuric acid alone may be used to trap the ammonium if quantitative analysis is not necessary; however, neither a large excess nor a deficiency of acid is desirable. An excess will interfere with hypobromite oxidation and a deficiency will cause ammonia loss during evaporation and isotopic fractionation. The resulting acidified distillate may be evaporated to dryness in open air, or dried more rapidly using an ammonia-scrubbed forced air system (Hauck, 1982; Lober et al., 1987). Distillation into H3BO3 solution is best suited for samples with high NH4 + concentrations, because the boric acid solution is evaporated and the end product is solid (NH4+)2SO4. The main difference between distillation of NH4 + and organic-N, is that for organic-N" 1) the sample must first be digested with sulfuric acid to transform the N to ammonia, (see section 15.4.2.3 - Digestion) and, 2) for NH4 +, MgO is used to raise the pH to between 9 and 10 rather than a strong base. This is because MgO has a lower propensity to hydrolyze organic-N to NH4 +. For the same reason, distillation times are kept to a minimum (Bremner, 1965b, c; Keeney & Nelson, 1982; Mulvaney, 1996). Note that one of the drawbacks of distillation is that heating promotes hydrolysis of DON. Steam which bubbles through the sample, carries the NH3 to a condenser from which it drips into the receiving solution at an average rate of about 7.5 mL/min. Distillation time varies with sample volume and concentration. Most of the distillation may be accomplished within the first 5 to 10 minutes of boiling (Hauck, 1982; Mulvaney, 1996). One major supplier of distillation equipment states that up to 20 min. (150 mL at 7.5 mL/min) of distillation may be necessary (Labconco commercial literature (see below)). Velinsky et al. (1989) used a distillation time of 36 minutes to distill low concentrations of NH4 + from 500 mL samples of estuarine waters from specially made 1.5 L distillation flasks. Two variations on the standard ammonium trapping method are offered by Garten (1992) and Velinsky et al. (1989). Garten (1992) used cation exchange resin to collect NH4 + from rain samples (see section 15.4.5 - Ion exchange). The NH4 + was eluted off the columns with K2SO4, and MgO was added to the solution. The solution was distilled into 2% boric acid and a small amount (N 50 mg) of cation exchange resin to adsorb NH4 § This slurry is mixed for 3 h on a shaker, filtered out, dried, and an ali1. see Appendix 15.A1.
318
Chapter 15 - C.C.Y. Chang, S.R. Silva, C. Kendall, G. Michalski, K.L. Casciotti & S. Wankel
quot is converted to N2 by Dumas combustion (see below). Velinsky et al. (1989) developed a method suited for extracting NH4 + from estuarine waters (NH4 + usually < 5/~M) by adsorption onto zeolite (Union Carbide W-85 molecular sieve, Tarrytown, NY). After adsorbtion onto zeolite, the zeolite is filtered out, dried, and the total amount is prepared by Dumas combustion. Unfortunately, this particular product is no longer available. Moran et al. (2002) describes a procedure for processing samples in H3BO3 solution that have undergone titration. The samples are acidified by the addition of 2.5 M KHSO4 (0.1 ~tL ~tgN-1) then evaporated to dryness on a hotplate (90~ To remove the H3BO3 5 mL of anhydrous methanol is added and the methanol remaining after formation of trimethyl borate [B(OCH3)] is removed by heating to dryness at 90~ Five mL of water is added, the petri dish swirled to dissolve any (NH4)2SO4) and the sample dried. The (NH4)2SO4 is dissolved in I mL water and transferred with a 1000 ~tL pipettor to a 1.5 mL microcentrifuge tube. The sample is dried, and redissolved in 200 - 1000 ~tL of water to obtain solutions containing at least 0.2 g N L-1 (optimally 0.9 to 1.25 g N L-l) and the tube heated to ensure complete dissolution. A 100/~L aliquot is pipetted into a tin capsule, and the sample freeze-dried for analysis by ANCA.
15.4.2.2 Kjeldahl distillation of nitrate (N03-) If only nitrate is desired and nitrite (NO2-) is present, nitrite can be eliminated by adding sulfamic acid (H3NO3S) and NH4 § can be removed as described above (steam distillation with MgO). If nitrite is not present, then no sulfamic acid is required. Quanitative recovery of NO3- and NO2- will be partial and variable unless preceeded by pretreatment with reduced Fe-KMnO4 or salicylic acid-thiosulfate (Bremmer, 1996). After isolation, NO3- can be converted to NH4 + by the addition of Devarda's alloy. The Devarda's alloy may be prepared by ball-milling until it passes a 100-mesh screen and 75% passes a 300-mesh screen (Bremner & Keeney, 1965); however, Mulvaney (1996) mentions that a satisfactory product is produced by Merck (Darmstadt, Germany). Various degrees of N contamination associated with Devarda's alloy have been reported in the literature in reference to diffusion techniques (Liu & Mulvaney, 1992; Stark & Hart, 1996; Sigman et al., 1997; Goerges & Dittert, 1998; Johnston et al., 1999). Therefore, reagent N contamination should be considered, particularly for samples of low NO3- concentration.
15.4.2.3 Digestion of organic-N for Kjeldahl distillation To sequester the nitrogen in organic-N, the sample must first be digested and heated with sulfuric acid (H2SO4) to convert organic nitrogen to ammonium (NH4 +) followed by the addition of excess base to the acid digestion converts the NH4 + to gaseous ammonia (NH3). The sample can then be boiled and distilled into a receiving flask. Additional substances are usually added along with the H2SO4, especially a salt, such as potassium sulfate (K2SO4), to increase the temperature of the digest, and a cat-
Preparation and Analysis of Nitrogen-bearingCompoundsin Water for Stable Isotope Ratio ...
319
alyst (Hg, Cu, or Se) to promote oxidation of the organic matter. Following digestion, the sample is transferred to a distillation unit and the pH is raised with a strong base to convert NH4 + to gaseous ammonia (NH3). The NH3 is then steam distilled from the sample and collected in an acidic receiving solution which converts the NH3 back to NH4 + (Bremner, 1965c,d, 1996; Bremner & Edwards, 1965; Stevenson, 1996). The extracted NH4 + is measured and/or dried to a salt which is subsequently converted to N2 for isotopic analysis by either hypobromite oxidation, Dumas combustion (Bremner, 1965b; Hauck, 1982; Shearer & Kohl, 1993; Kendall & Grimm, 1990), or automated on-line combustion during analysis by elemental analyzer-isotope ratio mass spectrometry (EA-IRMS). Digestion is accomplished mainly by addition of a few mL of concentrated H2SO4 (18 or 9N depending on the method used) (Bremner, 1996; Mulvaney, personal communication). The exact amount used may be adjusted according to sample type and size (Bremner & Mulvaney, 1982). The digestion rate can be increased by additions of a salt, commonly K2SO4 (also Na2SO4), which raises the boiling point of the H2SO4. However, there are trade-offs in adding K2SO4. Loss of H2SO4 can occur during digestion, which increases the salt concentration and elevates the temperature. The higher the concentration of K2SO4, the shorter the digestion time. Furthermore at high concentrations of K2SO4, the sample solidifies upon cooling (0.8 g / m L of H2SO4), and at still higher concentrations (1.3 to 1.4 g / m L of H2SO4) as temperatures approach 400~ volatile N compounds can be lost. Such temperatures are possible when K2SO4 is present at concentrations of 1.3 to 1.4 g / m L of H2SO4 (Bremner, 1996). Selenium, mercury, and copper have all been used as digestion catalysts. Mercury is considered most effective but forms a complex with ammonium. Mercury is precipitated by addition of sodium sulfide or thiosulfate after digestion and before distillation (Bremner, 1996). Use of mercury is falling into disfavor, due to concerns over health and waste issues. Selenium, copper sulfate + titanium oxide or commercially available catalyst mixtures are becoming increasingly popular1. For certain compounds, particularly those with N-N and N-O bonds (e.g. azo, nitroso, and nitro compounds, hydrazines, hydrazones, oximes, pyrazolones, isooxazoles, 1,2-diazines, 1,2,3-triazines, nitrites, nitrates), standard Kjeldahl procedure is not effective (Bremner & Mulvaney, 1982; Stevenson, 1982, 1996). Many pretreatment procedures have been developed to include these substances and also to include nitrate and nitrite for total N analysis (Bremner & Mulvaney, 1982; Stevenson, 1982, 1996); however, in many cases, Dumas combustion and EA-IRMS offer a simpler means of total organic N analysis. Though some refractory forms of organic-N, such as nicotinic acid, are not quantitatively recovered by Kjeldahl digestion, these compounds are typically not found in natural samples (Mulvaney, personnel communication).
1. Labconco commercial literature, http: / / www.labconco.com/ pdf / kjeldahl guide_kjeldahl.pdf
320
Chapter 15 - C.C.Y. Chang, S.R. Silva, C. Kendall, G. Michalski, K.L. Casciotti & S. Wankel
Mulvaney & Khan (2001) report that conventional steam distillation followed by diffusion with NaOH is an effective method for determining total hydrolyzable N. However, they recommend that hydrolyzable NH4-N can be more effectively recovered by diffusion with MgO and (NH4 + amino sugar) -N can be recovered by diffusion with NaOH after which amino acid-N is liberated by ninhydrin oxidation at pH < 1.8 and recovered by diffusion with NaOH.
15.4.2.4 Contamination and fractionation problems associated with distillation Difficulties with the reduction/distillation method have been described by several investigators (Hauck, 1982; Heaton & Collett, 1985; Mulvaney, 1993, 1986). Labile organic materials may be hydrolyzed to ammonium by reagents used to raise the solution pH. It is generally accepted that the use of MgO minimizes this effect. Furthermore, it is recommended that distillation time be kept to a minimum so as to avoid cross contamination from organic-N. Reagents, in particular Devarda's alloy, should be tested for N contamination (see above). Cross contamination during distillation of successive samples can be a serious source of error (Mulvaney, 1986). For this reason ethanol or, in the case of 15N-labeled samples, an acetic acid solution followed by ethanol is distilled through the apparatus between samples. Mulvaney (1986; see Volume II, Part 3, Table 5-2.1 for a reproduction of a table by Mulvaney) provides a comparison of cleaning procedures to avoid cross contamination, procedures which have since been updated from those original described by Keeney & Nelson (1982). Natural abundance samples are particularly prone to external contamination. For instance, laboratories in agricultural areas should be aware of the dangers of ammonium contamination through open windows. Volatile nitrogen-containing substances (e.g. ammonia or nitric acid), including some common cleaning products may also pose a significant contamination risk. In general, it is a good precaution to dry NH4 + samples in an NH4+-free environment. Isotopic fractionation may result from incomplete reduction of nitrate to ammonium due to improper preparation of Devarda's alloy, and from insufficient distillation time or leaks in the system. For natural abundance work, care must be taken to ensure that recovery is close to 100%, or fractionation is likely because of the ~ 30 %o fractionation between NH3 and NH4 +. Sample recovery typically decreases with sample size, but this can be offset by collecting at least half of the original sample volume in the receiving flask (e.g., a sample of 250 mLs would merit collecting 125 mL in the receiving flask). In summary, although regarded as time consuming, prone to fractionation, and contamination, the distillation method can be very reliable when carried out in controlled conditions by an experienced operator. It carries the extra advantage of great versatility and may be used, with only minor modifications, for a large range of sample types, including acid digests, soil extracts, and eluents prepared by the ion exchange methods described below, and natural waters.
Preparation and Analysisof Nitrogen-bearingCompoundsin Waterfor StableIsotopeRatio ...
321
15.4.3 Conversion of NH4 + to N2
Ammonium may be converted to N2 by either Rittenberg hypobromite oxidation (Rittenberg, 1948; Sprinson & Rittenberg, 1948, 1949; Hauck, 1982; Mulvaney, 1993), Dumas combustion (Fiedler & Proksch, 1972; Kendall & Grim, 1990) or by EA-IRMS. The latter two methods are currently preferred. The Dumas method produces higher yields with fewer interferences than the Rittenberg technique, and analysis by EAIRMS is by far the fastest and simplest. The Rittenberg technique uses a Y-shaped tube that holds the dry sample in one arm and Na or Li hypobromite solution in the other. The Y tube is evacuated on a vacuum line then rotated to allow the hypobromite solution to react with the sample. The reaction produces N2 from the NH4+; however, between 1.5 and 3.0% of the N forms N20. Following the reaction, the remaining reagent mixture is frozen and the N2 gas is cryogenically purified online and admitted to the mass spectrometer (Bremner, 1965c; Hauck, 1982; Mulvaney, 1993). Dumas combustion and EA-IRMS can be used as alternative methods to the entire Kjeldahl procedure for the conversion of total N to N2 or organic N to N2 if the inorganic fraction is insignificant. If organic N has been converted to NH4 + by the Kjeldahl method, Dumas combustion and EA-IRMS may be used in place of the hypobromite oxidation step. For the Dumas method, dried sample is loaded into quartz tubes with copper, copper oxide (CuO) and calcium oxide (CaO). The tube is evacuated on a vacuum line, flame sealed, and combusted in a muffle furnace at 650~ The CaO absorbs carbon dioxide and water (Kendall & Grim, 1990). The resulting pure N2 may be admitted to the mass spectrometer manually or by an automated tube cracker. EA-IRMS is one of a number of fairly recent developments that use preexisting analytical equipment as on-line preparation devices connected directly to mass spectrometers for isotopic analysis. In this case, samples (including NH4 + from Kjeldahl procedures or dried organic matter) containing just a few ~moles N are weighed and folded into tin capsules and loaded into an autosampler. The fully automated system drops the samples sequentially into the EA where they are combusted and reduced to N2, and CO2. The combustion gases are carried through the system in a stream of helium. Other gaseous combustion products such as water and SO2 are chemically removed. The N2 and CO2 are separated by a gas chromatograph (GC) column and passed to the mass spectrometer for individual analysis. 15.4.4 D i f f u s i o n of NH3
Methods suited for samples with high NH4 + concentrations Diffusion offers an alternative to distillation and has several advantages and has been used to recover 615N from soil extracts with high ammonia and DON concentrations. In distillation, there is the risk that DON will hydrolyze upon heating. Also, extended distillation times may promote carryover of DON, and memory effects.
322
Chapter 15 - C.C.Y. Chang, S.R. Silva, C. Kendall, G. Michalski, K.L. Casciotti & S. Wankel
As in distillation, MgO is added to the sample and the resulting NH3 diffuses into an acid trap, either an acid solution, usually a boric acid solution (Mulvaney et al., 1997; Mulvaney & Khan, 1999) or onto acidified disks (Brooks et al., 1989; Georges & Dittert, 1998; Khan et al., 1998). In solutions with high inorganic nitrogen concentrations (> 100 gM), only small volumes (10 to 20 mL) are required for diffusion. Brooks et al. (1989) developed a method for soil extracts that uses entirely disposable materials to avoid cross-contamination, and is suitable for automated combustion. A small volume (40 - 60 mL of soil extract is placed in a 140 mL specimen container (like a urine cup), and an acid-soaked disk (punched from glass-fiber filter paper) is suspended on a length of stainless steel wire in the container head space. The extract is made alkaline with MgO. Devarda's alloy may be added depending on the N-species of interest, and the lid quickly sealed. Containers are left for 6 days at room temperature before the filter paper is removed for analysis. Liu & Mulvaney (1992) suggested that drying of diffusion disks can be done with H2SO4, CaSO4 (Drierite), and silica gel with comparable results. Mulvaney et al. (1997) used a modified Mason jar in which to suspend the boric acid solution. Khan et al. (1997) refined this system by using a hot-plate to heat the jars to 45 - 50~ The resulting decrease in time needed to achieve full recovery of N has allowed them to increase the volume of water or soil extract processed to 100 mL. Even this large volume can be processed within an average working day. Khan et al. (1998) and Stevens et al. (2000) later modified this method so that the sample could be diffused onto an acidified disk. Of the three approaches, diffusion into the boric acid indicator solution gave better precision and accuracy than distillation into boric acid (Mulvaney & Khan, 1999), and distillation was superior to diffusion onto acidified disks (Khan et al., 1998) (Table 15.3). These methods have also been adapted for direct diffusion on soils (Khan et al., 2000), and can be analyzed either by Rittenberg Analysis, ANCA, or direct combustion (Moran et al., 2002). For more details about experiments of Mulvaney & Khan (1999) to investigate the effect of organic solutions on ammonia diffusion see the appendix. After diffusion, the procedure is similar to that of distillation. The sample can be titrated to determine the amount of NH3 sequestered, acidified with H2SO4 and the solution evaporated to dryness. Methanol is added to remove the boric acid, and the Table 15.3 - Accuracy of 615N-NH4 + values obtained from soil solutions treated with labeled inorganic nitrogen. Method
Accuracy
CV
Reference
Diffusion onto disk
no greater than 5.3% and usually 3% within 2.4% within 3.8% within 4% often 2%
<1%
Kahn et al., 1998
0.1 to 1.8% ___1.3% <1%
Kahn et al., 1997 Mulvaney et al., 1994 Mulvaney & Kahn, 1999
Diffusion into H3BO3 Distillation into H3BO3 Diffusion into H3BO3
Preparation and Analysis of Nitrogen-bearingCompoundsin Waterfor Stable Isotope Ratio ...
323
sample heated to remove the excess methanol, leaving behind solid (NH4)2SO4 (Khan et al., 1997). Methods that use boric acid to trap inorganic nitrogen are only suitable for solutions with high inorganic nitrogen concentrations, as solid (NH4)2SO4 is the final product. Sorensen & Jensen (1991) and Stark & Hart (1996) modified the trap used by Brooks et al. (1989). These approaches were successfully tried with both 15N-enriched and natural abundance samples. The procedure uses 25 mL scintillation vials with screw-on lids. Up to 17 mL of sample is added to the vial, along with MgO and, if necessary, Devarda's alloy. They sealed an acid-wetted strip of glass fiber filter inside a packet made from PTFE (Teflon) tape (the kind used as thread-sealing tape for plumbing). The tape is permeable to gaseous ammonia but is hydrophobic and the packet can be floated in the alkaline sample solution. Two layers of PTFE membrane are quickly placed over the top of the vial and an acid-wetted glass-fiber filter disk placed on top of this. A third layer of PTFE is put over this before screw cap is closed to hold and seal everything in place. The sealed bottle(s) can be placed on a horizontal shaker and agitated for the duration of the diffusion period, usually 72 h at 25~ If a shaker is not available, the diffusion period should be extended to 5 d. At the end of this period, the packets can be removed and the paper strip extracted for analysis. This method may not be suitable for large volume samples because of the small size of the vials. One problem is that water accumulates inside the Teflon packets as water vapor diffuses through the PTFE. This may occur because the ionic strength of the sample solution is lower than that of the acid trap. A remedy is to use KHSO4 instead of acid in the trap, and add KC1 or KHSO4 to the sample to increase its ionic strength (Goerges & Dittert, 1998). However, since extracts from resin columns prepared by the Downs et al. (1999) method, as well as soil extracts are at least I M in KC1, such measures may be unnecessary. As with the reduction/distillation method, Devarda's alloy and MgO are used to reduce nitrate to ammonia in individual sealed sample vessels. For a more detailed description of the Khan et al. (1998) method for diffusing samples with high NH4 + concentrations onto a disk, see the appendix. Whereas the precisions of these methods are best suited for samples with high inorganic-N concentrations these methods are mostly untried or unsuitable at natural concentrations without substantial modifications.
15.4.4.1 Diffusion of NH3 for samples with low inorganic concentrations In the late 1990's several papers describing methods to sequester inorganic nitrogen from samples with low inorganic-N concentrations were published. Many of these were designed to facilitate sequestering of nitrogen from samples with low concentrations of NH4 + and NO3-. For instance, Sigman et al. (1997) adapted the distillation method of Cline & Kaplan (1975) for measuring 815N-NO3- in waters with low nitrate concentrations ~ 5 gM in seawater. This procedure is also much less laborintensive since multiple samples can be processed simultaneously.
324
Chapter 15 - C.C.Y. Chang, S.R. Silva, C. Kendall, G. Michalski, K.L. Casciotti & S. Wankel
This method gives complete recovery of nitrate with standard deviations < 0.2%o for samples as small as 5 gM. Samples up to 500 mL can be processed. Later, Holmes et al. (1998) adapted the method of Sigman et al. (1997) for sequestering 615N-NH4 + in freshwater samples. In this case, Devarda's alloy is not added to the sample. Samples processed for 615N-NH4 + will experience fractionation, and is more pronounced when diffusion is incomplete. For this reason, standard solutions of the same volume are analyzed to correct for fractionation. Sigman et al. (1997) gave detailed descriptions of their procedure and undertook experiments to show the effects of size, time, pH, heating and the blank on the precision of 615N analysis. Briefly, MgO is added to elevate the pH to promote NH3 formation, and break down labile DON. The sample volume is reduced to 15 to 25% of initial volume by boiling or evaporation. Then the sample is diffused onto acidified disks (glass or paper), the disk is dried and put into sample boats or combustion tubes for analysis. A significant blank is associated with the Devarda's alloy, and affects samples processed for 615N-NO3 -. The correction for Devarda's alloy depends on the brand and lot number of alloy. For instance, Sigman et al. (1997) reported 615N values of -7.5 and -6 %0 for two different lots of alloy, which used on samples with 4 to 6 9M nitrate, resulted in the sample being lighter by ~ 1%0. Therefore the minimum amount of alloy needed to effectively reduce nitrate should be used. A more detailed description of this procedure is described in the appendix. The time needed to process a sample using this approach depends on the initial sample volume; more time is required for large samples to evaporate to 25% of their volume. If both 615N-NH4 + and 615N-NO3- are desired, Devarda's alloy is added to the sample after NH4 + diffusion. However, sequential diffusion from the same sample is recommended only when DON concentration is low. This is because the additional time necessary for NO3- diffusion can increase the likelihood of DON hydrolysis, particularly if heating is involved (Khan et al., 1998). Devarda's alloy also promotes DON hydrolysis (Sigman et al., 1997). Note that the rate of diffusion is related to the surface to volume ratio of the sample (Conway, 1947) and that shaking and heating hasten the diffusion process. While the amount of time to distill one sample is rapid, the diffusion technique may be more efficient when large number of samples are involved. This is because multiple samples can be diffused and because up to 3 days may be needed to prepare the sample for analysis after distillation. While decreasing the sample volume to 15 - 25% of the initial size aids nitrate reduction, reducing the volume below 12% had adverse affects. For seawater samples, when the residual volume is less than 6% of the initial sample volume, the resulting pH was only ~ 8.4, too low for Devarda's alloy effectively reduce nitrate (Sigman et al., 1997). Holmes et al. (1998) modified the Sigman et al. (1997) method to determine the i515N of ammonium from both fresh and saline water samples with ammonium concentrations as low as 0.5 gM. A large sample volume (4 L) is required because of the
Preparation and Analysis of Nitrogen-bearing Compounds in Water for Stable Isotope Ratio ...
325
low ammonia concentrations, and this resulted in incomplete diffusion and low 615N values. The fractionation appeared to be independent of ammonium concentrations. Standards with concentrations ranging from 0.5- 3.0 gM showed similar ~ 10 %0 fractionations. When the recovery was > 45%, there was an inverse linear relation (r2 = 0.97) between recovery and ~15N (observed- actual), for volumes up to 3 L (i.e., the larger the volume, the lower the 615N value). For example, the 615N of a 200 mL diffused sample was 0.2 %0 lower than the true value, whereas for a 3 L sample, the 615N was 10%o low. To correct for this fractionation standards with the same sample volume and ammonia concentrations are simultaneously diffused. After such fractionation corrections, the propagated standard error was very good for such small samples, ranging from 0.36 - 0.73 %0 ( n - 24). Other investigators reported little fractionation by diffusion (Jensen, 1991; Stark & Hart, 1996). Because the percent recovery declines faster than then total mass of N captured increases with increasing sample volume, diffusion of samples larger than 4 L is not recommended. While 4 L samples took about two weeks to diffuse, many samples can be processed at the same time. For samples with extremely low ammonia concentrations, the incubation time can be extended or the samples heated, but as previously mentioned, heating increases the chances of DON hydrolysis. An alternative is to spike the samples with of 615N- NH4 + and do standard additions (spike the sample with ammonium of known isotopic composition and back calculated the sample concentration). Measured 615N values for 24 samples had standard errors ranging from 0.36 to 0.74 %o. To analyze samples with even lower DIN concentrations, Holmes et al. (1998) offer three options: 1) Increase the incubation time or temperature to increase ammonium recovery, 2) Spike the samples with ammonium of known 615N and back calculate the sample ~15N, using a mass balance approach, or 3) Improve the mass spectrometer methods for measuring small samples. For example, methods such as cryo-focusing (Fry et al., 1996) can allow ~515N measurements on samples as small as 0.1 gmol. See the appendix for more details on the Holmes et al. (1988) method.
15.4.5 Ion exchange resins
Methods suited for samples with variable inorganic-N concentrations Ion exchange resins were first used for quantitative separation of NH4 + and NO3for stable isotope analysis to collect DIN in rain (Hoering, 1957; Moore, 1977). In the 1990s resin techniques became increasingly popular for concentrating NH4 + a n d / o r NO3- from fresh-water samples (Garten, 1992; Kendall et al., 1995a; Downs et al., 1999; Chang et al., 1999; Silva et al., 2000; Lehmann et al., 2001). The Garten (1992) and Downs et al. (1999) method are for both ammonium and nitrate. The Downs et al. (1999) and Lehmann et al. (2001) methods concentrate ammonium on cation exchange resin. The methods described by Kendall et al. (1995a), Chang et al. (1999), and Silva et al. (2000) describe different versions of a method for concentrating nitrate on anion exchange resin (see section 15.4.6.2).
326
Chapter 15 - C.C.Y. Chang, S.R. Silva, C. Kendall, G. Michalski, K.L. Casciotti & S. Wankel
The premise behind ion exchange methods is that ammonium is quantitatively adsorbed onto cation ion resins (resins with an exchangeable cation, usually H +, with a lower affinity for the resin than ammonium), and nitrate is quantitatively adsorbed onto anion resins (resins with exchangeable anions, usually C1- or OH-, with a lower affinity than nitrate). This is achieved by slowly dripping or pumping the pre-filtered water sample through plastic columns containing the resin. These methods are only suitable for low ionic strength waters because other ions can "out compete" the ion of interest for exchange sites, causing fractionation (Silva et al., 2000). It is important to know the DIN concentration and the water chemistry before deciding how much resin is needed and how much water can safely be passed through the column. It is good practice to check the column effluent of representative samples for nitrate breakthrough prior to full-scale sampling. Anion exchange methods offer a number of important advantages over previously described methods (Silva et al., 2000)" 1) Samples preserved on resins need only be refrigerated, eliminating the need for sample preservation through acidification or poisoning. 2) Ammonium and nitrate can be concentrated from more dilute waters than is practical by other methods. 3) Columns can be loaded in the field, eliminating the need to transport large volumes of water back to the lab. 4) Many samples may be processed at one time. The preparation procedure allows analysis of the nitrate for both 615N and c5180. The main disadvantages of the column method are the time required in the field, which varies greatly depending on the volume of sample needed, and the labor required during preparation as described below. The use of cation exchange resin for ammonium There are several differences between the three recent methods for concentrating and recovering NH4 +. The Garten (1992) and Downs et al. (1999) methods are similar in that the concentrated ammonium eluted from the cation columns is further purified and concentrated by steam distillation and diffusion, respectively. They differ in that the Garten (1992) method marries anion exchange procedures and to those of distillation. The Downs et al. (1999) method diffuses ammonium onto acidic disks; the resin and disks are combusted by sealed-tube combustion and EA-IRMS, respectively (see below). The Lehmann et al. (2001) method is designed for waters with low or moderate DOC concentrations; it requires no elution of the a m m o n i u m - the resins are directly combusted using EA-IRMS. More details of these methods are given below.
The Garten (1992) method was designed for collecting ammonium and nitrate from rain and throughfall samples in isolated locations by use of a cation exchange column in series with an anion column. Water collected in a bulk collector is then passed through a funnel covered with 8 mm polyethylene screen and through the exchange columns. Note that this approach cannot be used at temperatures below freezing. Cation resin (Dowex 50W-W8, 50 to 100 mesh) and anion resin (Dowex 1-8X, 50 to 100 mesh) (~ 2.5 g fresh weight) are put into tapered plastic columns, and the water allowed to pass through sequentially. The Dowex 1-8X, anion exchange resin is con-
Preparation and Analysis of Nitrogen-bearing Compounds in Water for Stable Isotope Ratio ...
327
verted from the C1- to the OH- form prior to use. The anion 1-8X resin has quaternary ammonium groups attached to a styrene and divinylbenzene copolymer. The cation 50W-W8 resin contains sulfonic acid groups attached to styrene and divinylbenzene copolymer and does not include N as part of its chemical structure; hence, it can be combusted without contaminating the sample. The exchange capacity for the cation and anion resin is 5.1 and 3.2 meq g-1 respectively. The columns are capped with glass wool plugs at the top and bottom to exclude particulate matter. The NH4 + is eluted from the cation column with 40 mLs of saturated K2SO4, MgO is added to the solution and the ammonia is steam distilled into 10 mLs of boric acid indicator solution. After distillation, a small amount of cation exchange resin (- 50 mg fresh weight) is added to the boric acid solution to adsorb NH4 +. This slurry is mixed for 3 h on a shaker, and then the resin recovered and washed with 5 mL deionized water and subsequently dried at 70~ and stored in a tightly capped glass vial. To analyze the sample, a small aliquot of the resin is combusted in a sealed-tube to yield N2 (Kendall & Grim, 1990). The NO3- is eluted from the anion column with 40 mLs of 2 M KCL followed by 10 mLs of water. The subsequent steps are identical to those described for ammonium, except that finely ground Devarda's alloy and MgO (200 mg each) are added to the eluant immediately prior to steam distillation, to convert the nitrate to ammonium. Nitrogen recovery for the Garten (1992) effluent (solution in carboys below the columns) indicated that the cation columns retained 95% of the ammonia, and the anion columns retained 97% of the nitrate. The retention of ammonia and nitrate in simulated rain solutions was even better, being 100%. The mean (+ SD) for NH4 + and NO3yields for steam distillation was 90 + 9% (n - 37) and 116 + 15% ( n - 34), respectively. Ammonia sorption from the boric acid solution onto the cation exchange resin (50 mg) was 97 + 5% ( n - 38). Interestingly, incomplete adsorption (- 30 - 70%) of the ammonium onto the cation resin in the boric acid caused only slight increases (a few 0.1%o) in the 615N of test samples. Hence, there was little evidence of isotopic fractionation associated with the entire procedure. The method of Downs et al. (1999) uses 5 mL anion (Dowex 1X-8) and cation (Dowex 50X-8) exchange columns to collect NO3- and NH4 + for ~15N analysis. The cation column is placed above the anion column. The sample flows through at a rate of 0.5 drops/sec, until 5 ~mol N (60 ~g N) is collected, but not enough to saturate the exchange capacity of the resin (2.3 meq/mL). The columns are eluted with 100 mLs of 2 N KCL at a rate of 0.5 drops/sec. The solution from the cation column, containing NH4 + is then prepared for diffusion. The NH4 + solution is converted to NH3 by increasing the pH with MgO. Ammonia diffusing out of the solution is trapped on Teflon enclosed glass fiber filter strips (see Downs et al., 1999 for details) that are acidified with 2 N H2SO4. Nitrate is reduced to ammonium using Devarda's alloy. The Teflon packets are placed into 250 mL wide-mouth polypropylene bottles and the tops sealed. When all solutions were ready for diffusion, excess MgO (- 0.2 g) is added to all containers and Devarda's alloy (~ 0.4 g) was added the NO3- solutions. The bottles
328
Chapter 15 - C.C.Y. Chang, S.R. Silva, C. Kendall, G. Michalski, K.L. Casciotti & S. Wankel
are quickly recapped and swirled. The samples are diffused for 7 days at room temperature, and the bottles swirled and inverted at least 3 times a day, or placed on a shaker. After 7 days, the packets are opened, the filter strips dried in a desiccator with an open container of concentrated H2SO4. After drying the filter strips are put in tin boats and stored in individual air-tight vials until analysis. When Downs et al. (1999) diffused solutions containing both NO3- and NH4 +, the NH4 + was diffused first, using only MgO. After 7 days, the traps were removed, replaced with fresh ones, and then Devarda's alloy was added to convert and trap the remaining NO3-. These authors compared the standard distillation method to the resin/diffusion method. They prepared samples from a fertilization experiment in which samples were enriched between 50 and 100 %o. For the diffusion samples, two sets of natural abundance standards were also prepared. The samples contained 100 9moles of N, from 4 different samples of NH4NO3 fertilizer, enriched with 15NO3- or 15NH4+. The 1 L solutions were passed through the 5 mL resin columns and eluted with 50 mL of 2 N KCL had a recovery efficiency of 102.2 _+2.75% (n = 3). For diffusion, 20 ~eq of NH4 + and 20 ~eq NO3- were diffused. The recovery was 98.4 + 1.7% and 99.0 + 2.8 % for NH4 + and NO3-, respectively. For the natural abundance standards, recoveries were consistent but slightly above 100%, due to the background N in the blank (0.16 gmol). The fertilizer samples that were concentrated on columns, then diffused or distilled had similar ~515N values, but were generally less than those of samples that had been directly distilled (see Downs et al., 1999 for details). For the natural abundance samples, the average standard error for 7 sets of triplicate and 10 sets of duplicate samples was 0.7%. The Lehmann et al. (2001) method uses 2 mL cation exchange columns (Biorad AG 50W-X8, 100-200 mesh) to concentrate ammonium. After the water sample passes through the cation exchange column, the resin is dried at 50~ and up to 25 mg of homogenized resin is put into tin capsules for analysis by EA-IRMS (see below). When preparing samples for analysis, it is critical that the sample resin is well homogenized since ammonium with low 615N values accumulates at the end of the cation exchange column. Lehmann et al. (2001) tested the effect of interfering cations (Mg +2 and Ca +2) which have a higher affinity for the resin than ammonium, and found that Mg +2 had a greater potential for inhibiting NH4 + adsorption than Ca +2, even though Mg +2 has a lower affinity for the resin than Ca +2. Although high concentrations of these cations cause incomplete adsorption of ammonium, they found that there was essentially no fractionation with yields > 25%; apparently 14NH4+ and 15NH4+ have almost identical affinities for the resin, similar to the finding of Garten (1992). To minimize the adsorption of DON onto the cation exchange resin, Lehman et al. (2001) placed an anion exchange column in front of the cation column to adsorb DON. The blank contribution from the anion exchange resin (which contains ammonium
Preparation and Analysis of Nitrogen-bearing Compounds in Water for Stable Isotope Ratio ...
329
functional groups) was negligible. This practice is consistent with the findings of Chang et al. (1999) who used anion columns to concentrate nitrate in water samples with low nitrate concentrations. These authors noted that DOC adsorbed onto anion exchange resin (AGIX and AG2X in the C1- form Biorad) could exclude or displace nitrate from exchange sites. 15.4.6 Nitrate Introduction
Nitrogen in terrestrial and aquatic ecosystems cycles among oxidized, reduced, organic, and inorganic species. It is currently of great interest because of its solubility and mobility, and because much of the nitrogen applied to crops is in the form of nitrate, ammonium, or urea. Applied in excess, ammonium and urea are quickly converted to nitrate in oxic surface environments. Because the isotopic fingerprint of atmospheric, fertilizer, soil, and manurederived nitrate, often have sufficiently distinct signatures to permit separation, nitrogen and oxygen isotope ratios of nitrate provide can be a powerful tool to investigate nitrate sources and cycling mechanisms (e.g. B6ttcher et al., 1990; Durka et al., 1994; Wassenaar, 1995; Kendall, 1998; Burns & Kendall, 2002; Campbell et al., 2002, Chang et al., 2002, Sickman et al., 2003). In the mid 90's, several new methods for recovering 18ONO3 were developed. It is now well established that the use of both N and O isotopes of NO3- can facilitate more confident identification of sources and processes than 15N alone (B6ttcher et al., 1990; Aravena et al., 1993; Kendall, 1998; Burns & Kendall, 2002, Campbell et al., 2002; Chang et al., 2002). The main difference between the various methods is the form of NO3- that the sample is converted to (KNO3, AgNO3, N20) and the way the NO3derived oxygen is converted to CO2 for analysis of 6180 by the mass spectrometer. Various methods have been developed for collection and preparation of nitrate1 for isotopic analysis. The technical hurdles include quantitative extraction of nitrate from water, isolation of nitrate from other nitrogen- and oxygen-bearing species for 15N and 180 analysis, and conversion of nitrate to suitable gases for 15N and 180 analyses. Each step must either be accomplished without isotopic fractionation or dilution, or these effects must be accounted for. The main technical challenge for nitrate-15N methods is ensuring complete isolation of the nitrate from the sample so that ~ 100% the resultant N2 is derived from ~ 100% of the nitrate in the sample. Unfortunately, it is very difficult to prevent other N-bearing species (in particular DON) from contaminating the nitrate. This problem can be overcome by removing all the non-nitrate Nbearing species from the solution, or by a procedure that isolates nitrate. The initial methods for recovering nitrate for stable isotope analysis were developed for 615N only and have been described above. Amberger & Schmidt (1987) first 1. Nitrite generally occurs in very low abundance in natural waters. For simplicity, nitrate plus nitrite will henceforth be referred to as nitrate (NO3-).
330
Chapter 15 - C.C.Y. Chang, S.R. Silva, C. Kendall,G. Michalski, K.L. Casciotti & S. Wankel
measured ~18ONO3 using KNO3 as a NO3- source and Hg(CN)2 as a combustion reagent. This method, in which samples were completely desiccated without preconcentration, was used for samples in which NO3- concentrations were > 40 ~M (Voerkelius, 1990; Durka et al., 1994). As detailed in the previous section, distillation and diffusion methods attempt to remove ammonium and DON from the solutions prior to distillation/diffusion, and provide further purification by the selective uptake of ammonium with an acidic solution, disk, or cation resin. Ion exchange methods try to eliminate non-nitrate species by adsorption onto resins. Procedures that isolate only the nitrate include the chemical reaction method of Johnston et al. (1999) and the microbial denitrification methods of Sigman et al. (2001) and Casciotti et al. (2002).
15.4.6.1 Evaporating NO3- samples Methods suited for samples with variable inorganic-N concentrations The earliest methods for concentrating sample waters for 615N analysis were evaporation and freeze drying. Sample preservation by acidification, addition of mercuric chloride, or freezing is required when there is a delay between collection and preparation. The resulting dry solids include nitrate and nitrite. A popular modification to the original Dumas process is the addition of CaO to the reagents to eliminate CO2 and water vapor so that the resulting gas can be admitted directly into a mass spectrometer (Kendall & Grim, 1990; Fiedler & Proksch, 1972). The limitations to simply drying water samples and combusting the residues are that 1) freeze-drying of large water samples requires considerable time and, at some volume, becomes impractical, 2) samples with high concentrations of dissolved solids produce large quantities of dried salts which are cumbersome to load into combustion tubes, 3) the quartz or Vycor combustion tubes often fail because of reaction of alkali metals, particularly Na, and 4) dissolved organic nitrogen (DON) is included in the 615N analysis. The nitrate procedure is essentially the same as that for ammonium except that the nitrate must first be converted to ammonium by use of Devarda's alloy (Bremner & Keeney, 1965; Keeney & Nelson, 1982; Mulvaney, 1996). Prior to nitrate reduction, native ammonium is purged from the sample by raising the pH with the addition of MgO, which converts sample ammonium to ammonia gas. Steam is then bubbled through the sample to remove the ammonia. If necessary, nitrite (NO2) may also be eliminated by addition of sulfamic acid. Following the removal of ammonium, nitrate is quantitatively reduced to ammonium by the addition of finely divided Devarda's alloy and the ammonium is distilled and trapped as previously described. The Devarda's alloy may be prepared by ball-milling until it passes a 100-mesh screen and 75% passes a 300-mesh screen (Bremner & Keeney, 1965); however, Mulvaney (1996) mentions that a satisfactory product is produced by Merck (Darmstadt, Germany). Various degrees of N contamination associated with Devarda's alloy have been reported in the literature in reference to diffusion techniques (Liu & Mulvaney, 1992; Stark & Hart, 1996; Sigman et al., 1997; Goerges & Dittert, 1998; Johnston et al., 1999). Therefore, reagent N contamination should be considered, particularly for samples of low NO3- concentration.
Preparation and Analysis of Nitrogen-bearing Compounds in Water for Stable Isotope Ratio ...
331
15.4.6.2 Ion exchange resins for NO3The use of anion columns to collect and concentrate nitrate was an important development, allowing sample concentration to occur in the field. This meant that samples could be collected from access-limited sites. It also made possible the sequestration of nitrate from fresh waters with extremely low nitrate concentrations, typical of snow and rain. Two anion exchange methods that were widely used in the late 1990's were those of Chang et al. (1999) and Silva et al. (2000). Both methods allow for analysis of both 615N and 6180 and nitrate is converted to AgNO3. The methods differ in that the Silva et al. (2000) method is suited for samples in which NO3- concentrations are relatively high, requiring only a few liters to obtain N 100 ~moles of nitrate and in which the total organic load is < 30 mg. In contrast, the Chang et al. (1999) method was designed to obtain samples from snow and streams in forested watersheds. In these samples, NO3- concentrations are low, (down to 0.7 ~M), larger volumes of water are required (up to 70 L), and the cumulative load of DOC may approach 30 mg of more. Both methods use anion exchange resin (BioRad AGIX8 or AG2X8). Water samples are first filtered through a 0.45 gm filter. An apparatus of flexible design is used to gravity-drip sample water through the columns at up to 2 L/hr. A resin column is suspended on tubing far enough below an appropriately sized vessel for the head pressure to force the sample through at a proper rate. Alternatively, a pump may be used to establish flow through the column. Before passing the sample through the anion column, it is important to calculate the volume of water that can be passed through the column to obtain a minimum of 100 gmoles of NO3-, without exceeding the exchange capacity of the column. This is important because if the exchange capacity is exceeded, NO3- will bleed out of the column with the isotopically heavy NO3- exiting first, resulting in fractionation (Silva et al., 2000). The volume of sample that can be passed through the exchange column is a function of the total exchange capacity of the column and the total load of anions in the sample. The exchange resin has an exchange capacity of 1.2 meq/L, therefore 2 mL columns can retain 2.3 meq (Silva et al., 2000) and 5 mL columns can retain 6 meq (Chang et al. 1999). Silva et al. (2000) described the approximate capacities and limits of small, prefilled columns containing 2 mL of resin, in terms of nitrate, chloride, sulfate, and dissolved organic matter (DOM) concentrations. The modifications made by Chang et al. (1999) were made to accommodate the large sample volume (10 to 70 Ls) and to prevent DOM accumulation on the anion column. DOM can occupy sites otherwise available for nitrate, can impede flow through the column, and can contaminate the 6180 portion of the sample if ~18ODOM a r e transferred. In the Chang et al. (1999) method, the sample is filtered through a filter capsule (Gelman, with pore size 0.45 gm) with a large surface area (600 cm2). Placement of a cation exchange (AG 50 W; mesh size 100 to 200) column in front of the anion
332
Chapter 15 - C.C.Y. Chang, S.R. Silva, C. Kendall, G. Michalski, K.L. Casciotti & S. Wankel
exchange (AG-2X mesh size 100 to 200) column minimized clogging of the anion column by DOC accumulation and served to minimize transfer of unwanted 618ODOC to the sample. The cation column absorbs cationic DOM and may protonate some anionic DOM thereby allowing it to pass through the anion column as a neutral species. Other modifications made to accommodate the large sample volume were to increase the mesh size of the exchange resin to 100 to 200 (the Silva et al. (2000) method used 200 to 400 mesh) and to use an anion resin (AG-2X) that was less aggressive in adsorbing DON (the Silva method used AG 1X). This increased the sample flow and decreased the possibility that DON would adsorb to the anion column.
Elution of nitrate from the anion column After samples are concentrated on columns, the columns are chilled and returned to the lab for processing. Nitrate is eluted from the columns with several successive aliquots of 3 M hydrochloric acid (HC1). The nitrate-bearing acidic solution is neutralized with silver oxide (Ag20), which precipitates chloride in the form of silver chloride (AgC1) and leaves the nitrate in solution as silver nitrate (AgNO3). The solid AgC1 and excess Ag20 are filtered (Whatman no.5) from the solution. If 6180 analysis is desired, then all oxygen bearing species other than nitrate must be removed. To remove barium sulfates and barium phosphates a saturated solution of barium chloride is added to precipitate sulfates and phosphate. The sample is refrigerated overnight to allow the precipitate to form. The next day, the sample is filtered to remove the precipitate, the sample is put through a cation column to remove excess barium and neutralized with a small amount of Ag20. After neutralization, the sample is filtered to remove the solid silver, after which the sample can be split for separate 615N and ~180 analysis. The 615N portion is freeze-dried and rehydrated in progressively smaller volumes of water as necessary to transfer the sample to either a quartz combustion tube for sealed tube combustion (Kendall & Grim, 1990) or to silver capsules for EA-IRMS (see below). In addition to automated analysis, EA-IRMS reduces the size of sample and therefore the volume of water that must be processed. For EAIRMS, silver capsules are used rather than the more common tin capsules because AgNO3 reacts with tin. A small amount of table sugar (about 2 mg) is added to the sample to aid in combustion for 615N analysis (Silva et al., 2000); however, Stickrod & Marshall (2000) show good precisions without sugar.
15.4.7 Hybrid approaches for nitrate The methods of Garten (1992) and Downs et al. (1999) combine ion exchange methods and distillation or diffusion. Garten (1992) used 2.5 g of anion exchange resin (Dowex 1-X8, 50 - 100 mesh) to sequester nitrate from rain and throughfall samples. The nitrate was eluted with 40 mL 2 M KCL. The method was similar to that previously described for NH4 + (see section 15.4.5) except that Devarda's alloy was added to the eluant directly prior to distillation. Note that neither of these methods recommend eluting the nitrate off the column with sequential aliquots of eluant as do Chang et al. (1999) or Silva et al. (2000). The reader is referred to section 15.4.5 for a description of the Downs et al. (1999) method, which marries exchange resins and diffusion to NO3- and NH4 + for 615N analysis.
Preparation and Analysis of Nitrogen-bearing Compounds in Water for Stable Isotope Ratio ...
333
15.5 Analysis of nitrate for ~180 There are currently three methods for obtaining 6 1 8 0 from nitrate: off-line combustion, in-line combustion and the denitrifier method. Amberger & Schmidt (1987) developed the first reliable off-line combustion method which has since been refined (Voerkelius, 1990; B6ttcher et al., 1990; R6v6sz et al., 1997; Silva et al., 2000). In this method nitrate (10's of/~moles of KNO3 or AgNO3) is combusted at 520 to 850~ with different C-bearing reducing agents in quartz or Vycor tubes to produce CO2. The CO2 is cryogenically transferred to sample tubes for ~180 analysis. In the on-line combustion method (Kornexl et al., 1999a) only a few gmoles (~ 3) of nitrate is required. Combustion occurs in a reaction tube at ~ 1400~ and a continuous flow of He transports the gas to a GC column where CO is separated from other gases (N2 is also produced). In the denitrifier method (Sigman et al., 2001) NO3- is microbially converted to N20 and as little as 20 nmoles of N can be analyzed. The gas passes through a cryogenic trap, into a gas chromatograph for separation and is analyzed by continuous flow spectrophotometery. The three approaches for obtaining ~180 from nitrate are described in more detail, below.
Off- line combustion The method developed by Amberger & Schmidt (1987) does not concentrate NO3and is suitable only for samples with high NO3- concentrations, as described in section 15.4.2. This method used activated carbon to remove DOM from the sample. While this may be acceptable for high concentration samples, the use of activated carbon for samples with low nitrate concentrations is not advised. Chang et al. (1999) determined that the amount of DOM removed from activated carbon is not only highly variable, but that nitrate is also indiscriminately removed. These authors also observed that nitrate was removed from waters in which the nitrate concentration was 0.7 gM (unpublished data). Amberger & Schmidt (1987) passed sample through a cation exchange resin column (Amberlite IR 129 in the H + form) then neutralized the sample with KOH. Sulfate and phosphate were precipitated by addition of an excess of BaC12 and filtered out of solution. The remaining solution was evaporated to dryness. An aliquot of the resulting solid KNO3/BaC12 mixture was combusted in a closed tube with mercuric cyanide (HgCN) to form CO2, which was then extracted cryogenically and analyzed for 6180. The drawbacks of this method are the toxicity of HgCN and low yields of CO2. Wassenaar (1995) found that silver cyanide gave more complete yields of CO2 and better precision than HgCN. Several other methods have been used for the preparation of nitrate for 6180 analysis. R6v6sz et al. (1997) combusted nitrate in the form of potassium nitrate (KNO3) with catalyzed graphite (C + P d + Au) in a sealed tube to form CO2, N2, and K2CO3. The fractionation caused by incomplete yields of CO2 is accounted for using the fractionation factor between the CO2 and K2CO3. With this method, both 615N and 6180 may be determined from the same preparation, with excellent precisions. Br~iuer & Strauch (2000) combust nitrate to CO2 in sealed tubes using guanidine hydrochloride.
334
Chapter 15 - C.C.Y. Chang, S.R. Silva, C. Kendall, G. Michalski, K.L. Casciotti & S. Wankel
This method is cheap, safe, and gives yields ~ 100%, comparable to the silver cyanide and graphite methods; however, ~15N cannot be measured. The graphite, silver cyanide, catalyzed graphite, and guanidine hydrochloride methods all give similar precisions, better than 0.5 %0. Br/iuer & Strauch (2000) suggest that more variation arises through collection and isolation of nitrate than through conversion to CO2, a conclusion shared by the authors of this chapter. On-line combustion
The ability to do on-line combustion for 6 1 8 0 (Kornexl et al., 1999a) was a significant advancement because continuous flow mass spectrophotometers needed much smaller samples (~ 3 to 15 ~moles) compared to dual inlet mass spectrophotometers. In this method the sample is combusted at high temperature, 1400~ a process referred to by these authors as pyrolysis, but is also referred to as Thermal Conversion with an Elemental Analyzer (TC/EA) (R6v6sz & B6hlke, 2002). The difference in nomenclature derives from the use of different manufacturers (GVI versus Finnigan MAT). One of the advantages of this method is that samples can also be analyzed for 615N and 613C, and 6180 at the same time. Both inorganic and organic substances can be analyzed, with some caveats (see below). Unlike previous methods (R6v6sz et al., 1997), smaller samples are required and the primary product is CO and N2. This eliminates the need for a second reaction step to convert sample 02 to CO2 for 6180 analysis. Kornexl et al. (1999a) initially described doing analysis on samples containing 50 to 100 ~moles, but R6v6sz & B6hlke (2002) report using less than 10 ~mole. In the high temperature combustion method, a glassy carbon tube is encased within a ceramic A1203 reaction tube. Encasement of the glassy carbon tube is necessary because sample oxygen can exchange with oxygen from the A1203 at high temperatures. The glassy tube is filled with glassy carbon grit, nickelized graphite, and nickel powder (see Kornexl et al., 1999a for details). Downstream from the tube is an ascarite trap (for CO2 removal) and GC column. The reaction furnace is heated to 1400~ and the system is flushed with He carrier gas. The standard deviation for 615N values for inorganic substances was somewhat higher than for conventional techniques (up to 0.8 %o as compared to 0.3 %o). The method appears less suited for organic substances because of the relatively small amount of N2. In general, samples with high CO/N2 ratios were less suited. For samples with C / O ratios ~ 1 ~13C values were in good agreement with conventional methods, though somewhat higher than for more traditional methods. For organic compounds such as caffeine, positional ~13C values for the oxygen bearing carbon atoms were found. Recovery for nitrates and ammonium sulphates were similar to the expected stoichiometric values. Some compounds (certain amino acids) higher CO/N2 ratios indicated that other nitrogen products beside N2 were formed. Recovery of silver phosphate and some carbonaceous materials was incomplete (90 and 70%). The incomplete recovery of the carbonaceous material indicated that only two of the three oxygen atoms were reduced to CO. The third O atom remains with the tube, presumably
Preparation and Analysis of Nitrogen-bearing Compounds in Water for Stable Isotope Ratio ...
335
in the form of an oxide.
Denitrifier method for nitrate ~15N and ~180 analyses The latest method for isotopic analysis of nitrate utilizes denitrifying bacteria to quantitatively convert NO3- to N20 for isotopic analysis (Sigman et al., 2001). The two bacterial strains used, Pseudomonas chlororaphis and Pseudomonas aureofaciens, lack nitrous oxide reductase and thus naturally produce N20 as the final denitrification product. The bacterial strains (obtained from the American Type Culture Collection) are initially cultured in tryptic soy broth amended with ammonium sulfate and potassium nitrate for 6 - 10 days, in which time all the initial nitrate is consumed. The bacteria are then concentrated by centrifugation, resuspended in 4 mL of the decanted supernatant, and divided into headspace vials. The sealed vials are degassed using pure N2 at ~ 15 m L / m i n for a minimum of 2 hours to remove N20 produced during initial culture growth and atmospheric 02 that would inhibit denitrification. Aqueous samples containing 10 - 20 nmoles of nitrate are then injected into the prepared vials and incubated overnight to allow for complete conversion of the NO3- to N20. The extraction and purification of N20 can be performed off-line (Sigman et al., 2001) but may also be performed on-line with automated sample extraction and isotopic analysis using a modified Finnigan GasBench or other appropriate apparatus (Casciotti et al., 2002). Mass interference by CO2 is eliminated in each case by using a chemical trap and chromatographic separation from N20, and the purified N20 passes into a continuous flow mass spectrometer. Masses 44, 45, and 46 are measured, and the 615N is calculated from the 45/44 mass ratio after applying a correction for 170 interference (assuming ~ 1 7 0 - 0.52 * ~180). The 6180 of the N 2 0 analyte is obtained from the 46/44 mass ratio. The denitrifier method can also be used for ~ 1 8 0 analysis of nitrate in seawater and fresh water samples (Casciotti et al., 2002). While mass balance for nitrogen requires that the final 615N of N20 equals the 615N of the nitrate sample, oxygen atoms are not quantitatively carried over into the N20 analyte. The calculation of 6180 of nitrate from 6180 of N20 thus involves additional considerations for the effects of fractionation during conversion and possible exchange of oxygen atoms with water during denitrification. Fractionation due to oxygen loss during conversion of nitrate to N20 is constant for a given batch of samples and therefore does not affect the 6180 of samples relative to a known nitrate standard analyzed in parallel. Water exchange contributes less than 10%, and frequently less than 3%, of the oxygen atoms in the N20 product for Pseudomonas aureofaciens. Because this exchange is constant for a given batch of samples, the analysis of appropriate standards can be used to accurately correct the measured 180/160 ratio of samples for the low levels of exchange catalyzed by P. aureofaciens. Therefore, the 6180 of nitrate can be reached by measurement of the 6180 of N20 using this method. The denitrifier method offers a number of distinct advantages over other conventional nitrate isotope analysis methods. First, it reduces the sample size by two orders of magnitude. This is helpful in watershed studies where sample volumes required
336
Chapter 15 - C.C.Y. Chang, S.R. Silva, C. Kendall, G. Michalski, K.L. Casciotti & S. Wankel
for conventional methods reach 10 - 20 liters, and is vital in applications where limited volumes of sample are available, such as ice core or sediment porewater analyses. Second, it eliminates problems associated with waters of high ionic strength. This is especially true for oceanic nitrate studies, where the concentration of nitrate in seawater using anion exchange resins is virtually impossible due to the high concentration of C1- that competes with nitrate for absorption on the resin. Third organic contaminants are avoided because the bacteria works exclusively on nitrate. Fourth, the preparation time is greatly reduced over other methods. Fifth, both 515N and 5180 analyses can be obtained from the same sample. There are aspects of the denitrifier method that still require further refinement at this point, namely a persistent blank (up to 5% of sample size for 10 nmol samples) and the correction of 515N for 1 7 0 in atmospherically-derived nitrate. The blank appears to be consistent among vials in any given run and is believed to arise from N20 absorbed onto the bacteria from the initial growth step. Because there is little variation in 515N of the synthetic nitrates that are used in the growth medium and because most 615N values for nitrates fall in a limited range, the blank appears to be of minimal concern and a correction can be applied. This correction becomes more significant as the sample c515N move farther away from "normal" values such as in strong denitrification zones or Antarctic nitrate aerosols (515N ~ 100 %o). The analysis of 515N in atmospherically-derived nitrate using this method is currently hampered by the large mass-independent fractionation (A170 ~ 2 0 - 30 ~ o ) that occurs in atmospheric nitrate, where 5170 = 0.52 * 6180. The assumed relationship of ~ 1 7 0 = 0.52 * 6180 used for correcting the ~15N of such samples can lead to miscalculation of their 615N values by I - 2 %o. Approaches for overcoming this limitation are currently being developed (M. Galanter Hastings, in prep.).
15.5.1 Comparisons between on-line, off-line and the microbial method for ~)180 As shown by R6v6sz & B6hlke (2002), off-line methods for 6180 of nitrate are subject to 5180 exchange between the sample and the combustion tube. Off-line combustion methods typically rely on one point calibration, and use a reference with an oxygen isotope composition close to that of atmospheric 02:5180 - 23.8 %o. However, as shown by R6v6sz & B6hlke (2002), when multiple reference samples are used, with 5180 values above and below that of atmospheric oxygen, there is an offset. The offset is more pronounced as sample size decreases, as the surface area of the tube increases, and as the 6180 of nitrate becomes heavier. The offset is negligible for samples with 6180 values between 10 to 20 %o. The isotope ratio scale factors for 50 ~mole samples were more compressed relative to 100 ~mole samples when compared to off-line techniques. This finding shows the importance of doing multiple point calibration with reference standards that have 5180 values that are below and above ~ +20 %o. The microbial method of Casciotti et al. (2002) should also give values similar to the offline combustion method because there is no contact with quartz or vycor tubes. This does not invalidate the conclusions of investigations that used off-line combustion methods, which were used by most studies published before 2003. The 6180
Preparation and Analysisof Nitrogen-bearingCompoundsin Waterfor Stable IsotopeRatio ...
337
values reported by these studies should be viewed as being operationally defined by the off-line combustion method. Since the offset is affected by sample size, and because few studies reported sample size, it would be difficult to correct the 6180 values of all off-line studies. 15.6 Analysis of nitrate for ~170 Only recently has an experimental method been developed to measure 6170 (A170) in nitrate (Michalski et al., 2002). Because of isobaric interference with 13C, standard conversion of NO3- to CO or CO2 is unsuitable for ~170 measurements and 02 is the requisite analytical gas. Conversion to 02 using well developed CO2 fluorination techniques are both time consuming, expensive and are impractical for routine analysis. Recently, the thermal decomposition of AgNO3 into 02 and NOx has been developed for both 6170 and 6180 measurements. The decomposition (AgNO3 ~ 02 + NOx (NOx - NO2 + NO)) is carried out under vacuum at 520~ in a quartz reaction tube and the NOx is cryogenically trapped as the gases evolve in two Pyrex traps at liquid nitrogen (as N203). The quartz reaction tubes are pre-cleaned in HNO3 and HF then annealed at 800~ to reduce labile O atoms on the quartz surface. AgNO3 samples are isolated and purified as described by Silva et al., (2000), then loaded into silver foil boats. The boats are slightly crimped to prevent volatilization of the silver salt at the decomposition temperature. Failure to crimp the boats can result AgNO3 vapor migrating out of the heated section of the reaction tube and re-condensing in cooler sections. This generates high ~180 values since AgNO3 with low 6180 values preferentially volatilizes, leaving AgNO3 with high 6180 values behind to decompose. The 02 generated is measured for yield, collected in a molecular sieve sample tube and transferred to the mass spectrometer for analysis. The fractionation during the partitioning of the oxygen atoms into 02 and NOx is fairly consistent (std. dev. 6180 + 1.1%0) and therefore 6180 values can be obtained by applying a correction factor after the decomposition line has been calibrated using known nitrate reference materials. Calibration is important because the 6180 values can vary widely depending on distance of the traps from the decompositions chamber and whether the NOx is directly trapped or allowed to equilibrate and therefore consistency of procedure is crucial. Since the decomposition is a kinetic process, and therefore mass dependent (e.g. 6170 = 0.52 * ~180) the fractionation leaves the A170 unaffected. This is reflected in a smaller standard deviation for A170 (+ 0.11%o) even for sample sizes as small as 5 ~mol. Sample sizes between I - 7 mg AgNO3 gave the most consistent result. The size of experimental blanks becomes relevant below ~ 5 ~mol sample sizes. The authors estimate the blank to be approximately 50 nmol due to silver oxides on the boat surface. The blank increased to 100 nmol when samples ~1 gmol are purified by ion chromatography because of impurities in the Ag20 and its solubility. The authors note they are currently exploring better purification techniques including Ag + exchange membranes to replace the Ag20 neutralization. Although this technique
338
Chapter 15 - C.C.Y. Chang, S.R. Silva, C. Kendall, G. Michalski, K.L. Casciotti & S. Wankel
was developed as an offiine method, it is readily adaptable to continuous flow/auto sampler technology.
15.7 Contamination by organic matter An unresolved issue in the preparation of nitrate for 6180 and (~170 analysis, and to a lesser extent 615N, is the contribution of O or N from organic matter. DOM contains roughly 30 to 50% oxygen compared to a few % of nitrogen. Silva et al. (2000) reported that most of the DOM measured in sample waters was eliminated during the collection and preparation process. About half of the remaining DOM was eliminated by shaking the sample with Norit* G-60 activated carbon. However, Durka (personal communication) and Chang et al. (1999) found that the adsorption of DOC by activated charcoal was highly variable, even within sample replicates, and sometimes incomplete. Furthermore, nitrate was also adsorbed by the charcoal, which would greatly affect samples with low nitrate concentrations. Given that different types of activated carbon have quite different behaviors for adsorption of organic compounds, the use of activated carbon to remove DOM should be done with extreme caution, if at all. Haberhauer & Blochberger (1999) described the use of poly(vinylpyrrolidone) for DOM removal and compared activated carbon and the solid phase extraction column material Isolute C18. The sample preparation procedure involved evaporating the sample to 50 mL, passing the sample through a cation resin column, adding poly(vinylpyrrolidone) (included heating and 12 hour contact time), filtering, and removal of sulfate with BaC12. The method showed excellent removal of DOM and recovery of nitrate for samples of high nitrate concentration (10 - 60 mg/L) requiring fairly small volumes. There are other methods to remove organic matter" 1) XAD-type resins have been used to adsorb organic matter (eg. Wassenaar, 1995). Their effectiveness varies greatly depending on the DOM composition. Initial cleanup procedures are rather lengthy but they do not adsorb nitrate. 2) Dialysis. 3) ultrafiltration. 4) ultra centrifugation (RNA thimbles). Johnston et al. (1999) took a slightly different approach to minimize DOM contamination. These authors developed a method that converts nitrate to 1-phenylazo-2napthol (Sudan-I) for combustion using EA-IRMS. This method was a modification of similar methods designed for 15N tracer studies (Schell, 1978; Kator et al., 1992). Sample sizes of 80 - 100 gg N per sample were required for adequate precision using EAIRMS. This is a very complicated method, requiring several chemical conversions and a final product Sudan-1 that contains one reagent-derived N and one N from nitrate. In brief, water samples are filtered, and the humic acids are removed by passage through prepacked, conditioned, reverse-phase columns (C18 resin) using a solidphase extraction manifold that holds 10 columns. Samples were adjusted to ~107 gM nitrate, buffered, and then slowly passed through columns packed with specially pretreated granular cadmium to quantitatively convert the nitrate to nitrite. The nitrite is then converted to Sudan-1 by a complicated reaction of aniline dye, naphthol, metha-
Preparation and Analysis of Nitrogen-bearingCompoundsin Water for Stable IsotopeRatio ...
339
nol, and assorted buffers. The Sudan-1 is then concentrated by passage through C18 columns, eluted with acetone, and then dried gently overnight. The Sudan-1 is then resuspended in acetone, pipetted into tin boats, freeze-dried, and analyzed by EAIRMS. This method provided accurate and precise 615N values for test samples. Spectrophotometric analysis showed complete conversion of nitrite-N to Sudan-1. About 93% of the dye formed could be put into the boats. Although the Sudan-1 contains one N from the aniline and one N from the sample, the resultant 615N is not the mean of the two values. Instead, the 615N was ~ 9 %o lower than the m e a n ~)15N, probably a result of a kinetic isotope effect involving an excess of the reagent (Johnston et al., 1999). However, the effect was consistent and by analyzing a set of samples with a range of known c515N values along with the samples, a regression line (r2 - 0.9999) was obtained that corrected the Sudan-1 615N values to provide precisions of 0.2 %0.
15.8 Sequestering dissolved organic nitrogen Dissolved organic nitrogen (DON) is usually operationally defined as the N in dissolved organic matter (DOM). Fractions of DOM rich in N include amino acids, proteins, and phenols. Because a wide variety of dissolved organic molecules contain nitrogen, exactly what is analyzed for DON-615N depends primarily on how the water sample is processed after particulate N is removed. Specific types of organic nitrogen can be converted to ammonium using modifications of the Kjeldahl method (see section 15.4.2), and then converted to N2 for ~15N analysis, using various methods. Methods for isolating the DOM for later combustion to N2 and analysis for ~15N range from the very simple freeze-drying of the sample, to the addition of several possible methods prior to freeze-drying, including removal of DIN by ion chromatography, ultrafiltration to isolate a specific molecular weight fraction followed by ion chromatography (Bronk & Glibert, 1991), roto-evaporation followed by dialysis to remove DIN (Feuerstein et al., 1997), and passage through various XAD resins to separate different fractions by their chemical properties (Aiken et al., 1979; Thurman &Malcolm, 1981). Procedures for isolating DOM that result in solid residues can be analyzed for 615N as well as 613C. Automated instruments for the analysis of DIC/DOC for 613C using a DOC analyzer connected to a mass spec are now commercially available (Gilles St. Jean, personal communication, 2003); these could be potentially be modified for 615N of DIN/ DON in the near future.
15.9 Methods for producing N2 from nitrogen-bearing species Sealed tube combustion
Solids resulting from any of the above concentration methods, including ammonium and nitrate, can be converted to N2 gas by Dumas combustion, in which the sample plus Cu and CuO are heated in an evacuated and flame-sealed tubes at various elevated temperatures (Kirshenbaum et al., 1947; Fraser & Crawford, 1963; and many others).
340
Chapter 15 - C.C.Y. Chang, S.R. Silva, C. Kendall, G. Michalski, K.L. Casciotti & S. Wankel
A commonly used modification to the original Dumas process involves the addition of CaO to the reagents to eliminate CO2 and water vapor (Fiedler & Proksch, 1972), with combustion at 850~ in quartz tubes followed by slow cooling, so that the resulting pure N2 gas can be admitted directly into a mass spectrometer with no further purification (Kendall & Grim, 1990). Some caveats include: 1) large fluffy samples are cumbersome to load into combustion tubes, 2) any powder remaining on the surfaces later flame sealed can weaken the glass, causing cracking of the tubes during cooling, 3) the combustion tubes often fail because of reaction of alkali metals, particularly Na, with the walls of the tubes, 4) the tubes must be allowed to cool slowly to produce pure N2 (Macko, 1981), and 5) if the CaO is not dried carefully to remove H20 and CO2, tubes often explode during combustion. Precisions better than 0.1%o are common (see Kendall & Grim (1990) for details). Off-line sealed tube combustion methods have been automated using elemental analyzers (see below). Nitrate solids resulting from the methods above can also be analyzed for ~ 1 8 0 using silver cyanide (Amberger & Schmidt, 1987), catalyzed graphite (R6v6sz et al., 1997), guanidine hydrochloride (Br/iuer & Strauch, 2000), and graphite (Silva et al., 2000). These methods are described in the previous sections. More recently such samples are reduced to CO using a glassy carbon furnace in an automated high temperature elemental analyzer (EA-IRMS), and analyzed for 6180 (see section 15.5.3).
Rittenberg method The "classic" Rittenberg method (Rittenberg et al., 1939; Rittenberg, 1948; Sprinson & Rittenberg, 1948, 1949) for analyzing ammonium produced by distillation (before or after Kjeldahl digestion) involves conversion of ammonium to N2 by oxidation with alkaline sodium hypobromite in an evacuated vessel. The reaction is performed in a special "Rittenberg" Y-shaped tube (see section 15.4.3) that can be tilted to bring the hypobromite in one arm, in contact with the liquid sample in the other arm. The tube must be carefully degassed prior to letting the liquids react. The reaction is usually represented as 2NH3 + 3OBr- ~ 3Br- + 3H20 + N2
[15.3]
However, the reaction is not quantitative and small amounts of N20 and NO3- are formed (Bremner, 1965a). The N2 produced must be purified on a vacuum line to remove hypobromite and water vapor. This is performed by freezing the sample container (Y-tube or vial) in a low-temperature bath, and passing the N2 through a liquidnitrogen filled dewar before admission into the mass spectrometer. Although this technique has been applied extensively in the agricultural and biological sciences, it is generally regarded as a tedious and labor-intensive procedure because of the time required to prepare the unstable reagent, clean the Y-tubes, degas the solutions in the Y-tubes, control effervescence during reaction, prevent leaks, and purify the resulting N2 (Mulvaney, 1993). Many modifications of this method have been published and are reviewed in Mulvaney (1993). Some particularly notable modifications are: use of vacuum stopcocks
Preparation and Analysis of Nitrogen-bearing Compounds in Water for Stable Isotope Ratio ...
341
and o-ring seals, to eliminate air leaks (Ross & Martin, 1970); use of lithium hypobromite instead of sodium hypobromite, especially for dry samples, because it is more stable (Ross & Martin, 1970); and the use of disposable reaction vials (Ross & Martin, 1970; Porter & O' Deen, 1977). The Rittenberg method has been successfully automated and commercial "Automated Rittenberg Analysis mass spectrometery" (ARAMS) units have been used by soil scientists for over 20 years (McInteer & Montoya, 1981; Mulvaney et al., 1990; Mulvaney & Liu, 1991). These ARA-MS system could handle small water samples with ammonium amounts as small as 25 gg N, and could analyze several hundreds of 15N tracer samples a day by automatically adding hypobromite to each sample, purifying the gas, and analyzing it for 615N (Mulvaney, 1993).
Automated combustion and analysis using an elemental analyzer Any of the above methods for preparing N-bearing samples for 615N analysis can be automated by placing the dry products into tiny foil boats, and loading them onto an automated C / N elemental analyzer connected to a continuous flow stable isotope mass spectrometer where the samples are combusted, purified, and the N2 is analyzed for 615N. Other combustion products, i.e. CO2, water, and SO2 are either trapped or separated from the N2 by the GC column. Such instruments, originally developed for 15N tracer samples (especially soils) and sometimes referred to as Automated Nitrogen and Carbon Analysis Mass Spectrometry (ANCA-MS) units, were developed almost 20 years ago (Otsuki et al., 1983; Preston & Owens, 1983; and many others), and have been commercially available for > 15 years. For a nice review of the history of ANCA-MS, see Barrie (1991). This general category of automated combustion is more typically referred to as "Elemental Analyzer- Isotope Ratio Mass Spectrometry", or EA-IRMS. Since the procedures are discussed in detail elsewhere (Chapter 8), only a brief description will be given here. The procedure for loading samples into foil boats depends on sample type. Ground and homogenized powders (from POM or biological samples) are weighed into small foil boats, which are crimped shut. Ammonium sulfate samples on glass filters are usually folded into larger boats, which are compressed and crimped. Tin boats are usually used, but silver boats are preferred for corrosive acidic samples. Samples and reference materials are placed into the wells of a carousel, which is mounted in an autosampler on an elemental analyzer. Under computer control, the samples are purged of air by a flow of helium (He), the samples are dropped one by one into a quartz tube, typically half filled with Cr203 and silvered cobalt oxide CO3/Ag generally between 120 and 1050~ The silver acts as a trap for halogens and sulfur. Helium continuously flows through the system. A measured pulse of oxygen is timed to enter the helium stream when the sample is dropped. The tin capsules oxidize at ~1700~ adding heat to the sample combustion. Combustion of organic samples generally produces N2, NOx, SOx, CO2, and H20. He carrier gas sweeps the gases through a reduction tube filled with Cu wire at ~ 600~ where the NOx is reduced to N2 and excess 02 removed. A dessicant removes water vapor. The N2 and CO2 peaks are separated by a GC column, and the sample flows into the mass spectrometer for analysis of 615N and 613C. Precisions of 0.1 to 0.2%0 for 615N and ~13C samples are routinely achieved, for sample sizes > 1 gmole of N or C.
342
Chapter 15 - C.C.Y. Chang, S.R. Silva, C. Kendall,G. Michalski, K.L. Casciotti& S. Wankel
If the ammonium ions have been collected on a sufficiently small volume of exchange resin or zeolite, then such samples can simply be added directly into the tin capsules. If samples have been trapped in acid and dried to form a salt, then the sample vials should contain a few mg of dry ammonium sulfate or other salt as a film at the base. If quantity allows it may be possible to scrape this off with a micro-spatula, and weigh it into tin capsules directly but some analysts have found it easier to add 100 ~L of distilled water to the vial to re-dissolve the salt and then dispense it with an adjustable (up to 25 mL) micro-pipette. The 6 x 4 mm tin capsules are prepared by adding I - 2 mg of a nitrogen free cross-linked dextran such as 'Ultrodex' (Pharmacia Biotech, Uppsala, Sweden), then an appropriate aliquot of the sample solution is pipetted in. The Ultrodex forms a paste with the sample and makes subsequent handling easier. If less than 20 mL of sample is added then it is possible, with care, to crimp the capsules closed over the wet absorbent but if more liquid is added then it is dried down into the capsules in an oven at 75~ Capsules may then be crimped and run as normal. Samples prepared using the diffusion techniques of Sorensen & Jensen (1991) and Stark & Hart (1996) are enclosed in the PTFE parcels. Once removed, the paper strips were of such a size that they could be simply rolled up and placed into a 6 x 4 mm tin combustion cup, and then loaded directly onto the EA-IRMS. However, as the paper strips had been wetted with an excess of sulfuric acid, they were transferred to the tin capsules immediately prior to analysis as the acid attacks the tin. The encapsulated samples cannot be stored for more than an hour without the capsules falling apart. Combustion of these samples seems to attack and shorten the life of the quartz tube forming the oxidation furnace and this will need changing more frequently. Use of silver boats avoids corrosion but the resultant mass of melted silver will eventually clog the combustion tube. 15.10 Methods for producing CO and CO2 from nitrates and other oxygen-bearing species Many groups are currently analyzing the resultant AgNO3 (or KNO3) using automated high temperature carbon reduction systems (sometimes erroneously called "pyrolysis systems") that produce CO and N2, which are separated by passage through a GC, and analyzed for 6180 and 615N, respectively. This will undoubtedly soon become the preferred method because it provides simultaneous analysis for 6180 and 615N. The first pyrolysis method of Koziet (1997) provided for simultaneous measurement of CO for 6180 and N2 for 615N using an automated elemental analyzer. Organic samples were pyrolyzed in a reactor with nickelized graphite and vitreous (glassy) carbon. This and similar procedures developed by Farquhar et al. (1997), Kornexl et al. (1999a), and others were described earlier in this chapter. 15.11 New Frontiers Since the late nineties significant progress in analytical techniques has been made for stable isotope analysis of materials since the original developments almost half a century ago. As a result, the number of stable isotope laboratories in the world increased rapidly, and an exploding number of isotopic analyses are being reported in
Preparation and Analysis of Nitrogen-bearing Compounds in Water for Stable Isotope Ratio ...
343
the literature. Overall quality control of isotopic data (e.g., sample documentation and tracking, information management for sample preparation, mass-spectrometric measurements, and data reduction) is becoming an important issue. These advances have included: 1) the ability to simultaneously determine more than one element in a single analysis, 2) the ability to processe samples in the field, 3) the ability to measure smaller sample sizes, 4) to shorten the sample preparation time, and 5) the ability to differentiate between different species of organic compounds and isotopomers of N20. The challenge will be to maintain accuracy and precision as the sample size decreases. For instance, recent developments of continuous-flow mass spectrometry are making it possible to determine both 615N and 6180 values of nitrate samples simultaneously (high-temperature carbon pyrolysis method, also the denitrifier N20 method). Such techniques are particularly useful for small-size samples. With decreasing sample size, possible isotopic fractionations associated with extraction from the matrix, contaminants in the reagents, and leaks will become more significant. Simultaneous measurement of 615N, ~170 a n d ~180 of nitrate requires some means for separating the isobaric interference of 170 within the analyte gas or gases (CO2, CO or N20). When analyzing compounds of multiple elements with multiple stable isotopes (e.g. carbon and oxygen in CO2) there is a certain amount of 'overlap' in the ion beams measured by the mass spectrometer- 'isobaric interference'. For example, when measuring M / Z 45 one cannot distinguish between 130160-160 and 120160170. While for mass dependent samples the effect of this interference is easily removed by measurement of the 6180 and the application of the "Craig correction," (Craig, 1957) this is not the case for mass independent fractionations (MIF; containing anomalous 170). A few recent approaches to measuring the ~170 of nitrate have been developed. Combined thermal decomposition and electrical discharge has been used effectively (Michalski, 2002) for measuring 6170 of NO3. Simultaneous measurement of N20 generated via the denitrifier method and N2 (reduced from the N20) as a means for solving for the isotope ratios seems to lack the precision necessary for most environmental research - but could potentially be improved (Wankel et al., 2002). With the analytical advance of the denitrifier method for analysis of the 615N and 6180 of nitrate, new doors have been opened for the analysis of atmospheric and porewater nitrate, as both are typically either low in concentration and/or small in volume. Furthermore, the merging of other new techniques should allow both the ability to simultaneously m e a s u r e ~15N, ~170 and 6180 of nitrate as well as a reduction in the amount of nitrate required for analysis of all three isotope ratios. This should open the way for research into the three-isotope ratio tracer system of nitrate, especially useful in atmospheric studies and in tracing N sources and cycling. Using purified non-microbial reductases to reduce nitrate also shows some promise. Such reductases that are unaffected by the presence of oxygen offer the advantage over oxygen sensitive microbial reductases and circumvent the need to grow the
344
Chapter 15 - C.C.Y. Chang, S.R. Silva, C. Kendall, G. Michalski, K.L. Casciotti & S. Wankel
microbes. For instance, soluble nitrate reductace (NaR), purified from corn leaves has been used to irreversibly catalyze the reduction of nitrate to nitrite with high specificity (Patton et al., 2002 and references therein). The measurement of isotopomers of N20 (the distribution of the 15N atom within the linear N-N-O molecule) carried out with multi-collector mass spectrometry is another rapidly advancing field (Yoshida & Toyoda, 2000; Chapter 19). While this method has proven to be useful in understanding the distributions and sources of N20, multi-collector technology may spawn other techniques useful for tracing nitrate as well. Multicollector IRMS systems are becoming increasingly common and the isotopic analysis of N20 produced via the denitrifier method along with the secondary fragments produced in the ion source of the mass spectrometer (N2+ and NO +) should allow measurement of the 1515N, ~1170, and 6180 of nitrate simultaneously. Finally, because of the interest in distinguishing different species of organic-N compounds, the diffusion method has been applied to differentiate types of organicN (Mulvaney & Khan, 2001). These efforts include procedures to determine urea615N in soil extracts (Marsh et al., 2003, in press), and the recovery of ~14C and 615N from alpha-amino acids in soil hydrolysates (Mulvaney, personal communication). This is of interest because there is some indication that cornfields may be insensitive to fertilizer application if amino acid sugar concentrations in the soil are elevated (Mulvaney et a12001). Clearly, new advances in stable isotope analysis (smaller sample sizes, multiple analysis of more than one isotope, automation and the ability to differentiate between species) are providing new tools for investigating sources, patterns, and transformations (denitrification, NOx to NO3- ). The ability to analyze water samples, in conjuction with atmospheric, biological, and soil samples will aid our understanding of how biologically sensitive elements such as C, N, O, and S cycle, are transformed, and stored.
Acknowledgements We thank David Velinsky, Richard Mulvaney and Bernhard Mayer for their comments and reviews. Use of firm, brand, and trade names in this manuscript is for identification purposes only and does not constitute endorsement by the U.S. Geological Survey. Thanks also to Darren Sleep for his contribution to an earlier version of this draft.
Preparation and Analysis of Nitrogen-bearing Compounds in Water for Stable Isotope Ratio ...
345
Appendix 15.A1 - H o w to distill a sample a soil solution for ammonia, as described by Mulvaney (1996) Pipet about 10 - 20 mL of the NH4 + containing solution into the distillation flask. Buffer the sample to p H 9.5 with (0.2 g) MgO and commence distillation. W h e n the distillate reaches 35 mL the process is stopped and the solution can be titrated to determine the a m o u n t of NH4 + recovered. To titrate, 5 mL of H3BO3-indicator solution is a d d e d to a 50 mL beaker, m a r k e d to indicate a volume of 35 mL prior to distillation. The beaker is placed u n d e r the distillation apparatus. After distillation the sample is titration with 0.0025 M H2SO4. The endpoint is reached w h e n the color changes from green to faint pink. Boric acid indicator solution: Dissolve 400 g of reagent grade boric acid H3BO3 is added to 18 L of deionized water in a 20 L bottle. Add 400 mL of indicator solution (0.495 g of bromocresol green and 0.33 g of methyl red in 500 mL ethanol) and bring the volume to 20 L with deionized water. With continuous stirring, adjust the pH to 4.8 or 5.0) or until the solution assumes a reddish purple, by careful addition of I M NaOH or single NAOH pellets. If excess NaOH is added, lower the pH with dilute HCL.
15.A2- How to diffuse a sample with high NH4 + concentrations onto a disk as described by Khan et al. (1998) 1 - A water sample or 2 M KCL soil extract (5 - 100 mL) containing 50 to 150 ~g of NH4 + is put into a 1 pint (473 mL) m a s o n jar. For fresh water samples, add e n o u g h KCL to make a 2 or 4 M KCL solution. 2 - A d d ca. 0.2 g of MgO, swirl, then place the acidified disk(s) into the disk holder. In this method, the disk is not encased b e t w e e n Teflon envelopes, rather it is susp e n d e d above the solution. See paper for details. 3 - Incubate sample at 20~ or on a heating plate at 45 - 50~ If the sample is heated, then a 4 M solution should be used. The a m o u n t of time to diffuse a sample d e p e n d s on the sample volume. A 20 mL sample heated at 45 - 50~ will take 5 hours, whereas a 100 mL sample will take 14 hours. 4 - Dry the disks in a v a c u u m desiccator containing Drierite. 15.A3- Summary of experiments by to Mulvaney & Khan (1999) to investigate the affects of organic solutions on ammonia diffusions Solutions contained 714 ~moles (10 mg) of a m m o n i a and ~ 11.4 mg of soluble organic-C. The lowest N concentration determined by this m e t h o d was ~ 15 mg L-1 or 1.07 mM. The sample volume was 10 mL and contained ~g of N labeled as (NH4+)2 SO4, KNO3, or NaNO2. Three grams of KC1 was a d d e d to each sample to give a 4 M KC1 concentration. The solutions were put into the m a s o n jars. Each m a s o n jar contained 7 mL of the boric acid indicator solution (see section 15.A1 or Kahn et al., 1997 for instructions on h o w to prepare the solution). The solution was placed in a small petri dish, s u s p e n d e d above the sample solution by a clamp (Khan et al., 1997). The samples were heated for 1.5 hours at 45 - 50~ or allowed to stand for 18 hours at room temperature.
346
Chapter 15 - C.C.Y. Chang, S.R. Silva, C. Kendall, G. Michalski, K.L. Casciotti & S. Wankel
To evaluate whether there were interferences caused by DON, a concentrated solution with (10 g N L-l) containing eight ammonia acids and two amino sugars was made. Organic interference was more pronounced for large sample volumes, samples that were heated, had Devarda's alloy) and in which the concentration of KC1 was less than 4 M. Small samples requiring shorter diffusion periods, and with low surface-to volume ratios had the least carryover. Carryover was minimized in samples with 4 M KC1, an effect attributed to the ability of salt promoting liberation of NH4+-N thereby minimizing the time needed to diffuse the sample. This is because larger samples take longer to diffuse. Organic carryover also occurred in samples with Devarda's alloy, presumably because the samples were heated. These results are consistent with those of Kahn et al. (1997) and Mulvaney et al. (1997) in which alkali-labile nitrogen compounds (300 gg of N as glucosamine or glutamine) were used as test compounds. No organic interference occurred during diffusion of NH4+-N, from 10 or 20 mL samples, but organic N was detected on 50 to 100 mL samples. Organic interference was also more pronounced in distilled samples because of the sample heating than samples diffused into H3BO3. 15.A4- Recovery of 815N-NO3- in waters with low nitrate concentrations (~ 2 to 5 /IM) as described by Sigman et al. (1997) 1 - Measure out a volume of water that will give the minimum amount of N need for mass spectrometry, - 2 to 5 ~mol N. For freshwater samples, 50 g NaC1 ~ L-1 (ashed 4 hours at 450 ~ must be added to increased to increase the ionic strength to ~ 35 ppt. This prevents the filter packs from bursting because of osmosis (Holmes et. al. 1998). 2- Add 300 mg MgO per 100 mL of sample (precombusted at 650 ~ to raise the pH to 9.7. 3- Reduce the sample volume by boiling or evaporation. Heating the sample promotes breakdown of DON. Reducing the volume will minimize the time needed to diffuse the sample. Since DON hydrolysis is promoted by heat, evaporating the sample at 65~ rather than boiling may be preferable. 4- Make diffusion packets out of I cm GF/D filter disks (Whatman #1823 010) or Whatman quantitative paper no. 41. Kahn et al. (1998) recommends the latter, because it is ashless, avoiding accumulation of glass in the combustion tube, and because the paper remains flexible after drying. The disks are acidified with 20 -30 ~L of 4 N H2SO4, H3PO4, or KHSO4 (the latter is less corrosive to tin capsules that the traps are put in) and sandwiched between two 2.5 cm diameter 10 ~M pore sized Teflon membranes. The reader is referred to the paper for details on making the diffusion packets. 5- Boil or evaporate the sample to reduce the sample volume (15 - 25%) and remove dissolved ammonia. 6 - Add the diffusion packet to the sample bottle. Add Devarda's alloy (75 mg/100 mL sample). Tighten cap immediately. 7- Swirl bottle and incubate at 65~ for 4 days. 8- Remove bottles from oven and place on reciprocating shaker for up to 7 days, depending on sample volume.
Preparation and Analysis of Nitrogen-bearing Compounds in Water for Stable Isotope Ratio ...
347
9 - Remove diffusion packet and dip into 10% HC1, then distilled water. Place packets in a dessicator in the presence of an open container of sulfuric acid (to remove trace ammonia). Leave packets for 1 - 2 days to dry. Alternatively, the samples may be dried in an oven, but this increases the chances of a m m o n i u m loss or fractionation. 10 -To store, place packets in individual air-tight vials with sealing caps. 11 - On the day of analysis, remove the filter from the Teflon membrane and put into tin boats immediately before analysis to minimize corrosion of the tin by the sulfuric acid in the filter. 12- Correct isotopic measurements for the blank effects caused by Devarda's alloy and DON.
Handbook of Stable Isotope Analytical Techniques, Volume 1 P.A. de Groot (Editor) 9 2004 Elsevier B.V. All fights reserved.
CHAPTER 16 815N Analyses of Ammonium-Rich Silicate Minerals by Sealed-Tube Extractions and Dual Inlet, Viscous-Flow Mass Spectrometry Gray E. Beboutl & Seth J. Sadofsky2 Department of Earth & Environmental Sciences, Lehigh University, Bethlehem, Pennsylvania 18015, U.S.A. e-mail: 1
[email protected], 2
[email protected]
Abstract
In this paper, we report progress on the development of techniques for the routine, quantitative extraction and isotopic analysis of trace amounts of nitrogen in micas and whole-rock silicate samples. Using relatively straight-forward sealed-tube extraction techniques and standard cryogenic purification methods and dynamic-vacuum, dualinlet, viscous-flow mass spectrometry, it is possible to routinely obtain highly accurate and precise 815N and N concentration data for such materials. For 815N values of samples containing > 100 ppm N, lo for 2-10 duplicate analyses typically approaches 0.1%o, and N concentration data nearly always have < 5% uncertainty. Blanks of these extraction techniques are < 0.025 ~mol N2, affording analyses of small N2 samples to 1.0 ~mole. The precision of these analyses is typically higher than that obtainable by static mass spectrometry techniques, with the latter methods commonly producing data with l a nearer 0.5%o. However, the requirement of viscous-flow conditions (in turn requiring maintenance of sufficient pressure in the inlet volume) and the geometry of the sealed-tube experiments (placing limits on the maximum sample size that can be fused without rupture of the tubes) impose a practical limit of ~ 10 ppm as the minimum N concentration that can reliably be analyzed by these techniques (for samples containing < 100 ppm N, lo ~ 0.30%o, using the same techniques). In many crustal igneous and metamorphic rocks, white-micas and biotite, both of which strongly partition N as NH4 + into their interlayer sites, contain the majority of the whole-rock N and commonly have N concentrations in the range of 10 - 1000 ppm. Despite the trace concentrations of N in most silicate rocks, the analytical precision obtainable by sealed-tube extractions and dual-inlet, viscous-flow mass spectrometery, and the degree of isotopic variation for N within individual suites, rival those for 8180 data in the same suites. This, and the particular ability of N isotope data to trace sediment-derived components in fluid-melt-rock systems and potassic metasomatic alteration in mafic and ultramafic settings, should make the N isotope system an attractive, more routine addition in field studies of crust-mantle mass transfer.
815N Analyses of Ammonium-Rich Silicate Minerals by Sealed-Tube Extractions ...
349
16.1 Introduction
Despite early pioneering work in the 1950s (Hoering, 1955; Mayne, 1957; Scalan, 1958), N isotopes have been underused in crustal petrologic studies, presumably largely due to the trace quantitites commonly present in silicate systems and the analytical difficulties inherent with the large potential for atmospheric contamination. After this early reconnaissance-style work on N isotopes in crustal silicate rocks, further development of the N system as a routine tool applied to petrologic problems appeared to have "fallen by the wayside" somewhat (with only a few exceptions; Becker & Clayton, 1977), as research on O, H, C, and S isotopes provided exciting new information in an increasingly broader range of earth-science disciplines. Over the last 30 years, N isotopes have been employed extensively in studies of biogeochemical pathways, meteorites and mantle materials (primarily diamonds), largely due to the ease of analysis in the relatively N-rich organic systems (see recent work by Williams et al., 1995; Ader et al., 1998) and the common use of high-vacuum, low-blank extraction (in some cases, stepped-heating) systems in laboratories conducting the meteoritical and mantle petrologic work (e.g., Kerridge, 1985; Boyd et al., 1987; Hashizume & Sugiura, 1995; Boyd, this volume). The common analysis of Ar isotopes in micas and feldspars by stepped heating (e.g., Wijbrans & MacDougall, 1986; Grove & Bebout, 1995), and the shared interlayer residency of Ar and NH4 + in micas, have provided the potential for some interesting intellectual synergy among Ar and N isotope geochemists, in particular, in considering release during incremental heating experiments (and implications for diffusive and other mechanisms of release; see discussions for N by Boyd et al., 1993a) and in assessing closure during cooling of higher-T rock suites (see Sadofsky & Bebout, 2000). More recent work indicates that the underexploited N isotope system shows great potential for elucidating processes such as fluid-rock interaction and ore formation, magma provenance, crystallization, and degassing, and crust-mantle cycling (see recent work by Bebout & Fogel, 1992; Boyd et al., 1993a; Mattey et al., 1994; Bebout, 1995, 1997; Boyd & Philippot, 1998; Bebout et al., 1999a, b; Sadofsky & Bebout, 2000). High concentrations of NH4 + in igneous and metamorphic minerals (Honma & Itihara, 1981; Duit et al., 1986; Hall, 1999) suggest that N may in some cases be an important fluid constituent. Nitrogen species (particularly N2) are commonly found in fluid inclusions (Kreulen & Schuiling, 1982; Kreulen et al., 1982; Bottrell et al., 1988; Darimont et al., 1988; Andersen et al., 1989, 1995). Nitrogen is a common constituent of volcanic gases (e.g., Fischer et al., 1998; Sano et al., 1998), is abundant in some magmatic/hydrothermal deposits (Junge et al., 1989; Hall, 1999; Krohn et al., 1993; Bebout et al., 1999a), and occurs in trace amounts in various mantle-derived rocks (see Javoy et al., 1986). It is an important constituent of natural gases from some sedimentary basins (Jenden et al., 1988) and deep crystalline rock environments (Sherwood et al., 1988). Previous N extraction techniques for isotopic analysis of rocks and minerals have, in general, involved wet chemical techniques (dissolution/distillation) or high-temperature combustions and fusions of samples, and have employed either dynamicvacuum/viscous-flow or static-vacuum mass spectrometry. The methods of Mayne
350
Chapter 16 - G.E. Bebout& S.J. Sadofsky
(1957; largely adopting the techniques of Rayleigh, 1939) employed long-duration (a 15 hours) heating of large samples (up to 30 grams) at 950~ in open-ended quartz tube linings in stainless steel vessels, and the elegant experimental study by Scalan (1958) tested a wide variety of extraction techniques including R. F. induction heating, resistance-furnace heating in stainless steel or Ni vessels with or without CuO, NaOH-fluxed fusions, and HF dissolutions/Kjeldahl determinations. The time-consuming Kjeldahl distillation techniques, which have continued to see some use in silicate studies (Haendel et al., 1986; Junge et al., 1989), involve numerous chemical steps and may not provide complete yields (Minagawa et al., 1984; Rigby & Batts, 1986), resulting in possible isotope fractionation and relatively poor precision (Rigby & Batts, 1986; Haendel et al., 1986). The fluxed-fusion techniques developed and applied by Zhang (1988; see description of earlier work with fluxes by Scalan, 1958), involving fusion of ~ 1.5 grams of sample using LiBO2-V205 flux mixtures, and dual-inlet, viscous-flow mass spectrometry, are relatively time-consuming but afford analyses (with reduced precision) of extremely low-N materials (e.g., basalts containing < 1 ppm N; blanks for this technique are ~ 0.01 gmoles). Scalan (1958) and Prombo & Clayton (1993) employed on-line heating in molybdenum and tungsten-wire crucibles, respectively, by R.F. induction (achieving temperatures of > 1500~ and obtained reliable N-isotope data for silicate and iron meteorite materials (Prombo & Clayton, 1993, report reproducibility of ~0.4%0 for NBS SRB 1098 steel standard). Static mass spectrometry (e.g., Boyd et al., 1993a; Boyd, this volume), in some cases coupled with lasers (for analyses of diamonds; Boyd et al., 1987) and stepped-heating experiments, has afforded analyses of extremely small N2 samples; however, the somewhat lower precision of these techniques (~ 0.5%o) relative to that obtainable by dual-inlet, viscous-flow mass spectreometry (~ 0.1-0.2%o; this study) can be problematic in studies of suites showing only several %0 overall variation. Combustion-gas chromatography-isotope ratio monitoring techniques (GC-C-IRMS; see Brand, 1995b), perhaps ultimately involving laser ablations, constitute a "next frontier" for N-isotope analyses of rock systems but have not yet been applied to analyses of silicates. Over the last nine years, at the Geophysical Laboratory (1990-1991; Carnegie Institution of Washington, Washington, D.C.) and at Lehigh University (1992 to the present), we have modified the sealed-tube combustion techniques initially used in biogeochemical studies (Macko, 1981; Minagawa et al., 1984; Rigby & Batts, 1986; Kendall & Grim, 1990; Boyd & Pillinger, 1990; Ader et al., 1998) and applied these modified techniques to the analysis of N isotopes in silicate minerals and whole-rock silicate samples. These techniques, which can be employed in practically any stable isotope laboratory (i.e., any laboratory operating off-line extraction vacuum lines and dynamic-mode, dual-inlet, viscous-flow mass spectrometers with reasonable sensitivity), with little or no additional extraction-line construction, involve heating of mineral separate and whole-rock samples to 910 ~ - 1250~ in sealed quartz tubes with CuO wire, Cu metal, with or without CaO (to remove H20 and CO2; Kendall & Grim, 1990; Bebout & Fogel, 1992; Boyd& Pillinger, 1990; Sadofsky & Bebout, 2000). The N2 gas samples we analyze are in the size range of I - 50/~moles and can easily be analyzed using the variable volume (bellows) and microvolume (cold finger) inlets on a viscous-flow mass spectrometer (in our case, on the Finnigan MAT 252 at Lehigh Uni-
815NAnalysesof Ammonium-RichSilicate Minerals by Sealed-TubeExtractions ...
351
versity). Nitrogen is known to strongly partition as NH4 + into micas (particularly biotite), in rocks containing these minerals, relative to other minerals capable of incorporating it into their structures (e.g., K-feldspar and plagioclase; see Honma & Itihara, 1981). Honma & Itihara (1981) reported that, in a rock containing biotite, muscovite, K-feldspar, and plagioclase, the muscovite contains on the average ~ 40% of the N concentration in the coexisting biotite, the K-feldspar ~ 40%, and the plagioclase (depending on Ca content) on the order of 10%. The tendency of muscovite to contain -- 40% of the N concentration in coexisting biotite has been confirmed in more recent studies by Boyd & Philippot (1998) and Sadofsky & Bebout (2000); however, the latter two authors have identified significant variation within individual suites that could be attributed to differential closure or varying effects of retrogradation reactions during cooling of higher-grade metamorphic suites. Our field-based studies (Bebout & Fogel, 1992; Bebout, 1995, 1997; Bebout et al., 1999a, b; Sadofsky & Bebout, 2000) have focussed primarily on metasedimentary suites lacking K-feldspar and in which biotite and/or muscovite are present, and on metamafic and metaultramafic suites in which K-feldspar is absent and one or both of the micas occur as metasomatic products (e.g., fuchsite in alkali-metasomatized ultramafic rocks; muscovite and biotite in K-metasomatized metabasalt; see Bebout, 1997). In such suites, N is extremely concentrated in the mica phases (> 95% in micaceous rocks lacking K-feldspar and containing minor plagioclase and/or hornblende). Our development of extraction techniques has largely focussed on the quantitative extraction of N from biotite with varying compositions and white-mica, the latter ranging from nearly endmember muscovite in higher-grade metamorphic rocks (Sadofsky & Bebout, 2000) and pegmatites to extremely celadonitic, in some cases, Cr-rich white-micas (fuchsites) in lower-grade suites (Bebout & Fogel, 1992; Bebout, 1997; Bebout et al., 1999a). 16.2 Discussion of analytical methods Clean, unweathered rock samples are crushed in a jaw crusher (or, for fine-grained samples, in a large mortar and pestle) and then (for some samples) in a disk mill to a size of 0.25 to 0.05 mm (smaller than most mica crystals in medium- to high-grade metamorphic rocks). These samples are then washed in deionized water to remove any adhered powder, and the minerals of interest are separated from one another by standard magnetic and gravitational techniques. Mechanically separable chlorite is removed from the biotite samples, however, it is possible that some chlorite is interlayered at a very fine scale and remains in the biotite samples (see Veblen & Ferry, 1983). No attempt has yet been made to separate white-mica phases (muscovite and paragonite); however, we have for the most part avoided sample suites in which paragonite is known to occur as a separate phase. Purity of mineral separates is determined by examination under a binocular microscope or by x-ray powder diffraction techniques, and only mica samples of greater than 95% purity are analyzed. Some samples are run as whole-rock powders, specifically, relatively fine-grained metasedimentary rocks (e.g., Bebout & Fogel, 1992; Bebout et al., 1999a, b) and rocks such as Kmetasomatized ultramafic rocks (e.g., fuchsite-bearing talc schists; see Bebout, 1997) and metamafic rocks (e.g., metaconglomerate cobbles, with gabbroic protoliths with
352 pseudomorphs of white-mica after plagioclase; Bebout & Barton, 1993; Bebout, 1997) in which mica is known to be the only likely mineral host for N. Samples of 20 to 1000 mg (occasionally larger for low-N samples) are loaded into quartz tubes (9 mm O.D., 7 mm I.D., 23 cm long) with Cu and CuO (2.5 g CuO, 4 g Cu). Some samples with higher N concentrations are loaded into quartz tubes with 6 mm O.D. and containing somewhat smaller amounts of reagent. The tubes are then evacuated for two hours while being warmed intermittently with a hightemperature heat gun and sealed under vacuum. The ends of the tubes containing the samples are heated in a resistance tube furnace (Deltech Inc. Model DT-28-HT, with heating elements capable of achieving > 1500~ for thirty minutes at 1200 ~ - 1250~ in order to fully fuse the micas and extract all N. Heating of the samples at these temperatures for periods longer than 30 minutes does not result in the release of Figure 16.1 - Calculated yield (in p p m N) of N from biotite mineral separates (all from Townshend Dam, Vermont) combusted to a variety of temperatures. Samples analyzed at 910~ were heated for three hours at that temperature; samples heated to higher temperatures were heated for thirty minutes in a tube furnace and then heated to 850~ with CuCuO reagents to ensure proper speciation of all gases. Heating at the higher temperatures for longer periods produces no additional N yield. The small numbers on the figures show the 615N values for each analysis (note the summaries of these data, comparing the results for the low-T and higher-T extractions, in Tab?e 16.1).
Chapter 16 - G.E. Bebout & S.J. Sadofsky
lO-ui_-u
II
II
I IIII
I II
IIII
I II
IIII
II
E
800 ...-"
"=n,.~,.n."
Z
2.7
II
I
.P
j"
...P-'
--'---
94 . 4
4.5
'--' J8 ,l
403
VT 9G 7a Bi ot ite
203 1";J30
....
1C00
:3.4 32
,,I.n.L.
E
III
,j-"l
1 ....""
>-
I II
....-"
8~n
n i n
I IIII
4.4
..N,.I.L.
n ....
I ....
I ....
I~1 ....
,
n nn n
'~" .3
___-8
.-
n l n nn n lip
04,5
..----'4 ..'5_,, 4 . 2 , 4 . 3
9
....-'-
45
800 3"2
t~uu
/
>-
400
Z
200
:3 2
i n
VT 96~ 71:., Bi otite
1400 .,=.1,..
E
.-,.~..,. /
~3
>-
,.
,:
I ,.
in
lln
Ii
In
I,
Inn
n.
il
i in
nl
,.n
nl
,
.in
O
_.._----~, , , - - " - - - 4 . 5
1200
e4. 3
._.~'
1000
=,--
2B
8C8 600
i /
Z
400
. . . I.I . .I. .I I.I I. . . . .I . .n l. . l.l l.l . . . . I. .I.I . .i l l l
II
12f'n 0 . A .
E
81~1_-I
e.--
I|
I III
...-,f-
4
.7
III I. .I I. . .I I.I . . |
4.5,
4B,
II
9
413
I "z
,
4.6 -.
+
_B
4.5
.=.==""=
6Ell
>-
4CEI
Z
20-1
n l
4.3/"
1Cn-uO
~
n
VT 96- 7c Biotite
21313
0
:-: .3
VT ,<-,~G7d Biotite
9_re . . . . ' . . 100,,' .. .... "1i/30
" . . .lz~JO' . . . . . ' . .1.'_~_0 ..
T ernperalure (C)
353
815N Analyses of Ammonium-Rich Silicate Minerals by Sealed-Tube Extractions ... Figure 16.2 - Calculated yield (in ppm N) of N from muscovite mineral separates (all from Townshend Dam, Vermont) combusted to a variety of temperatures. Samples analyzed at 91~ were heated for three hours at that temperature; samples heated to higher temperatures were heated for thirty minutes in a tube furnace and then heated to 850~ with reagents to ensure proper speciation of all gases. As with the biotite data, heating at the higher temperatures for longer periods produces no additional N yield. The small numbers on the figures show the 615N values for each analysis (note the summaries of these data, comparing the results for the low-T and higher-T extractions, in Table 16.1).
additional N and increases the risk of tube rupture due to melting, which weakens the quartz tubes, and differential thermal contraction of the silicate melt and quartz tube during cooling. The sealed tubes are then heated to 850~ for one hour in a programmable muffle furnace and cooled slowly to ensure proper speciation of gases (see Bebout & Fogel, 1992).
400
~-~
"~ ~) ~..
4.5
30Ci
,p
iI
VT 9G 7a M u s c o v i t e l
l
ll
. . . .
I
ll
I
. . . .
300
i
i
21:13
I
I
I
. . . .
l
ll
I
I
I
. . . .
I
l
l
II
l
l I
I . . . .
l
I
l
II
l
. . . .
I
l
I
. . . .
l
l
II
l I
l
l
. . . .
f
[
.=. 11,
__._-----~"- ::I.5
l
4.1
VT 9 6 71:: Musc,gvite
100
41_-r
I
"2_, "? a,_i ..."
f
Z
l
I
I "--
4 .4
= ~ -
1 El3! 400
E
3 ~___---99 -----
203
""'""'"""'"1'""'"'1'""'"'
--
i
~ 31-m3 =
i
i
. . . .
i . . . .
i
. . . .
9
:3 7 / - - - - -
i
. . . .
i
. . . .
I
4.3
4.2 I_._~Temperatures of > 1200~ have been determined as necessary in order " ~ 2Ci3 r 94 to achieve full release of N from higher- ._. grade metamorphic and igneous mus- ~-. covite and biotite separates (see Fig- Z 1 CO VT c ,-. 7d rn us; oov ite ures 16.1 and 16.2 and Table 16.1). This finding is consistent with the results of O_m_ 101ZlI-I 1100 1 :;130 1. ~ 0 vacuum stepped-heating studies of N release from micas (Boyd et al., 1993a; T ernp erat u re (C) see Figure 15.4 in Boyd, this volume), in which the bulk of the N release has been documented to be in the 900 ~ or 1000~ steps below the tem-perature at which sintering of the micas began (~ 1100~ Smykatz-Kloss, 1974) but significant amounts of N are retained to higher temperatures. This higher-temperature (> 1200~ procedure does not appear to be necessary for analyses of whole-rock, mica-rich samples and sam-ples containing more celadonitic and Cr-rich white micas from low-grade suites. For these materials, we have found that heating at 910 ~ - 950~ for three hours quantitatively releases the N and produces isotopic and concentration data indistinguishable from those obtained by the higher-T (1050~ - 1250~ heating methods (see discussions of yields for the 910~ experiments in Bebout & Fogel, 1992). The similarity in yield in our experiments on the samples containing more celadonitic white. . . .
.
. . . .
,
. . . .
,
. . . .
.
. . . .
.
. . . .
,
. . . .
|
. . . .
354
Chapter 16 - G.E. Bebout & S.J. Sadofsky
Table 16.1 -Demonstration of N-Isotope results discussed in text. Sample mean
815Nair (per mil) (std. dev.)
Concentration (ppm) mean (std. dev.)
SL-1 (fuchsite; see Figure 16.3) 910 (n = 5) ___1100 (n = 9)
2.24 2.32
0.13 0.23
2039 2052
45 (2%) 101 (5%)
WE-2 (fuchsite) 910 (n = 11)
1.63
0.10
2197
147 (7%)
0.15 0.18
624* 635
22 (4%) 20 (3%)
T54 (Silurian/Devonian Graptolite shale whole-rock) 910 (n = 7) 0.56** 0.14
4183"*
259 (6%)
20019 20168
715 (4%) 590 (3%)
WS-1 (metasedimentary whole-rock)f 910 (n = 40) 2.93* >1050 (n = 2) 2.83
Buddingtonite-2 910-950 (n = 29) ___1100 (n = 6)
-1.59 -1.22
0.23 0.21
Mica Separatestt- Townshend Dam, Vermont (from Sadofsky & Bebout, 2000) VT96-7a biotite 1250 (n = 6)
4.50
0.14
873
41 (5%)
VT96-7a white mica _>1200 (n = 2)
4.46
0.02
256
17 (7%)
VT96-7b biotite ___1100 (n = 6)
4.38
0.12
1134
30 (3%)
VT96-7c biotite _ 1200 (n = 2)
4.37
0.12
1286
44 (3%)
VT96-7d biotite ___1200 (n - 6)
4.54
0.05
1117
28 (3%)
VT96-7d white mica _ 1100 (n = 2)
4.25
0.07
299
12 (4%)
* K. Brauer reports +2.9 per mil mean ( l o - 0.1) and 550 p p m (lo = 30 ppm). ** K. Brauer & K. Hahne report +0.95 per mil (lo = 0.25) and 4000 p p m (lo = 300 ppm). f Combining the data from this lab with additional analyses obtained at the Geophysical Laboratory (during the period of 1990-1993). t data are presented in Figures 16.1 and 16.2.
355
815N Analyses of Ammonium-Rich Silicate Minerals by Sealed-Tube Extractions ... ;-II-II-II-I
I
i
I
-
E -,~,-9
I
i
"]
n
,n
2052 + 101 I:,F.rn I~%)
-
..- JUU
i
[" SL-1 (Fu ct]si te E,I:r~,datd)
275n-I ,-I1~
I
Io;i
Iq = L:! 9
,T...Ol-llI
.
..........................................................................
. . . . . . . .
~,)
1750
,
1.'__ql-ll-I
"1
% i,
3-hour headng
i
3 0 - r n i n u t e FiealJ rig, fol Io,~~ed h y 8 5 0 ' C : I"l_-II.,I ~1"18
1250 1 I_-II-II-I 91:10
I
I
1000
4.13
n
I
1100
I
n
"
3.5
12110
I
n
1300
I
n
'l
SL- 1 ,[R.lcil.si re_.E,t;~ldard)
., i
,=
3.0 i"
II
,::.. '-' 5
l~.,
.........................................................
2 .U
9
9 ............................ o
+2 ' 2 +
01.
'='1
II tl
...... ...o.....o......
+2.3+
.
.
.
0
9
. nC,
1.5
,0
900
i
i
i
1000
1100
1200
1300
Teml:.,erature (~C) of E.xperimer~t Figure 16.3 - Calculated yields (a) (in ppm N) and 615N values (b) for SL-1, a sample of fuchsite (Cr-rich, phengitic white-mica from the Catalina Schist; see Grove & Bebout, 1995) combusted to a variety of temperatures. Samples analyzed at 910~ were heated for three hours at that temperature; samples heated to higher temperatures were heated for thirty minutes in a tube furnace and then heated to 850~ with reagents to ensure proper speciation of all gases. The small numbers on the figures show the 815N values for the each analysis (note the summary of these data, comparing the results for the low-T and higher-T extractions, in Table 16.1).
356
Chapter 16 - G.E. Bebout& S.J. Sadofsky
micas at the lower (910~ and higher temperatures (a 1100~ is demonstrated in Table 16.1 for sample SL-1, a fuchsite-bearing sample of ultramafic schist (also see Figure 16.3), and for WS-1 (two higher-temperature extractions, at 1050~ and 1100~ a medium-grade metasedimentary sample provided by K. Brauer (University of Leipzig; see her data in a footnote in Table 16.1) and analyzed in her laboratory using the dissolution/distillation techniques employed by Haendel et al. (1986). Consistent with this observation, more celadonitic white-mica is known to release Ar at lower temperatures than those for Ar release from muscovite (M. Grove, personal communication, 1999; see discussions and Figure 5 in Wijbrans & McDougall, 1986, who documented enhanced stability of muscovitic white mica relative to celadonitic white mica during "in vacuo" incremental heating; also see Grove, 1993; Grove & Bebout, 1995). Fine-grained, whole-rock samples of micaceous metasedimentary rocks (even those containing more muscovitic white-mica) are (based on visual examination of the tubes following the extractions) believed to experience melting (in part reacting with the quartz tubes) at 910 ~ - 950~ thus promoting enhanced N release at the lower temperatures. At the Geophysical Laboratory, reaction products from the 910~ experiments for several whole-rock metasedimentary rocks (all showing obvious evidence of having undergone melting; samples 6-2-27a, 6-4-68H, and 6-3-41' with a wide range in their white-mica compositions from phengite to muscovite) were reloaded into 9 mm quartz tubes with reagents and reheated at 1000~ (90 minutes then slow cooling) to test for incomplete reaction in the initial experiments. The resulting amounts of N2 were similar to only slightly larger (all ~ 0.3 ~moles, most too small to analyze for their 615N) than those of the blanks (~ 0.1 ~moles at that time; note that the N2 samples extracted in the first experiments were all > 40 ~moles). One of the N2 samples from these reruns (for 6-3-41') barely large enough (~ 0.3 ~moles) to obtain reliable isotopic data yielded a 615N of +3.6%o similar to, but slightly lower than, the value obtained in the initial extraction (+4.0%o; see Table I in Bebout & Fogel, 1992). Interestingly, buddingtonite (an ammonium feldspar; see Voncken et al., 1993) does appear in our experiments to yield somewhat differing concentrations and ~15N values in extractions employing combustions at 910~ (for three hours) and in those employing combustions at > 1100~ (see data for Buddingtonite-2 in Table 16.1). The differences are consistent with the retention in the samples of a small amount of isotopically fractionated (likely isotopically heavy) N to temperatures exceeding 910~ as was indicated by Boyd et al. (1993a; also see Figure 4: in Boyd, this volume). In the stepped-heating runs for K-feldspar of Boyd et al. (1993a), less release of N occurred at temperatures below the melting point of the minerals (~ 1150~ and the bulk of the N was retained in the feldspar samples to melting temperatures. Above the melting temperature, N continued to be released with further heating (with all N apparently being released by ~ 1250~ as the viscosity of the feldspar melt decreased (discussion by Boyd et al., 1993a). For both the mica and the K-feldspar extractions, N release below the mineral melting temperatures is thought to be diffusive or perhaps more complex, involving a combination of diffusive and instantaneous release over a wide temperature range (M. Grove, personal communication, 1999; cf. Boyd et al., 1993a). Diffusive release, in general, favors the release of relatively light N and retention of heavy N in the residual solids - this retention to higher temperatures of an isotopically
815NAnalysesof Ammonium-RichSilicate Minerals by Sealed-TubeExtractions ...
357
heavy N component is demonstrated by the trends of N concentration and isotopic composition for micas showing varying degrees of N degassing in Figures 16.1 and 16.2. After combustion, N2 is cryogenically separated from the CO2 and H20 in the gas samples, using standard cryogenic techniques (at liquid nitrogen temperature; e.g., Macko, 1981; Bebout & Fogel, 1992; Boyd et al., 1993a). Purified samples of N2 gas are transferred by molecular sieve from a glass vacuum line to the inlet of the Finnigan Mat 252 mass spectrometer. Large N2 samples (> 7 ~moles) are transferred into the variable-volume (bellows) of the standard inlet system and run though the standard procedure. Small samples (0.5 - 2 ~moles) are frozen into a microvolume (cold-finger) inlet containing silica gel. Blanks for this technique have been reduced to < 0.025 ~moles N2 for routine large samples in the 9 mm quartz tubes and are thought to be related to static leakage during extractions (primarily on O-ring fittings; blanks are lower on the average for runs utilizing 6 mm quartz tubes and smaller, better-fitting 1/4-inch diameter O-rings) on the glass vacuum line. Variations in the isotopic composition of N are defined as: ~15N - I(15N/14N)spl - (15N/14N)std1103 (lSN/14N)st d
[16.1]
where the standard is atmospheric N2. Accuracy of our measurements of 615N is maintained by routine, repeated analyses of atmospheric N2 (615N - 0%0), two international N-isotope standards (NBS N2-ammonium sulphate with 615N = +20.41%o; USGS 32 potassium nitrate with 615N - +179.1%o), and one secondary, laboratory standard TCH-1 (ammonium sulphate from the Geophysical Laboratory with 615N = 0.15No). Low-grade, metasedimentary whole-rock, powdered sample WS-1 (with 615N - +2.9%o, provided by K. Brauer) has also been used, and 40 analyses of this sample over the last eight years, using the very different extraction lines and mass spectrometers at the Geophysical Laboratory (with the modified, double-focussing DuPont 491 mass spectrometer) and in the Lehigh University laboratory (using a Finnigan MAT 252) produced mean 615N - +2.93%0 (lcJ- 0.15; mean N content - 624 ppm, with lo of 22 ppm or 3.5%; see Table 16.1). Precision in 615N of the analyses of relatively high-N biotite and muscovitic whitemica separates (samples with >100 ppm N) ranges from lo - 0.02 to 0.14%o (n - 2 to 6; Table 16.1), but lo is < 0.35%o for analyses of lower-N samples (10 - 100 ppm N). Reproducibility of isotopic data (with lo nearly always ___0.2%o and commonly 0.1%o) and concentration data for several laboratory silicate standard materials and some higher-N micas (from an amphibolite-facies metasedimentary exposure at Townshend Dam, Vermont; Sadofsky & Bebout, 2000) are demonstrated in Table 16.1. For samples SL-1, WE-2, and WS-1, sample sizes used in the extractions varied by nearly an order of magnitude (e.g., for sample SL-1, 19-168 mg) and both the variablevolume (bellows) inlet and microvolume inlet on the mass spectrometer were used (each inlet with its own calibration for determinations of concentrations). Concentra-
358
Chapter 16 - G.E. Bebout& S.J. Sadofsky
tions are obtained by measurement of voltage on the m / z 28 peak for calibrated volumes in the mass spectrometer (either a variable-volume/bellows or microvolume inlet); voltage vs. gmoles N2 is calibrated by extractions and analyses of non-silicate standards with known N contents (usually combinations of ammonium sulphate and phenylalinine yielding calibration lines with r2 > 0.98). Uncertainties (again, expressed as lo) for N concentration data are nearly always < 5% (typical concentration data for various sample types are presented in Table 16.1). We have also constructed an all-metal, low-blank extraction line which is mounted on a cart and can be attached directly to the inlet of the mass spectrometer. This system is used for analyses of low-N materials for which experiments (fusions with and without addition of fluxes; cf. Zhang, 1988) involve large amounts of low-N sample (> 2 grams) and for extractions of N2 from fluid inclusions. Sealed-tube extractions of N from fluid inclusions in quartz veins, employing heating to 910 ~ - 1000~ for up to three hours (with Cu and CuO reagents), have been extremely successful in producing reproducible isotopic compositions (see earlier N-isotope work on fluid inclusions in rocks from the Dome de l'Agout, France, by Kreulen et al., 1982). In collaboration with J. Touret, M. Moree (both at Vrije Universiteit, Amsterdam) and M. Sintubin (Katholieke Universiteit Leuven, Belgium), we have performed analyses of the 615N of the generation of N2-rich fluid inclusions (a 93 mole % N2 with small amounts of CO2 and / or CH4) and coexisting biotite ("bastonite" with 2280 ppm NH4 +) in a quartz vein from these exposures. Three sealed-tube decrepitation experiments yielded 615N values of +3.7, +3.9, and +4.0%o for the quartz-vein fluid inclusion N2, and a single biotite analysis (using the techniques described in this paper) yielded a 615N value of +6.8%o. The difference in 615N of the N2 and biotite (mean vein ~15N is +3.9%o, thus making the inclusion-biotite difference ~ 2.9%o) is extremely similar to the A15NNH4+N2 calculated by Hanschmann (1981) for the petrologically inferred temperatures of formation of ~ 400~ (~ +2.8%o; indicated on Figure 16.4). 16.3 Outlook for future applications of N isotopes in studies of high-T silicate systems As in many other subfields of stable isotope biogeochemistry, upcoming advancements in both extractions and mass spectrometry will further widen the range of potential applications of N isotopes to an increasing number of petrologic problems. In particular, the application of lasers in heating experiments (Hashizume & Marty, Chapter 17 of this volume; Humbert et al., 2000) and the use of gas chromatographcombustion systems interfaced with carrier gas systems (see Brand, 1995b) should afford more routine and rapid applications of N isotopes (see early experimentation with elemental analyzer-continuous flow-IRMS methods by Jia & Kerrich, 1999; Kyser et al., 2000). The techniques we describe in this paper (and similar high-temperature "sealed-ampoule" extraction techniques employed by Boyd & Pillinger, 1990, and Boyd, 1997, in analyses employing static mass spectrometry) can easily be applied to a variety of other N-isotope petrologic pursuits, with (for example) extensions to facilitate extractions of N from feldspars, cyclosilicates (containing channel N2; Scalan, 1958, reported beryl with up to 234 ppm N and cordierite containing up to 125 ppm N; also see analyses of channel fluids by Damon & Kulp, 1958; Giuliani et al., 1997),
815N Analyses of Ammonium-Rich Silicate Minerals by Sealed-Tube Extractions ...
359
Figure 16.4 - Calculated fractionations (A15N - h15NNH4+- •15NN2,NH3 ) among N fluid species and NH4 +, the latter of which is bound structurally in silicate minerals in metamor-phic rocks (fractionations are from Hanschmann, 1981). The stippled horizontal and vertical lines indicate the excellent match of the calculated fractionations with the measured A15N for the "bastonite" (NH4+-rich biotite) and fluid inclusion N2 in quartz veins at Bastogne, Belgium (see description of veins and fluid inclusion chemistry in Darimont et al., 1988) for estimated vein formation temperatures.
and fluid inclusions (see Figure 16.4 and Kreulen et al., 1982). The precision obtainable for N isotopes using our methods (with lo commonly approaching 0.1%o; see Table 16.1) is similar to that obtainable for O in silicate systems (lo of 0.1 to 0.2%0 for n ~ 4; see 8180 data for garnet and quartz from various laboratories for laser probe methods and methods involving externally heated Ni reaction vessels in Valley et al., 1995), and the degrees of isotopic variation for the two systems in a given rock suite can also be quite similar (on the order of 5 - 10%o overall variation; see comparison of the O and N isotope systematics in the Catalina Schist, California, by Bebout, 1997; ~ 8%o 815N range in the Skiddaw aureole and granite, Bebout et al., 1999a; ~ 10%o 815N range at the Townshend Dam metamorphic locality, Sadofsky & Bebout, 2000). Two studies (Richet et al., 1977; Hanschmann, 1981) have published calculated Nisotope fractionation factors based on spectroscopic data. Of the two theoretical studies, only Hanschmann (1981) calculated fractionations involving NH4 + in solid phases. The calculated N-isotope fractionations among various N molecules based on spectroscopic data (see Figure 16.4) predict substantial N-isotope fractionation as a result of devolatilization, metasomatic alteration, and magmatic volatile release leading to 815N variation of the magnitude observed in recent integrated petrologicgeochemical studies employing N isotopes (see Javoy et al., 1986; Bebout & Fogel,
360
Chapter 16 - G.E. Bebout& S.J. Sadofsky
1992; Boyd et al., 1993a; Bebout, 1997; Bebout et al., 1999a,b; Sadofsky & Bebout, 2000). Conspicuously lacking for applications of N isotopes are experimentally determined fractionation factors for appropriate fluid-melt-mineral systems. The experimentally derived fractionation data and international N-isotope silicate standards will both be required in future applications of N isotopes to petrologic systems, regardless of the future direction of the analytical methods. However, one should consider applying N isotopes, particularly in studies of relatively potassic and organic-rich rock systems (e.g., metasedimentary and felsic-intrusive systems; K-metasomatized metamafic and metaultramafic rocks), to complement other geochemical data, just as one would consider the other more commonly used O, H, C, and S isotope systems in settings for which compositions are appropriate and there is some expectation that N isotopes will yield unique, useful constraints.
Acknowledgements
GEB acknowledges early support of this research by M. Fogel, the late T. Hoering, and the Geophysical Laboratory (Carnegie Institution of Washington, Washington, D. C.). Continued work has been supported by the National Science Foundation (grants EAR-9206679, EAR-9220691, EAR-9405625, EAR-9406135, and EAR-9727589), with some support also coming from the American Chemical Society Petroleum Research Fund (grant #25246-G2). GEB thanks M. Fogel, T. Hoering, S. Boyd, M. Grove, B. Idleman, and P. Zeitler for helpful discussions, and we thank Karin Brauer for providing several samples for interlaboratory comparison and Pier de Groot for undertaking this ambitious and extremely useful publication project.
Handbook of Stable Isotope Analytical Techniques, Volume 1 P.A. de Groot (Editor) 9 2004 Elsevier B.V. All fights reserved.
CHAPTER 17 Nitrogen Isotopic Analyses at the Sub-Picomole Level Using an Ultralow Blank Laser Extraction Technique Ko Hashizumel,2,4* & Bernard Martyl,3** 1 Centre de Recherches P6trographiques et G6ochimiques, 15 Rue Notre-Dame des Pauvres, B. P. 20, 54501 Vandoeuvre-16s-Nancy Cedex, France 2 Department of Earth & Space Sciences, Osaka University, Toyonaka, Osaka 560-0043, Japan 3 Ecole Nationale Sup6rieure de G6ologie, Avenue du Doyen Roubault, 54501 Vandoeuvre-16s-Nancy Cedex, France e-mail: *
[email protected]; **
[email protected]
Abstract We describe our recent achievement in determining the isotopic composition of nitrogen contained in extremely small samples. The aim of our development is to enable isotopic analysis of single mineral grains using the smallest quantity of nitrogen with a precision sufficient to resolve isotopic variations of nitrogen in extraterrestrial samples. The limiting factors are (i) residual nitrogen generated during extraction and purification of nitrogen, and (ii) mass interferences from hydrocarbons, N2H and CO at masses 28 and 29. The use of a defocused CO2 laser as a heating source allowed circumventing the first problem. Hydrocarbon (C2H5) is mass-resolved with our mass-spectrometer. Contribution of N2H was reduced dramatically by optimizing the ion source setting. CO is efficiently removed during purification and its residual contribution is corrected numerically using measurements at mass 30 in addition to masses 28 and 29. We are able to perform isotopic analyses of sub-picomole (< 30 picogram) quantities of N2 with a precision typically of + 10%o. The required amount of samples for nitrogen isotopic analyses, in the case of the lunar regolith, is reduced by a factor of-10-5 compared to the pioneer's works in the 1970's.
17.1 Introduction Nitrogen trapped in extraterrestrial materials exhibits large variations in the 15N/ 14N ratio, allowing to set strong constraints on the origin and evolution of the solar system and of the planets (e.g., Hashizume et al., 2000). The 615N values (which stand for the permil deviation of the 15N/14N ratio relative to terrestrial atmospheric nitrogen) range between-300 to +100%o among lunar grains (e.g., Kerridge, 1993; Hashizume et al., 2002), -200 to +1600%o among bulk meteoritic samples (see references in Hashizume et al., 2000), or even wider (log-scale variation of the absolute ratio, i.e., 5 4. Correspondence should be adressed to this author
362
Chapter 17 - K. Hashizume & B. Marty
< 14N/15N < 104) among the presolar grains contained in primitive meteorites (e.g., Zinner, 1998). Systematic studies of the nitrogen isotopic composition among the extraterrestrial materials started in the 1970's with the analyses of the regolith samples returned back from the Moon (e.g., Becker & Clayton, 1975), followed by bulk analyses of meteorites (Kung & Clayton, 1978). At that time, analyses were performed using dynamic type mass-spectrometers, which required 0.1 - 3 grams of the precious samples for each analysis. Later on, technical improvements were dedicated to reduce the sample size, by increasing the sensitivity of the mass-spectrometry system, as well as by reducing the blank levels, and to better resolve the indigenous nitrogen components from the terrestrial contamination. The sensitivity for nitrogen analyses increased dramatically following the development of static-type mass-spectrometers (e.g., Frick & Pepin, 1981; Wright et al., 1988; Hashizume & Sugiura, 1990; Murty & Goswami, 1992; Boyd et al., 1993b; Craig et al., 1993; Marty et al., 1995; Takahata et al., 1998; Yamamoto et al., 1998; Humbert et al., 2000; this study). The deconvolution of different nitrogen components has been improved by the introduction of the stepwise combustion technique (e.g., Frick & Pepin, 1981; Boyd et al., 1988; Boyd, Part L Chapter 13), or by using non-heating extraction methods such as vacuum crushing (e.g., Marty et al., 1995). In this paper, after presenting the general picture of the nitrogen isotope analysing system in CRPG-Nancy, we highlight our recent improvements concerning our ultralow blank extraction system and the mass spectrometry of sub-picomole quantity of nitrogen. 17.2 Overview of the N analyzing system The outline of the system and the analytical procedures are described here, although most of the details are described elsewhere (Humbert et al., 2000). The nitrogen isotope analysing system is described diagrammatically in Figure 17.1. It consists of a laser heating unit, an all-metal purification line, and a high-sensitivity static-type mass spectrometer.
Samples in the laser chamber are loaded in small pits machined in a stainless steel. The samples are heated by a defocused CO2 laser, working in a continuous mode with a wavelength of 10.6 gm. In case samples require stepwise analyses, several steps are performed before the final fusion step. The heating temperature is adjusted by inserting variable length of interval (e.g., 20 - 200 gsec) between laser emission (which last for 100 - 1000 ~sec per cycle). The temperature of the heated sample is estimated from its color and brightness. Samples are heated in oxygen atmosphere (P02 > 0.1 Torr) generated from CuO heated at > 720~ (combustion mode), or in vacuum (pyrolysis mode). Organic compounds released from the heated sample together with the sample gas (N2 and rare gases) are dissociated to oxide molecules (CO2, SO2, H20 etc.) in contact with the hot CuO, then are removed from the sample gas using the cold trap, a U-shaped glass tube cooled at -183~ The partial pressure of 02 is then lowered by slowly cooling down the CuO to 400~ Sample gases are normally split into two, N2 and Ar fractions, by volume dilution. These fractions are purified and introduced into the mass spectrometer sequentially. The impurities in the argon fraction are chemi-
Nitrogen Isotopic Analyses at the Sub-Picomole Level Using an Ultra-low Blank Laser ...
sorbed by two Ti-getters (GT1 and GT2) heated at 600- 700~ For the nitrogen fraction, no trap is used else than the CuO/ cold traps. Before introducing the gases into the mass-spectrometer, the gas pressure is monitored with an ion gauge having a tungsten filament working at a low filament current of 200 ~A. (The filament does not seem to pump out or emit measurable amount of nitrogen gas under the current condition, which is verified by the linear correlation between the nitrogen intensity measured by the massspectrometer and the pressure of standard air, adjusted by splitting the gas between known volumes.) The amount of nitrogen to be introduced into the mass-spectrometer is adjusted in order to get a signal compatible with the mass-spectrometer dynamic range. A rare gas mass spectrometer (VG 5400, Micromass 9 working in a static mode is used for the
363
Figure 17.1 - Diagram of the N / rare gas analytical system developed at CRPG-Nancy. The system is evacuated by ion pumps (IP) and turbo molecular pumps (TMP) connected to rotary pumps (RP). PG, IG, CT and GT denote Pirani gauge, ion gauge, charcoal trap and Ti-getter, respectively. The introduction line which includes the vacuum crushers is used for other extraction procedures and is not described in this contribution. Charcoal trap and helium cryotrap are used when analyzing light rare gases (He and Ne).
364
Chapter 17 - K. Hashizume & B. Marty
analyses. It has a Nier-type ion source, a bent tube of 0.54 m radius, across which a magnetic field is delivered by an electro-magnet stabilized with a Hall probe, and two collectors, a Faraday cup and an electron multiplier working at 2.5 kV. Its extended geometry allows a high mass resolution of 650 at 2% valley on the electron multiplier collector, whereas the resolution obtained on the Faraday collector is 250. On the Faraday collector, the output voltage is measured through a 1011 f~ resistor, whereas the count rates of electron burst output from the electron multiplier (dead time: 13 nsec) in response to the incident ions are measured by a pulse counting system (Orted~ 996). Comparing the count rates of ions detected by the two collectors (ion currents divided by the charge per ion, in the case of the Faraday collector), the yield of the electron multiplier collector was usually observed to be around 0.80 - 0.85 times that of the Faraday collector, which is normal for this type of collection. The source settings were 10 - 100/~A for the trap current, and 60 eV for the ionization energy. The trap current was reduced when we analyzed small quantities (e.g., sub-picomole) of N2, in order to avoid production of interfering N2H (See discussion in the "Nitrogen Mass-Spectrometry" section). Peak heights of N2 isotopes (14N14N, 14N15N and 15N15N) and interfering species are measured 1 0 - 15 times repeatedly. The total time required to complete the nitrogen analysis is 15 - 20 minutes. The mean sensitivity for N2 (the current measured on the collector divided by the partial pressure of the gas in the mass-spectrometer) ranged from 1.8 x 10-5 A/Torr at a trap current of 10 gA to 2.6 x 10-4 A/Torr at 100 gA. The half-life of N2 in the ion source was 350 and 25 min. for trap currents of 10 gA and 100 gA, respectively. In the standard procedure, hot blank (discussed in section 17.3) and standard gas are measured at least once a day, respectively before and after beside the sample gas analyses. Standard measurements are performed following the same procedure as for the sample gas analyses. A pipette (4 x 10-10 mol N2) of standard air, prepared from atmospheric air and stored in a vial (STD2), is taken for the measurement. The amount of the standard gas to be introduced into the mass spectrometer is adjusted by volume dilution to be comparable to that of the sample gas. 17.3 Low blank gas extraction using a laser A critical issue which controls the quality of analyses of small amounts of nitrogen is the hot blank level, which is the amount of gas released when the extraction and purification procedures are performed without a sample. Several sources can contribute to the blank, among which are the "hot spots" in the vacuum line, e.g., the extraction chamber, where the nitrogen contained in samples is extracted by heating, copper oxide, or the vacuum gauge. After the vacuum line had been well degassed by cyclic baking under high vacuum, the sample gas extraction part became the largest blank source. In order to reduce the blank generated during the extraction procedure, we used a defocused laser beam as a heating device. Franchi et al. (1986) also developed a laser system to extract nitrogen and other volatiles. They used a laser beam to heat a small area of a large sample, whereas we use it to homogeneously heat a small sample. The CO2 laser beam is absorbed by the silicate sample, while the part of the beam arriving at the surrounding stainless-steel surface is reflected and scattered. Thus only
Nitrogen Isotopic Analyses at the Sub-Picomole Level Using an Ultra-low Blank Laser ...
365
the silicate sample is heated, minimizing the source of hot blank. The variation of the hot blank level of our system is s h o w n in Figure 17.2. Periods w h e n a series of analyses were p e r f o r m e d after the clean-up procedure (described below) are s h o w n by the hatched bars. The sample chamber, after fusing all the samples loaded in the pits, is opened and w a s h e d with acids (which consist mainly of fluoric acid and a small a m o u n t of nitric acid), distilled water and acetone before reloading new samples. It normally requires 2 - 4 days, after the installation of the chamber to the v a c u u m line, to reduce the blanks to acceptable levels. The installed chamber is baked at 120~ at least overnight, often for two days. We often performed a "pre-combustion" procedure, introducing ~ 1 Torr of 02 into the hot chamber, connected to the cold trap (Figure 17.1). This procedure promotes removal of adsorbed air, organic contamination and water from the samples and from the stainless steel apparatus (At the baking temperature of 120~ most of the organic matter is not com-
Figure 17.2 - Hot blank level of nitrogen plotted as a function of time. Hot-blank analyses are carried out following exactly the same procedure as that applied to the samples. Data plotted on days within a hatched bar represent a series of analyses performed after the clean-up procedure subsequent to the reloading of samples. Intervals between analyses series are taken arbitrarily. The h15N values of the blank nitrogen were constantly lower than- 30%o during the first five series of analyses, as a probable result of contamination from nitric acid used in the cleaning procedure. Subsequently, the use of additional cleaning treatments (see text) reduced the N blank level, and resulted in blank h15N values close to 0%0.
366
Chapter 17 - K. Hashizume & B. Marty
busted. Nevertheless, we believe that removal of organic matter weakly adsorbed on the surface of the stainless steel is facilitated by the pre-combustion procedure). We verified that the major source of the blank nitrogen in combustion mode is not the CuO, which provides the pure oxygen. Indeed the "cold" blank level, the amount of nitrogen released by the purification procedure itself, was lower than 5 picogram, well below blank values observed during combustion experiments (Figure 17.2). We observed that the blank level was primarily determined by the cleanliness of the sample chamber on the one hand, and the mode of extraction, vacuum pyrolysis or combustion, on the other hand. Initially, we performed analyses in the combustion mode, heating the sample under an oxygen atmosphere ranging from 0.05 to 0.2 Torr. The blank level in this mode was generally higher than that obtained by heating the sample without 02 (pyrolysis mode), probably because the oxygen molecules mobilize residual nitrogen-bearing compounds adsorbed on the internal surface of the laser chamber. One of the problems encountered during the first five series of analyses was that ~515Nvalues of the blank nitrogen showed constantly negative values, -30%0 or lower. We suspect that it originates from the nitric acid added to the liquid to wash the sample holder. From the sixth series, the sample holder was boiled in distilled water several times after the acid treatment, and was washed again with a solution of an organic acid (COOH)2 to remove a possible chemisorption layer involving nitrogen. These additional treatments lowered the blank level by at least a factor of two (Figure 17.2). Subsequently, the 615N values of the blank nitrogen were observed to be constantly around 0%0. The blank nitrogen produced during vacuum pyrolysis was within 5 - 1 0 picogram (Figure 17.2). The amount of residual N obtained during sample heating may be higher than the blank level obtained when shooting the laser at an empty pit, because the heated sample can heat the sample holder by conduction. Though it is not easy to quantify the "true" blank level which involves sample heating (since no sample is granted to be perfectly nitrogen-free), we give an example suggesting that the radiation from the sample does not dramatically increase the blank level. Among the lunar grains we measured, several grains possessed extremely small amounts of nitrogen. When fusing such a sample in vacuum, the amount of nitrogen (including the blank) was 17 picogram, which is not dramatically higher than the hot blank range (5 - 10 picogram). In summary, we have achieved the lowest blank level among the systems used to analyze nitrogen isotopes. The literature hot blank levels range between 200 - 5000 picogram, while the typical blank level of our system is 5 - 10 picogram during pyrolysis, and 2 0 - 40 picogram during combustion, which enables nitrogen isotopic analyses of N2 in the picomole to sub-picomole range.
17.4 Nitrogen mass-spectrometry The nitrogen isotopic ratio of the extracted gas is determined basically by comparing the peak heights of N2 at masses 28 (14N14N) and 29 (14N15N). However, contribution of CO, N2H and hydrocarbons (C2Hx) interfering at these masses is critical, especially when the amount of N2 introduced in the mass spectrometer is small. The
Nitrogen Isotopic Analyses at the Sub-Picomole Level Using an Ultra-low Blank Laser ...
367
mass resolving powers (M/AM) required to discriminate these compounds from N2 at mass 29 are 5903 (13CO), 2688 (N2H) and 807 (C2H5). The best reproducibility for the 29/28 isotopic ratio is obtained when the N2 pressure exceeds 10-7 Torr (~ 300 picogram), i.e., when the contribution of the interfering species can be neglected. From the peak heights of N2 at masses 28 and 29 measured using the Faraday collector, the external reproducibility of the ~515Nvalue of standard air is typically 1 +_0.5%0 (Marry & Humbert, 1997; Dauphas & Marry, 1999). Below we describe the methods we developed to avoid or numerically subtract mass interferences at very low levels of N2.
Hydrocarbons: We are
able to separate partly the C2H5 peak from the 14N15N peak at a mass resolution of 650 using the electron multiplier collector. Figure 17.3 shows a mass-scan profile at mass 29 using a small amount of sample gas (which is actually ~ 3 picogram of blank N2 gas from the laser chamber). When > 3 x 10-9 Torr (7 picogram) of N2 is introduced into the mass-spectrometer working with a trap current of 100 gA, however, the ion counting system is saturated when analyzing N2 at mass 28. There are two countermeasures to cope with the problem. 1) Use of the Faraday collector for signals at masses 28 and 29, combined with additional analyses of two peaks, N2+CO+N2H and C2H5 at mass 29, using the ion
Figure 17.3 - Peak profile at mass 29 scanned over an interval of 0.1 amu for a blank N2 a m o u n t of -~ 3 picogram. The C2H5 peak is partly separated from 14N15N peak at the mass resolution p o w e r of 650 characterizing the electron multiplier collecter.
368
Chapter 17 - K. Hashizume & B. Marty
counting system. The ratio (28/29)N2+CO+N2H, corrected for interferences from hydrocarbons, is calculated as: (28 / 29)N2+CO+N2H - (28 / 29)Faraday X {1 + (29)CH/ (29)N2+CO+N2H}
[17.1]
where (29)CH and (29)N2+CO+N2H a r e the peak heights at mass 29 obtained in ion counting mode. 2) Reducing the trap current, for example to 10 gA, to enable the analysis of all N2 peaks with the electron multiplier. In the < 30 picogram range, the reproducibility of the nitrogen isotopic ratio of standard air analyzed in this mode was better than that obtained by using the first mode (The reason is described in the next subsection on N2H). The drawback of this mode is that it sometimes prevents simultaneous analyses of rare gases when their concentrations are low.
N2H: This species, along with CO, cannot be resolved from N2 by our mass-spectrometer. A small amount of N2H seems to be generated in the mass-spectrometer, although its occurrence depends largely on the ion source conditions. In this subsection, evidence for its occurrence is discussed in detail, because, unlike the interferences by hydrocarbon or CO, it is impossible to precisely correct for its contribution. In Figure 17.4, the 30/28 ratios of standard air N2 are plotted against the 29/28 ratios. This diagram exhibits a mixing relationship between N2, CO and N2H. With the trap current set at 10 ~A, all data lie on a mixing line between N2 and CO, suggesting that the contribution of N2H is negligible. In constrast, significant amounts of N2H are generated when the trap current is set at 100 ~A. The protonation rate can be reduced presumably by lowering the H2 partial pressure (e.g., Frick & Pepin, 1981). The dependence of the protonation rate on the trap current may be due either to the increase of the hydrogen partial pressure around the ion source, or to the increase of the reaction rate at 100 ~A, which are both likely to be a function of the filament temperature. Two observations suggest that the N2H observed in the standard air is produced around the ion source in the mass-spectrometer, not in the purification line or in the standard tank. First, the amount of N2H increases with increasing trap current, as stated above, and second, the 29/28 ratio is growing rapidly after introduction of the gas into the mass-spectrometer. The average growth rate of the 29/28 ratio was observed to increase with decreasing pressure of N2 in the mass-spectrometer (The growth of the ratio does not seem to be primarily due to mass-dependent isotopic fractionation processes, since the rate was observed to be much higher than the one expected by fractionation processes such as nitrogen consumption by the source filament). The average growth rate A(ln29/28)/dt was roughly proportional to the inverse square root of the N2 pressure, ranging from 1.5%o/min at PN2 - 2 x10-7 Torr to 15%o/min at PN2 - 2 xl0 -9 Torr. We interpret the linear correlation between PN2-1/2 and A(ln29/28)/dt as due to a rate determining process for the nitrogen protonation which involves the atomic form of nitrogen (e.g., N + H ---, NH, N + 1 / 2 H2 ---, NH or
HANDBOOK OF STABLE ISOTOPE ANALYTICAL TECHNIQUES VOLUME 1
Pier A. de Groot editor Economic Geology Research Institute, School of Geosciences, University of the Witwatersrand, Johannesburg, South Africa and Delta Isotopes Consultancy, Pastoor Moorkensstraat 16, 2400 Mol-Achterbos, Belgium (present address)
2004
ELSEVIER Amsterdam - Boston - Heidelberg - L o n d o n - New York - Oxford Paris - San Diego - San Francisco
- Singapore
- Sydney
- Tokyo
370
Chapter 17 - K. Hashizume & B. Marty
If the production of N2H was a linear function of residence time in the mass-spectrometer, the contribution of N2H could be corrected by taking the 29/28 ratio at time zero (introduction of N2 into the mass-spectrometer). Nevertheless, the contribution of N2H is still seen at time zero, which is demonstrated in Figure 17.4 where the ratios 29/28 and 30/28 extrapolated to time zero are plotted. Such effect seems to be due to the non-linear production of the N2H after introduction of the gas. The 29/28 ratio generally grew faster at the beginning, slowing down with time. The slower growth at longer time is possibly due to equilibration of generation and dissociation processes of N2H. The apparent protonation rate of N2 at time zero, represented by the N2H/N2 ratio, depends on the N2 pressure in the mass-spectrometer (Figure 17.5). At the N2 pressure range of ~ 10-8 Torr (~ I picomole), the protonation rate is variable and large enough to largely degrade the quality of nitrogen isotopic determination. For exam-
'
'
'
I
'
'
'
'
I
'
'
'
'
I
'
'
0
7-I
'
'
I
'
'
'
'
STD N2 (Under 100gA Trap Current)
@ 6 - ~@ 5
•
~
t'-,l
O
3 _0
Z
O O O
O O
0
,
0
,
,
,
I
5
,
,
,
,
I
10
,
,
,
,
I
,
i
i
15
i
I
20
,
,
,
,
2
N2 Pressure (10-8 Torr) Figure 17.5 - The protonation rate expressed as the N2H/N2 ratio, under the 100 ~A trap current condition, plotted against the nitrogen pressure in the mass-spectrometer. The amount of N2H in the standard air is calculated assuming that variations of the isotopic ratios 29/28 and 30/28 are caused by admixture of three components (N2, CO and N2H) with known isotopic ratios (i.e., obtained from the mixing diagram shown in Figure 17.4). Under the 100 ~A trap current condition, the contribution of the interfering species N2H cannot be neglected when the nitrogen pressure in the mass-spectrometer is below 10-7 Torr.
Nitrogen Isotopic Analyses at the Sub-Picomole Level Using an Ultra-low Blank Laser ...
371
ple, ambiguity in the protonation rate of + 2 x 10-4 results in an uncertainty of 30%o in the nitrogen isotopic ratio. Therefore, we applied a source condition of 10 gA trap current to avoid N2H interference when the amount of N2 to be measured was in the picomole to sub-picomole range. CO" The amount of CO interfering at masses 28 and 29 can be estimated by analyzing the N2+CO peak at mass 30, and solving the following mass-balance equations: Mass 28 = 14N14N + 1 2 C 1 6 0 Mass 29 = 14N15N + 13C160 (+ 12C170) Mass 30 = 15N15N + 12C180 (+ 13C170)
[17.2] [17.3] [17.4]
Species in parentheses are the ones which are normally negligible compared to the primary CO isotopes at each mass (They are ignored in the following formulae for simplicity). Assuming that isotopic equilibrium is achieved among the N2 molecules, the abundance ratios between their isotopes are: 14N14N:14N15N:15N15N=
1 : 2r : r2
[17.5]
where r = 15N/14N (Note that this assumption do not always hold in the case of extraterrestrial samples that contain extremely anomalous isotopic components, such as cosmogenic nitrogen. See Hashizume & Sugiura, 1992; Sugiura et al., 1995 for details). Then the mass balance equations are: Mass 28 = N2 + CO Mass 29 = 2r x N2 + 1 3 C / 1 2 C x CO Mass 30 = r 2 x N2 + 180/160 x CO
[17.6] [17.7] [17.8]
Assuming that the 1 3 C / 1 2 C and 180/160 ratios of the interfering CO are known, the remaining unknowns are (1) r (= 15N/14N), (2) the amount of N2 and (3) the amount of CO. Therefore, we obtain a unique solution by solving the three equations [17.6/8]. CO originates chiefly in the mass spectrometer ion source, since CO is efficiently removed during the purification procedure (Humbert et al., 2000), and because the calculated amount of CO is nearly constant, regardless of the nature of the sample gas. Therefore, we assume the 13C/12C and 180/160 values of the CO are equal to their standard terrestrial values even when we analyze extraterrestrial samples. However, to be conservative, we assume uncertainties for these ratios to be _+50%0, exceeding the range observed among bulk meteorites. These uncertainties are propagated on the uncertainty for the nitrogen isotopic ratio. When we introduce 10 picogram (0.3 picomole) of standard N2 in the mass-spectrometer, the contribution of CO at mass 28 is normally around 2%, and the ~15N is corrected typically by 10%o.
Nitrogen Isotopic Ratio: The overall performance of our nitrogen isotope analyses at the sub-picomole range is shown in Figure 17.6, which represents the 14N14N/14N15N ratios of various amounts of standard air N2, corrected for interference of CO. Since
372
Chapter 17 - K. Hashizume & B. Marty
Figure 17.6 - The 14N14N/14N15N ratios of sub-picomole quantity of standard air as a function of N2 amount in the mass-spectrometer (CO interference is corrected). This series of analyses is performed under 10 gA trap current condition, where the N2H interference can be neglected. The ion counting system was used for analyses of all peaks including N2 at 28 and 29. In this range of N2 amount, statistical counting error during measurements of the 14N15N peak height (plotted as dotted curves) appears to be the main cause of errors.
these analyses were performed with the trap current set at 10 t~A, we assume that the N2H interference is negligible. The absolute value of the 14N14N/14N15N ratio in this series was systematically higher by ~ 3% compared to the commonly accepted air ratio of 136, which we regard as a mass-discrimination effect. When the magnet and/ or the source conditions are changed, the absolute nitrogen isotopic ratio can vary from 136 to 141, however, the ratio is stable once these parameters are fixed. We adopt the observed mean ratio, e.g., 140.3 in this case, as the reference air value to calculate the 815N values. The 815N is formally defined by comparing the isotopic abundance ratio between 15N and 14N atoms relative to the atmospheric air value. However, we obtain an equivalent value by comparing the abundance ratio between 15N14N and 14N14N molecules with the corresponding reference value, provided that the isotopic equilibrium is achieved among the nitrogen molecules. 515N __ ((15N/14N)observed - 1 ) x 1000 - ( (15N 14N/14N 14N)~ (15N14N/14N14N)AIR (15N/14N)AI R
1) x 1000
[17.9]
Nitrogen Isotopic Analyses at the Sub-Picomole Level Using an Ultra-low Blank Laser ...
373
Error bars in Figure 17.6 represent lo e r r o r ((Jspectrometry) comprising the counting statistic error (Ocounting), the uncertainty (Ocorrection) caused by correction of interfering CO, and the e r r o r (ofitting) caused by fitting ratios to a line as a function of time and extrapolating it to time zero (i.e., when the gas is introduced into the mass spectrometer). We obtain the final uncertainty (Ospectrometry) concerning the mass-spectrometry by the following formula; (Ospectrometry) 2 = (Ocounting) 2 + (Ocorrection) 2 + (Ofitting) 2
[17.10]
The counting statistic e r r o r (Ocounting) for a given amount of nitrogen in the mass spectrometer is calculated to be; ((Jcounting) 2 - ((Jcounting(o)) 2 x [No/N]
[17.11]
where, N is the amount of nitrogen introduced in the mass spectrometer. The No and Ocounting(o) a r e the reference amount of nitrogen and its corresponding statistical counting error, respectively. When 10 picogram of nitrogen (No) is introduced in the mass spectrometer, the counting e r r o r (Ocounting(o)) is 4.4%0. The dotted curves in Figure 17.6 represent the counting statistic e r r o r (Ocounting). It is therefore shown here that the error on the isotopic ratio of sub-picomole quantities of nitrogen obtained by our procedure is limited primarily by counting statistics. 17.5 Total Performance- A case study
In the above sections, we described the technical options chosen for isotopic analyses of sub-picomole quantities of N2. Here we synthesize the performance of our analyzing system. Although practical results are shown elsewhere (e.g., single grain analyses of lunar grains; Hashizume et al., 2002), the grain size and the N concentration assumed here are typical of a natural sample analysis. We consider the case of a silicate sample weighting 0.03 mg (a 200 ~m-sized grain) loaded in the laser chamber. We assume that this sample contains 1 ppm of surface-correlated N, which is extracted at a heating temperature of ~ 1000~ and 1 ppm of volume correlated N, which is mostly extracted by fusion of the sample (Such a situation is often seen among the lunar grains. See for example, Kerridge (1993) for the release profile of lunar nitrogen by stepwise heating). First, the sample is heated at 1000~ in the vacuum pyrolysis mode (If the sample appears to be dirty with a significant amount of organic compounds at the surface, these contaminants can be removed by combusting the sample at lower temperature, e.g. at 600~ We monitor the color and brightness of the heated sample using a CCD camera (Humbert et al., 2000), which enables us to control the heating temperature with a precision of + 100~ Following the purification and analyses of nitrogen and rare gases of the 1000~ temperature step, the sample is melted at a temperature of 1600~ and the extracted gas is purified and analyzed in the same way. In each heating step, ~ 30 picogram of N2 is extracted. During the extraction procedure in vacuum pyrolysis mode, the blank contributes 5 - 10 picogram of N2 to the sample gas. We observed that the hot blank level is variable by + 20% relative to the
374
Chapter 17 - K. Hashizume & B. Marty
blank level measured on the same day. After the purification procedure, the sample gas is introduced into the mass-spectrometer. Since we split part of the gas for rare gas analysis, the amount introduced in the analyzer for nitrogen isotopic analysis is around 10 - 20 picogram. When analyzing 15 picogram of N2 in the mass-spectrometer in ion counting mode, the 615N value for the introduced gas (= sample + blank) is usually determined with a precision of + 7%o. Subtracting the blank contribution (which normally accounts for 15 - 25% of the total gas, in this case), the 615N value of the nitrogen gas is finally determined with a precision typically of + 10%o (1~). The 615N values with such uncertainties are still well informative to discriminate various isotopic components seen among extraterrestrial samples. 17.6 Conclusion
In this contribution, we have presented a new analytical procedure aimed to measure the isotopic composition of sub-picomole quantities of nitrogen. With respect to standard analytical procedures, two major points have been improved, the level of the hot blank and that of interfering species, especially that of N2H. The hot blank has been lowered by minimizing the hot area during the extraction procedure using a defocused laser beam as a heating device. The amount of the interfering species N2H arising from the mass-spectrometer ion source was lowered by optimizing the ion source condition. Having done that, our mass-spectrometry system performs nitrogen isotopic analyses within limits imposed only by the counting statistics. Using this system, we are able to measure the N isotopic ratio with a precision of + 10%o for a - 200 ~m sized single grain containing I ppm of nitrogen. Such system is useful for untangling the nitrogen isotopic puzzle exhibited among the extraterrestrial materials, that is, to identify fine-grained anomalous nitrogen isotopic carriers.
Acknowledgments
We are indebted to Laurent Zimmermann, Pascal Robert, Gregory Sauder and Jean-Claude Demange for their technical assistance. Discussions with Nicolas Dauphas, Frank Humbert, Stuart Boyd and Rainer Wieler greatly improved the quality of this work. We thank Ian A. Franchi and Haraldur R. Karlsson for their detailed and constructive reviews. This work was supported by Programme National de Plan6tologie/Institut National des Science de l'Univers. K.H. acknowledges support from the Japanese Ministry of Education, Science, Sports and Culture, and the Institut National des Sciences de l'Univers - Centre National de la Recherche Scientifique for his stay in Nancy. This work is CRPG-CNRS contribution 1510.
Handbook of Stable Isotope Analytical Techniques, Volume 1 P.A. de Groot (Editor) 9 2004 Elsevier B.V. All rights reserved.
CHAPTE R 18 Mass Independently Fractionated Ozone in the Earth's Atmosphere and in the Laboratory Jeffrey C. Johnston1 & Mark H. Thiemens2 Iterations, P.O. Box 590805, San Francisco, California 94159, USA Department of Chemistry and Biochemistry, University of California, San Diego, La Jolla, California 92093-0356, USA e-mail: 1
[email protected],2
[email protected] 1
2
18.1- Introduction The isotopic analysis of gases in the Earth's atmosphere provides a wealth of information regarding sources, sinks, chemical transformation, and transport processes that govern the budgets of atmospheric species. A non-random distribution of isotopes is observed in many atmospheric species because of a variety of kinetic and equilibrium isotope fractionation effects. This chapter examines the mass independent isotope fractionation observed in ozone (03), both in the Earth's atmosphere and in the laboratory. Many of the isotopic fractionations observed in nature are termed mass dependent because the magnitude of these fractionations scales with the relative mass difference between the nuclides of a specific element. As discussed in detail elsewhere, a mass dependent fractionation of oxygen isotopes will result in an isotopic value that closely follows the relation 6170 - 0.526180 (Thiemens, 1999). Equilibrium chemical processes - such as isotope exchange, diffusion, evaporation and condensation- are sensitive to isotopic mass, and different isotopes may thus be incorporated into the molecules at different rates. The oxygen isotope exchange between H20 and CaCO3, and vapor pressure isotope effects are two examples of mass dependent, equilibrium chemical processes. The sensitivity of a chemical rate constant to isotope substitution, termed a kinetic isotope effect, can generally be understood in terms of transition state theory. Isotopic substitution affects many terms in the transition state theory expression for the reaction rate, but the resulting fractionations are always mass dependent (Bigeleisen & Wolfsberg, 1958). In a system such as for oxygen isotopes, the isotopic fractionation for the 170/160 is always half that of that for 180/160. This is a direct result of the mass dependency of isotope effects. In the case of 170/160, the mass difference is 1 1. Correspondence should be adressed to this author.
376
Chapter 18 - J.C. Johnston & M.H. Thiemens
amu while for 1 8 0 / 1 6 0 is 2 amu. As a result, for example, a 5 per mil fractionation in the 1 7 0 / 1 6 0 ratio is accompanied by a 10 per mil effect in the 1 8 0 / 1 6 0 because of the dependency upon relative mass differences. Such isotope effects are termed mass dependent. In addition to these mass dependent processes, there are a significant number of gas phase fractionation processes that do not scale with differences in nuclidic mass. These so called mass independent fractionations (MIF) have been observed in a variety of chemical systems, including 03 formation (Morton et al., 1990; Thiemens & Heidenreich, 1983), $2F10 formation (Bains-Sahota & Thiemens, 1989), the photopolymerization of CS2 (Coleman et al., 1996), the O + CO reaction (Bhattacharya & Thiemens, 1989a; 1989b), OH + CO reaction (R6ckmann et al., 1998b), and H + 02 recombination (Savarino & Thiemens, 1999b). To date, mass independent isotopic compositions have been observed in meteorites, atmospheric 02, 03, CO, CO2, N20, H202, aerosol sulfate and nitrate, as well as solid samples from the Namibian desert, Antarctic dry valleys, Miocene volcanic ash and Death Valley varnishes (Thiemens et al., 2001). Thiemens (1999) and Weston (1999) have provided thorough reviews of mass independent fractionation processes. Although the mechanisms responsible for producing them are not completely understood, the identification of MIF in natural samples can provide information not available from concentration measurements and single isotope analysis alone (Cliff & Thiemens, 1997; R6ckmann et al., 1998a; 1998b). To date, most laboratory research on MIF has involved molecules incorporating the three stable isotopes of oxygen (160=99.762%, 170=0.038% and 180=0.200%). The Earth's atmosphere is an ideal place to search for MIF due to the wide variety of oxygen containing species and the wide range of isotopic fractionation processes occurring there. The deviation from a purely mass dependent fractionation is conveniently defined by the value A170, where A170=~5170-0.52~5180. Thus, when looking at oxygen isotopes, a mass independent fractionation is one in which A170 ~ 0. Sulfur (32S=95.02%, 33S=0.75%, 34S=4.21%, 36S=0.02%) is another isotopic system in which MIF have been found (Coleman et al., 1996; Cooper et al., 1997; Farquhar et al., 2000b), although to date less work has been done to search for these fractionations in the Earth's atmosphere. This chapter reviews the MIF observed in 0 3 in the Earth's atmosphere and in the laboratory. Chemically produced MIF processes were first observed in 03 (Thiemens & Heidenreich, 1983), and over the past 20 years a significant amount of research in many laboratories has been devoted to developing an understanding of this process. While there remain gaps in our understanding of the mechanisms responsible, recent experimental and theoretical work (discussed in section 18.4) has provided interesting new insights into the process involved. Section 18.2 briefly reviews the chemistry of 03 in the atmosphere, and section 18.3 discusses what is known about the isotopic composition of atmospheric 03. Section 18.4 examines what is known about the source of the MIF in 03 from laboratory experiments and theoretical models; conclusions and references are in sections 18.5 and 18.6.
Mass IndependentlyFractionated Ozone in the Earth's Atmosphereand in the Laboratory
377
18.2 - O z o n e in the Earth's a t m o s p h e r e
While 03 is found throughout much of the atmosphere, it has a very uneven distribution; a consequence of the fact that 03 is not emitted directly from the ground in any significant amount, but is formed photochemically in the atmosphere. Ozone is produced when an oxygen atom combines with an oxygen molecule:
(R1)
O + 02 + M --. 03 + M
where M symbolizes a third body species which is required to remove excess energy from the nascent 03 molecule. In the stratosphere, the oxygen atoms are produced largely by UV photodissociation (X ~ 240 nm) of 02, while in the troposphere the oxygen atoms are produced primarily by photodissociation (K ~ 430 nm) of NO2. Most of the atmosphere's 03 is located in the stratosphere between approximately 25-40 km in altitude. 03 in this region of the atmosphere is critical due to the fact that 03 absorbs UV radiation between 200-310 nm very strongly. Were it not for the stratospheric 03 layer, this biologically harmful radiation would penetrate to the Earth's surface. In contrast to the troposphere, which is characterized by decreasing temperature with increasing altitude, temperature increases with altitude in the stratosphere due to the presence and photolytic activity of 03. Only about 10% of atmospheric 03 is located in the troposphere, and its concentrations there are extremely variable. In very polluted regions it can rise to several hundred parts per billion by volume (ppbv), while in remote regions it can drop below 10 ppbv. The oxidizing potential of the Earth's atmosphere is largely maintained by the hydroxyl radical (OH), the nitrate radical (NO3), and 03. Reaction with OH is the predominant sink for a large number of reduced trace gases including carbon monoxide, methane and nonmethane hydrocarbons (NMHC), hydrofluorocarbons (HFCs), and hydrochlorofluorocarbons (HCFCs). The most important tropospheric source of OH begins with the ultraviolet photolysis of 03 (K < 320 nm). This produces oxygen atoms in the first electronically excited state, O(1D), 7-10% of which then react with water to form OH, the primary daytime oxidizing agent of the troposphere. 03 + hv --* O(1D) + 02 O(1D) + H20 ---, 2OH
(Z. < 320 nm)
[18.2] [18.3]
Reaction between 03 and NO2 leads to the formation of the nitrate radical (NO3), the primary oxidizing agent in the nighttime troposphere. For a number of biogenic and anthropogenic hydrocarbons, mostly alkenes, reaction with 03 represents the major loss process. Thus, to a large extent, 03 controls the oxidation state of the atmosphere. A direct and indirect influence on climate is the second important role played by 03. The direct effect is due to the fact that 03 is a greenhouse gas. 03 is optically active, absorbing solar radiation in the UV and visible regions, and absorbing and emitting terrestrial IR radiation in the 8-10 mm region. This fact means that a change in the distribution of 03 in the atmosphere will disrupt the radiative energy budget and possi-
378
Chapter 18 - J.C. Johnston & M.H. Thiemens
bly disrupt climate. The indirect climatic effect of 0 3 in the troposphere is a consequence of the strong chemical tie between 03 and several other important greenhouse gases, particularly CH4, HFCs and HCFCs which are removed from the atmosphere primarily through reaction with OH. As distributions of 03 change, distributions of OH also change, thus influencing the lifetimes of these greenhouse gases. Tropospheric 03 is also important because of its phototoxic effects on plants and animals. 03 is one of the primary components of photochemical smog, millions of people live in regions that regularly experience episodes of ground level 03 concentrations greater than what the U.S. Environmental Protection Agency considers safe. Effects on human health include reductions in lung function and increases in respiratory symptoms, airway reactivity, permeability, and inflammation. 03 exposure is also responsible for billions of dollars in crop losses each year (Adams, 1985). According to current understanding two tropospheric 0 3 sources, transport from the stratosphere and in situ photochemical production, are balanced by two loss processes, deposition at the Earth's surface and in situ chemical destruction. While these production and loss terms are all estimated to be of the same order of magnitude, spatially and temporally quantifying them is difficult. As with many other atmospheric components, the isotopic characterization of source and sink terms could help to provide important, and otherwise unobtainable, information about the budget of atmospheric 03. 18.3 - The isotopic c o m p o s i t i o n of atmospheric o z o n e
Cicerone & McCrumb (1980) published the first paper on a possible non-statistical distribution of isotopes in stratospheric 03, suggesting that 3402 (160180) may be preferentially dissociated in the stratosphere and mesosphere due to selective absorption in the Schumann-Runge bands (175-205 nm) when compared to 3202 (160160). The proposed effect was an isotopic self-shielding arising from the separate and differential adsorption of the isotopically substituted molecular isotopic species. The individual absorption lines coupled with the differences in natural abundances of the isotopes was suggested to produce a variation with altitude in the photolysis of molecular oxygen, leading to increased production of 180, which would in turn produce more 5003. In 1981, the first measurements were published of stratospheric 0 3 between 22 and 38 km using a balloon borne mass spectrometric system (Mauersberger, 1981). The balloon payload consisted of a high-speed differential pumping system that formed stratospheric gases into a molecular beam. This beam traversed the ion source of a Mattauch-Herzog magnetic instrument, with a mass range of 11-90 amu, without wall collisions. The ion detector was a low noise counting multiplier which permitted long sampling intervals (Mauersberger, 1977; Mauersberg & Finstad, 1979). The results showed a pronounced enhancement in 5003 (160160180 and 160180160) reaching a maximum of over 400%o (relative to the expected statistical distribution of natural 160
Mass IndependentlyFractionated Ozone in the Earth's Atmosphere and in the Laboratory
379
and 180 in the atmosphere) at 32 km and decreasing toward higher and lower altitudes. As only a single isotope ratio was measured in this work, it was not possible to distinguish between mass dependent and mass independent isotopic compositions. Kaye & Strobel (1983) showed that the rapidity of the isotopic exchange between O and 02 would dilute any increased concentration of 180 from preferential 3402 photolysis, and thus concluded that a paradox existed between the measurement of large 5003 enrichments in the stratosphere, and theory which predicted no enrichment. The rapid isotopic exchange between O and 02 has important implications on the isotopic composition of atmospheric 03 and makes it important to distinguish between the chemical and isotopic lifetimes of the 03 molecule. While the chemical lifetime can be on the order of months (Liu & Trainer, 1988), the isotopic lifetime will generally be much shorter. The photochemical lifetime of 03 is extremely variable, but an average value of on the order of hours can be calculated using standard actinic flux data. Because the dominant fate of the O(3P) produced during 03 photolysis is to recombine with 02 to produce 03, the effective chemical lifetime is longer than the lifetime against photolysis. However, because isotopic exchange between O and 02 is approximately 300 times faster than recombination to form 03, the isotopic composition of the of the original 03 has been lost. Thus, rather than providing an integrated picture regarding a variety of sources and sinks, as is found when looking at the isotopic composition of relatively long lived species like CO2, N20, or CH4, the isotopic composition of 03 should be viewed as providing information regarding processes that are typically on the order of hours old. Results from two additional flights of the balloon-borne in situ mass spectrometer payload confirmed the large and variable enrichments in 5o03, and showed that the concentration of 4903 (160160170 and 160170160) is also enhanced (Mauersberger, 1987). Due to a very low abundance, quantitative analysis of 4903 profiles from the in situ mass spectrometer data was not possible. Thus, Mauersberger and co-workers developed a balloon-borne cryogenic sampler capable of returning samples for laboratory analysis (Schueler et al., 1990). Cryogenic sampling utilizes the differences in the thermodynamic properties of gases to separate them. A gas will condense on a surface when the partial pressure of the gas is significantly higher than the equilibrium vapor pressure of the gas, which is determined by the temperature of the surface. By flowing a gas stream through a series of cold traps at progressively colder temperatures, the components of the gas stream can be isolated as a function of vapor pressure. Thus, more abundant gases like H20 and CO2, which condense at higher temperatures, can be separated from 03. Equilibrium vapor pressures of most atmospheric gases are well known, and precise vapor pressure measurements were made over liquid and crystalline 03 by Hanson & Mauersberger (1986). The stratospheric collection system (Schueler et al., 1990; Stehr et al., 1996) involved a series of liquid nitrogen cooled traps, the first at 80 K to condense H20 and CO2, and the second at 63 K to condense 03. These two traps were followed by a small zeolite pump which is used to remove the last traces of non-condensable gases (primarily N2, 02, and Ar)
380
Chapter 18 - J.C. Johnston & M.H. Thiemens
after the collection had ended. The trap temperatures were controlled by regulating the N2 pressure above the liquid. At 63 K, over 95% of the 03 passing through the system condensed, while the major atmospheric gases did not. For stratospheric flight, the balloon platform carried three 03 samplers connected by a common air inlet line, allowing multiple samples to be collected on a single flight. Prior to analysis, the 03 samples were allowed to warm to room temperature to convert the 03 into 02. The samples were analyzed with a mass spectrometer attached to a beam system, as described in (Anderson & Mauersberger, 1981; Mauersberger, 1977; Mauersberger & Finstad, 1979) using atmospheric 02 as the isotopic standard. Results from three flights were published by Schueler et al. (1990), and results from a fourth flight of the same system in Mauersberger et al. (1993). More recently the results from a second series of flights were reported in Krankowsky et al. (2000) and Mauersberger et al. (2001). This data showed that stratospheric 03 is mass independently enriched in 180 and 170; see Table 18.1. The relationship between 5o03 and 4903 is complex; at times the enrichments are nearly equal while at other times 4903 enrichments are significantly less than those in 5o03. While there is a significant amount of variability in the data, the enrichments measured with the sample return system never approach the very high enrichments measured with the in situ system (Mauersberger, 1981; 1987). Mauersberger et al. have attempted to resolve this discrepancy by stating that the older, high values of 5o03 are suspect, and should thus be disregarded (2001). The source of the errors has not yet been identified, however. There have been a variety of ground, atmosphere, and space-based studies of the isotopic composition of stratospheric 03 using infrared emission and absorption spectroscopy. Optical spectroscopy has an advantage over mass spectroscopy in that it is able to distinguish between the different isotopomers of 03, i.e., 160160180 from 160180160, but a disadvantage in that the measurement uncertainties are generally large. Rinsland et al. (1985) reported column enrichments of 50 + 50%o in the symmetric isotopomer and 110 _+ 110%o in the asymmetric isotopomer using ground based FTIR solar absorption measurements. Abbas et al. (1987), using a balloon-borne far infrared spectrometer, reported a larger enrichment in the symmetric isotopomer (690 +_200%0 at 33 km for example) than in the asymmetric isotopomer (200 _+200%0 at 33 km for example). This was a surprising result considering the laboratory (Anderson et al., 1989) and atmospheric data (Rinsland et al., 1985) which indicated that the enrichment was primarily in the asymmetric species. This result is also in disagreement with subsequent measurements of stratospheric 03. Goldman et al. (1989) used a balloonborne FTIR absorption spectrometer to measure isotopic enrichments in the 03 column above 37 km. The first flight, over Fort Sumner, New Mexico (34 ~ N, 104 ~ W), showed enrichments of 200 +_140%o and 400 _+180%o for the symmetric and asymmetric isotopomers, respectively. A second flight over Palistine, Texas (32 ~ N, 96 ~ W), showed slightly smaller enrichments, 160 + 80%0 and 250 _+ 120%o in the symmetric and asymmetric isotopomers, respectively. Meier & Notholt (1996) used a groundbased FTIR instrument to measure solar absorption at Spitsbergen, Norway (79 ~ N, 12~ E), and found a column enrichment of 112 + 14%o and 154 + 9%0 in the symmetric
T
a
b
l
e
1
8
.
1
-Summary
of the isotopic composition
5 0 0 3 (%o) 160180160
of atmospheric
03
(see text and original references for details)
4 9 0 3 (%o)
160160180
Total
Total
Method
Altitude
Location
Reference
t~
(km) ................................................................
-
-
-
-
-
-
-
-
-
-
Str
-
400+115"
-
-
410+70"
-
-
141.7+7.0
!0
phere
..............................................................................
balloon borne mass spec.
102+11
sample
return mass spec.
S c h u e l e r e t al., 1 9 9 0
"
33.5-29.0 30.6 37.1-29.8
-
88.6+3.6
80.1+5.7
"
-
76.6+3.0
71.8+4.8
"
-
88.6+3.3
80.5+5.2
"
-
96.3+3.5 89.5+4.1
82.3+5.5 82.5+5.2
" "
-
-
balloon borne emission
based absorption
200+140
400+180
-
-
balloon borne absorption
160+80 112+14 85+25 f
250+120 154+9 90+33 f
-
_
ground
" based absorption ,. .
100+70
150+60
-
space based absorption
-
ground
130+50
170+4
-
-
102+43
73_+60
airborne emission
- ............
91~--2 . . . . . . . . .
7i!-3 ........
collection, mass spec.
-
82+7
69+7
.
-
86+6
66+6
.
.
-
90+4
78+5
.
.
enrichment
observed
measurements.
during
IR IR
.
Sweden
Krankowsky
28.9-20.9
France
"
30.7-25.6
Sweden
"
31.8-23.3 33.2-25.9
France "
" Mauersberger
column
Arizona
Rinsland
33 k m
Texas
A b b a s e t al., 1 9 8 7
IR .
> 37
.
New
Mexico
Texas Norway .
25-41
global
column
California
18-37
various
~,,L.
o
.
22.4-21.6
" column .
IR
.
" Mauersberger
column
IR
.
New Mexico "
IR
IR
based absorption
TroFosphere
* Maximum
"
" "
200+200*
f Night-time
33.9-27.5
86.5+18
690+200*
-
1987
83.0+8.0 110+29
ground
-
Mauersberger,
85.0+14
-
-
1981
"
91.0+5.0 158+28
-
.... - .............
Texas
-
110+110
-
Mauersberger,
42.5
-
50+50
-
32
"
e t al., 1 9 9 3 e t al., 2 0 0 0
e t al., 2 0 0 1
e t al., 1 9 8 5
Goldman
> 9
e t al., 1 9 8 9
" Meier & Notholt,
m" t~
1996
. I r i o n e t al., 1 9 9 6 " Johnson
e t al., 2 0 0 0 a
t9
..............................................................................
.
Heidelberg
Krankowsky
.
La Jolla
Johnston
.
.
Pasadena
"
.
.
New
"
.
ground
flight, see original reference for enrichment
level
profile.
9 N O
Mexico
e t al., 1 9 9 5
& Thiemens,
1997
O
382
Chapter 18 - J.C. Johnston & M.H. Thiemens
and asymmetric isotopomers, respectively. These authors also used lunar spectra to measure the enrichment within the polar vortex during the polar night. These results show slightly lower enrichments, 85 + 25%0 and 90 + 33%0 for the symmetric and asymmetric isomers, respectively. These results are interesting in that the lower enrichments are consistent with the known temperature dependence of the 03 formation process (Morton et al., 1990). However, during the polar night no 03 can be formed, so the significance of this correlation between enrichment and temperature is unclear. The fact that the difference between symmetric and asymmetric enrichments is lower in the dark is also interesting, and currently unexplained. The ATMOS (Atmospheric Trace Molecular Spectroscopy) FTIR spectrometer has also been used to measure distribution of 5003 in the stratosphere, with measurements being made both from space and from the ground (Irion et al., 1996). The results from four Space Shuttle missions show the globally averaged 5003 enrichment between 2.6 and 26 mbar (~25-41 km) is 130 + 50%0, and the enrichment for 160160180 and 160180160 are 150 + 60%o and 100 + 70%o, respectively. A series of ground-based measurements from Table Mountain, California (34.4 ~ N) resulted in an average total column 160160180 enrichment of 170 + 4%o, with no discernable seasonal variation. The Smithsonian Astrophysical Observatory far-infrared spectrometer (FIRS-2) has been used to measure the isotopic composition of stratospheric 03 from a balloon altitude of around 37 km down to 18 km. The average enrichments from 7 flights (1 flight at 68 ~ N, 6 flights between 30 and 35 ~ N) are 102 + 43%0 and 73 + 60%0 for 5003 and 4903, respectively (Johnson et al., 2000a). Every technique used to measure the isotopic composition of 03 in the stratosphere has detected a substantial and variable enrichment, see Table 18.1. There does not appear to be any correlation between enrichment and time of day, season or altitude, although the larger enrichments tend to be at the higher altitudes (Mauersberger et al., 2001). Laboratory measurements (see section 18.4, below) show that at stratospheric temperatures and pressures, stratospheric 03 should be enriched by about 8090%o, slightly less in 4903 than in 5o03. While the current understanding of isotopespecific stratospheric 03 chemistry is able to explain the bulk of the stratospheric measurements (provided the very large enrichments are disregarded as suggested by Mauersberger et al. (2001)), it is possible that there remain questions regarding the variability in both enrichment magnitude and in ~170/~180 values. Additional measurements of the relative reaction rates of other isotopically substituted species, particularly those of 170, along with their parameter dependencies, will be important. In contrast, the isotopic enrichment of 5003 and 4903 in the troposphere is close to what is expected based on laboratory measurements (see section 18.4, below). Two groups have used cryogenic collection techniques to measure the isotopic composition of ground level 03 (Johnston & Thiemens, 1997; Krankowsky et al., 1995). The operational premise behind the collection is the same as discussed above for the stratospheric collection system, however, the requirements of the system are more
Mass Independently Fractionated Ozone in the Earth's Atmosphere and in the Laboratory
383
extreme due to the fact that mixing ratios of 03 in the troposphere are about 100 times lower than in the stratosphere. The Krankowsky et al. (1995) collection method utilizes a preliminary trap at 77 K for H20, CO2, and N20, followed by the 03 trap at a temperature of 54.9 + 0.1 K and a pressure of about 6 mbar. Following the collection, the 03 is converted to 02 and separated from Xe, prior to analysis on a Mattauch-Herzog magnetic instrument (Krankowsky et al., 1995; Stehr et al., 1996). Johnston & Thiemens (1997) collected 0 3 at a temperature of 55.0 + 0.1 K and a pressure 7.67 mbar (5.75 Torr). At 55 K the vapor pressure over crystalline 03 is lx10 -u bar (8x10-9 Torr) (Hanson & Mauersberger, 1986). To prevent 02 from condensing along with 03, the 02 partial pressure must be kept below 2.0 mbar (1.5 Torr), its vapor pressure at 55 K. At a total pressure of 7.67 mbar, the partial pressure of 03 will be higher than ~1x10-10 bar (lx10-7 Torr), and the partial pressure of 02 will be 1.6 mbar (1.2 Torr). Thus, by maintaining a total pressure of less than 7.67 mbar in the 03 trap, 03 should trap effectively, without any concomitant trapping of 02. In this temperature and pressure regime, 03 and Xe (87 ppbv in air) are the only atmospheric components collected in the trap. The 03 trap is preceded on the collection system by four liquid N2 traps at 77 K to remove H20, CO2, and N20. Following the collection, the 03 trap is warmed to ~82 K and the 03 and Xe recondensed on molecular sieve at 77 K. Following the transfer, the sample is allowed to warm to room temperature, quantitatively converting the 03 to 02 for mass spectroscopic analysis using a Finnigan MAT 251 isotope ratio mass spectrometer. The fact that Johnston & Thiemens (1997) did not separate Xe from the 0 3 derived 02 prior to mass spectroscopic analysis is a potentially important difference between their method and that of Krankowsky et al. (1995), who did perform this separation. There is some evidence that the presence of Xe can enrich the measured ~180 values by more than 10%o (S. Chakraborty, unpublished data, 1998). However, as discussed below, the mean values reported by Krankowsky et al. (1995) fall within the range of values reported by Johnston & Thiemens (1997). The effect of Xe remains to be quantified. Both groups reported the enrichment of tropospheric 03 relative to air 02 (6180 23.5%o, ~170 = 12.2%o V-SMOW). Krankowsky et al. (1995) reported mean enrichment values of 91 + 2%0 in 5003 and 71 + 3%o in 4903 (2 standard deviations) for 47 measurements of ground level 03 from an urban environment near Heidelberg, Germany (49 ~ N, 8 ~ E). The enrichments vary by 27%o and 46%o in 6180 and 6170 respectively, but show no systematic variation. The scatter is statistical in both axes with variances compatible with statistical errors. Johnston & Thiemens (1997) reported the isotopic composition of ground level 03 from three environments in the western United States: La Jolla, California (33 ~N, 117~ W), Pasadena, California (34 ~ N, 118~ W), and White Sands Missile Range (WSMR), New Mexico (32 ~ N, 106 ~ W). The mean values of the enrichments for 5003 and 4903, respectively, in La Jolla (n=29) were 82 + 7 and 69 + 7; for Pasadena (n=6) 86 + 6 and 66
384
Chapter 18 - J.C. Johnston & M.H. Thiemens
+ 6, and for WSMR (n=7) 90 + 4 and 78 + 5, where the stated uncertainty is the standard deviation in the measurements. Significant isotopic variability was observed at each location, in addition to potentially important differences between the sampling locations. While the measured isotopic variability in ground level 03 showed no correlation with 0 3 or NOx mixing ratios, meteorological parameters, or time of da35 there did appear to be a correlation between the pattern of isotopic fractionation and degree of photochemical control over the local 03 budget at each sampling location (Johnston & Thiemens, 1997). Before these differences can be really understood, a much larger data set is required, both of the isotopic composition of 03, as well as many other species that could be interacting with 03. The MIF originating in atmospheric 0 3 has been shown to work its way into a wide variety of other important atmospheric species, including CO (R6ckmann et al., 1998a), CO2 (Johnston et al., 2000; Yung et al., 1997), 0 2 (Luz et al., 1999), H 2 0 2 (Savarino & Thiemens, 1999a), and sulfate deposits on the Earth's surface (Bao et al., 2000b). A recent review by Thiemens et al. provides a thorough review (2001). Lyons utilized a photochemical equilibrium model to demonstrate that the mass independent isotopic composition of 0 3 c a n be transferred to HO2, NO2, and a variety of other species (2001). 18.4 - The origin of mass independent enrichment in 03: Experiment and theory
The fractionations associated with the formation and decomposition of 03 have undergone intensive experimental and theoretical scrutiny in a variety of laboratories since the mass independent enrichment in 03 was first reported (Thiemens & Heidenreich, 1983). Initially~ the mass independent enrichment in 03 was believed linked to the 02 dissociation process (Thiemens & Heidenreich, 1983). Optical shielding by the major isotopic species 1602 will result in the preferential dissociation of 170160 and 180160 with subsequent formation of isotopically heavy 03. As pointed out by Kaye & Strobel (1983) and Navon & Wasserburg (1985), however, this mechanism cannot be responsible because the isotopic exchange between O and 02 (equation [18.4]) is significantly faster than 03 formation. 180 + 1 6 0 1 6 0 <--->160 + 1 6 0 1 8 0
[18.4]
Thus, any increase in 170 or 180 atom concentrations relative to 160 will be removed before they can be incorporated into the 03 molecule. This exchange is mass dependent and, based on the value for the reduced partition function, resets the isotopic composition of the O(3P) to 6180 = -81.0%o at 298 K, relative to the 02 (Urey, 1947). Since the initial experiments in which 02 was dissociated via electron impact in the radio frequency region (Thiemens & Heidenreich, 1983), experiments in the microwave region (Bains-Sahota & Thiemens, 1987), and with UV light (Morton et al., 1990; Thiemens & Jackson, 1987; 1988; 1990) have all produced 03 with A170 ~ 0, provided the formation occurs in the gas phase. At low pressures (< ~3 Torr), wall effects become important; 03 formed under these conditions is mass dependently depleted in heavy isotopes (Bains-Sahota & Thiemens, 1987; Morton et al., 1990).
Mass Independently Fractionated Ozone in the Earth's Atmosphere and in the Laboratory
385
Using a "photolysis/recycling" technique, Mauersberger and coworkers demonstrated that the mass independent enrichment in 03 arises during the O(3P) + 02(3~g) (ground state), recombination reaction (Mauersberger et al., 1993; Morton et al., 1990). These experiments show that the enrichment is not tied to chemistry of excited states, thus eliminating a mechanism of Valentini and coworkers that involved nonadiabatic collisions between different electronic states of symmetric linear molecules (Valentini, 1987; Valentini et al., 1987). While not relevant for explaining MIF in 03, the mechanism observed by Valentini and co-workers remains an interesting symmetry-induced isotope effect for electronically excited species. Bates considered the 0 3 isotope anomaly in several publications. In the first, symmetry numbers were used to conclude that the O + 02 recombination reaction is faster for the substituted than the unsubstituted 03 (Bates, 1986). Anderson & Kaye (1987) pointed out that isotopic exchange and recombination reactions share a common energized collision complex, which can dissociate to products that are isotopically distinct from the original reactant, thus invalidating Bates' contention. It was then suggested that the role of symmetry in the 03 + energized complex is linked to the process of energy randomization and dissociation (Bates, 1988). However, this mechanism produced a pure 180160160 enrichment, which is inconsistent with experimental observations (Anderson et al., 1989). To accommodate this deficiency, an intermolecular "flip" was invoked to rearrange the metastable 03 + complex (Bates, 1990). This rearrangement occurs during a bond stretch that takes place before the excess energy of the 0 ~ 0 2 association is completely randomized. The bond to the terminal atom is stretched sufficiently to initiate bond formation with the other 03 terminal atom, thus converting 160160180 into 160180160. A problem with this mechanism is that in order to produce the enrichments observed in the laboratory (~85%o), the randomization frequency for the molecular rearrangement must be on the order of 1013 s -1 (Bates, 1990). This is equivalent to the fastest vibrational frequency, so is physically unrealistic. When a more plausible randomization frequency was used, the predicted enrichment increased to ~430%o. Experimental artifacts were invoked to explain this discrepancy between model prediction and observation (Bates, 1990). As pointed out by Thiemens (1992), however, experimental artifacts are insufficient to explain these differences. A mechanism involving highly vibrationally excited 0 2 w a s developed by Miller et al. (1994) and Houston et al. (1996) to address both the 03 deficit problem (the fact that model predicted 03 concentrations in the upper stratosphere were lower than observed 03 concentration levels, until satellite observations were revised downward and the rate of OH+C10 was measured to be faster than previously believed) and the large isotope enrichments in stratospheric 03. In this scheme, (1) a fraction of 03 photodissociation events produce 02 (v a 26) + O, (2) the 02 (v >__26) reacts with ground state 02 to produce 03 + O, and (3) each O atom recombines with 02 to form 03. Step (1) is more likely for heavy 03 than for normal 03, and since one of the original 03 oxygen atoms is incorporated into a new 03 molecule through step (2), heavy oxygen is distilled into the 03 pool. As 03 is cycled by photodissociation many times (ranging
386
Chapter 18 - J.C. Johnston & M.H. Thiemens
from about 50-400 times at altitudes of 50-32 km) before destroyed by other processes, even a very small enrichment factor can be strongly amplified. This mechanism can explain only about 3% of the heavy 03 enrichment in the stratosphere, however. Another theory regarding the source of the mass independent enrichment in 0 3 is that it is related to the symmetry of the different isotopically substituted 03 molecules, 160160160, 170160160, and 180160160 (Heidenreich & Thiemens, 1986). The rate of a three-body recombination reaction is very sensitive to the lifetime of the metastable species, 03 + in this case. The greater this lifetime, the greater the probability that the metastable molecule will be collisionally quenched by the bath molecule (M). The fact that 170160160 and 180160160, both Cs, will have a slightly longer lifetime than the C2v species, 160160160, due to the appearance of alternate rotational states in the asymmetric species is taken into account by standard 3-body reaction rate theory. Thus, for this lifetime argument to be valid, some additional factor must be responsible. Additionally, the lifetime mechanism requires the isotopic enrichment to be independent of pressure until the falloff region of the O + 02 recombination is approached. Measurements of this falloff curve show that the deviation from low pressure, third order behavior does not occur until about 8 atmospheres, and that the high pressure limit is not reached until greater than 400 atmospheres (Croce de Cobos & Troe, 1984; Hippler et al., 1990). Laboratory measurements show that the isotope enrichments are quite sensitive to pressure and do not obey the kinetics of the observed falloff curve. The room temperature enrichment reaches a maximum at approximately 10 Torr, with 5 1 8 0 - 130%o and 5 1 7 0 - 100%o, and begins to fall off quickly at pressures greater than ~ 1/2 atmosphere (Morton et al., 1990; Thiemens & Jackson, 1988). By 56 atmospheres the enrichment has disappeared entirely, and between 56-87 atmospheres the 03 shows a small mass dependent depletion in 180 and 170 (Thiemens & Jackson, 1990). It is therefore unlikely that theories involving differential lifetimes of the metastable state are the source of the observed mass independent fractionations. As discussed below, however, Marcus and co-workers have recently proposed a mechanism by which a slightly higher density of states in the asymmetric vibrationally excited 03 molecule can reproduce the observed MIFs in a way consistent with observed pressure dependencies (Gao & Marcus, 2001; 2002; Hathorn & Marcus, 1999; 2000). It is established that absolute rate coefficients for 0 3 formation are strongly influenced by the bath gas composition, and vary by as much as one order of magnitude, depending on the quenching efficiency of the third body (Steinfeld et al., 1987). Guenther et al. have shown, however, that the composition of the third body does not alter the relative rates of the isotope specific formation reactions (Guenther et al., 2000). While the enrichment mechanism based on differential lifetimes appears to be invalid, experimental observations and theoretical models continue pointing to the importance of molecular symmetry. Anderson et al. (1989) used a tunable diode laser system to measure the isotopomeric distribution in 03 produced in an electric discharge and found that 80% of the enrichment is in the asymmetric molecule, while
Mass IndependentlyFractionatedOzonein the Earth's Atmosphereand in the Laboratory
387
only 20% is in the symmetric molecule. A purely statistical argument predicts that only 66% of the enrichment should reside in the asymmetric isotopomer. In natural oxygen, the abundances of 180 and 170 are small enough that concentrations of multiply substituted species, such as 180160180 and 180170160, are too low to contribute to the observed enrichments. However, by forming 03 from 02 that is artificially enriched in 180 and 170, all isotopomers of 03, from 4803 to 5403, can be formed. Mauersberger and co-workers produced 03 via the photolysis/recycling method and found that, relative to 4803, the symmetric isotopomers (170170170, 180180180) are slightly depleted (Mauersberger et al., 1993; Morton et al., 1989). This result agrees with the 03 isotopic formation theory developed by Kaye & Strobel (1983), that predicts a small mass dependent depletion. For the asymmetric isotopomers, large enrichments are observed. A 180%o enrichment is seen in 160170180, and approximately 2/3 of that in the other asymmetric species. These results seem to indicate that the enrichment lies in the asymmetric molecules alone, and that the symmetric molecules show the depletion predicted from recombination, theory (Kaye & Strobel, 1983; Kaye, 1986). Gellene (1996) applied a theory of symmetry induced kinetic isotope effects (SIKIEs) to the O + 02 recombination reaction and was able to quantitatively explain the enrichment and depletion pattern in isotopomers 4803 - 5403 observed by Mauersberger and co-workers (Mauersberger et al., 1993). The general theory (Gellene, 1992) was developed from observations of a variety of ion-molecule SIKIEs, including the formation of 04 + (Griffith & Gellene, 1992), He2 + (Gellene, 1993), and (CO2)2 + (Yoo & Gellene, 1995). In the case of 03, symmetry restrictions arise for homonuclear diatomics (i.e., 160160 and 180180) involved in the O + 02 collision because only a fraction of their rotational states correlate with those of the corresponding 03 molecule. In contrast, all of the rotational states of the heteronuclear (i.e., 160180) 02 molecules correlate with those of the resulting 03 molecules. While the theory developed by Gellene and co-workers remains relevant to other SIKIEs, the rate coefficient measurements of Mauersberger et al. (1999), (discussed below) show that this mechanism is not relevant to 03. A number of studies have focused on characterizing the adiabatic energies, binding energies, lifetimes and geometries of the low-lying metastable electronic states of 03 in an effort to isolate the source of the mass independent enrichment (Anderson et al., 1992; 1993; 1995; Anderson & Mauersberger, 1995). The motivation being that these metastable states may be populated during 03 formation, and subsequent symmetry selective pathways to the ground state could then enhance the asymmetric molecules over the symmetric ones. The spectroscopy of 03 is very complex, eight electronically excited states with adiabatic energies below 8 eV are known, at least half of which lie below 2 eV. Despite these efforts, no connection between the isotope effect and the metastable states has been made to date.
388
Chapter 18 - J.C. Johnston & M.H. Thiemens
Kinetic studies have provided a great deal of information about the mechanism of 03 formation and recently, the source of the MIF. Studies using 160 and 1802 showed that 03 is formed via a simple end on addition, rather than through a more complicated insertion process (Larsen et al., 1991). Recent work by Mauersberger and coworkers indicates that the MIF occurring during the O + 02 reaction is determined by the nature of the collision rather than by molecular symmetry. Anderson et al. (1997) used isotopically pure 02 ( 1 6 0 1 6 0 and 1 8 0 1 8 0 ) , a filtered deuterium lamp, and a molecular beam mass spectrometer gas analyzer to investigate the kinetics of four 03 formation channels: 160 180 180 160
+ + + +
160160 160160 180180 180180
+ M --* 160160160 + M + M --* 180160160 + M + M --* 180180180 + M 4- M --* 160180180 + M
[18.5] [18.6] [18.7] [18.8]
Reactions [18.5 - 18.7] all have similar rates of formation, while [18.8] has a rate coefficient 50% faster than the first three. If molecular symmetry was the controlling factor, reactions [18.6] and [18.8] should display equal rate constants. Using a slightly different technique, Mauersberger et al. (1999) confirmed the results of Anderson et al. (1997) and extended the studies by including reactions involving 170. By combining these measured rate coefficients with other laboratory observations (i.e., Anderson et al., 1989 and Mauersberger et al., 1993), it is possible to infer information about the rate coefficients, which cannot be directly measured For example, atmospheric 5003 is formed by reaction [18.6] and by: 160 + 160180 + 160 + 160180 +
M M
--* 1 6 0 1 6 0 1 8 0 + M ~
160180160 + M
[18.9] [18.10]
The rate coefficient determined by Mauersberger et al. (1999) indicates that [18.6] cannot be responsible for the large enrichment in atmospheric 03. Using a diode laser Anderson et al. (1989) showed 80% of the enrichment is carried in the asymmetric 5003 isotopomer. Thus, the rate coefficient for reaction [18.9] must be ~43% faster than the rate coefficient for reaction [18.5] (Mauersberger et al., 1999). In describing the observed pattern in isotopic enrichment, a collision between a light atom and a heavier molecule will result in a rate constant that is higher than the rate constant from reaction involving only one isotope, while a collision between a heavy atom and a lighter molecule will not yield the rate enhancement (Mauersberger et al., 1999). Janssen et al. (1999) used tunable diode laser and mass spectrometry to measure the relative formation rate coefficients of each of the four channels that can form 5003 and 5203 from mixtures of 160 and 180. The results, consistent with those of Anderson et al. (1997) and Mauersberger et al. (1999), show that molecular symmetry plays no apparent role in the 03 formation process, and that the isotopic enrichment in 5003 arises from an enhanced rate of one formation reaction [18.9] (Janssen et al., 1999).
Mass Independently Fractionated Ozone in the Earth's Atmosphere and in the Laboratory
389
Following up on the earlier work of Hathorn & Marcus (1999; 2000), Gao and Marcus developed a statistical (RRKM)-based model with a hindered-rotor transition state that explains most of the laboratory observations (Gao & Marcus, 2001; 2002). The theory assumes (1) an "~l-effect" which can be thought of as a small deviation from the statistical density of states for the symmetric versus the asymmetric isotopomers, and (2) weak collisions in the deactivation of the vibrationally excited 03 (Gao & Marcus, 2002). A partitioning effect, arising from small differences in zero-point energies of the two exit channels of dissociation of an asymmetric 03, controls the ratios of the recombination rates. These small differences are magnified into large differences in numbers of states in the two competing exit channel transition states. In isotopically unscrambled systems this second effect dominates, leading to the large, unconventional mass-dependent effects in the rate constants reported by Mauersberger and coworkers (Anderson et al., 1997; Janssen et al., 1999; Mauersberger et al., 1999). In contrast, in the scrambled systems the partitioning factor disappears exactly, leaving the ~l-effect responsible for the observed fractionations. The calculated isotopic enrichments are consistent with a wide variety of laboratory experiments examining 03 formation (Bains-Sahota & Thiemens, 1987; Heidenreich & Thiemens, 1986; Morton et al., 1990; Thiemens & Heidenreich, 1983; Thiemens & Jackson, 1987; 1990; Yang & Epstein, 1987). 18.5 - Conclusions
Ozone carries a large and variable mass independent isotope fractionation, which has been studied extensively in the atmosphere and in the laboratory for more than 20 years. Measurements of MIF in tropospheric 03 (Johnston & Thiemens, 1997; Krankowsky et al., 1995), and a recent assessment of MIF in stratospheric 03 (Mauersberger et al., 2001) indicate that the atmospheric variability is largely in agreement with laboratory measurements. The recent work of Gao & Marcus (2001; 2002) has provided a theoretical model to understand a wide variety of laboratory results. The development of this model is timely, as it will undoubtedly contribute to the interpretation of the wide variety of natural species that have now been shown to possess a mass independent fractionation.
Acknowledgements The National Science Foundation and NASA is gratefully acknowledged for their support for the many facets of work reported in this paper.
Handbook of Stable Isotope Analytical Techniques, Volume 1 P.A. de Groot (Editor) 9 2004 Elsevier B.V. All fights reserved.
CHAPTE R 19 Site-specific Nitrogen Isotope Analysis in N20 by Mass Spectrometry Sakae Toyodal,4* & Naohiro Yoshidal,2,3,4** Department of Environmental Chemistry and Engineering, Frontier Collaborative Research Center, and Department of Environmental Science and Technology, Tokyo Institute of Technology, 4259 Nagatsuta, Midori-ku, Yokohama 226-8502, Japan, and SORST, Japan Science and Technology Corporation (JST), Kawaguchi, Saitama, Japan 1 Department of Environmental Chemistry and Engineering, Tokyo Institute of Technology 2 Frontier Collaborative Research Center, Tokyo Institute of Technology 3 Department of Environmental Science and Technology, Tokyo Institute of Technology 4 SORST,JST (Japan Science and Technology Corporation) e-mail'*
[email protected]; **
[email protected]
Abstract A novel method has been developed for site-specific nitrogen isotope analysis in N20, which is an asymmetric linear molecule and an important trace gas in the atmosphere. The method makes use of mass analyses of the molecular (N20 +) and fragment (NO +) ions of N20 in an isotope-ratio mass spectrometer equipped with a special ion collector system. The fragmentation of N20 in the electron impact ion source is stable, and the precision of isotope ratio measurements of the fragment ion is better than 0.1%o for pure N20 samples introduced from a conventional dual-inlet system. Although the observed isotope ratio of the fragment ion is affected by rearrangement reactions in the ion source, a correction can be applied using an experimentally determined rearrangement fraction. This technique has been shown to supply useful information on environmental N20 when it is coupled with a continuous-flow technique to accommodate highly sensitive analyses.
19.1 Introduction N20 is one of the important trace gases in the atmosphere that affect radiative balance and atmospheric chemistry. Since it has a long lifetime of ca. 120 years, and absorbs infrared radiation, it is a major greenhouse gas, following CO2 and CH4 (IPCC, 2001). In the stratosphere, it is decomposed by ultraviolet light, and chemically active species (atomic oxygen and nitrogen oxides) are produced, which then react with ozone. The tropospheric concentration of N20 is increasing by 0.2 - 0.3% per year presumably due to human activity, but the global budget and cycle of N20 has not been well resolved because it has a variety of sources and sinks. Isotopic studies of N20, which are essential to the understanding of its origins, chemical/physical processes, and fate, have not been extensively conducted, while there have been a num-
Site-specific NitrogenIsotopeAnalysisin N20 by Mass Spectrometry
391
ber of extensive works based on concentration or flux measurements in various fields. One of the reasons that it is difficult to analyze isotope ratios of N20 is its low concentration in nature" for example, ~50 L of atmospheric or water sample are required for the conventional dual-inlet isotope analysis. However, recent developments in instrumentation and analytical techniques allow us to measure isotopic ratios of N20 with a smaller sample size and a less time-consuming procedure. In earlier studies, N20 was converted to N2 and 0 2 / C O 2 / H 2 0 , and then introduced into the conventional dual-inlet system of an isotope-ratio-monitoring mass spectrometer (Yoshida & Matsuo, 1983; Wahlen & Yoshinari, 1985; Yoshinari, 1990; Kim & Craig, 1990, Thiemens & Trogler, 1991). Later, direct introduction of N20 into the ion source of the mass spectrometer was performed, and found to be applicable to obtaining both nitrogen and oxygen isotope ratios simultaneously, although separation and purification of N20 from isobaric CO2, which is much more abundant in the environment, are critical for the accuracy of the measurement (e.g. Kim & Craig, 1993; Tanaka et al., 1995; Rahn & Wahlen, 1997). This technique can be combined with a continuous-flow technique for on-line analysis of smaller size samples (Yoshinari et al., 1997; Dore et al., 1998). Although previous studies for the isotopic characterization of N20 were based on the bulk element contained in the molecule (nitrogen and oxygen), N20 has more isotopic information owing to its asymmetric molecular structure (N-N-O), i.e. the intramolecular distribution of nitrogen isotopes. Since formation and cleavage of N-N and N-O bonds of N20 take place in naturally occurring processes, an intramolecular site preference for nitrogen isotopes is expected (e.g. Yung & Miller, 1997). We (Toyoda & Yoshida, 1999) attempted to monitor the isotope ratio of the fragment ion of N20, as well as the molecular ion formed in the ion source of a mass spectrometer, and reported for the first time that the site-specific nitrogen isotope analysis in N20 can be performed with high precision. Brenninkmeijer et al. (1999) also reported a mass spectrometric technique basically the same as ours, and an infrared spectroscopic one has been developed (Esler et al., 2000a; Uehara et al., 2002), although the precision of the latter technique is not as high as that of the mass spectrometric technique at present. In this manuscript, a notation for the intramolecular distribution of isotopes is briefly discussed, and instrumentation, experimental techniques, and applicability are described.
19.2 Notation for the isotopomers of N20 Let us define an isotopomer as one of a set of molecules that contains isotopically substituted atoms. The number of isotopomers of a certain compound depends on the number of elements that constitute the compound, the number of isotopes of each element, and the symmetry of the molecular structure. In the case of N20, 12 isotopomers can exist, but only the following five are significant at the natural abundance level: 14N14N160, 15N14N160, 14N15N160, 14N14N170, and 14N14N180.
392
Chapter 19 - S. Toyoda & N. Yoshida
If the nitrogen atoms at the center and end positions are denoted as N~ and N~, respectively, the nitrogen isotope ratio for each site is expressed as the isotopomer ratio as follows (Toyoda & Yoshida, 1999): 15Ra = 14N15N16 O / 14N14N16 O 15R~- 15N14N16 O / 14N14N16 O
[19.1] [19.2]
The conventional nitrogen isotope ratio, which is hereafter referred to as the bulk isotope ratio, corresponds to the average of the two isotopomer ratios, since it does not distinguish N-isotopomers: 15Rbulk = (15Rc~+ 15R~) / 2
[19.3]
In delta notation 615Ni- {15Ri / 15Ri (std)- 1} 1000
[19.4]
for i = a, [3, and bulk, where std means the standard or reference material. If atmospheric N2 is chosen for the standard, equations [19.3] and [19.4] lead to the following equation: 515Nbulk- (615NR + 615N~ ) / 2
[19.5]
Note that equation [19.5] does not strictly hold when N20 is used as the reference (Toyoda & Yoshida, 1999).
19.3 Experimental methods 19.3.1 Natural and 15N labeled N20 samples Commercial N20 in a cylinder (Showa Denko Co., Ltd., purity > 99.999%) was used as the working standard gas without further purification. The isotopomer/isotope ratios for the gas were determined to be -4.0, 0.2, and 23.3%o for 515N% 515Nf~, and 6180, respectively, by our nitrogen-isotopomer calibration technique (Toyoda & Yoshida, 1999) and by conversion of N20 to N2 and CO2 in the presence of graphitic carbon (after Yoshinari, 1990), where the standards for nitrogen and oxygen are atmospheric N2 and SMOW (standard mean ocean water), respectively. Labeled samples of 15N14NO and 14N15NO, whose 15N purity at the labeled and unlabeled positions are > 99% and 0.37% (natural abundance level) by atomic fraction, respectively, were purchased from Icon Services Inc., USA. From each of the materials three samples were prepared by static dilution with the working standard so that they have different 15N concentrations at the labeled position. Their 15N concentration was calculated from the manometrically determined dilution factor: 0.582, 0.758, and 1.06% for a-labeled N20 and 0.524, 0.749, and 1.09% for the 13-labeled N20. Accuracy was estimated to be better than 1% (relative error against calculated 15N concentration), although uncertainty of the 15N content of the working standard gas was not
Site-specific Nitrogen Isotope Analysis in N20 by Mass Spectrometry
393
included because it canceled out in the relative measurement between the labeled and working standard gases.
19.3.2 The principle of site-specific nitrogen isotope analysis for N20 by mass spectrometry In conventional mass spectrometric isotope analyses for CO2, N2, H2, etc., molecular ions produced by electron-impact ionization are separated and detected. Fragmentation of the molecule also takes place in the ionization chamber of the mass spectrometer, although their fraction is smaller than for molecular ions. For N20, the formation ratio of N20 + to NO +, one of the fragment ions, is about 3 91. Considering the asymmetric structure of the N20 molecule, one would expect that the nitrogen atom in NO + originates mostly from the R-N atom of the parent molecule (there is a minor contribution from the f~-N atom as shown later), while N20 + contains both c~and ~-nitrogen atoms. Therefore, mass analyses of both ions make it possible to deduce the intra-molecular distribution of nitrogen isotopes. In this study a sector type, isotope-ratio mass spectrometer (Finnigan MAT 252, Thermo Quest K. K.) was used throughout the experiments. The multi-collector system was specially designed and modified to measure the isotope ratios for both molecular and fragment ions of N20 without changing the Faraday collector cup configuration and the amplifiers associated with the cups (Figure 19.1). The modified collector system consists of five cups, three of which (cups no. 1, 3, and 5) have larger collector slits than the standard cup in order to allow the measurement of other gases, such as N2, 02, etc., as well as N20 and NO. In the N20+-measuring mode, cups no. 1, 2, and 4 were used to monitor ions of m / z 44, 45, and 46, respectively, and in the NO+-measuring mode, cups no. 1, 3, and 5 monitor m / z 30, 31, and 32, respectively. Taking account of the isotopomers of N20 that contribute to the m / z 44, 45, and 46 molecular ion beams and the m / z 30, 31, and 32 fragment ion beams in the mass spectrometer, the 45R, 46R, 31R, and 32R isotope ratios for molecular and fragment ions relative to the most abundant ones, are given by: 45R = 46R = 31R = 32R =
15R~ + 15R~ + 17R 18R + (15R~ + 15R~) 17R +15Ra15R~ 15R~ + 17R 18R +15R~ 17R
[19.6] [19.7] [19.81 [19.9]
Since the observed reproducibility of 32R is much worse than for the other ratios, probably because of an unstable instrumental background level of 02 and the 02 + produced from ionizing reactions of N20, we assumed the mass-dependent fractionation of oxygen contained in the sample N20 and used the following equation (Craig, 1957) instead of equation [19.9] to obtain 15R~, 15R~, 17R, and 18R from 45R, 46R, and 31R of the sample" 18R / 18R(std)= {17R / 17R(std)}2
[19.10]
394
Chapter 19 - S. Toyoda & N. Yoshida
Figure 19.1 - Schematic diagram of the measurement of N20 isotopomers on a mass spectrometer (A modified version of Figure 1 in Toyoda & Yoshida, 1999). IC: ionization chamber, C: cathode, ET: electron trap, EP: extraction plate, AN: analyzer, MC: multi-collection system.
This assumption is not valid when a mass-independent oxygen isotope fractionation occurs. Such an 170 anomaly is observed in stratospheric and tropospheric N20 (Cliff & Thiemens, 1997; Cliff et al., 1999). However, the magnitude of the anomaly is about 1%o for 6170 and equation [19.10] brings about an overestimate of about 0.1 and 0.05%o for 615N~ and 615Nbulk respectively. These differences are small compared with the current analytical precision for atmospheric N20 by a continuous-flow technique (Yoshida & Toyoda, 1999), although corrections will be needed if the 170 a n o m aly is independently determined for the same sample. 19.4 R e s u l t s a n d d i s c u s s i o n
19.4.1 Fragment pattern of N20 and precision of the measurement The critical point of the mass fragmentation analysis is that the fragment ratio, which refers to the ion beam intensity ratio of fragment ion to molecular ion, and isotope fractionation in the fragmentation, should be constant in a series of measurements in which sample and reference are introduced alternately into the ion source. Figure 19.2 shows the variability of the fragment ratio when the electron energy is
Site-specific Nitrogen Isotope Analysis in N20 by Mass Spectrometry
395
Figure 19.2 - Relationship between the fragment ratio of N20 and ion source conditions. The parameter'ext.' is the scale on the potentiometer that modulates the extraction plate voltage. A larger number corresponds to a lower potential difference between the ionization chamber and the extraction plate.
intentionally changed over a wide range. Typically, experimental parameters of the ion source of the mass spectrometer were optimized so that high linearity (isotope ratio is not dependent on sample pressure) can be obtained, although this tuning does not give the maximum sensitivity that the machine can produce. For the instrument used in this study, the optimum setting values were between 0 and 5 for the extraction plate voltage parameter, and between 60 and 100 eV for the electron energy. The fragment ratio was about 0.3 under typical conditions, and its variability was within _+0.2% over several hours. The effect of the total pressure, or matrix effect, in the ionization chamber was evaluated by introducing He continuously using the interface for isotope-ratio-monitoring gas chromatography-mass spectrometry. As shown in Figure 19.2, pressure
396
Chapter 19 - S. Toyoda & N. Yoshida
dependence of the fragment pattern was small under optimum conditions. The precision of the ten sets of measurements and variation among nine independent measurements between two identical samples obtained using the dual inlet system are shown in Table 19.1. For the molecular ion, precision was nearly equal to that for the other gases such as CO2 and N2, although in the case of the fragment ion it became a little worse because the beam intensity of the latter was weaker than that of the former.
19.4.2 Rearrangement reaction of N20 In order to check the stability of isotope fractionation during electron impact fragmentation of N20, 15N-labeled samples were analyzed relative to the working standard. The analytical precision of the samples was better than 0.2 - 1.2%o, depending on the 15N concentration. Bulk nitrogen isotope ratios of the six samples observed in the N20-measuring mode agreed well with calculated values (~15Nbulk(obs) - 1.01615Nbulk(calc) - 3.78, R = 0.99998), indicating that there is no isotope fractionation in the sample preparation and that the fractionation factor for the formation of N20 § is constant within a range of 1000%o (Toyoda & Yoshida, 1999). However, the isotopomer ratios obtained from the fragment NO measurements were not identical with the calculated values, although a linear relationship was found. As shown in Figure 19.3, 615N was lower than the calculated value at the labeled position and higher at the unlabeled position. From the magnitude of the 6 deviation and 15N purity of the purchased material validated by near-infrared absorption spectroscopy (Uehara et al., 2002), we attributed this result to the rearrangement reaction of N20 under electron impact ionization. Here, we define "rearrangement" or "scrambling" as inclusion of the [3-N atom Table 19.1- Analytical precision of the isotopomers' determination. Experiment
Average + l o
615Nbulkair (%o)
615N%ir (%o)
-1.896• -1.895• -1.887• -1.888• -1.915• -1.921• -1.899• -1.963• -1.882• -1.905+0.025
-3.981• -4.020• -4.034• -3.980• -4.021• -4.037• -4.027• -4.009• -4.015• -4.014+0.021
615Nf~air (%o)
618OsMow (~o)
0.188• 0.229• 0.261• 0.204• 0.191• 0.195• 0.229• 0.082• 0.251• 0.203+0.053
23.308+0.038 23.313+0.040 23.336+0.045 23.343+0.024 23.342+0.038 23.297+0.062 23.341+0.033 23.219+0.024 23.330+0.019 23.314+0.040
About 2 ml STP of N20 identical with the working standard is introduced into the mass spectrometer through a conventional dual inlet system. Average of 10 set of relative measurements is used for 6-calculation and the standard deviations of molecular and fragment ion analysis for ~)15Nbulk and ~)15Ncz are listed, while square mean errors are estimated for 615Nfi (= 2 x 615Nbulk - 615Ncz).
Site-specific Nitrogen Isotope Analysis in N20 by Mass Spectrometry
397
Figure 19.3 - Relationship between the observed nitrogen isotopomer ratio of N20 and the calculated value (A modified version of Figure 2b in Toyoda & Yoshida, 1999). Calculated values refer to the labeled position. in the parent N20 molecule in the fragment NO + ion, irrespective of what the mechanisms are. The formation mechanisms of NO + after electron impact of N20 are considered to be spontaneous and collision induced dissociation of the N20 + (Begun & Landau, 1961) and ion-molecule reaction between O + / N + and N 2 0 (Derwish et al., 1964; Ryan, 1972), and the formation of N~O + has been observed from the ionization of highly 15N-enriched N20 (Friedman & Bigeleisen, 1950; Begun & Landau, 1961). Although the relative contributions of these reactions may depend on the geometry of the ionization chamber, gas pressure, etc., their effect on the overall isotopic fractionation of NO + should cancel out when the sample and reference gases are analyzed under the same instrumental conditions.
398
Chapter 19 - S. Toyoda & N. Yoshida
If the r e a r r a n g e m e n t fraction y is defined as the ratio of N O + b e a r i n g 13-N of the p a r e n t N 2 0 to total N O +, the o b s e r v e d isotope ratio of N O + can be e x p r e s s e d by the following e q u a t i o n 15R%bs - (1 - y) 15R~ + y 15RI3
[19.11]
F r o m e q u a t i o n s [19.3], [19.4], a n d [19.11], y is expressed as a function of the u n s c r a m bled isotopomer, b u l k isotope, a n d o b s e r v e d i s o t o p o m e r ratios" y - A(a15N~ - 815NC%bs) / [2{(1 - A) 815N~obs - 815N bulk + A815N~}]
[19.12]
where A ~ (15R~ / 15Rbulk )std
[19.13]
In the p r e s e n t s t u d y A w a s d e t e r m i n e d to be 0.998 by c o m p a r i n g the w o r k i n g stand a r d to the calibration s t a n d a r d p r e p a r e d from thermal d e c o m p o s i t i o n of NH4NO3, w h i c h forms N 2 0 w h o s e c~-N a n d f3-N atoms are d e r i v e d from nitrate a n d a m m o n i u m ions, respectively (Toyoda a n d Yoshida, 1999). U s i n g the calculated a n d o b s e r v e d i s o t o p e / i s o t o p o m e r ratios of 15N-labeled N 2 0 , y w a s e s t i m a t e d to be a b o u t 0.08, w h i c h is in g o o d a g r e e m e n t w i t h other r e p o r t e d values ( F r i e d m a n & Bigeleisen, 1950, B e g u n & L a n d a u , 1961). A l t h o u g h this p a r a m e t e r m a y d e p e n d on the i n s t r u m e n t used, the variation of y w a s f o u n d to be v e r y small u n d e r typical e x p e r i m e n t a l conditions, even if m e a s u r e m e n t s w e r e p e r f o r m e d on difTable
19.2
-
Reproducibility of the rearrangement fraction y.
...Exper!m.ent . a..........................................................Ionsourc.e..p.a!~.ametersb ........... .............................................................Va!ue....of.yC ............ ..................... Trape (V) Extractiond Electron energy (V) 1 2 3f 4 5 6 7
4.64 2.46 2.46 0.00 0.00 0.00 0.00
54.9 37.4 37.4 50.0 50.0 50.0 50.0
66.9 86.6 86.6 86.5 86.5 86.5 102.1 Average & lo
0.0803 0.0810 0.0813 0.0813 0.0805 0.0812 0.0823 0.0811+0.0006
All experiments were performed on different days except for no. 4 and 5, which were performed at a 6-hour interval on the same day. b These parameters were optimized so that both high linearity and high sensitivity could be obtained, but optimum values were different from day to day. c Calculated from eq 12 in the text. Sample was b-labeled gas (15N14NO, 15N = 0.744 at.%). d See the caption in Figure 19.2 e Potential difference between the ionization chamber and the electron trap. f Sample was introduced into the ion source with He. [A modified version of the Table 2 in Toyoda & Yoshida, 1999]. a
Site-specific Nitrogen Isotope Analysis in N20 by Mass Spectrometry
399
ferent days and the pressure in the ionization chamber was changed by introducing He (Table 19.2). Once the rearrangement rate is determined, the nitrogen isotopomer ratio of the sample can be obtained from the observed ratio: 615N~ = ~)15Naobs + 2y(615N%bs- 615Nbulk) / {A(1- 2y)} 615N~ = 1515Nbulk+ A(615Nbulk - ~)15Na) / (2 - A)
[19.14] [19.15]
In the above equations y is assumed to range between 0 and 0.5, otherwise one could not obtain the intramolecular nitrogen isotope distribution by mass fragment analysis.
19.5 Summary and applications of the technique Site-specific nitrogen isotope analysis for N20 has been made possible by mass fragmentation analysis on a conventional isotope-ratio mass spectrometer for a sample size at the micromolar level. The precision of the analysis is almost as high as that of the bulk isotope analysis. This technique, combined with continuous-flow analysis, reduces the required sample size to the nanomolar level, which enables us to analyze N20 in the natural environment. For instance, tropospheric N20 has been found to have the ~-site preference for 15N (~)15N~- ~115Nf3)of about 20%0, while it is as much as 90%o in the stratosphere (Yoshida & Toyoda, 2000; Toyoda et al., 2001). Measurements of the N20 isotopomers from various sources and sinks are now in progress and several results have shown that the 15N-site preference is characteristic of each production/ consumption process (e.g. Yamulki et al., 2001; R6ckmann et al., 2001; Toyoda et al., 2002; Sutzka et al., 2003). Another approach to site-specific nitrogen isotope analysis, which utilizes infrared absorption spectroscopy of the N20 isotopomers, has also been developed (Esler et al., 2000a; Uehara et al., 2002). Although the sensitivity and precision of the measurement is not as high as that of mass spectrometric determinations at present, it will serve effectively as an alternative technique in the near future, since it has some advantages over the mass spectrometric technique (e.g. non-destructive analysis, less need for purification of N20 from other species).
Handbook of Stable Isotope Analytical Techniques, Volume 1 P.A. de Groot (Editor) 9 2004 Elsevier B.V. All fights reserved.
CHAPTER 20 Fluorination Methods in Stable Isotope Analysis Bruce E. Taylor Geological Survey of Canada, Ottawa, Ontario KIA OE8, Canada e-mail: btaylor@nrcan, gc.ca
20.1 Introduction Fluorine, is the most oxidizing element known (one of the few elements more oxidizing than oxygen), and possesses the highest electronegativity of all elements (e.g., Pauling, 1964). Consequently, fluorine or fluorine-bearing compounds have been the principal means by which to successfully facilitate the extraction of oxygen from oxygen-bearing substances for stable isotope analysis. Among most earth scientists, silicates and oxides have historically dominated the materials of interest treated by fluorination. However, fluorination techniques have also been successfully extended to other minerals, including phosphates, sulfides and selenites, in addition to elemental S, Se and U, among others). The purpose of this chapter is to review the most common fluorination methods currently applied in stable isotope geochemistry, incorporating both conventional and the more-recently applied laser-induced heating techniques, and to describe the construction and utilization of the basic apparatus used for fluorination employing fluorine (F2), or interhalogen fluorides (e.g., BrF5, C1F3) with or without the hydrogen halide, HF. It is not feasible to cover all possible variations in the details of customconstructed apparatus or of their application. Rather, this paper attempts to summarize the basic components and methods deemed reliable and functional. Although primarily intended for those not previously familiar with the fluorination techniques used in stable isotope geochemistry, I hope that this contribution will also be of use to the more-experienced. Fluorination is a robust technique for oxygen isotope analysis of silicates, oxides, and phosphates, as well as for sulfur isotope analysis of sulfides, one that is likely to continue to be used into at least the near future, whether for macro- or micro-analysis. As a result, it is important to incorporate health and safety issues in our discussion, as well as provide as much guidance as possible, based largely on our own experience, for the actual application of fluorination methods in isotope extraction.
Fluorination Methods in Stable Isotope Analysis
401
20.2 Background and overview
Early efforts by Manian et al. (1934) employing CC14 at 1000~ and producing CO as the analyte, for oxygen the isotope analysis of silicates were not encouraging, largely due to the lack of sufficient sensitivity (+ 2.5%o) to detect isotopic variation among the samples analyzed. Subsequently, two principal techniques were developed to extract oxygen from silicates and oxides: (1) reduction by graphite (carbon); and (2) oxidation by fluorine (+ HF), or one of its interhalogens. Carbon reduction of silicates and oxides was attempted before the adoption of fluorination as the method of choice, and met with some success in selected applications. Fluorination procedures dominated through the second half of the 20th century, but recent developments in automated, on-line reactors employing high-temperature furnaces coupled with continuous-flow mass spectrometric inlet techniques, have prompted some renewed interest in carbon reduction. Accordingly, a brief review of carbon reduction is also given below for completeness. Recent attempts at high-temperature pyrolysis of samples for oxygen isotope analysis (e.g., Kornexl et al., 1999a; Werner et al., 2001) indicate that replacement of silicate/oxide fluorination by carbon reduction, or a carbon-reduction related process, however tantalizing, may not be on the immediate horizon. 20.2.1 Carbon Reduction
Following the seminal work of Urey (1947) on the thermodynamic properties of isotopically-substituted compounds, several investigators sought techniques for the isotopic analysis of minerals (especially oxygen-bearing minerals) and water. Early attempts at carbon reduction of silicates produced CO, but determined the 180/16 0 ratio in one case by mass spectrometry on electrolytically recovered 02 (Manian et al., 1934), and, in the other case, by the measurement of the density of H20 formed upon reaction of the CO with H2 (Vinogradov & Dontsova, 1947). The reported uncertainties of the relative 6-values were, respectively, 2.5%o and 1.5%o. Subsequently, Adams (1949-1950) reduced quartz to CO for determination of its 180/16 0 ratio directly by mass spectrometry. However, accurate isotopic analysis was prevented by the presence of N2 and hydrocarbons, and, consequently; Adams converted the CO to CO2 by reaction with copper oxide at 455~ to facilitate cryogenic purification. Schwander (1953) published the first detailed description of the carbon reduction method, sources of uncertainty, and analytical results for various materials, from waters to different rock types. A graphite resistance furnace, mounted in vacuo and externally water-cooled, with power supplied by a 220V (primary)/6V, 3.5-4.0kW (secondary) transformer in combination with a variable transformer (secondary), permitted out gassing and carbon reduction at temperatures up to ca. 2200~ (Schwander, 1953). Isotopic measurements were made mass spectrometrically on CO, reportedly to an uncertainty in the relative &value of 0.4%o. The nominal C-reduction reaction is: Metal oxide(s) + C(s) ~ Metal carbide(s) + CO
[20.1]
Which, in the case of quartz, can be written as" SiO2(s)+ 3C(s) ~ SiC(s) + 2CO
[20.2]
402
Chapter 20- B.E. Taylor
Throughout this chapter, reactants and reaction products are assumed to be in a gaseous state during the reaction as written, unless indicated to be either liquid (1) or solid (s). Typically, sub-equal weights (ca. 60mg) of finely pulverized sample and graphite were mixed together and placed in a cylindrical, capped graphite container of ca. 0.4 cm3 volume. Following high-temperature out gassing of the furnace, and simultaneous heating of the sample to ca. 900~ the sample was inserted further into the furnace so that it out gassed at temperatures up to ca. 1600~ (measured by optical pyrometer) to remove all water and any traces of carbonate (Schwander, 1953). The temperature was then raised, and carbon reduction began generally above about 1700~ (from thermodynamic data for silicates, Clayton & Epstein (1958) estimated ca. 1727~ for quartz and 1200~ for iron oxides, and was usually complete in about 10 minutes, yielding virtually no more CO (from quartz) above ca. 2000~ The measured 6180 value was found to increase with yield, and that yields of at least 70-80% were required for reliable isotopic measurement. In cases where a silicate melt phase formed upon heating, and separated from the graphite, Schwander (1953) noted that yields were greatly reduced (10-30%) and c5180 values were too low. This was avoided by adding more graphite. Cryogenic purification using liquid air readied the sample of CO for mass spectrometry. Clayton & Epstein (1958) used a radio frequency (RF) induction furnace to heat a capped, 2g graphite crucible in vacuo surrounded by a platinum shield, to simplify the heating process. Use of the RF furnace avoided the presence of reactive, heated components in the vacuum system other than the graphite capsule and sample. To aid purification for mass spectrometry, product CO was converted to CO2 by means of nickel powder at 450~ as a catalyst. A pressed pellet containing ca. 10mg of sample and 100mg of graphite powder was prepared to ensure intimate contact between graphite and sample, and out gassed at a temperature, between ca. 1000-1250~ depending on sample composition. The tendency of graphite to absorb gases required pre-reaction out gassing of the crucible as well. For yields of ; 97%, Clayton & Epstein (1958) indicated a reproducibility of ca. 0.5%0. Dontsova's (1959) procedures and apparatus closely followed those of Schwander (1953). Reactions proceeded measurably above about 1000~ and were complete by 1900~ after 15-18 minutes. At this temperature, yields of ~ ca. 95% for a range of rock types (granites to dunites) and minerals (quartz and olivine) were achieved. Reproducibility was indicated to be between 0.2 and 0.4%o. Schwander (1953) noted that alkali metals and some metal carbides, volatile under the conditions of the carbon reduction, often condense as metallic films on the cooler walls of the reaction vessel and vacuum line. Clayton & Epstein (1958) suggested that reaction between released CO and such metal films can cause a marked reduction in yield (e.g., reduced to 70% in one case), with an attendant isotopic fractionation that depletes the residual CO in 180. Indeed, they found poor comparison (differences in 6180 of up to 5.2%o) between analyses for the same orthoclase by C-reduction and flu-
Fluorination Methods in Stable Isotope Analysis
403
orination, and attributed this to reaction between CO and metal films precipitated from volatized sample. The isotopic effect of the CO-metal film reaction was opposite to that reported by Schwander (1953) for cases where yields were low due to incomplete C-reduction. In contrast, Dontsova (1959) indicated in her study that CO-metal reaction was insignificant owing to the presence of dispersed graphite. Franchi et al. (1986) demonstrated the possibility of preparing CO by carbon reduction/ laser heating of small aliquots of silicates. They used a Nd-glass laser to heat a mixture of quartz and graphite pressed into a pellet. The product CO had a carbon isotope composition similar to that produced by combustion of the same graphite in an oxygen atmosphere, but the oxygen isotope composition of the CO was not determined. Using a Nd-YAG laser, Sharp & O'Neil (1989) heated small (3mm) pressed pellets consisting of sub-equal weights (< 1 to 3mg) of sample and graphite to temperatures above 1727~ in a glass vacuum line. Product CO was converted to CO2 by platinumcatalyzed, high-voltage discharge. They found that, with the exception of potassium feldspar, isotopic analyses were as precise as those by fluorination, and independent of yield. Erratic results for feldspar noted by Clayton & Epstein (1958) were also encountered by Sharp & O'Neil (1989), but Sharp & O'Neil (1989) found less positive, rather than more positive measured values of 6180 compared to accepted values. Again, a kinetic reaction may be the cause. Since CO was absorbed onto molecular sieve during the laser heating, selective decomposition rather than reaction between CO and precipitated metals was suspected (Sharp & O'Neil, 1989). The success of fluorination procedures described below in facilitating oxygen isotope analysis of even the most resistant of minerals with acceptably low uncertainty soon led to the abandonment of the carbon reduction method. Nevertheless, the apparent simplicity of reaction [20.1], safety, and the apparatus required, would seem to make this type of extraction still attractive in certain applications, providing that the sources of uncertainty (e.g., temperature of reaction; metal volatility; extent of reduction; etc) noted in Schwander (1953) and Clayton & Epstein (1958) could be overcome. With regard to modern laser-heating techniques, and on-line, continuous flow preparation procedures, it is encouraging that Schwander (1953) noted that rapid heating, although resulting in a low yield of CO (probably from the reaction of primarily grain surfaces), made no detectable difference in isotopic composition of the product CO. Consequently, the potential of this reaction, especially when assisted by a small amount of a fluorination reagent (M. Gehre, pers. comm., 2001) to assist in metal-oxygen bond breakage, has, perhaps, not yet been fully tested using recently developed, high-temperature, glassy-carbon furnaces (cf., Kornexl, et al., 1999a). High-temperature reaction with carbon in sealed quartz tubes (O'Neil et al., 1994) provides a convenient, precise alternative to fluorination for the oxygen isotope analysis of purified macro phosphate samples that, if properly calibrated in each lab, is accurate. This method is relatively straightforward, and requires simple vacuum apparatus found in many isotope laboratories. Moreover, a number of samples may
404
Chapter 20 - B.E. Taylor
be processed at one time. The sealed-tube method of O'Neil et al. (1994) can, by extension, be readily adapted to automated techniques employing high-temperature pyrolysis, with reduction in the size of the analyzed sample (e.g., 0.1mg) and an increase in throughput. An additional advantage of this sealed-tube method is the apparent lack of dependency of 6180 on reaction yield (O'Neil et al., 1994). Yet, as with silicates, oxides, and sulfides, analysis of phosphates by fluorination still provides the best accuracy and precision. Nevertheless, oxygen isotope analysis of phosphates by fluorination still provides the best accuracy and precision. Sulfates have proven to be less amenable to fluorination, although Pickthorn & O'Neil (1985) have demonstrated that a correction, required owing to incomplete, selective fluorination, may be used for improved accuracy. Carbon reduction, either by classical resistance heating (e.g., Nehring, et al., 1977), or external, radio frequency (RF) heating methods using macro samples of ca. 20mg, combined with spark discharge conversion of CO to CO2 (Nehring, et al., 1977), typically results in a combined accuracy and precision of ca. 0.1 to 0.2%o. High-temperature (1400~ pyrolysis of sulfate to CO, aided by nickelized graphite, can be carried out in an automated fashion considerably faster than by traditional methods (e.g., Nehring, et al., 1977), but at a cost of comparatively lower accuracy and precision ( ~ 0.5%0; Kornexl et al., 1999a). As noted in some detail below, fluorination is also the method of choice for the analysis of silicon isotope ratios, and for sulfide minerals, where precision and accuracy are of principal concern. The earliest isotopic analyses of selenium were carried out by fluorination (see below), but very recently instrumental methods have taken over, proving to be comparable in speed and precision. 20.2.2 Fluorination 20.2.2.1 Silicates and oxides
Successful and reliable oxygen isotope analysis of silicate and oxide minerals by fluorination was first demonstrated by Baertschi (1950), who found that sedimentary rocks contained more 180 than did igneous rocks. Detailed description of the fluorination method and apparatus by Baertschi & Silverman (1951) comprised the foundation of the fluorination technique, on which later refinements were based. Baertschi & Silverman chose C1F3 (+ HF) as the fluorinating reagent due to " ... the omission of certain necessary precautions .... " which prevented a successful test of F2 (+HF). Silverman (1951) used C1F3 to analyze a variety of rocks and meteorites. Others (e.g., Tudge, 1960; Taylor & Epstein, 1962; Clayton & Mayeda, 1978; Borthwick & Harmon, 1982) soon followed suit refining the fluorination method and expanding the range of reagents and applications. Two principal methodologies were developed: one method relying on fluorine gas (F2) derived from a high-pressure cylinder, and another method utilizing an interhalogen fluoride (e.g., BrF5 or C1F3) derived from a low-pressure cylinder. Though similar in many respects, each of these methods has its own advantages and disadvantages, as noted below, and both are in use today for macroand micro-analysis.
Fluorination Methods in Stable Isotope Analysis
405
Both conventional vacuum lines ("macro'-systems by today's standards) and laser-based ("micro"-) vacuum systems are used to release oxygen by fluorination. Each system offers particular advantages, and many laboratories find these "macro-" and "micro-" scale analytical tools complement each other. Conventional systems typically require 5-20mg of mineral or whole-rock powder, reacted in externally-heated vessels, generally for ca. 12-16hrs, whereas laser-assisted fluorination utilizes considerably less sample, typically on the order of 0.1mg (e.g., several grains or small fragment from a thin-section), or involve small in situ reaction volumes ("spots" or craters) some 150-500 mm in diameter, and of similar depths, and occurs often within seconds or fractions of a second. Thus, in some cases, the desired scale of analytical resolution, nature of the sample, or purpose of the analysis may guide the choice of analytical system. The diameter of the incident ion beam and the depth of sputtering determine the volume of a sample analyzed by ion-beam analytical methods (SIMS, or Secondary Ion Mass Spectrometry; see Ireland, T., Chapter 30, this volume). In contrast, volumes of minerals analyzed in situ by laser-assisted fluorination may be considerably larger than the nominal laser beam diameter, owing to the fact that the resultant volume is the result of a fluorination reaction. The size of the volume (i.e., the spatial resolution of analysis) depends upon a number of factors, including duration of reaction, temperature reached (in those cases where absorption of laser radiation produces the requisite heating), and pressure of fluorinating reagent. Isotopic analysis by SIMS can be accomplished at a higher spatial resolution relative to that by laser-assisted fluorination. However, one must accept a lower analytical precision and accuracy (owing to drift), and present intractability of some minerals to analysis (e.g., S: Eldridge et al., 1987; Chaussidon & Lorand: 1990; O: Ricuputi & Paterson, 1994), although the use of multi-beam collection, analytical precision for some SIMS analyses is nearing that capable by fluorination-based methods (Valley, pers. commun., 2003; see also Ireland, T., Chapter 30, this volume). As even many early SIMS-based investigations showed, however, the isotopic variations on a microscopic scale may readily exceed the analytical uncertainty (e.g., O, C, and H: Valley et al., 1998; S: Eldridge et al., 1987 and McKibben & Riciputi, 1998; Si: Zinner et al., 1987). In these cases, the high spatial resolution can be of great advantage.
20.2.2.2 Silicates for silicon isotopes Silicon isotope analysis of silicate minerals requires their fluorination to silicon tetrafluoride (SiF4) as the analyte used for the measurement of 30Si/28Si ratios (e.g., Taylor & Epstein, 1970). Samples containing a1% of carbonates, carbon, phosphates, boron-bearing minerals, sulfides and sulfides must first be purified using HC1 (carbonates, sulfides), high-temperature oxidation (carbon), or fluxing with NaOH and Na202 followed by dissolution in HC1 and firing at 1000~ (e.g., Ding et al., 1996). Allenby (1954) was the first to measure silicon isotope ratios in rocks, preparing SiF4 by fluorinating rock samples with HF at room temperature to 100~ Although his analytical procedure and the reporting of silicon isotope ratios as 28Si/30Si are no longer followed, his observation is still valid that rocks exhibit a comparatively smaller range in silicon isotope composition than found for oxygen isotopes. Success-
406
Chapter 20- B.E. Taylor
ful fluorination procedures for the isotopic analysis of purified silica or silicates used today employ either F2 (Taylor & Epstein, 1962; Douthitt, 1982; De La Rocha et al., 1996), or BrF5 (Ding et al., 1988; 1996; and 2003), and generally follow the procedures for oxygen isotope extraction from silicates and oxides. Purification of the SiF4 for mass spectrometry, as discussed in a later section, differs slightly from routine methods in oxygen isotope analysis. Reynolds & Verhoogen's (1953) early investigations of silicon isotope variations in rocks, sinters, and cherts made use of a rather involved wet chemical sample purification procedure that culminated in the precipitation of the stable compound BaSiF6. Compared to other methods of fluorination (discussed later), their procedure has the distinct advantages of producing not only a form of silica that is pure and decomposes under vacuum heating according to the reaction: BaSiF6(s) ~ BaF2(s) + SiF4
[20.3]
but also obviates the need for fluorination by gaseous or liquid fluorinating reagents, as is currently done. This simple reaction, in spite of the time required to purify and convert the sample to BaSiF6 obviates the need for an elaborate vacuum fluorination apparatus and purification protocol for the SiF4 produced. In fact, reaction [20.3] can be performed in sealed tubes, individually or as a batch, and stored for subsequent mass spectrometry (Reynolds & Verhoogen, 1953). Reynolds & Verhoogen found that the ca. 2.3% silicon isotope exchange between the SiF4 and the walls of the glass sealed tubes had a negligible effect on the measured isotopic composition of the sample. As found by subsequent workers, Reynolds & Verhoogen (1953) reported a small (ca. 3%0) variation of 630Si (reported in terms of the 30/28 mass ratio) among natural samples, which, they suggested was related to the temperature and mode of formation and warranted further investigation. Tilles (1961a), using the Reynolds-Verhoogen technique, extended the range of natural variations in 630Si to 5.3%0 (reported in terms of the 30/(28+29) mass ratio), and published the first measurements of permil silicon isotope fractionations between quartz and feldspar from a pegmatite that suggested a potential for silicon isotope geothermometry, and also indicated (Tilles, 1961b) a tendency for the order of enrichment (in 30Si) among common rock-forming minerals. Some minerals (e.g., olivine) were found to give low yields (ca. 80%), however, and, accordingly, results for these minerals are not reliable by the Reynolds-Verhoogen technique. Silicon isotope studies have been limited in number (but increasing recently), and largely focused on extra-terrestrial material because the variation in reported terrestrial 30Si/28Si ratios is comparatively small (Douthitt, 1982). Recent isotopic determinations of dissolved riverine silica Ding et al. (2002 and 2003) have demonstrated that the terrestrial variation of 630Si is as large as 3.0 (0.4 to 3.4%0), compared to that of dissolved marine silica (+0.06 to 2.2%0; De La Rocha et al., 2000), or to biologically-deposited silica (-1.2 to -3.7%o; De La Rocha, 2003). The routine measurement precision of 0.1%o nevertheless allows for the investigation of a number of phenomena of interest, especially among surficial processes, despite the fact that the terrestrial variation of
Fluorination Methods in Stable Isotope Analysis
407
silicon isotopes is not large by comparison with isotopes of hydrogen, carbon, nitrogen, oxygen and sulfur. 20.2.2.3 Phosphates Fluorination as a process for isotopic analysis is, of course, not restricted to oxygen isotopes in silicates and oxides. As is also the case for sulfides, fluorination of phosphates provides the most accurate and precise means of (oxygen) isotope analysis (Vennemann et al., 2002; L6cuyer et al., 1993; see L6cuyer, Ch., Chapter 22, this volume)). Fluorination of phosphates can be employed in both macro- and micro-techniques. However, the fact that oxygen can reside in several sites (PO4 -3, CO3- and OH), in biogenic apatite (Ca5(PO4)3-x(CO3)xOHyFl-y), and that additional, organic compounds, are present requires the isolation and purification of the phosphate-oxygen (as Ag3PO4 or BiPO4) from biogenic apatite prior to fluorination. The purification step imparts no isotopic fractionation to the results from biogenic phosphate (e.g., Karhu and Epstein, 1986), and is not required for analysis of abiogenic (e.g., hydrothermal or magmatic) apatite (Conway & Taylor, 1969; Fortier & L~ittge, 1995; Rhodes & Oreskes, 1999). Unlike Ag3PO4 or BiPO4, however, fluorination of abiogenic apatite [Ca5(PO4)3(OH,F,C1)] by BrF5 requires high temperatures (e.g., 650~ for complete yields (Rhodes & Oreskes, 1999; B. Taylor & Mirnejad, unpub.).
Micro-analysis of biogenic phosphate by direct, laser-assisted fluorination (e.g., Kohn et al., 1998; Jones et al., 1999; Lindars, et al., 2001), without purification of contained phosphate-oxygen, limits the accuracy and precision of the results, but relative isotopic variations may still be of great value owing to the high spatial resolution afforded by laser sampling. The same may also be said of direct laser heating of tooth enamel in a He stream (Cerling & Sharp, 1996) which provides an alternative, albeit still relative, method of micro-analysis that does not involve fluorination. The principal issue in the fluorination of biogenic phosphate remains the choice between direct, laser-assisted fluorination of the phosphate mineral and classical fluorination of a purified phosphate (Ag3PO4 or BiPO4) prepared from the original phosphate mineral by dissolution and re-precipitation. The former, albeit on a microscopic scale, may yield only a relative analysis of lower accuracy~ whereas the classical technique is preferred for best accuracy and precision. Currently, the sampling resolution of biogenic phosphate for classical fluorination is ultimately limited by the techniques of purification. 20.2.2.4 Sulfides Although the sulfur isotope ratios of sulfide minerals have mostly been determined from SO2 produced by oxidation, using either gaseous 02 (commonly used in elemental analyzers) or a solid source of oxygen mixed together with the sample (e.g., V205 or mixture of CuO/Cu20), fluorination of sulfide minerals to produce SF6 can facilitate the most precise and accurate analysis. The reasons for this include both (1) mineral/reaction-specific factors, especially in micro-analysis (e.g., mineral composition dependent fractionation; Crowe et al., 1990; Kelley and Fallick, 1990), and (2) cross-contamination in the ion source of the mass spectrometer. In the former
408
Chapter 20- B.E. Taylor
instance, accuracy of the results is compromised by the mineral composition data. In the latter case, cross-contamination or memory in the ion source results in variable contraction of the 6 scale in different laboratories. Fluorination of sulfides requires apparatus similar to that used for oxides and silicates. However, owing to the mass of SF5+ (127), the principal ion measured in the mass spectrometry of SF6, an isotope ratio mass spectrometer that can measure to at least a mass of 130 is required, and above 131, if 636S is to be measured. Although high-mass capable models are readily available, only a few labs today utilize fluorination for sulfur isotope analysis and gas-source isotope ratio mass spectrometers more commonly purchased today have a mass range on the order of 2-80 and, with regard to sulfur isotope analysis, are used for measurement of SO2. Although conceived nearly 40 years ago (Hulston & Thode, 1965b), sulfur isotope analysis by fluorination has not, with few exceptions (e.g., Rees & Thode, 1977; Rees, 1978; Ding et al., 1985), commonly been employed, and has remained virtually dormant as a technique from the mid-70's to early 90's. The superior capabilities of SF6 in microanalysis (and for determination of 633S: Hulston & Thode, 1965b; Thode & Rees, 1971; Ding et al., 1987), however, have caused a recent renaissance in sulfide fluorination, almost exclusively by laser-assisted methodologies (e.g., Rumble et al., 1991, 1993; Beaudoin & Taylor, 1994; Farquhar & Thiemens, 2000; Farquhar et al., 2000a; Taylor et al., 2001, in press). 20.2.2.5 Water
The classic CO2-equilibration technique (Epstein & Mayeda, 1953) has been the method of choice because it is relatively easy and precise. On the other hand, fluorination offers several advantages, especially for the oxygen isotope analysis of microlitersized samples. Uncertainties of many types can be introduced in the isotopic analysis of water, from the calibration of internal laboratory (gas and water) standards and the experimentally determined mineral-water and CO2-H20 fractionation factors, to control of the conditions governing equilibration and analysis in an individual laboratory. Therefore, direct analysis of water would be preferable, at least theoretically, by either fluorination (O'Neil & Epstein, 1966b), reduction by carbon (Majzoub, 1966), or reaction with guanidine hydrochloride (Dugan et al., 1985).
20.3 Fluorinating reagents A number of physical and chemical properties of the commonly used fluorination reagents (HF, F2, BrF3, BrF5 and C1F3) have influenced both the choice of reagent and the development of different fluorination methodologies. Other sources of fluorine (both solid and gaseous) have been variously used either as fluorinating regents, or as an aid to laser ablation or pyrolysis (e.g., Kornexel et al., 1999). The solid compound XeF2 has, apparently, been used with some success as a source of fluorine by Dmitry Krylov, a Russian scientist (S. Hoernes, pers. commun., 2003). His procedure, in brief, comprises weight ratio of XeF2/sample of ca. 10"1 (e.g., 10-15mg of sample added to ca. 100rag of XeF2), loaded together (in a dry box) and reacted, as in the BrF5 procedure, in nickel tube at temperatures up to 600~ (above 600~ the XeF2 dissociates; it melts at 140~ This procedure has received little attention, but the ease of handling small amounts of the fluorinating reagent might prompt some interest in its applica-
409
Fluorination Methods in Stable Isotope Analysis
tion in microanalysis. Other substances, such as Freon R-134a (CH2FCF3), for example, have been used recently to increase yield of the chemical and Pb-isotope analysis of zircon by laser ablation using the ICP-MS (Hirata, 2003). The application of such fluorocarbons in more traditional methods of stable isotope analysis has not been widely pursued, however, and is likely to be fraught with problems owing to mass spectrum interference and poor pumping characteristics (c.f., Rumble & Hoering, 1994). Differences in freezing points (Table 20.1), in particular, have played a large role in how the interhalogen fluorides on the one hand, and fluorine gas, on the other, have been utilized. Whereas, interhalogen fluorides can be transferred cryogenically in the vacuum line, fluorine gas cannot under normal circumstances. Cryogenic transfer of a fluorinating reagent facilitates the measurement of an appropriate aliquot size that may then be condensed into each sample tube without pre-mature reaction. The ability to cryogenically transfer the interhalogen fluorides also facilitates the post-reaction separation of residual reagent from extracted oxygen. Other physical properties may influence the choice of one interhalogen fluoride over another. For example, the vapor pressure (Table 20.1, cylinder pressure; psig, or pounds-per-square-inch gauge pressure) of BrF3 is sufficiently low as to render its use in typical vacuum fluorination apparatus rather impractical, and, for this reason, it is not commonly used today. Awareness of its properties is important, however, as BrF3 can occur as an impurity in BrF5 and create some difficulties.
Table 20.1 - Selected physical and chemical properties of fluorinating agents and fluorine compounds I Property
HF
F2
BrF3
BrF5
C1F3
CoF3
CaF2
KF
XeF2
State (S.T.P.)
gas
gas
liquid
liquid
liquid
solid
solid
solid
solid
Molecular Weight
20
38
136.9
174.9
92.5
1 1 5 . 9 3 78.08
58.1
169.29
Boiling Point @ 100 kPa (~
19.54
-188.1
135
41.3
11.3
1400
ca2500
1505
114.35
Freezing Point @ 100 kPa (~
-83.1
-219.6
8.8
-62.5
-83
ca1200
1423
858
129.03
Cylinder Pressure (kPa) @ 20 ~
6.21
2760
0.93
43.2
46.92
Critical Temperaure
188
Specific Gravity (H20 = 1.000)
0.99
(oc)
197
1.69
2.49
2.48
1.77
1 Lide, D. R. (ed.), 1996, Chemical Rubber Comapany Handbook of Chemistry and Physics: CRC Press, Boca Raton, Florida, U.S.A., 77th ed.
410
Chapter 20 - B.E. Taylor
Although all fluorinating reagents pose potential (but manageable) health and safety risks, the use of F2 requires particular attention to safety and handling precautions as a commercially available tank or cylinder of F2 is typically under considerable pressure (e.g., ca. 400psig; see Table 20.1). The cylinder pressures of the interhalogen fluorides, on the other hand, are all less than one atmosphere. Minor leakage of a cylinder containing a halogen fluoride would initially result in migration of air into the tank, rather than halogen fluoride out of the tank. Purity also distinguishes the commercially available fluorinating reagents. Halogen fluoride compounds can be obtained in rather high purity, and further purified by vacuum distillation. A tank of commercially available F2, however, usually has a purity of only ca. 97%. Contaminants, among others, typically include oxygen. Its presence contributes, at least in a small way, to the uncertainty of each analysis, in spite of the fact that such a "blank" of oxygen can be quantified, and for which a corFigure 20.1 - Conventional fluorination line of metal and glass construction for the extraction of oxygen isotopes from silicates, oxides and phosphates utilizing either halogen fluoride reagents (e.g., BrF5 or C1F3) or F2 gas, and employing top-loading (see Figure 20.5), externally-heated reaction tubes that facilitate rapid, blank-free (i.e., without water) loading of samples. Components utilized with F2 gas, either commercial or laboratory generated and/or purified are shown in insets F and J. Details of construction may vary from lab to lab, but may include the following components: (A) Hg (or other) diffusion pump, backed by roughing pump to yield vacuum of ___10-3Torr; (B) Argon and drying agent (C; e.g., 4A mol. sieve) to ensure dry inert gas used primarily during sample loading; (D) KBr (or other salt), externally heated to 150-200~ used to passivate excess reagent (F2 or halogen fluoride compounds) by reaction to form KF and Br (trapped at LN2 temperature on P-style trap; see Figure 20.2); (E) halogen fluoride storage, and Kel-F and nickel tubes for distillation, visual inspection, and storage of working aliquot for 2-300 analyses; (F) commercial F2 supply and/or Asprey-type F2 generator for purification/production of pure F2; (G) nickel reaction tube and cooling collar fitted with rapid loading top (see Figure 20.3; one 12 tubes attached to manifold of stainless steel tubing); (H) furnace, thermocouple (tc) and temperature controller, or variable transformer, for the external heating of reaction tubes (G); (I) Pt-catalyzed carbon rod reactor to convert 02 to CO2 (see Figure 20.4); and (J) in-line reactors (2; not shown in detail) to convert F2 to Br (or other halogen) allowing separation of 02 from F2-based fluorination. Vacuum is monitored by G-1 and G-2 (capacitance manometers, e.g., Setra | 0-10psi absolute (G-l) and 0-25psi absolute), and thermocouples or other, similar range gauges, G-3 to G-5. Cold traps T-l, T-2, and T-3 in stainless steel portion of line are held at liquid nitrogen temperature and utilized during, respectively, sample loading, post-reaction conversion of excess halogen fluoride reagent, and during extraction of oxygen. Cold trap T-4, is held at liquid nitrogen temperature to insure against loss of CO2 formed during conversion of 02 (most of which is trapped in converter vessel; see Figure 30.4). U-trap T-5 facilitates measurement of the 02 reaction yield (as CO2) and cryogenic purification of the CO2 sample if required. All valves are high-vacuum valves. Those in the metal section of line are a 'diaphragm' style valves (e.g., Nupro| SSDLS-4 and SS-DLS-6, with Swagelock @ tube fittings) except for the reaction tubes, which utilize toggle-style bellows valves (Nupro| SS-4BKT; see Figure 20.3). Pyrex tubing (1 / 4in or 6mm) is attached to the vacuum line below G-5 using O-ring connectors. CO2 is collected in these tubes with liquid nitrogen then sealed in lengths of 5-6in with a torch. Note that the exhaust of both rotary (roughing) vacuum pumps is vented outside the laboratory. A copper wool-filled trap between the Hg diffusion pump and roughing pump acts to reduce transmission of any Hg vapor.
Fluorination Methods in Stable Isotope Analysis
411
412
Chapter 20- B.E. Taylor
rection can be adopted (Taylor & Epstein, 1962). Alternatively, relatively pure F2 can be generated (or, purified from another primary source) in the laboratory at low pressure using the method of Asprey (1976). This method is in use with both conventional fluorination lines for macro scale analysis (e.g., University of Bonn) or, in laserassisted fluorination apparatus for micro-scale analysis (e.g., Rumble, et al., 1993; Taylor & Beaudoin, 1993; Beaudoin & Taylor, 1994), as described briefly below. Reagent cost and availability further separate F2-based from interhalogen fluoridebased techniques. Bromine pentafluoride (BrF5), a reliable reagent for many years, was once available rather inexpensively when it was used in industrial applications. More recently, BrF5 has been difficult to obtain, and then only at a cost of several thousand dollars (U.S.) per kg. Chlorine trifluoride (C1F3) is a cheaper alternative, and can be more readily obtained than BrF5. In contrast, a tank of F2 costs even less than either of these interhalogen fluorides, and is available from specialty gas companies. Be forewarned that the availability and shipping fluorine and some of its compounds may strictly controlled in some countries. The sealed integrity of reagent cylinders should be of utmost concern, and care should be taken that their shipment does not expose the cylinders to environmental extremes such as freezing conditions that could lead to valve leakage. Inspection of the cylinder valves should be made for stains or deposits (red-brown in the case of BrF5, for example) which would provide some indication of leakage.
20.3.1 Purification of fluorinating reagents The interhalogen fluorides BrF3, BrF5 and C1F3 can all be purified by vacuum distillation, although the low vapor pressure and relatively high freezing point of BrF3 renders its purification by this means more time-consuming. Trace amounts of noncondensable gases such as oxygen, nitrogen, argon, fluorine, SiF4, CF4, SF6, etc. can be readily separated by cryogenic, vacuum distillation using a -70~ trapping temperature (e.g., frozen/liquid ethanol mixture made by adding liquid nitrogen to ethanol). The process comprises cryogenic transfer of an aliquot of reagent from one small storage vessel to another, followed by pumping away of all gases not condensed with the reagent. Typically, this involves 1-2 distillations, although some investigators prefer triple distillation. Cylinders of reagent used infrequently may develop a headspace of non-condensable gases such that the first few aliquots of reagent removed may have relatively more associated contaminants. It is advantageous to be able to actually see the liquid reagents. For this reason, semi-translucent Kel-F, a fluorinated plastic, is often used to manufacture a storage vessel for a working aliquot of the halogen fluoride (see Figure 20.1). A vessel of the same dimensions as the nickel reaction tubes is convenient. In the case of BrF5, purity can, in part, be assessed from its color. A clear, pale yellow liquid is characteristic of pure BrF5, whereas the presence of a reddish or brownish hue indicates contamination by Br (and, in this case, possibly some BrF3 as well). Bromine has a freezing point o f 7.2~ allowing for its cryogenic separation from BrF5. Exposure of the working aliquot of liquid BrF5 to gaseous F2 apparently can be carried out in order to fluorinate any excess Br and BrF3. This was apparently routine in the laboratory of Prof. S.
Fluorination Methods in Stable Isotope Analysis
413
Epstein, California Institute of Technology (G. Yoshiba, pers. commun., 2003). Fluorination of BrF5 consisted of overnight exposure of the BrF5 to F2 at ca. l atm pressure and room temperature. The fluorination reactions are exothermic, and it is advisable to work with small quantities (e.g., 20-30 ml) of BrF5, preferably in a Ni vessel. The efficiency can be visually assessed after excess fluorine has been pumped away and neutralized (see below), and the reagent transferred to a Kel-F vessel. For example, the conventional fluorination line illustrated in Figure 20.1 has both a Kel-F vessel and a Ni vessel, each of similar size to the Ni reaction tubes, and is constructed in such a way as to facilitate reagent 'pre-treatment' upon connection to a source of F2. Alternatively, a reaction tube could be utilized. Commercial F2 can be readily purified in small batches by means of the K2NiF6,KF- filled F2 generator ('fluorine pump') originated by Asprey (1976). When heated, K2NiF6.KF decomposes, releasing pure F2 according to the following reaction: 2[K2NiF6. KF](s)~ 2[K3NiF6](s) + F2
[20.4]
This reaction is reversible, and re-fluorination of solid K3NiF6 can be accomplished by exposure to F2 at about 125~ Below this temperature, impurities such as oxygen, nitrogen, argon, etc. can be pumped away. Asprey (1976) showed that thermal cycling of the F2 generator, followed by cooling below 125~ and additional evacuation, could result in the production of relatively pure F2 at a low pressure. The pressure of F2 that develops as the generator is heated should be closely monitored because, although relatively low, the pressure increases markedly with temperature as reaction [20.4] proceeds. The required K2NiF6. KF can be purchased commercially, or made in the laboratory. For example, the procedure followed at the University of Bonn (Prof. S. Hoernes, pers. commun., 2003) comprised mixing together 3mol NiF2 and 9mol KE and loading into a nickel (or, Monel) vessel sealed with a Cu-gasketed Conflat~ flange to construct a 'fluorine generator' suitable for a standard fluorination line (such as shown in Figure 20.1). K2NiF6.KF is quite hygroscopic and the 'fluorine generator' (and also any commercially-purchased K2NiF6.KF similarly contained) must be activated by several fluorination cycles as follows: evacuate and heat to at least 300~ to liberate any HF which formed that must be pumped away. The generator is exposed to a raw fluorine pressure of some 3-5b while heating to 500~ and which point the F2 tank is closed and the fluorine generator allowed to cool slowly to ca. 250~ during which the K2NiF6.KF is fluorinated. Any gases not absorbed by the K2NiF6.KF are then pumped away. Asprey (1976) recommended a F2 pressure of 10b. However, 5b is a typical output limit on commercial, 2-stage F2 tank regulators, and this is sufficient for the purpose. The cycle is repeated (perhaps several times) until the F2 pressure drops below the starting pressure. The 'fluorine generator' is then ready for service as described in the previous paragraph.
414
Chapter 20 - B.E. Taylor
Figure 20.2- Pyrex glass Pstyle trap used to collect Br (if KBr is used to passivate a fluorinating reagent, or C1 in the case of NaC1, for example) at liquid nitrogen temperature and to evacuate the KBr reactor (component D, Figure 20.1) and vacuum line. After thawing, the Br is transferred to a vent line with the aid of a few psi of dry Argon (components B and C, Figure 20.1), or other inert gas. Vacuum seal between custom stainless vessel top (dimensions as in Figure 20.3) and metal/glass seal is made with Teflon ferrules. Valves are stainless steel, high vacuum valves (e.g., diaphragm type, Nupro | SS-DLS-6; Monel stem tips where adjacent to KBr reactor, otherwise Kel-F stem tips) that permit valve part replacement, if needed, without removal from line.
Asprey's (1976) method has garnered considerable attention and is becoming the method of choice for safe production or purification of very pure fluorine. Another method, described by Jacob & Christie (1977), requires the removal of 02 and other trace impurities from commercial F2 in two stages. Trace impurities (except N2 and Ar) having no measureable vapor pressure just above LN2 temperature (-196~ can be removed by distillation at this temperature. Others require colder trap temperatures (e.g., -210~ achieved only with some difficulty in standard stable isotope labs (e.g., by streaming He through LN2). Trace amounts of 02, on the other hand, can be effectively removed from F2 according to Jacob & Christie (1977) by the following reaction carried for 2 hours in a closed Monel vessel at ca. 190~ 0 2 + F2
nSbF5 ~ S b O 2 F 6 . (n-1)SbF3
[20.5]
The Monel reaction vessel is then cooled to-183~ and the O2-free F2 distilled to another vessel held at-196~ Reaction [20.5] might be used with F2 purified according to Asprey (1976) as an extra precaution in micro-analysis for oxygen isotopes.
Fluorination Methods in Stable Isotope Analysis
415
Figure 20.3 - Rapid-loading reaction tube "chimney" caps from stainless steel. The vacuum seal of the Ni reaction tube to the "chimney" cap is made with Teflon ferrules and a 3/4" Swagelock | nut (not shown; see Figure 20.4). The top (0.375in O.D.) of the tube ("chimney") is closed with a Swagelock @ cap (SS-600-C), also fitted with Teflon ferrules (see Figure 20.4).
20.3.2 Reagent storage The storage of F2 and other fluorinating reagents often presents a problem for the researcher in carefully regulated working environments. The fear of an accident involving stored reagent can be acute among those charged with safety and regulatory responsibilities. This is not to downplay such concerns, which should be foremost in all of our minds, but the requirements placed upon laboratories can, at times, create other potential sources of worry. Typically, BrF5 and other interhalogen fluorides are kept in the cylinder in which they were received from the commercial source, and securely supported and attached to the vacuum fluorination line. This makes good sense inasmuch as the means of neutralization is at hand, and there is no concern regarding moving or relocation of such cylinders. In some cases, the cylinder of reagent is kept in a fume hood, and connected to the vacuum line via metal tubing. The connecting pipe work between such cylinders and the vessels used for storing small quantities (e.g., Kel-F vessels) should be as short as possible. Lengthy pipe work simply slows the process of reagent transfer and increases the possibilities of leaks via needed connections, etc. In rare cases
416
Chapter 20 - B.E. Taylor
where the cylinders and valves received are poorly manufactured (as is the case from one supplier of BrF5 on the international market), cryogenic transfer to a pre-treated vessel having a reliable valve is most advantageous, if not absolutely necessary. As noted before, the less than one atmosphere cylinder pressure of BrF5 typically minimizes the danger of direct leakage into the laboratory. The presence of a reddish brown stain or white deposit around the cylinder valve, or its connections, however, provides evidence of leakage and transfer to a more reliable vessel is needed. Transfer should be carried out slowly on the vacuum fluorination line (where the means for reagent neutralization is accessible if need be), and monitored by a vacuum / pressure gauge. Steel, Monel| or nickel are suitable cylind e r materials for storage of fluorinating reagents, Such cylinders should be obtained new from a commercial source, and carefully leak-checked under vacuum before requisite pre-treatment with small amounts of
Figure 20.4 - (A) The Ni reaction tube, with attached "chimney" cap, is connected to the manifold of fluorination line (inset G, Figure 20.1) with a toggle-style, bellows high vacuum valve with Kel-F stem tip (Nup r~174SS-4BKT-KF). A 60mm stainless steel chromatographic filter (Swag el~174 p / n 696-530), inserted in the reaction tube side of the valve, prevents pieces of Ni- or other metal fluorides and sample particles from entering the valve seat. Note the direction of gas flow, as indicated by an arrow on the valve body, which ensures that the valve stem tip and not the bellows of the valve, is exposed to the fluorination reagent during the reaction. The toggle-design exerts sufficient pressure to guarantee a vacuum-fight seal, and insures against over tightening. (B) Stainless steel loading tool, with highly polished interior end, closed by a spring-loaded rod soldered to a stainless-steel bearing, permits rapid sample loading through the "chimney" of the reaction tube's "chimney" cap (see also Figure 20.3). The reaction tube and manifold are filled with slightly greater than latm Ar. The Ar flows continuously out of the "chimney' cap during loading, which prevents entry of atmospheric moisture to the reaction tube. A weighed sample is placed in the loading tool, and released at the bottom (only) of the reaction tube by depressing the spring-loaded central rod.
Fluorination Methods in Stable Isotope Analysis
417
Figure 20.5 - Water-cooled, C-rod converter forms CO2 from extracted 02, which is condensed with liquid nitrogen at the bottom of the converter vessel. The platinum coil simultaneously heats the ca. 30mm x 3mm O spectrographic graphite rod, simultaneously, with the surrounding platinum shield, catalyzes the oxygen-limited reaction (based on a design by T. K. Kyser; pers. commun.). The temperature of the carbon rod can be visually monitored through the water-cooled glass reaction vessel. This design requires but a very small amount of liquid nitrogen, and facilitates conversion of ca. 100mm of 02 in less than 5 minutes. A socalled "back diffusion trap" (T-4, Figure 20.1) prevents loss of CO2 not condensed initially at the bottom of the converter.
reagent (at room temperature, followed by mild heating) until no m o r e 0 2 is evolved. Under no circumstances should cleaning and/or re-use of older cylinders be attempted- potentially dangerous residues of low vapor pressure compounds (e.g., BrF3) may reside in these cylinders. Similarly, no reagent cylinder should ever be opened later to the atmosphere, but disposed of properly (typically returned to its source or another qualified receiver). Leakage of cylinders of F2 is another matter. Where possible, the storage of highpressure (ca. 400psig, or 2760kPa; Table 20.1) cylinders of F2 in the laboratory itself should be avoided to minimize health and safety concerns. One solution is to pipe the F2 to the laboratory from a storage room. However, such piping must be accessible for
418
Chapter 20 - B.E. Taylor
monitoring, protected and leak-free. Exposure to environmental factors (e.g., by extremes in temperature) could compromise the leak-free integrity of the gas-delivery system. An alternative, safe-storage solution for a cylinder of F2, particularly viable where smaller cylinders are used to recharge Asprey-type "fluorine pump", is to utilize a mobile, specialty gas storage. When not in use, the cabinet can be stored in an appropriate and safe location. At the Geological Survey of Canada (Ottawa), for example, a commercially available specialty gas storage cabinet, designed for use with hazardous gases and fitted with appropriate piping, regulator, valves, and HF getter (NaOH) is mounted on wheels and fitted with an exhaust port for connection to a vent system when in use in the laboratory. An HF sensor and auto-dialing mechanism provide unattended monitoring of the cabinet and its contents while in a separate, appropriately labeled storage room.
20.3.3 Neutralization and disposal of reagents With the same regard for safety in the storage and use of fluorine and fluorinating agents, the waste products of fluorination reactions and excess reagent require safe handling in their neutralization and disposal. In systems where F2 is the chosen reagent, on-line neutralization is required during the extraction and separation of oxygen. Neutralization of both F2 and halogen fluoride compounds can be readily accomplished by reaction with crystalline KBr (ca. 2-Smm grain size) at ca. 150-200~ to yield Br. The reagent passivation section of the fluorination line in Figures 20.1 and 20.2 illustrates one way in which this can be conveniently and safely accomplished. A stainless steel cylinder containing ca. 500g of coarsely crystalline KBr, closed at each end with Cu-gasketed flanges (e.g., Conflat~) and fitted with fine Monel screen at the inlet and outlet, is externally heated and insulated. A thermocouple affixed to the cylinder mid-way between flanges permits monitoring of the temperature. Coarse scrap KBr (cut-offs and broken single crystals) from the optics industry is particularly well suited as it is of high purity, contains a minimum of water, and can be appropriately crushed, and sieved to a narrow size range so as to maximize the life and permeability of the reactor. Too fine a grain size and/or too large a range in grain size can lead to clogging and inefficient throughput and pumping. The apparatus and methodology for BrF5 neutralization employed at the Geological Survey of Canada (Ottawa) involves on-line collection of excess reagent, neutralization (conversion to KF and Br) and storage of Br after each set of 12 fluorinations. The procedure entails collection of excess reagent and volatile waste products (e.g., Br) in a liquid-nitrogen cooled trap (%2, Figure 20.1) during an hour-long evacuation of the reaction tubes while heating (typically to 550-650~ depending on the next samples to be run; a slightly-higher-than-reaction temperature is preferred for out gassing). The collected excess reagent, waste products and KBr reactor are isolated together from the remainder of the vacuum line for ca. 20 minutes by closing valves 10 and 11 (Figure 20.1). The contents of trap T-2 are allowed to thaw naturally and react with the heated KBr, after which, the product Br is condensed into a glass P-trap maintained throughout the week at liquid nitrogen temperature and normally left
Fluorination Methodsin StableIsotopeAnalysis
419
open to the KBr reactor, which degasses Br slowly, via valve 11 (Figure 20.1). The small amount (< 10ml) of Br, accumulated from reaction of 48-60 samples in one week, is vented from the trap and lab area through valve 13 to power vented ducting by a ca. 10 psi stream of Ar. The Br is first thawed and then valves 13 and 14 are opened simultaneously. Warming the Br storage vessel (e.g., with a heat gun) while flushing with Ar ensures complete removal of Br. This procedure avoids exposure of the lab or its occupants to any Br vapor. In some labs, the Br is condensed into a removable vessel that is taken to a fume hood, under which the liquid Br may be collected and disposed of by others. In an alternative procedure, Ding et al. (1996) describe neutralizing residual BrF5 and Br by bubbling the thawed waste products through a Ca(OH)2 solution, using Ar as a carrier gas under a fume hood, to produce harmless CaF2, CaBr2 and 02. A similar method for passivation of Br (and C1 from NaC1 or KC1 reactions) that allows for their safe disposal via the drain has long been in use at the University of Bonn (S. Hoernes, U. Bonn, pers. commun., 2003). This method entails dissolving in 2N Na(OH) (80g NaOH + 21 H20), followed by neutralization of the alkalic solution with sodiumthiosulfate. The neutralized solution can then be safely flushed down the drain. On-line neutralization of F2 is necessary not only for the disposal of excess F2, but also to effect separation of F2 from oxygen. Because F2 is not condensed at LN2 temperature at the low pressures involved, a mixture of F2 and 02 is released from each reaction vessel after fluorination, while Br and other reaction products are retained in the reaction tubes submersed in liquid nitrogen. As in the above methods, KBr has traditionally been used to yield Br upon reaction with F2, which may then be cryogenically separated from 02 for later disposal. To overcome the additional reaction time necessar?4 and to facilitate an efficient collection and conversion of oxygen, Taylor & Epstein (1962) devised two heated KBr reactors (e.g., inset J, Figure 20.1), one of which has a small volume to allow efficient conversion of the final remaining F2 and accelerate the collection/purification of oxygen. 20.4 Fluorination reactions 20.4.1 Fluorination of silicates and oxides by F2 gas The F2-based technique and apparatus described by Baertschi & Silverman (1951) became the backbone procedure for a number of labs, especially those at the California Institute of Technology (Taylor & Epstein, 1962) and in laboratories subsequently established elsewhere by students of Professors H. P. Taylor, Jr. and S. Epstein. Compared to halogen-fluoride based techniques, widespread use of F2 as the fluorinating reagent of choice in conventional (macro-) vacuum lines was perhaps hindered by some of the analytical and safety. However, the advent of micro-scale analytical techniques employing lasers has seen a renaissance of the use of purified F2 as the preferred fluorinating reagent owing to the simplicity of the fluorination reaction (Table 20.2).
420
Chapter 20 - B.E. Taylor
A typical reaction between an oxygen-bearing mineral (a metal oxide, where M metal) and F2 can be represented by" MxyOy(s) + yF2 ~ xyMF2/x(S) + Y / 2 0 2
[20.61
which yields solid metal fluorides (MF), oxygen gas (02) and, in the case of silicates, volatile fluoride compounds such as SiF4 (and others). For the mineral titanite (CaSiO5), for example, this becomes: 2CaTiSiO5(s) + 10F2 ~ 2CaF2(s) + 2TiF4 + 2SiF4 + 5 0 2
[20.7]
From the stoichiometry of reaction [20.7], and others involving halogen fluoride reagents discussed below, two moles of fluorine (F2; or halogen fluoride reagent) are required per mole of oxygen (02) for complete reaction. In practice, a fluorine/oxygen ratio of ca. 2.5-3.3 (or, 1.2 to 1.6 for the excess/stoichiometric fluorine ratio) was used by Taylor & Epstein (1962). Reaction of fluorine with the hot walls of the reaction vessel claims some of the loaded reagent, but this is not a major concern at the typical reaction temperature of 450~ (Haimson & Knauth, 1983). When insufficient reagent is present, the amount of oxygen released varies directly with the fluorine/oxygen ratio or, "fluorine excess" (Haimson & Knauth, 1983; Figure 20.6). The effect on 6180 values is variable, depending on the fluorinating reagent used. Epstein & Taylor (1971) and Haimson & Knauth (1983) found partial extracFigure 20.6 - Plot of fractional oxygen yield (as % of total) vs. fluorine tion of oxygen with excess (molar F2/02) in partial fluorination reactions at constant temperF2 gas occurred ature (450~ and time. A linear relationship between yield and fluorine without fractiona- excess for both quartz and opal A exists until F2/02 ratio is greater than tion of oxygen iso- 2 (after Haimson & Knauth. 1983). Notably the measured values of 6180 topes. This was true are constant for quartz-derived oxygen, but vary for opal A vary, owing whether the F2 pres- to progressive dehydration with fluorination. In contrast, isotopic analyses of oxygen derived from incomplete fluorination with BrF5 are inaccusure was varied at rate (Garlick & Epstein, 1967). constant temperatu-
Fluorination Methods in Stable Isotope Analysis
421
re (Haimson & Knauth, 1983), or the reaction duration was reduced at constant temperature and F2 pressure (Epstein & Taylor, 1971). Oxygen yields from quartz may be inconsistent when the amount of fluorine present is <2 times the stoichiometric requirement (Haimson & Knauth, 1983; Figure 20.6). Similarly, when using BrF5, the values of 6180 are often too high whenever oxygen is not quantitatively extracted (Garlick & Epstein, 1967). The above findings on the effects of variable F 2 / O 2 may have implications for micro-scale analysis. The thermal aureole (melt zone) surrounding a reaction crater from in situ analysis may be a zone of partial reaction. A change in the stoichiometry of the sample in the thermally affected aureole about a reaction crater has been presumed to indicate such a fractionation in oxygen isotopes (e.g., Elsenheimer & Valley, 1992) when using BrF5, and is often assumed a priori when using F2 (e.g., Wiechert & Hoefs, 1995; Rumble & Sharp, 1998). This is discussed further later on. The inability to condense F2 and 02 at the temperature of either liquid air or liquid nitrogen provides a set of technical issues to overcome, including the safe neutralization of any remaining F2, and the management of such an aggressive reaction as [20.6]. Care must be taken to neutralize and remove F2 from the 02 in order to ensure it is not introduced to the mass spectrometer, either as F2 or, where 02 is converted to CO2, as CF4 or COF-type compounds. The presence of F2 (or CF4) in the mass spectrometer source typically leads to poor vacuum pumping characteristics and to inaccurate (often too low) oxygen isotope ratio measurements. Separation of 02 from excess F2 after a fluorination reaction is readily carried out by neutralization of the F2 by reaction with a salt such as potassium bromide (KBr) or sodium chloride (NaC1) in a flow-through reactor (inset J, Figure 20.1). F2 is neutralized (typically at ca. 150-200~ by a reaction similar to: F2 + 2KBr(s) ~ 2KF(s) + 2Br
[20.8]
whereby Br can be cryogenically separated from 02 using liquid nitrogen. Here, the volume and the grain size of KBr (e.g., 2-5mm; careful crushing, cleaning and sizing are essential) in the neutralization reactor constrain the reactor's conductance and the time required for complete neutralization. Two such reactors were utilized by Taylor & Epstein (1962) to increase the rate of transmission of the 02 + F2 reaction products through the vacuum line, thereby reducing the overall extraction time for one sample. Purity of the KBr can be another factor. In the case of small-sample preparation lines, such as those employing laser heating, optical grade KBr or NaC1 (e.g., available as scrap material from the optics industry) may be used. Alternative means of F2 neutralization (but for small or trace quantities) in the sample preparation stream include also reaction with metals such as steel wool or mercury (e.g., Akagi et al., 1995; Sharp, 1990). Oxygen is not the only byproduct of reaction [20.6]. Indeed, a number of fluorinebearing compounds result, some of which are condensable, and others not, at liquid
Table 20.2-Selected macroscopic fluorination techniques Sample Type
Typical Size
Reagent
Pretreatment Sample
Pretreatment Sample + vessel /line
MS Analyte
Rxn Temp. (~
4~ to to Rxn. duration
r= r to o !
Table 20.2 continued >
O
> Table 20.2 c o n t i n u e d =
o Yield
Blank
% T.Y. Yield1
Blank/other correction
(%o)
+6X1 +6180
+6X2 +6170
(%0)
(%0)
~o
Additional Notes
9
Reference
=r o r
r~
9 9 > v~ ~,.L~
Table 20.2 c o n t i n u e d >
4a to
4~ t,a 4a
> Table 20.2 continued
Sample Type
Typical Size
Reagent
Pretreatment Sample
Pretreatment Sample + vessel /line
MS Analyte
Rxn Temp. (~
Rxn. duration
Table 20.2 continued >
r= r t,~ o !
~q O
> Table 20.2 c o n t i n u e d o
Blank/other Yield
Blank
% T.Y.
Yield1
correction
(%0)
2.
+6X1 +634S
+6X2 +6170
(~o)
(~o)
Additional Notes
9
Reference
=r 9
r 9 9
v,e r
Table 20.2 continued
>
4a to
4a to
> Table 20.2 continued
Sample Type
Typical Size
Reagent
Pretreatment Sample
Pretreatment Sample + vessel /line
MS Analyte
Rxn Temp. (~
Rxn. duration
r
t,o O !
Table 20.2 continued > O
> Table 20.2 continued 9
Blank/other Yield
Blank
% T.Y. Yield1
correction
(~(,)
+6X1 +63osi
Additional Notes
Reference
o
(%0) 9
(/2 r
9 9
>
Table
202
continued
> 4~ to -..1
4~ to
> Table 20.2 continued
Sample Type
Typical Size
Reagent
Pretreatment Sample
Pretreatment Sample + vessel /line
MS Analyte
Rxn Temp. (~
Rxn. duration
Table 20.2 continued >
r
to O !
o
> Table 20.2 continued
Yield
Blank
9
% T.Y. Yield1
Blank/other correction (%o)
+6X1 +882Se (%0)
~o
Additional Notes
Reference
o 9 r 9 9
> m
1_ percent of theoretical yield mineral abbreviations: o1., olivine; qtz, quartz; fsp, feldspar; hem, hematite; mt, magnetite; g, garnet; ka, kaolinite; il, illite; mu, muscovite; bt,
2_
3_ 4_ 5_ 6_ 7_ 8_ 9_
biotite; phi, phlogopite; chl, chlorite; br, brucite; d, dickite; gb.gibbsite w.r. pwdrs.: whole-rock powders except for w.r. pwdrs, fsp, micas "tube": C-reduction in sealed silica tube following O'Neil et al. (1994) EA: high-temperature, carbon reduction by elemental analyzer and on-line irm-GCMS following Kornexl et al. (1999a) following method of Clayton and Mayeda (1963) apparatus and operating protocols similar to H.P. Taylor and Epstein (1962) R. Krouse, pers. commun., 2003
4~ tO
430
Chapter 20- B.E. Taylor
nitrogen temperatures (-197~ These may include, for example, SiF4 that can be cryogenically separated for the measurement of silicon isotopes, and as a monitor of the purity in the analysis of oxide minerals. The usual analyte for mass spectrometry is CO2. Complete conversion of the 02 is important in order to avoid isotopic fractionation owing to a kinetic isotope effect in the reaction of 02 with hot (500750~ graphite (Sakai & Honma, 1966; Figure 20.7). Taylor & Epstein (1962) ensured complete conversion in a step-wise fashion. The kinetic isotope effect accompanying this O2-1imited reaction is illustrated in Figure 20.7. This effect, longknown in conventional analysis (Taylor & Epstein, 1962), and, although reduced by the catalytic action of hot platinum, is the underlying cause of problems recently described in laser-assisted microanalysis (e.g., Mattey & Macpherson, 1993; Wiechert & Hoefs, 1995).
Figure 20.7- Plot of the oxygen isotope composition (logRo2 x 1000) vs. log f, where f = fraction of 02 remaining during its oxidation over hot graphite, and R = ( 1 8 0 / 1 6 0 ) O 2 (after Sakai & Honma 1966). This plot illustrates a kinetic isotope effect accompanying the oxygen-limited conversion of 02 to CO2 that mirrors a Rayleigh distillation process, becoming more marked as the 02 pressure, or fraction of 02 remaining (f) decreases. The isotope effect varies with size of 02/CO2 reactor, pressure of 02, size of 02 sample, and extent of reaction, and is of concern to both conventional and laser-aided micro-analytical processes that include this oxidation reaction.
Where 02 is chosen as the analyte, the 02 gas sample must be transferred (in the case of "off-line" preparations) to the mass spectrometer via a sample vessel. The sample is collected either by pumping (e.g., Toepler pump), or by adsorption on molecular sieve (e.g., 5~, 13X, etc.; see Karlsson, H., Chapter 36, this volume) at liquid nitrogen temperature. Silica gel is a good substitute for the molecular sieve, and avoids a potential source of contaminant 02 should F2 be present (e.g., Young et al., 1998a; Jones et al., 1999). For classical isotope ratio mass spectrometry, an absorbent-filled (mol. sieve or silica gel) micro volume is required for small samples. Liquid He, or a He cryostat, is otherwise required to cryogenically concentrate the 02 gas sample in an empty micro volume (e.g., R. Socki, pers, commun., 1998). Introduction of the 02 sample to the mass spectrometer in a carrier gas (e.g., He) via a split interface is a feasible alternative, and particularly well suited to small samples. Cryo-focusing, with or without combined gas chromatography, can be used to facilitate the carrier-gas introduction of very small samples (e.g., Young et al., 1998a; Wiechert, et al., 2002; Table 20.3).
Table 20.3 -Selected microscopic fluorination techniques o
Sample Type3 Powder/grains (P / G); in-situ (I)
Size Sample: (mg: P / G) (diameter, tim: I)
Laser type ( p w r ; ;~ in 12m)1
Mode Pulse (P) Continuous (CW)
Fluence
Reagent
9
~,,,i ~
(j/cm2)1o r~
S9
~,,,i ~
9 9
>
Table 20.3 continued >
4~
4a
> Table 20.3 continued
tO
Pretreatment Sample Sample and chamber (pre-fluorination)
MS Analyte
Yield
Blank
% T.Y. Yield1
Blank/Prep. correction (%o)
{' 3 =r to o !
Table 20.3 continued > O
> Table 20.3 continued o +8X1(~o)
+8~8o(%o)
+6X2(%o)
MS 7
+6170 (%o) IRMS/irm-GCMS
Additional Notes
Reference 9
=r o
r 9 9
>,
Table 20.3 continued
> #a L,o LaJ
> Table 20.3 continued
4a 4a
Sample Type
Powder / grains (P / G); in-situ (I)
Size Sample: (rag: P / G) (diameter,/am: I)
Laser type (pwr; ~. in tlm)l
Mode Pulse (P) Continuous (CW)
Fluence (J/cm2)10
Reagent
Table 20.3 continued >
r
=r
b,a O !
o
o
> Table 20.3 continued 9
Pretreatment Sample Sample and chamber (pre-fluorination)
MS Analyte
Yield
Blank
% T.Y. Yield1
Blank/Prep. correction (%0)
e~
o
9
r
o o r~
>
v~
Table 20.3 continued >
4~
> Table 20.3 continued
_+6X1(%o) 4-6180(%o)
+6X2(%o) MS 7 +6170 (%o) I R M S / i r m - G C M S
4a Additional Notes
Reference
Table 20.3 continued >
C3 r~ t,o O !
O
> Table 20.3 continued
Sample Type
P o w d e r / grains (P / G); in-situ (I)
o Size Sample: (rag" P / G) (diameter, tim: I)
Laser type (pwr; X in ~m)l
Mode Pulse (P) Continuous (CW)
Fluence (J / cm2) 10
Reagent
9 =r or
r o 9 >
Table 20.3 continued >
#a ...1
> Table 20.3 continued
Sample
Pretreatment Sample and chamber (pre-fluorination)
4~
MS Analyte
Yield
Blank
% T.Y. Yield1
Blank/Prep. correction (%0)
r r~ t,~ O !
Table 20.3 continued > o
>Table 20.3 continued +6X1(%o) q-&180(~o)
+6X2(%o) + 6 1 7 0 (%o)
=
9 3. MS 7
Additional Notes
Reference r
IRMS/irm-GCMS
o ~,,d.
o o
> v< ~,,Lo
Table 20.3 continued >
ga
4~ 4~ o
> Table 20.3 continued Sample Type
Powder / grains (P / G); in-situ (I)
Size Sample" (mg: P / G) (diameter, tom" I)
Laser type (pwr; ;~ in/xm)l
Mode Pulse (P) Continuous (CW)
Fluence (J / cm2) 10
Reagent
Table 20.3 continued > e3 to O !
O
>Table 20.3 continued +6X1(%o) _+6180(%o)
+6X2(%o) MS7 _+~34S(%0) IRMS/irm-GCMS
9
Additional Notes
Reference
~.
o S" o
r~ r 9 9 1_ pwr: power, in watts (W) 2_ Pwdr., w.r.: powder, whole-rock 3_ Mineral abbreviations: ol, olivine; g, garnet; mt, magnetite; ep, epidote; zr, zircon; tm, tourmaline; en, enstatite; sp, spinel; di, diopside; aug, augite; ky; ky; kyanite; mu, muscovitie; bt, biotite; pyx, pyroxene; p> pyrite; cp> chalcopyrite; gn, galena; spl, sphalerite; po, pyrrhotite; tr, troilite; acan, acanthite; grn, greenockite 4_ Absorbed on 5 ~ molecular sieve for transportation to mass spectrometer 5_ Indicated in Figure 20.4 of Farguhar & Rumble (1998) 6_ t: torr 7_ MS: IRMS, classical, dual-inlet isotope ratio mass spectrometry; irm-GCMS: isotope ratio monitoring-gas c h r o m a t o g r a p h y - m a s s spectrometry 8_ 0.1 atm = 10.13 kPa 9_ Ca. 100 m m in crater diameter 10_ j/cm2 on sample surface, but not reported by all (often laser power only is quoted, measured either above sample chamber or from exit port of laser) 11_ n.a., not available or not applicable
>
4~ 4x
442
Chapter 20 - B.E. T a y l o r
20.4.2 Fluorination of silicates and oxides by interhalogen fluorides Fluorination of silicates by interhalogen fluorides, principally BrF5 (Clayton & Meyeda, 1963) and C1F3 (Baertschi & Silverman, 1951; Borthwick & Harmon, 1982), can be described by the following example reactions (Table 20.2): KA1Si308(s)+ 8BrF5 ~ KF(s) + A1F3(s) + 3SiF4 + 402 + 8BrF3 3Mg2SiO4(s) + 8C1F3 ~ 6MgF2(s) + 3SiF4 + 4C12 + 602
[20.91 [20.10]
Similar to fluorination by F2, metal fluorides (some of which are volatile) can also comprise reaction products, in addition to oxygen (and, under certain conditions, new oxides; see below). In particular, fluorination by interhalogen fluorides produces BrF3, Br, or C1 and their fluorides as additional reaction products in the reaction tube(s). These are readily condensable, facilitating separation from the 02 gas in the reaction tube, but also add to the corrosion of the reaction tubes, or form hygroscopic deposits. Neutralization of excess of BrF5 (or C1F3) takes places after all oxygen has been removed from the reaction vessels. Consequently, less time is required for the extraction, collection, and conversion of oxygen from each sample when using BrF5 (or C1F3; e.g., often ca. _~10 minutes/sample) as compared with the F2 fluorination procedure. Moreover, the ability to condense excess BrF5, reaction-related waste products, and the CO2 converted from the released 02, permits cryogenic purification of the gas sample. Duration of the reaction is often determined by routine laboratory work schedules, and, accordingly, reaction times are typically on the order of 16 hours (i.e., overnight). This is a common time span between the end of the workday, when the reaction tubes have been loaded and furnaces turned "on", and the morning of the next day when the extractions are performed, and is often more a matter of convenience. Vennemann & Smith (1990) showed, for example, that fluorination with C1F3 at temperatures 450~ can be significantly shorter (e.g., 2hrs) for quartz, feldspar, biotite and pyroxene, and require only 4-8hrs at 600~ Nevertheless, from personal experience, it seems best (for a number of reasons related primarily to the reactivity of the reaction products) to keep the reaction vessels at their reaction temperatures until preparing the vacuum line for oxygen collection. Some minerals, as a reflection of their bonding and crystallographic structure, are especially resistant to fluorination, and their oxygen yields, which are not always 100%, depend on the fluorination reagent, and reaction time and temperature (Taylor & Epstein, 1962; Kyser et al., 1981; Vennemann & Smith, 1990). These include silicates such as high Mg-olivine, pyroxenes, garnets, and aluminosilicates, and oxides such as magnetite and corundum. For these minerals, reaction is promoted (but, not necessarily to 100%) by very fine grain size (few micrometers), longer reaction times, and higher temperatures (close to, but less than 700~ at which Ni begins to oxidize). In some cases (e.g., magnetite) higher oxygen yields are obtained when using BrF5 instead of F2. Unlike reactions with most minerals, however, a correct isotopic analysis of Mg-rich olivine can be obtained even with only an 80% theoretical yield of oxy-
Fluorination Methodsin StableIsotopeAnalysis
443
gen. In some cases (e.g., garnet) analysis is greatly facilitated by preparing a glass from the pure mineral (e.g., Taylor & O'Neil, 1977). Reuter et al. (1965; and also Garlick & Epstein, 1967) achieved 100% yields and accurate isotopic analyses from olivine by fusing known amounts of olivine and quartz (with a known isotopic composition) to form a pyroxene glass. Reuter et al. (1965) employed yet other methods, such as heating the sample to 550~ in a stream of H2, to remove oxygen and adsorbed H20 from slightly 'weathered' meteorites. Iron metal and sulfides were removed by reaction with dry C1 gas.
20.4.2.1 Yield and accuracy of ~180 values Garlick & Epstein (1967) found that the measured values of 6180 of quartz were lower when incomplete fluorination was due to insufficient reagent (BrF5). Although a 100% reaction yield from the sample as loaded is traditionally used as a criterion of the reliability of the reaction, there are exceptions, as noted above. This suggests that fluorination reactions do not all proceed in the same way. In the case of silicates (and quartz, in particular), partial, isotopically selective replacement of oxygen by fluorine in the SiO2 tetrahedral structure may be at work. Or, as suggested by Kohn et al. (1998) for micro-analysis of phosphate, some variability of 6180 may be due to formation of P-O-F compounds in cases when yields are low (e.g., 80%). In this instance, a correction must be applied. If reaction temperatures are allowed to exceed ca. 700~ NiO may also form in the reaction tube, reducing the yield, and (by virtue of the gasoxide fractionation) increasing the 6180 value of the residual 02. The formation of NiO has been suspected also at lower temperatures (S. Sheppard, pers. commun., 2003), with similar affects. 20.4.3 Additional fluorination procedures and comments 20.4.3.1 Step-wise fluorination, or "fluorination stripping" Partial fluorination reactions can be used to document isotopic inhomogeneity in minerals, whether of an intra-grain, or intra-crystalline nature, and to selectively analyze silicate-bonded oxygen. Step-wise fluorination, or fluorination "stripping", in a conventional fluorination line using F2, is accomplished by interrupting the fluorination reaction prior to completion. This technique has been used to detect, among others, grain-scale isotopic inhomogeneties and to infer isotopic zoning in quartz phenocrysts from hydrothermally altered rocks (Magaritz & Taylor, 1976).
Partial fluorination is more problematic with BrF5. Too little reagent, or too low a temperature, leads to 6180 values that are, respectively, too high or too low (Garlick & Epstein, 1967). These affects are apparently not so pronounced when F2 is used according to Epstein & Taylor (1971). However, Hamza & Epstein (1980) found that partial fluorination of kaolinite by F2-1imitation at 180~ yielded increasing 6180 values with extent of reaction. Using BrF5, Javoy et al. (1973) and Javoy & Fourcade (1973) found that initial 6180 values from partial extraction of oxygen from lunar fines were low, and interpreted this as due to contamination by adsorbed, low-180 atmospheric water. Similar results from a Hawaiian basaltic glass (control sample) led Javoy & Fourcade (1973) to suggest that partial fluorination could be accomplished
444
Chapter 20 - B.E. T a y l o r
using BrF5 without isotopic fractionation. Outgassing of the sample in the reaction tube, prior to fluorination, has been used (by Javoy et al. (1973), and others) in an attempt to rid the sample of adsorbed water. Inasmuch as nickel fluoride, which coats much of the inner wall of the reaction tube, and bromine, which may deposit at the cool end of the reaction tube and/or at the Swagelock ferrule on the top of the 'chimney' in the cap shown in Figures 20.3 and 20.4, are very hygroscopic, sample degassing in the reaction tube should be used with caution. Periodic checking and cleaning of the reaction tubes and caps is recommended. Reaction may proceed at any surface accessible during fluorination, including cracks, grain boundaries, crystal faces, as well as the grain surfaces, and, therefore, extreme caution should be used in interpreting the origin and isotopic composition of step-wise released oxygen. For example, the assumption of simple rim-to-core reaction (e.g., Epstein & Taylor, 1971; Magaritz & Taylor, 1976) was brought into question by in situ, high-spatial resolution isotopic analyses of quartz phenocrysts from Isle of Skye intrusions by Ion Probe, or Secondary Ion Mass Spectrometer (Valley & Graham, 1996) that demonstrated isotopic inhomogeneties caused by hydrothermal alteration followed inter-grain fractures. Where possible, knowledge of the textural context of any micro-scale isotopic variations is preferred, and, indeed, necessary to a correct interpretation.
20.4 3.2 Selective analysis by partial fluorination
Partial fluorination, which exploits the differences in reactivity, or reaction rate of various minerals, has been suggested as a useful technique for selective analysis of mineral mixtures. For example, Clayton & Mayeda (1963) noted that a one-hour reaction at 100~ between BrF5 and a difficult-to-separate quartz-feldspar mixture was sufficient to preferentially release the feldspar oxygen. Pumping away this oxygen while retaining the remaining BrF5 by cooling the reaction tube with LN2, allowed for reaction of the quartz in the usual way. The general procedure of separation by partial reaction seems to work best on simple mixtures of components with markedly different reactivity (e.g., analogous to procedures for mixed carbonates: A1-Aasm et al., 1990), but, again, interpretation must proceed with caution. Urey (1947) suggested, on the basis of a thermodynamic framework, that the oxygen isotope ratio of chemically and structurally different sites in minerals should differ. This raised the exciting possibility that intra-mineral isotopic fractionations could provide the basis for single mineral geothermometry. Several authors have investigated this potential (e.g., hydrous silicates: Savin, 1967; Hamza & Epstein, 1980; Bechtel & Hoernes, 1990; Girard & Savin, 1996; alunite: Pickthorn & O'Neil, 1985). Partial fluorination of pure minerals for such studies requires the use of F2. Savin (1967) and Hamza & Epstein (1980) attempted partial fluorination of hydrous minerals using F2 as a means for site-selective analysis in order to determine whether the OH-framework oxygen isotope fractionation could be measured. For kaolinite, Hamza & Epstein's (1980) procedure entailed an initial low-temperature (0~ reaction to extract OH-oxygen, followed by a high-temperature (100-450~
Fluorination Methodsin Stable IsotopeAnalysis
445
reaction to yield the remaining framework oxygen. Similar, two-step extractions were used for biotite, chlorite, muscovite, and phlogopite, differing only in the temperature of the second step (600~ in the case of phlogopite). Monitoring of recovered SiF4 assured that no framework oxygen was released during the first, 0~ reaction step. Intra-mineral oxygen isotope fractionations between OH- and framework oxygen were estimated to range from 5.2 (muscovite) to 12.6 (kaolinite), illustrating the expected increase with decreasing formation temperature and suggesting perhaps a fruitful field for further investigation (Hamza & Epstein, 1980). Bechtel & Hoernes (1985; 1990) devised a procedure for Fe-free minerals involving dehydroxylation and partial fluorination. Reduction of water released during vacuum dehydroxylation obviates this method for Fe-bearing minerals (Bechtel & Hoernes, 1990). Their methodology yielded an improved reproducibility over earlier results based on partial fluorination alone. Comparison of results from dehydroxylation and fluorination of the dehydrated residue and the original mineral indicated a consistent isotopic mass balance (and, therefore, complete analysis). Thus, using illite as single mineral geothermometer, one need only measure the 6180 values of the whole mineral and of the dehydrated residue. Internal fractionations of ca. 10.0 to 15.0%o were found for test samples. These fractionations correspond with temperatures of 200300~ (according to illite crystallinity studies), and are consistent with fractionations calculated from Sch~tze's increment method (Richter & Hoernes, 1988). Subsequently, Girard & Savin (1996) supported the results of Bechtel & Hoernes (1990) showing that the thermal dehydroxylation step provides more consistent results than partial fluorination for analysis of OH-oxygen. For reasons unclear, Girard and Savin (1996) were less successful in their attempt to determine and intra-mineral oxygen isotope fractionation, suggesting that factors such as crystal size, temperature of formation, etc. may be important to consider. In any case, caution in the application of partial fluorination and interpretation the results is advised. Acquisition of accurate and consistent oxygen isotope analyses of hydrated silica by step-wise extraction was used by Haimson & Knauth (1983), who demonstrated that step-wise fluorination successfully removed both adsorbed and absorbed water from hydrated silica. Yields of oxygen are, up to a fluorine excess of ca 2, linearly dependent on the abundance of fluorine present (Haimson & Knauth, 1983; Figure 20.6). Knowledge of the isotopic composition of different oxygen sites in single minerals can yield information on provenance or genesis. Supergene alunite (formed during weathering) can be distinguished from hypogene alunite (hydrothermal or magmatic) (e.g., Rye, 1993). As is the case with alunite, however, techniques other than partial fluorination are sometimes more suitable to site-selective, or mineral-specific analysis. For example, a simple, feldspar can be removed at room temperature from quartzfeldspar mixtures by pre-treatment with HF or fluorosilicic acid in Teflon beakers, followed by careful washing and separation of the residual, largely unreacted quartz. Although more time-consuming than the previously mentioned partial fluorination procedure of Clayton & Mayeda (1963), the accuracy of isotopic results from pure
446
Chapter 20- B.E. Taylor
minerals is assured.
20.4.4 Fluorination of phosphates and sulfides by halogen fluorides and F2 gas Tudge (1960) originally described the fluorination of biogenic phosphate (pre-purifled as BiPO4) for oxygen in PO42- using BrF3. The reaction: BiPO4(s) + 8 / 3BrF3 ~ BiF3(s) + 4 / 3Br2 + 202
[20.11]
was quantitative and took place for 30 minutes at ca. 100~ in an externally heated Teflon reaction tube. BrF3 is liquid below ca. 125~ and was considered the fluorinating reagent of choice because the reaction could be carried out at low temperature. Until BiPO4, was replaced by Ag3PO4 (e.g., O'Neil et al., 1994; see also Firsching, 1961) as a more stable compound, reaction [20.11] represented the classic technique. More commonly, however, fluorination was accomplished using BrF5, C1F3, or F2 (Longinelli & Nutti, 1973a; Vennemann et al., 2002). Fluorination of BiPO4 (and presumably using Ag3PO4) by BrF5 carried out at high temperature (e.g., 500~ 9Vennemann et al., 2002) yields no systematic differences from results obtained using BrF3 (Longinelli & Nutti, 1973a). The reaction likely proceeds as follows" Ag3PO4(s) + 4BrF5 ~ 3AgF2(s) + PF5 + 3BrF3 + 1/2Br2 + 202
[20.12]
indicating the need for a molar fluorine/oxygen ratio of at least 5 for complete reaction. Fluorination of abiogenic apatite [Ca5(PO4)3(OH, F,C1)] by BrF5 proceeds via a reaction analogous to [20.12], but requires high temperatures (e.g., 650~ to give complete oxygen yields (Rhodes & Oreskes, 1999; B. Taylor & Mirnejad, unpub.). Conversion of sulfide minerals and elemental sulfur to SF6 permits the most accurate determination of 634S, in addition to 633S and ~36S. Unlike SO2, which is polar and readily absorbed on surfaces, SF6 is non-polar and chemically inert. Consequently, its memory in the ion source, though finite, is minimal in comparison to that of SO2. Differences in absorption and pumping characteristics are largely to blame for discrepancies in analytical results using one or the other gas (e.g., Taylor et al., 2001a, b; Beaudoin & Taylor, 1994; Rees, 1978). Moreover, fluorine is mono-isotopic, unlike oxygen in SO2, which obviates any corrections for isobaric interferences in the ion spectrum. Hence, the four SF5+ ion currents measured as masses 127, 128, 129, and 131, are attributed to the stable sulfur isotopes 32, 33, 34, and 36, respectively. Puchelt et al. (1971) described the preparation of SF6 by conventional fluorination of sulfur and sulfide minerals using BrF3, following the reaction: 2FeS2(s) + 10BrF3 ~ 4SF6 + 2FeF3(s) + 5Br2
[20.13]
carried out in a nickel reaction tube, externally heated to 200~ overnight. Repeated
Fluorination Methods in Stable Isotope Analysis
447
cryogenic distillation pre-purified the SF6 prior to final purification on a 5A mol sievepacked 5ft x 0.25in column. The measured 634S values were reproducible to 0.1%o. Whereas, elemental sulfur, greenockite (CdS), sphalerite (ZnS), galena (PbS), cinnabar (HgS), covellite (CuS), argentite (Ag2S), and pyrite (FeS2) reacted readily, giving yields of 96-98%, low yields were obtained from pyrrhotite (FeS), troilite (FeS), and digenite (Cu9S5). The latter minerals were converted to CdS for analysis. Thode & Rees (1971), Hoering (1990), and Gao & Thiemens (1991) have reported fluorination of sulfur compounds with BrF5. Gao & Thiemens (1991) used a conventional fluorination line, similar in many aspects to that shown in Figure 20.1, for overnight (16h) fluorination of sulfide minerals (previously converted to CdS) at 450~ with a 150X stoichiometric excess of BrF5. Thode & Rees (1971), on the other hand, found that a 300~ 16 h reaction with a 20X molar excess of BrF5 sufficed to quantitatively convert 3-7mg aliquots of Ag2S to SF6. In either case, accurate measurement of ~36S following cryogenic distillation requires further purification of the SF6 by gas chromatography. This process is time-intensive. For example, GC purification described by Gao & Thiemens (1991) required as much as an hour or more per sample, plus a 5h bake-out (100~ of the GC column (80-100 mesh Porpak Q in a 12ft, 1/8 in OD column) between samples. Hulston & Thode (1965b) prepared SF6 by fluorination of Ag2S (prepared from each sample) with F2 gas because it was found that fluorination with BrF3 by reaction [20.13] resulted in yields of only 75%, albeit without apparent isotopic fractionation. Although not noted specifically, yields were presumably improved with the use of F2 gas. The SF6 was purified on a 5~, mol sieve-packed column held in a temperature gradient (150~ to room temperature). Contaminants in the column were removed after ca. 6 samples by an overnight bake-out at 220~ Largely because of the time-consuming nature of sulfide fluorination in this manner, fluorination of sulfides did not gain wide use, in spite of the mass spectrometric advantages of SF6 (Puchelt et al., 1971; Rees, 1978) and the ability to measure 633S and 636S. Simpler, but less accurate techniques that produced SO2 for measurement of 634S have been more widely used (Taylor et al., 2001a,b). Unlike SO2-based laser methods, fluorination of sulfides to form SF6 occurs without isotopic fractionation, and, consequently, no sample-dependent correction factors are needed (Beaudoin & Taylor, 1994). The advent of laser-assisted fluorination (e.g., Sharp, 1990), combined with capability to safely generate pure F2 in the laboratory (Aspre~ 1976), has begun to re-vitalize fluorination of sulfides for isotopic analysis (e.g., Rumble et al., 1993; Beaudoin & Taylor, 1993; Taylor & Beaudoin, 1993; Beaudoin & Taylor, 1994). Fluorination by F2 may follow a simple reaction such as" FeS2(s) + 7.5F2 ~ FeF3(s) + 2SF6
[20.14]
448
Chapter 20- B.E. Taylor
that yields no halogen waste products such as form in reaction [20.13]. In practice, other fluorine-bearing compounds may form (e.g., HF, CF4, and complex fluorocarbons) by reaction with contaminants in the sample chamber or the sample. It is the fluorocarbons, which pose the greatest problem in sample purification, but these can be generally separated in a timely manner. Considerable purification, sufficient for precise and accurate measurement of 633S and 634S, can be accomplished without gas chromatography by use of a variable temperature trap (Taylor & Beaudoin, 1993; Beaudoin & Taylor, 1994; Taylor & Beaudoin, 2000; Coleman, M. Part 2, Chapter 44, this volume; see Figure 20.8). A variable temperature trap offers greater accuracy and precision in cryogenic separation than is possible using traditional, fixed-point cooling mixtures (e.g., dry ice + acetone), and fewer potential health and safety hazards. Sulfur isotope anomalies have previously been associated with cosmogenic studies (e.g., Hulston & Thode, 1965b; Farquhar & Thiemens, 2000). However, the recent discovery of terrestrial 633S anomalies and their significance for evolution of the Earth's early atmosphere (e.g., Farquhar et al., 2000a) have re-generated interest in sulfide fluorination. Such anomalies can only be detected by isotope ratio measurements made using SF6. Also, the increased the accuracy of the SF6 method (Taylor, et al., 2001, in press), will likely lead to wider spread use of fluorination procedures in sulfur isotope analysis. 20.4.5 Fluorination of selenium and selenides for selenium isotopes
Fluorination of elemental selenium was the original method of choice used to measure the isotopic ratio of 82Se/76Se in various types of samples, from native selenium and from selenides, to rare selenite and selenates, and including trace concentrations of selenium in sulphides, organic materials, and in solution (Krouse & Thode, 1962). Figure 20.8 - (A) High-surface area P-trap ensures efficient cryogenic trapping of even trace amounts of a condensable gas, whether from a high-throughput He-carrier stream or, for example, during evacuation of a sample-chamber filling F2 atmosphere. High surface area is provided by the 60mm filter element (Swagelock| SS-6TF-60) used as a "foot" in the P-trap (see Figure 20.8A). The P-trap design is achieved by arc-welding a 1 / 4" tube to 3/8" OD sleeve, providing for a vacuum fit inside bored-through 3/8" stainless "tee"; (B) Variable temperature trap (VTT; modified after design of DesMarais, 1978b), provides for cryogenic trapping and separation with 1~ accuracy. Temperature of the VTT may be varied from ambient to nearly -197~ by combined heating of Ni-chrome heating wire, with variable transformer connected to low voltage (e.g., 24 V.A.C.) transformer via power feed-through, and simultaneous submersion of VTT in liquid nitrogen. VTT may be dismantled for repair at indium-wire sealed flange if ever necessary. Internal Cajon flexible metal hose, with factory-supplied 1/4" tube ends, is connected to the flange via welded Swagelock @ components and to the vacuum line shown in Figure 9 via 1/4" stainless steel Swagelock | run-tee-components having female pipe thread on one arm. A thermocouple inserted inside flexible hose, and vacuum-sealed in a reducing fitting with Teflon ferrules, provides for monitoring of temperature (slight differences in placement of thermocouple tip may result in slight variation in indicated temperature). One way, 10psirated poppet check valve (e.g., Nupro | B-4C-10) provides for safe release of condensed oxygen should outer envelope of VTT leak while submersed in liquid nitrogen. Modified after Taylor (2003) and Taylor & Beaudoin (1993).
Fluorination Methods in Stable Isotope Analysis
449
450
Chapter 20- B.E. Taylor
Krouse & Thode (1962) extracted selenium from a range of sample types using either hydroxylamine hydrochloride for solutions, or a hydrobromic acid-bromine extraction procedure (referenced in their paper) for whole-rock samples and minerals. This extraction procedure provides a way to measure selenium isotope compositions in sulfide and sulfosalt minerals in which selenium may commonly occur as a substitution for sulfur. The fluorination of selenium by F2(g) evidently produces fewer contaminants than when carried out using CoF3(s) (c.f., Krouse & Thode, 1962; Webster & Warren, 1981), yet the uncertainties in measured ~82Se values in these studies was similar (ca. 0.5 %0). Liquid air and liquid oxygen cryogenic traps were particularly useful in removing a fluorocarbon contaminant of some type that caused a mass interference (R. Krouse, pers. commun., 2003). Where F2 is the fluorinating reagent, the reaction is relatively simple, and could be ideally written ignoring products of incomplete fluorination, such as Se2F2 and SeF4, and contaminants (e.g., CF4) as: Se(s) + 3F2 = SeF6
[20.15]
Fractionations among natural samples up to 15%o were discovered using the fluorination process, with the largest isotopic deviations found in organic compounds, presumably due to kinetic isotope effects during selenate reduction (Krouse & Thode, 1962). Krouse & Thode (1962) noted that although not as large in magnitude as known for sulfur isotopes, the association of larger fractionations with biologic processes in was, indeed, analogous to the behavior of sulfur isotopes. In the case of selenium isotopes, preparation of samples via the fluorination technique will likely give way to instrumental methods. Thermal ionization mass spectrometry (TIMS) employing negative ions, or N-TIMS, was shown by Wachsmann & Heumann (1992) to permit determination of the isotopic ratio of 80Se/76Se to a precision of 1-3%o (i.e., poorer than by fluorination) for g80Se. A double-spike technique devised by Johnson et al. (1999) currently yields a precision of 0.2%o. The N-TIMS technique requires the concentration and purification of selenium from the sample, but facilitates isotopic measurements on samples with considerably lower Se contents than previously achieved. Multi-collector ICP-MS (inductively coupled plasma mass spectrometry; see Rehk/imper, M., et al., Chapter 31, this volume) provides yet another step in the direction of permissible lower Se concentrations, and has recently been used to measure 82Se/76Se ratios in a variety of rock types, soil, sediment, and naturally-deposited silica (Rouxel et al., 2002). Although this technique has its own associated potential errors of mass interference and contamination, and requires a pre-analysis acid sample digestion, the precision achieved thus far of ca. 0.25%0 (2~J) for ~82Se surpasses the original results by fluorination with far greater ease and economy of time (Rouxel et al., 2002). Most importantly, the multi-collector ICP-MS instrumentation can determine Se isotopic compositions in samples containing nanogram quantities (gg/g) of Se, with little, if any sample matrix related problems. The multi-collector ICP-MS approach is certain
Fluorination Methods in Stable Isotope Analysis
451
to comprise a principal research tool. The N-TIMS technique is viable, as well, but the current reporting of selenium isotope compositions by N-TIMS as 680Se values rather than traditional 682Se values complicates comparison of results obtained by other techniques. 20.4.6 Fluorination of silicates for silicon isotopes SiF4 is produced during fluorination of quartz, for example, using BrF5 as follows:
SiO2(s) + 2BrF5 = SiF4 + 2BrF3 + 02
[20.16]
The product SiF4 may be cryogenically separated and purified at-80~ (e.g., using dry ice+acetone), or, ca.-70~ or colder, using liquid + frozen ethanol (ethanol-liquid nitrogen mixtures) monitored with a thermocouple. Any trace amounts of active fluorine compounds produced by reaction [20.16] may be removed by reaction with Zn granules at 50-60~ in a Cu tube to form condensable ZnF2 and ZnBr2 in three distillation stages (Ding et al., 1996). These include two successive stages of trapping and vacuum pumping at ca.-80~ followed by reaction with Zn, the trapping of the purified SiF4 at-197~ (liquid nitrogen temperature) and removal, by vacuum pumping, of any non-condensable gases released during the hot Zn treatment. This step constitutes an improvement over previous methods using F2 (Epstein & Taylor, 1971), or BrF5 (e.g., Clayton & Mayeda, 1978; Douthitt, 1982). When silicates or purified silica are fluorinated via a reaction analogous to reaction [20.16] using F2, reaction products do not, of course, include Br and Br fluoride compounds. Rather, these form in the waste line during neutralization of the F2. Ding et al. (1996) note, in particular that purification of the BrF5 reagent by distillation at-70~ may be necessary to attain high precision. 20.4.7 Fluorination of water by halogen fluoride The fluorination method described by O'Neil & Epstein (1966b) was designed to analyze but a few milligrams of water. Although miniaturization of the CO2-H20 equilibration technique (e.g., Kishima & Sakai, 1980; Ohba, 1987; Socki et al., 1999) facilitates indirect analysis of several milligrams of water, the uncertainty in measured 6180 values is, except for the study by Socki et al. (1999) utilizing 10rag of water, larger than by fluorination. Usually at least 1 ml of water is required for analysis by conventional CO2/H20 exchange procedures. Fluorination of water with a 3- to 4fold excess of BrF5 takes place rapidly at 80-100~ according to the reaction:
BrF5 + H20(1) ~ Br + 3/2F2 + 2HF + 1/202
[20.17]
with a yield of 100% (O'Neil & Epstein, 1966b). Although the possibility of fluorination by F2 was acknowledged (O'Neil & Epstein, 1966b), a method for water analysis by this means is not published. Reaction [20.17] proceeds rapidly. Loading of sample and reagent, plus evacuation, occupy the bulk of the analytical time. The principal difficulty is the transfer of water (in this case, small quantities), in a vacuum line containing abundant hygroscopic fluorides on the walls of the reaction vessel. Both of the two methods, capillary and direct cryogenic transfer devised by O'Neil & Epstein (1966b)
452
Chapter 20- B.E. Taylor
avoided this potential problem and yielded similar results. This method not only offers direct analysis with excellent precision (<_0.1%o in 1966!) of ca. 2mg of water, but provides a bulk oxygen analysis. Isotopic analysis of non-neutral, non-dilute waters (e.g., brines) by the equilibration technique can be difficult and fraught with additional uncertainty. Fluorination would seem a better choice, provided that care is taken to ensure 100% efficient loading (transfer) of the sample by guarding against precipitation of hydrated residues in the vacuum line, for example. Except for the analyses of fluid inclusion waters by Rye & O'Neil (1968), oxygen isotope analysis of waters by fluorination has not received much attention. 20.4.8 M e a s u r e m e n t and correction of an oxygen blank
Potential contribution of oxygen from sources other than the sample needs to be quantitatively assessed. Sources may include the reagent, residual sample, or leaks. The total blank (i.e., from all such sources) can be measured by performing a set of extractions, carried out in a way identical to the normal extraction procedure, but without samples. A repeat of this procedure will identify whether a contribution of oxygen comes from the sample tube (e.g., from reaction of the residuum of the previously loaded sample). If a residuum is present, then the size of the blank should be reduced with each set of "blank" extractions. A relatively constant "blank" size would rather suggest a reagent source a n d / o r a leak (when the leak rate and length of exposure to the leak are relatively constant). A leak can be discerned by isolating and vacuum-monitoring selected segments of the extraction line, and by employing the usual leak-checking procedures, keeping in mind that leaks could be present in the line a n d / o r sample tube. A first order assessment of leaks in the vacuum line and/or across the seats of reaction tube valves can be made by collection and analysis of any gas that may be present in the vacuum line after an over-night reaction. Blanks due to memory effects, such as the exchange of oxygen on the surface of silver solder if used in a vacuum system, may also interfere with accuracy and precision. The affect of such blanks on the measurements can sometimes be corrected, but blanks not related to the reagent are often not consistent and these corrections are difficult to make. It is best to discover and eliminate the source(s) of such blanks if possible. Lee et al. (1980) miniaturized a BrF5 fluorination line (reaction tubes ca. 1/4 of standard size) in order to analyze small (ca. lmg) samples with a reduced blank. They found, that the total 0.6 gmole blank 02 comprised at least two components, one if which was about 0.03 gmole that formed even in the absence of reagent. Similar blank sizes are typical of normal routines in at least some standard-size conventional lines (e.g., <0.5 gmole; Geological Survey of Canada, Ottawa). Management (correction) of reagent-related blanks is discussed below. The ability to cryogenically distill an aliquot of BrF5 or C1F3 allows for the removal of any dissolved oxygen, and virtual elimination of reagent derived blank. Moreover, this distillation step can also be utilized when loading an aliquot of reagent into each sample tube. Purification of commercial fluorine using an Asprey-type fluorine pump virtually eliminates any oxygen contribution from the reagent. Reagent-derived oxy-
Fluorination Methods in Stable Isotope Analysis
453
gen should be an issue, then, only with commercial fluorine. The quantity and isotopic composition of any oxygen added along with commercial F2 reagent to each sample tube can be measured when the line is known to be leak-free (the desired condition in general !), and the sample tubes are free of any residual sample. Once determined, the contribution of this blank, and its isotopic effect, to each sample can be accounted for as a routine correction (see Taylor & Epstein, 1962). Lee et al. (1980) ascribed a standard size and isotopic composition to the blank in their BrF5 system based on analyses of a quartz standard, somewhat analogous, in effect, to the normalization procedure utilized by Taylor & Epstein (1962). Lee et al. (1980) also found, however, that the isotopic composition of the actual blank was not the same as that assumed from analysis of the quartz standard. Unavoidably, the uncertainty of the 6value of each analysis is raised slightly as a result of such a correction because of slight variations in aliquot size, etc. 20.5 Conventional fluorination apparatus
Traditionally, pure minerals and whole-rock samples, have been analyzed in ca. 540 mg. aliquots by fluorination in what is nowadays referred to as a 'conventional' vacuum fluorination line (Figure 20.1) primarily owing to the fact that inlet systems to early mass spectrometers, and the sensitivity of these first-generation instruments, required larger volumes of gas than are commonly analyzed today. Accordingly, the scale of analysis (e.g., a hand-sample) was sufficiently large that averaging of sometimes genetically significant small-scale isotopic variations necessarily occurred. Mechanical micro sampling and step-wise fluorination (stripping) have been used, as previously discussed, to reduce somewhat the scale of analysis. In spite of the increase in number and variety of laser-based micro-analytical apparatus, and the new capabilities they offer for high spatial resolution sampling and analysis, conventional extraction lines are unlikely to be replaced in the very near future, owing to the need, at the least, for analyses of hydrous whole-rock samples. Both the presence of structural water and the need for an 'average' composition preclude microanalysis. In short, conventional preparation lines complement laser-assisted extraction. 20.5.1 Construction
A number of components are common to both halogen-fluoride compound-based and F2-based apparatus. These components are discussed below in the context of one or the other types of fluorination lines. The discussion draws largely on our own experience, complemented by that of many others, acquired over a number of years. Because fluorination apparatus are not commercially available, vacuum fluorination lines in various laboratories may differ slightly in detail, reflecting the experience, logistical constraints, and individuality of the builder.
20.5.1.1 F2-based apparatus The fluorination reaction [20.6], driven by heat, is extremely aggressive and, hence, requires a special apparatus (e.g., Figure 20.1) constructed in large part of nickel (reaction tubes) and stainless steel (extraction line, tubing, valves and connections). Some workers have also used Monel. A glass section to the vacuum line is permissible where virtually the only gas in contact with the glass tubing is the extracted oxygen
454
Chapter 20 - B.E. Taylor
and CO2. The nickel reaction tubes are externally heated to a temperature (e.g., 450690~ and for a duration (e.g., 12-18hrs) depending on the mineral/substance analyzed. Nickel is the preferred metal for construction of the reaction tubes primarily because of its relative resistance to attack by fluorine or interhalogen fluorides (Myers & DeLong, 1984; Table II in Seboldt et al., 1993). In practice, an inner layer of nickel fluoride develops which armors the tube, slowing further attack. Moreover, nickel does not oxidize readily below 700~ and thus poses no threat to the isotopic integrity of sample 02 produced. The tubes are typically custom made from 3/4in. (ca. 19mm) outer diameter nickel rod (>95% purity is desired here), with a length of ca. 12in. The tubes are bored to a 0.5in. (12.7mm) inner diameter. Shorter or narrower vessels are also possible. In any case, a round inner bottom promotes longer life, especially compared with tubes having welded bottoms, and honed and polished inner surface helps minimize the surface area and eliminate potentially more reactive sites to fluorine attack. Furthermore, a polished inner surface is thought to reduce variability and increase reliability of isotopic results. Tubes made with welded bottoms, although initially easier and less expensive to manufacture, provide an active site for fluorine attack and have a markedly shorter life. The square, welded bottoms inside are also more difficult to clean. Re-honing of the bored, round bottom tubes can be carried out from time to time as required. Nickel tubes maintained in this way have served well under heavy use for more than a decade at the G.S.C. (Ottawa). Connection of the nickel tubes to the vacuum line can be accomplished in one of several ways, but perhaps the most convenient is via a Teflon seal. Swagelock| components are particularly handy in this regard (e.g., Figures 20.3 and 20.4). Teflon ferrules are advantageous because they do not swage the neck of the tubes, and thereby ensure a long life. The neck of the reaction tube, i.e. that portion in contact with the vacuum fittings, should be free of scratches and highly polished to ensure vacuum integrity. A cooling coil, consisting of 2-3 turns of 3/16" (ca. 4.8mm) O.D. copper tubing in tight, mechanical contact with the tube (spring tension of the coil itself is usually sufficient, but a band clamp can be used where necessary; both avoid silver solder) is required to protect the Teflon ferrules during both heating and liquid-nitrogen cooling of the tubes. Circulated, 15~ water is usually adequate for this purpose. At the G.S.C. (Ottawa), a thermostatted circulator (15~ is used, thereby avoiding seasonal variations of water temperature that can cause analytical errors (occasionally referred to as the "Chicago effect" by those who directly or indirectly benefited from the scientific tutelage of Prof. R. N. Clayton and careful guidance of T. Mayeda at the U. of Chicago). If the circulated water is too cold, (e.g., as when tap water is used in the late Fall to early Spring), BrF5 and Br other products condense at the cooled end of the reaction tube. This can cause inaccurate (e.g., low) 6180, due to incomplete reaction or to some other isotopic fractionation effect. Stainless steel seamless tubing, type 304 or 316, usually 3/8" or 9mm O.D. is used for general purposes, and 1/4" or 6mm for short runs. Earlier lines (e.g., Tudge, 1960; Taylor & Epstein, 1962; Clayton & Mayeda, 1963; Longinelli, 1965) were often con-
Fluorination Methods in Stable Isotope Analysis
455
structed with silver-soldered joints. During inter-lab comparison of phosphate analyses, a marked discrepancy in 6180 values (3.5%0) for the same sample was discovered when comparing results from two labs that used differently constructed fluorination lines (Longinelli & Nuti, 1973a). The cause of the discrepancy was the silver solder which evidently caused some sort of oxygen isotope memory or isotopic exchange process (due to absorption or reaction with atmospheric oxygen) was suspected. A full explanation was not found, however, due in part to the lack of any discrepancy in oxygen yields. Yet, any variation in oxygen yield associated with a large isotope effect might have gone undetected with the Hg manometers typically used at the time of this discovery. Nowadays, the preferred construction method, where the entire line is not welded, is to connect the metal tubing using Swagelock| or similar components, provided care is taken to first polish tube ends to remove pits and longitudinal scratches, and that a small notch is cut in the end of the tube to ensure against any dead volumes or "virtual leaks". Though designed for use as pressure fittings, Swagelock| components make very serviceable vacuum fittings when properly used. Use of mini flanges or gaskets (e.g., Cajon VCR| fittings) can be used where zero-clearance disconnection/ removal is occasionally desired. In this case, however, inert gas arc welding of the tubing and VCR component is required.
20.5.1.2 Interhalogenfluoride-based apparatus Vacuum lines used with halogen fluoride compounds (e.g., BrF5) contain many of the same features found in lines used with F2 gas, but arranged and used in a different sequence. The principal difference (apart from the fluorination reactions themselves) is that neutralization of excess of BrF5 takes place after all oxygen has been removed from the reaction vessels, rather than during the removal of 02 and it's conversion (if carried out) to CO2. As a consequence, considerably less time is required for 02 extraction and collection when using a BrF5-based line. A sequence of extraction, collection, and conversion of oxygen typically requires less than 10 minutes per sample. Moreover, the ability to condense BrF5 and reaction-related waste products, plus the CO2 converted from the released 02, permits cryogenic purification of the sample for mass spectrometry. Figure 20.1 illustrates the conventional vacuum line in use at the Geological Survey of Canada. In principle, the operational components of the line in Figure 20.1 are similar to those found in vacuum fluorination lines used in other laboratories, but with inevitable differences in detail representing variations on basically the same theme. A Kel-F tube (between valves 2 and 3; Figure 20.1) and a nickel tube (attached via valve 5), both of similar size to the individual reaction tubes (inset G, Figure 20.1; see also Figures 20.3 and 20.4), provide for double distillation of a working aliquot of BrF5 (or, C1F3; in this case enough for ca. 250-300 samples) from the storage cylinder (inset E). As previously noted, the Kel-F material permits the purity of BrF5 to be estimated, whereas the nickel vessel below valve (5) is used for storage of the working aliquot of distilled reagent. Accurate measurement of single-sample aliquots of BrF5 is made using (G-2), a 0-25psi stainless steel absolute gauge, or capacitance manometer
456
Chapter 20- B.E. Taylor
(e.g., Setra| An oxygen service-ready bourdon tube gauge, which has been typically used in the past provides a serviceable, though less precise, alternative to the capacitance manometer. The capacitance manometer (G-2; or, bourdon tube gauge) provides a means by which to monitor both the pressure of argon (slightly greater than one atmosphere) used during sample loading, and the small, single-sample aliquots of reagent and general state of vacuum. Excess BrF5 is neutralized (converted to bromine) in the KBr reactor (inset D, Figure 20.2), and the product bromine cryogenically trapped in the adjacent glass trap described earlier. Rapid, contamination-free loading of 12 samples is facilitated by a cap (inset "G", Figure 20.1) and loading tool system (Figures 20.3 and 20.4) that permit samples to be loaded into reaction tubes continuously flushed with a slight positive pressure of argon. Samples can be rapidly loaded even on days with extreme humidity without the slightest adverse effect. Nickel reaction tubes (inset "G", Figure 20.1) are externally heated (inset H, Figure 20.1). The temperature of the reaction is monitored and/ or controlled, e.g., by a temperature controller and thermocouple (typically Cr-A1, type K) attached to the nickel reaction tube using thin wire (thin Ni-Chrome| works well), or by a variable transformer, calibrated for each furnace according to the output of the dedicated thermocouple. The thermocouple is positioned so that the tip of the thermocouple is roughly adjacent to the sample at the bottom of the vessel (i.e., ca l cm from the bottom of the reaction tube). The furnaces are hung such that the hottest zone of the furnace is adjacent to the tip of the control thermocouple. Simple, dedicated variable transformers provide an alternative to temperature controllers, but require that each transformer/furnace pair be calibrated to permit precise control of the reaction temperature. Variable transformers or SCR-type temperature controllers extend the life of the furnace relative to an on/off type temperature controller. Extracted oxygen is typically converted to CO2 (inset I, Figure 20.1; Figure 20.5) for ease of handling and transport, and the quantity of CO2 (yield) recovered is measured with (G-4), an absolute capacitance manometer (e.g., Setra| 0-10psi gauge), a recommended replacement to Hg manometers of the past. The in-line u-tube configuration of the manometer permits a left-to-right sample flow that simplifies and facilitates a speedy extraction process, as well as cryogenic purification if required. 20.5.2 Maintenance
Certain aspects of maintenance are unique to vacuum fluorination lines, due to the aggressive properties of fluorinating agents. Rotary vacuum pumps, especially those used in the "waste" line where direct pumping of the line or KBr furnace and trap occurs, require frequent oil changes (e.g., every 3 months), as indicated by marked darkening of the oil. Chemically resistive oils (e.g., Alcatel 200| are recommended. Pre-fluorinated, corrosive-resistant oils are available which would need to be changed less frequently, but these oils are expensive and still become contaminated with bromine and/or chlorine over time. Solid waste products of the fluorination process (e.g., fluorides of nickel and other, sample-derived metals) must be removed periodically in order to maintain reliable
Fluorination Methods in Stable Isotope Analysis
457
results. Failure to do so can lead to variable and often higher measured 6180 values. Several options for removal of waste products are available, but the simplest method is that which avoids removing the nickel tubes from the line and maintains the integrity of the vacuum connections. For this purpose, a simple vacuum system comprised of a bench-top vacuum pump (e.g., bellows-type), connected to an evacuating flask with a few milliliters of rotary vacuum pump oil in the bottom, and a 30cm length of stainless tube connected to the flask by means of Tygon| tubing is very efficient. While a low pressure (ca. 2-5psi) of argon is streamed through the line and reaction tube, acting as a barrier to contamination by atmospheric moisture (as when loading a sample), the vacuum pump is turned on, and the tube insert down through the "chimney cap" to the bottom of the nickel tube. Gently tapping the sides of the nickel tube with a soft object (e.g., piece of wood) will knock loose NiF from the walls without causing undo damage to the tube. A vacuum cleaning process of this sort is required about once every 12-15 reactions (per tube), depending on the amount and type of sample, fluorinating reagent, and reaction temperature. Typically, 1-2 over-night fluorinations are required before the reaction tube can be re-loaded with confidence. Nickel fluoride is the only residual product of quartz fluorination, and several more reactions may therefore be possible before cleaning of the reaction tube is required. In any case, it is good practice to fluorinate and analyze an internal laboratory with each set of samples, using a different reaction tube each time, in order to monitor the state of reliability of the reaction tubes and need for cleaning. A more thorough cleaning can be performed on an annual basis (or, more frequently if needed) that should include re-polishing of the interior of each tube using a metalworking lathe. Honing of the tube interior may be required also in order to produce as smooth a finish as possible, although this has been found to be necessary only after several years. Where BrF5 is used as the fluorinating reagent, care must be taken to remove any bromine deposits that accumulate as a ring at the top of the "chimney", around the sealing area of Teflon ferrule. Bromine deposits such as these are very hygroscopic and absorb sufficient moisture during sample loading (these deposits are not protected by the argon flow) to contaminate the analyses. The deposits can be readily removed with a water-dampened laboratory tissue, followed by a wipe with an acetone or ethanol-dampened tissue. Cap cleaning can conveniently be monitored and performed when vacuum cleaning the reaction tubes. Similarly, cleaning and re-polishing of the sample-loading tool (Figure 20.4) should be done every several months, as needed. Storing the loading tool in a dessicator extends the period between re-polishing. In systems where the extracted oxygen is converted to CO2 using a converter of the type shown in Figure 20.5, the carbon rod must be replaced about every 80, or so samples. A 2-cm length of spectrographically pure 3mm O.D. graphite rod is broken off for this purpose, and the outer surface roughened (e.g., using a small file) to increase the surface area for reaction, and dust removed. The converter should be vented, and the outer (optionally water-cooled) jacket removed and its interior cleaned. Aqua
458
Chapter 20- B.E. Taylor
Regia can be used to remove any precipitated metal (Pt) films, followed by distilled water, then acetone or ethanol (to remove the water). Replacement of the carbon rod and alignment/centering check of the Pt wire coil are followed by open-air heating of the rod for several minutes, before the outer jacket is repositioned and evacuated, to remove any surficial carbon dust and increase the rod's surface area. 20.6 Laser-assisted micro-scale fluorination Laser-assisted fluorination systems are undergoing such rapid evolution and development that, as emphasized by Rumble & Sharp (1998) in their excellent review, any summary of laser-assisted techniques and apparatus is very soon out of date. The aim, here, is to focus on selected features that seem either to be achieving some sort of 'standard' status a n d / o r are worthy of incorporating in current and future systems. The high-resolution sampling afforded by laser-assisted extraction has prompted both the investigation of many important problems and the evolution in the manner of sample introduction to the mass spectrometer.
Fluorination processes and general procedures applied in microanalysis are, except in UV laser-based systems, rather similar to those described previously in "macro-" or "conventional" fluorination systems. Miniaturization is, of course, the underlying theme. On the one hand, the volumes of the samples analyzed and the vacuum extraction lines used have been greatly reduced, indeed. On the other hand, a number of associated new problems and sources of errors have arisen that can adversely affect the size, precision and accuracy of the analysis. Laser/sample interaction can be summarized as either (1) cold desegregation of the sample comprising coarse fragmentation or, in the case of the ultra-violet (UV) excimer laser, atomic-scale ablation, or (2) intense heating due to an increase in the vibrational energy of the lattice as it absorbs the laser energy (Farquhar & Rumble, 1998). CO2 lasers, for example, emit a characteristic infrared (IR) wavelength of 10,600nm (10.6gm) that is readily absorbed by a wide range of silicate and oxide minerals. The temperature that the sample may ultimately attain (within a matter of seconds or fractions thereof) can reach more than 1400~ The efficiency and magnitude of the heating depends on the wavelength of the radiation and the mineral under investigation. Optimum coupling of the laser wavelength and absorption characteristics of minerals ensures controlled reactions, shorter reaction times, and better yields. Consequently, some lasers are better suited to certain applications than others. See Kyser (1995) and Rumble & Sharp (1998) for a summary of some laser and mineral energy-related characteristics. In contrast to irradiation by IR lasers, minerals irradiated with a UV laser undergo electronic excitation that results in bond breakage and ionization without intense heating whenever the mineral-specific ablation threshold is exceeded (Haglund & Itoh, 1994; Wiechert & Hoefs, 1995; Farquhar & Rumble, 1998). Fluorination occurs, in this case, in a cold plasma, rather than at the heated mineral surface. Some heating can occur in isolated cases (e.g., quartz irradiated with a 248nm wavelength ArF excimer laser; Wiechert & Hoefs, 1995), but this is more the exception than the rule.
Fluorination Methods in Stable Isotope Analysis
459
Whereas, oxygen isotope analysis by laser-assisted fluorination of single grains and small domains cut with thin-bladed saws (e.g., Elsenheimer & Valley, 1992), or broken from thin wafers of rocks and/or minerals (e.g., Sharp & Moecher, 1994) is relatively reliable, with a precision and accuracy commensurate with conventional analysis of larger samples, in situ analysis has proven difficult. Difficulties may stem from several factors, including, room-temperature reaction of the sample (or, constituents of the sample) with the fluorinating reagent, possible partial isotope extraction and fractionation adjacent to the analytical spot, and the energy absorption-dependent nature of laser / mineral interaction.
20.6.1 Spatial resolution and precision High-spatial resolution sampling is, of course, the driving force behind laser applications. The thin-saw blade technique of sampling offers a resolution on the order of 500-1000~m (Elsenheimer & Valley, 1993), and, if fortuitously broken in the desired locations, manual fragmentation of taped, thinly sliced specimens (Sharp & Moecher, 1994) can permit sampling at a resolution of perhaps 200~m. Although both methods require careful dexterity in the extraction and handling of the micro samples, they offer the advantage of examining samples for inclusions, thus avoiding unseen contamination. The size of the area (or, "spot") analyzed using microbeam techniques such as the electron microprobe or SIMS, is determined by the diameter of the incident electron or ion beam. An aperture placed in the path of the UV excimer laser determines the size of the analytical area (e.g., 250~m O: Wiechert et al., 2002). In contrast, the size or volume of the area analyzed in non-UV laser-based systems is variable, and markedly exceeds the nominal laser beam "diameter" due to a number of factors. These factors include the type of mineral and laser, wavelength and fluence of the laser energy, duration of reaction, temperature of the sample, and the pressure and type of fluorinating reagent. For example, the nominal diameter of the CO2 laser beam used by Beaudoin & Taylor (1994) was ca. 65~m, whereas the typical reaction crater or "pit" was about 130~m, with a depth on the order of 150-200~m. If resolution is viewed as the closest distance between analyzed "spots" possible without overlap, then resolution is approximately equivalent to the radius of the average "spot". In this case, the current resolution for in situ laser sampling, based on the size of reaction craters reported by various authors (e.g., Table 20.3) is on the order of 75~m (infrared laser/ sulfides) or ca. 100/~m (infrared laser/silicates), and 125~m (excimer laser/silicates). Alternatively, the resolution is lower by a factor of two if the diameter of the analyzed area is chosen as the criterion. The amount of sample gas produced in laser-assisted extraction typically is some 0.1 to 30% of that associated with conventional fluorination, i.e., on the order of 0.1 to a few ten's of micromoles; smaller quantities from in situ analyses, larger quantities from analysis of grains, chips and powders. This variability may be gleaned from Table 20.3. The practical limit to sample size is determined by both the nature and precision of the mass spectrometer and type of sample inlet used, as well as by the relative size of any blank analyte in the system. When continuous-flow (carrier gas) inlet
460
Chapter 20- B.E. Taylor
techniques are used to introduce the sample into the mass spectrometer, the size of the actual gas sample measured may be on the order of nanomoles, and the origin and size of any blank become a significant concern (e.g., Young et al., 1998a). However, the nature of a split interface, which may admit an aliquot of only a few nanomoles, and the sensitivity of the mass spectrometer, typically require preparation of a considerably larger sample (e.g., 300nm, or 0.3gm; Wiechert et al., 2002). The variation of precision for 8180 with sample size for each of several instrumental techniques is illustrated in Figure 20.9, assuming for a hypothetical yield of 10 g m / m g of sample having the density of calcite. The smallest possible sample size for each technique is also shown for comparison. For a desired uncertainty on 8180 of 0.2%0, a maximum resolution of ca. 40gm is possible, based solely on the basis of gas analysis using a helium carrier gas inlet system, and the presumed absence of additional causes of uncertainty such as the blank (Young et al., 1998a). This precision is degraded by additional factors, including contribution from a blank, isotopic fractionation of the sample during preparation, etc.
Figure 20.9- Plot of sample size limits (as nanomoles of CO2) and theoretical precision (i.e., best-case, or shot-noise limit; in permil) of 8180 versus size of micro-sample, measured using (1) dual inlet mass spectrometry (IRMS with micro-volume); (2) continuous-flow, or carrier-gas inlet, isotope ratio monitoring (GC-IRM-MS); (3) static mass spectrometry; and (4) ion probe. Micro-sample dimensions are given in mm, as the diameter (and height) of a cylinder of material (e.g., calcite) yielding 10mmoles 0 2 / m g (adapted from Rumble & Sharp, 1998; Young et al., 1998a).
Fluorination Methods in Stable Isotope Analysis
461
20.6.2 Micro-scale fluorination with F2
The quantity of oxygen extracted, means of sample heating (in the case of infrared lasers), and the length of time allotted to a fluorination reaction, are the principal aspects that distinguish laser-assisted fluorination protocols using F2.
20.6.2.1 Separation of analyte from fluorinating reagent
The small size of analyte produced in laser-assisted micro-analysis requires marked attention to the purity of the analyte eventually introduced into the mass spectrometer. In cases where halogen fluorides are used, or where the analyte is condensable (e.g., SF6), cryogenic separation can be used (provided that recovery of the analyte from cold traps is 100%). Multiple traps in series, high-surface area miniature traps (e.g., Taylor, 2004; Figure 20.10), and/or variable temperature traps (Taylor & Beaudoin, 1993; Beaudoin & Taylor, 1994) can be used efficiently. In systems where F2 is the fluorinating reagent, investigators have generally resorted to miniaturization of technology proven in a conventional vacuum line, but with some differences. KBr of high purity, or high-purity NaC1 such as may be had from the optics industry can be used as F2 getter (converter) material. As noted earlier, other in-line reaction traps (e.g., hot mercury: Sharp, 1990; stainless steel wool: Akagi et al., 1995) have also been used to remove F2.
20.6.2.2 Grains, glasses, and powders Fluorination of mineral grains or mineral/rock powders by laser-assisted methods resembles reaction [20.6], above, but with several caveats. The first caveat, is that greater attention must be paid to the "blank" (i.e., sources of oxygen, for example, unrelated to the reaction site), for it can easily comprise a much larger fraction of the derived gas sample than customary in conventional analysis. In the case of oxygen, sources unrelated to the sample could include an oxygen impurity in the F2 reagent, moisture in the vacuum chamber or adsorbed to surfaces of the sample, and/or to reaction of other sites in the sample. Contributions of oxygen from the F2 reagent can be quantified as noted in a preceding section, and adequate (e.g., overnight) evacuation, combined with moderate heating of the sample and chamber can reduce much of the adsorbed moisture in the sample and sample chamber. Room-temperature pretreatment(s) with F2 are necessary, however, prior to analysis until oxygen (or sulfur, in the case of sulfides) is no longer derived from the sample chamber. Reaction of minerals such as feldspars, clays (e.g., kaolinite) and carbonates with fluorine and/or its compounds at room temperature can occur even at 0~ Hamza & Epstein (1980), resulting in cross-contamination and large blanks. A cold stage, with which the temperature of the sample can be maintained at -100~ can be used to render mineral grains (or, thin slab; see below) virtually non-reactive, and avoid room-temperature reaction (even from highly reactive substances such as synthetic Ag2S: Taylor, 2004; Figure 20.11). Alternative methods can be used to avoid cross contamination that involve devices to introduce only one sample at a time to a fluorine-rich atmosphere (e.g., Spicuzza et al., 1998b), but these are generally limited to analysis of grains or whole-rock powders.
462
Chapter 20 - B.E. Taylor
Figure 20.10 - Schematic diagram of MILES (Micro-Is| Laser Extraction System) stainless steel sulfide fluorination line constructed using 1/4" and 1/2" (heavier lines) tubing. Looped tubing is of 1 / 8" OD stainless steel. System includes a cold stage (item 3; see Figure 20.11) and P-type cryogenic trap (item 5; see Figure 20.8A) for efficient separation of microquantities (e.g., < 1.0mmole) of SF6 from excess F2. Single-temperature cryogenic traps (all at-197~ include finger trap T1, open to sample chamber and built into base of valve, and cryogenic traps T2 to T6. Other components are: (1) binocular zoom microscope, camera and monitor; (2) 25watt CO2 laser and beam delivery (lense/mirror) system; (3) small-volume sample chamber and integral cold stage (separate modules for in situ analysis on micro-slabs, or bulk analysis of mineral grains/powders; see Figure 20.11); (4) flexible metal hose (1/4" OD); (6) VTT (see Figure 20.8B); (7) calibrated volume and capacitance manometer (including finger trap, T7, for samples larger than a few mmoles); (8) 6mm O.D. glass collection tube; (9) inlet for commercial F2 tank to recharge Asprey-type F2 pump (item 10; K2NiF6-filled Monel vessel, fitted with fine Monel screen at outlet, and external furnace); (11) excess F2 storage volume; and (12) activated charcoal trap for passivation of F2 to CF4. High vacuum is provided by corrosive resistant turbo-molecular pump. Valves include: (1) sealed, Monel tipped (Nupro | SS-4UW-TW) and welded to tubing in the F2 storage/generation section; (2) stainless steel, Kel-F tipped, 1 / 4" toggle valves (Nupro 0, SS-4BKT) with Swagelock | end connections in the majority of the vacuum line, and (3) pneumatically-actuated, Kel-F tipped, 1/2" valves (Nupro@). After Taylor (2003), Beaudoin & Taylor (1994) and Taylor& Beaudoin (1993). Addition of an F2 passivation module (miniaturized version of inset J, Figure 20.1), and either a C ~ CO2 converter module (miniaturized version of Figure 20.5), placed between (6) and (7), or (better) a cry| carrier gas inlet system for use with 02 directly (e.g., Figure 20.11; Young et al., 1998a; Wiechert et al., 2002), adapts this vacuum line for oxygen isotope analysis.
Fluorination Methods in Stable Isotope Analysis
463
Figure 20.11 - (A) Schematic exploded view of small-volume sample chamber fitted with cold stage and ZnSe window for use with infrared CO2 lasers (modified from Taylor & Beaudoin, 1993). Other window materials can be selected according to the laser's wavelength, and readily removed for cleaning by unscrewing snugly (but not tightly!) installed brass retainer ring. The sample chamber is mounted on a motorized (or manual) X-Y stage, and fitted with a manually operated stage in the Z-direction. The chamber is opened at the KF-type flange in the base for sample loading, and vacuumed sealed with a Kal-Rez | or (dark) green Viton | O-ring. (B) Schematic view of cold stage for in situ analysis on 0.5 to 1.0mm thick wafers (after Taylor, 2003). A separate sub-stage module, made to accept 0.5" x 1/4" OD x 1/8" ID nickel crucibles for analysis of grains/fragments or powders, is also available, modified in Taylor (in press) after Taylor & Beaudoin (1993) by addition of similar cold-stage. In either case, cooling of the sample (typically maintained at -100~ on the thermocouple) is facilitated by immersion of Cu rod in liquid nitrogen. Heating of the sample during overnight out gassing is accomplished by placing a small tube furnace around the Cu rod of the cold stage.
A second important difference between laser-assisted and conventional fluorination is the effect of sample/laser interaction, which directly impacts the yield. These effects include physical loss of sample, loss of analyte, and loss or reduction in susceptibility to fluorination. Scattering of grains of quartz under the laser beam is perhaps the most notorious of the physical interruptions to complete fluorination. Although the exact cause is a matter of speculation, energy absorption at the c~-f3transition and/ or by marked gradients in power density at the sample surface would seem to be involved. Numerous schemes to avoid these frustrating difficulties, including mechanical restraints of the quartz grains (e.g., Kirschner & Sharp, 1997), control of grain size (Fouillac & Girard, 1996), and rapid irradiation under a defocused beam
464
Chapter 20- B.E. Taylor
protocol (Spicuzza et al., 1998a) have met with various degrees of success. Processes of absorption of the laser energy by the product gas can reduce the yield by dissociation of the analyte gas molecule (e.g., SF6: Beaudoin & Taylor, 1994). Such a dissociation process is inevitably associated with an isotopic fractionation of the residual SF6 (Beaudoin & Taylor, 1994). Rapid collection (exposure to a liquid nitrogen trap while lasing), appropriate pressure of F2 in the sample chamber (Beaudoin & Taylor, 1994; Rumble & Hoering, 1994), and a fluorination reaction of minimal duration can usually rectify this potential problem. The products of fluorination reactions by F2 under a CO2 laser can be both solid (e.g., FeF3 or ZnF2) and liquid (e.g., MgF2 or AgF2). The metal-fluorine melts appear to have various energy absorption characteristics with respect to the incident laser energy. Mg-fluoride and Ag-fluoride melts, for example, become transparent to the 10,600 nm characteristic wavelength of the CO2 laser. The end result is a lower-thanexpected yield of 02 from Mg-olivine (Farquhar & Rumble, 1998), for example, or of SF6 from Ag2S (Taylor, in press). In the case of Ag2S, the analyses are unaffected for yields >60% (Taylor, 2001), but problem can be more acute for olivine (Farquhar & Rumble, 1998).
20.6.2.3 In-situ analysis The use of lasers to provide a routine means of in situ isotopic analysis somewhat analogous to the point-source, high-spatial resolution chemical and isotopic analyses facilitated by the electron microprobe and SIMS, has been the dream of many. On the one hand, analysis of mechanically-separated, micro-samples (e.g., individual grains, drilled powders, grain fragments) of silicates and oxides by means of laser-aided fluorination is, if not routine, a reality. On the other hand, success in the arena of in-situ oxygen isotope analysis has not kept pace, whereas acceptable in-situ analyses of sulfide minerals, whether by oxidation using 02 (e.g., Crowe et al., 1990), or fluorination (e.g., Beaudoin & Taylor, 1994; Rumble, et al. 1993; Taylor & Beaudoin, 2000) has been clearly demonstrated. In the case of in situ sulfide fluorination, reproducibility of 634S of 0.03%o can be readily attained with sample sizes on the order of 0.1~mol (Beaudoin & Taylor, 1994; Taylor & Beaudoin, 2000). Indeed, similar precision (< 0.05%o) can be attained even when several poly-mineralic samples are in the sample chamber. Oxygen and sulfur isotope fractionation during in situ fluorination has been ascribed to 'edge-effects' for some time, beginning with the sulfur isotope study of Crowe et al. (1990) and the oxygen isotope study of Elsenheimer & Valley (1992). In both studies, a lack of stoichiometry in rims about laser-induced reaction craters formed by 02 oxidation in Crowe et al. (1990) and by fluorination in Elsenheimer & Valley (1992) which has been taken to indicate a zone of oxygen or sulfur isotope fractionation. The textures and chemistry of these melt zones is certainly distinct (reviewed in Rumble & Sharp, 1998). Although stable isotope analyses of reaction rims have yet to be performed, such 'edge-effects' have been considered among the chief obstacles to accurate in situ analyses by fluorination.
Fluorination Methodsin StableIsotopeAnalysis
465
The large size of some melt zones and the observed effects were promoted in the studies noted above by relatively long duration of laser/sample interaction and the absence of any ability to cool the sample during fluorination. Sulfides are especially thermally conductive, yet the use of a cold stage can markedly reduce the size of the reaction craters and any such edge effects, and facilitate consistent isotopic results (Taylor & Beaudoin, 2000). As noted earlier, experimental evidence on partial fluorination using F2 gas, plus a lack of isotopic measurements of material in such aureoles, leaves the question open as to whether such isotopic fractionation occurs in crater rims produced during F2-based fluorination. It seems fair to say, that, although melting in the rim of a reaction crater may create a potential issue (e.g., Rumble & Sharp, 1998; among others), the purported deleterious isotopic effects need to be critically reexamined. Other aspects of the fluorination process can also be of concern. For example, sulfur isotope fractionation during simultaneous fluorination of both a sulfide and a silicate grain might spoil in situ sulfur isotope analyses in a silicate matrix due to the formation of SOF2 (or similar compounds such as SO2F2). The presence of SOF § ions can, as a 'check', be routinely monitored by peak jumping when samples large enough for classical isotope ratio mass spectrometry. Rumble & Hoering (1994:) suggested that sulfur isotope analyses of 30 gm diameter pyrite grains in a silicate matrix were in error owing to the formation of S-O-F compounds. On the other hand, analyses of 75100gm grains of pyrrhotite grains in fine-grained quartz arenite and siltite matrices by Taylor & Beaudoin (2000) appear to have been unaffected by the presence of adjacent quartz caught up in the in situ reaction. In an attempt to understand the S/O constraints under which sulfur isotope fractionation might occur, sphalerite-quartz mixtures with weight fractions of quartz up to 50% were analyzed without apparent isotopic shifts (B. Taylor & Ross, unpub, data). The consistency of correction factors for ~170 and g180 analyses derived from laser fluorination of barite (BaSO4; Bao & Thiemens, 2000) and Bao & Thiemens' assessment of the relative significance of SO3 and SO2F2, also suggests that a limited amount of oxygen may be incorporated during in situ fluorination of sulfides without markedly adverse effects. 20.6.3 B l a n k s a n d o t h e r s o u r c e s o f error
The isotopic integrity of the small quantities of 02 gas produced in laser-aided microanalysis can be readily compromised by sources of oxygen ('blanks'; e.g., air) other than the sample. The danger is more pronounced the greater the difference in the isotopic composition of contaminant and sample, and the greater the complexity and number of steps in sample processing. The process of cryofocusing a small sample (or aliquot) on molecular sieve at liquid nitrogen temperature does not result in any measurable isotopic fractionation (Wiechert et al., 2002). However, adsorption of 02 gas in a stainless steel cryogenic trap at liquid nitrogen temperature such that the 02 sample does not quantitatively pass through these traps, can form a blank that could isotopically alter successive 02 samples (Young et al., 1998a).
466
Chapter 20- B.E. Taylor
20.7 Laser fluorination apparatus 20.7.1 Laser Selection
Early workers chose Nd-glass or Nd-YAG lasers (e.g., Elsenheimer & Valley, 1992), in part because of a perceived need for a high fluence. The short wavelength of the Nd-YAG laser (1,064nm), however, is best absorbed by transition metal (e.g., Fe) bearing minerals; quartz is essentially transparent to this wavelength. Thus, it is the wavelength and absorption characteristics of the sample that are prime factors, rather than simply the power (often quoted in watts) of the laser. Appreciation of the absorption spectra of minerals in light of the laser's characteristic wavelength (Sharp & O'Neil, 1989; Sharp, 1990; Rumble & Sharp, 1998; graphically illustrated in Kyser, 1995) soon led to a wider spread employment of infra-red lasers. Metal-oxygen bonds readily absorb the 10,600nm wavelength of the CO2 laser, leading to efficient heating of most silicates and oxides. In fact, many minerals that were known to be difficult to analyze using conventional vacuum fluorination techniques (e.g., olivine, garnet, aluminosilicates, zircon, etc.) can be readily analyzed under a CO2 laser beam (e.g., Valley et al., 1994; Sharp, 1995; Tennie, et al., 1998; Krylov, et al., 2002), whereas analysis of a few minerals (e.g., sapphire grains; Fe-free sphalerite by in situ analysis) pose greater difficulties. More recently, harmonic generators have been added to Nd-YAG laser systems to alter the output wavelength frequency of 1,064nm. This so-called "frequency summing" or "sum frequency mixing" results in wavelengths of 532nm (frequency-doubled, o r 2 nd harmonics; near infrared), 266nm (frequency-quadrupled, o r 4 th harmonics; UV), 213nm (frequency-quintupled, o r 5 th harmonics; UV). Silicate minerals and oxides more readily absorb the lower frequencies (266nm and, especially, 213nm), and this has revitalized the early interest in Nd-YAG lasers (e.g., Jackson & Ryde, 2001; Rumble & Sharp, 1998). Much of the development of in-situ sampling using such modified Nd-YAG lasers has been carried out by those utilizing UV laser-induced ablation in ICP-MS applications for chemical and isotopic analysis, promoting efficient sample delivery to the ion source (e.g., Durrant, 1999; for ArF excimer lasers see also: G~inther et al., 1997; G~inther & Heinrich, 1999; Bizzarro, et al., 2003). Reduction in size of analyzed area and increased coupling efficiency for different minerals have driven much of the interest in application. A frequency-quintupled Nd-YAG laser was recently (Hirata, 2003) utilized to increase the yield from a 16~m spot size (on zircon) analyzed for chemical and Pb isotope composition. Actual ablation (fragmentation) of a mineral in response to absorption of laser energy differs markedly from the effects of localized heating due absorption of (comparatively longer) infra-red (e.g., CO2) laser energy; and both chemical and isotopic fractionation can be problematic (e.g., Jackson & G~inther, 2003). These latter issues must be addressed in each application involving ablation. A useful descriptive overview of the laser-induced ablation process can be found in Russo et al. (2002). The KrF excimer laser has been successfully applied to in situ analysis of silicates by Wiechert & Hoefs (1995), with little or no difference (0.1 to 0.4%o) in reported 6180 values of spot analyses compared to conventional analyses of nominally homoge-
Fluorination Methods in Stable Isotope Analysis
467
neous minerals. However, not all minerals absorb radiation in the UV spectrum (e.g., quartz is transparent to the KrF laser). Moreover, fluorination reagent/laser combinations are not all equally viable. For example, neither ArF excimer lasers, whose 193nm wavelength can be absorbed to some degree by quartz, nor XeC1 excimer lasers facilitate fluorination by C1F3 (Wiechert & Hoefs, 1995). Fluorination using BrF5 yields values of (~180 that vary systematically with reagent pressure when using a KrF excimer laser (Rumble et al. 1997; Farquhar and Rumble, 1998). Thus, for many purposes, the sealed-tube CO2 laser is still a good, affordable choice, and one that is economical to maintain in comparison to UV excimer and other lasers requiring constant replacement of a fill gas.
20.7.2 Optical system It is beyond the scope of this paper to describe in detail the many variations of beam delivery systems in use. Suffice it to note that beam delivery systems (i.e., laser plus lenses) basically can be divided between those that are stationary, with the sample chamber translated beneath the laser beam, and those systems with moveable lasers, in which case the sample chamber remains in a fixed position. All such systems require various turning (reflecting) mirrors, beam expanders and condensers arranged so as to bring the laser beam (which may exit the laser with a diameter of 6mm or more) to the sample surface as a fine "spot" as the wavelength of the laser and budget for high-quality optics permits. For example, standard, reasonably priced optics used with the CO2 laser at the G.S.C. (Ottawa) result in a minimum spot diameter of ca. 64~m (but, as previously noted, the analytical "spot" is larger due to reaction). Specialized optics such as multi-component/multi-faceted lenses are available that attempt to "flatten" the power distribution in such a way that the laser peak is more "flat-topped" than the usual Gaussian peak. Whereas Nd-YAG laser beams may be directed through a sample-viewing microscope (owing to the transparency of optical components to 1,064nm wavelength), CO2 laser beams (10,600nm) must be reflected into the sample chamber externally from a viewing microscope. Many systems (e.g., Elsenheimer & Valley, 1992) use to great advantage a CCD camera and monitor that often provide a better means of sample observation. Safety shielding against stray reflections along the beam path and appropriate personnel safety accessories are, of course, to be strictly adhered to.
20.7.3 Vacuum system construction Vacuum systems, including, sample chambers, and sample collection, purification, and introduction to the mass spectrometer are undergoing continual evolution, with miniaturization perhaps the most common denominator. The small sample sizes involved, especially those derived from micro sampling (e.g., in situ, or physically separated) demand a high degree of integrity of the vacuum system. Both high- and low-vacuum pumps used with F2-based systems should be of a corrosive-resistant nature, with pre-fluorinated lubricants. The ready absorption of CO2 laser radiation (in wide use among laboratories) by most minerals, obviates the use of clear glass as a window through which to pass the
468
Chapter 20- B.E. Taylor
laser radiation into reaction chambers. Consequently, other materials must be used to construct vacuum tight, non-reactive, sample chambers that transmit the externallysourced laser beam. Barium fluoride (BaF2) or zinc selenide (ZnSe) windows are typically preferred for use with CO2 lasers, whereas other materials (quartz, pyrex, topaz, diamond) are chosen for Nd-YAG laser applications. However, ZnSe is amber in color, whereas BaF2 is clear. This, and the fact that ZnSe windows are a bit more expensive, has lead some workers to initially choose BaF2. Nevertheless, ZnSe should be considered where high-powered CO2 lasers are used because of its characteristically higher transmission and greater "strength". For use with CO2 lasers, the ZnSe window is coated with magnesium fluoride, which, in addition to low reaction rates at room temperature, yields a robust, essentially non-reactive window for fluorination. BaF2 is softer than ZnSe, and may also crack in response to the high fluence of CO2 laser energy, releasing of F2 or other fluorinating reagent to the laboratory. The vacuum seal between the lenses or windows and the (often stainless steel) housing of the sample chamber are typically made using Kalrez| O-rings, a fluorinated "soft" Viton| This material is rather expensive, but resistant to attack by BrF5 and C1F3 (e.g., Sharp, 1990) and F2 (e.g., Taylor & Beaudoin, 1993). Recently, Miller et al. (1999) reported using (dark) green Viton| O-rings (available in Europe) to seal a BaF2 window to a vacuum chamber used for fluorination reactions employing BrF5. Green Viton| O-rings cost only about 1 / 10 that of Kalrez| Recent tests at the Geological Survey of Canada (Ottawa) confirm the lack of reaction between the dull, dark green variety of Viton| and F2 gas. A lighter, shinier green variety, which can be readily distinguished from the darker green variety, as well as the usual black and brown varieties of Viton| are not suited to the purposes here, and react to varying degrees with F2. Reduction of surface area of the interior of the sample chamber by electro- or mechanical polishing, is advantageous for the reduction of the background in the vacuum system and time required for evacuation to an acceptable level. This constitutes good practice for high-vacuum systems in general, often with readily measurable effects (e.g., Fouillac & Girard, 1996).
20.7.4 Mass spectrometer inlet techniques for small samples Two techniques are available for dynamic mass spectrometry (see Pillinger, 1992, for a discussion of static mass spectrometry), and the choice is essentially controlled by the size of the sample and, to a lesser degree, by the sensitivity of the mass spectrometer. These are (1) classical isotope ratio mass spectrometry using a micro-volume; and (2) isotope ratio monitoring mass spectrometry using carrier-gas (usually He) sample introduction (e.g., Matthews & Hayes, 1978; Merritt & Hayes, 1994). For the former, samples as small as 0.1gmol can be analyzed with a precision on par with that for large samples (e.g., 0.1%o for 634S of SF6: Beaudoin & Taylor, 1994; Taylor & Beaudoin, 2000). Transport in a He-carrier gas stream via a split interface results in the introduction of a sample of ca. 3-15nmol, depending on the sensitivity of the mass spectrometer (Wiechert et al., 2002), whose isotopic ratio is measured in a single pulse.
Fluorination Methodsin Stable IsotopeAnalysis
469
The uncertainty of the isotopic ratio measured in this fashion is, in the best case, ca. 3 times larger than by classical means (Figure 20.9). Where permitted by the size of the prepared sample, the averaging of measurements from multiple aliquots (i.e., single "pulses") can reduce the uncertainty of the isotopic ratio, and also the difference in uncertainty relative to classical techniques. What remains unchanged is the fact that measurements by carrier-gas technology require markedly less time (e.g., 10 minutes vs. 20-30 minutes) than by classical, micro-volume inlet techniques. Herein, lies an advantage to the carrier gas inlet technique, in spite of the larger uncertainty. Much smaller samples can be measured in laser applications where non-fluorinating analytical techniques are used in conjunction with a continuous flow of He through the sample chamber and to the mass spectrometer (e.g., Cerling & Sharp, 1996); a second advantage of the He-carrier gas inlet method. The size of the actual sample or aliquot delivered to split interface is larger, as previously noted, than that which enters the mass spectrometer. For a split ratio of 1:3, ca. 33% of the actual sample is admitted. The sample prepared by the laser system, on the hand, may be even larger. In the case described by Wiechert et al. (2002), their 300nmol samples (0.3gmol) are large enough to be measured by classical mass spectrometry using a micro-volume inlet and sensitive mass spectrometer. Inasmuch as a lower uncertainty obtains when measured by classical methods, plus the opportunity to ascertain the presence of one or more contaminants, careful consideration needs to be given to the inlet technique. As noted previously, the presence of blank contributions to a single small sample (such as from an in situ analysis) may be such as to require production of sample sizes sufficiently large for classical measurement. The inlet system described by Wiechert et al. (2002) for small (to 0.3~mol) samples of 02, prepared by fluorination using F2 with a KrF excimer (UV) laser, is shown in Figure 20.12. Following passivation of the excess F2 (traps T1 and T2 condense the released halogen), an aliquot of 02 is expanded into T3. At this juncture, the sample can be further cryogenically purified by vacuum pumping if T3 is filled with silica gel and, at the same time, filled with He. The sample (thawed if required) is switched to a cryofocusing trap in T4 (filled with 5A molecular sieve held at liquid nitrogen temperature). Warming T4 rapidly with boiling water transmits the sample to the mass spectrometer. Although gas chromatographic valves are convenient, they are not essential to the principles of the inlet system (S. Hoernes, Univ. Bonn, pers. commun.). Addition of two prime/purge type valves shown in Figure 20.12, (plus the ability to provide make-up He not shown in Figure 20.12) would permit the monitoring of, among others, the cryofocusing process for that system. A well-designed system incorporating admission of a test 02 gas through the same system, plus the capability to monitor the cryofocusing process, was described by Young et al. (1998a). Removal of water is essential, and this is often accomplished by use of a Nation| membrane (see also Volume IL Part 3, Appendix C7.6 for details). Nation| is a Teflonbased synthetic ion-conductive polymer developed by DuPont that has the very useful property of being highly selective and permeable to water. Thus, Nation| tubing can be used, as here, to great advantage for water removal in carrier-gas applications.
470
Chapter 20 - B.E. Taylor
Figure 20.12- Schematic layout, flow scheme, and valve-operation of a collection and cryo-focusing, He-carrier inlet system for analysis of small 02 samples (adapted from Wiechert et al., 2002; see also Young, et al., 1998a). Traps T1 and T2 are held at liquid nitrogen temperature to contain chlorine or bromine produced by reaction of F2 with NaC1, KC1, or KBr. T3 contains aliquot of 02 sample trapped (cryo-focused) on molecular sievefilled sample loop (T4). He flow is indicated for valve positions of V1 and V2 to effect sample collection and injection modes. Addition of prime/purge-type valves (3-way; gray) with tees and make-up He flow (not shown) can be added to permit background monitoring and monitoring of trapping process (cryofocusing) with the mass spectrometer (cf., Young et al., 1998a).
FluorinationMethodsin StableIsotopeAnalysis
471
Gas chromatography can be readily used as a final step of sample purification (e.g., N2 and other constituents) after passivation of excess F2 and cryofocusing. For this, a 15m-long PLOT column (Porous Layer Open Tubular fused silica) containing 5A molecular sieve has been successfully utilized both at 50~ (Young et al., 1998a) and at room temperature (Wiechert et al., 2002). 20.8 Conclusions
Fluorination, in spite of special apparatus and safety precautions required, is a reliable process for oxygen isotope analysis of silicates, oxides, and phosphates and for sulfur isotope analysis of sulfides. The traditional conventional and newer laserassisted micro-analytical techniques are complementary. Both are needed to manage the full spectrum of analytical challenges, especially in earth sciences. Similarly, the halogen fluoride compounds (BrF5 and C1F3) and F2 gas fluorinating reagents share common ground in some applications, yet occupy particular niches in others. Halogen fluoride compounds are the more commonly used fluorinating reagent in conventional apparatus, although F2 gas, whether directly as commercially supplied (97% pure), or laboratory-purified, is also used. Being able to be produced safely in the laboratory in high purity, F2 gas is particularly well-suited to many micro-analytical applications, simplifying fluorination chemistry and avoiding absorption of laser energy. In addition, partial fluorination reactions are best-accomplished using F2. On the other hand, particularly resistant minerals (e.g., olivine) can more readily be analyzed using BrF5 (but is problematic with C1F3: Vennemann & Smith, 1990) when using conventional applications. Laser-assisted micro-analysis by fluorination facilitates high spatial resolution sampling, permitting isotopic analysis of minerals within their textural context. Questions regarding knowledge of isotopic zoning and scales of isotopic equilibrium can more readily be pursued. However, the marked reduction in sample size over conventional techniques demands close attention to sources of blanks, adsorption, and isotopic fractionation in an effort to produce samples of high purity. Gas chromatographybased purification, and helium carrier gas introduction of very small samples to the mass spectrometer have pushed technical innovations and have taken the field of laser-assisted micro-analysis beyond simply a miniaturization of conventional methodology. Although the sealed-tube CO2 IR laser has proven itself to be generally useful, and other lasers, particularly in the UV spectrum, exhibit some advantages in certain (but not all) cases, laser selection and application remains an active area of investigation. Whereas, the goal of in situ oxygen isotope analysis has not been yet fully attained, in situ analysis of sulfides has been significantly advanced by fluorination. Moreover, laser-assisted sulfide fluorination with pure F2 gas has made feasible improved accuracy (and re-calibration) of the sulfur isotope scale, and of ~33S measurements.
472
Chapter 20- B.E. Taylor
Acknowledgements It is, indeed, very difficult to write a comprehensive review that mentions all possible contributions to this analytical subject during the last nearly 50 years of development. I have, instead, tried to generalize some aspects sufficiently so as to offer the reader an overview that points the way to "tried and tested" aspects of fluorination methods and the necessary apparatus. However, I hasten to add that this review should not be used as a "how-to" procedural guide. Instead, it would be far wiser for the interested researcher to spend some time in the laboratory of an experienced practitioner, for there are innumerable aspects regarding vacuum lines in general, and fluorination procedures in particular, that should be learned. This review reflects not only information available in the literature, but also the guidance, training, collaboration, and contributions of many (suspecting and unsuspecting) along the way, as well as my own. In particular, I have learned a great deal from, among others, Drs. J. R. O'Neil, H. Friedrichsen, S. Hoernes, T. K. Kyser, G. Landis, R. T. Gregory, G. Beaudoin, Z. Sharp and D. Rumble, III. Unfortunately, many other colleagues and students remain unmentioned, but not forgotten. Dr. S. Mirnejad was of great assistance in the last stages of trying to put this together, and P. De Groot exhibited unusual tolerance and patience well into the 11th hour. I especially appreciate the time and care expended by Drs. C. L6cuyer, S. Mirnejad, S. Sheppard, and M. Wilson in the review of this chapter. Their suggestions erased the typos, oversights, and general 'bloopers' and, hopefull~ helped render the final version more tractable to the reader.
Handbook of Stable IsotopeAnalyticalTechniques,Volume 1 P.A. de Groot (Editor) 9 2004 ElsevierB.V. All fights reserved.
CHAPTER 21 Oxygen Isotope Analysis of Plant Water Without Extraction Procedure Kim S. Gan, S. Chin Wong & Graham D. Farquhar*l Environmental Biology Group, Research School of Biological Science, Institute of Advanced Studies, Australian National University, GPO Box 475, Canberra, ACT 2601,Australia e-mail: *
[email protected]
21.1 Introduction Measurement of the isotopic composition of water within plants (mainly xylem, phloem and leaf water) is gaining importance as it reflects plant-environment interactions and this isotopic signal is eventually incorporated into plant organic matter. Previous work has shown that there is no isotopic fractionation of oxygen isotopes of water during water uptake by the roots and water transport through the xylem (Ehleringer & Dawson, 1992 and references therein). When xylem water reaches leaf tissues, evaporative enrichment in the heavier isotopes of hydrogen and oxygen occurs due to the preferential loss of lighter water molecules through the stomata during transpiration (Dongmann et al., 1974). The extent of this enrichment in leaf water is dependent upon air temperature and humidity, as well as leaf temperature, a parameter affected by plant transpiration rate. The isotopic composition of leaf water could be described by the freely evaporating water surface model developed by Craig & Gordon (1965) and modified by Farquhar et al. (1989b) and Farquhar & Lloyd (1993). As there exists oxygen isotope exchange between enriched leaf water and the carbonyl oxygen of triose phosphates (DeNiro & Epstein, 1979; Farquhar et al., 1998), the plant organic matter could be a surrogate for climatic information. This is provided the complexities of isotope fractionation in evaporating leaf water and biochemical processes are well understood, necessitating the need to analyse leaf, xylem and phloem water. Until recently, this type of research required the isolation of water from the plant parts (leaf and stem) for equilibration with CO2 for ~180 analysis. For 62H analysis, the extracted water can be equilibrated with H2 using a platinum catalyst (Coplen et al., 1991) or be reduced to H2 by hot zinc/uranium (Bigeleisen et al., 1952; Coleman et al., 1982; Kendall & Coplen, 1985). The most commonly used extraction methods are azeotropic distillation and cryogenic extraction (Dewar & Mcdonald, 1961; Ehleringer & Osmond, 1989). These methods are time-consuming as the extraction procedure has to continue to completion owing to an isotopic fractionation between the evolving 1. Correspondenceshould be adressed to this author.
474
Chapter 21 - K.S. Gan, S.C. Wong & G.D. Farquhar
vapour and the remaining liquid (Ingraham & Shadel, 1992). Thus, the bottleneck in water purification by vacuum or azeotropic distillation would not complement the speed of isotope analysis provided by continuous-flow IRMS (Isotope-Ratio Mass Spectrometry), especially when large data sets are needed for statistical calculations in biological studies. Further, a substantial amount of plant material is needed for azeotropic distillation and leaf samples would invariably include the unfractionated water from the veins. There has only been one report of an attempt to analyse plant water without undergoing extraction (Scrimgeour, 1995). Twig and stem samples were directly equilibrated with CO2 for 6180 measurement and with H2 in the presence of a platinumon-alumina catalyst for 62H analysis of plant sap. One major drawback of the direct equilibration technique is the liberation of CO2 from metabolic processes in plant samples during equilibration. Scrimgeour (1995) observed that when 5% CO2 in N2 was being used as an equilibrating gas for plant samples, a two-fold increase of CO2 or more was noted after equilibration. To dilute out the isotopic signature of the respired CO2, pure CO2 was recommended for use as the equilibrating gas which would require larger sample size for a higher water content. The stem water undergoing isotopic exchange with the equilibrating gas would represent not only the xylem sap in transit but also the tissue water which could be relatively more enriched. Thus, the best way of analysing xylem sap in the transpiration stream is to use a root-pressure chamber (Yong et al., 2000) to bleed sap from the stem or leaf petiole and directly pyrolyse the sap in a continuous-flow IRMS. To measure sub-microlitre amount of leaf mesophyll water without the need for extraction, we here describe a mass-balance approach where a small disc of fresh leaf is cut and pyrolysed in an elemental analyser for conversion to the analyte gas, carbon monoxide. The latter is separated from other gases by gas chromatography before isotopic analysis in continuous-flow IRMS. By comparing with results from pyrolysis of the dry matter of the same leaf, the 180 content of leaf water can be determined without extraction from fresh leaves. This approach is made possible with a specially designed leaf punch that allows the cut disc to fall directly into a tin cup that is immediately sealed by pincers controlled by pneumatics. The whole process of sampling and sealing the tin cup takes seconds and eliminates handling errors due to evaporative fractionation. As all leaf water isotopic measurements found in the literature were obtained either by the azeotropic or vacuum distillation, we here present a comparison of our new approach of direct pyrolysis of fresh leaf with the azeotropic distillation method. To validate our new technique, fresh leaf samples are simulated by combining cellulose and water of known isotopic composition to give mixtures of variable moisture content.
Oxygen Isotope Analysis of Plant Water Without Extraction Procedure
475
21.2 Experimental
21.2.1 Plant material and sample collection Cotton plants (Gossypium hirsutum L. vat. Deltapine 90) were grown under fullsunlight in a well-ventilated glasshouse (28 + 2~ day, 20 + 2~ night; relative humidity 35 + 10%) until they were 48 to 53 days old. Leaf discs (diameter 3 mm) were cut from fully expanded leaves using a specially designed leaf punch (Figure 21.1). The punch allows the disc to fall directly into a pre-weighed smooth-walled tin cup (9 x 3.5 mm) with the aid of an argon burst and to be sealed immediately with the push of a button. The argon burst also serves to flush out air (and hence exclude nitrogen and oxygen) from the cup. To check for potential leaks through the seal, the filled cups were weighed at regular time intervals before isotopic analysis. After punching out leaf discs from one half of the leaf, the other half was trimmed, avoiding primary veins, and immersed in toluene (80 ml) for water extraction by azeotropic distillation using a specially designed receiving funnel as described by R6v6sz & Woods (1990). Approximately 0.5 ml of leaf water was collected per sample and any trace of toluene was removed by adding wax and warming it in a closed sample bottle. Cooling the bottle right-side up will solidify the wax on top of the water preventing evaporation during storage and handling. Before analysis, warming and cooling the bottle upside down allows water to be decanted easily. Small leaf segments from which leaf discs
Simplified diagrammatic representation of the leaf punch. The punch is designed to cut a leaf disc and flush the collecting cup with argon at the same time, and to seal the cup with the push of a button. Figure 21.1 -
476
Chapter 21 - K.S. Gan, S.C. Wong & G.D. Farquhar
were previously sampled were dried in a 70~ oven for isotopic analysis of organic matter.
21.2.2 Oxygen isotope analysis Oxygen isotopic analyses of water, dried and fresh leaf samples were all performed using the continuous-flow pyrolysis technique described by Farquhar et al (1997). For water, samples of 0.7/~1 were transferred to smooth wall tin cups (4.5 x 2 mm) with a Gilson micro-pipetteman (2 ~1) and sealed under argon with a modified Carlo Erba liquid encapsulator equipped with pneumatics control. For dried leaf samples, 1.0-1.5 mg of material were weighed into tin capsules that were then crimped manually. Smooth-walled tin cups containing fresh leaf samples were directly pyrolysed after recording the fresh leaf net weight. Tin cups were found to have negligible contribution to CO formation (Farquhar et al., 1997). 21.2.3 Standardization Elemental standards As fresh leaf comprises substances in both liquid and solid forms, the pyrolysis system must be capable of converting quantitatively both organic matter and water to the analyte gas. From the thermal conductivity detector (TCD) output of the gas chromatography, yield of CO from encapsulated water is 98-99.5 % of that from beet sucrose. Due to this slight discrepancy, both beet sucrose and water were used as elemental standards in generating the calibration plot when determining the oxygen elemental composition of fresh and dry leaf samples from the TCD output. As water is the main constituent of fresh leaf, water is used as a correction reference in computing percentage oxygen of fresh leaf samples. Isotopic standards Isotopic standardisation of the pyrolysis system was performed using the international water standards, Vienna Standard Mean Ocean Water (VSMOW), Standard Light AntarcticgPrecipitation (SLAP) and Greenland Ice Sheet Precipitation (GISP). All 5180 reported here are expressed relative to VSMOW on a scale such that the oxygen isotopic value of SLAP is-55.5 %o (Coplen, 1994; Gonfiantini, 1984). Internal laboratory standards (water standards: ANU-C1, ANU-P1, ANU-A1, ANU-LP, ANU-HP, ANU-OW; solid standards: beet sucrose, ANU sucrose) were also analysed together with the international standards to check on the calibration agreement between water and solid samples. The internal water standards were supplied by Research School of Earth Sciences, ANU and their assigned 5180 were based on long-term mean values obtained by the conventional H20-CO2 equilibration method. 6180 assigned to beet sucrose and ANU sucrose were previously reported in Farquhar et al. (1997). Calibration of the CO working gas was done periodically in a run using internal standard beet sucrose. Based on nitrogen analyses of representative plant samples, a standard of beet sucrose containing 5% nitrogen was used for correcting dry leaf samples.
21.2.4 Method validation using cellulose-water mixtures A range of cellulose-water mixtures simulating the constituents of fresh leaf was prepared by doping cellulose (Whatman No 1 filter paper) with waters of varying
Oxygen Isotope Analysis of Plant Water Without Extraction Procedure
477
8180, ranging f r o m - 55%0 to + 50%o. The doping waters included SLAP, GISP, ANULP, ANU-HP, ANU-OW and a series of enriched waters prepared by mixing Canberra water with 1.7 atom% 180 water (Amersham, UK). The mixtures generally had a water content of 75 _+3% and total mass of around 0.9 mg. The mass of each component in the mixture was recorded to allow calculation of the percentage oxygen in the mixture and determine 8180 of the doping water from 8180 of the mixture. Another set of cellulose and ANU-OW mixtures was also prepared in a similar way with water content ranging from 44% to 93% and a total mass of 0.66 - 1.12 mg, keeping within the linearity limits of the IRMS. Isotopic composition of the doping waters and cellulose was directly measured in the same run as the mixtures.
21.2.5 Calculation of 6180 of leaf water in fresh leaf sample Direct pyrolysis of fresh and dry leaf samples gives the measured values of 818OF and 818OD from which the 180 content of leaf water (818OLw) present in the fresh leaf sample can be calculated by isotopic mass balance,
818OF = X 818OLW + (1 - x) &18OD
[21.1]
where x refers to the proportion contributed by leaf water in the total oxygen pool of the fresh leaf sample.
21.2.6 Calculation of water fraction in fresh leaf sample To determine x, oxygen elemental composition of fresh and dry leaf samples, OF and OD respectively, were obtained from the TCD output of the gas chromatograph in the same acquisition as the 180 analysis. By oxygen mass balance, OF - y OLW + (1 - y) OD
[21.2]
where y is the water fraction in the fresh leaf sample and OLW is the oxygen fraction of leaf water with a fixed value of 16/18. Thus, x -
y OLW/OF
[21.3]
Overall,
~)18OLw
_
16 &18 OF(1 - T~OD) 818 18OD OD+ OF_OD [ OF-8 ]
[21.4]
For the method validation experiment using cellulose-water mixtures, y and OF were also calculated from the known mass of individual component added into the mixture, assuming OD has a fixed value of 80/162 for cellulose. This exercise allowed us to assess the precision of obtaining percentage oxygen from the TCD output.
478
Chapter 21 - K.S. Gan, S.C. Wong & G.D. Farquhar
Figure 2 1 . 2 - 618OvsMow calibration curve of international reference materials (VSMOW, GISP, SLAP) and some internal laboratory standards. Measured 6180 refer to the machine raw values while s.d. and n denote standard deviation and n u m b e r of replicates respectively. 21.3 Results and discussion
When accepted 6180 values of international water standards and internal laboratory water and solid standards were plotted against their measured raw 6180 values (Figure 21.2), a perfect linear correlation was obtained with a slope of 0.27% deviation from unity. This excellent calibration agreement between solid and water samples in our pyrolysis system opens up a new approach for determining the 180 content of leaf water from the direct pyrolysis of fresh leaves.
Samples
Accepted Measured 618OvsMow, %o 6180, %o
SLAP ANU-A1 GISP ANU-LP ANU-C1 ANU-HP ANU-P1 VSMOW ANU-OW Beet sucrose ANU sucrose
-55.50 -32.10 -24.82 -11.94 -6.10 -5.54 -4.60 0.00 0.50 30.81 36.40
-54.27 -31.09 -24.04 -11.30 -5.32 -4.91 -3.67 0.75 1.22 31.78 37.23
s.d.
n
0.19 0.14 0.17 0.07 0.12 0.16 0.09 0.09 0.16 0.12 0.17
5 6 8 4 7 6 8 6 6 8 7
To assess the feasibility of the technique, cellulose (the main organic constituent in the leaf) was doped with waters of wideranging 6180 to simulate fresh leaf samples. 180 content of the doping water calculated from the direct analyses of the cellulose-water mixtures showed a very good correlation with that directly measured from the pyrolysis system (Figure 21.3). A slope very close to unity (drift of < 0.7%) was also obtained and any offset observed was not more than 0.2%0. There was no significant difference in the 6180 of doping water calculated from y and OF which were determined either from the TCD output or the weight of individual component added into the mixture. Thus it is analytically reliable to obtain sample oxygen elemental composition from the TCD output of the gas chromatograph in the same acquisition as the 180 analysis. For cellulose-water mixtures of varying water content (Figure 21.4), their measured ~5180 values agreed reasonably well (drift of < 1%, R2 > 0.98) with that predicted based on the known 6180 and mass contributions of the individual components in the mixture. The technique could therefore be applied to samples of wide-ranging water content.
Oxygen Isotope Analysis of Plant Water Without Extraction Procedure
479
Figure 21.3 - h 1 8 O v s M o w of doping water calculated from the pyrolysis of cellulose-water mixtures, versus measured values from direct pyrolysis of the doping water. In calculating 6180 of doping water using equation [21.4], OF was determined either from the TCD output (O) or from the weight of individual component added into the mixture (D) with the aid of equation [21.2]. Using data from the TCD output gives a linear regression of 618Ocalculated = 1.007 hl8Omeasured + 0.199 (R2 = 0.9998) while that from component weighing gives 618Ocalculated = 1.0024 h18Omeasured + 0.089 (R2 = 0.9999). The dashed line represents a 1:1 relationship.
~180 determination of leaf water by fresh leaf pyrolysis was compared with the conventional extraction method of azeotropic distillation (Table 21.1). As expected, azeotrope extracted leaf water generally showed a lower 6180 value than that obtained from pyrolysis of the same leaf. This discrepancy arises from the different sample size needed in the two methods which affects sample homogeneity. For azeotropic distillation, a larger pool of plant material is needed to produce adequate amounts of water for easy sample handling and storage. This inevitably includes nonenriched water from the secondary vein network, the contribution of which is difficult to quantify, and thus obscures the actual isotopic signal of mesophyll water at the leaf evaporating site. Attempts to avoid minor veins in sampling would require the pooling of multiple small leaf discs, a process that could subject the leaf water to evaporative enrichment due to the large exposed cut surface. On the other hand, the fresh leaf pyrolysis technique needs only very small quantities of plant material (one leaf disc with 3 mm diameter) containing sub-microlitre amounts of water and therefore uncertainties such as inclusion of unfractionated vein water will no longer be an issue in leaf water measurements. Indeed, it was observed that whenever the leaf punch hit a fine vein during sampling, ~180 of leaf water could be 2 - 6%0 lower, depending on vein size, and these results were excluded from the data analysis in Table 21.1. How-
480
Chapter 21 - K.S. Gan, S.C. Wong & G.D. Farquhar
Figure 21.4 - The relationship between predicted and directly measured f)18OvsMow of mixtures of cellulose and ANU-OW with varying water content shown as % next to the plot points. Prediction was based on equations [21.1-3], with y and OF determined from weighing the individual component added into the mixture, and cellulose and ANU-OW having measured 618OVSMOW of 30.1%o and 0.7%o respectively. The dashed line represents a I :1 relationship.
Table 21.1 - Comparison of ~)18OLwof cotton leaf water extracted by azeotropic distillation, and indirectly determined from fresh leaf pyrolysis. For the leaf pyrolysis method, oxygen fractions (OF and OD) and 6180 of the fresh and dry leaves are also presented, together with water fraction in the fresh leaf, y. Subscripts LW, F and D refer to leaf water, fresh leaf and dry leaf respectively. All 6180 are expressed relative to VSMOW. The values shown are mean + standard error. Number of replicates is 2 for azeotropic distillation and 6 - 8 for the leaf pyrolysis method. Sampling period *
Azeotropic distillation
618OLw, %0 0 4 0 0 - 0630 14.0 + 0.6 1030 - 1200 14.9 + 0.4 1300- 1500 15.5 + 1.3
Leaf pyrolysis
OF ~)18OD' %o OD ~)18OF' %0 18.3 + 2.1 0.764 + 0.007 31.4 + 1.3 0.334 + 0.007 19.8 + 2.3 0.777 + 0.006 34.7 + 1.4 0.345 + 0.002 21.0 + 2.9 0.763 + 0.024 35.6 + 1.3 0.342 + 0.001
* Australian Eastern Standard Time. Dawn is at around 0630 h. For leaf pyrolysis method, 618OLw is calculated from equation [21.4].
y 618OLw,%o # 0.776 + 0.014 16.9 + 2.5 0.793 + 0.011 18.0 + 2.7 0.769 + 0.043 19.2 + 3.3
OxygenIsotopeAnalysisof Plant WaterWithoutExtractionProcedure
481
ever, results from leaf punch sampling generally showed larger standard error, as isotopic composition of leaf water is non-homogeneous over the leaf surface. It was observed that mesophyll water at and near the leaf margin is significantly more enriched (up to 6%o for g180, results not shown) than other parts of the leaf. Therefore, only leaf discs sampled from the inner leaf section were included in the data analysis presented in Table 21.1. Spatial variation of leaf water ~180 and g2H values has previously been noted in several studies (Bariac et al., 1994; Helliker & Ehleringer, 2000; Luo & Sternberg, 1992; Wang & Yakir, 1995). The small sample size provided by the fresh leaf pyrolysis technique will also be an asset in conducting further detailed studies of leaf water enrichment as a function of leaf length or distance from This will facilitate our understanding of how progressive enrichment comes about and assist in applying an appropriate leaf water model gross photosynthetic productivity (Farquhar et al., 1993; Yakir & Wang, Dole effect (Bender et al., 1985; Bender et al., 1994).
the mid-vein. of leaf water to studies of 1996) and the
The pyrolysis technique described here could potentially be applied to 62H analysis of plant water as well, with the same pyrolysis set-up as for ~180 analysis. Also, the technique need not be confined to analyses of leaf water but could also be used for a variety of biological fluids and foodstuffs. Some examples include the analysis of phloem water from phloem sap, fruit juice water and even possibly wine. The pyrolysis technique described by Farquhar et al. (1997) and used in this study, is basically a powerful and robust approach that could measure the isotopic content of water or any volatile present in a homogeneous solution or solid hydrous substance.
Acknowledgments We thank Peter Groeneveld and Jim Neale for excellent assistance in designing and constructing the leaf punch, Michael Bird and Mike Gagan, Research School of Earth Science (ANU) for supplying internal water standards. We also wish to acknowledge the helpful comments of Kinga R6v6sz. The support of Micromass U.K. Ltd is appreciated.
Handbook of Stable Isotope Analytical Techniques, Volume 1 P.A. de Groot (Editor) 9 2004 Elsevier B.V. All fights reserved.
CHAPTER 22 Oxygen Isotope Analysis of Phosphate Christophe L6cuyer Laboratoire CNRS UMER 5125, "Pal6oenvironnements & Pal6obiosph6re", Batiment <
, Campus de la Doua, Universit6 Claude Bernard Lyon 1, 27-43 Boulevard du 11 Novembre 1918, 69622 Villeurbanne cedex, France e-mail: [email protected]
Abstract
During the past decade, considerable progress has been made in measuring oxygen isotope ratios of phosphates. Wet chemistry procedures based on the isolation of phosphate radicals using ion-exchange resins and precipitation as stable crystals of silver phosphate have ensured a comparison of oxygen isotope ratios between biogenic phosphates characterized by oxygen-bearing compounds (CO32-, SO42- , O H - ) present in varying proportions. Extraction of oxygen from Ag3PO4 as CO2 or CO can be performed either by fluorination (BrF5) or by graphite reduction. In both cases, high-precision oxygen isotope measurements can be made (~ + 0.2%o). These allow for high-sensitivity estimates of paleotemperature, paleohumidity and isotopic compositions of water. Oxygen can also be extracted from untreated biogenic phosphates by in situ laser analysis. The result is a high-spatial resolution (100 ~m) isotopic analysis with a good internal reproducibility (~ 0.5%0; lo) despite the analysis of bulk samples.
22.1 Introduction
Phosphorus is a critical nutrient for biological processes, occuring mainly as a tetrahedral PO43- chemical group present in minerals, organic matter and natural waters. Oxygen isotope ratios of PO43- from these materials have been used widely in the study of paleoenvironments for several decades, because the oxygen bound to phosphorus is very resistant to isotopic exchange by inorganic processes at low temperatures (Tudge, 1960; Kolodny et al., 1983; Shemesh et al., 1988; L6cuyer et al., 1999). Kolodny et al. (1983) showed that temperature and the isotopic composition of ambient water are recorded in the isotopic composition of phosphate from fish bones and teeth. The initial relation between the isotopic compositions expressed in the conventional 6-notation in parts per thousand (%o) as 618Op values of phosphate and 618Ow values of water, and temperature (T~ proposed by Longinelli & Nuti (1973a) were confirmed by Kolodny et al. (1983) who provided the following equation: T~ - 111.4 - 4.3 (618Op - 618Ow)
[22.1]
OxygenIsotopeAnalysisof Phosphate
483
Oxygen isotope fractionations in the phosphate-water system are characterized by some interesting properties including 1) an apparent better resistance of strongly mineralized phosphatic tissues, such as tooth enamel, compared to calcium carbonates that display higher solubilities in natural waters than calcium phosphates, 2) the existence of a unique equation that describes temperature-dependent oxygen isotope fractionations between ectotherm fauna and ambient waters (Longinelli & Nuti, 1973a; Kolodny et al., 1983; L6cuyer et al., 1996a; Blake et al., 1997), and 3) the possibility to combine analyses of coexisting ectotherms (e.g. fishes) and homeotherms (e.g. mammals) may yield a unique solution for both temperature and isotopic composition of the ambient water (Longinelli, 1984:; L6cuyer et al., 1996b). The usefulness of oxygen isotope compositions of PO43- from biogenic phosphates to determine either changes in temperatures of past oceans (e.g. Longinelli & Nuti ,1973b; Kolodny et al., 1983; Luz et al., 1984; Karhu & Epstein, 1986; Kolodny & Raab, 1988; L6cuyer et al., 1993; Anderson et al., 1994; Picard et al., 1998; Vennemann & Hegner, 1998) or isotopic compositions of waters ingested by terrestrial mammals has been well documented (e.g. Longinelli, 1984; Luz et al., 1984:; Luz & Kolodny, 1985; D'Angela & Longinelli, 1990; Ayliffe et al., 1992; Fricke et al., 1995; 1998a,b; Genoni et al., 1998). In any case the selection of biogenic phosphates must be made very carefully. Indeed, numerous studies performed on phosphatic remains from terrestrial and marine sedimentary deposits have shown post-depositional isotopic modifications due to exchange reactions with surrounding waters (e.g. Longinelli, 1966; McArthur et al., 1980; McArthur & Herczeg, 1990; Shemesh, 1990; Sanchez Chillon et al., 1994; Kolodny et al., 1996; Sharp et al., 2000). The loss of original oxygen isotope compositions of phosphates does not appear to be really age-dependent but more related to the post-depositional environment and taphonomy (Shemesh, 1990; Kolodny & Luz, 1992). The goal of this paper is to review the progress made in the analysis of oxygen isotope ratios of phosphates. Over the last decade, most of the efforts have focused on reducing the time and increasing the reproducibility of isotopic analyses with the precipitation of silver phosphate that progressively replaced bismuth phosphate as the final stage of the wet chemical procedure. Progressive replacement of the hazardous bromine pentafluoride (Crowson et al., 1991) by graphite (O'Neil et al., 1994) as the reducing agent to extract oxygen made phosphate analysis available to a larger scientific community. Finally, in situ laser-based analytical techniques (e.g. Sharp, 1992; Sharp & Cerling, 1996; 1998) opened the field of high spatial resolution without loosing too much precision, leading for example to the measurement of isotope ratios of individual conodonts (Wenzel et al., 2000). All these technical developments now make possible the acquisition of large series of measurements that constitute a prerequisite for paleoclimatic and paleoenvironmental reconstitutions.
22.2 Wet chemistry: unscrambling the oxygen isotope signal in apatites Most studies of oxygen isotope ratios in apatites start with isolating PO43- using acid dissolution and anion-exchange resin. The reason is the complexity and diversity of chemical compositions displayed by sedimentary phosphate minerals. The pres-
484
Chapter 22 - C. L6cuyer
ence of oxygen-bearing compounds such as carbonate, sulfate, organic matter, and hydroxyl radicals in various amounts in the structure of apatites makes it difficult to compare 6180 values among samples or to apply empirical and experimental fractionation equations for thermometric applications. Moreover, oxygen from carbonate and hydroxyl radicals are more sensitive to isotopic exchange than PO4 3- during diagenetic processes. The possible effect of CO3 2- contribution upon the bulk oxygen isotope composition of apatite may be easily evaluated by combining a mass balance equation [22.2] with the oxygen isotope fractionation between carbonate and phosphate [22.3]"
6180(apatite) = [1 - X(c03)] 981 80(PO4 ) 4- X(CO3 ) 96 180( CO3)
[22.2]
A - 818O(CO3) - 6180(PO4)
[22.3]
leading to the following relationship:
6180(apatite) = ~180(PO4 ) + X(co3 ) 9A
[22.4]
from which has been derived a set of curves that illustrate the difference between the 5180 values of the bulk apatite and its PO43- radicals as a function of the phosphatecarbonate fractionation and the amount of CO32- in the crystal lattice (Figure 22.1). For example, considering a fish tooth characterized by a difference in 6180 values between carbonate and phosphate of + 8%o and a carbonate amount of 6 wt%, the bulk apatite would have a h180 value of 0.48%o higher than that of PO4 3- alone. This offset of the bulk apatite oxygen results in the calculation of paleotemperatures systematically too low by about 2~ 22.2.1 The BiPO4 method
The first method used to isolate PO43- from natural apatites was described by Tudge (1960) who first explored many aspects of apatite. Tudge (1960) emphasized the removal or chloride and organic matter from apatite samples through the various steps of the wet chemistry, the slow precipitation of BiPO4 to avoid the inclusion of bismuth nitrates, and a complete drying of BiPO4 at 140~ during 3 hours. Once prepared, BiPO4 is fluorinated under vacuum. This method was refined by Kolodny et al. (1983) and used extensively for many years. Considering this method in more detail, samples containing organic matter are crushed and soaked in hydrogen peroxide then dried. Phosphate powders are dissolved at 60 ~ to 70~ in a 250 ml Erlenmeyer flask with 15 ml of 10 M HNO3. The solution is filtered to remove any insoluble residues. The remaining fraction of organic carbon is oxidized by using a few drops of 0.3 M KMnO4 to the solution heated again in the range 60 ~ - 70~ A black precipitate of MnO2 must form and the sample
Oxygen Isotope Analysis of Phosphate
485
Figure 22.1 - Oxygen isotope shifts that result from the bulk analysis of natural apatites as a function of variable carbonate content. Computed curves represent the difference between the ~180 values of the bulk apatite and its PO43- radicals as a function of the phosphate-carbonate fractionation and the amount of CO3 2- in the crystal lattice.
remains slightly heated overnight. The excess MnO2 is dissolved with 1M NaNO2. The PO43- is first precipitated as an ammonium phosphomolybdate ((NH4)2HPO4 (12MOO3) by combining ammonium molybdate with ammonium nitrate added to the phosphate solution. The yellow precipitate is filtered, washed with ammonium nitrate and dissolved again in a 250 ml Erlenmeyer flask with ammonium citrate reagent. The next step of purification is the precipitation of magnesium ammonium phosphate (MgNH4(PO4)(6H20)) obtained with 25 ml of magnesia reagent (50 g of MgC12(6H20) with 100 g of NH4C1 dissolved in 500 ml of H20) added to a I 91 ammonia solution. The precipitate is washed with a large amount of the same solution and dissolved with 25 ml of I 94 HC1. The magnesium ammonium phosphate is reprecipitated, filtered and washed with 1 920 ammonia solution until complete removal of chloride ions. The precipitate is dissolved in a 250 ml Erlenmeyer flask with 6.5 ml of 5.4 M HNO3 before precipitation of BiPO4 by addition of a solution of 2M HNO3 and 0.2 M Bi(NO3)3. Crystal growth is facilitated by a thermostatic bath whose temperature is set at 80~ Crystals of bismuth phosphate are washed with deionized water, dried at 105~ and stored in a desiccator.
486
Chapter 22 - C. L6cuyer
22.2.2 The Ag3PO4 method An alternative to the precipitation of BiPO4 is the precipitation of Ag3PO4. The principle of the Ag3PO4 method is to isolate the PO43- groups from sedimentary apatites as solid crystals of Ag3PO4 (Crowson et al., 1991). The main advantage of this method is that it is less time-consuming (3 days) compared to the
Ag3PO4 method HF dissolution 2M 12 - 24 h
Neutralisation with KOH 2M
Complexation of PO43on an anion-exchange resin
Figure 22.2- Scanning electron microscope (SEM) image of a chemically-precipitated crystal of silver phosphate. The scale bar is 60 tim BiPO4 method (6 days), and it produces Ag3PO4 crystals (Figure 22.2) which are practically not hygroscopic. They may thus be easily degassed and dried in a vacuum line.
22.2.2.1 The protocol of wet chemistry Elution of PO43with NH4NO3
Precipitation of Ag3PO4 crystals at 70~ using nitrate ammoniacal silver nitrate Figure 22.3 - Flow chart summarizing the chemical protocol to isolatePO43- radicals from apatites.
The protocol presented below was directly derived from the original method published by Crowson (1991) and slighltly modified by L6cuyer et al. (1993). It is summarized in Figure 22.3. The samples are washed in distilled water and air dried at 50~ After dissolution of 15 to 30 mg of powdered sample in 2M HF at 25~ for 24 hours, the CaF2 that precipitates is separated from the phosphate solution by centrifugation. The CaF2 precipitate is rinsed three times using double deionized water (DDW) and the rinse water added to the phosphate solution which is finally neutralized with a 2M KOH solution. Before use, the AmberliteWM-IRA-400(OH) or Amberjet TM anion-exchange resin is flushed with several liters of DDW to eliminate chloride and
Oxygen Isotope Analysis of Phosphate
487
avoid precipitation of AgC1. A 2 ml aliquot of cleaned resin is then added to the neutralized solution in a polypropylene tube. The tubes are placed on a shaker table for 12 hours to promote the ion exchange process. Vanadium molybdate color indicator is used on one test sample to check that all the phosphate ions are adsorbed on the resin. Excess solution is discarded and the resin is washed again three times with DDW to remove the last traces of ionic contaminants. To elute the phosphate ions quantitatively from the resin, 2 5 30 ml of 0.5M NH4NO3 were added to bring the pH of the solution to 7.5 8.5, and the tubes are gently shaken for about 5 hours. The resin and the phosphate were then separated on a 63 ~m stainless steel sieve and the resin rinsed with 15 ml of DDW to transfer possible remaining phosphate ions to the solution. Silver phosphate was then precipitated from the solution following the method of Firsching (1961). The solution was placed in a 250 ml Erlenmeyer flask and about 1 ml of concentrated NH4OH were added to raise the pH to 9 - 10; 15 ml of ammoniacal AgNO3 solution were then added to the flask. Upon heating this solution to 70~ in a thermostated bath, millimetersize, yellowish crystals of Ag3PO4 were quantitatively precipitated. The volume of solution was maintained constant durFigure 22.4 - Frequency histograms of yields obtained during the ing precipitation by regu- precipitation of silver phosphate after acid digestion of the NIST lar addition of a few standard NBS120c (natural phosphorite from Florida). a) routine drops of DDW. The crys- chemistry performed with the ion-exchange resin Amberlite TM IRA-400(OH) and b) routine chemistry using the AmberjetTM iontals of silver phosphate exchange resin. were then collected on a
488
Chapter 22 - C. L6cuyer
millipore filter, washed three times with DDW and air dried at 50~ Repeated analyses of the NBS120c phosphorite have been made following this wet chemical procedure closely. We emphasize that the NBS120c phosphorite is an international standard for its phosphorus content (P205 - 33.34 wt%) and not for its oxygen isotope ratio. Inter-laboratory comparisons of measured 6180 values are given at the end of section 22.3.2. The data base is used in a further section of the paper to discuss 1) the chemical yields depending on the anion-exchange resin used, 2) the influence of chemical yields upon the 6180 reproducibility of the NBS120c standard and 3) the comparison of isotopic compositions between the fluorination and graphite methods.
22.2.2.2 Phosphate yields and the measurement of 6180 values High chemical yields of Ag3PO4 are a prerequisite to obtain reproducible oxygen isotope ratios. A compilation of more than one hundred analyses of the standard NBS120c (Florida phosphorite of Miocene age) has been performed to examine the effects of sample size and phosphate yields upon the reproducibility of oxygen isotope ratios. Poor yields are mainly caused by either too short elution time or-massive precipitation of very small (< 10 mm) silver phosphate crystals that remained on the glass walls of the Erlenmeyer flasks and trapped in the filter. Yields obtained using the resin Amberlite TM IRA-400(OH) may be accidentally as low as 62% compared to the average value of 78% (n = 54) for a mean sample weight of 35 mg (Figure 22.4a). The best yields obtained with this resin have been 87%. The best recovery of has been recently obtained with the strongly anionic resin Amberjet TM which provided an average phosphate yield of 86% for a mean sample weight of 30 mg (n - 20; Figure 22.4b). The reasons for the observed scattered values of phosphate yields are not yet clearly identified. It certainly depends on the proportion of fine powder of precipitated silver phosphate that are washed away during the filtration step. It could also result from the function of the resin itself. Chemical yields lower than 70% are responsible for a significant scattering of ~)180 values (SMOW) ranging from 21.3%o to 22.3%o (Figure 22.5). The choice of the resin has no effect on both average and standard deviation values of NBS120c that are of 21.7%o and + 0.17%o, respectively (Figure 22.5).
22.2.2.3 Problems in producing Ag3P04 from biogenic phosphates The most common problems encountered include: 1) solutions turn yellow to red during neutralization with KOH. This typically indicates the presence of exotic metallic elements such as Fe, Mn, or A1 that precipitate as hydroxides as pH increases. 6180 values of such samples must be interpreted very carefully since they commonly result in anomalous values. 2) reproducible low chemical yields for duplicated samples likely reflect a phosphate loss during diagenetic processes. The phosphorus component may be replaced by carbonates, silica or metallic oxides.
Oxygen Isotope Analysis of Phosphate
489
Figure 22.5 - Variations of the 8180 value of Ag3PO4 derived from the standard NBS120c as a function of the percentage of silver phosphate recovery (chemical yield). Note the absence of correlation for the data obtained using both types of ion-exchange resins
3) instead of well-crystallized yellowish silver phosphate crystals, precipitation of PO43- in the presence of silver nitrate leads to greenish-brownish platy crystals. This reflects contamination by organic molecules that potentially modify both carbon and oxygen isotope compositions of CO2 produced by reduction of silver phosphate with graphite. The removal of organic matter may be achieved by the use of: 1) a powerful oxidizing reagent during the wet chemistry such as the KMnO4 (see the BiPO4 method above and Schwarcz et al., 1985) 2) carbon oxidation using a cold plasma furnace filled with an oxygen atmosphere 3) hydrazine to remove proteins or a strongly basic resin such as Amberlite TM IRA900 4) pre-treatment of samples with NaOC1 and NaOH (Stephan, 2000) The use of hydrogen peroxide is relatively inefficient to remove organic matter from phosphates and high-temperature ashing may significantly modify oxygen isotope compositions during the formation of pyrophosphates. Unfortunately, no comprehensive and detailed investigations of the best techniques adapted to the removal or organic matter from biogenic phosphates have been published yet. It is thus obvi-
490
Chapter 22 - C. L6cuyer
ous that future efforts should focus on this aspect that is critical for obtaining highquality data on fresh phosphatic tissues using the silver phosphate method.
22.2.2.4 Producing BiP04 and Ag3P04 from dissolved and organic P compounds The chemical methods presented above have been initially designed for the oxygen isotope analysis of phosphate minerals. However, oxygen isotope measurements can be carried out from dissolved phosphates or from organic phosphate compounds following various chemical procedures (Longinelli et al., 1976; Paytan et al., 1990 ; Colman et al., 2000). Within abstracts, Paytan et al. (1990) mention the use of chitosan ammonium molybdate beads to extract dissolved phosphate. More recently, Colman et al. (2000) indicate a protocol involving a ~ sequence of magnesium hydroxide precipitations, ion exchange resin treatments, and evaporative reductions ~, to concentrate dissolved inorganic phosphate from a water sample. The basic principles of the method presented by Longinelli et al. (1976) have been derived from techniques developed by Lal et al. (1964) to extract trace elements from seawater. Dissolved phosphate from seawater (about 2 gmol.l-1) is concentrated using acrylic fibers soaked with iron hydroxide that mimics an ion-exchange resin. Pieces of acrylic fibers of about 100 to 150 g each are treated with 10 M HNO3 for about two hours at room temperature. Most of the iron hydroxide is dissolved along with the adsorbed chemicals while the fibers are mechanically crushed. The solution is filtered and heated to boiling for several hours to oxidize a large portion of the dissolved part of the acrylic matter. Hydrogen peroxide (about 500 ml) is repeatedly used to ensure oxidation of the organic matter. The addition of 100 ml of 4 M NH4NO3 promotes the flocculation of the remaining acrylic matter. After filtration, addition of KMnO4 eliminates any residual traces of organic matter then the solution follows the BiPO4 method presented in section 22.2.1. A similar procedure has been used by the authors to separate phosphate compounds from organic matter such as fish flesh. The best results seem to be obtained when the organic matter is dissolved with 10 M HNO3. 22.3 Gas extraction 22.3.1 The fluorination and bromination methods 22.3.1.1 The use of BrF3 or BrF5
Once phosphate salt such as BiPO4 or Ag3PO4 is produced, oxygen is liberated for isotopic analysis. This may be achieved by several different ways. Tudge (1960) presented the technique of oxygen extraction from bismuth phosphate by reaction with bromine trifluoride according to the following reaction: BiPO4 + 8/3BrF3 ~ BiF3 + PF5 + 4/3Br2 + 202
[22.5]
Such a technique has been extensively used by Longinelli and co-workers (e.g. Longinelli, 1965; 1966; Longinelli & Nuti, 1968a,b; 1973a, b) whilst Kolodny et al. (1983) preferred to use bromine pentafluoride for the decomposition of bismuth phosphate.
Oxygen IsotopeAnalysisof Phosphate
491
Tudge (1960) and Kolodny et al. (1983) heated the hygroscopic BiPO4 in vacuum at about 130~ for 3 hours prior to fluorination. Karhu & Epstein (1986), however, suggested that this temperature is not high enough to achieve dehydration of BiPO4 crystals. They proposed to heat samples to 420~ Above this temperature, BiPO4 is converted from a monazite-type structure to a high-temperature structure (MooneySlater, 1962) that is less sensitive to hydration. Shemesh et al. (1988) compared standard ~180 values using both techniques and observed that BiPO4 heated to 420~ is in average 180-enriched by 0.5%o relatively to BiPO4 degassed at 130~ Shemesh et al. (1988) finally proposed that BiPO4 should be dehydrated at 130~ then immediately loaded into reaction vessels before fluorination. Reactions of silver phosphate with BrF5 have been presented by Crowson et al. (1991) then by L6cuyer et al. (1993) who provided repeated analyses of phosphate standard NBS120c. Crowson et al. (1991) found consistent 6180 values of 21.33 + 0.05%0 for NBS120c when Ag3PO4 reacted with BrF5 at temperatures above 425~ These authors used between 10 and 15 mg of silver phosphate in presence of 6 91 to 10 91 excess of BrF5 for a 22 h period. The procedure given by L6cuyer et al. (1993) recommends the use of aliquots of between 12 and 15 mg of Ag3PO4 crystals weighed into nickel reaction vessels. Samples are degassed 4 hours at room temperature and an additional 4 hours at 150 - 200~ This degassing procedure assures that atmospheric water is efficiently desorbed from the samples. A 5/1 mole excess of BrF5 is reacted with the samples at 600~ for 12 hours. Oxygen yields of 100% were obtained routinely at the University of Rennes with the NBS120c Florida phosphate standard. A few experiments were done under various experimental conditions: 1) degassing up to 400~ for 1 hour, and 2) fluorination at 500 ~ 550 ~ 600 ~ and 650~ for between 10 and 20 hours. No difference in the isotopic analyses were detected through various combinations of these conditions. Twenty measurements of NBS120c using BrF5 lead to an average 6180 value of 21.64 + 0.18%o (Figure 22.6a). 22.3.1.2 The Br2 method Stuart-Williams et al. (1995) have shown that it is also possible to react silver phosphate with Br2 instead of BrF5 for extracting oxygen that is further converted into CO2 with a heated carbon rod. Samples from 40 mg to 50 mg of silver phosphate are placed into quartz tubes and reacted with Br2 at temperatures of 550~ according to the following reaction of bromination:
2Ag3PO4 + Br2 ~ Ag4P207 + 1/202 + 2AgBr
[22.6]
The yield of this reaction of bromination is quite low compared to the quantitative fluorination reaction, about 17.5% but leads to high-precision measurements (~_ + 0.3%0) only if the temperature of the reaction may be carefully controlled (+ 1~ to be able to apply adequate fractionation corrections. Stuart-Williams et al. (1995) used a copper cored furnace built to ensure uniform temperatures within the reaction vessel and less than 3~ of variation in temperature. The authors emphasized some advantages over the BrF5 method such as reduced costs for building an extraction line, time
492 Figure 22.6- Frequency histograms of the 6180 values of Ag3PO4 derived from the standard NBS120c obtained using a) bromine pentafluoride reagent and resin Amberlite TM IRA-400(OH), b) graphite reagent and resin Amberlite TM IRA-400(OH), and c) graphite reagent and resin AmberjetTM ion-exchange resin. Note the similarity of computed statistics. Isotopic ratios obtained with the fluorination method have been measured at the University of Rennes (France) with a VGTM Sira 10 mass spectrometer; those obtained with the graphite method have been measured at the Ecole Normale Sup6rieure de Lyon (France) with a VGTM Prism mass spectrometer. savings for oxygen extraction, and less hazardous conditions.
22.3.2 The graphite method The m e a s u r e m e n t of 180/ 160 ratios from silver phosphates has been simplified by O'Neil et al. (1994) who set up a rapid and precise m e t h o d of CO2 production using graphite reagent. Samples of Ag3PO4 are mixed with pure graphite p o w d e r in proportions of about 0.3 m g of C to 22 m g of Ag3PO4. The mixture is loaded into silica tubes that were previously heated for a few hours at 800~ in a furnace to remove potential organic contaminants. O'Neil et al. (1994) emphasized that graphite must be intimately mixed with the silver phosphate and that excess graphite may cause CO formation responsible for 6180 shifts of the measured CO2 as m u c h as several tenths per mil. The sil-
Chapter 22 - C. L4cuyer
Oxygen Isotope Analysis of Phosphate
493
ica tubes are linked to a vacuum line and heated at 550~ and pumped for a few minutes to remove adsorbed water, then the tubes are sealed with a torch. The sealed silica tubes are placed in a horizontal tube furnace and the reaction takes place at 1200~ during 3 minutes. Tubes are withdrawn and immediately quenched in water to avoid back reactions and further isotopic exchange during slow cooling. CO2 is the major gas (along with minor amounts of CO) present in the silica tube that is easily connected to the inlet system of an isotope ratio mass spectrometer then cracked to liberate the gas towards the source. The quality of the reaction is also monitored with the 613C value of the CO2 that should not differ by more than a few tenths per mil from that of the graphite reagent. Unfortunately, the oxygen yield is only 25% but provides generally enough gas for a high-precision oxygen isotope analysis. The reproducibility of the technique developed by O'Neil et al. (1994) is equal to or less than + 0.2%o in most cases and the mean 6180 value found for NBS120c is 21.8%o. It is noteworthy that ~180 values inferred from silver phosphates are higher than those obtained by the BiPO4 method. Comparative measurements suggest an average difference of 1.3 + 0.8%o (see Table 3 in O'Neil et al., 1994). L6cuyer et al. (1998) slightly modified the O'Neil et al. (1994) method to save time by performing the Ag3PO4-C reaction directly on a vacuum extraction line. Aliquots of between 12 and 15 mg of Ag3PO4 crystals are mixed with 0.6 to 0.8 mg of pure graphite, weighed into tin reaction capsules and loaded into quartz tubes degassed for 30 minutes at 110~ in vacuum. The sample is heated to 1100~ for 1 minute to promote the redox reaction catalyzed by tin. Tin capsules also allow silver phosphate crystals and graphite powder to remain mixed together during the reaction. The CO2 produced during this reaction is directly trapped in liquid nitrogen avoiding isotopic exchange with silica at high temperatures. High-quality redox reactions of C conversion into CO2 in the quartz tubes are reflected by a normal distribution of the measured 613C values of the CO2 with a standard deviation less than 0.1%o (Figure 22.7). Applying this method to 40 samples of Ag3PO4 derived from the NBS120c standard, a mean 6180 value of 21.75 + 0.18%o (Figure 22.6b) was obtained using the resin Amberlite TM IRA-400(OH) during the chemical procedure. It is interesting to note that indistinguishable results have been obtained using the resin Amberjet TM with a mean 6180 value of 21.68 + 0.16%o (n - 21; Figure 22.6c). This m e a n 6180 value for NBS120c is consistent with other statistics published by O'Neil et al. (1994) and Fricke et al. (1998a) who recommend a 6180 value of 21.8 + 0.25%o as well as the value of 21.5 + 0.2%o given by Vennemann & Hegner (1998). With the three above methods, the produced CO2 is analyzed using Gas Isotope Ratio Mass Spectrometry (GIRMS) techniques. Using either the BrF5 or C method, reproducibilities ~ 0.2%o are routinely obtained for gas amounts of a few tens of micromoles.
494
Chapter 22 - C. L6cuyer Figure 22.7- Frequency histogram of the 613C values of the analyzed CO2 obtained during the reduction of silver phosphate (NBS120c source) with a standard graphite (mean ~13C value of C = 23.6 + 0.05%o with the method of CuO oxidation). Note the standard deviation less than 0.1%o (lo at 95% confidence). Measurements have been performed using a VG T M Prism mass spectrometer located at the Ecole Normale Sup6rieure de Lyon, France.
22.3.3 Mass spectrometry: Gas Isotope Ratio Mass Spectrometry (GIRMS) versus Negative Thermal Ionization Mass Spectrometry (NTIMS) Both GIRMS and NTIMS analytical techniques are now able to provide high-precision measurements for small phosphate samples. Micromass TM recently developed a new high-temperature (1300~ pyrolysis technique for the oxygen isotope analysis of phosphates. The application is described in a technical note published by J. Morrison (Application Brief AB10, August 1999) and more extensively presented in Kornexl et al. (1999a). The Elemental Analyser is interfaced in continuous flow mode to a Micromass T M Isoprime isotope ratio mass spectrometer. The pyrolysis reactor is packed with glassy carbon and 30% nickelised carbon. The reactor temperature is set to 1260~ Phosphate samples such as Ag3PO4 and KH2PO4 that contain about 80 ~g of oxygen are weighed into tin cups. The samples are loaded onto a carousel then sequentially dropped into the pyrolysis reactor producing both N2 and CO. The gases are separated through a gas Chromatographic Column. The oxygen isotope ratios are obtained by integrating the ion beam areas at masses 30 (C180) and 28 (C160). O x y g e n yields are above 90% and reproducibility is better than _+0.5%0. Measurement of oxygen isotope ratios in phosphates has also been developed using NTIMS techniques by Holmden et al. (1997). A thin layer of colloidal Pt powder is loaded on a Pt filament. An aliquot of ammoniacal silver phosphate solution is put onto the Pt powder under a 0.2 A current. Crystals of Ag3PO4 precipitate over the Pt powder as the solution evaporates. To reduce the work function of the filament, a uniform BaC12 layer is deposited over the Ag3PO4 precipitate. The amount of BaC12 added corresponds to a I 91 molar ratio of Ba 9PO4, thus leading to intense PO3- ion beams without C1- or Br- production. The ionization efficiency larger than 10% provides an exceptional sensitivity relatively to GIRMS. This corresponds to the analysis of less than 20 nmoles of CO2 with an internal reproducibility of + 0.5%o.
Oxygen Isotope Analysis of Phosphate
495
22.3.4 Laser-based analysis of silver phosphate Wenzel et al. (2000) opted for the laser ablation of small amounts (0.1 to 0.5 rag) of Ag3PO4-C mixtures. Even though there is an apparent constant offset of- 2.5 _+0.5%o relative to specified isotopic compositions, this method allows for a precision of _+0.2 (lo) comparable to conventional techniques. The observed constant offset is attributed by the authors to a simultaneous formation of CO and CO2 during the laser heating of the mixture of silver phosphate and graphite. This method offers the possibility to obtain high-precision oxygen isotope data on a few conodont specimens only, reducing the sample size to less than I rag. 22.4 In situ laser-based analysis of apatites Although oxygen from all O-bearing compounds present in apatites are released using laser-based analysis, several methods have been described to analyze small samples. Laser-based systems for stable isotope analysis have been extensively developed in the last decade (Dickson, 1991; Powell & Kyser, 1991; Sharp, 1992). Sharp & Cerling (1996) presented a laser GC(Gas Chromatography)-IRMS technique for in situ oxygen isotope analysis of natural phosphates. In summary, the beam generated by a CO2 laser passes through a ZnSe window and is focused on the surface of the sample placed in a Pyrex TM reaction chamber (see Fig. I in Sharp & Cerling, 1996). The carbon dioxide gas released during the decomposition of the sample is carried by a helium flux and separated from other gases through a chromatographic column. The purified CO2 is then introduced into a mass spectrometer via a split interface. Sharp & Cerling (1996) used a CO2 laser with a power of 20 W and a focal length of 12.7 cm that permits spot sizes down to 100 ~m. With this power, a single laser pulse of 10 ms generates enough gas for a high-precision isotopic analysis using a GC column and a He flow rate between 30 and 40 ml.min-1. This technique presents two major advantages" a single analysis is complete in less than three minutes and a high spatial resolution. Isotopic ratios are measured using a dynamically pumped mass spectrometer designed for small CO2 samples issued from a gas chromatograph following the method given by Matthews & Hayes (1978). Sharp & Cerling (1996) argue that the 6180 values of tooth enamel obtained from this method of laser heating are equal to those from the PO43- ions. They argue that during the very high-temperature reaction, the oxygen fractionation between CO3 2- and PO4 3- should be very small, and consequently the oxygen contribution from the carbonate does not affect the i5180 value of the phosphate. However, Wenzel et al. (2000) observed that laser heating of chemically untreated phosphates leads to poorly reproducible oxygen isotope measurements. These authors suspect that oxygen isotope exchange takes place between CO2 and some water derived from the hydroxyl site in the apatite before separation by cryogenic traps. In situ oxygen isotope analyses by UV laser fluorination using both isotope ratio mass spectrometry and isotope ratio monitoring gas chromatography mass spectrometry (IRM-GCMS) microanalysis have been performed on tooth enamel (Jones et al., 1999). The 6180 values found for Pleistocene and Pliocene samples are similar to those obtained using the Ag3PO4 method with a mean precision of + 0.4%0. Samples of tooth
496
Chapter 22 - C. L6cuyer
enamel were gold coated to reduce the amount of contaminant CF4 and NF3 that were likely derived from reaction of F2 with organic compounds present within the enamel. All samples were initially degassed under high vacuum for at least 24 hours before prefluorination to ensure desorption of atmospheric water. Jones et al. (1999), indeed, note that the presence of water in the reaction chamber forms HF or H3PO4 upon addition of F2. Successful oxygen isotope measurements of biogenic apatites were obtained using 330 mW - 500 mW of UV light with a spot size of at least 80 mm. This minimum size for craters is imposed by the I nmol oxygen blank determined by Jones et al. (1999) during their experiments. One major advantage of this method is the possibility to make oxygen isotope analyses adjacent to one another because no damage surrounds UV laser pits.
22.5 Conclusion: geological applications Technical developments that occurred during the two last decades now offer the possibility to acquire high-precision oxygen isotope ratios from small apatite samples. The use of laser-based systems may also improve the spatial resolution of sampling. Similarly to biogenic carbonates, seasonal variations may be inferred from oxygen isotope ratios recorded in vertebrate tooth enamel (Fricke & O'Neil, 1996; Stuart-Williams & Schwarcz, 1997; Fricke et al., 1998a, b; Kohn et al., 1998). Our knowledge of the cooling history of metamorphic and plutonic rocks can also be improved using oxygen isotope compositions of mineral pairs including apatites (Fortier & L~ittge, 1995; Santos & Clayton, 1995). The phosphorus cycle could be investigated through a quantification of the various residence times in aqueous reservoirs, rates of biological recycling and source contributions. Early studies by Longinelli et al. (1976) showed that isotopic measurements of dissolved marine phosphate are possible. This was subsequently explored further by Blake et al. (1997 ; 1998) and L6cuyer et al. (1999) that, respectively, explored biologically mediated reactions of phosphate and kinetics of isotopic exchange under inorganic conditions. In conclusion, we can expect the emergence of many exciting results in the future involving the stable isotope geochemistry of apatites in life and earth sciences.
Acknowledgements
The author is grateful to T. Vennemann, H. Fricke and an anonymous reviewer who made constructive and thorough reviews that significantly improve the scientific quality of this work. P. Grandjean is especially thanked for her assistance in the chemistry lab.
Handbook of Stable Isotope AnalyticalTechniques, Volume 1 P.A. de Groot (Editor) 9 2004 Elsevier B.V. All rights reserved.
CHAPTER 23 Pyrolysis Techniques for Oxygen Isotope Analysis of Cellulose M. Saurerl & R. Siegwolf2 Paul Scherrer Institute, CH-5232Villigen PSI, Switzerland e-mail: 1 [email protected], 2 [email protected]
23.1 Introduction Oxygen isotope analysis of organic matter has many potential applications in ecology, plant physiology and climate reconstruction. Isotope fractionations during evaporation are causing variations in 8180 of precipitation, soil water and leaf water. This signal is transferred to the cellulose, whereby imprinting a time-averaged information on water sources and evaporative conditions (Sternberg et al., 1986; Yakir, 1992). Despite the complexity of the isotope fractionation in plants, there is good evidence that the oxygen isotope ratio in cellulose is mainly determined by the water absorbed by the tree and that the tree rings record the 8180-variations in the precipitation (Burk & Stuiver, 1981; Saurer et al., 1997a). The first attempts to relate climate conditions to oxygen isotope variations in trees have already been made in the late 70's (Gray & Thompson, 1976; Libby et al. 1976), but the method has not been very wide-spread ever since due to the laborious nature of the off-line pyrolysis techniques (Wong & Klein, 1986). Unlike in carbon isotope analysis, the organic samples cannot be cornbusted because this would introduce non-sample-derived oxygen. Among the pyrolysis techniques (= decomposition by heating), the nickel-tube pyrolysis (Thompson & Gray, 1977) and the mercuric-chloride pyrolysis (Rittenberg & Ponticorvo, 1956) were the most commonly used. There have been numerous attempts to improve the reliability and precision of these techniques (e.g. Schimmelmann & DeNiro, 1985; Brenninkmeijer 1983), but still no convenient and easy method was available until the early ninety's. However, the situation was rapidly changing with the development of on-line pyrolysis techniques (Werner et al. 1996; Koziet 1997). Most of these techniques are based on the pyrolysis of the organic material in an elemental-analyser or in a high-temperature device, whereby a helium flow carries the pyrolysis products to the isotope ratio mass spectrometer (thus the term on-line or continuous-flow). The precision achieved with these techniques is sufficient for natural abundance measurements and the analysis time is reduced by a factor of about ten compared to the offline techniques. It is possible to measure not only cellulose, but also nitrogen-containing plant material (Farquhar et al., 1997). In spite of some technical difficulties, related to blank and memory effects, with this progress an increasing number of studies using 6180 in organic matter for environmental and climatic research can be expected in the near future. The use of oxygen isotopes, so well established for water and carbonates,
498
Chapter 23 - M. Saurer & R. Siegwolf
could then more often be successfully applied in the study of organic matter. In this paper, we first give an overview on the features of the most commonly used off-line methods for the analysis of organic matter, in particular for cellulose, and we then discuss the on-line approaches. Finally, some questions related to calibration are addressed. The off-line methods are not completely obsolete, though most laboratories planning to measure oxygen isotope analysis in organic matter will nowadays favour a continuous-flow system. However, if carefully applied, the off-line methods should still yield a higher precision than on-line methods. Further, the off-line methods could be useful, in particular, for calibration purposes. There is no well-established organic standard for 6180 at present (Buhay et al., 1995). Because CO2 can be more easily related to VSMOW than CO, the off-line techniques producing CO2 could be useful in defining a reliable standard, whereas the on-line methods mostly use CO as measuring gas.
23.2 Off-line pyrolysis techniques Off-line techniques typically involve the use of vacuum lines with valves that are operated manually. The purpose of these systems is to quantitatively convert the oxygen of the samples into the measuring gas CO2. The gas is purified and cryogenically transferred to a sample tube that is later attached to the isotope-ratio mass spectrometer for analysis with the "classical" dual-inlet system. Off-line methods can be described as a two-step process: 1) The sample conversion (pyrolysis) resulting usually in the production of a mixture of CO and CO2 and 2) the conversion of CO to CO2 to concentrate all of the oxygen of the sample in the measuring gas.
23.2.1 Sample conversion 1) Nickel-tube pyrolysis: Cellulose decomposes when heated in vacuo to temperatures above 1000~ resulting in a complex mixture of gaseous species, like H20, CH4, H2, CO2 and CO. During pyrolysis in a nickel-tube, as proposed by Thompson & Gray (1977), H2 diffuses through the nickel-tube walls (the diffusivity depending on the temperature, pressure and wall thickness). By exchange reactions between the hydrogen-bearing compounds and H2, finally all of the hydrogen disappears, and a mixture of CO and CO2 remains. The rate constant of the Boudouard-equilibrium determines the relative amounts of CO and CO2: 2CO ~ CO2 + C + 172,58kJ
[23.1]
The conversion of CO to CO2 and subsequent isotope measurement on the combined CO2 is usually considered necessary (see below). Important modifications to increase sample throughput have been suggested by Brenninkmeijer (1983), including in particular the use of a batch method that allows multiple samples to be processed in one run. A clear disadvantage of the nickel-tube method is that impurities in the cellulose adversely affect the pyrolysis. Consequently, it is not possible to analyse nitrogen-containing matter. Further, a memory effect has been found related to oxygen deposited on the walls inside the nickel-tube (Borella et al. 1999). When samples are measured in a newly manufactured nickel-tube, the oxygen yield - given as per-
PyrolysisTechniquesfor OxygenIsotopeAnalysisof Cellulose
499
centage of oxygen of the cellulose sample recovered as CO and CO2 after pyrolysisis relatively low, indicating that the inner wall of the tube is oxidised. Therefore, an exchange of oxygen of the samples with the oxygen on the surface is likely to take place (Borella et al., 1999). 2) Mercury(II) chloride pyrolysis" This method is based on heating of the cellulose with excess mercury(II) chloride (Rittenberg & Ponticorvo, 1956). At temperatures above 350~ mercury(II) chloride dissociates, yielding chlorine for the oxidation of the cellulose, whereby mercury is the catalyst. After the reaction is complete, the hydrogen of the sample is found as hydrogen chloride and thus the formation of water is prevented (which otherwise would absorb oxygen from the sample). However, the hydrogen chloride has to be removed prior to the isotope analysis. For this purpose, Rittenberg & Ponticorvo (1956) suggested to use quinoline or 5,6-benzoquinoline. But the use of quinoline is cumbersome as it is toxic and has to be purified by multiple distillation shortly before use. Therefore, alternatives have been proposed for the removal of hydrogen chloride. Zinc was used by Sauer & Sternberg (1994), whereas Amberlyst A-21 (a macroreticular ion exchange resin) by Field et al. (1994). After the pyrolysis, the oxygen of the sample is found in both carbon monoxide and carbon dioxide in varying proportions, depending on the pyrolysis temperature and the heating rate. 23.2.2 CO to C02-conversion
In a second step after the pyrolysis, CO is converted to CO2. This step is usually considered necessary to ensure identical isotope ratio of the sample and the measuring gas, although Edwards et al. (1994) found that the conversion step may be omitted, thereby reducing the sample preparation time. To convert CO to CO2, a highvoltage discharge reactor, partially submerged in liquid nitrogen, was proposed (Aggett et al., 1965; Hardcastle & Friedman, 1974). In this process, the CO2 formed by the discharge is trapped in the liquid nitrogen trap and thus the Boudouard-equilibrium shifted toward the formation of additional CO2. The discharge method can be used in both the Ni-tube and the mercury(II) chloride method. Alternatively, the conversion of CO to CO2 can be achieved catalytically by the use of nickel-powder at about 350~ (Brenninkmeijer, 1983). The catalysis is based on the fact that by adsorption of CO on the nickel surface the activation energy for the formation of CO2 is reduced. The mercury(II) chloride method may also be used for nitrogen-containing samples. However, in this case the discharge method for the conversion of CO to CO2 cannot be used because of the formation of nitrogen oxides from N2 (NO2 and N20 conflict with the isotope measurement of CO2). Alternatively, when metallic nickel is used to disproportionate CO, the method is also applicable to nitrogen-containing compounds (Schimmelmann & DeNiro, 1985). When applied with care a reproducibility of about 0.2%o for cellulose can be achieved with the off-line methods and thus the methods yield reliable results at the natural abundance level. However, the techniques are time-consuming and the large number of steps involved can result in operator-influenced data (Mullane et al. 1988). The results of an interlaboratory comparison for a short tree-ring series involving the
500
Chapter 23 - M. Saurer & R. Siegwolf
mercury(II) chloride and the nickel tube method are shown in Figure 23.1. The correlation between the results obtained by the two methods is highly significant (r2 = 0.93, p < 0.001), but the slope of the regression line is 0.59 only. This could be caused by a memory effect of the nickel-tube method (see above) resulting in reduced sensitivity.
23.3 On-line pyrolysis techniques Oxygen elemental analysis (but not oxygen isotope analysis) is since long time routine based on the Schiitze-Unterzaucher reaction (Sch/itze, 1939; Unterzaucher, 1940). In this technique, the material is thermally decomposed in an oxygen-free atmosphere and all of the sample oxygen is converted to CO on hot carbon. The use of excess carbon in the reactor is necessary to minimise the formation and deposition of tarry products that might otherwise bind some of the sample oxygen. By use of a commercial elemental analyser this technique allows the automated and rapid analysis of micro-amounts of oxygen. The first attempts to apply this method to isotope analysis were done by Doering & Dorfman (1953) and Taylor & Chen (1970), but the results were not promising for natural abundance measurements. The introduction of carbon into the reactor is one important reason why the method is not easily applicable for oxygen isotope analysis. The hot carbon can react with the reactor walls, which usually consists of quartz (SiO2) or a ceramic material (A1203), resulting in the production of non-sample derived CO (in particular at temperatures above 1100~ This is the "blank problem" or "background problem". Another problem arises from memory effects, which is the contamination of a sample with oxygen from the previous analyses. The memory is caused by incomplete conversion of the sample oxygen to CO resulting in the production of an involatile oxide on the reactor surface that may exchange oxygen with the pyrolysis products of the subsequent sample or may slowly desorb as CO from the surface (Santrock & Hayes, 1987). A disadvantage of the abovementioned techniques (Taylor & Chen, 1970; Santrock & Hayes, 1987) is the use of iodine pentoxide for the conversion of CO to CO2 after the pyrolysis step, prior to the isotope analysis. This results in the introduction of oxygen from iodine pentoxide into the measurement gas, which necessitates correction procedures. Al-though Santrock
Figure 23.1 - 6180-values of a series of tree ring cellulose samples measured with two off-line methods: The nickel-tube method and the mercury(II) chloride method. Data replotted from Saurer et al. (1998). The mercury(II) chloride values were determined at the Godwin Institute for Quaternary Research, Cambridge, UK.
Pyrolysis Techniques for OxygenIsotope Analysis of Cellulose
501
& Hayes (1987) were successful in applying this method and thus were probably the first to measure 6180 at natural abundance level in continuous-flow mode, the use of CO as measuring gas is more convenient (see below). In recent years, efforts have been renewed to overcome the problems related to blank memory and CO/CO2-conversion. 23.3.1 CO as measuring gas An important step for improving on-line pyrolysis methods was to use CO instead of CO2 as measuring gas. This means that the isotope ratio 180/16 0 is inferred from the mass ratio 30/28 (Schmidt et al., 1992; Brand et al. 1994). Objections against the use of CO have been made because fractionation may occur when the CO-yield is not 100% and because CO cannot be trapped and purified as easily as CO2. If some CO2 is produced during pyrolysis, then CO+-ions are produced in the ion source from disintegration of the CO2 (with a production efficiency of about 10% compared to CO2 +ions), which would adversely affect the analysis. The Unterzaucher procedure, however, should give CO-yields close to 100% under favourable conditions, i.e. temperatures > 1100~ or with the use of catalysts (see below). Further, traces of CO2 can be eliminated with a soda lime trap. Another problem with the use of CO as measuring gas potentially is the interference with N2. N2 has a very low 30/28 ratio (thus very different to natural abundance CO), and could be produced from nitrogen-containing samples and from atmospheric air (background or leak). As separation of CO and N2 is easily achieved with a molecular sieve, the use of CO as measuring gas appears to be justified. Precautions have to be taken when pumping off CO (standard gas) because it is highly toxic. 23.3.2 Memory and blank Various approaches were proposed to reduce memory and blank. Gygli (1993) used a glassy carbon tube for elemental oxygen determination, which necessitates, however, protecting the tube from the atmosphere (by flushing the outside of the tube with the carrier gas). In a similar line, Werner et al. (1996) proposed to use glassy carbon instead of elemental carbon in the pyrolysis reactor of an elemental-analyser. This proved to reduce the blank because the relatively inert carbonaceous material does not significantly react with the quartz, but all the same has the effect to produce high yields of CO. Koziet (1997) performed the pyrolysis at 1300~ in a vitreous carbon tube which is virtually free of oxygen. However, this procedure requires substantial modification of the elemental analyser and is not completely free of memory effects. High-temperature pyrolysis (T ~ 1400~ has been proposed for organic as well as inorganic substances (Kornexl et al., 1999b). The high temperature ensures complete conversion of the sample to CO for many substances (including water), but necessitates the use of a special furnace (elemental analysers can usually not be operated above 1100~ The introduction of nickelised carbon (30% or 50% Ni) in the reaction tube allows it to lower the temperature and still results in the complete conversion of the organic material to CO (Farquhar et al., 1997; Breas et al., 1998), although conversion of water may not be complete. The relatively low temperature (1080~ will substantially reduce the exchange reactions with the quartz or the ceramic tube. This exchange can further be reduced by insertion of a nickel foil preventing the contact of
502
Chapter 23 - M. Saurer & R. Siegwolf
catalyst and sample with the tube. The method with nickelised carbon in the reactor has been successfully applied to nitrogen-containing matter and water. An approach based on nickel-catalysis has also been proposed for the simultaneous analysis of h180 and 6D (Begley & Scrimgeour, 1997), whereby water and volatile organic compounds (in sub-microliter amounts) are pyrolysed to H2 and CO. Here, the memory effect is reduced by inserting a precisely determined amount of hexane as source of carbon (deposited on the nickel). By use of a GC capillary column prior to the pyrolysis the approach by Begley & Scrimgeour (1997) enables compound-specific oxygen isotope analysis. This methodology, referred to as GC-pyrolysis-IRMS, was already proposed by Brand et al. (1994) and further evaluated by Hener et al. (1998), but will not be discussed here in detail. 23.3.3 Water adsorbed on the samples A problem encountered in the on-line methods unknown in the off-line methods is the water adsorbed on the samples. Cellulose is highly hygroscopic, and the water content can be around 7% of the total weight in cellulose equilibrated with laboratory air (assuming the relative humidity of air to be about 50%). Drying the cellulose does not really circumvent this problem as moisture is taken up within seconds after exposure to air. Whereas all the water is p u m p e d off by the vacuum-system in the off-line methods, pyrolysis with the on-line methods could also produce CO with oxygen from the adsorbed water. This adversely affects the measurement because the oxygen isotope ratios of water and cellulose are very different: the 6180 of water can be expected to be in the range from a b o u t - 5%0 to - 15%o, whereas the/5180 of most cellulose samples will be in the range from + 20%o to + 40%o. A water content of 1% could shift the measured 6180 by as much as 0.4%0 (assuming for this calculation a difference of 40%o between the isotope ratios of water and cellulose). This effect is demonstrated in Figure 23.2, where it is shown that differing exposure time of cellulose samples to atmospheric moisture does affect the measured oxygen isotope ratio (using the method described in the next section). For practical reasons we suggest to measure samples and standard both in equilibrium with air moisture. This equili-bra-
Figure 23.2 - 8180-values of cellulose samples (extracted from wood of 4 different tree species) measured with the online method described in the text. The exposure time to air moisture of one part of the samples was kept as short as practicable ("dry"), whereas the other group of samples was equilibrated with laboratory air ("humid").
Pyrolysis Techniques for OxygenIsotope Analysis of Cellulose
503
tion could be done in a closed system where a constant isotopic composition of the vapour is maintained. If samples and standard have about the same water content, then the contaminating effect will be less significant. To avoid any contamination with water, a solution could be to first weigh the samples into tin or silver capsules, heat the closed capsules to dryness in an oven, and for the analysis place them in a specially adapted autosampler which is continuously flushed with an inert gas such as Argon (G. Schleser, pers. communication). Further, it might be important to always analyse the same amount of sample and standard, which also should be of about the same particle size.
23.3.4 On-line method: An example In the following section, results based on the method of Saurer et al. (1998) and Werner et al. (1996) are presented. An elemental analyser (Carlo Erba 1108, Italy) is linked to an isotope ratio mass-spectrometer (Delta-S, Finnigan MAT, Germany) via an open split interface (Conflo II, Finnigan MAT). The quartz pyrolysis tube in the elemental analyser is filled to half of its length with glassy carbon (SIGRADUR G; 3150 to 4000 mm grit; HTW GmbH, Germany) and is kept at a temperature of 1080~ The samples (ca. 1.5 mg) are filled in tin or silver capsules (which should be pretreated to remove oxides) and then are dropped into the pyrolysis furnace by an autosampler (without addition of oxygen). The produced gases (mostly CO, H2) are swept by the helium carrier gas (99.9999% pure; flow rate ca. 80 ml/min) through a CO2-trap (carbosorb, Elemental Microanalysis, United Kingdom) and a water trap (magnesium perchlorate). For cellulose samples the separation column used is a Poropak QS at about 40~ whereas for nitrogen-containing samples a molecular sieve 5A at about 60~ is used. The time period between consecutive samples should be adapted such that the 30/28-ratio has reached the background level after each analysis. This time depends on impurities slowly evolving from the column and can be expected to be in the range from 500s to 1000s. The isotope ratios are calculated from the time integrals of the peak areas of the ion intensities m / z 30 and 28 (i.e. 12C180+ and 12C160+). For every sample, the 30/28-ratio is related to the 30/28 ratio of a CO standard gas (99.97% pure; Linde, Germany). The calibration vs. VSMOW is done by cross-calibration with standard cellulose the 6180 of which has been determined on CO2 produced by the nickel pyrolysis method (see also next section "Calibration"). The CO standard gas is measured against CO produced by pyrolysis of the standard cellulose and from this difference the oxygen isotopic composition of the CO standard gas is calculated. As the 13C170+-ions have the same m / z as the 12C180+-ions, a "170-correction" has to be done, in an analogous way as for CO2 (Craig, 1957; Mook & Grootes, 1973), although the correction is small (the natural abundance of 13C170 is ~ 4.3 x 10 -6 and thus ~ 500 times smaller than that of 12C180). An assumption for the relationship between 180/160 and 170/160 is made, which is often taken as: 17Rsal
= 18Rsa
17RstJ
18Rst
[23.2]
504
Chapter 23 - M. Saurer & R. Siegwolf
where Rsa and Rst are isotopic ratios of sample and standard, respectively (17R = 170 / 160 and 18R - 180/160). With this expression, a relatively simple correction term can be calculated (Farquhar et al., 1997; Saurer et al., 1998). More generally, the relationship between 18R and 17R is taken as 17R - k18Ra (where k and a are empirically derived constants) which necessitates an iterative correction algorithm (Ricci & Douthitt, 1998). In Figure 23.3 data for cellulose samples measured with this method are shown. The results demonstrate that the method yields a reproducibility of about 0.3%o. No memory effect is detectable for this range of h-values, although the first 3 - 4 samples in a series are sometimes increasing due to an "equilibration effect". Comparison of the results with the nickel-tube and the mercury(II) chloride method were also promising. 24 different samples were measured with both the on-line and the mercury(II) chloride method (Saurer et al., 1998). The correlation between the results of the two methods is highly significant (r2- 0.83; p < 0.001), whereby the slope is 0.85 + 0.08 (using the mercury(II) chloride method as the reference method). Further, the 6180value for the IAEA-C3 cellulose measured with the mercury(II) chloride method (32.14%o + 0.20%0) is consistent with the value measured with the on-line method (32.52%o + 0.33%o). Figure 23.3 - Two series of consecutive measurements of cellulose samples with the on-line method described in "On-line method: An example". The technique is based on pyrolysis in an elemental analyser at 1080~ The values are given as "raw data", i.e. the relative deviation of the 30/ 28-ratio of the sample from the CO standard gas is shown. The data show the absence of memory effects for samples differing by up to 20%o.
Pyrolysis Techniques for Oxygen Isotope Analysis of Cellulose
505
In summary, a surprising diversity of differing continuous-flow approaches has evolved in only a few years. The choice of the "best" method for someone entering the field is not easy and depends on the materials he is interested in, taking into account also the ease or complexity of the different techniques. The reported range of reproducibility is from about 0.2%o to 0.6%o, but could be sufficient for most methods, provided that the details of the procedure are carefully considered. A simple means for addressing memory problems could be to measure every sample two or three times. Memory and blank have to be carefully observed and therefore pyrolysis may be less routine than combustion for a considerable time in the future. 23.4 Calibration
At present there is no well-calibrated organic standard material available. The IAEA had initiated an interlaboratory comparison (at a time when only off-line methods were in use), but the results must be regarded as preliminary (Buhay et al., 1995). 5 labs have been involved in this test all using either the nickel-tube or the mercury (II) chloride pyrolysis. A cellulose sample was analysed (IAEA-C3) and the results were in the range from 31.3%o to 32.7%o vs. VSMOW. This is quite a large range comparing for instance with intercomparison results of calcite materials, where the stan-
Figure 23.4 - Results of an interlaboratory comparison for on-line 180/160 analysis of cellulose. Laboratories involved in the ring test (in alphabetical order): ANU Research School of Biological Sciences, Australia (G. Farquhar); Department of Biology, University of Utah, USA (J. Ehleringer); Finnigan MAT, Bremen, Germany (H. Avak); Paul Scherrer Institute, Switzerland (M. Saurer); UFZ Centre for Environmental Research, Leipzig, Germany (B. Kornexl); Weizmann Institute of Science, Israel (D. Yakir). Data courtesy of J. Ehleringer.
506
Chapter 23 - M. Saurer & R. Siegwolf
dard deviation of results from different labs is usually smaller than 0.3%o for ~180 (Buhay et al., 1995). A ring test aimed at comparing different on-line methods was organised recently by J. Ehleringer (University of Utah, USA). A number of cellulose samples was analysed by 6 different laboratories. The on-line methods used are all based on pyrolysis in an elemental-analyser or in a high-temperature device, but there are significant differences in the exact details of the pyrolysis set-up (Werner et al., 1996; Farquhar et al., 1997; Kornexl et al. 1999b). The results spanning the range from about 20%o to 40%o vs. VSMOW were generally very consistent, yielding correlation coefficients r > 0.97 for the crosswise correlation of the results from different labs. However, there is still considerable uncertainty about the absolute calibration, reflected in systematic differences of up to 2%o between the different laboratories (see Figure 23.4). A reason for the problems with the ring tests may be that cellulose is not an ideal reference material for ~180 because it is highly hygroscopic (see above). Therefore, it could be better to rely on a different material. The IAEA proposed to measure the IAEA-CH6 sucrose (referenced as ANU sucrose earlier) for 6180 calibration in addition to the IAEA-C3 cellulose. However, there are not yet many data available for this material (besides, sucrose is hygroscopic too). A synthetic material, like polyester or polycarbonate, might be a better intercomparison material. Calibration using SLAP and VSMOW waters is an alternative and promising way for those systems that allow the complete pyrolysis of water (Kornexl et al., 1999a). A disadvantage of this approach is that the standard and the samples would thus not be of the same material. Pyrolysis of organic material may result in fragments of mass 28 other than CO, depending on the structure of the pyrolysed compounds (Werner et al. 1996). Using differing sample and reference materials is thus only feasible if the method has been proven to yield 100% conversion for both sample and reference. Nevertheless, the calibration should also be validated for every new group of organic compounds by using a reference of similar structure. This problem is particularly urgent for compoundspecific pyrolysis (Hener et al., 1998).
Handbook of Stable Isotope Analytical Techniques, Volume 1 P.A. de Groot (Editor) 9 2004 Elsevier B.V. All fights reserved.
CHAPTER 24 Sample Homogeneity and Cellulose Extraction from Plant Tissue for Stable Isotope Analyses Silvio Borellal, Guillemette M6not 2 & Markus Leuenbergerl,3 Climate and Environmental Physics, Physics Institute, University of Bern, Sidlerstrasse 5, 3012 Bern, Switzerland 2 Institute of Geology, University of Bern, Baltzerstrasse 1, 3012 Bern, Switzerland present address: Woods Hole Oceanographic Institution, Department of Geology and Geophyics, Massachusetts, USA e-mail: 3 [email protected] 1
Abstract Stable isotopes have been measured on organic samples for several decades. Increasing interest was given to measurements in tree rings in the field of climatic and environmental reconstruction, since it represents a well-dated, climate sensitive archive. Although several studies have shown that hardly any loss of information is found when analyzing whole wood, most of the studies have been performed on cellulose. In this chapter we present and discuss the principal steps necessary between wood sampling and isotopic analysis on cellulose. We found that fine milling the samples is very important for (1) avoiding sample inhomogeneity leading to increased measurement uncertainty, and (2) for permitting complete R-cellulose extraction. For both reasons, we find that a wood grain size of approximately 0.15 mm is optimal when analyzing subsamples of I - 1.5 mg. Then, we describe the method for ~-cellulose extraction from tree ring wood samples. Additional considerations concerning the extraction of cellulose from shrubs, annual plants and mosses are also presented. This technique has proved to be highly reproducible for carbon and oxygen stable isotope measurements, and allows rapid throughput (up to 120 samples in two weeks, the time intensive milling not included). Finally, we show that the R-cellulose extraction does not induce any visible scatter in the 813C and 8180 values, at least with present isotope analysis techniques, which allow a measurement precision of 0.1%o and 0.2%o for 813C and 8180, respectively (see Chapter 23).
24.1 Introduction During the last few decades there has been an increasing interest for stable isotope measurements from plant material, as tracer, as indicator for plant physiological processes, and for environmental and climatological studies. As a subfield of the dendro3. Correspondence should be adressed to this author.
508
Chapter 24 - S. Borella, G. M6not & M. Leuenberger
ecology and dendroclimatolog~ stable isotope measurements in tree rings have a large potential, since they combine the principal advantage of the dendrochronology, the precise and accurate dating, with the now well known sensitivity of the stable isotope ratios of carbon, oxygen, and hydrogen to meteorological and other environmental variations (see for example Griffiths, 1998 or Leuenberger et al., 1998 for a review of this topic). The recent development of on-line methods for the stable isotope determination in organic material based on gas chromatography mass spectrometer techniques opens new research opportunities (Barrie, 1991). This technique leads to a significant increase in sample throughput and a potential to decrease the sample size, which are both necessary conditions to acquire long time series from limited wood amounts in tree-ring studies. As a consequence of this technical improvement, sample preparation (e.g. cellulose extraction) becomes the time-limiting step, and one has to take greater care of the sample homogeneity. Therefore, it is justified to question, whether the extraction of a single wood component is necessary for stable isotope analysis, and if this is judged necessary, a rapid and reproducible extraction technique, which produces homogeneous samples, is required. The question, whether stable isotopes can be analyzed on whole wood has been indirectly addressed by several authors for ~}13C. Francey (1986), Leavitt & Long (1991), Schleser (1990), Yildiz (1995), and Livingston & Spittlehouse (1996) found that interannual or intra-annual signal patterns of 613C values measured on cellulose and wood are highly correlated. Wilson & Grinsted (1977) and Mazany et al. (1980) found similar results for cellulose and lignin. However, despite these results, the postulated necessity of extracting cellulose for isotope measurements (Wilson & Grinsted, 1977; Grinsted et al., 1979; Livingston & Spittlehouse, 1996) had not been questioned until very recently, when Borella et al. (1998) stated that 613C analyses can be performed on whole wood instead of cellulose, without significant loss of information, at least for Quercus and Fagus. Analogous studies for 6180 are much more scarce, and rather point to that cellulose contains significantly better climatic information than wood. Indeed, Gray & Thompson (1977) and Borella et al. (1999) both found better correlations between meteorological parameters and 6180 of cellulose than with 6180 of wood, for Picea glauca and Quercus, respectively. Nevertheless, uncertainties of the 6180 values of wood due to variable N-content (interference between CO + and N2 + in the mass spectrometer) is also a possible cause of the reduced correlations values (Borella, 1998). Moreover, some ~180 analysis methods involving conversion of CO to CO2 on hot Nickel do not function with impure cellulose (Brenninkmeijer, 1983; Borella, 1994). And last, the necessity of extracting R-cellulose or holocellulose prior to 62H measurements results from the necessity of removing the hydroxyl H-atoms, which can exchange with water (Heuser, 1944). In principle two methods exist for accessing the non-exchangeable H-atoms in cellulose, (1) nitration of cellulose (Epstein et al., 1976; DeNiro, 1981) and (2) control of exchangeable hydrogen by equilibrating the cellulose with water of known isotopic composition (Wilson & Grinsted, 1975; 1977; Schimmel-
Sample Homogeneity and Cellulose Extraction from Plant Tissue for Stable Isotope Analyses
509
mann, 1991, Feng et al., 1993). Cellulose nitration is more often used and gives good reproducibility but its time consuming and requires a rather complicated (somewhat dangerous) procedure. Furthermore, Brenninkmeijer (1983) experienced that incomplete removal of lignin interferes with the nitration and leads to erroneous 82H values. Therefore, it is clear that cellulose extraction and sample homogeneity are still of actuality. The aims of this chapter are to discuss influences of sample homogeneity / heterogeneity on stable isotope ratios and to present one of the numerous methods for R-cellulose or holocellulose extraction, which has proved very reliable, and which allows to process up to 120 samples in parallel. Additionally, other techniques are also briefly presented in section 24.3.8.
24.2 Sample homogeneity 24.2.1 Theoretical considerations
Here we like to recall that the original wood used for cellulose extraction should be mixed from four cores of different directions and at least four to six trees per site when acquiring long time series for different sites (see for example Leavitt & Long, 1984; McCarroll & Pawellek, 1998). This will reduce the amount of sample preparations significantly and enhances the representativeness of the measurements for the corresponding site. Besides the sample selection, an important step in the preparation of the samples is the milling process, especially for measurements on very small subsamples. A very coarsely milled sample could lead to an overrepresentation of some parts of the ring. This effect can be limited when the isotope analysis is restricted to late wood only. However, it is still worthwhile to estimate the fineness, which is required to assure that subsamples do represent the whole sample, we model the sample as consisting of two parts with distinct isotopic values 8 + A and 8 - A (This is of course extremely simplified, since a large part of the wood will have intermediate 8-values. However, it permits simple calculations and gives an upper limit for homogeneity needs). We set A to 1%o or 1.5%o, respectively, in accordance with intra-annual differences for different tree species (e.g., Pinus radiata, Pinus strobus, Acer saccharinum, Juniperus, Pseudotsuga, Quercus robur)( Wilson & Grinsted, 1977; Leavitt & Long, 1991; Leavitt, 1993a, b; Switsur et al., 1995). If the measured subsample is made up of n pieces of cellulose (or wood) with the same weight, a binomial distribution gives the following results: k pieces with 8 + A and n - k pieces with 8- A gives a mean value of 8m - 8 + (2k- n)/nA instead of 8. The probability, P, that one measures a value (2k- n)/nA higher than expected is p(hm_8 +2kA_A) _ p(k)_ n
(~) n
n! k!(n - k) !
[24.1]
Summing equation [24.1] for ~ m - ~ - ~ to 8 + ~ gives the probability to measure a 8m with a maximum deviation of ~ from the correct value 8; it is given by
510
Chapter 24 - S. Borella, G. M6not & M. Leuenberger
n k<-
n!
)+p(k_2)
[24.2]
n
2
For n > 36, this formula can be approximated by a Gaussian distribution (~ = n / 2 ; o = ~/(n/2))(see any mathematical table). As the reproducibility of today's on-line methods is in the range of 0.1 to 0.2%0, we calculate the n u m b e r of pieces of a subsample necessary to restrict the potential deviation to less than 0.1%o (r < 0.1%o, 0.15%o, 0.2%0, respectively) with a probability of better than 68% ( l o = ~) or 95% (2 o = ~), respectively (The approximation with a Gaussian distribution mentioned above leads to n _>(~xA/~)2, where c~ = ~ / o). Table 24.1 shows that - 250 pieces are required to restrict the deviation to less than 0.1%o with a 68.3% (1 o) probability for k - 1.5%o. This is probably an extreme case because a real sample is not constituted of two pools but covers a continuous isotopic spectrum. These 250 pieces of a sample of 1.5 mg correspond to pieces of _<6 mg. Note that a sample weight of only 100 mg would be sufficient for the mass spectrometric measurements but we mainly use larger samples (1.5 mg) because of a better sample homogeneity. With a cellulose content of 40% and a wood density of about 600 k g / m 3 the fineness of the milled wood has to be better than ~ 0.025 mm3, i.e. cubes of less than 0.29 m m side length (Wood is milled prior to cellulose extraction, because (1) the extraction is not well achieved on very large pieces and (2) cellulose is extremely difficult to grind). Using a sieve, it has to be taken into account that longer pieces with the same cross section could pass, so we would suggest using a sieve of 0.15 - 0.2 m m to be sure that all pieces are small enough. Measuring wood instead of cellulose, the sieve must be ~/2.5 times finer, i.e. ~ 0.10 mm. An alternative to a sieve to control the fineness of a sample is to weigh the largest pieces after milling. These theoretical results show that one must pay attention to the milling. For example, we calculated the scatter caused by the inhomogeneity of samples "ground to 20 mesh" as described by Leavitt & Lara (1994). We find a 1 o error of ~ ~ 0.2%o, which is not negligible in comparison with their "usual overall precision of 0.15 0.20%0". Table 24.1 - Homogeneity requirements.
All these calculations rest on Probability \ ~ 0.1%o 0.15%o 0.2%o the assumption that the subsample is taken from an infinite sample. 68.3% 110 / 239 53 / 110 29 / 68 95.4% 410/914 187/410 104/232 This assumption is approximately valid in our case, as we measure subsamples of 1 - 1.5 mg out of Number of pieces in a sub-sample as required in order to measure a value which deviates less than ~ (0.1, 0.15, samples of 10 - 50 mg (in most 0.2%o respectively) from the correct value with a probacases). In the extreme case that all bility of 68.3% or 95.4%, respectively. The first value of the sample material is used for applies to k = 1%o, the second to A = 1.5%o, see text.
Sample Homogeneity and Cellulose Extraction from Plant Tissue for Stable Isotope Analyses
511
the measurement, homogeneity is unimportant, of course. As an intermediate case, we calculated the probabilities for subsamples making up one half of the sample. Our results show that one needs about half as much pieces in a subsample than in the "infinite case" to obtain the same precision.
24.2.2 Experiments Homogeneity depends on the fineness of a sample, as seen in the previous section. To test this dependence, we measured twice or more times the 13C/12C isotope ratio of some coarse-milled samples, from which we weighted single pieces. The standard deviations of the samples are then compared with that of our commercial standard cellulose, which can be assumed as highly homogeneous in comparison to our measurement precision, to split the standard deviation into effects from sample inhomogeneity and analysis uncertainty, respectively. In Figure 24.1 the 613C deviations from the accepted mean value are given versus the number of cellulose pieces for the online and off-line techniques. Despite the fact that the number of samples in this study is too low for a statistical analysis, we can see that the samples with fewer cellulose pieces tend to show more scatter. The larger deviations are in agreement with our theoretical calculations shown as dashed lines. Another important result is the generally lower 813C values of the samples with larger cellulose pieces (i.e., fewer pieces). This shows that the cellulose extraction is incomplete when processing too large wood pieces, and that impurity (lignin) in the cellulose lowers the 613C value (see also section 24.4.2 and Figure 24.2). 24.3 Description of holo- or c~-cellulose extraction methods
In the following section we describe the method which is used at our department. It is based on the protocol introduced by Brenninkmeijer (1983). The method consists of four principal steps" First, grinding of the wood. Secondly, removal of lipids and Figure 24.1 - Difference from the accepted value for various sub-samples with k n o w n number of cellulose pieces. The mass of the subsamples is almost kept constant within each method: 1.3 - 1.5 mg for the on-line method (full circles) and 2.8 - 3.5 mg for the off-line method (open circles). The error bars to the right represent the standard deviation of the online (larger) and offline method (smaller), as determined with measurements of our commercial cellulose. The dashed lines show the expected uncertainty as a function of the number of cellulose pieces as calculated according to subsection 2.1 for k = 1%o and a probability of 68.3%.
512
Chapter 24 - S. Borella, G. M6not & M. Leuenberger
Figure 24.2 - (a) 613C-increase during c~-cellulose extraction, i.e. during lignin removal (along the lines) and hemicellulose removal (between both lines), respectively. Symbol size characterises the measurement uncertainty. (b) Measurements on different complete extraction-series from our internal standard shows that a bad extraction can lower the 613C-value by almost 0.2%0. But we can consider our extraction technique as suitable for 613C measurements, as there is no significant shift after a second extraction (compare third and first points from the right). The errors of the means (error bars) and the numbers of measurements (in brackets) are given in panel b. The lines connecting the two figures should emphasise that the variations between different complete extractions are much smaller than the difference between wood and cellulose.
resins with toluol-ethanol. Thirdly; bleaching of the wood with NaC102 and acetic acid, and finally, purifying of the R-cellulose with NaOH. In between and after the last two steps rinsing with de-ionized water is essential. In the following, we will detail the successive steps. For holocellulose one can principally stop after the NaC102 extraction. 24.3.1
Milling
For the extraction of R-cellulose from wood samples, it is necessary to finely mill the wood. This is an important prerequisite for a complete cellulose extraction (see section 24.4.2), as well as for yielding homogeneous samples (Borella et al., 1998)(see also section 24.2, above). Different commercial milling devices are available, most of which are quite expensive. Moreover, many large machines are not well suited for the small tree ring samples of usually a few tenths of mg. Therefore, we use a simple coffee-mill (the model we use is Moulinex 980, 60 g), with minor modifications (see Appendix, Picture 24.A1). The milling device consists of two rotating cutting blades. We only added a deflection plate to direct the wood pieces towards the rotating blades. The angle of this plate has to be optimized experimentally. Without this plate, or with a plate fixed in a wrong angle, the wood is pressed to the cover by the rotational movement, without coming in contact with the blades.
Sample Homogeneity and Cellulose Extraction from Plant Tissue for Stable Isotope Analyses
513
After drying for a 4 hours at ~ 100~ the wood samples are processed in this modified commercial coffee-mill. As shown in section 24.2, above, a fineness of 0.1 to 0.15 mm seems ideal if subsamples of ~ 1 mg are measured. It is possible that the extraction will be incomplete if the pieces are too large, resulting in impure R-cellulose. This has an impact on the isotopic composition, because the different constituents of the cellulose have not the same isotopic composition (see Figures 24.5 and 24. 2, and section 24.4.2). For the chemical extraction, the samples are placed into glass tubes of 100 mm height, 12 mm outer diameter, and 10 mm inner diameter, with ground up glass at their bottom as filter (see Figure 24.3 and Picture 24.A2 in the Appendix). The tubes are then moved into a cylindrical Teflon holder of 350 mm height and 900 mm in diameter perforated with 30 holes, which allows processing 30 samples in parallel (see Appendix, Picture 24.A3). Additional apparatus necessary for the extraction are: A Soxhlet of sufficient size to contain the Teflon holder with the 30 glass tubes; a water bath that can be heated to 70~ and a drying stove for drying the samples at ~ 100~ The samples are weighted (normally 50 - 100 mg) for control and placed into the glass tubes (see Figure 24.3). Smaller samples (< 10 mg) can be processed if care is taken to minimize sample loss. Much larger samples can also be problematic when the different solutions cannot reach the entire sample. Therefore, larger samples may have to be processed longer, and stirred with a glass stick during steps 3 to 5 (see below, section 24.3.7: Remarks). For the next steps we reached a good efficiency by processing 30 or 60 samples in parallel. This is done with one or two Teflon holders (as described above) containing each 30 glass tubes. But processing even more groups of 30 samples in parallel is also possible, and Teflon holders designed with another geometry may contain even more samples, provided that the Soxhlet extractor is large enough.
24.3.2 Toluol-ethanol step Dissolution of the extractives is performed with a Toluol-ethanol (1 91) in a Soxhlet distillation apparatus for 5 hours. The temperature is chosen such that the solution is automatically renewed every 15 - 30 minutes by siphoning. To avoid boiling of the toluol-ethanol mixture, we add a few small glass-pieces into the solution. The power of the heating device is
Figure 24.3 - Glass tube for cellulose extraction. The filter at the bottom of the tube is made of small glass particles sintered together, its porosity is PO; it is permeable for the liquids used, but not for the wood (respectively cellulose) particles. The placement of this filter is important since it improves the availability of chemicals during the extraction. We ordered our glass tubes (with filter) by Trabold & Cie, Konsumstrasse 28, CH-3007 Bern. (N22SFr/ piece ~ 16US$ / piece).
514
Chapter 24 - S. Borella, G. M6not & M. Leuenberger
controlled with a potentiometer. Toluol-ethanol is flammable and carcinogen, and this process must be carried out in a ventilated chemical chamber.
24.3.3 NaCl02 step The second extraction step involves sodium chlorite (NaC102) and acetic acid (CH3COOH). The aim of this step is removal of lignin. Following proportions should be used: 0.5% (by weight) NaC102 and 0.15% (by volume) acetic acid in de-ionized water (e.g. 1 liter of de-ionized water, 5 g of NaC102, 1.5 ml of acetic acid). The samples (still in the Teflon holder) are placed into the solution, and the solution is kept at 70~ in a water bath. This step lasts for 9 - 15 hours and is repeated 3 - 5 times (daynight-day-...). The solution is renewed between the repetitions. For large samples (e 100 mg), or for non-wood samples (e.g. leaves), this step may have to be repeated more than 5 times. The criterion for finishing this step is the whiteness of the samples (a rather subjective criterion, but comparison with commercial cellulose and experience could help).
24.3.4. NaOH step Then, the hemicelluloses are removed with a 4% (by weight) sodium hydroxide solution (40 g of NaOH in I liter of de-ionized water) to obtain a pure (x-cellulose. The dissolution of NaOH is exothermic and the solution should be stirred until all NaOH is dissolved. As during the NaC102 step, the samples are kept at 70~ in the solution. But this step is performed only once, during 24 hours. It often occurs that the solution becomes yellowish at the beginning of this step. Generally, it becomes transparent again after a few hours. If this is not the case, replace the solution after 12 hours. A persistent yellow coloration could also be an indicator for incomplete rinsing after the NaC102 step, or even an incomplete NaC102 step itself. Some stable isotope laboratories work with holocellulose (i.e. hemicellulose and c~cellulose), and do not proceed to this last step (Leavitt & Danzer, 1993).
24.3.51-120 rinsing In order to purify the extracted c~-cellulose from all chemicals utilized in the extraction procedure, the samples are kept for 24 hours in de-ionized water, at 70~ 24.3.6 Drying Finally, the samples are dried for _>5 hours at ~ 100~ and then weighed. The percentage of c~-cellulose in the wood depends on the material (wood or leaves) and the tree species (See section 24.4.1 and Table 24.2). 24.3.7 Remarks 9Prior to each solution step the samples are rinsed with de-ionized water. Between the NaC102 and the NaOH steps, and after the NaOH step, the samples are rinsed to closely match pH ~ 7 (3 - 5 rinsing processes are required). ~ Samples larger than ~ 100 mg may need longer extraction. One has also to be aware that softwood (conifers), as well as leaf material, is less dense than hardwood
515
Sample Homogeneity and Cellulose Extraction from Plant Tissue for Stable Isotope Analyses Table 2 4 . 2 - Cellulose yields. species
10 10' 11 11' 11' 11' 12' 14 14 14
Fagus Quercus Quercus Quercus Fagus Quercus Quercus Quercus Quercus Quercus Quercus Quercus Quercus Quercus Quercus Quercus Quercus Fagus Betula
site
Li Li Li Li Li Sa Sa Sa Sa Sa Sa Sa Sa, Li Sa Li Li Sa Kr WSL
date of extraction
1 / 96 2/96 3 / 96 4 / 96 4 / 96 5 / 96 6 / 96 6 / 96 8 / 96 11 / 96 11 / 96 12 / 96 12/96 12 / 96 12 / 96 4 / 97 1 / 99 1 / 99 1 / 99
mean yield [%]
o
slope
+ slope
intercept
+ intercept
[%]
[%]
[%]
[mg]
[mg]
39.0 41.6 43.9 41.9 39.7 35.5 36.4 29.3 32.4 32.0 35.7 35.7 39.0 36.7 40.1 41.6 32.1 39.1 25.9
0.9 2.4 1.7 1.9 1.2 4.7 3.4 5.1 3.8 6.2 4.2 4.7 4.1 3.8 4.0 3.3 1.1 1.5 2.7
40.3 46.1 49.8 49.5 41.4 39.8 44.9 39.5 32.2 33.5 30.3 43.1 42.8 45.7 40.3 41.8 32.8 40.5 57
0.7 0.6 1.0 3.8 2.2 0.9 1.7 1.2 2.1 7.8 2.4 0.7 1.4 0.7 2.7 1.7 4.1 0.9 12
-1.58 -1.05 -4.99 -8.25 -2.88 -1.84 -3.72 -6.74 -0.02 -4.17 1.20 -2.94 -1.92 -4.95 -0.07 -1.98 -0.72 -0.84 -31.7
0.92 0.19 0.88 4.23 3.99 0.64 0.85 1.43 1.13 4.14 1.23 0.51 0.91 0.50 1.48 0.60 4.35 0.71 12.1
n
r2
26 27 28 8 20 28 28 20 30 30 29 30 30 10 19 19 8 11 6
0.993 0.995 0.991 0.966 0.950 0.986 0.963 0.983 0.937 0.396 0.859 0.993 0.972 0.998 0.931 0.971 0.914 0.995 0.855
Typical cellulose yields obtained with our extraction method for different tree species. Because of losses during the extraction and of the non-quantitative removal of the extracted cellulose from the glass-tube, the real yield cannot be determined directly as the ratio of the w o o d and cellulose masses measured before and after the extraction. In order to take these losses into account, we determined the yield by the use of the linear regression method demonstrated in Figure 24.4 (see text, and especially equation 24.2) The first column of the Table gives the number of the extraction; the second is the tree species; the third is the site name: "Li" is a site in Northern Germany, "Sa" and "Kr" are sites on the Swiss ,,Mittelland", and "WSL" stands for a 1-year old birch grown in a controlled chamber. Column 4 gives the date of extraction (month and year). Columns 5 and 6 are the mean values and the standard deviations (o), respectively, of the ratio of measured cellulose to wood mass. Columns 7 to 12 give the results of the linear regression described above: slope (~ real cellulose yield), error of the slope, intercept (- minus lost mass), error of the intercept, number of samples in the series (n), and correlation coefficient of the regression (r2). Obvious outliers have been excluded prior to the calculations.
(broad-leaved trees), which influences the duration of the extraction (softwood, and leaves, needing longer extraction time, due to the larger volume of the samples). 9In order to improve the extraction, and its rapidity, we suggest to stir the samples approximately once every hour with a glass stick; this is especially necessary for large samples. 9 The drying temperature for organic samples should never exceed 100~ partial combustion which could create large isotopic fractionation.
to avoid
9The whole procedure lasts one to two weeks (excluding milling), depending on the number of NaC102 steps necessary until the samples are white. The total duration
516
Chapter 24 - S. Borella, G. M6not & M.
Leuenberger
will depend on whether one has several Soxhlet extractors to run in parallel.
24.3.8 Other methods Several other procedures for R-cellulose extraction from wood samples have already been described during the last decades. Our method is primarily derived from the technique of Brenninkmeijer (1983), with major practical improvements. Older methods used benzene-methanol instead of toluol-ethanol as organic solvent for the first step. Toluol-ethanol works as well as benzene-methanol mixture, and reduces health risks associated with the use of benzene and methanol. Here we like to briefly mention other techniques for R-cellulose extraction published in the last decades. A review of several older methods can also be found in (Green, 1963). Sheu & Chiu (1995) evaluated cellulose extraction procedures for stable carbon isotope measurements in tree ring research. The method they finally recommend involves two main steps: (1) extraction with benzene-ethanol in a Soxhlet extractor for 12 hours, followed by (2) bleaching and soaking in NaC102-CH3COOH solution. At the end of the procedure they also soak their cellulose in distilled water at 70~ for 6 hours. They recommend this method as being more accurate for 613C measurements than when the NaOH step is added, which means that they work with holocellulose instead of ~-cellulose. However, a closer look at their data shows that the accuracy of their holocellulose ~13C results is not significantly better than their Rc e l l u l o s e ~13C results. Loader et al. (1997) also presented a method which is quite similar to ours. The main differences are: (1) They use organic solvents (2" 1 toluenemethylated spirit azeotrope) only for softwood, (2) instead of milling their samples, they cut them into fine slivers (~ 40 mm), and (3) the different steps take place in an ultra-sonic bath, which enhances the removal of extracts. Recently, Brendel et al. (2000) presented a new rapid and simple method for extracting R-cellulose. With this method they obtained a very constant c~-cellulose content for samples weighing between 10 - 100 mg, indicating that the extraction protocol is highly reproducible. The advantage of this method may be its rapidity of cellulose extraction (single day process), when devoting the whole working time to it. There are also other laboratories which do not extract R-cellulose for their stable isotope analyses, but use holocellulose (Leavitt & Danzer, 1993), or lignin, instead. The method used by Leavitt & Danzer (1993) is mainly the same as ours, except that their organic solvent step involves two sub-steps (one with 2 91 toluene-ethanol and one with pure ethanol), and that they don't extract the hemicellulose with NaOH at the end of the procedure.
24.3.9 R-cellulose extraction from shrubs, herbaceous and non-vascular plants After collection, plants are dried and milled with a modified coffee-mill similar to the one described in section 24.3.1. Shrubs, annual plants and mosses have a much lower cellulose content than the 30 - 50% cellulose typically found in woody plants. For example, cellulose extraction yields are typically around 25% for Calluna vulgaris and Vaccinium uliginosum, around
Sample Homogeneity and Cellulose Extraction from Plant Tissue for Stable Isotope Analyses
517
15% for
Eriophorum vaginatum and around 10% for two non-vascular plant species, Sphagnum magellanicum and S. capillifollium.Therefore, a much larger quantity of plant material is initially required. We started cellulose extraction of sphagnum, for example, with 500 mg of dried material. To handle this larger quantity we used glass tubes of 100 mm height but with a larger outer diameter of 24 mm, compared to 10 mm for the tubes used for cellulose extraction from tree rings (Figure 24:. 3 and Picture 24.A2). The number of samples treatable in one run is then limited by the size of the Soxhlet apparatus. In our case, we can perform 7 sample extraction in parallel. Because the chemical composition differs for vascular plants and for non-vascular plants, they require a slightly different extraction treatment. For shrubs and herbaceous plants, the extraction protocol is quite similar to that described for wood. We use an extraction method adapted from that of Brenninkmeijer (1983). Lipids, resins and waxes are first dissolved by a Soxhlet extraction with a solution of toluol-ethanol (1 91). This step, not really decisive for wood, is much more important and requires more time in the case of non-woody plants. The duration of this step is not completely fixed, we stop when the toluol-ethanol solution, renewed every 15 - 20 min., remains colorless. This may take, for example, up to 14: hours for Andromeda polifolia and about 8 - 10 hours for the grass-like plants Carex pauciflora and Eriophorum vaginatum. Samples are then dried over night at 100~ The lignin fraction is removed next with a solution of NaC102 mixed with acetic acid at 70~ with proportions of 5 g NaC102 and 1.5 ml acetic acid for I 1 water, as for extraction from wood. Again, the only difference between the two protocols is the required time. The NaC102 step may last 8 - 24: hours depending on plant type. It is very important to stir the samples and to renew the solution often (every 1/2 hour and every 2 hours, respectively) during the first hours of extraction, else the solution can not reach the whole sample, due to the large volume of the sample. This "bleaching" step is repeated until the samples turn completely white. Samples are then washed under vacuum filtration and by immersion in de-ionized water until the pH of this rinsing solution reaches neutral values. At this point, samples consist only of holocellulose. The extraction of the hemicellulose fraction is performed in a 4% NaOH solution at 80~ Once again, compared to the wood protocol, this step takes much longer. As a final step, a-cellulose is washed for one day with de-ionized water at 80~ For cellulose extraction from mosses, the protocol used is based on (Wise et al., 1946) and adapted from (Price et al., 1997). The resin and lipid fractions are removed, also by Soxhlet extraction, but using three different organic solvents" first a mixture of chloroform and ethanol (2" 1), then a pure ethanol solution and finally de-ionized water. As previously mentioned, the timing of each step can not be completely fixed, and is dependent on the species. For example, the chloroform / ethanol extraction lasts 12 hours for S.cuspidatum but 20 hours for S.capillifollium. Again, the extraction is complete when the solution is clear. For the 2nd and 3rd Soxhlet extraction steps, the timing is less species dependent. Each of these steps lasts around 10 hours. Between each step samples are dried over night in an oven at 90~ The two next steps (NaC102 and NaOH) are identical to those used for vascular plants. However sphagnae are really difficult to rinse, especially after the NaC102 step. One reason may be that as
518
Chapter 24 - S. Borella, G. M6not & M. Leuenberger
long as plant structures are not completely destroyed, the large hyaline cells which constitute the sphagnum structure fill up with the various solutions, making it very hard to remove the solution within the cellular structure. Each rinsing step lasts at least one day. The de-ionized water is replaced several times and samples are washed by vacuum filtration. The use of an ultra-sonic bath may decrease this "rinsing" time. In both cases, extraction products are dried and the non-vascular plants samples are milled again prior to stable isotope measurements.
24.4 Reproducibility of the c~-cellulose extraction One can think of two independent methods for testing the reproducibility of the c~cellulose extraction. Firstly, by comparing c~-cellulose yields of different extractions of the same type of material (because of different cellulose content for wood and leaves, or for different tree species)(Brendel et al, 2000). Secondly, by comparing the isotope ratios of c~-cellulose obtained from different extractions of the same (well homogenized) wood. The first method is less precise due to sample loss which is difficult to quantify.
24.4.1 (x-cellulose yield The determination of the cellulose yield based on the measured sample weights (before and after the extraction) is only moderately accurate, since material loss is not avoidable and difficult to quantify. One way to take account of this loss is a linear regression analysis between wood and cellulose mass (Figure 24.4). Indeed, the measured cellulose mass of a wood sample is the real cellulose mass minus the lost cellulose mass: Cm = C r - n
[24.3]
where C is the cellulose mass, L the lost cellulose mass, and the subscripts m and r assign the measured and the real value, respectively. The real cellulose mass can be written as the product of the wood mass times the percentage of cellulose in the wood, which gives: Cm = ar 9 W r - L
[24.4]
where W is the wood mass, c~ is the percentage of cellulose in the wood sample, and the other symbols are defined as in equation [24.3]. Hence, a linear regression analysis of the measured cellulose mass as a function of the wood mass gives us the cellulose percentage and the mean loss, under the assumption that the loss is constant, and for each tube the
Figure 24.4 - Cellulose mass as a function of the wood mass. The real cellulose yield and the mean loss can be estimated from the slope (46.1%) and the intercept (1.05mg) of this linear correlation, respectively (see text).
Sample Homogeneityand Cellulose Extraction from Plant Tissue for Stable Isotope Analyses
519
same. This is of course not the case, but since we have no direct determination of this loss, this method gives us at least an estimate of it. To document the usefulness of such a treatment we present results of the regression analyses for fourteen extraction series obtained with our cellulose extraction method (Table 24.2), together with the mean cellulose yield calculated as the ratio of the measured cellulose and wood masses. It would be helpful for other institute that such quality tests are mentioned in publications. The results shown in Table 24.2 suggest that the cellulose yield is generally underestimated by up to 9%, when the material loss (estimated to 1 to 8 mg) is neglected. Nevertheless, one has to be careful in interpreting the results of such correlation analyses, since the loss could be mass and tube dependent, and outliers could strongly falsify the results. Moreover, a sample serie could be inhomogeneous in that earlywood and latewood, or different trees of the same species may have a significantly different cellulose content. This warning is well justified by some extreme results of Table 24.2, like the positive intercept obtained for extraction series number 10', which would correspond to a negative loss! However, this method allows us at least to get an estimate of the loss and to revise the cellulose content accordingly. We found in this way that beech and oak wood contain 40 + 5% R-cellulose. For comparison, we give the values obtained some years ago with the same extraction method for different materials by Borella (1994)" 29.0 + 2.2% for wood of a 1-year old birch; 33.8 + 1.6% for poplar wood; 37- 44% for beech wood; and 6 - 11% for birch leaves.
24.4.2 Stable isotope values In order to test the influence of the cellulose extraction on the isotopic ratio of the obtained R-cellulose, we processed the same wood sample (birch wood) several times and determined the ~13C values of the different subsamples. The results of these tests are shown in Figure 24.5. The extraction series 8 and 9 (Figure 24.5: the three points to the right) cannot be compared with others, since the extraction has been incomplete (series 8) or the material was not well homogenized (series 9). The lower 613C value of series 8 shows that the completeness of the extraction is an important criterion (see also Figure 24.2). After eliminating series 8 and 9 (Figure 24.5), we can see that the inhomogeneity due to different cellulose extractions does not exceed the uncertainty of the ~13C determination (~ 0.1%o for on-line techniques, down to ~ 0.02%0 for off-line techniques, as shown by the error bars to the right of Figure 24.5 for extremely fine - and hence homogeneous - commercial cellulose; see Chapter 23). Even more important to show the reliability of our cellulose extraction method are the results of the samples that have been extracted twice: no significant shift in 613C is observed between the single and the double extractions (compare the triangles and the squares in extractions number 2 and 9, in Figure 24.5). The i513C gradients following the several extraction steps (Figure 24.2) also clearly documents the importance of a complete extraction. Indeed, the difference in 613C between pure a-cellulose and wood is of the same order of magnitude as the year-to-year variations in tree ring cellulose (see for example Borella, 1998). Even the shift caused by skipping one or two NaC102 steps is
520 comparable to the 613C measurement reproducibility. However, the flattening of the trend between the 2nd and 4th NaC102 steps confirms the results of Figure 24.5, that the R-cellulose extraction is complete. It is important to note that these conclusions must not be true for 6180 and 62H. Indeed, Brenninkmeijer (1983) found that a prolonged cellulose extraction reduces the cellulose yield and could produce a shift in the oxygen and hydrogen isotopic composition of the remaining R-cellulose. However, prelimnary results from analogous tests for 6180 show no significant difference between two R-cellulose subsamples extracted from the same well homogenized wood sample.
Chapter 24 - S. Borella, G. M6not & M. Leuenberger
Figure 24.5 - 613C of different extractions of our home-made standard cellulose (,,WSL", wood from a 1-year old birch), average values for each extraction and measurement technique: off-line (full symbols) or on-line (open symbols) techniques (see chapter from Saurer). The error bars represent the standard deviation (1 o), as determined from multiple measurements; symbols without error bar represent a single measurement. The triangles represent cellulose that has been extracted twice. The circles represent cellulose extracted from coarse milled wood. The error bars at the right handside of the figure represent the standard error of the single measurement, as determined for the off-line (solid line) and on-line (dashed line) method with extremely fine (< 0.01mm) commercial cellulose. The different symbols (circles and triangles, extraction numbers 1, 2, and 9) are shifted to the right to enhance the clearness of the figure. The ,,WSL" cellulose from extractions 7 and 10 has not been analyzed for ~13C.
Furthermore, we like to note here that for certain applications requiring not the highest possible reproducibility it is possible to measure the isotopic composition directly on wood. Year to year variations in lignin and holocellulose are approximately the same (Mazany et al., 1980), although lignin seems to be produced later during the growing season than cellulose by the trees (Wilson & Grinsted, 1977; Tans et al., 1978). At least for broadleaved trees, we could show that it is not necessary to extract cellulose, nor lignin prior to ~13C measurements; wood reflects the same year to year variations within the measurement uncertainties (Borella et al., 1998). However, similar results for ~180 values were not so conclusive. These findings have not to be true for other tree species. 24.5 Conclusions
The method for ~x-cellulose extraction from wood presented in this chapter is highly reproducible in regard to the carbon and oxygen isotopic composition. Moreover, it allows a high sample throughput (up to 120 samples in two weeks are possible, excluding the time necessary for milling). In order to achieve reproducibility of 0.1%o for 613C and 0.2%o for ~i180, it is important to finely mill the samples prior to the extraction, and to perform the NaC102 steps until the samples are white. We found
Sample Homogeneity and Cellulose Extraction from Plant Tissue for Stable Isotope Analyses
521
that a w o o d g r a i n size of a p p r o x i m a t e l y 0 . 1 5 m m h a s to be r e a c h e d , if s u b s a m p l e s of 1 - 1.5 m g are a n a l y z e d . M o r e o v e r , fine milling of the w o o d is n e c e s s a r y to p e r f o r m a c o m p l e t e c~-cellulose extraction, a n d e v e n l o w c o n c e n t r a t i o n of lignin in the c~-cellulose can l o w e r the ~)13C b y u p to 0.2%0.
Appendix
Picture 24.A1 - Coffee-mill used for milling the wood samples. We can see the double blade in the ,,body" of the mill (above), and the additionally mounted deflection plate in the cover (below).
Picture 24.A2- Filter glass tube used for the extraction of cellulose. The ground up glass at the bottom of the tube serves as filter; it lets the solution through, but not the wood (respectively the cellulose). The tube is 10cm high, has an outer diameter of 12mm and an inner diameter of 10mm. (For the dimensions, see also Figure 24.3). The number seen at the top of the glass has been engraved to avoid any hazard of mixing samples. [Picture 24.A2 on next page!]
522
Chapter 24 - S. Borella, G. M4not & M. Leuenberger
[Picture 24.A2 - see caption on f o r m e r page]
Picture 24.A3 - Teflon holder for the filter tubes (see Picture 24.A2) used for cellulose extraction. The holder has a diameter of 90mm and height of 35mm. There is place for 30 filter tubes. The holes for the tubes are a little smaller at the bottom to avoid the tubes to go through. Only the central hole goes down to the bottom without diameter reduction. The holder with the 30 sample glasses can be hold by putting the metal piece shown at the bottom of the picture into the central hole. Pressing the two long metal bars together permits the teflon holder to sit on the two curved parts of the metal piece, seen at the bottom left of the picture
Handbook of Stable Isotope Analytical Techniques, Volume 1 P.A. de Groot (Editor) 9 2004 Elsevier B.V. All fights reserved.
CHAPTER 25 Analytical Methods for Silicon Isotope Determinations Ding Tiping Institute of Mineral Resources, Chinese Academy of Geological Sciences, Beijing 100037, P. R. China e-mail: [email protected]
25.1 Introduction Four isotopes of silicon exist in the natural environment: 28Si, 29Si, 30Si and 32Si. The first 3 isotopes are stable isotopes and the last one is radiogenic. The relative abundance of 28Si, 29Si and 30Si is 92.23%, 4.67% and 3.10%, respectively (Barnes et al., 1975). By the 1920's, all three stable silicon isotopes had been discovered. Mass-spectrometric studies on silicon isotope variation in the natural environment started in the 1950's (Reynolds & Verhoogen, 1953; Allenby, 1954). In the 1970's, extensive studies on silicon isotope compositions of meteorites and rocks were made (Epstein & Taylor, 1970; Yeh & Epstein, 1978; Clayton et al., 1978, Clayton 1986). Douthitt (1982) reported a number of data on silicon isotope composition of terrestrial samples. Since 1988, the author and his co-workers have carried out a long-term studies on the silicon isotope composition of a variety of natural materials (Ding et al., 1988, 1990; Ding 1991; Ding et al., 1994, 1998; Jiang et al. 1992, 1993; Wu et al. 1997; Song & Ding, 1990). Their results were summarised in a book entitled in" Silicon Isotope Geochemistry"(Ding et al., 1996). As for the isotopes of other elements, silicon isotope compositions of a sample (Sa) are expressed as the 6 values related to a standard (St), i. e.:
629Si (%o)= [(29Si/28Si)sa/(29Si/28Si)st- 1] x 103 630Si (%o)= [(30Si/28Si)sa/(30Si/28Si)st- 1] x 103
[25.1] [25.2]
The standards used for silicon isotopes are described in section 25.2.4. There are two methods presently available for silicon isotope analyses" gas source isotope ratio mass-spectrometry (IRMS), and secondary ion microprobe mass-spectrometry (SIMS). The former is a routine method and suitable for all kinds of samples. It provides excellent precision, but needs relatively large samples. SIMS is able to determine Si isotopes on a small spatial scale but its precision is low, so that it is only applicable to the study of meteorites at present.
524
Chapter 25 - T. Ding
25.2 Gas source mass-spectrometric analysis of silicon isotopes Reynolds & Verhoogen (1953) firstly reported a mass-spectrometric analysis of silicon isotope ratios. Since then many modifications to the method have appeared. The common aspect of these methods is that samples of different types are transformed to SiF4 before their silicon isotope compositions are determined in a gas-source isotope ratio mass-spectrometer. There are two main reasons for using SiF4 as the gas for mass-spectrometric analysis. Firstly, SiF4 is relatively easy to prepare and is stable under laboratory conditions. Secondly, there is no need to make isotope corrections with SiF4, as fluorine has only one stable isotope. Three methods for SiF4 preparation are presently used: a) The BaSiF6 decomposition method (Reynolds & Verhoogen, 1953); b) The direct fluorination method using F2 + HF as fluorination reagent (Taylor & Epstein, 1962); c) The direct fluorination method using BrF5 as a fluorination reagent (Clayton & Mayeda,1963). A brief discussion on the advantages and disadvantages of different methods will be given in the following section. The method developed in Beijing will be described in detail, which is an improvement on the method of Clayton & Mayeda (1963).
25.2.1 A brief introduction of SiF4 preparation methods for mass-spectrometry In order to make reliable isotope ratio measurements, the gas samples produced for isotope determination must satisfy a basic requirement, i. e., the gas sample must reproducibly represent the isotopic composition of the original sample. There are two ways to satisfy this basic requirement. The first is to quantitatively transform the element in the original sample to a gas for mass analysis without any observable contamination. The second is to ensure that there is a fixed relationship between the isotopic composition of the gas sample and that of the original sample. The former is the most reliable of two and has been adopted in the preparation of SiF4 for isotope analysis. In this case, the key points are to guarantee 100% yield and to protect the sample from contamination by foreign silicon and other impurities that might interfere with the isotope analysis. 25.2.1.1 BaSiF6 method This method was firstly developed by Reynolds & Verhoogen (1953) and consists of two basic steps. The first step is to convert the silicon in the sample to BaSiF6, and the second step is to decompose the BaSiF6 to form SiF4 for mass-spectrometric analysis. The complete procedure is described below. The chemical reaction begins with the conversion of silicon in mineral specimen to SiO2, which is done through standard analytical procedures. Then 0.5 g of this silica is mixed with 2.5 g of sodium carbonate. This well-mixed combination is fused into a 20 ml platinum crucible, yielding a clear melt. After cooling, the solidified melt is extracted as a pellet by pressing the crucible gently around the base. This pellet dissolves in 15 ml of water to give a clear solution. To this is added 30 ml of concentrated perchloric acid and the solution is heated on a hot plate (swirling constantly to avoid bumping) until copious dense fumes of perchloric acid have boiled off for ten min-
AnalyticalMethodsfor SiliconIsotopeDeterminations
525
utes. When cooled, the solution solidifies to a gelatinous mass. It is then taken up in 200 ml of water and brought to boil. The precipitate of gelatinous silica is then filtered and transferred to a polyethylene beaker with 60 ml of water. Concentrated HF is added dropwise until the solution just clears reflecting total conversion of the silica gel to fluorosilicic acid, H2SiF6. To the clear solution 20 mL of a solution containing 0.2 g of BaC12.2H20 per mL is added. Under these conditions BaSiF6 precipitates while BaF2 does not. To ensure complete precipitation of the silicon, the beaker is allowed to stand overnight before filtering and drying the BaSiF6 at 110~ A series of analytical tests on the above procedure were done to determine the chemical yield. A typical test consisted of starting with an accurately weighed amount of silica, precipitating BaSiF6 with a calibrated barium solution, weighing the precipitate, and extracting and weighing the unreacted barium in the filtrate as BaSO4. Within the accuracy of these tests (+ 2%) the following observations were made" a. BaSiF6 is the correct formula of the barium chloride precipitate when the above procedure is adhered to strictly, i. e., occluded water and coprecipitated BaF2 are not present in the final precipitate. b. The yield of BaSiF6 always exceeds 95% of the theoretical yield calculated from the amount of silica added at the start of the experiment. This yield was sufficiently satisfactory for the method used in preparing samples for mass spectrometric analysis. The final step is to weigh out 150 mg of the dried precipitate into a short length of 10 mm Pyrex tubing, closed at one end. This tube is sealed into a 75 ml Pyrex sample bulb fitted with a stopcock. The bulb is evacuated with a diffusion pump for 12 hours or more, flaming gently from time to time in order to facilitate the removal of water vapour. Then with the stopcock closed, the BaSiF6 is decomposed by heating the bottom of the 10 mm tube vigorously with an oxygen-gas flame until there is no further change visible in the solid material. The temperature of the glass is brought to such a point in this heating that the glass softens and begins to melt around the solid BaF2 residue. Lastly, the lower half of the 10 mm tube is sealed with a torch from the rest of the sample bulb. This can remove the BaF2 residue and prevent any recombination with the silicon that might be isotopically selective. Control experiments show that the percentage decomposition of the BaSiF6 by this scheme is 90%. The SiF4 obtained from above steps in placed directly into the mass-spectrometer for isotope analysis. This method uses procedures relatively common in wet chemistry. The equipment used in this method is simple and no elaborate techniques are needed. Correct and meticulous operation of this method yields reliable analytical results. However, there are problems with this method that need to be addressed. For example, it needs relatively large amounts of sample (gram level). The procedure is also time consuming and requires great care to prevent silicon isotope fractionation if thermal decomposition of BaSiF6 is not complete. It is also possible that the produced SiF4 may change its Si isotope composition by reacting with the glass during the high temperature decomposition stage. In view of these disadvantages, the BaSiF6 method is rarely used in routine analysis of silicon isotopes except in some special cases. For example, De
526
Chapter 25 - T. Ding
Bievre and his colleagues have been using this method to calibrate silicon isotope ratios of their silicon isotope reference materials and determine molar mass (atomic weight) of silicon (De Bi~vre et al., 1994, De Bi6vre & Valkiers, 1994 and 1995). In their studies, Cs2SiF6 was chosen as the form of silicon compound to prepare synthetic mixtures used in calibration of absolute isotope ratios. Then the Cs2SiF6 was converted to BaSiF6, which was used in preparation of SiF4 for mass-spectrometry analyses.
25.2.1.2. Direct preparation of SiF4 The main difference between this method and that described above is that no intermediate step of BaSiF6 preparation is needed. Instead, the silicon-bearing rocks and minerals are reacted directly with a fluorinating reagent to produce SiF4. Baertschi & Silverman (1951) used firstly direct fluorination for sample preparation for stable isotope analysis. Allenby (1954) was the first to use the direct flourination method to prepare SiF4 for Si isotope analysis. He used HF as a fluorination reagent. However, due to the inconvenience of using HF and the unsatisfactory results, this reagent is no longer used. The most widely used methods today are those developed by Taylor & Epstein (1962) and Clayton & Mayeda (1963). Taylor & Epstein (1962) used a mixture of F2 and HF as the fluorination reagent in their experiments to extract 02 from silicate rocks for oxygen isotope analysis. Using this method the silicon in silicate was converted quantitatively to SiF4. However, the SiF4 products were initially pumped away as waste as the main aim of their studies was to prepare sample for oxygen isotope analysis. Epstein & Taylor (1970) made some modifications to this method to allow collection of SiF4. This method was then used in silicon isotope analyses of lunar samples and meteorites (Epstein & Taylor, 1970). Clayton & Mayeda (1963) developed an alternative method for preparation of 02 from silicates and oxides for oxygen isotope analyses by using BrF5 as the fluorination reagent. BrF5 has a low vapour pressure (about 33342 Pa at room temperature), so it is easily transferred and handled by evaporation and condensation in a metal vacuum system. Furthermore, BrF5 is an active fluorination compound that reacts with all kinds of silicates and oxides just as effectively as a mixture of F2 and HF. In addition, the wastes and residues of the reaction process are easily removed from the apparatus. For these reasons the BrF5 fluorination method is most popular in preparing samples for oxygen isotope analyses of silicates and oxides. Clayton et al. (1974) made some innovations to their extraction line by adding cryogenic parts for collecting and purifying SiF4. Subsequently, they have largely applied this technique to silicon isotope determination of meteorites and lunar rocks (see Molini-Velsko et al., 1986). Compared with the BaSiF6 method, the direct preparation method of SiF4 reduces the sample size to mg level, simplifies the operation procedure, increases the speed of sample preparation, and improves the precision of the analyses. Hence, this method has replaced the BaSiF6 method and has become the routine method for preparation
527
Analytical Methods for Silicon Isotope Determinations
of SiF4 for silicon isotope analysis. It should be noted that only pure silica and silicate rocks and minerals could be directly fluorinated in the manner described above. For samples containing impurities of C, S and B, pre-treatment is needed before fluorination.
25.2.2 The SiF4 preparation method used in the institute of mineral resources, Beijing 25.2.2.1 The sample preparation line The sample preparation line used in Beijing is shown in Figure 25.1. It is similar to the line of Clayton & Mayeda (1963), but improvements have been made in several key aspects. The whole line is made of metal. The left side of Figure 25.1 shows the section involved in SiF4 preparation and extraction. It consists of reaction vessels, the BrF5 storage vessel, a cold trap for waste, cold traps for separation, pressure gauge, Whity valves, connecting tubes, tubes for input of Ar and output of waste. The right side of Figure 25.1 shows the SiF4 purification and collection section, which includes a copper tube containing Zn particles (0.5 - 1.0 mm size), cold traps, resistance gauge, Whity valves, connecting tubes and a sample tube.
~
v9 LV V1
:~-1V2
Ar
l
V3
~-I V19
[ [
T1 V4
V5
T2
T3
BrF5 V6
V8
l Vll ~~ V12
HV
'7-',
~qV10
T l |
V20 CuT
V13~
l l V14 ~'e V15 "
v
S~F V21
V
'v22 0 T4 R1
R2
R3
R4
R5
Figure 25.1 - SiF4 sample preparation line.
R6
LV
V24
THV
ST
".m
T5
528
Chapter 25 - T. Ding
The reaction vessels are made of pure nickel. The CuT is made of 2 m long pure copper and is filled with pure Zn particles. The sample tube is made of Pyrex glass. Other tubes, cold traps, storage vessels and the pressure gauge are all made of stainless steel. All the valves are Whity two way metal ball valves. The low vacuum of the system is obtained by using mechanical pumps and the high vacuum is generated using a turbo-molecular pump. The dynamic vacuum of the system is about 6.6 x 10-4 Pa, and the static vacuum of the system can be kept in the range of 3~4 Pa in the 24 hours after pumping has stopped. The major differences between this line with that of Clayton & Mayeda (1963) are that a CuT tube is used for SiF4 purification and a device for waste disposal has been added to this line.
25.2.2. 2 Sample preparation method The sample preparation method adopted in Beijing is the BrF5 method similar to that of Clayton et al. (1963), i. e., the silicon-bearing sample is reacted with BrF5 to produce SiF4. SiO2 + 2BrF5 = SiF4 + 2BrF3 + 0 2 KA1Si308 + 8BrF5 = KF + A1F3 + 3SiF4 + 402 + 8BrF3
[25.3] [25.4]
However, in comparison with the method of Clayton et al. (1963), the method used in Beijing is superior in several aspects; namely in the purification of BrF5 and the SiF4 products and the treatment of waste.
Figure 25.2 - Mass-scanning diagram showing the impurities in the BrF5 agent, which can be frozen by liquid N2 but not by dry ice-acetone.
Analytical Methods for Silicon Isotope Determinations
529
Purification of BrF5 reagent The BrF5 reagent (produced by the Wuhan Institute of Chemical Products, China) contains some impurities of CF4, SiF4 and SF6 (Figure 25.2). These impurities will interfere with the isotope determination of SiF4, so purification of the BrF5 is necessary. The method for BrF5 purification consists of cryogenic vaporization and condensation. At 1 atm (101325 Pa) the melting and sublimating points of SiF4 are -90~ and 95.1~ respectively. The melting and boiling points of SF6 are -80.4~ and -63.7~ respectively; and the melting and the boiling points of CF4 are -184~ and -128~ respectively. In contrast, the melting and boiling points of BrF5 are -61.3~ and 40.5~ respectively. Under vacuum, these temperatures will reduce, but their relative order does not change. Therefore BrF5 can be purified through cryogenic vaporization and condensation. Dry ice-acetone (-80~ and ethanol-liquid nitrogen mixtures (its temperature is adjustable and can be as low as -100~ were tested as cryogenic liquids. At the temperatures of these cryogenic liquids, BrF5 remains in the solid phase, but the impurities, such as SiF4, SF6 and CF4, are present in the vapor phase and can be pumped to waste. In general the ethanol-liquid nitrogen mixture of-70~ is used as the cryogenic liquid and the BrF5 reagent is purified several times to ensure good purification. This is a key step to improve the precision of the Si isotope analyses.
Pre-treatment of samples The SiF4 preparation method can be used for any sample containing silicon for isotopic analysis. However, different pre-treatment procedures are needed for the different types of samples. A. For silica and high purity silicates of containing little or no C, S and B, no pretreatment is needed. The sample can be fluorinated directly after it is grounded. B. For samples containing significant contents of S, C and B (>1%), chemical pretreatment of the sample is necessai:y. a. When the impurities in the sample are carbonates or acid soluble sulphides, the HC1 dissolution method is used. b. When the sample contains only graphite or organic carbon, high temperature oxidation and evaporation are used to remove the contaminants. c. When the sample contains boron compounds, sulfates or sulphides insoluble in HC1, wet chemical pre-treatment is required. In this instance there are several methods for the preparation of SiO2. The procedure adopted at Beijing is as follows; the sample of 0.1--0.5 g (depending on the Si content) is mixed with NaOH and Na202 and is placed in graphite crucible. This mixture is melted in a furnace and the cooled melting products are dissolved with a 1"1 HC1 solution. This solution is dried by low temperature evaporation. Concentrated HC1 of 10 mL is then added to the dried product. The solution is warmed for 10 minutes in a water bath and animal glue is added into the solution to precipitate SiO2. The solution is filtered and the precipitates rinsed several times with 1% HC1 followed by distilled water. The filter paper containing the precipitates is placed in a platinum crucible, and ashed at 700~ then calcined at 1000~ The SiO2 obtained
530
Chapter 25 - T. Ding
is then fluorinated.
SiF4 preparation and purification A. Sample loading and reaction a. Sample loading: The Ni reaction vessels are filled with Ar after the treatment with BrF5 and left over residues of last reaction. The reaction vessels are opened, the solid products of the previous fluorination are turned over, and the new dried sample entered. Then the reaction vessels are connected back to the vacuum line. The sample loading is normally 5 mg of silicon, but it can be reduced as small as I mg of silicon. b.Adding BrFs" Cool the waste cold trap T1 with liquid nitrogen and evacuate the manifold and reaction vessels by mechanical pump. Waiting for a few minutes, cool the separation traps T2 and T3 with liquid nitrogen, pump the line to high vacuum by a turbo-molecular pump. When the ion gauge near molecular pump show a vacuum of ~ 2 x 10-3 Pa (indicating the reaction vessels are properly connected), the BrF5 is added. Let BrF5 diffuse into the manifold of the line and reach its saturation vapor pressure (33342 Pa at room temperature). Freeze the BrF5 in the manifold into R1 (at liquid nitrogen temperature). Fill the manifolds with BrF5 again. Freeze BrF5 into R2. Repeat above steps to transfer BrF5 into R3, R4, R5 and R6 one by one. Then pump the manifold and reaction vessels with mechanical pump for ten minutes while these vessels are frozen with liquid N2. Then close V3 and open V17, pump the manifold and reaction vessels with turbo-molecular pump. When the ion gauge by the molecular pump indicates a vacuum higher than 2 x 10-3 Pa, close Vii-V16, remove the liquid nitrogen cups from the reaction vessels, and warm in a water bath. In general, the amount of BrF5 added is 4~5 times more than that needed for complete reaction. c.Reaction: A cooling water device is placed around the upper part of R1-R6, just below the valves. These vessels are then heated with electric furnaces. The temperatures of the furnaces are adjusted using adjustable transformers. The reaction temperature varies according to the type of sample. For samples which react easil3r such as" quartz, silica sinter, feldspars, micas and granite rocks, a temperature of 550~600~ is used. For samples that do not react so easily, such as olivine, pyroxene, actinolite and topaz, a temperature of 650~700~ is needed. The reaction time is normally set to be more than 14h (overnight). The above procedure is almost the same as that for oxygen isotope analyses, and hence, sample preparation for Si and O isotope analyses are often done simultaneously. BO Extraction and purification of SiF4
After the reaction is completed, the electric furnaces are switched off, removed from the reaction vessels, and cooled with cold water. Freeze the reaction vessels (R1-R6) and cold traps (T1 -T3) with liquid nitrogen cups. Then open V11-V16, pump the line with the turbo-molecular pump to release 02. In the case of simultaneous preparation of 02 and SiF4 for isotope analysis, the extracting procedure is different. Instead of pumping out the 02, the 02 in each
Analytical Methods for Silicon Isotope Determinations
531
reaction vessel is extracted separately and converted to CO2 before extraction of the SiF4. In the processes of SiF4 extraction, three stages of distillation and condensation at dry ice-acetone and liquid nitrogen temperatures are used for separating SiF4 from 02, N2, BrF5 and BrF3. Through these processes, the SiF4 obtained will contain no impurities of 02 and N2, but may contain trace amounts of BrF5 and other active fluorine compounds still. These impurities will react with the glass and grease of the sample tube to produce contaminants, such as SiF4 and CF4, which will interfere with the isotopic determination. For this reason, a new step for SiF4 purification has been added by using a Cu tube containing pure Zn particles (CUT). Heated Zn particles (50~176 react with BrF5 and other active fluorine compounds (except SiF4) to form ZnF2 and ZnBr2. This step is simple and very efficient. Very pure SiF4 gas can be prepared, as shown in the mass-scanning plot in Figure 25.3. Freeze SiF4 into the sample tube at liquid nitrogen temperature. Then the sample tubes are brought to the mass-spectrometer for isotopic analysis. A set of 6 samples can be prepared each day using this method. CQ Waste treatment After completing preparation and purification of SiF4, the waste left in the reaction vessels and cold traps are treated as follows. Condense the BrF5 and other wastes left in reaction vessels into T1. Fifteen minutes later fill Ar gas into T1 through V4, V5 and V3. Remove the liquid N2 cup from T1 and warm it with a water bath to defrost the BrF5 and other wastes. Open V3 and V2, blow the wastes with Ar into a bucket containing Ca(OH)2 solution. The BrF5 and other active fluorine compounds in the waste gas react with Ca(OH)2 to produce CaF2, CaBr2 and 02. After the reaction in the bucket has stopped, keep Ar blowing for 5 more minutes. Then close the Ar tank, V4 and V2 in order. This waste treatment is simple, secure, and pollution free.
Figure 25.3 - Mass-scanning diagram of purified SiF4gas.
532
Chapter 25 - T. Ding
25.2.3 Laser probe extraction method for SiF4 preparation The laser probe extraction method for SiF4 preparation was reported by De La Rocha et al. (1996). Their sample preparation line is shown in Figure 25.4:. It consists of two portions" a metal section for fluorination and a glass section for yield determination and sample collection. Purified fluorine for sample reaction is generated inside the vacuum line by heating potassium hexafluonickelate, which decomposes at ~350~ to produce F2. A tank of F2 provides the commercial fluorine for charging the pure fluorine generator. A tank of N2 plumbed into the metal section is used to pressurize the sample chamber for sample loading and melting. Purified silica samples containing 15-100 mmol of Si, are loaded into 0.5 cm deep wells drilled into a cylindrical nickel plate. After sample loading, the reaction chamber and vacuum line are evacuated for several hours. Pieces of pure quartz can be fluorinated directly, but the finely powdered silica that has been purified through precipitation may sputter during fluorination. Sputtering is eliminated by melting the powder under I atm of N2 into lumps of glass that then fluorinate in a controlled fashion. A CO2 laser (22 W) coaxial with a He-Ne sighting laser (3 mW) is mounted on a motorized x-y translation stage and set to fire through a BaF2 window on the top of reaction chamber. The CO2 laser is set for a beam width of 0.8 ms and a pulse period of 0.9 ms. The intensity of the laser beam at this setting is raised from zero to near maximum levels slowly during lasing. Melting the silica with a continuous beam is avoided as partial vaporization of the sample can occur. After the melting of samples the reaction chamber is evacuated and pumped for several hours. Samples are fluorinated under 0.1 atm of purified F2 that has been passed through a liquid nitrogen trap (cold trap 1 on Figure 25.4). Silica samples react with F2 upon being heated by the laser set to fire a continuous beam: SiO2 + 2F2 --* SiF4 + 02
[25.5]
Figure 25.4 - Schematic of SiF4 preparation line with laser microprobe device (After De La Rocha et al., 1996).
Analytical Methods for Silicon Isotope Determinations
533
Lasing of 1-3 mg samples takes approximately 10-20 min, during which the intensity of the continuous laser beam is varied between zero and maximum intensity to maintain controlled, continuous fluorination. When lasing has been completed, the resulting SiF4 is collected in a coil cooled to -195~ with liquid nitrogen (cold trap 2 on Figure 25.4). 02 and F2 are then pumped away through the coil. Any water present either in the silica or in the reaction chamber during fluorination will form HE which may in turn form SiF4 by etching the walls of the glass section of the line. Traps cooled to liquid nitrogen temperature will collect both SiF4 and HE and so will not serve to separate them. For effecting a separation, the SiF4 (and any HF) is transferred from cold trap 2 into the variable-temperature trap, which has cooled to at least -183~ with liquid nitrogen. Remaining noncondensibles are pumped away through the trap. The variable-temperature trap is then heated t o 140~ distilling SiF4 to the glass side of the line, where it is collected at liquid nitrogen temperature in the multitrap. Tests indicate that SiF4 quantitatively distils out of the trap at this temperature but HF remains behind. The liquid nitrogen on the multitrap, where the sample is frozen, is replaced by a dry-ice-2-propanol slush. Distil SiF4 from the multitrap to the capacitance manometer, where the micromoles of sample gas are determined. Samples are distilled into borosilicate tubes, sealed, and are then ready for mass-spectrometer analyses.
25.2.4 Standards for silicon isotopes So far 6 samples have been used or proposed as silicon isotope standards: i. e. NBS-28, Rose Quartz, IRMM 017, IRMM 018, GBW 04421 and GBW04422. NBS-28 is a sample of quartz sand distributed by National Bureau of Standard of the United States (now NIST) as an oxygen isotope reference materials for silicates. It has been used in laboratories of the Chicago University, the Institute of Mineral Resources, CAGS (Beijing) and other institutions as a silicon isotope reference material. Rose Quartz is a quartz sample that has been used at the California Institute of Technology as a reference material for oxygen and silicon isotopes. IRMM 017 (a silicon metal) and IRMM 018 (a sample of silica) are two samples preparedat the Institute for Reference Materials and Measurements (IRMM) (De Bibvre et. al. 1994a). GBW 04421 and GBW 04422 are national reference materials for silicon isotopes in China. The 630SINBS-28 values of these reference materials are listed in Table 25.1. De Bi6vre et al. (1994a) determined the absolute isotopic ratios of silicon in IRMM 017 and IRMM 018. IRMM 017 has 29Si/28Si value of 0.0507715(66) and 30Si/28Si ratio of 0.0334889(78). IRMM 018 has 29Si/28Si value of 0.0508442(48) and 30Si/28Si ratio of 0.0335851(66).
25.2.5 Mass-spectrometry analysis The SiF4 obtained from the above process is analyzed in a gas mass-spectrometer for its isotopic composition. At Beijing a MAT-251 EM mass-spectrometer with multiple collectors is used.
534
Chapter 25 - T. Ding
Table 25.1 - The ~)30SiNBS-28values of some reference materials for silicon isotopes Sample No.
Description of sample
NBS-28 Rose Quartz IRMM-17 IRMM-18 GBW-04421 GBW-04422
Quartz Quartz Silicon metal Quartz Quartz sand SiO2 chemical agewnt
630SINBS-28 (%o) 0 -0.28 -1.3 0.0 -0.02 -2.68
Reference
Molini-Velsko et al. (1986) This study This study Wan et al. (1997) Wan et al. (1997)
As mentioned above, silicon has 3 stable isotopes" 28Si, 29Si and 30Si. The isotopic compositions are commonly expressed as ~)30Si and ~)29Si values. However, in routine investigations of terrestrial samples, only 630Si value is measured. 629Si determinations are done in special cases, w h e n meteorite Si isotope anomalies or n o n - m a s s d e p e n d e n t isotope fractionation are studied. The most abundant SiF3 + ions are normally used in the isotopic determination. Two collectors are simultaneously used to collect 28SIF3+ and 30SiF3+ at mass numbers 85 and 87. A 10 kV accelerating voltage is used with a magnetic field intensity of 0.5587 T. The ion beam intensity is measured on the 6 V scale. The working standard used is the SiF4 prepared from NBS-28. The results are expressed as 630Si values related to NBS28. As fluorine has only one isotope, there is no need to make an isotopic correction for the 630Si determination of SiF3 +. Six sets of data are collected for each analysis. The precision of the mass-spectrometry analysis is + 0.05 to + 0.10 %o (lo). During mass-spectrometry, the 630Si values will be altered if the SiF4 is not pure or if there are trace amounts of air leaking into the sample tube. If the carbon in the sample is not removed out before fluorination, COF3 + ions will appear in the mass-spectrum and cause unusual low ~)30Si values. In Beijing the present arrangement of collectors in the mass-spectrometer does not allow simultaneous collect ions of SiF3 + of masses 85, 86 and 87, so that 630Si and ~)29Si can not be determined simultaneously. When it is needed, ~)29Si is determined in the form SiF4 +, at masses of 104 and 105. The intensity of SiF4 + is m u c h lower than that of SiF3 +, but reliable data can still be obtained.
25.2.6 Analytical precision Several samples, such as the NBS-28, GBW 04421 and a sample of SiO2 prepared from a diatom (88-43) have been fluorinated and measured repeatedly at Beijing. In addition, a meteorite sample and a tektite sample have been analyzed in duplicate; the results are listed in Table 25.2. From these results the standard deviations of the analyses is estimated to be _+0.03%o (lo) to _+0.10%o (lo).
535
Analytical Methods for Silicon Isotope Determinations Table 25.2 -
~)30SiNBS-28
values and analytic uncertainties for several samples. Average and uncertainty of ~)30SiNBS-28(%o)
Sample No.
Sample type
~)30SiNBS-28(%o)
NBS-28
Quartz sand
GWB-04421
Quartzite
-0.04 + 0.07
88-43
SiO2 prepared from a diatom Thailand tektite
0.03, 0.03, 0.02, 0.05, -0.04,-0.01, 0.00 0.0, 0.0,-0.01, 0.04, -0.13, -0.11, -0.09 -1.18, -0.94, -1.05, -0.06 -0.23, -0.29
-0.26 + 0.03
88-44
0.0 + 0.06
-1.06 + 0.1
Since 1988, all analyses made in Beijing have been monitored by the analyses of NBS-28, and the precisions have always been better than _+0.10%o (lo).
25.3 Ion Microprobe mass-spectrometer analyses The secondary ion microprobe mass-spectrometer (SIMS) developed in the 1970"s differs from traditional instruments, such as gas isotope mass-spectrometers and thermal ionization mass-spectrometers. It measures secondary ions instead of the primary ions. In the ion microprobe, secondary ions are produced by ion bombardment on the surface of the sample to be analyzed. Focusing of the bombarding ions into a fine beam allows the in situ analysis of individual mineral phases. The main advantages of the technique are high sensitivity~ small sample size and measurement of elements that are difficult to analyze by other techniques. However, there are also problems with this technique and these include interference of molecular ions with the atomic ions of interest, large variation of the ionization efficiencies of different elements and a matrix-effect. Several kinds of ion microprobes, such as the Cameca IMS 3f and the SHRIMP (Sensitive High Mass Resolution Ion Microprobe) have been applied to isotopic measurements on extraterrestrial material and on terrestrial rocks. The application includes U-Pb dating of individual zircons and the study of the distribution of Pb, S, H, C, O, Mg, Si, Ca and Ti isotopes in a number of minerals. A schematic diagram of a SIMS is shown in Figure 25.5. The general aspects of isotopic measurement with SIMS are described by McKeegan et al. (1985) and Clement & Compston (1989) and Chapter 30 of this book. Silicon isotopes have been measured as positive (Huneke et al., 1983; Clayton et al., 1991; McKeegan et al., 1985) and negative ions (Zinner et al., 1987). The silicon signal per primary beam current for a given sample is higher than for negative secondary ions. However, since the sputter rate (number of sputtered atoms per incident ion) of Cs + is roughly 40 times that of O-, the ionization efficiency (number of secondary ions
536
Chapter 25 - T. Ding
Figure 25.5 - Schematic diagram of an ion microprobe mass-spectrometer (After Eldridge et al., 1989). P - pump.
per sputtered atom of a given species) is much higher for positive ions. The instrumental mass fractionation, however, is much larger for positive (~32%0/amu) than for negative ions (~8%o/amu; amu is atomic mass unit). Stone et al. (1991) measured the silicon isotope compositions of SiC grains from carboniferous chondrites and enstatite chondrites with a Panurge IMS-3f ion microprobe mass spectrometer. For the majority of the analyses, silicon was analyzed as positive secondary ions, produced by a 0.2~0.5 nA 160- primary beam rastered over a 20 mm square region surrounding each grain. The low sputter rate, <0.5mm/h, ensured preservation of material for C and N isotope studies. An 8mm field aperture in the sample image plane transmitted only those ions originating at the centre of the rastered field, suppressing background Si§ emission. A mass resolving power (MRP) greater than 3000 is required to effectively separate hydride interferences (28Sill)+ and (29Sill)+ from isobaric Si+ peaks. However, in order to maximize secondary ion transmission when analysing individual SiC grains, which yielded lower (28Sill)+/29Si+ ratio, the resolving power was reduced to ~2400.
Analytical Methods for Silicon Isotope Determinations
537
Additional Si isotopic measurements were made in conjunction with subsequent C isotopic analyses. In these runs, Si was analyzed as Si- ions, sputtered by a ~0.1 nA Cs § primary ion beam defocused over a 20 mm diameter area around the sample grain. The higher MRP of ~3500 required for concurrent C isotopic measurements ensured that the Si- peaks were free of isobaric interference. As for positive ion analyses, a field aperture was employed to minimize Si- background contribution. The precisions of measurements at present are: < + 2%o for positive secondary ions and < + 4%0 for negative secondary ions.
25.4 Concluding remarks The excellent work by many researchers over the past tens of years, has resulted in a remarkable progress in the field of silicon isotope analysis. The SiF4 method has become the conventional extraction method with precision of + 0.1%o. The laser fluorination method has been successfully used in the sample preparation for silicon isotope analyses of micro sample. The secondary ion probe mass spectrometry also became a powerful tool for in situ micro-field analyses of silicon isotopes, although the precision of measurement is still relatively low (+ 2%0 to + 4%0). These methods will be improved in the future, providing better and better technique base for silicon isotopic study.
Acknowledgements Thanks are going to my colleagues Wan D.F., Li Y. H., and Li J.C. for their co-operation in the methodology study of silicon isotope determination. I wish to express my sincere thanks to the National Natural Science Foundation of China, the Ministry of Science and Technology, and the Ministry of Land and Resources for the financial support to related research.
Handbook of Stable Isotope Analytical Techniques, Volume 1 P.A. de Groot (Editor) 9 2004 Elsevier B.V. All fights reserved.
CHAPTER 26 Procedures for Sulfur Isotope Abundance Studies Bernhard Mayer1 & H. Roy Krouse2
Department of Geology and Geophysics, University of Calgary, 2500 University Drive NW, Calgary, Alberta, Canada T2N 1N4 2 Department of Physics and Astronomy, University of Calgary, 2500 University Drive NW, Calgary; Alberta, Canada T2N 1N4 e-mail: I [email protected] 1
26.1 Introduction Isotopic studies of the sulfur cycle are analytically more complex than those of many other element cycles. This is because sulfur occurs in nature in a variety of inorganic and organic compounds with valence states ranging f r o m - 2 to +6. The three essential steps in sulfur isotope analysis are (1) the extraction of the sulfur compound(s) of interest, (2) the preparation of a measurement gas, and (3) isotope ratio mass spectrometry. In this chapter, we first discuss available techniques for gas preparation and mass spectrometric measurements of sulfur compounds (section 26.2). Subsequently, we elaborate on the most suitable recipes for extracting sulfur compounds from gaseous, dissolved, and solid samples for isotopic studies of the sulfur cycle in the atmosphere, hydrosphere, lithosphere, pedosphere, and biosphere (section 26.3). Sulfur, element 16, has four stable isotopes 328, 338, 348, and 368. The abundance ratio of the two most abundant stable sulfur isotopes [348/328] Of a sample is measured relative to that of a reference material and expressed as: x 1000 ~348 (in~o) _ { }[348/328]reference_1 [34S/32S]sample
[26.1]
where [348/328] is the ratio of the numbers of 348 to 328 atoms in a sample and reference, respectively. ~338 and 6368 scales are defined similarly in terms of [338/328] and [368/328]. The historical reference for the 6348 (and 6338 &: 6368) scale was CDT troilite (FeS) from the Cation Diablo meteorite. Russian scientists tended to use troilite with a very similar isotope composition from the Sikhote Alin meteorite. Natural reference materials tend to be inhomogeneous a n d / o r not readily available globally in sufficient
539
Procedures for Sulfur Isotope Abundance Studies
quantities. Since the 1960's, variability in the 634S value of CDT was reported to exceed 1%0 (Jensen & Nakai, 1962). However, it was not clear to which extent this was due to extraction techniques versus inhomogeneity. Beaudoin et al. (1994) were able to conclude that variations in ~34S values of CDT due to inhomogeneity were at least 0.4 %0. Consequently, an advisory committee of the International Atomic Energy Agency (IAEA, Vienna, Austria) established a V-CDT scale in 1993, on which an Ag2S reference material IAEA-S1 (formerly called NZ-1) was defined as having a 634S value of-0.30 %0 (Coplen & Krouse, 1998). Unfortunately, 634S measurements on IAEA reference materials, such as IAEA-S2 and IAEA-S3, among different laboratories have displayed variabilities an order of magnitude larger than that obtainable in a given laboratory (often < +0.2 %o). Interestingly, reported 634S values seem to differ dependent upon whether SO2 or SF6 was used as a measurement gas (e.g. Rees, 1978; Taylor et al., in press). This discrepancy is thought to result mainly from cross-mixing effects during SO2 based analyses with dual inlet isotope ratio mass spectrometers (e.g. Meijer et al., 2000) suggesting that the commonly used SO2 based methods for determining 634S values have an inherent error resulting in a scale contraction (Taylor, in press). 634S values for international reference materials have been measured using gravimetric isotopic mixtures by Ding et al. (2001). Sulfur isotope ratios and 634S values for currently available reference and intercomparison materials based on measurements using SF6 and SO2 are listed in Table 26.1. Procedures to correct new and old sulfur isotope compositions based on mass spectrometry with SO2 to the new SF6 based V-CDT scale are currently developed (Taylor et al., in press). For a given mass spectrometer this correction should be achievable by a linear regression similar to that proposed by Taylor et al. (in press)" 6
34
SSF 6 -
1.029 6
348
S O 2 --
0 15
[26.2]
.
Until final recommendations on how to adjust SO2 based measurements to the accurately determined SF6 based V-CDT sulfur isotope scale are available, we suggest that laboratories compare samples on a scale calibrated with IAEA-S1, IAEA-S2, and IAEA-S3 (Table 26.1) and report the obtained values for those reference materials in their publications. Absolute isotopic abundances are much more difficult to determine than
Table 26.1 - 634S values of IAEA reference materials determined using SO2 or SF6 as a measuring gas (data from Coplen et al., 2002; Ding et al., 2001; Taylor, in press; Taylor et al., in press). Note that there is a marked discrepancy between SO2 and SF6 derived 634S values indicating a scale contraction for SO2 derived data possibly due to cross-contamination in the ion source. Reference material
V-CDT IAEA-S-1 IAEA-S-2 IAEA-S-3 IAEA-SO-5 IAE A-SO-6 NBS 127 NBS 123
Compound
FeS (troilite) Ag2S Ag2S Ag2S BaSO4 BaSO4 BaSO4 ZnS
SO2 based
SF6 based
834Sv_CDT
834Sv-CDT
[%0]
[%ol
0.00 -0.30 21.8 -31.95 0.7 -32.54 20.39 17.1
0.00 -0.30 22.66 -32.24 0.49 -34.05 21.08 17.47
540
Chapter 26 - B. Mayer & H.R. Krouse
relative isotopic abundances. Unfortunately, early measurements reported in the literature were not done with samples of known 6-values (e.g. MacNamara & Thode, 1950). Based on measurements of gravimetric mixtures of separated isotopes, Ding et al. (2001) reported recently the following abundances for a material with ~34S = 0 %o relative to V-CDT: 32S: 95.03957(90) %; 33S: 0.74865(12) %; 34S: 4.19719(87) %; 36S: 0.01459(21) %. This is notably different to the early determinations by MacNamara & Thode (1950) reporting 95.02%, 0.75%, 4.21%, and 0.02% for 32S, 33S, 34S and 36S respectively. Sulfur isotope methodology has been applied in basically four ways" 1. Isotope dilution measurements, 2. isotopic labeling or tracer studies, 3. determination of isotopic fractionation by mass independent conversions, and 4. determination of isotopic fractionation by mass dependent conversions. Isotope dilution is analogous to adding a known number of red marbles (spike) to an unknown number of green marbles to determine the latter by counting the green to red number ratio (see also Chapter 37 on IDMS). Very low S contents have been determined in a variety of materials using a 34S enriched spike and thermal ionization mass spectrometry (e.g. Kelley & Fallick, 1990). Isotope dilution has the advantage of correcting for highly variable recoveries of analyte during extraction, if the spike is added and well mixed with the sample before procedures begin.
Historically tracer studies were carried out with radioactive 35S (e.g. Fossing, 1995). Because of its relatively short half life time (87 days), radioactive 35S has been mainly used in short-term laboratory studies and occasionally in field experiments (e.g. Dhamala & Mitchell, 1992). Separated stable isotopes of sulfur would be certainly more suitable for long-term studies in the field, but artificially enriched 33S, 34S, and 36S is very costly. Consequently, the number of labeling studies with tracer compounds enriched or depleted in 34S reported in the literature is orders of magnitude lower compared to those conducted with enriched 13C and 15N. Sometimes, variations in the 6 values of natural S-containing substances can be used in tracer studies (e.g. Mayer & Krouse, 1996). Significant departures from the mass dependent relationship: 636S - 2
x ~34S - 4 x 633S
[26.3]
result usually from mass independent processes such as nuclear reactions. An example is a 33S anomaly in the Allende meteorite (Rees & Thode, 1977). Recently, Farquhar et al. (2001) reported mass independent S isotope fractionation during photolysis of SO2. This concept has been applied to better understand S cycling in the Earth's early history (Farquhar et al., 2000a). Farquhar and co-workers have defined departures from mass dependent fractionation by the following relationships:
Procedures for Sulfur Isotope Abundance Studies
A33S- 1000 9 [(1 + ~ ) 3 3 S / 1 0 0 0 ) - (1+ ~)34S/1000)0-518- 1] A36S- 1000 9 [(1 + 636S/1000)- (1+ 634S/1000) 1.91 - 1]
541 [26.4] [26.5]
Mass dependent conversions in nature and the laboratory invoke a number of questions, which must be answered to properly interpret the isotopic data. Does a system approach isotope exchange equilibrium conditions or do unidirectional processes with associated kinetic isotope effects dominate? If the latter, is the system closed or subjected to additions and losses of reactants, intermediates, and products? If it is a closed unidirectional conversion, have steady-state conditions been realized whereby the step with the lowest rate constant is fully controlling the conversion? In that case the overall isotope fractionation is the kinetic isotope effect of the slowest step. If steady-state is not realized, the isotopic composition of the product does not bear a simple relationship to that of the disappearing reactant because of accumulation and reaction of intermediates. Whereas the above comments apply to any element, their relevance is particularly apparent in sulfur isotope investigations, because S occurs in numerous different organic and inorganic compounds with valence states ranging from-2 to +6. It follows that the protocol for sampling is very important with emphasis upon analyses of as many different S compounds as feasible. During chemical conversions, there should be consecutive sampling of product, reactant, and if possible, intermediates to ascertain how the isotopic composition of these compounds depend upon the extent of conversion. In many systems it is also essential to sample S from different components. For example, a conclusive study of the impact of industrial sulfur emissions on the environment should involve atmospheric constituents, biological materials such as lichens (some of which acquire sulfur mainly from the atmosphere), foliage (which may acquire sulfur from both the atmosphere and the pedosphere), soil horizons, as well as surface water and groundwater. In some cases, the spatial distribution of sulfur isotope ratios in areas as large as states or provinces or as small as a mineral surface may be of interest. Needless to say that successful stable isotope investigations require determining how 634S values relate to physical, chemical, and biological parameters, and to 6 values of other elements. In the following, we discuss strategies for measurements of sulfur isotope abundances and for sampling and extraction of sulfur compounds from gaseous, aqueous, and solid sample materials. The high resolution required spatially, temporally, physically, and chemically to better understand S cycling will be addressed. Some topics in this chapter are presented in detail elsewhere in this book. In those cases, we will discuss the rationale and potential problems with the methodology. Unpublished refinements to techniques developed in the Isotope Science Laboratory (ISL) at the University of Calgary (Alberta, Canada) will also be described. 26.2 Measurement of sulfur isotope abundances Dependent upon the application and the accuracy required, there are a number of ways to measure sulfur isotope abundances (Figure 26.1). The most frequently practiced procedure is analysis using SO2 in an electron impact source isotope ratio mass spectrometer (IRMS). Whereas SO2+ currents at masses 64 and 66 are usually determined, some laboratories have used SO + (masses 48 and 50), which has the advantage
Figure 26.1 - Flow chart summarizing techniques for gas preparation and mass spectrometric analyses of sulfur isotope ratios of S-containing materials.
4~ t~
r
t,~ !
:=
O
Procedures for Sulfur Isotope Abundance Studies
543
of a lower resolution requirement for the mass spectrometer. Other methods include continuous flow isotope ratio mass spectrometry (CF-IRMS) with on-line combustion of samples to SO2, dual inlet isotope ratio mass spectrometry (DI-IRMS) with SF6, thermal ionization mass spectrometry (TIMS) using As2S3, and more recently inductively coupled plasma isotope ratio mass spectrometry (ICP-IRMS) and secondary ion emission mass spectrometry (SIMS). The traditional IRMS features a dual inlet (DI) system, which permits introduction of the sample and a working standard gas alternately into the ionization chamber. Sulfur dioxide has the advantage of rapid preparation by many techniques including oxidation of metal sulfides such as Ag2S and thermal decomposition of BaSO4 alone or in mixtures with other reagents (see Volume IL Part 3, Chapter 8-3 for a detailed description of many preparation systems). However, SO2 has also many disadvantages. It tends to adsorb to surfaces particularly when traces of H20 are present. In fact, this poses one limitation on the minimum of sample size, which can be confidently analyzed. There are also memory effects whereby a sample acquires portions of SO2 from other samples (e.g. Meijer et al., 2000). One potential trouble area in DIIRMS is the common tube between the dual inlet valves and the mass spectrometer source. Gas adsorbed from the reference can subsequently contribute to the sample gas and vice versa. This can be minimized by mild heating of this tube. Another solution is based upon the valve system described by Halas (1979). In most systems, there is a waste line to which sample gas is diverted during analysis of the reference gas and vice versa. The Halas design omits a waste line and features valves alternately blocking the ends of the sample and reference capillaries. Pressure increases in the blocked capillaries and after opening the valve to the ion source there is a transient over-pressure in the common inlet tube. It is believed that this transient effectively displaces the adsorbed gas from the previous measurement, since any memory effect is too low to be detected. In our experience the data obtained with this valve system using SO2 are closer to those obtained using SF6 than indicated by equation [26.2] (see also Halas & Szaran, 2001). The original Halas design used pneumatically pushed teflon sheets against the end of the capillary. In the Isotope Science Laboratory at the University of Calgary, Nupro valves have been modified for this purpose. A further problem with SO2 preparation is the generation of other S-containing gases. There are two effects. There may be interfering ions from the other gases and/ or isotope exchange occurs so that the measured 634S value is not representative for that of the sample. One contaminant is SO3, which is enriched in the heavier isotopes in comparison to SO2 (Eriksen, 1972). Further, SO3 also contributes to SO2 + and SO + currents. Its production can be minimized by preparing SO2 at high temperatures and/or low 02 pressure. The first option was initially practiced in the early 1950's in Harry Thode's laboratory at McMaster University, Hamilton, Canada (Thode et al., 1961), where samples were combusted in an 02 stream at 1350~ The higher temperatures also realized a lower equilibrium constant for the S isotope exchange reaction: K 32S 0 2 + 34S 0 3 34SO 2 4-32803 <=~
[26.6]
544
Chapter 26 - B. Mayer& H.R. Krouse
K has been calculated to be 1.0046 and 1.0035 at temperatures of 1000 and 1200~ respectively (Sakai & Yamamoto, 1966). It is possible to calculate the shift in ~34Svalue for the SO2 as a function of yield and temperature. For example, production of 95% SO2 and 5% SO3 at 1200~ shifts the measured 634S value for the SO2 by -0.175 %o (Rees & Holt, 1991). However, the higher temperatures typically result in faster deterioration of furnace windings and combustion tubes. Therefore, SO2 generation is performed in many laboratories at temperatures <1200~ At lower temperatures, metal sulfate can form particularly if 02 pressures are high. Preparation of SO2 at low 02 pressures has been pursued with N2-O2 mixtures (e.g. Oana & Ishikawa, 1966; Sakai & Yamamoto, 1966) and the use of the solid oxidant Cu20 rather than CuO, V205, etc. (e.g. Robinson & Kusakabe, 1975). In this case, a suitable approach to solve the SO3 problem is to subsequently reduce it to SO2 by passage over Cu at 800~ (e.g. Bailey & Smith, 1972). Li (1996) used a Nd-YAG laser to combust a variety of sulfide and sulfate powders mixed with V205 wrapped in tin foil. Yields exceeded 90 % and the reproducibility of 634S values was usually better (< +0.1%o) than with other combustion techniques. This may reflect the fact that SO3 was undetectable in the prepared gas. However, there were mineral dependant ~34S shifts of typically-2 %o in comparison to data realized with other combustion techniques. For further discussion of solid oxidants for preparing SO2, see Volume II, Part 3, Chapter 8. Another minor complication with oxygen-poor combustion systems is the possibility of production of the analogue OS2 (private communication, E. Roth, 1993). Additional to S isotope partitioning between OS2 and SO2, there is the complication of mass overlaps of $2+ and SO2+. The latter would be absent if SO + currents were measured. Also, Halas (1987) noted that production of HSO2 + ions in the mass spectrometer source causes interferences in the isotopic currents. Addition of H to ion species during ionization of many compounds proceeds readily. Therefore, sources of hydrogen should be minimized during SO2 preparation and subsequent IRMS. Some of the above problems can be addressed by purifying the prepared SO2 in a gas chromatograph as will be discussed under continuous flow techniques. Another complication with SO2 is the isotopic composition of its oxygen. In typical IRMS, 34S1602+ is not resolved from 32S160180+, etc. Therefore, corrections for the oxygen isotope contributions are necessary as described by Coleman for dual inlet IRMS (Part 2, Chapter 44) and Leckrone & Ricci for continuous flow IRMS (Part 2, Chapter 45). Further, the oxygen isotope composition of SO2 produced from all samples and reference materials should be identical. This requires constant combustion conditions, but it has been reported that some sulfide minerals do not combust as quantitatively as others (e.g. Fritz et al., 1974; Robinson & Kusakabe, 1975). Therefore, many laboratories prefer to convert sulfur from all raw materials to Ag2S because of its reliable combustion. However, this is not an option for sulfate if its oxygen isotope ratio is also of interest for the particular study. Thermal decomposition of BaSO4 is another approach for SO2 generation (e.g. Bailey & Smith, 1972; Holt & Engelkemeir, 1970). Pure BaSO4 requires a temperature
Procedures for SulfurIsotopeAbundanceStudies
545
near 1400~ for effective decomposition, which is close to the softening temperature for quartz. Production of SO3 and variable oxygen isotope composition of the reactant BaSO4 are additional problems. Holt & Engelkemeir (1970) addressed the latter problem by measurements with both SO + and SO2+ spectra to correct for different oxygen isotope contributions. Haur et al. (1973) found that the decomposition temperature for BaSO4 could be lowered using the mixture BaSO4:V205:SiO2 in the weight ratio 1:3:2. Yanagisawa & Sakai (1983) refined this technique and proposed that a 1:10:10 mixture realized effective decomposition at 900~ Coleman & Moore (1978) used Cu20 instead of V205 and a reaction temperature of 1120~ Another approach is reaction of BaSO4 with sodium metaphosphate (NaPO3) at temperatures below 1000~ (e.g. Halas & Wolacewicz, 1981): BaSO4 + NaPO3 ~ NaBaPO4 + SO3
[26.7]
With all methods, SO3 is produced and must be reduced to SO2 over hot Cu (e.g. Bailey & Smith, 1972; Halas & Szaran, 1999). A potential benefit of the above reaction mixtures is that variations in the oxygen isotope composition of BaSO4 have a minor influence on the 180/16 0 ratio of the produced SO2 because of the other dominating sources of oxygen. This is particularly true for the method described by Yanagisawa & Sakai (1983) and that of Halas & Szaran (1999; 2001). Because of the necessity of preparing SO2 with constant oxygen isotope composition, laboratories tended to convert all gaseous, aqueous, or solid sulfur compounds either to BaSO4 for thermal decomposition or to Ag2S for combustion. Historically, comparisons of 634S values of SO2 produced from BaSO4 and Ag2S from the same sample revealed significant differences (e.g. Jensen & Nakai, 1962). To a large extent this was probably due to difficulties in properly correcting for different oxygen isotope compositions in SO2 generated by the different methods. Ueda & Krouse (1986) found that reaction with V205:SIO2 mixtures could yield SO2 with the same O isotope composition for sulfides and sulfates. Many SO2 preparation lines are discussed in Volume IL Part 3, Chapter 8. Any of these techniques can realize very good reproducibility in 634S values, particularly if the same S mineral is used. However, in view of the many factors discussed above it is not surprising that the variability for data obtained on the same sample at different laboratories is much larger than the reproducibility in one laboratory. Sulfur hexafluoride (SF6) has also been used with DI-IRMS. It has been prepared using F2 (e.g. Hulston & Thode, 1965b) or bromine trifluoride, BrF3 (e.g. Puchelt et al., 1971; Thode & Rees, 1971). The distinct advantage of pure SF6 is that its hexagonal structure is chemically very stable, similar to inert gases. Further SF5+ constitutes typically over 90 % of all species formed upon ionization whereas SO2§ is about 50 % of the ion species derived from SO2. Since fluorine has only one stable isotope 19E the SF5+ spectrum gives the S isotope abundances without corrections for isobaric interferences as needed for oxygen in SO2 (Part 2, Chapters 44 and 45). The combined
546
Chapter 26 - B. Mayer & H.R. Krouse
advantages of SF6 serve to realize better data from much smaller samples (< 1 mg S) than is possible with SO2 and DI-IRMS (~ 5 mg S). Another advantage of SF6 is the possibility of also directly measuring 633S a n d ~36S values. Measurements based on SF6 have tended to be more consistent than those obtained using SO2 among different laboratories (e.g. Rees, 1978). There are also some disadvantages to SF6. A minor one is the need for a higher resolution (1 in 131) mass spectrometer than is required for SO2 (1 in 65). Fluorinating compounds are chemically very reactive and have caused fatalities. This reactivity makes it difficult to prepare pure SF6 safely. Pure sulfide minerals are the best reactants. Traces of sulfate lead to the formation of sulfur oxyfluorides. Readily produced organic fluoride compounds can cause interferences with the SF5+ spectrum. It was found that some fluoro-polymers used for gaskets and valve seats could generate such contaminants (H. R. Krouse, personal experience, 1958). Not surprisingly, it is very difficult to obtain pure fluorination compounds. One purification method is reaction of F2 to produce minerals such as K3NiF6 followed by its regeneration (Taylor & Beaudoin, 1993). The prepared SF6 can also be purified using gas chromatography (e.g. Thode & Rees, 1979). This procedure is preferred if fluorinating agents such as BrF3 and BrF5 are used. Since isotope fractionation occurs in chromatographic columns, complete recovery of SF6 is required. Throughout the last decade, continuous flow isotope ratio mass spectrometry (CFIRMS) has become an increasingly popular technique for sulfur isotope analyses. Sulfur-containing samples are thermally decomposed in an elemental analyzer (Figure 26.1). Subsequently, the generated SO2 is purified chemically, separated in a GC column, and swept by an inert carrier gas via an open split through the ion source of a mass spectrometer, where the changing ion currents are recorded with multiple collectors (e.g. Giesemann et al., 1994; Pichlmayer & Blochberger, 1988). The major advantage of this technique is that it can realize rapid sulfur isotope analyses (< 15 min per sample) on small samples (< 100 ~g S) of sulfate and sulfide minerals. However, it is not trivial to achieve accuracy and precision comparable to that obtained by DI-IRMS (< + 0.2 %o), since the oxygen isotope composition of the produced SO2 may change depending on the condition of the chemicals in the elemental analyzer (Grassineau et al., 2001). Fry et al. (2002) suggested an SO2-SIO2 buffering system consisting of a tube filled with quartz chips as a measure to minimize the need for oxygen corrections. Nevertheless, frequent analysis of sulfide and sulfate reference materials of known 634S values is essential for calibrating the measurements including an empirical oxygen isotope correction. Direct thermal decomposition of various sulfur-containing raw materials in an elemental analyzer followed by CF-IRMS has also been attempted in recent years (e.g. B6ttcher & Schnetger, Chapter 27, (Fry et al., 2002; Kester et al., 2001)). To facilitate such measurements in our laboratory, we have installed a variable open split with moving capillaries (Con-Flow) to prevent large quantities of CO2 from entering the ion source prior to the arrival of SO2. For organic samples with S contents > 0.1% we are able to obtain reproducible ~34S values (< + 0.5 %o), but normalizing the data to the
Procedures for Sulfur Isotope Abundance Studies
547
international V-CDT scale is somewhat problematic. Utilizing pure sulfide and sulfate minerals with known 634S values as standards for calibrating the sulfur isotope ratios of the unknown raw material with high organic matter content may lead to erroneous results. A possible reason is that water produced during the thermal decomposition of organic raw materials exerts a significant influence on the oxygen isotope ratio of the generated SO2. In our experience, using inorganic S standards for calibrating sulfur isotope ratios obtained via CF-IRMS of organic raw materials yields typically 634S values, which are between 0.5 and 3.0 %0 higher than the true value. Thus, it seems more desirable to use standard materials, which are similar in their biogeochemical composition to the samples analyzed. For many sample types, such standard materials with known 634S values do currently not exist. Therefore, it is recommended that a subset of such S-containing materials is converted to BaSO4 or Ag2S followed by traditional IRMS to determine the extent of potentially differing 634S values using on-line and conventional techniques. For many small samples e.g. from food web studies, direct conversion of bulk material in an elemental analyzer followed by CF-IRMS may represent the only option to obtain sulfur isotope ratios. It is however important to keep in mind that this technique provides only 634S values for total S. Such data are typically meaningful for materials, in which total sulfur is dominated by one particular S fraction (e.g. carbon-bonded S in foliage). However, total sulfur in a sedimentary sample may be comprised of 50 % pyrite with a 634S value of-20 %o and 50 % sulfate with a 634S value of +20 %0. The ability to obtain a 634S value for total sulfur (~ 0 %0) of such a bulk sample via CF-IRMS is clearly of limited use for a meaningful interpretation of this sedimentary system. Hence, further improvement of the CF-IRMS technology will not eliminate the necessity for extracting individual S compounds from bulk samples and converting them to pure inorganic sulfides or sulfates. Compound specific carbon isotope analyses (CSIA) have been successfully applied to a wide range of organic carbon compounds since more than a decade (Hayes et al., 1990). In contrast, compound specific sulfur isotope analyses on organo-sulfur compounds have not been accomplished. From the analytical viewpoint, numerous organic sulfur compounds can be separated by chromatographic techniques (e.g. Witter & Jones, 1999). However with a few exceptions, the S:C ratio is quite small. Consequently with combustion techniques, the CO2 production greatly exceeds that of SO2. Hence it is challenging to generate a sufficient quantity of pure SO2 during chromatographic separation for subsequent isotope ratio mass spectrometry. This in turn can promote generation of interfering species such as COS + in the ion source. In principle it would be possible to remove the SO2 from CO2 by reaction of the former sequentially with aliquots of a metal such as silver. However, it would be very difficult to achieve compound specific resolution with this approach and subsequent preparation of SO2 from the metal sulfide would make the overall procedure very time consuming and expensive. Alternate approaches, where sulfur functional groups rather than compounds are separated on the basis of chemical reactivity, have been more successful (see section 26.3.4). For high spatial resolution of S isotope analyses, laser ablation systems (LA-IRMS) generating SO2 or SF6 have been used successfully (e.g. Crowe et al., 1990; Kelley &
548
Chapter 26- B. Mayer & H.R. Krouse
Fallick, 1990; Rumble et al., 1993; Sharp, 1992; Taylor & Beaudoin, 1993). Such systems allow in situ isotope analyses on sulfur minerals with a resolution of ~ 100 gm with a ~534Sreproducibility often better than _+0.5 %0. They typically require cryogenic purification of the produced S gases and mineral-dependent corrections of the sulfur isotope ratios. Hence, such systems are not intended for high throughput and routine sample analysis. Li (1996) thoroughly studied the use of a Nd-YAG laser for combusting sulfide minerals in an 02 atmosphere. Some of his findings are consistent with those of other investigators as described in Volume II, Part 3, Chapter 8-1.7. Parameters affecting yields and 634S values included mineral properties, 02 pressure in the reaction chamber, energy density of the laser beam, firing sequence, and sampling location. In the 02 pressure range of 100 to 1500 torr, the SO2 yield rate was in the order stibnite > argenite > galena > chalcopyrite. Yields were much lower for pyrite and other sulfides. Low yields can result from the mineral being highly reflective of the laser beam, being transparent to the laser beam, and/or having high thermal conductivity so that heat quickly migrates away from the target area. SO2 yields near the edge of the specimens are higher than in more central regions attesting to more energy loss from the target area by thermal diffusion in the latter case. For the same mineral and 02 pressure, SO2 yields decreased with both laser power density and the total energy received by consecutive firings. It was found that above 500 torr 02 pressure, yields of SO2 and ~34S values remained constant with change of pressure. It was also found that with N2-O2 mixtures, total pressures above 500 torr realized highly reproducible data provided that the 02 pressure exceeded a minimum value (50 to 100 torr dependent on mineral). The data were consistent with the requirement that SO2 was produced by gas phase reactions. Hence, the mineral specimen has to reach the boiling point, which is dependent upon the total pressure of the surrounding gas. As observed by others (e.g. Crowe et al., 1990; Kelley & Fallick, 1990) there were mineral dependent shifts of a few per mil towards lower ~34S values than found by traditional analyses. The shifts decreased with increasing 02 pressure but no consistent correlations were found between SO2 yields and ~34S values among different minerals. Li (1996) designed an intriguing experiment to better understand the ~534Sshifts. A piece of thin glass (0.5 mm thickness) was placed 4 mm above a galena surface in the reaction chamber. Enough solid PbS was deposited on the underneath side of the glass for isotopic analysis after 12 laser firings. Whereas the PbS specimen had a conventionally determined ~34S value near 0 %o, laser produced SO2 and the deposit on the glass had ~34S values o f - 1 and +11%0 respectively. From these results, it was concluded that the shifts resulted from kinetic isotope effects favoring lighter isotopes in the vapor phase (probably sulfide molecules, ions, and $2) and preferential removal of heavier isotopes from the vapor phase during sublimation. The ~534Sshifts associated with laser SO2 production appear to be absent or insignificant when SF6 is formed. Another suitable technique for studying ~34S variations over sub-mm distances is secondary ion mass spectrometry (SIMS). The instrumentation and data reduction are somewhat complex and few facilities have attempted S isotope abundance measurements (e.g. Deloule et al., 1986; Eldridge et al., 1988; Eldridge et al., 1987; McKibben and Eldridge, 1995; Riciputi et al., 1996; Ireland, Chapter 30). Two manufacturers of
Procedures for Sulfur Isotope Abundance Studies
549
suitable instruments are Cameca and Australia National University (SHRIMP, Sensitive High Resolution Ion Microprobe). A high energy ion beam is directed at a mineral surface and secondary ions are emitted from a sputtered pit of typically 20 gm diameter and 5 gm depth. For conducting minerals such as pyrite, the specimen must be Aucoated, polished, and a primary beam of positive ions (e.g. 14.5 keV, 113Cs+) is used to generate negative secondary ions, S- (Riciputi et al., 1996). Au-coating is used to reduce static charge build-up on the mineral surface. Because of their insulating nature, sulfates are analyzed using a negative ion beam (e.g. 14.5 keV, 160-) and positive secondary ions (S§ are generated and measured (Riciputi et al., 1996). Since the secondary ions have a wide range of velocities (both direction and magnitude), a double focussing mass spectrometer must be used and only a portion of the ion current reaches the collector. Since the secondary ion currents are low, an electron multiplier is used in counting mode (usually <106 counts per second for the minor ion beam). Magnetic peak jumping has been used to monitor the ion beams and the 34S ion beam is counted at least 5 times longer than the 32S ion beam to have comparable counting statistics for both beams. A spot is usually sputtered for a few minutes before starting the measurement and the data is collected over the order of 100 counting periods for a total analysis time of about 30 minutes. The main challenge of analyzing S isotopes with SIMS is correcting for the very large shifts in 834S values of up to several hundred per mil incurred during secondary ion generation and transmission in the mass spectrometer optics. The shifts are mineral dependent and the linear/nonlinear behavior over large ranges of 834S values must be established. This requires isotopically homogeneous materials each with different 834S values to be used for calibration. Fortunately, the morphology of a given mineral does not seem to have a significant effect on the instrumental fractionation, e.g. a cubic pyrite could be used to analyze framboidal pyrite (McKibben & Eldridge, 1989). The instrumental isotope fractionation relates to the bond strengths involving sulfur in the minerals and tends to be the lowest for "soft" sulfides such as galena and highest for sulfates (Eldridge et al., 1987). However uncertainties with sphalerite are higher than for other sulfides (up to 4 %0) and do not seem to be related to iron content, any other transition metal impurity, or make of instrument (MacFarlane & Shimizu, 1991; McKibben & Eldridge, 1995). Another source of error is unintentional analyses of mixtures of minerals in the specimen. Although the technique has uncertainties in 834S values of I to 5 %0, natural variations in 834S values of up to 50%0 over sub-mm distances have been identified with SIMS (McKibben & Riciputi, 1998). Hence, this technique has provided greater insight to understanding ore formation, e.g. identification of several successive depositions, which would be difficult to determine with traditional techniques. Additional information on this technique is available inVolume IL Part 3, Chapter 8-1.8.
Thermal ionization mass spectrometry (TIMS) constitutes a technique for performing isotope measurements on samples with S quantities in the microgram range. Paulsen & Kelly (1984) used positive ion thermal ionization mass spectrometry for isotope dilution analyses of trace sulfur in steel and copper based on AsS + ions generated from As2S3, which had been dissolved in a As3+-NH3 solution and loaded on a bed of silica gel on a Re filament (for details see Volume IL Part 3, Chapter 8-5.3). The tech-
550
Chapter 26 - B. Mayer& H.R. Krouse
nique has seldom been used for natural S isotope abundance determination, one example being a study of sub-micrometer aerosol particles over the Pacific Ocean (Calhoun et al., 1991). Wieser (1997) modified the technique of Paulsen & Kelly (1984). A micro reduction apparatus was used with a 5 mL reduction flask. The carrier gas was argon and a septum on the reduction flask was used for injection of 2 mL of Thode reduction mixture (Thode et al., 1961; Volume II, Part 3, Appendix D). Generated H2S was transferred through short teflon tubing into a pipette tip inserted in a 1.5 mL centrifuge tube containing 1 mL of an As3+-NH3 solution (1 ~g As mL-1 solution prepared from dissolving As203 in concentrated NH4OH) kept at 0~ to prevent loss of NH3. Unlike the procedure of Paulsen & Kelly (1984), no As2S3 was precipitated at the completion of the S reduction by dropwise additions of HC1 to the As3+-NH3-S solution. Rather than precipitating and washing As2S3 and dissolving again, the S was left dissolved in As3+-NH3 solution and an aliquot was deposited on the filament hence minimizing S blank and losses of S. Analyses were made on a Finnigan MAT 262 with ion counter. The sample size requirement of 1.5 gg S by Paulsen & Kelly (1984) was reduced to 200 ng. Wachsmann & Heumann (1992) used negative ion TIMS for S isotope abundance measurements. A double filament source was used with 30 gg of Ba deposited on an ionization filament and I to 5 gg S as (NH4)2SO4, H2SO4, or Na2S on the ionization filament followed by a 10 gL silica gel coating. The ionization filament was heated to 1100~ and its radiant heat was also used to indirectly heat the sample filament. Both S- and SO2- species were produced, the former being more abundant (major ion currents ~ 5 x 10-12A). With both positive and negative ion TIMS, the precision is about +2 %0 mainly because of selective evaporation of the lighter isotopic species from the sample filament. The reproducibility can likely be greatly improved by double spiking techniques. For specimens with less than 5 gg S, TIMS seems to be the currently preferable analytical option. However, the gain in sensitivity not only realizes poorer precision but also increased analytical time per sample. The most recent development for measuring sulfur isotope abundance ratios is inductively coupled plasma isotope ratio mass spectrometry (ICP-IRMS). This instrumentation features a source consisting of a plasma torch followed by conical extraction plates (sampling cone and skimmer) and an extraction electrode, which introduces ions to the isotope ratio mass spectrometer. The range of kinetic energies for a given ion species is quite large (~ 20 eV). Earlier instruments used quadrupole or time of flight analyzers. Precise isotope abundance ratio measurements require multiple collector double focussing mass spectrometers (see Chapter 31 by Rehk~imper et al. for details) developed in the 1990's such as the VG Plasma 54 with 9 Faraday collectors plus a secondary electron multiplier (SEM), the Nu Instruments Plasma with 12 Faraday cups and three SEM's, and more recently the Thermo Finnigan Neptune. Quadrupole analyses are also incorporated into the optics of the instruments to achieve additional focussing, e.g. vary the dispersion of ion beams so that consecutive masses fall into fixed collectors. A carrier gas such as argon brings particles or droplets (< 4 gm diameter) to a quartz tube surrounded by a coil of water-cooled copper tubing. The tubing is connected to a radio-frequency generator (~1 kW, 27.12 MHz), which maintains an intense plasma (~ 8000 K). The system is very versatile in that samples can be introduced to the plasma region by vaporization, nebulization, or
Procedures for Sulfur Isotope Abundance Studies
551
laser ablation. However, the ICP-IRMS technique has some inherent problems (e.g. Heumann et al., 1998). Although the bulk of the ions are atomic, there are troublesome amounts of molecular ions generated from the sample a n d / o r water a n d / o r carrier gas. Thus isotopic species of 02 + interfere with S+. This problem can be addressed by trying to minimize 02 + production (Mason et al., 1999b) or going to sufficiently high mass resolution to separate the interfering species. With reference to the latter, a recent brochure on the Thermo Finnigan Element 2 high resolution ICP-MS shows an analysis for Zn in H2SO4, where 64Zn is clearly resolved from 32S32S and 32S1602. The other problem is large mass discrimination, which in theory should only be of concern for absolute abundance measurements. However, there is uncertainty as to how the mass discrimination might vary with signal and concentrations of other elements in the sample. This is attributed to two mechanisms. One is a space charge effect whereby positive ions leaving the skimmer mutually repel each other, deflecting the lighter ions preferentially outwards. The second mechanism is nozzle separation. Ions are generated at a high pressure in the source, enter a region pumped down to ~ 1 Torr between the sampling cone and the skimmer and a region pumped down to ~ 10-4 Torr between the skimmer and the extraction electrode. Again, the lighter ions are pumped away preferentially. The linearity of mass discrimination can be evaluated using isotopic reference materials. Ion currents in the same mass range can also be collected for elements, which do not vary significantly in isotopic composition. However, there is the question as to whether the mass discrimination would be the same for chemically different elements. It seems that the best approach would be double spiking of the sample, which would be easily realized if it were introduced in solution. One attractive feature of ICP-IRMS is the capability of introducing raw solutions to the plasma source with concentrations of sulfur < 1 ppm. The technique is limited to total S since all species are dominantly converted to S+ ions. Analysis time is a few minutes with ~ 1 ~g S consumed. In principle it should be possible to construct a valve system to alternately introduce reference and sample solutions analogous to DIIRMS systems for gases. Currently, a precision of the order of + 1.0 %o is achievable with ICP-IRMS, but in all likelihood significant improvements be possible in the near future. In summar~ there are a variety of techniques available for precise determinations of 34S/32S ratios. Choosing the appropriate technique should be governed by the study objectives. SIMS or laser ablation techniques can achieve 634S measurements on pure mineral specimens with the highest spatial resolution. Isotope analyses on samples with less than 5 gg S are best performed with TIMS and possibly ICP-IRMS. Rapid determination of 634S values for total sulfur on large sample sets can be accomplished by CF-IRMS. The highest precision and accuracy for sulfur isotope measurements is accomplished by dual inlet mass spectrometry using either SO2 or SF6 gas produced via high temperature decomposition of pure sulfate (e.g. BaSO4) or sulfide minerals (e.g. Ag2S). Since few of the above-described techniques can handle raw sample specimens, extracting various forms of sulfur from raw materials and converting them to pure sulfate or sulfide precipitates is a prerequisite and a major challenge in sulfur isotope abundance studies. The remainder of this chapter is devoted to the discussion of appropriate techniques for extraction of various sulfur compounds from
552
Chapter 26 - B. Mayer & H.R. Krouse
raw materials. 26.3 Extraction of sulfur compounds Sulfur occurs in the atmosphere, biosphere, pedosphere, hydrosphere and lithosphere in gaseous, liquid and solid forms. Different sulfur compounds may have radically different 634S values because of diverse sources and distinct formation mechanisms. Hence, it is highly desirable to perform isotope analyses on separate individual sulfur compounds. Some major inorganic and organic S compounds are described below. A more comprehensive summary of the chemistry of sulfur and its compounds is provided by Greenwood & Earnshaw (1997), Mitchell (1996), and Richard (2001) among others.
Inorganic Sulfur Sulfur exists in inorganic molecules and ions with oxidation states ranging from -2 to +6 (Table 26.2). More than one oxidation state may exist in compounds containing two or more sulfur atoms, e.g. thiosulfate ($2032-). For a given molecular or ionic form, there may be several species present (e.g. dissolved species of sulfate, forms of elemental S). This means that attention must be given to parameters such as pH (ion chromatography) and temperature (cryogenic separation). In contrast to 02, $2 exists only in the gas phase at high temperatures. Rhombic sulfur, containing stacked eight membered rings ($8) is the most stable form at room temperature. Upon heating, $6 rings and S2 form. If heated above 120~ and slowly cooled, monoclinic sulfur forms, which contains chains as long as 104 atoms, and also $8 rings, which are stacked differently than in the rhombic form. Sulfur monoxide (SO) is very unstable. SO2 (gas) has a much longer lifetime and oxidation to SO3 is very slow in the absence of a catalyst. Gaseous SO3 condenses to a colorless liquid at 44.5~ and freezes at 16.8~ to three solid forms, $309 rings and two (SO3)x chains. Table 26.2 - Examples of gaseous, aqueous, and solid inorganic sulfur compounds. Oxidation State gaseous
Inorganic S compounds dissolved
solid
+6
CaSO4x2H20 (gypsum) SO3, SF6 H2SO4, HSO4-, SO42CaSO4 (anhydrite) ........................................................................................................................................................................................................................... BaSO4 (bar!te) ..................................... +4 s02, SF4 H2SO3, HSO3-, SO320 $8 rings and other forms of elemental S -1 FeS2 (pyrite) -2 H2S HS-, S2PbS (galena) ..............................................................................................................................
~. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
................................................................................................................................
~. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
........................................................................................................................................................................................................................... Ag2S
Mixed i
$2032- (+6 and-2) S4062- (+6 and -2)
...............................................................................................
553
Procedures for Sulfur Isotope Abundance Studies
Sulfurous acid (H2SO3) formed by the reaction of SO2 and H20 is almost non-existent in solution, the major form being dissolved SO2. Salts of SO3 2- (sulfites) and HSO3- (hydrogen sulfites) are well known. SO3 reacts violently with water to give sulfuric acid (H2SO4). It is unique in that it dissociates to HSO4- as a strong acid but the second dissociation step to sulfate ion (SO42-) is consistent with a weak acid. Thiosulfate ion ($2032-) is essentially a sulfate ion with one oxygen atom replaced by sulfur. Thus the central and replacement sulfurs have oxidation states of +6 and -2 respectively. There are other species with sulfur in different oxidation states such as polythionates, e.g. tetrathionate ($4062-). With selective chemical and bacterial reactions, it is possible to release only sulfur in the -2 oxidation state from these compounds (McCready et al., 1980). The -2 oxidation state exists in metal sulfides (PbS, Ag2S, etc.) and in hydrogen sulfide (H2S). In aqueous solution, H2S behaves as a strong reducing agent.
Organic Sulfur
There are hundreds of different organic S compounds of which a few types (or classes) are summarized in Table 26.3. A given compound may have more than one name. Organic chemical structures can be envisaged as analogues of well-known inorganic S compounds. If one of the hydrogens in H2S is replaced by an R group (CH3, C2H5, etc.), a mercaptan (or thiol) is formed as in methyl mercaptan (H-S-CH3). Phenyl mercaptan, H-S-C6H5 is obtained by replacing a hydrogen by a phenyl group. Although mercaptans are sulfides, the latter term tends to be used where both hydrogens in H2S are replaced by R groups such as an important minor constituent of the atmosphere, dimethyl sulfide (CH3-S-CH3). Note that the two R groups need not be the same (R, R'); an example is methyl ethyl sulfide (CH3-S-C2Hs). There are also polysulfides, one atmospheric constituent being dimethyldisulfide (CH3-S-S-CH3). One can imagine that with the loss of one or more hydrogens, the ends of a R-S-R' chain might be joined in some cases to give a stable cyclic compound. Examples given in Table 26.3 are pentamethylene sulfide and thiopene. Very high molecular weight compounds can contain several cyclic structures. Sulfoxides can be equated to one doublebonded oxygen of SO2 being replaced by two single bonded R groups, e.g. dimethyl sulfoxide: If the similar replacement is carried out with SO3, sulfones are generated e.g. dimethyl sulfone"
0
H
H.~C--S---CH.~ O
II
HsC--S--CH~
II
0
Replacement of one hydrogen in sulfuric acid, H2SO4, by a R group gives a sulfonic acid. Replacement of both hydrogens by R groups yields alkyl sulfates (Table 26.3).
554
Chapter 26 - B. Mayer & H.R. Krouse
Table 26.3 - Examples of organic sulfur compounds. Types of organic S compounds
Generalized formula
Examples gaseous
solid
Mercaptans (or thiol)
R- S - H
CH3SH (methylmercaptan)
Cysteine
Organic sulfides
R- S - R
CH3SCH3 (DMS: dimethylsulfide) CH3SC2H5 (methyl ethyl sulfide)
Methionine
CH3SSCH3 (DMDS: dimethyldisulfide)
Cystine
R - S - R'
Organic disulfides
Cyclic sulfides
R- S - S - R
H2 C H2C "~' "~CH2
I
Thiacyclohexan Pentamethylene sulfide
I
H2C.~,S j-CH2 Alkyl sulfates
R~O
O
\//
S
R' o ' % Sulfonic acids
R - SO3-
Sulfoxides
O
II
R~S~R' Sulfones
O
II
R ~S---R'
II
O Thiopenes
CH~CH
II
II
HC,,~sfCH Polycyclic sulfides
~
Benzothiophene
S=C=O
Carbonyl sulfide
S=C=S
Carbon disulfide
Sulfonates Sulfate esters
R-
OSO3-
Heparin
Procedures for Sulfur Isotope Abundance Studies
555
Replacement of one O atom in CO2 by an S atom gives carbonyl sulfide (COS), whereas replacement of both oxygens gives carbon disulfide (CS2). Comparison to RS-S-R' shows that the term disulfide is used for two quite different S functional groups. Amino acids contain the
NI-I~ ~ - CI~ C- O O H
(A) grouping. Sulfur-containing amino
I
H acids are quite abundant in biological materials, e.g. L-cysteine, HS-CH2-A L-methionine, CH3-S-CH2-CH2-A L-cystine, A-CH2-S-S-CH2-A Note that cystine is the disulfide product of joining two cysteine molecules. For many sulfur isotope studies it is desirable to extract selected sulfur compounds or functional groups for isotope ratio determinations. The challenges common to all sulfur extraction schemes are: 9 extraction of a given compound with minimal contamination from other components, 9 avoiding isotopic fractionation during extraction, 9 avoiding interconversions of S compounds during extraction, 9 generating a sufficient amount of preferably Ag2S or BaSO4 for subsequent isotope ratio mass spectrometry. The following is a summary of techniques, which have proven useful in determining isotope ratios of sulfur compounds from gaseous, liquid, and solid samples. The objective is not necessarily to recommend any one of these techniques as superior to the others, but to aid in the selection of procedures that are best suited to the specific needs of any potential research project.
26.3.1 Sulfur in the atmosphere Gaseous (SO2, H2S, (CH3)2S (=DMS), (CH3)2S2 (=DMDS), COS, CS2 among others), dissolved (SO42-, SO32-, etc.) and solid (particulates, So, etc.) forms of S-containing compounds can be found in the atmosphere. Particulates are very complex. They may occur as inorganic inhomogeneous dust particles, minerals, viscous liquid droplets, and biological debris (e.g. seeds, pollen, insects). Some minerals may be soluble in water (e.g. (NH4)2SO4) or organic solvents (e.g. elemental sulfur). Further, SO2, H2SO4, and other compounds may be adsorbed to particulate surfaces. This section summarizes analytical techniques for isotope analyses on major atmospheric S compounds. A simplified summary chart for this section is shown in Figure 26.2.
26.3.1.1 Dissolved sulfate in atmospheric deposition Historically, dissolved sulfate in bulk deposition has been the most commonly studied atmospheric sulfur compound (e.g. Herut et al., 1995; Mizutani & Rafter, 1969; Nakai & Jensen, 1967; Nriagu et al., 1987). Because of the generally low sulfate concentrations in atmospheric deposition, it is typically necessary to collect large quanti-
556
Chapter 26 - B. Mayer & H.R. Krouse
Figure 26.2- Flow chart summarizing analytical procedures for chemical pretreatment of some atmospheric S compounds and their conversion to BaSO4, Ag2S, or As2S3 for subsequent sulfur isotope analysis (dw - distilled water).
ties (1 to 20 L) of rain or snow for isotopic analysis, particularly if DI-IRMS is performed. Special metal-free high-volume precipitation collectors with effective sampling areas of more than I m2 are often designed for this purpose (e.g. Tanaka et al., 1994; Wadleigh et al., 1994:). It is also important to exclude dry deposition by using an appropriate cover during non-precipitation periods. After membrane filtration (0.45 ~m) of the sample, dissolved sulfate is usually converted to BaSO4 by addition of BaC12 solution. Since BaSO4 has a solubility of 2.3 mg L-1 under standard state conditions, pre-concentration of sulfate is typically necessary to achieve maximum conversion to BaSO4. Ion exchange techniques have been successfully employed for this purpose (see section 26.3.2.1). Alternately, it is possible to evaporate part of the water sample to increase the sulfate concentration. If oxygen isotope analyses on the sulfate are desired it should be noted that there is a risk of oxygen isotope exchange during high temperature treatment particularly of acid water samples (Chiba & Sakai, 1985). Atmospheric sulfate is derived from both anthropogenic and natural sources. The latter include seasalt sulfate (SSS), which is mechanically produced at the ocean surface, and which has a 634S value similar to that of sea water sulfate (+21%o). The sul-
Procedures for Sulfur Isotope Abundance Studies
557
fur isotopic composition of non-sea-salt SO42- (NSS) can be determined using conservative chemical seasalt tracers (e.g. Wadleigh et al., 1994). This technique relies on the precise determination of the sulfur isotope ratio of sulfate in bulk precipitation and the concentration of SO42- and a conservative tracer such as CI-, Na +, or Mg2+ (Keene et al., 1986), which is presumed to be entirely ocean-derived. Using CI-, non sea-salt derived (NSS) or excess sulfate is defined as: [SO42-]NSS = [SO42-]total- 0.14[C1-]total
[26.8]
where 0.14 is the weight ratio of SO42-/C1- in ocean water (Drever, 1982). The isotope balance of sulfate in atmospheric deposition is as follows: [SO42-]total 9 ~34Stota 1 = [SO42-]SSS 9 i~34Sss S + [SO42-]NSS 9 ~34SNs S
[26.9]
or ~34SNss = (1/fNSS) " [~34Stotal - (fsss " ~34Ssss)]
[26.10]
where fsss and fNSS identify the sea-salt and non sea-salt derived sulfate fractions, respectively. Interestingly, ~)34SNSS values derived using these principles fall in a comparatively narrow range between-1 and +6 %0 independent of study site (Cortecci & Longinelli, 1970; Herut et al., 1995; Mizutani & Rafter, 1969; Moerth & Torssander, 1995; Wadleigh et al., 1994; Wakshal & Nielsen, 1982).
26.3.1.2 Atmospheric sulfur gases NSS is usually derived from a combination of anthropogenic and biogenic sources. The former include SO2 derived from fossil fuel combustion, which is subsequently oxidized to sulfate via heterogeneous or homogeneous pathways (Tanaka et al., 1994). Natural sources comprise biogenic dimethylsulfide (DMS), which is primarily emitted to the atmosphere from the ocean surface (Andreae & Raemdonck, 1983), and other biogenic S gases such as COS, CS2, and H2S. Determination of the isotopic composition of these biogenic atmospheric S gases has been challenging because of their generally low concentrations and their short residence times in the atmosphere. The ability to collect sufficient volumes of sample for isotopic analysis is often the limiting factor for isotope ratio determinations of gaseous atmospheric S compounds. Historically, several methods have been used to extract atmospheric sulfurous gases including exposure of chemically treated surfaces, high volume (HiVol) passage of air through chemically treated filters, bubbling of air through reactant solutions, and gas chromatographic separation of gases, which have been cryogenically trapped and adsorbed to molecular sieves. For traditional DI-IRMS, the major challenge is to accumulate at least 1 mg of S. This situation has recently improved because of advances in CF-IRMS, TIMS, and ICP-IRMS reducing sample size requirements.
558
Chapter 26 - B. Mayer & H.R. Krouse
The sulfurous gas, which has been used most frequently for S isotope analysis is SO2. It has seldom been used directly for mass spectrometric measurements because of its variable oxygen isotope composition. Rather, it is oxidized to sulfate, precipitated as BaSO4 or Ag2S, which is subsequently decomposed to SO2 under controlled conditions (see section 26.2) for mass spectrometric analysis. Research conducted in our laboratory in the 70's and 80's found that PbO2 cylinders exposed for one month produced sufficient sample for traditional DI-IRMS (Krouse, 1980; Legge et al., 1988). These cylinders consisted of PbO2 paste on paper, which was either wrapped around the sides of a vertical can, or placed on the inside wall of a vertical can with the ends removed. This approach is not recommended from the environmental health viewpoint. Since this method gives the 634S values of an "integrated" sample collected over one month, desirable information such as dependence on wind direction, and operational schedule of an industrial emitter were unattainable. SO2 has been successfully collected by drawing air through filters treated with KOH mounted in high volume air samplers. Cellulose filters are impregnated with 20% K O H - 10% triethanolamine (TEA) solution (Huygen, 1963). Cellulose filters are often preferred because deterioration by KOH is less severe than for glass fiber filters. Filter paper S blanks can be reduced by pre-washing with distilled water and by drying the impregnated filters for 3 hours at 100~ washing out the KOH, and re-drying the filter paper before they are again impregnated with KOH-TEA (Forrest & Newman, 1973). Subsequently, the filter paper is dried again for 5 minutes. Filters treated with other chemicals such as K2CO3 have also been used (e.g. Holt et al., 1972; Saltzman et al., 1983). The coating with KOH or K2CO3 should be preferably conducted in a SO2 free glove box to avoid contamination with laboratory air, and filters should be stored in closed containers or plastic bags prior to use. After collection, the filters are removed from the HiVol sampler and immediately returned to the glove box or clean room for extraction with distilled water, preferably under ultrasonic agitation. Addition of 1% H 2 0 2 is recommended to oxidize sulfur species in the filtrate to SO42-. Some researchers have boiled the extraction water after addition of bromine (strong oxidant) to convert all S compounds to SO42- (Holt et al., 1972). Subsequently, the solution is evaporated to increase the sulfate concentration, and sulfate is precipitated as BaSO4 by adding BaC12 solution. The produced BaSO4 can be used directly for conversion to SO2 and subsequent isotope mass spectrometry, but some researchers have preferred to convert BaSO4 to Ag2S prior to IRMS (Forrest & Newman, 1973; Van Everdingen et al., 1982). Determination of an analytical blank on not exposed filters is essential for correction purposes. Forrest & Newman (1973) report that the use of two KOH-TEA filters yields satisfactory measurements of SO2 concentrations and isotope ratios at normal air humidity (> 30%). Most commercial HiVol samplers use 20 x 25 cm filters, which are supported mechanically by resting on a down-stream stainless steel mesh. Air flow rates of 2 m3 min-1 are obtainable without undue sacrifice of collection efficiency of SO2 (Holt, 1975). At these rates, sufficient sample for DI-IRMS (> 1 mg S) can be obtained on the time scale of hours. This makes sampling of specific wind directions or industrial emission events feasible.
Procedures for Sulfur Isotope Abundance Studies
559
Dequasie & Grey (1970) were able to obtain milligram quantities of atmospheric sulfur dioxide for DI-IRMS by adsorption of SO2 on a molecular sieve. Since removal of water vapor and separating CO2 and SO2 were not straightforward and quite laborintensive, this technique has never received widespread use. Atmospheric SO3 is rapidly converted to sulfate and hence concentrations are typically too low for isotopic analyses. However, its concentration in flue gas can be quite high. Forest & Newman (1973) sampled flue gas by passage through a quartz wool filter and condensing of SO3 in a coil at a temperature just above the dew point of water. SO2 was not condensed and subsequently oxidized to sulfate in a solution of alkali salts and hydrogen peroxide. If isotopic equilibrium between SO2 and SO3 has been achieved (equation [26.6]), the g34S values of the two gases provide information about the temperature conditions during combustion in oil or coal fired power plants. Where H2S emission from springs or wetlands are significant, sampling of H2S prior to its oxidation in the atmosphere might be feasible. Van Everdingen et al. (1982) used sheets of filter paper soaked in saturated cadmium acetate solution to collect airborne H2S near sulfurous springs. The same authors also used silver acetate treated H2S test paper for the same purpose, reporting 634S values as low a s - 3 0 %o. Other researchers inferred indirectly from low ~34S values in atmospheric SO2 or rainwater sulfate that biogenic S gases released from soils, marshes, and wetlands must have contributed to atmospheric S (Grey & Jensen, 1972; Hitchcock & Black, 1984; Nriagu et al., 1987). Dimethylsulfide (DMS) is a major source of atmospheric S. It is mainly produced via assimilatory sulfate reduction in the oceans and can be transported into the atmosphere through the sea-air interface. DMS can be oxidized to either sulfate or methane sulfonate (MSA). Gold surfaces have been used for selective adsorption of nmol quantities of DMS (e.g. Ammons, 1980). Calhoun (1990) up-scaled such a gold collection system to obtain micromole amounts of DMS for isotopic analysis using TIMS. The collection tube consisted of an ultrapure quartz tube filled with 200 meters of loosely packed gold wire. This setup had a collection capacity of approximately 80 nmoles of DMS. The collected DMS was subsequently converted to As2S3 via thermal desorption of the DMS in an H2 atmosphere to generate H2S, which was trapped as As2S3 in an As3+/NH3 solution. A minimum of 25 collection and purge cycles was used to obtain micromole quantities of DMS for isotopic analysis. Using this system, Calhoun (1990) was able to collect a single seawater-derived DMS sample (10 ~g S) from the remote southeastern Pacific Ocean. TIMS analysis revealed a ~34S value of +17 %0, which is consistent with the sulfur isotope ratios of submicrometer NSS aerosols and MSA reported below. Hence it was concluded that DMS can be the source of NSS in the remote South Pacific (Calhoun, 1990). This is evidence that NSS from fossil fuel combustion is often isotopically distinct from NSS formed by oxidation of biogenic marine S sources such as DMS. This makes sulfur isotope ratio measurements on atmospheric S compounds a useful tool for differentiating between anthropogenic and biogenic sources (e.g. Nriagu et al., 1991).
560
Chapter 26 - B. Mayer & H.R. Krouse
26.3.1.3 High volume atmospheric sampling A typical high volume atmospheric sampler is a motor driven impeller, which draws air through a system of filters at a rate of 1.41 to 1.83 m3 min-1. Filter sheets (8" x 10") are mounted with regular gaskets (rubber or similar material) leaving an effective filtering area of 7" x 9". Motors are usually 115 V AC with 24 V DC versions for mounting on aircrafts. The flow can be maintained reasonably constant as material collects on the filters by using a sensor and a feedback circuit to control the motor's speed. Sometimes the volume flow is kept constant by using manometer sensors. Alternately, mass flow can also be regulated using a combination of velocity and temperature sensors. If the filter paper becomes excessively loaded with particulate matter, the motor may overheat. This is more problematic if the pore size of the paper is too small. For most commercial units, if the flow is < 20 ft3 min-1, there is insufficient air cooling of the motor. Brush life is another limitation, being typically ~ 500 hours for motors run at 115 V. It can be extended to ~ 1500 hours if the motor is run at 90 V. Brushless motors are available from some suppliers. One problem in sampling particulates is that large amounts of material may be collected, which are a nuisance since they load the filter papers but are not the object of the study. Examples are wind blown dust from coal, ore piles or road construction. To some extent, the peaked roof design of high volume sampling chambers reduces the intake of large particles. A better approach is the use of a cyclone pre-separator, which collects particles typically > 5.5 gm. Cascade impactors (discussed below) fail with larger particles because they may bounce or roll through. A cyclone pre-separator has an inlet vent with a vane to align it with the wind direction. Air enters the cyclone body tangentially and sets up a vortex flow pattern. Larger particles migrate to the walls or to the top of the chamber. The Andersen cascade impactor has a different design. There are geometrical patterns of holes in circular metal plates. These differ for adjacent plates but are the same for every other plate.
26.3.1.4 Aerosol sulfate and particulate sulfur High volume samplers have been used to collect both particles and aerosol sulfate. Forrest & Newman (1973) tested the suitability of glass fiber filters and recommended Whatman 81 filters because of their high collection efficiency and little tendency to convert SO2. The glass fiber filter with the S-containing particulates can be subject to Thode reduction (Thode et al., 1961) to convert all S to H2S (see Volume II, Part 3, Chapter 8-3.4), which is trapped as CdS and subsequently converted to Ag2S (Forrest & Newman, 1973; Van Everdingen et al., 1982). Other researchers have collected aerosol and particulate S by passing several thousand m3 air through Whatman 41 cellulose filters (McArdle & Liss, 1995; Nriagu et al., 1991). Sulfate and methanosulfate can be extracted from the filter with deionized water in an ultrasonic bath. The sample should be irradiated with ultraviolet light. Subsequently, dissolved sulfate is converted to BaSO4 by adding BaC12 solution. Using this procedure, McArdle & Liss (1995) were able to generate 1 mg S for DI-IRMS. Other researcher have preferred teflon filters for the collection of aerosol and particulate S (Hitchcock & Black, 1984;
Procedures for Sulfur Isotope Abundance Studies
561
Quinn & Bates, 1989). Calhoun (1990) collected particle phase non sea-salt sulfate (NSS) and methanesulfonate (MSA) from marine air using a 1.0 ~m pore size teflon filter. Aerosol sulfate collected on these teflon filters was converted to As2S3 using the method of Paulsen & Kelly (1984), which is described in Volume II, Part 3, Chapter 8-5, but MSA did not convert to As2S3 using this technique (Calhoun et al., 1991). Subsequently, 1.5 ~g S was loaded as As2S3 on a filament and sulfur isotope ratios were determined by TIMS. This technique was chosen because of its low sample volume requirements and the lack of memory effects and isobaric interferences, yet achieving reasonable precision of ~ 2 %o for ~34S measurements including sample collection, handling, and conversion to As2S3 (Calhoun, 1990). To determine the sulfur isotope composition of both aerosol sulfate and methanosulfate (MSA), a subsection of the teflon filter was wetted with I mL spectrophotometric grade methanol in 10 mL distilled water and centrifuged for 30 minutes. The extract was transferred into a Carius tube (Paulsen & Kelly, 1984) and frozen together with 10 mL of 16 M HNO3 and 4 mL of 11 M HC1. The tube was flame-sealed and heated to 240~ for 16 hours to convert MSA to sulfate. Total sulfate (MSA-derived and aerosol sulfate) was subsequently converted to As2S3 and sulfur isotope ratios were determined by TIMS as described above. The isotopic composition of MSA was determined by mass and isotope balances, and was found to be similar to that of NSS (~34S = +15.6 + 3.1%o) from submicrometer aerosol particles in the remote southeastern Pacific Ocean (Calhoun et al., 1991). Aerodynamic sizing of atmospheric particulates can be carried out with a five or six stage cascade impactor fitted to a high volume sampler. One design by Sierra Instruments Inc. consist of stacked plates, which alternately have 9 or 10 parallel slots. Consequently, slots on a given plate occur midway between those of the plates above and below. As the air flow bends going through successive plates, particulates of decreasing size are deposited on slotted glass-fiber filter paper (dimensions 5.625" x 5.375"). The width of the slots decreases with successive plates (stages) and the finest particulates, which pass through the cascade, are trapped on a 8" x 10" "back-up" glass-fiber filter paper mounted above the fan of the high volume sampler. Aerodynamic sizing not only depends on the size of the particle, but also on its shape and density. On a given plate, the physical size of a particle has an inverse dependence on the density. The particle size cut-offs increase with flow; they are chosen on the basis of human inhalation. Those above 7 gm are deposited in the throat. Smaller particles are deposited in the bronchia and the smallest reach the lungs. Sizing of atmospheric particulates for obtaining 634S values seems to have been first reported for a study near two sour gas processing operations (Krouse, 1991). Larger particles had higher 634S values near +20 %o consistent with industrial emissions in the study area. The smallest fraction had ~34S values closer to 0 %0 implying a different source. The collection of size-segregated aerosols for S isotope analyses has also been reported in some recent studies (e.g. Patris et al., 2000b; Turekian et al., 2001). Multiple-stage high volume samplers with glass fiber substrates and backup filters can separate size fractions from more than 5 ~m to less than 0.2 ~m geometric
562
Chapter 26 - B. Mayer & H.R. Krouse
mean radius. With flow rates of approximately I m3 min-1, sufficient S can be sampled for each size fraction in less than 48 hours provided that CF-IRMS is used for the sulfur isotope measurements (Turekian et al., 2001). Using these techniques, it was found that submicron radius NSS over the North Atlantic Ocean is mainly derived from fossil fuel combustion (Patris et al., 2000b). In the above studies, the particulates were not distinguished on the basis of solubility. In a study on Bermuda, Norman & Krouse (1992) separated bulk particulate matter on the basis of solubility in water. They found insoluble particulates to have a v e r a g e ~)34S values near +3 %o, which interestingly is the range reported by Turekian et al. (2001) for the smaller size fraction.
26.3.1.5 Combined sampling of 802 and aerosol~particles High volume samplers can be fitted with several filter papers allowing for the simultaneous collection of SO2 and aerosol/particle S for subsequent isotope analysis. Forrest & Newman (1973) employed a glass fiber prefilter to collect aerosol and particle S followed by two KOH-TEA coated cellulose filters for the collection of SO2. They reported a reproducibility of 634S measurements of _+ 0.2 %o. Newman et al. (1975) used such a setup to sample the plume of an oil-fired power plant with a single engine aircraft. Quinn & Bates (1989) used tandem filters, which separated particle phase non sea-salt sulfate (NSS) and methanesulfonate (MSA) collected using a 1.0 gm pore size teflon filter, from gas phase SO2 collected on K2CO3 coated filters as described above. The reproducibility of sample collection, handling, and conversion to As2S3 for TIMS analysis was reported as better than +_2 %0 (Calhoun, 1990). 26.3.2 Sulfur in the hydrosphere In the hydrosphere, sulfur occurs predominantly in dissolved form either as SO42-, HS- or $2-, with dissolved organic sulfur (DOS) typically being of minor importance. Additionally, gaseous S compounds such as H2S may be present in water samples obtained from reducing environments. This section describes methods to quantitatively recover dissolved and gaseous S compounds from water samples for subsequent isotope ratio mass spectrometry (see Volume IL Part 3, Chapter 15 for additional details). A simplified summary chart for this section is shown in Figure 26.3. The selection of the appropriate procedures to collect water samples for sulfur isotope analyses is critically dependent on knowledge about the existing sulfur species in the sample and their concentrations. It is strongly recommended to fix dissolved or gaseous reduced inorganic sulfur species in the field, since these compounds may otherwise rapidly outgas or oxidize to sulfate. Dissolved sulfate in oxidized water samples is typically fairly stable. Hence preservation of such water samples may not be necessary if they are further processed immediately upon return to the laboratory. If storage of sulfate-containing water samples is necessary, the addition of a bacertiocide (e.g. HgC12) and storage in a refrigerator at 4~ in darkness is recommended. 26.3.2.1 Dissolved sulfate Under oxidizing conditions, SO42- is usually the dominant S species in water samples and reduced inorganic S compounds are typically not present. It is recommended
Procedures for Sulfur Isotope Abundance Studies
563
Figure 26.3 - Flow chart summarizing analytical procedures for chemical pretreatment of various S compounds from water samples and their conversion to BaSO4 or Ag2S for subsequent sulfur isotope analysis.
to filter such water samples with 0.45 ~m membranes either in the field or immediately after return to the laboratory. Subsequently, the samples should be acidified to a pH value between 3 and 4 with dilute sulfate-free HC1 to convert dissolved carbonate species into CO2 and hence avoid co-precipitation of BaCO3. Acidification to pH values of less than 2 should be avoided since HSO4-will be the dominant S species and precipitation of BaSO4 will be slow. Note that excess chloride may pose problems if ion exchange techniques are subsequently used. If sulfate concentrations are above 20 mg L-l, BaSO4 can be precipitated directly. After acidification of the water sample, 0.25 M BaC12 solution (typically 10 mL) is added in excess and a white precipitate of BaSO4 will form and settle overnight. Some researchers prefer to heat the water sample prior to adding the BaC12 solution (e.g. Carmody et al., 1998) since this facilitates the outgassing of dissolved CO2 and enhances the rapid formation of a coarse BaSO4 precipitate. However, if oxygen isotope analyses on the sulfate are to be performed it is important not to heat samples with low pH values (< 2) excessively, since oxygen isotope exchange between water and sulfate is promoted under high temperature low pH conditions (Chiba & Sakai, 1985). A 1 L water sample with a sulfate concentration above 20 mg L-1 will yield
564
Chapter 26 - B. Mayer & H.R. Krouse
more than 40 mg BaSO4, which is sufficient for both continuous flow and dual inlet isotope ratio mass spectrometry. Samples with very low concentrations of sulfate (e.g. atmospheric precipitation samples) must be pre-concentrated prior to precipitation of BaSO4 because of its solubility of 2.3 mg L-1. This can be achieved by reducing the volume of the water sample via slow evaporation in a beaker on a hot plate or in a rotary evaporator. Depending on the chemistry of the sample, non sulfate-containing precipitates may form towards the end of the heating process and these must be removed prior to the addition of BaC12 to precipitate BaSO4. Patris et al. (2000a) successfully evaporated up to 2.5 L meltwater from Antarctic firn cores under partial vacuum in a rotary evaporator to a final volume of only 2 mL. Rather than precipitating BaSO4, these authors introduced the remaining sulfate-containing liquid into ChromosorbWM-filled tin cups, where sulfate precipitated upon further heating. Subsequent CF-IRMS analysis yielded precise sulfur isotope ratios for micromolar levels of sulfate from polar ice samples. Another technique of collecting and concentrating sulfate from dilute solutions is the use of anion exchange resins (e.g. Mizutani & Rafter, 1969). Suitable commercially available products include Dowex TM Amberlite IRA-400 or BioRad TM AG 1-X8 and AG 2-X8, which are all in chloride form. Large quantities of water can be passed through such resins in flow-through mode either via gravimetric dripping or by using a peristaltic pump. Since divalent ions have a higher retention affinity than CI-, sulfate will be quantitatively retained at low flow rates (less than 10 mL min-1 recommended). In selecting the size of the ion exchange column it is important to consider the maximum exchange capacity (typically 1.2 meq per mL resin) and the fact that other negatively charged ions such as bicarbonate, nitrate, and some DOC will also be retained hence occupying adsorption sites. High concentrations of chloride may also interfere with the retention process (Carmody et al., 1998) and thus excessive use of HgC12 or HC1 for preserving or acidifying the sample to remove bicarbonate should be avoided. Ion exchange resin columns of any desired size can be made by filling the resin in glass tubes with glass wool on both ends. Preconditioning includes eluting of traces of sulfate with 3M KC1 solution and subsequent rinsing with deionized water to remove excess chloride (Carmody et al., 1998). A convenient alternative are BioRad TM PolyPrep columns pre-filled with 2 mL of anion exchange resin, which need in our experience no pre-conditioning. After passing a sufficiently large sample volume to generate a few mg of BaSO4 through the anion exchange resin, the moist resin can be stored in a refrigerator until further processing. To remove the sulfate quantitatively, a small volume (e.g. 15 mL per 2 mL of resin) of concentrated chloride solution (e.g. 3 M KC1 or HC1) is passed through the resin and the sulfate-containing eluant is collected in a beaker. Subsequently, BaSO4 is precipitated as described above. BaSO4 produced by any of the above techniques is left to settle overnight, filtered off with a pre-weighed 0.45 gm membrane, and subsequently washed thoroughly with deionized water to remove C1- from the precipitate. Thereafter, the precipitate is either air or oven dried. Re-weighing of the BaSO4 containing filter paper allows for an approximate determination of the weight of the precipitate and hence yields an
Procedures for Sulfur Isotope Abundance Studies
565
estimated sulfate concentration for the sample. If organic contaminants appear to be present e.g. in DOC-rich solutions, baking of the BaSO4 precipitate at 800~ in a muffle furnace is recommended in case that only sulfur isotope ratio are to be determined. However, this procedure is not suitable if oxygen isotope measurements are planned, because of the risk of oxygen isotope exchange at high temperatures. Direct precipitation of BaSO4 as well as pre-concentration of sulfate e.g. via ion exchange techniques and subsequent isotope ratio mass spectrometry can yield accurate ~34S values for dissolved sulfate with an uncertainty of < +0.2 %o. Incomplete recovery of sulfate from ion exchange resins tends to yield slightly too low 634S values (e.g. Carmody et al., 1998).
26.3.2.2 Dissolved organic sulfur Some solutions (e.g. forest floor seepage water) contain significant amounts of dissolved organic carbon (DOC). It is a common procedure to oxidize the organic carbon either with H202 or with a saturated bromine solution to avoid contamination of BaSO4 precipitates with organics (e.g. Carmody et al., 1998). However, it must be noted that some amino acids of the DOC fraction may be S-containing. This carbonbonded organic sulfur is oxidized to sulfate and will contribute to the sample. One of the co-authors has attempted to determine the isotope composition of dissolved organic sulfur via mass and isotope balances (Mayer, 1993). A DOC-containing water sample was split in half, and dissolved sulfate was precipitated as BaSO4 from one aliquot by addition of BaC12 solution. The second aliquot of the water sample was subject to H202 oxidation under UV light to oxidize DOS to SO42-. Subsequently, total sulfate was precipitated as BaSO4 as described above. Sulfate concentrations in the second aliquot increased only by less than 2% and 634S values of the sulfate from both samples were found to be identical within the uncertainty of the method, hence preventing a conclusive determination of the sulfur isotope composition of the dissolved organic sulfur (DOS). Alternate techniques comprise isolation of a sufficient quantity of DOC via resin techniques (e.g. Fluka XAD 8) and subsequent isotope analysis of total S of the dried DOC sample e.g. by the Eschka method (see section 26.3.3.1). Using this technique, fulvic and humic acids in the Gorleben aquifer (Germany) were found to have 634S values varying between-3 and +17 %0 (Wang et al., 1998).
26.3.2.3 Dissolved sulfide One of the challenges of obtaining reduced inorganic S species for isotope analyses is to prevent their oxidation during sampling. Hence, exposure to atmospheric oxygen must be avoided wherever and whenever possible. Therefore, we do not recommend filtering of sulfide-containing water samples prior to further processing. Generally, there are two different procedures for collecting reduced inorganic S species for isotope analysis" (1) direct precipitation or (2) degassing of H2S from a water sample with a N2 stream. Direct precipitation is particularly suitable for water samples with > 1 mg L-1 dissolved sulfide. The technique involves adjustment of the pH value of the water sample to either strongly acidic conditions under which H2S is thermodynamically stable
566
Chapter 26 - B. Mayer& H.R. Krouse
or strongly basic conditions, under which $2- is prevalent (see Figure 26.3). Sodium or potassium hydroxide have been used to create alkaline solutions, to which either cadmium or zinc acetate is added to precipitate CdS (Van Everdingen et al., 1982) or ZnS (Rye et al., 1981), respectively. Zinc solutions are environmentally more desirable, whereas Cd solutions tend to react faster and more quantitatively (e.g. McKibben & Eldridge, 1989). Direct precipitation of Ag2S by addition of a AgNO3 solution is typically unsatisfactory because of co-precipitation of AgC1. Acidification of the water sample with cadmium acetate and trapping of the evolving H2S as CdS has also been successfully employed (e.g. Fouillac et al., 1990). We recommend this technique for rapid sampling of dissolved sulfide in the field using sampling bottles pre-filled with circa 100 mL of cadmium acetate solution. After pumping the water sample for a few moments and discarding the initial sample, the end of the pump hose should be inserted into the cadmium acetate solution at the bottom of the sampling bottle. This ensures that the water sample is not exposed to atmospheric oxygen prior to mixing with the cadmium acetate solution, in which dissolved sulfides will rapidly precipitate as bright orange-yellow CdS. Upon return to the laboratory, the CdS is filtered off with a 0.45 ~m membrane filter. In case of significant contamination of the CdS with other solids (e.g. silt) it may be necessary to purify the sulfide precipitate by acidification with 6M HC1 and collection of the evolved H2S in a Cd acetate trap (see section 26.3.3). Wet CdS is subsequently converted to Ag2S by titration with 0.1 M AgNO3 solution containing a 3% NH4OH solution to complex C1- and avoid undesirable AgC1 precipitation. The Ag2S precipitate is subsequently filtered, thoroughly washed with deionized water, and air-dried prior to mass spectrometric measurements. As a note of caution we emphasize that Cd acetate is extremely poisonous and advise that utmost caution is necessary during transport and handling of Cd acetate-containing sampling containers. An alternate technique for obtaining dissolved sulfide species for isotope analyses particularly from large water samples (> 20 L) with low sulfide concentrations (< 1 mg L-l) was suggested by Moncaster & Bottrell (1991). The technique involves the transfer of the water sample into a large carboy and subsequent acidification with concentrated HC1 (1 mL L-I) to drive the dissolved sulfide species equilibrium towards H2S. Care must be taken to minimize exposure of the water sample to atmospheric 02 during the filling procedure. The evolving H2S is subsequently purged with a N2 stream (0.1 - 0.3 L min-1) and trapped as Ag2S in a NH4 + containing silver nitrate solution (see Volume II, Part 3, Figure 14.5). To maximize recovery of H2S, the degassing procedure should be performed for several hours. The Ag2S precipitate is subsequently filtered on a pre-weighed 0.45 gm membrane. Re-weighing the filter paper with the dried precipitate enables gravimetric determination of dissolved sulfide contents. Note that Ag2S collected by N2 stripping will be isotopically inhomogeneous with early-formed Ag2S being enriched in 32S relative to the later-formed Ag2S (Carmody et al., 1998) since there is small isotope fractionation between H2S(gas) and H2S(aq) favoring 32S in the former (e.g. Carmody et al., 1998; Szaran, 1996). To avoid erroneous results it is therefore desirable to approach complete recovery and to remove and homogenize the entire Ag2S from the filter paper prior to isotope analysis. Carmody
Procedures for Sulfur Isotope Abundance Studies
567
et al. (1998) report that dissolved sulfide collected by N2 stripping yields typically 0.2 to 0.5 %0 lower ~)34Svalues than sulfide obtained by direct precipitation presumably due to incomplete recovery.
26.3.2.4 Simultaneously occurring reduced and oxidized sulfur compounds Many water samples contain both reduced and oxidized forms of sulfur. The challenge is to quantitatively sample the sulfide fraction before processing the sulfate. Since dissolved sulfide may have a 634S value more than 30 %0 lower than that of the associated sulfate in the same water sample, oxidation of even small portions of dissolved sulfide may bias the isotopic composition of dissolved sulfate. The choice of sampling procedures is dependent on both the concentration of dissolved sulfide and the sulfide to sulfate ratio in the sample. For samples with high sulfide concentrations (> 1 mg L-l), the most effective technique to trap dissolved sulfides is using sampling bottles pre-filled with Cd acetate solution in the field, in which sulfide will precipitate as CdS prior to exposure to atmospheric 02 (see section 26.3.2.3). However in samples containing both dissolved sulfide and sulfate, the contribution of co-precipitated pale yellow CdSO4 may be substantial and failure to remove it gives erroneous gravimetric and isotope composition determinations. After recovering the precipitate on a membrane filter, it is therefore recommended to subject the precipitate to reduction with 6 M HC1 in a distillation apparatus as described in section 26.3.3.2 in order to recover S from the sulfide fraction only. To the remaining sulfate-containing water sample, BaC12 solution is added in the laboratory to produce BaSO4, which is subsequently recovered by filtration and subject to isotope ratio mass spectrometry. There are a variety of procedures to deal with samples with low sulfide concentrations and they are somewhat dependent on the sulfide to sulfate ratio in the sample. One option is to conduct the above described N2 stripping technique in the field, but this requires the hauling of heavy equipment (e.g. N2 tank) and several hours of time per sample. By using flow rates between 1.1 and 2.1 L N2 min-1, Carmody et al. (1998) achieved > 80 % sulfide recovery and 634Ssulfide values only 0.25 %o lower than expected for a groundwater sample from the Floridian aquifer. However, if sulfide recovery is less than 50 %, the obtained ~34Ssulfide value may be more than 1%0 lower than expected (Carmody et al., 1998). In some studies only the isotopic composition of dissolved sulfate is of interest. It has been suggested that under such circumstances rapid precipitation of BaSO4 may yield satisfactory results despite the presence of trace amounts of sulfide. At high sulfate concentrations (> 20 mg L-l), BaC12 solution can be added and the rapidly forming BaSO4 precipitate can be filtered off after 5 minutes. Carmody et al. (1998) showed that this technique provides reliable 634S value for sulfate, if the sulfate to sulfide ratio exceeds 40. For sulfate to sulfide ratios below 40, driving off H2S with the N2 stripping technique without capturing the H2S is an alternate option for ensuring accurate isotope analyses on the dissolved sulfate. Carmody et al. (1998) recommend N2 flow
568
Chapter 26 - B. Mayer & H.R. Krouse
rates of 8 L min-1, covering of the sampling container to avoid contact with atmospheric 02, and continuation of the procedure until sulfide concentrations in the water sample decrease below 0.01 mg L-1. In most of their experiments this was achieved in less than one hour and little sulfate was produced via sulfide oxidation during the outgassing process (SO4 2- concentrations < 0.5 mg Lq). However, since sulfate produced via sulfide oxidation has a sulfur isotopic composition similar to that of the sulfide, it may nevertheless modify the isotopic composition of sulfate in the water samples with sulfate to sulfide ratios below 10.
26.3.3 Sulfur in the lithosphere There are numerous forms of sulfur in minerals, rocks, and sediments. They include sulfates (barite, anhydrite, gypsum, etc.), elemental sulfur, mono-sulfides (sphalerite, pyrrhotite, galena etc.), di-sulfides (pyrite etc.), as well as numerous different forms of organic S. Total sulfur refers to the sum of all these individual S compounds in a sample. Total S contents of lithospheric materials may vary from more than 30 % in some sulfide ores to less than 0.005 % in some plutonic rocks such as granites. In this section we summarize techniques suitable for extracting sulfur from magmatic, metamorphic, and sedimentary rocks. Many of the described techniques are also suitable for extracting S from peat deposits or coal (see also section 26.3.4 and 26.3.5). Pretreatment procedures for lithogenic material depend on sample type and the scope of the study. In some cases it is possible to handpick individual sulfur minerals such as pyrite. More typically, rock samples are broken and finely ground in a mill to mesh size 4:0 prior to analysis. In cases where loss of reduced inorganic S compounds such as H2S is anticipated (e.g. fresh lake sediment), sample pretreatment in a glove box under N2 or Ar atmosphere may be necessary (Lasorsa & Casas, 1996).
26.3.3.1 Total sulfur Many studies reported in the literature have determined ~)34Svalues of total sulfur (e.g. Nriagu & Coker, 1983) using a variety of different techniques (see also Volume IL Part 3, Chapters 8 & 13-2). Chemical oxidants may be employed either as high temperature fusion mixtures or as wet acid treatments for extracting total sulfur from geological samples. In both cases all S compounds in the sample are converted to SO4 2- for subsequent precipitation of BaSO4. The amount of raw material required to generate enough BaSO4 for subsequent mass spectrometry is dependent on the S content of the sample and may vary from less than 100 mg to as much as 50 gram. Eschka mixture is suitable for oxidizing S compounds in coal and coke to sulfate (e.g. Smith & Batts, 1974), but has also been used for total S analysis in whole rocks and sediments (e.g. Fry, 1986; Nriagu & Soon, 1985). Eschka mixture consisting of MgO and anhydrous Na2CO3 in a weight ratio of 2:1 is mixed with dry and ground sample material in a crucible and heated to 800 + 25~ After slow cooling, the generated sulfate is rinsed with hot distilled water into a beaker (ASTM, 1993), where it can be precipitated as BaSO4 by adding BaC12 solution (see Volume II, Part 3, Chapter 132.8). Since the amount of Eschka mixture used exceeds the weight of the sample by a
Procedures for SulfurIsotopeAbundanceStudies
569
factor of four, it is important to use an Eschka mixture with the lowest possible S blank. Although commercially available Eschka mixtures (e.g. from MERCK) have S blanks of less than 5 gg g-l, blank corrections may be necessary in determining the true ~34Stotal value of S-poor materials such as granites. Wet chemical techniques for oxidizing lithogenic S compounds to sulfate have the advantage that the S blanks of the utilized chemicals are often negligible, and that there is typically no limit in sample size. This is particularly convenient if mg quantities of BaSO4 must be generated for DI-IRMS from S-poor materials. Digestion of raw materials with a HNO3-Br2 mixture has proven satisfactory (Krouse & Tabatabai, 1986; Zhabina & Volkov, 1978). This technique involves soaking a sample overnight in open beakers in a fumehood with conc. HNO3 and liquid Br2 and subsequent heating to dryness (see Volume II, Part 3, Chapter 13-2.8). The generated sulfate is dissolved in dilute HC1 and after removing undissolved matter by filtration the sulfate-containing solution is transferred into a beaker, where BaSO4 is precipitated by addition of BaC12 solution. Many laboratories prefer Ag2S for conversion to SO2 or SF6 for subsequent isotope ratio mass spectrometry. An alternative to oxidizing all S compounds to sulfate and generating BaSO4 is their reduction to H2S and subsequent formation of Ag2S. This is most commonly achieved by the Kiba technique, which was initially developed for the reduction of sulfate (Kiba et al., 1955). The Kiba solution, which consists of waterfree phosphoric acid mixed with tin(II)-chloride dihydrate (Kiba et al., 1955), reacts in a closed and de-oxygenated reaction flask with geological samples between 120 and 280~ to evolve H2S, which is swept with a N2 carrier gas through a washing solution into a chemical trap to form ZnS or CdS (see Volume II, Part 3, Chapter 8-3.6). These compounds can be subsequently converted to Ag2S by titration with AgNO3 solution. Soon after its invention, it was however found that the Kiba technique also converts organic S compounds (Ohashi, 1955) and sulfide minerals such as pyrite (Kiba et al., 1957) to H2S. Hence, the Kiba technique is often used for total S analysis, although it has been reported that some common sulfide minerals (e.g. arsenopyrite) and elemental S are only partially recovered (Sasaki et al., 1979). If Ag2S is preferred as the final reaction product, it is also possible to convert all S compounds in a geological sample to sulfate, followed by the conversion of the latter to H2S by reduction with Sn(II) strong phosphoric acid (Sasaki et al., 1979) or various hydriodic acid mixtures as described by Johnson & Nishita (1952) and Thode et al. (1961), among others. The produced H2S is trapped as CdS or ZnS in acetate solutions, and these compounds can be subsequently converted to Ag2S by titration with AgNO3 solution. Recently, generation of SO2 from geological materials in an elemental analyzer followed by CF-IRMS has also been successfully attempted (B6ttcher & Schnetger, Chapter 27). This technique is only suitable for samples with sufficient Stotal contents (e.g. > 100 ppm), since commercial autosamplers accommodate only samples of less than 100 mg. It has also been reported that matrix effects may result in erratic results partic-
570
Chapter 26 - B. Mayer & H.R. Krouse
ular with samples containing pyroxene, amphibole or biotite (Studley et al., 2002). Further limitations of this technique have been discussed in section 26.2. If a lithospheric sample is dominated (> 98 %) by one particular S compound, the use of any of the above-described extraction techniques may be justified to asses the 634S value of the predominant form of sulfur. This is because minor S components would presumably not change the result by more than 1%o even if they are isotopically distinct. However, most geological materials contain a mixture of several S compounds with often quite different 634S values. In this case, extraction of individual S compounds is strongly recommended, since total S analysis would provide an often meaningless average 634S value dependent on the quantitative proportions of the respective S compounds and their sulfur isotope ratios.
26.3.3.2 Inorganic sulfur compounds In some cases, it is possible to hand-pick pure sulfur minerals such as pyrite or barite or to use gravimetric techniques (e.g. heavy liquids) for their separation. After isolating these minerals, they may be used for direct conversion to SO2 or SF6 for subsequent sulfur isotope measurement (e.g. Ueda & Krouse, 1986) or for laser ablation (Volume II, Part 3, Chapter 8-1.7). More commonly however, sulfur compounds are finely interspersed in the rock matrix and have microscopic or sub-microscopic size. In these cases, wet chemical extraction procedures are suitable for recovering individual S compounds from geological materials. For this purpose, it has become common practice to describe lithospheric S compounds in broad groupings based on their similarity in chemical properties. It is important to note that these groupings are operationally defined and include terms such as acid volatile sulfur (AVS), chromium reducible sulfur (CRS), and hydriodic acid reducible sulfur (HI-red S), among others. Ideally) methods extracting discrete S compounds should do so without affecting other S forms present, although this is not always the case.
Acid volatile sulfur (AVS) Since sulfide is liable to oxidation once exposed to air, careful sample handling in the field is essential to prevent loss of S compounds. Where possible handling of material under a N2 atmosphere is recommended, and samples should be preferably stored at 4~ or frozen and analyzed within 2 weeks of collection (Lasorsa & Casas, 1996). An alternate solution is the addition of a 20% zinc acetate (ZnAc) solution (w/v) to the sample thereby fixing dissolved sulfide and some acid volatile S by forming more stable ZnS (Duan et al., 1997b). Upon return to the laboratory and after thawing the sample, acid volatile sulfur (AVS) is liberated as H2S by treatment with 6M HC1 (e.g. Tuttle et al., 1986) in a distillation apparatus similar to that shown in Volume II, Part 3, Figure 13-1.10. It is widely assumed that AVS comprises predominantly monosulfide minerals such as sphalerite (ZnS), galena (PbS), pyrrhotite (magnetic Fe0.9S), mackinawite (Fel.IS), greigite (magnetic F3S4), and amorphous monosulfides of other elements. Up to 100 gram of sample can be placed in a 200 mL reaction flask. The flask is attached to the distillation apparatus and the entire system is de-oxygenated for at least 15 minutes by a stream of N2. Thereafter, 40 to 80 mL of 6 M HC1 are introduced with a syringe via the rubber septum and the reaction is allowed to pro-
Procedures for Sulfur Isotope Abundance Studies
571
ceed at room temperature. The released H2S is swept by the carrier gas through a washing solution (e.g. pyrogallol) into a trap with Zn or Cd acetate solution, where ZnS or CdS precipitates. After 25 minutes at room temperature, the reaction vessel may be heated to 70~ for circa 15 minutes since minerals such as crystalline pyrrhotite and greigite will not completely react with cold HC1 (e.g. Tuttle et al., 1986). Rinsing the residual sample with hot distilled water yields the operationally defined fraction of acid soluble sulfate (see below). Unfortunately, the selectivity and recovery of this extraction technique is somewhat dependent upon the nature of the sample and maturity of the respective reduced S compounds. One potential problem with AVS extractions may arise if acid soluble Fe(III) is present, which can oxidize generated H2S to elemental S within the reaction flask. Since this oxidation will likely proceed with negligible sulfur isotope selectivity, the 634S values determined for the recovered Ag2S may still be representative for acid volatile S, but gravimetric determinations will underestimate the acid volatile S fraction. A potential solution to this problem is the addition of tin(II) chloride to the sample prior to the reaction. Tin(II) rapidly reduces Fe(III) to Fe(II) and thus prevents it from reacting with H2S (Pruden & Bloomfield, 1968) usually without mobilizing significant amounts of pyrite during treatment with cold HC1 (Chanton & Martens, 1985). However, hot HC1 + SnC12 digestion has been shown to liberate some pyrite S (Cornwell & Morse, 1987; Fossing & Jorgensen, 1989; Rice et al., 1993).
Chromium reducible sulfur (CRS)" Chromium reduction converts sulfur from monosulfide minerals, elemental S, and di-sulfides such as pyrite (cubic FeS2) and marcasite (ortho-rhombic FeS2) to H2S (e.g. Fossing & Jorgensen, 1989; Zhabina & Volkov, 1978) and is hence widely accepted as a means to determine total reduced inorganic sulfur. Chromium reducible sulfur can be extracted from powdered rock samples (< 100 mesh) using the same distillation apparatus as described above and shown in Volume II, Part 3, Figure 13-1.8. After the sample is placed into the reaction apparatus together with 10 mL ethanol, a continuous stream of N2 is established and 40 - 60 mL of 1M CrC12 solution (obtained by reduction of CrC13) and 20 mL 6M HC1 are introduced with a syringe via a rubber septum (Canfield et al., 1986). The sample and the solution are boiled for 1 to 2 hours and the released H2S is trapped as ZnS or CdS and can be subsequently converted to Ag2S for isotope analysis as described above. Canfield et al. (1986) reported recoveries between 92 and 97 % for elemental S, monosulfides, and disulfides using this technique, while sulfate minerals and organic S compounds were shown to be unreactive (exception cystine with 2 % recovery). The method is specific for disulfides (mainly pyrite) if monosulfides and elemental S are removed prior to chromium reduction (see section 26.3.3.4). However, Newton et al. (1995) documented that 634S values are typically shifted by +0.55 %o as a result of incomplete conversion of pyrite (usually 95-96 % recovery) using this technique. Some researchers have attempted to separately extract less mature pyrite (synthetic or recently formed) with cold CrC12 solution (1 hour) followed by extraction of mature pyrite with hot CrC12 solution (1 hour) as described above (Duan et al., 1997b; Fossing & Jorgensen, 1989). Others have reacted rock samples with lithium aluminum hydride (LiA1H4, short LAH) to convert pyrite S into H2S (see Volume IL Part 3, Chapter 8-1.5), which was subsequently trapped as CdS or Ag2S (e.g. Smith et al., 1964; Westgate & Anderson, 1982).
572
Chapter 26 - B. Mayer & H.R. Krouse
We do not recommend this technique since the reagent can be very explosive and yields for LAH reduction of pure pyrite are often low (e.g. Tuttle et al., 1986). Also, extracting pyrite (and sulfate) with 2M nitric acid + bromine (Riley et al., 1990) is not recommended since this treatment is known to also mobilize some organic sulfur (Bottrell et al., 1994; Smith & Batts, 1974).
Elemental S: Elemental (or native) sulfur in rocks and sediments was rarely extracted as a separate fraction prior to the 90's (e.g. Zhabina & Volkov, 1978) but has recently gained increased attention since the significance of bacterial disproportionation of elemental S for the sedimentary sulfur cycle has been detected (e.g. Canfield & Thamdrup, 1994). Measurement of isotope ratios of elemental S usually involves extraction with organic solvents such as benzene, methylene chloride, CC14, dichloromethane (CH2C12), or acetone (Duan et al., 1997b; Hall et al., 1988; Smith & Batts, 1974; Wieder et al., 1985; Zaback & Pratt, 1992). Some researchers have also used carbon disulphide (CS2), but traces of CS2 may decompose upon heating (Fossing & Jorgensen, 1989) and hence we do not recommend this solvent for extraction. Elemental S can be extracted by adding 70 to 150 mL of the organic solvent (e.g. acetone or dichloromethane) to the sediments and shaking for 16 hours (Duan et al., 1997b; Wieder et al., 1985) or by ultrasonic agitation of the sample in a beaker (Hall et al., 1988). Zahbina & Volkov (1978) report that acetone extraction also recovers part of the bituminous organic sulfur from the samples. The elemental S containing solution is filtered through 0.45 ~m PTFE membranes. Subsequently, the solvent with the elemental S is either transferred to the above described distillation apparatus, and reduced to H2S via chromium reduction (Canfield et al., 1986), or evaporated to dryness followed by oxidation to sulfate and precipitation of BaSO4 (e.g. Hall et al., 1988; Smith & Batts, 1974). An alternate technique described by Zaback & Pratt (1992) features extracting approximately 50 gram of rock powder in a Soxhlet apparatus for 48 hours using methylene chloride. Granular copper was added to the collection flask to remove elemental S, which was solubilized during the extraction. The resulting copper sulfide was converted to SO2 for subsequent dual inlet isotope ratio mass spectrometry. Sulfate: Sulfate occurs in rocks as minerals such as barite, anhydrite, gypsum, or less commonly as hydroxy-sulfate minerals (alunite, jarosite, basaluminite etc.). Sulfate is also found in pores in water-soluble form or is occasionally adsorbed on iron and aluminum oxides and hydroxides, particularly under acidic conditions. Carbonate rocks contain sulfate as trace constituents (Staudt & Schoonen, 1994) in what is often referred to as structurally-substituted sulfate. There are numerous techniques described in the literature for removing sulfate from geological samples. Extraction in an open beaker under ultrasonic agitation or in an overhead shaker for several hours at 10 rpm at a solution to sample ratio of 10:1 by weight has been found satisfactory for most procedures described in this section. Extraction with distilled water yields the fraction of water-soluble sulfate (e.g. Hall et al., 1988), which comprises pore water sulfate and some mineral sulfate (e.g. gypsum). Sulfate minerals such as barite, anhydrite, or gypsum can be leached from geological samples by reacting with a 5% Na2CO3 solution (Breit et al., 1985). It is noteworthy that the high pH value of the lat-
Procedures for Sulfur Isotope Abundance Studies
573
ter reaction solution may promote conversion of labile organic sulfur to sulfate, but the content of labile organic sulfur in mature geological samples is typically negligible. Reacting a geological raw sample with 6 M HC1 and subsequent rinsing with hot distilled water yields the fraction of acid soluble sulfate. In all cases, dissolved sulfate can be converted to BaSO4 by addition of BaC12 solution. To yield pure BaSO4 it may be necessary to increase the pH of the solution e.g. with ammonia or NaOH and remove Fe(OH)3 by filtration prior to the addition of BaC12 (e.g. Hall et al., 1988). An alternate technique to extract sulfate-S from geological samples is its reduction to H2S. Sulfate-S can be converted to H2S by the Thode reduction mixture consisting of HIH3PO2-HC1 in a ratio of 4:2:7 (Thode et al., 1961), or by the Johnson-Nishita reduction mixture, which comprises HI-HCOOH-H3PO2 in a ratio 4:2:1 (Johnson & Nishita, 1952). The Kiba reduction technique (Kiba et al., 1955) is not recommended for this purpose since it is not specific for sulfates (Sasaki et al., 1979). The selective analysis of structurally substituted sulfate in carbonate specimens requires samples, which are free of other sulfate or sulfide minerals. Pre-treatment with a 5.25 % sodium hypochlorite (NaOC1) solution has been suggested as suitable for removal of organic matter, soluble sulfates, and metastable sulfide minerals associated with the sample (Burdett et al., 1989). Subsequently, the powdered carbonate sample can be digested with 6 M HC1 in a N2 stream as described above, removing sulfide S as H2S and liberating the structurally bound sulfate from the calcite lattice as acid soluble sulfate (e.g. Burdett et al., 1989; Kampschulte et al., 2001; Ohkouchi et al., 1999). Following filtration of the insoluble residue, Hurtgen et al. (2002) raised the pH to values between 3 and 5 by adding NaOH. Subsequently, they added 10 to 15 mL of saturated bromine water to facilitate the precipitation of iron oxyhydroxides. After their removal by filtration, BaSO4 was precipitated by addition of BaC12 solution. The in vacuo Kiba technique constitutes an alternate approach for analyzing trace sulfates in carbonate rocks. This method uses Kiba solution (Kiba et al., 1955, see also section 26.3.3.1) with one tenth (1/10) of the original Sn2+ concentration and extractions are carried out under vacuum (Ueda & Sakai, 1983). Whereas sulfide-S is converted to H2S, sulfate is only reduced to SO2, which is cryogenically trapped and subsequently used for isotope analysis. One problem with this technique is that a small fraction of the sulfate (< 5 %) does convert to H2S accompanied by a slight enrichment of 32S (Krouse & Ueda, 1987). Bottomley et al. (1992) found that this technique is very effective in releasing sulfur from barite as SO2, but is not quantitative in reducing pyrite to H2S.
26.3.3.3 Organic sulfur compounds
In many geological materials, the amount of organic S is minute compared to that of inorganic S. To our best knowledge, there are no reliable techniques, which can extract individual organic sulfur compounds from fresh geological samples for isotope ratio measurements (see also sections 26.2, 26.3.4 and 26.3.5). Hence, the isotopic composition of organic sulfur in geological samples is typically determined by removal of all inorganic S constituents followed by conversion of total S in the sample residue to BaSO4 or Ag2S by any of the previously described techniques (e.g. Eschka method, digestion with HNO3-Br2, etc.). Subsequent isotope ratio mass spectrometry
574
Chapter 26 - B. Mayer& H.R. Krouse
yields a ~34S value which is thought to be representative for organic S. Another option is to obtain the kerogen fraction of sedimentary rocks by treatment with HC1 and HF under N2 atmosphere (e.g. Hitchon, 1974; Zaback & Pratt, 1992) and subject it to a total S isotope ratio determination (section 26.3.3.1).
26.3.3.4 Extraction schemes for isolating individual sulfur compounds In most geological samples, several inorganic and organic sulfur compounds do
occur simultaneously in variable quantities. There are two fundamentally different approaches to extract different S compounds from a powdered rock sample: 1. to extract the S compounds of interest (e.g. acid volatile S, acid soluble sulfate, chromium-reducible S, remainder = organic S) sequentially from one sample, 2. to extract individual S compounds (e.g. total S, acid volatile S, acid soluble sulfate, chromium reducible S) from fresh aliquots of the same sample. The latter approach has often been used for content analyses with some sulfur compounds being calculated by difference (e.g. organic S - total S - 21 inorganic S compounds). However, obtaining sulfur isotope ratios for the calculated S compounds requires isotope and mass balances, which can result in large uncertainties particularly for S compounds representing a small part of the total sample. Hence, sequential extraction schemes are commonly preferred for obtaining ~)34S values for different S fractions of geological samples. It has become common practice to combine a number of the previously described extraction procedures to a sequential extraction scheme (e.g. Bates et al., 1993; Hall et al., 1988; Rice et al., 1993; Zhabina & Volkov, 1978), which recovers the individual S fractions either as Ag2S or BaSO4 to allow for both gravimetric content determination and isotope analyses. Selecting the appropriate extraction scheme for a given sample depends on the relative abundance of organic and inorganic sulfur compounds and on the lability of these compounds. Most rock samples contain sulfur predominantly in form of sulfate, monosulfide and disulfide minerals, and as organic S. For such samples, an extraction scheme was suggested by Tuttle et al. (1986), which is in our view suitable for most magmatic, metamorphic, and mature sedimentary rocks (Figure 26.4). To avoid or minimize the oxidation of reduced inorganic sulfur compounds, it is advisable to extract these compounds at the beginning of a sequential extraction procedure. Monosulfide minerals are initially recovered from the powdered sample by releasing acid volatile sulfur (AVS) as H2S using hot 6 M HC1 in a distillation apparatus described above. Hot HC1 should be used to facilitate complete recovery of minerals such as crystalline pyrrhotite and greigite. Additions of stannous chloride is only recommended if the pyrite fraction is well-crystallized and coarse grained (Chanton & Martens, 1985; Rice et al., 1993), since other pyrite forms may be partially liberated by the vigorous hot HC1 +SnC12 treatment (Cornwell & Morse, 1987; Fossing & Jorgensen, 1989). The evolving H2S is carried with a N2 stream into a cadmium acetate trap, where S precipitates as CdS, which is further converted to Ag2S. Subsequently, the residual rock sample is removed from the reaction vessel and is thoroughly rinsed with hot distilled water to remove acid soluble sulfate, which is precipitated as BaSO4 for subsequent sulfur isotope analysis; some researchers prefer the reduction of sulfate to H2S and subsequent conversion to Ag2S (Zhabina & Volkov, 1978) for sulfur
Procedures for Sulfur Isotope Abundance Studies
575
Figure 26.4 - Flow chart summarizing analytical procedures for extracting monosulfide S (acid volatile S), acid-soluble sulfate, elemental S + disulfide S (chromium-reducible S), and organic S from magmatic, metamorphic and sedimentary rocks for subsequent isotope analysis (dw = distilled water).
isotope ratio determinations. The remaining sample is returned to the distillation apparatus and undergoes CrC12 reduction yielding H2S, which is converted to Ag2S. This precipitate is representative of the elemental S and di-sulfide (pyrite and marcasite) fractions of the sample, since monosulfide minerals have been previously removed. In case that all sulfates in the sample were acid soluble, the above-described procedures removed all inorganic S compounds from the sample, and only organic S compounds remain in the sample residue. The sample residue is thoroughly washed with distilled water and subsequently subjected to a total S extraction (e.g. Eschka method, HNO3-Br2 digestion, etc.), which yields either BaSO4 or Ag2S (see section 26.3.3.1), which is typically representative for organic sulfur. Tuttle et al. (1986) tested the above described extraction scheme using oil shales and associated rocks from the Eocene Green River Formation (Colorado, Utah, Wyoming) providing evidence for quantitative recovery of the individual S fractions by 57Fe M6ssbauer spectroscopy. The above-described extraction scheme is not suitable for samples containing acidinsoluble sulfates such as barite. Since acid-insoluble sulfates are non-reactive during HC1 and chromium reduction, they will be part of the residual sample together with organic sulfur. From such residues, Smith & Batts (1974) extracted organic S by adding 40 % NaOH solution followed by repeated additions of 100 % H202. After no more
576
Chapter 26 - B. Mayer & H.R. Krouse
residue dissolves upon boiling, the solution is filtered and sulfur compounds in the filtrate are oxidized to sulfate with aqua regina and precipitated as BaSO4 (Smith & Batts, 1974). The predominantly inorganic residue, which contains the acid-insoluble sulfate minerals such as barite and possibly some recalcitrant organic S, subsequently undergoes a total S extraction (section 26.3.3.1). Whereas Smith & Batts (1974) used the Eschka method, Hall et al. (1988) preferred the Kiba reduction technique because of its excellent yield for barites. Modern sediments usually contain significant amounts of organic S and elemental S and the reduced inorganic S fractions may be considerably more labile than those of ancient rocks. Hence, procedures to sequentially extract S compounds from such samples must be adjusted accordingly. A potential approach is depicted in Figure 26.5. The fresh sediment sample should be immediately treated with a 20% zinc acetate (ZnAc) solution (w/v) to fix dissolved sulfide and some acid volatile S by forming more stable ZnS (Duan et al., 1997b). Upon return to the laboratory and after thawing and potentially freeze drying the sample, elemental S is extracted with acetone or another solvent (e.g. Rice et al., 1993). Duan et al. (1997b) provided evidence that initial extraction of elemental S with dichloromethane does not affect the AVS pool, if AVS was pre-fixed using ZnAc. Subsequently, acid volatile sulfur (AVS) is liberated from the remaining sediment sample as H2S by treatment with cold 6M HC1 (e.g. Rice et al., 1993). This comparatively mild AVS treatment is usually successful in collecting the majority of the S from monosulfides without liberating S from disulfide minerals such as pyrite. The use of stannous chloride in conjunction with AVS extraction from modern sediments is strongly discouraged, since it would almost certainly liberate S from the disulfide (pyrite) fraction. The sample residue is thoroughly washed and filtered with hot distilled water. Acid soluble sulfate is precipitated as BaSO4 from the filtration solution by adding BaC12. The solid sample residue may contain some elemental S, which was generated during AVS extraction via reaction of H2S in the presence of Fe(III). Therefore, it is recommended to conduct an additional solvent extraction (acetone, dichloromethane, or others) on the solid sample residue after AVS liberation. The elemental S containing solvent is evaporated and the Ag2S generated via chromium reduction is representative for part of the acid volatile S fraction (Rice et al., 1993). Subsequently; disulfide sulfur is recovered from the sample residue by chromium reduction (Canfield et al., 1986). After thoroughly washing the sample residue with distilled water, a total S extraction (e.g. Eschka method, HNO3-Br2 digestion, etc.) is performed, which yields either BaSO4 or Ag2S representative for the organic S fraction of the sample, since barite is typically non-existent in modern sediments.
26.3.4 Sulfur in fossil fuels
Sulfur in petroleum and coal specimens occurs as complex assemblies of organic and inorganic compounds. In some cases, organic coatings on inorganic compounds may interfere with their extractions. A potential health hazard is gaseous or dissolved H2S in some samples.
Procedures for Sulfur Isotope Abundance Studies
577
Figure 26.5 - Flow chart summarizing analytical procedures for extracting elemental S, monosulfide S (acid volatile S), acid-soluble sulfate, disulfide S (chromium-reducible S), and organic S from modern sediments for subsequent isotope analysis (dw = distilled water).
578
Chapter 26 - B. Mayer & H.R. Krouse
Historically, three techniques have been used to extract total sulfur from liquid and solid fossil fuels for sulfur isotope analyses. 1) In the quartz combustion method (ASTM, 1958), chemically scrubbed air or 02 is passed over the sample in a horizontal porcelain or quartz combustion boat at temperatures above 950~ The combustion products are passed through a H202 solution, which adsorbs sulfur as dissolved sulfate. This technique has never gained widespread popularity for sulfur isotope work. 2) In the bomb combustion method (ASTM, 1964), all sulfur compounds in a sample are converted to sulfate in a closed high temperature reactor under high pressures (2.5 MPa) of 02 (Siegfriedt et al., 1951). Suitable bombs are manufactured by the Parr Instrument Company (Moline, Illinois, USA). In Parr bomb model 1108, up to I gram of material can be placed in a sample cup. One mL distilled water is added to the bomb and oxygen gas is admitted slowly prior to closing. Subsequently, the bomb is placed in a water bath and the sample is ignited by passing a current through a thin Ni alloy wire placed above the cup. Sulfur in the sample is quantitatively converted to SO3, which reacts with the water in the bomb to H2SO4. Circa 15 min after ignition, unexpended 02 and produced CO2 are released and the bomb is opened. The dissolved sulfate is carefully transferred into a beaker through a filter paper (Whatman # 1) and converted to BaSO4 by adding BaC12 solution. A 1 gram sample with a total S content of 0.1% should yield 7.3 mg of BaSO4, sufficient for both dual inlet and continuous flow IRMS. 3) Eschka mixture (ASTM, 1993) has also been successfully used to extract total sulfur from coal and coke (see section 26.3.3.1), but one challenge with this reagent is finding a supply with sufficiently low sulfur blank. Isotope data for total sulfur in oil and bitumen have provided information on their origin, migration, and alteration processes (e.g. Thode, 1981; Thode et al., 1958). Thode & Monster (1970) demonstrated that oil accumulations in the Tertiary and Cretaceous of Northern Iraq had a common origin with extensive vertical migration in contrast to postulating four different epochs of oil formation. Other studies identified changes in composition of oils and condensates with H2S during maturation (Manzano et al., 1997; Orr, 1974). In these and other studies, the isotopic data proved more diagnostic when used in combination with data for other parameters. In a study of the Bolivar Coastal Fields (Venezuela), Manowitz et al. (1990) found an inverse relationship between ~34S values and pristane/phytane ratios of crude oils. They concluded that the five major oil classes were derived from a reasonably uniform source rock and that minor isotopic variability resulted from alteration processes. Despite the successful studies cited above, it is important to realize that isotope data for total S tend to have limited usefulness. This is particularly true for coals with biogenic pyrite having 634S values more variable and quite different from those of organic sulfur (Dai et al., 2002; Lei et al., 1994; Tang et al., 2001).
Procedures for Sulfur Isotope Abundance Studies
579
26.3.4.1 Coal and bitumen
Separation of different forms of sulfur from coal and bitumen specimens is not straightforward. Occasionally S-containing minerals such as pyrite or elemental sulfur can be handpicked for subsequent isotope analysis. Various analytical schemes are used for disseminated S compounds, which may involve extraction of most inorganic S first, most organic S first, or alternate extraction of inorganic and organic components. One of the fundamental papers on the isotopic composition of sulfur in coal was published by Smith & Batts (1974). They sequentially extracted elemental sulfur, acid soluble sulfate, pyritic S, organic S, and barite if present. One problem with their analytical scheme, as noted by the authors, is that pyritic S was extracted with bromine and nitric acid, which also mobilizes some organic sulfur. Bottrell et al. (1994) found that the liberated organic S is enriched in 32S compared to the total organic S. Therefore, we recommend chromium reduction for extracting pyrite S from coals, since this procedure does not mobilize significant amounts of sulfate or organic S (Bottrell et al., 1994; Canfield et al., 1986). For a basic characterization of four major S forms in coal (AVS, acid soluble sulfate, CRS, organic S), the extraction scheme shown in Figure 26.4 has proven satisfactory (e.g. Chu et al., 1994; Westgate & Anderson, 1982). A more rigorous procedure for extracting various forms of S from organic-rich sediments was suggested by Zaback & Pratt (1992) and is summarized in Figure 26.6. It features a 48 hour Soxhlet extraction using methylene chloride to divide the sample into a soluble and an insoluble fraction. The soluble fraction comprises elemental sulfur, which is reacted with granular copper added to the collection flask. The copper sulfide is subsequently removed for isotope analysis. The remaining sulfur compounds in the soluble fraction are termed bitumen S, which is converted to BaSO4 using a Parr bomb for subsequent isotope analysis. The insoluble sample residue undergoes extraction of AVS with hot 6 M HC1 + SnC12 followed by extraction of acid-soluble sulfate and chromium-reducible sulfur as described in section 26.3.3.4. For barite-free samples, the residue after chromium reduction is rinsed, dried, and treated with 48% HF in order to isolate kerogen S, which is converted to BaSO4 by Parr bomb combustion (Zaback & Pratt, 1992). If barite minerals are present, extraction of organic sulfur with hydrogen peroxide (Hall et al., 1988; Smith & Batts, 1974) followed by Kiba reduction to convert barite-S to H2S appears to be a preferable procedure (Hall et al., 1988). Oxygen isotope analyses on sulfate in coals have also been attempted, but special precautions (e.g. N2 atmosphere) are necessary to ensure extraction of sulfate with negligible effects on other sulfur compounds in the sample (McCarthy et al., 1998). The organic component remaining after removal of inorganic minerals including carbonates and silicates is defined as kerogen (e.g. Hitchon, 1974), which is converted to crude oil and natural gas during maturation. Comparison of the carbon isotope compositions of different solvent extracts of petroleum and kerogen has been successful in identification of oil source-rock relations (e.g. Stahl, 1977). Kerogen is interesting in that as it matures, the N / C ratio does not alter much but the S/C ratio can decrease by I to 2 orders of magnitude (Durand et al., 1972). Few studies have been carried out on its S isotope composition (e.g. Werne et al., 2003). A study by Hitchon & Krouse
580 Chapter 26 - B. Mayer & H.R. Krouse
Figure 26.6 - Flow chart summarizing analytical procedures for extracting bitumen S, elemental S, monosulfide S (acid volatile S), acid-soluble sulfate, disulfide S (chromium-reducible S), and kerogen S from organic-rich samples such as shales and coals (after Zaback & Pratt, 1992) for subsequent isotope analysis (dw = distilled water).
Procedures for Sulfur Isotope Abundance Studies
581
cited in Krouse (1977) revealed the following: whereas pyrite associated with bitumen at Windy Knoll (Derbyshire, England) was found to have ~34S values consistently near-10 %o, those of co-existing kerogen varied between 0 and-30 %o with an inverse dependence on 613C values. Purnell & Doolan (1983) recommended the use of Kiba reducing reagent (Sn2+ in H3PO4) under N2 flow at different temperatures for rapidly determining the concentrations of different forms of inorganic sulfur in coal. Initially, HC1 was added to the sample-containing flask to evolve H2S from non-pyritic sulfides at temperatures up to 120~ Then, Kiba reagent was added. H2S, which evolved between 170 and 210~ and 200 and 270~ were attributed to reaction with sulfate and pyrite, respectively. To test the suitability of this technique for isotope analysis, sulfate minerals and pyrite of widely different isotopic compositions were added to a low sulfur coal in our laboratory. Despite using finely ground mineral mixtures and different temperature-time programming profiles, the isotopic data showed considerable overlap in the H2S evolution peaks due to sulfate and pyrite. It may be possible to collect and analyze many H2S aliquots to properly characterize these evolution peaks, but clearly the simplified approach as published (Purnell & Doolan, 1983) is not recommended for obtaining S isotope data. Purnell & Doolan (1983) also found no evidence of reaction between several aromatic S compounds and Kiba reagent. Whereas their conclusion that organic S in coal does not interfere with the S determinations for inorganic sulfate and sulfide is probably valid, some organic S compounds can be converted to H2S with Kiba reagent (Sasaki et al., 1979). Although it is known that the contents of organic S within individual macerals of coal samples may vary significantly (Demir & Harvey, 1991), few researchers have attempted to analyze the isotopic compositions of different organic S compounds in coal or bitumen. Monster (1972) carried out S isotope analyses on oil sand bitumen from the McMurray Formation (Alberta, Canada). The oil sand bitumen was separated into asphaltenes, saturates plus aromatics, dark oils, and three classes of resins by liquid-solid chromatograhy on activated clay using various organic solvents as depicted in Figure 26.7, followed by converting the individual compounds to BaSO4 with the Parr bomb technique. The S isotope compositions were remarkably uniform among the different S fractions with deviations of less than 1.2 %o from the 634S value of total S in bitumen (Monster, 1972). However, these traditional extraction techniques do not separate different S functional groups. When this is done using the technique described below, the ~34S values may vary by over 20 %o in a single sample (unpublished data, Isotope Science Laboratory, University of Calgary). Attar and coworkers (e.g. Attar, 1979) have developed a method for analyzing different sulfur functional groups in coal and heavy oil. It is based on the premise that all organic S functional groups can be reduced to H2S with strong reducing agents. Their rates of reduction are characterized by unique activation energies. If a sample containing many S groups is reduced during gradual increase of temperature, H2S from an individual functional group evolves as a peak in a characteristic temperature range. Experiments with model compounds containing aliphatic thiols, thiopenes, and aryl
582
Chapter 26 - B. Mayer & H.R. Krouse
Figure 26.7- Flow chart summarizing analytical procedures for extracting asphaltenes, saturates + aromatics, dark oils, and various resins from bitumen samples for subsequent sulfur isotope analysis (after Monster, 1972). More recently, many laboratories have replaced benzene with less carcinogenic toluene
sulfides realized 94 to 99 % S recovery (Attar & Dupuis, 1979). Recovery from pure crystalline pyrite was low, ~ 1%. One complication was considerable overlap of the H2S evolution peak from pyrite with that of aliphatic sulfides. Krouse et al. (1987b) used the findings of Attar and colleagues as an approach to determining the S isotope compositions of different functional groups in bitumen and coal samples. The use of organic reducing agents was not as successful as linear temperature increasing (3~ min-1) pyrolysis with N2 flushing. The results were consistent with those of Attar's group with the following temperature range assignments: 200~ to 325 ~ C: elemental sulfur; 325~ to 375~ thiols and disulfides; 375~ to 500~ saturated sulfides; 500~ to 650~ thiopenes; and > 650~ benzothiopenes. Kinetic analyses of the peaks showed that above 500~ H2S was evolved from a solid phase whereas evolution from a liquid phase occurred below 500~ Since this initial work by Krouse et al. (1987b), pyrolysis has been replaced by hydrogenation, i.e. flushing with a 1:20 He:H2 mixture. In the earlier work, 2 to 5 gram of sample (liquid or solid ground to 40 mesh) to a height of a few centimeters were placed in a tube above a horizontal ceramic frit (Figure 26.8). There was concern that H2S released at lower temperatures might participate in sulfurization reactions before evolution from the sample. With introduction
Procedures for Sulfur Isotope Abundance Studies
583
of CF-IRMS, it was possible to reduce the sample size to the order of 100 mg and a height of only few millimeters. The revised apparatus (Figure 26.8) tends to give better resolved H2S peaks and the data are consistent with the earlier work. A number of extractions were conducted with sulfide and sulfate minerals of widely different ~34S values added to low-S coal and bitumen. There was no isotopic evidence of H2S evolved from these minerals. In contrast, Krouse et al. (1987b) showed that addition of elemental S to a bitumen sample produced a distinct H2S evolution peak in the 200~ to 325~ region. Comparison of data by Monster (1972) with the programmed H2S evolution technique is quite revealing. Using whole bitumen from the Fort McMurray oil sands, two resolved H2S evolution peaks were found; the larger in the temperature range identified with saturated sulfides and the smaller at higher temperatures associated with thiopenes. In addition, a minor unresolved peak occurred on the low temperature edge Figure 26.8- Apparatus for extracting sulfur from of the major peak. The 634S values of different functional groups in a coal or heavy oil sample. these S sources were found to be +6, 0, and +7 %o respectively. An asphaltene extract had the same H2S evolution peaks with the one at 0 %o being comparatively slightly more abundant. This is consistent with Monster's finding that asphaltene had a slightly lower bulk sulfur 634S value (Monster, 1972). With a sufficient number of consecutively evolved H2S samples, the temperature programmed reduction procedure is potentially a reliable tool for determining the S isotope composition of different S functional groups. It is limited to solids and high viscosity liquids. Materials with higher vapor pressures are likely to distill out of the reactor before H2S is produced. 26.3.4.2 Oil
Traditional evaluation of petroleum quality was based on separation and measurement of the content of different fractions as defined by a combination of distillation and solvent extraction techniques. Monster (1972) carried out S isotope analyses on fractions prepared from oil samples of the Mission Canyon Formation (Saskatchewan, Canada). The oils were divided into low and high boiling fractions. From the former,
584
Chapter 26 - B. Mayer& H.R. Krouse
Figure 26.9 - Flow chart summarizing analytical procedures for extracting asphaltenes, saturates, aromatics, and polar compounds for subsequent sulfur isotope analysis (after Monster, 1972). More recently, many laboratories have replaced benzene with less carcinogenic toluene. asphaltenes were precipitated by mixing with pentane and subsequent filtration. The de-asphaltened sample was further subdivided into saturates, aromatics, and polar compounds by liquid-solid chromatography on an activated F-20 Alcoa alumina column. After placing the sample on top of the column, saturates were eluted with npentane, aromatics with benzene, and polar compounds with a 1"1 mixture of benzene and methanol as depicted in Figure 26.9 (note that more recently benzene has been replaced in many laboratories by the less carcinogenic toluene). Subsequently, the individual fractions were Parr bombed to produce BaSO4 for sulfur isotope analysis. Despite increasing 634S values with biodegradation, the different solvent extracts from a given sample were markedly uniform, the greatest deviations being found for asphaltene with 634S values ~ 1%o lower than those of total S (Monster, 1972). It would be interesting to ascertain whether the ~)34Svalues of individual S functional groups changed during biodegradation. 26.3.4.3 Natural gas Sulfur in natural gas occurs predominantly in the form of H2S (sour gas). Hydrogen sulfide may be generated by bacterial sulfate reduction (BSR), thermochemical
Procedures for Sulfur Isotope Abundance Studies
585
sulfate reduction (TSR) and desulfurization reactions. The highest concentrations of H2S (up to 95%) typically arise during TSR. Sulfur isotope analyses have proven useful in identifying and quantifying these processes (e.g. Krouse, 1977). Natural gas samples are collected from exploratory or recovery wells and refineries using stainless steel cylinders with high pressure valves on each end. Expensive cylinders are available with glass or polymer linings to reduce adsorption of H2S. However, if these linings break down, retention of H2S may be worse than if the interior is not lined. Field sampling of natural gas is frequently done by service companies and most stable isotope laboratories simply receive the filled cylinders. A reliable service company will heat and evacuate these cylinders between samplings. The practice of attaching the cylinder to the field or refinery piping and flushing with the gas is not recommended and highly dangerous. Some gas samples may have pressures of tens of atmospheres and up to 90% H2S. During field sampling, an evacuated cylinder should be attached to the "plumbing". Appropriate valves are opened and left open for a few minutes to ensure that the cylinder contents have the same physical and chemical properties (e.g. temperature) as the system. Because H2S strongly adsorbs to surfaces, two sets of sampling cylinders are desirable for gases with low and high H2S contents respectively. If cylinders arrive in a sealed box, it should be opened in a fume hood in case that H2S is present from leakage. In the laboratory, the natural gas is slowly passed through a cadmium acetate solution, where H2S is quantitatively converted to CdS. This precipitate is converted to Ag2S for subsequent sulfur isotope ratio mass spectrometry as described above. A suitable apparatus for sampling H2S from natural gas and converting it to Ag2S is shown in Volume II, Part 3, Figure 17.13 (Chapter 17.6). Other sulfur compounds present in smaller quantities in natural gas include mercaptan and alkane sulfides. To our knowledge, no sulfur isotope ratios have been reported in the literature for these compounds. In our experience, GC separation of Scontaining gases does not proceed satisfactorily using Poropak Q packed columns. H2S desorbs very slowly producing memory problems from sample to sample and poor reproducibility of 634S values. 26.3.5 Sulfur in peat and soils Sulfur in peat and soils occurs in both organic and inorganic forms. Organic S constitutes often more than 80 % of total S. Organic S is commonly subdivided in two major groups: (1) organic (or ester) sulfates, which constitute a very labile fraction of the organic S pool (Freney, 1986), and (2) carbon-bonded sulfur, which includes S-containing amino acids (e.g. cystine, cysteine, methionine) as well as sulfonates (Biederbeck, 1978). Inorganic S may occur in both oxidized and reduced forms. Inorganic sulfate may exist in water soluble, adsorbed, or insoluble forms (Bohn et al., 1986). In water-logged soils, various reduced inorganic S compounds including elemental sulfur, monosulfide and disulfide minerals may also be present (see section 26.3.3.2). Total S contents in peat and soil vary from more than 0.1% to less than 0.01% depending on peat type or soil horizon.
586
Chapter 26 - B. Mayer & H.R. Krouse
Proper sample handling is essential for obtaining meaningful isotope ratios for peat and soil sulfur compounds. If reduced inorganic S compounds are expected, samples should be taken and transported under N2 or Ar atmosphere to prevent their oxidation. Extraction of these sulfur compounds should occur immediately after return to the laboratory while minimizing the exposure of the field-moist samples to oxygen. In aerated soils, contents of reduced inorganic S compounds are typically negligible. In that case, samples can be transported in plastic or paper bags, while temperatures should be kept < 4~ Upon return to the laboratory, soil material from aerated horizons should be dried at temperatures below 40~ Drying at higher temperatures may result in conversions or loss of some S compounds (Amaral et al., 1989; David et al., 1989; Wieder et al., 1985). Dried soil samples should be sieved (< 2 mm) to remove rocks and roots and ground where appropriate to provide a homogeneous representative sample. In principle, all techniques described in section 26.3.3 are also applicable to isotope analyses on peat and soil S compounds. However, the special characteristics of these materials including their high organic S contents, significant contents of organic sulfates, and occasionally large amounts of adsorbed sulfate, require special considerations in selecting the most appropriate extraction procedures.
26.3.5.1 Total sulfur Several studies have determined isotope ratios of total sulfur in peat and soils (e.g. Bottrell & Novak, 1997; Kusakabe et al., 1976). A variety of techniques are suitable to convert total S to either Ag2S or BaSO4. For example, Ag2S can be generated by converting all S compounds to sulfate via alkaline oxidation (Tabatabai & Bremner, 1970) followed by Johnson-Nishita reduction (Johnson & Nishita, 1952) in a distillation apparatus (see Part 3, Volume-II, Figure 8.1-13) as described by Schoenau & Bettany (1988). Complete conversion particularly during the reduction step is essential to avoid isotope fractionation and thus erroneous results. An alternate method is to convert all peat or soil S compounds to sulfate followed by precipitation of BaSO4. Alkaline oxidation (Tabatabai & Bremner, 1970) results in high DOC contents in the sulfatecontaining solution making precipitation of pure BaSO4 challenging. Therefore, wet chemical oxidation using HNO3/Br2 or other mixtures (Krouse & Tabatabai, 1986; Zhabina & Volkov, 1978) is often preferred (see section 26.3.3.1). Another widely accepted procedure is fusion with sodium carbonate and an oxidizing agent (Tabatabai, 1992). Whereas Kusakabe et al. (1976) relied on a 10:1 mixture of NaHCO3 and Ag20, the commercially available Eschka mixture (see section 26.3.3.1) has been more widely used. All these techniques generate SO4 2-, which is subsequently precipitated as BaSO4. The precipitate is subsequently filtered, weighed, and converted to SO2 or SF6 as described in section 26.2. Although many studies have evaluated isotope ratios of total sulfur in peat and soils (e.g. Kusakabe et al., 1976), more insight can be gained by determining the 634S values of individual S compounds (Chae & Krouse, 1986; Mayer et al., 1995; Schoenau & Bettany, 1989). Selection of the appropriate analytical extraction scheme for peat and soil samples depends upon whether reduced inorganic S compounds are present
Procedures for Sulfur Isotope AbundanceStudies
587
or not. For the former sample type, sequential extraction procedures are often preferred (see section 26.3.5.2). For aerated soils with negligible contents of reduced inorganic sulfur, extraction of individual sulfur compounds from fresh sample aliquots is often more satisfactory (see section 26.3.5.3). Also, a variety of combinations between sequential and individual extractions from fresh sample aliquots are reported in the literature (e.g. Mandernack et al., 2000; Morgan & Mandernack, 1996; Novak et al., 2003a,b; Wieder & Lang, 1988), too numerous to be fully described in this chapter. In the following, we provide a general outline of some of the available analytical options in anticipation that this provides the reader with sufficient information to design a custom-made extraction scheme, which is best suited for the respective sample material of interest.
26.3.5.2 Sequential extraction of individual sulfur compounds Peat and soil samples containing reduced inorganic S compounds should be immediately processed after return to the laboratory. This can be accomplished by using a sequential extraction procedure, which recovers reduced inorganic sulfur compounds prior to further processing of the sample material (Figure 26.10). Fieldmoist samples are transferred to the distillation apparatus shown in Part 3, Volume-IL Figure 8.1-13, while minimizing exposure to atmospheric 02. One option is to extract
Figure 26.10 - Flow chart summarizing analytical procedures for sequential extraction of various sulfur compounds from peat and soil samples (dw = distilled water).
588
Chapter 26 - B. Mayer & H.R. Krouse
the sum of all monosulfides, elemental S, and disulfides by chromium reduction (Canfield et al., 1986) as described in section 26.3.3.2 and shown in Figure 26.10. Alternately, these sulfur compounds can also be extracted individually as described in section 26.3.3.2. In both cases it is difficult to avoid that some organic S compounds are also converted to H2S (Amaral et al., 1993; Brown, 1986). Rinsing the residual sample with hot distilled water yields the fraction of acid soluble sulfate. It is noteworthy that this fraction may contain sulfate generated by hydrolysis of organic sulfates during the prior acid treatment. The residual sample material can be either treated with further sequential extraction steps as shown in Figure 26.10, or as described in section 26.3.5.3. For further sequential extractions it is essential that all inorganic sulfate, including the adsorbed fraction, is completely recovered to avoid erroneous carryover into subsequent fractions. It is therefore recommended to subject the residual sample to an additional sulfate extraction with 16 mM KH2PO4 (see section 26.3.5.3). The remaining sample material after this extraction step contains only organic sulfur compounds. Determination of the 634S values of the total organic sulfur fraction (organic sulfates and carbon-bonded S) can be achieved by processing the remaining sample material with any of the techniques described in section 26.3.5.1. Alternately, organic sulfates, which may constitute between less than 10 to more than 80 % of the organic soil S fraction (e.g. Bettany et al., 1979), can be extracted via reduction with a mixture of hydriodic acid (HI), formic acid (HCOOH), and hypophosphoric acid (H3PO2) in a ratio of 4::2:1 in a distillation apparatus (Part 3, Volume-II, Figure 8-1.13 ). This process is often referred to as Johnson-Nishita reduction (Johnson & Nishita, 1952). The Johnson-Nishita (or HI) reduction mixture converts both inorganic and organic sulfate to H2S (Freney, 1961), a fraction which is often referred to as total sulphate or HIreducible S (SHI-red). Since in the sequential extraction scheme inorganic sulfate has been quantitatively removed from the peat or soil sample (Figure 26.10), H2S generated via HI reduction is solely released from organic sulfates and its 634S value can be directly determined (Spratt & Morgan, 1990; Wieder et al., 1985). The sample residue remaining after quantitative extraction of inorganic S (reduced inorganic S and sulfate) and HI reduction contains only organic sulfur in carbon-bonded form. To determine its 634S value, the remaining sample material can be subject to any total S treatment described in section 26.3.5.1 yielding BaSO4 or Ag2S for sulfur isotope ratio measurement (Figure 26.10). Further subdivision of the carbon-bonded S fraction has been seldom attempted for isotope analysis. Reduction with Raney-Nickel alloy supposedly differentiates between S-containing amino acids, which are converted to H2S, and other carbon-bonded non-reducible organic S compounds (DeLong & Lowe, 1961; Lowe, 1965). However, it has been found that the amount of Raney-Nickel alloy affects the results and that Fe and Mn may interfere with the determination of the carbon-bonded S fractions (Freney et al., 1970). Therefore, this technique has never gained widespread popularity. Schoenau & Bettany (1988) extracted humic and fulvic acid fractions from soils using HC1 and NaOH solutions and converted S in the extracts by alkaline oxidation followed by Johnson-Nishita reduction (Tabatabai &Bremner, 1970) to Ag2S for subsequent isotope analysis.
Procedures for Sulfur Isotope Abundance Studies
589
Figure 26.11 - Flow chart summarizing analytical procedures for extracting individual sulfur compounds from fresh aliquots of peat and soil samples.
590
Chapter 26- B. Mayer & H.R. Krouse
26.3.5.3 Extraction of individual sulfur compounds using fresh sample aliquots Extraction of individual sulfur compounds from fresh aliquots of the same sample is an attractive alternative for peat and soil samples, particularly if they have negligible contents of reduced inorganic S compounds. This approach has the advantage that incomplete recovery of inorganic sulfate does not cause erroneous results for subsequently extracted sulfur compounds as is the case during sequential extractions (see section 26.3.5.2). Also, the determination of organic sulfur fractions is presumably more accurate. Some options for extracting sulfur compounds from peat and soil materials are summarized in Figure 26.11. Field-moist samples can be pressure-filtered (e.g. Wieder & Lang, 1988) and, after filtration of the pore water, dissolved sulfate is precipitated as BaSO4 as described in previous sections. Reduced inorganic sulfur compounds including elemental S (Maynard & Addison, 1985), mono-sulfides, and disulfides can be extracted together or individually as described in section 26.3.3.2 on fresh aliquots of compressed peat or soil yielding Ag2S for mass spectrometric determinations. Complete extraction of inorganic sulfate in aerated soils without mobilizing organic S compounds is a major challenge. Dry and sieved samples are typically mixed with an extraction solution and shaken in an overhead shaker for 18 hours (~10 rpm) at a solution to sample ratio of 5:1 by weight for mineral soils and 10"1 for litter horizons. Subsequently, the extraction solution is separated from the soil material by centrifugation and filtration with 0.45 ~m membranes, and sulfate is converted to BaSO4 by adding BaC12 solution. Contents of inorganic sulfate-S in many peat and soil samples vary between 10 and 100 mg kg-1, and therefore up to 100 g of soil must be extracted to yield sufficient quantities of BaSO4 for isotope analysis using dual inlet IRMS. The choice of extraction solution is dependent on soil horizon and soil properties. Different concentrations of phosphate, carbonate, chloride, and acetate solutions, as well as deionized water have been recommended in the literature for concentration analysis of soil sulfate (Tabatabai, 1982), but not all of these are suitable for isotope measurements. An important consideration is that conversion of organic soil S to sulfate should be minimized during the extraction. Since 634S values of organic soil S compounds may differ significantly from those of inorganic sulfate, conversion of even small parts of the often large organic S pools may result in e r r o n e o u s ~34S values for inorganic sulfate. For litter horizons, extraction with distilled water has proven satisfactory. For mineral soils, extraction with distilled water often recovers only a small part of the inorganic sulfate pool. This is because a large fraction of inorganic sulfate may be adsorbed to clay minerals and Fe and A1 oxides or hydroxides, particularly in acid soils. Therefore, slightly alkaline extraction solutions with high charge densities to release sulfate from the sorption sites are preferred for mineral soils. It is widely believed that 16 mM KH2PO4 or NaH2PO4 are the most suitable solutions for quantitative extraction of sulfate from mineral soils (Ensminger, 1954), while minimizing mobilization of organic S. However, one problem with these solutions is that addition of BaC12 results in co-precipitation of BaSO4 with large quantities of Ba3(PO4)2. Hence, the precipitate must be further treated with either Kiba, Thode, or JohnsonNishita reduction to convert sulfate-S into H2S and subsequently Ag2S (Johnson &
Procedures for Sulfur Isotope Abundance Studies
591
Nishita, 1952; Kiba et al., 1955; Thode et al., 1961) as described in section 26.3.3. Extraction of peat and soil samples with NaHCO3 (Kilmer & Nearpass, 1960) is not recommended, because the high pH value of this solution may facilitate the conversion of organic S to sulfate. Since there is no significant isotope fractionation during sulfate adsorption and desorption (Van Stempvoort et al., 1990), incomplete recovery of soil sulfate does not prevent the determination of ~34S values representative for the entire inorganic sulfate pool in soils. In our experience, extraction with 0.1 M LiC1 solution yields very reliable 634S values for inorganic soil sulfate, although it is known that this solution only partially removes adsorbed sulfate because of the low charge density of C1- (Krouse et al., 1996). An advantage of this extraction solution is that the Li ion may act as a metabolic inhibitor, minimizing the microbial conversion of organic S to sulfate during the extraction procedure. In carbonate containing soils, sulfate co-precipitated with calcium carbonate can be released by extraction with 1M HC1 (Roberts & Bettany, 1985), but conversion of some organic S to sulfate is difficult to avoid (e.g. Amaral et al., 1993). Reduction with a mixture of hydriodic acid (HI), formic acid (HCOOH), and hypophosphoric acid (H3PO2) in a ratio of 4:2:1 in a distillation apparatus (Part 3, Volume-IL Figure 8.1-13) reduces both inorganic and organic sulfate to H2S (Freney, 1961). The H2S from this so-called Johnson-Nishita reduction is converted to Ag2S. This precipitate represents total sulfate, comprised of both the organic sulfate and inorganic sulfate fractions. Knowledge of the sample masses, which have undergone extraction, and precise weighing of the obtained Ag2S or BaSO4 precipitates provides an approximate estimate of the contents of the individual sulfur fractions. There are also alternate chemical techniques for precise determination of the contents of many of the previously described sulfur compounds (e.g. Tabatabai, 1992). If the concentrations (C) and isotope ratios (634S) of all sulfur compounds shown in Figure 26.11 are precisely determined, it is possible to calculate contents and ~34S values of sulfur fractions, which can not be directly extracted from fresh aliquots of peat and soil samples. For example, the content of organic sulfate is the difference between total (HI-reducible) sulfate and inorganic sulfate: Corg sulfate = CHI-red - Cinorg sulfate
[26.11]
value of organic sulfate can be subsequently determined by mass and isotope balances according to equation [26.12]"
T h e 634S
~34 S
_ CHI - red ~ ~34SHI _ red - C inorgsulfate " ~34S inorg sulfate org s u l f a t e C org sulfate
[26.12]
In case that concentration and ~34S values of total sulfur (see section 26.3.5.1) were accurately determined, it is also possible to calculate contents and isotope ratios of the organic sulfur and the carbon-bonded sulfur fractions. Organic sulfur is the difference
592
Chapter 26 - B. Mayer & H.R. Krouse
between contents of total S and all inorganic S fractions (including reduced species and inorganic sulfate). [26.13]
CSorganic - CStotal - CSinorg
In many aerated soils with negligible amounts of reduced inorganic S, inorganic S is equivalent to the inorganic sulfate fraction. The 634S value of the entire organic sulfur pool can be determined as follows: 34 9 ~)34S Cstotal 9 ~) S s t o t a l - Csinorganic Sinorganic (~34Sorganic = Corganic
[26o14]
Contents of carbon-bonded sulfur can be calculated as the difference between contents of total sulfur and HI-reducible S (total sulfate) and all reduced S species (SCRS): [26.15]
CScarbon-bonded = CStotal - CS-HIred - CS-CRS
In aerated soils with negligible amounts of reduced inorganic sulfur compounds, carbon-bonded S equals: [26.16]
CScarbon-bonded = CStotal - Cs-HIred
In this case, the ~)34Svalue of carbon-bonded S can be calculated as follows:
~)34Scarbon- bonded -
Cstotal
9
34
(~ S s t o t a l - C s - H I r e d
9
634Ss-HIred
[26.17]
Cscarbon- bonded
One disadvantage of this approach is that the uncertainties of the calculated ~34S values become relatively large, if the respective organic sulfur pools are small, as may be the case in mineral horizons of acid soils.
26.3.6 Sulfur in plants Sulfur contents in plants vary from less than 0.05 % to more than 0.2 %. Typically, most of the sulfur in plants is bound organically, e.g. in S containing amino acids of proteins, whereas a smaller fraction of the total S occurs as sulfate (Blanchar, 1986). Interestingly, higher levels of total S in plants seem to be associated with an accumulation of sulfate. In those cases, sulfate may become the dominant S compound in plants. It is widely believed that organic sulfates do not occur in significant quantities in plant tissues. Proper sample handling prior to analysis should include initial storage of freshly sampled plant tissue at < 4~ removal of dust and surface contaminants by washing with deionized water, and subsequent drying at a maximum temperature of 65~ for several days (Jones & Steyn, 1973). Thereafter, samples should be finely ground using
Procedures for Sulfur Isotope Abundance Studies
593
a screen size of 40 mesh or finer. It is recommended to re-dry the samples again after grinding at 65~ for 24 hours to remove residual moisture prior to storage in sealed containers.
26.3.6.1 Total sulfur So far, most investigations of the sulfur isotope composition of plants have been based on the determination of 634S values of total sulfur (e.g. Chukhrov et al., 1980; Takala et al., 1991; Trust & Fry, 1992; Wadleigh & Blake, 1999; Yun et al., in press). A variety of techniques are suitable to convert total plant S to either Ag2S or BaSO4 for subsequent isotope ratio mass spectrometry. For instance, Ag2S can be generated by converting all S compounds to sulfate via alkaline oxidation (Tabatabai & Bremner, 1970) followed by Johnson-Nishita reduction (Johnson & Nishita, 1952) as described by Schoenau & Bettany (1988). Alternate techniques comprise the Eschka method (ASTM, 1993), wet chemical oxidation using HNO3/Br2 or other mixtures (Krouse & Tabatabai, 1986; Zhabina & Volkov, 1978), and Parr bomb oxidation (Siegfriedt et al., 1951), which have all been described in earlier sections (26.3.3.1, 26.3.4, and 26.3.5.1). All these techniques generate SO4 2-, which is subsequently precipitated as BaSO4 by adding 0.25M BaC12 solution. The precipitate is subsequently filtered, weighed, and converted to SO2 or SF6 as described in section 26.2. Total S in plant material can also be converted to SO2 in an elemental analyzer followed by CF-IRMS. One problem is that most biological samples produce at least 50 times more water and CO2 than SO2 and the former should be completely removed prior to SO2 entering the mass spectrometer. Monaghan et al. (1999) obtained reliable sulfur isotope ratios for wheat plants with > 1 mg S g-1 dwt. using an on-line continuous flow system. To avoid tailing of the SO2 peak, these authors used removable quartz liners in the combustion tube, PTFE couplings to avoid metal surfaces, and a Nation drying tube instead of the usual magnesium perchlorate water trap. They also prevented CO2 from entering the ion source by using two GC columns. A maximum of 10 mg sample was used together with 20 mg vanadium pentoxide to aid the combustion/oxidation process. Using these modifications, Monaghan et al. (1999) were able to obtain reliable (~34S values for total S in wheat plant samples containing between 10 and 25 ~g S, but samples much below 1 mg S g-1 were not suitable for analysis with their technique. Yun et al. (in press) used a similar technique to determine 634S values of total S in lichen samples. These authors obtained excellent results by weighing up to 15 mg of sample equivalent to ~ 10 ~g S in ultra-light Sn capsules together with V205. However, we reiterate that it is not trivial to normalize the obtained data to the international V-CDT scale (see section 26.2).
26.3.6.2 Individual plant sulfur compounds Few studies have attempted to analyze sulfur isotope ratios for distinct sulfur compounds in plants (e.g. Schoenau & Bettany, 1989). Sulfate-S can be converted to H2S by Johnson-Nishita reduction of plant material without prior oxidation (Johnson & Ulrich, 1959). The generated H2S can be trapped as CdS or ZnS, which is subsequently converted to Ag2S for isotope analysis. Removal of plant sulfate with hot
594
Chapter 26 - B. Mayer & H.R. Krouse
water, CH3OOH, or 0.1 M HC1 solutions has also been attempted, but these extraction solutions do not always achieve sulfate yields comparable to those from JohnsonNishita reduction. Hence, we recommend the latter method as a reliable procedure for accurately determining the content and isotope ratio of plant sulfate. Work in our laboratory has revealed that 634S values of sulfate in pine needles can deviate by more than 3 %o from those of total sulfur. Using the residue of the Johnson-Nishita reduction for subsequent total S analysis is a suitable method to determine the isotopic composition of organic plant sulfur. Alternately, if the contents and ~534Svalues of both the total S and inorganic sulfate fractions are known for a sample, the ~)34Svalue of plant organic S can be calculated by mass and isotope balances according to equations [26.13] and [26.14]. Few investigators have attempted to extract specific organic S compounds from plant material. For example, allylisothiocyanate from mustard plants has been isolated from plant material for sulfur isotope analysis (Remaud et al., 1997c).
26.3.7 Sulfur in humans, animals, and other materials Sulfur isotope analyses can be extremely useful in food web studies, since they often allow the identification of food sources e.g. of marine versus terrestrial origin (e.g. MacAvoy et al., 1998, 2000; Petersen et al., 1986). Sulfur isotope ratio measurements are also increasingly used in tracing origins and migration of wildlife (Hobson, 1999). Many tissues, fluids, and minerals in biological specimens contain sufficient sulfur for isotopic measurements (e.g. Katzenberg & Krouse, 1989). So far, many studies have focused on the isotopic analysis of total sulfur in biological materials including but not limited to muscle (Hesslein et al., 1991; Kwak & Zedler, 1997), liver (Hesslein et al., 1993), blood and skin of fish (MacAvoy et al., 2001), muscle and liver tissues of rats (Hobson et al., 1999), eggs (Hobson et al., 1997), feathers and muscle tissue of birds (Kwak & Zedler, 1997), and hair (Krouse et al., 1987a). Many of the previously described extraction techniques for total sulfur are applicable, including the Eschka method (section 26.3.3.1), HNO3-Br2 oxidation (section 26.3.3.1), and Parr bombing (section 26.3.4). Isotope ratios of total sulfur in tissues can also be analyzed by CF-IRMS techniques (e.g. Hesslein et al., 1991). Recently, ion microprobes have been employed to determine 634S values of total sulfur in fish otoliths with external precisions ranging between I and 5 %o depending on the instrument used (Weber et al., 2002). The reader is referred to previous sections of this chapter for analytical details. Hair is attractive to study because the animal is not sacrificed, the S content is high (~ 4 % mainly in form of keratin), and it is resistive to degradation. Seemingly reliable data have been obtained with 12,000 year old specimens from Siberian mammoth remains and 3,000 year old human hair (Aufderheide et al., 1994). Feathers and claw/ finger nails have similar advantages. All these materials can be easily Parr bombed to generate BaSO4 or converted to H2S and subsequently Ag2S with Kiba reagent (Sasaki et al., 1979). It appears that a number of S-containing amino acids and other compounds in tissues and fluids are capable of conversion to H2S with Kiba reagent (Kiba et al., 1957; Ohashi, 1955).
Procedures for Sulfur Isotope Abundance Studies
595
Cystine is relatively insoluble. In some animals, kidney stones may form because of lack of the enzyme to convert cystine to more soluble cysteine. Total sulfur in these stones has been analyzed using the Parr bomb technique (Krouse et al., 1987a). The invacuo Kiba technique (section 26.3.3.2) was applied to phosphorous-containing minerals in kidney stones, bladder stones, and teeth (Krouse et al., 1987a; Krouse & Ueda, 1987). Sulfate in fluids such as blood and urine is difficult to isolate because S-containing organic matter is trapped in BaSO4 precipitates. In our laboratory, one technique used was to dilute the fluids and add BaC12 solution with vigorous stirring. The dried precipitate was further treated by heating at 500~ in an 02 stream. Sulfate in some body fluids is derived from oxidation of organic S. Therefore, dilution should be carried out with deoxygenated H20 and BaSO4 should be precipitated in a N2 atmosphere to minimize organic sulfur oxidation during sample processing. There are many reports in the literature describing chemical extraction of individual organic S compounds or compounds containing the same sulfur moiety (e.g. Mestres et al., 2000). In some cases, the yields are very low. However, if the molecules are large, isotope fractionation associated with the chemical procedures may be acceptably small. Interestingly, very few ~34S values have been obtained for such extracts. One example is methionine-bound S in milk casein (Pichlmayer et al., 1998).
26.4. Summary Traditionally, sulfur isotope ratio determinations comprised three steps" 1) extraction of S from the sample and conversion into BaSO4, Ag2S, or other pure S-containing compounds, 2) preparation of a measurement gas such as SO2 or SF6, and 3) isotope ratio mass spectrometry in dual inlet mode. In the days of off-line gas preparation and dual inlet mass spectrometry, sulfur isotope ratio determinations were cumbersome and labor intensive. Throughout the last 15 years, sulfur isotope abundance studies have benefited from the advent of new technologies. These have resulted in higher sample throughput and greatly reduced sample size requirements. For example, coupling of elemental analyzers to isotope ratio mass spectrometers in continuous flow mode (CF-IRMS) has automated the gas preparation step and amalgamated it with fully computer controlled isotope ratio determinations. Using pure inorganic sulfur compounds such as BaSO4 or Ag2S, it is possible to achieve a reproducibility of better than + 0.2 %o for sulfur isotope measurements within a given laboratory. However, these technical improvements have not eliminated the necessity of extracting the respective sulfur compounds from gaseous, aqueous, or solid samples. Reliable extraction of individual sulfur compounds from often complex sample matrixes remains the key for the successful use of sulfur isotope techniques in many case studies. Complete recovery of the sulfur compound of interest without mobilizing other sulfur fractions is often challenging. Some of the extraction procedures are operationally defined and many of them remain cumbersome and labor-intensive.
596
Chapter 26 - B. Mayer & H.R. Krouse
Repeated extraction of a specific sulfur compound from a given sample often reveals uncertainties in excess of +0.5 %o associated with the analytical procedure. Therefore, it is essential that researchers determine and report the overall reproducibility of (1) sulfur extraction, (2) gas preparation, and (3) mass spectrometric measurements in their publications. In many cases, the overall reproducibility of the entire analytical procedure may be 2 to 3 times larger than that of the mass spectrometer measurement. Fortunately, determinations of 634S values of specific sulfur compounds with an uncertainty of _+ 0.5 %o is more than sufficient for most case studies, since sulfur in nature is characterized by a wide range of sulfur isotope ratios. Unfortunately, there are also uncertainties regarding the values of international reference materials (see Table 26.1). Among different laboratories, discrepancies between measurements on reference materials with very high or very low 634S values have been detected, particularly if different measuring gases such as SO2 or SF6 are used. Therefore, it is essential that researchers include careful descriptions of methodologies used in their publications, and that they report ~34S values obtained for international reference materials to allow for a meaningful comparison of sulfur isotope data reported in the literature. There is no doubt that the number of stable sulfur isotope abundance studies will increase in the future as analytical instrumentation becomes more automated and sophisticated. This expansion will not be restricted to the traditional research themes in Geochemistry and Hydrology since there is a tremendous potential for new applications of sulfur isotope techniques in disciplines such as Ecology and Atmospheric Chemistry, among many others. We anticipate that the key for advancing sulfur isotope studies in the future will not primarily be the development of better instrumentation for sulfur isotope ratio measurements. Improving our ability to reliably extract specific sulfur compounds from complex sample matrixes is more critical for an improved understanding of the sulfur cycle in the atmosphere, biosphere, pedosphere, hydrosphere, and lithosphere. In particular, techniques for extracting different forms of organic sulfur for isotope ratio determinations are still in their infancy. Better chemical resolution of individual sulfur functional groups or advancements in the field of compound specific isotope analyses particularly of organic S compounds coupled with sulfur isotope ratio determinations appear highly desirable.
Handbook of Stable IsotopeAnalyticalTechniques, Volume 1 P.A. de Groot (Editor) 9 2004 ElsevierB.V. All fights reserved.
CHAPTER 27 Direct Measurement of the Content and Isotopic Composition of Sulfur in Black Shales by Means of Combustion-Isotope-Ratio-Monitoring Mass Spectrometry (C-irmMS) Michael E. Bfttcherl & Bernhard Schnetger2 Max-Planck-Institute for Marine Microbiology, Department of Biogeochemistry, Celsiusstr.1, D28359 Bremen, Germany 2 Carl-von-Ossietzky University, Institute for Chemistry and Biology of the Marine Environment (ICBM), P.O. Box 2503, D-26111Oldenburg, Germany e-mail: 1 [email protected], 2 [email protected] 1
Abstract
The content and sulfur isotopic composition of black shales (down to 0.1 wt.% S) were directly measured by means of combustion-isotope-ratio-monitoring mass spectrometry (C-irmMS), and the results are compared to the Kiba reagent method for sulfur isotope preparation and the coulometric method for determination of the concentration. The C-irmMS measurements were not disturbed by the common combustion of sulfur- and carbon-bearing compounds up to 13 wt.% C. The C-irmMS method was successfully applied to a set of different pure synthetic and natural sulfur-bearing compounds and natural shale geostandards. The results show a good agreement, indicating that C-irmMS is a powerful analytical tool both precise and fast in sample preparation, which needs only small amounts of sample material. 27.1 Introduction
The determination of the contents and concentrations and sulfur isotopic composition of sulfur species in modern and ancient sediments is of fundamental interest for the evaluation of biogeochemical reactions in the coupled sedimentary element cycles (e.g., Hartmann & Nielsen, 1969; Goldhaber & Kaplan, 1975; Chanton et al., 1987) and our understanding of the paleo-environment and the evolution of life (e.g., Schidlowski et al., 1983; Ohmoto, 1992; Strauss, 1997). Most analytical schemes developed for the separation of sulfur species from recent and ancient sediments for isotope analysis (e.g., Sasaki et al. 1979; Allen & Parkes, 1995) are based on the early work of Kiba et al. (1955) and Zhabina & Volkov (1978). Due to their different environments of formation, metal sulfides or sulfates are typically occurring as the main respective sulfur-bearing phases in these sediments and, therefore, a number of sulfur isotope studies were based on the analysis of total sulfur (e.g., Brumsack, 1980; Vet6 et al., 1994; Calvert et al., 1996; Bfttcher & Lepland, 2000). The classical off-line scheme for sample preparation is time consuming (e.g., Giesemann et al., 1994; Sasaki et al., 1979) and the
598
Chapter 27 - M.E. B6ttcher & B. Schnetger
preparation may cause some of the previously reported uncertainty in sulfur isotope determination. With the development of modern on-line techniques using a combination of an elemental analyzer directly coupled to a gas isotope mass spectrometer (Pichlmayer & Blochberger, 1988; Giesemann et al., 1994; B6ttcher et al., 1998a) most previous problems related to sulfur isotope measurements (e.g., isotope effects due to chemical conversion of sulfate to sulfide or memory effects in the mass spectrometer) were minimized. Combustion-isotope ratio-monitoring mass spectrometry (C-irmMS) was successfully applied to determine the sulfur isotopic composition of pure barium sulfate (e.g., Giesemann et al., 1994; B6ttcher et al., 1998a, b) and metal sulfides (B6ttcher et al., 1998c, 2001). In the present study we apply C-irmMS to determine the contents and stable sulfur isotopic composition of sulfur in Devonian black shales. The measurements are compared to those obtained with classical off-line preparation methods using the Kibareagent and a coulometric method for stable isotope and content analyses, respectively. The results show a good agreement, indicating that C-irmMS is a powerful analytical tool both precise and fast. 27.2 Materials and methods
34S/32S ratios were measured on-line by means of combustion isotope-ratio-monitoring mass spectrometry (C-irmMS) (Pichlmayer & Blochberger, 1988; Giesemann et al. 1994; B6ttcher et al., 1998a) using a Carlo Erba EA 1108 elemental analyzer connected to a Finnigan MAT 252 mass spectrometer via a Finnigan MAT Conflo II split interface as described by B6ttcher et al. (1998a). Sample amounts equivalent to 20 to 50/~g sulfur were wrapped together with reagent grade V205 as a catalyst in pure tin capsules. All samples were weighed with a high-precision micro balance (Sartorius MC1 Research RC 210P). Natural samples were carefully ground and homogenized prior to further analysis. Except for ZnS, the natural sulfides were drilled by a microdrilling device from polished sections. ZnS corresponds to the NBS-123 reference material, with a sulfur isotopic composition of +17.3 + 0.3%o vs. V-CDT (B6ttcher et al., 1997b). The Sn caps were combusted in a pulse of 02 (grade 4.6, Messer Griesheim) at 1100~ oven temperature, leading to a short increase of temperature in the reaction zone of about 1800~ (Giesemann et al., 1994; B6ttcher et al., 1998a). The liberated sample gas was transported in a continuous stream of He (5.0 grade; Messer Griesheim). Water was removed from the gas stream by a water trap filled with magnesium perchlorate, and SO2 was separated from other gas impurities by a chromatographic column (0.8 m length; PTFE tubing; HekaTech) at 80~ A split of the total gas stream was introduced into the gas mass spectrometer via a fused silica capillary using a Finnigan MAT Conflo II interface, and the ion currents of masses 66 and 64 in the sample gas were compared to the corresponding ion currents of external in-house standards which were combusted every 10 samples. Comparison was done via a commercial SO2 gas (3.8 grade; 666S1 close to CDT composition; Messer Griesheim) which was introduced via the Conflo II into the mass spectrometer. The ion currents were
1. with 666Sis meant the m/z peak 66, including both the sulfur and oxygen isotopes.
Direct Measurement of the Content and Isotopic Composition of Sulfur ...
599
Table 27.1 - Measurements of sulfur recoveries for synthetic and natural sulfur compounds by means of C-irmMS. Absolute range was between 20 and 200 ~g sulfur. Calibration substance to calculate the recovery was synthetic reagent grade sulfanilamide (Carlo Erba). Number of measurements in parenthesis. Composition
Recovery + S.D. (%)
Cometic
BaSO4
104.0 + 10.3 (41) 99.8 + 8.1 (2) 98.9 + 11.9 (16) 103.5 + 1.9 (2) 97.4 + 11.7 (9) 102.8 + 4.3 (3) 94.7 + 9.6 (6) 101.5 + 7.8 (10) 95.4 + 3.0 (6) 100.9 + 19.5 (6) 103.3 + 11.5 (2) 101.5 + 1.7 (2)
synthetic synthetic natural anhydrite natural langbeinite natural celestite synthetic synthetic natural sphalerite (NBS-123) natural galenite natural pyrite natural chalcopyrite natural elemental sulfur
(NH4)2SO4
CaSO4 K2Mg2(SO4)3 SrSO4 CdS Ag2S ZnS PbS FeS2 (CuFe)S2 S~
r e c o r d e d as a f u n c t i o n of time a n d i n t e g r a t e d for m a s s e s 64 a n d 66 u s i n g i n t e g r a t i o n time steps of 0.25 sec w i t h the F i n n i g a n MAT Isodat 5.2 software. The i n t e g r a t e d signal for m a s s 64 w a s c o m p a r e d to a calibration curve d e r i v e d from synthetic s t a n d a r d s w i t h k n o w n sulfur c o n t e n t (CdS, sulfanilamide, BaSO4) to o b t a i n the sulfur contents of black shale samples. Isotope ratios are g i v e n in the 6-notation v e r s u s the C a n y o n Diablo troilite (CDT) s t a n d a r d a c c o r d i n g to: ~34S[%o] - {(34S/32S)sample/(34S/ 3 2 S ) C D T - 1} 103. Replicate a n a l y s e s on p u r e sulfate or sulfide s a m p l e s g e n e r a l l y a g r e e d w i t h i n + 0.2%o. Synthetic b a r i u m sulfate a n d c a d m i u m sulfide s a m p l e s previously m e a s u r e d w i t h the off-line m e t h o d (see below) w e r e u s e d for calibration of the m a s s s p e c t r o m e t e r a g a i n s t the C D T - s t a n d a r d to m a k e b o t h black shale d a t a sets, o b t a i n e d by off-line a n d on-line p r e p a r a t i o n , c o m p a r a b l e . It s h o u l d be noted, h o w ever, that a n e w sulfur isotope s t a n d a r d (V-CDT; V i e n n a - C a n y o n Diablo troilite) h a s b e e n i n t r o d u c e d by the IAEA w h i c h is n o w r e c o m m e n d e d for i n t e r n a t i o n a l calibration of sulfur isotope ratios ( C o p l e n & Krouse, 1998). The a b s o l u t e 32S/34S ratio of VCDT is 22.6436 + 0.0020 (Ding et al., 2001). The isotopic c o m p o s i t i o n s of i n t e r n a t i o n a l rock reference materials g i v e n in Table 27.2 are, therefore, r e p o r t e d v e r s u s V-CDT. Table 27.2 -Comparison of the determination of total sulfur contents by means of C-irmMS with the coulometric method, and sulfur isotope ratios of total sulfur measured by C-irmMS. TW-TUC is an inhouse standard (schist), and SR-1 and Jet-Rock I are international oil shale standards. Number of measurements in parenthesis. Data are taken from B6ttcher et al. (1998b). Note: 634Svalues are given vs VCDT. Standard
S + S.D. (wt.%) Coulometry
S + S.D. (wt.%) C-irmMS
~)34S (%0) C-irmMS
TW-TUC SR-1, 23 Jet-Rock 1, 23
0.37 + 0.02 (32) 1.28 + 0.00 (2) 7.14 + 0.10 (2)
0.34 + 0.03 (2) 1.29 + 0.06 (2) 7.31 + 0.10 (2)
+1.2 + 0.2 (2) +0.9 + 0.1 (2) -17.5 + 0.4 (2)
600
Chapter 27 - M.E. B6ttcher & B. Schnetger
Table 27.3 - Natural black shale samples from the Wismuth location (Germany): Comparison of the determination of the total sulfur contents by means of C-irmMS with the coulometric method, and of sulfur isotope ratios measured by means of C-irmMS with the off-line method after Kiba-reagent-preparation. 634S values are given vs. CDT. nd: not determined, na: not applicable. All C-irmMS and coulometric measurements were run in duplicate. Disseminated total sulfur consists mainly of pyrite sulfur. In the SA samples, sulfidic sulfur was removed by high temperature oxidation (1 h at 1000~ and the remaining sulfur is believed to mainly consist of barite sulfur. Total organic (TOC) and inorganic (TIC) carbon contents were measured by coulometry. Sample
6003-428-S 6003-430-S 6003-433-S 6003-438-S 6175-218-S 6175-249-S 6175-253-S 6175-383-S 6529-952-S 7512-590-S 7515-428-S 6003-430-SA 6003-433-SA 6003-438-SA #1-1 #1-2 #1-3 #1-4 #1-5 #2-1 #2-2 1 in
wt.%;
2
S1
S1
6348 2
Coul.3
MS 4
MS 4
6348 2 off-line
2.1 1.2 3.4 4.5 11.4 3.8 2.8 4.6 4.9 4.0 0.5 nd nd nd nd nd nd nd nd nd nd
2.2 1.2 2.9 3.8 12.5 3.4 2.7 4.4 4.9 3.5 0.5 0.2 0.1 0.1 nd nd nd nd nd nd nd
-18.6 -3.5 -12.2 -7.7 +6.1 -16.4 -9.0 -14.3 -7.7 -6.5 -36.2 -2.1 -3.4 +4.6 -1.7 +0.7 +1.4 +2.1 -0.4 +13.1 +10.8
-19.3 -3.4 -11.8 -8.1 +6.2 -16.9 -8.3 -15.2 -6.8 -5.9 -38.3 -3.2 -3.1 +4.3 nd nd nd nd nd nd nd
in %o; 3 Coul. = Coulometry;
4
TOC (wt.%)
TIC Comment (wt.%)
4.9 3.2 nd nd nd 5.5 12.5 nd 5.7 nd nd na na na na na na na na na na
1.3 7.0 nd nd nd 1.1 0.6 nd 1.0 nd nd na na na na na na na na na na
disseminated total sulfur disseminated total sulfur disseminated total sulfur disseminated total sulfur disseminated total sulfur disseminated total sulfur disseminated total sulfur disseminated total sulfur disseminated total sulfur disseminated total sulfur disseminated total sulfur disseminated barite sulfur disseminated barite sulfur disseminated barite sulfur idiomorphic pyrite crystal idiomorphic pyrite crystal idiomorphic pyrite crystal idiomorphic pyrite crystal idiomorphic pyrite crystal massive band of pyrite massive band of pyrite
MS = C-irmMS.
For off-line preparation, black shale samples were prepared by the Kiba reagent method (Kiba et al., 1955; Sasaki et al., 1979) and the isotopic composition of SO2 was measured on a Finnigan MAT 251 mass spectrometer at the Geochemical Institute of GSttingen University. Replicates agreed within about + 0.5%o. Sulfur measurements were carried out by coulometry (Heinrichs & Herrmann, 1990) using a Str6hlein coulomat with a reproducibility of about 10% (Lange & Brumsack, 1977). 27.3 Results and discussion
In the present investigation we demonstrate the applicability of the elemental analyser- isotope mass-spectrometer connection (C-irmMS) to measure directly the contents and sulfur isotope ratios of sulfur in black shales on small samples without any chemical pre-treatment of the samples. The black shale samples contain total sulfur contents between 0.1 and 12% (Tables 27.2 and 27.3).
601
Direct Measurement of the Content and Isotopic Composition of Sulfur ...
As a first part of the study the dependence of the combustion procedure from the chemical composition of pure sulfur-bearing compounds at a constant temperature of 1100~ (rising up to 1800~ during flash combustion) was investigated. This sample set was build of different anhydrous synthetic and natural sulfates, sulfides, disulfides, and elemental sulfur (Table 27.1). It was found that the signal for mass 64 generally varied linerarily with the sample amount combusted between equivalents of 10 and 200 gg sulfur. The recovery of sulfur as SO2 was always complete and independent from the chemical composition and the oxidation state of sulfur in the samples (Table 27.1). The average recovery was 101%, 99%, 102%, and 101% for sulfates, sulfides, disulfides and elemental sulfur, respectively. Most of the standard deviation observed in the recovery is believed to be due to the weighing procedure. In the second step, both coulometry and C-irmMS where applied to natural samples covering the range between 0.5 and 12 % total sulfur (Tables 27.2 and 27.3) and a good agreement between the two methods was found (Figure 27.1). Again, no influence of the chemical bonding environment of sulfur in the fossil sediments was 15
i
I
I
I
i
I
i
I
i
.
.F
~um
12 ,-" "'~
O
F" F"
9
."
9
F
9
0 6
3
,O
S (Coul.) - 0.92 S (C-irmMS) + 0.39 (r2 = 0.99; n = 14) I
0
3
I
I
6
I
I
i
9
I
12
i
15
S (wt%; C-irmMS) Figure 27.1 - Comparison of the determination of total sulfur of black shale samples by coulometry and C-irmMS (Tables 27.2 and 27.3). Dashed line indicates the 1: I relationship.
602
Chapter 27 - M.E. B6ttcher & B. Schnetger
observed and there was no influence of the absolute sulfur content on the reproducibility and precision of the analytical methods applied. The sulfur isotopic composition of the sulfur fraction in the natural samples was found to range between-36 and +13%o (Table 27.3), and the replicate measurements generally agreed within 0.2%o for samples with sulfur contents exceeding 0.1 wt.% and within 0.3%0 for the two samples with low sulfur contents (Table 27.3). No influence of the formation of carbon dioxide due to the common combustion of sulfur compounds with organic and inorganic carbon (up to 13%; Table 27.3) from the natural samples is observed, indicating a sufficient separation of SO2 from CO2 on the chromatographic column of the elemental analyzer and no later interference in the ion source of the mass spectrometer. Varying primary diagenetic conditions during sediment formation and subsequent overprints are reflected especially by the shift to heavier sulfur isotope values in the idiomorphic grains and massive bands of pyrite when compared to disseminated sulfur. Coexisting sulfate (barite?) sulfur was gener-
10
'
I
'
I
'
I
'
I
'
,"
..."
O
r or,-~
a
-10
O
9
,,O
9
-20
r
-30 m
634S (off-line) - 1.04 634S (C-irmMS) + 0.22 (r 2 - 0.996; n - 14)
,..
,~ a'mdmOn
-40 -40
i -30
I -20
i
I -10
,
I 0
i 10
~)34S (~ Figure 27.2 - Comparison of the determination of the sulfur isotopic composition of total sulfur of black shale samples by the off-line and C-irmMS method (Table 27.3). Dashed line indicates the 1" 1 relationship.
Direct Measurement of the Content and Isotopic Composition of Sulfur ...
603
ally isotopically heavier than sulfide sulfur as expected from known thermodynamic and genetic relationships (Ohmoto & Goldhaber, 1997). These results are in agreement with other geochemical signatures found in the different black shale samples (Schnetger, unpublished). The agreement between the off-line and the on-line sulfur isotope determinations can be regarded as very good (Figure 27.2) and independent from the sulfur concentrations in the samples used here (Table 27.3). This indicates that the CirmMS method which requires only small amounts of sample material is a precise and fast (with respect to sample preparation) analytical tool for sulfur isotope and content determination of sulfur in natural black shales with sulfur contents down to 0.1% (Table 27.3). Caution has to be taken in more recent sediment with high porosity, where the contribution of residual pore water sulfate may have to be taken into account. It should finally be noted, that for a detailed analysis of microbial reactions leading to an isotope fractionation between different sulfur bearing-species (acid volatile sulfides, pyrite, elemental sulfur, organic sulfur), especially in recent marine sediments or bacterial cultures, the chemical separation of the different sulfur phases may be necessary.
Acknowledgments
The authors wish to thank the German Science Foundation (DFG, Bonn) and Max Planck Society (Munich) for financial support, H. Avak (Finnigan MAT, Bremen) and A. Giesemann (FAL, Braunschweig) for stimulating discussions, and A. Giesemann and V. Reppke for the supply of sample material. J. Hoefs (University of G6ttingen) kindly allowed access to the Finnigan MAT 251. The constructive comments of reviewers M. O. Jedrysek and M. A. Tabatabai helped to improve the manuscript.
Handbook of Stable Isotope AnalyticalTechniques, Volume 1 P.A. de Groot (Editor) 9 2004 Elsevier B.V. All fights reserved.
CHAPTER 28 Summary of Methods for Determining the Stable Isotope Composition of Chlorine and Bromine in Natural Materials H. G. M. Eggenkamp Department of Geochemistry, Faculty of Earth Sciences, Utrecht University, P.O.Box 80021, 3508 TA, Utrecht, The Netherlands e-mail: [email protected]
Abstract
During the past 80 years many attempts have been made to measure natural variations in the stable isotope compositions of C1 and Br. These variations are quite small; variations in the C1 isotope compositions of natural samples were first measured in 1982, and natural variations in Br isotope compositions were first measured in 1997. In this chapter I describe several methods that have been applied over the past 80 years both to separate C1 and Br from natural materials and to measure their isotope ratios by mass spectrometry. It is hoped that this chapter increases interest in the stable isotope geochemistry of halogens. 28.1 Introduction
It was shown by Aston (1919) that C1 consists of two different isotopes with masses 35 and 37. In later years, the C1 isotope ratio was determined quite often (e.g. Curie 1921, Gleditsch & Sandahl 1922, Harkins & Stone 1925, von Kallman & Lasareff 1932, Nier & Hanson 1936, Graham et al. 1951, Shields et al. 1962), but because in nature the differences in C1 isotope ratios are small, no measurable variations were found. Variations in C1 isotope ratios were found in chemical experiments (e.g. Bartholomew et al. 1954, Klemm & Lund6n 1955, Lund6n & Herzog 1956, Herzog & Klemm 1958, Hill & Fry 1958, Howald 1960) and it was found that the diffusion coefficient of 35C1 w a s about 1.0012 to 1.0022 times that of 37C1 (Madorsky & Strauss 1948, Konstantinov & Bakulin 1965). After the development of a new mass spectrometer with double ion collectors (Nier et al. 1947, Nier 1947, 1955, McKinney et al. 1950) it was possible to measure the C1 isotope ratio variation with a precision of +1%o. Hoering & Parker (1961) measured 637C1 values of 81 samples. They found no significant variations from the standard, which was NBS C1 isotope reference standard NBS 105. Two samples of formation water had relatively large (although non significant) deviations from the standard (-0.7 and -0.8%o), but they were not considered to be significantly different relative to the analytical precision. Hoering & Parker (1961) also measured 3 samples of Chilean perchlorate. Although Urey (1947) had predicted that, if hydrogen chloride and perchlorate are in equilibrium the fractionation could be as much as
Methods for Determiningthe StableIsotopeCompositionof Chlorine and Bromine ...
605
92%o, Hoering & Parker (1961) did not find any difference in C1 isotope ratio between their perchlorate and chloride samples and concluded that the perchlorate samples were not formed in equilibrium with the chloride in these deposits. Recently Ader et al. (2001) reported new methods for determining C1 isotope compositions in chlorates and perchlorates, but they also found only small variations, hardly significantly different from the ocean water isotope composition. Morton & Catanzaro (1964) measured the C1 isotope composition of apatites and found no variations larger than their analytical precision of +1%o. Since the early 1980s it has been possible to measure C1 isotope ratio variations sufficient to resolve natural variations. As a result of foresight of Austin Long and his research group in Tucson, Arizona that it was realised that C1 isotopes could probably be measured at these precisions. Kaufmann (1984) published the first thesis in which measurable variation of C1 isotopes in natural materials was shown. The precision of these analyses was better than +0.24%o and improved in later years. In this period, the first results were presented at several congresses (Kaufmann et al. 1983, 1984a, Kaufmann & Long 1984, Campbell & Kaufmann 1984) and published (Kaufmann et al. 1984b). In the years that followed several additional studies on the geochemistry of the stable isotopes of C1 were presented by the Arizona group (Kaufmann et al. 1987, 1988, 1992, 1993 Kaufmann & Arn6rsson 1986, Kaufmann 1989, Desaulniers et al. 1986, Eastoe et al. 1989, Eastoe & Guilbert 1992, Gifford et al. 1985). It was also the group from Arizona that proposed to use the C1 isotope composition of ocean water as an international standard reference for chlorine isotope measurements. This could be justified as they showed that the C1 isotope composition of several ocean water samples did not vary outside the analytical error (Kaufmann, 1984). Considerable progress in C1 isotope geochemistry occurred after 1990. Long et al. (1993) published in detail the method which was developed in Arizona. The following year Eggenkamp (1994) published his thesis, discussing C1 isotope data from a large range of environments. In the same period it was shown that C1 isotope compositions can be measured by positive ion thermal ionisation mass spectrometry (Xiao et al. 1992; Magenheim et al. 1994). This was an important development, following unsuccesful studies attempting to measure C1 isotope compositions using negative ion thermal ionisation mass spectrometry (Vengosh et al. 1989 Gaudette 1990). Besides methods to measure stable C1 isotope compositions in inorganic samples, methods for measuring C1 isotopes in organic compounds were developed and published (Tanaka & Rye 1991; Van Warmerdam et al. 1995; Holt et al. 1997; Jendrezejewski et al. 1997). Only very few studies are known in which the (geo)chemistry of Br isotope variations is described. Although it has long been known already that Br has two stable isotopes (79Br and 81Br, Aston, 1920) no large natural isotope variations were expected due to the small relative mass difference between these two isotopes. Early studies by Cameron & Lippert (1955) showed no Br isotope variations beyond their analytical precision. In later years, however, fractionation due to diffusion was shown in molten
Chapter 28 - H.G.M. Eggenkamp
606
lead bromide (Cameron et al., 1956) and zinc bromide (Lund6n & Lodding, 1960). Willey & Taylor (1978) showed that is was possible to measure the Br isotope composition with a method comparable to the method for C1 isotopes, using bromomethane. Xiao et al. (1993) described a method to measure the Br isotope composition using positive ion thermal ionisation mass spectrometry, which is analogous to the method for measuring C1 isotopes. Eggenkamp & Coleman (2000) described a method to measure Br isotopes in natural samples, which includes a method to separate Br and C1 from samples which contain only (very) small Br concentrations. 28.2 Notation and standards
Chlorine has two stable isotopes, 35C1 and 37C1, with natural abundances of 75.771 and 24.229% respectively (Rosman & Taylor, 1998; abundances still from Shields et al., 1962!). No international standard for C1 isotope ratios has been defined yet officially, but in all recent studies Standard Mean Ocean Chloride (SMOC) is used. Kaufmann (1984) showed that no variations in the C1 isotope composition of ocean water from different locations and depths are found. All laboratories in the world use seawater as their standard. Chlorine isotope data are reported as 637C1 which is defined as: (37C1]
_ (37C1]
1~37C1 _ ~35C1)sample /35C1)standard x 1000 37C1] 35C1)
[28.1]
standard
Bromine also has two stable isotopes, 79Br and 81Br, with abundances of 50.686 and 49.314%, respectively (Rosman & Taylor, 1998; abundances still from Catanzaro et al., 1964!). No formal international standard for Br isotope ratios has been defined, neither have tests been done to check for variations of the Br isotope composition in ocean water. However, as the residence time of Br in the oceans is even larger than that of chlorine, it is assumed that oceanic Br does not show variations, and thus can be used as an isotopic standard reference material. This standard can be called Standard Mean Ocean Bromide (SMOB, Eggenkamp & Coleman, 2000). Bromine isotope data are reported as 681Br which is defined as:
(81BF/ _(81BF/
a81Br _
79Br)sample
79Br)standard
81Br/ 79Br)
standard
x 1000
[28.2]
Methods for Determining the Stable Isotope Composition of Chlorine and Bromine ...
607
28.3 Basic techniques for measuring chlorine isotopes 28.3.1 Chlorine isotope measurements by gas-source ratio mass spectrometry Many studies report attempt to measure C1 isotope compositions in gas source mass spectrometers. Many different gases have been applied in these studies. In this chapter attempts using hydrogen chloride, C12 gas and chloro methane will be described. In early studies, several other gasses were used, including carbonyl dichloride (COC12, phosgene; Aston, 1942), boron trichloride (BC13, Osberghaus, 1950), phosphorus trichloride (PC13, Kush et al., 1937), arsenic trichloride (AsC13, Kush et al., 1937), and antimony trichloride (SbC13, Kush et al., 1937). These studies will not be discussed here. 28.3.1.1 Hydrogen chloride used as m a s s s p e c t r o m e t e r g a s Hydrogen chloride was used in several C1 isotope studies (Nier & Hanson, 1936; Madorsky & Strauss, 1947; Johnston & Arnold, 1953; Hoering & Parker, 1961). Of these, Hoering & Parker (1961) presented the first major study of C1 isotopes. They measured the C1 isotope composition of 81 samples in all types of geological environments. They were able to measure the C1 isotope ratio of HC1 with a precision of 0.8%0. The reason for using HC1 gas was that it can be prepared quantitatively, and because it has a simple cracking pattern. A large disadvantage of HC1 is that it sticks to the walls of the vacuum system, and as a result it has a large memory effect. Hydrogen chloride has a relatively simple mass spectrum in a mass spectrometer, with four ion species formed (Table 28.1). Samples of chloride were precipitated as silver chloride (AgC1). For rock and mineral samples this was produced according to the method described by Kuroda & Sandell (1953). Samples were dissolved in ammonium hydroxide, to which magnesium metal was added in excess. The magnesium displaces the silver, and an ammonium chloride solution was formed from which solids were removed by filtration. The remaining solution was evaporated to dryness (in a vacuum oven). The solution from which the ammonium chloride was precipitated is basic, and care must be taken that it would not react with CO2 from the atmosphere, as this can not be separated from the hydrogen chloride, and as such it interferes during isotope analysis. The ammonium chloride residue was converted to HC1 by reaction with sulphuric acid under vacuum. The gaseous HC1 could then be trapped onto frozen phosphorus pentoxide to remove traces of water, and the dried HC1 was frozen into a sample container for introduction into the mass spectrometer. 28.3.1.2 Chlorine g a s used as m a s s spectromTable 28.1 - Cracking pattern for HC1 (Hoereter g a s Chlorine gas has been used only rarely as ing & Parker, 1961). mass spectrometer gas. Owing to its simple Species m/ z Relativeintensity cracking pattern, however, it could be useful. Unfortunately, it also produces large mem35C1+ 35 17.0 ory effects. H35C1+ 36 100 37C1+ 37 5.4 Chlorine gas has a relatively simple cracking H37C1 + 38 32.5 pattern, which should be approximately as is
608
Chapter 28 - H.G.M. Eggenkamp
shown in Table 28.2.
Table 2 8 . 2 - Estimated cracking patters for chlorine gas, based upon the abundances for the isotopes 35C1 and 37C1, and assuming that the total intensity for monoatomic species is 10% of the total intensity for diatomic species.
Bartholomew et al. (1954) used C12 as the mass spectrometer gas in their studies on the isotope effect in reactions of tert-butyl chloride (2-chloro-2-methyl-propane). Hydrogen Species m/z Relative intensity chloride was produced from silver chloride 35C1+ 35 13 in a way comparable to that of Hoering & 37C1+ 37 4 Parker (1961). However, the ammonium 35C12+ 70 100 chloride was heated with concentrated sul72 64 phuric acid, and the product HC1 was 35C137C1+ 37C12+ 74 10 trapped in a bubbler (by means of a stream of nitrogen gas) containing a little cold water. The aqueous hydrogen chloride was then oxidised to C12 gas by persulphate oxidation (Brown et al. 1953) in a stream of helium. The C12 gas was frozen in a cold trap, and the helium pumped away. The C12 then could be transferred to the mass spectrometer. 28.3.1.3 Chloro-methane used as mass spectrometer gas
The most commonly used gas for isotope ratio mass spectrometry of C1 is now chloro-methane (methyl chloride). Several methods have been developed to produce quantitative yields of this gas, which is a lot less reactive than HC1 or C12, and it gives no memory effects in inlet and vacuum systems of mass spectrometers. This gas is generally measured at m / z 50 and 52 at which positions the main peaks are found. It's cracking pattern is relativity complex with many peaks as both chlorine and carbon have a significant minor abundant isotope (Table 28.3). Chlororomethane can be produced from several chloride compounds. In the literature on chlorine stable isotopes three of these have been proposed: chloromethane produced from ammonium chloride Table 28.3 - Cracking pattern from chloromethane (Owen & Schaeffer, 1954; Herzog & D6rnenburg, 1958), chloromethane (Taylor & Grimsrud, 1969). produced from silver chloride (LangSpecies m/z Relative vad, 1954; Hill & Fry, 1962; Taylor & intensity Grimsrud, 1969; Kaufmann, 1984; Long et al., 1993; Eggenkamp, 1994), 35C1+ 35 2.8 and chloromethane produced from H35CI + 36 1.2 copper chloride (Holt et al., 1997). 37C1+ 37 1.0 H37C1+ C35C1+ CH35C1+ CH235C1+C37C1+ CH335 C1+ CH37C1 + 13CH335Cl+CH237C1 + CH337C1+ 13CH337C1+
38 47 48 49 50 51 52 53
0.4 7.7 3.2 9.6 100 3.4 31.4 0.5
Chloromethane produced from ammonium chloride Chloromethane can be produced from ammonium chloride, and this technique was used in studies from the 1950s (Owen & Schaeffer, 1954; Herzog & D6rnenburg, 1958). Silver
Methods for Determining the Stable Isotope Composition of Chlorine and Bromine ...
609
chloride was precipitated from a sample in dissolved form, which was redissolved in ammonia. Silver then was removed as the sulphide. After drying pure ammonium chloride was present. This could be reacted into chloromethane following a procedure described by Blatt (in Owen & Schaeffer, 1954). 240 ml of concentrated sulphuric acid was diluted with 40 ml of distilled water, to which was added 350 ml of methanol. The temperature was kept below 70~ at all times. Approximately 2 ml of this solution was added to 50 mg of ammonium chloride under vacuum. The following reactions can occur: (CH3)2SO4 +2NH4C1 ~ 2CH3C1 + (NH4)2SO4 CH3OH ~ (CH3)20 (Catalysed by sulphuric acid) 2NH4C1 + H2SO4 --~ 2HC1 + (NH4)2SO4
[28.3] [28.4] [28.5]
Reaction [28.4] can be minimised by keeping the temperature of the mixture below 50~ A weighed sample of ammonium chloride was placed into a bulb and evacuated. Approximately 2 ml of sulphuric acid and methanol were added through a stopcock. The mixture was warmed with an infrared lamp until the reaction ceased. The reaction tube was chilled with dry ice and the chloromethane distilled into an evacuated sample bulb. A potassium hydroxide trap removed the hydrogen chloride formed during this reaction. A disadvantage of this technique is that the yield will not be higher than approxiately 35%. Several tests were done to determine errors caused by this low yield and it was found that data were reliable within an experimental error of 2%0. However, as it is now known that the large majority of 637C1 data are within 2%o of SMOC, an error this large is not acceptable for most natural C1 isotope measurements.
Chloromethane producedfrom silver chloride In most modern studies chloromethane is produced from geological samples via silver chloride. The method was originally described by Langvad (1954), and improved by Hill & Fry (1962) and later by Taylor & Grimsrud (1969). Ultimately a very effective technique was published by Kaufmann (1984). The method described here is taken from Eggenkamp (1994), which is based upon the method from Kaufmann (1984). Long et al. (1993) published a comparable version, also based upon the one described by Kaufmann (1984). Their method was developed to measure C1 isotope compositions by accelerator mass spectrometry for 36C1 measurements, and was designed to remove traces of sulphur from the sample. As such that technique is more complex than the one presented by Eggenkamp (1994). Eggenkamp (1994) used the following three steps to produce unfractionated chloromethane of sufficient purity for isotope measurement: 1) precipitation of silver chloride, 2) reaction of silver chloride with iodomethane, 3) separation by gas chromatography.
610
Chapter 28 - H.G.M. Eggenkamp
The procedure to prepare the silver chloride depends slightly on the amount of chloride in the solution. The method aims at precipitating silver chloride from solutions of fixed C1- amount, fixed ionic strength and fixed pH. Kohnen (1988) found that the best results are obtained when the amount of silver chloride formed is about 100 gmole (or 14.3 mg AgC1, corresponding to 3.5 mg C1-). This was confirmed by later studies. Therefore the amount of chloride solution needed is: 3000 p p m chloride
= ml necessary
[28.6]
However, if the amount of sample available is limited, with amounts down to 20 gmole reliable measurements can be made. It is recommended however that samples within a mass spectrometric run all contain comparable amounts of chloromethane. If the amount of solution is less than 10 ml, the following standard procedure is used" 4 ml of a 1 M KNO3 solution and 2 ml of a Na2HPO4-citric acid buffer solution are added to the chloride solution. The purpose of the KNO3 solution is to reach a high ionic strength. Taylor & Grimsrud (1969) found that using a less than 0.4 M KNO3 solution leads to very low chloromethane yields; for instance a 0.2 M KNO3 solution gives only 45% yield. The reason for this effect probably is that smaller crystals form at a high ionic strength. These small crystals can react completely whereas larger crystals form a coating of silver iodide that prevents the inner part of the crystals from reacting. Incomplete reaction inevitably leads to fractionation; Taylor & Grimsrud (1969) found a fractionation of +0.43%o due to this effect. The Na2HPO4-citric acid buffer solution is used to buffer pH at 2.2. This is necessary to remove small amounts of sulphide which otherwise precipitate as Ag2S (Kaufmann 1984), and also to prevent precipitation of other silver salts such as phosphate and carbonate (Vogel 1951). We used a buffer solution after McIlvaine (1921) which contains 0.71 gr (0.004 mole) Na2HPO4.2H20 and 20.6 gr. (0.098 mole) HOC(CH2CO2H)2CO2H. H20 (citric acid) per litre. After adding the KNO3 solution and Na2HPO4-citric acid buffer, the mixture is placed on a boiling ring and heated to about 80~ Then I ml of a 0.2 M AgNO3 solution is added and AgC1 starts precipitating instantaneously. The solution is not stirred because the newly formed AgC1 will coagulate and it is difficult to remove from the stirrer. The suspension then is filtered over a Whatman| glass fibre filter, type GF/F with a retention of 0.7gm and a standardised filter speed of 6 ml/sec. During filtration the suspension is rinsed with a dilute nitric acid solution (1 ml concentrated HNO3 in 500 ml water). When the silver chloride precipitate is rinsed with pure water, it occasionally will become colloidal and pass through the filter. Therefore the rinsing solu-
Methods for Determining the Stable Isotope Composition of Chlorine and Bromine ...
611
tion must contain an electrolyte; nitric acid is chosen because it has no reaction with the precipitate and leaves no residue upon drying (Vogel 1951). After filtration, the filter with the precipitate is dried at 80~ overnight. Care must be taken to protect the silver chloride from exposure to light. Silver chloride decomposes under the influence of light according to the reaction: 2AgC1 ---, 2Ag + C12
[28.7]
Therefore the filter with silver chloride is covered with aluminium foil. The aluminium foil must not be in contact with the silver chloride otherwise the aluminium will reduce the silver chloride: 3AgC1 + A1 ~ A1C13 + 3Ag
[28.81
which may cause isotope fractionation. The filter is weighed before precipitation and after drying, so that the amount of silver chloride is known. Samples with a chloride content below 300 ppm are treated in a slightly different way because more sample solution is needed. For these samples, KNO3 and the pH buffer are added as dry chemicals, otherwise the amount of solution would become too large. Per 100 ml of sample solution 6.00 g (0.06 mole) KNO3, 2.06 g (0.0098 mole) citric acid and 0.07 g (0.0004 mole) Na2HPO4.2H20 are added. The reaction of AgC1 to CH3C1 takes place in evacuated Pyrex tubes sealed at both ends; the tubes are 8-10 cm long and have an inner diameter of 8 m m and an outer diameter of 12 mm. The filter with AgC1 is loaded in a tube sealed at one end, a capillary drawn at the other end, and the tube is evacuated to a pressure less than 2 x 10-1 mbar. The tube is then filled with nitrogen gas and sealed with a rubber stopper to prevent air coming in. In a fume-hood, 200 ~1 (3.21 mmole) of iodomethane (CH3I) is added. Back on the vacuum line the CH3I is frozen on the AgC1 with liquid nitrogen and the tube is p u m p e d to less than I x 10-1 mbar. The tube is then sealed at the site of the capillary. The sealed tube is placed in an oven at a temperature of 70 to 80~ for 48 hours so that the following reaction takes place: AgC1 + CH3I ~ AgI + CH3C1
[28.9]
This is an equilibrium reaction so the CH3I must be added well in excess to get good CH3C1 yields. If the reaction temperature is too high the CH3I will partly decompose: 2CH3I --+ C2H6 + I2
[28.10]
see Eastoe et al. (1989). When CH3I decomposes, the colourless liquid will become yellow to brown. Samples that have been overheated can give much less accurate 637C1 values. Decomposition of CH3I can also been detected in a background scan that is routinely made after the isotope measurement. Overheated samples then show
612
Chapter 28 - H.G.M. Eggenkamp
increased background peaks at m / z 29, 45 and 46. CH3C1 and CH3I are separated by gas chromatography on two 75-cm long, 1/4" OD SS (stainless steel) columns, filled with Porapak-Q 80-100 mesh. Because the columns are easily overloaded with the large amount of excess CH3I, and the CH3C1 must be very pure, the gases are separated in two successive runs. The carrier gas is helium, at a pressure of 3 atm and a gas flow of about 100 ml min-1 (this rate should be adjusted such that a good separation between the three gases is obtained). The column temperature is 140~ The CH3C1 peak is detected by a thermo conductivity detector using a Carle 100 Micro Detector Control. A schematic drawing of the set-up is shown in Figure 28.1. The procedure for processing a series of samples is as follows: the gas chromatograph is first back flushed to remove the excess CH3I from the previous separation, to minimise contamination of the column and detector (3-way valve 3 set to position A, and 3-way valve 2 set to position B, while the openclose valve is closed). The borosilicate glass reaction tube is scratched by a glass cutting knife, and placed in the tube cracker. The tube cracker is evacuated (3-way valve 1 to position A) and liquid N2 is placed around the first coldtrap. 3-way valves 1 and 2 are closed and the reaction tube is broken. At the same time 3-way valve 3 is turned to position B and the open-close valve is opened. After 30 seconds 3-way valves 1 and 2 are turned to position B and the open-close valve is closed. After 3 minutes the liquid N2 around trap 1 is replaced by warm water and liquid N2 is now placed around trap 2. The recorder is started and a CH3C1 peak will be seen after about 2 minutes. As soon as the recorder signal has returned to the base line, and before the CH3I peak would be detected (after about 6 more minutes), 3-way valves 1 and 2 are turned to position A, the open-close valve is Figure 28.1 - Schematic drawing of the gas chromatograph opened and the liquid N2 around (see description in text).
Methods for Determining the Stable Isotope Composition of Chlorine and Bromine ...
613
the second cold-trap is replaced by warm water. Liquid N2 is now placed around the sample vessel. Valve 5 is closed and just before the expected arrival of CH3C1, 3-way valve 4 is turned to position B. After the pressure in the sample vessel has increased to approximately 3 psi above atmospheric level, valve 5 is turned to position A so that the over-pressure of helium is released. The CH3C1 is trapped in the sample vessel. When all the CH3C1 has been trapped, as indicated by GC, valve 5 is closed and the other valves must be returned to the starting configuration. Helium is pumped out of the sample vessel that should now contain pure CH3C1. The column is then backflushed, so that remaining CH3I is removed. The broken reaction tube can be replaced by the next one to be extracted. CH3CI producedfrom CuCI Holt et al. (1997) produced copper(I) chloride in the process of reducing chlorinated organic material (as summarised below). This copper(I) chloride is reacted in vacuum with iodomethane at 300~ for two hours. This procedure leads to ~100% yields for chloromethane. Holt et al (1997) separated chloromethane from iodomethane cryogenically.
Cryogenic separation of chloromethanefrom iodomethane Holt et al. (1997) developed a method to separate the mixture of chloromethane and iodomethane cryogenically. The mixture of chloromethane and iodomethane, that sits in a scratched tube, is put into a vacuum line that is evacuated. The location of the tube is cooled with dry ice-acetone slush to -79~ The tube where the iodomethane is frozen too is cooled with n-pentane-liquid nitrogen slush (-130~ and the tube where the chloromethane is to be frozen to is cooled with liquid nitrogen (-196~ The tube containing chloromethane and iodomethane is cracked open. As the mixture is cooled with dry ice-acetone slush the chloromethane and iodomethane only evaporate slowly, thus enhancing the cryogenic separation. Ten minutes after breaking the tube the dry ice-acetone slush is removed, and after 10 minutes more the chloromethane is moved to a third cold trap, by moving the liquid nitrogen to this trap. The n-pentaneliquid nitrogen slush is moved to the former chloromethane trap to trap iodomethane that has slipped through the first trap. This whole transfer may take up to 15 minutes. After separation the chloromethane can be frozen into a sample vessel for measurement in the mass spectrometer, while the excess iodomethane can be frozen into a waste tube for subsequent disposal. This is an advantage above the method to separate the two gasses by gas chromatography, where the waste iodomethane is vented to the air.
Measuring chloro methane samples on a stable isotope ratio mass spectrometer Chloromethane can be measured on all normal isotope ratio mass spectrometers. 637C1 is determined from the beams of mass 52 (CH337C1 +) in collector 3 and mass 50 (CH335C1+) in collector 1. The isotope ratio of C1 is much higher than for the light elements for which these mass spectrometers were built. Thus, beam 52 will be off scale at small working pressures. Working with very low pressures gives isotope fractionation in the inlet system. For this reason the ion source should be made less sensitive. This is done by reducing the trap current to 100~A. In this case the minor beam is
614
Chapter 28 - H.G.M. Eggenkamp
brought to a value smaller than 10-10A while still maintaining sufficient gas pressure in the inlet system. At these conditions, the results are highly reproducible.
28.3.2 Cl isotopes measurements by P-TIMS Apart from methods to determine C1 isotope ratios using gas-source isotope ratio mass spectrometry they are also being measured successfully by positive ion thermal ionisation mass spectrometry. Xiao & Zhang (1992) first described a method to measure C1 isotope ratios using this technique, followed by Magenheim et al. (1994) who also described several techniques to extract C1 from a variety of geological materials for isotope measurements. Chlorine is used in the form of HC1. Xiao & Zhang (1992) produced this from seawater by cation exchange chromatography, where resin in the Ba2+ form was used. The hydrogen chloride solution was then diluted to a concentration of 3 mg C1/ml. Tantalum filaments were treated with with 3 ~1 graphite slurry (100 ~g graphite, 80 vol% ethanol/20 vol% water), so that it was completely coated. This is almost dried completely, and the sample, I ml of the HC1 solution, neutralised with cesium carbonate is added. This is then dried for two minutes using a current of 1.1 A passing through the filament. After loading the treated filaments into the mass spectrometer isotope analysis begins as the pressure in the instrument is between 2 x 10-7 and 3 x 10-7 Torr (1 Torr = 1,33 mbar). Current on the filament is increased to 1.1 A in ten minutes. The Cs2C1+ ion current is monitored and used to focus the instrument. Its intensity is adjusted to 5 x 10 -12 to 8 x 10 -12 A by adjusting the filament current, which is typically 1.15 to 1.25 A. The data are collected by switching between the masses 301 (133Cs235C1 +) and 303 (133Cs237C1+). The baseline is determined at m / z 300.5. The data are acquired for 1.5 hours in high precision runs. Magenheim et al. (1994) applied a comparable method. They only removed ions such as fluoride and sulphate by ion exchange from the HC1 solution (to diminish interferences in the mass spectrometer), and the CsC1 was produced from the hydrogen chloride solution by ion exchange. This has the advantage that all acidity of the solution is removed.
28.3.3 Comparison of TIMS with IRMS Rosenbaum et al. (2000) compared positive ion thermal ionisation mass spectrometry with gas isotope ratio mass spectrometry and found that samples measured by both techniques gave comparable data. This indicated that data obtained and published from the two methods can be compared to each other.
28.3.4 Cl isotopes measurements by fast atom bombardment (FAB) mass spectrometry Westaway et al. (1998) proposed to measure the C1 isotope composition using a FAB-mass spectrometer. In this type of instrument silver chloride is mounted on a silver plate, heated and bombarded by Xe atoms. Negative C1 ions are then formed, and measured in Faraday cups at masses 35 and 37. The advantage of this method is that conversion of silver chloride to methyl chloride is not necessary. The analytical error
Methods for Determiningthe StableIsotopeCompositionof Chlorine and Bromine ...
615
of this method is 3 to 5 times larger than IRMS of methyl chloride, but within this error results are in good agreement.
28.4 Preparation techniques for different sample types All techniques described above require to get chloride in aqueous solution. Thus, preparation techniques are necessary to convert non-aqueous C1 (such as in organic compounds or rock samples) to aqueous chloride.
28.4.1 Preparation for organic samples Four different techniques have been proposed in the literature to prepare organochlorine compounds for chlorine stable isotope measurements.
28.4.1.1 LiCl-technique In the very first paper on the potential measurement of the C1 stable isotope composition of organochlorine compounds Tanaka & Rye (1991) proposed to react the organic compound with lithium metal. Samples were prepared by reacting the sample with lithium metal in a sealed quartz tube. The lithium chloride was dissolved in water and the chloride reacted to chloromethane according to the method described by Taylor & Grimsrud (1969). Replicate analyses of samples yield a precision of 0.15%o.
28.4.1.2 Parr Bomb technique Van Warmerdam et al. (1995) oxidised the organochlorine compounds in a socalled "Parr Bomb", according to the standard ASTM method D808-91. Five ml of a 5% CaCO3 solution is injected to the bottom of the bomb to absorb combustion products. The sample capsule is then sealed with cellophane tape to prevent loss of sample by volatilisation, after which sample (max. 50 ~1) is introduced by injecting through the tape using a gas tight syringe. The fuse wire loop is then brought in contact with the centre of the tape cover. The bomb is sealed and oxygen is added to a pressure of 20 atm, placed in an ice-water bath and the fuse is ignited. After the combustion, the bomb is depressurised and combustion products are removed using ultra pure distilled water. The yield of combustion products is only 65 to 75%, but no practical effects are reported on the C1 isotope ratio. The combustion products are then converted to chloromethane for isotope measurement by the method described in Long et al. (1993). The analytical reproducibility of this method is fairly low, with a standard deviation of +0.29%o based on 21 analyses of a 1,1,1,-trichloroethane sample. Standard deviations of 10 other chlorinated solvent samples range between +0.10 and +0.68%o.
28.4.1.3 Break seal technique Holt et al. (1997) proposed a method to produce both CO2 and chloromethane from the same chlorinated volatile organic sample. A borosilicate glass tube with a length of 20 cm is prepared with 4 contractions 2.5 cm apart starting from the open end. These tubes contain I g CuO wire, which is preheated for one hour at 550~ This tube is attached to a vacuum line and the sample, containing 10 to 70 ~mol C1, is frozen onto it. The tube is then sealed at the first constriction. Then the tube is heated for 2 hours at 550~ After this period the capsule is put with one end into a Watlow fur-
616
Chapter 28 - H.G.M. Eggenkamp
nace, and with the other end into a liquid nitrogen container. When heating the capsule to 750~ for 45 minutes, the formed CuC1 will evaporate and precipitate in the liquid nitrogen part of the tube. Hereafter the remaining CuO is collected at the end of the capsule where the copper chloride is not precipitated, and the capsule is scratched and put into a tube cracker. After evacuation the capsule is broken and the carbon dioxide is transferred to a sample vessel. Water remaining in the capsule is coldtrapped using a dry ice-acetone slush. The CO2 is subsequently stored and can later be measured for the C isotope ratio. Iodomethane is then frozen onto the remaining part of the capsule containing the copper chloride, evacuated and reacted to form chloromethane as described above.
28.4.1.4 Sealed tube technique Jendrezejewski et al. (1997) proposed to measure both the C and C1 isotope ratios from organochlorine compound using a sealed tube method. Copper oxide is added to a quartz tube, to which (via a septum) I to 3 ~1 of the sample is frozen. The tube is then evacuated and sealed. The capsule formed is heated in a furnace between 720 and 820~ for at least one hour. After this period the capsule is allowed to cool slowly. The capsule is then scratched, put in a tube cracker and after evacuation it is broken. The CO2 that was formed is then transferred cryogenically into a sample vessel. All solid residues (glass plus copper oxide and copper chloride) are recovered and transferred to a small glass beaker, with 15 ml high quality water. This is allowed to dissolve the copper chloride for 15 hours. After this period the solution is separated from the solids and AgC1 is precipitated and reacted with CH3I as described above to produce chloromethane for C1 isotope analysis.
28.4.2 Preparation for rock samples
A few techniques, summarized below, have been described in the literature to obtain a solution that contains enough chloride for isotopic analysis.
28.4.2.1 NaOH fusion technique
Eggenkamp (1994) described a method to extract chloride from silicate rocks based upon a technique which was formerly used by Behne (1953). X grams of powdered rock in which X is an amount of rock containing enough chloride to perform one or more ~37C1 measurements, is heated together with 10X grams of NaOH pellets for about 30 minutes in a nickel crucible. The rock dissolves in the molten NaOH and SiO bonds are partly destroyed. The temperature must not be too high since NaOH will evaporate at high temperatures (e.g vapour pressure is I mmHg (1.33 mbar) at 739~ and 10 mmHg (13.3 mbar) at 897~ (Stull, 1947)). After cooling, the sample is dissolved in 35X ml H20. This is achieved by putting the nickel crucible in a beaker with water on a magnetic stirrer. After two hours of stirring the content of the crucible is dissolved or suspended in the water. This suspension must be exposed to air for some time (overnight) to oxidize Fe2+ and other ions. Because the solution has an extremely high pH, which would cause precipitation of Ag20 after addition of Ag + ions, the solution must be acidified. This is
Methods for Determiningthe Stable Isotope Compositionof Chlorine and Bromine ...
617
done by adding 17.5X ml HNO3 65% to produce a colloidal "solution" of silica gel according to" SiO44- + 4HNO3 --+ H2SiO4 ~ + 4NO3-
[28.11]
This colloidal solution can not be filtered and, therefore, 3X ml HF 40% is added. The silica gel reacts with the HF to form a combined silica oxyfluoride. H4SiO4 + mHF --+ H4-mSiO4-mFm + mH20 m_<4
[28.12]
This reaction does not go to completion. The oxyfluoride precipitates slowly (overnight) to the bottom of the glass container. Then 10.9X ml of a 2.5 tool kg-1 Mg(NO3)2.6H20 solution is added (64.1 gram Mg(NO3)2~ and 35.9 gram H20, density 1.25 gr ml-1). Excess toxic HF will react to poorly soluble MgF2. The resulting suspension is filtered. The residue contains the reacted silicagel, the MgF2, the insoluble hydroxides and rock components that were not dissolved. Chloride is concentrated in the filtrate. The filtrate has a low pH and a high ionic strength, so that the silver chloride can be precipitated directly from this solution without adding a buffer or KNO3.
28.4.2.2 Ion exchange technique Musashi et al. (1998) developed a method in which the chloride from rocks was extracted using ion exchange resins, to get a solution with less possibly interfering ions. Musashi et al. (1998) decompose rock samples in concentrated hydrogen fluoride solution (1 g sample plus 1 ml dionised water and 4 ml 17M HF). Samples were decomposed in a PTFE centrifugation tube (40 ml). After decomposition of the sample a calcium hydroxide slurry is added drop by drop to precipitate fluoride from the solution. After settling, the solution was centrifuged. The solution was then filtered over a fibre glass filter (Whatman| GF/F). This was done several times. The final fluoride was removed by passing the solution through an acryl column, packed with strongly acidic cation exchanger with a flow rate of I ml/min. The purified chloride bearing solution then was passed through a quartz column with a strongly basic anion exchange resin to trap chloride on the column. The chloride was finally released from the column by rinsing with 0.5 M nitric acid, and the chloride bearing effluent was sampled in 4 ml sample bottles. The C1 isotope composition could then be determined as described above (Long et al., 1993; Eggenkamp, 1994).
28.4.2.3 Pyrohydrolysis method To obtain chloride solutions from rocks, Magenheim et al. (1994) applied a pyrohydrolysis technique (e.g. Dreibus et al., 1979). The pyrohydrolyses was conducted by inductively heating a sample (1300-1400) in a Pt boat. The induction furnace is com-
618
Chapter 28 - H.G.M. Eggenkamp
posed of a vertically oriented fused silica combustion tube with PTFE connectors. The carrier gas (water) and the analyte were condensed in a water-cooled fused silica condensor and collected in a polyethylene bottle. Prior to each sample extraction, the Pt boat and apparatus were cleaned by heating the boat to the running temperature with a flow rate of 3.0 g water per minute for 2 hours and then rinsed with deionised water. Extraction was performed for 50 minutes at a flow rate of 0.3 g water per minute. The RF field was then turned off and the system was flushed with a water flow of 2.5 g per minute for 20 minutes. Finally, the interior of the combustion tube was rinsed with water for quantitative recovery. The total amount of solution recovered was 90-95 gram. Magenheim et al. (1994) then determined the isotope composition using thermal ionisation mass spectrometry.
28.4.3 Techniques for other sample types Eggenkamp (1994) described preparation techniques for a few other types of samples, including evaporites, samples containing high sulphide concentrations, samples with a high pH and carbonatites.
28.4.3.1 Evaporites Evaporites were prepared by simple dissolution of a known amount of sample in water (Eggenkamp et al. 1995).
28.4.3.2 High sulphide samples Samples which contain sulphides will precipitate silver sulphide instead of silver chloride when treated with silver nitrate, which has the disadvantage that dimethyl sulphide can be formed in the reaction with iodomethane, which will interfere with the isotopic measurement. If the sulphide content is low enough it will evaporate from the sample solution after adding the citric acid/phosphate buffer as hydrogen sulphide (Kaufmann, 1984:). However, this is not sufficient for samples with high sulphide concentrations. To remove the sulphide the sample is heated with a few ml of 30% hydrogen peroxide to a temperature of approximately 80~ for approximately 1 hour. The sample can then be treated as a normal sample.
28.4.3.3 High alkaline samples Samples with a high pH will precipitate silver oxide when treated with silver nitrate. This also will cause problems during measurement of the gaseous sample in the mass spectrometer (formation of dimethyl ether?). If the pH is only moderately high, it will be lowered by adding the citric acid/phosphate buffer. However, this is not sufficient for very high pH samples. The pH of these samples can easily be lowered by adding one or two ml concentrated nitric acid to the sample. It can further be treated as a normal sample.
28.4.3.4 Carbonatites Most carbonatites, with the exception of the natro-carbonatites, contain little chloride. Eggenkamp & Koster van Groos (1997) extracted chloride from the carbonate fraction of carbonatites by the following method. Five gram of carbonatite was mixed with 100 ml distilled water and stirred at a temperature of 50~ To this mixture, con-
Methods for Determiningthe Stable IsotopeCompositionof Chlorine and Bromine ...
619
centrated HNO3 was added (drop by drop), until the solution stopped effervescing (pH is then as low as 0.5). Next, the solution was filtered. Only chloride present in carbonates and apatites is liberated with this method. The sample can further be treated as a dilute water sample. In the case of natro-carbonatites, which contain much more chloride (Keller & Krafft, 1990) a much smaller amount of sample is required (Eggenkamp & Koster van Groos, 1997). 28.4.3.5 Other non-silicate and non-carbonate minerals Other minerals such as phosphates and hydroxide chlorides can generally be dissolved in dilute (1:5) nitric acid (Hintze, 1915). Minerals of this type, especially those originating from oxidised sulphide ores can show extreme 637C1 values (Eggenkamp & Schuiling, 1995). 28.5 Methods for the measurements of stable Br isotopes Only very few methods have been described for the measurement of Br stable isotope compositions. As the relative mass difference between 79Br and 81Br is only 2.5%o it is expected that isotope variations should be very small, and, as the (geo)chemistry of bromine is comparable to chlorine, the isotope data could be expected to show the same information.
28.5.1 Separation of bromide from a large excess of chloride One of the biggest challenges in measuring Br isotope compositions is the large excess of chloride in most geological samples. This makes it necessary to separate the chloride from the bromide, which might be a problem due to the comparable chemistries of the two elements. Eggenkamp & Coleman (2000) describe a method, based upon classical techniques, able to produce nearly pure bromomethane that can be used as mass spectrometer gas. The method used is based upon the well-known differences of oxidation-reduction behaviour of the halogens. With increasing atomic number the halide ions are increasingly more easily to oxidise to the native elements. Bromide is oxidised by a boiling solution containing K2Cr207 and H2SO4, while chloride is not. Separation of Br from C1 is performed in a simple distillation apparatus (Figure 28.2; after Friedheim & Mayer, 1892). The left hand 500 ml round bottom flask (a) is filled with the sample (either in solution or as a water-soluble salt), containing 2-10 mg bromide. Then 10 grams of K2Cr207 is added, and the flask filled with approximately 100 mL distilled water. To facilitate gentle boiling anti-bumping granules are added to this flask. Through the separating funnel (b) 20 ml of a 1"1 H2SO4:H20 mixture is added, after which the fluid in flask (a) is diluted to approximately 200 ml with distilled water. The right hand flask (c) is filled with 100 ml of a solution containing I g KOH. This flask is put into a bath containing cooling water. The water is not stirred, so that it becomes warm while the distillation takes place. The left hand flask (a) is then heated. When the contents of this flask start to boil, yellow-brown vapour is evolved and is distilled over to flask (c) through connection (d). This is Br2-gas. A small part (<1%) of the C1- in flask (a) is transferred to flask (b) in the form of HC1 (due
620
Chapter 28 - H.G.M. Eggenkamp Figure 28.2 - Distillation equipment for the separation of Br from C1. 'a' and 'c' are round-bottom flasks (500 ml), 'b' is a separating funnel and 'd' is the connection between round-bottom flasks 'a' and 'c'.
to the acid conditions in flask (a)), and a small portion reacts with the K2Cr207 to form chromyl chloride (CRO2C12), which is not carried over in the distillation (Dechan, 1886). Br2 will react with KOH to form KBrO and KBr, as this solution slowly increases in temperature the KBrO slowly decomposes to KBr and 02 (Bremner, 1965b). Although the bulk of Br is transferred over in the first few minutes, the distillation is continued for 20 minutes to ensure as complete as possible a yield of Br. After 20 minutes the boiling is stopped and the distillation set-up is disassembled and the contents of flask (b) is transferred into a 500 ml Erlenmeyer flask, approximately 2 g Zn dust (Aldrich, <10 micron, 95+ %) is added, and the mixture is boiled for 10 minutes. Any residual BrO- left in solution will then be reduced to Br- that can be processed further. As a few tenths of a percent of the originally present C1 is also transferred the procedure should be repeated for samples with high (>1000) initial C1/Br ratios. The resulting solution is stored in HDPE bottles. 28.5.2 Gas-IRMS method Willey & Taylor (1978) showed that bromomethane is a gas that behaves very well in a gas source mass spectrometer, and that the isotope ratio of this gas is measurable with a precision that is at least as good as that for chloromethane. The chemical procedure described above is able to remove approximately 99.6% of chloride from a solution. That still leaves about equal amounts of chloride and bromide in a solution. The remainder of the chloride is removed during the gas chromatography stage. Bromomethane is formed by the same procedure as that for chloromethane (Willey & Taylor, 1978). Both the remaining chloride, and the bromide are converted (after addi-
Methods for Determiningthe Stable Isotope Composition of Chlorine and Bromine ...
621
tion of silver nitrate) in their respective silver salts. The mixture of silver chloride and silver bromide is then reacted with iodomethane as described above. It is now very important that the gaschromatographic separation it is done in two steps. During the first run, the first two peaks that arrive are chloromethane and bromomethane. After passing trough this peak valves are switched to avoid iodomethane to go through the column. During the second run the chloromethane is not trapped into the sample vessel, and valve 4 is only switched to send the gas to the sample vessel after the chloromethane peak has passed. Then the only gas that is trapped into the sample vessel is bromomethane. If during measurement of a sample in the mass spectrometer the current of m / z 50 is measured, this is generally 1/100th of the current on m / z 94. That means that in these two procedures together more than 99.995% of all chloride is removed from a sample. The resulting bromomethane is purified enough to be measured accurately, with an analytical precision (standard deviation) of +0.18%o (Eggenkamp & Coleman, 2000). 28.5.3 Bromine isotope measurements using thermal ionisation mass spectrometers 28.5.3.1 Negative ion measurements The first attempts to measure Br stable isotope variations were made using solid source mass spectrometry. The first study to determine natural Br isotope variations (Cameron & Lippert, 1955) used elemental Br, prepared by commercial suppliers. The bromine was reacted with ammonium hydroxide in an ice-bath, evaporated to dryness and a calculated amount of sodium hydroxide was added to form sodium bromide. This was added to a nickel filament and installed in a mass spectrometer. Ion beams with masses 79 and 81 were subsequently measured. The same technique was applied by both Cameron et al. (1956) and Lund6n & Lodding (1960). 28.5.3.2 Positive ion measurements Xiao et al. (1993) developed a method to measure stable Br isotopes using positive ion thermal ionisation mass spectrometry. They used a technique comparable to the measurement of chlorine isotopes (Xiao & Zhang, 1992), and measured the ratio of the masses 345 (Cs279Br+) en 347 (Cs281Br+). No method to extract Br from natural samples was proposed, and the isotope composition of Br samples provided by factories was determined. They found variations up to 1.0%o, and reported an anylytical e r r o r of +0.11%o. 28.5.4 Bromine stable isotope measurements with ICP-MS In a few cases, researchers have attempted to measure Br stable isotope ratios by inductively coupled plasma mass spectrometry have been reported (Janghorbani et al., 1988; Nirel & Lutz, 1991; D'Allessandro et al., 1997). But as errors in the 81Br/79Br ratio are up to 1 to 5% (10 to 50%o!) this method is not yet feasible for natural variations, and could only be applied to tracer studies (in which 79Br or 81Br enriched material is added to the system). 28.6 Final remarks
- Conclusions
In this review recently developed and described analytical methods to measure the stable isotope ratio variations of the halogens C1 and Br have been described. All tech-
622
Chapter 28 - H.G.M. Eggenkamp
niques have their own stronger and weaker points, and it will depend upon the researcher, and the available equipment in their laboratories, as to which technique will be used in a specific application.
Acknowledgements Without the pioneering work of the isotope hydrology group at the University of Arizona, Tucson, including Austin Long, Ronald Kaufmann and Chris Eastoe the knowledge of the isotope geochemistry of the halogens would certainly not be as well developed as it is today. Very important was, and probably still is, the funding, supplied by research councils and foundations in various countries, as without their courage to invest money in this new field of research development would certainly have been delayed considerably. Many thanks also to the reviewers Profs Austin Long and Neill Sturchio who supplied useful suggestions to improve both the contents, and the English of this paper. I also would like to thank Dr. Pier de Groot for the invitation to write this chapter and his enormous amount of patience to get me revise this chapter. Pier Bedankt!
Handbook of Stable Isotope Analytical Techniques, Volume 1 P.A. de Groot (Editor) 9 2004 Elsevier B.V. All rights reserved.
CHAPTER 29 Selenium, Iron and Chromium Stable Isotope Ratio Measurements by the Double Isotope Spike TIMS Method Thomas M. Johnson1 & Thomas D. Bullen2 Geology Department, 245 Natural History Bldg., MC-102, University of Illinois, Urbana-Champaign, Urbana, IL 61820, U.S.A. 2 Water Resources Division, MS-420, U.S. Geological Survey, 345 Middlefield Road, Menlo Park, CA 94025, U.S.A. e-maih 1 [email protected]; 2 [email protected] 1
Abstract We present double-spike calibrated TIMS methods for measurement of massdependent isotope fractionation in Se, Fe, and Cr. Current measurement precision is approximately + 0.2 per mil (95% confidence) on 80Se/ 76Se, 56Fe/ 54Fe, and 53Cr / 52Cr. Sample size requirements are 500ng, 1/~g, and 250ng for Se, Fe, and Cr respectively. These measurements have been developed recently, and further improvements in precision and sample size are likely. We present purification procedures for these elements and review the geochemical applications of the measurements.
29.1 Introduction Recently, measurement of stable, non-radiogenic isotope ratios by thermal ionization mass spectrometry (TIMS) has increased greatly. TIMS analyses tend to be more difficult than gas-source analyses because of the difficulty in controlling a n d / o r monitoring instrumental discrimination. Conditions within samples ionizing on hot TIMS filaments are difficult to reproduce exactly from one sample to the next, and each sample is isotopically fractionated continuously as it is consumed from the filament during a run. Thus, instrumental discrimination varies from sample to sample and with time within each run. There is no way to rapidly and effectively compare sample and standard, and thus one must take pains to keep the discrimination stable over time or monitor the discrimination via tracer isotopes (i.e., double spikes) that are completely mixed with the sample. This chapter focuses on the latter approach, as it is applied to Se, Fe, and Cr isotope measurements. With each of these elements, we are fortunate in having four or more stable isotopes, so that two can be spiked and two or more can be used for measurement of natural isotopic variation. In many cases, measurements of radiogenic isotope ratios such as 87Sr/86Sr do not require the double isotope spike approach, as instrumental discrimination is monitored by measuring two natural stable isotopes. For example, with 87Sr/86Sr measure-
624
Chapter 29 - T.M. J o h n s o n & T.D. B u l l e n
ments, discrimination is monitored by comparing measured 86Sr/88Sr to an assumed value of 0.1194, then correcting 87Sr/86Sr data for the observed discrimination. This correction not only removes instrumental discrimination, but also removes natural mass-dependent isotopic fractionation, which could be significant given the high precision of 87Sr/86Sr determinations. However, with Pb isotope measurements, this cannot be done and the double spike approach has been developed and used for many years (Cumming, 1973; Dallwitz, 1970; Dodson, 1970; Gale, 1970; Hamelin et al., 1985; Hofmann, 1971; Russell, 1971; Russell, 1972). Current interest in low-temperature geochemical applications of stable isotope ratios of many elements necessitates the double spike approach as a means of measuring the mass-dependent isotope fractionation that is so carefully removed in the radiogenic isotope measurements. Se, Fe, and Cr have multiple oxidation states; this leads to complex biogeochemical behavior. Each element can be highly soluble or highly insoluble, depending on redox conditions. For Se and Cr, solubility increases with increasing oxidation state, whereas for Fe the opposite is true. In general, the reduction reactions are sluggish at circumneutral pH and are likely mediated by bacteria (Blum et al., 1998; Dowdle & Oremland, 1998; Dungan & Frankenberger, 1998; Lovley & Phillips, 1994; Lovley et al., 1987; Oremland et al., 1989; Tebo & Obraztsova, 1998; Zehr & Oremland, 1987). In contrast, oxidation reactions at circumneutral pH tend to be rapid, and can be driven by both microbial and inorganic processes. Redox reactions can result in isotopic fractionation, and thus the isotope data may provide information about redox processes in complex natural settings where mass balance studies may fail. Furthermore, microbial mediation of a particular redox process may result in a different amount of isotopic fractionation relative to the inorganic counterpart (e.g., microbially-mediated vs. abiotic precipitation of ferrihydrite), although the extent of difference is difficult to predict. Isotope data may thus be exploited as monitors of redox reactions as well as indicators of microbial activity, as has already been done with sulfate and nitrate reduction in groundwater (B6ttcher et al., 1990; Strebel et al., 1990) and in other settings (Kaplan, 1983; McMahon & Bohlke, 1996; Zaback & Pratt, 1992). Work on Se, Fe, and Cr isotopes is not very far advanced, and experiments are currently in progress to elucidate basic isotopic fractionation systematics. Selenium is chemically similar to sulfur and can be found in +6, +4, 0, and -2 valences, and in a variety of organic compounds, in natural settings (Elrashidi et al., 1987; McNeal & Balistrieri, 1989). The Se(VI) and Se(IV) oxyanions are highly soluble, but Se(IV) adsorbs to solids more strongly (Neal & Sposito, 1989). Se0 readily precipitates as elemental Se. Selenide (Se=) and other highly reduced forms of Se may substitute into sulfides (Coleman & Delevaux, 1957), or may be incorporated into proteins and other organic molecules. Se is an essential nutrient at low concentrations and a toxin at higher concentrations (Cooper & Glover, 1974; Ganther, 1974; Skorupa, 1998). Shales rich in organic matter are often rich in selenium, and Se toxicity problems have occurred where seleniferous shales are exposed in arid regions (Presser, 1994a; Presser, 1994b; Seiler, 1998). Se is of interest in ore deposits, where it may serve as a redox indicator (Coleman & Delevaux, 1957) and occurs in precious metal bearing minerals.
Selenium, Iron and Chromium Stable Isotope Ratio Measurements ...
625
Se isotope fractionation appears to be broadly similar to sulfur isotope fractionation, according to research completed so far, which is summarized in a recent publication (Johnson et al., 1999). Previous work has demonstrated that reduction of Se(VI) and Se(IV) induces significant isotopic fractionation (Krouse & Thode, 1962; Rashid & Krouse, 1985; Rashid et al., 1978; Rees & Thode, 1966; Webster, 1972). Accordingly, Se isotopic shifts observed in nature should be useful as indicators of Se sources and redox transformations. This is particularly valuable with Se because reduction of soluble oxyanions to insoluble Se0 decreases the mobility and bioavailability of Se (Elrashidi et al., 1987; McNeal & Balistrieri, 1989; Tokunaga et al., 1994). Iron is the fourth most abundant element in the earth's crust. It is found in most surficial environments, where it has a relatively simple redox chemistry and occurs in the +2 and +3 valence states. Under reducing conditions Fe is soluble, occurring either as the Fe(II) ion or complexed with anions such as chloride, carbonate and hydroxyl. In oxidizing environments, Fe(II) is readily oxidized to Fe(III), which rapidly precipitates from solution generally as a variety of oxyhydroxides. An additional complexity of Fe chemistry is that the mixed-valence state mineral magnetite (Fe2+Fe3+204) is stable under environmental conditions, and can be formed either through oxidative (Blattner et al., 1983; O'Neil &Clayton, 1964) or reductive (Lovley et al., 1987) pathways. Metallic iron is found in planetary cores, meteorites, and lunar basalts, but is extremely rare near the earth's surface. Fe plays an important role in many biological processes. Bacteria are known to use Fe as an electron donor/acceptor or as an electron transfer agent, and in doing so they catalyze redox processes (e.g., microbially-mediated precipitation or dissolution of hydrous ferric oxides, intra-cellular production of magnetite). Fe is critical to life as a component of enzymes and other basic compounds such as hemoglobin. A small but flowering body of recent work has demonstrated that the Fe isotopes can be, but are not necessarily fractionated by both microbially-mediated (Beard et al., 1999; Mandernack et al., 1999) and abiotic (Anbar et al., 2000; Beard & Johnson, 1999; Beard et al., 1999; Bullen et al., 2001) processes. The amount of fractionation recognized thus far in natural samples is on the order of a few per mil, and fractionations associated with abiotic and biotic processes appear to be of similar magnitude. Chromium is somewhat similar to S and Se chemically (it is a group VIB element, whereas Se and S are in group VIA). Chromium may be found in +6 and +3 oxidation states. Under oxidizing conditions Cr is soluble and mobile as the oxyanions CrO4 = (chromate) and HCrO4- (bichromate). Cr(VI) contamination derives from electroplating and other industries and is commonly found in urban groundwater (Fetter, 1994; Perlmutter et al., 1963). Chromate is a known carcinogen and may be otherwise toxic to fish and other wildlife (Barnhart, 1997; Katz & Salem, 1994). In reducing environments, Cr(VI) is reduced to Cr(III), which is likely to adsorb onto surfaces, precipitate as Cr(OH)3 or coprecipitate with ferric oxyhydroxides (Davis & Olsen, 1995; James, 1994). Bacteria are known to reduce Cr(VI) (Ehrlich, 1996), and in at least one case, a bacterium can grow with Cr(VI) as a terminal electron acceptor (Tebo & Obraztsova, 1998). We have recently completed a series of experiments demonstrating a 53Cr/52Cr
626
Chapter 29 - T.M. Johnson & T.D. Bullen
fractionation of a few per mil during reduction of Cr(VI) (Ellis et al., 2002). Cr isotope fractionation in natural settings was tentatively detected a few years ago (Ball, 1996), and we have recently found several per mil fractionation in Cr(VI) contaminated groundwater from locations where Cr(VI) reduction is thought to occur (unpublished data). 29.1.1 Previous Se, Fe, and Cr mass spectrometry 29.1.1.1 Selenium The first work on Se isotopes was done by Krouse & Thode (1962). They employed gas-source mass spectrometry using SeF6, following techniques developed for S. Se was extracted from samples, converted to Se0, fluorinated by reaction with fluorine gas, and measured in a specially modified mass spectrometer. Se from several ore deposits showed little variation in 82Se/76Se, but two plant samples and one soil sample spanned a range of 15%o. This study and a few later studies investigated isotopic systematics during reduction of Se(VI) and/or Se(IV) (Rashid & Krouse, 1985; Rashid et al., 1978; Rees & Thode, 1966), and concluded that significant fractionations occur. Webster (1972) studied Se isotope variation in a uranium ore deposit. Little work was done on Se isotopes in natural samples, perhaps in part because it is difficult to obtain enough Se to make measurements of this type in most earth materials. The gas source technique remains a workable option, especially with the advent of continuous flow techniques capable of measuring very small samples. However, it is likely that the sample requirement will remain much larger than that of the TIMS method presented here, and this is a critical issue in all but the most Se-rich environments. Also, fluorination is unavoidable, as SeO2 is solid up to 350~
One of the major recent developments in thermal ionization mass spectrometry (TIMS) is the development of methods for measuring negative ions. The electronic properties of Se result in a very low yield of positive ions during thermal ionization, and until recently, TIMS instruments were usually limited to measurement of positive ions. Recently, techniques have been developed that measure the abundant negative ions produced by several elements such as osmium and boron (e.g., Creaser et al., 1991; Hemming & Hanson, 1994:; Wachsmann & Heumann, 1992). Se forms the Se- ion readily by thermal effects, and work in Klaus Heumann's laboratory (e.g.,Wachsmann & Heumann, 1992) showed that mass spectrometry can be carried out on 500 ng Se. However, a double spike calibration was not employed, and the precision attained was + 1 - 3%o (80Se/76Se). Better precision is needed to apply Se isotope measurements in most natural settings. 29.1.1.2 Iron The first attempts to precisely measure Fe isotope compositions were made in studies of potential nucleosynthetic effects in meteorites (Voelkening & Papanastassiou, 1989). The predicted amount of variation was small, and thus the TIMS platform was generally used to obtain the best precision. A major strength of this approach is that the Fe isotopes can be measured simultaneously on the multi-Faraday collector arrays of the new-generation TIMS instruments, significantly reducing fluctuations of the measured ratios caused by ion beam instability. However, the substantial instru-
Selenium, Iron and ChromiumStable Isotope Ratio Measurements ...
627
mental discrimination usually encountered requires correction to obtain meaningful results. Prior to the recent application of the double-spiking technique, workers relied on two general methods to attempt to correct for machine discrimination during mass spectrometric analysis. The first approach, referred to as internal normalization, adjusts measured ratios of the Fe isotopes (e.g. 56Fe/57Fe, 58Fe/57Fe) to an accepted value for another measured reference isotope pair (e.g. 54Fe/56Fe) according to a mass-dependent fractionation relation (Voelkening & Papanastassiou, 1989). As mentioned above, this technique can be highly precise and has been successfully applied to the determination of Sr and Nd radiogenic isotope variations, but implicitly loses any information about mass-dependent isotopic variations in natural samples. The second approach involves correction of measured Fe isotope ratios according to the average instrumental discrimination estimated by isotopic analysis of gravimetrically-prepared Fe isotope standards. This technique has been successfully applied to the determination of isotope variations of heavy elements such as Pb. However, the technique requires quantitative recovery of the element during chemical purification procedures, and demands that all standards and natural samples are analyzed under precisely identical conditions (i.e., filament temperature, ion beam current, sample size, filament amendment proportions, etc.). Due to the inevitable variability of these conditions, the precision of this technique in the Fe mass range is probably on the order of I to 2 per mil per atomic mass unit (amu) difference. Analysis at low temperature (~ 1200~ and ion currents using Daly collector/ion counting instrumentation (Dixon et al., 1993) significantly reduces thermal fractionation effects, presumably because the fraction of Fe consumed is small. This allows for precise analysis of pure standards, but does not improve measurement precision for natural samples. 29.1.1.3 Chromium
Cr isotope ratios are easier to measure than those of Fe on the TIMS platform due to the lower ionization potential. As with the Fe isotopes, Cr isotope ratios were first studied in meteorites as indicators of early solar system evolution using TIMS (e.g., Birck & Allegre, 1988; Lee & Tera, 1986). Measured 50Cr/52Cr ratios were used as the internal normalization parameter to correct for instrumental discrimination, and 53Cr and 54Cr anomalies, relative to terrestrial materials, were extracted. This method implicitly loses any information about mass-dependent isotopic variations in natural samples. Ball et al. (2000) were the first to attempt to develop a TIMS technique that could be used to precisely measure natural stable isotope variations of Cr. This work centered on establishing both a chemical purification technique that resulted in near quantitative recovery of Cr from natural samples, as well as a reproducible filament loading and mass spectrometry method. Analyses were performed on a dual collector TIMS instrument which provided internally precise data, but had no capability to accurately report filament temperature. A statistically significant variation of a few per mil in the 53Cr/52Cr ratio was recognized for a suite of reconnaissance samples.
628
Chapter 29 - T.M. Johnson & T.D. Bullen
However, late in the study it was recognized that variation of filament temperature on a single sample over a range of ~ 100~ resulted in a similar isotopic shift, thus casting uncertainty on the range observed in the natural samples.
29.2 Double isotope spike methods In double-spiked isotope measurements, isotopic fractionation during mass spectrometry and sample preparation is measured and corrected for through addition of two stable spike isotopes to each sample. The spike isotopes are mixed to form a single solution with a constant ratio of the two spike isotopes that is used for all samples. During mass spectrometry, the raw measured ratio of the two spike isotopes depends on the instrumental discrimination, and thus the discrimination can be extracted from this ratio and used to correct the measured ratios of the unspiked isotopes. The technique was first employed in attempts to detect radiogenic Mo and Ba isotope anomalies in meteorites (Eugster et al., 1969; Wetherill, 1964), and has been used extensively in Pb isotope studies (Cumming, 1973; Dallwitz, 1970; Dodson, 1970; Gale, 1970; Hamelin et al., 1985; Hofmann, 1971; Oversby, 1969; Russell, 1971; Russell, 1972). More recently, double spiked Ca isotope measurements have been used to measure slight variations in 40Ca/44Ca in solar system materials and in biological cycling (Russell et al., 1978; Skulan et al., 1997). The spike isotopes were 42Ca and 48Ca, and a precision of +_0.2%0 was achieved in the more recent study.
29.2.1 Double spike data reduction In practice, double spike data reduction is somewhat complicated because the natural element contains significant amounts of the two spike isotopes, and the spike isotopes contain significant amounts of the other isotopes. However, the sample and spike can be "separated" mathematically, as long as their compositions are well constrained and the relative proportions of sample and spike are known. The latter information is contained in the measured ratios, so uncertainty in the amount of spike solution added to a sample does not affect the final result. The instrumental discrimination, the isotope ratio interest in the sample, and the relative proportions of sample and spike must be solved for simultaneously. It follows that three independent isotope ratios must be measured to achieve this" 1. The ratio of interest- two unspiked isotopes. 2. The ratio of the two spike isotopes, which will be used to calculate the instrumental discrimination. This ratio need not be measured directly, but can be calculated from ratios of the two spike isotopes to other isotopes of the element. For example, in our method for Se, the spike isotopes are 74Se and 82Se, and the 74Se/76Se and 8 2 S e / 8 0 S e ratios are measured. 3. A ratio of a spiked isotope to an unspiked isotope, which is used to determine the relative proportions of double spike and sample. The isotope ratio of interest in the sample must then be extracted from the raw ratios through a data reduction procedure. Various approaches have been taken to these calculations. If a linear fractionation correction is used (i.e., if the fractionation per a.m.u. mass difference is assumed to be the same for heavier and lighter isotope pairs), then
629
Selenium, Iron and Chromium Stable Isotope Ratio Measurements ...
a set of linear equations can be solved algebraically. This approach was used for Pb isotope studies (Cumming, 1973; Dallwitz, 1970; Gale, 1970; Hamelin et al., 1985). However, isotopic fractionation is known to be less for heavier isotope pairs than for lighter isotope pairs (e.g., Johnson et al., 1999; Russell et al., 1978) and linear fractionation is inadequate for Se, Fe, and Cr isotope determinations. A non-linear expression must be used to model the isotopic discrimination. An exponential discrimination law appears to be adequate (Beard & Johnson, 1999; Johnson et al., 1999; Russell et al., 1978), though the fit is empirical and not necessarily perfect. The measured ratio, r, is related to the unfraction- I Measure 3 Isotope Ratios ated ratio, ro, by ml) p r -
ro G
1. Subtract natural to obtain double spike ratio
[29.1]
where m l and m 2 are the masses of the two isotopes and p is the fractionation exponent that determines the discrimination. The accuracy of the exponential model in approximating fractionation by processes other than thermal ionization has not been established, but this is a negligible source of potential error relative to current precision (Johnson et al., 1999). Data reduction using an exponential discrimination law has been approached in two ways. The non-linear equations can be solved iteratively, as described below. Alternatively, equation [29.1] can be approximated by a first order expansion, which yields a linear equation and "closed form" algebraic solutions (Johnson & Beard, 1999; Russell et al., 1978). The closed form equations used for Fe isotope determinations are given in a recent publication (Johnson & Beard, 1999), and will not be repeated here. They can be adapted for Se or Cr readily. The error introduced by the linear approximation to the exponential fractionation law is minor, and the mathematical effort needed to set up the data reduction is less than the iterative approach. Setting up an iterative data reduction routine is somewhat more involved, but the extra effort is minimal compared to that required for other steps in developing a double spike. The routine
2. Compare double spike ratio to " k n o w n " value, calculate discrimination 3. Correct measured ratios for discrimination
~/"
disCrrlmic~-?otin~
NO-
YES 5. Subtract spike to obtain natural isotope ratio 6. Update natural isotopic composition, using exponential fractionation law
<,,
composition
.t> NO m
YES Calculate 6 Figure 29.1 - Flow chart for double spike data reduction by the iterative method. Step numbers correspond to those used in the text.
630
Chapter 29 - T.M. Johnson & T.D. Bullen
can be implemented in a computer spreadsheet, and recent desktop computers can complete the calculations instantaneously. Spreadsheet routines can be obtained from the authors upon request. The iterative routine described in a recent selenium isotope study (Johnson et al., 1999), which directly follows earlier studies (Eugster et al., 1969; Russell et al., 1978; Skulan et al., 1997) serves to illustrate the procedure. In this case, the spike isotopes are 74Se and 82Se and the naturally varying ratio is 80Se/76Se. The measured ratios are 76Se/74Se, 80Se/76Se, and 82Se/80Se. This routine is directly applicable to both Fe and Cr isotope measurements, where the spike isotopes are 57Fe and 58Fe for Fe and 50Cr and 54Cr for Cr, and the naturally varying ratios are 56Fe/54Fe and 53Cr/52Cr. The calculations follow a nested iteration scheme whereby the instrumental discrimination and natural isotope composition are successively refined (Figure 29.1)Q
A trial value for the 82Se/74Se ratio of the double spike, with the natural Se mathematically removed, is calculated:
745eI
745eI (8OSe)m_ (8~ 745el 82Ses,trial 82Se) (82S (80Se)m- ~808~)
825
x
82Sei
11- (80S~)m/(8__~Se)s] 74S I x - (8~
[29.2]
m/(74Se18-~Se)sl
where the subscripts m, n, and s refer to the measured mixture, the assumed natural Se composition, and the "known" double spike composition (see below), respectively. The measured 74Se/80Se ratios are obtained by multiplying 74Se/76Se by 76Se/80Se. The reader will note that this equation takes into account the significant 80Se impurity in the double spike. 0
The trial 82Se/74Se ratio of the double spike is compared to the "known" ratio to estimate the discrimination. Equation [29.1] is solved for p, the discrimination exponent.
3. Using the calculated value for p, the measured ratios are corrected for discrimination. 4. Step I above is performed using the corrected ratios; steps I through 3 are repeated iteratively until the values converge. This usually takes fewer than ten iterations. 0
A trial value for the sample's 80Se/76Se ratio is calculated by mathematically subtracting the double spike:
Selenium,Iron and ChromiumStableIsotopeRatioMeasurements... 765 I(765 +nc 8~ 80:)nat,trial_ (805~) 76
631
745el (745el 1 8~ nlc 7~SS~~80Ge) s nc- (765 ' 8~ s7~Se~ BOSe)
[29.3]
BOSe)nc
where the subscript "mc" refers to the measured ratios, corrected for discrimination, after step 4. 0
0
0
0
Step 5 gives a trial value for the sample's 80Se/765e ratio. This is not the final value because step I is calculated using an assumed isotopic composition for the sample. Now that an estimate of the actual composition has been made, the assumed natural isotopic composition must be updated. Using the result from step 5, new values for the sample's 745e/80Se and 825e/80Se are calculated according to the exponential fractionation law. The reader will note that this assumes the exponential fractionation law gives an adequate model for natural isotopic fractionation. Using equation [29.1] and the results from step 5, a value for the fractionation exponent, p in equation [29.1], is calculated. This is then used to update the natural 745e/80Se and 825e / 80Se. A new series of iterations (steps 1 through 3) that refine the instrumental discrimination are then performed using the sample's revised isotopic composition. Step 7 is repeated, usually less than 5 times, until the sample's isotopic composition converges to a precision far better than the measurement precision. The final ratio is converted to a delta value by comparison with results from a standard analyzed using the same double spike.
It is important to be sure the routine converges for each sample, as variations in sample composition, instrumental discrimination, and spike amount cause variation in the number of iterations required for convergence. However, provided the routine converges fully, the calculation error is orders of magnitude smaller than the analytical errors. The routine given above is essentially the same for Cr and Fe double spike analyses; the only difference is the ratios that are entered into the data reduction. In the above routine, the "known" 82Se/74Se ratio of the double spike is highly uncertain, but this does not significantly affect the final results when expressed as per mil deviations of samples from standards. It is impossible to measure the 82Se/74Se ratio accurately by mass spectrometry because of the unavoidable instrumental discrimination. It is possible to calibrate the double spike gravimetrically, by weighing the enriched spikes before they are dissolved and mixed, but the results will be highly uncertain if the amount of spike purchased is small and/or the spikes are purchased as hygroscopic salts. Alternatively, the concentrations of the spike solutions can be
632
Chapter 29 - T.M. Johnson & T.D. Bullen
measured precisely via isotope dilution, but instrumental discrimination limits the precision of these determinations. The final uncertainty in the double spike 82Se/74Se ratio is likely to be several per mil, and thus the adoption of an uncertain value likely imparts a bias to the final ratios. However, this bias is propagated equally to standard and samples alike. As with gas-source mass spectrometry results, the absolute ratios are not accurate, but the per mil deviations are. The assumed isotopic composition of the sample used in the data reduction routine can be obtained by measurement of an unspiked aliquot of the sample, but usually this is not necessary unless radiogenic, cosmogenic or non-exponential massdependent fractionation effects are expected. In meteorites, radiogenic isotope anomalies resulting from processes in the early history of the solar system have been measured (e.g., Birck & Allegre, 1988; Podosek et al., 1991). However, terrestrial Se, Cr, and Fe should be constant in their radiogenic contents, as the parent nuclides became extinct before planetary differentiation. A possible exception is materials with large meteoritic components, such as deep-sea spherules or sediments deposited after meteorite impacts (Herzog et al., 1999; Shukolyukov & Lugmair, 1998). In most cases, it should be sufficient to measure the isotopic composition of representative unspiked samples to be sure that there are no radiogenic effects and that they are different only in the degree of mass-dependent fractionation (Johnson & Beard, 1999; Russell et al., 1978). The results of unspiked runs clearly do not give the true isotopic composition of the samples, but give ratios that are systematically shifted due to instrumental discrimination and randomly distributed over some range determined by the uncertainty. The values obtained for terrestrial materials form fractionation arrays on ratioratio plots (e.g., 74Se/ 76Sr v. 82Se/ 80Se, 54Fe / 56Fe v. 58Fe/ 56Fe, 50Cr / 52Cr v. 54Cr / 52Cr) that conform to the exponential mass dependent fractionation relation given in equation [29.2]. As long as the initial assumed natural isotopic composition chosen lies anywhere along that array, the relative deviation of the measured isotopic composition of samples from that of the standard will be consistent. To standardize methodology, we propose that currently the best natural values to use are those of Wachsmann & Heumann (1992) for Se, Johnson & Beard (1999) for Fe, Shields et al., (1966) for 52Cr/50Cr, and Rotaru et al., (1992) for 53Cr / 52Cr and 54Cr / 52Cr. These values are given in Table 29.1.
29.2.2 Planning and Developing Double Spikes The first step in establishing a new double spike is selection of spike isotopes if there are multiple options. The most important issue is the natural abundance of the isotopes. It is best to spike using the naturally rare isotopes and leave the more abundant isotopes unspiked, as this gives the largest ion currents and best measurement precision possible by measuring ratios as close to one as possible. In addition, when the natural contribution to the spiked isotopes of the sample/spike mixture is small, the effects of uncertainty in the assumed natural composition are minimized. Another issue to consider is sensitivity to the choice of fractionation law. If the spiked and measured isotope ratios have greatly different mean masses (e.g., spiking 74Se and 76Se
633
Selenium, Iron and Chromium Stable Isotope Ratio Measurements ... Table 29.1 - Compositions of natural Se, Fe and Cr and currently used double spikes (atom %).
Natural Se 82 / 74 Spike
74Se 0.889 31.757
76Se 9.366 1.9260
77Se 7.635 0.9159
Natural Fe 58 / 57 Spike
54Fe 5.843 0.640
56Fe 91.758 16.553
57Fe 2.118 44.352
Natural Cr 54 / 50 Spike
50Cr 4.3450 48.517
52Cr 83.785 2.232
53Cr 9.5059 0.527
78Se 23.772 2.2908
80Se 49.607 4.0870
82Se 8.731 59.023
58Fe 0.281 38.455 54Cr 2.3637 48.724
and measuring 82Se/80Se), the choice of fractionation law is more important than the case where the spike ratio and the measured ratio have the same mean mass (e.g., spiking 74Se and 82Se, and measuring 76Se/80Se). This follows from the fact that, regardless of the chosen fractionation law, the fractionation per a.m.u, is approximately a function of mean mass. Another consideration in choosing spike isotopes is the cost of the spikes. Separated isotopes can be obtained from various sources 1. The cost of a purified isotope is directly related to its rarity, and prices vary from a few U.S. dollars to several hundred U.S. dollars per milligram. Despite the cost of spike isotopes, the amount purchased should be as large as possible because calibration of each double spike solution is time-intensive and ideally this should be done only once in a lifetime. The next steps are to measure the impurity isotopes in the spike isotopes and calibrate the solutions. In some cases, it may be possible to calibrate the double spike gravimetrically, if the purified isotopes are not hygroscopic and can be purchased in large enough quantities to allow highly precise weighing. Generally though, one dissolves the purified isotopes and calibrates the solution concentrations via isotope dilution (Faure, 1986) of both solutions against a standard solution of normal isotopic composition. The accuracy of the isotope dilution determinations depends on the precision of the ratio measurements (i.e., on variation in instrumental discrimination) and may only be accurate to a few parts per thousand. Thus, when the spike isotopes are mixed, the uncertainty in the spike isotope ratio may be several per mil. The mixing ratio of the two spike solutions influences the precision of the isotopic determinations, and it is thus advantageous to determine the optimal ratio before the solutions are mixed. With the iterative data reduction method, this can be done by generating synthetic data for a range of spike mixtures and observing the degree to which errors are propagated from all of the inputs, through the data reduction routine, to the final results. With the closed form algebraic solution, error propagation 1. including the Isotope Production and Distribution group at the U.S. Dept. of Energy's Oak Ridge National Laboratory (ORNL; http://www.ornl.gov / isotopes/)
634
Chapter 29 - T.M. Johnson & T.D. Bullen
equations can be derived and plotted to find the optimal spike mixing ratio (Johnson & Beard, 1999). It is possible, in some cases, to improve the precision of the double spike by adding a small amount of a third enriched isotope (e.g., Johnson & Beard, 1999). This is because the effects of the impurity isotopes in the double spike (i.e., the nominally unspiked isotopes in the targeted natural ratio) are minimized if the ratio of the impurity isotopes is close to the natural ratio. Once the optimal double spike composition has been found, the solutions can be mixed and stored in a manner that avoids any isotopic fractionation. The isotopic composition of the final double spike solution should be measured to the best possible precision, because errors in the abundances of the impurity isotopes cause systematic errors in the final results. Another consideration is the mixing ratio of sample to spike. Again, there is an optimal ratio that minimizes error propagation, and this can be determined using synthetic data. As a general rule, the mixture is within reasonable proximity of the optimum when greater than 90% of the spike isotopes in the mixture derive from the spike solution and greater than 90% of the unspiked isotopes in the mixture derive from the sample. To attain the optimal mixing ratio one must measure the concentration of each sample before spiking. Gross over- or under-spiking (e.g., by a factor of two or more) increases propagated uncertainties significantly, and gross overspiking causes systematic errors due to uncertainty in the double spike composition. Sequentially analyzing grossly over-spiked, under-spiked and optimally-spiked standards can provide considerable confidence in the data reduction algorithm and identify systematic errors to be expected in the occasional case of improper spiking.
29.2.3 Currently used Se, Cr, and Fe double spikes Given the isotopic composition of Se (Table 29.1), we chose 74Se and 82Se as the spike isotopes. This choice has several advantages. First, the wide mass spread of the spike isotopes maximizes the ability of the double spike to detect and correct for instrumental discrimination. Second, 74Se is a logical spike isotope because the natural abundance is low and thus uncertainty in the natural abundance causes little error in the final results. Third, the mean mass of the 82Se/74Se ratio is the same as the mean mass of the 80Se/76Se ratio, and thus the final 80Se/76Se results are highly insensitive to the choice of discrimination law. Also, this spiking scheme enables simultaneous determination of 80Se/76Se, 80Se/77Se, 80Se/78Se ratios. These ratios all utilize the major Se isotope, 80Se, giving the highest beam currents and thus the best precision possible. Finally, a dynamic peak-hopping routine measuring 82Se / 80Se, 76Se / 80Se, and 74Se/76Se as described below minimizes the sensitivity of the measurements to amplifier gain drift. Because many isotope combinations can be measured with Se, expanded notation is necessary for unambiguous reporting of selenium isotope data. Measurements can be reported as 680/76Se, 680/77Se, and 680/ 78Se, the per mil deviations of the indicated ratios from those of a standard. The isotopic composition of the double spike we use currently is given in Table 29.1. The double spike 82Se/74Se ratio, measured by mass spectrometry, was checked via isotope dilution of the 82Se and 74Se solutions against a standard solution. The two results agree within the _+1% uncertainty of the isotope dilution determination (gravi-
Selenium, Iron and Chromium Stable Isotope Ratio Measurements ...
635
metric calibration was not attempted because the 74Se spike is costly, and the small amount purchased could not be precisely weighed). There are two options for Fe spike isotopes; the isotopic composition of natural Fe is given in Table 29.1.58Fe is a logical choice for one, because it is rare in nature. One can then choose between 54Fe and 57Fe for the other spike isotope. If 54Fe is spiked, 56Fe/57Fe, with only one mass unit difference is the measured natural ratio. If 57Fe is spiked, 54Fe/56Fe is the measured natural ratio, with the advantages of a greater mass difference in the measured natural ratio and greater abundance of the lesser isotope by a factor of 2.5. However, this approach is less able to precisely correct for instrumental discrimination because there is less mass difference in the spiked ratio. This approach is also more prone to problems with the mass fractionation law used, because the mean masses of the spiked and measured natural ratios are quite different. For our work, we have chosen 57Fe and 58Fe as the spike isotopes. The isotopic composition of our current Fe double spike is given in Table 29.1. Similar precision has been obtained with a 54Fe-58Fe double spike (Johnson & Beard, 1999). The options for Cr are very similar, as the relative masses and abundances of the Cr isotopes are close to those of the Fe isotopes. The isotopic composition of natural Cr (Birck & Lugmair, 1988) is given in Table 29.1. The key difference between Cr and Fe is that 53Cr is of relatively high abundance and is thus attractive as a measured natural isotope and less attractive as a spiked isotope. The logical choice for Cr is thus spiking of 54Cr and 50Cr, and measurement of 53Cr/52Cr. The composition of our current Cr double spike is given in Table 29.1. 29.3 Purification
Preparation of samples for mass spectrometry must succeed in separating the target element from all the major elements and certain interfering trace species in the sample matrices: Because the double spike method corrects for fractionation incurred during sample processing as well as thermal ionization, the spike should be added as early as possible in the sample preparation process. This removes the requirement of 100% yields, which are otherwise needed to ensure a lack of fractionation during sample processing steps known to fractionate isotopes (e.g., Anbar et al., 2000; Russell & Papanastassiou, 1978). However, it is critical that the spike is in the same chemical form(s) as the sample Se, Fe or Cr, so that any fractionation that occurs affects sample and spike in exactly the same way. It is useful to have multiple forms of the double spike, e.g., Se(VI) and Se(IV), so that spiking can be carried out as early as possible in sample preparation. 29.3.1 Selenium
The two most important interferents that must be removed during Se processing are phosphorus and organic compounds. Isobaric interferences have been observed from PO3- ions and from ions derived from organic molecules. Fortunately, the major mass of the PO3- ion is 79 a.m.u; it is only the 170-bearing species at 80 a.m.u, that causes problems. Also, PO3- ions appear only at temperatures greater than 1000~
636
Chapter 29 - T.M. Johnson & T.D. Bullen
The organic interferences are more problematic. Organic residues can severely retard ionization, and an array of fragments appear in the 74 to 82 a.m.u, mass spectrum when organics are present on the filament or in the ion source vacuum. For example, a peak centered at 76.02 a.m.u, has been observed, which appears to be C6H4. The fact that this species has a mass greater than 76.00 indicates that it contains hydrogen. It is likely that this is a cyclic hydrocarbon, many of which can be resistant to oxidative cleanup steps (e.g., Anbar et al., 1997). Fortunate134 this particular species is almost completely resolved from 76Se (75.92 a.m.u.) and only causes trouble when its intensity is quite high. Species at 77 and 78 a.m.u, seem to cause similar problems. The one at 77 a.m.u, appears to affect many samples and standards significantly, and so far has rendered the 80Se/77Se ratios useless. No interferences have been observed from SO3ions or from 74Ge and 76Ge. One general warning that must be heeded in processing of Se for mass spectrometry is that significant Se losses occur as acidic solutions are evaporated to dryness. Chau & Riley (1965) measured Se losses of 82%, 27%, and 73% during evaporation to dryness in Add double spike I HC1, HNO3, and H2SO4 solutions, respectively. However, they also reported that the presence of Convert Se to Se(IV): sodium, and presumably other cations, in the 6 M HC1 at 100~ 30 min. HNO3 evaporation decreases the Se loss to essentially zero. In our experience, evaporation of React with NaBH4 to generate H2Se HNO3 matrix solutions causes no loss of Se, but the evaporation of HC1 matrix solutions causes large losses. Scrub H2Se from solution, collect in conc. HNO3 trap
29.3.1.1 Batch Hydride Generation Selenium can be purified for mass spectrometry via a rapid "hydride generation" procedure, in which Se(IV) is converted to H2Se by reduction, scrubbed from solution, and recovered in an oxidizing trap (Tanzer & Heumann, 1991). Purification via hydride generation is facilitated by the double spike technique, because the process involves chemical reduction, the yield is typically less than 90%, and significant isotopic fractionations are expected. We have observed 80Se/76Se fractionations greater than 4%0. However, many double spiked standards processed through hydride generation have yielded results that are indistinguishable from those of unprocessed standards. A flow chart for Se purification via hydride generation is given in Figure 29.2. Hydride generation has been used extensively to enhance Se concentration analysis, is compatible with strongly acidic soil digests (Dedina & Tsalev, 1995), and has
Evaporate to dryness Add graphite Destroy organics: 100gL H202, heat on 70~ hot plate to dryness
small and
NO-
,~ YES Load on top of Ba(OH)3 Figure 29.2 - Flow chart for Se purification.
Selenium, Iron and ChromiumStable Isotope Ratio Measurements ...
637
a large body of supporting literature. Sb, As, Bi, Ge, In, Pb, Te, Sn, and T1 may also form volatile hydrides (Dedina & Tsalev, 1995), but these do not adversely affect Se mass spectrometry by TIMS. The hydride technique is Se(IV)-specific and does not recover Se(VI). Se(VI) can be converted to Se(IV) by adding HC1 to attain a 6N HC1 solution and heating for 30min at 100~ (Brimmer et al., 1987). Until recently, we used a batch-type procedure developed by Tanzer & Heumann (1991). The sample HC1 concentration is adjusted to 4M, and a NaBH4 (4% w/V, stabilized with 0.1% NaOH) reductant solution is metered with a peristaltic pump (lmL/min.) into a stripping vessel containing the entire sample solution. H2Se is scrubbed from the sample solution by N2 or Ar and trapped in concentrated nitric acid. This procedure was used to purify Se in three studies that applied negative ion TIMS mass spectrometry to measure Se concentration by isotope dilution (Heumann & Radlein, 1989; Heumann & Wachsmann, 1989; Tanzer & Heumann, 1991). We have used the Tanzer & Heumann (1991) procedure with three modifications. First, ferric iron is removed via cation exchange resin whenever the Fe(III) concentration in the reaction vessel would be greater than 100mg/1 because it can interfere with Se reduction (Dedina & Tsalev, 1995). Second, organic compounds transferred along with the H2Se are destroyed by oxidation (see below). Some dissolved organic compounds in the samples apparently are volatilized by the reduction/scrubbing procedure, as is evidenced by suppressed ionization in the mass spectrometer. Finally, the entire hydride generation apparatus is constructed of PFA plastic (Savillex Corp., Minnetonka, MN, USA) and is rinsed with 8M HNO3 between samples. The Tanzer & Heumann (1991) method used a glass vessel, but this is more difficult because the glass should be coated with silane periodically.
29.3.1.2 Continuous-flow Hydride Generation We have recently developed a continuous flow purification method that simplifies and improves sample processing. The sample, in a 6N HC1 matrix, the reductant (1% NaBH4 solution), and a small flow of carrier gas are pumped via a peristaltic pump into two mixing junctions, where they meet. The flow rates we use currently are 7 mL / min., I mL / min., and 100 mL / min. for sample, reductant and carrier gas, respectively. The reacting solution is then transport via PFA tubing to a glass stripping vessel (Figure 29.3) similar to that described in a recent publication on As analysis by atomic emission (Brindle et al., 1992). The bottom of the chamber is a porous glass frit through which a carrier gas is pumped at approximately 100 mL/min, to strip the H2Se from the solution. The gas is then passed to an oxidation trap, and the reacted solution is pumped out of the reaction vessel continuously. A 4N HC1 solution is passed through the apparatus between samples to clean it. This continuous flow apparatus requires fewer steps between samples; the batch method requires opening, cleaning, and reassembling the reaction vessel after each sample. More importantly, the continuous flow method uses less reagent for small volume samples and is less sensitive to interference by ferric iron and other oxidizing species in sample matrices. Large amounts of boron are introduced with the NaBH4 into the sample solutions in either hydride process, and some of this boron passes, possibly as volatile BC13, to
638
Chapter 29 - T.M. Johnson & T.D. Bullen
the oxidation trap. Very large BO2- ion beams (> 10qOA) are observed during analysis of samples and hydride-processed standards, but these apparently do not degrade the quality of the analyses. Also, significant BO2- is always observed when Ba loads are used, presumably as a result of impurities in reagents. 29.3.1.3 Anion chromatography Anion exchange chromatography could also be used in some cases but is problematic for soil or rock digests. The usual form of Se in these digests is Se(IV), which forms an uncharged species under the highly acidic conditions needed to stabilize dissolved Fe, A1, and other highly charged elements. With other types of samples that can tolerate higher pH conditions, it is possible to separate Se(IV) Figure 29.3 - Glass H2Se stripping vessel u s e d for c o n t i n u o u s from other species using flow purification of Se. an anion exchange resin and a weak acid eluent such as formic acid (Tanzer & Heumann, 1991). Another approach is to oxidize the Se to Se(VI), which is strongly retained on anion exchange resins under acidic conditions. It must be noted, however, that Se(VI) will reduce slowly to Se(IV) in strong HC1 at room temperature. Also, separation of sulfate and selenate could be difficult because of their chemical similarity. In general, anion exchange procedures are more time-consuming than hydride generation and we have favored the hydride techniques. 29.3.1.4 Preconcentration via anion exchange or copreciptation Se concentrations in water samples are usually below lmg/1, and large volume samples must be processed for mass spectrometry. It is helpful to preconcentrate the Se before the hydride generation step to decrease the volume to be processed. In fresh
Selenium, Iron and ChromiumStableIsotopeRatio Measurements...
639
water samples, all of the anions in a large volume of water can be collected onto an anion exchange resin column. At roughly neutral pH, this method quantitatively recovers selenite and selenate, but likely loses most of the dissolved organically bound Se. The separation of Se species in water samples is the subject of many papers in the analytical chemistry literature (e.g., Dedina & Tsalev, 1995) and is beyond the scope of this chapter. With brackish and saline waters, preconcentration of Se by anion exchange is not feasible because the amount of anions in solution is much greater than the capacity of the resin (typically about 1 milliequivalent of ions per mL resin bed). Se (IV) can be preconcentrated from fresh and saline waters via coprecipitation with ferric hydroxide (Chau & Riley, 1965). 50 mg Fe(III) (as FeC13 solution) is added per liter of water, and the pH is adjusted to 4.0 to precipitate ferric hydroxide. The precipitate is recovered by centrifugation or filtration and dissolved in 0.2N HNO3. Fe is then removed by passing the solution through cation exchange resin and washing with 0.2N HNO3. The solution can then be evaporated to dryness and processed further.
29.3.1.5 Oxidation of organic compounds The most common problem in Se mass spectrometry by this method is suppressed ionization caused by organic species. Dissolved organic compounds in natural waters and sediment extracts appear to move with the H2Se in the hydride generation process and appear as a white spot when the HNO3 solution from the trap is evaporated away. Organic contaminants have been also been observed in the HNO3 itself. Currently, we oxidized most or all of the organics by adding 100 m130% H202 to the dried sample and heating on a 70~ hotplate to dryness. 29.3.2 Iron
For effective TIMS isotopic analysis, Fe must be separated from all major cations, anions other than chloride, and Cr and Ni in the sample. As is typical in TIMS analysis of other elements, the presence of any major cation other than Fe and certain anions makes filament loading difficult and generally causes suppression of the ion beam. 54Cr directly overlaps 54Fe, and 58Ni directly overlaps 58Fe, and although Cr and Ni typically occur at trace levels compared to Fe they may ionize far more efficiently, depending on filament loading technique, and may cause significant interference. In addition, we have observed significant probable organic-hydride peaks in the 52 to 60 a.m.u, mass spectrum that generally burn off during the course of TIMS analysis. However, efforts should be made to remove organics from the sample prior to filament loading as their presence greatly suppresses the ion beam and their burnoff increases analysis time. To date some of these organics have proven resistant to treatment with both H202 and HC104 following purification steps, and as with Se we are currently exploring ultra-violet irradiation techniques to eradicate them. The purification procedure for Fe is straightforward, and relies on the strong affinity of the Fe(III)C14- complex to strong-acid anion exchange resins. The procedure is given in flow chart form in Figure 29.4. Throughout the procedure for Fe (and Cr) Teflon-distilled acids and Barnstead distilled water are used in order to minimize
640 blank contribution. As a routine, sufficient fluid sample to provide I ~tg Fe is mixed with an optimal amount of double-spike solution in a Teflon beaker, and the mix is taken to dryness on a hot plate to ensure sample-spike homogenization. Meanwhile, approximately 2 mL of AG-1-X8 resin (BioRad Industries) is loaded into a small glass or Teflon column that is closed at the bottom with a porous Teflon frit. The resin is rinsed with 10 mL of H20, 10 mL of 2% HNO3, another 10 mL of H20, and finally equilibrated with 10 mL of 6N HC1. The solid residue of the sample is dissolved in 1 mL of 6N HC1 and loaded onto the prepared anion exchange column. The resin is then rinsed with 10 successive 1 mL additions of 6N HC1 to strip cations other than Fe and anions other than chloride from the sample load. We find that by doing successive additions and allowing each addition to completely flow through the resin bed, we minimize the ability of undesirable cations and anions to be redissolved into the fluid volume and thus remain near the top of the resin bed with the Fechloride complex. The Fe-chlorides are then stripped from the resin bed with 7 successive I mL additions of H20 and collected in a Teflon beaker. At this point, the sample can be either taken to dryness on a hot plate, re-dissolved in I mL of 6N HC1 and passed through a prepared resin column for a clean-up step, or prepared for loading on the filament for TIMS analysis.
Chapter 29 - T.M. Johnson & T.D. Bullen
Add double spike
]
Load sample, in 6 N HC1 onto AG 1-X 8 column Rinse with 10 x I mL 6 N HC1; FeC14- retained, other species eluted Elute Fe with 7 x I mL H20
]
Add 100 ~tL 0.15 N H3PO4
I
Evaporate to dryness Destroy organics: Add 100 ~tL 30% H202; evaporate
NO-
~ YES Add 10 ~tg SiO2 + 100 ~tL HC1; evaporate + load Figure 29.4 - Flow chart for Fe purification
When purification with the exchange resin is complete, 100 ~tL of 0.15 N H3PO4 (Ultrex) is added to the sample solution, which is then taken to dryness on a hot plate. When dry, the remaining small spot of H3PO4 that contains the Fe at the bottom of the beaker is usually dark-colored. As a final cleanup step, approximately 100 gL of H202 (Ultrex) is added to the sample and taken to dryness. The remaining small spot of H3PO4 should at this point be clear; if not, the treatment with H202 is repeated until it is. Finally, 10 gg of silica as a colloidal suspension and 100 gL of Teflon-distilled HC1 are added to the beaker, and the mixture is taken to near dryness to complete the purification process. 29.3.3 Chromium For effective TIMS isotopic analysis, Cr must be separated from all major cations, anions other than chloride, and Ti, V and Fe in the sample. 50Ti and 50V directly overlap 50Cr and 54Fe directly overlaps 54Cr. In addition, as with Fe, we have observed sig-
641
Selenium, Iron and Chromium Stable Isotope Ratio Measurements ...
nificant probable organic fragment peaks in the 48 to 56 a.m.u, mass spectrum that generally burn off during the course of TIMS analysis. To date, some of these organics have proven resistant to treatment with both H202 and HC104 following purification steps, and we are currently exploring ultra-violet irradiation techniques to eradicate them. The purification procedure for Cr (Figure 29.5) is somewhat more complicated than that for Fe, due to the occurrence of Cr as Cr(VI) and Cr(III). The procedure exploits this dualit)r and consists of a two-stage approach that relies on the strong affinity of the Cr(VI) anions for strong-base anion exchange resins, and the ability of Cr(III) cations to pass through the same resins. Sufficient fluid or digested solid sample to provide 250 ng Cr is mixed with an optimal amount of double-spike solution in the proper oxidation state. If necessar3r the Cr in the sample is converted to Cr(VI) form as follows: The solution is taken to dryness, and Cr in the sample is oxidized to Cr(VI) by adding 100 mL each of NH4OH and H202 to the beaker and taking the fluid volume to dryness on the hot plate. The sample is then dissolved in I mL of 0.1N HC1. Add double spike Convert H2SO3 to H2SO4: Heat on 100~ hot plate for 20 minutes YES Pass sample through second AG 1-X8 column (sulfate is retained)
m~
Convert Cr to Cr(VI): 100 gL NH4OH + 100gL 30% H202, evaporate to dryness Dissolve in I mL 0.1 N HC1
Rinse Cr(III) through column with 5 mL 0.1 N HC1
Load onto AG 1-X 8 column
Destroy organics: Add 100 gL 30% H202; heat on hot plate at 70~ to dryness
Evaporate to dryness
Rinse: 20 mL 0.1 N HC1 Reduce Cr(VI) to Cr(III): Add 1 mL 0.1 M H2SO3 to column
NO-
Elute Cr(III)" 5 mL 0.1 N HC1 ] YES
Dissolve in 3 ~L 0.1 N HC1 Add 15 gg SiO2 + 0.6/JL saturated boric acid load on Re filament, heat to red Figure 29.5 - Flow chart for Cr purification.
642
Chapter 29 - T.M. Johnson & T.D. Bullen
Meanwhile, approximately 2 mL of AG1-X8 resin (BioRad Industries) is loaded into a small glass or Teflon column that is closed at the bottom with a porous frit. The resin is cleaned to remove organic residues, rinsed, and equilibrated with 5 mL of 0.1N HC1. The sample containing all Cr as Cr(VI) is loaded onto the prepared anion exchange column. The resin is then rinsed with 20 mL 0.1N HC1 to remove all cations and the weak acid anions such as phosphate and acetate. Cr(VI) is then reduced to Cr(III) by 2 additions of lmL 0.1M H2SO3, eluted with 5mL 0.1N HC1, and collected in a teflon beaker. Strong acid anions are retained by the column, and at this point, only the Cr from the sample and the H2SO3 are in the beaker. H2SO3 is converted to H2SO4 by heating on a 100~ hot plate for 20 min. After the solution has cooled, it is passed through a second AG1-X8 column (2ml bed volume), on which the sulfate is retained. The Cr(III) moves through this column quickly; and any residual Cr is rinsed through with 5 mL 0.1N HC1. The sample is evaporated to dryness and the resin columns are cleaned with 50 mL 6N HC1. In order to destroy most of the organic residues originating from the columns, approximately 100 gL of H202 (Ultrex) is added to the sample and taken to dryness. The sample should dry to a very small brown or green spot; if not, the treatment with H202 is repeated until it does. It may be possible to speciate Cr from liquid samples using this resin column technique (e.g., Strelow, 1973). For example, the sample containing both Cr(III) and Cr(VI) can be loaded directly onto the prepared anion exchange resin column, and the eluent collected into a suitably-sized beaker. The Cr(VI) retained on the column can be purified as described above. The Cr(III) collected in the column effluent can then be converted to Cr(VI) and purified separately. 29.4 Mass spectrometry 29.4.1 Selenium A Finnigan MAT 261 mass spectrometer is used for our work with all three elements. For Se, the instrument is operated in negative ion mode, following general procedures worked out by Klaus Heumann and his group several years ago (Wachsmann & Heumann, 1992). Positive ion techniques for Se that generate polyatomic ions may be possible, but have not produced workable methods. A method for S isotope measurement via generation of AsS + ions exists (Paulsen & Kelly, 1984), but adaptation of this method to Se isotope measurement is difficult because of the chemical properties of AsSe. Thus, despite the difficulties of negative ion techniques, they are presently superior to the alternatives. Negative ion TIMS suffers in many cases from the deleterious effects of large electron beams generated on the filament. The abundant electrons are desirable in that they help to generate negative ions, but often, milliamp-size electron beams are accelerated along with the desired ions. These beams are intercepted to varying degrees by the plates in the ion source and then flow into the focussing electronics. Depending on the size of the electron beam, its path through the ion source, and the design of the source electronics, variable effects are observed. An
Selenium, Iron and Chromium Stable Isotope Ratio Measurements ...
643
apparent "smearing" of the ion beam may be encountered, where a sharp maximum current as a function of the focus parameters is not seen and beam currents are lower than expected. Also, the electron currents may alter the potentials of the source plates to the extent that a maximum cannot be found within the range of the focus controls. Finally, large electron beams can cause arcing and possible damage in the source or the electronics boards. To date, we have taken the approach of avoiding very large electron beams and have worked to maximize ionization efficiency through means other than increasing electron density at the filament. The negative ion TIMS technique worked out by Wachsmann & Heumann (1992) is a double filament technique in which ionization occurs on a hot filament onto which 30 gg Ba as Ba(OH)2 is loaded to increase production of electrons and thus enhance Se ionization. The sample, in the form of H2SeO3, is held in a silica gel matrix on a relatively cool filament 3mm away. This technique was developed because the volatility of Se causes rapid Se losses at the temperatures required for efficient ionization (SeO2 and Se0 vaporize at 350~ and 610~ respectively). However, a weakness of this approach is the loss of Se that evaporates from the sample filament but does not migrate to the ionization filament. Wachsmann & Heumann (1992) reported ion beams of greater than 10-11A 80Se- on 0.5 gg loaded Se with their technique. We were unable to obtain more than 3 x 10-12A 80Se-, except when filaments were oxidized with nitric acid. Under these conditions, a rhenium oxide coating apparently enhanced production of electrons; we observed enhanced Se beams, larger rhenium oxide beams, greater degradation of focussing by electron beam effects, and often, electron beams large enough to cause serious electric arcs in the ion source. We have recently published a new filament loading technique (Johnson et al., 1999) that has advantages over the Wachsmann & Heumann technique. Samples are loaded onto a single Re filament. One microliter of a saturated barium hydroxide solution (24 gg Ba) is placed on the filament and dried at the lowest practical temperature. The sample, containing 500 ng Se, is mixed with approximately 1 mg colloidal graphite, which is best dispensed as a dilute suspension. We currently use a few microliters of a suspension made by diluting "Aquadag" paste (Acheson Industries, Ltd.), with deionized water by factor of 5,000. The sample-graphite mixture is placed on top of the dried Ba(OH)2, and gently evaporated to dryness. The colloidal graphite serves to control the volatility of Se, and it may also decrease loss of Se as more volatile oxidized species a n d / o r control the formation of rhenium oxides. The behavior of the Se on the filament is highly sensitive to the graphite, and the amount must be carefully controlled. If too little is added, the beam intensity diminishes rapidly. If too much is added, lesser beam intensity is observed. This technique is easier to perform than the Wachsmann & Heumann (1992) technique and in our experience produces stronger ion beams (approximately 10-11A 80Se-). It generally produces smaller electron beams and better beam focus than the double filament technique. No significant Se contamination in the graphite has been observed. With the single filament graphite technique, filaments are heated to 900~ over a 5 minute interval. At 900~ Se- ions are observed, along with rapidly diminishing Br-
644 beams at masses 79 and 81 arising from traces of Br in the samples and reagents. Maximum signal intensity for Se- occurs between 950~ and 1000~ though data are usually collected at lower temperature to decrease drift in instrumental discrimination. Measurements of 82Se / 80Se, 78Se / 80Se, 77Se / 80Se,
Chapter 29 - T.M. Johnson & T.D. Bullen Table 29.2 - Optimal measurement sequence for Se mass spectrometry. Step Number 1 2 3 4
1
76Se 74Se
Collector Number 2 3 4
5
80Se
82Se
78Se 77Se
80Se 80Se 76Se
76Se / 80Se, a n d 74Se / 76Se ratios are
made in a sequence of three steps. This is necessary given the limited travel of the detectors in the Finnigan 261; it possible that with other instruments, the ratios could all be measured simultaneously. This instrument is equipped with 5 faraday collectors, and the two ion beams for each ratio are measured simultaneously to eliminate errors caused by temporal variations in intensity. The detectors cannot be positioned far enough apart to measure the 80Se/ 74Se ratio, and each ratio requires a unique detector spacing that usually cannot be used to measure other ratios. Accordingly, the ratios are measured in a series of steps given in Table 29.2. Use of this sequence minimizes the effects of drift in instrumental discrimination and amplifier gain on the 80Se/76Se determinations. Measurements are averaged and outliers removed after ten measurement cycles, and the results are fed into the data reduction routine. This averaging gives the same result as does reducing the data from each measurement cycle individually. It is helpful to monitor the PO3- beam (major mass at 79 a.m.u., minor masses at 80 and above), which may be very large for phosphate-contaminated samples. Fortunately; the PO3- beam intensity is low below 1000~ If a large signal at 79 a.m.u, is observed, comparison with the 81 a.m.u, beam is necessary to distinguish PO3- from Br-. 29.4.2 Iron and Chromium
The fundamentals of Fe and Cr analysis using TIMS are very similar, with the exception that Cr is much more readily ionized in the TIMS source. In previous studies of Fe isotope systematics, purified Fe samples ranging in amount from 0.1 to 10's of micrograms were loaded as a chloride complex onto single Re filaments with silica gel, and positive ions were generated between 1150~ and 1600~ The small sample size, low temperature technique was developed by Dixon et al., (1993) to produce small ion beams (< 10-13A) that could be accommodated by an ion counting device. The large sample size, high temperature technique was used by Voelkening & Papanastassiou (1989), to produce sufficiently large ion beams (> 10-11A) such that the minor isotopes 57Fe and 58Fe could be measured acceptably by Faraday collectors. Isobaric interferences by Cr and Ni were occasionally encountered, and were monitored by measurement of the 52 and 60 a.m.u, peaks. Similarly, in previous studies of Cr isotope systematics approximately one microgram of Cr as a nitrate complex was loaded onto single Re filaments with silica gel, and positive ions were generated between
Selenium, Iron and Chromium Stable Isotope Ratio Measurements ...
645
1100~ and 1600~ (Lee & Tera, 1986). Beam currents of approximately 10-11A were obtained. Isobaric interference by Fe was occasionally encountered, and was monitored by measurement of the 57 a.m.u, peak. We have attempted to improve on these previous techniques by developing filament loading methods that require relatively small sample load sizes (< 1 gg) and provide for strong ion beams that can be measured by Faraday collectors. The requirement of small sample size reflects the cost of the double spike, while measurement by Faraday collector results in greater analytical precision and the ability to simultaneously monitor multiple ion beams using the multi-collector configuration of the TIMS instrument. Experience has shown that the key to producing strong, longlived ion beams is simply the ability to retain the Fe or Cr sample on the Re filament surface at the high temperatures required for ionization. Post-analysis examination of filaments that provided short-lived and/or unstable ion beams revealed that little of the sample load remained on the filament surface. The following procedures were developed for Fe, but can be used equally well for Cr. We find that by using a small amount of phosphoric acid, as we do for the analysis of Sr, Nd, Pb, U, and Th isotopes, the Fe and Cr sample filaments can be raised to higher temperatures while retaining the sample on the filament surface and thus providing for greater ionization efficiency. Through experimentation we have also found that a small amount of aluminum deposited onto the filament prior to sample loading results in greater ion beam stability for both Fe and Cr. For consistency, the typical sample size is 1 gg for Fe, and 250 ng for Cr. The filament loading techniques for Fe and Cr isotope analysis are essentially identical. Approximately I ~g of A1 (~ 5 gL as a nitrate solution) is deposited drop-wise onto the center of the previously out-gassed Re filament and taken to near dryness at a current of 1.5A. The sample bead (consisting of the purified Fe or Cr, H3PO4, and colloidal silica) is dissolved in ~3 mL of 6N HC1 and deposited drop-wise onto the near-dry A1 slurry. The mixture is allowed to go to dryness, and the filament current is raised to 1.8A for 30 seconds during which the H3PO4 gives off a small amount of smoke. The samples are then loaded into the TIMS source, which is evacuated to a pressure of < 10-8 Torr. Most of the previous mass spectrometry of Cr has been done using boric acid and silica gel (Lee & Tera, 1986; Birck & Lugmair, 1988; Ball et al, 2000). This technique gives results similar to those of the technique given above, and may produce more stable ion beams for use with single collector instruments. The sample is dissolved in 3 gL 0.1N HC1. 15 mg of silica as a colloidal suspension and 0.6 gL of a saturated boric acid solution are added, and the mixture is placed on the filament and dried at 1.1A. When the spot has dried, the filament is heated to redness for 3 seconds. Prior to analysis the samples are out-gassed in the mass spectrometer by sequentially taking the filaments to a temperature of ~ 1200~ over a two-minute period for the phosphoric acid technique. Our experience is that little sample material is lost from the filament during this short out-gassing procedure. A 30 minute warm-up period is generally used for the boric acid technique for Cr. Ion beams typically ini-
646
Chapter 29 - T.M. Johnson & T.D. Bullen
tiate at ~ 1200~ for Cr and -- 1300~ for Fe. The filaments are allowed to sit at this condition for - 10 minutes. Filament currents are then increased until the ion beam intensity begins to increase without further current increase. Analysis is initiated when the major ion beam (either 56Fe or 52Cr) reaches an intensity of 0.5 x 10-11A. Generally the ion beam intensity continues to increase to > 10-nA for both Fe and Cr, then steadies for at least I hour. Data is collected as scans using both simultaneous measurement and peak-hopping capabilities of the TIMS instrument. Each scan consists of four progressive subscans: 1) the 57Fe (or 53Cr) ion beam is centered and measured simultaneously with the 56Fe and 58Fe (or 52Cr and 54Cr) ion beams; 2) the 56Fe (or 52Cr) ion beam is then centered and measured simultaneously with the 54Fe (or 50Cr) ion beam; 3) for Fe, the 52Cr ion beam, if present, is then centered and measured on the ion counting channel; and 4) the 60Ni (or 56Fe) ion beam, if present, is then centered and measured on the ion counting channel. The process is then repeated, and data is accumulated in blocks consisting of 10 scans. Measurement of baseline counts is performed between each block. If significant interferences are recognized: 1) 54Fe is corrected for 54Cr contribution using 54Cr/52Cr - 0.0282; 2) 58Fe is corrected for 58Ni contribution using 58Ni/ 60Ni = 2.616; and 3) 54Cr is corrected for 54Fe contribution using 54Fe/56Fe = 0.06368. The data are immediately processed using the spike subtraction algorithm described above; analysis continues until the calculated 56Fe/54Fe ratio of the sample is consistent to within 0.3%o for five blocks. Using this approach, the standard deviation of our reported values for Fe and Cr isotope compositions is at maximum 0.25%o and generally is significantly better. We have continually recognized significant probable organic interferences within both the Fe and Cr mass spectra. The organic component in the sample load is apparently highly recalcitrant and persists through both H202 and HC104 treatment during sample preparation. The interferences occur for both natural samples and standards, and thus probably derives from the exchange resins used for sample or colloidal silica purification. Other researchers have noted this previously (Anbar et al., 1997). We are currently exploring ultra-violet radiation techniques to eradicate the organics prior to sample loading. There have been informal reports of suspected organic contamination on the Re filament material, though this should not persist through the standard outgassing procedures used for the filaments. Regardless, the interferences decrease to negligible levels over time periods ranging from a few minutes to in some cases hours. Because the reduced data fail to converge until the interferences become negligible, the rule of isotope ratio consistency for at least five data blocks as outlined above is strictly adhered to.
29.5 Standards and analytical precision 29.5.1 Selenium
For Se isotope analyses, we use a provisional internal standard, MH495, for the zero point on the delta scale. An interlaboratory standard has not yet been established, but we are in the process of developing one through the U.S. National Institute of Standards and Technology (NIST). Although the Canyon Diablo Troilite (CDT)
Selenium, Iron and Chromium Stable Isotope Ratio Measurements ...
647
standard was used for some past Se isotope work (Krouse & Thode, 1962), we chose a synthetic standard because of the scarcity and potential heterogeneity of the CDT material. Based on variability in our standard analyses (more than half of these processed through the hydride generation systems) and duplicate analyses from laboratory experiments and field samples, we estimate our long-term measurement reproducibility (95% confidence) to be + 0.2%0 in ~80/76Se (Johnson et al., 1999). The Se double spike routine successfully corrects for large variations in instrumental mass discrimination. Uncorrected 80Se/76Se ratios drift several per mil during most analyses, but the final natural isotope ratios do not correlate with discrimination. Similar precision was attained in a recent study of Ca isotope ratios using a double spike (Skulan et al., 1997). Uncertainty of ~80/78Se is somewhat greater than that of ~80/76Se, and that of ~580/77Se is approximately + 1%o, much greater than that of 680/76Se. Apparently, an isobaric organic fragment interferes with 77Se, even though little interference is observed on 76Se under most circumstances. A similar problem seems to exist to a much smaller extent with 78Se. Precision is approximately + 0.2%0 on g80/78Se, despite the greater abundance of 78Se relative to 76Se. A very small interference of this type may be present at 76 a.m.u, and may be the limiting factor on improving precision at this point. Because the isotope ratios are not all measured at the same time, errors are introduced by changes in discrimination between the four steps of the measurement cycle. However, use of the measurement sequence given in Table 29.2 eliminates most of this error when discrimination changes linearly with time. Thus a steady filament temperature should be maintained during data collection to minimize non-linear discrimination drift, and precision is degraded if changes in instrumental discrimination are highly non-linear with time (e.g., after abrupt changes in filament temperature). The mass of Se recovered in blank solutions processed through the hydride generation procedure is approximately 0.2% of the mass of the previously processed sample, for both the batch reactor and the continuous flow system. The Se blank did not vary in response to changes in the amounts of HC1 or NaBH4 solution used. Sample carryover appears to be the source of the blank, in the form of a recalcitrant SeO precipitate that survives the washing steps. This level of cross contamination is insignificant if successive samples contain roughly the same mass of Se; sample aliquot sizes are chosen to ensure this is the case. When the carryover must be eliminated, overnight soaking of the apparatus in concentrated HNO3 reduces the carryover by at least a factor of ten. Se contamination on filaments loaded with Ba(OH)2 and graphite, and Ba(OH)2 alone, was minimal, producing Se- ion beams 104 smaller than the smallest sample beams. 29.5.2 Iron
An international standard, IRMM-14, is available from the JRC Institute for Reference Materials and Measurements (IRMM), located in Geel, Belgium. For a natural rock-matrix standard we use Fe separated from the U.S.G.S. rock standard BIR-1, an
648
Chapter 29 - T.M. Johnson & T.D. Bullen
Icelandic basalt. Our logic is that mantle-derived basalts provide the most isotopically homogeneous pool of Fe available for sampling at the earth's surface, as long as only fresh material is sampled. This pool should provide an inexhaustible supply of standard material. A variety of industrially-purified Fe solutions are likewise available for use as laboratory standard materials. However, we point out that industrial purification procedures may potentially fractionate the Fe isotopes to a certain extent. For example, fractionation of ~ 10 per mil from the near-uniform natural composition has been recognized in industrially-purified Ca (Russell et al., 1978). On the other hand, we routinely analyze a purified ICP-MS standard Fe solution and obtain the same Fe isotope composition as that of BIR-1. Likewise, Beard & Johnson (1999) report that their long-term average measured isotope composition of an industrially-purified Fe standard is identical to that of lunar and terrestrial igneous rocks which represent a very homogeneous Fe isotope pool (to which BIR-1 belongs). We feel that one advantage to using BIR-1 as the laboratory standard is that the basalt is a chemically complex material that must be purified prior to mass spectrometric analysis, and thus provides us with a monitor of the effectiveness of the purification procedure. As described in the section on Fe and Cr mass spectrometry, individual analyses of standards and samples have uncertainties of + 0.2%o in 656Fe (95% confidence; 2 times the standard error of at least 5 data blocks). Reproducibility of total procedural replicates is consistent with this; agreement is usually within + 0.1%o and almost always within 0.2%o. For example, our long-term average value of 656Fe for BIR-1 is - 0.03%o + 0.16%o (2o, n - 35) relative to our assumed isotopic composition of "natural" Fe. Similarly, our long-term average value of 656Fe for an ICP-MS high-purity Fe standard is + 0.05%o + 0.20%o (2o, n - 20). Blank levels on total procedural preparations are typically on the order of 1-5 nanograms, and are negligible. 29.5.3 Chromium
For our laboratory standard we use the National Institute of Standards and Technology (N.I.S.T.) standard reference material 979, a chromium chloride. As mentioned above, the certified isotopic composition of this material (Shields et al., 1966) provides the values we use for our initial estimate of "natural" Cr in the double-spike subtraction routine. As with Fe, we could likewise use a Cr-bearing natural material such as basalt BIR-1 as the laboratory standard. Based on 3 analyses of the Cr isotopic composition of BIR-1 using our technique, we find that Cr in BIR-1 is isotopically indistinguishable from that in N.I.S.T. 979. Using BIR-1 or a similar natural material as the laboratory standard would provide a monitor of the effectiveness of the Cr purification procedure. As with Fe, individual analyses of standards and samples yield uncertainties of + 0.2%o in 653Cr (95% confidence) based on variation among at least 5 accumulated data blocks. Our long-term average value of 653Cr for N.I.S.T. 979 is + 0.04%o + 0.20%o (2o, n - 10) relative to our assumed isotopic composition of "natural" Cr. Similarly; the average value of 653Cr of Cr purified from basalt BIR-1 is + 0.05%o + 0.18%o (2o, n - 3). Blank levels on total procedural preparations are typically on the order of a I - 5 nanograms, and are negligible.
649
Selenium, Iron and Chromium Stable Isotope Ratio Measurements ...
29.6 Current knowledge of Se, Fe, and Cr isotope systematics 29.6.1 Selenium
Several laboratory studies completed over many years of sporadic work with the gas source method (Krouse & Thode, 1962; Rashid & Krouse, 1985; Rashid et al., 1978; Rees & Thode, 1966; Webster, 1972) combined with our recent work (Herbel et al., 2000; Johnson et al., 1999) provide a reconnaissance understanding of Se isotope systematics. Se isotopic fractionations accompanying various biogeochemical transformations are summarized in Table 29.3. The most important conclusion is that the dominant isotope fractionating processes are selenate and selenite reduction. Thus, if Se isotope fractionation is observed in nature, it provides strong evidence that reduction is occurring. We envision applications similar to those developed for sulfate and nitrate reduction in groundwater (B6ttcher et al., 1990; McMahon & Bohlke, 1996; Strebel et al., 1990) and in other settings (Br~ichert & Pratt, 1996; Dickman & Thode, 1990; Kaplan et al., 1963; Zaback et al., 1993). Alternatively, isotopic tracing studies could be useful in settings where isotopically distinct sources of Se must be distinguished. 29.6.2 Iron
To date, several studies using either a double-spiking TIMS technique or multi-collector ICP-MS analysis have been completed that shed some light on Fe isotope systematics. Two field studies have provided an initial assessment of Fe isotope systematics at pristine and contaminated sites. At a ferrihydrite-precipitating spring in New Zealand (Bullen et al., 2001), Fe isotope evolution of spring water follows a Rayleigh fractionation trend, indicating an instantaneous fractionation of approxiTable 29.3 - Instantaneous kinetic isotopic fractionation, e, caused by Se transformations. Reference
Tr ans fo rmati on
Reacting A gent
Ellis et al. (2003) Ellis et al. (2003) Herbel et al. (1998) Herbel et al. (1998) Rees & Thode (1966) Krouse & Thode (1962), Rees & Thode (1966), Rashid & Krouse (1985) Johnson (in press) Johnson et al. (1999), Zawislanski & Zavarin (1996) Johnson et al. (1999) Herbel (1997) Hagiwara (2000) Johnson et al. (1999) Johnson et al. (1999)
Se(VI) Se(IV) Se(VI) Se(IV) Se(VI) Se(IV)
Sediment Slurry Sediment Slurry Bacterial Cultures Bacterial Cultures HC1, 25~ NH2OH or Ascorbic acid
-2.8 -6.5 -0.9 to-4.8 -7.3 to -8.3 -12 -7 to-13
Se(IV) to Se(VI) Se(0) oxidation
Alkaline H202 Incubated Soil
<0.5 <0.5
Se(IV) adsorption Plant uptake Algal uptake Se volatilization Se volatilization
Fe(OH)3 on H20 Barley C. reinhardtii Algal Culture Soil (Microbes)
<0.5 <1%o ca.-1%o <1.1 <0.6
to to to to to to
Se(IV) Se(0) Se(IV) Se(0) Se(IV) Se(0)
* Some values are converted from 82Se/76Se in this column.
e(880/ 76Se,%o)*
650
Chapter 29 - T.M. Johnson & T.D. Bullen
mately 1%o. This fractionation accompanies the oxidation of dissolved Fe(II) and abiotic precipitation of isotopically-heavy ferrihydrite. Four bacterial experiments have been completed, three using single microbe cultures and one using a naturally-occurring microbial consortium cultured from an aquifer sediment. In one experiment (Beard et al., 1999), dissimilatory reduction of Fe(III) in a ferrihydrite substrate by Shewanella algae induced a 1.3%o fractionation, resulting in isotopically light dissolved Fe(II). In another experiment (Brantley et al., 2001), microbial dissolution of a hornblende substrate by Arthrobacter sp. and Streptomyces sp. induced a 1%o fractionation, again resulting in isotopically light dissolved Fe(II). In a third experiment (Mandernack et al., 1999), intra-cellular formation of the mineral magnetite by two strains of magnetotactic bacteria (Magnetospirrilum magnetotacticum strain MS-1 and strain MV1) induced essentially no fractionation. In a fourth experiment, dissimilatory reduction of Fe(III) in an amorphous Fe gel substrate by a natural microbial consortium induced a 3%0 fractionation, again resulting in isotopically light dissolved Fe(II). Finally, one abiotic experiment (Anbar et al., 2000) demonstrated that Fe isotopes can be fractionated on an ion exchange column, an observation which has obvious implications for a variety of natural exchange processes. Based on these few studies, it appears that fractionation occurs during reduction and oxidation of Fe, and during microbial and abiotic processes. 29.6.3 Chromium
Work on mass-dependent fractionation of Cr isotopes is in a very early stage. One study measuring several groundwater samples found differences of several per mil in 53Cr/52Cr (Ball, 1996), but did not use a double spike calibration. Attempts were made to control instrumental discrimination through consistent operating procedures, but the study concluded with concern over the success of this approach. More recently, we have completed a few experiments in which Cr(VI) reduction induced 53Cr/52Cr fractionation of about 3.5%o with magnetite and two different sediment slurries as the reducing agents (Ellis et al., 2002). This establishes that chromium isotope fractionation is large enough to be useful, and that it should serve as a good indicator of Cr(VI) reduction in groundwater. This reduction reaction is if critical importance, as Cr(VI) reduction decreases the mobility and toxicity of Cr in the environment. In fact, Cr(VI) reduction via injection of reductants into aquifers is a common remediation technique. 29.7 Future direction
MC-ICP-MS An important future direction in mass spectrometry of Se, Fe, and Cr, is multi-collector inductively coupled plasma- mass spectrometry (MC-ICP-MS) as described by Rehk~imper et al. (Chapter 31). The detector ends of these instruments are similar to TIMS instruments, but the ions are generated in a radio frequency radiation-induced plasma. This very recent technique has a number of advantages over TIMS techniques, and may eventually dominate. Samples are introduced into the plasma as solutions. This eliminates the need to dry samples and load them onto filaments, and allows faster changeover from one sample to the next. For some elements, MC-ICPMS provides greater ionization efficiency and thus smaller sample requirements.
Selenium, Iron and ChromiumStable Isotope Ratio Measurements ...
651
With Se, high ionization efficiency can be achieved by direct injection of H2Se created by hydride generation into the plasma (Rouxel et al., 2002). Instrumental discrimination is very large with MC-ICP-MS and must be monitored and modeled correctly. The double spike routines presented in this chapter may thus be carried over into the MC-ICP-MS techniques. However, rapid sample-standard comparison is an option with these instruments, though careful matching of sample and standard matrices is critical. Recent work on Se isotope measurement has used this approach with great success (Rouxel et al, 2002). Another promising approach is the "elemental spike" technique, which has produced Fe isotope measurements of precision similar to that of the double spike TIMS technique (Anbar et al., 2000). A spike containing 63Cu and 65Cu is used to monitor instrumental discrimination much as the double spike is used in our measurements. Plasma-type ion sources create polyatomic ions that can greatly interfere with isotope ratio measurements. For example, Ar2 + and other Ar-based interferences are common because the plasma is generated in Ar gas, and all of the Se isotopes, with exception of 82Se, are interfered with by these species. Fe and Cr and similarly interfered with by polyatomic ions combining Ar, O, and N. In recent work (Anbar et al., 2000), these interferences have been rendered sufficiently constant and adequately characterized such that high precision is achievable. Recently, "reaction cells" that decrease polyatomic interferences by interaction of the ions with He or H2 have been invented, and have already proven highly effective with the Se interferences (Rouxel et al., 2002).
Handbook of Stable Isotope Analytical Techniques, Volume 1 P.A. de Groot (Editor) 9 2004 Elsevier B.V. All fights reserved.
CHAPTER 30 S I M S M e a s u r e m e n t of Stable Isotopes Trevor R. Ireland Research School of Earth Sciences, The Australian National University, Canberra ACT 0200, Australia e-mail: [email protected]
Abstract Secondary Ion Mass Spectrometry uses a primary ion beam to sputter a solid sample and produce secondary ions. SIMS offers an in situ analytical capability with spatial resolution of down to 50 nm for imaging and around 10-30/~m for high precision isotopic analysis. Isotopic compositions can be measured from the secondary ions of a variety of elements including those of the common stable isotopes (H, C, N, O, S) and rock-forming elements (Li, B, Mg, Si,). Electropositive elements (metals) are best analyzed as positive ions from negative oxygen primary beams, whereas electronegative elements (non-metals) are sputtered with positively charged cesium and have higher negative ion yields. For negative ion analysis, sample charging of insulating targets must be controlled by introduction of electrons to the sputter site. Recent advances in ion microprobe design have allowed in situ negative ion measurements at the same level of precision as for positive ions. Measurement errors down to 0.1%o are achievable for major elements, but at this stage, the external reproducibility for ion probe analysis is of the order of 0.5 permil because of variability in instrumentally induced fractionation. The goal of current work is to see this external reproducibility improve to a level commensurate with the internal measurement errors. SIMS is best used when in situ analysis of small domains in a heterogeneous target is required. 30.1 Introduction Secondary ion mass spectrometry (SIMS) is a versatile technique for analyzing solid materials. A primary ion beam sputters or erodes a sample causing the emission of ionic species from the target. These secondary ions are transported through a mass spectrometer where the mass separated beams are measured. In the geosciences, SIMS is accomplished with ion microprobes or ion microscopes which use focused primary ion beams either in static or scanning mode to analyze samples. A typical analysis consumes only a few nanograms of material leading to the labeling of this technique as virtually non-destructive. Stable isotope analysis requires some special conditions owing to the nature of information that geochemists are trying to obtain. In particular, geological materials require primary beam (spot) resolution of the order of 10 gm and in situ analysis is
SIMS Measurement of Stable Isotopes
653
preferred. This is the basis for ion microprobe mass spectrometry, a static spot allowing measurement of isotopic abundances in a given volume. With such a small volume of material, obtaining high precision is not always possible. However, the strength of ion microprobe analysis is not in the ultimate precision that can be obtained on any given spot, but rather the ability to assess heterogeneity of samples on a microscale. If large variations are present in a sample, then ion microprobe analysis offers petrographic control of measurements that is unsurpassed by any other technique. Furthermore, ion imaging with spot resolutions down to 50 nm of such a sample may give a visual display of isotopic distributions. Stable isotope analysis is often thought of as the analysis of O, C, H, and N because it is these elements that have proven to be highly versatile in characterizing biological and geological processes. These elements are also important in stable isotope research by ion microprobe, but isotopic analysis of any element can provide important information regarding formation conditions such as temperatures, chemical species of reactants, reaction pathways, and so forth. The limitations for stable isotope analysis on the ion microprobe are governed typically by the amount of material available for analysis, the concentration of the target element, and the ion yield of that element under the chosen analytical conditions. Intrinsic to ion microprobe analysis is measurement of isotope abundances and a very wide range of elements can be analyzed. In general, electropositive elements (metals of the periodic table) can be analyzed as positive ions from a negative O ion beam, whereas electronegative elements (nonmetals) ionize best as negative secondary ions from positive Cs-ion bombardment.The relative ionization efficiencies for different elements are shown in Figure 30.1. A great deal of stable-isotope technique development has been associated with analysis of extraterrestrial materials. Such samples can have natural advantages over terrestrial materials; the effects that are being sought are often large and analytical precision is therefore not so much an issue. Also, non-mass-dependent effects are common, such as the 4 % 160 anomaly in refractory inclusions (e.g. McKeegan et al. 1998) and so limitations imposed from the determination of instrumental massdependent fractionation are not so stringent. The drive for higher precision and accuracy comes from the requirements of terrestrial geochemists who are used to the high precision of bulk methods. Unfortunately, measurement of stable isotope compositions in terrestrial samples is more than obtaining intense secondary ion beams. Mass fractionation produced during sputtering and within the ion microprobe must be well characterized and stabilized if routine determinations of intrinsic fractionation of stable isotopes are to be made. SIMS isotope measurements are always referenced to standards and the reproducibility of the standard is paramount in accurate stable isotope measurements. Ion microprobes are complicated instruments with a multitude of lenses, slits, defining apertures and other components that are used to define the trajectories of the ion beams. As such it is difficult to give specific recipes for conditions of use. The following discussion is offered if only to bring to mind some preferred options and potential pitfalls, but SIMS is a highly versatile technique and different targets may
654
Chapter 30- T.R. Ireland
Figure 30.1 - Relative secondary ionization efficiencies for different elements. Plotted are the relative ion yields for (a) positive ions relative to Si + (O-primary), (b) negative ions relative to Si(Cs + primary), and (c) M - / M + for negative ions normalized to Si-/Si + = 1.0. For positive secondary ions, the ionization efficiency is largely a function of the filling of the outer electron shell i.e. the most electropositive elements have the highest ionization efficiencies. The converse is true for negative ions with the most electronegative showing the highest yields albeit with some exceptions (for example, note the anomalously low ionization yield of N-). The most common stable isotopes are typically best measured as negative ions. Of the light elements commonly measured, H, Li and B are analyzed as positive ions. Data from Wilson (1995) for Si metal matrix. Note that Na + ionization approaches 100 % under Obombardment (Hinton, 1995) and allows an indication of the relative ionization efficiency for other positive secondary ion species. Similarly, F- and C1- ionization yields approach 100% under Cs + bombardment (ASU webs i t e - see Table 30.1).
require completely different conditions to those offered below. Over the past 10 years there have been substantial developments in stable isotope analysis by ion microprobe. In particular, measurement of negative ions from insulators is now possible through charge neutralization by way of electron injection schemes. This chapter covers general features of stable isotope analysis by SIMS. A more detailed review of historical developments in terms of geological analysis can be found in Shimizu & Hart (1982a), Zinner (1989), Ireland (1995), Hinton (1995) while instrumental aspects are covered in more detail by Benninghoven et al. (1987). Besides the primary references cited, additional useful source material and discussion should be consulted from the compilations of Harmon & Hinton (1992) and McKibben & Shanks (1998).
SIMS Measurementof Stable Isotopes
655
30.2 Secondary Ion Mass Spectrometry- SIMS A primary ion beam is focused to erode, or sputter, the target. The energy of the primary ions (typically 10-20 keV) causes a fraction of the atoms and molecules to become ionized. These secondary ions are electrostatically transferred into a mass spectrometer where they are separated according to mass and energy before being passed into a detector. The principles governing secondary ionization are not well understood, and it has not been possible to derive a quantitative model that predicts the abundances of the secondary ion species relative to the actual concentrations in the matrix. For stable isotope analysis, this is not a fundamental limitation. Of primary concern is that a stable secondary ion beam can be generated and that the composition of the secondary ion beam reflects the composition of the target. Of the sputtered material, most of it is neutral atoms and molecular fragments, and a fraction that depends on the element and the matrix (as in Figure 30.1), is ionized. A characteristic feature of SIMS is that the observed isotope ratios of the secondary ion beam is enriched in the light isotopes relative to that of the target. Both experimental and theoretical considerations suggest that this observed enrichment is a result of more efficient momentum transfer through lighter atoms (and molecules) so that they are more likely to become ionized. Alternatively, mass fractionation could be produced by neutralization processes where the heavier, slower ions have a greater probability of neutralization (for further discussion see Williams, 1979, 1982). The exact site of ionization is not clear. However, observations such as the matrix dependence of isotope fractionation as well as inter-element ionization variability, suggest that it must occur while the target can have an influence over the ionization probability. The accurate assessment of this instrumentally induced fractionation is essential to quantitative stable isotope measurements with an ion microprobe.
30.2.1 Ion microprobes An ion microprobe consists of a primary column, to generate, accelerate and focus the primary beam to a spot on the target, a source chamber, where the primary beam interacts with the target and the secondary ion beam is formed, and the mass analyzer where mass and energy separation of the secondary beam is produced and the massseparated ion beams are measured (Figure 30.2). Ion microprobes require ultrahigh vacuum (_<10-8mbar) to reduce the effects of gas scattering on the ion beams and limit neutralization through charge exchange. In some cases deposition on the sample from the residual gas is also of concern, especially if the sputter rate is very low. 30.2.2 Primary beam considerations The primary ion beam is typically composed of O or Cs because these elements produce a significant enhancement in the ionization yields of electropositive and electronegative species, respectively. Thus, most lithophile metal elements are best analyzed with an oxygen primary beam, while the halogens and other electronegative species are better analyzed with Cs. Cesium can only be utilized as Cs +, while oxygen can either be O- or O +. Of further consideration is the choice of polarity for the pri-
656 Chapter 30- T.R. Ireland
Figure 30.2 - Schematic Cameca ims 3f ion microscope (adapted from Ireland, 1995).
SIMS Measurementof Stable Isotopes
657
mary and secondary extraction fields. In general, the secondary extraction voltage is given the opposite polarity to the primary. Selection of like voltages requires the banking of power supplies such that +10 kV primary and +10 kV secondary would require a potential difference of 20 kV to real ground at some position in the ion probe. As such, the most common electrical configurations used are O- primary with positive secondary ion extraction and Cs + primary with negative secondary extraction. Accelerating voltages for primary and secondary ion beams are typically 10-20 kV and 5-10 kV respectively. The oxygen primary beam is typically generated in a duoplasmatron discharge. The most common type of duoplasmatron consists of a cold hollow cathode (cold indicating it is heated by the plasma, rather than a hot filament as occasionally used), an intermediate electrode for geometrical confinement, a surrounding magnetic field for magnetic confinement, and the anode extraction aperture where the ion beam is formed. The axial beam formed in the discharge is dominated by electrons with an annulus of O-ions. Thus the extraction aperture is offset from the axial position for extraction of the O- beam. This is not the case for an O + beam because the electrons are repelled from the extraction aperture and so the extraction aperture is axial for O + beams. The negative plasma consists predominantly of O- and 0 2 - w i t h a higher proportion of the atomic species than the molecular (O-/O2-~ 3 - 10). The atomic species is commonly used but the advantage of the molecule is that double the mass is delivered to the sample per unit charge, thus increasing the ionization yield per nanoamp of primary. Hinton (1995) has demonstrated that 02- generates almost exactly a factor of two higher yield than O-. There is also less potential for charge build up on the target when using the 02- ion to generate the secondary ions. In addition, it is possible that matrix effects on the secondary ion beam are ameliorated by using the heavier species. However, sufficient 02- must be generated to produce the required secondary ion beam and if high yields are required the molecular beam may not be an option. It should be noted that the optimal tuning for the atomic species is not necessarily the best for the molecular ion beam and some degree of retuning can improve 02- delivery relative to O-. The Cs beam is generated in a thermal ion source. A reservoir of Cs (either as metal or Cs-bearing compound) is heated to cause vaporization of Cs metal. The Cs vapor is then passed through a super heated frit causing ionization of Cs. While the use of Cs metal produces a lot of Cs vapor for the ion source, and hence brighter ion beams, the handling of Cs metal is rather onerous. Once the Cs metal is exposed to the vacuum, it must be maintained in vacuum until exhaustion. Cs compounds such as chromate, carbonate, or silicate have the benefit of being inert in air and so once the source is cooled it can be safely vented to air. The primary ion beam is accelerated away from the ion source and shaped with various ion optical elements. A standard feature on most ion probes is a mass filter, either a sector magnet or a Wien filter, that allows a specific species such as O- or 02to be selected, or for contaminants to be eliminated from the beam. The main contam-
658
Chapter 30- T.R. Ireland
inants to the beams are hydrides, the 16OH- peak is prevalent after servicing the duoplasmatron, so a mass resolution of ca. 32 is advisable to be able to separate 0 2 - f r o m O2H- if 02- is to be used. Introduction of H to the sputter site is undesirable because of the additional complication to the secondary ion speciation. Other minor contaminants (e.g. hydrocarbons, metal ions from the cathode discharge) will also be removed at this mass separation. Cameca ion microscopes generally have both duoplasmatron and Cs source permanently mounted on the primary column and so the primary magnet allows the selection of ion beams from the two alternate sources. Along with computer controlled primary and secondary voltage systems, it allows rapid switching between the different polarities and settings. A Wien filter is used on the SHRIMP primary column because the duoplasmatron and Cs gun are interchangeable and no ion-source switching capability is required. The Wien filter is therefore a more effective option when no primary deflection is required. Both systems perform equally well in terms of the requirements of the respective instruments. m
m
The Wien filter uses crossed electrostatic (E) and magnetic (B) fields such that only an ion with a specific velocity v will be transmitted undeflected when v = E x B. The Wien filter operates as a mass filter because the velocity of an ion passed through an electric potential V is mass dependent, i.e. qV- 1/2mv2, where q is the fundamental charge (for singly charged species), V is the extraction potential, m is the mass of the ion and v is its velocity. The primary column is generally at an angle of 20-45 ~ away from normal where it has been shown that there is optimal secondary ion emission for normal (i.e. to the sample surface) secondary-ion extraction. As such ion probe spots are typically ellipsoid although these can be shaped back to circular images if required. An exception is the Cameca nanoSIMS where the primary ion delivery and secondary ion extraction axes are coaxial and normal to the surface. This is one of the parameters that enables the extremely small spots (order 50 nm) on the nanoSIMS to be obtained coupled with high sensitivity. Spot size plays a fundamental role in experimental design on an ion microprobe. Spot size can be as small as the 50 nm spots on the on the Cameca nanoSIMS, or as large as possible to extract the most secondary ions. Spot size is generally selected on the basis of the smallest domains to be analysed with the tradeoff of the reduction of primary beam current with spot size. The smallest spot on the Cameca nanoSIMS has an intensity of only 10 pA and so the total secondary ion yield will be far less than when using a 30 ~m spot with 10 nA. Fewer ions means lower precision and so there is a tradeoff in the intensity of the primary beam to be used and the required precision. On the other end of the scale, the largest spot that can be used will be governed by the refocusing ability of the primary column, but also the useful transmission of the secondary extraction system. If the secondary extraction system cannot accept the
SIMS Measurement of Stable Isotopes
659
ions from the larger diameter, there is no point in operating with a larger spot. In effect, a useful upper limit to spot size is around 30 gm for isotope analysis. This holds for negative as well as positive primary ion species. Even for negative ions, the spot will charge, affecting the secondary ion characteristics and in turn, instrumental fractionation stability. Theoretical counting statistics can be calculated from the number of atoms in the target, and this can then be translated into a volume, and a depth for a given spot diameter. As an example, to achieve 0.1%o counting precision for oxygen isotopes in quartz (SiO2) assuming a 10 % O- ion yield, a volume of 10 gm 3 of SiO2 is required. For a 10 gm diameter spot, this requires a spot depth of approximately 0.1 gm, but for a 1 gm spot the depth would need to be 10 gm. The most stable and best analytical conditions are obtained when the spot depth is substantially smaller than the spot diameter. Consequently there is a fundamental limitation on the use of the smallest spots depending on the precision required. For a 1%o oxygen isotope analysis, a I gm spot with depth of 0.15 gm is sufficient. These are optimal calculations for a major element and the required sample could be substantially larger depending on the elemental concentration and other conditions. The nature of the focusing of the primary ion beam on to the surface can have an effect on the composition of the secondary ion beam. A common method of beam transport is to produce a spot on the target that is the demagnified image of the anode extraction aperture (for the case of an oxygen beam from a duoplasmatron). However, such an image is often less than perfect with the summing of aberrations produced through the various lens elements. If the final spot on the target is Gaussian (in the extreme), then there is differential sputtering across the spot from the center to the rim. Such behavior is of obvious detriment if depth resolution is required, but could also affect the stability of instrumentally induced isotope mass fractionation. In order to ameliorate differential sputtering across a static spot, a method analogous to the optical microscopy method of Kohler illumination (Liebl, 1983) can be used. The specific analog is that the defining aperture is placed at the focal length of the final lens of the primary column. This configuration theoretically also removes the effects of any upstream aberrations and the final spot on the target is simply the demagnified image of the defining aperture (Figure 30.3). In a modified approach, the aperture can be placed at a position intermediate between single and double focal length depending on the desired (de)magnification, but the illumination at this aperture must be uniform. In practice even the true Kohler method is improved by having a uniform intensity distribution at the aperture. The drawback to Kohler illumination is that it is best used with quite large spots (of order 15-30 gm). If high spatial resolution is required there may be no option but to use highly demagnified spots. In this case, Fitzsimons et al. (2000) show that measurement uncertainties approaching counting statistics can be obtained provided standards and unknowns are measured under identical conditions. Even illumination can also achieved by rastering of a finely focused beam over a given area. This method
660
Chapter 30 - T.R. Ireland
Figure 30.3 - The primary ion spot allows the selection of sites for analysis that are free of contamination through inclusions, cracks, or other possible sources of isotopic heterogeneity. Importantly, a well-defined primary spot with even illumination is essential for maintaining stable isotopic mass fractionation during analysis. Plotted are 634S/32S isotopic compositions in a heterogeneous sulfide.
also offers the possibility of electronic gating where only a selected part of the area is accepted. This method is widely used in depth profiling methodologies. The impact of the primary-ion beam causes energy and charge transfer to the sample. If the net flow of charge is not dealt with, the spot can charge up, the sample potential can change, and the secondary ion beam may no longer be focused through the mass analyzer. For analysis with O- primary ion beams (and positive extraction), the resultant charge build up can be quite satisfactorily handled with the use of a thin conductive layer (C, Au, Au-Pd, etc) which allows the dissipation of electrons. This enables the potential of the sample to remain uniform and hence a stable extraction field is maintained. The extraction field geometry is also improved by the use of wellpolished samples. Under such conditions stable secondary ion signals can be produced that last several hours. Furthermore, stability of isotopic mass fractionation can be achieved over reasonable analytical times as well. In the circumstance of a long
SIMS Measurement of Stable Isotopes
661
analysis, some charge build up may be produced because of progressive removal of the conductive coating in proximity to the spot. The charge build up can be monitored by measuring the energy distribution, and a suitable offset on the sample accelerating voltage or steering can be applied such that the energy distribution passes through the same trajectory through the mass analyzer. In this case, compensation for movement of the energy distribution over a few tens of volts can be accomplished satisfactorily. With non-normal primary beam incidence, the ion probe crater floor will not only move progressively into the sample, it will move across it. Thus, in the extreme, the floor of the crater could be optically masked from the extraction system, and the maximum ion transmission will likely fall out of optimization. Furthermore, the crater walls will likely charge up, also changing secondary ion tuning conditions. For these reasons, the optimal configuration for an ion microprobe spot is for it to be relatively large and the analysis time (i.e. penetration depth) limited such that crater geometry is not an influence on the effective composition of the measured secondary ions. In general, with 15-30 gm spots and analysis times of less than ca. 20 minutes, this will not be an issue (c. 2 gm depth for 5 nA). However for very small spot sizes (order 1 gm or less) the penetration depth in 20 minutes could indeed exceed the hole diameter. For the Cameca nanoSIMS, the primary beam incidence is normal to the sample surface and so tuning conditions will not change geometrically to the same degree. At this stage there is no published information on high precision isotopic analyses with the nanoSIMS. The use of the Cs + primary beam with negative secondary ions presents the greatest analytical difficulties. The impact of Cs + at the sputter site requires neutralization. At low Cs currents, there may be a significant degree of self-neutralization, but in most cases for stable isotope analysis on non-conductors, insufficient electrons can get to the sputter site from the conductive coating. Furthermore, the extraction system responsible for acceleration of negative secondary ions also removes secondary electrons from the sputter site, further exacerbating the charge build-up. The resulting charge effectively changes the potential of the crater and so the tuning through the mass analyzer degrades quickly, especially if energy dispersion is limited through defining slits or apertures. This problem is particularly important here, because stable isotope measurements of O, C, and S are best performed with negative secondary ion beams. McKeegan (1987) ameliorated charging by taking fragments of interplanetary dust particles and pressing them into gold foil. For fragments smaller than ca. 15 gm, there was only a minimal charging affect and stable negative secondary ion beams could be obtained for oxygen isotopic analysis. However, such a method is not desirable where petrographic constraints need to be placed on the material analyzed, which is of course the strength of in situ analysis. The alternative approach developed has been to deliver electrons to the sputtering region so that the positive charge is balanced by the incoming electrons. The most
662
Chapter 30- T.R. Ireland
straightforward approach instrumentally is to inject electrons in a similar fashion to the p r i m a r y beam, that is, an electron gun is aimed at the sample directly (Figure 30.4). The electrons must have sufficient kinetic energy to traverse the secondary accelerating potential, between the sample and the extraction lens, and arrive at the sputter site. If the electrons have insufficient energy, they cannot reach the sample, and the b e a m is simply deflected away from the sample. The impact of these electrons produces secondary electrons, the n u m b e r of secondary electrons p r o d u c e d being d e p e n d e n t on the material and the energy of the impacting electrons. At low energies,
Figure 30.4 - Schematic of secondary-ion extraction region for high-energy charge compensation used in SHRIMP ion microprobe. The electron beam is directly focused on to the sample to overlap with the primary beam. Owing to the non-normal incidence and the extraction field, the electron beam is deflected through the extraction region as it approaches the sample. The electron source must be held at sufficiently high potential to overcome the extraction gap potential which in this case for the SHRIMP extraction geometry is only 700 V. High-energy charge compensation was originally developed on a Cameca ims-3f (Hervig et al., 1992) where the extraction gap potential difference is much higher.
SIMS Measurement of Stable Isotopes
663
there is insufficient energy to knock a secondary electron out, and at high energies, the electrons penetrate the target also resulting in fewer secondary electrons escaping than primary electrons arriving. But at intermediate energies, more secondary electrons are produced than there are primary electrons being delivered to the target and as such the electron beam cannot neutralize the charge build up on the target. This situation is also affected by the composition of the conductive coating material. Gold is a very strong secondary electron emitter, but C coatings have very low secondary electron emission characteristics and so should be used for this methodology. This method does not necessarily compensate for the charge build-up on the spot. More likely it swamps the sample with electrons, with excess electrons being dissipated through the conductive coating. As such, the sample potential is not self regulating and the use of a fixed energy window, for example, could effect mass fractionation. The use of high-energy electrons can also cause deterioration of the sample through heating. Hervig (1992) developed this technique on a Cameca ims-3f. For a 2-3 nA Cs § primary beam, a high-voltage, high-current electron gun was used to neutralize the charge. The filament was held at-7.5 keV with 10 gA of electrons striking the target with a net energy of approximately 3.4 keV. Under these conditions, stable secondary ion beams could be obtained but Hervig (1992) resorted to energy filtering in an attempt to better stabilize instrumental mass fractionation (see below). The problems with a high-energy electron gun can be ameliorated through instrumental designs that include a low secondary-ion extraction field (Isolab 54, SHRIMP) such that the impact energy of the electrons can be reduced. Lyon et al. (1995) describe the use of the Isolab 54 for stable isotope measurements with high-energy charge compensation; Cs § beams of 0.1 to 5 nA could be used producing up to 5 x 10 -11 A of 160-. Saxton et al. (1996) document the difficulties of focusing the electron beam through the extraction potential, with resultant electron beams 100 microns wide by several hundred microns long. On SHRIMP instruments, the electron gun floats on the sample potential and so the electrons only see a 700 V or so potential difference allowing well focused electron beams to strike the target. An alternative approach developed by Slodzian et al. (1986) for Cameca ion microscopes is to transport the electrons down the secondary extraction axis such that the electron beam is normally incident on the sample (Figure 30.5). The energy of the electrons can then be matched to the surface potential of the sample and because they have no horizontal momentum, the electrons essentially stall out in a cloud above the spot. Thus there is no impact with the target and no secondary electrons are produced. The build up of charge through Cs sputtering causes electrons to move to the spot and so a self-regulating charge-neutralization system can be achieved.
30.2.3 Secondary beam considerations The secondary ion beam is electrostatically accelerated and focused to the source (or entrance) slit of the mass analyzer. In ion microscopes, masking apertures can be used to select ions only coming from a particular location on the sample (for example to exclude the spot walls) or from a particular part of an ion optical crossover if aberrations are present. For the ion microprobe mode proper this is not possible because of
664
Chapter 30- T.R. Ireland
Figure 30.5 - Schematic of extraction region for low-energy charge compensation used on Cameca ion microscopes. The electron beam is inserted along the secondary ion axis with the use of an insertion magnet (A). The extraction energy of the electron source is adjusted such that the electrons stall out in a cloud just above the sample producing a self-regulating charge compensation. Any charge build-up will attract electrons from the cloud. Because the secondary ion beam is also deflected by the insertion magnet (A), its trajectory must be compensated by two correction electromagnets (B,C). Figure adapted from Cameca promotional materials (http: / / www.cameca.fr/).
astigmatic focusing to increase transmission. The secondary ion beam is then passed into the mass spectrometer.
30.2.4 Mass analyzer Geological materials are not chemically simple compounds, and SIMS produces a plethora of non-stoichiometric species. These two factors combine to produce complicated secondary ion mass spectra. The function of the mass analyzer is to produce sufficient mass dispersion to eliminate, to the highest order, all possible isobaric interferences from the isotopes of interest, and to produce ion beams suitable for highprecision measurement. There are a variety of mass spectrometer types (e.g. quadrupole, time-of-flight) that will achieve sufficient mass resolution, but magnetic sector instruments are preferred for isotope ratio measurements of high precision. The magnetic sector deflects ions through a specific radius depending on the ion's mass, charge, and energy. Because sputtering produces secondary ions with a variety
SIMS Measurement of Stable Isotopes
665
of energies, double-focusing mass spectrometers are required such that secondary ion beams of disparate energy can be refocused. The essential elements of a double-focusing mass analyzer are the electrostatic analyzer and the magnet. The ESA disperses the ions according to energy and the magnet refocuses them on the basis of the energy and mass. The net result is the cancellation of the energy term to yield a mass focus. The parameter most often sited in terms of mass analysis on ion microprobes is mass resolution. The American Society for Testing and Materials Standard Terminology (ASTMST) for Surface Analysis- SIMS defines mass resolution as M / D M where DM is the full width at half maximum of the peak at mass M, but in practice ion probe analysts use a mass resolution definition where DM is defined as the full peak width at the 10 % or 1% height of a peak (Figure 30.6a). This is essentially the width of the transferred source slit of the mass spectrometer (typically demagnified but with aberrations) convoluted with the collector slit width. Mass resolving power is frequently used interchangeably with mass resolution. The specific ASTMST definition of mass resolving power is the peak to valley ratio between adjacent equal sized peaks separated by one mass unit. It can also be formulated like the mass resolution calculation except DM is defined as only the magnified source slit image ( plus aberrations) and is represented by the 10-90 % rise on one side of the peak. Thus some care should be exercised in evaluating figures of merit concerning mass resolution and mass resolving power because of the different definitions. Note that the ASTMST definition of mass resolving power would be more readily equated with abundance sensitivity, the peak to tail signal strength at a distance of one mass unit. In practice, peak separation at full mass resolution may not be required if it can be shown that the interfering signal does not contribute to the measurement point of the mass of interest. If the interfering species is small, it may not contribute to the analysis even though a valley between the peaks is not defined (Figure 30.6b, c). If the interfering species is large with respect to the peak of interest, it may be fully resolved at the 10 % level, but could still contribute to the measurement. In this case higher mass resolution is required (Figure 30.6d). Isobaric interferences can be either atomic or molecular, but isotopes with atomic isobars are often avoided for isotope ratio work because they require extremely high mass resolving power. For most elements, atomic isobars can be avoided by selection of an alternative isotope. There are no atomic isobaric interferences of consequence for the common stable isotope systems. The molecular interference of most consequence for stable isotope analysis is the hydride. A proton has a mass excess relative to 12C of 0.0078 AMU. The isotopic mass deficits of elements with atomic mass less than 100 systematically increase a few mAMU per AMU and so the hydride appears as a heavier species than the atomic. For instance, 16OH (15.9949 + 1.0078 - 17.0027) is heavier than 170 (16.9991) by 3.6 mAMU and therefore requires a minimum of 17/ 0.0036 - 4,400 M/AM for resolution. However, the abundance of 16OH-relative to 170- often necessitates the use of higher mass resolution. Separation of 180 from 17OH only requires 2300 R, and the abundance of 17OH-is low compared to 180-. Such resolutions are well within the capabilities of most ion microprobes involved in the geosciences. Increased resolution is obtained by closing down slits which decreases trans-
666
Chapter 30- T.R. Ireland
,',all 100 %
90 %
. . . . .
,--
.:.- -
' ) 9
C.~,
2
-
,r ,--" '-
/
,!
.-"
r
1 C/'
1 0" l 0 "~
,,ii ~"
1 Cr' "1"Y I ,::.e
,,,r~1,_
!' o
I
t
1
,I
II
l
I
iI
(d)
r l-'
l O
1 0-'
ru1 ....-;,
3 _ ]
I,C
1
log 10(coun t mte) 6 t,41
c
0 ~;4
9
(b)
I
,45__ ,l'§
50
ll.'l
M
1 c.'.e 1 c)" I (F'
1 Cr~ 1 C-~
-
r.13
1 cY 1 Cr' I c:r' 1 c'r" 1 c::d
Figure 30.6 - Peak schematics showing mass resolution, mass resolving power, and estimation of the contribution to a peak from an interference. (a) The linear representation of peak shape allows an assessment of the degree of flat-top which is essential for determination of best-focus for the mass spectrometer. Under these conditions the peak shape is trapezoidal. The collector slit width can be determined at the FWHM through the mass dispersion formula for the mass analyzer. Mass resolution, in this case defined at the 10% level (AM1) is given by M/AM1. This is essentially the source-slit image-width at the collector (s) added to the collector-slit width (c). Mass-resolving power is often cited for mass spectrometers. This is commonly formulated as the ratio of M/AM2 where AM2 is the demagnified source slit image defined at the 90 % - 10 % level of the slope of the peak. (b) The logarithmic representation of the peak shape allows the degree of contribution from interferences to be ascertained. In this case Peak M2 is not clearly resolved from Peak M3. (c) The overlay of a clean peak (M1) shows that M3 does not contribute to M2 above the 0.1 permil level even though a valley of resolution is not achieved between the peaks. The converse does not hold with M2 contributing to M3 at the permil level. This difference is due to the differing intensities of these peaks. (d) In order to effect complete resolution of M2 from M3 requires higher mass separation with resolution in excess of that calculated simply from the relative masses of the interferences.
SIMS Measurement of Stable Isotopes
667
mission and also increases the prospect for instrumental mass fractionation. As such, the measurement of 170 is likely to be at the expense of signal as well as reproducibility of fractionation measurements. Molecular interferences can also be excluded by the energy filtering method (Shimizu & Hart, 1982a; Zinner & Crozaz, 1986). This is based on the observation that polyatomic ions have a narrower energy distribution than atomic ions. Energy filtering is achieved by a suitable energy window at the energy focus of a double-focusing mass analyzer that selects high-energy ions. This is the preferred method for trace element analysis in the region of the rare earth elements because of unresolvable (in terms of mass) isobaric interferences. But, it is at the expense of a large fraction of the beam (generally greater than 90%) and this can hinder high precision isotopic measurements. However, there are benefits to energy filtering that offset the potential loss in sensitivity that will be discussed below. 30.2.5 Detection systems Ion beam measurement is fundamental to mass spectrometry, but ion microprobe measurements present conditions that are quite often at the edge of operation of the different systems used. While a stable isotope measurement in a gas source mass spectrometer can be tailored to give satisfactory signal strength, SIMS analysis has limitations associated with analytical conditions including concentration of the target element in the sample, primary beam strength, and mass resolution requirements. As such a variety of detection methods is required that cover many orders of magnitude of signal strengths. Until recently, ion microprobe measurements have been carried out on single electron multipliers operating in a pulse-counting mode. An incoming ion hits the first dynode and generates 1 or more electrons. The potential drop across the multiplier accelerates the electron(s) into succeeding dynodes typically with an amplification of two or more per stage. Hence the charge delivered by a single ion striking the first dynode can be amplified by factors of 107 or more. This charge burst is passed into the detection circuitry mainly consisting of an amplifier-discriminator and a counter. The discriminator is used to block zero-point noise produced by ambient electrons in the circuitry such that only pulses of a specific minimum height are counted as real events. A single event takes a finite amount of time to be processed through the counting system and during this time an incoming ion may not trigger the multiplier resulting in a lost count. This loss of counts is referred to as dead time and must be corrected in order for accurate analyses to be made. It can be shown that for a retriggerable counting system" Cmea s =
CtrueeTctrue
[30.1]
where Cmeas is the measured count rate, Ctrue is the actual number of events per second, and 9is the system dead time. The true count rate can then be expressed as"
668 Ctrue ~ Cmeas exp [Cmeas~ exp([Cmeas'~ exp([Cmeas~ exp(...) ) )]
Chapter 30- T.R. Ireland
[30.2]
and approximated by Ctrue ~ Cmeas exp [Cmeas~ exp([Cmeas~ exp([Cmeas~) )]
[30.3]
The typical dead time for the pulse-forming process in the multiplier is of the order of 5 ns in terms of full-width half maximum, although pulse pair resolution at the discrimination level is longer and the total system dead-time is longer still because of amplification and discrimination. Typical dead times for counting systems may range from 10 to 30 ns depending on the specific setup. For instance, a higher discriminator level can produce a shorter dead time, but at the expense of pulses under the discriminator level. If this is suspected, the efficiency of the multiplier can be obtained by comparing a suitable signal from the Faraday Cup with the multiplier response. The main limitation on precision of a pulse counting system is the total counts (N) on the minor isotope for which Poisson counting statistics dictates a variance of N, and hence the limitation on precision to be 1/~N. Thus 106 counts of the minor isotope will yield 1%o precision (lo), whereas 108 counts are required for 0.1%o precision. Clearly higher precision is preferable and hence the highest number of counts is required. But the nature of electron multipliers dictates that count rates be kept under ca. 2 MHz so that the electron multiplier life and stability is not compromised. Hayes & Schoeller (1977) have shown that there are limits imposed on an analysis due to the uncertainty in the dead-time. When ratios differ significantly from unity the uncertainty in the dead time can be the limiting factor. However, when only two peaks are being analyzed it can be argued that the dead time effects from the pulse counting system are the same for all analyses provided the count rates can be kept close, and thus are removed in the normalization process. The situation is slightly different when mass fractionation is monitored from one isotopic ratio and used to determine the residual in another isotope ratio(s). In this case, the uncertainty in the residual can be quite small and not limited by the standardization process and so the dead time must be accurately determined. This can be accomplished through analysis of standards at a variety of count rates. Where dead time variability is found, it is often associated with the muliplier gain being off the plateau. The gain of a multiplier slowly degrades with time and if the gain falls off the plateau rapid changes in gain are apparent, resulting in changes in efficiency and dead time. One methodology to remove uncertainty from the dead time is to gate the pulse counting system electronically such that it will not retrigger within a set period. This is generally accomplished by taking a gate period longer than the system dead time (i.e. 30-50 ns). The correction for dead time will be larger, but theoretically there is no uncertainty on the correction. Slodzian et al. (2001) note another phenomenon that is similar to the dead time effect in that it affects the high-abundance isotope preferentially. The Quasi Simulta-
SIMS Measurementof Stable Isotopes
669
neous Arrival (QSA) effect is related to the efficiency of secondary ion collection and conversion of primary ions to secondary ions. A high efficiency of production and collection will therefore mean a higher chance of a simultaneous arrival of two ions from a single incident primary ion. For instance with Si isotopes, there is a much higher probability of collecting two 28Si ions than there is of two 29Si atoms or two 30Si ions. Therefore QSA produces an undercount of the most abundant species and is independent of count rate and dead time. However, in most experimental setups where dead time is a variable, the effect of QSA will be masked by the dead time measurement, that is, the underproduction of 28Si will be compensated by a longer deadtime. This effect will however be very apparent when a gated dead time is used. For two isotope systems, the effect will be normalized through comparison of standards at given count rates. Pulse counting is the preferred method for low count rates up to approximately 1 MHz. In most instances, a 1 permil precision can be obtained in a straightforward manner, better than 0.7 permil is possible with care, but obtaining precisions at the 0.1 permil level, although theoretically possible (Slodzian et al. 2001), is simply not practical with an ion counter. Essentially only a few analyses could be achieved in a day thus inhibiting the determination of analytical reproducibility. The main limitation of pulse counting is the maximum allowable count rate. This is circumvented by the use of Faraday cup measurement. The secondary ion beam is captured in the cup and the current can be measured in an electrometer across a highresistance feedback resistor (typically 1010 - 1012 ~ ) . The impact of an ion directly transfers one charge unit (1.6 x 10-19 C) and so there is no problem associated with dead time. The limitation to Faraday cup measurement is the minimum current that can be collected. This is a result of Johnson noise from thermal electrons moving across a high impedance resistor. The current can be expressed as: INRMS-
~/4kTBw/R
[30.4]
where k is the Boltzmann constant, T is temperature (K), Bw is the frequency bandwidth of the noise to be examined, and R is the resistance of the circuit. At T = 300 K a frequency bandwidth of 0-10 Hz, and a resistance of 1011 ~, the Johnson noise (RMS) is 1.2 x 10-15 A or roughly 8,000 c/s. This is comparable to the analytical noise of commercial electrometers (Keithley TM Model 6517A is listed at 0.75fA). Thus a Faraday cup is not the optimal device for measuring low intensity secondary ion beam currents. There is no explicit upper limit to the current that can be collected on a Faraday cup, but potential problems can occur with the linearity of the system where intense ion beams are to be compared to weak ones. The dynamic range of the counting system can also limit the inherent precision possible as well. Typically the current arriving at the Faraday cup is translated into a pulse train where the frequency of the pulses is proportional to the frequency of arrival of secondary ions at the detector. A typical upper limit to the frequency of this
670
Chapter 30- T.R. Ireland
signal is 1 MHz for a full-scale deflection of the electrometer. If both large and small signals are collected on the same range, then the resolution of the smaller signal can be compromised, particularly for small beams such as 180 vs 160 which is a ratio of 1 / 500. If the 160 signal generates a full scale deflection = 1 MHz, the 180 signal will be 2000 c/s. If a precision of 1%0 is required, there is only a 2 c/s window left to ascertain precision. Clearly this situation can be ameliorated by changing the range of the electrometer, but the two ranges must be calibrated against one another, and another background measured, thus introducing other uncertainties. The minimum count rate that can be measured can be substantially improved if the feedback resistor is replaced with a capacitor. In this case, charge is collected on the capacitor and the Johnson noise across the resistor is not measured during the analysis. Rather, an analysis of signal strength consists of allowing the capacitor to charge up for a specified period of time and then determining the amount of charge that has accumulated on the capacitor. Measurements of this type have not been successfully carried out until recently on thermal ionization mass spectrometers (Esat, 1995) and background currents of 10-17 A may be possible. Such a technique may benefit data collection where signal strengths are intermediate between multiplier and Faraday cup. Often in stable isotope analysis, the situation arises where there is insufficient beam for the Faraday Cup on (one of) the minor isotope(s). For example, in oxygen isotope analysis, 160 might be readily measured on the Faraday cup, but not the 180. In this situation, the 180 can be measured on the ion counter in the same cycle with the 160 measurement on the Faraday cup. The relative sensitivity factor between the electron multiplier and Faraday cup can be ascertained through the measurement of the standards provided the count rates are kept uniform throughout the measurement period. Recently, multiple collectors have been added to large ion microprobes. A multiple collector has two or more multipliers or Faraday cups or a combination of the two that simultaneously detect the isotopes of interest. Multiple collection offers great advantages in ion microprobe analysis. The collection efficiency is improved because all peaks are being collected simultaneously. Temporal variations in the secondary ion signal strength due to primary beam instability or through heterogeneity of the target element in the sample do not affect the determination of an isotope ratio. However, multiple collection devices for most mass spectrometers have used Faraday Cups exclusively. Secondary ion signal strengths on ion microprobes are not always conducive to Faraday cup measurement and multiple collection with electron multipliers has presented a number of problems. The main issues have been the size of the multipliers and cross talk between multipliers causing erroneous signal detection. To address these issues, the Cameca 1270 utilizes very small discrete dynode electron multipliers (Hamamatsu Photonics K. K.) that are contained within the Faraday cups. The SHRIMP II multiple collector utilizes Sjuts TM channeltrons (continuous dynode electron multipliers) the design of which also limits electron cross talk. These detectors can be stacked at the very small spacing between adjacent Pb isotope peaks.
SIMS Measurement of Stable Isotopes
671
Channel plate detectors, such as used for direct ion imaging, can be used for isotope ratio measurement although they have low dynamic range, and unstable gain characteristics (see for example, Saxton et al., 1996).
30.3 Measurement of stable isotope ratios 30.3.1 Mass fractionation The goal of SIMS analysis is to obtain an isotopic measurement that can be related to the absolute composition of the target. The main process affecting this determination is mass fractionation, the dependence of the measured isotopic composition on the mass of the isotope. Isotopic mass fractionation is intrinsic to SIMS (see discussions by Shimizu & Hart, 1982a; Ireland, 1995). Thus a measurement requires the separation of the instrumental factors that might be variables during an analytical session, from the fractionation produced in the sputtering process, which must at some level be assumed to be constant for the course of the analytical session. The constancy of instrumental and sputtering fractionation (i.e. total measured fractionation) is generally taken as constancy for both components. For any given element with multi-isotopic composition, sputtering will give a lighter isotopic abundance than expected. In practice, the measured composition may be more variable owing to other instrumental parameters. The difference between the actual composition and the measured fractionation can be large, commonly percent level for low masses and tending to smaller fractionations for heavy isotopes. However, the measurement of a relative fractionation difference is not affected by the absolute fractionation i.e. it doesn't matter if there is 20 % fractionation relative to actual composition provided it doesn't change. The constancy of measured fractionations to a level concomitant with measurement errors indicates that relative fractionation measurements between standards and unknowns are viable and can give accurate results at the level of cited precision.
30.3.2 Isotope Analysis SIMS isotope measurements can be divided into two types. Where only two isotopes are measured, mass fractionation is calibrated to an external standard of the same mineralogical composition and known isotopic composition. In this case, mass fractionation cannot be known to a level better than that determined for the standards. This is the methodology commonly used for stable isotope measurements of C, O, and S where the mass dependent fractionation effects are sought as indicators of physical processes. Where three or more isotopes are measured, one of the ratios can be used for an internal calibration such that the deviations of the third isotope ratio from a mass fractionation law can be used to derive a residual. The feature of this method is that an analysis can continue provided the internal reference ratio stays within limits prescribed by experimental determination. Analytical uncertainty in the normalized ratios can often be limited by counting statistics. This method can be used to deter-
672
Chapter 30- T.R. Ireland
mine mass independent effects such as the abundance of a radiogenic decay product, or mass independent fractionation such as that observed in 33S. The fractionation law is generally derived empirically, but based on the measured isotopic composition of an element over a range of mass dependent fractionation. The simplest of these laws, the linear law, simply expresses the mass fractionation law as a linear dependence of the mass differences. For example, the 33S/32S deviation will be half that of 34S/32S. Other normalization laws are used that are based on more elaborate formulations e.g. Rayleigh, power, exponential laws. Esat (1984) has shown that these laws all have the same basic form. If we define the ratios Rik and Rjk, where Rik = iA / kA and Rjk = jA / kA, then power law: Rayleigh law: exponential law:
Rik- (Rjk) (mi-mk)/mj-mk) Rik- (Rjk) (Vrmi-V'mk)/~/mj-V~mk) Rik- (Rjk) log(mi/mk)/log(mj/mk)
[30.5a] [30.5b] [30.5c]
from which the generalized form of the fractionation laws can be seen to be Rik = [Rjk]~ [30.6] For example, the ~, values for the Mg isotopes (24Mg, 25Mg, 26Mg) a r e 1power = 1.996, ~/Rayleigh = 1.976, and yexponential - 1.957. Therefore the different laws can be seen to be simply imparting different degrees of curvature to the mass fractionation function for any given element. In practice, the real functional relationship must be determined for each element on each mass spectrometer. If an element in a sample experienced complete ionization, rather than having only a fraction of the sputtered atoms ionized, there would be no fractionation associated with the ionization process (although fractionation in the mass spectrometer is also present). Elements that experience high degrees of ionization to positive species (e.g. Ca, Mg) have relatively low degrees of isotopic mass fractionation. A relationship is apparent between an element's first ionization potential (for positive ions) and the degree of isotopic fractionation as measured on an ion microprobe. But measured mass fractionation is quite variable owing to a number of parameters and generalizations concerning ion yields and fractionation have many exceptions. Indeed, the observation between ionization and mass fractionation is not so apparent for negative ion production on the basis of electron affinity. The real problems associated with SIMS stable isotope analysis are in achieving stable analytical conditions for sustained periods of time such that the uncertainty of a single analysis is consistent with the reproducibility of a population of analyses of that same material (assuming a constant composition of the target) (Fitzsimons et al., 2000). Or at least, the reproducibility of the standards has to be at a level commensurate with the required precision. The goal is to achieve a constant measured isotopic mass fractionation in the standard [that doesn't seem to be enough, one also has to be sure that it is the same on
SIMS Measurement of Stable Isotopes
673
standard and unknown. This is the analytical requirement; note also mineralogical differences in standard and unknown, standard heterogeneity, will also affect the measurements], and hence from the knowledge of the isotopic composition of the standard, calculate the instrumental mass fractionation component that can then be subtracted from the unknown. A typical analytical session would consist of a series of measurements of the standards (std) and unknowns (unk) comprising std - u n k - std or std - u n k - u n k - std such that there is at least one standard in close proximity to the unknown. The std measurements can be assessed for internal consistency (e.g. temporal changes of composition) as well as consistency over the whole analytical session (drift, reproducibility). If the std analyses are consistent, the mean fractionation can be used to calculate the instrumental fractionation and can be removed from the individual unk analyses. The errors of the means from both standard and unknown must be combined suitably because the absolute fractionation for the unk cannot be known any better than the standard. Such a treatment still only involves random errors from measurements and does not address the possibility of systematic errors, such as from matrix effects, or from intrinsic variability of the standard itself. A potential systematic effect may be induced by having standards and unknowns in different plugs, or even different mounts or otherwise mounted in dissimilar fashion. Geometric effects may play a role in the instrumental mass fractionation, such as mounting of grains too close to a discontinuity (holding ring, gap between plug and holder), and these might affect the extraction field geometry, as could different conductivities of coatings on std and unk. Uncertainties for stable isotope measurements are usually limited by external effects, not internal measurement errors. It could be argued then that there is no point in reducing measurement errors significantly below reproducibility if the parameters affecting reproducibility are not going to be addressed. Thus time spent on obtaining internal measurement errors of 0.1 permil is wasted if the reproducibility is limited to 1 permil. But, for advancement of the analytical technique to higher levels of overall precision, the ability to obtain measurement errors significantly below reproducibility is required if the factors governing the limitations are to be addressed. 30.3.3 Matrix effects
The matrix is possibly the most important determinant in the relative isotopic mass fractionation of secondary ions from a given element given stable analytical conditions. The chemistry of a specific mineral is known to produce changes in ionization yields as indicated by correction schemes that are related to chemistry (Deloule et al., 1991). Structure can also be important, for example, different polymorphs of Ti oxides give different Ti isotopic mass fractionation factors that are correlated with the Ti-O bond strength (Ireland, 1986). In most cases, the matrix effect is probably a combination of chemistry and structure (mineralogy) and so suitable standards must be used for calibrating mass fractionation. It appears however, that such relative mass fractionation effects can be calibrated over a variety of minerals and only one phase need be analyzed during the analytical period, although some error propagation would be necessary.
674
Chapter 30- T.R. Ireland
These effects indicate that the target can play an important role in the measured isotopic mass fractionation and so beam overlap .onto surrounding mineral species must be treated with caution. However, Riciputi & Greenwood (1998) have documented S and C isotope ratios in synthetic mixed matrices and have found that the presence of another mineral phase does not influence the mass fractionation from a given mineral species provided that element is not present in the other phase. Poorly defined matrices where an inappropriate instrumental mass fractionation factor is applied can also result in residual artifacts. Eiler et al. (1997) have conducted an extensive study of matrix effects during O isotope analysis covering 40 silicate and phosphate minerals and glasses. Trends and correlations of fractionation were found for mean atomic mass of the target, chemistry of specific cations (e.g. Mg/Mg+Fe in olivine, Na/Na+K in feldspathic glasses, structural differences between minerals and glasses of the same composition, and sputter rate in silicate glasses. Correlations are sufficient to allow correction for different suites of particular type, but Eiler et al. (op. cit.) conclude that mineral standards close to the composition of the unknowns must still be characterized and measured. An important parameter in spot locations must be the complete characterization of samples prior to analysis and also after analysis. Electron-beam imagery (back-scattered electron, secondary electron, and cathodoluminescence) provides important information for assessing spot placement. Cracks, mineralogical imperfections (such as exsolutions, alteration), small inclusions, can all affect the measured isotopic mass fractionation and should be avoided. 30.3.4 Factors affecting instrumental fractionation In general application to stable isotopes, the object is to maintain the measured fractionation of the standards (hence instrumental fractionation) as constant as possible such that any deviations for unknowns can be reliably construed to be intrinsic fractionation in the unknown. In order to achieve this constancy, it is worth noting some of the instrumental parameters that can affect the measured fractionation. Then, it is worth avoiding changing those parameters that cause great change in fractionation for a small change in the given parameter. Ion microprobes are complicated instruments with many operational variables all of which may have the potential for changing the instrumental fractionation. For each isotope of a single element the method of analysis and instrument setup can change the absolute number of ions that are formed, the proportion that are transmitted through the mass spectrometer, and the relative proportion that are counted in the detector. Implicit in this, is an appreciation that different isotopes have different mass, and possibly different energy distributions, and that unintentional physical separation on the basis of mass and/or energy occurs within all instruments.
The primary beam affects the absolute fractionation because it is responsible for the sputtering event. Thus it is clearly worthwhile having a well-controlled primary beam. The stability of the primary beam is important because any noise on the primary beam will be immediately transferred to the secondary beam as well in terms of
SIMS Measurementof Stable Isotopes
675
intensity fluctuations. While it might be argued that a multiple collector would remove all these instabilities, there is a tendency for a poorly performing primary beam to also affect mass fractionation. That is, a fluctuating beam does not give fractionation measurements that are consistent with the expected measurement errors. The primary beam spot itself is also important. The spot is essentially the manifestation of the ion beam density within the primary column as projected on to the sample surface. The primary spot should be well focused with no internal aberrations that might form transient sputtering loci. In this regard, the Kohler-illumination method (Liebl, 1983) affords the best stability in terms of continuous sputtering. For routine ion microprobe analysis, best (i.e. most reproducible) results are obtained from a ca. 15-25 gm spot. However, stable isotope analysis is not limited to such spot sizes and analysis can be carried out to much smaller spots provided due care is exercised. Once the primary ion beam is setup for an analytical session, it is generally not changed and so the sputtering fractionation might be regarded as constant. During the course of the analytical session, a variety of secondary ion extraction and shaping lenses, as well as steering plates may be changed, apertures inserted and removed, etc. Thus, it is the control of the secondary ion system that provides for the most factors affecting measured fractionation. If the secondary ion beam is truncated anywhere, it can cause fractionation in the isotope ratios. This is particularly the case if there is isotopic heterogeneity in the distribution of the ion beam paths in the secondary ion beam, which is almost certainly the case. Generally this means the instrument must be reasonably tuned for analysis, and be reproducibly tuned in the same way for all analyses. Shimizu & Hart (1982b) reported a number of ion probe parameters that affect isotopic mass fractionation. These observations still hold for the operation of the small Cameca ion microscopes. For example, they demonstrated the necessity of proper and stable alignment of the primary beam spot with the secondary ion extraction axis to achieve reproducible isotopic mass fractionation measurements. The measured fractionation is a function of the energy of the secondary ions and is thus affected by the width and position of the energy slit. Essentially, accurate mass fractionation measurements require that the analytical conditions remain as fixed as possible during the analytical session. The maxim for stable isotope measurements is beam truncation produces fractionation. The corollary is therefore that if transmission is maximized and beam truncation minimized, possible instrumental fractionation is minimized. Thus apertures and closed slits should be avoided if at all possible. For the large ion microprobes, this condition is automatically ameliorated because narrowly closed slits are not required to obtain the mass resolution required for stable isotope analysis. A further source of analytical variability lies in the methods for charge compensation/ neutralization from Cs + sputtering of insulators. These methods can change the sample potential and in doing so can affect the tuning and by extension instrumental mass fractionation. The general philosophy is to achieve equilibrium conditions for a
676
Chapter 30 - T.R. Ireland
sample such that instrumental effects are stabilized but under certain conditions the charge compensation can be unstable due to instabilities in the electron beam, ripple on the electron gun potential, or poor focusing of the electron gun.
30.3.5 Specific operating conditions There are a number of laboratory factors that can have a significant impact on data collected. For example, of oft cited concern is the need for stable temperature in the laboratory and while not quite as disturbing, stable humidity as well. These parameters can upset Hall effect probes for magnet calibration, and cause drifts in electronics and high voltage power supplies. Obviously, if the lab is not in order, the data will not be either. In the first instance, the mass spectrometer is operated to exclude all unwanted species. This is generally accomplished by sufficient closure of source and collector slits to produce the requisite mass resolution. The objective then is to produce a stable analytical configuration. High mass-resolution is generally preferred to energy filtering in isotopic analysis because of the higher transmission and hence better precision that is ultimately available. However, isotopic mass fractionation is found to be quite dependent on the energy of the secondary ions selected. The form of the relationship is that for low secondary ion energies there is a steep slope to the function whereas at higher energies the relationship flattens out (Figure 30.7). For example, in the case of low-energy secondary ions illustrated in Figure 30.7, a 10 eV change in the sample potential could cause a 7 %o shift in instrumental fractionation, while measurements taken at very high energy offsets should show a lower degree of dependence on variability in secondary ion energy. However, the instrumental mass fractionation is substantially higher for the high-energy offsets. Riciputi (1996) has evaluated the two techniques (energy filtering vs. high mass resolution) for S-isotopic analysis in conductive phases on a Cameca ims-4f and determined that energy-filtering is favored in terms of precision and accuracy for a given analysis on this instrument. In particular, energy filtering produced more consistency for instrumental mass fractionation between mounts and this is important if the standard cannot be loaded onto a particular mount. It is worth noting that in order to obtain high mass resolution conditions in the ims-4L substantial masking of the secondary ion beam is required at the entrance slit as has already been indicated. As such, the low-resolution conditions and the selection of high-energy ions is a better configuration in terms of instrumental mass fractionation reproducibility on the small Cameca instruments. However, for large sector instruments (Cameca 1270, SHRIMP), the source slit can be operated wider open for a given mass resolution and therefore the proportion of the secondary beam stopped by the slit is low and hence the possibility of instrument induced fractionation lower. The energy slit performs several functions in the setup of a double focusing mass spectrometer. For high mass resolution operation on the Cameca ion microscopes, an
SIMS Measurement of Stable Isotopes
677
Figure 30.7 - The measured instrumental mass fractionation (IMF) for O isotopes is a strong function of secondary ion energy. For secondary ions measured close to the extraction potential (i.e. zero initial kinetic energy), fractionation changes rapidly with sample charging requiring well-constrained analytical conditions. At higher energies, there is less dependence so analytical uncertainties from variation in sample potential are less critical. As illustrated, the mean of thirteen analyses with initial kinetic energy ~_350 eV gives a statistically coherent mean. However, the use of the high-energy ions is at the expense of a large proportion of signal. Data from Hervig et al. (1992).
energy window is defined (e.g. 30 eV) to facilitate obtaining high mass resolution in the case of both extreme energy filtering (e.g. Eiler et al., 1997; Riciputi & Greenwood, 1998) and low-initial-energy analysis e.g. (e.g. Greenwood et al., 2000). Interestingly; the SHRIMP ion microprobes do not utilize the slit in this way and obtain satisfactory refocusing with no energy window selected. It is not clear whether this relates to the nature of the ion optics of the mass analyzer (tunability) or if there is an energy selection produced at, for example, the source slit of the SHRIMP instruments. The energy window is however, useful on all instruments for the selection of a specific energy of ions to be analyzed such as in energy filtering and high-energy offset for stable isotope measurements. If a specific energy window is specified, the sample voltage needs to be maintained to a high degree to ensure stable instrumental mass fractionation. Recent results clearly indicate that this can be achieved (see examples below). If an electron multiplier is used as the detector, its condition can have a substantial effect on the data (Slodzian et al., 2001). Eiler et al. (1997) report that drift in isotopic mass fractionation was associated with changes in multiplier detection efficiency and when the gain of the multiplier was recalibrated after a session showing drift, the
678
Chapter 30- T.R. Ireland
mass fractionation data were stable. The detection efficiency is a function of both gain of the multiplier and discrimination of the counting system. The fact that the multiplier itself can induce shifts in isotope ratios indicates that it is a source of variability and so a stable configuration must be established. Often, the count rate of a particular isotope is initially specified (e.g. 500,000 c/s) to avoid problems with dead time. However, this alone may not be sufficient to guarantee data is free of counting artifacts. The multiplier should be clearly on the gain plateau of the multiplier (count rate vs. voltage), and a valley established in the discrimination of counts from the background noise (grass) in the counting system. In addition, in setting the count rate for the start of an analysis, it should be established that the parameter being changed does not itself induce a change in instrumental mass fractionation. For example, changing the entrance slit setting is commonly used to control the secondary ion count rate in dead time measurements. Under certain conditions this could have an effect on isotopic mass fractionation because of the trimming of the beam. 30.3.6 Instrumentation
The general features of ion microprobes have been covered above. Following are specific instrumental features that can aid in stable isotope measurements on individual types of instrument. Analytical protocols are usually established for a specific instrument, but the caveats that hold for any given laboratory are still relevant for the design of new techniques in any laboratory. That said, specific conditions that exist on one type of instrument may not be apparent on another. URL locations for manufacturers and some SIMS laboratories carrying out stable isotope measurements are given in Table 30.1. Cameca ion microprobes are by far the most common instrument used for stableisotope analysis. The small Cameca ims-nf (where n=3,4,5,6) have proved to be reliable and capable instruments despite the narrow entrance and exit slits required for isotopic analysis and concomitant loss of transmission. The Cameca ims-nf operate as stigmatic ion microscopes and allowance is made for masking apertures to be positioned to exclude undesirable parts of the ion beam. In this way, for example, residual surface water sputtered from the edge of the spot can be excluded such that only secondary ions from the center of the crater are passed, thus minimizing the OH- interference in O isotope measurements. The ion microscope mode can also aid in spot selection in terms of avoiding areas of non-optimal behavior (cracks, inclusions, etc). Possibly the most important feature related to stable isotope research is the ability, introduced in the Cameca ims 4f, to produce stable negative secondary ion beams through the normal-incidence electron gun/charge compensation system. This allows stable secondary ion beams of electronegative elements to be produced from insulating samples. The Cameca ims 1270 uses the same extraction system (i.e. primary column and source chamber) as the ims-nf models, but it has a large mass analyzer allowing high transmission (high sensitivity) analysis at high mass resolution, and has wide dispersion allowing multiple collection. The Cameca 1270 can operate either in a microscope mode (i.e. stigmatic with direct ion imaging) or a microprobe mode where the second-
SIMS Measurement of Stable Isotopes
679
Table 30.1 - URL for SIMS Laboratories Manufacturers
Cameca (France) http://www.cameca.fr Australian Scientific Instruments http://www.anutech.com.au / asi / Users
Edinburgh University - Cameca ires 4f http: / / www.glg.ed.ac.uk / research / ionprobe / ionprobe.htm CRPG, Nancy, France - Cameca ires 3f, Cameca 1270 http://www.crpg.cnrs-nancy.fr / Sonde / intro-sonde.html Arizona State University - Cameca ims 3f, Cameca ims 6f h t t p : / / w w w . a s u . e d u / clas / csss / SIMS / University of California, Los Angeles - Cameca 1270 http: / / oro.ess.ucla.edu / ionprobe / home.html Washington University in St Louis - Cameca ims 3f, Cameca nanoSIMS http: / / presolar.wustl.edu / work / instrumentation.html The Australian National University - SHRIMP I, II, RG http://shrimp.anu.edu.au Stanford University- SHRIMP RG http://shrimprg.stanford.edu / Manchester University - VG Isolab 54 http: / / www.man.ac.uk / Geology / research / isotopes / instrum / ionprobe.html
ary optics are retuned to maximize transmission (astigmatic- high transmission mode). The benefit of the high transmission is that there is little truncation of the secondary ion beam and sufficient beam is available for Faraday cup measurement (McKeegan et al., 1998; Mojzsis et al., 2001). Multiple collection with Faraday cups allows high precision analysis of O and S in a matter of minutes (examples below). Cameca has recently commissioned their nanoSIMS instrument. This differs from previous Cameca instruments in having a Mattauch-Herzog geometry that offers refocusing over a wide mass range. The feature of this instrument is the ability to produce extremely small primary beam spots of order 50 nm. This is accomplished by having the primary beam normally incident on the sample surface, that is, bringing the primary beam down through the secondary extraction optics. This allows the immersion lens to be placed very close to the sample surface allowing high degrees of demagnification and also improves secondary ion extraction efficiency. SHRIMP ion microprobes have been somewhat neglectful of stable isotope analysis although some of the first reliable S isotope results were produced on SHRIMP I (Eldridge et al. 1987). Despite the lack of application, there are a number of features that make SHRIMP well suited to stable isotope analysis. One of the main features is a shallow-depth-of-field optical microscope that essentially places the sample into the same position with respect to primary and secondary ion-optical axes each time the sample is brought into focus. Such a simple operation therefore affords great reproducibility in terms of analytical conditions. The potential between the sample and first
680
Chapter 30- T.R. Ireland
extraction plate is only a few hundred volts and so perturbation of potential field lines has less of an effect on the total extraction system than with the Cameca extraction geometry. SHRIMP typically operates with no energy filtering while still maintaining a high mass resolution capability. Thus, there is less opportunity for truncation at the energy slit and hence influence on isotopic mass fractionation. For charge compensation, a high-energy electron gun is used with impact energies of 1-1.5 keV. The Isolab 54 is another ion microprobe, similar in concept to SHRIMP, with a much lower extraction field than the Cameca source geometry (England et al., 1992; Saxton et al., 1996). For stable isotope measurements (e.g. Lyon et al., 1995), a high energy electron gun is used for charge compensation, and a multiple collector is used for ion detection incorporating Faraday cup for the light isotope (e.g. 160 ) and conversion dynode/channel plates for the heavier isotopes (e.g. 170 and 180). This system requires extended deadtime for the conversion dynode and channel plate to avoid double pulsing.
30.4 Recent examples of stable isotope measurements Ireland (1995) included some details of specific ion microprobe analytical arrangements for stable isotope configurations. Much has changed in the past few years such that capabilities that were cutting edge at the time of that review are now commonplace. 30.4.1 Oxygen Oxygen is perhaps the quintessential element for stable isotope analysis in geochemistry. Oxygen is a difficult element to analyze because of the large difference between the abundance of 160 and the minor isotopes 170 and 180 and because of the large 16OH interference on 170. In extraterrestrial O-isotope analysis it is essential to measure all three O isotopes whereas in terrestrial applications measurement of the 1 8 0 / 1 6 0 will suffice because the main parameter of interest is mass-dependent fractionation. At mass 18, the interferences of 17OH and 16OH2 are smaller than 180 and readily resolved at around 2,300 R. However, the hydroxide species are sputtered not only from the primary spot but also the area surrounding the spot, causing degradation of the OH- peak shape. Tailing of the OH- species is therefore a potential problem and should be monitored during the course of analysis. The OH- is apparently due to surface water migrating on the sample; water is a residual to the vacuum and the 16OH-/170- falls with time after sample insertion and hence pumping time. Hence methods for excluding water from the analysis are required. For example, water can be minimized through a variety of sample preparation methods such as extended pumping in a sample lock prior to insertion, or heating, sample under vacuum conditions. During analysis, the edge of the spot can be masked on Cameca ion microscopes to accept only the center of the spot where water contamination is minimal. O is best analyzed as a negative ion through Cs + bombardment. Other schemes involving F- primary ion beams have been tried but have rather grave consequences for the duoplasmatron (Hervig, 1992). As discussed above, the main problem with
SIMS Measurement of Stable Isotopes
681
Cs + bombardment is the charge build up on the surface of the sample. Electrons can be delivered with either low or high energies, and similarly secondary ions can be selected for either low or high energies. As an example, Graham et al. (1996) used Cameca ims 4f with a Cs + primary beam (1-4 nA) defocused to 20-30 gm on goldcoated samples with a net impact energy of 14: kV. The normal-incidence electron gun was used for charge compensation with compensation indicated by adjusting the electron flux to yield zero net sample current. Secondary ions of 325-375 eV were extracted with count rates kept close to 5 x 105 for 160-. McKeegan and coworkers (e.g. McKeegan et al., 1998; Leshin et al., 1997; etc.) have developed techniques on the Cameca 1270 to use lower-energy secondary ions for stable isotope analysis. This methodology requires that the instrument be kept in a very stable configuration in order to avoid instrumental fractionation variability caused by the secondary ion energy dependence. The Cameca 1270 can operate at high mass resolution without a severe loss of sensitivity. Leshin et al. (1997) used a I nA Cs + beam defocused to 25-30 gm. Secondary ions within a 95-120 eV window were selected and the mass resolution was set at ~ 6100 M/AM, sufficient to completely resolve any hydride contribution of 16OH- from 170-. Secondary ions were measured on an electron multiplier. Ion probe data were compared with conventional fluorination data on mineral separates and showed good agreement to within a permil or so following matrix corrections of < 1%o. The main limitation of electron-multiplier measurements is the total number of counts that can be collected during the analysis time. Leshin et al. (1998) and Mahon et al. (1998) used ions with initial kinetic energies of 0-30 eV to increase count rates and allow a mixed detection mode for the analysis of carbonates whereby 160- was measured on a Faraday cup (3 sec integration time) and 180- was measured on the electron multiplier (5 secs per integration). In this configuration, 180- count rates up to 250,000 c/s could be utilized enabling internal precisions of 0.5 %o for 180/160. At the relatively low count rates for 180- dead time corrections are not problematical. The Faraday cup measurements must be calibrated to the electron multiplier, which requires an additional measurement for accurate calibration. Once calibrated, any variability in the detector efficiencies will be reflected in a change in the measured fractionation. Engrand et al. (1999) also measured 170- o n the electron multiplier in a study of Antarctic micrometeorites. Quite recently, the Cameca 1270 multiple collector has been used. Mojzsis et al. (2001) analyzed zircons from 4.3 billion year old zircons from Western Australia. Both 180 and 160 w e r e collected on two Faraday cups at effective count rates of 2 GHz for 160- and 4 MHz for 180- allowing measurement precisions of 0.1 to 0.2 permil (2o) to be obtained in less than 6 minutes. These data allow an evaluation of the SIMS O isotope technique for analyses at subpermil level. Mojzsis et al. (2001) analyzed two standard materials, 91500 (6180 - 9.8 + 0.2 %0) and KIM5 (5.04 + 0.05 %o) with 15-18 standard analyses presented for each of the two
682
Chapter 30- T.R. Ireland
sessions. Analytical uncertainties are better in the first session than in the second with mean uncertainties of 0.30 %o and 0.12 %o for 91500 and 0.20 %o and 0.12 %o for KIM5 respectively. The scatter of the data as determined from MSWD is 3.9 and 4.1 for 91500, and 24 and 8 for KIM5. Even with a higher analytical uncertainty in the first session, the MSWD values for the standards are no better in the first session than in the second. The scatter for 91500 is better than that for KIM5 and suggests that 91500 is better behaved than KIM5 at the microscale. If the excess scatter is assumed to be a Gaussian source of noise (errors inflated by V'MSWD), the mean analytical uncertainties (per spot) are increased to 0.60 %0 and 0.24 %0 for 91500, and 0.94 %o and 0.33 %0 for KIM5 (lo). The errors in the means of the analyses are 0.10 %o and 0.05 %0 for 91500, and 0.13 %0 and 0.07 %o for KIM5. This treatment assumes all data are equal and no outliers have been excluded. In summary, the reproducibility of the standard data indicates a source of external variability is present. Even allowing for that variability, individual spots can be determined with precisions down to 0.4 %o (2o) and errors in the mean down to 0.1%o (2o). These data suggest that 0.1%o data are possible, but sufficient standards must be analyzed to ascertain the level of variability beyond that predicted from the measurement errors. In the above example, KIM5 can be examined relative to 91500 to assess accuracy, i.e. we can use 91500 as the standard and examine the KIM5 composition. Using the augmented errors, the mean compositions for 91500 are 8.31 + 0.10 %0 and 7.99 + 0.05 %0, and for KIM5 are 4.01 + 0.13 %0 and 3.60 + 0.07 %0 (lo). These data are offset from the absolute compositions noted above, but if the 91500 data is shifted back to 6180 of 9.80 %0 for each session, the compositions for KIM5 are corrected to 5.50 + 0.33 %0 and 5.42 + 0.09 %o. The analyses for KIM5 agree well, but the composition is elevated above the 5.04 + 0.05 %o reported as the KIM5 composition. This can be interpreted several ways but at this level of precision it is difficult to find a definitive interpretation. Possibly both of the compositions of the standards at the microscale do not agree with the bulk analyses or there are subtle heterogeneities within these samples. Reference to additional standard materials would be useful. The Jack Hills zircons show a range of compositions and also show some variability within grains. The data presented by Moizis et al. (2001) represents data combined for individual spots and grains. Analyses with higher uncertainties are quite likely due to sample heterogeneity (zones of metamict zircon for instance). The measured 6180 of the zircons ranged from 5.4: %o to 7.7 %o for optically continuous grains thus stressing the requirement of high precision for single measurements as well as high spatial resolution for these measurements. For comparison, similar systematics were determined by Wilde et al. (2001) in a similar study but in this case they used energy filtering (offset of 350 eV) and a total of 2 x 106 counts of 180 were counted to yield precision close to 0.7 permil (lo). In this case the oxygen isotopic composition of the oldest zircon (at 4.4 Gyr) was resolved from the oxygen isotopic composition of zircons from direct mantle-derived rock by approximately 3o (Figure 30.8). A number of the zircons with anomalously high 6180 are characterized by relatively young U-Pb ages (as opposed to Pb-Pb ages) and suggest that the zircons have exchanged with meteoric water, removing Pb and disturbing the O isotopes.
SIMS Measurement of Stable Isotopes
683
F i g u r e 3 0 . 8 - Oxygen isotopic compositions in the world's oldest zircons. The solid symbols are data from Mojzsis et al. (2001) while the open symbols are from Wilde et al. (2001). (a) The zircons show a wide range in oxygen isotopic composition with 207Pb/206Pb age but all are above 6180 of +5 %o, the maximum level expected for derivation of magma from a primary mantle source. Rather these compositions are indicative of being sourced from material that has undergone low-temperature interaction with a liquid hydrosphere. (b) Plotted against 206Pb / 238U, the heaviest zircons (in terms of 6180) show the highest levels of discordance suggesting that late stage alteration is responsible for the more extreme O isotopic compositions. Nevertheless, the most concordant zircons still lie above the limit for a mantle source.
684
Chapter 30- T.R. Ireland
Figure 30.9 - Correlated carbon and oxygen isotopic compositions in carbonate cements of differing composition from the San Joaquin Basin, California (Fayek et al., 2001). Early dolomite was precipitated soon after deposition (~-10 ~ at -- 7 Ma), calcite between 4 and 5 Ma at 5 0 - 65 ~ and Fe dolomite near 100 ~ in response to the pore-pressure reduction following exploitation of the gas cap.
30.4.2 Carbon and Nitrogen The developments in analytical capability have also benefited carbon isotope analysis. Carbon is generally analyzed as negative ions with a Cs + primary ion beam. A mass resolution of ~3500 M/AM is required to separate 12CH- f r o m 13C-. Nitrogen is a difficult element to analyze by secondary ion mass spectrometry. It does not form Nsecondary ions and yields N + ions a thousand times less productively than Si. However, Zinner et al. (1987) noted that nitrogen in the presence of carbon formed a very intense and stable CN- beam. The high ion yield of CN- makes this species attractive in Biological SIMS where organic materials can be imaged (e.g. Larras-Regard & Mony, 1997). Furthermore, living materials (e.g. cells) can be doped with isotopically labeled (13C, 15N) drugs and the effects of these drugs on the cells can be directly imaged through the 13C and 15N abundances. The 50 nm spatial resolution of the Cameca NanoSIMS 50 allows an unprecedented view of intracellular structures.
SIMS Measurement of Stable Isotopes
685
686
Chapter 30- T.R. Ireland
Figure 30.10 - S isotopic compositions of terrestrial standards measured with a multiple collector (Greenwood et al., 2000). Data from three standards are shown in plots of 534S (relative deviation of 34S/32S from that in Cation Diablo Troilite), against A33S, the fractionation-normalized 33S/32S ratio (terrestrial = 0.0). The open stars indicate the conventional (bulk) compositions of the standards. Despite the high precision afforded by the multiple Faraday Cup measurements, all three standards show analytical precisions limited by reproducibility at the ca. 0.5-0.7 %0 level. This indicates there are instrumental limitations to the measurement of the fractionation not related to signal intensity. The A33S measurements suggest the presence of non-linear fractionation or the possibility of analytical artifacts. Balmat pyrite shows the expected terrestrial ratio but with higher scatter than expected from the measurement errors. The other two standards show systematic deviations from the terrestrial composition at a level of 0.1%0. Mojzis et al. (in press) have used the same technique to document non-linear mass fractionation of A33S in Archean sedimentary sulfides at levels of-2 to + 2 %o. Similarly, e v i d e n c e for biological fossils w a s f o u n d w i t h in situ h i g h p r e c i s i o n C isotopic a n a l y s e s of graphitic c a r b o n inferred to be 3.9 Ga-old r e m n a n t s of early life (Mojzsis et al., 1996). M a h o n et al. (1998) m e a s u r e d c a r b o n isotopes in c a r b o n a t e c e m e n t s w i t h the s a m e analytical s e t u p for c a r b o n as o x y g e n a n d u s i n g the electron m u l t i p l i e r as a single collector. Precision in the 13C/12C better t h a n 1%0 (2o) c o u l d be o b t a i n e d . Similar t e c h n i q u e s are r e p o r t e d in F a y e k et al. (2001)(Figure 30.9). H a r t e et al. (1999) describe t e c h n i q u e s for c a r b o n isotope analysis a n d n i t r o g e n a b u n d a n c e s in d i a m o n d s u s i n g a single collector C a m e c a ims-4f.
SIMS Measurementof Stable Isotopes
687
30.4.3 Sulfur and Selenium The abundances of S isotopes are of great interest in economic geology where the origin of sulfide deposits can be addressed. In particular, biologic activity can cause quite extreme variability in S isotope ratios. Sulfur perhaps has the most varied range of possible primary and secondary ion species. Many combinations of primary O and secondary S polarities have been measured (e.g. O+/S -, Pimminger et al., 1984; O-/ S+, Eldridge et al., 1987; O-/S-, Macfarlane & Shimizu, 1991). However, the high ion yield from Cs + and S-makes this the preferred analytical setup.
Greenwood et al. (2000) studied S-isotopic compositions of Martian meteorites and were particularly interested in the abundance of 33S, which requires high precision and accurate determination of mass fractionation behavior. This study demonstrates the use of the Cameca 1270 with multiple collection in Faraday cups. The primary beam was ca. 2.5 nA Cs + with low-energy electron charge compensation, and selection of secondary ions of 0-25 eV with respect to the accelerating potential. Under these conditions, individual ~)34S analyses of standards (ca. 5 minute analyses) have 0.02 to 0.08 %o measurement uncertainties (2o) with ca. 0.5 - 0.7 %o reproducibility (1 standard deviation of pooled analyses; Figure 30.10). Measurements of A33S yielded analytical uncertainties of 0.07- 0.2 %o (2o). These standards show excess scatter with an MSWD of 3.7. In this case the errors can be augmented by V'MSWD thereby indicating measurement uncertainties of 0.1 to 0.2 %o, or four apparent outliers can be rejected to yield a satisfactory MSWD. Without any indication of external influences, the former is the more robust method. Greenwood et al. (2000) reported the detection of A33S anomalies in two separate ALH84001 pyrite grains (A33S = -0.74 + 0.39 %o and -0.51+ 0.38 %o, 2o); none were detectable in Nakhla pyrrhotite (total range in A33S =-0.4 + 0.5 %o to -0.07 + 0.5 %o, 2o). Mojzsis et al. (2003) used the same technique to determine a range of A33S in Archean sedimentary sulfides f r o m - 2 to + 2 %o and are well resolved from the analytical uncertainties. Selenium is geochemically associated with sulfur and also shows isotope variability (Johnson et al., 1999; 2000b). Like S it can be efficiently ionized as a negative ion. However, Se typically remains as a trace element in most S-bearing minerals, but notably occurs as native Se which may be of interest in environmental geochemistry. 30.4.4 Hydrogen Being the lightest element, the propensity for isotopic mass fractionation in D / H is the highest of any element (because of the large fractional mass difference between D and H). Hydrogen can be analyzed as either positive or negative ions: Hinton et al. (1983) used an O-primary beam with hydrogen isotopes measured as H + and D+ while Zinner et al. (1983) used a Cs + primary beam and collected negative secondary ions. Zinner et al. (1983) found that this technique has the advantage over using positive secondary ions of producing far less H2- and this species was less than 0.5 % of the D- signal for all samples. They also found that the isotopic mass fractionation of
688
Chapter 30- T.R. Ireland
the hydrogen isotopes was far less than for positive ions. Deloule et al. (1991) were interested in measuring H isotopes in terrestrial samples and so high precision and accuracy is imperative. They used an O- primary beam and measured H + and D + at a mass resolution of 1300 M / D M to separate H2 + from D +. They took care to remove moisture from the sample surface by baking the sample and ion probe at 120~ and used a liquid nitrogen cold trap to fix residual water in the vacuum. Measurements commenced when the H2 + / H + ratio was lower than 8 x 10 -4. Rather than simply comparing D / H ratios of standard and unknown, Deloule et al. (1991) found that the instrumental fractionation could be further calibrated by the measurement of Si +, Ca +, Ti +, and Mn + ion intensities on the same materials. The error of the best-fit calibration is around 7 %o and the 6D can be measured to a precision of around + 10 %o. Watson et al. (1994) used similar methodology as Deloule et al. (1991) and measured D / H ratios in hydrous amphibole, biotite, and apatite from SNC (Martian) meteorites. They found that the 6D values in these minerals ranged up to +4,000 %0, consistent with interaction with Martian crustal fluids having near atmospheric D / H . Leshin (2000) used the Cameca 1270 for D + / H + measurements of apatites from the Martian meteorite QUE94201. The water contents of the apatites were 0.2 - 0.6 wt %, as determined from H +/42Ca+ comparison to a terrestrial apatite standard, and the D / H is strongly correlated with water content. This indicates the mixture of two pos-
Figure 30.11 - Hydrogen isotopic composition and water concentration in apatites from the Martian meteorite QUE94201 (Leshin, 2000). The data suggest a mixture of water in the apatites between the Martian atmosphere at 6D of +4200 %0and magmatic water at ca. +900 %0.
SIMS Measurement of Stable Isotopes
689
sible endmembers on Mars: atmospheric water with 6D of ca. +4,200 %0, with water intrinsic to the Martian crust with 6D ca. +900 %0 (Figure 30.11). 30.4.5 Lithium, Boron These elements are typically measured as positive ions with an O- primary beam. Some minerals occur with these elements as major constituents and so high precision analyses can be carried out quite quickly. For example, boron can be measured in tourmalines at ca. 10 wt % (e.g. Chaussidon & Uitterdijk Appel, 1997; Smith & Yardley, 1996; Nakano & Nakamura, 2001). In general however, in situ analysis of these elements requires long counting times because of low B concentrations in rock-forming minerals. Chaussidon et al. (1997) report methods for B isotope analysis in meteorites and mantle rocks with B concentrations less than I ~g/g. Besides the low concentration, surface contamination and matrix effects also present major difficulties. Samples were rinsed in ultra-pure water and then sputter cleaned to reduce surface contamination to the level of 0.01 ~g/g. Standard glasses showed reproducible analyses of less than 2 %o over a period of several years. Chaussidon & Jambon (1994) measured B isotopic compositions and contents in oceanic basalts (Figure 30.12). Chaussidon & Robert (1998) used similar techniques to examine Li and B isotopes as well as Be concentrations in the Semarkona chondrite. Lithium isotopic compositions were measured from implanted solar wind in lunar soils by Chaussidon & Robert (1999). McKeegan et al. (2000) measured B isotopic compositions in a meteoritic refractory inclusion to infer the presence of live 10Be in the early solar system. Williams et al.
Figure 30.12- Boron content and isotopic composition of Hawaiian basalt glasses show a strong correlation with MgO concentration of the magma that is interpreted as the result of assimilation- fractionation processes in seawater-rich oceanic crust (Chaussidon & Jambon, 1994).
690
Chapter 30- T.R. Ireland
(2001b) used both SIMS and TIMS techniques to elucidate diagenetic reactions in marine sediments. Decitre et al. (2002) describe Li isotope methodologies for characterization of serpentinized peridotites. In the bulk samples determined by TIMS, Li concentrations range from 0.6 to 8.2 ppm, while whole rock 66Li values range from-2.9 to-14%o. In situ analyses display a greater range in both Li concentration (0.1-19.5 ppm) and Li isotopic composition (-27 to +19%o), with the serpentinized portions having higher Li concentrations than the associated relict phases. 30.5 State of the Art 30.5.1 Precision vs Accuracy vs Reproducibility
One of the main issues confronting SIMS analysis at present is the increasing analytical precision that can be obtained in a given analysis as opposed to the reproducibility of a given set of analyses. The pooled results may be accurate, but the scatter of analyses from a pool may not be commensurate with that expected from the measurement precision. The goal is to obtain reproducibility and accuracy commensurate with that precision. This paper has not addressed in any detail the methods of combining data and obtaining reliable statistics for measurements. Data can be assessed at various levels for reliability indicators from the analytical error in a single measurement through to the final grand mean of the unknown referenced to the standard. In a truly robust method, the analytical scatter would be consistent at all these levels. SIMS analysis has made great strides in reducing analytical error in the past few years. In large part this is due to the use of Faraday cup analysis, particularly in multiple collector mode, allowing higher levels of analytical precision to be obtained from large ion beams. With that higher level of precision, scatter in pooled analyses becomes more apparent, as is seen in the S isotopic analyses in Figure 30.10. However, with this level of precision, the scatter of data can be quickly assessed from a given analytical protocol and will likely lead to commensurate improvements in achieving accurate and reproducible data. At present, precisions (internal measurement error only) are achievable for O and S at a level of 0.1%0 with multiple faraday-cup collection. The accuracy of a given measurement is not quite as good with reproducibility indicating a level of ca. 0.5 %o. 30.5.2 Standards
The ability to assess analytical precision and accuracy is dependent on having standards that are homogeneous to a sufficient level. Such a condition is very difficult to establish analytically, in large part because the ion probe samples such a small volume. Microscopic heterogeneity does not necessarily reveal itself in the substantially larger samples required for other techniques. The issues are important in terms of interpretation because scatter could be interpreted as due to a geological or biological phenomenon. Only when the level of homogeneity in a material can be reliably ascertained can the level of analytical reproducibility be established.
SIMS Measurement of Stable Isotopes
691
Good standards are therefore very valuable both in calibrating instrumental mass fractionation, but also in assessing measurement statistics. Obtaining standards is no easy task because most are based on single crystals where there is no guarantee that the next crystal from that source is as homogeneous as the previous one. As such, great care and effort is expended in calibrating and assessing standards for SIMS analysis.
30.6 Concluding remarks Analytical protocols are improving such that highly precise and reproducible analyses are now readily forthcoming. However, the ease of placing a spot on a target is still mitigated by problems associated with determining instrumental mass fractionation for specific mineral compositions and other matrix effects that would otherwise be ascribed to a change in isotopic composition in the target. As such, ion microprobe analysis still requires careful measurement conditions and caution in interpretation of results where target chemistry and structure might be an issue. Real advances in SIMS have been in stable isotope research allowing in situ negative ion analysis and these areas are only beginning to be exploited. SIMS offers an unparalleled capability for in situ analysis with high sensitivity and high spatial resolution. In the coming years, it can be expected that accuracy commensurate with precision will be obtained allowing in situ capability at a similar level to bulk methods. However, natural materials are seldom as homogeneous as we would like to believe and analysis at the microscale reveals the true complexity. The rapid feedback of data allows siting of analyses based on the observed characteristics of the material. As such SIMS allows real-time assessment of data and responsive control of the analytical program.
Acknowledgments
This work is largely based on the compiled knowledge of many practitioners through their published and unpublished work. I am grateful for useful discussions with Kevin McKeegan and Ernst Zinner, and for reviews and useful comments for this work by Richard Hinton, Ian Hutcheon, and Ernst Zinner.
Handbook of Stable Isotope Analytical Techniques, Volume 1 P.A. de Groot (Editor) 9 2004 Elsevier B.V. All fights reserved.
CHAPTER 31 Stable Isotope Analysis by Multiple Collector ICP-MS Mark Rehk~imperl, 2*, Frank Wombacher2 & J. K. Aggarwal2 Institute of Isotope Geology and Mineral Resources, ETH Ztirich, NO C61, CH-8092 Ztirich, Switzerland 2 Institute of Mineralogy, Mtinster University, Corrensstrasse 24, D-48149 Mtinster, Germany e-mail: *[email protected] 1
31.1 Introduction Inductively-coupled plasma mass spectrometers with quadrupole mass analyzers (Q-ICP-MS) are widely used for the accurate and precise determination of the chemical abundances of a wide range of elements. The particular advantage of such instruments lies in the combination of the extremely high ionization efficiency and ease-ofuse of a plasma source with the excellent signal-to-noise ratio and precision offered by mass spectrometry (Jarvis et al., 1992). Such ICP-MS systems have also been used with success for isotope-dilution (ID) concentration measurements and, coupled with laser-ablation instrumentation, for in-situ analyses of geological materials (Jarvis et al., 1992). In addition, Q-ICP-MS has been applied to the determination of isotopic compositions, as an alternative to established techniques such as thermal ionization mass spectrometry (TIMS). In all instances, however, the precision of the results is significantly inferior to TIMS. The precision and accuracy of isotope ratio measurements by ICP-time of flight mass spectrometers (ICP-TOF-MS) and high-resolution sector field ICP-MS instruments with a single collector (HR-ICP-MS) are superior to Q-ICP-MS, but still inferior to TIMS (Turner et al., 1998). For the precise determination of (stable) isotope ratios, TIMS, however, is associated with low sample throughput, necessitates careful attention to sample preparation and measurement protocols, and is difficult for elements with high ionization potentials. The development of multiple-collector ICP-MS (MC-ICPMS) follows directly from the wish to combine the ionization efficiency of an ICP ion source with the superior precision attainable with the magnetic sector analyzer and multiple detector array of a TIMS instrument. The results that were obtained with the first prototype MC-ICPMS (built by VG Elemental in 1990) were impressive. They demonstrated that the isotopic compositions of elements such as Sr, Nd and Pb could be measured with an accuracy 3. Correspondence should be adressed to this author.
Stable IsotopeAnalysisby MultipleCollectorICP-MS
693
and precision comparable to TIMS (Walder & Freedman, 1992; Walder & Furuta, 1993; Walder et al., 1993b, c; Halliday et al., 1995). Furthermore, isotope ratio measurements of elements with high ionization potentials, such as Hf or W, were more precise and obtained at much greater speed and ease (Walder et al., 1993b, c; Halliday et al., 1995; Lee & Halliday, 1995). The first commercial MC-ICPMS instruments were primarily used for the study of radiogenic isotope systems or the acquisition of high-precision ID concentration data, with both research avenues focusing on elements difficult to analyze by TIMS. The potential of MC-ICPMS for stable isotope ratio determinations, however, has recently attracted significant attention and the number of studies being conducted in this field is growing rapidly. With only a few exceptions, past stable isotope studies have been restricted to light elements such as H, C, O, S and a few others. Most other (heavier) elements are virtually unexplored, because of the analytical difficulty of resolving small natural variations in isotope composition. The advent of MC-ICPMS has ameliorated this limitation, because it permits stable isotope ratio measurements of "heavy" elements at a level of precision that is sufficient for the resolution of small natural variations. This paper summarizes the most important aspects of this new field of stable isotope research. It provides an introduction to MC-ICPMS instrumentation, its analytical capabilities and measurement techniques. This is followed by a summary of stable isotope work conducted by MC-ICPMS until August 2001. 31.2 Multiple collector ICP-MS 31.2.1 Limitations of other ICP-MS systems The application of Q-ICP-MS instruments to isotope ratio measurements is limited by the comparatively poor precision of the analytical data. Even under ideal conditions the precision of isotope ratio measurements is no better than -~ 2%o (+ 2o), and for many applications it is > 5%o (Russ, 1987; Jarvis et al., 1992; Barnes, 1998; Taylor et al., 1998). Somewhat better results (2o of + 1 - 4%o) have been reported for ICP-TOFMS (Vanhaecke et al., 1999a). Isotope ratio data obtained by HR-ICP-MS have even better reproducibilities (Vanhaecke et al., 1996; 1997; Hamester et al., 1999). Such instruments have variable mass resolution. At a mass resolution of 300, reproducibilities (2 o) of as low as + 1%o and + 0.3%o have been achieved for the measurement of 7Li/6Li and 207Pb/206Pb, respectively (Hamester et al., 1999), approaching the precision of TIMS.
The analytical capabilities of such ICP-MS systems are limited by the following shortcomings. In Q-ICP-MS the mass analyzer is unable to transmit more than one mass at a time and HR-ICP-MS instruments have only a single collector. The analytical performance of these instruments is thus restricted by the sequential measurement of the unstable ion beams produced by plasma ionization. The precision of isotope ratio measurements is further limited by the peak shapes produced by the mass analyzer, which are rounded for Q-ICP-MS and triangular for HR-ICP-MS at higher mass resolutions of > 2000. Time of flight ICP-MS systems have very high repetition rates
694
Chapter 31 - M. Rehk~imper,F. Wombacher & J.K. Aggarwal
and thus nearly simultaneous ion detection capabilities. This should aid in the elimination of noise related to plasma flicker, but such systems have the r o u n d e d peak shapes characteristic of quadrupole mass analyzers.
31.2.2 The concept of MC-ICPMS MC-ICPMS instruments were designed specifically to overcome the limitations of other ICP-MS techniques for the m e a s u r e m e n t of isotope ratios. To achieve this goal, they combine an ICP-source with the magnetic sector analyzer and multiple-Faraday cup array of TIMS. The ICP source ensures excellent ionization efficiency, which is > 80% for most elements of the periodic table (Table 31.1). The magnetic sector analyzer (in conjunction with the slits of the ion-optical system) provide the flat-topped peaks necessary for high-precision isotope ratio measurements. The simultaneous collection of the ion beams with multiple Faraday cups cancels out the degradation of analytical precision due to plasma flicker. Interfacing a plasma source, which is at atmospheric pressure, and a magnetic sector mass spectrometer is not straightforward. First, this necessitates a powerful differential p u m p i n g system, to attain a v a c u u m of ~ 10-8 to 10-9 mbar in the analyzer. Such v a c u u m conditions are essential in order to obtain good peak shapes and an abundance sensitivity of less than 5 - 10 ppm. Second, the ions m u s t be accelerated into the mass spectrometer, by establishing a large potential difference (about 5 - 10 kV) between the ICP source and the detectors. Third, the broad circular ion b e a m produced by the ion source must be reshaped to fit the rectangular analyzer entrance slit at minimal loss of transmission. Fourth, the large energy spread of the ions p r o d u c e d in the plasma, generally ~ 20 V, needs to be reduced by about an order of m a g n i t u d e or, alternatively, energy focusing with an electrostatic analyzer m u s t be applied.
31.2.3 Commercial MC-ICPMS instrumentation The first commercial MC-ICPMS instrument built was the model Plasma 54 from VG Elemental (Figure 31.1). The Plasma 54 uses an ion source that is held at ~ 6 kV. A DC q u a d r u p o l e lens reshapes the circular ion beam and an electrostatic analyzer (ESA) serves to match the energy dispersion of the source to that of the magnetic secTable 31.1 - Ionization energies and degree of ionization for singly charged ions in the argon ICP. Ionization energy ( e V )
Degree of ionization
<7
___95%
7- 8
___90%
8 -9 9- 10 10 - 11 11 - 1 3 >13
~ 60- 90% ~ 30- 75% ~ 15 - 40% ~ 5%
<1%
All data are from Jarvis et al. (1992).
Elements Li, Na, A1, K, Ca, Sc, Ti, V, Cr, Ga, Rb, Sr, Y, Zr, Nb, In, Cs, Ba, REE, Hf, T1, Th, U Mg, Mn, Fe, Co, Ni, Cu, Ge, Mo, Ru, Rh, Ag, Sn, Ta, W, Re, Pb, Bi B , Si, Pd, Cd, Os, Ir, Pt Be, Zn, As, Se, Te, Au P, S, I, Hg C, Br, Xe H, He, N, O, F, Ne, C1, Ar, Kr
Stable Isotope Analysis by Multiple Collector ICP-MS
695
Figure 31.1 - Schematic diagram of the VG Elemental Plasma 54 (after Halliday et al., 1998).
tor. The instrument could be equipped with a wide flight tube and an off-axis highmass collector for the simultaneous collection of 6Li and 7Li, or 238U and 204Pb-208Pb, for example. An optional ESA at the back end permitted a reduction in abundance sensitivity by about a factor of 10 to < 0.5 ppm. The Plasma 54 has been succeeded by four "second-generation" MC-ICPMS instruments, the Micromass IsoProbe, the Nu Instruments Nu Plasma, the ThermoElemental Axiom, and the Finnigan MAT Neptune. Improvements relative to the Plasma 54 include the use of laminated magnets (to achieve rapid mass scanning) and multiple ion counting using arrays of either channeltrons (IsoProbe, Neptune) or discrete dynode multipliers (Nu Plasma, Axiom). The Micromass, Nu Instruments and Finnigan MAT instruments allow an optional retardation lens for the measurement of a low-intensity ion beams with an ion counting system at an abundance sensitivity of < 0.5 ppm. The IsoProbe source is at ground potential and the analyzer floats at - 6 kV. It is the only single-focusing MC-ICPMS instrument that uses a hexapole collision cell (Figure 31.2) to reduce the energy spread of the ions to ~ 1 - 2 V (Turner et al., 1998). A "collision gas" is fed into the center of the hexapole ion guide, and the analyte ions collide with the gas particles. The energy transfer that occurs in these collisions serves to thermalize the analyte ions and this reduces their energy spread. Reactions between the ions and the introduced gas can also be exploited to eliminate or reduce critical isobaric interferences (Feldmann et al., 1999b; Bandura et al., 2001). For the analysis of heavy elements (atomic mass > 40 amu) Ar is generally used as the collision cell gas, because it provides efficient thermalization. For the measurement of light elements (e.g., Li, B) He is used as the main collision gas, because Ar significantly reduces the
696
Chapter 31 - M. Rehk/imper, F. Wombacher & J.K. Aggarwal
Figure 31.2 - Schematic diagram of the plasma interface and hexapole collision cell of the Micromass IsoProbe (after Rehk~imper et al., 2001b).
Figure 31.3 - Schematic diagram of the Nu Instruments Nu Plasma MC-ICPMS (after Rehk~imper et al., 2001b).
Stable Isotope Analysis by Multiple Collector ICP-MS
697
transmission of the light elements. The IsoProbe collector block can be equipped with up to 9 Faraday cups and 8 channeltrons for the simultaneous measurement of masses with ~ 15% mass difference. The Nu Plasma (Figure 31.3) utilizes a source that is held at ~ 4 kV. An ESA matches the energy dispersion of the ions produced in the plasma source with that of the magnetic sector. The Nu Plasma has a variable dispersion zoom lens system behind the magnetic sector. Because the collector coincidences for static measurements of multiple ion beams are adjusted by changing the dispersion of the zoom lens system only, moveable Faraday cups are no longer required. The instrument can be equipped with 12 Faraday buckets and three discrete dynode electron multipliers for the simultaneous measurement of ions with a mass difference of up to ~ 15%. The Axiom has an ICP-source at ~ 6 kV. The double focusing ion optical system has a Nier-Johnson geometry (the ions pass through the electrostatic sector prior to the magnet) andthe instrument can alternatively be used as a single collector HR-ICP-MS. The collector block is the Plasma 54 design, with nine moveable Faraday cups plus an optional fixed-position off-axis bucket that extends the mass range of the collector array up to ~ 20%, for static measurements of isotope ratios such as 7Li/6Li. The Axiom can be equipped with up to 3 discrete dynode multipliers. The Neptune plasma source is at ground potential whereas the mass analyzer and collector array are floated a t - 10 kV. The instrument provides double focusing with a Nier-Johnson geometry and the capabilities for high mass resolution with flat-topped peaks, albeit at reduced transmission. The collector block features 8 motorized detector carriers that can be equipped either with Faraday cups or channeltron ion counters and permits the measurement of masses with up to 17% mass difference. A dynamic zoom lens system is furthermore positioned between the magnetic sector and the detector block, to support analyses at high mass resolution and to enhance multidynamic measurements by applying variable zooming for the different sequences. At the time of writing, Nu Instruments is in the process of building a second MCICPMS system, the Nu Plasma 1700. This instrument will provide high mass resolution, flat-topped peaks and high sensitivity by using a dispersion much greater than in previous MC-ICPMS systems (Halliday et al., 2000b). High mass resolution and multi-collection is also possible with the Neptune, but only with an order of magnitude lower transmission (Hamester et al., 1999). The Nu Plasma 1700 is calculated to provide a peak flat of ~ 80 ppm at a mass resolution of > 2500 for the axial collector with a transmission similar to that of the conventional Nu Plasma instrument.
31.2.4 Instrument performance In order to obtain wide, flat-topped peaks (with a peak flat width of > 400ppm) for the collection of precise isotope ratio data, MC-ICPMS instruments are generally operated at a mass resolution of about 400 (M/AM, 10% valley definition). Under such conditions, all present MC-ICPMS instruments achieve an abundance sensitivity of < 5 - 10 ppm at mass 237 relative to 238U. Definitions of these (and other) technical terms
698
Chapter 31 - M. Rehk~imper, F. Wombacher & J.K. Aggarwal
can be found in Habfast (1997) and Rehk~imper et al. (2001b). The transmission of MC-ICPMS is high relative to Q-ICP-MS instruments. Sensitivities of > 5 - 10 x 10-11 A / p p m (> 3 - 6 x 108 cps/ppm) are generally achieved for elements such as Nd, HL Pb and U with conventional aspiration systems (this term is used in the following for the combination of a concentric nebulizer with a glass double-pass spray chamber). Assuming a solution uptake rate of 400 gl/min, this is equal to a transmission efficiency of > 0.002% for Pb. The transmission efficiency is defined here as the ratio of the number of ions registered by the collector to the total number of ions available for analysis. Using a low-flow nebulizer with membrane desolvation (Cetac models MCN 6000 or Aridus), sensitivity can be increased considerably, to > 5 x 10-10 A / p p m (> 3 x 109 cps/ppm). At a solution uptake rate of N 75 ~1/min, this translates to a transmission efficiency of > 0.05%. Such sensitivities permit precise isotopic analyses on samples sizes of < 30 - 50 ng. Sensitivities of > 20 x 10-10 A / p p m (> 10 x 109 cps/ppm) have been achieved for Pb with both the IsoProbe and the Nu Plasma using a MCN 6000, at a transmission efficiency of > 0.3% (Rehk~imper & Mezger, 2000; M. Rehk~imper, unpublished results). Low-mass elements, however, have significantly lower sensitivities. With the Nu Plasma, the sensitivity for Zr or Cd is typically about 50% lower than for T1 or Pb. Sensitivities of about 8.5 and 14 x 10 -11 A / p p m were reported for Mg and Ca, respectively, using a Nu Plasma with a desolvating nebulizer (Halicz et al., 1999; Galy et al., 2001). Tomascak et al. (1999a) obtained ~ 6.5 x 10-11A/ppm (4 x 108 cps/ppm) for Li with the Plasma 54 and the MCN 6000. Further reviews and discussions of MC-ICPMS instrumentation and general performance parameters can be found in a number of recent publications (Walder, 1997; Halliday et al., 1998; Halliday et al., 2000a; Rehk~imper et al., 2001b). 31.3 Instrumental mass discrimination and analytical protocols for stable isotope ratio measurements A significant feature of plasma source mass spectrometry is the large instrumental mass bias, which is related to the preferential extraction and transmission of the heavier ions. For the high-mass elements Nd, Hf, Pb, and U, the mass bias is about 0.5 - 1.5% / a m u (Figure 31.4), approximately an order of magnitude larger than the mass fractionation that is observed during TIMS measurements. For light elements such as B, the mass bias is much larger at > 15%/amu (Figure 31.4). The large instrumental mass discrimination of plasma source mass spectrometry is ascribed by many to "space-charge effects" in the plasma interface and the focusing lens region (Jarvis et al., 1992; Douglas & Tanner, 1998). While this model can explain many of the effects seen in Q-ICP-MS, it is presently unclear if space-charge effects are also the primary source of mass bias in MC-ICPMS (Mar6chal et al., 1999; Rehk/imper et al., 2001b). Regardless of the origin, the large mass bias associated with plasma ionization clearly necessitates that mass discrimination is carefully controlled during isotope ratio measurements, if precise and accurate analytical results are to be obtained.
699
Stable Isotope Analysis by Multiple Collector ICP-MS
100
-
I
I
n
@ -B -A@
'
I
'
I
r~ ~9
10
m
-
A
Ca
Mg A
~D Cd
A~
FeCU
[~
Plasma 54
A
Nu Plasma
Nd i
~A
_
SB
~
-
or3 r
c~
0.1
Hf
,
0
I
Sr Zr
0r3
c~
'
@ IsoProbe
m
I
50
,
I
100
,
I
150
,
T1 Pb ~
m
@: U-
I
200
Atomic Mass Figure 31.4 - Plot of mass bias per amu (in %) vs. atomic mass for various elements measured with different MC-ICPMS instruments. Most data were acquired by the authors and colleagues at Mfinster (IsoProbe) and Zfirich (Nu Plasma); other results are from references summarized in section 31.4.
31.3.1 Mass bias correction by internal normalization Measurements of radiogenic isotope compositions (e.g, 87Sr/86Sr, 143Nd/144Nd) are corrected for the effects of instrumental mass fractionation by normalization of the measured ratio to an invariant isotope ratio of the same element. This procedure, termed internal normalization, most commonly employs either a linear, power, or exponential law correction (Russell et al., 1978; Wasserburg et al., 1981; Hart & Zindler, 1989; Habfast, 1998). All three "laws" were originally derived empirically, for the correction of TIMS data for evaporation-induced fractionation. The lack of a causal law that is able to describe accurately the static mass discrimination effects of plasma source mass spectrometry has led to the use of the same laws in ICP-MS, and their application is detailed below. It is assumed that rA and rB are the values of two isotope ratios measured by mass spectrometry with nuclides of the masses M1 and M2 (for rA) and M3 and M4 (for rB), where the isotopes with the masses M2 and M4 are in the denominator. RA and RB are the "true" isotope ratios, and RB is known because it is not affected by radiogenic ingrowth. It is defined that AMA- M1- M2 and AMB - M3- M4. The linear law states that R A - rA[1 + (xAAMA] where r the mass bias per atomic mass unit, is obtained by:
[31.1]
700
Chapter 31 - M. Rehk~imper, F. Wombacher & J.K. Aggarwal
[RB/rB]- 1 ~B =
[31.2]
AM B
The power law states that: R A - rA[1
-
(IA]AMA
[31.3]
with O~B -
[RB/rB] 1/(AMB)- 1
[31.4]
The exponential law differs from the linear and the power law, because the mass bias is assumed be a function of the absolute mass: RA -
[31.5]
rA[M~113A
The mass fractionation coefficient 13is obtained by: ln[RB/rB] [3B = ln[M3/M4 ]
[31.6]
All three "laws" thus assume that the fractionation observed for one isotope ratio (and quantified by the parameters aB, 13B),can be used to correct for the unknown fractionation of the other isotope ratio. This implies that OtA -- (1 B
and
~A-
f3B
[31.7]
for the linear/power and exponential laws, respectively. The choice among these laws will depend on which law best reproduces the true isotopic composition RA following correction for mass discrimination, regardless of the magnitude of c~. All three laws give similar results if ~ is sufficiently small. Detailed studies of the correction schemes for the internal normalization of TIMS data have shown, however, that the power law, and particularly the exponential law, are preferable for the correction of larger mass fractionation effects (Russell et al., 1978; Wasserburg et al., 1981; Hart & Zindler, 1989). Despite the fundamental difference in the physical origin of the mass fractionation in TIMS and ICP-MS, application of the exponential law for the internal normalization of isotopic data acquired by MC-ICPMS has been quite successful. Use of the exponential law is indicated by the fact that plasma ionization is associated with large mass fractionation effects compared to TIMS. Further support for this conclusion is derived from the observation that the mass bias a is clearly a function of mass in
Stable Isotope Analysis by Multiple Collector ICP-MS
701
plasma source mass spectrometry (Figure 31.4), whereas ~ has been shown to be approximately constant (at ~ 2) over a wide range of masses (Mar6chal et al., 1999). The application of the exponential law for the internal normalization of MC-ICPMS data was thoroughly investigated for a number of radiogenic isotope systems, and has been demonstrated to provide reliable results. Strontium and Nd isotope measurements conducted by MC-ICPMS display reproducibilities similar to TIMS analyses. For elements with high ionization potentials such as Zr, Hf or W, MC-ICPMS provides superior precision. The ability of MC-ICPMS to produce accurate isotope ratios for various elements (e.g., Sr, Nd, Hf) where precise TIMS reference values are available has also been established by a number of studies (Rehk/imper et al., 2001b). 31.3.2 Mass bias correction for stable isotope ratio measurements Internal normalization cannot be used in general for stable isotope ratio measurements, because this eliminates both instrumental and natural mass fractionations. If a double-spike is added to the samples prior to the isotopic measurement this "rule" does not appl~ however, and internal normalization can be used correct for instrumental mass discrimination. Double-spike procedures have been used in the past in conjunction with TIMS for the acquisition of stable isotope data for various elements (e.g., Ca, Fe and Cd) at high precision (Russell et al., 1978; Rosman & De Laeter, 1988; Johnson & Beard, 1999). Siebert et al. (2001) were the first to adopt a double-spike technique for stable isotope analyses by MC-ICPMS, to precisely measure natural Mo isotope fractionations. A particular advantage of double-spike methods is that they afford very precise and robust control of instrumental and laboratory-induced mass fractionation, if the spike is added prior to the chemical separation procedure. Double-spike procedures are, however, labor-intensive and complicated to establish and the target element must have at least four isotopes. A carefully calibrated spike and optimized spike-sample ratios must be used to minimize the propagation of analytical uncertainties. Currentl~ double-spiking is used routinely in only a few laboratories mainly for high-precision Pb and Fe isotopic measurements. With MC-ICPMS, the precise correction of only instrumental mass discrimination is also possible without double-spiking, using measurement techniques that are not applicable in TIMS. These techniques are based on two distinct characteristics of plasma ionization. First, the ICP source operates at steady-state. Thus mass fractionation is not primarily a time-dependent process, as it is in TIMS where the measured isotopic composition changes with time due to the progressive evaporation process. This is beneficial for the correction of instrumental mass bias by external standardization, where the isotope data obtained for a sample are referenced to the value obtained for an isotopic standard. Second, the mass discrimination associated with plasma ionization is, to a first order, a relatively simple function of mass, such that elements with similar or overlapping mass ranges display a nearly identical mass bias (Figure 31.4). Using a solution containing a mixture of two elements with similar masses, the mass discrimination observed for an element of known isotopic composition can be used to determine the unknown isotopic composition of the second element. This procedure of external nor-
702
Chapter 31 - M. Rehk~imper, F. Wombacher & J.K. Aggarwal
malization was first suggested by Longerich et al. (1987) to improve the precision of Pb isotopic measurements by Q-ICP-MS (using T1 as the reference element) and has since been applied in a number of Pb isotope studies conducted by MC-ICPMS (Hirata, 1996; Belshaw et al., 1998; Rehk~imper & Halliday, 1998).
31.3.2.1 External standardization
For heavy elements such as U, where the mass bias is ~ 0.5%/amu, the mass discrimination generally varies by < 0.2%0/hr. For light elements below mass 40, where the mass discrimination is > 5 % / a m u (Figure 31.4), the drift of mass bias is significantly larger at up to ~5%o/hr (Tomascak et al., 1999a). Changes in mass discrimination with time thus appear to be more severe for the lighter elements, such that optimized application of external standardization will require different analytical protocols for different elements. At present, the most precise data have been collected by alternating standard and sample measurements, such that each sample is referenced only to the mean of the standards measured immediately before and afterwards. This technique of "standard-sample bracketing" (Figure 31.5) is similar to the standardization method used in gas-source isotope ratio mass spectrometry. For light elements, where the absolute drift in mass bias is particularly severe, the precision of sample measurements can be improved by performing multiple short analytical runs that are each bracketed by standard measurements. Switching between samples and standards can be very rapid, if long washout protocols are not required, and mass spectrometric runs of ~4 min or less have been used to maximize the precision of B and Mg isotopic measurements by MC-ICPMS (Aggarwal et al., 1999; Galy et al., 2001). Longer data acquisition periods are more applicable for heavier elements for which the drift in mass bias is less severe. 31.3.2.2 External normalization
External normalization can be used for stable isotope analyses because it corrects only for instrumental mass discrimination, whereas internal normalization also removes the effects of any natural mass dependent fractionation processes. The procedure involves application of the same mass fractionation "laws" (equations [31.1-6]) originally developed for TIMS and internal normalization. In the case of external normalization, however, RA, rA are ratios of the analyte element and RB, rB are isotope ratios of an admixed element of similar mass that is used for mass discrimination correction. External normalization thus assumes that the c~ or [3 values for the isotope ratios of the two elements are identical (Equation [31.7]). A number of studies have shown that a linear law is least suitable for the external normalization of MC-ICPMS data, but both the power and the exponential law have been used for this purpose with some success (Taylor et al., 1995; Hirata, 1996; Belshaw et al., 1998; Rehk~imper & Halliday, 1998). Mar6chal et al. (1999), however, noted that strict application of the exponential law was unable to provide a precise correction scheme for Cu and Zn stable isotope measurements (using Zn for external normalization and vice versa), such that an empirical procedure was adopted instead. If the mass discrimination of Cu and Zn were to show ideal exponential law behavior, the raw measured 65Cu/63Cu and 68Zn/64Zn ratios should show a linear relationship
Stable Isotope Analysis by Multiple Collector ICP-MS
703
in a plot of ln(65Cu/63Cu) vs. ln(68Zn/64Zn)(Mar6chal et al., 1999). From equations [31.5,6] it follows that the slope SE of this linear correlation is defined by ]3Cu ln(M65/M63) S E - ]3Zn x ln(M68/M64 )
[31.8]
where M65 represents the atomic weight of the isotope 65Cu, etc. Strict application of the exponential law (in a manner identical to internal normalization procedures) requires that 13Cu= [3Zn, such that ln(M65/M63) S E = ln(M68/M64 ) = 0.5156
[31.9]
Mar6chal et al. (1999) noted that their analytical data indeed defined linear correlations in plots of ln(65Cu/63Cu) vs. ln(68Zn/64Zn). The regression lines, however, varied from day to day with slopes that were always significantly different from the theoretical slope that would be expected for exponential mass fractionation (Figure 31.6a). Based u p o n this observation, it was concluded that the exponential law could not be used for the precise correction of mass discrimination, because ~Cu [3Zn. The observation that a well-defined correlation was obtained for each measurement session, however, showed that the ratio ~Cu/~Zn was constant and this permitted the application of an empirical correction procedure (Figure 31.6b). For each measurement session, multiFigure 31.5 - Schematic plot of 7Li/6Li vs. time that illustrates the appliple standard runs cation of the external standardization technique (standard-sample were performed to bracketing) for the correction of instrumental mass discrimination. The define the slope of plot is based on the Li isotope techniques of Tomascak et al. (1999a). The raw data obtained in each sample run are corrected relative to the mean the ln(65Cu / 63Cu) of the adjacent standard runs only, to optimize analytical precision; each vs. ln(68Zn/64Zn) sample, however, is analyzed several times. correlation for that
704
Chapter 31 - M. Rehk~imper, F. Wombacher & J.K. Aggarwal
particular day. Extrapolation of the regression line to the "true" isotopic composition of the admixed Zn was used to obtain the fractionation corrected Cu isotope ratio of the standard (Figure 31.6b). Extra-polation of the sample results with the same slope and to the same Zn reference value then permitted the precise calculation of differences in the Cu isotopic composition between standards and samples (hCu, Figure 31.6b). An important feature of this technique is that it does not yield accurate absolute Cu isotopic compositions, but it accurately resolves small isotopic differences (MarOchal et al., 1999). Note that the calculated isotopic difference hCu does not depend on the choice of the Zn reference value. Rehk/imper & Halliday (1999) and Hirata (1997) applied external normalization for stable isotope ratio measurements of T1 and Ge, respectively, using the power law. The use of a power law for the correction of T1 isotope data was based on the observation that this procedure produced more precise data than an exponential law correction (Rehk/imper & Halliday, 1999), but it is likely that this was fortuitous. The suitability of the power law correction, however, follows directly from a thorough analysis of the analytical data. In a diagram of ln(205T1/203T1) vs. ln(208Pb/206Pb)(Figure 31.7) the raw measured T1-Pb data for the standard (M. Rehk/imper, unpublished results) define a regression line (r2- 0.9998) with a slope S = 1.0001 + 0.0026 (1 o uncertainty), and this is significantly different from the slope that would be expected for exponential law behavior (SE = 1.0147). If mass discrimination follows a power law, the measured T1-Pb isotope data should show a linear correlation in a ln(205T1/203T1) vs. ln(208Pb/206Pb) plot, with a slope given by
Sp -
(M205- M203) (M208 - M206)x
ln(1 + O~T1) ln(1 + ( I p b )
(see equations [31.3, 4]). If the power law is strictly applicable, (M205 - M 203 ) Sp = ( M 2 0 8 _ M 2 0 6 )
= 0.9999
[31.10] RT1 -- C~Pb,
such that [31.11]
Figure 31.6 - Schematic plots of ln(65Cu/63Cu) vs. ln(68Zn/64Zn) that illustrate the empirical external normalization technique used by Mar6chal et al. (1999) for mass bias correction of Cu. Both diagrams are based on results published by these authors. (a) On each of the four separate measurement sessions, the uncorrected isotope data for a mixed Cu-Zn standard solution defines a regression line with a distinct slope (full lines). The slopes of these regression lines are different from the slope that would be obtained if the mass discrimination (m.d.) of both Cu and Zn displayed ideal exponential law behavior (dashed line). Shaded fields encompass the range of raw data obtained in the MC-ICPMS measurements as well as estimates of the true Zn isotopic composition. (b) Application of an exponential law correction for mass discrimination (which assumes f3Cu = [3Zn; dotted lines) would result in poor external precision. Use of the empirical correction scheme (which assumes that [3Cu/[3Zn is constant for one measurement session and identical for standards and sampies; full lines) generates data with superior precision (see text).
Stable Isotope Analysis by Multiple Collector ICP-MS
705
706
Chapter 31 - M. Rehk~imper, F. Wombacher & J.K. Aggarwal
The theoretical slope for power law mass fractionation is thus identical, within error, to the observed slope of the regression line for the analytical data (Figure 31.7). Use of the power-law and 208Pb/206Pb - 2.16701 for NIST SRM-981 Pb (Todt et al., 1996), however, generated a mean corrected 205T1/203T1 ratio of 2.38908 + 11 (2 o) for NIST SRM-997 T1 (Rehk/imper & Halliday, 1999). This is significantly higher than the certified T1 isotopic composition of this reference material which is 2.3871 _+10 (Dunstan et al., 1980). The latter value is considered reliable, within the given errors, because it is based on an absolute measurement of the atomic weight. The discrepancy between the true and the measured isotopic composition is probably related to the fact that the most precise correction procedure does not necessarily generate accurate isotopic data, because the regression line through the measured data does not intersect the true isotopic compositions of the standards at a = 0. Nonetheless, Rehk/imper & Halliday (1999) were able to show that the correction technique was suitable for the accurate and precise measurement of T1 isotopic differences between samples and an isotopic standard. These observations indicate that external normalization using either the approach of Mar6chal et al. (1999) or the exponential/power law do not provide a sufficiently correct description of the isotope fractionation process that is associated with plasma ionization. These procedures therefore are not suitable for the accurate and precise determination of absolute isotopic compositions. Internal normalization of MCICPMS data using an exponential law, however, appears to be able produce isotopic results that are accurate to within < 50 - 100 ppm for a number of elements. This indicates that chemical or physical properties other than mass may cause different elements to display a unique behavior during instrumental isotope fractionation. External normalization, however, can be used to resolve small isotopic differences between samples, even if the mass discrimination of the analyte and normalizing element is (slightly) different. The primary requirement is that the two elements show a highly correlated response to changes in instrumental mass bias but this condition is probably only met by elements with very similar or overlapping mass ranges.
31.3.3 Caveats and potential problems In the following we summarize some of the most important problems that may be encountered in acquisition of precise and accurate stable isotope ratio data by MCICPMS. Further treatments of this subject can be found in a number of publications (Jarvis et al., 1992; Habfast, 1997; Horlick & Montaser, 1998).
31.3.3.1 Spectral interferences Isotopic measurements are particularly vulnerable to artifacts generated by spectral interferences, because they aim to resolve very small differences in isotope composition. Such artifacts are a major concern, particularly below 100 amu, because MCICPMS instruments are generally operated at a mass resolution of N400, to produce peak flats that are sufficiently wide for high-precision isotope ratio measurements. At such operating conditions, the mass spectrometer is unable to resolve interferences between the analyte and other ions with similar mass to charge ratios. Spectral interferences are primarily generated by ions derived from the sample matrix (and the
Stable Isotope Analysis by Multiple Collector ICP-MS
707
Figure 31.7 - Plot of ln(205T1/203T1) vs. ln(208Pb/206Pb) for the raw data obtained by MCICPMS for a mixed solution of NIST SRM-997 T1 and NIST SRM-981 Pb. The data, obtained over a period of several months with a Plasma 54 (M. Rehk~imper, unpublished results), define a single linear correlation with a slope that is statistically identical (at the 99% confidence level) to the slope expected for power law mass fractionation.
sample solvent) or by molecular ions and refractory oxides that form in the plasma. The collection of accurate isotope ratio data necessitates that spectral interferences are either insignificant, or reduced to tolerable levels, such that an accurate correction can be applied. In most cases this can be achieved by (1) chemical separation of the analyte element from the sample matrix a n d / o r (2) the application of appropriate measurement protocols.
Isobaric interferences from other elements can be corrected for, by monitoring an isotope of the interfering element and application of a suitable interference correction (Halliday et al., 1995). Nonetheless, it is imperative to evaluate the correction scheme, to determine the maximum levels of contamination that can be tolerated for accurate isotope ratio measurements. To avoid the measurement uncertainties that are associated with such interferences, the majority of the published MC-ICPMS isotope data for geological samples have been collected either (1) on "clean" samples that were processed through a chemical separation step prior to the mass spectrometry, or (2) by direct analyses of major elements in a single phase (e.g., Cu in ore samples, Ca in carbonates). Correction for spectral interferences from polyatomic ions such as hydrides, oxides, nitrides, or argides is less straightforward. In many cases, such interferences (e.g., from element-argides) can be avoided by the chemical separation of the analyte prior to the mass spectrometry. The production of hydrides, oxides and nitrides is further-
708
Chapter 31 - M. Rehk~imper, F. Wombacher & J.K. Aggarwal
more greatly reduced if a desolvating nebulizer is applied for sample introduction (Montaser et al., 1998). The development of methods for the measurement of Fe isotopic compositions by MC-ICPMS has been particularly difficult, due to interferences from various argides. The published data (section 31.4), however, demonstrate that precise results can nevertheless be obtained for this element, if appropriate analytical protocols are adapted. Interferences from refractory oxides and hydroxides (such as REE oxides) are often not a severe problem in MC-ICPMS because they can be avoided by chemical separation and because oxide formation can be greatly reduced in many cases by the use of a desolvating nebulizer. Collision cells, which are used in some Q-ICP-MS systems (Feldmann et al., 1999b; Bandura et al., 2001) and in the IsoProbe MC-ICPMS, have also been used to reduce polyatomic and isobaric interferences. Experiments conducted with Q-ICP-MS systems, however, have also shown that interfering molecular ions can also be produced in collision cells, for example by reaction of a He-H2 collision gas mixture with residual water from the solvents (Feldmann et al., 1999a).
31.3.3.2 Matrix effects Correction of instrumental mass bias by external standardization requires that the mass discrimination of the analyte element is identical for both the sample and the standard. Similarly, external normalization assumes that the relative mass bias encountered by the analyte and the reference element are indistinguishable for the samples and the isotopic standard. A number of studies conducted by both Q- and MC-ICPMS have shown, however, that instrumental mass discrimination can vary considerably with the analyte matrix and the response of two elements of similar mass to a different matrix may not be sufficiently correlated (Douglas & Tanner, 1998; Horlick & Montaser, 1998; Galy et al., 2001). This indicates that precise measurements of natural stable isotope fractionations are particularly prone to the generation of artifacts and should preferably be conducted on samples that have been chemically isolated from the sample matrix. Because no chemical separation is perfect, a number of MC-ICPMS studies demonstrated that remaining matrix constituents do not have a detrimental effect on data quality, either by influencing the mass bias behavior or through the formation of spectral interferences (Halicz et al., 1999; Zhu et al., 2000a). It is furthermore possible, that the accuracy of stable isotope ratio measurements may be affected by the concentration of the analyte (or the relative concentrations of the analyte and normalizing element) as well as the type and strength of the acid matrix of the sample and standard solutions. The effect of such factors on the accuracy of stable isotope ratio measurements at the 20 - 200 ppm level has yet to be investigated systematically. Future studies should consider such factors, because matrix effects on mass discrimination may vary considerably for different elements a n d / o r instruments.
Stable IsotopeAnalysisby MultipleCollectorICP-MS
709
31.3.3.3 Other factors
A number of MC-ICPMS stable isotope studies have noted that instrumental operating conditions (gas flows, focusing lens settings, acceleration lens potentials) must remain constant during an analytical session (Hirata, 1997; Rehk~imper & Halliday, 1999) because even small changes in these parameters may result in changes in mass bias. Memory effects have been noted as a particular problem for elemental concentration measurements by Q-ICP-MS, particularly for the analyses of samples with highly variable analyte abundances. For (stable) isotope ratio measurements by MC-ICPMS, memory problems are often less severe, because most analyses are conducted with "clean" solutions of separated elements, and large variations in the concentrations of the analyte solutions should be avoided in any case (Rehk~imper & Halliday, 1999). The "instrumental blank" problem is further demagnified, because samples and standards typically display only small differences (generally < 10%o) in isotopic composition. This condition does not apply if a double-spike is used for mass bias correction, such that a more rigorous control of instrumental memory may be required. It has been observed by many that memory effects are typically somewhat larger for desolvating nebulizers in comparison to other sample introduction systems. This disadvantage, however, can generally be overcome with appropriate washout protocols. Particularly severe memory problems for isotope ratio measurements by plasma source mass spectrometry have been encountered for a only a few elements (e.g., B, Os) which are thought to "stick" well to the surfaces of the sample introduction system. Isotopic analyses of these elements are nevertheless possible by ICP-MS, but require the application of designated sample introduction systems such as direct injection nebulizers (Montaser et al., 1998), and/or special cleaning procedures.
31.4 Applications In the last ten years, the technique of MC-ICPMS has found application in a continuously growing number of analytical laboratories wold-wide. The interest of the geochemical community in MC-ICPMS is largely based on the high precision and versatility of this mass spectrometric technique, which has been applied for isotopic measurements of numerous elements (Rehk~imper et al., 2001b). The following section provides a brief review of the application of MC-ICPMS to stable isotope ratio measurements in geo- and cosmochemistry. Stable isotope data for geological samples have already been collected for more than a dozen elements by MC-ICPMS (Table 31.2). In sections 31.4.1 to 31.4.10, we provide a review of the literature concerning those elements for which peer-reviewed papers have been published until August 2001. Section 31.4.11 briefly summarizes results presented in abstract form for additional elements. 31.4.1 Lithium
Lithium has the two stable isotopes 6Li (7.5%) and 7Li (92.5%). Due to the low ionization potential (5.4 eV), the isotopic composition of Li is readily measured by TIMS, but the acquisition of precise data is difficult due to the large instrumental mass fractionation. Significant Li isotopic variations have been detected in natural samples. This has spurred the development of analytical techniques for Li, employing TIMS
710
Chapter 31 - M. Rehk~imper, F. Wombacher & J.K. Aggarwal
(Chan, 1987; You &Chan, 1996; Moriguti & Nakamura, 1998a), Q-ICP-MS (Gregoire et al., 1996) and HR-ICP-MS (Hamester et al., 1999). Lithium isotopic measurements by TIMS can achieve precisions of + 0.7- 2.5%0, but the most precise methods require long measurement protocols and either large sample sizes or extremely clean samples for precise control of instrumental mass fractionation. Tomascak et al. (1999a) were the first to develop analytical techniques for the measurement of Li isotopes by MC-ICPMS and these have been applied in a number of geochemical studies (Tomascak & Langmuir, 1999; Tomascak et al., 1999b, c). The Li isotopic data were acquired with a Plasma 54: MC-ICPMS using a Cetac MCN 6000 desolvating nebulizer. Individual analyses required ~ 40 ng of Li to achieve an internal precision of + 0.5%0 (2 o mean). Sample purity was shown to be a less critical factor for MC-ICPMS than for TIMS, thus allowing for a simpler chemical separation procedure. Control of mass discrimination was achieved by external standardization, with sample runs being interspersed between measurements of an isotopic standard. Short analysis times (~ 8 min) were used to minimize sample consumption and to achieve optimal correction for mass bias, because the latter was found to display a significant, but smooth, drift (up to 60%0 over 12 hr). An external reproducibility of + 1.1%o (2 o) was achieved for multiple measurements of samples and isotope standards (Tables 31.2, 3). Accuracy was evaluated by replicate analyses of two well-characterized standard samples: seawater and basalt JB-2 (Table 31.3).
31.4.2 Magnesium Magnesium has the three stable isotopes 24Mg, 25Mg, and 26Mg, with relative abundances of 79.0%, 10.0% and 11.0%, respectively. The first ionization potential of Mg is 7.6 eV. Investigations of mass dependent isotope fractionation of Mg, conducted by TIMS, reported measurement uncertainties (2 ~) of about + 1 - 2%o/amu (Catanzaro & Murphy, 1966; Catanzaro et al., 1966; Wasserburg et al., 1977). Galy et al. (2001) utilized a Nu Plasma MC-ICPMS instrument for Mg isotope analyses. A Cetac MCN 6000 desolvating nebulizer was used for sample introduction. The mass bias was reported to be ~ 7.5%/amu and was found to be sensitive to variations in instrumental operating conditions. The measurements of Mg stable isotope compositions applied a standard-sample bracketing technique for mass discrimination correction, with repeated short (200 s) analyses of both sample and standard solutions. An external reproducibility (2 o) about + 0.06%o/amu was obtained for several samples over a period of 15 months. The addition of Na, A1, and Ca to sample solutions was observed to give rise to higher 26Mg/24Mg isotope ratios, relative to a pure Mg standard. This was thought to result from a mass-dependant matrix effect that was most serious for Ca, followed by A1 and Na. The isobaric interference of 48Ca2+ on 24Mg was found to be significant at Ca/Mg ratios of a 0.5. The Mg isotope compositions of several terrestrial metal, mineral and chlorophyll samples (Figure 31.8) were reported to vary by up to about 2%o/amu (Galy et al., 2001). Results obtained for chondrules from the Allende meteorite were used to infer the conditions at which these objects formed in the early solar system (Galy et al., 2000).
Element Isotope ratio(s)
~
Li B Mg S Ca Fe cu Zn Ge Se Mo Cd Sn Sb Hg T1
Sample Introductiona
~
Mass Bias Correctionb
Mass Bias (70 lamu)
Precision (20) of MC-ICPMS measurements (700 1amu)
~~
Sol Sol Sol LA Sol Sol, LA Sol Sol Sol Sol Sol Sol Sol LA VG Sol
Terrestrial variation measured by MC-ICPMSC (%lamu) ~~
~
ES ES, EN ES EN ES EN, ES EN, ES EN EN ES EN, DS EN EN EN EN EN
-1
- 0.3 - 0.7 0.1 - 0.2 - 0.3
15 - 40 -8 -8 -5 -3 -3 -3 -1.5 -2 -2
- 1.7 - 0.9
- 0.7
0.05 - 0.1 0.03 - 0.15 0.04 - 0.05 0.1 - 0.5 0.15 0.02 - 0.1 0.07 - 0.15 0.05 - 0.1
-4 -5 0.5 (- 1.5") -1 0.3 0.2 0
-
-
-
-
-2
-
0.05 (?) 0.03 - 0.05
-
Stable Isotope Analysis by Multiple Collector ICP-MS
Table 31.2 -Summary of stable isotope ratio measurements conducted by MC-ICPMS.
-
-
- 0.25 -1
LA = laser ablation, Sol = solution nebulization, VG = vapor generation. b DS = double spike, EN = external normalization using an admixed element, ES = external standardization, including standard-sample bracketing. c No data are given for Li, B, and S because large terrestrial variations have already been determined with other techniques; * variation observed in Fe-meteorites. All data are from references summarized in section 31.4 or unpublished results of the authors. a
711
712
Chapter 31 - M. Rehk~imper, F. Wombacher & J.K. Aggarwal
Table 31.3 - Comparison of Li isotope data for standard samples acquired by MC-ICPMS and TIMS. Instrument Type
Ref.
Sample Type
MC-ICPMS
(a)
JB-2 basalt Pacific seawater NIST-688 Li
TIMS
(b) (b) (c) (d)
JB-2 basalt Pacific seawater Atlantic seawater Seawater
Sample Size
67Li + 2o
n
40 ng 40 ng 40 ng
5.1 + 1.1 31.8 + 1.9 2.8 + 1.1
13 15 4
> 250 ng > 250 ng < 100 ng 4 mg
4.9 + 0.7 30.0 + 0.7 31.4 + 1.0 33.3 + 1.2
5 5 6 5
67Li represents the relative deviation of 7Li/6Li for a sample from the L-SVEC Li isotope standard in %. References: (a) Tomascak et al. (1999a); (b) Moriguti & Nakamura (1998a); (c) You &Chan (1996); (d) Chan (1987). n = number of measurements.
31.4.3 Calcium
Calcium has six stable isotopes, 40Ca (96.9%), 42Ca (0.65%), 43Ca (0.14%), 44Ca (2.1%), 46Ca (0.004%), and 48Ca (0.19%). Thermal ionization of Ca is straightforward, due to the low ionization potential of the element (6.1 eV), but the acquisition of precise isotopic data by TIMS is difficult, because of the large instrumental fractionation
Figure 31.8 - Three-isotope diagram in 6-notation for Mg (after Galy et al. (2001). 62XMg is the relative deviation of the 2XMg/24Mg ratio of a sample from the standard in permil. The bold line is the terrestrial mass fractionation line of Galy et al. (2000). Data are shown for two Mg metal samples (open circles), two Mg solutions (filled circles), magnesia (open square), two magnesites (diamonds), and two chlorophyll samples (triangles). Error bars are 2o.
Stable Isotope Analysisby Multiple Collector ICP-MS
713
effects. The most precise TIMS measurements of Ca isotopic compositions (external precision ~ + 0.1%o/amu) were performed by the double-spike technique and small terrestrial variations in Ca isotope ratios were identified (e.g., Russell et al., 1978). Ca isotope ratio measurements have also been conducted by Q-ICP-MS using "cool plasma" conditions (Patterson et al., 1999) and HR-ICP-MS (Sttirup et al., 1997). Halicz et al. (1999) determined Ca isotope compositions in carbonate samples by MC-ICPMS. The measurements were performed with a Nu Plasma instrument and a MCN 6000 desolvating nebulizer. The presence of 40Ar+ prevented the measurement of 40Ca, but backgrounds at higher masses were reported to be sufficiently low to permit accurate data acquisition for 42Ca, 43Ca, and 44Ca. Potential interferences by 84,86,88Sr2+ were corrected, by monitoring 87Sr2+ at mass 43.5. Mass discrimination for Ca was ~ 5%/amu, with a drift of about 0.1 - 0.05%o/amu per hour. A sensitivity of 0.3 x 10-11A/ppm was reported for 44Ca, and analyses were generally performed on Ca solutions with concentrations of ~ 15 - 30 ppm. External standardization by a standard-sample bracketing technique was used to correct for instrumental isotope fractionation during the measurements, and an external precision (2 o) of ~ + 0.05%o/amu was reported. The measured isotope ratios were found to be unaffected by the presence of Mg at concentrations up to twice that of Ca. Several carbonate samples were analyzed, and the data indicate a natural variability of 0.7%o in 42Ca/44Ca isotope ratios. 31.4.4 Iron
Iron has four naturally occurring isotopes, 54Fe (5.8%), 56Fe (91.8%), 57Fe (2.1%) and 58Fe (0.3%). Precise measurements of Fe isotope ratios by TIMS are difficult because of the high ionization potential of the element (7.9 eV), its low mass, and isobaric interferences from 54Cr and 58Ni. With double-spike TIMS procedures the Fe isotopic compositions of geological samples can be measured with an external precision (2 o) of about + 0.1 - 0.3%o/ainu (Beard, 1999; Mandernack et al., 1999). TIMS studies have identified natural variations in Fe isotope ratios for both meteorites (V61kening & Papanastassiou, 1989) and terrestrial materials (Beard, 1999; Beard et al., 1999). Various ICP-MS methods have also been developed for Fe isotopic measurements (Whittaker et al., 1992). The isobaric interferences of ArN + and ArO + on masses 54 and 56, however, are a major obstacle to precise Fe isotopic studies by plasma source mass spectrometry. Anbar et al. (2000) measured 56Fe/54Fe and 57Fe/54Fe ratios with a VG Elemental Plasma 54 using a microconcentric nebulizer coupled to a membrane desolvation system for sample introduction. The use of a desolvating nebulizer (operated without N2) reduced O and N based molecular species, such that the interferences of 40Ar14N and 40Ar160 were typically less than 6%o and 1%o for 54Fe and 56Fe, respectively. Prior to each run, the intensities of 40Ar14N and 40Ar160 were determined, and these data were used to correct the measured Fe isotope ratios. The intensities of sample and standard ion beams were furthermore matched by appropriate dilution. Correction for mass discrimination was achieved by monitoring 63Cu/65Cu of admixed Cu. The
714
Chapter 31 - M. Rehk~imper,F. Wombacher & J.K. Aggarwal
accuracy and precision of the techniques were verified by analyses of a standard solution enriched in 54Fe relative to an isotopically normal standard. An external longterm reproducibility of + 0.1 to 0.15 %o/amu was obtained for the 56,57Fe/54Fe ratios in these measurements, each consuming ~ 1 tog of Fe. Analyses of biogenic FeCO3 samples, produced by in-vitro microbial growth, showed shifts in Fe isotopic composition of ~ 0.4%o/amu relative to the starting material (Anbar et al., 1999). The fractionation of Fe isotopes during elution of Fe from an anion-exchange resin at the 2No/ amu level was also reported (Anbar et al., 2000). Zhu, Belshaw and coworkers reported the measurement of 56,57Fe/54Fe ratios with a Nu Plasma MC-ICPMS instrument for a variety of geological samples and meteorites (Belshaw et al., 2000; Zhu et al., 2000b; Zhu et al., 2001). Sample introduction utilized a MCN 6000 desolvating nebulizer, operated without a flow of N2, to minimize formation of ArN. The isotopic measurements were conducted at typical ion beam intensities of ~ 5 x 10-11A for 54Fe. With such signal levels, the molecular interferences at masses 54 and 57 were typically < 200 ppm; the Cr interference on mass 54 was < 25 ppm. Mass discrimination was ~ 3 % / a m u , and was corrected for during sample measurements by external standardization using a standard-sample bracketing technique. This procedure, however, required that the Fe concentrations of samples and standards were matched to within better than 20 - 50%. An external precision (2 o) of ~ + 0.03%o/amu was achieved for m i n i m u m sample sizes of 20 gg. A time-series analysis of a ferromanganese crust revealed large variations in Fe isotope composition that
Figure 31.9 - Fe and Pb isotope data for a time-series (0-2.5 Ma) of a Fe-Mn crust from the North Atlantic (Zhu et al., 2000b). 657Feis the relative deviation of the 57Fe/54Fe ratio of a sample from the standard in permil. Error bars are 2o. The iron isotope variations are coherent with those of 206Pb/204Pb"
715
Stable Isotope Analysis by Multiple Collector ICP-MS
were found to be correlated with changes in radiogenic Pb isotope ratios (Figure 31.9). Thus, the variations in 57Fe/54Fe were thought to be caused by changes in the continental input into the oceans (Zhu et al., 2000b). Analyses of various terrestrial minerals and meteorites for 56,57Fe/54Fe revealed differences of up to about 1%o/amu in isotope composition. All samples were found to fall on a single mass-fractionation line when plotted in a three-isotope diagram, indicating the homogenization of Fe isotopes in the early solar system (Zhu et al., 2001). Hirata and Ohno (2001) used a Plasma 54 coupled to a UV laser ablation system for in-situ Fe isotope analysis. An "on-peak" baseline subtraction procedure (similar to that used by Anbar et al., 2000) was applied to correct for interfering signals from ArN +, ArO +, and ArOH + and mass bias correction utilized sample-standard bracketing. External normalization to Cu offered no improvement in measurement precision. The laser beam was rastered over a 250/am2 area during each 50 s-analysis, which consumed about 30 ng of Fe for a total Fe ion current of ~ 6 x 10-11A. With these techniques, internal precisions (2 o mean) of better than 1%o were reported to be achievable for both 54Fe/56Fe and 57Fe/56Fe. The analysis of several natural materials (Feminerals and Fe-meteorites) showed no significant variations in Fe isotope compositions relative to the NIST SRM 665 Fe standard, within the analytical uncertainties of the method.
31.4.5 Copper Copper possesses two naturally occurring isotopes, 63Cu and 65Cu, with relative abundances of 69.2% and 30.8%, respectively. Cu displays a first ionization potential of 7.7 eV. Recent Cu stable isotope data, acquired using low temperature TIMS techniques, display external uncertainties (2 o) of + 0.2 - 0.3%o/amu (Gale et al., 1999). Isotopic analyses of Cu have also been performed by Q-ICP-MS (Lyon & Fell, 1990; Lyon et al., 1996) and HR-ICP-MS (Hamester et al., 1999). Mar6chal et al. (1999) performed the first Cu (and Zn, see below) isotope analyses of geological samples by MC-ICPMS. All measurements were performed with a Plasma 54, using a Glass Expansion nebulizer for sample introduction. Isobaric interferences from polyatomic ions (e.g., 40Ar23Na+, 48Ca160 + and 50,52Cr160+) required chemical separation prior to isotopic analyses of geological samples. The purified Cu fractions were doped with Zn for external normalization using 6 6 Z n / 6 4 Z n and an empirical correction technique (see section 31.3.2). A mass bias of ~ 3.3% / amu was reported for Cu. Repeated measurements of standard solutions and samples were found to display an external
Table 31.4 - Cu and Zn isotopic results obtained by MC-ICPMS for multiple dissolutions of sediment trap material (Mar6chal et al., 1999). Sample No
665Cu %oa
1 2 3 4 5 6 Meanc
0.05 0.06 0.06 0.09b 0.06b 0.16b 0.08 + 0.08
666Zn %oa 0.49 0.48 0.45 0.37 0.42 0.42 0.44 + 0.09
a 6-values expressed relative to NIST SRM-976 Cu and JMC Zn, respectively, b Cu isolated twice by anion-exchange chemistry, c Uncertainties are 2o.
716
Chapter 31 - M. Rehk/imper,F. Wombacher& J.K. Aggarwal
reproducibility of + 0.04%o/amu for 65Cu/63Cu (Table 31.4). As little as 200 ng of Cu were sufficient for one analysis. Data reported for oceanic sediments, biological materials and ore samples indicate the existence of terrestrial Cu isotopic variations with 665Cu values ranging from - 3 to + 5%o, confirming the results of previous TIMS measurements (Shields et al., 1965). 665Cu is the relative deviation of the 65Cu/63Cu isotope ratio of a sample from NIST SRM-976 Cu in %o. Gale et al. (1999) measured the Cu isotopic compositions of ores and archeological artifacts with a Plasma 54. Correction for instrumental mass discrimination of Cu was obtained by external normalization to an admixed Zn solution, using the exponential law. The mass bias of ~ 3.7%/amu was reported to be constant to within 0.8%o (2 o) over 10 h. Repeated standard measurements displayed a 2 o precision of ~ + 0.04%o/ amu. For the analysis of samples, the Cu was isolated from the matrix by a combined anion-exchange/electro-deposition technique. Data obtained for Cu minerals displayed 665Cu values ranging f r o m - 1.6 to + 7.7%o, whilst archeological samples ranged from + 0.2 to + 4.3%o in 665Cu (all data are relative to NIST SRM-976 Cu). Ore samples analyzed by both MC-ICPMS and TIMS displayed identical Cu isotopic compositions, within error. Zhu et al. (2000a) measured Cu isotope ratios of ore samples without chemical separation, to avoid isotope fractionation during ion-exchange chromatography. A Nu Plasma combined with a Cetac MCN 6000 desolvating nebulizer was used. A mass bias of ~ 3% / a m u was reported, and this was observed to vary with changing operating conditions during the course of an analytical session. Two methods of mass discrimination correction were evaluated: (1) external normalization to admixed Zn, and (2) external standardization by standard-sample bracketing. "Zn-doping" was reported to be faster and more precise (2 o external precision ~ + 0.02%o/amu), but required the removal of other elements that could interfere with the mass discrimination correction. The external standardization technique was chosen for analysis, because samples were not subjected to chemical purification. An external reproducibility (2 o) of ~ + 0.03%o/amu was obtained with this method. Isobaric interferences from NaAr +, NaOH+ and MgAr + were found to be negligible for the ore samples and Cu isotope ratios were unaffected by the presence of Fe in the sample solutions at levels up to 15 times higher than the Cu contents. Ore samples that were formed in association with low-temperature aqueous solutions were reported to display somewhat larger differences in 665Cu (- 0.3 to + 1.6%o) than chalcopyrites from igneous rocks (665Cu ~ - 0.6 to + 0.4%o). This was thought to indicate that Cu isotopic variations arise principally through low temperature processes, rather than as a result of source heterogeneities. 31.4.6 Zinc
Zinc has five stable isotopes (64Zn: 48.6%, 66Zn: 27.9%, 67Zn: 4.1%, 68Zn: 18.8%, 70Zn: 0.6%) and displays a very high first ionization potential of 9.4 eV. Stable isotope ratio measurements of Zn using TIMS and a double spike method have achieved an analytical precision of ~1.5%o/amu (Rosman, 1972). Isotopic analyses of Zn have also been conducted by Q-ICP-MS, e.g., for medical and archeological studies (Roehl &
Stable Isotope Analysisby Multiple Collector ICP-MS
717
Gomez, 1995; Budd et al., 1999). The first stable isotope measurements of Zn by MC-ICPMS were conducted by Nishizawa et al. (1998). A Plasma 54 instrument was used to investigate the Zn isotope effects of chemical exchange reactions performed in the laboratory. A precision of better than + 0.5%o was noted for the measurement of the isotopic enrichment factor of 66Zn relative to 68Zn, but no further analytical details were given. Stable isotope measurements for Zn in geological materials were performed by Mar6chal et al. (1999) using analytical techniques identical to those described for Cu. Despite the use of Ni cones on the Plasma 54 MC-ICPMS, only a negligible (<1 ppm) contribution of 64Ni (0.9%) to the 64Zn ion beam was noted; other spectral interferences were precluded by chemical separation of Zn from the geological samples. Correction for mass discrimination of Zn was achieved by external normalization relative to the 65Cu/63Cu ratio of admixed Cu using an empirical correction technique (see section 31.3.2). The mass bias for Zn was similar to that of Cu at ~ 3.3% /amu. Repeated measurements of standard solutions and samples were found to display an external reproducibility of about + 0.05%o / amu for 66Zn / 64Zn (Table 31.4), and as little as 200 ng of element were sufficient for one analysis. Zinc isotope data for samples with significant Cu isotope variations (665Cu ~ - 3 to + 5%o) were observed to be less variable, with ~66Zn values ranging from - 0.2 to + 0.8%o (~}66Zn is the relative deviation of the measured 66Zn/ 64Zn ratio from a JMC Zn solution in %o). More recently, Mar6chal et al. (2000) reported Zn data for Fe-Mn nodules and marine sediments. These samples were reported to have variable but uniformly high ~66Zn values (+ 0.2 to + 1.2%o). The depletion of the light Zn isotopes in the Fe-Mn nodules and marine particles was inferred to result from biological activity in the water column. 31.4.7 Germanium
Germanium has the naturally occurring isotopes 70Ge (20.8%), 72Ge (27.5%), 73Ge (7.7%), 74Ge (36.3%) and 76Ge (7.6%). Due to the high first ionization potential of 7.9 eV, thermal ionization of Ge is difficult. Previous studies of Ge isotopic compositions in natural materials utilized Q-ICP-MS (Xue et al., 1997) or TIMS in conjunction with double spiking (Green et al., 1986) to achieve precisions (2 o) of about + 0.5 - 1%o/ amu. Natural terrestrial variations in Ge isotopic compositions have hitherto not been detected, but fractionated Ge was reportedly observed in samples from the Canyon Diablo meteorite (Xue et al., 1997). Hirata (1997) measured the Ge isotopic compositions of iron meteorites by MCICPMS. The measurements were performed with a Plasma 54, using both a conventional sample introduction system and a microconcentric nebulizer. An instrumental mass bias of ~ 1.5%/amu was observed for Ge. Isotopic measurements of Ge were corrected for this effect by external normalization to the 71Ga/69Ga ratio of admixed Ga. The most problematic isobaric interferences, 36Ar40Ar and 38Ar2 on mass 76, prevented measurement of 76Ge. Interference from 56Fe160 on 72Ge necessitated the isolation of Ge by solvent extraction. These techniques permitted the determination of Ge isotopic compositions on ~ 400 ng of Ge with an internal precision (2 o mean) of ~
718
Chapter 31 - M. Rehk~imper,F. Wombacher & J.K. Aggarwal
Figure 31.10 - Ge isotopic compositions of six iron meteorites and one pallasite determined with a Plasma 54 MC-ICPMS (after Hirata (1997). Most of the meteorite data fall within error of the theoretical mass fractionation line (slope = 1/3). This supports the conclusion that the variability of Ge isotopic compositions was generated by cosmochemical fractionation processes. 0.02 - 0.1%o/amu, but no estimate of reproducibility was given. The data obtained for three different Ge standard solutions, however, indicate an external precision (2 o) of 0.1 - 0.5%0/amu (Table 31.2), if it is assumed that the Ge isotopic compositions of all standards are identical. Analytical data were reported for 6 iron meteorites and one pallasite (Figure 31.10). All samples were found to be characterized by negative 670Ge and 672Ge values, with fractionations of up to ~ 1.5%o/amu (67XGe represents the permil deviation of the 7XGe/73Ge isotope ratio of a sample from a Merck standard solution). In a diagram of 672Ge vs. ~570Ge (Figure 31.10), most of the meteorite data plot within error of a single mass fractionation line. The variability of Ge isotopic compositions was thought to be due to the isotopic fractionation that occurred during the evaporative loss of Ge.
31.4.8 Molybdenum M o l y b d e n u m possesses 7 naturally occurring isotopes" 92Mo (14.8%), 94Mo (9.3%), 95Mo (15.9%), 96Mo (16.7%), 97Mo (9.6%), 98Mo (24.1%) and 100Mo (9.6%). The first ionization potential of the metal is 7.1 eV. Natural isotope variations of Mo have previously been studied using TIMS (e.g., Qi-Lu & Masuda, 1994) and resonant ionization mass spectrometry (Nicolussi et al., 1998). Further isotopic measurements of Mo have been conducted by HR-ICP-MS (Giussani et al., 1997). Lee & Halliday (1995) used a Plasma 54 MC-ICPMS to determine the atomic weight of Mo, but they applied internal normalization for mass discrimination correction.
Stable Isotope Analysis by Multiple Collector ICP-MS
719
Anbar et al. (2001) used a Plasma 54 for Mo stable isotope ratio measurements. Using an MCN 6000 desolvating nebulizer, sensitivities of about 8 x 10-11A/ppm Mo were obtained. For the correction of mass discrimination, the sample solutions were doped with either Zr or Ru and 90Zr/91Zr or 99Ru/101Ru was applied for the external normalization of the Mo ratios using the empirical correction method of Mar6chal et al. (1999). With these techniques, sample measurements typically displayed an external precision (2 ~J) of better than _+0.1%o/amu for the 95Mo/97Mo isotope ratio. The accuracy of the method was confirmed by analyses of Mo solutions enriched in 95Mo. Subsequent analyses of molybdenite samples provided the first evidence for the natural fractionation of Mo isotopes with an observed fractionation of ~ 0.3%o/amu relative to a laboratory standard (JMC Mo). Dauphas et al. (2001) measured Mo isotope compositions with an IsoProbe MCICPMS. To determine the isotopic composition and atomic mass of Mo, they applied external normalization to 88Sr/86Sr using an admixed solution of NIST SRM 987 Sr. Siebert et al. (2001) used a Nu Plasma in conjunction with a low-flow nebulizer and a cyclonic spray chamber for measurements of Mo isotope fractionations in natu-
Figure 31.11 - Plot of 6-values for four Mo isotope ratios and three geological samples, where 6Mo is the relative deviation of the 98Mo/9XMo ratio of a sample from the JMC Mo standard in permil (after Siebert et al., 2001). The samples are a hydrothermal molybdenite (open square) and two sediments (open circle and filled diamond). All data were acquired with a double-spike technique. The 6-values are plotted vs. the mass difference of the corresponding Mo isotope ratios. The results of each sample are reasonably consistent for all Mo isotope ratios. The larger error bars for some ratios are due to isobaric interferences from Zr (92Zr+ and 94Zr+).
720
Chapter 31 - M. Rehk~imper,F. Wombacher& J.K. Aggarwal
ral samples. Correction for instrumental mass discrimination was performed with a 97Mo-100Mo double-spike. The double-spike was added prior to the separation of Mo from the samples by anion-exchange chromatographN such that a quantitative chemical yield was not required. Sample analyses typically consumed 1 to 3 gg of Mo. The external reproducibility for standard measurements was about 0.02 - 0.03%0/amu for the 98M0/95M0 ratio. Three natural samples (molybdenite, sediments) showed fractionations of up to 0.1%o/amu relative to a JMC Mo standard (Figure 31.11). 31.4.9 Mercury Mercury has the naturally occurring isotopes 196Hg (0.2%), 198Hg (10.0%), 199Hg (16.9%), 200Hg (23.1%), 201Hg (13.2%), 202Hg (29.9%), and 204Hg (6.9%). Due to the extremely high first ionization potential of Hg (10.3 eV) and its high volatility, isotopic measurements by TIMS are impossible. Previous studies of Hg isotopic compositions in natural materials used NAA (neutron activation analysis; e.g., Jovanovic & Reed, 1976), gas source mass spectrometry (Zadnik et al., 1989), Q-ICP-MS (Jackson, 2001) and HR-ICP-MS (Klaue & Blum, 2000). Several reports of isotopically anomalous Hg in meteorites have been published in the past, but these results were based on NAA analyses, which have large uncertainties (Lauretta et al., 1999). Jackson (2001) furthermore reported fractionated Hg isotopes for lake sediments and organisms, but the data have a precision of only about 3%o.
Lauretta et al. (2001) used a Plasma 54 for the measurement of Hg isotopic compositions in meteorites. Following acid-digestion of the samples, Hg was preconcentrated at high yield by cold-vapor generation. Samples were introduced into the plasma as elemental Hg (that was produced in a second vapor-generation step) entrained in a flow Ar. Mass discrimination correction was achieved by external normalization to T1, that was admixed as a dry aerosol produced with a desolvating nebulizer. Differences in the stable isotope compositions of samples and standards were reported to be resolvable down to the + 0.05%o level, if sufficient quantities of Hg (> 1 /~g) were available for analysis. Isotopic measurements were conducted for two carbonaceous chondrites, Allende and Murchison. The relative abundances of all seven stable Hg isotopes in both samples were shown to be identical to the terrestrial standard (elemental Hg from the Almaden mine, Spain) within 0.2 to 0.5%o. Previous reports of large anomalies in 196Hg/202Hg for these meteorites were thus thought to be erroneous. Fractionated Hg isotope compositions with isotopic variations of up to 0.25%o/amu were reported for samples of cinnabar and native mercury from various mines (Klaue & Blum, 2000). The variability of Hg isotopic compositions in these samples was attributed to evaporation/condensation processes during ore deposition. 31.4.10 Thallium Thallium has the two naturally occurring stable isotopes, 203T1 and 205T1, with abundances of 29.5% and 70.5%, respectively. Thermal ionization of T1 is straightforward (ionization potential = 6.1 eV), and TIMS measurements of T1 isotope compositions have achieved precisions (2 o) of ~ + 0.3-0.4%o/amu for terrestrial materials and meteorites (Arden, 1983a, b).
721
Stable Isotope Analysis by Multiple Collector ICP-MS i
i
i
i
I
i
i
i
i
I
i
t
i
i
I
~1
i
i
,
i
Figure 31.12- T1 isotope data for hydrogenetic FeMn crusts (Rehk~imper et al., 2001a), igneous rocks, standard solutions and the Allende meteorite (Rehk~imper & Halliday, 1999). The Fe-Mn crusts have 8205T1 values that are uniformly higher than the igneous rocks and Allende by about 1 to 1.5%o. This is thought to be due to the fractionation of T1 isotopes during the scavenging of T1 from seawater. The Fe-Mn crust data have smaller error bars (+ 0.05%o), because these analyses applied analytical techniques that were improved compared to the methods of Rehk~imper and Halliday (1999) which have uncertainties of about + 0.15%o.
Hydrogenetic Fe-Mn Crusts tO~
~~-4
,,
t---C.v--q ,,
V
I
-0.5
I
I
I
Igneous Rocks
Std. Solutions
Allende Chondrite I
0
I
I
I
I
I
I
I
0.5
i
i
I
1
,
i
1.5
6205T1 (%o) Rehk~imper & Halliday (1999) investigated the use of MC-ICPMS for the measurement of T1 isotopic compositions in natural samples. The analyses were performed with a Plasma 54 and a Cetac MCN-6000 desolvating nebulizer. For T1 isotope analyses of geological samples, both Pb and T1 were separated from the matrix by anionexchange chromatography. Correction for instrumental mass discrimination of T1 was achieved by external normalization to the 208Pb/206Pb isotope ratio of admixed Pb. The natural Pb that had been isolated from the sample was used for this purpose. The accuracy and precision of the analytical techniques were verified by the analyses of "synthetic" rock samples from which the natural T1 had been quantitatively removed by chemical techniques, and which were subsequently doped with the NIST SRM-997 T1 standard. Multiple measurements of such synthetic rock solutions yielded an average result of 8205T1 - - 0.03%0 _+0.08%0 (6205T1 is the deviation of the 205T1/203T1 isotope ratio of a sample relative to NIST SRM-997 T1 in %o) and this was within error of the expected result, 6205T1 - 0. The T1 isotopic compositions of several terrestrial rock samples and commercial standard solutions were reported (Rehk/imper & Halliday, 1999). The largest fractionations were observed for two ferromanganese crusts, which displayed 8205T1 values of up to + 1.2%o. This result has been corroborated by a detailed T1 isotope study of recent ferromanganese crusts growth layers, which were found to have 8205T1 values of between + 1.0 and + 1.4%o (Figure 31.12). These latter
722
Chapter 31 - M. Rehk~imper,F. Wombacher & J.K. Aggarwal
HF / Mannitol Water
Washout with Direct Injection Nebulizer"
A v
Figure 31.13 - Graph illustrating the memory problems associated with B analyses by MCICPMS (J.K. Aggarwal, unpublished results). Using a Cetac MCN 6000 desolvating nebulizer even prolonged rinsing was unable to reduce the B signal to << 1% of the original signal level, regardless of the composition of the wash solution. Similar results were obtained for a conventional glass aspiration system (data not shown). With a direct injection nebulizer (Cetac Microneb 2000), memory effects were effectively eliminated, regardless of the composition of the wash solution. measurements, which were conducted with a Nu Plasma MC-ICPMS and a desolvating nebulizer, were reported to have uncertainties (2 o) of about 0.025%o/amu (Rehk~imper et al., 2001a). The heavy T1 isotope compositions of the Fe-Mn crusts were thought to be the result of isotopic fractionation that occurs during the formation of the deposits by precipitation and scavenging from seawater. 31.4.11 Other Elements Boron has the two stable isotopes 10B and 11B that display relative abundances of 19.9% and 80.1%, respectively. A detailed B isotope study by MC-ICPMS has not been published to date, but preliminary results were presented by Aggarwal et al. (1999) and L6cuyer et al. (2000). Aggarwal et al. (1999) used a Micromass IsoProbe operated with helium as collision cell gas. Application of a direct injection nebulizer (Cetac
Stable Isotope Analysis by Multiple Collector ICP-MS
723
Microneb 2000) was found to effectively eliminate memory problems (Figure 31.13), while delivering a reasonable transmission efficiency (~ 3 x 10-11 A / p p m of B). A mass bias of ~ 40% was reported, with a drift of up to ~ 1%o during the course of a single measurement. A satisfactory correction for mass bias was achieved by external standardization with rapid alteration between samples and standards. Matrix effects on mass bias were found to be significant, such that chemical separation of B from water samples was necessary prior to analysis. L6cuyer et al. (2000) conducted their measurements with a Plasma 54 MC-ICPMS following chemical separation of B from the analyte matrix. Mass discrimination correction was achieved by external normalization to 26Mg/24Mg. A sensitivity of about I x 10-10 A / p p m was reported, such that at least 2 gg of B were required to achieve optimum reproducibilities (2 o) of about 0.3%o for sample analyses. Results were obtained for a number of carbonate and seawater samples.
Sulfur has four stable isotopes: 32S (95.0%), 33S (0.8%), 34S (4.2%) and 36S (0.02%), but most studies focus only on the determination of 34S/32S ratios. In a conference abstract, Mason et al. (1999a) presented a method that applied MC-ICPMS (Micromass IsoProbe) coupled with a UV laser ablation system for in-situ S isotopic measurements. The isobaric interferences of 02 + on 32S+ and 34S+ were reduced by the introduction of a He-Xe mixture (Mason et al., 1999b) into the hexapole collision cell of the IsoProbe. Chlorine was admixed to the ablated sample vapor by simultaneous aspiration of an HC1 solution. Correction for mass discrimination of S (~ 8%/amu) was achieved by external normalization to C1 isotopes. With this procedure, an internal precision of < + 0.15%o/amu (2 o mean) was reported for the analysis of sulfides, sulfates, and elemental S, prepared as pressed powder pellets. The accuracy of the MC-ICPMS technique was said to be good (+ 1%o) when sulfide standards or different pyrite samples where calibrated against each other, but matrix effects prevented the intercalibration between sulfides, sulfate and native sulfur. Selenium has 6 stable isotopes" 74Se (0.9%), 76Se (9.4%), 77Se (7.6%), 78Se (23.8%), 80Se (49.6%) and 82Se (8.7%). Rouxel and co-workers (Rouxel et al., 2000; Rouxel et al., 2001) recently developed methods for Se isotope analyses of natural samples using an IsoProbe MC-ICPMS. The standard-sample bracketing technique was used for mass discrimination correction. An external analytical precision (2 o?) of better than 0.3%o and 0.8%o was obtained for 82Se/ 80Se and 82Se/76Se, respectively (~ 0.15%o/ amu). The variation of Se isotope compositions in massive sulfides and active chimneys from different submarine hydrothermal fields was reported to be about 1%o/amu and distinct Se isotopic fingerprints were obtained for different types of deposits.
Cadmium possesses eight stable isotopes: 106Cd (1.3%), 108Cd (0.9%), 110Cd (12.5%), 111Cd (12.8%), 112Cd (24.1%), ll3Cd (12.2%), ll4Cd (28.7%), and 116Cd (7.5%). The Cd isotope compositions of both terrestrial samples and meteorites have been measured by MC-ICPMS (Wombacher et al., 2000; Wombacher et al., 2001). The analyses were conducted with a Nu Plasma instrument, using the 107Ag/109Ag ratio of admixed Ag and the empirical normalization procedure of Mar6chal et al. (1999) for mass discrimination correction. Samples of Cd metal, which were residues of evaporation experi-
724
Chapter 31 - M. Rehk~imper, F. Wombacher & J.K. Aggarwal
Figure 31.14 - Comparison of Cd isotope data (in 6Cd/amu-notation) for three chondritic meteorites acquired by MC-ICPMS (Wombacher et al., 2000; Wombacher et al., 2001) with TIMS results obtained with the double-spike technique (Rosman et al., 1980; Rosman & De Laeter, 1988). 6Cd is the relative deviation of the 11xCd/114Cd ratio of a sample from a terrestrial Cd standard in permil. The excellent agreement of the results endorses the accuracy of the MC-ICPMS analyses.
ments, were found to be enriched in the heavy isotopes compared to the starting material by up to 10%o/amu. Repeated measurements ( n - 28) of a fractionated metal sample yielded a long-term reproducibility of _+0.07%0/amu (2 0). Terrestrial minerals and rocks displayed no or only small (< 0.25%0/amu) mass-dependant variations of Cd isotope compositions. Data obtained for chondritic meteorites, however, revealed evaporation/condensation related Cd isotope fractionations ranging f r o m - 2.7 to + 3.9%o/amu relative to the terrestrial standard (JMC Cd). The agreement of the MCICPMS data with previous Cd isotope results obtained by TIMS and the double spike technique for three meteorites was reported to be excellent (Figure 31.14).
Tin has ten stable natural isotopes: 112Sn (1.0%), 114Sn (0.7%), 115Sn (0.3%), 116Sn (14.5%), 117Sn (7.7%), 118Sn (24.2%), 119Sn (8.6%), 120Sn (32.6%), 122Sn (4.6%), and 124Sn (5.8%). Yi et al. (1999) used a Plasma 54 MC-ICPMS to measure the Sn isotopic compositions of archeological artifacts and experimentally prepared bronze alloys. Correction for instrumental mass discrimination was obtained by external normalization to Sb and measurement uncertainties of N + 0.08%o/amu (2 c~) were obtained on 124Sn/ 127Sn for multiple measurements of a standard. The Sn from archeological samples was separated from the matrix by ion exchange chromatography prior to analysis (Yi et al., 1995). No significant Sn isotope anomalies were identified in the bronze artifacts or could be detected as the result of laboratory-induced tin evaporation.
Stable IsotopeAnalysisby MultipleCollectorICP-MS
725
Antimony has the two stable isotopes 121Sb and 123Sb, with relative abundances of 57.2% and 42.8%, respectively. In-situ isotopic analysis of antimony in samples of Sbbearing native Ag and sulfide minerals by laser-ablation MC-ICPMS were attempted by Jackson et al. (1999b). The measurements were conducted with a Nu Plasma coupled to a UV laser system. Tin was introduced into the sample carrier gas (as a dry aerosol) for the correction of Sb mass discrimination by external normalization. Variations of up to 2%0 in the raw 121Sb/123Sb ratios during laser ablation analysis of a single (presumably homogeneous) sample were reported. The measured Sb isotope ratios were negatively correlated with signal intensity; in contrast the 124Sn/122Sn ratios displayed only minimal or no correlation with signal size. These observations indicate that precise and accurate correction for mass discrimination using an "external" correction procedure may not straightforward during laser ablation analyses if the elements are derived from different sources and/or variable elemental ratios are observed during a run. 31.5 Conclusions
MC-ICPMS offers a number of key advantages over existing methods for stable isotope ratio measurements. The ionization efficiency of the plasma source permits the study of a wide range of elements, the high sensitivity allows the analysis of small samples and the precision achievable by MC-ICPMS is suitable for the resolution of the small isotopic differences that occur in geological samples. In-situ stable isotope studies are, in principle, also possible by MC-ICPMS with a laser-ablation sample introduction system. Compared to state-of-the-art TIMS techniques, MC-ICPMS achieves similar or superior precision, but provides simplified sample preparation and significantly faster sample throughput. By allowing the measurement of natural isotopic variations that were previously too small to be resolved readily with existing methods, MC-ICPMS is opening up new avenues of stable isotope research in both geo- and cosmochemistry. A survey of the recent literature suggests that the field of "heavy-element" stable isotope research will continue to grow in the coming years. Highly desirable further developments in MC-ICPMS include improvements of instrumental sensitivity to permit the analysis of even smaller samples, and a better understanding of the mass discrimination process a n d / o r a reduction of the instrumental mass bias to provide a more robust physical framework for accurate and precise data acquisition.
Acknowledgments We thank K. Mezger, C. Mfinker, E. Scherer and S. Weyer from M/.inster and H. Baur, M. Fehr, M. Frank, A. Halliday, D.-C. Lee, S. Levasseur, R Oberli, M. Sch6nb~ichler, C. Stifling, N. Teutsch, T. van der Fliert, U. Wiechert, and S. Woodland from Z~rich for unpublished data, informal manuscript reviews as well as insightful discussions. Saskia Goes is gratefully acknowledged for mathematical support. Alex Halliday and an anonymous reviewer provided critical and very helpful formal manuscript reviews. This paper also profited immensely from numerous discussions with and support from many other (MC-) ICPMS users.
Handbook of Stable Isotope Analytical Techniques, Volume 1 P.A. de Groot (Editor) 9 2004 Elsevier B.V. All fights reserved.
CHAPTER 32 Different Isotope Ratio Measurement Applications for Different Types of ICP-MS: Comparative Study of the Performance Capabilities and Limitations Christophe R. Qu6tell & Jiirgen Diemer European Commission, Joint Research Center, Institute for Reference Materials and Measurements, Retieseweg, B-2440Geel, Belgium e-mail: 1 [email protected]
32.1 Introduction Inductively coupled plasma mass spectrometry (ICP-MS) can be reliably used for the measurement of isotope ratios. The argon ICP ionisation source operated at atmospheric pressure is very flexible, but plasma flicker is a noticeable source of ion beam instabilities and can also be seen as a limitation for this mass spectrometry technique. Other instrumental sources of potential measurement bias include sample introduction systems, instrumental background, mass spectroscopic interferences and matrix effects, and the way ions are focused in energy, transmitted and detected. Instruments of very different designs and characteristics are commercially available. The most widely used designs are the quadrupole (ICP-QMS) and the double focusing magnetic sector (ICP-SMS) instruments. The quality of the performance of these various types can be evaluated relative to each other from the comparison of the respective measurement combined uncertainties. Although important, the instrument repeatability (i.e. within-run relative standard deviation of the mean) is not necessarily the unique and often not even the most important source of uncertainty for the overall measurement. This depends on the nature of the study and the application foreseen as well as on the type of sample and sample preparation involved. Typically, the instrument repeatability on isotope ratio measurements that can be achieved in a reproducible manner is 0.1 - 0.35% for ICP-QMS instruments (Qu6tel et al., 1997) and down to ~ 0.05% for single detector ICP-SMS (Qu6tel et al., 2000a). Many of the instabilities mentioned above can be reduced by the combination of a double focusing action on the ion beams and the simultaneous acquisition of the concomitant ion signals using an array of several detectors. Hence, multiple Faraday collector instruments (MC-ICPSMS) have the potential to perform isotope ratio measurements with repeatability values down to 0.01% or below (Taylor et al., 1995). These different types of instruments do not represent the same financial investment (difference of a factor ~ 3 to 6 or more), they cannot be operated equally with the same flexibility and they do not necessarily require the same operating skills.
Different Isotope Ratio Measurement Applications for Different Types of ICP-MS ...
727
In this chapter, the performance exhibited by four different commercial ICP-MS instruments of the three different types mentioned above are compared, using identical samples, first for isotope ratio measurements and second for isotope dilution mass spectrometry (IDMS) measurements. Finally, the specific measurement capabilities of a MC-ICP-SMS instrument are illustrated. The experimental results reported include Cu and U isotope ratio measurements. Obviously uranium is not a 'stable isotope' element and it is likely that spectral interferences will be much more common during ICP-MS measurements for'stable isotope' elements than for uranium (heavier). Moreover, MC-ICP-SMS isotopic measurements for very light elements such as boron and lithium might require specific technical features (large flight tube) for an optimum performance. However, most of the concepts described throughout the uranium experiments as well as most of the conclusions reached can be transferred to light 'stable isotopes'.
32.2 Experimental The experimental results introduced here are partly based on uranium (Qu6tel et al., 2000a, b) and copper (Diemer et al., 2002) measurements published elsewhere. Experimental conditions are described in detail in these references and will only be summarised here. 32.2.1. Instrumentation and measurement procedures The measurements were performed on two quadrupole ICP-MS and two magnetic sector ICP-MS instruments: respectively a "PQ2+" (VG Elemental, Winsford, England) equipped with turbo pumps, an "Elan 6000" (Perkin-Elmer Sciex, Norwalk, USA), an "Element2" (Finnigan MAT, Bremen, Germany) and a "Nu Plasma" (Nu Instruments, Wrexham, Wales).
The main characteristics of all instruments are described in Table 32.1. The ICPQMS instruments are of different generations and manufacturers, and hence of different design. The "PQ2+" is fitted with an extraction lens followed by a stack of several symmetrical electrostatic lenses and a photon stop. In comparison, the ion focusing system of the "Elan 6000" is somewhat simpler as it is composed of a photon stop and one cylinder lens with only one adjustable parameter. The "AutoLens" feature on the "Elan 6000" (for dynamic adjustments of the lens potential) was always disabled during the measurements. Both ICP-SMS instruments use double focusing Nier Johnson geometry (forward for the "Nu Plasma" and reverse for the "Element2"). With the "Element2" isotopic ratio measurements are performed by rapidly scanning a combination of the acceleration and electrostatic filter high voltages (ion energy scanning over a limited mass-range) at constant magnetic field. This instrument has variable mass resolution capabilities (3 settings, up to ~ 12000) whereas the others can only be operated at low resolution. Measurements with the "PQ2+", the "Elan 6000" and the "Element2" are sequential (i.e. only the signal of one isotope is measured at a time and the same single detector is used for all isotopes) and, usually, 'digital'. A dual mode detector is available on the last two instruments that allows comparisons with 'analog' acquisitions. This con-
728
Table 32.1 - Main instrumental characteristics between three different types of ICP-MS (four instruments all together) ICP
PQ2+
27MHz crystal :ontrolled
Elan 6000
40MHz free running
Element2
27MHz crystal :ontrolled
RF power Scan type
- 1400 - 1100
- 1230
Points per mas5 peak
Sampling Ni cone (mm)
RF-DC voltages 12 Peak jump mode Quadrupole
3
1.0
0.7
-5
- 0.8 u
RF-DC voltages 0.2 Peak jump mode Quadrupole
1
1.1
0.9
-5
- 0.8 u
~
27MHz -1300 crystal :ontrolled
Noscan
5 to 10
1 High voltages
-
-
2 Magnet
1.1
Skimmer Abundance Resolution Ni cone sensitivity (mm) (mass 237 / 238U) 106 (PPm)
0.8
- 12
300
Detector
Dead time (ns) *
Channeltron (Galileo8 Kore Tech., Cambridge, UK) Discrete dynode EM (Em Ringwood, Australia) Discrete dynode EM (Mascom, Bremen, Germany)
59
16
. .
1.0
0.7
-5
- 400
Faraday cups
+
Discrete dynode EM (ETP, Ringwood, Australia)
* Dead time values were determined by means of methods described in Nelms et al. (2001) (methods 2 and 4 herein were used).
49
Chapter 32 - C.R. Qu6tel & J. Diemer
Combination acceleration + electrostatic filter high voltages (magnetic field
~.
Nu Plasm6
Settle time (ms)
Different Isotope Ratio MeasurementApplications for Different Types of ICP-MS ...
729
trasts with the "Nu Plasma" where the ion beams are simultaneously focused, via variable dispersion optics, onto a fixed array of 12 Faraday cups ('analog' acquisition) a n d / o r a fixed array of 3 ion counting detectors ('digital' acquisition). Correction for dead time effects (Knoll, 2000) were made whenever necessary. Dead time values and their associated standard uncertainties were determined by means of methods described elsewhere (Nelms et al., 2001). The acquisition parameters were optimised (Koirtyohann, 1994; Begley & Sharp, 1994, 1997; Qu6tel et al., 1997) to ensure sufficiently small and reproducible repeatabilities. Routine experimental settings were applied throughout the measurements. Rinsing followed by a blank measurement was performed prior to every sample measurement in order to monitor and correct for sample to sample memory effects.
32.2.2 Reagents and solutions All samples were diluted in 2% HNO3 and sample dilution level was adjusted depending on instrument sensitivity and mode of detection. The isotopic reference materials (IRM) IRMM-072/1 to IRMM-072/10 (except IRMM-072/5 and IRMM/072/8) were measured. The IRMM-072 series have been described in detail previously (Rosman et al., 1987). The following dilutions levels were prepared: ~ 10 ng U 9 g-1 for the ICP-QMS experiments, ~ 1 ng U 9 g-1 for the single detector ICP-SMS experiments and ~ 1000 ng U 9 g-1 for the MC-ICP-SMS experiments. More details about the samples and their preparation can be found elsewhere (Qu6tel et al., 2000b). For copper IDMS measurements, water samples with a matrix close to drinking water were blended gravimetrically with aliquots of the 65Cu enriched IRMM-632 IRM. The natural Cu IRM IRMM-633 was used to correct for mass-discrimination. Identical blend samples at different dilution levels were used" ~ 5 to 30 ng Cu 9 g-1 for the ICP-QMS, ~ 3 to 30 ng Cu 9 g-1 for the single detector ICP-SMS experiments and ~ 30 to 150 ng Cu 9 g-1 for the MC-ICP-SMS experiments. The sample preparation is explained in more details elsewhere (Diemer et al., 2002). Finally, for the last two sets of experimental results reported, 500 to 1000 ng U 9 g-1 solutions were prepared both for the nuclear safeguards samples and the IRMM184 "natural like" uranium certified IRM1 used to correct for mass-discrimination effects. 32.3 Mass-discrimination effects The possible or demonstrated sources of mass-discrimination during ICP-MS measurements were reviewed recently (Heumann et al., 1998; Qu6tel et al., 2000b). Common to all types of ICP-MS instruments are the effects induced by electrostatic interactions between ions of like charge, e.g. space-charge effects in the high ion den-
1. Provisional revised version (30 January, 2001) of the certificate of isotopic composition of IRMM-184 (original version available at www.irmm.jrc.be).
730
Chapter 32 - C.R. Qu6tel & J. Diemer
sity beams leaving the skimmer cone. Other effects, specific to the design and operation of the mass spectrometer, might also have a significant impact on the measurement results, as shown later for the single detector ICP-SMS. Two basic ways of correcting for mass-discrimination will be illustrated. In both cases, a ratio (K factor) between the certified (RCert) and the measured (RMeas) values of a known isotope ratio is calculated (equation [32.1]): K = RCert / RMeas
[32.1]
The first generic approach is an internal correction. A known isotope ratio is measured together with the unknown isotope ratio to be corrected (same run). A K factor is calculated according to equation [32.1] and a normalised mass-discrimination factor is derived as a function of K and the masses of the isotopes of the known ratio. The normalised value is then used to correct the measured value of the unknown ratio for mass-discrimination. Ideally, the isotopes of the known and the unknown ratios belong to the same element (n(146Nd)/n(144Nd) ratio to correct other Nd ratios, for instance: Platzner et al. (1997)). Otherwise, an element of known isotopic composition and close in mass to the element of interest is added to the sample (unless already present at measurable levels): for instance, the n(203T1)/n(205T1) ratio has been proposed to correct Pb isotope ratio measurements (Longerich et al., 1987). For normalisation, Taylor et al. (1995) described three possible models of which two, the well known linear law and power law, are reported in equations [32.2] and [32.3]. The third model proposed by Taylor et al. (1995) appears to give nearly identical results to those from the second model and hence is not reported here. The exponential model described by Mar6chal et al. (1999) is preferred instead (equation [32.4]): Kli n - 1 + (m 2 - ml) " 81in
[32.1]
Kpo w = (1 + Epow)(m2- ml)
[32.2]
K exp - (ma/ml)~ exp
[32.3]
in which ml and m2 are the atomic masses of the two isotopes 1E and 2E involved in the ratio n(1E)/n(2E), ~;lin, Epow and eexp are the normalised mass-discrimination factors for the three models respectively. Clearly, the internal approach is valid only if there is no significant mass or element dependency of the mass-discrimination factors (at least for a limited mass-range). The second generic approach consists of correcting externally for mass-discrimination, using an IRM measured before a n d / o r after the unknown sample. Ideally, the same pair of isotopes from the same element is measured in both solutions and both ratio values are almost identical. The unknown ratio is corrected by multiplication with the K factor (equation [32.1]) calculated using the IRM measurement result. In case only the IRM of another element is available, the K factor is calculated by application of one of the models described in equations [32.2-4:]. Element concentration and
Different Isotope Ratio MeasurementApplications for Different Types of ICP-MS ...
731
sample matrix conditions in the unknown sample must also be matched in the IRM, although in the case of very low element concentration in the sample higher element concentration in the IRM to achieve sufficient counting statistics might also be advisable. Obviously, applying this approach will depend on the availability of an appropriate IRM, adapted to the characteristics of the unknown sample measured. Moreover, mass-discrimination and any other effect monitored by the IRM must remain as stable as possible over the time spent at measuring the IRM and the unknown sample (incl. rinsing times, blank acquisitions etc...). Despite these possible limitations, external calibration remains an attractive approach particularly when it is difficult or impossible to apply internal calibration a n d / o r the mass-discrimination effects are not sufficiently mass or element independent.
32.4 Uncertainty calculations Combined uncertainties attached to the measurement results were calculated according to the ISO/GUM guide (GUM, 1995) by applying an uncertainty propagation procedure to individual uncertainty contributions. This is very different from common practice in isotope ratio work, where often repeatabilities or at best reproducibilities are quoted. For the results presented here, a dedicated software program1, based on the numerical method of differentiation described by Kragten (1994), was used. Unless otherwise specified, all calculated combined uncertainties are expanded with a coverage factor k - 2. "Additive" corrections commonly applied to the individual isotope signal intensities measured include the instrumental background, the isobaric interferences and the dead time effects to some extent ("semi-additive" correction). They do not mutually cancel when isotope ratios are calculated and the uncertainty contributions arising from these corrections can be significant as will be shown later. However, the common practice is that these uncertainties are not taken into account possibly because propagating them directly would require involving the repeatabilities of the measurements of the individual isotope signal intensities and not those of the ratios directly. It is well known that the repeatabilities of the intensities of a ratio are significantly worse than those of the ratio itself. Hence, there is a risk of exaggerating the combined uncertainty of the evaluated measurements. In our experiments, in order to avoid this risk, "additive" corrections on intensities were translated into multiplicative corrections on ratios for the combined uncertainty calculations according to a method described elsewhere (Qu6tel et al., 2001).
32.5 Uranium isotopic ratio measurements using ICP-MS: a comparison From IRMM-072 / 1 to IRMM-072 / 10 the n(233U) / n(235U) and n(233U) / n(238U) isotope ratios varied from ~ 1 to ~ 2.10-3 (certified to within 0.03%, k - 2). For each sample the n(235U)/n(238U) ratio was always close to unity (certified to within 0.02%, k = 2) and was measured to enable the measured values of the other two ratios to be corrected internally for mass-discrimination effects. Similar types of calculations were performed for all four instruments (only the model differed, e.g. linear - equation 1. GUM Workbench| Metrodata GmbH, D-79639Grenzach-Wyhlen, Germany.
732
Chapter 32 - C.R. Qu6tel & J. Diemer
Different Isotope Ratio Measurement Applications for Different Types of ICP-MS ...
733
Figure 32.1 - Data "PQ2+" (reproduced from Qu6tel et al., 2000b). Cf. comments Figure 32.4 below. Figure 32.2- Data "Elan 6000" (reproduced from Qu6tel et al., 2000b). Cf. comments Figure 32.4 below. Figure 32.3 - Data "Element2" (reproduced from Qu6tel et al., 2000a). Cf. comments Figure 32.4 below. Figure 32.4- Data "Nu Plasma" (reproduced from Qu6tel et al., 2000b). Comments Figures 32.1 to 32.4: Difference between the certified and the experimental rl(233U)/n(235U) isotope ratio values for IRMM-072 samples in the range N 1 to N 2.10-3. Measurement results are corrected for mass-discrimination effects, using the linear law model for the "PQ2+", the "Elan 6000" and the "Element2", and using the exponential law model for the "Nu Plasma". The vertical bars and the horizontal dotted lines represent the expanded uncertainties (k = 2) for the experimental and the certified values respectively. [32.2] - for the three single detector ICP-MS a n d exponential - e q u a t i o n [32.4] - for the MC-ICP-SMS). Three u n c e r t a i n t y c o m p o n e n t s w e r e c o m b i n e d together to build u p the final u n c e r t a i n t y on the corrected m e a s u r e m e n t results. For the n(233U)/n(235U) ratio a n d for the n(233U)/n(238U) ratio, the repeatability of their respective m e a s u r e m e n t s w a s c o m b i n e d together w i t h the n(235U)/n(238U) ratio m e a s u r e m e n t repeatability a n d the s t a n d a r d u n c e r t a i n t y of the certified value of the s a m e n(235U)/n(238U) ratio. As illustrated in Figures 32.1 to 32.4 for the n(233U)/n(235U) ratio, all the corrected m e a s u r e d values agreed well w i t h i n e x p a n d e d uncertainties w i t h the certified values (similar results for the n(233U)/n(238U) ratio). E x p a n d e d uncertainties ( k - 2) on m e a s u r e m e n t s p e r f o r m e d w i t h the multiple collector i n s t r u m e n t varied from + 0.04% to + 0.24% for the n(233U) / n(235U) ratio (from + 0.08% to + 0.27% for the n(233U)/n(238U) ratio). Typically, they w e r e ~ 1 to 5 times larger w i t h the single detector m a g n e t i c sector i n s t r u m e n t , a n d ~ 10 to almost 25 times larger w i t h b o t h q u a d r u p o l e instruments. The closer to u n i t y the ratio, the better the p e r f o r m a n c e of the MC-ICP-SMS relatively to those from the other types of ICP-MS
734
Chapter 32 - C.R. Qu6tel & J. Diemer
instrumentation (and the smaller the combined uncertainties of the measurements). For the smallest ratios, the combined uncertainty of the measurements was very similar for both magnetic sector instruments. These results illustrate the capacity of an instrument like the "Element2" to measure a ratio between a 235U signal which equivalent concentration in solution is ~ 500 pg 9 g-1 and a 233U signal which equivalent concentration in solution is only ~ 1 pg 9 g-1 (n(233U)/n(235U) ~ 2,10 -3 in IRMM-072/ 10 sample). For all the instruments, repeatability of the n(233U)/n(235U) ratio measurement was the main source of uncertainty of the three taken into consideration. For ratios close to unity, the counting statistics represented only 33% of the single detector ICP-SMS measurement repeatability. For the smallest ratios the same theory could explain 100% of the observed repeatability (0.16%), thus indicating an optimum performance of this instrument and leading to an estimated combined uncertainty between 0.30% - 0.35%. Therefore, this range could be considered to be the best achievable uncertainty values for measurements of ratios of nearly 3 orders of magnitude performed with the single detector ICP-SMS instrument. An average value of the normalised mass-discrimination factor (~lin) w a s calculated from the four series (i.e. using the linear model for every instrument) of eight n(235U)/n(238U) ratio values successively measured (Table 32.2). These average ~lin values probably reflect the similarities and differences between the various ICP-MS designs. Results from the "PQ2+" and the "Elan 6000" are almost identical. Both magnetic sector instruments exhibit larger (N 2.5 times for the "Element2" and ~ 5 times for the "Nu Plasma") but more reproducible apparent mass-discrimination than the quadrupole instrument over approximately 4-6 hours of operation. Results obtained with the "Nu Plasma" can be compared with results obtained on a "Plasma 54" (first generation of MC-ICP-SMS) by Taylor et al. (1995) for the same IRMM-072 samples. An apparent normalised mass-discrimination factor of 0.97% is calculated (linear model) for sample IRMM-072/1 from the n ( 2 3 3 U ) / n ( 2 3 5 U ) ratio data reported in their Table 1. It is larger than the average 0.67% calculated for the "Nu Plasma" using the n(235U)/n(238U) ratio data, and thus the difference with ICP-QMS mass-discrimination values is confirmed. Burgoyne et al. (1997) formulate the hypothesis that the size of mass-discrimination decreases with the acceleration voltage. If space charge effects were the main source of mass-discrimination for all the instruments considered, and knowing that the ions are transferred at a slower pace in ICP-QMS than in ICP-SMS instruments, our results are in contradiction with this hypothesis. It must also be noted that theoretical considerations on quadrupole mass spectrometry predict that the heavier the ion the longer the time spent in the fringing fields and, therefore, the greater the dispersion in the quadrupole field (Dawson, 1995). This is opposite to the mass-discrimination induced by the space charge effects. In our results, the relative standard uncertainty measured for the apparent mass-discrimination effects appears to be smaller on magnetic sector than on quadrupole ICP-MS instruments. Finall~ it can be seen in Figure 32.3 that the results obtained with the "Element2" were not random and were always negative relative to the certified values. This indicates the presence of a measurement bias not properly corrected for, and thus not properly taken into account in the estimation of the combined uncertainty budget.
Different Isotope Ratio Measurement Applications for Different Types of ICP-MS ...
735
Possible sources of measurement Table 32.2 - Comparison of the average apparent normabias were examined (Qu6tel et al., lised mass-discrimination factor calculated (linear model) 2000a). Calculations to simulate 1 from eight independent measurements of the n(235U)/ ns change in the dead time value n(238U) ratio for three different types of ICP-MS (four (15 ns instead of 16 ns) showed instruments: "PQ2+", "Elan 6000", "Element2" and "Nu Plasma") u p wa r d shifts on ratios (up to 0.1% on ~ 2.10-3 ratios). The pos+ rel. SD* (%) + SD* l~lin sibility of a contamination was (n=S) (n=S) examined. The simulation con24.5 0.143 0.035 sisted in subtracting the signal PQ2+ 16.7 0.023 0.138 Elan 6000 corresponding to the addition of 5.1 0.017 0.332 Element2 only 100 pg natural uranium to 1.5 0.010 0.673 Nu Plasma the measured samples. The ratios were shifted by up to 0.2% but, * SD stands for Standard Deviation. because it only increased the difference from the certified values, such hypothetical contamination could be ruled out. Correction for mass-discrimination was also carefully examined. Using the exponential model, average values of the normalised mass-discrimination factor versus average mass for all three uranium ratios are compared for the three types of instrument (Figure 32.5a-c). Data from the ICP-QMS instrument almost "perfectly" fall on a horizontal line (slope = 8,10-4, R2 = 0.99). This is also the case for the MC-ICP-SMS instrument (slope - - 71 ~ 10-4, R2 - 1.00). The data calculated for the single detector ICP-SMS also fall on a straight line but the slope is 64 times larger (absolute) than for the ICPQMS instrument (slope - - 51 ~ 10-3, R2 - 1.00). These results seem to indicate that there is more than only one major source of mass-discrimination with an instrument of the type of the "Element2", with variable size depending on the ratios measured. This might be explained by the Liouville's theorem (Brunn6e, 1987) according to which the product of the aperture angle of the ion beam and the square root of the acceleration voltage is constant for a given ion optical geometry. During isotope ratio measurements, decreasing the combination of high voltages (ion energy) at constant magnetic field to select heavier masses increases the aperture angle and can lead to a reduction of the total transmission (when slits and apertures in the mass analyzer are illuminated in the x and y directions). Thus, the result is a mass-discrimination in favor of light mass particles and opposite to the general trend observed in a typical instrument "response curve" across the full mass spectrum (Qu6tel et al., 2000a). Hence, when performing isotope ratio measurements with an instrument like the "Element2", internal correction for mass-discrimination using either another pair of isotopes of the same element or of another element can easily lead to erroneous results, i.e. combined uncertainties of measurement results can be seriously underestimated. In this case, the safest approach (but more complicated) is to apply external calibration, using isotopic reference materials certified for the same ratio value, with the same pair of isotopes and prepared under identical conditions of element concentration and sample matrix as for the sample measured.
736
Chapter 32 - C.R. Qu6tel & J. Diemer
32.6 Copper IDMS measurements using ICP-MS: a comparison The copper mass fraction in drinking water samples was determined, based on the measurement of isotope ratios, using "direct" IDMS. Beside the comparison for this application of the same three types of ICP-MS instruments, the respective merits of the different modes of acquisition ('digital'/'analog') were evaluated for the 2 types of single detector instruments (Diemer et al., 2002). "Direct" IDMS, described in equation [32.5] below, requires a "spike" material for the blending certified for its own element mass fraction. It is experimentally more straightforward to implement than "double IDMS" where a "back-spike" material is needed to characterize the "spike" material. More details can be found elsewhere (De Bi6vre, 1994).
my (Ry-(acert/aMeas)'RB) Cx - Cy " G "
(Rcert~):G'--'G
z(ai) X ablank
"2(Ri)y mx
[32.5]
Cx and CY a r e the element mass fractions in the sample and the spike solutions; ablank is the absolute mass of the element measured in the procedural blank; mx and my are the masses of the sample and the spike respectively; Rx, Ry and RB are the sample, spike and blend ratios and X(Ri)x and X(Ri)y are the sum of the ratios in the sample and the spike solutions (all referenced to the same isotope of the element). All (Ri)Y values are certified and, in the case of Cu, all (Ri)x values can be calculated from data in column 9 of the IUPAC table (Rosman & Taylor, 1998). The blend isotope ratio was corrected externally for mass-discrimination (comparison between the certified ratio and the measured ratio of the IRMM-633 reference material, Rcert and RMeas respectively). In practice, measurements of the blend samples were bracketed with measurements of the IRMM-633 and time weighed average values were calculated for the correction for mass-discrimination. Uncertainties associated to all the variables and parameters described in equation [32.5] were taken into account (repeatability of the ICP-MS measurements were divided by square root of the number of respective measurement replicates). Other sources of uncertainty considered included those applied to the individual isotope signal intensities described earlier e.g. the correction for instrumental background and the correction for dead time effects. Sodium was found to be present in the samples at a concentration exceeding 70 times the Cu concentration. Under these conditions significant amounts of ArNa + are formed during the ICP-MS measurements. Instead of correcting the resulting isobaric overlap on mass 63 directly, the overall effect on the n(63Cu)/n(65Cu)blend ratio was investigated using two independent approaches (Diemer et al., 2002) and a correction factor (0.9994 + 0.0005) was eventually applied to all Cu blend ratios measured. Whereas in many cases sample preparaFigure 32.5a-c - Average values of the normalised mass-discrimination factor (exponential model) for the n(233U)/n(235U), n(233U)/n(238U) and n(235U)/n(238U) ratios plotted against their corresponding average masses for each type of instrument. Vertical bars correspond to the relative standard deviation for the first six samples (IRMM-072/1 to IRMM-072/7)(reproduced from Qu6tel et al., 2000b).
Different Isotope Ratio Measurement Applications for Different Types of ICP-MS ...
737
738
Chapter 32 - C.R. Qu6tel& J. Diemer
tion represents a major source of uncertainty that has an overruling effect irrespective of the type of ICP-MS instrument used during the measurements, this did not apply here since the sample did not require particular treatment. As shown in Figure 32.6, there was a very good agreement, within uncertainty, between the results obtained using different instruments/instrumental settings. The measurements performed with the "Nu Plasma" consistently produced the smallest combined uncertainty although the difference to the results of the single detector machines is not very important, particularly when they are operated in 'analog' mode. In Figure 32.7, the distribution of the various uncertainty contributions are compared for the following experimental conditions" 5 ng Cu 9 g-1 _ 'digital' mode with the "Elan 6000", 3 ng Cu 9 g-1 _ 'digital' mode with the "Element2" and 150 ng Cu 9 g-1 _ 'analog' mode with the "Nu Plasma". Radical differences can be observed between the three types of ICP-MS instruments. The contributions from the repeatability on all isotope ratio measurements (blend samples and reference material used for the determination of the mass-discrimination) account for almost 30% in the case of the ICP-QMS but are nearly negligible for the magnetic sector instruments. Although often ignored or neglected, the correction for dead time effects can represent a major source of uncertainty as it is illustrated in this study with both single detector systems (~25% for the "Elan 6000" and ~60% for the "Element2"). With the "Nu Plasma" the significant sources of uncertainty are not instrumental and the main ones are by far (~80%) the uncertainties on the spike CRM components (mostly Cy). These contributions are also important for the "Elan 6000" (~40%) and the "Element2" (~30%). Finally, the relative impact of the correction for the ArNa interference remains limited and identical for the three series of uncertainty calculations. Thus, the quality of the spike material appears to be absolutely critical and the use of an expensive instrumentation such as a MC-ICP-SMS for "direct IDMS" applications appears to be not justified unless high-quality CRM's are available and the sample matrix does not require a high uncertainty chemical treatment. It must also be noted that for "double IDMS" measurements the quality of the "back-spike" material used to characterize the spike material is primordial. 32.7 Characterization of small isotope ratio differences using a double focussing MC-ICP-SMS Double focusing MC-ICP-SMS has the potential of delivering measurement results of very high quality (e.g. small combined uncertainty) providing that uncertainty contributions other than the measurement repeatability itself are sufficiently low. This is illustrated with the following example taken from the area of environmental sampling nuclear safeguards. Accurate nuclear accounting of uranium and the capacity to detect undeclared nuclear activities have become crucial for the international community over the last decade. One of such nuclear safeguarding activities consists in measuring uranium isotope ratios in "environmental" samples collected inside or outside a given facility. Being able to measure "quasi-natural" uranium isotopic signatures in a reliable way can be used to identify the presence and the origin of nuclear contami-
Different Isotope Ratio Measurement Applications for Different Types of ICP-MS ...
739
Figure 32.6 - Concentration of copper in the water sample measured by IDMS using three types of ICP-MS instruments and under different modes of acquisition. Vertical bars correspond to expanded uncertainties (k =2). D: 'digital' acquisition mode (pulse counting); A: 'analog' mode; MR: medium mass resolution mode (reproduced from Diemer et al., 2002).
nation in areas possibly linked to non declared nuclear activities. In this context, very important political decisions can derive which will have to be based on measurement data of non-disputable uncertainty. For quality control of their environmental sampling programme, the International Atomic Energy Agency (IAEA) asked IRMM to prepare a set of uranium Certified Test Samples (CTS) with isotopic abundances close to "natural" uranium and forming a sequence of increasing 235U enrichments with relative increments of only ca. 1%. These CTS with n(235U)/n(238U) ratios certified to - 0.05 % ( k - 2) were made up of samples prepared by isotopic mixing of certified uranium IRMs in the gas phase as UF6 (De Bolle et al., 1999). Gas-source electron impact mass-spectrometry (GSMS) was used for the certification of the n(235U)/n(238U) ratios (~ 0.007 range). Solid source thermal ionisation mass-spectrometry (TIMS), after hydrolysis of the UF6 sample and transforming to the nitrate form by treatment with nitric acid, was used for the certification of the n(234U)/n(238U) (~ 0.00005 range). For these samples, all ratio values were verified independently outside IRMM by TIMS measurements, as well as inside IRMM with the "Nu Plasma". With the MC-ICP-SMS, signal intensities from 234U, 235U a n d 238U were acquired simultaneously using a combination of, respectively, one ion counting detector and two Faraday cups. The correction for mass-discrimination was performed externally by running the IRMM-184 uranium certified IRM. The contributions to the combined uncertainties of the measurement results included the repeatability of the measured ratios in the samples, the repeatability of the measure-
740
Chapter 32 - C.R. Qu6tel & J. Diemer
Figure 32.7- Main contributions to the expanded uncertainty (k = 2) on the measurement of the copper concentration for three types of ICP-MS instruments. Copper concentration (ng ~ g-l) in the blend sample and mode of acquisition are reported between square brackets. D: 'digital' acquisition mode (pulse counting); A: 'analog' mode (reproduced from Diemer et al., 2002).
ment of the similar ratios in the IRMM-184 sample and the uncertainty on the IRMM184 certified values for the same ratios. The results for n(235U)/n(238U) and for n(234U)/n(238U) are displayed in Figure 32.8 and Figure 32.9 respectively as normalised values (the certified values remain privileged information of IAEA1). Within combined uncertainty of the measurements, the agreement with the certified values and the external and independent TIMS results was excellent. For the n(235U)/n(238U) ratio, the bias from certified ranged between 50 and 250 parts per million (ppm), and the combined uncertainty (k = 1) ranged between 0.034% and 0.059%. Thus, using a double focusing MC-ICP-SMS, six n(235U)/n(238U) ratios could be differentiated within < _+ 2.8%. For the n(234U)/n(238U), the bias from certified ranged between-5000 and -114:00 ppm, and the combined uncertainty ( k - 1) was for all results ~ 0.55% (with measurement repeatabilities always between 0.25% and 0.3N). 1. Perrin R (IAEA), personal communication.
Different Isotope Ratio Measurement Applications for Different Types of ICP-MS ...
Figure 32.8 - Comparison between IAEA n(235U)/n(238U) certified values and experimental results obtained using the double focusing MC-ICP-SMS at IRMM, and using TIMS outside IRMM. Vertical bars represent expanded uncertainties (k = 2).
Figure 32.9 - Comparison between IAEA n(234U)/n(238U) certified values and experimental results obtained using the double focusing MC-ICP-SMS at IRMM, and using TIMS outside IRMM. Vertical bars represent expanded uncertainties (k = 2).
741
742
Chapter 32 - C.R. Qu6tel & J. Diemer
This systematic negative difference may be pointing at a bias on the n(234U)/n(238U) value indicated in the IRMM-184 certificate and used to correct the MC-ICP-SMS measurement results, and this possibility is currently being investigated. It must also be noted that MC-ICP-SMS n(234U)/n(238U)measurement results that could be completely differentiated from each other correspond to the samples (1 - 3), of which the certified values are themselves entirely different from each other (e.g. no uncertainty overlap). 32.8 Accurate measurement of isotope ratios over more than 6 orders of magnitude using a double focusing MC-ICP-SMS fitted with an energy filter The magnitude of the signal ratios potentially measurable between neighbouring isotopes depends on the abundance sensitivity of a given mass spectrometer. Abundance sensitivity is the measure of the tailing on either side of the peak of an isotope signal. For instance, abundance sensitivity in the high mass range is evaluated using uranium by measuring the ratio "signal at mass 237 / 238U signal", and is expressed in ppm. This effect results from the dispersion in energy of the ions during their transmission throughout the mass spectrometer and is mostly due to collisions with residual gas molecules. Inappropriate tail corrections can cause large inaccuracy in isotope ratio determination by MC-ICP-SMS as illustrated in the paper from Thirwall (2001). By creating an energy barrier before a detector, ions that have lost energy and that are the cause of peak tailing on the low mass side are blocked. Hence, the abundance sensitivity is enhanced, and the magnitude of the ratios potentially measurable with the mass spectrometer is increased. The abundance sensitivity in the "Nu Plasma" on the high mass range is ~ 5 ppm (Table 32.1). The introduction of an energy filter on the ion
Figure 32.10 - Intensity of 238U signal and background signal at mass 237 as a function of the energy filter potential
743
Different Isotope Ratio Measurement Applications for Different Types of ICP-MS ...
path to the central ion counting system in the instrument (so called "ICI") allows an improvement by a factor ~ 10 (to ~ 0.53 ppm) of the abundance sensitivity (Figure 32.10). Clearl~ the further away from an important peak signal the larger the attenuation of the effect of the tailing of this important peak signal. However, two mass units away an effect remains measurable and the usefulness of the energy filter in minimising this effect was evaluated by measuring the IRMM-184 for its n(236U)/n(238U) ratio (certified value = 0.000 000 138 _+0.000 000 015, k = 2). The 12th Faraday cup (so called "Fll" at the extremity of the array on the low mass side) was physically inserted in the slit between the 9th and the 10th Faraday cups (e.g. between "F8" and "F9"), thus masking the way to the first ion counting detector (so called "IC0"). With this preliminary adjustment the measurement could be performed over 3 successive acquisition cycles as described in Table 32.3. IC1 was used to measure the 236U signal in cycle 2, and the correction for mass-discrimination was achieved internally by measuring the n(235U)/n(238U) ratio during the same cycle (using the exponential model, equation [32.4]; IRMM-184 certified n(235U)/n(238U) = 0.007 262 3 + 0.000 002 2, k = 2). The gain of IC1 was obtained by dividing the n(235U)/n(238U) ratio measured in cycle 2 by the n(235U)/n(238U) ratio measured in cycle 1. Measurements on IC1 were corrected for dead-time effects. Cycle 3 was necessary to quantify the hydride formation process from the signal of 238U on mass 239. It was measured to be approximately ~ 13 ppm, and this fraction was applied to the 235U signal in cycle 2 to correct the signal at mass 236 from the 1H235U isobaric interference. Finally, all signals were corrected for background by subtracting the noise measured beforehand for each detector (preliminary block of 5 replicate measurements of "zero" values) by deflecting completely the incident ion beams. Another way of estimating the background consists in replacing the "zero" values obtained as described above by "'chemical blank" values measured beforehand using separate solutions in order to reflect the possibility of contamination problems during both the sample preparation stage and the instrumental step (so called "memory effects"). This method is particularly suitable for elements likely to be already present as contaminants in the reagents used or later in the instrumentation, as the acquisition of "pure noise" values only could lead to an under estimation of the "background". Although this is unlikely to be the case for uranium and the isotopes Table 32.3 - Description of the method used with the Nu Plasma for the calibrated measurements of n(236U)/n(238U) ratios. The method involves three successive acquisition cycles and allows the internal correction for mass-discrimination (using the n(235U)/n(238U) ratio), as well as the correction for 1H235U formation on the signal at mass 236. The "zero cycle", for the measurement of the instrumental background is not represented. F8 to F l l represent Faraday cups 9 to 12, and IC1 represents the "axial" ion counting system. Fll, originally on the low mass side, was deliberately moved across and inserted in between F8 and F9. F8 Cycle 1 Cycle 2 Cycle 3
Fll
238 238
F9
IC1
FIO
235 236 239
235 238
744
Chapter 32 - C.R. Qu6tel & J. Diemer
under scrutiny during these experiments, this method was also tested by running an aliquot of the 2% HNO3 solution used to dilute the IRMM-184 samples. Figure 32.11 illustrates the results obtained by applying appropriate and inappropriate voltages to the energy filter (i.e. equivalent to energy filter on and off). For 4022 V and 4023 V (corresponding to the bottom right region of Figure 32.10) there is a good agreement within uncertainties between the measured ratios and the corresponding IRMM-184 n(236U)/n(238U) certified value. The relative expanded uncertainty ( k - 2), resulting from the combination of all the standard uncertainties arising from the measurements and the various corrections described above, was found to be 6.8 % for the measurements at 4023 V. This is a very satisfactory uncertainty result considering the difficulty of the exercise. The largest contributor was by far (97.1% of total) the standard error of the mean on the acquisition of the n ( 2 3 6 U ) / n ( 2 3 8 U ) ratio in cycle 2, and the second largest contributor (2.1% of total) was the correction for hydride generation 1H235U o n mass 236 also in cycle 2. Very close results are obtained for the 2 voltage values tested for the energy filter. Moreover, the result obtained at 4022 V with a correction for a "chemical blank" is nearly identical to the others, which indicates that the simplest way of correcting for background using "pure noise" values remains sufficient for these kind of experiments. Finally, at 3922 V (corresponding to the upper left region of Figure 32.10) the apparent n(236U) / n(238U) ratio observed is more than 6 times higher than the certified value, suggesting that significant peak tail2.0E-07 1.9E-07 n (236)/rt (238U) =
1.8E-07
8.4E-07 3
1.7E-07 m r
1.6E-07
I~ 1.5E-07 m
1.4E-07 1.3E-07 1.2E-07 1.1E-07 1.0E-07 energy filter at 4022 V energy filter off (3922 V) Certified IRMM-184 energy filter at 4022 V, energy filter at 4023 V "chemical blank" subtracted |
|
|
|
Figure 32.11- Comparison of the IRMM-184 n(236U)/n(238U) certified value with the experimentally determined n(236U)/n(238U) ratios using different instrumental settings. Vertical bars represent expanded uncertainties (k = 2).
Different Isotope Ratio Measurement Applications for Different Types of ICP-MS ...
745
ing effects from the 238U signal render the measurement of the 236U signal impossible. 32.9 Conclusions
These results demonstrate that accurate isotope ratio measurements can be carried out with ICP-MS of very different design and generation. Accuracy is not a function of the cost of the instrument, nor of its size. It depends more on the efforts that were made to understand the nature and the magnitude of the various quantities influencing the measurement result. Measurement results obtained with different instruments using identical samples (different dilution factors) are comparable only if they are traceable to the same stated system of reference. In the case of the results presented in this chapter, they all have been corrected for mass-discrimination effects as well as for other relevant effects possibly more specific to each of the instrumentation employed. The different types of ICP-MS instruments might require different calibration strategies depending on their performance characteristics and their mode of operation (e.g. "Element2" versus the "Nu Plasma" and both ICP-QMS instruments tested). The combined uncertainty of the measurement is a direct indicator of the quality of the result, and in that respect significant differences can be observed between the instruments. Quadrupole ICP-MS are certainly the easiest instruments to operate, but magnetic sector ICP-MS have the potential to provide the best quality and particularly the multi-collector ICP-SMS instrument tested for ratios close to unity. This instrument offers another unique advantage over the others for the measurement of extreme ratios (6 orders of magnitude or more). However, for applications such as the determination of an element concentration by IDMS the differences level off as other sources of uncertainty than purely those of instrumental origin can be largely dominant. Thus, resorting to a multi-collector ICP-SMS for IDMS applications is mostly not appropriate and unjustified.
Handbook of Stable Isotope Analytical Techniques, Volume 1 P.A. de Groot (Editor) 9 2004 Elsevier B.V. All fights reserved.
CHAPTER 33 Isotope Ratio Analysis Techniques using Photoionization as a Source of Ions Ian Lyon Department of Earth Sciences, The University of Manchester, Manchester, M13 9PL, United Kingdom e-mail: [email protected]
33.1 Introduction
There are many techniques which use photons to ionize or excite atoms for isotopic analysis, but all are relatively specialized. The main application occurs when sample size is the limiting factor and these situations arise if the sample is exceedingly small (for example cosmic dust particles), exceedingly rare (for example cosmogenically produced radioactive noble gas isotopes in the environment) or the analytical site is inaccessible to conventional high precision mass spectrometers, (remote sensing, spacecraft). This chapter will describe these techniques and the important niches they fill, particularly with respect to the study of stable isotope ratios of the lighter elements. The principal applications of these techniques are, however, usually aimed at the analysis of rare or radioactive isotopes, particularly of heavy elements and these applications fall outside the main scope of this volume. These applications will, therefore, only be described in summary with extensive referencing and bibliography for further reading. Isotope analytical methods, where high precision is required, are well covered by other chapters in this volume but all require abundant sample quantities (abundant being a relative term but meaning here samples of typically nanomoles or more). There are however fundamental and well known limitations on conventional stable isotope analytical techniques which limit their usefulness, when applied to the study of smaller sample sizes; principall)~ ionization efficiency, contamination or background. Typical ionization efficiencies for a number of ionization mechanisms are illustrated in Table 33.1 and show that even the most efficient ionization mechanisms (e.g. Secondary Ionization) rarely exceed 1% and are typically much smaller for most elements. There are, therefore, many niche areas where the ability to improve this ionization efficiency can open new scientific areas. This paper will discuss laser resonance and non-resonance multi-photon ionization and show how these techniques may be coupled with mechanisms for removing
Isotope Ratio Analysis Techniques using Photoionization as a Source of Ions
747
Table 33.1 - A comparison of ionization mechanisms for producing ions required in mass spectrometers Ionization Mechanism
Ionization Efficiency
Comments
Electron impact
Approx. 0.01%
Cheap and easy but difficult to improve sensitivity because of space charge effects in electron beam
Plasma source
upto 100%
For heavy elements, lighter elements have lower efficiencies
Secondary ionization
upto 1%
For specific elements, other elements may be orders of magnitude worse ionization efficiency
Thermal ionization
upto few %
As high as 30% for some negative ions (N-TIMS) but most elements cannot be ionized using thermal ionization
Laser resonance ionization
upto 100%
Almost all elements can be ionized but ionization is limited to one element at a time
Multi-photon ionization
upto 100%
Requires very high power lasers. Achieves multi-element ionization, but chance of producing multiply charged ions
atoms from samples, particularly with very high spatial resolution, to give versatile and powerful methods for isotopic analysis. Until recently however, the measurement of isotopic ratios using laser ionization has been dogged by isotope selection effects, which have reduced accuracy. These effects will be discussed along with new solutions to this problem. Also discussed will be fluorescent techniques which can be used for in-situ isotopic analysis in hostile and testing environments such as spacecraft or in instruments landed on extra-terrestrial bodies. A review paper such as this necessarily relies substantially upon the work of others and the author expresses his appreciation of this excellent work. Any omissions or incomplete accreditation are accidental and regretted. 33.2 Laser Resonance Ionization
There are several ways in which photons can either excite or ionize atoms and so be used for isotopic analysis. These methods have been considerably aided by the improving power, versatility, reliability and availability of lasers over the last 10 years, dramatically increasing the number of published applications. Probably the most familiar and widespread technique is resonance ionization.
748
C h a p t e r 33 - I. L y o n
This review will concentrate upon resonance ionization of lighter atoms as heavier elements are outside the scope of this book. Fortunately, understanding the quantum interaction between photons and light atomic mass atoms is much easier than for heavier atoms as electrons in light atoms generally do not couple strongly together and simple quantum selection rules are more rigorously obeyed. Molecules will also be ignored as, for the purposes of this paper, the principal use of this technique is for the formation of atomic ions as a prelude to mass spectrometry. Resonance ionization, looked at in the simplest and most naive way, consists of a single laser tuned in wavelength to the energy difference between a ground-state and excited state of a selected elemental species. This produces a population of atoms in the excited state. Absorption of a second photon by the excited atoms results in ionization of the atom which may then be mass analysed and detected in some form of mass spectrometer. Looked at on this simplistic level, this ionization mechanism has several advantages over more traditional ionization mechanisms such as electron impact, plasma discharge or secondary ionization: a) Resonance ionization is element selective - if the laser wavelength is tuned off the energy required to excite the selected atomic species then ionization ceases. Similarly, isobars are not excited and ionized. We thus have a method of separating isobars, such as 87Sr and 87Rb, which cannot be separated by mass spectrometry. This "physicist's" method of in-situ chemistry has instant appeal to those for whom extensive chemical preparation of samples is impractical or impossible (either because of lack of skill or facilities, because the desired species is too small compared to blanks introduced in chemical processing, or too intergrown with other minerals / materials to be efficiently separated). b) The ionization may be extremely efficient - up to 100% within the laser beam since laser powers can be high enough to saturate ionization and focused laser beams do not suffer from problems which limit ionization efficiency of other methods such as space charge in electron impact ionization. This review intentionally aims to give a simple, even simplistic, overview of the physical principles involved in order that non-specialists may gain an understanding of the applications and applicability of the technique. There are many excellent reviews of the theoretical basis of resonance ionization and applications. For example Hurst & Payne, 1988; Fairbank 1996; Wendt et al., 1999 are recommended. 33.3 Theoretical basis of resonance ionization
The description of the resonance ionization process given below is intentionally pared to the minimum number of equations and treats only the simplest cases in order to indicate the basis of the technique. Electrons in atoms may only exist in discrete quantum states and the laws of conservation of energy, angular momentum and parity determine which state an electron may move to during a transition between states. These conservation laws are
Isotope Ratio Analysis Techniques using Photoionization as a Source of Ions
749
expressed as a series of fairly simple rules known as selection rules. An understanding of selection rules is fundamental to understanding the interaction between light and atoms and so these are elaborated below. Simple selection rules are also assumed to be rigorously obeyed which is a reasonable assumption for light atoms. 33.3.1 Selection rules Selection rules reflect conservation of momentum and parity in the atom/photon system. Electrons carry spin (S) and orbital (L) angular momentum to give a total angular momentum J by vector addition. In light atoms with a single valence electron, absorption or emission of a photon (which carries one atomic unit h) of angular momentum is usually (but not always) between states with orbital angular momentum differing by this amount - i.e. transitions are between s and p states, p and d states etc. (AL= + 1).
More rigorously, the magnetic moment of an electron coupled with the component of the total angular momentum J in the magnetic field of the atom gives rise to socalled magnetic sub-levels (mj) which produce fine splitting of the energy levels. The shift in energy produced by this fine structure splitting is small compared with differences between energy levels (a typical wavelength difference between fine structure states is of the order of 0.1nm (1A)). Formally, these selection rules are written as AJ = 0, + 1, Amj = 0, + 1 (AL = + 1 is not true generally, but works for most of the simpler cases considered here). Ordinary unpolarized light is an equal mixture of left and right handed circularly polarized light (+ h) so the deliberate choice of one polarization state will allow only transitions with Amj - +1 or -1 (depending upon the polarization). Where there are odd-mass number isotopes, the nucleus also has net nuclear spin I (usually 0.5 h), whereas for even mass number isotopes the net nuclear spin is almost always zero. The nuclear spin may also couple with the J vector of electrons to shift the energy levels of these electrons further although the energy shifts are approximately 2000 times smaller than the fine structure splitting. The total angular momentum of electrons is then F, which results from the vector addition of J+I. F may take different components in the atomic magnetic field to produce further splitting of energy levels known as hyperfine structure. The quantum number which describes this splitting is analagous to mj and is designated mF. Selection rules are then AmF - 0, +1. Since odd-mass isotopes are affected by hyperfine structure but not even isotopes and the typical energy shifts are comparable to laser bandwidths, this effect can have serious consequences for the measurement of isotope ratios during photoionization and this will be discussed below. 33.3.2 Resonance Ionization Electrons in an atom exist in discrete eigenstates (discrete quantum states) which determine their energy. When an electron absorbs the energy and angular momentum of a photon it may therefore make a transition to an excited eigenstate if the energy difference matches the photon energy and the two energy levels have opposite parity
750
Chapter 33 - I. Lyon
and an angular momentum difference of h, (the angular momentum carried by the photon). As excitation has been achieved by absorption of a single photon, the excited state can then decay emitting a single photon in an 'allowed transition' back to its original state. This 'spontaneous emission' occurs typically on a time scale of nanoseconds. If the photon flux is sufficiently high that there is a significant probability of a second photon of the same energy as the first arriving at the atom whilst the atom is in the excited state then the excited state may be stimulated to emit a second photon coherently with the first. At high photon flux this will happen on a time scale which is much faster than the spontaneous decay time. This stimulated emission will be followed by repeated absorption and further stimulated emission leading to the electron in the atom oscillating between the ground and excited state at high frequency (known as the Rabi frequency). This frequency is determined by the intensity of the photon flux. If the excited state energy level is more than half way to the continuum (see Figure 33.2) then absorption of another photon by the electron in the excited state can lead to ionization. The processes of stimulated emission and ionization compete and the probability of the electron taking the ionization branch will be determined by the photoionization cross section of the excited state. Thus we have two necessary conditions for resonance ionization to take place with high probability, the so-called 'flux' and 'fluence' conditions, (Hurst & Payne, 1988). The flux of photons (photons cm-2 s-l) must be high enough that the Rabi frequency is much higher than the inverse of the spontaneous emission time, so that a significant excited state population builds up. The fluence (photons cm-2) must be high enough that the photoionization factor 1-e-no approaches unity where n is the photon fluence and o the photoionization cross section in cm2. If an atom is in an excited state then this condition determines that it will almost certainly be ionized. Some practical examples which use this single photon ionization scheme are detailed below. For atoms where the energy difference between states or from a discrete state to the continuum (for ionization) is greater than can be easily reached by a single photon there are alternative resonance ionization schemes. These are illustrated in Figure 33.1. Ionization schemes for most elements have been designed and often tested and these are collated and described in several publications (e.g. Saloman, 1990, 1991, 1992 and 1993), references in Tables 1 and 3 of Fairbank, 1996 and Web sites (e.g. search web site of National Institute of Science and Technology (NIST), Gaithersburg, Maryland, USA). The basis of some of the more complicated resonance schemes and also multi-photon ionization may be understood when it is remembered that even if the energy of the photon does not exactly match the energy difference between the ground and
751
Isotope Ratio Analysis Techniques using Photoionization as a Source of Ions
~tmaum
l~ll I
Confm~
~t~nuum
Con~~
~m~~
hvlelhv2 bv~ Ir~'l bv'l
Gzo~ds~a~
RIS~I
G~unds~ RLS~II
Gzo~ds~ RIS ~ H l
b,,'l
Gzo~ds~ RLS ~ I V
Gzo~ds~ RIS Scheu~ V
Figure 33.1 - A schematic diagram of basic resonance ionization schemes. Scheme I is the simplest involving resonant absorption to an excited state which is more than half way to the continuum. Absorption of a second photon ionizes the atom. Scheme 2 is more typical and involves 2 tunable lasers resonantly exciting the atom to two sequential excited states. This arrangement is usually required as it is rare to find an excited state more than half way to the continuum (as in scheme I) that can be easily reached using available lasers. Scheme 3 is very similar to scheme II except that the second photon is tuned to take the atom to a 'Rydberg' state (very high principle quantum number n). From this excited state it is relatively easy to ionize the atom using a non-photon technique such as field ionization. Scheme 4 involves very high power densities to populate a virtual state (dotted line). The laser is tuned so that the electron is excited to a virtual state which is exactly half-way between the ground state and the real excited state. Absorption of a second photon promotes the electron into the excited state. This state is meta-stable with respect to the ground state as there is no allowed single photon spontaneous decay back to the ground state. This technique is used particularly for noble gas atoms where a single photon excitation to the lowest excited state would require a photon energy beyond the capabilities of available commercial lasers. Scheme 5 is multi-photon ionization and similar in principle to scheme IV except that the power density in the focused laser beam is so high (typically >1012Wcm -2) that excitation may proceed by a series of virtual states right up to the continuum and ionization. No resonant intermediate level is involved in this scheme. This form of ionization is therefore largely indiscriminate and offers the potential for simultaneous ionization of all different elements.
excited state, a very short-lived transition (typically 10-16s) to the excited state is possible with the energy deficit or excess accounted for by the Uncertainty Principle. The electron may 'borrow' the energy it needs to make the transition to an excited state for a short time (At) determined by the Uncertainty Principle: AEAt-h
[33.1]
where AE is the difference in energy between the photon energy and the energy required to excite the electron from the ground to the excited state. Whilst the electron is in this temporary state it is said to have been excited to a virtual state.
752
C h a p t e r 33 - I. L y o n
If a second photon which matches the energy deficit between the virtual state and excited state is absorbed by the electron during this time then the electron can make a real transition to the excited state. This excited state is moreover meta-stable relative to the ground state as two photons (total angular momentum change of 2 h or 0) have been required to excite the electron and there is therefore no simple allowed transition back to the ground state. The electron can therefore usually exist in this state for relatively long times (typically microseconds) and so be readily ionized by lower intensity laser beams. This is the basis of 2-photon excitation (scheme IV of Figure 33.1) and also for multi-photon ionization. In the latter case, the photon flux is so high that the electron may repeatedly absorb photons travelling via several virtual states all the way to the continuum and ionize. To try and give some very general indication of equipment requirements to achieve the resonance ionization schemes outlined in Figure 33.1 we need to look at typical spontaneous decay times of levels and photoionization cross-sections. Numerical values given here are estimates chosen to indicate typical equipment parameters. For simple single-photon excitation, where the spontaneous decay time (~) back to the ground state is typically a few nano-seconds, the stimulated excitation/emission rate must be greater than 1/~ or >109 Hz. The rate of stimulated emission/excitation is given by the expression pBif where p is the photon energy density Hz-1 m-3 and Bif the Einstein B (initial-final state) coefficient for stimulated excitation/emission (e.g. Atkins, 1988).
eaf
[33.2]
Bif = 4eohmevi f
where vif is the frequency of the photon (= c/K) emitted in a transition between the initial and final state and f is known as the oscillator strength (0 ~ f ~ 1 for a single electron). The oscillator strength is a measure of the strength of the allowed transition and for simplicity I shall here assume f = 1. e - electron charge, me - electron mass, ~o permeability of free space and h - Planck's constant. The photon energy density is given by the energy per laser pulse (or per second for cw lasers) m-3 Hz-1. The volume swept out during the time of a laser pulse At with a beam radius r is z~r2cAt and if the bandwidth of the laser is Av then: P -
E
~r2cAtAv
T
J m-3Hz
-1
Thus putting these together, we find that for the Rabi frequency to exceed 1/x:
[33.3]
Isotope Ratio Analysis Techniques using Photoionization as a Source of Ions
e2fE ~ 1/1: P B i f - 4J~csohmer2vifAtAv
753
[33.4]
Substituting values for the constants and typical atomic parameters we therefore find that 1.276 x
1027 2
E ~ 10 9 r vi/AtAv
[33.5]
This equation is still quite general and typical operating parameters for different laser types may be substituted to estimate the necessary pulse energy and focusing required to meet this condition. Thus, to take a typical example for lasers operating in the blue to ultra-violet, vif- ~ 1015 Hz and if we have a 'Q' switched laser so At might be 5ns and if we assume a (poor) bandwidth of ~10n Hz, then: E r
~ 0.4
[33.6]
so for r = 1 mm, E > 0.4 ~J. This energy per pulse is well within the range of commercially available lasers and may be achieved over the whole spectrum from infra-red to vacuum ultra-violet using techniques such as 4-wave mixing (Hurst & Payne, 1988) to extend the available wavelength range. 33.3.3 Problems
One of the main problems in measuring isotope ratios using resonance ionization has been the so-called odd-even effect. As described above, odd isotopes have net nuclear spin which causes splitting of atomic energy levels into hyperfine structure. Nuclei with an even number of nucleons usually have no net nuclear spin and so no hyperfine splitting. There are therefore more excitation pathways for odd isotopes and this can alter the excitation and ionization probabilities. Even if the bandwidth of the laser is large enough that it covers all of the hyperfine transitions, the excitation and ionization probabilities for different isotopes can still be different. Furthermore, variations in laser power a n d / o r polarization between pulses or over time can then vary the excitation/ionization probabilities between isotopes. A further complication may be that non-resonant ionization may also occur which may cause further isotope selection effects. Some excitation schemes such as scheme IV of Figure 33.1 do not suffer from this problem. An example is the resonance ionization of xenon in which a two-photon excitation 5p 6 3P1 -> 5p 5 6p 3P1 has no hyperfine splitting since transitions are from an
754
Chapter 33 - I. Lyon
jn r
3/2
5/2
11/2
9/2
';/2
1 D 2 F = 11/2 1 P I F = 9/2
&InF = +1 a+
/ is o
~
F=
r/2
l
mitt. = +1
Figure 33.2- (From Figure 2 of MUller et al., 1998). Optical pumping of hyperfine components in the excitation of 41Ca and 43Ca. Circularly polarised light of only one 'handedness' is used (a+) for which only AmF = +1 transitions for excitation (thick arrows) are allowed by selection rules. Decay can occur with A m F = + I or 0 (thin arrows) but subsequent excitation drives the electrons to states with higher mF values until they accumulate in the mF - 7/2 state and further excitation can only occur to the mF = 9/2 state. This situation is then equivalent to that for even isotopes for which there is no hyperfine splitting and so measured isotope ratios are equal to the abundances in the material under analysis. S -> S state. As this state has zero orbital angular momentum, there is no hyperfine splitting and hence measured isotope ratios are as expected from natural abundances, (Gilmour et al., 1991, 1994b). Other measured isotope ratios were found to be anomalous however. An attempt to measure Sn isotopes (Fairbank et al., 1989) exhibited variable o d d / e v e n fractionation and an attempt to develop a in-situ method of Rb/Sr dating by sputtering minerals with high spatial resolution and resonantly ionizing the strontium atoms also suffered from this problem. The excitation transition 5s2 1So -> 5s5p 1P1 in neutral strontium was shown to suffer from the o d d / e v e n effect with measured 87Sr/88Sr ratios lower by approximately 10% from the true value (Perera et al., 1995). An elegant solution to the o d d / e v e n problem has recently been demonstrated (Mtiller et al., 1998) using a technique known as optical pumping. M~ller et al. used the scheme shown in Figure 33.2 to measure Ca isotope ratios. Ca has a nuclear spin of 7/2 giving rise to complex hyperfine structure in the energy levels of the odd isotopes 41Ca and 43Ca. The plethora of ground and excited levels accessible with AmF = 0, + 1 selection make it clear that with unpolarized light it is exceedingly difficult, if not impossible to accurately measure the ratios of these isotopes relative to the even isotopes. However, Mtiller et al. (1998) used laser light of one polarization only (designated o +) and this allows only transitions AmF - +1. Referring to Figure 33.2, absorption of a photon will therefore drive electrons into an excited state with mF one unit higher than they were already in. Radiative decay transitions can occur from these excited states with any polarization (indicated by the 3-way branching decays from each level in Figure 33.2) to a ground state with mF +1, -1 or 0 relative to the excited state. After repeated absorption and radiative decay, there is thus a net movement towards the highest mF state (+7/2) and electrons accumulate in this configuration of the ground state. The process is analogous to climbing a ladder which,
Isotope Ratio Analysis Techniques using Photoionization as a Source of Ions
755
although one may slip back a rung occasionally, will result in climbing to the top. Even isotopes have only the single ground state and optical pumping exclusively populates a single hyperfine level of the odd isotopes. Subsequent double resonance excitation by 2 o + photons leads to the mF - +9/2 state of the first excited state and the mF - 11/2 state of the second as illustrated in Figure 33.2. Thus, electrons in the ground state of the odd Ca isotopes become concentrated in a single mF state, just like the even isotopes and so have an excitation probability equal to the even isotopes.
33.3.4 Applications We are looking here at resonance ionization as a means to an end: ionizing atoms as a prelude to mass spectrometry. It is therefore appropriate to now look at some real applications. Since the ability to measure extremely small sample sizes is the modus operandi of resonance ionization, real scientific advances can come from situations in which this ability is exploited. The main way in which this is realised is through laser or ion sputtering of samples with high spatial resolution.
33.4 SIRIS and SIRIMS (Sputter Initiated Resonance Ionization Spectroscopy and Sputter Initiated Resonance Ionization Mass Spectrometry) Photon ionization becomes an extremely powerful method of measuring isotope ratios in exceedingly low element concentrations or where one isotope is extremely rare. Here a solid sample is sputtered with a high energy ion beam, usually with very high spatial resolution (down to 50nm). Most (typically > 98%) of the sputtered species are neutral atoms which are in stoichiometric ratios representing the composition of the sample (unlike the highly variable sputtered secondary ion composition). An ionizing laser passing through the sputtered neutral cloud can then be tuned to ionize a selected atomic species with high efficiency - the efficiency usually being governed by the spatial overlap between the neutral atom cloud and the laser beam. This technique has been covered by several recent review papers, Arlinghaus et al., 1994; Nicolussi et al., 1996; Mathieu & Leonard, 1998; Wendt et al., 1999; Wiley et al., 1999. An excellent example of the potential of this combination is illustrated by the work of Nicolussi et al., (1997a, b) who measured Zr and Mo isotope abundances in individual interstellar grains (typical size ~ l~m). Despite the relatively low abundance of these elements in minute grains, they were able to show that the grains showed isotopic patterns which were consistent with s-process nucleosynthesis from their inferred formation in Asymptotic Giant Branch stars. Other examples of low abundance detection are given below which, although a resonance ionization scheme and feasibility study for the low mass elements has usually been conducted or studied, has not yet been applied to real practical problems: Hydrogen - Ultra-sensitive detection of hydrogen and measurement of hydrogen isotope ratios is described by Miyake et al. (1998) with a detection limit of 104 atoms cm-3. This paper describes the feasibility of the method by detecting H and D atoms evolved from a hot tungsten surface in ultra-high vacuum and so demonstrates the
756
Chapter 33 - I. Lyon
capability of the technique for measurements of D / H ratios in hydrogen sample sizes limited only by the vacuum blank. Tritium and muonium may be also detected by the same means and this is where the method has so far found practical expression. A useful variant is given by Yoruzu et al. (1999), who show that resonant laser ablation mass spectrometry (RLAMS - in which the laser ablates a solid sample as well as resonantly ionizing the released hydrogen) may be used to detect and measure D / H ratios in solid samples. Existing literature demonstrates the feasibility of the technique but not as yet, practical application. Lithium - A theoretical and experimental study of ionization and isotope selectivity effects during resonance ionization of lithium has been made by Suryanarayana et al. (1998). This showed good agreement between theoretical predictions and observations opening the way to the measurement of lithium isotope ratios in natural samples although none has yet been conducted. Carbon - Downey et al., (1992) demonstrated the use of resonance ionization mass spectrometry to depth profile device substrates for carbon but no practical applications measuring carbon isotope ratios have been reported. Oxygen - Orlando et al., (1991) demonstrated the detection of ground-state atomic oxygen above an electron irradiated surface although no isotopic measurements utilizing RIS are known. N i t r o g e n - A demonstration of a resonance ionization scheme for nitrogen has been made by Debeer et al., (1992) although no isotope measurements using this method are known. S u l p h u r - Studies of resonance ionization schemes for sulphur have been made by Woutersen et al., (1997) and Venkitachalam & Rao, (1991) although no isotopic measurements are known. Other e l e m e n t s are summarized in Table 33.2. 33.5 A b s o r p t i o n and Fluorescence Techniques
The previously described optical techniques have been concerned with isotopic measurements of rare isotopes or elements in sample sizes down to the ultimate limit of single atoms. Absorption or fluorescence techniques are alternatives that are aimed at larger sample sizes and remote sensing or robustness in hostile environments (e.g. atmospheres of other planets). The full details of this technique are too involved to expand upon here and a full explanation may be found in Irving et al., (1986). Basically infra-red radiation at an absorption wavelength is partially absorbed by a cell containing 12CO2. The pressure in this cell is mechanically modulated so that the intensity of the infra-red beam is modulated due to modulated absorption. The infrared beam passes through a second cell containing 13CO2 which is pressure modulated at a second frequency similar but not equal to that of the first cell. The infra-red beam after passing through the two cells then has a bi-modulation (essentially the beat frequency between the two pressure modulation frequencies). This bi-modulated infrared beam is then passed through the CO2 sample of unknown composition which is pressure modulated at a third frequency. The complex heterodyne absorption signal may be analysed to recover the effective absorption due to 12CO2 and 13CO2 in the unknown sample and the ratio of 13C/12C determined. The method has been mainly applied to bio-medical applications such as measuring the 12CO2/13CO2ratios in res-
757
Isotope Ratio Analysis Techniques using Photoionization as a Source of Ions Table 33.2 - Other elements Element
Reference
Comments
Lead (Pb)
Arlinghaus et al., 1996
Acquisition of 206pb, 207Pb and 208Pb in zircons using SIRIMP. Also describes general SIRIMP technique
Uranium (U Plutonium (Pu))
Trautmann, 1994
Includes other very long lived radioelements. 238U and enriched 235U, not disequilibria Development of instrument, also plutonium Review paper Environmental tracing of Pu isotopes -- 106 atoms
Young et al., 1994 McMahon et al., 1998 Wendt, 1998 Erdmann et al., 1998
Magnesium (Mg)
Koumenis et al., 1995
Laser desorption, not RIS, isotopes of Cu, Ca, Mg, Fe and Zn
Calcium (Ca)
Mtiller et al., 1998 N6rtersh~iuser et al., 1998 Bushaw et al., 1996
41Ca 41Ca Isotope selectivity > 1015 for 4K41Ca
Strontium (Sr)
Perera et al., 1993, 1994
Bushaw & Cannon, 1998
Attempt to develop in-situ Rb / Sr dating 89,90Sr in environmental samples 90Sr
Zinc (Zn)
Hansen et al., 1996
Zn isotopes in silicon
Samarium (Sm)
Park et al., 1998
RIS scheme only
Gadolinium (Gd)
Blaum et al., 1998 Rhee et al., 1998 Jeong et al., 1998
Gd isotopes RIS scheme CW-RIS scheme
Helium (He)
Lancoursi6re et al., 1994
RIS + VUV synchrotron radiation
Argon (Ar)
Xenakis et al., 1996
XUV RIS scheme
Krypton (Kr)
Thonnard & Lehmann, 1994
Rare Kr isotope dating groundwater + ice Ultra-sensitive detection
Wendt et al., 1996
Lassen et al., 1994 Xenon (Xe)
Gilmour et al., 1991, 1994a, 1994b, 1995b Meyer et al., 1996
Ultra-sensitive Xe mass spectrometer + applications RIS + VUV synchrotron radiation
758
Chapter 33 - I. L y o n
piration as necessary sample sizes are large. Accuracy can by typically < 1%o which is of acceptable accuracy for clinical applications. The above method did not require monochromatic radiation, but using tunable dye lasers can give selective absorption at particular molecular vib-rotor transitions which are sufficiently resolved that isotopic measurements may be made. Such solidstate and small instruments are currently under development for existing (including one on the ill-fated Mars Polar Lander) and future Mars, Venus, Titan and Europa missions (Webster et al., 1999). 33.6 Conclusions
Techniques using photons to measure stable isotope ratios are many and varied although not of widespread use. The techniques described fill many valuable niche areas which are growing in size and application. Of particular importance is the capability for measuring isotope ratios in ever decreasing sample sizes with exceedingly high spatial resolution and this is one area in particular which is expected to grow, for example in such diverse areas as analysis of interstellar grains to analysis and depthprofiling of semiconductor devices.
Acknowledgements I am indebted to some of my colleagues, past and present for joint work which is referenced here, particularly J. D. Gilmour, G. Turner, and I. K. Perera. Other work referenced in this review is as complete as I could achieve and I express my appreciation to many international colleagues for their assistance and discussions, especially to K. Wendt in addition for permission to reproduce Figure 33.2. I finally thank Dr R M611er and an anonymous reviewer for helpful comments.
Handbook of Stable IsotopeAnalyticalTechniques,Volume 1 P.A. de Groot (Editor) 9 2004 ElsevierB.V. All rights reserved.
CHAPTER 34 Isotope Ratio Infrared Spectrometry Erik Kerstel Center for Isotope Research, Department of Physics, University of Groningen, The Netherlands e-mail: [email protected]
34.1 Introduction This chapter deals with optical techniques to accurately measure isotope abundance ratios as alternatives to isotope ratio mass spectrometry (IRMS). The optical techniques discussed here all involve excitation in the infrared region of the spectrum, associated with molecular rotational-vibrational motions. We thus ignore recent measurements on atomic systems like 81Kr and 85Kr in magneto-optical traps (Chen et al., 1999; Bailey et al., 2001; Lu & Wendt, 2003). Also not considered here are techniques in which lasers merely serve to prepare the sample (e.g., by ablation), subsequently to be analyzed by IRMS. The case of photoionization followed by mass spectrometry analysis is the subject of Chapter 33. Different approaches are discussed in terms of the optical detection technique employed, rather than the molecular system being studied. Each technique is presented with an example of experimental work. Where possible, these examples are selected because of their successful application in the field of isotope ratio measurements. Thus, many demonstrations of optical techniques that miss the precision to compete with IRMS or the connection to accepted international standard materials are not treated. However, we do include a listing of relevant studies reported in refereed international journals, including parameters such as accuracy achieved and sample amount (Table 34.1). Traditionally, isotope ratio measurements use isotope ratio mass spectrometers (IRMS). These have now evolved to the point where commercially available machines are able to attain an extremely high measurement precision and high sample throughput. In fact, this handbook may be considered a tribute to the success of IRMS. Still, a number of fundamental problems remain with IRMS. Most predominantly there is its practical inability to deal with condensable gases and the mass-overlap of interesting isotopomers. The common solution has been to resort to chemical pretreatment of the sample. Usually, different pathways are required for the different isotope ratios in the same molecule. For example, in the case of water (arguably the most important environmental molecule), one resorts to oxygen isotope exchange between water and CO2 and reduction of water to H2, yielding end-product molecules that are easily analyzed
molecule isotope sample condition / ratio measurement range
technique / laser
setup1
analysis time
760
Table 36.1 - Overview of optical isotope ratio measurement reported in the international (refereed) literature. lo-precision / accuracy2
reference
10% relative error
Lehmann et al., 1977
-
pure C2H2 200 pmol [13C]/[W]=l% to 12%
-
Optoacoustic, C02 laser line coincidence
C
CH4
13c
pure CH4, natural abunc 2 pmol
Direct absorption, tunable 3.3 pm Pb-salt laser 2f-detection
R2
30 min, inc evac / refill
<0.6%oA
Bergamaschi et al., 1994
CH4
13c
natural air, 1.7 ppmv CH4 -5 nmol CH4
cavity-locked CRDS CO - tunable sideband laser
C
4 min excl. evac / refill
13%oP
Dahnke et al., 2001
CH4
13c
pure CH4, natural abund. -1 pmol
Direct absorption, 8.1 pm cw-QC laser 1f-detection
C
5 min.
-1O%oP
Gagliardi et al., 2002
CH4
13c
pure CH4, natural abund. -200 pmol
Direct absorption, 8.1 pm QC laser transient absorption detection
C
-5O%oP
Kosterev et al., 1999
CH4
13c
pure CH4, natural abund. 100 pmol
Direct absorption, 1.66 pm DFB diode laser 2f-detection
R1
0.3%oP
Uehara et al., 2001
CH4
13c
in-situ statospheric CH4
Direct absorption, tunable Pb-salt laser
C
9070oP
Webster et al., 1992
20 min, incl. evac/ refill
Table 36.1 continued >
E. K e r s t e l
13c
Chapter 34-
C2H2
molecule isotope sample condition / ratio measurement range
technique / laser
setup1
analysis
lo-precision /
time
accuracy2
30 min, incl. evac / refill
5%oA
Bergamaschi et al., 1994
2.5%oP
Lee & Majkowski, 1986
4%"P
Becker et al., 1992
0.25% P
Bowling et al., 2003
reference
-
CH4
2H
pure CH4, natural abund. 80 pmol
Direct absorption, tunable 3.3 pm Pb-salt laser 2f-detection
R2
co
180
pure CO, natural abund. -50 pmol
Direct absorption, tunable 4.7 pm Pbsalt laser, transient
C
co2
13c
pure C02, natural abund. 10 pmol
Direct absorption, tunable 4.3 pm Pbsalt laser, transient
c02
13c
natural air, 350-700 ppmv Direct absorption, C02 (613C=-16%mto -6760) tunable 4.3 pm Pb230 ml / min flow-through salt laser, transient
R2
coz
13c
pure C02 -4 mmol
Direct absorption, 1.6 pm DFB laser, 1f-detection
C
-5'roP
Chaux & Lavoral, 2000
c02
13c
exhaled breath, -5% C02 0.4 liter
Direct absorption, 1.6 pm DFB diode laser 2f-detection
C
-10%0P
Cooper et al., 1993
c02
*3c
volcanic air samples, 1000 ppmv C02 at 100 Torr (flow-through)
Direct absorption, 4.3 pm DFG
R2
-0.8%,0P
Erdelyi et al., 2002
10 sec
5 min
Isotope Ratio Infrared Spectrometry
> Table 36.1 continued
Table 36.1 continued > 761
762
Table 36.1 continued > molecule isotope sample condition / ratio measurement range
technique / laser
setup1
analysis time
Imprecision accuracy2
1
reference
c02
13c
natural air, 350 ppmv C0.2 FTIR, 1 cm-1 resolution 1 atm total pressure
R1
6 - 15 min
-0.15%0P
Esler et al., 2000b
c02
13c
exhaled breath, 5% C02 1 atm
FTIR, 1 cm-1 resolution
R1
6 min
-O.l%oP
Esler et al., 2000b
c02
13c
exhaled breath, -4-5% c02 200 ml (ca. 40 pmol C02) A613C--25%0 to +40%0
NDIR, photoacoustic FANci 2TM
R2
1-2 min
< 0.3yOoP
Fischer Analysen Instruments, http: / / www.fan-grnbh.de
-270oA
13c
exhaled breath, 3-5% C02 0.5 liter A6W--25%0 to +25%o
NDIR, photoacoustic
R2
4 min
0.3%oA O.6%oA 0.8%0P;1.1%0A
Haisch et al., 1996, Schadewaldt et al., 1997 Hildebrand& Beglinger, 1997
c02
13c
exhaled breath, 3% C02 0.4 liter 6*3c=-100%~ - 320%~
NDIR, heterodyne detection
R2
4 min
0.4%oP 1.3%oA
Irving et al., 1986
c02
13c
10 Torr C02 in 10 bar N2 -20 pmol [ W ]/ [12C]=O.l%to 5%
FTIR, press. broad. 1 cm-1 resolution
R1
8%oP
Kindness & Marr, 1997
c02
13c
diluted exhaled breath and natural air, 380 ppmv C02, ca. 5 pmol C02
Direct absorption, 4.3 pm Pb-salt laser, transient detection
R1
-0.2%0P
McManus et al., 2002
c02
13c
pure C02
NDIR, photoacoustic
R2
>20%0P
Milatz et al., 1951
8 min
~
Table 36.1 continued >
Chapter 34- E. Kerstel
c02
technique 1 laser
molecule isotope sample condition / ratio measurement range ~
co2
13c
~~
~
~~
setup1
lo-precision /
analysis time
exhaled breath, 3-5% COz 5 ml at 3 torr, < 100 nmol
~
~~
~
Optogalvanic LARA'lM
R2
Direct absorption, tunable 4.3 pm Pbsalt laser, transient
R1
accuracy2 ~~
1-2 min
~
~
13c
pure CO2 2 pmol natural abund.
co2
13c
natural air, 350 ppmv CO2 FTIR, 8 cm-1 resolution
coz
13c
exhaled breath, 0.5-5%C02 50 ml (ca. 10 pmol C02) A613C--25%o to +25%o
c02
c02
Murnick & Peer, 1994 Murnick et al., 1998 Murnick, 2001
<2%0P
Sauke & Becker, 1998
R1
- 60 min
-0.7%oP
Soderholm et al., 1998
NDIR, photoacoustic IRIS 2TM
R2
2 min
< 0.3%oP
Wagner Analysen Technik, http: / / www.wagnerbremen.de
13c
Direct absorption, tunable Pb-salt laser
R
-10 min
3yOoP
Wong, 1985
13c,
Direct absorption, pure CO2 4.3 pm Pb-salt laser, -100 pmol transient detection [13C]/[12C]=lyhto 12% [17O]/[*60]=0.04%,to 2.6%) [180]/ [160]=0.2%to 51%
C
102, relative error
Lehmann et al., 1977
exhaled breath, 3-576 C02 5 ml at 3 torr, < 100 nmol
R2
0.5c%,oP
Murnick, 2001
170, ' 8 0
co2
~
~~~~~~
co2 c02
reference
'80
Optogalvanic LARATM
1-2 min
Isotope Ratio Infrared Spectrometry
> Table 36.1 continued
co2 ~
~
~~~~
~~
Table 36.1 continued > 763
764
> Table 36.1 continued molecule isotope sample condition 1 ratio measurement range
technique
1 laser
setup1
analysis time
10-precision / accuracy2
reference
170, gas phase H20 180,2H -1Opl H20 (1) 6 1 7 0 = - 1 0 0 ~to ~ +30o%~ 6180=-200%0to +1,200%0 62H=-500%0to +12,000%0
Direct absorption, color center laser at 2.7 pm AM-detection
R2
20 min, incl. evac/refill
<0.4%0P,0.5% to 2%0A (170); <0.5%oP,0.5%0 to 3.5%0A(180); 0.6%oP, 1%0 to 6O%oA(2H)
Kerstel et al., 1999 Trigt et al., 2001,2002a,b
H20
170, gas phase H20 ISO,2H -10111 H20 (1)
Direct absorption, DFB diode laser at 1.39 pm If-detection
R2
20 min, incl. evaclrefill
0.5%0P('70); 0.2%0P (180); 0.5%oP (2H)
Kerstel et al., 2002 Gianfrani et al., 2003
Hz0
180
gas phase H20 -5 p1 H20 (1)
photoacoustic, pulsed dye-laser at 720 nm
R2
4 min, excl. evac / refill
16%0P
Matsumi et al, 1998
H20
2H
liquid H20 -300 1.11 [D]/ [HI = 0.002 - 0.08%
4 pm band pass filtered IR bulb. PZT-photothermal detection
R1
-4 min.
-
N D I / [HI) 0.002% (-100%0P)
Annyas et al., 1999
H20
2H
FTIR, liquid H20 60 pl 8 cm-1 resolution 62H=-500%0to +16,000%0
R1
-12 min
-3%*P 3740 to 45%0A
Fusch et al, 1993
H20
2H
liquid H20 -100 pl [D]/ [HI > 0.09% (62H > 5,00070)
R2
-2 min
5%0 to 30%ooP
Fusch, 1985
integrated vibrational band near 4 pm with grating spectrometer
Table 36.1 continued >
Chapter 34- E. Kerstel
H20
~~~~~~
~
molecule isotope sample condition / ratio measurement range
setup1
analysis time
1 o-precision / accuracy2
reference
R2
30 min
2% relative error
Gaunt, 1956
FTIR 0.012 cm-1
R1
30-60 min
1-2%0I'(15N); 2.5-4%oP (180)
Turatti et al., 2000
Direct absorption, 8.1 pm cw-QC laser 1f-detection
C
5 min
--3OyOoP
Gagliardi et al., 2002
30-60 min
4.4%0P
Esler et al., 2000a
technique
/ laser
integrated liquidH20 vibrational band -3 ml [D]/ [HI = 0.015% to 0.08% near 4 pm with grating spectrometer
H20
2H
N20
15N ( a pure N20 and fl), 6-9 pmol 180
N20
15N (p), pure N20 180 0.5umol
N20
170
pureN20 6-9 ymol
FTIR 0.012 cm-1
R1
N20
180
pureN20 25 pmol
Direct absorption, tunable 8.2 pm Pb-salt laser If-detection
R23
0.4yOoP
Wahlen & Yoshinari, 1985
0 3
180
laboratory generated pure Direct absorption, 0 3 tunable 10 pm I'd-salt laser
C
estimated -5o%oP
Anderson et al, 1989
R1 indicates a true ratio measurement with respect to a reference material, with sequential measurement of sample and reference gas in one and the same gas cell; R2 indicates a true ratio measurement with simultaneous measurement of a sample and a reference gas cell spectrum; C refers to a concentration measurement in a single gas cell. 2 The precision (P) and accuracy (A) are 1-0standard deviation values. The accuracy refers to calibration with respect to IRMS. Both figures are based on claims presented by the authors and should be treated with the necessary skepticism. Not always are the data presented that ought to support the claims, and precision and accuracy are sometimes confused. 3 The laser is alternatively passed through the reference and the sample cell.
Isotope Ratio Infrared Spectrometry
> Table 36.1 continued
1
765
766
Chapter 34- E. Kerstel
for respectively 6180 and 62H using IRMS. However, due to the prevalence of the 13CO2 molecule, which shows up at the same mass as 17OCO, the accurate determination of the 170/160 isotope ratio in water is difficult, if not impossible. The chemical preparatory steps not only severely limit the over-all throughput, but often also compromise the attainable accuracy, and can even be potentially hazardous (in the case of reduction over hot uranium). This is particularly true for 62H, as the water to hydrogen gas reduction step is accompanied by a very large fractionation, which is not always easy to quantify. The accuracy for 62H is further reduced by the substantial correction that has to be made for H3+-formation in the ion source of the IRMS. Similar problems exist with IRMS measurements on methane and other molecules. Yet another problem is posed by molecules such as NNO: the inability of IRMS to directly measure site-selective isotope-ratios, without resorting to chemical pretreatment or the need to make assumptions regarding the fragmentation pattern in the ionization source; in this case, to distinguish between 15N isotopic substitution at the c~ (center) or 13(end) site. Strongly depending on the application, less fundamental, but more practical limitations of IRMS are associated with the size and weight of the instrumentation (even though this has not prevented some from operating an IRMS on a balloon-born platform), as well as with its requirement of operation by a trained technician and high capital and maintenance costs.
34.2 Infrared spectrometry The near to mid infrared is often referred to as the "finger print" region of the optical spectrum: most small molecules exhibit highly characteristic rotational-vibrational bands in this region, associated with rotational and vibrational bending and stretching motions of the nuclei (Wilson et al., 1980). The usually strong fundamental vibrational bands are found towards longer wavelengths (typically from 2.5 gm and onwards) and correspond to the excitation of a molecule from the vibrational ground state to the first vibrationally excited state. Overtone and combination bands involve excitation of more than one quantum of vibrational energy. Both are typically at least one order of magnitude weaker than the fundamental transition, and occur at wavelengths shorter than 2 gm. The changes in the rotational motions of the molecule that accompany the absorption or emission of an infrared photon give rise to the fine structure observed at sufficiently high resolution (resolving power) (see Figure 34.1). The resulting spectra are highly sensitive to isotopic substitution of the molecule. This can be exploited to measure isotope ratios using a number of different optical detection schemes. Although these methods can be quite different in their underlying physical principles, the basic principles of optical isotope ratio measurements are the same. In general, all optical techniques rely on the registration of spectral features that can be uniquely assigned to the isotopic molecular species ("isotopomer") of interest, whether they are (vibrational-) rotational transitions in a high resolution spectrum, or the complete vibrational bands in a liquid, a high-pressure gas phase, or an otherwise (instrument limited) low-resolution spectrum.
Isotope Ratio Infrared Spectrometry
767
Figure 34.1 - The near to mid infrared spectrum of a natural abundance water sample at room temperature. The spectrum is a simulation based on literature data (Rothman et al., 1998; Toth, 1994). The line widths are assumed to be infinitely small. The band structures around 3 and 6 ~m are fundamental vibrational bands associated with, respectively, stretching and bending motions of the molecule. The band structures near 1.8 and 1.4 ~m are associated with combination and overtone vibrational excitations. The fine structure is due to rotational motions.
The last decade has shown a fair number of successful applications of infrared absorption techniques to the measurement of isotope ratios, somewhat trailing the development of tunable infrared laser sources. However, the idea to use the infrared spectrum for this purpose is by no means new. We are aware of at least one publication in 1950, describing the use of the 4 ~m infrared absorption of HOD to measure the deuterium concentration in liquid D 2 0 / H 2 0 mixtures (Thornton, 1950). Already in 1951, Milatz and co-workers (1951) had built a dual beam nondispersive infrared absorption instrument, using photoacoustic detectors, for the 13C/12 C determination in CO2, which in its essential features was very similar to its more modern counterparts. It should be noted that the isotope ratio thus determined is in fact a molecular, rather than an atomic abundance ratio. Although in general the molecular quantity is not equal to its atomic counterpart (e.g., 8(H1602H)~ 82H), in all practical cases the difference is very much smaller than the measurement precision, principally owing to the very low abundances of the rare isotopes, as demonstrated in the Appendix.
768
Chapter 34- E. Kerstel
34.3 Absorption spectroscopy In absorption spectroscopy the frequency dependent dimensionless absorbance c~ of the spectral feature is assumed to be proportional to the abundance of the species (Smith et al., 1985): R(v) : S 9 y (v-Vo)
9n 9l
[34.1]
Here, S represents a line strength factor depending on the transition dipole matrix element connecting the ground and exited states. In general, S is a property of the isotopomer, as well as temperature (see equation [34.4]). The function f(v-vo) represents the (normalized) line shape, with v the frequency of the light source and Vo the centerline frequency. The function f is a convolution of the instrumental line width (spectral resolution) and the molecular absorption profile. It is determined by experimental, as well as molecular factors (such as collision broadening). The concentration of the isotopomer is given by the number density n (molecules per unit volume). Finally, l represents the effective optical path length. The product of S and f has the dimension of 'm2' and may be viewed as a cross section o(v) for absorption of a photon. The absorbance a is in general not measured directly, but obtained by a proper transformation of the measured quantity (e.g., through Beer's law, as we will see in section 34.3.1). To see how one obtains the isotope ratio from a spectroscopic measurement, it is instructive to consider the data presented in Figure 34.2. In this generalized isotope ratio measurement, two spectra were recorded in the same spectral region encompassing spectral features belonging to the most abundant isotopomer (subsequently identified by the superscript a) and at least one rare isotopomer (superscript x). One spectrum is that of a reference material (subscript r; traceable to one of the internationally accepted standard materials, such as VSMOW in the case of water), the other of the sample being studied (subscript s). The isotope ratio is given by xR=xn/an. It is however customary to use xS, the relative change in the isotope ratio with respect to that of a standard material. Without loss of generality we can take our reference material to be this standard, in which case the 5-value is defined as:
X~)r(S ) =-XRs 1 - (Xn/an)s - 1 XRr ( Xn / an )r
[34.2]
If one realizes that f0) is inversely proportional to the width F of the transition, provided the exact same lineshape function applies, this yields together with equation [34.1]: x ~)r(S)
- (x c~/ a(~)s.(XF/ aF )s . ( a S /XS)s . (al/Xlls -1 (xo~/ao~)r (xF/aF)r (aS/xSlr (al/Xl)r
[34.3]
where c~ is short for the peak intensity c~(Vo). Thus, the 6-value is obtained from the "super ratio" of the center-line intensities a(Vo) in the two spectra. Alternatively, the
Isotope Ratio Infrared Spectrometry
769
Figure 34.2 - A very short section of the infrared spectrum showing two rotational-vibrational transitions in a sample and a reference material. In this particular case, both are natural abundance, gaseous water absorbing near 3663 cm-1 (2.73 ~m). The weakest transition occurs in H18OH, the strongest in H16OH. These lines are separated by 0.12 cm-1 and at the gas pressure of 13 mbar, exhibit a half-width-at-half-maximum (hwhm) of 0.008 cm-1. The area under the individual lines or, as indicated here, the peak intensities can be used to calculate the 6-value. integrated intensity may be used instead of the product of peak intensity and line width, in which case the result is valid also for differing lineshapes. In order to calculate the actual peak or integrated intensity, several approaches are possible. One can carry out a direct integration under the measured curve, if the lines are well separated, or first fit a sum of the proper line profiles (like the Voigt profile) to the spectrum. It is also possible to fit appropriate sections of the sample spectrum to the reference spectrum, with the ratio of the line intensities (X~s/XRr and a~s/a~r) as a variable (see, e.g., Bergamaschi et al., 1994). Whichever procedure is adopted, it is essential that sufficient baseline is available in order to arrive at a correct estimate of the (ratio of) absorbances. With a proper choice of experimental conditions, the three factors multiplying the super ratio of the intensities in the second half of equation [34.3] reduce to unity. In most practical cases, the width of the spectral features is the same for both isotopomers, or at least their ratio and temperature dependence is the same in the reference and sample spectra. Also, l (the optical path length, in the case of direct absorption) can be made the same for each isotopic species, or for the sample and reference spectra, or for both. The line strength S depends on the number of molecules in the lower level of the transition and is therefore in general temperature dependent (it also includes the effect of induced emission, which however is negligibly small in the infrared)" a change in temperature will redistribute the population over the rotational levels of the ground vibrational state. Allowing for a small temper-
Chapter 34- E. Kerstel
770
ature difference between sample and reference, AT = Ts - Tr, equation [34.3] is in first order equal to" Xc~r(S )
[
(x(x/a(x)s
1 ( xe~/ aO~) r 9 1-XS(rr) (xa/aa)~ (xlx / a(x )r
~(OT]
1
r - aS(Vr)
} "~;"Z") r " A Z
- 1
9{ 1 - [ x ~ - a~]. A T } - 1
= X~3r(S)*" { 1 - [x~ _ a~]. AT} - [x~ _ a~]. AT
[34.4]
where ~ is the (relative) temperature coefficient of the intensity of the spectral feature associated with, respectively, the rare and the most abundant isotopomer. For the case of a static gas cell (constant number density) the coefficients may be calculated from (see, e.g., Rothman et al., 1998): E" kT
E'- E" kT
Q(TF) e 1-e S(T) - S(TF)" Q(T) ~" ~,_~,, 9
[34.5]
,
e
kT
T
1-e
kT
T
Here, E" and E' represent the lower and upper state energy, respectively. The penultimate term in equation [34.5] accounts for the ratio of Boltzmann populations, while the last term accounts for the effect of stimulated emission, which, however, may usually be neglected in the infrared. If the temperature coefficients need to be evaluated at constant pressure, rather than at constant volume, the right-hand side of equation [34.5] should include the term Tr/T, which is proportional to the number density. Equation [34.5] ignores the fact that in a static gas cell, the amount of material absorbed on the walls may also be temperature dependent. This may be a significant effect, e.g., in the case of water. The ratio of total internal partition functions is most easily calculated from a parametrization of Q(T) as carried out for the HITRAN molecular spectroscopic database (Rothman et al., 1998). Often a good estimate may be obtained by ignoring the temperature dependence of the vibrational partition function, as well as rotation-vibration interactions, and approximating the rotational partition function by its classical form (i.e., the rotational level spacing is small compared to kT):
3 Q ( T ) ~ Q r o t ( Z ) . Q v i b ( T ) = QClassical(rot
T) ~
~Id31IAIBIC1
(2kT~h2 J ~2
[34.6]
where o equals the number of symmetry operations of the rotational subgroup of the molecule. Thus, e.g., o - 2 for H20 and o - 12 for CH4. The moments of inertia of the molecule are given by IA, IB, and Ic. The relation with the rotational constant A (or B, or C), expressed in wavenumbers (cm-1), is given by hcA = h2/(2Ia). Around room
Isotope Ratio Infrared Spectrometry
771
temperature the temperature coefficients are typically of the order of several per mil per degree Kelvin. A temperature difference between sample and reference may thus results in a significant scale error and zero-offset. In general, it is imperative to design the experiment such that either (a~_x~) or AT is minimized, or both. 34.3.1 Direct absorption spectroscopy As mentioned above, Beer's law of linear absorption (also known as the BouguerLamber-Beer law) relates the absorbance a of equation [34.1] to the measurable quantity of relative absorption, or absorptance, of monochromatic light of frequency v in a gaseous medium (see, e.g." Demtr6der, 1982)" = e
or"
c~(v) - - l n
[34.7]
It should be noted that in practice equation [34.7] does not apply exactly, due to the finite resolving power of the spectrometer. The observed spectrum is always distorted by the finite instrumental resolution. Several techniques have been developed in the past to correct for errors introduced by the finite resolution, such as the Wilson-Wells method (Wilson & Wells, 1946) and the "curve-of growth" method (see for a discussion: Smith et al., 1985). Using modern high resolution laser techniques, the need for these corrections is almost absent. In addition, in an isotope ratio measurement, the errors will usually cancel to a large extent in the final super-ratio of equation [34.2] or [34.3]. Still, care has to be taken to correctly locate the 100% transmittance level. In general there will be an optimal value of the transmittance It/Io maximizing the signal-to-noise ratio of a. If the transmittance is very close to unity, the difference between two almost equal quantities will carry a large measurement error, and inversely, if the sample is nearly black, the relative error on the transmitted light level will explode. We can quantify the preceding, for the simple case in which the measurement of I0, as well as the intensity It transmitted through the gas cell, are each inflicted with a measurement error ~I that is independent of the signal level (this will be the case if detector a n d / o r amplifier noise is the limiting noise factor). In this case it is straightforward to show that the S / N of the absorbance a(Vo) equals:
S/N
(x(v0) Ioi t (I~) - Ac~(v0) = ~i(I ~ + it ) ln
[34.8]
The maximum S / N is obtained for a transmittance It(vo)/I0(vo) = 0.28, corresponding to an absorbance c~ of 1.28. In addition, if we demand that the S / N be larger than 50% of this maximum, It(vo)/I0(vo) should be between 0.048 and 0.71 (i.e., the attenuation should be between 95% and 29%, or the absorbance between 3.0 and 0.33). In experiments that measure I0 and It directly, the optical path length a n d / o r the sample density (concentration) should be adjusted to yield absorbances that on average are in this range, if possible.
772
Chapter 34- E. Kerstel
The ratio of transmitted intensity It to incident intensity I0 can be measured most directly using a transient recording technique, or by amplitude modulation in combination with phase sensitive detection. In fact, Lee & Majkowski (1986), in one of the earlier demonstrations of an isotope ratio measurement by infrared laser spectroscopy, used a transient recording technique to measure the C180/C160 abundance ratio in a 3 mbar pure CO sample with a reproducibility of about 3%0. In this case, a narrow line-width, single mode, 4.7 gm lead-salt diode laser was rapidly scanned across two nearby but well-resolved rotational lines of 12C160 and 12C180. While recording the laser intensity transmitted through the dual-path gas cell and impinging on a dc-coupled infrared detector, the strong 12C160 absorption was detected in the short arm, the weaker 12C180 (as well as an even weaker 12C170) absorption in the long arm of the instrument. The value of I0 was inferred from a measurement with an empty gas cell. No attempt was made to compare the result to a reference material (traceable to a primary standard). In this sense, the measurement should be categorized as a concentration, rather than an isotope ratio, measurement. In our laboratory we developed a direct absorption spectrometer aimed at the measurement of water isotope ratios using amplitude modulation of the laser light and phase sensitive detection (Kerstel et al., 1999, 2001). The transmittance It /Io is determined by measuring It and I0 simultaneously for both the sample and the reference channels. In this way, coherent laser noise, present on both signals, is divided out, significantly improving the quality of the spectra. The laser source is a color center laser operating near 2.73 ~m. Four transitions occurring within a 1 cm-1 range were identified, each corresponding to a different dominant isotopomer (H16OH, H18OH, H17OH, or H16OD), with approximately equal absorption in the case of a natural abundance water sample. The fact that the selected transitions are relatively strong, together with modest multiple passing of the laser beam between two spherical mirrors inside the Herriott-type gas cells (Herriott et al., 1964), assures that the sample size can be rather small (5 to 10 ~L of liquid water in a I L volume), while the absorbance is still close to optimal (cf. equation [34.8]). The apparatus has been extensively calibrated against international standards (provided by the International Atomic Energy Agency) and local reference materials (back-traceable to IAEA standards through repeated IRMS measurements), over a very wide range of isotopic compositions (62H from -428%o to +15,000%o; ~180 from-55%o to +1,200%o; 6170 from -100%o to +300%0). The accuracy for 62H (better than 1%o at natural abundance) is competitive with IRMS over the entire range and limited by isotope fractionation in the sample handling and by memory effects. For ~180 the accuracy is similar to that of IRMS at high enrichment levels. But, at a level of 0.5%o, the accuracy is much lower than what can be obtained by IRMS in the natural abundance range. One analysis, the simultaneous measurement of all three isotope ratios, takes about 20 min and does not require any sample preparation. The technique has been successfully applied in biomedicine (van Trigt et al., 2001, 2002b), ice-core research (van Trigt et al., 2002a), and the determination of the 170 content of the D20 used in a solar neutrino experiment (Kerstel, 2001).
Isotope Ratio Infrared Spectrometry
773
34.3.2 Frequency modulation spectroscopy The sensitivity of absorption spectroscopy can be dramatically increased by the use of frequency modulation of the laser wavelength in combination with phase sensitive detection at the modulation frequency, one of its higher harmonics, or even the difference of two modulation frequencies (two-tone modulation). Normally, the term wavelength modulation spectroscopy (WMS) is used for the case of modulation at relatively low frequencies (in the kilohertz range; small compared to the typical gas phase absorption line width of the order of 100 MHz), while the term frequency modulation spectroscopy (FMS) is reserved for modulation at very high (RF) frequencies. Although the practical implications for the experimental implementation are considerable, there is no fundamental difference between the two extremes and both can be described within the same theoretical formalism (Silver, 1992; Supplee et al., 1994). In all cases, the contribution of 1/f (laser source) noise is drastically reduced, as well as the baseline slope a n d / o r curvature associated with direct absorption. Such a detection technique is particularly well suited for application with diode lasers, which can be easily frequency modulated through the injection current. The line shapes recorded in a WMS experiment are so-called derivative line shapes, as they closely resemble the first (lf) or higher (2f, 3f, ...) derivative of the direct absorption line profile (see Figure 34.3). An advantage of WMS is that not only the S / N of a frequency modulated spectrum is generally (much) higher than that of its corresponding amplitude modulated spectrum, but also that a fit to the data of a derivative line shape function over a relatively short section around the line center will recover the peak intensity more reliably. A complication is that the measured signal is no longer proportional to the transmittance, but rather to the absorptance AI/Io =- 1- It /Io. Consequently, the signal is
Figure 34.3 - First-derivative spectra of two identical water samples (reference and sample) recorded with a frequency modulated 1.39 ~m DFB diode laser.
774
Chapter 34- E. Kerstel
only linearly proportional to the absorbance for small fractional absorptions. At higher absorptance, deviations from linearity of the Beer law become visible: c~--ln
(A,)
AI l(A,)2 1(AI)3
1-T~~ -T~0+~_ To ~
[34.91
+~ ~-0 +""
For example, at a fractional absorption (absorptance) of 1%, the deviation from linearity is 5%o, while at 10% it is already 53%0. This may lead to sizable errors in the case of a large disparity in isotope concentrations between sample and reference material. This situation should be either avoided by selecting weaker transitions, using less material, a shorter path length, or a combination of these (often not by free choice!). Or it should be corrected for, conform equation [34:.9]. In order to be able to carry out this correction, the direct absorption signal has to be recorded after all (and preferably simultaneously). For example, by low-pass filtering the dc-output of the detector, as indicated in the experimental setup of Figure 34.4. As an example of WMS, we mention the work of Bergamaschi and co-workers, who successfully used a tunable lead-salt diode laser set-up for isotope ratio measurements on methane (Bergamaschi et al., 1994; Schupp et al., 1993). Two lasers were required to access the two spectral regions used for the 62H and 613C determinations. The optical path length was about 200 m, obtained with a commercial multiple-reflec-
Figure 34.4 - Typical experimental arrangement using a frequency modulated, single mode diode laser. Shown are the distributed feed-back diode laser (DFB), optical isolator, a second collimating lens, multiple-pass gas cells, detectors, low-pass filters, and lock-in amplifiers (LIA).
~
.....................
|~........... ............................ , ~ |~ "~ ~ ~ | L ".................. ........ "" ............... ~"~i i i~:::::~:~84 ..... ::"
.......................... u
....................
Isotope Ratio InfraredSpectrometry
775
tion gas cell of the design by White (1976). The instrument was compared directly to IRMS (and later used on a routine basis) for a variety of atmospheric methane samples from natural wetlands and landfills. The overall accuracy achieved is 0.6%0 for 613C (with a minimum sample amount of 2 ~mol CH4) and 5%0 for ~2H (requiring 80 ~mol CH4). The throughput (one ~5-measurement per 15 min.) is significantly higher than (was) possible with IRMS preceded by methane conversion to CO2 and H2. Uehara and co-workers (2001) used modern distributed feedback (DFB) semi-conductor lasers in the 1.7 ~m and 2 ~m spectral range for ~13C and ~15N determinations in CH4 and N20, respectively. Without doubt the most innovative feature of their approach is the use of different lasers, each locked to the transition of one of the isotopomers. Thus, instead of tuning the laser repeatedly over spectral features belonging to 13CH4 and 12CH4, two lasers are wavelength stabilized at the center of the respective transitions. Similarly, in the case of N20, three lasers are used for detection of the 14N15N160 (c~-isotopomer), 15N14N160 (~-isotopomer), and 14N14N160 isotopomers. This has the advantage that the spectral features used in each case do not have to be near to each other (i.e., not necessarily within the tuning range of the laser). At the same time, two optical paths of 100 m and I m are realized within the same multiplepass gas cell, that effectively compensate for the 1-100 natural abundance ratio of 13CH4 and 12CH4, and almost compensate for the 0.4"100 15N:14N abundance ratio in N20. The resulting freedom is used to select transitions that originate from ground state levels with very nearly the same energy. This drastically reduces the difference in the temperature coefficients of the line intensities (a~/x~ of equation [34.4]). The transitions also have approximately equal strengths, such that the absorbances are balanced. The precision of the method was demonstrated to be 0.3%o for 13h(CH4), requiring about 100 /~mol CH4, and estimated to be about the same for 15~(N20), albeit at the cost of a larger sample size of about 2 mmol. The ability to measure the site-selective isotope ratios in N20 at this level of precision, without chemical sample pretreatment, is certainly attractive for studies of N20 dissolved in the oceans, trapped in Antarctic snow, and of the global atmospheric N20 budget, but the currently required amount of sample is at least a factor of thousand too high (Rahn & Wahlen, 2000; Yoshida & Toyoda, 2000; R6ckmann et al., 2001). In Groningen, we have used an alternative multiple-laser solution to the problem of selecting an optimal combination of water isotopomer absorption features. Two diode lasers were scanned over two different spectral regions near 1.39 ~m (about 0.5 to 1 cm-1 wide, but separated by N15 cm-1), one chosen for the presence of favorable H16OH, H17OH, and H18OH ro-vibrational lines, the other for a combination of H16OH and H16OD lines. Both lasers traversed exactly the same beam path through sample and reference gas cells. The sample and reference spectra corresponding to each laser were recovered by wavelength modulation and phase sensitive detection at incommensurate frequencies. This dual-wavelength multiplexing scheme enabled us to record all 3 water isotope ratios of interest truly simultaneously. A precision of 0.2%o for ~180 and of 0.5%o for ~170 and i~2H were demonstrated (Gianfrani et al., 2003).
776
Chapter 34- E. Kerstel
34.3.3 Cavity Ring Down Spectroscopy In essence, cavity ring down spectroscopy (CRDS) and its derivatives are absorption techniques in which the effective optical path length is cleverly increased to far beyond what is feasible with standard multiple-passing. Due to this ability to obtain very long effective absorption paths, it is particularly suitable for trace gas detection and isotope ratio measurements on exceedingly small sample sizes. In CRDS narrow bandwidth laser radiation is coupled into a cavity composed of high reflectivity mirrors and enclosing the sample gas. The laser radiation is then abruptly turned off, after which the light inside the cavity will decay exponentially in time, due to scattering and absorption losses inside the cavity in addition to the loss of light at each reflection from the mirrors. This decay can be followed by recording the intensity of the light leaking out of the cavity as a function of time: t I(t)
-
[34.10]
Ioe ~
Provided the absorption in the cell obeys Beer's law, the decay time constant % known as the ring-down time, is given by: d 9(v) - c(1 - R + c~(v))
[34.11]
Here, d is the mirror separation, c the speed of light, R the mirror reflectivity (assumed to be nearly unity), and c~(v) - ~J(v).n.d the frequency dependent molecular absorbance. It then follows that the molecular absorption spectrum is obtained by plotting the decay rate 1/~ as a function of frequency: o(v)'n - -
-
4/
[34.12]
In the above equation, "to is the decay time of the empty cavity or with the laser tuned far from resonance (the base line). The effective path length is increased by a factor of 1/(l-R) with respect to single pass absorption. Since commercially available mirrors attain reflectivities up to 99.999% (depending on the wavelength- generally somewhat lower in the infrared), the effective path length can approach 104 m. Since equation [34.12] does not contain the length of the cavity, or other instrumental parameters, CRDS provides an absolute measurement of the molecular absorption. Also, because no laser light is entering the cavity during the measurement of the decay time curve, the technique is largely immune to laser amplitude noise. CRDS is based on early techniques to measure mirror reflectivities (Herbelin et al. 1980; Anderson et al., 1984). O'Keefe & Deacon (1988) developed it into a successful spectroscopic technique, using a pulsed dye laser as the light source. Since then, applications of the CRDS technique have proliferated (see, e.g., Scherer et al., 1997; Wheeler
Isotope Ratio Infrared Spectrometry
777
et al., 1998; Berden et al., 2000). In the quest for a higher spectral resolution, a higher signal-to-noise level, and a faster data acquisition rate (limited by the repetition rate of the pulsed laser), the technique has been extended to the use of cw lasers (cwCRDS) (Lehmann, 1996; Engeln et al., 1996; Romanini et al., 1997a), including more compact, single mode diode lasers (Scherer et al., 1995; Romanini et al., 1997b,c). Paldus and co-workers (1998) devised a clever way to lock the cavity mode to the diode laser frequency, further improving the performance by strongly reducing baseline oscillations, as well as the shot-to-shot noise on the decay time (down to 4%0), but at the cost of much greater complexity. Meijer and co-workers (Engeln et al., 1998) developed the very promising alternative of Cavity Enhanced Absorption Spectroscopy (CEAS), an intermediate between direct absorption spectroscopy and CRDS. Instead of measuring the temporal profile of the radiation leaking out of the cavity, the total time-integrated intensity is measured, which is inversely proportional to the decay time. There is no need to switch the laser output on and off, and no fast detector and data acquisition system are needed, greatly reducing the complexity of the technique. A drawback is that, as for direct absorption, the signal is no longer independent of the laser intensity. In CEAS the cavity is not locked to the laser frequency. Rather, in order to optimize the efficiency of coupling light into the cavity, a cavity is constructed with a very dense mode structure. Light is coupled into as many modes of the cavity as possible, relying on accidental overlap of the narrow band laser radiation with one of the multitude of longitudinal and transverse cavity modes. The cavity mode structure can be optimized (made more dense) by construction of a non-confocal cavity (Engeln et al., 1996), or by off-axis coupling into the cavity (Baer et al., 2002). An absorption spectrum is obtained by rapidly sweeping the laser over the spectral region of interest and recording the detector output (alternatively, the cavity modes are scanned by modulating the length of the cavity with a piezo-element). The single-sweep spectrum is the convolution of the cavity mode structure and the actual absorption spectrum. The influence of the mode structure can be removed by 'random interleaved sampling', induced by mechanical instability of the optical set-up, thus flattening the spectral response by co-adding repeated scans. This reliance on the mechanical instability of the cavity actually makes the technique inherently robust! Recently, Romanini and co-workers have shown that optical feedback from a (Vshaped) cavity can be used to enhance the injection of diode laser light into the ringdown cavity (OF-CRDS). The resulting setup is almost as simple, robust, and low-cost as CEAS, but with the additional advantage of measuring the decay time directly, thus providing an absolute calibration of the absorption (Romanini et al, 1999; Morville et al, 2001; Morville et al., 2003). Practically the same optical feedback- cavity injection technique has been used in a cavity enhanced scheme (OF-CEAS), reducing the number of optical components and complexity of the system even further (Romanini 2003). The advantage of instantaneous absolute absorption calibration is lost, but may be restored by temporarily switching the instrument to the OF-CRDS mode. In addition, this particular OF-CEAS scheme yields an absorption spectrum sampled at frequencies that are spaced by an integer multiple of the cavity free spectral range. This
778
Chapter 34- E. Kerstel
may be important, e.g., in a single gas cell setup, when the sample spectrum cannot be compared directly to a simultaneously recorded reference spectrum. In such a case, frequency linearization by means of an external etalon would otherwise be required in order to reduce frequency-induced amplitude noise. Dahnke and co-workers (2001) used cw-CRDS to measure the 13C isotopic composition of CH4 in natural air. They produced tunable radiation at 3.3/~m by the generation of weak sidebands on the output of a CO gas laser. The tunable sideband radiation is mode matched into the ring-down cavity. The cavity length is controlled in a feed-back loop in order to lock a single longitudinal mode of the cavity to the sideband frequency. Since two sidebands are generated, one on each side of the CO laser line, a monochromator is needed to separate the two. In all, the experimental setup is fairly complex. The 12CH4 and 13CH4 absorbances are measured in a natural air sample (about 2 ppmv CH4 and a total amount of 4 nmol CH4) with an estimated precision of 6%o and 12%o, respectively. Consequently, the precision (reproducibility) of an isotope ratio measurement would at best be of the order of 13%o; to be compared to a precision much better than 1%o required in most atmospheric methane studies (Mroz, 1993; IAEA, 2001). Crosson et al. (2002) applied the cavity-locked cw-CRDS technique first described by Paldus et al. (1998) to the measurement of the 13CO2/12CO2 ratio in exhaled human breath. Given the elevated CO2 concentration of the breath samples (2 to 5%, or about a hundred times the concentration in natural air), a precision of 0.3%o could be obtained with an industrial prototype instrument. CRDS has clearly proven to be able to achieve extremely high detection sensitivities. However, in order to measure isotope ratios sufficiently accurate to be useful for environmental studies, one needs above all a good signal-to-noise ratio, long term instrument stability (reduced complexity), and the ability to calibrate the measurements with respect to a reference material on time scales short compared to those associated with instrumental drifts.
34.3.4 Non-dispersive techniques The most common form of non-dispersive infrared (NDIR) absorption spectroscopy uses an amplitude modulated broad band light source (lamp) and gas-specific detector units. Generally a dual-beam configuration is used in which one channel measures the background signal in an empty or inert gas filled cell, the other the sample gas of interest. In order to measure isotope ratios, this set-up is duplicated, with one dual-beam channel equipped with a rare isotopomer detector, the other with a abundant isotopomer detector. Gas- and isotopomer specificity of the detectors is achieved by filling a small chamber with the pure isotopomer gas. Heating of the gas is detected with a microphone (photoacoustic detection), or alternatively, the transmission through the chamber is measured with an infrared detector. NDIR is by definition a spectrally low-resolution technique yielding a signal that is the integrated absorption over a relatively large fraction of the absorption spectrum of the molecule of interest (in practice limited by a narrow band-pass filter). As described here, it
Isotope Ratio Infrared Spectrometry
779
requires that at least part of the isotopomer absorption is well separated from that of the most abundant molecule. The selectivity can be high for 13C-analysis in CO2, as the asymmetric stretching band used in this scheme is shifted considerably (66 cm -1 ) u p o n 13C substitution. An entirely different implementation of NDIR is possible, that actually depends on the overlap of the isotopomer bands. In this heterodyne scheme, the signal is derived from the absorption of amplitude modulated infrared radiation by a pressure-modulated gas (Dimeff, 1972; Irving et al., 1986). A disadvantage of NDIR over higher resolution techniques is its relative sensitivity to foreign gas interferences. Moreover, since the detailed information contained in a high-resolution spectrum is not available, one has no built-in check on sample contamination. On the other hand, the technique is relatively simple, low-cost, and robust. This explains why NDIR has become the optical method of choice for 13CO2/12CO2 measurements on exhaled breath for detection of Heliobacter pylori infections (Hildebrand & Beglinger, 1997; Schadewaldt et al., 1997, and references therein). The instrument built by Haisch and co-workers on the basis of a commercial (Hartmann & Braun) NDIR gas analyzer (using photoacoustic detectors), achieved an accuracy of 0.3%0 on 0.5 L exhaled breath samples containing between 3% and 5% CO2, calibrated against IRMS over a 50%0 range in gl3C (Haisch et al., 1996; Koletzko et al., 1995). Validation studies carried out by others in a clinical setting revealed somewhat reduced accuracies of the order of 0.6-1%o (Hildebrand & Beglinger, 1997; Schadewaldt et al., 1997). The discrepancy is most likely related to sample handling and instrument calibration difficulties in the clinical versus the laboratory environment, a n d / o r foreign sample gas interference (like water). One of the commercial implementations of this apparatus (the IRIS 2TM by Wagner Analysen Technik, Bremen, Germany) claims a precision < 0.3%o for a sample size of only 50 mL and CO2 concentrations in the range of 0.5 to 5%.
34.3.5 Fourier-transform infrared spectroscopy In a Fourier transform infrared (FTIR) spectrometer a broadband light source illuminates the sample which is held in a gas cell inside the stationary arm of a Michelson interferometer. As the length of the second arm is changed in a periodic manner, an interferogram is recorded that is the Fourier transform of the sample gas absorption spectrum. The resolution of the instrument is proportional to the path length difference traveled by the mirror in the second arm. In practice this limits the resolution of an FTIR instrument to about 0.01 cmq. At this level of resolution the instrumentation quickly becomes rather expensive and bulky. Fortunately, most small molecules of environmental interest (excluding H20 ) have an average rotational line spacing of the order of 1 cm-1, allowing for a relatively compact FTIR apparatus. Esler and coworkers (2000a,b) have shown that excellent results can be obtained with a commercial instrument with a spectral resolution of I cm -1. Retrieval of the molecular concentrations requires a sophisticated data analysis procedure, involving the calculation of synthetic spectra and a multivariate calibration algorithm. Their results for the measurement of ~)13C in atmospheric CO2 and on exhaled human breath (a precision of about 0.1%o), are probably the best that has so far been demonstrated with optical methods. Using a higher resolution spectrometer it is possible to resolve more closely spaced rotational structures and thus achieve sufficient isotope selectivity also for
780
Chapter 34- E. Kerstel
determination of ~)170 and 6180 in CO2 and CO, as well as site-selective 615N in N20 (Turatti et al., 2000; Griffith, 2001 pers. com.1). 34.4 Indirect spectroscopic techniques For detection methods other than absorption spectroscopy, the relation between the observed signal and the molecular density is often linear (at least in the small signal limit), but usually depends on other parameters than seen in Beer's law (equations [34.1] and [34.7]). For the techniques discussed in this section, the most prominent of these is the intensity of the laser radiation. As the signal is directly proportional to the number density of the absorbing molecules, techniques like photoacoustic and optogalvanic spectroscopy have zero background. This makes them potentially very sensitive detection techniques.
34.4.1 Laser photoacoustic spectroscopy In photoacoustic (PA) spectroscopy a sound wave, generated when a sample gas in a closed volume absorbs amplitude modulated infrared radiation, is detected by a sensitive microphone. In combination with powerful CO2- or CO-line lasers extremely high detection limits have been demonstrated, often exceeding those of CRDS (Harren & Reuss, 1997; Harren et al., 2000). With the advent of powerful and tunable diode laser in the (near) infrared (Paldus et al., 1999), PA spectroscopy should become more widespread, especially in the trace gas detection field (Boschetti et al., 2002). So far, the only isotope ratio measurement performed with laser PA spectroscopy, showed a disappointingly low precision (approximately 16%o) on the measurement of 6180 in natural water, largely due to the choice of exciting very weak third overtone OH stretching transitions near 721 nm (Matsumi et al., 1998).
34.4.20ptogalvanic spectroscopy The optogalvanic signal is the electrical response of a low pressure RF gas discharge to optical excitation. When the laser emission is in resonance with a transition that couples a lower state to the ionization continuum the resulting change in the density of charge carriers can be detected as an impedance or voltage change across the discharge (Barbieri et al., 1990). Murnick & Peer (1994) developed a 13CO2/12CO2 isotope ratio spectrometer based on this optogalvanic effect. They demonstrated a precision of better than 1%o for 613C. A potential problem of the method is that the relation between optogalvanic signal and molecular density is not only dependent on the laser power, but also on the discharge conditions and other experimental details, such as the modulation frequency (Rusak et al., 1999). It is also different for the 13CO2 and 12CO2 isotopomers. The required level of precision and accuracy can therefore only be maintained through extensive and repeated (re-) calibration of the instrument and care has to be taken that contamination, e.g., by water, is avoided. 34.5 Discussion and conclusions So far, the precision of optical measurements has not been able to match that of IRMS. Except for a single case in which highly enriched water samples were mea1. Private communication with D.W.T. Griffith, Yokohama, August 2001.
Isotope Ratio Infrared Spectrometry
781
Table 34.2 - Absorption band strengths for CH4, CO2, H20, and N20. The data are taken from a compi-
lation by Pugh & Rao (1976) and Smith et al. (1985). molecule
vibration
wavenumber
wavelength
b a n d strength
[cm-l]
[mm]
[atm-lcm -2 at
300 K]* CH4
v4 v3 v3+v4 2v3
1306 3019 4321 6005
7.66 3.31 2.31 1.66
130 270 10 1.8
CO2
V21
667
15.0
200
V3
2349 3715
4.26 2.69
2450 44
4978
2.01
0.88
6228 6973
1.61
0.012 0.035
6.3 2.73 / 2.66
254
Vl / v3
1595 3657 / 3756
v2+v3 2v2+v3
5331 7150
1.88
1.4
20 15
Vl v3 2Vl
1285
7.78
230
2224
4.5
1400
2563 3481
32 46 1.8 1.5 0.013
Vl+V3 v1+2v20+v3 V1+4v20+v3 3v3 H20
N20
v2
1.43
Vl+V3 3Vl 2v3
3836 4417
3.90 2.87 2.61 2.26
2v1+4v20
4911
2.04
v21+2v3
4978
2.01
3v1+2v20
5026
1.99
4Vl
5106
1.96
134
0.07
* 1 atm-lcm-2 at 300 K corresponds to 4.088.10-20 cm.molecule-1.
sured (and in particular for the 82H measurements for which isotope fractionation effects during sample handling determine the ultimately attainable accuracy), IRMS is more precise than optical alternatives. However, using optical techniques, isotope ratio instrumentation can be build that potentially is cheaper, compacter, and easier to operate (air-borne, or at a remote location, if needed) than IRMS instrumentation. In some cases it may be necessary to perform a preconcentration of the sample. But, due to the extreme selectivity possible with infrared spectroscopy, chemical sample pretreatment is usually not required. If
782
Chapter 34- E. Kerstel
the need exists, the sample can be recovered, as the measurement is non-destructive. A high sample throughput can be obtained and many techniques are adaptable to online, even real-time, measurements. Techniques based on (direct) absorption spectroscopy have the additional advantage of being conceptually simple. This translates in the ability to carry out isotope ratio measurements that require a much smaller normalization and scale correction than is often necessary with IRMS measurements. It should be noted that isotope ratio infrared spectroscopy is not necessarily a matter of high detection sensitivity (minimal detectable absorption). Rather, what is important is the ability to measure with good signal-to-noise, and with a linear response over a large dynamic range. Still, as there is a need to handle ever smaller sample sizes, as well as molecules present at very low concentrations in a carrier gas (e.g., atmospheric air), techniques originally designed for trace gas detection, such as CRDS, will find their way into the optical isotope laboratory. Widespread application of optical isotope ratio techniques based on laser spectrometry depends on the easy availability of relatively powerful, compact narrow band laser sources with good mode structure and tunability. Ideally, such lasers operate at room temperature and can be fabricated at modest cost to lase at the desired wavelength. Currently, single-mode tunable diode lasers, operating continuous wave near room temperature, are commercially available only at wavelengths up to 2 gm. At these short wavelengths, the lasers probe the relatively weak overtone and combination bands in the small molecules that are of environmental interest. An absorption spectrometer designed to detect C 0 2 based on a 1.6 ~m diode laser, would show an improvement in signal-to-noise of almost a factor of hundred if the laser is replaced by a 2.0 gm laser, all else being equal. Replacing the 2.0 gm laser with a 2.7 gm laser would result in an additional gain with a factor of roughly 50. Table 34.2 provides a comparison of vibrational band strengths of four molecules of prominent interest in environmental research. The band strengths are given for the most abundant isotopomer only. In order to evaluate the sensitivity of a possible isotope ratio spectrometer, the band strength of the rare isotopomer and the isotope shift have to be taken into account as well, but the data of Table 34.2 give the general trend. It turns out to be very similar for most small molecules. The strongest bands are observed in the 3 gm region, not much is gained by moving towards longer wavelengths, while at shorter wavelengths the vibrational bands show a rapidly decreasing intensity. Even though a single mode diode laser made from group III-V compounds (including antimonide) operating near 2.7 gm at 180 K was demonstrated as long ago as in 1996 (Martinelli, 1996), it has not yet made it to the market place. The same is true for the III-V type diode lasers operating continuous wave between 2.2 and 2.4 gm at room temperature with output powers up to 50 mW developed by Alibert et al. (2001). The color center laser used in our laboratory has the advantage of being continuously tunable over the very wide range from 2.4 gm to 3.4 gm, delivering over 10 mW in a single mode. Although it is cryogenically cooled, it does not depend on temperature tuning of the emission wavelength, in contrast to diode lasers. But the color center crystals require an ion-laser pump and are probably better replaced by the
Isotope Ratio Infrared Spectrometry
783
room-temperature, Erbium fiber-laser pumped, chromium-doped chalcogenides, such as Cr2+:ZnSe (Sennaroglu et al., 2000; Sorokina et al., 2001). In the mid-infrared region above 3/~m, lead-salt lasers made from group II-VI materials have been the standard choice. These lasers are, however, difficult to work with as they show frequent mode hops during scanning, have a poor side-mode suppression, and deliver very little power (tens of microwatts). On top of this, they require a stabilized and tunable temperature around 77 Kelvin. Increasingly they will face the tough competition from recently developed Quantum Cascade lasers, which span the wavelength region from just over 4 ~m to well beyond 10/2m. They combine excellent single mode characteristics and ease of operation with high output powers. So far, they still need to be cooled to liquid nitrogen temperatures in order to work in the continuous wave (cw) regime, but the developments are taking place at a high pace. Interestingly, Tittel and co-workers have shown that even in pulsed mode a QCL-based direct absorption spectrometer may be able to achieve the high precision levels usually associated with cw operation (Weidmann et al., 2003; Roller et al., 2003). Moreover, QC lasers are already commercially available (e.g., from Alpes Lasers, Neuchatel, Switzerland), and can in principle be fabricated for any desired wavelength between roughly 5/~m and 12 ~m. There are certainly other ways to generate radiation in the near infrared, such as lasers based on optical parametric oscillation and difference frequency or microwave sideband generation, but these are generally too complex for application outside a spectroscopy laboratory. Moreover, the cost of such laser systems is incompatible with the ultimate goal of developing a viable alternative to mass spectrometry. For a good discussion of the various diode laser types mentioned above, with their distinct operating principles and characteristics, the reader is referred to Werle (1999). Tittel et al. (2003) present an excellent overview of diode and other solid-state laser sources for trace gas detection in the mid-infrared. In selecting a particular (diode) laser for an isotope ratio spectrometry instrument, it is important to also consider the performance of the available detectors in the appropriate wavelength range. Detector performance can be conveniently characterized by the detectivity (D*), which is a measure of the signal-to-noise ratio under standard conditions of 1 Watt incident radiation, a 1 cm2 active area, and a noise bandwidth of 1Hz. At the longer wavelengths, above 5/~m, the detector of choice is a HgCdTe device operating at liquid nitrogen temperature. Its detectivity is of the order of 1010 cm.Hzl/2W-1. In the 3 ~m to 5 ~m range, liquid nitrogen cooled InSb would be a good choice (with D*-values reaching 1011 cm.Hzl/2W-1), followed by HgCdTe and PbSe operated at 200-250 K; the higher temperatures being easily obtained by thermoelectric (TE) cooling. From 1.5/~m to 3 ~m, liquid nitrogen cooled InAs and TE-cooled PbS and InAs exhibit detectivities up to 1012 cm.Hzl/2W-1. Finally, between I ~m and 1.7 ~m InGaAs and Ge show excellent performance levels when cooled, with detectivities up to 1014 cm.Hzl/2W -1, but both may also be operated at room temperature. Thus, in choosing the wavelength range for detection, there is a clear trade-off between laser and detector performance, as well as absorption strengths. Ultimately,
784
C h a p t e r 34 - E. K e r s t e l
the choice of light source, detector, and measurement technique, will have to be weighted against other important issues such as sample condition (mixing ratio, quantity, level of isotope enrichment), measurement time, desired precision and accuracy; and cost and complexity of the instrumentation. It may be clear that the sample isotope ratio measurement should be carried out with respect to a reference material. This is best done in a dual (sample and reference) channel set-up, in which instrumental drifts may be compensated for. Even in the case where the intensities of the dominant and rare isotopomer features have the same temperature coefficient, a sequential measurement (h la Uehara et al., 2001) always risks being compromised by a memory effect: the current measurement is affected by gas left behind from the previous measurement. In the sequential case, the effect is directly comparable to that of cross-contamination in an IRMS instrument, and would result in a non-linear (quadratic) scale contraction (Meiier et al., 2000). Finally, we would like to quote one of the recommendations made by the IAEA Advisory Group meeting on New approaches for stable isotope ratio measurements (IAEA, 2001): During the past few years a variety of new techniques have appeared. Here we discuss in particular clinical applications of isotope measurements and laser spectrometry techniques. The requirements for clinical use are similar to those required for field deployment - easy of use, self diagnostic, robust, and moderate cost.
...] For such new techniques to be accepted by the isotope community they should be adequately compared with existing state-of-the art equipment and checked against isotope reference materials that are back-traceable to (IAEA) primary standards. Publications should distinguish between precision, accuracy; as well as instrument stability. In addition they should clearly state the procedure followed for calibration (defining the scale zero) and normalization (scale multiplication factor). In practice this means, e.g. for the case of water, that VSMOW, SLAP, GISP and a number of local standards should be measured. Unfortunately, the terms precision (a measure of the reproducibility of the result) and accuracy (a measure of the deviation from the true value) are sometimes used too loosely in the literature. Moreover, since measurements should be made with respect to a local reference material, instead of the international standard material (which is simple too expensive and rare to be used on a daily basis), the total variance can be decomposed into at least two components: the variance of the local reference material with respect to the international standard, and that of the sample with respect to the local reference (Jasper, 2001). Both should be reported. Conformation with the procedures outlined above is essential if optical isotope ratio instrumentation is to move from the physics laboratory to the isotope ratio work floor.
Isotope Ratio Infrared Spectrometry
785
Acknowledgement I am indebted to past and present colleagues for joint research, of which the results are referenced in this chapter, and in particular to Harro Meijer, who was instrumental in starting the optical line of research at the Groningen Center for Isotope Research. Substantial financial support from the Dutch Foundation for Fundamental Research on Matter (FOM), the University of Groningen, and a 5-year fellowship of the Royal Netherlands Academy of Arts and Sciences (KNAW) are greatly appreciated. Finally, I wish to thank Kevin K. Lehmann and an anonymous reviewer for valuable comments on the manuscript.
786
Chapter 34- E. Kerstel
Appendix: Molecular and Atomic Isotope Ratios In order to investigate the relation between molecular and atomic isotope ratios and 6-values, we first define the atomic ratio, taking the deuterium ratio as an example:
2RA=[D]
[H]
[34.A1]
Similarly, we define the molecular ratio for the case of the deuterated water molecule" [34.A2]
2RM= [H160D] 2[H160H]
In the above, the factor of two accounts for the two equivalent positions of the hydrogen atom in the water molecule. Ignoring isotopomers containing more than one rare isotope in addition to the one being considered, we can write:
2RA -
[D] _ ([H160D] + [H180D] + [H170D] + 2[D160D] + ...) [H] (2[H160H ] + 2[H180H ] + 2[H170H] + [H160D ] + ...)
-1 1 + [H180D] + [H170D ] + [D160D ] 9( [H16OD]I [H160D ]
[He60D] [HI60D] [He60H ] ~2i7-]-I%H])
2[H160H]
1 + [H180H]
+ [H170H] + [H16OD] +... [H16OH] [H160H] 2[H160H]
)
) [34.A3]
The first term on the right-hand side of the final expression in equation [34.A3] is the molecular abundance ratio, such as determined in a spectroscopic experiment. The second is a first-order correction term. We can therefore make the assumption, correct to first order, that the molecular abundance ratios can be replaced by atomic abundance ratios in the following manner:
2RA _ [H160D] . (1 +
2[H160H]
18
R+
17
R+(
2R 2 (2
) 9 R)
-1
(1 + 18R + 17R + 2R + ... )
60
+ ...)= [H 1
D]
_
aRM
[34.A4]
2[H160H]
It is thus implicitely assumed that the binding of a particular atom in the water molecule does not depend on the isotopic composition of the remainder of the molecule. The 6-value is defined as the relative deviation of the abundance ratio of a sample from that of a reference material. The definition is similar for the atomic and molecular cases:
787
Isotope Ratio Infrared Spectrometry
2r
2r
2 A =__ 2 A R t" 2 M =__ 2aM
Rs 1
Rs
-
([D]/[H])sample
-
1
[34.A5a]
([D]/[H])reference 1
-
([H160D]/2[H160H])sample
-
1
[34.A5b]
( [ H160 D ]/ 2[ H160H] )reference
When rewriting the result of equation [34.A4] for the case of the h-values of equation [34.A5], the same approximations are made in the nominator as in the denominator. The final result is that to an excellent approximation the atomic and molecular 6-values are equal: 26A - 2 ~ M . The reasoning to show that 176A = 1 7 ~ M and 186A = 186M is completely analagous. Also, the conclusion is completely general and not restricted to the case of the water molecule.
Handbook of Stable Isotope Analytical Techniques, Volume 1 P.A. de Groot (Editor) 9 2004 Elsevier B.V. All fights reserved.
CHAPTER 35 Glow Discharge Mass Spectrometry: Fundamentals and Potential Applications in Stable Isotope Geochemistry David M. Wayne NMT-15, Pit Disassembly and Nuclear Fuels Technologies, MS E 530, Los Alamos National Laboratory, Los Alamos, NM 87545, USA e-mail: [email protected]
Abstract The application of glow discharge mass spectrometry (GDMS) in the Geosciences has been limited largely to the elemental analysis of soils, meteorites, and ceramics. Commercially built GDMS instrumentation includes single-collector, double-focusing mass spectrometers (e.g. VG 9000) designed for rapid determinations of trace and major elements at high mass-resolving powers (M/DM - 4000 to 10,000). The analytical capabilities of the technique are broadly similar to those of laser ablation ICPMS, although the spatial resolution of sampling via GDMS is limited to the millimeter scale, at present. However, its direct solid sampling capability and very low matrix dependence makes GDMS a convenient technique for the rapid analysis of highly refractory materials, or for samples where compositional depth profiling is needed. Very little work has been done to optimize GDMS for isotope ratio determinations, however. A summary of recent literature on isotope ratio determinations by GDMS indicates that the data have similar, or slightly lower, precision to those obtained using single-collector double focusing ICPMS. Mass bias due to space charge effects across the differentially pumped sampler interface is significantly lower for GDMS than for ICPMS, and can be modeled using fractionation equations based on either power, or exponential laws. Previous investigations are limited to isotope ratio determinations in metals and other conducting matrices by direct current (dc) GDMS. Direct isotope ratio determinations are also possible on non-conducting materials through the use of tantalum secondary cathodes. Direct measurements of trace thallium, lead and uranium isotopes in a well-characterized glass standard are similar in quality to isotope ratio data obtained from conducting matrices.
35.1 Introduction The application of ion and electron beams can provide quantitative elemental and isotopic data directly from the sample surface. The development of the 'SHRIMP' double-focusing, secondary ion mass spectrometer (SIMS)(Compston, 1996; DeLaeter & Kennedy, 1998) has proven to be a highly significant achievement. The subsequent emergence of double-focusing magnetic sector inductively-coupled mass spectrome-
Glow Discharge Mass Spectrometry" Fundamentals and Potential Applications ...
789
ters (ICPMS) has permitted the acquisition of accurate and precise isotope ratio data on most elements in the periodic table (Walder & Freedman, 1992; Walder et al., 1993a; Halliday et al., 1995; Lee & Halliday, 1995). In this paper, I briefly describe the theory, application and practical aspects of a direct-sampling ion source similar to SIMS, but not well known to the geological community: the glow discharge. Glow discharge (GD) ion sources sputter the sample's surface, and thereby permit direct atomization, ionization and analysis. The typical operating pressure of the GD ion source (~1 to ~10 Torr) is intermediate between the ICP (ca. atmospheric pressure) and thermal ionization (TI) sources, and most GD analytical instrumentation is adapted from either ICP- or TI-based instruments. Glow discharge ion sources can be powered by direct current (dc), pulsed (Hang et al., 1996; Steiner et al., 1997; Harrison, 1998), or radio frequency (rf) potentials, and are used as line sources for spectroscopy and as ion sources for mass spectrometry. Unlike other direct sampling techniques, the GD exhibits relatively uniform elemental response, and can rapidly produce quantitative trace and ultra-trace (low ppb) analyses of solid materials without the need for strictly matrix-matched standards. High resolution (e.g., magnetic sector, time-of-flight, etc.) glow discharge mass spectrometry is also capable of measuring the isotope ratios (and elemental concentrations) of many elements (e.g., Li, B, S, Sr, Nd, Pb) in a single sample. The chief limitation of GD analysis for most geoscientists is that samples must conduct electricity. For meteorites, native elements (including graphite) and many sulfide and oxide minerals, GDMS can be applied directly. For soils, silicate rocks and other insulating materials, bulk analyses are typically performed by mixing the powdered sample with a conducting metal binder. However, recent developments in GD ion source technology include a radio-frequency-powered source (Donohue & Harrison, 1975; Winchester et al., 1991; Marcus, 1993) which can directly sputter insulating materials, and secondary cathodes (Milton & Hutton, 1993; Schelles et al., 1995; Betti et al., 1996a; Schelles & Van Greiken, 1996; Schelles et al., 1996; Wayne et al., 1999) similar to those employed in SIMS analysis. Each of these approaches provides convenient means for the direct quantitative analysis of insulating materials. Another limitation of the GD ion source is its lack of lateral resolution during sampling. Cathodic sputtering, the manner in which samples are atomized from the solid state in the GD ion source, does not lend itself to aiming, focusing, or collimating as can be done for a laser or ion beam. Although the sputtered area can be physically restricted to a small (> 2 ~m2) portion of the sample, as in the 'Grimm-type' (or 'obstructed') GD ion source (Grimm, 1968), the attainment of mm2-scale sampling capabilities has not yet been demonstrated. Commercial GD instrumentation is not supplied with real-time imaging or sample manipulation capabilities, however such features could be rapidly developed if there was sufficient interest. Some glow discharge analytical applications overlap better developed, and widely accepted geoanalytical techniques such as XRMF, SIMS, and laser ablation ICPMS. Consequently, the utility of GDMS in geochemistry, and particularly in the realm of
790
Chapter 35 - D.M. Wayne
isotope ratio analysis, has not yet been fully explored. Most GDMS instruments are single collector, double-focusingmagnetic sector instruments, such as the VG9000 GDMS (VG Elemental, Winsford, Cheshire, UK). These are designed to provide quantitative elemental analyses by scanning the entire periodic table at high mass resolutions (M/AM = 4:000 to 10,000). The focus of this paper will be on the characteristics of the GD ion source which are relevant to geochemical applications in general, and isotope ratio measurements in particular. 35.2 G l o w discharge ion sources
The history and development of the glow discharge (GD) source is comprehensively reviewed in several other publications (Coburn & Harrison, 1981; Harrison et al., 1986; Harrison & Bentz, 1988; Harrison, 1988). Until the early 1930s, glow discharge (or gas discharge) devices were widely used as ion sources for early mass spectrometric studies (Aston, 1942) but were replaced by spark source sources in the mid1930s (Dempster, 1935). Until the late 1960s and early 1970s, glow discharges were rarely used for analytical purposes. Research activity in the late 1960s and early 1970s (Grimm, 1968; Coburn & Kay) 1971; Harrison & Magee, 1974) culminated in the development of commercial analytical instrumentation based on the GD ion source by several manufacturers in the mid-1980s. Currently, only VG Elemental (Winsford, Cheshire, UK) and LECO (St. Joseph, MI, USA) continue to manufacture GD-based analytical instrumentation. However, GD is widely utilized in commercial and industrial laboratories, and fundamental GD processes and applications are studied in university and government laboratories worldwide. The GD source (Figure 35.1) consists of an electrode couple maintained in a lowpressure (ca. 0.1 to 10 Torr) noble gas (high-purity Ar or Ne) atmosphere. The appeal of noble gases, particularly Ar, in analytical GD applications is the ease with which they attain a long-lived highly energetic, or 'metastable' state. Metastable Ar atoms possess energy levels that exceed the excitation and ionization energies for most elements in the periodic table, with the exception of other noble gases, such as neon (Fang & Marcus, 1993). The sample to be analyzed forms the cathode, while the anode material (usually tantalum, steel, copper, etc.) is of little consequence. When an electric field (about 0.3 to 0.5 kV) is applied to the electrode, the working gas breaks down and electrons, ions and other species are formed in the cathode - anode gap (Figure 35.2). Bombardment of the cathode (sample) surface by incident particles (dominantly working gas ions) erodes atoms from the exposed sample surface. Most of the sputtered atoms remain neutral and redeposit on to exposed surfaces within the ion source, but a small percentage diffuse into a collision-rich zone (the negative glow region) in the cathode - anode gap. The ions in the negative glow can then be sampled via physical transfer into a mass spectrometer.
35.2.1 Glow discharge plasma structure The glow discharge (i.e., the region between the cathode and the anode) is comprised of eight discrete zones, each with different electrical properties, radiation intensities and particle population (Chapman, 1980). The size and importance of each region is dependent on the carrier gas species, pressure, GD potential, sample current,
Glow Discharge Mass Spectrometry: Fundamentals and Potential Applications ...
791
Figure 35.1 - Schematic diagram of the Grimm-type glow discharge ion source used in the Kratos sector GDMS. 1) tantalum ion exit plate, 2) tantalum anode, 3) argon inlet, 4) heating coils, 5) sample (cathode), 6) sample holder, 7) thermocouple, 8) quartz insulator, 9) quartz anode insulator, 10) copper spring, 11) stainless steel cap, 12) boron nitride insulator, 13) Macor insertion probe insulator, 14) stainless steel cell.
GD cell configuration, and cathode-anode gap (Chapman, 1980; Coburn & Harrison, 1981; Harrison, 1988; King & Harrison, 1993; Fang & Marcus, 1993; Bogaerts & Gijbels, 1998). The most important regions are the cathode dark space (also termed the cathode fall), the anode dark space, and the negative glow (Figure 35.2). The cathode dark space (CDS) is a narrow, non-luminescent zone located immediately above the cathode surface. The entire electrical potential difference between the anode and the cathode occurs within this region. Positive working gas ions are accelerated across the CDS and collide with the cathode surface. These collisions cause the continuous release of neutrals and electrons from the cathode surface. The electrons released from the cathode accelerate across the CDS into the negative glow region where they can participate in ionization reactions, and sustain the GD plasma. The
792
Chapter 35 - D.M. Wayne
Figure 35.2- Schematic depiction of plasma structure (cathode dark space, negative glow and anode dark space) and different ionization mechanisms in a typical analytical glow discharge plasma. Metal and Argon atoms: Mo, Aro; Metal and Argon ions: M +, Ar+; Argon metastables: Ar*; electrons: e-. Diagram shows several possible interactions: 1) collision of Ar + with cathode releasing Mo + e- (metal atom or ion may redeposit on cathode), 2) ionization of Aro and Mo via electron impact, 3) ionization of Mo via interaction with Ar* (Penning ionization).
negative glow (NG) is a broad, light-emitting region adjacent to the CDS. The NG is essentially field-free, charge-neutral and equipotential, as electrons are slowed by numerous collisions with carrier gas and sample atoms. The majority of ionization processes occur in this collision-rich zone, but only ~0.001% of the original sputtered atom population are ionized. The anode dark space (ADS) lies between the NG and the anode (Bogaerts & Gijbels, 1998), and is analogous to the CDS except that electrons are accelerated towards the anode, and cations are repelled.
35.2.2 Cathodic sputtering Sputtering occurs when high energy (e.g. ~100 to 200 eV for a 0.5 kV cathode fall) Ar ions collide with, and penetrate several ~ngstroms into, the cathode surface (Chapman, 1980; Harrison, 1988; King & Harrison, 1993; Fang & Marcus, 1993; Betz & Wein, 1994; Bogaerts & Gijbels, 1998). The kinetic energy of the incident ions is transferred to the cathode material via collision cascades, and if the energy imparted to the lattice atoms exceeds the lattice binding energy, a variety of particles may be ejected from the cathode. The vast majority of these particles are neutral atoms (Fang & Marcus, 1993) which may redeposit on the cathode and other surfaces, or diffuse into the negative glow. Most importantly, cathodic sputtering creates an atom population that
Glow DischargeMass Spectrometry:Fundamentalsand PotentialApplications ...
793
constitutes a representative sampling of the cathode itself. The number of sample atoms produced by a single incident ion is known as the sputter yield, a property dependent on the kinetic energy, mass, angle of incidence, and species of the projectile ion, and the mass, lattice binding energy, surface properties and temperature of the target material (King & Harrison, 1993; Fang & Marcus, 1993; Betz & Wein, 1994; Bogaerts & Gijbels, 1998). Sputter rate, or the number of sample atoms ejected per unit time, is a function of sputter yield and of the operating current of the glow discharge (ca. 0.5 to 3 mA) which is, itself, dependent on carrier gas pressure and GD potential. Factors that alter the sputtering behavior of a material include the presence of oxide layers or other surface contamination, impurities in the carrier gas, phase inhomogeneities at the surface, and heating of the cathode. Compared to volatilization yields, which may vary over several orders of magnitude, sputter yields vary by less than an order of magnitude across the periodic table for a wide range of incident ion energies. Thus, GD processes have little matrix dependence, relative to thermal ionization, and ionization by laser and ion beam bombardment (Betz, 1980; Fang & Marcus, 1993; Bogaerts & Gijbels, 1998). For alloys, mixtures, and compounds, complications may arise due to 'differential sputtering', where elements having higher sputter yields (e.g., Cu, Ag, Au) are sputtered from the surface at a quicker rate than those with lower sputter yields (e.g., C, Si, Nb, Ta). However, a steady state is quickly reached within the glow discharge plasma, and the composition of the sputtered atoms then becomes a predictable function of the surface composition(Fang & Marcus, 1993). 35.2.3 Ionization processes
Ionization in the glow discharge source is spatially and temporally removed from the site of atomization. This separation further contributes to the relative lack of matrix effects observed in analytical glow discharges (Chapman, 1980; Coburn & Harrison, 1981; Harrison et al., 1986; Harrison & Bentz, 1988; Harrison, 1988; King & Harrison, 1990, 1993; Fang & Marcus, 1993; Bogaerts & Gijbels, 1998, 2000; Bogaerts et al., 2003). Working gas atoms, and neutrals derived from the cathode, may be ionized through a variety of collisional interactions in the negative glow region (Chapman, 1980; Harrison, 1988; King & Harrison, 1993; Fang & Marcus, 1993; Bogaerts & Gijbels, 1998). The dominant processes by which carrier gas atoms become ionized in the GD plasma are electron impact, fast Ar ion (or Ar atom) impact, and via interactions with metastable (excited) Ar atoms ('Penning' ionization). Advanced modelling and mathematical simulation of processes within the glow discharge by Bogaerts & Gijbels (1998, 2000) and Bogaerts, et al. (2003) indicate that, in addition to electron impact and fast Ar ion impact, Penning ionization and asymmetric charge transfer are also significant ionization mechanisms for sputtered neutrals.
794
Chapter 35 - D.M. W a y n e
35.3 Glow discharge analysis of insulating materials 35.3.1 Glow discharge spectra As in ICPMS, there are numerous interfering peaks in a typical GDMS spectrum. Most of these are multiply charged Ar and analyte species (absent from most ICPMS spectra), carrier gas clusters (e.g., [ArArl+), and metal - gas clusters (e.g., [FeAr]+ is abundant when sputtering steel samples). For silicate- or oxide-based based ceramics and geological materials, metal oxide clusters are also present. Unlike ICPMS, solvent- and atmospheric-derived peaks are very minor (Feng & Horlick, 1994). However, small amounts of water vapor, carbon, oxygen and various hydrocarbons may be present as contaminants in the GDMS source, and on the sample surface. Such contamination can be minimized prior to analysis by baking the source and sample in vacuo (Wayne et al., 1996; Wayne, 1997), by sputter cleaning the sample, and by cryocooling the GD source (Hall et al., 1989; Ohorodnik & Harrison, 1993). Analysis of insulating materials by dc-GDMS is typically accomplished by combining a powdered sample with a high-purity, electrically conducting metal powder, or binder, such as Ta, Ag, Cu or A1 (Mei & Harrison, 1991; Winchester et al., 1993; Teng et al., 1995; Wayne, 1997). This approach is used routinely for bulk and trace analysis of non-conducting solids, and permits the addition of an internal standard. However, the volume ratio of sample to binder is restricted by the amount of metal binder needed to make good electrical contact, and by the oxygen content of the nonconductor (Teng et al., 1995). Complete homogenization of the analyte in the binder may also be difficult. Compacted powders also have a high net surface area, thus surfaceadsorbed species and occluded atmospheric gases are not routinely sputtered or baked away prior to analysis (Wayne, 1997). There are two alternative methods for the analysis of non-conductors by GDMS. Direct sputtering of non-conducting solids has been demonstrated in radio-frequency (rf) glow discharge sources (Marcus, 1993; Marcus et al., 1994; Becker et al., 1996). Currently, rf sputtering is the preferred method for the analysis of non-conductors via GD atomic emission spectroscopy (GDAES). High-resolution rf-GDMS instrumentation is not available commercially, and the if-powered source must be assembled and interfaced in-house. The use of secondary cathodes (Milton & Hutton, 1993; Schelles et al., 1995; Betti et al.,1996a; Schelles & Van Greiken, 1996; Schelles et a1.,1996; Wayne et al., 1999), while not necessarily superior to the rf-GDMS technique, provides an inexpensive and convenient alternative. In this method, metal atoms sputtered from the secondary cathode redeposit and form an electrically conductive thin film on the exposed surface of the non-conducting sample. Sputtered metal atoms continuously redeposit and replenish the surface film, permitting direct depth profiling and quantitative analysis of the underlying non-conductor.
35.3.2 Quantitative analysis by GDMS One of the primary benefits of GD plasmas for analytical chemistry is the relative lack of matrix effects due to the temporal and spatial separation of atomization and ionization processes. However, GD plasmas are not entirely free from matrix effects, therefore simple calibration procedures are required for quantitative analysis. Relative
Glow DischargeMass Spectrometry:Fundamentalsand Potential Applications ...
795
sensitivity factors (RSFs) are used to relate the peak intensity of an element of unknown concentration to the peak intensity of a reference element of known concentration: RSFx/R = (Ix/Cx) / (IR / CR)
[35.1]
where Ix and Cx are the peak intensity and concentration, respectively, of the unknown element, and IR and CR are the peak intensity and concentration, respectively, of the reference element (Vieth & Huneke, 1990; King & Harrison, 1990, 1993). The RSF values for a series of elements in a given sample are functions of glow discharge processes, and may vary from one sample type to another. Part of this variation is due to differences in the sputtering characteristics and major element chemistry between samples (Vieth & Huneke, 1990; King & Harrison, 1990, 1993). Other studies have shown that changing GD parameters (gas pressure, GD potential, gas composition) can alter RSF values (Smithwick et al., 1993; Saito, 1995). Differences in ionization properties of the sputtered atoms are perhaps even more significant (King & Harrison, 1993; Bogaerts & Gijbels, 1996), and are more difficult to control. Although strict matrix matching is not required for the acquisition of quantitative analytical data by GDMS, calibrations should be performed using standard materials that are chemically similar to the unknown. For example, RSFs determined from steel (or other similar metal) standards are best suited to the quantitation of unknown Fe-CrNi alloys, and a soil sample is best calibrated using a soil, clay, or powdered rock standard (King & Harrison, 1990). Despite the complications summarized above, the relative sensitivities for most elements in typical GD samples vary by less than an order of magnitude, and yield linear calibration curves. Magnetic sector GDMS is a high-sensitivity analytical technique, with a linear dynamic range comparable to that of magnetic sector ICPMS. Numerous studies have shown that magnetic sector GDMS is routinely capable of quantitative analysis down to the nanogram per gram (part per billion) range (e.g., Becker & Dietze, 2003).
35.4 Isotope ratios by GDMS 35.4.1 Previous work
Isotope ratio measurements by GDMS have been the topic of only a handful of publications (Table 35.1) that describe data collected using the VG9000 double-focusing, magnetic sector instrument (Donohue & Petek, 1991; Shimamura et al., 1993; Duckworth et al., 1993; Barshick et al., 1994; Riciputi et al., 1995; Betti et al., 1996b; Itoh & Hasegawa, 1998). Others have reported isotope ratio measurements made using GDMS instrumentation adapted in-house from a conventional magnetic sector TIMS (Ecker & Pritzkow, 1994), and from a double-focusing spark source mass spectrograph (Chartier & Tabarant, 1997). All of these investigations have focused on metallic samples (solid and pressed powders), solutions dried onto metal powders, and powdered insulating materials dispersed in metal powders. In a novel study, Pajo, et al. (2001) applied GDMS to examine oxygen isotope ratios in pressed pellets formed from powdered uranium dioxide. Their results, however, proved to be too imprecise (0.5% to 4% RSD) to reliably detect variations in oxygen isotopes in the samples.
796
Chapter 35 - D.M. Wayne
35.4.2 M a s s bias in the GD source
Mass bias in mass spectrometry arises when physical phenomena during ionization, sampling or analysis result in the preferential loss of either light or heavy ions from an ion population. Thus, the measured isotope ratio is different from the actual value. As the filament is heated during TIMS analysis, thermal effects favor the evaporation of lighter isotopes, relative to heavier isotopes, causing a bias of ~0.1% per atomic mass unit (amu). However, if the sample is completely consumed, as in total evaporation TIMS (Callis & Abernathey, 1991), mass bias is reduced to negligible level. Isotope ratio measurements made using plasma-source mass spectrometry (e.g., GDMS, ICPMS) are inherently less accurate than TIMS due to greater (>1.0% per amu) mass bias. Mass bias in plasma-source magnetic sector MS appears to be a predictable function of element mass, and is adequately described by a power law function (Russ & Bazan, 1987; Walder & Freedman, 1992): Rtrue = Rmeas (1 + F)n
[35.2]
where Rtrue and Rmeas correspond to the true and measured isotope ratios, (m2/ml) where m~ is the lighter isotope, F is the fractionation factor, and n is the difference between the 2 masses, ml and m2. An exponential fractionation law (Taylor et al., 1995) produces similar results (e.g. Lee & Halliday, 1995). For samples having similar chemistries, the isotopic composition of one element can be normalized to the isotopic composition of an element of similar mass (Longerich et al., 1987; Ketterer et al., 1991; Walder et al., 1993) that can be added directly to the sample solution. For direct solids analysis by GDMS, such mass corrections may be complicated somewhat if the necessary calibrant element (e.g. T1 for Pb) is not present in sufficient quantities to provide a precise determination of its isotope ratio. A calibration on a separate sample may be necessary in some instances, although most isotope systems of geochemical interest have at least one stable pair (e.g. 86Sr/88Sr) which can be used as an internal standard.
For plasma-sourced MS, mass bias is primarily the result of space charge effects that arise following ionization and sampling of the quasi-neutral plasma (Gillson et al., 1988; Li et al., 1995; Niu & Houk, 1996; Chen & Houk, 1996; Heumann et al., 1998; Hang et al., 1999). As positive ions are extracted from the ion source through the skimmer cone, charge separation occurs and lighter ions are repelled farther from the beam axis than heavier ion. The magnitude of the space charge effect in ICPMS is related to the sample matrix, the type of instrumentation (Heumann et al., 1998), and to the total ion current (typically ~ 6 ~A) passing through the skimmer (Gillson et al., 1988; Li et al., 1995; Niu & Houk, 1996; Chen& Houk, 1996; Hang et al., 1999). Experimental studies and theoretical calculations relevant to ionization and sampling processes in GD sources (Hang et al., 1999) indicate that space charge effects may be much less severe in GDMS, due to its lower total ion current (~ 0.02 gA). These calculations further suggest that post-skimmer beam divergence across the GD interface is a function of the square root of the ion mass (Hang et al., 1999). Thus, if space charge is the dominant cause of mass bias in GDMS, theory predicts that isotope ratios so obtained will favor the lighter isotope. However, the improvements in iso-
hlatrix
Type
Pdmetal rl:ld Fe met ecl r i t e -10% soil in Ag pcllder
sollJ.tii:iii
iii
Ag
pc1vder
Re:100% ApJ.. 5 yo
&metal Sr:SrC0, in Ag, Cupowder Re:m e t a1 Ag A$JO, on C u p ~ ~ d ~ r Eb:pwemetal, in Cumetal
B:boric acid in Ag p c d e r sb:in Cumetal 5:in Almetal Umetal
B: i i i r;teel 0s: n,a, U:fiuel rod B: in Zrmetal t a m e t a1
Glow Discharge Mass Spectrometry: Fundamentals and Potential Applications ...
Table 35.1 - Summary of published isotope ratio data acquired using sector GDMS.
797
798
Chapter 35 - D.M. Wayne
tope ratio accuracy relative to that of the ICP source will be difficult to assess using existing commercial GDMS instrumentation. Commercial GDMS instruments are adapted from existing platforms designed specifically for use with ICPMS, TIMS, or a variety of ionization and atomization techniques used to identify organic molecules. Subsequent evaluation of the isotope ratio capabilities of GDMS would best be carried out using an instrument fitted with an interface optimized for sampling ions produced by GD, rather than by ICP or TI (Hang et al., 1999). Isotope ratio studies to date (Table 35.1) indicate that scanning (i.e. single collector) magnetic sector GDMS provides equal, or slightly inferior, precision relative to that obtained using scanning sector ICPMS (Heumann et al., 1998). The cause of this disparity is not yet known, as the ion signal provided by the GDMS source is as stable as the signal supplied by conventional ICP sources (~2% over 10 min.; Finnigan MAT prom. lit., ca. 1995). Only two studies (Riciputi et al., 1995; Chartier & Tabarant, 1997) have examined mass fractionation trends in GDMS isotope ratio measurements. Riciputi et al. (1995) noted no correlation between GD potential or sample current and the severity of mass fractionation effects, but did observe a systematic increase in 65Cu/63Cu from 0.451 to 0.4:55 with increasing Ar pressure in the ion source region. By contrast, Chartier & Tabarant (1997) found no isotopic effects with changes in ion source pressure. The latter study (Chartier & Tabarant, 1997) was conducted using flat samples in a Grimm-type discharge cell, while Riciputi et al. (1995) used pin-shaped samples in a conventional GD cell. Significantly, Riciputi et al. (1995) speculated that the observed mass fractionation effects could be due to pressure-induced changes in the geometry of the negative glow relative to the ion extraction slits. Although considerable work remains to be done to thoroughly investigate mass bias in GD ion sources, the results of Chartier & Tabarant (1997) suggest that the effects of changing Ar pressure on the shape and position of the negative glow region may not be as severe for flat samples. 35.5 Direct measurement of isotope ratios by GDMS 35.5.1 Experimental
Isotope ratio measurements were performed on two NIST standard reference materials (SRMs) using a Kratos 'Concept' (Kratos Analytical, Manchester, UK) magnetic sector, double focusing, scanning GDMS (Smithwick et al., 1993; Wayne et al., 1996; Wayne, 1997) equipped with Mach 3 software. Standards used in this study include SRM 1264a (modified high-carbon steel), and SRM 611 (trace elements in glass, wafer form). The steel and glass standards were cut to I cm diameter discs, cleaned, affixed to a direct insertion probe using a disposable steel clip, and introduced into the GD source (Figure 35.1). Samples were baked in vacuo at ~200~ for ~30 minutes, cooled, and then pre-sputtered for N20 minutes. As a result, the water content of the ion source and sample surface is minimized, thus reducing the possibility of isobaric interference from metal hydrides. Samples are routinely examined for potential metal hydride (or other) interference at high resolution prior to the collection of isotope data at lower resolution. For all analyses, the distance of the sample surface from the anode was set initially at I mm, and all samples were run using a GD potential of 1.5 to 2.0 kV, with a sample current of 1.0 to 1.5 mA, at an accelerating voltage of 8 kV. Mass resolution, M/
799
Glow Discharge Mass Spectrometry: Fundamentals and Potential Applications ...
AM, (peak width defined at 10% of peak height) was maintained at 1500, in order to obtain a flat-topped peak. All samples and standards were run in 'single ion monitoring' (SIM) mode (Wayne et al., 1996) by scanning either the magnetic field ('magnet scan') or the accelerating voltage ('ESA scan'). Further description of operating conditions for the Kratos GDMS can be found elsewhere (Wayne et al., 1996; Wayne, 1997). 35.5.2 Results NIST SRM 1264a: Isotope ratio analyses of W (0.102%) and Pb (0.024%) in NIST
SRM 1264a are summarized in Table 35.2. All errors and per cent RSDs are based on 2o standard deviations from the mean. Each analysis was performed by continuously sputtering the central region of the sample disc, resulting in the formation of a sputter crater 3 to 4 mm in diameter and several 10s of ~tm deep. Isotopic data was collected in sets of 8 to 15 scans over the mass range for each element, and repeated 8 to 10 Table 35-2 - Thallium, lead and uranium isotope ratios in NIST SRM 611 'Trace Elements in Glass'.
Run Name
205T1 203T1
F (%)
207Pb 206Pb
208Pb 206Pb
235U 238U
03/29/99 Isotop 13 %RSD Isotop 14 %RSD Isotop 16 %RSD Isotop 18 %RSD
2.42 (5) 2.1 2.42 (13) 5.5 2.38 (7) 2.9 2.40 (11) 4.7
-0.7
Trace 51 %RSD Trace 52 %RSD Trace 53 %RSD Trace 54 %RSD Trace 55 %RSD
2.42 (15) 6.1 2.34 (12) 5.1 2.41 (9) 3.6 2.41 (3) 1.4 2.36 (5) 2.1
-0.6
-0.6 0.1 -0.2
0.903 (9) 1.0 0.905 (16) 1.8 0.906 (35) 3.9 0.910 (13) 1.4
2.177 (19) 0.9 2.172 (43) 2.0 2.175 (43) 2.0 2.176 (48) 2.2
0.00220 (21) 9.3 0.00230 (5) 2.3
0.906 (23) 2.6 0.914 (9) 1.0 0.911 (18) 2.0 0.903 (12) 1.4 0.916 (18) 2.0
2.173 (55) 2.5 2.241 (31) 1.4 2.189 (29) 1.4 2.174 (34) 1.6 2.231 (25) 1.1
0.00247 (16) 6.3 0.00236 (14) 5.9 0.00227 (8) 3.7 0.00237 (19) 8.0 0.00236 (19) 7.9
0.9096 (8) 0.9095 (10) 0.9095 (30)
2.1670 (18) 2.170 (2) 2.160 (6)
0.002382
06/30/99 0.9 -0.4 -0.4 0.5
Reference Values MC-LAICPMS MC-TIMS MC-SIMS
2.3871
Pb and U ratios corrected for mass bias using factors (F%) calculated using an exponential law, based on 205T1/203T1 measured in-run. Parenthetical numbers are 2-s standard deviations reported to the last significant figure. Reference value for the T1 isotope ratio is from DeBievre & Barnes (1985). MC-LAICPMS data for Pb from Walder et al. (1993). MC-TIMS and MC-SIMS data for Pb from Belshaw et al. (1994). TIMS data for U in NIST SRM 611 from Barnes et al. (1973).
800
Chapter 35 - D.M. Wayne
times per analysis. Total elapsed time required for each analysis (i.e., 8 to 10 sets of 8 to 15 scans) was 30 to 40 minutes for magnet scanning, and 4 to 6 minutes for ESA scanning. For the steel SRMs (Table 35.2), blank corrections for Hg (on 204Pb) were found to be unnecessary. Even so, 204Pb peaks were not intense enough to provide statistically useful Pb isotope ratios. Mass fractionation was estimated using the departure of the 186W/183W ratio from an accepted value (Volkening et al., 1991) and an exponential fractionation law (Lee & Halliday~ 1995; Taylor et al., 1995)" F1,2-[ln (Rtrue / Rmeas )] / [ml(ln (m2/ml))] Rtrue - Rmeas (m2/ml)~
[35.3] [35.4]
where 13- F1,2(m2), and ml is the lighter isotope. For magnet scanning, F1,2varies from -0.8% to +0.4% per amu. Mass bias is significantly greater for ESA scanning (2.4% to 6.0% per amu), and the fractionation on 186W changed considerably between runs 7 and 8 (Table 35.2) in response to changing lens settings during refocusing. Although the internal precision of W isotope ratios collected using ESA (N0.5% to 1.3% RSD) and magnet (~0.4% to 1.5% RSD) scanning was similar, the data obtained using ESA scanning was much less accurate (Table 35.2). External precision for raw W isotope ratios was 1.7%. Normalized W isotope ratios were more consistent (0.41% RSD) between runs. Lead isotope ratios were corrected for mass bias using fractionation factors derived from in-run 186W/183W. The 207Pb/206Pb ratios obtained from ESA scans agree reasonably well with those obtained via magnet scanning. However, the 208Pb/206Pb ratios obtained by ESA scanning have a consistent negative bias relative to measurements made by magnet scanning. The precision of the Pb isotope ratios acquired using ESA scanning (~1% to 2.5% RSD) is similar to that observed for magnet scanning (~1.5% to 3.6% RSD). Between-run variation for Pb isotope ratios was 1% to 2% RSD for 207Pb/206Pb, and 2% to 4% RSD for 208Pb/206Pb. NIST SRM 611" A Na-Ca-A1 silicate-based glass-ceramic standard (NIST SRM 611)
contains trace amounts of Pb (426 ppm), T1 (62 ppm), and U (461.5)(Barnes et al., 1973), and was analyzed by dc-GDMS using a tantalum (Ta) secondary cathode. Isotope ratios of Pb (208Pb/206Pb and 207Pb/206Pb) and uranium (235U/238U) in SRM 611 have been characterized using TIMS (Barnes et al., 1973; Belshaw et al., 1991) multicollector ICPMS (Walder et al., 1993c) and multi-collector SIMS (Belshaw et al., 1991). Isobaric interference from (TaNa) + clusters (m/z - 203.9378 amu) prevented the collection of 204Pb at a mass resolution of 1500. All the NIST SRM 611 data (Table 35.3) were collected using magnet scanning, and mass bias on Pb ratios was corrected using fractionation factors calculated from in-run determinations of 205T1/203T1 (Table 35.3), as described by Walder et al. (1993c). Mass bias using ESA scanning was greater by an order of magnitude (Figure 35.3), therefore only data acquired using magnet scanning (mass bias <+ 1%) is reported here. Higher GD potential (2.0 kV) and sample current (1.5 mA) were used for the glass analyses, as the T1 signal is enhanced relative to Pb and U ion signals by a factor of ~2 at these conditions. The internal precision (Figure 35.4) of the single-collector GDMS measurements (-~1.0% to ~2.5% RSD) is inferior to that attainable by multi-collector LA-ICPMS (Walder et al., 1993c) and TIMS (Belshaw
~
182W 183W ~~
~~~~
~~
~~
182w* 183W*
1.857 (7) 0.40 1.882 (28) 1.48 1.834 (16) 0.85 1.857 (11) 0.57 1.850 (8) 0.46 1.85129
~~
~
~
2.145 (16) 0.73 2.190 (18) 0.83 2.127 (16) 0.76 2.148 (15) 0.72 2.144 (26) 1.23 2.14078
1.842 1.853 1.851 1.847
1.962
2.094 (17) 0.80 2.087 (21) 1.02 2.158 (29) 1.33 2.141 (16) 0.77 2.059 (19) 0.90 2.14078
1.975 1.843 1.863 2.019
~~
186W 183W ~~
1.850
1.866 (11) 0.57 1.879 (14) 0.76 1.799 (16) 0.90 1.808 (9) 0.47 1.901 (9) 0.47 2.85129
184W* 183W*
~
~~~~~~
Magnet Run 1 %RSD Run 2 % RSD Run 3 %RSD Run 4 %RSD Run 5 %RSD Ref. Value ESA Run 6 %RSD Run 7 %RSD Run 8 %RSD Run 9 %RSD Run 10 %RSD Ref. Value
184w 183W
~~
~~
~~
2.135
~~~~~
1.996 (15) 0.75 2.031 (17) 0.83 1.972 (19) 0.94 1.993 (7) 0.36 1.990 (19) 0.97
2.140 2.145 2.143 2.138
208Pb 206Pb
207Pb 20Wb ~~
~
-0.17
-0.8 0.4 -0.25
-0.1
0.882 (15) 1.65 0.877 (23) 2.62 0.876 (27) 3.08 0.879 (21) 2.33 0.857 (13) 1.48
2.133 (40) 1.88 2.113 (76) 3.55 2.150 (69) 3.24 2.131 (63) 2.93 2.237 (61) 2.74
0.864 (18) 2.22 0.878 (13) 1.59 0.868 (17) 2.02 0.873 (11) 1.34 0.874 (12) 1.47
2.057 (24) 1.28 2.083 (20) 1.03 2.035 (28) 1.44 2.056 (25) 1.29 2.092 (46) 2.47
1.98594
2.200
1.722 (20) 1.15 1.709 (15) 0.87 1.847 (24) 1.29 1.817 (23) 1.26 1.661 (20) 1.21
2.193 2.211 2.206 2.186
5.0
5.0 2.4 3.0 6.0
G l o w D i s c h a r g e M a s s Spectrometry: Fundamentals and Potential Applications ...
Table 35.3 - Tungsten (0.102%)and lead (240 ppm) isotope ratios in NIST SRM 1264a Steel.
1.98594 ~~
~~
* - denotes data normalized to 186W/183W = 1.98594. Pb ratios corrected for mass bias using factors (F%) calculated using an exponential law, based on in-run 186W/ *83W. Parenthetical numbers are 2-sigma standard deviations reported to the last significant figure. Reference values for W isotope ratios from Volkening et al. (1991). 801
802
Chapter 35 - D.M. Wayne 0,10
-
0 mOB -
m F'tacticalatir
])/la"~ e t Scarrning
l ~ Fxactic,nation - ZSA Scan'rlir, g
9e
0 mO~ --
e e
0 m04 -
F (%) 0,02
0,00
,02
e
-
m
mm
m
mm
m
mm
,-,tic
0 "o
"~~T1/"~
m
-m
= 2.3871
-
'
0
l
2
'
l
4
'
l
6;
'
l
'
l
'
l
8 10 12 R~m N~mb er
'
I
14
'
l
16
'
l
18
Figure 35.3 - Plot of fractionation factors determined for T1 in NIST SRM 611 'Trace Elements in Glass'. Isotope ratios (205T1/203T1) were corrected for mass bias to an accepted value (2.3871) from DeBievre & Barnes (1985) using an exponential fractionation law (Taylor et al., 1995). Mass bias for isotope ratios acquired by ESA scanning is strongly positive, and is 5x to 10x higher than mass bias based on data collected via magnet scanning. ESA scan data are not shown in Table 35.3.
et al., 1991)(both ~0.09% RSD), or by multi-collector SIMS (~0.3% RSD)(Belshaw et al., 1991). External (between-run) precision was worst for T1 (2.3% RSD) and U (6.8%), but markedly improved for 207Pb/206Pb (1.0% RSD). The external precision of the 208Pb/ 206Pb data is 2.3%, but improves to 0.5% if two anomalous measurements ('Trace 52', 'Trace 55', see Table 35.3) are omitted. The accuracy of Pb isotope ratios determined by GDMS, is estimated by quantifying the departure of the GDMS measurements from the multi-collector TIMS data: B (%)= 100" [(RGD- RTI) / RTI]
[35.5]
where B is the percent bias relative to TIMS data, RGD is the ratio determined by GDMS, and RTI is the ratio determined by multi-collector TIMS. The percent difference between the GDMS and TIMS data is less than 1.0% for all Pb ratios except for two 208Pb/206Pb measurements (see above), which are ~3% greater than the TIMS values. The uranium isotope ratio (235U/238U) w a s also measured in the glass standard. The extremely low abundance of 235U (<1 ppm) resulted in low signal intensities, thus the 235U/238U ratios obtained by GDMS were less precise (2.3% to 9.3% RSD) and less
Glow Discharge Mass Spectrometry: Fundamentals and Potential Applications ...
803
Figure 35.4 - Plot of Pb isotope ratio measurements from NIST SRM 611 acquired using single collector, magnetic sector GDMS. Error bars (2o) on individual GDMS measurements indicate that the GDMS data are less precise than those acquired using multicollector techniques (Belshaw et al., 1991; Walder et al., 1993c). Relative to the multicollector data, GDMS data are accurate to within 1%, with the exception of 2 measurements. accurate (-8% to +4%) than the Pb isotope ratio measurements. Some of the loss in accuracy may be due to the lack of a near-mass isotope pair to use as a calibrant. Interestingly, the GDMS data indicated that the U used in the fabrication of the glass standard was depleted in 235U. The presence of depleted U in SRM 611 was subsequently confirmed by published data (Barnes et al., 1973). 35.6 Conclusions The glow discharge ion source is a mechanically simple, highly efficient ion source for mass spectrometry and optical spectroscopy. As an ion source for magnetic sector mass spectrometers, its performance stability, elemental coverage, and versatility is comparable to that of the ICP ion source. Glow discharge MS has been used to provide trace and major element analyses of soils (Duckworth et al., 1993; Teng et al., 1995; Betti et al., 1996a), meteorites (Shimamura et al., 1993), relevant standard reference materials (De Gendt et al., 1995), and various ceramic materials (Hall & Sanderson, 1989; Mei & Harrison, 1991; Winchester et al., 1993; Ohorodnik & Harrison, 1993; Schelles et al., 1995; Wayne et al., 1996; Wayne, 1997). Limitations to the wider application of GDMS in geochemistry are easily overcome in m a n y cases, or could be mitigated by minor changes in instrument design. For example, the application of secondary cathodes, or of a radio-frequency GD, is sufficient for the analysis of non-
804
Chapter 35 - D.M. W ay n e
conducting materials. Glow discharge MS is also capable of performing rapid depthprofile analyses, and isotope ratio measurements. However, very little work has been done to optimize GDMS for routine use in isotope ratio determinations. Isotope ratio determinations by single-collector GDMS (Donohue & Petek, 1991; Shimamura et al., 1993; Duckworth et al., 1993; Barshick et al., 1994; Ecker & Pritzkow, 1994; Riciputi et al., 1995; Betti et al., 1996b; Chartier & Tabarant, 1997; Itoh & Hasegawa, 1998) are somewhat less precise than those acquired using single-collector double focusing ICPMS (Heumann et al., 1998), although mass bias due to space charge effects is less severe in GDMS (Hang et al., 1999). The results presented here show that total mass bias is routinely less than 0.7% on isotope ratios measured in glass and steel standards. Although space charge effects should be less important at the high accelerating potentials used in magnetic sector MS, mass discrimination often exceeds 1% for isotope ratio measurements by sector ICPMS (Heumann et al., 1998). Unlike GDMS, the optimization of single collector, magnetic sector ICPMS for the determination of precise and accurate isotope ratio has been a major research focus for several years (Furuta, 1991; Kim et al., 1991; Vanhaecke et al., 1996; Becker & Dietze, 1998; Gwiazda et al., 1998; Hinners et al., 1998). If the same level of effort were to be invested in GDMS, it is highly likely that the quality of isotope ratio determinations by GDMS would increase dramatically. Until multi-collector GDMS becomes available, however, the overall quality of isotope ratio measurements by single collector GDMS is sufficient for rapid 'survey'-type determinations on elements present at concentrations >10 ppm in a wide range of materials. Promising recent developments include the development of GD ion sources fitted to compact, robust time-of-flight (TOF) mass analyzers capable of isotope ratio precision and accuracy in the 1% to 5% range (Guzowski, et al., 2000; Pisonero, et al., 2001). Barnes, et al. (2002) describe a GD-sourced Mattauch- Herzog mass spectrograph that acquired data for Ti isotopes with 1% to 3% accuracy, and precision of about 0.2% RSD.
Acknowledgements Pre-submission comments by Dr. Vahid Majidi (LANL), and further comments by Dr. S.H. Sie and Dr. Margaret Ricci are appreciated. This research was supported by the Stockpile Support and MOX Fuels Programs at Los Alamos National Laboratory. This publication is LAUR 99-5014.
Handbook of Stable Isotope Analytical Techniques, Volume 1 P.A. de Groot (Editor) 9 2004 Elsevier B.V. All fights reserved.
CHAPTER 36 The Use of Molecular Sieves in Stable Isotope Analysis1 Haraldur R. Karlsson Department of Geosciences, and Department of Chemistry and Biochemistry, Texas Tech University, Box 1053, Lubbock, TX 79409, USA e-mail: [email protected]
36.1 Introduction High-precision stable isotope analysis typically requires a good vacuum in preparation systems and a clean working gas in the mass spectrometer (H2, He, 02, N2, CO2, SiF4, SO2, SF6). Molecular sieves and zeolites play an important role in achieving these goals by trapping undesirable components such as hydrocarbons and preventing back-diffusion between roughing pumps and the main vacuum line working volume. In this review, I first discuss the properties of zeolites and molecular sieves that make this possible and then consider specific applications of these materials to stable isotope analysis. Although molecular sieves are also commonly used as catalysts (see e.g., Rabo, 1976) this aspect of molecular sieves will not be discussed. The International Union for Pure and Applied Chemistry (1999) classifies pores in porous substances based on their diameters: micropores (< 20 ~), mesopores (20 to 50 ~), and macropores (> 50 A). Molecular sieves are natural and synthetic crystalline solids that have a microporous crystal structure. The pore dimensions are fixed for a given molecular sieve under specific operating conditions so that only vapor species in a specific size range can pass through. The term molecular sieve includes zeolites (natural and synthetic) and other microporous compounds such as borosilicates, aluminophosphates, and carbon molecular sieves (CMS e.g, carbosphere). Table 36.1 depicts a classification scheme for molecular sieves. 36.1.1 Zeolites
The first known molecular sieves were naturally occurring zeolites such as analcime, chabazite and mordenite. Primarily R. M. Barrer and his group at the Imperial College in London extensively researched the properties of these zeolites in the 1930s. In the 1940s after the first synthetic zeolites were produced, emphasis shifted to syn1. I dedicate this chapter lovingly to the memory of Mrs. Toshiko K. Mayeda, who in the words of Jim O'Neil, was "the first lady of stable isotope geochemistry". Tosh was Harold Urey's laboratory assistant and later Bob Clayton's right hand at the University of Chicago. She advised a number of students and post-docs in the stable isotope laboratory in matters of the lab as well as of life.
806 thetic molecular sieves (Breck, 1974), which can be synthesized in higher purity and designed for specific applications. Currently natural zeolites comprise only a small portion of total molecular sieve production.
Chapter 36- H.R. Karlsson Table 36.1 - Classification of Molecular Sieves (Modified from Szostak, 1989) Silicas Titanosilicates Metallosilicates Zeolites (aluminosilicates) Gallosilicates Chromosilicates Borosilicates Ferrisilicates Metalloaluminates Germanium-aluminates Aluminophospates* A1PO4, SAPO, MeAPO, MeASAPO, E1APO, E1SPO Other Gallogerminates, Gallophosphates, Arsenates Carbon molecular sieves (CMS); molecular sieve carbon
Natural and synthetic zeolites consist of a framework made of silica and alumina tetrahedra (SiO4-4 and A104 -3 units), such that all apices (oxygens) of the tetrahedra are shared with neighboring tetra(MSC) hedra. The packing of oxygen atoms forms a regular pattern * Other than for A1PO4 acronyms are derived from the cation occupying the tetrahedral site. O= oxygen, A = aluminum, of cages and pores. These cages P = phosphorous, S = silicon, Me = metal ion, E1 = some and pores are connected to other element. form a system of channels (Figure 36.1). The size of each pore opening is determined by the number of oxygen atoms surrounding it. In zeolites pore openings consist of 6, 8, 10, and 12 oxygen atoms. Assuming a 2.7 A diameter for the oxygen atom, the openings for a perfect ring are 2.7, 4.4, 6.0, and 7.7/~ for 6, 8, 10, and 12-membered rings, respectively. Substitution of A13+ for Si4+ leads a positive charge deficiency in the crystal structure, which is satisfied by extra-framework cations located in the cages and/or channels (Figure 36.1). In natural zeolites, those cations are typically Ca2+, K + and Na + and less often Ba2+, Cs + or Fe 2+. In synthetic zeolites, charge balance is accommodated by one of these cations, other cations not generally encountered in significant concentrations in natural zeolites (e.g., H +, Ag +, Cu +, rare earth elements) or cation complexes such as NH4 +. Thermally stable and acid-resistant zeolites generally have high Si/A1 ratios (McDaniel & Maher, 1976) and thus require fewer charge-balancing cations. Terms like "zeolite-like" or "zeolitic" refer to crystals whose structures are similar to those of zeolites but lack the Si-A1 framework. The term molecular sieve is therefore more inclusive than zeolite. In natural zeolites, the cations are hydrated by water molecules that are readily removed by heating and/or evacuation. Upon dehydration the framework responds either with little or no distortion or with total collapse (Breck, 1974), the former being preferred in commercial applications. However, with proper selection and careful use, zeolitic molecular sieves can undergo repeated hydration and dehydration up to 1000 times (Hersh, 1961).
The Use of Molecular Sieves in Stable Isotope Analysis
Figure 36.1 - A polyhedral representation of the structures of zeolites A and X (and Y). The truncated cuboctahedron (a) is the basic building block in zeolites A and X (and Y). Silicon or aluminium atoms are located at the corners of the polyhedron. Oxygen atoms are situated approximately midway along the vertices (e.g., O1-O4 in c). In zeolite A, the cuboctahedra are linked to each other along the square faces with rectangles (b). What results is a cubic structure with 4, 6, and 8-membered rings of oxygen atoms. Also shown are the approximate extra-framework cation positions (SI and SII). In zeolite X (and Y) the cuboctahedra are connected at the hexagonal faces. As a result, cages are accessed through 4, 6, 8, and 10-membered rings of oxygen atoms. Approximate extra-framework cation position are also indicated (SI, SI', SII, SII', SIII). Adapted from Wortel (1979).
807
808
Chapter 36- H.R. Karlsson
36.1.2 Other molecular sieves Zeolites and their analogs are probably the best-known and most utilized molecular sieves today. The past four decades have, however, seen a drastic increase in the synthesis and commercial production of non-zeolite molecular sieves such as A1PO4types and their derivatives (Feng et al., 1997), and carbon molecular sieves (CMS or MSC; also see Table 36.1). M a n y of these non-zeolitic molecular sieves are niche applications. For example, CMS is rapidly replacing zeolites in air purification systems (Szostak, 1989; K/irger & Ruthven, 1992). For stable isotope work, however, the author has found few examples of use of CMS (e.g., Carbosphere) and no examples of some other molecular sieves such as a l u m i n o p h o s p h a t e although both have been available for some time. A brief overview of these substances will nevertheless be given as a point of contrast with the zeolites and because these materials m a y well prove suitable in future stable isotope w o r k A l u m i n o p h o s p a t e s are group of zeolite-like substances in which the tetrahedral sites are occupied either by A13+ or p5+. Only A1PO4s with P/A1 - 1 are thermally stable, whereas the Si/A1 ratio in stable zeolites can be varied over a considerable range. On the other hand, the A1PO4 f r a m e w o r k is electrically neutral and there are no charge-balancing cations in the channels. This leads to a decreased preference for polar over non-polar adsorbate molecules. Carbon molecular sieves (CMS) or molecular sieve carbon (MSC) differ from ordinary porous carbon substances in that they have a n a r r o w range of pore widths. CMS pore sizes range from approximately 3 to 5 ~ (see Table 36.2). The CMS pore structure is generated by acid etching of graphite leading to a n e t w o r k of pores or slits running along the basal planes (see e.g., Seaton et al., 1997). Although not well understood, the Table 36.2 - Examples of Synthetic Molecular Sieves Name*
Structure
Linde NaA LTA Linde CaNaA LTA Linde KA LTA Linde NaX FAU Linde NaCaX FAU Linde Y FAU Mordenite MOR ZSM-5 MEL A1PO4-5 AFI VPI-5 VFI CMS graphite
Chemical composition Na12[A112Si12048] 27H20 Na86Ca[A112Si12048] 27H20 (Na, K)86[A112Si12048]27H20 Na86[A186Si1060384] 264H20 Na86Ca[A186Si1060384] 264H20 Na56[A156Si13602841 250H20 Na8[A18Si40080] 24H20 Nan[AlnSi(96-n)O153] 518H20 [Al12P12048] R qH20** [A118018PO72] 42H20 C (anthrasite)
Pore diameter (A) 4.1 4.2-4.4 3.3 7.4 7.8 7.4 6.5 5.3 7.3 12.1 3-5
Product name 4A 5A 3A 13X 10X Y ZSM-5 MSC-4A,-5A
* Linde NaA is also known as simply "A". Similarly Linde NaX is known as "X". Linde of Union Carbide manufactures ZSM-5. Mobil has made a zeolite with the same structure and similar chemistry. It is marketed as silicalite, which is almost A1 free. ** R and q stand for (C3H7)4NOH and number of water molecules, respectively. Pore diameters are for room temperature. Sources: zeolites; Breck (1974) and Int. Zeol. Council Web site (1999): CMS; MacElroy et al. (1999).
The Use of Molecular Sieves in Stable Isotope Analysis
809
pore network most likely consists of cavities linked by smaller pores. The size and length of the pores (MacElroy et al., 1999) govern the diffusion and adsorption properties of the material. CMS behave in many respects like small-pore zeolites (e.g. analcime and zeolite A) although they lack discrete pore openings and cages (Chihara et al., 1978). 36.2 Properties of Zeolites and Molecular Sieves
36.2.1 Molecular Sieving The term molecular sieve was coined by McBain (1932). Unlike other porous solids, molecular sieves are characterized by continuous channels with a narrow range of pore sizes (Figure 36.2). The restricted range of pore sizes in molecular sieves allows these materials to separate guest molecules according to size and shape. Currently, about 150 different zeolite framework topologies have been synthesized (International Zeolite Council WEB-site, 19991) with pore sizes ranging from 2.6 ~ to 7.4 ~. Small-pore zeolites such as analcime (2.6 ~) allow only small molecules (e.g., H2, He, H20) to enter whereas large-pore zeolites such as faujasite, zeolite-X, and zeolite-Y
Figure 36.2 - Approximate pore size distribution in molecular sieves and other porous absorbents. Note the logarithmic x-axis. Individual zeolites, aluminophospates and carbon molecular sieves have much more restricted ranges in pore sizes than other absorbents such as silica gel or activated carbon. Modified from Breck (1974).
1. See WEB-site: http: / / www-iza-sc.csb.yale.edu / iza-sc /
810
Chapter 36- H.R. Karlsson
allow molecules up to 7.4 A to enter. Larger pores occur in aluminophospates (A1PO4), e.g., VPI-5 has 12.1 A wide channels, but nature still holds the record (14.2 A pore diameter in the phosphate mineral cacoxenite; Moore & Shen, 1983). The restricted range in pore size for each individual molecular sieve means that it is possible with careful selection to separate many different sized molecules. Cation identity, temperature, and pre-adsorption all influence the effective pore diameter in zeolites. Consider, for example, the synthetic zeolite Linde A. The Na endmember (Linde 4A) has an effective pore diameter of 4.1 A and will adsorb molecules smaller than that diameter (see Table 36.3). Replacement of Na + by K+ (Linde 3A) narrows the pore opening to 3.3 ~ because K+ is larger than Na +. Both forms will adsorb H20 (is 2.6 A) but the adsorption of 02 (3.5 A) and CO2 (3.2 A) drop off with increased K exchange so that 20% K exchange leads to negligible adsorption of 02. Exchange of Ca2+ for Na + (Ca = 2Na) leading to Linde 5A has the effect of widening the pore diameter by unblocking the pores (fewer crystallographic sites are occupied and "nonblocking" crystallographic sites are populated preferentially at lower site occupancies). Both Linde 3A and 4A exclude N2 (3.6 riO. At liquid nitrogen temperatures it can be sorbed on the Ca zeolite Linde 5A (Breck, 1974). Increasing temperature has two basic effects on the sorption properties of zeolites, effectively opening up the pores allowing larger molecules entrance and enhancing the diffusion rates of adsorbants. The effect of temperature on the properties of molecular sieve Linde 4A is dramatic. At room temperature N2, 0 2 , and Ar are all adsorbed very slowly. As the temperature is decreased the adsorption capacity initially Table 36.3 - Useful properties of some molecules encountered in stable isotope work Species He H2 H20 NH3 Ar N2 NO N20 02 CO CO2 CH4 CF4 SO2 H2S SiF4 SF6 NF3
Kinetic diameter (A) 2.6 2.9 2.6 2.6 3.4 3.6 3.2 3.3 3.5 3.8 3.3 3.8 4.7 3.6 3.6 4.9 5.5
* Sublimation temperature.
Dipole moment (D)
1.854 1.471
0.159 0.161 m 0.110 m
1.633 0.97
0.235
Boiling point (~ -268.9 -252.8 100 -33.35 -185.7 -195.8 -151.8 -88.5 -183.0 -192.0 -78.5* -164 -128 -10 -60.7 -86 -63.8* -128.8
The Use of Molecular Sieves in Stable Isotope Analysis
811
Figure 36.3 - Adsorption of 02, N2, and Ar on zeolite 4A as a function of temperature. Note that N2 is most effectively sorbed at around-78~ (dry ice), whereas Ar is most effectively sorbed around-160~ At-195~ (liquid nitrogen) 02 very efficiently sorbed whereas N2 and Ar are not. Adapted from Breck (1974).
increases for each of these species as illustrated in Figure 36.3.02 adsorption increases steadily with further decreases in temperature, whereas N2 and Ar adsorption peak a r o u n d - 100~ a n d - 150~ respectively. Accordingly, the most efficient adsorption of 02 would be at liquid nitrogen temperatures or lower while Ar and N2 are more effectively trapped at higher temperatures. Breck (1974) attributes this temperaturerelated selectivity to narrowing of the effective pore aperture in Linde 4A by 0.1 - 0.2 ~ at low temperatures. Apparently, this narrowing does not inhibit entry of 02 (3.5 ~) but it results in rejection of the slightly larger N2 (3.6 ~). This view is probably overly simplistic since Ar (3.4 ~), which has a smaller pore diameter than 02, has dimin-
812
Chapter 36- H.R. Karlsson
ished adsorption relative to 02 b e l o w - 150~ The fact that the activation energy for diffusion of N2 (24.3 kJ/mol) is higher than that for Ar (11.3 kJ/mol) in Linde 4A might explain this difference. The adsorption capacity of zeolites can also be influenced by the presence of molecules in the pores. Polar molecules such as H20 and NH4 are particularly effective in reducing the sorption capacity of small-pore zeolites such as Linde A. These polar molecules are strongly adsorbed and not easily displaced by non-polar molecules such as 02. The adsorption capacity of 02 on Linde 4A falls drastically, even with a small amount of preadsorped polar molecules, because cation-adsorbate complexes forming in the zeolite channel act as a blockage (Breck, 1974). In large-pore zeolites (e.g., zeolites X and Y) similar effects are seen when large inorganic cation complexes are formed (e.g. Cu pyridine).
36.2.2 Sorption Depending on grain size, the external surface areas of zeolites range from 1 to 3 m2/g and internal surface areas lie between 700 and 1000 m2/g (e.g., Hersh, 1961). Molecules small enough to enter the pores of the sieve are overwhelmingly adsorbed internally; whereas molecules that are too large will be adsorbed externally. Similarily the diffusion rates of molecules through a bed of molecular sieve will be affected. Molecules migrating through the sieve will diffuse slower than those migrating around the sieve (micro- vs. macro-pore diffusion) because they encounter higher electrostatic barriers. The amount sorbed onto a molecular sieve, or the sorption capacity depends on the maximum available void volume, the size of the sorbate molecule(s), the nature of the gas, pressure and temperature. The void volume is greatest in activated molecular sieves or zeolites, which are generated by either heating, typically in the range 300- 500~ or a combination of heating and evacuation, and varies with sorbed species. For example, in Linde 4A the void space in terms of cm3/g is 0.289, 0.252, and 0.213 for H20, CO2 and 02, respectively. In zeolite Linde 13X, the corresponding values for those molecules are 0.36, 0.33, and 0.31. Provided that the dimensions of an adsorbate molecule are similar to or smaller than the effective pore diameter, it will be adsorbed until an equilibrium state is reached. The equilibrium state varies with temperature and pressure but the sorption capacity is limited by the available void volume as discussed above. Generally, more of a vapor species is adsorbed at higher pressures and lower temperatures, although slow kinetics may make the equilibrium values impractical measures of effective adsorption. The sorption of gas mixtures onto zeolites is still poorly understood and not easily predicted. For example (Figure 36.3), pure 02 is readily adsorbed a t - 183~ on Linde 4A but pure N2 is not. In an N2 - 02 mixture, however, little 02 absorption occurs presumably because N2 hinders absorption of 02. A similar effect is seen at higher temperatures (- 78~ where pure N2 is adsorbed in greater abundance than 02. The overall adsorption of both gas species in a mixture drops significantly.
The Use of Molecular Sieves in Stable Isotope Analysis
813
High-silica zeolites (silicalite or ZSM-5), A1PO4s and CMS have molecular sieving properties similar to that of zeolites but differ in their sorption capacity. Since these substances have either few cations (silicalite, ZSM-5) or none (A1PO4, CMS) in the pores, the channels are unobstructed. Furthermore, due to charge-balanced lattice the channels show little preference for polar over non-polar adsorbates. In fact, these substances are hydrophobic compared to other zeolites. For example at 24~ Linde 4A and silicalite adsorb 0.3- 0.35 cm3/g and 0.0- 0.04 cm3/g H20, respectively. CMS are similarly hydrophobic but A1PO4 are intermediate between silicalite and Linde 4A (Szostak, 1989). 36.2.3 Commercial molecular sieves
Commercial zeolites are either natural or synthetic. Natural specimens are usually fine-grained crystals from sedimentary deposits mined primarily for the agriculture and construction industries. Most natural zeolites are not adequately characterized in terms of their chemical composition and purity, and are thus unsuitable for sorption and molecular sieving work. Additionally, large-pore zeolites such as faujasite are extremely rare in nature and therefore not widely available commercially. Synthetic zeolites have well-known chemistries and physical properties and are thus well suited for aiding stable isotope work. Synthetic zeolites are available in two f o r m s - as pure crystals and as aggregates or pellets made from zeolites and an inert binder. The binder is often a clay, silica or silica-alumina mixtures and can make up 10 to 40 wt% of the material (Breck, 1974). Similarly, carbon molecular sieves are made with a binder composed of tar or polymeric material (K~irger & Ruthven, 1992). It is important to know whether or not a binder is present because the binder changes the overall sorption capacity and isotopic exchange properties of the molecular sieve. In this paper I will refer to binder-less materials as crystals and binder-containing materials as pellets. The term granular is non-specific. 36.3 P r o c e s s e s
Various processes can alter the original isotopic compositions of the gas of interest. As the gas diffuses through (e.g., GC) or into (e.g., trap) a molecular sieve, isotopic fractionation will undoubtedly take place. The fractionation can be due to diffusion, chemisorption, and/or isotopic exchange between the molecular sieve and host gas. Kinetic effects due to differences in the diffusivites of isotopic species are well known but should not be a problem when recovery of the gas is complete. Chemisorption can be a problem when the gaseous species reacts with the host. For example, NO sorbed on the zeolites chabazite, A, and X breaks down to form N20 and NO2 or N203. The rate of disproportionation increases with decreasing temperature and is nearly complete in 1 hour a t - 78~ for chabazite (Barrer, 1978). Chemisorption, however, occurs mostly in noble-metal and transition-metal ion-exchanged zeolites. Isotopic exchange could take place between the guest molecule and the host lattice and/or other guest molecules. It is well established that oxygen isotopic exchange takes place between the framework oxygen and molecules such as 02, CO2, and H20
814
Chapter 36- H.R. Karlsson
at temperatures as low as 0~ (see e.g., Karlsson & Clayton, 1990; Karlsson, 1995, and references therein). However, since the exchange rate increases with temperature and time, it is unlikely that significant isotopic exchange will occur when these gases are desorbed from small traps or cold-fingers in a matter of minutes with gentle heating (< 200~ A worst case scenario will occur when large quantities of sieve material (e.g. sampling traps and GC columns) are heated at high temperature (> 200~ for longer periods of time (hours) to desorb gases. Conceivably H isotopic exchange could take place between adsorbed H2 or H20 and terminal OH groups in zeolites. However, since the concentration of OH is normally small compared to the guest hydrogen-bearing species, it is unlikely to have a significant effect. Exchange between guest molecules is a serious consideration when using molecular sieves in isotopic studies. For example, extensive oxygen isotopic exchange can take place between CO2 and H20 if these coexist in a sieve, either inside the sieve or upon exiting. It is therefore critical that molecular sieves be reactivated between samples. 36.4 Applications of Molecular Sieves to Isotope Analysis - Examples
Table 36.3 lists the properties of gaseous species commonly encountered in stable isotope work. Shown are relevant properties such as atomic or molecular diameter and polarity. The sizes are kinetic diameters. Breck (1974) derived these values from the Lennard-Jones and Stockmayer potentials and by considering the adsorption of species onto zeolites with well-constrained pore openings (e.g., Linde A). Dipole moments (Debye units) were taken from the 75th edition of the CRC Handbook of Chemistry and Physics (Lide, 1994). Breck (1974) evaluated species typically used in light stable isotopic analysis (e.g., H2, 02, CO2, N2, SiF4, SO2, SF6) and some other gases that can interfere with analysis. Simple cryogenic trapping, with the aid of dry ice (- 78~ and liquid nitrogen (- 195~ can be used to transfer and separate gases. However, this procedure does not work in many instances. For example, 02 and N2 cannot be transported quantitatively without the aid of liquid He. Even liquid He cannot transfer H2 and it requires a Toepler pump (or U-chemisorption). Similarly, Ar, N2, 02, and CO cannot easily be separated cryogenically from each other and so are the pairs H2S~SF6, CH4~NO, and CO2~SiF4. Many of these separations can, however, be accomplished with molecular sieves because the sieves allow one to discriminate on the basis of molecular size (kinetic diameter) and polarity (dipole and quadropole moments).
36.4.1 Attainment of good vacuum and trapping of hydrocarbons Molecular sieves such as zeolites can be used to improve vacuum and keep unwanted components from back-diffusing from roughing pumps into the main vacuum line work space. By placing a large quantity of zeolite between the roughing pump(s) and the high-vacuum pump(s) two goals are achieved. First, once activated, the molecular sieve will trap undesirable compounds such as hydrocarbons and prevent them from reaching the main vacuum system. Such fore-traps are common on VG instruments but usually contain activated alumina rather than a molecular sieve
The Use of Molecular Sieves in Stable Isotope Analysis
815
(D. Bourne, pers. comm. 1999). Second, by cooling the molecular sieve to liquid nitrogen temperatures it effectively acts like a pump (cryopump) and ~helps to produce high vacuum faster than would otherwise be obtainable. Dr. Irving Friedman's Niertype mass hydrogen isotope mass spectrometer contains such a trap and good vacuum can be obtained within half-an-hour after the machine has been down (I. Friedman, pers. comm. 1998).
36.4.2 Gas sampling Zeolites have been used to collect air samples for carbon isotope analysis. Bol & Harkness (1995) used 13X pellets (7- 8 g, 1/16") packed in quartz tubes for field sampling of CO2 in air. The zeolite cartridges were activated in the laboratory and subsequently filled with clean nitrogen. In the field, the tubes were opened and roughly 8 liters of air passed through each with the aid of a light battery-powered pump. The trapped CO2 was subsequently recovered in the laboratory by heating to 500~ in a vacuum system and separated from other gases by conventional cryogenic trapping. Comparison with CO2 samples obtained by conventional expansion into 2 1 glass bottles showed no appreciable fractionation in carbon isotopes. Bol & Harkness (1995) obtained a precision of 0.2%o or better. Recovery of CO2 from the sampled air was estimated to be > 97%. Neither the pumping rate (200 - 1000 ml air/min) or the amount of air collected (6 - 18 liters) had a significant effect on the carbon isotope values. The technique of Bol & Harkness (1995) may thus prove very useful in collecting gas samples in which virtually the entire isotope of interest resides in one species (e.g., C isotopic composition of CO2 in air). When the sieve retains multiple species, it is likely that isotopes would have been fractionated. If for example, one desired to analyze the oxygen isotope composition of CO2 in air, the procedures of Bol & Harkness (1995) would have been inappropriate because the amount of H20 collected in the trap would have been many times greater than for CO2. Upon heating, oxygen isotopic exchange would have ensued between these two species. Indeed, R Bol found that there was significant oxygen isotopic fractionation in the CO2 (pers. comm. 1999). Schwarz et al. (1996) collected air moisture in the Antarctic using an automatic sampler. Their device featured zeolite molecular sieve 5A (Wolfen Zeosorb A5; G. Schwarz, pers. comm., 1999) to trap the moisture for hydrogen isotope analysis. The device has twelve zeolite cartridges containing 50 g zeolite "balls" (1.5 to 2.0 mm diameter), each of which can be used to collect a sample at a specified time interval. The zeolite was activated by heating to 400~ under vacuum. In the field, air was pumped through the cartridges with the aid of a battery-powered micro-vacuum pump. Sampling times were 6 - 150 hours with airflow rates of 100 - 1201/hour. Thus, at least I g of water was collected onto the molecular sieve. The water was recovered by heating the zeolite cartridge to 400~ under vacuum (Schwarz et al., 1998a) and the evolved H20 reduced to H2 by reaction with hot Cr (Schwarz et al., 1998b). 6D values of moisture collected using the molecular sieves are within 5%o of data obtained by cryogenic methods, which is quite good considering that sampling was done some days apart.
816
Chapter 36 - H.R. Karlsson
36.4.3 Gas separation Cheng & Bremner (1965) report that Linde 5A used in a GC column is effective in separating N2, 02, NO, CO, CH4, N20 and CO2. At room temperature, the sieve retains N20 and CO2 but N2, 02, and NO are separated. N20 and CO2 are released upon heating but no mention was made of possible isotope effects. Meier-Augenstein et al. (1994b) tested the use of zeolite molecular sieves in removing moisture from CO2 produced during breathing. The samples were obtained from human subjects and a 200 ml aliquot passed through columns containing molecular sieves at room temperature. Zeolites 3A, 4A and 5A were tested and compared with results obtained by conventional methods (dry ice trapping). Two 1/4_inch columns were u s e d - the short one was 152 mm and the long one was 994.8 mm. Breath samples were run sequentially through the sieves without reactivation between samples. CO2 passed through the columns containing 4A and 5A experienced large fractionations in both C and O isotopes and the use of these zeolites was therefore not investigated further. More extensive work was carried out with zeolite 3A including tests of memory effects. It appears that C and O are not fractionationed substantially (> 0.5%o) especially in the case of 3A short column. Zeolite 3A may therefore be used as simple way to remove moisture from breath samples without introducing significant isotopic fractionation. The observed isotope effects may readily be explained when the pore sizes of these zeolites are considered. In the case of 3A, little CO2 enters the pores but H20 is readily adsorbed. The CO2 thus flows past the molecular sieve. In the case of 4A and 5A however, both molecules enter the zeolite structure and interact with each other and with the zeolite framework generating significant isotopic effects.
36.4.4 Gas transfer and purification Oxygen (02): Clayton & Mayeda (1983) used zeolite 13X to separate NF3 from 02 prior to isotopic analysis of oxygen isotopes in meteorites. Invaluable information can be obtained by analysis of the three isotopes of oxygen in meteorites and the analysis is best done using 02 rather than CO2. During fluorination of meteoritic samples, NF3 is produced in such large quantities that erroneous 6170 results (NF + has a m / e of 33, the same as 170160) may be obtained. The mixture is, therefore, adsorbed onto zeolite 13X pellets (Anasorb" 30 - 60 mesh) at liquid nitrogen temperatures. Subsequently, the temperature of the sieve is raised t o - 115~ (using a solid/liquid ethanol slurry) releasing the 02 (and Ar) but retaining the NF3 and N2. The purified 02 is then transferred to a mass spectrometer using a liquid He cold-finger. The 13X was regenerated between experiments by heating under vacuum to 250~ for 30 minutes (R.N. Clayton & T. K. Mayeda pers. com. 1999). Miller et al. (1999) used a slight variation of Clayton's & Mayeda's method. In addition to using 13X to purify 02, Miller et al. (1999) used a second 13X cold trap for transporting the 02 to the mass spectrometer inlet system. In order to release 02 into the mass spectrometer, they raised the temperature of the 13X trap to 50~ for 6 minutes. Although, it is certain that oxygen isotopic exchange occurred between the
The Use of Molecular Sieves in Stable Isotope Analysis
817
molecular sieve and the 02 gas, the exchange appears to have been negligible due to the low temperature used to heat the sieve and the short duration of heating. The 13X was activated between samples by heating it to 105~ for about 40 minutes (M. F. Miller, pers. com. 1999). MacPherson et al. (1999) used zeolite 5A for storage of 02 used for combustion in their major volatile extraction system. Details of reactivation or trapping on the sieve were not given. Wassenaar & Koehler (1999) employed molecular sieve 5A to purify 0 2 from air, soil gas and water for isotopic analysis in a continuous-flow mass spectrometer (CFIRMS). Sample purification was carried out on an on-line modified Carlo Erba NA1500 elemental analyzer. The sample was contained in a He-carrier stream. After the initial purification steps, that removed H20 and CO2, the sample was passed through a I m GC column containing 5A. The GC column, held at 35~ was used to separate 02 from N2. Finally, the 02 was carried directly into an IRMS utilizing a He carrier gas where it was measured isotopically.
Nitrogen (N2): Nitrogen has been transferred to mass spectrometer inlet systems for isotopic analysis using zeolite molecular sieves. Macko (1981) used the closed-tube combustion technique to produce N2 from organic samples. The reaction also produced CO2, which was retained by freezing the sample tube in liquid nitrogen. The N2 was then transferred onto a zeolite 5A crystals (30 - 60 mesh) frozen at liquid nitrogen temperature for ten minutes (S. A. Macko, pers. comm., 1999) and released into the mass spectrometer by heating the sieve to 300~ for another ten minutes. For shorter periods of either freeze-down or bake-out of the sieve, the N2 experienced isotopic fractionation. During the freezing process, 15N was preferentially condensed onto the molecular sieve relative to 14N and thus too short a freezing period led to increased 615N values. Shorter bake-out period can lead to complex behavior with increasing time. 615N values first decreased, then increased and finally, decreased until the true value was reached. It thus appears that nitrogen is being released from different crystallographic sites within the molecular sieve. The fractionation observed by Macko (1981) resulted in 615N values that were within 0.5%0 of the true value. More recently, Macko (pers. com. 1999) reactivates the sieve by heating at 180~ under vacuum for a couple days. After transferring the N2 to the 5A cold-finger, he now releases the gas into the mass spectrometer at 150~ rather than 300~ This ensures that contaminants that come out of the sieve at higher temperatures are not released. Bebout & Fogel (1990), working with silicate minerals and rocks, compared Macko's molecular sieve method with a gas expansion technique that did not involve a molecular sieve. They found that the sieve trapped CO2 and H20 in addition to N2 and that the N2 blank was higher by 0.1 - 0.2 gmole when the sieve was used. However, heating the sieve to 400~ for 3 hours reduced the blank. Boyd et al. (1988) used a molecular sieve of an unspecified type in a stepped combustion/pyrolysis system designed to produce nmole quantities of N2 for isotopic
818
Chapter 36 - H.R. Karlsson
analysis in a static mass spectrometer. A cold-finger containing pellets of the 5 A pore molecular sieve was cooled in liquid nitrogen (- 196~ to collect gases produced during the combustion or pyrolysis procedure. The sieve was then heated to 200~ in one minute to release all trapped gases (N2, CO2, CO, CH4, N20 etc.) which were then oxidized over hot CuO (850~ to produce H20, CO2 and SO2 (N20 was converted to N2 and 02). Excess oxygen was readsorbed onto the CuO by lowering its temperature (600~ Following the removal of 02, the N2 is purified cryogenically by freezing the other gases onto a liquid nitrogen-cooled cold-finger. The molecular sieve was maintained at 300~ under vacuum between runs (typically overnight). Blanks for the pyrolysis/combustion system were reportedly better than 0.04 nmole suggesting a very low N2 blank for the molecular sieve.
Sulfur (S): Puchelt et al. (1971) employed zeolite 5A to remove Br2 from SF6 prior to sulfur isotope analysis. This was done by passing the gases through a 5 feet by 3 / 4 inch GC column packed with the zeolite and held at 100~ He gas carried the components through the GC and a liquid-nitrogen-cooled trap removed the SF6 from the carrier gas. Tests run on the GC with pure SF6 of known isotopic composition gave quantitative yields and unfractionated sulfur isotopes.
Silicon (Si): Molini-Velsko (1983) determined Si istotope ratios in meteorites using SiF4 as a working gas. The SiF4 was extracted from the samples using the BrF5 method of Clayton & Mayeda (1963). Purification of the SiF4, however, was necessary because contaminants such as SF6, SO2F2 and CF4 were also created. One of the purification methods that Molini-Velsko (1983) examined involved using a molecular sieve column in a gas chromatograph. Four types of molecular sieves were tested in the 3 feet by 0.85 inches chromatographic column: 4A, 5A, 13X (30 - 60 mesh) and CMS (Carbosphere, 60 - 80 mesh). The column was located inside a furnace, which allowed temperature control. Samples were carried through the column using He carrier gas. The 4A sieve failed to separate SiF4 and SF6 since they were too large to enter the zeolite pores. Attempting to open the 4A pores by heating up to 350~ had no effect. SF6 passed easily through 5A and 13X but SiF4 did not emerge even after several hours at elevated temperatures. The CMS (~ 13 ~ pores) eluted SiF4 ahead of SF6 but there was too much overlap between the two gases. According to Breck (1974) SiF4 is to some extent irreversibly sorbed onto 4A and 13X at 0~ suggesting that chemisorption took place. However, if these zeolites are held at 200~ it appears that SiF4 is reversibly sorbed.
Methane (CH4): Jackson et al. (1999a) used molecular sieves in a preparation system designed to handle small quantities of atmospheric CH4 for isotopic analysis in a static mass spectrometer. Air samples were carried through a gas chromatograph containing CMS (Carbosphere, 80/100 mesh) using a high-purity He stream, which had been cleaned by running it through a molecular sieve 13X (IMS-100) to remove H20 and CO2. As CH4 exited the chromatograph it was run through zeolite13X held in liquid nitrogen. Once CH4 collection was complete, non-condensable He was pumped away, CH4 and other trapped gases released from the 13X trap by heating, and the
The Use of Molecular Sieves in Stable Isotope Analysis
819
gases transferred to a 13X cold-finger held at liquid nitrogen temperatures in the inlet section of the mass spectrometer. Once transferred, the CH4 was desorbed from the 13X and cleansed of N2 using getters (the eluted gas was only 10% CH4, the balance being N2). The purified CH4 was finally exposed to a trap held a t - 188~ to remove any trace of H20 that might have remained and admitted to the mass spectrometer. Analysis of a CH4 mixed in with the appropriate concentration of N2 yielded a precision of 0.3%o indicating that the 13X had little effect on the isotopic composition of the CH4.
Hydrogen (H2)"The author found no examples of H2 collection or trapping. Because of its low boiling point (Table 36.3), it is unlikely that H2 could be trapped completely on a molecular sieve unless liquid He was used. Preliminary experiments by the author suggest that only 90 - 95% of H2 would condense on zeolites (Linde Corp. ET200 and Oxysiv-5) at liquid nitrogen temperatures. Zeolites are thus similar to activated charcoal (see e.g., Halas & Durakiewiz, 1995) in terms of sorption capacity. According to Breck (1974), D2 is preferentially sorbed over H2 onto zeolite 4A. If this holds true for other zeolites and hydrogen gas adsorption is incomplete even a t 198~ then fractionation of hydrogen isotopes will take place in marked contrast to activated charcoal. 36.5 Conclusions
Molecular sieves have been used to 1) improve vacuum performance, 2) collect gases in the field, 3) separate gases, and 4) transfer gases within vacuum systems. Except for the first of these processes, it is essential that the isotopic composition of the gas of interest remain intact at the completion of the operation. Thus far, only a handful of the available molecular sieve materials have been utilized in isotope work. Rarely have molecular sieve a n d / o r trapping processes been used in combination although such an approach could be fruitful. For example, a small pore sieve could be used to separate small molecules from large ones (molecular sieving). A large pore sieve would then be used to separate the larger gas species according to boiling point (trapping). It is surprising that molecular sieves such as zeolites are not more widely used considering that they are very sturdy and non-toxic substances. Future studies are needed to investigate the use of other molecular sieves such as the high silica zeolites (silicalite, ZSM-types) and non-zeolitic material such as A1PO4s and CMS.
Acknowledgements I am grateful to Mrs. Mayeda and Drs. Clayton, Macko, Miller, Meier-Augenstein and Schwarz for providing further details in their use of molecular sieves. Early versions of the manuscript benefited greatly from reviews by Dr. John R. Beckett. Reviews by an anonymous reviewer and Drs. Bol and Schwarz improved the final version.
Handbook of Stable Isotope Analytical Techniques, Volume 1 P.A. de Groot (Editor) 9 2004 Elsevier B.V. All fights reserved.
CHAPTER 37 Introduction to Isotope Dilution Mass Spectrometry (IDMS) Michael Berglund European Commission- Joint Research Centre, Institute for Reference Materials and Measurements (IRMM), Retieseweg, 2440 Geel, Belgium e-mail: [email protected]
37.1 The beginning The purpose of this chapter is to give the reader a short introduction to IDMS. The basic principles will be outlined and explained. The IDMS equations will be derived and explained. The main benefits of IDMS will be discussed as well as some pitfalls. There will be no detailed discussions on element, or technique specific subjects because the sheer number of topics would be impossible to cover in this chapter without being terribly incomplete. The use of IDMS by Inductively Coupled Plasma Mass Spectrometry (ICP-MS) has been covered to some extent in "Inductively Coupled Plasma Spectrometry and its Applications" (Vanhaecke et al., 1999b). Thermal Ionization Mass Spectrometry (TIMS) and IDMS has been discussed by Fassett & Paulsen (1989). The isotope dilution technique is used in many fields of analytical science, from biochemistry to geology. Despite the very varied applications of isotope dilution in existence they all have one thing in common, the addition of a spike to a sample material. In the most common case the spike consists of a known amount of an isotopically enriched element. The amount of this element in the sample material is then the measurand of the IDMS measurement. From a principal point of view the element in the spike only need to have a different isotopic composition relative to the same element in the sample material. This means that IDMS can only be used on elements with more than one stable isotope, and if these isotopes are measurable on a mass spectrometer, IDMS is in principle possible. So, what does this addition of an isotopically enriched element lead to? To answer this I would like you to make a mental switch and imagine Stanley meeting Dr. Livingstone in the village Ujiji at the Tanganyika lake in Africa in 1871. Little did they know that some 60 years later an entymologist named C. H. N. Jackson were going to study the density of tsetse flies in this very region (Jackson, 1933), and as a tool to count tsetse flies he used fly dilution. This is to my knowledge the first time the principle of what is now named isotope dilution was documented. Etienne Roth (1997) mentions that the dilution method was first used for evaluating populations of rare bird species on islands but no date or references are given. First or not first, let us look at a small scale simplified model of the 1933 African fly dilution.
821
Introduction to Isotope Dilution Mass Spectrometry (IDMS)
In Figure 37.1 below a number of marked flies, coloured black, are added to a normal population of white flies. I am quite sure Jackson did not colour his tsetse flies black but for the sake of argument in this example, normal tsetse flies will be white and marked, or coloured, flies will be black. After waiting for a suitable amount of time to allow for the black flies to blend into the white fly population a sample is taken. This is indicated in Figure 37.1 by the dotted line and seen in Figure 37.2. After Jackson collected the sample blend of flies in Figure 37.2 he calculated the true number of flies using equation [37.1]:
marked flies in sample unmarkerd flies in sample
total marked flies total unmarked flies
[37.1]
which after input of the example data gives: 3
10
total unmarked flies-
6 total unmarked flies
6.10
3
= 20
[37.2]
Is not it beautiful and simple? Do note that in the example above the worries of statistically representative samples and counts have been left out. The criteria for achieving a true count of tsetse flies using fly dilution is given in Table 37.1. Translate these criteria from flies to nuclides and you get the only criteria you need to master IDMS. It is not always easy and there can be numerous difficulties to overcome when applying IDMS. These difficulties can be either in assaying the spike material, i.e. knowing exactly how many black flies you added or achieving a homogenous blend after adding the spike. When you think about it, how did Jackson make sure he got a homogenous mix of marked and unmarked tsetse flies? He does not discuss this, he was more concerned with diffusion of flies into and from his fairly self-contained fly community. Jackson had his difficulties, in IDMS we have ours. The actual count-
. . . . . . . . . . .
......
:~
:.:::~
~ v ~;:~iii!i~i~':
g g ~i!i!!!~i!!~::
~i~
~
i'~i~
~ ~:.v::G::i:
.....
.............
....
..................&::,::::
:~i;::
§
....
Figure 37.1 - Fly dilution experiment. Z ~: ~(s i~i::!:.i!i~:
t
Figure 37.2 - Sample taken from the blend of flies. --)
.... ~=..~.
..... ,,~,N:~:~,:,,,~,~ ~.
.
.
.
~
822
Chapter 37 - M. Berglund
ing of nuclides is not always Table 37.1 - Criteria for an accurate fly dilution experiment straightforward. There can for example be isobaric interfer- 1 you need to know how many flies you have marked and released ences, which can be seen as 2 you need to know that there is a homogenous blend of additional species of flies, non white and black tsetse flies tsetse flies, getting caught in 3 when you start counting flies in your fly mix you must be the sampling process and able to identify and separate black and white tsetse flies wrongfully counted as either 4 you must not loose your count white or black tsetse flies. Another effect that will render an erroneous count of the sampled blend, is the blank contribution. In the fly case this means that you are counting more white flies than you actually sampled in Figure 37.2. They can have entered your little net while you happily walked back to the camp in the African jungle, passing areas with a normal tsetse fly population. Maybe a few landed on your table, drawn there by the light in your tent in the pitch-black African night while you were busy counting.
37.2 At the highest metrological level IDMS has been recognized as a primary method of measurement by the Comit6 Consultatif pour la Quantit6 de Mati6re (CCQM)I or as it is named in English the Consultative Committee for Amount of Substance. In every day life everyone refers to it as CCQM. CCQM is part of the Comit6 International des Poids et Mesures (CIPM) whose main function is to ensure the propagation and improvement of the SI system. The primary method was defined by CCQM in 19952 and slightly redefined in 19983. A primary method of measurement is a method having the highest metrological qualities, whose operation can be completely described and understood, for which a complete uncertainty statement can be written down in terms of SI units. The two terms primary direct method and primary ratio method were introduced in 19983 to better incorporate IDMS since it is not a direct method like coloumetry. In the explanatory notes we read: A measurement traceable to the SI can be made using a primary ratio method in combination with a reference of the same quantity that is itself traceable to the SI. However, a method whose operation cannot be completely described and understood cannot be a primary ratio method. The fact that IDMS has been recognized as a primary method of measurement is not only due to its transparency, which is the qualifying criterion, it has also proven itself in the line of duty. In Figure 37.3 we see CCQM's second key comparison (Papadakis et al., 2001)4 exercise where all participating national metrology institutes (NMI) were using IDMS. In the same figure, measuring identical samples, we see results from the participants to IMEP-9 (Papadakis et al., 1999)5. IMEP is a tool with which field labo1. 2. 3. 4. 5.
http: / / www.bipm.fr / enus / 2_Committees / CCQM.shtml Report of the 1st meeting of CCQM, 1995 Report of the 4th meeting of CCQM, 1995 http://kcdb.bipm.org/BIPM-KCDB / default.asp http://www.irmm.jrc.be / imep
Introduction to Isotope Dilution Mass Spectrometry (IDMS)
823
Figure 37.3
824
Chapter 37 - M. Berglund
ratories can compare their results against SI traceable values. The participants to IMEP-9 work under normal conditions of their choice, with respect to technique, instrumentation etc. The grey bar in Figure 37.3 is the certified range for IMEP-9. The agreement between the NMI laboratories in CCQM-K2, and other CCQM key comparisons, reflects the state of the art in chemical measurements. IDMS has become the most important technique for accurate measurements of amount content and at the highest metrological level, which the key comparisons represent, it is the method of choice. Even though IDMS has the capacity to yield very small uncertainties it must be pointed out that it does not come automatically. IDMS can be misused as any other technique. It will not surprise you, but very careful planning and correctly applied measurement procedures are needed for optimum performance. For further information on accuracy of IDMS look at De Bi6vre (1990) who has written an article on the accuracy aspect of IDMS. OK, IDMS is transparent and can be very precise, but what makes it such a solid technique is the fact that it works even if we have non-quantitative recoveries from for example digestion or separation steps. It is also independent of instrument sensitivity. This is due to the fact that, with or without sample losses, whether the instrument is perfectly optimized or not, the measured amount ratio of two nuclides of the same element, is always the same. This of course provided that adequate mixing of sample and spike has occurred.
37.3 Deriving the IDMS equation The IDMS equation derived below represents the simplest case. On the other hand it is the building stone for all other variations. We are interested in determining the number of atoms nx(E) of a specific element, E, in a sample, denoted by the subscript x. An isotopically enriched spike with subscript y will be used to prepare a blend of sample and spike. This blend will have a subscript B. The term nuclide will now be dropped in favor off isotope, which is less correct, but is the commonly used term in the mass spectrometric community. So, let us first define some basic relationships"
nx(E ) - ~ nx(iE)
[37.3]
where the superscript i denote an isotope of the element. In the same way the number of atoms in the enriched spike, denoted by the subscript y, will be:
ny(E) - ~ ny(iE)
[37.41
One isotope, usually the most abundant in the sample, is selected as reference isotope and all isotope amount ratios are expressed relative to it. The reference isotope and the most abundant isotope in the spike, are then normally selected to give the master amount ratio. If we now make a blend of sample and spike we can define an isotope
Introduction to Isotope Dilution Mass Spectrometry (IDMS)
825
Figure 37.4a-c- See text for explanation.
amount ratio of two well-chosen isotopes, aE and bE, as:
RB -
nx(bE) + ny( bE)
nx(aE) + ny(aE)
[37.5]
The reference isotope is thus aE and the master amount ratio to monitor is n(bE)/ n(aE). Let us look at an example. In Figure 37.4ac we see the isotope amount fractions of a hypothetically prepared blend for an IDMS measurement using magnesium as an example. In Figure 37.4a-c we see the sample '4a', with natural isotopic composition, and the spike '4b', enriched in 26Mg, and the blend '4c', which is then a mixture of sample and spike. The most suitable master amount ratio in this example would then be n(26Mg)/n(24Mg). In most cases it is advisable to strive for an RB close to 1 because this can cancel out possible non-linearities in the detector system. However, in many cases this is not possible because of for example, low enrichment in the spike available or a specific choice of master amount ratio is desirable. In these cases a
826
Chapter 37 - M. Berglund
thorough a priori uncertainty analysis should be performed. An investigation of the influence of measurement precision on the choice of blend ratio has been done by De Bi6vre (1994). To find out the exact number of atoms of element E in the sample we will solve for nx(E). We would like to express it in terms of isotope ratios, because those we can measure. Useful as a start are then the following definitions:
Rx Ry
_
nx(bE)
[37.6]
nx(aE) (bE
n
Y
)
[37.7]
ny(aE) ~ny(iE) l
~Riy -
aE
[37.8]
ny( ) ~nx(iE) Rix - ' nx(aE)
[37.9]
where Rx and Ry are the amount ratios of isotopes bE and aE in sample and spike respectively. YRix and YRiy a r e the sums of amount ratios for all isotopes of an element relative to the reference isotope for sample and spike respectively. Now we have all the necessary ingredients. Now it is just a matter of an algebraic rearrangement. Inserting equations [37.6], [37.7], [37.8] and [37.9] into equation [37.5] gives:
RB -
nx(aE) " Rx + ny(aE) " Ry
~nx(iE) t
~ny(iE) +
~iix
aB -
t
~iiy
[tlx(aE) " ax 4- Fly(aE )" ay] " ~ a i x " ~ a i y
~nx(iE)" ~Riy + ~ny(iE)" ~Rix l
RB'I~nx(iE)'~Riy l
l
+~ny(iE)'~aixl
- [nx(aE).Rx+ny( aE ).Ry]. ~ Rix " ~ Riy
l
inserting equations [37.8] and [37.9] into the equation above gives"
827
Introduction to Isotope Dilution Mass Spectrometry (IDMS)
R B'I~nx(iE)'~Riy +
iE . RixI -
l
I~ Flx(iE) ~Fly(iE) [ ~-a2 "Rx+ aiy
1
_
ix
and if we now insert equations [37.3] and [37.4] in the resulting equation above we get:
nx(E) ny(E) ~ RB'[nx(E)" Riy 4-ny(E).~aix ] - I ~-R~;'Rx+ ~aiy'Ryl" ~Nix" ~aiy now rearrangements and simplifications will give us the IDMS equation
[nx(E) . ~ R i y . R B + ny(E) . ~aix
. aB] - [nx(E) . ~Riy. ax 4-ny(E) . ~aix . ay]
nx(E)'[~Riy'RB-~Riy" Rx] - Fly(E) . [~Rix . Ry- ~Rix. RB] nx(E ) _ ERix. i y - ERix. i B Ry-RB
nx(E)
-
ny(E)-RB_ Rx" ~iiy
[37.10]
In equation [37.10] we have managed to express nx(E) in terms of measurable or given quantities, isotope amount ratios, R, and amount of added spike, ny(E). In the easiest case where the masses of spike and sample are determined gravimetrically, mx and my in (kg), and cy the amount content (mol/kg) of the spike is given, which is the case if a certified spike is used, it will turn equation [37.10] into:
cx
_
cy
.my.[Ry-RB]
mx [RB--Rxx]
~Rix 9
[37.11]
In scientific journals it is common to see equation [37.11] with added input quantities, correcting for various effects, and of course other names for the input quantities. Every now and then you will also see a different looking IDMS equation like in Fassett & Paulsen (1989). This equation can, however, with some simple algebra be turned into equation [37.11]. There is a drawback with the Fassett & Paulsen (1989) equation and that is that amount fractions are used as input quantities. Amount frac-
828
Chapter 37 - M. Berglund
tions of isotopes of the same sample are correlated and this complicates the uncertainty calculation. 37.4 Mass fractionation
correction
In the equation above Cx is the amount content of element E in the sample. Equation [37.10] or [37.11] is what you will find in most articles and textbooks covering IDMS. It may look simple, but there is more to this equation than meets the eye. Following the derived IDMS equation it is clear that the ratios, Ri, given in equation [37.10] and equation [37.11] are absolute amount ratios and not measured ratios. Mass discrimination is a well-known effect in mass spectrometry and is caused by, for example, evaporation, diffusion and electrostatic effects. These effects are mass and time d e p e n d e n t a n d the result is, that in the mass spectrometer, different isotopes do not race on equal terms. Heavier, or lighter, isotopes will, due to the effects mentioned, be favoured which will lead to non-absolute isotope amount ratios. This correction factor, K, is measured using a certified isotopic reference material. The assumption is that the quotient of the measured ratios of sample and reference are equal to the quotient of the true amount ratios of sample and reference. This is a valid assumption if the mass fractionations, fms and fmR, for the measured ratios rs and rR respectively are equal. The subscript R denotes a certified reference material and the subscript S the sample in question. To calibrate a measured ratio an isotopic reference material traceable to the mole must be used. Tables of available reference materials meeting this criterion are listed in Part 2, Chapter 40. For the elements not present in this table the IUPAC Table (Rosman & Taylor, 1998) can be used. This does not mean it is impossible to do IDMS on elements not in the table. We just introduce a small exception. For the IDMS equation to work a mass fractionation correction factor must be applied. An exception can be when the K-factor is the same for two or more measured ratios where some cancellations are possible. fmS" rs fmR rR 9
Rs RR ~ Rs
~
RR
rs ~ Rs - K r s rR ,
,
if
RR K - rR
[37.12]
In equation [37.12], and henceforth, a small r will denote a measured amount ratio and a capital R the absolute isotope amount ratio. This would turn equation [37.11] into Cx _ Cy . m y . [ KRy 9ry - KRB 9rB] 9~ (K i 9rix ) mx [ K R B ' r B - KRx "rx] ~ ( K i ' r i y )
[37.13]
Every measured isotope amount ratio for element E has its own K factor. In equation [37.13] KRB, KRx and KRy are all correction factors for the measured master ratio nx(bE) / nx(aE) but it does not necessarily mean that these mass fractionation correction factors are identical. In ICP-MS mass fractionation can be very time dependent and matrix dependencies of K are seen for both TIMS and ICP-MS. In the sum of ratios, all amount ratios have different K-factors. Since the same reference isotope has to be used
829
Introduction to Isotope Dilution Mass Spectrometry (IDMS)
for all amount ratios the master ratio, in our case nx(bE)/nx(aE) will enter the sum of ratios as well.
37.5 An IDMS example If we take a look at equation [37.13] using a real element, chromium, as an example we first need to define the reference isotope and the master ratio. The most abundant isotope of chromium is 52Cr so let us use this one as reference isotope. If we use a spike isotopically enriched in 50Cr the master ratio would be n(50Cr)/n(52Cr). This would turn equation [37.13] into equation [37.14]" \
my.
Cx - Cy" mx
IKRy'r50-KRB'rB501
~ (Kx50 9r 50 + 1 + Kx53 9r 53 + Kx54" r 54} x~ x5~ x~ /
IKRB'rn50-KRx'rx501
~ (Ky50 9r 50 + 1 + Ky53. r 53 + Ky54" r 541 Y~-~ Y~-~ Y~-~"
Y5-2
~
[37.14] Here the indexes x,y and B denotes sample, spike and blend respectively. The indexed ratios (i.e. 50/52) indicates the isotope ratio. For the K-factors the indexes indicates the measured ratio they are correcting. Now someone might say: but look, this is a lot of ratio
measurements, not only do I need to measure rx, ry and rB, every K-factor also carries a ratio measurement. Well, normally you never measure all of these because some are given in a certificate, if you use a certified spike material and often IUPAC values (Rosman & Taylor, 1998) are used for isotope amount ratios where the assumption is that the amount ratio in question has a natural isotopic composition. Do not forget that assumptions normally carry an uncertainty. Assuming that a sample has an isotopic composition that is natural and represented by the composition given by IUPAC can be wrong. As a contrast let us look at the simplest case. If we in this example use a certified chromium spike and assumes that the sample material has a natural isotopic composition equation [37.14] would turn into:
cx
Cy my I1 L R y - K R B ' r 50_ . . . . . . ~J ~ mx IKRB " r 50 - Rxl ~
R
(Riy )
where
KRB -- r'~5R
[37.15]
K-~
where only the weighings of mx and my and the measurements of the amount ratios rB and rR need to be performed. The amount content Cy, Ry, and all Riy are taken from the spike certificate and Rx and all isotope ratios for the sample, Rix, are taken from IUPAC. From my experience it is always beneficial to measure rx, rix and relevant Kfactors, even if the isotope amount ratios can be assumed to be close to values given by IUPAC. First of all, mistakes will be avoided and there will also be fewer assumptions to worry about w h e n it comes to uncertainty budgeting. However, there may be cases with severe interference on some masses with the result that all rix cannot be
830
Chapter 37 - M. Berglund
measured without introducing very large uncertainties. In these cases it may be better to use IUPAC values. 37.6 D o u b l e I D M S
Equation [37.10] and the derived equation [37.11], require a well-defined spike. As you may remember it is a must to know how many atoms of the enriched spike you have added, see criterium 1 in Table 37.1. The availability of such certified reference material, enriched in a specific isotope with a certified amount content and isotopic composition is somewhat limited. To overcome this, 'double' IDMS is frequently used. The idea here is to use a material of natural isotopic composition as primary assay standard. Do note that there is no true natural isotopic composition of an element. The isotopic composition of an element varies in nature, unless it is mono-isotopic of course. Sometimes this variation is measurable and sometimes it is not. With every generation of mass spectrometric instrumentation the trend is that isotopic variations are detected for more and more elements. To perform double IDMS we need to make another blend, here called blend B'. For blend B', we use a well-characterised primary assay standard with the amount content Cz and our enriched spike material y. The subscripts x, y and z denotes sample material, spike material and assay material respectively. Let us first set up an equation, equivalent to equation [37.10] but with the assay material (z) instead of the sample (x). c z - cy m ' y . [ R y - R B , ] 9 mz iG;---R-z]" ~ R i y
[37.16]
All quantities in equation [37.16] are equivalent to those in equation [37.11] except that it concerns assay and spike (z, y) instead of sample and a spike (x, y). The mass of spike y for the blend assay and spike is denoted m'y to distinguish it from the mass of spike my used in equation [37.11] If we now divide equation [37.11] with equation [37.16] we get:
CX
Cz
9m y . [ R y - RB] " ~ , R i x Cy m x [ R B - R ] x ~ R i y
Cy"
m ' y . [Ry - RB,]. mz
[R B , - R z ]
~Riy
We can directly see that the amount content of the spike, Cy, and the sum of spike ratios, l~Rip are cancelled out. Reconstructing gives:
C x _ Cz"
m .m
z. m x 9m y [R B - R K ] ' [ R y -
. RB, ] I ~ R i z )
[37.171
Introduction to Isotope Dilution Mass Spectrometry (IDMS)
831
Very often equation [37.17] is used with the ratio of the sum of ratios equal to unity. This can only be done if the isotopic composition of the sample and assay materials is identical. In most cases when this model is used the sample and assay materials have a natural isotopic composition. The question is can we cancel out the sum of ratios without introducing a bias? To decide that the isotopic compositions are equal, without measuring, requires a good knowledge of natural isotopic variations of the particular element and knowledge of the origin of the sample and assay materials. If in doubt leave it in the equation and measure YRix and ZRiz. In equation [37.17] only calibrated ratios are used and mass fractionation correction factors need to be applied, see equation [37.12] and [37.13].
37.6 Isotope specific IDMS In nuclear mass spectrometry very often the target is not the total amount of an element. Here the total amount of a specific isotope is more interesting. Combining equation [37.5], [37.6] and [37.7] we get equation [37.18] or equation [37.19] depending on preferred master ratio. The master ratio will of course consist of the enriched isotope in the spike and the sought for isotope in the sample.
Ry-RB. 1
Flx(aE) - Fly(bE). G
~
-ay
[37.18]
n x (bE) - rly(aE) 9R G y- R- xR B . Rx
[37.19]
37.7 Some difficulties with IDMS IDMS will work perfectly unless you fail one of the criteria in Table 37.1. Let us look at these criteria in more detail and transfer them to mass spectrometry and see where possible problems may show up.
37.7.1 How many enriched atoms have been added? There are some spikes (Table 41.2, Chapter 41), certified for both isotopic composition and amount content available that are made for IDMS measurements. If you do not have access to any of these, or are measuring on an element where none is available you will have to follow the procedure outlined in Double IDMS.
37.7.2 The necessity of a homogenous blend of sample and spike. This is regarded as one of the most critical steps in IDMS, and this for a good reason. Let us assume we have a digestion procedure and it is not 100% efficient, which is not that unlikely. If the blend is digested, some of the sample atoms of element E may still be bound in a complex of some kind resulting in a measured blend ratio that does not represent the true amount ratio of n(bE)/n(aE) in the blend. There could also be a disproportionate amount of a volatile complex of element E formed due to the initial matrix differences between sample and spike, which would also render an erroneous blend amount ratio. Even if digestion is the most common pre-treatment in inorganic chemistry this problem holds for any pre-treatment of the blend. To achieve an accurate result, losses from blend pro-
832
Chapter 37 - M. Berglund
cessing have to affect element E in the added spike exactly the same way it affects element E in the sample. The best way to identify a digestion problem is to look at the blend to blend reproducibility. If this is alarmingly poor in comparison to the uncertainty of a measurement, it is an indication of a problem in your measurement procedure and the digestion is where you should look first.
37.7.3 Mass spectral interferences and contamination. Mass spectral interferences can be tricky and possible interferences have to be checked for every mass used in the measurements. There are only two ways of handling mass spectral interferences and that is either by estimating the interference, e.g. by measuring, and apply a proper correction, or remove the interfering compound. In IDMS it is also very important to be aware of the contamination risks when working with enriched material. Not only is it easy to contaminate the enriched material but also to change the isotopic composition of sample, blends and standards. It is crucial to realise that contamination on an isotopic composition level is a real danger in IDMS. 37. 7.4 Counting ions. Counting ions is usually done by secondary electron multipliers (SEM) or faraday cups backed up with the appropriate electronic gadgetry. While faraday cups are known to be precise, stable and linear they are much less sensitive than the SEM. The SEM on the other hand, sensitive as it is, is of a more delicate nature. There are three important parameters that need to be controlled when it comes to SEMs and that is dead time, trigger level and the plateau. The dead time is the time the SEM is blind to incoming ions because it is already busy handling a signal pulse. Dead time has been thoroughly investigated by Rameb~ick et al. (2001) and practical ways of correcting for dead time has been handled by Nelms et al. (2001) and Appelblad & Baxter (2000). The trigger level, or discrimination level, is the signal or peak height needed to trigger a count. It is important to set the trigger or discrimination level so that double counting caused by ringing is avoided. For more information look at Hunter & Gray (1993). The plateau is the flat region of the characteristic count rate vs. high voltage curve where a change in the output count rate is small relative to variations of the high voltage of the multiplier. An example of the sensitivity of the plateau can be seen in Figure 37.5. In Figure 37.5 we see change of K-factor versus SEM high voltage (circles) and the plateau curve, signal counts for 187Re versus SEM high voltage (rectangles). The multiplier used for obtaining Figure 37.5 was a nine month old MC-Z-19-TRITON from MasCom, Germany. The instrumental settings were similar to the settings in Rameb/ick et al. (2001). Please note that the data points in Figure 37.5 were obtained in random order. Plateau checks are a convenient tool to determine the optimum high voltage setting for the SEM detector system and it is obvious from Figure 37.5 that the high voltage need to be stable to avoid drifts in the K-factor.
Introduction to Isotope Dilution Mass Spectrometry (IDMS)
833
Figure 37.5 - SEM high voltage vs. mass fractionation and ion counting (H. Rameb~ick,
pers. com.).
37.8 U n c e r t a i n t y B u d g e t
Because of its transparency IDMS can easily produce traceable results. The way to demonstrate traceability is to use an uncertainty budget, where all parameters influencing the final result are presented together with their uncertainties. Table 37.2 is an example of an uncertainty budget. In this case it is an IDMS measurement of the amount content of C1 in a 37C1 enriched material. In this measurement (M. Ostermann, pers. com.) the reference material IRMM-641, which has a natural isotopic composition, was used as spike. The model equation used is described in equation [37.20]. 9my. [ R y - K . r B ] 1 + K'rx35/37 9 c x - cy mx [ K . r B - K . r x ] 1 + Ry35/37
where
K -
Ry ry
[37.20]
In this example Cy and Ry are taken from the IRMM-641 certificate and rB, rx35/37, ry35/37, mx and my are measured. An uncertainty budget must contain the model equation used in the calculation of the result as well as the input quantities and their uncertainties. The uncertainty is calculated using the concept of error propagation and there are a couple of ways to do this. There is the simple to use spreadsheet approach described by Kragten (1994) but for larger model equations the spreadsheet can get difficult to manage. The help can then be special uncertainty software. There are a few around and it can be the way to go. There is of course the strict mathematical approach, which is used by people who either love partial derivatives and have nothing better to do, or old school know-it-alls. The level of detail in modern IDMS equation modeling, with replicates and blanks, is way beyond paper and pen uncertainty
834
Chapter 37 - M. Berglund
budgeting so save Table 37.2- Uncertainty budget yourself the trouble Value and use one of the Quantity other options. GUM1 was published in 1.13624 1995 and is the referrB 2.50.10-5 mol/g Cy ence work on estiRy 3.1272 mating and reporting rx35/ 37 0.018321 uncertainty. A more ry35/37 3.11291 2.51340 g easy to read guide on mx 1.02090 g uncertainty is the my 4.434-10-6 mol/g EURACHEM guide2 cx which is made for chemists by chemists.
Standard Uncertainty 0.0014 5.50.10-9 mol/g 0.0041 0.00037 0.0031 0.00029 g 0.00012 g 0.012-10-6mol/g
Contribution to total uncertainty 48.1% 0.7% 12.7% 6.5% 31.6% 0.2% 0.2%
Acknowledgements The author would like to acknowledge discussions and exchange of ideas with previous and present colleagues at the isotope measurement unit at IRMM and especially Henrik Rameb~ick, Markus Ostermann, Philip Taylor and Paul De Bi6vre. Helpful comments on the manuscript were made by Bob Loss and Wolfgang Frech.
1. GUM, Guide to the expression of uncertainty in measurement, ISO 1993, ISBN 92-67-10188-9 2. Quantifying Uncertainty in Analytical Measurement, 2nd edition (2000) EURAcHEM: http://www.eurachem.bam.de
PART 2 Calibration and Correction Procedures, Standards, Mass Spectrometers, Experimental Isotope Fractionation Determination and General Information
This Page Intentionally Left Blank
Handbook of Stable Isotope Analytical Techniques, Volume 1 P.A. de Groot (Editor) 9 2004 Elsevier B.V. All fights reserved.
CHAPTER 38 Mass Spectrometer Hardware for Analyzing Stable Isotope Ratios Willi A. Brand Max-Planck-Institute for Biogeochemistry, PO Box 100164, 07701 Jena, Germany e-mail: [email protected]
Abstract Mass spectrometers and sample preparation techniques for stable isotope ratio measurements, originally developed and used by a small group of scientists, are now used in a wide range of fields. Instruments today are typically acquired from a manufacturer rather than being custom built in the laboratory, as was once the case. In order to consistently generate measurements of high precision and reliability, an extensive knowledge of instrumental effects and their underlying causes is required. This contribution attempts to fill in the gaps that often characterize the instrumental knowledge of relative newcomers to the field.
38.1 Introduction Since the invention of mass spectrometry in 1910 by J.J. Thomson in the Cavendish laboratories in Cambridge('parabola spectrograph'; Thompson, 1910), this technique has provided a wealth of information about the microscopic world of atoms, molecules and ions. One of the first discoveries was the existence of stable isotopes, which were first seen in 1912 in neon (masses 20 and 22, with respective abundances of 91% and 9%; Thomson, 1913). Following this early work, F.W. Aston in the same laboratory set up a new instrument for which he coined the term 'mass spectrograph' which he used for checking almost all of the elements for the existence of isotopes. Aston not only confirmed the neon findings, he also discovered 21Ne which has only a 0.3 atom% abundance. During his scientific career, Aston discovered 212 out of the total 287 naturally occurring isotopes (Aston, 1942). This work brought new order into the periodic table of the elements which had previously been troubled by irregularities between atomic weight and chemical properties of the elements. Aston showed that the isotopic masses are not simple integral masses of a basic nucleon but rather that there is a mass defect that is related to the binding energy of the nuclei. Both J.J. Thomson and F.W. Aston were awarded Nobel Prizes for their achievements (Physics in 1906 and Chemistry in 1922, respectively). In general, a mass spectrometer is used to make a quantitative assessment of the contents of a given sample. The quality of the analysis thus depends on the ability of the mass spectrometer to detect all components of a sample with the same constant
8 36
Chapter 3 8 - W.A. Brand
sensitivity, irrespective of the complexity and chemical nature of the sample.This ideal mass spectrometer does not exist. Instead, the contents of a given sample have to be transformed into something which can be manipulated, separated and detected. In mass spectrometry, ions serve this purpose. The ability to quantitate the contents of a given sample is facilitated if sample complexity is reduced through separation of the individual chemical components prior to the measurement. This principle has led to the extensive use of separation devices (chromatographs) combined with mass spectrometers (as detectors) in chemical analysis. This combination has more recently been used for determination of the stable isotope ratios of the bio-elements (C,N,O,S and H)(Brand, 1996), alongside the more familiar method of isotope ratio measurement by high precision comparison of purified gases in the dual inlet system. A stable isotope ratio mass spectrometer consists of an inlet system, an ion source, an analyzer for ion separation, and a detector for ion registration. The inlet system is designed to handle pure gases, principally CO2, N2, H2, and SO2 but also others such as 02, N20, CO, CH3C1, SF6, CF4, and SiF4. Neutral molecules from the inlet system are introduced into the ion source, where they are ionized via electron impact and accelerated to several kilovolts, and then separated by a magnetic field and detected by Faraday cups positioned along the image plane of the mass spectrometer (Nier, 1940). The principles guiding the design and operation of each of these individual sections of the mass spectrometer are described and discussed in sequence.
38.2 Inlet System Design Inlet systems for gas isotope mass spectrometers are rather simple and clean devices consisting of valves, pipes, capillaries, connectors, and gauges. Home made inlet systems are often made of glass, but commercially available inlet systems are mostly designed from stainless steel components that have no cavities. All components and surfaces are carefully selected for maximum inertness towards the gases to be analyzed. The materials used as components of the valves deserve special attention. The highest quality valves are of 'all-metal' design, with all wetted surfaces made either from stainless steel (the body and membranes) or from gold (the gaskets or seals and the valve seat). The heart of the inlet system is the "Changeover Valve' (Figure 38.1). It was first described in 1947 by B.F. Murphey, who was studying thermal diffusion in gases. The Changeover Valve allows the inlet system to alternately switch within a couple of seconds between two gases which enter in turn into a vacuum chamber (e.g. the mass spectrometer). The gases are fed from reservoirs to the Changeover Valve by capillaries of around 0.1 mm i.d. and about lm in length with crimps for adjusting gas flows at their ends (Honig, 1945; Nier, 1947; Halsted & Nier, 1950). While one gas flows to the vacuum chamber, the other is directed to a vacuum waste pump so that flow through the capillaries is never interrupted. Without capillaries, a flow directly from a reservoir through an orifice into the mass spectrometer would be a direct effusion into vacuum, which would result in a change in isotopic composition over time. The forward flow of gas in the viscous flow regime through the long capillaries prevents the isotopic diffusion profile from penetrating from the crimp back into the sample reser-
Mass Spectrometer Hardware for Analyzing Stable Isotope Ratios 837
Figure 38.1 - Essential components of a 'Dual Inlet System' for gas isotope ratio mass spectrometry. For clarity the pumping infrastructure for the variable volume reservoirs and for the pipework has been omitted from the figure. The 'Changeover Valve' is an arrangement of four valves that allows gas from one reservoir to flow to the mass spectrometer while the other goes to waste. The design of the changeover valve deserves special attention in order to minimize cross talk between the gases.
838
C h a p t e r 38 - W . A . B r a n d
voir (Halsted & Nier, 1950; Habfast, 1997). In 1950, C.R. McKinney et al. applied the Changeover Valve principle to isotope ratio measurements. With their system, McKinney and coworkers were able to measure the stable isotopes of oxygen in 02 and of both carbon and oxygen in CO2 with a precision of about 0.1 per mill (g-notation1). To achieve such high precision, instrumental drifts occurring during measurement need to cancel almost quantitatively. To achieve this goal, the gas reservoirs on either side of the Changeover Valve are normally stainless steel bellows (formerly, mercury pistons) that allow the ion current signals of the two gases to be precisely balanced. Any non-linearity, temperature dependence of electronic components, or changes in sensitivity of the ion source or the magnetic field thus tend to cancel. By comparing the two gases several times within minutes it was possible for McKinney et al. to achieve the reported high precision. The McKinney instrument provided the basis for the "classical" procedure for high precision stable isotope ratio measurements. Its principles have survived for 50 years with little change and they will provide the basis for ultimate precision isotope ratio determination into the foreseeable future. The smallest amount of sample that can be analyzed using the dual inlet system is limited by the requirement to maintain viscous flow conditions. As a rule of thumb, the mean free path of a gas molecule should not exceed 1/10 th of the capillary dimensions. With the capillary dimensions of 0.1 mm i.d., the lower pressure limit for viscous flow and thus accurate measurement is about 15 to 20 mbar. When trying to reduce sample size, it is necessary to concentrate the gas of interest into a small volume in front of the capillary. For practical reasons, such a volume cannot be made much smaller than 250 gl. For condensable gases, it is shaped into a cold finger to be operated as a cryotrap at liquid nitrogen temperature under molecular flow conditions. Using the ideal gas law, the product of pressure and volume yields the smallest sample amount that can be accurately analyzed in a microvolume inlet system to be about 5 bar~l or 220 nmol of clean gas. Because real life samples rarely are the clean gas species used in the dual inlet system, each sample, be it a carbonate, a water sample, a lentil or a piece of tree ring, must be converted into the required simple gaseous form prior to analysis. There is a wide variety of specialized sample conversion and inlet systems including manually operated devices whose output must be manually introduced into the inlet reservoir and automated devices that deliver the final product gas directly to the inlet system under computer control. Other chapters in this book cover the various forms and experimental challenges of sample preparation for high precision isotope ratio determination.
1. ~ [%o1= (Rsa/ Rref - 1) 91000 {for 13C: RSa= 13C/12Cion current ratio of sample gas}
[38.1]
839
Mass Spectrometer Hardware for Analyzing Stable Isotope Ratios
38.3 The Ion Source: Electron Impact Ion Production
Wishful thinking:
If we could only sit and watch the molecules directly distinguishing their different weight through some colorful property, we could calculate isotope ratios just by counting1. This would, however, be a tedious and time consuming task because of the large number of particles required for high precision. 38.3.1 Basic principles
Sample molecules enter the ion source of the mass spectrometer from the inlet system in gaseous form. Here, some of them are ionized by b o m b a r d m e n t with electrons
(Electron Impact, El): M+e-~M+,
+2e -
[38.2]
The efficiency of this process determines the sensitivity of the mass spectrometer. It depends on the ionization cross section, the number of electrons, and the number of molecules presenting themselves to be ionized. Following ionization, the M + ~ molecular radical cation can further fragment into several pieces (e.g. CO2 +o ~ CO + + O.), depending on the internal energy the ion has acquired during the ionization process and the possible reaction pathways. The result of such unimolecular reactions is the mass spectrum of a chemical compound. More specifically, the fragments that form in the ion source within about one microsecond following ionization comprise the mass spectrum. Later reactions give rise to what we refer to as 'metastable' ions. As an example, there is a broad peak at mass 17.8. CO2 +o molecular ions which were accelerated as mass 44 but decayed to CO+ (mass 28) in front of the magnet arrive at the detector plane at mass position m* - m22 / m l . Here, m* is the apparent mass position (17.8), m2 and ml are the mass positions of the daughter (28) and parent ion (44), respectively. 38.3.2 Ion Source Schematics Figure 38.2 is a schematic representation of an electron impact ion source. Electrons are released from a hot filament made from tungsten, rhenium or thoriated iridium and accelerated by electrostatic potentials to an energy between 50 and 150 eV before entering the ionization box. Their velocity, v, can be calculated according to"
v-
,,/2eU/m
[38.3]
where e - elementary charge, U - accelerating potential, m - mass of the particle. The velocity of 100eV electrons is about 6 ~ 108 cm/s. Thus they traverse the ion box in about 2 nanoseconds. The molecules appear virtually motionless because they 1. To avoid confusion: Ion counting is also a special technique using fast secondary electron amplifiers with amplification up to 108 together with time and threshold discrimination techniques. It indeed is a powerful tool e.g. for measuring small abundances of isotopes in thermal ionization mass spectrometry.
840
Chapter 38 - W.A. Brand
Figure 38.2- Schematic layout of an Electron Impact (EI) ion source for gas isotope ratio mass spectrometry. The insulating spacers that also provide an enclosure for the whole source are omitted for clarity.
are moving in the ion source at thermal velocities of only about 3 9 104 cm/s. A homogeneous magnetic field of 100 to 500 Gauss is used to keep the electrons on a spiral path (to increase the ionization probability) through the ionization box effectively confining the ionization region to a diameter of <1 millimeter. At the end of the ion box, the electrons hit the electron collector or "trap", sometimes after a gentle post-acceleration. The electron current at the trap is measured and is kept constant by the emission regulator circuitry. The properties of the emission regulator are important when making peak area measurements which are converted into abundance measurements of weight %C and%N). Ions are extracted from the ion box perpendicular to the direction of the electron beam by a field generated by either a repeller potential inside the ion box, an outside extraction lens, or a combination of the two. In order to minimize the translational energy spread of ions entering the ion optics of the mass spectrometer, the voltage
Mass Spectrometer Hardware for Analyzing Stable Isotope Ratios
841
drop across the ionization region should not be too large. On the other hand, the potential gradient across the ionization region should not be too small either because ion-molecule reactions need to be suppressed by a fast acceleration out of the ion box. Ions are rather reactive chemical species. They like to react, for instance, with hydrogen-bearing molecules to form protonated MH + ions. This is particularly important in isotopic analysis because these ions fall onto the same mass positions as the isotopic species of interest at the m + l mass position. For instance, an isobaric interference from 12C1602H+ iS collected at m / z 45, which should be 13C1602+ and 12C160170 exclusively. Likewise, for hydrogen isotope ratios, H3 + ions (mass 3) hit the detector plane exactly where the Faraday cup for collecting HD + ions is mounted. After extraction from the ion box, the ions are further accelerated and shaped into a beam which passes through the ion exit slit into the analyzer. Mass 44 ions with 5 keV translational energy will travel through the analyzer with a speed of 1.5 9 107 cm/s, hitting the detector after just a few microseconds.
38.3.3 Ionization efficiency In modern isotope ratio mass spectrometers, the electron emission current is typically around 1 mA or 6 9 1015 electrons per second moving through a cross sectional area of about 1 mm2. With a typical ionization cross section (Kiser, 1965; Platzner, 1997) of about 3 9 10-16 cm2 or 3/~2 and the given geometry of the ionization region, the number of ions produced per second and thus the m a x i m u m sensitivity can be calculated (example)" Length of ionization region Cross section of ionization region Density of gas at 10 -7 mbar: ( f l o w - i nmol/sec) Total ionization cross section in volume: (2.7 ~ 107 particles x 3 ~ 10 -14 m m 2)
10mm =1 mm2
=
= 2.7 ~ 1012 particles / 1 = 2.7 9 107 particles in ion volume = 8 9 10 -7 m m 2
Thus, out of a total cross sectional area for the electrons of 1 mm2, only 8 9 10 -7 mm2 will typically be effective for ionization producing an ion current of 8 9 10-10 A from a I mA electron emission current. In order to improve on gas utilization, the pressure in the ion source is raised by using a 'closed source' design. The sole openings in such a source are the entrance and exit holes for the electrons and the exit slit for the ions. The rest of the source is gas tight which enhances the number of neutrals in the ion volume by a factor of about 100 resulting in an ion current of 8 ~ 10-8 A or 80 nA in the example given. In practice, not all ions produced enter the mass spectrometer and reach the detector. The burn marks visible on the ion exit slit and on the aperture in the flight tube are traces of those ions that do not make it into the detector. Also, molecular ions that fragment in the ion source are lost for the isotopic analysis. Overall, an allowance for roughly 50%
842
Chapter 38 - W.A. Brand
loss must be made. Concluding the example: Out of the total flow of about I nmol gas per sec, an ion current of 40 nA is created. This corresponds to 2.5 ~ 10-11 ions from 6 ~ 1014 molecules or 2400 molecules are needed for one ion at the detector. Typical sensitivity specifications of commercially available stable isotope mass spectrometers are 1000 to 2000 molecules per ion. 38.3.4 Problem areas in ion source design
The major objectives in isotope ratio mass spectrometry include: a linear relation between the ion current intensities and the measured ratios - no memory between subsequent introductions of sample and reference gases in the mass spectrometer. Memory sources include gas adsorbtion on welds, copper gaskets (SO2) and polymer gaskets (CO2, H2) - chemical inertness of the hot filament - highly stable ion currents over a time range much longer than required for the measurement of a single sample
-
Some of these requirements are challenging and not all of them have completely met the satisfaction of the analyst. Some are very difficult to tackle. 1. Deviations from linear behavior can come from chemical or physical effects in the ion source. The major source of chemical non-linearity is the appearance of protonated species at the same mass position as the ion of interest due to incomplete suppression of ion-molecule reactions. The keys to avoiding this problem are rapid extraction of ions from the ion volume and utmost cleanliness of the samples. Where unavoidable, exact quantitative measurement of the effect is necessary in order to correct for it (e.g. H3 + correction for deuterium measurement). The presence of the collimating magnetic field in the ion source can lead to non-linear ratio response. If it is too strong, the field can lead to dispersion of the ions at the entrance slit prior to entry into the analyzer. Changing the number of ions influences the space charge in the ion source and thus the extraction conditions. As a consequence, the amount of pre-dispersion can vary with signal height. This physical effect scales with the relative mass difference, e.g. for CO2 the 46/44 ratio shows a larger deviation than the 45/44 ratio. It can only be avoided by careful design of the ion optical components of the ion source. This is a major challenge for designers of mass spectrometers. 2. It is important to minimize the time required for gas exchange. The very high precision required for a number of investigations makes fast comparisons between sample and standard a necessity. During gas exchange the ion currents do not contribute to the measured signal. Possible changes in instrument response can be best compensated for by fast gas exchange. On the other hand, any gas left over after an analysis will mix with the new sample and will thus adulterate the measured isotope ratio. The measured difference in isotopic composition will be smaller than the true difference ('eta effect').
Mass SpectrometerHardware for Analyzing Stable Isotope Ratios
843
Some gases exchange easily (N2, N20, H2), while more polar molecules are more "sticky" (CO2, SO2) due to enhanced surface activity. For sulfur isotope ratio measurements, some mass spectrometers have a window in the closed ion source which can be opened and closed from the outside to enhance the pumping speed. The quality of the wetted surfaces, the avoidance of dead volumes that cannot be efficiently pumped and thus act as virtual leaks, and the minimization of surface area that is sputtered by charged particles must all be taken into consideration to ensure proper gas exchange behavior. Similar considerations also apply to the changeover valve and the connection to the ion source. The latter should be short and wide, i.d. 4 m m or larger. The ion source should be chemically inert with respect to the analytical gases. Welds for instance have been identified as contributors to chemical memory and reactivity. Thus, wetted surfaces should be weld-free wherever possible. Several gasket materials have been identified as significant contributors to chemical memory (e.g. Cu gaskets in the analysis of SO2). They can also act as virtual leaks (e.g. polymers with microporosity). 3. The hot filament can alter the gas composition in the ion source. Reaction of CO2 with the hot tungsten for instance will form carbides that can be partly burned off when an 02 pulse enters the ion source. The CO2 released by this reaction typically has a rather positive 613C value as can be seen in chromatographic runs when oxygen enters the ion source as a pulse. In addition to carbide build up, tungsten oxide is formed at the filament surface by interaction with CO2, 02 or H20. When the instrument is switched from carbon dioxide to hydrogen measurement, it is observed that ion currents and isotope ratios need some time (up to one hour) to stabilize. This is a consequence of reaction of hydrogen gas with the filament. Hydrogen reacts with the oxide layers on the filament to produce traces of water that temporarily give rise to extra H3 + ions at mass 3. Once the filament is conditioned for the gas to be measured, a steady state is reached with no further interference with the isotope ratio determination. 4. In order to reliably measure isotope ratios, ion currents need to be highly stable. More specifically, the conditions in the ion source need to remain optimized for ion production and high linearity over long periods of time. The stability of ion source conditions is governed by several design aspects: - the electrostatic potentials applied to the different lenses need to be stable to about 2 9 10-5.
- The accelerating potential must have a similar stability. Any ripple on top of this voltage leads to a reduction in the steepness of the peak flanks. - Insulating surface layers that could lead to a charging up of surfaces in the ion source need to be strictly avoided. If such surface layers build up over time, it is necessary to clean the ion source.
844
Chapter 38 - W.A. Brand
38.4 Separation and Detection of Ions in the Mass Spectrometer 38.4.1 Magnetic sectors High precision isotope ratio mass spectrometers employ magnetic sectors for the separation of ions almost exclusively. Both permanent magnets and electromagnets are used in commercial instrumentation. The ions normally possess a kinetic energy between 2.5 and 10 keV, varying from instrument to instrument. Smaller instruments have lower acceleration potentials and smaller magnets. Larger instruments with higher accelerating voltage have higher sensitivities, higher resolutions and better peak shapes than instruments with lower accelerating voltages, all other things being equal. Ions entering a homogeneous magnetic field are deflected perpendicular to their flight direction and perpendicular to the magnetic field according to the Lorentzian rule. The result is a circular path with the radius"
r - 1/B 9d ( 2 m U / z e )
[38.4]
with B being the magnetic field, e is the elementary charge (= 1.6 9 1 0 -19 Coulomb), z is the number of charges, U is the accelerating potential and m is the mass of the ion. As an example, a singly charged ion of 44 amu (= atomic mass units or Daltons) that has been accelerated to 5 keV will describe a radius of 13.5 cm when travelling through a homogeneous magnetic field of 0.5 Tesla. It should be noted that mass and translational energy are equivalent in equation [38.4]. Any inhomogeneity of the kinetic energy of the ion beam will lead to a broadening of the ion image at the detector. This is especially important in light of the discussion above about the necessity to suppress ion-molecule reactions by applying a high draw out field to the ionization chamber. The effect is more severe for smaller mass spectrometers since the relative energy spread AE/U is larger for lower acceleration potentials U. For high precision molecular mass determination in organic mass spectrometry, double focusing arrangements that reverse the dispersion due to the energy spread are commonly used (Mattauch & Herzog, 1934:; Dempster, 1935). Early design isotope ratio mass spectrometers had the ions entering the magnetic field perpendicular to the field boundaries (Figure 38.3). The magnet angle (hence the term sector) was either 60 ~ a design based on the early mass spectrometer proposed by A.O. Nier (Nier, 1940), or 90 ~ which allowed similar ion optical properties (Herzog, 1936) to be realized in an instrument with a smaller footprint. These mass spectrometers were characterized by single direction focusing properties and 1:1 imaging. A magnetic sector can be treated like an optical prism in geometrical optics: Monoenergetic ions of different mass are dispersed through the magnetic field. Light ions follow a path with a small radius whereas heavier ions describe circular paths with larger radii. Ions with identical mass coming from a focal point with a certain lateral spread c~will be focused again after exiting the magnet. The focal points of different masses lie on a focal plane. When entrance and exit drift lengths are identical, the image at the detector plane will be the same size as the original beam at the entrance slit.
Mass Spectrometer Hardware for Analyzing Stable Isotope Ratios
845
Unfortunately, the focusing properties apply only in the x-direction, the direction of the deflection of the ion beam. In the y-direction (along the height of the slit) no focusing takes place (Figure 38.3). As a consequence, the early mass spectrometers had a transmission considerably smaller than unity. Other deleterious effects of not having y-focusing include ions striking the flight tube in the y-direction, which leads to an accumulation of static charge on the surface of the flight tube which can deteriorate peak shapes of the ion beams; ions that are partially reflected can be deflected further in an unpredictable manner. They may also enter the wrong detectors. Secondary electrons released in scattering events can reach the detectors where they create false
Figure 38.3 - Schematic representation of a 90~ magnetic sector mass spectrometer for isotope ratio measurements. In case of equal length of l'm and l"m the image width at the focal plane is identical to the width of the entrance slit. However, in y-direction no focusing takes place.
846
Chapter 38 - W.A. Brand
currents. A n additional last d i s a d v a n t a g e is that the detector entrances in the y-direction are smaller t h a n the ion beams, so that b e a m fluctuations can lead to r e d u c e d precision. All of these limitations were largely o v e r c o m e w i t h the i n t r o d u c t i o n of " d o u b l e direction" or stigmatic focusing, first p r o p o s e d in 1951 (Cross, 1951)(Figure 38.4), in w h i c h the ions enter a n d exit the m a g n e t at an offset of 26.5 ~ from n o r m a l entry. In this case, the fringing field of the m a g n e t at the pole gap acts as a focusing e l e m e n t in the y-direction leading to coincidence of the focus in y- a n d x-directions. The first commercial isotope ratio i n s t r u m e n t to e m p l o y stigmatic focusing w a s the MAT 250 intro-
Figure 38.4 - Illustration of the effect of 'Stigmatic Focusing'. A non-normal entry and exit of the ions of 26.5~ provides double direction focusing at the focal plane. The fringing fields at the magnet gap act as focusing elements in y-direction. No ions are lost to the walls, so transmission of close to 100% is possible.
Mass Spectrometer Hardware for Analyzing Stable Isotope Ratios
847
duced in 1977; today, almost all commercial instruments have incorporated it in some way. The choice between a permanent magnet and an electromagnet needs to be discussed in terms of the limitations that a particular choice imposes on the analytical performance. From a theoretical (i.e. ion optical) point of view, both types of magnets are identical. However, permanent magnets often do not reach the same homogeneity as electromagnets. In practice, the difference between the two types of magnets shows up in the way a particular mass is selected. The field of an electromagnet may be changed by applying a variable current to the coils thus allowing a scan of the complete mass spectrum (equation [38-4]). Because the field of a permanent magnet cannot be changed, mass selection is made by altering the accelerating voltage. Usually, the magnetic field strength of a permanent magnet mass spectrometer is selected to ensure that the isotopes of nitrogen can be measured with an accelerating voltage close to the maximum (e.g. 5 kV). Then, the CO2 + isotope triplet can be brought into focus with 28/4:4: ~ 5kV - 3.18 kV. For SO2, the voltage must be lowered even further to 2.19 kV. As a consequence of reducing the accelerating voltage for mass selection, the sensitivity and the focusing conditions of the ion source change with mass. Because the isotopes of nitrogen are measured at the maximum accelerating voltage, ions with masses lower than 28 cannot be measured in such a system. This is quite unfortunate because water, mass 18, is a potent source for protonation in the ion source and should be routinely monitored (Leckrone & Hayes, 1998). This is of particular importance for the isotopic characterization of water. The isotopes of hydrogen are measured on H2, hydrogen gas (masses 2 and 3), on a 'spur' which can not access mass 18 either. Incomplete conversion can lead to contamination of the ion source with water and the formation of extra H3 +. In the case of CO2, water contributes to an isobaric interference at mass 45 through formation of CO2H +. When on-line high temperature pyrolysis is used for sample conversion (Burgoyne & Hayes, 1998) mass 15 is indicative of CH4 formation. Methane is also a very potent source for protonation. Therefore, it needs to be monitored routinely in order to avoid its presence in the ion source.
38.4.2 Multiple Faraday Cup Detectors The use of multiple detectors to simultaneously monitor and integrate the ion currents of interest was introduced by A.O. Nier in 1947. Two ion currents, e.g. masses 44 and 45 from CO2, hit a collector plate mounted behind a grounded slit and a pair of secondary electron suppressor shields. The advantage of the simultaneous measurement with two separate amplifiers is that fluctuations of the ion currents due to temperature changes, electron beam instability etc. cancel completely. The magnet and the accelerating voltage remain constant. No peak jumping is required, which eliminates the corresponding settling times. Moreover, each detector channel can be fitted with a high ohmic resistor appropriate for the mean natural abundance of the isotope ion current of interest. This principle of static multicollection is still in use but the collector plate has been replaced by deep Faraday cups, in order to minimize false detector currents generated from secondary electrons. The interior of the Faraday cups are sometimes lined with graphite in order to minimize errors arising from reflections
848
Chapter 38 - W.A. Brand
and sputtering effects. Modern isotope ratio mass spectrometers have at least three Faraday collectors. For CO2, the mass 46 collector gathers, almost exclusively, the 180 information whereas the mass 45 collector has a contribution from both 13C and 170. In order to measure the 1 3 C / 1 2 C ratios of CO2, three collectors are necessary, because t h e 1 7 0 correction is made from the measured 8180 value and the known terrestrial relation of the two oxygen isotopes (Craig, 1957). The Faraday cups are carefully positioned along the focal plane of the mass spectrometer. Because the spacing between adjacent peaks changes with mass, and because the scale is not linear, each set of isotopes requires its own set of Faraday cups. A compromise can be employed by making the outer two cups wider in order to cover the dispersion range between the N2 and the C O 2 triplet (Figure 38.5). No obvious deterioration in performance resulting from this compromise positioning of the Faraday cups has been detected yet. Alternatively, the Faraday cups can be made moveable by mounting them on a variable support which allows the cup positions to be adjusted to the relative positions of the ions. Here, special care needs to be taken to shield the Faraday cups and their leads from stray electrons. 38.5 Instrumental effects requiring correction Despite continuing efforts to improve mass spectrometer hardware, high precision and high accuracy determination of isotope ratios from samples with terrestrial isotope abundance still requires that a number of corrections be made. 38.5.1 H3 +factor
Mass 3 forms during the analysis of H2 gas via the ion-molecule reaction: H2 +e + H2 ~ H3 + +
H-
[38.5]
From reaction [38.5] it can be seen that the amount of H3 + formed is directly proportional to the number of neutral H2 molecules and to the number of H2 +o ions, which for a given sensitivity, is also directly proportional to the number of H2 molecules in the ion source. Thus: [H3 +] - k l 9 [H2] 2
[38.6]
With the ion current ratio expressed as 3 R - {[H3+] + [HD+]} / [H2+] and substituting we have: 3R = [HD +] / [H2 +1 + k2 9[H2+]
[38.7]
with k2 being the H3 + factor. Thus, 3R is a linear function of the mass 2 ion current. By measuring 3R with varying amounts of H2 flowing into the mass spectrometer, k2 can
Mass Spectrometer Hardware for Analyzing Stable Isotope Ratios
Figure 38.5 - Schematical arrangement of Faraday cups positioned along the focal plane of an isotope ratio mass spectrometer. The two outer cups are wider than the middle cup in order to accommodate a wider dispersion range. The isotopomer triplets of CO2 (44-46), 02 (32-34), SO2 (64, 66) and N2 or CO (28-30) can be measured with the same set of Faraday cups.
849
850
Chapter 38 - W.A. Brand
be determined. Usually, k2 is about 10 p p m / n A , the units being selected to allow convenient comparison with the measured ion currents. The international standard water VSMOW has 156 ppm of deuterium, so the HD + ion current is 312 ppm of the total current. For a typical mass 2 ion current of 5 nA, the contribution of [H3 +] to the total mass 3 ion current is 50 ppm. Thus, the H3 + ion current measured for H2 from VSMOW is 16% of the true mass 3 ion current, that which represents the deuterium content of the sample. Clearly the H3 + factor must be rather constant during a series of measurements if determination of the D / H ratios is to be made with a precision <0.5 per mill. This discussion assumes that the original content of deuterium in the sample is available in the hydrogen gas for analysis. For water reduction methods, this is usually true. However, when H2 gas is equilibrated with water in the presence of a platinum catalyst (Horita, 1988), the equilibrium fractionation factor between H2 and H20 dictates that at 25~ the H2 gas will have only 25% of the deuterium concentration in the liquid. Hydrogen gas equilibrated with VSMOW thus has only 39 ppm of deuterium or 78 ppm HD (= -750 %0 !) resulting in an HD + ion current of 4 9 10 -13 A (mass 2 current 5 nA) which needs to be measured with a precision of better than 10-3. The requirement for a stable H3 + factor is consequently enhanced by a factor of four relative to analysis of H2 from water reduction. Because the H3 + factor cannot be determined with the precision required for isotope ratio determination, it is advisable to measure a series of samples with the same H3 + factor. This has the advantage that errors in the precise knowledge of the H3 + factor can be corrected for by scaling, i.e. adjusting the measured differences to the known difference of a pair of standards, in general SLAP and VSMOW (Coplen, 1988).
38.5.2 Scaling of 6 values Scaling is mandatory for hydrogen and oxygen isotope ratio determination (Coplen, 1988) mainly because of difficulties and inconsistencies in sample preparation. However, it has been shown in a recent laboratory intercomparison study (Brand & Coplen, 2001) using pure hydrogen gas (i.e., there was no sample preparation) that adjustment of measured differences to precisely known differences is also required to correct for instrumental artifacts. These can have many causes including non-perfect removal of the H3 + contribution, errors in determining the hydrogen background1, predispersion at the ion entry into the mass spectrometer etc. The effects are particularly important for D / H measurements due to the large relative mass difference between HH and HD, but the need to have laboratories intercalibrated at high levels of precision points to an increasing need for scaling of measured 6 values for gases other than H2.
1. H y d r o g e n gas dissolves in the rotary p u m p oil. It can diffuse back into the mass spectrometer due to an insufficient compression ratio of the turbo pump. This effect may vary with time.
Mass SpectrometerHardware for Analyzing Stable Isotope Ratios
851
38.5.3 Abundance Sensitivity Abundance sensitivity refers to the contribution at one mass arising from an ion current at the neighboring mass. The contribution from the mass 44 ion current to mass 45 can be treated with:
imeas -itrue + iab.sens.
[38.8]
o r I t r u e - i m e a s (l-a) with a representing the portion of the ion current that is generated through the abundance sensitivity effect.
Inserting into the h-equation [38.1] and transforming yields: 15= (Rsa -Rst) / (Rst + a) 9 1000
[38.9]
For modern isotope ratio mass spectrometers operating at pressures in the analyzer of better than 10-7 mbar, abundance sensitivity usually is <10 -5 on mass 45. Its value is influenced by the ion optical properties of the instrument. Large dispersion here means small abundance sensitivity. Careful alignment of the magnet and thus peak shapes optimized for best flank steepness help keeping the abundance sensitivity small. For an abundance sensitivity value a - 10-5 the difference between equation [38.1] and equation [38.9] amounts to 10-3 per mill and can thus be neglected safely. However, if abundance sensitivity increases by one or two orders of magnitude, as can be the case in on-line measurements of isotope ratios, it cannot be neglected (see below). 38.5.4 Linearity Correction The precision of a measurement using the changeover technique often depends upon the quality of the pressure matching between the variable bellows reservoirs. This is especially the case when the mass spectrometer is not operating in the optim u m range for linearity. This may occur when the application of a high drawout potential results in a drop in sensitivity that cannot be tolerated. Alternatively, the sample preparation may simply fail to provide the required cleanliness of the sample. It may also be caused by design deficiencies of the ion source causing physical nonlinearities. In such cases, a linearity correction can improve the results, provided the conditions are monitored carefully and are sufficiently described by a linear relationship between the major ion current and the measured isotopic ratio. When fully applied the correction is similar to the H3 + correction; the measured ratios of both sample and reference gas are corrected, being normalized to an identical major ion current before the 8 value is calculated. A simplified correction works as follows: If the measured non-linearity k of the ratio is given in %o/nA, then the correction to be applied to the final result is:
6corr = ~)meas- A 9 (isa- iref)
[38.10]
85 2
Chapter 38 - W.A. Brand
As an example, the measured 6 value is +10%o. The reference gas has been measured with a major beam intensity of 5 nA, the sample gas with 3 nA. The non-linearity of the mass spectrometer was determined as +0.1%o/nA. Then the corrected 6 value is 10.20%o.
38.5.5 Background correction In isotope ratio mass spectrometry, the background has two components: an electronic offset usually referred to as baseline and a chemical part that represents the ion current on the mass positions of interest when there is no gas added to the mass spectrometer (the blank). The treatment of the background is normally taken care of by the mass spectrometer operating software. The standard technique is to measure a background during a routine measurement by first closing the changeover valve to the mass spectrometer and then waiting an appropriate time for the gas to be p u m p e d away. After this the residual signal is measured. This signal is then stored and taken as background for the subsequent measurements until the next background measurement overrides the previous one. This standard procedure relies on a small and stable background condition and on an assumed ideal balance of ion currents between sample and reference gas. The errors associated with a false background determination are small: of the order 0.01 per mill.
38.5.6 Memory effects In general memory effects can occur during the automated or manual preparation of the measurement gases. They are particularly important when the sample material is water or carbonate (especially when the carbonates are reacted in a common acid vessel). But when the samples differ considerably from natural isotopic abundance, i.e. in tracer studies, almost every preparation is prone to memory effects. Each sample preparation method thus needs to be checked carefully and quantitatively for such artifacts. If necessary, a mass balance correction can be applied that assumes a certain percentage of the measured total ion current imeas is due to the previous sample: 6Sa = (~)meas 9 imeas - ~)mem " imem)/(imeas - imem)
[38.11]
proportion of memory signal, e.g. for a memory of 2%, imem =0.02 9 imeas. Equation [38.11] assumes that the samples are similar in size, otherwise imem needs to be evaluated from the integrated ion current of the previous sample. w i t h imem -
In practice, memory effects are tested by measuring two samples of sufficiently different isotopic composition in series, e.g. 5 times sample a followed by 5 times sample b, and watching the transition between them. Applying equation [38.11] should give a quantitative correction to the data. If necessary, several of the previous samples need to be considered.
853
Mass Spectrometer Hardware for Analyzing Stable Isotope Ratios
38.5.7 Isobaric interferences Isobaric interferences are ion currents in Faraday cups that belong to ionic species other than those being examined. The most prominent examples are the aforementioned correction for the 170 moiety at mass 45 of CO2 and the correction for the H3 + contribution to the mass 3 ion current. Other isobaric interferences and contributions include: - 170 and 13C on mass 46 of CO2 - CO + interference on mass 28 when measuring nitrogen - N2 interference on mass 28 when measuring 180 on line on the CO masses 30 and 28 - N20 interference on the masses 44, 45 and 46, and - 180 and 170 contributions on masses 66 and 65 of SO2. The corrections for interferences that are due to another isotopic form of the sample gas have been discussed in the literature1. A detailed discussion is not presented here. 38.6 Instrumental
aspects
of isotope
ratio monitoring
('irm')
Isotope ratio monitoring ('irm')(Matthews & Hayes, 1978) is a relatively recent innovation in isotope ratio mass spectrometry (Brand, 1996) which does not require the use of the dual inlet system or the Changeover Valve. This is possible because the requirement for viscous flow is achieved by transporting the sample gas into the mass spectrometer entrained in a stream of helium carrier gas. Isotope ratio monitoring is ideally suited for the coupling of chromatographic techniques with isotope ratio mass spectrometers. The sample size required for an isotope ratio measurement is dramatically reduced, with reported lower limits in the low picomolar range (Merritt & Hayes, 1994). The traditional dual inlet system approach which served isotope mass spectrometers well for many years can not be adapted to handle on-line chromatography for a number of reasons" - Ion currents are measured in the order in which they emerge from a GC column, without significant capability of modifying their intensities relative to a reference gas. A linear response of the entire mass spectrometer system is therefore of utmost importance. - The time for measurement of the isotopic signals is restricted by the width of the chromatographic peak. For good gas chromatography (which means sharply defined narrow peaks), this can mean less than 5 seconds. - All signal intensities vary in time. The mass spectrometer must be capable of handling transient signals (Brand, 1998) acquired on multiple channels. - It must be possible to measure the analytical ion currents in a large surplus of He carrier gas without loss in precision for the isotope ratio determination. This is of particular importance for hydrogen isotope measurements. - Absolute sensitivity is much more important than with the dual inlet system. Standard gas chromatographic techniques with capillary columns have capacity limits that require high sample utilization. Because it is the total number of ions containing 1. For a recent comprehensive discussion see: IAEA-TECDOC-825, 'Reference and intercomparison materials for stable isotopes of light elements', IAEA, Vienna 1995, ISSN 1011-4289.
854
Chapter 38 - W.A. Brand
the minor isotope that determines the statistical limit of precision (Merritt & Hayes, 1994), and because sample sizes required for good chromatography are significantly smaller than for the dual inlet system, it is important to achieve the highest level of sample usage possible, within the constraints of linear response. - Chromatography not only separates different chemical species. It even separates at the level of isotopomers, i.e. molecules with identical chemical composition but different isotopic content (Gunter & Gleason, 1971; Brand, 1996). For carbon, the 13Cbearing compounds generally precede the 12C only compounds by 50 to 300 msec. This isotopic separation within the GC column means that the isotopic composition of a compound will vary across the peak after elution, which has extremely significant consequences for data acquisition, background determination, and data reduction. - The long term stability is important, at least for the time it takes for the compounds of interest to elute from the chromatographic column. Often, the chromatograms have a huge number of peaks, making the proper timing of reference gas pulses difficult. In order to cope with the new requirements, a new generation of mass spectrometers has been designed with a number of features that specifically address the challenges of isotope ratio monitoring. Because abundance sensitivity becomes an issue due to the high load of He carrier gas, differential pumping1 has been introduced to alleviate this problem considerably. A special problem in measuring hydrogen isotope ratios is caused by tailing of large mass 4 (He + 9 onto the mass 3 channel. This tail is so large that it saturates the HD channel. The problem has been solved only recently by measuring the hydrogen ion currents with a large dispersion (Prosser & Scrimgeour, 1995) to increase the abundance sensitivity and, at the same time, use an energy discrimination filter in order to prevent ions of lower than nominal energy from entering the mass 3 Faraday cup (Hilkert et al., 1999). The fast flow of data and the fact that the isotopomers are separated in time through chromatography means that peaks cannot be sampled partially without sacrificing precision and accuracy. Rather, uninterrupted, continuous integration of all ion signals over the entire chromatographic peak is required. Isotope ratio monitoring mass spectrometers have the means for handling such transient signals in a quantitative manner. Both sensitivity and linearity need to be optimized simultaneously. Sensitivity should be better than 2000 molecules per ion reaching the detector and, at the same time, linearity should be better than 0.1%o/nA.
Contrary to the dual inlet system, standardization in isotope ratio monitoring measurements should be done exclusively using the principle of 'Identical Treatment of reference and samle material' or "IT Principle'. Mostly, isotopic referencing is made with a co-injected peak 1. Differential pumping is achieved by dividing the recipient into two separate volumes individually pumped by two high vacuum pumps. The opening for having ions pass between the volumes should be made as small as possible in order to achieve a high difference in working pressure. Typically,the pressure in the flight tube and detector section should be a factor of 10 lower than the pressure in the ion source region.
Mass Spectrometer Hardware for Analyzing Stable Isotope Ratios
855
of standard gas (Merritt et al., 1994). However, the analytical history of the sample is more complex. In order to avoid systematic errors, isotope reference materials must be run frequently between sample runs thus rendering the standard gas pulses a mediator between the different chromatograms, including the reference runs. Alternatively, reference compounds may be added to the sample solution, provided there is no chromatographic interference when the reference peak elutes (internal referencing). 38.7 Statistical limits to precision Although the number of ions sampled for isotope ratio evaluation is larger than state-of-the-art counting techniques are able to resolve, the basic process still is that of counting of particles. The counting of ions follows the laws of Poisson statistics (Matthews & Hayes, 1978; Merritt & Hayes, 1994) where the limit of precision is given by" lo-
1/d(N)
[38.12]
with N being the number of ions counted or sampled. As an example, a peak of 10 sec width and 10 nA maximum intensity has a total charge of about 50 nAsec or 3 ~ 1011 ions. Out of those, about 3 ~ 109 ions contain the 13C isotopic information. Thus, such peak can be measured with a precision of 2 9 10-5 or 0.02%o. A reference peak suffers from similar limitations, and also the major ion beams play a small role. Therefore, the statistical limit is about 0.05%0, rather close to the typical goal of the measurement of about 0.1 per mill.
38.8 Conversion equations for isotope ratio reporting Isotope ratios are generally expressed as deviations from a reference value (equation [38.1]), not as absolute ratios. The reason behind this practice is that relative measurements can be made rather precisely with moderate efforts whereas absolute measurements are much more difficult to make. Mostly, deviations from international standards are known with a precision that exceeds the absolute knowledge of the isotope abundance ratio of the international standard by more than one order of magnitude. In practice, measurements are made against a working or laboratory reference that has been calibrated carefully against an international reference material. Thus, the measured delta values need to be converted to another scale before reporting. From equation [38.1] it can be shown that ~)Sa / St - ~)Sa / WS + 8WS / St +
10-3(~)Sa/ WS " 6WS / St)
[38.13]
with WS - working standard, St denoting the international standard material and Sa the measured sample. In a typical measurement series reference compounds are processed the same way as the sample ('IT Principle', see above). In such cases, the term 6ws/st is measured in its reverse form, 6st/ws. It may be easily converted using: ~)BA - - (1 / (~AB +
10-3)-1
[38.14]
856
Chapter 3 8 - W.A. Brand
38.9 Conclusions
Variation of stable isotope ratios in nature mostly are small. They are, however, important tracers that can reveal a wealth of information about processes that are happening or that have happened in the past. In order to read this information, the underlying principles need to be understood, and, most importantly, the measurements need to be made with the appropriate high precision. This has become possible with the development of the special instrumentation now common in the laboratories that have specialized on isotope ratio analysis. In order to measure and maintain high precision, the instruments need to be understood by the analyst in a quantitative way. All instrument designs have their merits and pitfalls which must be weighed in order to produce a consistent set of isotope ratio values. The role of differential pumping, the design principles of inlet systems, and the selection of permanent versus electromagnets have been discussed. The importance of software in dealing with ion corrections, the need for quantitative integration of transient signals, and the importance of recognizing and mitigating ion molecule chemistry all have been stressed in this contribution. The future in isotope ratio mass spectrometry clearly belongs to the chromatographic techniques (isotope ratio monitoring, irm). The central role of excellent chromatography for high precision data has been pointed out. The best utilization of isotope ratio monitoring techniques will require good working knowledge of chromatography, to ensure the best separations, and good working knowledge of the isotope ratio mass spectrometer, to maintain high precision analysis capability in routine operation on a daily basis.
Acknowledgments
I am indebted to Roland A. Werner and Jon Lloyd for helpful discussions and a critical review of the manuscript. Charles B. Douthitt has improved the first draft considerably with his critical editing of the manuscript. Most of the experience on which this contribution is based was acquired at Finnigan MAT in Bremen, Germany during 14 years working as a designer of mass spectrometric equipment. Here, I am particularly grateful to Karleugen Habfast for the many fruitful discussions and for being a wonderful teacher.
Handbook of Stable Isotope Analytical Techniques, Volume 1 P.A. de Groot (Editor) 9 2004 Elsevier B.V. All rights reserved.
C H A P T E R 39
Techniques of Ion Current Stabilization in Isotope Ratio Mass Spectrometry S. Halasl & T. Durakiewicz2 Uniwersytet Marii Curie-Sklodowskiej, Instytut Fizyki, Pracownia Spektrometrii Mas, P1. M. CurieSklodowskiej 1, 20-031 Lublin, Poland 2 Los Alamos National Laboratories, Condensed Matter & Thermal Physics Group, Mailstop K764, Los Alamos, NM 87545, U.S.A. e-mail: 1 [email protected]; 2 [email protected] 1
39.1 Introduction The quality of isotope ratio measurements made with a mass spectrometer primarily depends on the stability of the ion currents. In this chapter we discuss the most typical sources of the instability of the measured ion currents. At the beginning we will assume that vacuum conditions are good and the magnetic analyzer (this refers to either a permanent magnet or an electromagnet) produces constant field and the high voltage supplied to the ion source is stable. When both conditions are fulfilled simultaneously, then it guarantees a permanent position of the ion beams with respect to the collector slits. Hence, if the ion beams have no tendency to drift, the average ion current should be constant over a long period of time. If this is not the case, then the most likely the ion current variations may be due to the unstable electron beam (in gas ion sources with electron impact ionization) or unstable filament temperature (in thermal ionization sources). These two sources of the ion beam instability can be nearly completely removed by the methods described in sections 39.6 and 39.7. Fluctuations of the ion current of the beams having stable positions with respect to the collector system are usually generated in the ion source. It is characteristic in this case, that the fluctuations observed on any pair of the Faraday cups are correlated. One can easily recognize whether the ion beams fluctuate due to instabilities of high voltage and/or magnetic field or due to variation of the ion production rate. A test of this is to set the major beam at the edge of the collector slit in the position at which the ion current drops to about half of its maximum value (plateau). If the stability of the ion current is apparently worse at this position than at the plateau, then this is due to poor stability of the high voltage supply or, if no permanent magnet is employed, the electromagnet power supply. The high voltage supply may be tested by precision measurement of a portion of the high voltage (between ground and a selected point on the high voltage divider). A precision digital voltmeter (DVM) is necessary for this test. If the DVM reading indicates fluctuations of the high voltage
858
Chapter 39 - S. Halas & T. Durakiewicz
on the ion source, then the high voltage supply should be tested after disconnecting it from the ion source. If the fluctuations disappear, then cleaning of the ion source is mandatory, see section 39.5. The instability of the magnetic field of the analyzer can be detected either by precision measurement of the electromagnet current or by observation of the m / e indicator which displays a value proportional to square of the magnetic field (B2). Simple, but precision m / e indicators are described in section 39.4.
39.2 Instabilities generated in the collector assembly The collector assembly of an isotope ratio mass spectrometer (IRMS) contains two or more Faraday cups mounted on the high quality ceramic insulators. These cups are surrounded by a common electrode for suppressing the secondary electrons. Although the collector assembly of any IRMS is installed inside the high vacuum system, it can be troublesome when a noise is generated by unwanted leaks of the ion currents to ground or by the unstable secondary electron suppressor voltage. The electric leaks are due to presence of various kinds of impurities on the collector ceramics and feedthrough insulators. In order to test the insulation quality of the vacuum feedthroughs and the ceramic spacers of the collector assembly one should observe (record) the noise of the amplifiers on the most sensitive ranges, i.e. at the same conditions as for their zero-setting. Then the amplifiers should be set to the least sensitive range and their inputs disconnected from the collector feedthroughs. Setting the amplifiers to the least sensitive range will help to prevent their high-ohm input circuits from damage during removal. Subsequently, the amplifier box should be closed, or firmly shielded, and the zero-lines recorded again on the most sensitive ranges. If an essential improvement is observed, i.e. the amplitude of the noise is lowered, then the above mentioned insulating surfaces have to be cleaned. First, the accessible surfaces (outside the high vacuum system) are cleaned by washing them with ethyl alcohol and drying. Then the above test is repeated. If the enhanced noise is still observed, then prior to eventual cleaning of the collector assembly and inner surfaces of the feedthroughs, a uniform heating up to 100~ of the flange with the collector assembly is recommended. In most cases such heating performed for a long period (e.g. overnight) significantly reduces the noise due to impurities on the inner insulating surfaces. Nota bene it is a good praxis to keep this part of the vacuum chamber always warm, at 50~ to 60~ except during the measurement time. In the worst case, when the electric leaks are detected after thorough heating of the flange with the collector assembly, the collectors themselves have to be cleaned after venting the analyzer tube. The best method of cleaning of the ceramic spacers is an acid treatment (HC1, HF) followed by boiling twice in distilled water, drying and baking in a quartz tube furnace at 600~ in open air. Deposits on the metallic parts of the collector assembly can be cleaned in the similar manner as the lenses of the ion source, see section 39.5. The second reason for noise in all the Faraday cups can be unstable suppressor voltage, which is supplied to a common electrode (grid) installed in the collector
Techniques of Ion Current Stabilization in Isotope Ratio Mass Spectrometry
859
assembly for repelling the secondary electrons. Such a correlated noise can be observed without ion beams. The secondary electrons are produced by high energy ion beams when they pass through the collector slits and hit the surface of the cups. Without any suppressing voltage, for instance if the grid is firmly grounded, it is possible to observe characteristic negative peaks (valleys) adjacent to each positive peak. These valleys are formed due to electrons ejected from the edges of the collector slits. Because the voltage at the suppressing grid should be extremely stable, it is advised to replace any electronic unit (DC supply) by a battery which can be simply arranged from 4 to 6 commercially available 9V batteries. The batteries should be closed in a grounded box and the negative potential of-40V to-60V (depending on the high voltage applied to the ion source) should be connected to the suppressing grid by a shielded cable. In parallel to the output voltage a high-quality RC filter is recommended (e.g. a 1MQ resistor and l gF capacitor). The above comments refer to spectrometers having the suppressing grid installed in the collector system. Some modern machines have a permanent magnet installed for bending of the secondary electron trajectories. A small magnet is able to bend the trajectories of the electrons, but not the ion beams being composed of high energy massive particles. The best combination, used in some modern machines, is magnet and suppressing voltage. 39.3 Conditions for stable work of the ion source
A typical ion source comprises numerous electrodes being kept at precisely selected potentials with respect to the ground (GND). These potentials are set by means of a quality voltage divider which is usually mounted in the same box as the electron emission controller (see section 39.6) or the filament temperature controller, see section 39.7. Firstly, we will consider the electron impact (EI) ion source, which is used in analysis of gases and vapors. The molecules are admitted from an inlet system into a ionization chamber (cage). A schematic diagram of a Nier-type ion source is shown in Figure 39.1. Note that the ionization chamber has the highest potential in the whole mass spectrometer, so a gas has to be introduced to the cage by a piece of insulating tube, usually made of ceramic or quartz glass. The electrons emitted from a hot filament are accelerated by a voltage (from 20 to 100 V, typically 70 V) between the filament and cage. Most of them hit the wall of the ionization cage but a considerable fraction passes through the cage with constant kinetic energy and terminate on a plate (called the electron trap) which is fixed at the opposite side of the cage. The trap potential is somewhat higher (ca. 30 V) than the cage in order to suppress the secondary electrons produced as primary electrons pass through the cage and hit the trap. The electron emission controller keeps the trap current constant by precise adjusting of the filament temperature. The trap current (also called as electron emission current, Ie) ranges from a few microamps to about I mA, typically in IRMS it is set at 50 gA. Two solid-state magnets are usually used to form the magnetic field parallel to the electron beam. In such a magnetic field electrons move on elongated spiral paths, which enhances the ionization efficiency substantially.
860
Chapter 39 - S. Halas & T. Durakiewicz
Schematic diagram of a Nier-type ion source with a typical potential distribution on the electrodes. The extracting plates and beam defining slits are expanded for better insight into construction. Figure 3 9 . 1 -
ca.
If gas pressure in the ionization cage is properly controlled, then only a minor fraction of electrons passing through the cage collides with the molecules causing their ionization and dissociation in the following most common reactions: AB + e ~ A B + + 2e A B + e - - * A + + B + 2e
[39.1] [39.2]
where A and B denote atomic species or fragments a molecule AB. Other species, like doubly charged ions, are produced with significantly lower efficiency; thereby they are not useful in IRMS. The rate of ion production at constant Ie depends on electron energy, i.e. on the electron accelerating voltage. A broad m a x i m u m of ionization efficiency is about 100 V for most of molecular species. The ions formed along the electron beam are driven off by a weak electric field which penetrates into the cage from the space between accelerating and extracting plates, see Figure 39.1. In some EI sources a plate (called as ion repeller) is fixed in the cage in order to enhance the ion extraction. Efficient extraction results in lowering the rate of ion-molecule reactions in the cage, which is particularly important in hydrogen isotope analysis, because species formed in the cage interfere with HD +. The species are formed due to following reactions: H2 + + H ~ H 3 + H2 + H + --* H3 +
[39.3] [39.4]
Techniques of Ion Current Stabilization in Isotope Ratio Mass Spectrometry
861
where H2 +, H + and H are formed from H2 by electron beam. Most of ions extracted from the cage should pass subsequently through the slits of the extracting plates. One half of the first set of these plates has a slightly adjustable potential for optimum positioning of the ion beam at the narrowest grounded slits, called beam defining slits. Another adjustment is made by selecting potential (e.g. from - 100 V to + 100 V) of the correction plate. In the case of ideal geometry and clean plates these adjustable potentials will be identical to those of the opposite electrodes. The setting of the adjustable potentials in an ion source (extraction voltage, focusing voltage and correction voltage) should be done alternately several times in order to attain optimum potential distribution. This will result in the most intense and stable ion beam. In thermal ionization mass spectrometers (TIMS), used in IRMS of solids, the electron beam is replaced by an ionizing filament. The analyzed substance is either placed directly on this filament or on a side filament used for its controlled evaporating. Phenomena, evaporation and surface ionization are strongly (exponentially!) temperature dependent. Therefore a high quality temperature controller has to be employed in order to reduce instabilities of the ion beam, see section 39.7. 39.4 m/e indicator A device capable of precisely displaying the m / e ratio of a selected ion beam is extremely useful, not only in identification of impurities in the background of a mass spectrum and in the adjustment of the distance between collector slits, but also in detection of instabilities of the magnetic field produced by an electromagnet. The reason of spatial instability of the ion beam can be easily recognized by use of the m / e indicator, which has a short-term precision of at least 0.01 mass unit.
Halas & Sikora (1987) described a simple dual Hall probe device that indicates m/ e ratio with the required short-term precision of 0.01 mass unit. The principle of its action is as follows. The first Hall probe, H1 in Figure 39.2, is fed by a constant current I1. Thus the output voltage, V1, is proportional to the product V1 ~- IIB
[39.5]
where B is the magnetic field intensity experienced by this Hall probe. The output voltage V1 is repeated by the operational amplifier with 100% feedback coupling. The second Hall probe, H2, is fed by current I2 directly proportional to V1 from a voltageto-current converter. If the second Hall probe experiences the same magnetic field, then its output signal, V2, will be proportional to the product I2B. Then, from equation [39.5] and the proportionality of I2 to V1 we obtain that V2 ~ B 2
[39.6]
which means that V2 will be proportional to the m / e value in the magnetic scanning operation mode.
862
Chapter 39 - S. Halas & T. Durakiewicz
I 11=constant
V2 H2 /
V1
VtoI converter
Figure 39.2 - Schematic diagram of the dual Hall probe method.
By adjustment of the I] value, one can obtain the voltage V2 (in milivolts) exactly equal to m / e ratio, which is very useful. Two small electrically insulated Hall probes may be placed in parallel between the magnet poles to determine the magnetic field intensities. In our m / e indicator two inexpensive CdHgTe thin-film Hall probes were used with 60f~ resistance each. The maximum current applied to feed the probes was less than 20 mA, so that no significant power was dissipated. In the case of mass spectrometers with permanent magnets, B is a constant, and so: m / e ~ 1 / Vacc
[39.7]
where Vacc is acceleration voltage of the ion beam, which is the highest potential in the ion source with respect to ground. Hence, to indicate the m / e ratio it is necessary to reverse the high voltage (or a portion taken from a voltage divider). The reversed voltage can be displayed by means of a so-called double integrating DVM. Such voltmeters are commonly used in laboratories and DVMs with 4.5 digits are commercially available in panel versions with price below 30 US $. These voltmeters compare the measured voltage with a reference voltage produced by a specialized circuit inside the DVM. By reversing the inputs of the "measured voltage" with the "reference voltage" one can simply obtain the reversed voltage on the display. By adjustment of the voltage divider, a selected fraction of the high voltage is reversed in order to obtain the value of m / e ratio on the display. 39.5 Cleaning of the ion source
After a long-term operation of any ion source the insulating ceramics and the lenses became coated by a thin semiconducting layer. The deposits are formed by the decomposition of the analyzed substances as well as metal sputtered off the lenses by the interaction with high speed ions. These layers may produce episodic breakdowns
Techniques of Ion Current Stabilization in Isotope Ratio Mass Spectrometry
863
of the high voltage in the ion source or long-term instability of the ion beam due to permanent charging/discharging of the semiconducting layer on the surfaces of the ion lenses. During high voltage breakdowns the ion currents are seen to vary rapidly and with unpredictable recurrence (the characteristic effect is a sudden loss of ion beam signal followed by a return to the previous strength over a matter of seconds). Such an erratic behavior of the mass spectrometer is an immediate signal that the ion source has to be cleaned. The cleaning procedure described below is not difficult and can be done by any competent operator. The ion source can be disassembled from the analyzing tube after venting the instrument through a filter to avoid entry of any solid particles, the most dangerous of which are small ferromagnetic pieces which can be attracted by magnetic field of the analyzer. It is therefore advised that the magnet be removed before venting the analyzer tube. If an electromagnet is used then its power supply should be turned off before venting. Having the ion source in a clean working environment, while the end of the analyzer tube protected against dust by a piece of aluminum foil; disassemble the ionization cage and the lenses. Usually the ionization cage is the most dirty part of a gas ion source. It can be cleaned by removing the deposits by use of a delicate abrasive paper, e.g. 1/600 mm. The lenses can be cleaned in a similar manner. Use of distilled water for periodical rinsing of the parts is recommended. Finally, the parts are rinsed in acetone and boiled in distilled water. At this stage the lenses can be electropolished (see Peele & Brent 1977). The electropolishing is a reverse electrolysis in a viscous strong electrolyte. It is a simple, effective and inexpensive technique. For electropolishing of stainless steel the following recipe for the electrolyte can be used: H2SO4 (1.84 specific gravity) 1000 ml water 370 ml glycerin 1370 ml Stirring, add the acid slowly to the water; avoid overheating. Cool to room temperature, then add the glycerin and stir well. The electrolyte can be used for many cleaning procedures during which the cleaned stainless steel part is immersed as a positive electrode whilst a larger piece of stainless steel foil surrounding that part is connected to the negative pole of a 24V/ 25A rectifier. The use of any other metals or alloys as negative pole is not recommended. The temperature of the electrolyte should be about 60~ The ceramic parts can be cleaned by abrasive paper, rinsed with distilled water, and finally by boiling twice in distilled water. After thorough drying of all the parts in a clean furnace (at 70 to 100~ overnight). Baking of the ceramic parts at 700~ in air may be the next step, if their surfaces still seem to be colorated by a dark contaminant. The ion source can be reassembled using clean tools. Ion sources of modern mass spectrometers often have lenses, which are mounted
864
Chapter 39 - S. Halas & T. D u r a k i e w i c z
on sapphire balls for the best insulation and adjustment. This ensures long-term operation of the source without cleaning and fast outgassing of the surfaces. In contrast, the outgassing of ceramic rods and wobbly ceramic spacers used in older ion sources is time consuming. Sparks due to trapped gas inside ceramic parts may appear when the high voltage supply is turned on too early. The tolerance of the sapphire balls can be limited to _+3 mm, which guarantees identical location of the lenses after cleaning or replacement. 39.6 Stabilization of electron e m i s s i o n current
Fluctuations of the electron emission current transform into fluctuations of the ion current immediately. If the local pressure variations inside the ionization chamber are neglected, the variance of the ion current may be considered as the linear combination of the variance in both electron emission current and ionization cross-section of gas molecules (which varies with electron energy). It is nowadays relatively easy to obtain the electron acceleration voltage (Ve) stabilized with the precision better than 0.01%. Hence, it is only the electron emission current (Ie) that may disturb the ion current stability. There are two general types of electron emission regulators: the series regulator and the switching mode regulator. Regulators, which utilize a switching mode, can offer several advantages over conventional, continuously controlled, series type regulators. In contrast to a well known series-pass regulator of filament current (Halas & Sikora 1990) the comparator in the switching mode circuit has also a positive feedback loop formed by a resistor connecting the noninverting input of an operational amplifier (OA) with its output. This results in pulsation of the output voltage from - 15V to + 15V when the signal at the inverting input is slightly modulated around the reference voltage. Such a design leads to a self-oscillating stabilizer (Durakiewicz 1996) the simplified diagram of which is shown in Figure 39.3. A number of improvements have been made to the original circuit (see Figure 39.4), namely: (1) The electric potential distribution on the filament, cage, ion repeller and trap have been redesigned in order to allow the precise control of the average value of the electron current passing through the cage. The average trap current instead of total emission current is stabilized. (2) The protection of the filament against vacuum break down has been simplified. The respective fragment of electronic circuit comprises only a simple capacitanceresistance filter (C2 = 100 nF and R2 = 2.2 MQ). (3) A modern solid state relay (SSR) for the power supply of the filament was used; thereby the number of electronic components has been reduced. The inverting input of the Schmitt trigger (OA2 with positive and negative feedback loops) is fed from the output of OA1 (trap current-to-voltage converter). Positive feedback loop is sustaining the self-oscillating action of the circuit, whereas the negative feedback is responsible for stabilization. A square wave generated at the output of the trigger passes through the C2-R2 filter to the input of OA3 (voltage follower).
865
Techniques of Ion Current Stabilization in Isotope Ratio Mass Spectrometry
Ve
Vr
T
,p
I,F / -)-
I
IeI
"%.
=
,"9
! VREF
'!,.%
[
R3
__.,.r"
I
T Figure 39.3 - Simplified diagram of a self-oscillating electron emission stabilizer. Ie is the electron emission current, Ve is electron acceleration voltage, VREFis the reference voltage, and Vf is the filament supply voltage.
The output voltage from OA3 feeds the light emitting diode of the SSR. Due to the switching action the electron beam is modulated, i.e. the trap current varies periodically between "Hi" and "Lo" values. The switching frequency for a typical filament installed in the ion source ranges from 0.1 to I kHz. The maximum positive voltage is generated on the OA2 output in the case of bad vacuum or any other break in the negative feedback loop. Use of a capacitor (C2 in Figure 39.4:) prevents the filament from burning out; such a protection may not be applied in conventional stabilizers.
(tlA'J ~_T I"-.R1 -.-.-.-] - ~
l +1
Trap l+ Voltage I
CAGE I --
'
+15V ION REPELLER r
TRAP
R~6 ~) ~
r k1 [ ~ P I C A MEN T ~
5 k fa klo
J R7
-15V
I vl
Figure 39.4 - A detailed schematic diagram of the self-oscillating electron emission controller.
866
Chapter 39 - S. Halas & T. Durakiewicz
Although Ie is slightly modulated, its mean value remains extremely stable. Even after a few days of continuous operation the mean Ie value does not vary more than + 1%. The stability is not affected even by large variations of electron acceleration voltage (AV) from 40 to 100 V. The short-term variations of ion current induced by Ie modulation are negligible as long as the ion current detectors have the response time of the order of 100 ms or larger. The long-term stability of the ion current is improved significantly, due to better stability of Ie. The mean value of Ie is not influenced by network supply voltage instabilities of amplitude _+20%. A comparison of results obtained by electron emission stabilizers of various designs is given in Table 39.1. All the designs listed here, except that by Durakiewicz (1996), are generally based on the proportional regulation principle. The advantages of pulsed heating for electron emission control may be summarized as follows: (i) low power consumption - the circuit may be used in a battery operated system (e.g. in space technology), (ii) self-protection against breaks in the negative feedback loop, including the inner part of this loop between filament and electron trap (e.g. due to a bad vacuum), (iii) electron emission current is independent of the electron acceleration voltage and vice-versa, (iv) one pole of the electron acceleration voltage source is connected to the ground of the system, thus Ie does not influence the acceleration voltage itself, (v) the circuit is simple and convenient in operation. The slight modulation of Ie with frequencies above 100Hz is not a drawback in mass spectrometry, where ion current detecting devices with only slow response time are used. Table 39.1 - Comparison of operation parameters of selected electron emission stabilizers. Chapman)
Close & Yarwood (1972)
(1972)
Herbert (1976)
Shaw & Lue (1980)
Halas & Sikora (1990)
Durakiewicz (1996)
operation principle
proportional proportional proportional temperature regulator regulator regulator stabilizer
doubleproportional regulator
selfoscillating
stabilization coefficient
0.1%
relative cost
medium
filament safety
-
0.1%
I low -
0.1%
>1%
1%
0.1%
very high
high
low
low
safe start
-
safe start & loop break
loop break
...........................................................................................................................................................................................................................
Ie/Ue independence
-
, .........................................................................................................................................................................
+
-
~..........................................................................................
-
+
<50%
<80%
l
.................................................................................................................................................................................
energetic efficiency
-
<50%
t...........................................................................................................................................................................................................................
I<50% i
<50%
<30%
Techniques of Ion Current Stabilization in Isotope Ratio Mass Spectrometry
39.7 Stabilization of filament temperature In thermal ionization mass spectrometry the ions hot metal with high melting point (usually W, Ta, which can be drawn from the surface at equilibrium well-known Saha-Langmuir formula (Benninghoven 2001)" .+
867
are produced on the surface of a Re, Ir, Pt). The current density, conditions, is determined by the et al. 1987, Valyi 1977, de Laeter
ve
J = 1
g~ E~-q3 + --~exp kT g
[39.8]
where v is the flux density of particles supplied to the surface, e is elementary charge, g+ and gO are statistical weights of ions and neutral particles, respectively, Ei is the ionization energy of the particle, q~is the work function of the metal, k is Boltzmann constant and T is the absolute temperature. The ratio of gO/ g+ is equal 2 and 1/ 2 for alkali and alkaline earth metals, respectively. Numerical values of ionization potentials of elements and statistical weights of ground state of atoms and ions are given in Uns61d (1968). The formula [39.8] is strictly valid only when the surface coverage is small. For negative ions the respective formula is: -
J -
0
ve
[39.9]
1 + g exp q9- A
g
-
kT
where A is the electron affinity. Numerical values for electron affinities of some elements and their oxides can be found in Valyi (1977) and Wachsmann & Heumann (1991, 1992). Inasmuch as both nominator and denominator in formulae [39.8] and [39.9] is strongly temperature dependent, a precise filament temperature stabilization is demanded for production of stable ion beams by the surface ionization process. In the case of a single filament ion source the flux density, v, is driven by the surface diffusion from cooler ends of the filament towards its center. The filament center should be kept at a temperature at which the analyzed atoms can be effectively ionized, which means that the surface coverage is far below the monolayer at the filament center. The filament length may be also crucial in obtaining stable ion currents. It is so because a temperature plateau exists in the case of too long filaments. In such a case, the fluxes of ions are emitted from relatively large surface with high fluctuations due to variations in surface coverage at the plateau region. However, steady conditions for ionic emission can be obtained for relatively short filaments, for which no temperature plateau exists (details on defining long and short filaments may be found in Halas & Durakiewicz, 1998a).
868
Chapter 39 - S. Halas & T. Durakiewicz
The following methods may be used for the filament temperature stabilization: (1) A commonly used method of current stabilization, (2) Filament resistance stabilization, (3) Filament voltage stabilization, (4) Stabilization of power supplied to the filament. Method (1) is relatively easy to apply but it has a significant drawback: the filament temperature depends highly nonlinearly on the heating current. Therefore we describe below the temperature controllers based on methods (2) and (3) which are characterized by the linear dependence of filament temperature on its resistance or on the voltage drop between its ends, respectively. Method (4) will not be considered here because it requires rather complicated electronic circuits. The filament resistance stabilizer can be made on the basis of the Halas-Kaminski bridge (Halas & Kaminski 1995, Halas et al. 1993, Durakiewicz & Halas 1995). A conceptual diagram of the stabilizer is shown in Figure 39.5. The bridge contains a DC excitation voltage source, Ve. The excitation current passes through a diode and the bridge resistors but not through the transistor T. RF and Rv denote, respectively, the resistance of the filament and a variable resistor used for temperature setting. The filament is made of a pure metal (e.g. W, Ta, Re, Pt) which have positive and significantly large temperature coefficients of electrical resistivity. If Rv < RF then the bridge signal supplied to the noninverting sample-and-hold amplifier (S/H) is negative or zero and the transistor does not conduct. The noninverting input of the OA is connected to the output of a S/H amplifier whilst the inverting input of OA is fed by a positive signal from a triangular wave form generator, G. In the reverse situation, i.e. Rv > RF, the imbalance signal generated by the bridge is positive and has a defined magnitude. This signal is amplified by the S/H and held at the S/H output, when the negative potential taken from the output of OA turns to positive. The comparator OA produces a positive output voltage if the generator signal is below the voltage held by S/H. In this fraction of the cycle, transistor T conducts. Hence the power supply, V1, is connected in parallel to RF (through T, which plays the role of a switch). No significant current from V1 can pass through the remaining parts of the bridge due to presence of the diode D. Because V1 considerably exceeds Ve, the excitation current cannot pass through the low-resistance legs of the bridge (Rv and RF) when RF is supplied from V1. This is the reason for using a sample-and-hold amplifier instead of a normal operational amplifier. As result of the action of the circuit shown in Figure 39.5; a series of heating pulses is supplied to RF with frequency driven by the generator but their duration is driven by the output voltage of the S/H amplifier. The bridge is always kept close to the balance state, i.e. RF - Rv. If for some reason (e.g. voltage V1 starts to diminish) RF resistance becomes somewhat lower than Rv, then the bridge imbalance signal becomes somewhat higher, which results in a longer duration of the heating pulses. A complete circuit diagram is described by Halas & Durakiewicz (1998b).
869
Techniques of Ion Current Stabilization in Isotope Ratio Mass Spectrometry Figure 39.5 - Conceptual diagram of using the Halas-Kaminski bridge for constant-resistance operation of a filament RF. S / H is a sample-and-hold amplifier; G is a triangular wave generator.
N_
S/H
,," ,,,,-"
o,,,d",.,",,d,,,,,c,,d ,. .,."
R
/'.
;x./
-',.-"%,%
"--,.\.
+ ~ "
I
No
",,% ,,, ",. "%",% ", ,, %., "',,,
"de"
/
I G A/V'~]
% ",,%
Ve
---tl ,.
"%:q
I b,l
I",J
"'.....
,,,,,,".--
D ,z' .,."
F- 4 ~
Vl
lt---~ '1
+
T
The use of a simple voltage stabilizer for temperature control of a filament was described by Halas et al. (2001). Voltage may be stabilized typically by use of the same two wires to feed voltage to the filament and measure the voltage value (2-wire method) or, as it is described below, by use of a separate pair of wires for voltage measurement. In such a 4-wires configuration the voltage drop along the supply lines does not influence the value measured by the stabilizer circuit, hence the fluctuation of the ion beam is significantly reduced. The schematic diagram of the circuit is shown in Figure 39.6. The MA741 operational amplifier with the negative feedback loop constitutes the basis of the stabilizer circuit. The reference voltage supplied to the noninverting input of the amplifier may be set either manually by use of the 10k potentiometer, or digitally by use of the digital-to-analog converter (AD7243), or manually by use of the potentiometer. All the fluctuations of the filament voltage are minimized by the negative feedback loop what allows for the temperature stabilization of the filament. In the 4-wires method the resistance variations of filament power supply wires do not affect the filament voltage measurement. The output signal of the MA741 drives the Darlington circuit comprising the 2N3055 and BD439N transistors. This circuit keeps the filament voltage constant and therefore stabilizes its temperature. Since ions have to be formed into a beam, the filament and the whole supply circuit is fixed on the 2kV potential with respect to the ground. Because of the high potential of the source it was necessary to construct the three channel optical relay between the master computer and slave digital-to-analog converter in the filament power supply. In this way the master computer is protected against high voltage. The remaining components identified in the schematic diagram are used to allow smooth switching between the computer and manual control, and to protect the filament against burning during power-on of the supply.
870
Chapter 39
- S. Halas & T. Durakiewicz
SYNC SDIN LK SYNC ~.x.. ~::'S :,, '~' ~ l SDIN \470K D K"I ~.]N SCLK \.,. ", -1- "~ CLR 470KN\470 i i i rlVref d====++5V\ ;,.X, ./ / ..a,k14v \~-- I MAA I12v _ I c12J, liAD7243{ ZlmF 2.2M~ GND T,, ~ , . , ' T 17812 I i ~ +~5 I I [ +5V
'3
~ .~
1",-.t/ I I FI"L IJ'-It 1
_
I MAA
I
I
_
I I ] 7805 l i i35k "+5V 10mP ~ ~ ~mP,ll~P~ 0 ~10 k Vref
=-15V
Vout
Comp_ Adjust"~ == ~I~F
~--_ MA741 L~~__
----~+2kV
///"N\+ 3.8V "
Il
T+V -3x i2 mF
r
2N3055
91
Figure 39.6 - Schematic diagram of the filament temperature controller based on the stabilizer.
wires voltage
)
9N 4-
The voltmeter indicated on the schematic diagram enables the visual control of the filament temperature. This is possible because the temperature-voltage characteristics of the filament are almost linear, as demonstrated by Halas et al. (2001). The voltmeter was calibrated directly in temperature units what allows the operator immediate information on the filament status. The stabilizer is fully computer controlled via the digital-to-analog converter AD7243 (see Halas et al. 2001 for details). In order to demonstrate the superiority of the 4-wires over the 2-wires approach, we have continuously recorded the ion current of 39K+ for 20 minutes. Potassium was loaded onto a platinum filament, 0.05mm x 0.63mm x 10mm, as 50% K3PO4 solution. After drying in air and degassing under vacuum, the filament was incrementally heated to 900K. Ion emission was measured at a filament temperature about 1000K. Comparison of 39K ion currents obtained using the above-described voltage stabilize with 4-wires and 2-wires configuration is shown in Figure 39.7. 39.8 Integrating of the ion currents The ion detection system of an IRMS consists of a collector assembly containing two or more Faraday cups (as mentioned in section 39.2). Although the ion counting
871
Techniques of Ion Current Stabilization in Isotope Ratio Mass Spectrometry
Figure 39.7 - Comparison of 39K ion currents obtained by the authors for filament voltage stabilization by use of 2-wires and 4-wires methods, respectively. technique is used in specialized instruments for detection of very weak ion beams (like 3He, 230Th), we will not consider these devices here. The ions collected in each cup produce electric currents, which flow through a high-value resistor (i.e. 109 - 1012 f2) to ground. The voltage produced on the resistor is not amplified but rather repeated by a circuit arranged as a voltage follower, or, more frequently, it is reversed by an amplifier with the resistor forming a negative feedback loop. A typical detector system of an IRMS is shown schematically in Figure 39.8. The ion current is converted to a voltage by an operational amplifier with ultra-low input bias current and the high-value resistor. The output voltages are then converted to trains of short pulses by so-called voltage-to-frequency converters (V/F), the repetiN 109Q I
t---"
V/F Faraday cups
,
Counter
N 1011Q I t---
Stop
V/F
Counter
1J Display Figure 39.8 - A typical arrangement for digital measurements of current ratios
872
Chapter 39- S. Halas & T. Durakiewicz
tion rate of which is linearly proportional to the input voltage. These pulses are fed to separate counters which are set to zero at the start of measurement. When the major beam counter reaches a count of 106, the minor beam counter is stopped and its value displayed. The value of the full six-digit display is therefore equal to the current ratio. Unfortunately the detection system described above introduces, its own noise, predominantly from the high-value resistors. This is a fundamental phenomenon which cannot be eliminated by technological improvement of the production of resistors (see Felgett & Usher, 1980, for example). Moreover, the resistors also suffer from fluctuations of their value due to variations of the potential drop along the resistor and temperature changes (Habfast, 1960). The key to the improvement of isotope ratio measurements is to replace each highvalue resistor by a capacitor (Jackson & Young 1973, Halas & Skorzynski 1980, McCord & Taylor 1986). This replacement converts the ion detector from an ion "amplifier" to an "integrator" where the voltage on the capacitor raises in time proportionally to the charge collected: T
V - c f l d t --
I.T
C
[39.10]
0
where C is the value of capacitance, I is the ion current and T is the integration period. The value of the capacitor is selected in such a manner that the final voltages are of order of 10 volts. The integration period may be estimated from the following statistical considerations. Let us assume that the 180/160 ratio is measured using CO2 gas. Typical currents obtained by a Nier type ion source for mass 44 and 46 are 2.5 x 10-9A and 1 x 10-11A, respectively. Hence, after time T the number of electrons collected on each capacitor is: n~
I.T
[39.11]
e
where e is the elementary charge 1.602 x 10-19As. According to the general principles of statistics, the relative uncertainty of An/n is equal n-l/2; hence for a desired uncertainty of about 10-5 (i.e. + 0.01%o) the number of ions collected has to be 1010. From equation [39.11] one obtains for the minor beam: T~ 101~
1.602.10-19As 1.6. 102s 10
-11
A
i.e. the integration time has to be of the order of 100 seconds.
[39.12]
873
Techniques of Ion Current Stabilization in Isotope Ratio Mass Spectrometry
The obvious inconveniences of the integrating system are that the "---. 2 ""--, 1 "-... capacitors have to be discharged periodically, prior to each charging C cycle, and the system cannot be used I I directly for the instrument adjust" I I ments by continuous monitoring of the ion current. Both difficulties, however, can easily be overcame today by use of the computer controlled high quality reed switches for discharging the capacitors and their instant switching to the resistors. Such a solution was described by Halas & Skorzynski (1980). One pair Figure 39.9 - Schematic diagram of the capacitance/ of reed switches is required for each resistance system. amplifier. The switches are connected in parallel to C and R as shown in Figure 39.9. In the integrating mode switch 2 is closed and switch I is used for periodical discharging of the capacitor. For adjustment of the mass spectrometer, recording the mass spectra, etc., switch I is closed whilst switch 2 is open.
i
R
39.9 Final remarks
Good quality isotope ratio results are worth taking the effort. The ideas and general remarks presented in this chapter are certainly not the only solutions that guarantee success. They are, however, tested by many years of operation and maintenance of several instruments in our Mass Spectrometry Laboratory. In the final section we should make our reader aware, that having the ion beams stable enough to produce satisfactory precision of 0.05 permil or better for standard versus standard measurement does not guarantee identical precision for sample versus standard measurement. If this is not the case, one should check the purity of the sample and/or the gas flow conditions through the inlet system, as well as the linearity of the system. Having stable ion beams, pure samples and a high quality inlet system one has a chance for good and long-term performance provided that settings of the IRMS are favorable. The optimum setting of the ion source should assure maximum ion current and best peak shape at minimum electron beam. The gas flow rate through the capillaries should be selected depending on the geometry of the ion source. The pressure of the analyzed gas (as estimated on the basis of major beam current) during gas flow should not exceed significantly the 100-fold background pressure. Too high a pressure leads to ion scattering by the gas molecules and thereby to peak broadening.
Handbook of Stable Isotope Analytical Techniques, Volume 1 P.A. de Groot (Editor) 9 2004 Elsevier B.V. All fights reserved.
C H A P T E R 4O International Stable Isotope R e f e r e n c e Materials Manfred Grtning International Atomic Energy Agency, Agency's Laboratories Seibersdorf, Isotope Hydrology Laboratory, A-1400 Vienna, Austria e-mail: [email protected]
Abstract The availability and the proper use of suitable reference materials is one of the basic preconditions to ensure the comparability of stable isotope ratio measurements as performed by different laboratories. In this contribution the most important international stable isotope reference materials are introduced and described in some detail, which define 6 scales for the elements hydrogen, carbon, nitrogen, oxygen and sulfur. In addition brief comments are given on isotope reference materials for the elements lithium, boron and chlorine. The recent situation, demands and future trends for new isotope reference materials are discussed. Some information is provided on the calibration of measurements using these reference materials. The problem of a proper terminology for reference materials is presented and some terms proposed for general use with the intention to achieve better consistency in discussions, descriptions and publications. 40.1 Stable isotope ratios and reporting scales 40.1.1 Isotopic abundance and conventional scales For reporting an isotope ratio R of isotopes a and b of an element, commonly the ratio is expressed as the abundance [b] of the minor isotope b divided by the abundance [a] of the most abundant isotope a, both abundances being calculated from the same measurement process. For some elements (e.g. B, S), the inverse notation is used to express isotope ratios, in order to start the numerical value with a leading non-zero integer value. For several elements, including boron or lithium, this direct reporting of isotope ratios is commonly used. In most applications of stable isotopes in earth sciences, however, it is of much more of interest to know the differences in isotopic ratios between samples, than to know the "absolute" isotopic ratios of the samples ("absolute" meaning a ratio of amounts of isotopes, traceable to the S.I. system). The improvements of mass spectrometry for gas samples in the 1940s and 1950s (Nier, 1947; McKinney et al., 1950) have facilitated very much the precise measurement of isotope ratios for all those elements that can be converted to a gas for mass spectrometric analysis (e.g. to H2, N2,
International Stable Isotope Reference Materials
875
CO2, CO, CF4, 02, SO2, and SF6). Therefore, the isotope ratios of elements such as hydrogen, carbon, nitrogen, oxygen and sulfur have been commonly measured for decades by dual inlet mass spectrometers (McKinney et al., 1950) and the isotope ratios are reported as deviation from the isotope ratio of an artificially selected reference (~ scale). In this approach the knowledge of the mole fraction of an isotope of a given reference material is not necessary, since these isotope ratio ~ scales are defined completely arbitrarily relative to the isotope ratio of a selected primary reference material. This primary reference material functions as the end of the traceability chain (Chapter 41, this Volume). This concept as described is the realization of a "conventional" scale. It is not traceable back to the S.I. system, since it is not based on fundamental constants, but on arbitrarily selected properties, e.g. the stable isotope ratio in a sample of a chosen primary reference material (the pH scale being another example for a conventional scale). The choice of the primary reference material is completely arbitrary; as we will see, in several cases the selected material is even virtual / hypothetical and does not exist itself physically. In such a case its isotopic composition is just defined in measurable manner in terms of another real existing material and a statement on the isotopic difference of both. Certain basic requirements for the used reference materials have to be fulfilled, e.g. regarding homogeneity and stability (ISO Guide 34, 2000c). Special care should be taken in the isotopic characterization of reference materials; the mole fractions of the isotopes comprising the material should be precisely determined to allow the transition in the future, to a different scale based on another reference material. In this respect, precise determinations of mole fractions of the isotopes have distinct advantages. Firstly, in case a primary reference material is not more available, then the knowledge of mole fractions of its isotopic composition allows to maintain the original scale by a similar mole fraction determination for a suitable new material. Secondly, direct mole fraction determinations using primary methods (traceable back to the S.I. system) will allow to safeguard the linearity of the scale, because such measurements are much less susceptible to non-linearities than for differential measurements. Absolute isotope ratios of selected reference materials - that are discussed later in section 40.3 - are presented in Table 40.1. For the five elements H, C, N, O and S, the abundance of the major and the minor isotopes is different by orders of magnitude, resulting in very small ratio values. Since the natural variations in isotopic composition due to fractionation processes are rather small, the stable isotope ratios of naturally occurring materials on Earth do not differ considerably from each other. For example, the total range of the variation of 180/160 in water and ice on earth is in the order of only nine percent (Coplen et al., 2002; see Table 17 and Fig. 6) and increases to about seventeen percent if all natural oxygen bearing materials are considered (Coplen et al., 2002). Similar statements are true for most other elements except hydrogen (where the variations are exceptionally large). For that reason, an alternative way was introduced of reporting isotope ratios (McKin-
876
Chapter 40- M. Gr6ning
Table 40.1 - Absolute isotope ratios and associated standard uncertainties at 1c-level for selected reference materials discussed in this chapter. Name
Material
Isotope ratio
Isotope ratio value
Reference
Table 40.1 continued >
877
International Stable Isotope Reference Materials > Table 40.1 continued Name
a
b
Material
Isotope ratio
Isotope ratio value
Reference
Hypothetical carbon isotope mole fraction for VPDB calculated from the carbon isotope ratio of NBS 19. Hypothetical sulfur isotope mole fraction for VCDT calculated from those isotope ratios of IAEA-S-1, IAEA-S-2 and IAEA-S-3
ney et al., 1950; Epstein & Mayeda, 1953) using 6 scales. There, only the deviation of a sample isotope ratio Rsample is reported relative to that of an arbitrarily selected reference material Rreference:
6
-
R
sa
mple- Rreference Rreference
[40.1]
with 6 (e.g. 62H, 813C, 615N, 6180, and ~)34S) being the normalised difference of the isotope concentration ratios R (2H / 1H, 13C / 12C, 15N / 14N, 180 / 160, 34S / 32S) of the sample and the reference (e.g. ~15N with atmospheric nitrogen as reference). ~5-values are therefore unitless numbers, like the isotope ratios itself. As the differences between a sample and reference are normally very small, the ~ values are usually expressed as per mill difference (parts per thousand - per mill - %0 - 10-3). Thus, we have" 6 -
Rsample- Rreferencex 1000 %o Rreference
[40.2]
This modified equation [40.2] results in more convenient numbers being reported. The values can be positive or negative, with e.g. negative numbers indicating a lower abundance of the minor isotope in the sample than in the reference. Note that the per mill sign (%o) is part of the notation and may not be omitted. For a very detailed discussion of all features of the 6 notation as realization of a conventional scale, see Gonfiantini (1981). The reference in equation [40.2] is often chosen to represent the isotope ratio of a major sufficiently homogeneous reservoir of the element (e.g., ocean water for the elements hydrogen and oxygen, atmospheric N for nitrogen or marine carbonate for carbon). This way of reporting isotopic abundances has several advantages (Mook, 2000). First, the mass spectrometers typically used for measuring isotope abundances in naturally occurring materials are not really suitable for obtaining reliable absolute isotope ratios (or mole fractions of isotopes). Secondly, in most applications, the
878
Chapter 40- M. Gr6ning
differences of isotope ratios among samples are far more important than the values of the absolute ratios themselves. Moreover, due to the small variations in isotope ratios (most samples showing values close to the reference), all reported R values would need five to six digits after the decimal to be significant. Finall~ for the precision desired in many applications, the mass spectrometric dual-inlet mode or continuous flow mode are most appropriate, where a reference is necessary; because 6 values (differences in isotope ratios) can be determined about an order of magnitude more accurately that absolute isotope ratios or mole fractions of the isotopes of an element in a substance. The necessity for defining an arbitrary material with a well-known isotope ratio to realize the proper calibration of relative isotope data is not a disadvantage per se. That requirement is true for any transfer of properties from instrument to instrument, regardless of the measurement method used. The major disadvantage with this approach is that the property "isotope ratio scale" is defined using a physical material with its limited quantity available. Therefore, the isotope ratio scale, as defined by that material, is at risk to change when an exhausted primary material is replaced by a new one. The comparison of the old with the new material will introduce an extra uncertainty. Comparing isotope ratio values across decades is therefore not obvious. As we will see, such transitions have already taken place several times during the last four decades. It seems, however, that the advantages of using relative isotope ratios, by far exceed the disadvantages, as will be discussed in the next sections. The isotopic abundances of the elements stated above and their natural variations are reported elsewhere (Coplen, 2001b; Coplen et al., 2002).
40.1.2 Historical reporting scales As discussed before, commonly isotope ratios are reported as 6 values using arbitrary reference isotope ratios. The 6 value of the reference is zero by definition (equations [40.1] and [40.2]). Careful investigations were carried out in the early days of stable isotope work in the 1950s and 1960s to define useful reference materials. As a main criterion, the selected reference should represent a major pool of the element under consideration. Therefore, a logical choice for hydrogen and oxygen stable isotope ratios could be ocean water. Indeed, a well-mixed mean ocean water was proposed as the reference in view of the rather small variations in isotopic composition throughout the oceans. This purely hypothetical water was called SMOW (Standard Mean Ocean Water) (Craig, 1961) and its isotopic composition was defined in terms of an isotopic difference from an existing water (NBS 1). This is a good example for a scale based on a virtual material, not realised in nature. SMOWs proposed isotope ratio was compiled by averaging the isotope ratio values of measurements of different sea water compartments available at that time (McKinney et al., 1950; Horibe & Kobayakawa, 1960; Craig, 1961). Unfortunately the same term SMOW was used for a water subsequently prepared by H. Craig, Scripps Institution of Oceanography, La Jolla, USA, and which was isotopically adjusted to match that virtual reference SMOW, as described above. Just to mention, even a third SMOW concept existed, related to a Rose quartz sample. The IAEA then decided to rename the water prepared
International Stable Isotope Reference Materials
879
by H. Craig to VSMOW (Vienna SMOW) in order to remove the confusion. Since that time this material VSMOW is used as primary reference material to define the hydrogen and oxygen isotope ratio scale. Similarly, for stable carbon isotopes, the largest and relatively homogenous carbon reservoirs are marine carbonates. Therefore, the carbon reference was chosen to be the carbonate of the rostrum of a Cretaceous belemnite (Belemnitella Americana) collected in the Peedee formation of South Carolina, USA. This reference was called PDB (Peedee Belemnite) and was used for reporting carbon isotope ratios and additionally as reference for reporting oxygen isotope ratios in carbonates. For nitrogen, atmospheric N2 is the only logical, and nearly ideal, choice, but it needs purification from other gases before being usable. For sulfur, several references were proposed, the most common one was Canyon Diablo Troilite (CDT) iron sulfide (FeS) from an iron meteorite. It was chosen due to the expected primordial isotopic composition in the meteorite sulfur. The major drawback of all these early references (except air N2) was the limited quantity of the selected material physically available to realise a calibration of these isotope-ratio scales and the problem of material homogeneity and long term preservation. Indeed, PDB and CDT have been exhausted for three decades; and even worse, as indicated above, SMOW was just a virtual reference material and never existed physically. The developments associated with the historical isotope-ratio scales are discussed in section 40.3. Also discussed are the consequences and problems of achieving consistency in reporting data using conventional scales, and the problem of transferring the calibration to another successor material and successor scale.
40.1.3 Measurement techniques The main international efforts on standardization of stable isotope ratio measurements so far have concentrated on problems of mass spectrometric methods. Up to now, mass spectrometry has been used almost exclusively for the analysis of stable isotope ratios at environmental levels. Concerning the elements H, C, N, O and S, until the last decade, most measurements on gas samples were performed by using the dual-inlet configuration for mass spectrometers. That involved considerably large amounts of samples being prepared, converted into suitable gases and transferred into steel bellows for inlet into the mass spectrometer. The great advantage was the direct comparison of each sample with one gas sample used as transfer gas or working standard over one full measurement day, which facilitated to achieve a high measurement precision for each individual sample. During the last decade, continuous flow techniques emerged, allowing the preparation of very tiny sample amounts and transfer via a helium carrier gas into the mass spectrometer. There, new on-line sample preparation techniques have been developed
880
Chapter 40- M. Gr6ning
to combust samples and to separate effectively the different gases produced during that process via gas chromatographic techniques. New requirements for reference materials emerged on issues like homogeneity and available compounds, due to the smaller sample sizes used and the different standardisation technique applied (injection of reference gas only before or after the samples, no direct simultaneous measurement possible; treatment and behaviour of reference materials should be similar as for samples). Several non mass spectrometric analytical methods are applied for isotopically enriched materials, using often cheaper instruments and robust analytical techniques, but with lower sensitivity (Roth, 1997). These methods are of great advantage, where ultimate precision is not required, and often provide faster and easier measurement possibilities. Even for isotopic compositions at environmental level with their rather small variations in isotope ratios, some other promising techniques are emerging, such as optical techniques using infrared absorption spectra (Kerstel et al., 2001) and the optogalvanic effect (Murnick, 2001). Since the requirements in terms of standardisation and calibration are somewhat comparable and similar to mass spectrometric techniques, no distinction is made in the following considerations on those different analytical techniques. 40.2 Terminology for different kinds of reference materials In the literature different definitions and terms are used for internationally distributed reference materials and for local laboratory standards. Different authors often use the same terms in varying context and meaning, and some expressions are used in an ambiguous manner (e.g. the term 'standard'). So far no clear guidelines exist on the definitions for reference materials to be used in the field of stable isotope ratio measurements. The ISO International Vocabulary of Basic and General Terms in Metrologyl provides a consistent general set of definitions, but unfortunately is not in agreement with well established terms used throughout the stable isotope measurement community, consistently causing misunderstandings.
An effort was made originally to define the various categories of reference materials (Gr6ning et al., 1999) in accordance to the practice in the field of stable isotoperatio measurements (Gonfiantini et al., 1995). The terms below are a slightly modified and updated version of those in Gr6ning et al. (1999). Some examples on real materials are given to illustrate the definition. Those definitions used for the various kinds of materials should be clearly distinguished from each other:
Primary reference material (or international standard 1).
a natural, synthetic or virtual
material, which, by general agreement, serves as the substance against which iso1. International Vocabulary of Basic and General Terms in Metrology (1983), International Organization for Standardization, 2nd ed., Geneva, Switzerland.
International Stable Isotope Reference Materials
881
tope ratios of samples are expressed. In the context of this publication, it is used to define a conventional scale (arbitrarily defined by agreement) for reporting variations of stable isotope ratios. For most of these materials information on their stable isotope abundances as molar fractions are available. Recently, the term 'primary conventional reference material' was suggested to point to the nature of the defined scale, being conventional due to the artefacts used for their definition. Example" 62H: VSMOW, 613C: VPDB. Note: For 613C the primary reference material is a virtual (non existing) material !
Calibration material (Gonfiantini et al., 1995) (or primary standard 1): a natural or synthetic compound, which has been carefully calibrated against the primary reference material. It is used in case the primary reference material is not available to calibrate measurements and instruments (or as substitute if the primary material is not existing at all). Each physical existing primary reference material can be referred to as a calibration material as well. Example" 62H: VSMOW, 613C: NBS 19. Note: VSMOW is both calibration material and primary reference material, NBS 19 is not a primary reference material, as ~13C values are expressed versus VPDBisotopic composition as defined zero-point of the 6-scale.
Reference material (RM): a natural or synthetic compound which has been carefully calibrated against the primary reference material and property values of which are sufficiently homogeneous, well established, and associated with determined uncertainties. It is used to calibrate laboratory equipment and measurement methods for analysis of materials of a composition different from that of the primary reference material. The available reference materials cover a broad spectrum of chemical compositions and a wide range of stable isotope ratios. Most existing stable isotope ratio reference materials were investigated in interlaboratory comparison exercises and were first distributed as intercomparison materials. For the reporting of stable isotope measurements, the isotopic value of the reference material used for their calibration should be stated. Example" 62H and ~13C: several materials, see Tables 40.3, 40.5, 40.6, 40.7, all calibrated against respective primary reference materials.
Certified reference material: see under reference material, with the additional requirement of a detailed description of its calibration and measurement in a certificate and certification report. Example: BCR-IRMM materials in Table 40.3.
Intercomparison material (quality control material)" a homogeneous natural or synthetic compound that provides the means to check the overall quality of measurements performed in comparison with that of other laboratories. It is useful to identify the relative bias of results between laboratories. The second term is used at IAEA for ongoing distribution of materials having been used previously in interlaboratory comparison exercises, since they are used like a quality control material (see next term below).
882
Chapter 40- M. Gr6ning
Example: 82H and 813C" various test materials circulated in interlaboratory comparison exercises, since it is used like a quality control material (see below) (e.g. IAEA-OH-1 to IAEA-OH-8 water samples).
Quality control material (or internal standard / internal laboratory standard) (or reference standard1): This term describes a homogeneous material being of similar composition as normal samples, but carefully calibrated versus reference materials and used routinely day by day in a laboratory to transfer the calibration to samples and to check measurements and the measuring instruments. The second term 'internal standard' is commonly used in stable isotope mass spectrometry. Instead of the expression 'internal' similar terms like 'local' are also used. Example at IAEA: 62H: ST7 to ST10 water samples calibrated against VSMOW, 813C: Carrara marble calibrated against VPDB. 1)- The first term is to be preferred, describing an arbitrarily chosen gas used as a reference for analysis of isotope ratios of samples in a dual-inlet mass spectrometer (see section 40.1.3). The second term 'working standard' is well established in stable isotope mass spectrometry, but is somewhat misleading since it is not describing at all a real 'standard'. All dual-inlet measurements of prepared samples and reference materials are made relative to this transfer gas and results are later converted and expressed on an international accepted 8-scale. For continuous flow measurements, these terms should not be used, but rather the expressions as stated directly below.
Transfer gas (or working standard) (or transfer standard
Reference gas (or reference injection gas) (or laboratory reference gas)" a gas used as a reference for analysis in continuous flow isotope ratio mass spectrometry. The gas of known isotopic composition is injected in the carrier gas stream alternate to sample gas.
Standard: The single term 'standard' is a quite ambiguous expression and is not well defined due to its application for many different purposes. Its use should be strictly limited, and used only exceptionally as general expression for a material meeting one of the classifications above. The more specific terms as stated above should be strongly preferred and used whenever possible. According to the Commission on Atomic Weights and Isotopic Abundances of the International Union Standard of Pure and Applied Chemistry (IUPAC) the use of the term "standard" should be reserved in chemistry solely for use in standard states and standard deviation (oral communication).
40.3 International stable isotope reference materials The presentation here focuses on 'international' reference materials being easily accessible and relevant on the international scale. It discusses available reference materials with property values traceable to internationally agreed scales. No attempt 1. International Vocabulary of Basic and General Terms in Metrology (1983), International Organization for Standardization, 2nd ed., Geneva, Switzerland.
International Stable Isotope Reference Materials
883
was made to evaluate comprehensively the market of suppliers and stable isotope reference materials to its whole extend. Many materials are available from providers, where the direct link to the international scales (traceability, see Chapter 42 in this Volume) is not evident, i.e. not fully documented and not accompanied by reliable uncertainty statements. Those materials are not considered here. Please note that so far none of the IAEA stable isotope reference materials is called 'certified' since not all the ISO requirements for certification are fulfilled. 40.3.1 Common characteristics of available stable isotope reference materials The reference materials discussed here can be classified in two categories: first synthetically produced or refined substances, such as carbonates, sulfates, sulfides, nitrates, graphite, sugar or polyethylene, being chemically pure; secondly, natural materials like distilled water, carbonate rock, silicates, refined oil, cellulose and similar compounds, selected and tested for their purity and isotopic homogeneity.
Most of these materials have been prepared with the intention of using them as reference materials. Therefore, much care was taken in the initial purification and homogenisation of the raw materials. In most cases, the recommended values for stable isotope ratios in these materials were determined by interlaboratory comparison exercises. For some of the materials issued twenty or more years ago, the number of participating laboratories was rather limited (in some cases, less than ten), constraining the validity of any statistical evaluation. All of those materials were originally prepared for use with dual-inlet mass spectrometers with off-line sample preparation systems using relative large sample amounts. Therefore, an urgent need exists to assess the homogeneity of those materials at the sub-milligram level, which is the usual sample size for modern continuous flow systems. In this contribution, the recommended isotopic 6 values for available internationally distributed reference materials are reported with associated standard uncertainties as derived during the reference material calibration, in most cases from the reported uncertainties of individual laboratories. All uncertainties are reported at the l o-level. However, limited information is on record on the uncertainty assessment methods used for data reported by laboratories more than a decade ago. The recommended 6 values for the reference materials, as stated in this contribution, are derived from published literature with a few exceptions of data from publications in preparation. All available raw data for existing and currently distributed reference materials have been compiled and re-evaluated using a consistent approach for identifying and excluding outliers. This statistical approach, called 'exclusive sigma-test,' is a variation of the well known 2o-outlier test and was first suggested by H. Meijer, CIO, University of Groningen, Netherlands, during an IAEA experts meeting on stable isotope reference materials in the year 2000 in Vienna. A detailed description of this outlier test method will be given elsewhere together with a discussion on the resulting (slight) changes for some recommended isotope ratio values.
884
Chapter 4 0 - M. Gr6ning
Most of the reference materials discussed in this chapter are distributed by the International Atomic Energy Agency (IAEA), the U.S. National Institute of Standards and Technology (NIST) and the EU Institute for Reference Materials and Measurements (IRMM). The IAEA and NIST have cooperated since the 1960s in the distribution of stable isotope reference materials by sharing the available materials for the elements hydrogen, lithium, carbon, nitrogen, oxygen, sulfur, and silicon. In Table 40.2, the corresponding names for the reference materials, as used by IAEA and by NIST, are presented. 40.3.2 Hydrogen and oxygen stable isotope reference materials
40.3.2.1 Historical development of scales Hydrogen and oxygen are discussed in a common section due to the correspondence for water stable isotope ratio reference materials. First the historical development is presented on the establishment of the scales for reporting hydrogen and oxygen isotope ratios, primarily intended for measurements on water samples and later extended to other hydrogen and oxygen bearing materials. By 1953 "average ocean water" was suggested and used as a reference point for isotope-ratio measurements (Epstein & Mayeda, 1953). Because no 'average ocean water' existed, H. Craig refined this concept in 1961 by defining the hypothetical Standard Mean Ocean Water (SMOW) as zero-point of that conventional scale in terms of Table 40.2 - Synonyms used by IAEA and NIST for jointly distributed stable isotope reference materials. Other materials not listed in the table are not distributed by both organisations. The following reference materials are also known under synonyms: IAEA-S-4 as "Soufre de Lacq", IAEA-CH-6 as "Sucr.Anu', IAEA-CH-7 as "PEF-I", IAEA-S-1 as "NZI", IAEA-S-2 as "NZ2". IAEA-name
NIST-name
compound
IAEA-name
NIST-name
compound
VSMOW GISP SLAP NBS 30 NBS 22 IAEA-CH-7 USGS24 IAEA-CH-6 NBS 18 NBS 19 LSVEC NBS 28 IAEA-N-1 IAEA-N-2 IAEA-NO-3 USGS25 USGS26
RM RM RM RM RM RM RM RM RM RM RM RM RM RM RM RM RM
water water water biotite oil polyethylene graphite sucrose carbonatite limestone Li2CO3 silica sand (NH4)2SO4 (NH4)2SO4 KNO3 (NH4)2SO4 (NH4)2SO4
NSVEC IAEA-S-4 IAEA-S-1 IAEA-S-2 NBS 123 NBS 127 USGS32 NGS1 NGS2 NGS3 USGS34 USGS35 USGS40 USGS41
RM RM RM RM RM RM RM RM RM RM RM RM RM RM
nitrogen gas sulfur Ag2S Ag2S sphalerite BaSO4 KNO3 hydrocarbon gas hydrocarbon gas hydrocarbon gas KNO3 NaNO3 L-glutamic acid L-glutamic acid
8535 8536 8537 8538 8539 8540 8541 8542 8543 8544 8545 8546 8547 8548 8549 8550 8551
8552 8553 8554 8555 8556 8557 8558 8559 8560 8561 8568 8569 8573 8574
International Stable Isotope Reference Materials
885
a real reference water (Craig, 1961). Its isotopic composition was defined as a weighted average of the available measurements of the isotopic composition in the main oceanic water masses. But because SMOW was just the concept of a hypothetical water and never existed as a real water sample, it couldn't be used directly for calibration of laboratory measurements. Thus, the isotopic ratios of SMOW were defined with respect to an existing water standard1 NBS-1 (Mohler, 1960) distributed by the US National Bureau of Standards. This water was used earlier for an water stable isotope interlaboratory comparison. So for the first time a physically existing material was used to calibrate different laboratories to the SMOW scale. NBS-1 water was readily available for worldwide distribution together with a second water standard1 called NBS-1A, which was obtained from melted snow with a lower abundance of the heavier isotopes. During an IAEA interlaboratory comparison in 1965, serious doubts were confirmed concerning the preservation of the NBS-1 water standard and possible changes of its isotopic composition over time. For this reason, it was unsuitable to be used as standard anymore. At an IAEA Panel Meeting in 1966, it was therefore recommended to establish a pair of two new primary water reference materials, the first one being as close as possible to the defined SMOW and the other one with an abundance of the heavier isotopes close to the lowest limits observed in natural water. The new material with an isotopic composition as close as possible to SMOW was prepared by R. Weiss and H. Craig, Scripps Institution of Oceanography, La Jolla, USA. It was obtained by mixing distilled ocean water with small amounts of other waters in order to adjust its isotopic composition as close as possible to that of the defined SMOW. This task was complicated due to the required adjustment of both the isotopic composition of hydrogen and of oxygen. This reference water was ready in 1968 and was called Vienna Standard Mean Ocean Water (VSMOW). According to the control analyses performed by H. Craig, VSMOW has the same 180/16 0 ratio as the defined SMOW, but a slightly lower 2H/1H ratio or respectively a slightly negative 62H value (-0.2%o) relative to SMOW. However, this slight difference is about a factor of four to five lower than the measurement uncertainty of most laboratories. Absolute isotope ratios of VSMOW were determined for 180/160 (Baertschi, 1976), for 170/16 0 (Li et al., 1988) as well as for 2H/1H (Hagemann et al., 1970; De Wit et al., 1980; Tse et al., 1980). The second water reference material was obtained by E. Picciotto, Universit6 Libre de Bruxelles, Begium, from melting a firn sample at Plateau Station, Antarctica. This material was named Standard Light Antarctic Precipitation (SLAP). The absolute isotope ratios of SLAP were determined only for 2H/1H (Hagemann et al., 1970; De Wit et al., 1980; Tse et al., 1980). Absolute 180/160 ratios for SLAP in this report are com1. Here the general and vague term 'standard' is used, since the exact status in the 1960s is not anymore known for the two materials NBS-1 and NBS-1A. This also applies to some other materials mentioned in the text.
Chapter40- M. Gr6ning
886
puted from those of VSMOW and by using its assigned relative 5 value of 5180 = -55.5 %o relative to VSMOW (see below). All four materials, VSMOW, SLAP, NBS-1, and NBS-1A, were then distributed by the IAEA Isotope Hydrology Laboratory (the latter two materials were transferred to the IAEA from the National Bureau of Standards, today the National Institute for Standards and Technology (NIST), Gaithersburg, Maryland, USA). Nowadays, most stable isotope reference materials are distributed both by the IAEA and NIST. In 1976, an IAEA Consultants' Meeting was convened in order to discuss the isotopic results on these reference materials and to advise on future actions on standardization of stable isotope ratio measurements (Gonfiantini, 1978). The recommendation of the experts concerning data of the water reference materials was to express all future results for hydrogen and oxygen isotope ratios as 5 values relative to VSMOW in order to resolve confusion on results expressed in non-corresponding scales. Thus, VSMOW water was recommended to serve as a new primary reference material. It was stated that the coherence between 5 values reported by different laboratories could be improved by adopting a fixed 5 value for a second water reference material. The experts recommended that SLAP be adopted for this purpose and that both the 180/160 and 2H/1H 5 values be normalized relative to VSMOW water, the primary reference material. In Table 40.3 the 5 values of selected reference materials are listed with those of VSMOW being by definition at zero and those of SLAP established by assessing the close agreement of three direct determinations of hydrogen isotope mole fractions as well as the weighted mean of relative isotope ratio measurements of different laboratories. The definition of VSMOW as zero-point for the oxygen and hydrogen 5 scales and the adoption of fixed 5 values for SLAP is therefore a slight modification of the original definition of the 6 scale in equation [40.2]" -
RSAMPLE-RvsMow" RVSMOW
(~3SLAp/RSLAp~RvsMow I 1000%o •
RVSMOW
[40.3]
,
with 5 referring to ~2H and 6180, R being the corresponding 2H/1H and 180/160 ratios and the additional term in the bracket being the normalization of the respective VSMOW-scale in terms of pre-defined hydrogen and oxygen isotope ratios of the two primary reference materials (Table 40.3) (Gonfiantini, 1981). The two scales defined in equation [40.2] for SMOW and in [40.3] for VSMOW coincide only if RSMOW - RVSMOWand if the adopted 5 values for SLAP in Table 40.3 correspond to the true ones as defined by equation [40.2] for both hydrogen and oxygen. From the reported measurements of NBS-1 and VSMOW, a slight offset of the zero-point of the two scales could be concluded (offsets of 0.05%0 and 0.5%0 for 5180 and 62H, respectively), but these offsets are well within the limits of measurement
887
International Stable Isotope Reference Materials
Table 40.3 - Oxygen and hydrogen 6-values versus VSMOW assigned to the existing major water reference materials with associated standard uncertainty at lo-level (where applicable). Available materials are marked in bold. 6-values marked with * are reported versus SMOW. Given references are valid both for 6180 and 62H. CM stands for "Calibration material" and RM for "Reference material". Name
Material
Status
Distribution
6180 [%o]
VSMOW SLAP GISP
water water water
CM CM RM
IAEA, NIST IAEA, NIST IAEA, NIST
0 -55.5 -24.78_+0.08
NBS-1
water
62H [%o] 0 -428 -189.73+0.87
Gonfiantini et al. (1995) discontinued
-7.94*
-47.6*
Craig (1961) -7.89+0.12
-47.1+1.2
Gonfiantini (1978) NBS-1A
water
discontinued
-24.33*
-183.3*
Craig (1961) -24.29+0.25
-183.2+0.7
Gonfiantini (1978)
uncertainty of most laboratories (Gonfiantini, 1977). Both offsets were a bit larger than evaluated before by H.Craig. However, due to the scatter of the individual results, the stated mean offset should not be applied for conversion from one scale to the other. A third water reference material was proposed during the same meeting in 1976 (Gonfiantini, 1977) with an isotopic composition intermediate between that of VSMOW and SLAP. This material was obtained by W. Dansgaard, University of Copenhagen, Denmark, from Greenland firn in 1978 and was called GISP (Greenland Ice Sheet Precipitation). GISP is intended to demonstrate successful calibration performed with VSMOW and SLAP. Results of two interlaboratory comparisons investigating GISP are published in IAEA reports (Gonfiantini, 1977; Gonfiantini et al., 1995). It was noted, that the coherence of results from different laboratories improved by a factor of more than two, when data were normalized using SLAP as second primary reference material. The normalization of oxygen and hydrogen isotope ratio data using the pair of VSMOW and SLAP primary reference materials was recommended by an IAEA panel group (Gonfiantini, 1977) and later by the Commission on Atomic Weights and Isotopic Abundances of the International Union of Pure and Applied Chemistry (IUPAC) (Coplen et al., 1996). 40.3.2.2
Necessary changes of the VSMOW/SLAPscale
The original prepared amount of SMOW was about 70 litres in 1968. Today (in the year 2003), only about 10 litres in total remain at the two distributing institutions, the International Atomic Energy Agency in Vienna, Austria, and the National Institute of Standards and Technology in Gaithersburg, Maryland, USA. In December 1996, a small meeting was convened at IAEA to discuss the possibilities of preparing a successor material before the stocks of VSMOW were totally
888
Chapter 40- M. Gr6ning
exhausted. The unanimous recommendation was to try to prepare a successor material isotopically indistinguishable from VSMOW within measurement uncertainty of the best analytical techniques used for routine measurements, in order to minimize any problem of scale conversion. It was decided to prepare the replacement as a mixture of three distilled natural waters, having an initial isotopic composition close to that of VSMOW. The use of water enriched in 180 was rejected due to the aim to reproduce not only 62H and 6180, but also ~170 values. Materials enriched in 180 often show a non mass-dependant fractionation of 170 versus 180. It took three years until sufficient amounts of three suitable water samples were supplied to the IAEA from Lake Bracciano, Italy (G.-M. Zuppi, University of Venice and R. Gonfiantini, Istituto di Geoscienze e Georisorse, Pisa), from Lake Genezareth, Israel (E. Adar, Ben-Gurion University), from a well near Cairo, Egypt (F. Hussein, University of Cairo and I. Nada, Egyptian Atomic Energy Authority, Cairo). All three samples were initially distilled at the IAEA. A careful calibration of the raw waters made directly against VSMOW was performed by five laboratories (T. Coplen, USGS Reston, Virginia, USA; H.A.J. Meijer, CIO Groningen, Netherlands; W. Stichler, GSF Neuherberg, Germany; R Dennis, UEA, Norwich, UK; IAEA IHL Vienna, Austria). The results of all five stable isotope laboratories were in accordance by better than _+0.02%0 for ~180 and by better than +0.3%0 for ~52H(see Figure 40.1). As a result of the calibration, it was decided to produce an amount of 300 litres of this new material. Provisionally, the working name for the new material is set to "NEW VSMOW", subject to change after completion of the production. Special stor-
Figure 40.1 - Calibrated measurements of five laboratories for three raw waters used for preparation of a mixture as a replacement for VSMOW (VSMOW being at the origin of the coordinate cross) and the result of the mixing of the "NEW VSMOW" (big square close to the coordinate origin of the plot), see data in Table 40.4.
889
International Stable Isotope Reference Materials
age containers with a volume of 300 litres were purchased to enable the transfer of water in and out of the container without contact to the atmosphere in order to avoid isotopic fractionation of the water. Due to the particular isotopic compositions of the available three water samples, direct mixing of such a large amount of NEW VSMOW was not possible. Therefore, a deuterium enrichment method was specially designed to enrich several litres of the Egyptian water in its deuterium content without disturbing its natural oxygen isotopic composition. This resulted in the 62H value of this sample water being enriched in 2H by about 40%0, while no significant shift in 6180 was observable (Figure 40.1). Finally appropriate fractions of the three raw waters were mixed gravimetrically. The isotopic composition of the mixture "NEW VSMOW" was assessed by three laboratories (T. Coplen, USGS; H.A.J. Meijer, CIO; IAEA Isotope Hydrology Laboratory) by directly measuring aliquots of NEW VSMOW against aliquots of VSMOW. For 62H determinations, both the equilibration method and the zinc reduction method were used. The results are shown in Table 40.4 and in Figure 40.1. Details of the preparation of this material will be presented elsewhere. The results show no significant deviation within the stated estimated standard error (standard deviation divided by square root of measurements) for all performed measurements in the three laboratories. It is believed that virtually no laboratory with current state-of-the-art techniques will be able to detect this deviation in its routine measurements. Meanwhile, as second project, the preparation of a successor material for SLAP has started. After some efforts, two water samples substantially depleted in 2H and 180 were obtained from the U.S. Antarctic Station at the South Pole (initiated by T.B. Coplen, USGS, Reston, Virginia, USA, with kind support of the U.S. NSF Polar Sciences Program) and from the Antarctic Vostok drilling program (water supplied by J.R. Petit, LGGE-CNRS, Grenoble, France). The two samples are just under calibration at the IAEA. With the experience gathered during the NEW VSMOW production, it is believed that one can reproduce SLAP with only a small difference in isotopic composition from that of SLAP. At least 200 litres of water are planned to be produced from these samples. The provisional working name "NEW SLAP" will be used for this material during its production process. Table 40.4 - Estimate of the deviation of NEW VSMOW from VSMOW for 62H and 6180, expressed in per mill vs. VSMOW for a series of measurements using three mass spectrometers at the IAEA Isotope Hydrology Laboratory. Deviation of NEW VSMOW vs. VSMOW [%0]
6180 62H
0.002 -0.12
Estimated standard error of the mean [%o] for NEW VSMOW (no. of analyses) +0.007 (125) +0.09 (118)
Estimated standard error of the mean [%0] for VSMOW (no. of analyses)
+0.006 (109) +0.08 (1~5)
890
Chapter 40- M. Gr6ning
After c o m p l e t i o n of the p r o d u c t i o n a n d bottling of b o t h materials, it is p l a n n e d to p r e s e n t all the results to the scientific community. For the future it is e x p e c t e d to realize any calibration relative to the established V S M O W / S L A P scale by u s i n g the n e w pair " N E W V S M O W " / " N E W SLAP" by assigning precise ~5 values to them. It is h o p e d that the a g r e e m e n t a n d m a t c h i n g of the t w o scales will be convincing, so that no long-lasting debate on the scale issue will e m e r g e as it w a s the case in the p a s t during the transition from S M O W to VSMOW. As an a d d i t i o n a l measure, an absolute calibration of b o t h n e w materials for their h y d r o g e n a n d o x y g e n isotopic c o m p o s i t i o n (isotope mole fractions) s h o u l d be u n d e r taken. Suitable m e t h o d o l o g i e s a n d institutions will h a v e to be identified for this p u r pose. Ideally, this w o u l d p i n d o w n the u n c e r t a i n t y on the isotopic a b u n d a n c e of N E W V S M O W a n d N E W SLAP a n d allow to m a i n t a i n a h i g h degree of consistency of 6180 a n d 62H scales, regardless of the calibration materials used.
40.3.2.3 Other hydrogen and oxygen stable isotope reference materials In Table 40.5 a few m o r e reference materials u s e d for 62H a n d 6180 analyses are listed. Two of these materials, NBS 28 silica s a n d a n d NBS 30 biotite, w e r e p r e p a r e d b y I. F r i e d m a n of the U.S. Geological Survey. They p r o v i d e a link to &180 m e a s u r e m e n t s on silicates. A l t h o u g h no absolute silica isotope a b u n d a n c e m e a s u r e m e n t of NBS 28 has b e e n p e r f o r m e d , NBS 28 has served for several decades as the p r i m a r y reference material for relative 30Si/28Si m e a s u r e m e n t s ; thus, 630SINBS28 = 0%o. H y d r o g e n isotope ratios for NBS 30 biotite s h o u l d be d e t e r m i n e d on the w a t e r fraction (3.5% b y weight). Table 40.5 - Recommended 6-values relative to VSMOW for additional oxygen and hydrogen reference materials with associated standard uncertainties at lo-level. Available materials are marked in bold. For references see text. RM denominates "reference material", CRM stands for "Certified reference material" Name
Material
Status
Distribution
6180 [%o]
62H [%o]
630Si
NBS 28 NBS 30 IAEA-302A
silica sand biotite water
RM RM RM
IAEA,NIST IAEA,NIST IAEA
+9.58 + 0.09 +5.24 + 0.25 -
-65.70+ 0.27 508.4
0 -
IAEA-302B
water
RM
IAEA
-
Parr & Clements (1991)
996
-
Parr & Clements (1991)
IAEA-304A
water
RM
IAEA
251.7
-
-
-
-
-
-
-
-
Parr & Clements (1991)
IAEA-304B
water
RM
IAEA
502.5 Parr & Clements (1991)
BCR-658
water from wine
CRM
BCR(IRMM)
-7.19+0.04 Guillou et al. (2001)
BCR-659
water from wine
CRM
BCR(IRMM)
-7.18+0.02 Guillou et al. (2001)
International Stable Isotope ReferenceMaterials
891
The two reference material sets, IAEA-302 and IAEA-304, consist of two reference materials each, both isotopically enriched at different levels with deuterium and 180, respectively (Parr & Clements, 1991). These materials were prepared by P. Klein, USDA/ARS, Houston, Texas, USA and I.I. Lesk, MSD Canada Ltd., with the aim of providing reference materials for medical and biological applications and for applications needing materials enriched in 180 and 2H. Two certified reference materials, BCR-658 and BCR-659 synthetic wines, were produced by M. Lees, Eurofins Scientific, Nantes, France, in cooperation with C. Guillou and G. Remaud, EC Joint Research Centre, Ispra, Italy, in order to provide the means for proper analysis of wines in the European Community according to officially approved methods (Guillou et al., 2001). Both materials consists of water - ethanol mixtures (7 and 12 %vol.) with added chemicals to mimic the composition of wine. The parameter to be certified was the 6180 composition in the water phase. Recently, two more organic materials were prepared as candidates for ~)180 reference materials. These are benzoic acids of different oxygen isotopic compositions (one of them is enriched in 180), produced within a joint project of A. Schimmelmann, University of Indiana, USA, and W. Brand & R. Wernerl, Max Planck Institute for Biogeochemistry, Jena, Germany. Those materials are prepared with the aim to serve as reference materials for 6180 analysis using online combustion techniques. Initial isotopic measurements were performed using both on-line and off-line combustion techniques. They resulted in provisional isotope values, which have to be verified and further refined by calibration measurements performed in other experienced laboratories. Two nitrate materials were prepared at USGS, Reston, USA by J.K. B6hlke and T. Coplen. One is depleted in 170 and 180 with normal 170/180 ratios. The other is enriched in 170 and 180, with anomalously high 170 content. Those materials could be used for the normalization of oxygen-bearing substances and for calibration of nitrogen-bearing materials, especially atmospheric materials that show 170 anomalies.
40.3.3 Carbon (plus oxygen and hydrogen) stable isotopic reference materials The number of stable isotope analyses of carbon-bearing materials has increased drastically with the development of new mass spectrometric systems and analytical techniques, involving the use of elemental analyzers, gas chromatographic columns and continuous flow systems. Isotope techniques are incorporated in an increasing number of scientific disciplines and new applications emerge continuously. Therefore, suitable isotopic reference materials are requested continuously. The variety of applications has resulted in a change of the priorities for new materials and demanded an improved characterization of existing ones. The assessment of the isotopic homogeneity especially needs further attention due to the smaller and smaller amounts needed for sample analysis in modern on-line mass spectrometric techniques. 1. Now at ETH, Ziirich, Switserland
892
Chapter 40- M. Gr6ning
40.3.3.1 Inorganic carbon isotopic reference materials In the early 1950s, 6180 measurements were often expressed against the so-called PDB primary reference material. The PDB material originated from the CaCO3 of the rostrum of Cretaceous belemnites (Belemnitella Americana) collected in the PeeDee formation of South Carolina, USA. The CO2 obtained from the reaction of PDB with 100% H3PO4 (McCrea, 1950) was used for calibration of oxygen isotope paleotemperature measurements (Epstein et al., 1953) and for carbon isotope ratio variations in natural compounds (Craig, 1953). Its 613C and 6180 values were close to the average isotope ratio values of marine limestone. The PDB primary reference material has long been exhausted. The calibration of measurements with respect to PDB in practice was performed by using the reference material NBS 20 (Solenhofen limestone) (Craig, 1957). Doubts on the isotopic stability of NBS 20, especially for 6180, due to potential exchange of this finely ground material with air moisture and CO2 have lead to the discontinuation in the use of NBS 20. As replacement, the carbonate material NBS 19 was produced. NBS 19 is derived from white marble of unknown geological origin (Friedman et al., 1982). It was provided by I. Friedman, U.S. Geological Survey, using a slab of marble representing the form of a toilet seat, and for this reason it is also referred to as TS limestone. The calibration of NBS 19 relative to PDB was performed using NBS 201. It was recommended at two Advisory Group Meetings organised at the IAEA (Gonfiantini, 1984; Hut, 1987) that the use of the PDB scale be discontinued and that it be replaced by a new scale called the VPDB scale (Vienna-PDB), which would be anchored by assigning a fixed 6 value to NBS 19 carbonate, as stated in Table 40.6. VPDB itself never existed as material, but it is defined through NBS 19 in such a way that the VPDB scale corresponds nominally to the previous PDB scale, i.e. the isotopic values of NBS 19 expressed versus PDB are nominally exactly the same as those expressed versus VPDB (Table 40.6). Because no measurement can be performed without an associated uncertainty, the calibration value of NBS 19 versus PDB (via NBS 20) inherently contains some additional uncertainty. Some problems related to the scale transition are discussed in Coplen (1996c). The reason for introduction of the new scale, was to eliminate this (unknown) uncertainty component for all reported data using NBS 19 as a calibration material. Stating a 613CVPDB or 618OVPDB value, therefore, implies the calibration being performed through the existing NBS 19 calibration material and its defined isotopic values. This minimises confusion on data reporting, as long as the supply of NBS 19 is sufficient (at least for the next 20 years). Complementary to NBS 19, an additional calcite material named NBS 18 is used as reference material with slightly more negative (~13C and 6180 values. NBS 18 is a carbonatite from Fen, Norway, and was collected by B. Taylor, University of California, Davis, USA. It was prepared by H. Friedrichsen, University of Ttibingen, Germany and I. Friedman, J.R. O'Neil and G. Cebula, USGS (Friedman et al., 1982). NBS 23, a strontium carbonate, was prepared by I. Friedman, but its distribution was discontinued due to doubts on its isotopic homogeneity. For studies of methane and other 13C depleted materials the LSVEC Lithium carbonate is used. LSVEC was originally pre1. Note: While the NBS 19 isotopic composition was actually measured versus PDB, the values are defined versus VPDB and therefore are used to anchor the VPDB scale.
893
International Stable Isotope Reference Materials
Table 40.6 - Carbon 6-values versus VPDB of the inorganic carbon isotope reference materials with
associated standard uncertainties at lo-level. Available materials are m a r k e d in bold. 6-values m a r k e d with * are reported versus PDB. References given in the 613C column are relevant for both 613C and 6180 values. CM stands for "Calibration material" and RM for "Reference material". Name
Material
PDB NBS 20
carbonate limestone
Status
Distribution
613C [%0]
6180 [%o]
exhausted exhausted
0* - 1.06"
0* Isoto pi cally exchanged
Craig (1957) -1.08+0.06
Hut (1987) NBS 19 IAEA-CO-1
limestone calcite
CM RM
IAEA, NIST IAEA, NIST
carbonatite
RM
IAEA, NIST
+1.95 +2.48 + 0.03
-2.20 -2.44 + 0.07
Gonfiantini et al. (1995) NBS 18
-5.01 + 0.06
-23.00 +_0.07
Gonfiantini et al. (1995) IAEA-CO-8
calcite
RM
IAEA, NIST
-5.75 + 0.06
-22.67 + 0.19
Gonfiantini et al. (1995) NBS 23
SrCO3
discontinued
-35.32+0.16
-29.98...-30.54
Hut (1987) LSVEC
Li2CO3
RM
IAEA, NIST
-46.48 + 0.15
-26.64 _+0.25
Gonfiantini et al. (1995) IAEA-CO-9
BaCO3
RM
IAEA, NIST
-47.12 + 0.15
-15.28 + 0.09
Gonfiantini et al. (1995) NBS 16
CO2
exhausted
-41.64+0.17*
-25.75_+0.37*
Gonfiantini (1984) -41.59_+0.06
-25.86_+0.35
Hut (1987) NBS 17
CO2
exhausted
-4.48+0.10"
-8.37+0.25*
Gonfiantini (1984) -4.45_+0.05 RM8562
CO2
-8.51+0.15
Hut (1987) RM
IAEA, NIST
-3.76 + 0.03
-8.45 + 0.11
Verkouteren (1999) RM8563
CO2
RM
IAEA, NIST
-41.56 + 0.06
-23.72 + 0.11
Verkouteren (1999) RM8564
CO2
RM
IAEA, NIST
-10.45 + 0.04
+0.19 + 0.10
Verkouteren (1999) NGS 1
Hut (1987) NGS 2
Hut (1987)
NGS 3
Hut (1987)
Natural hydrocarbon gas Natural hydrocarbon gas Natural hydrocarbon gas
NIST
NIST
NIST
(CH4): -28.6...-29.16 (C2H6): -25.42...-26.4 (C3H8): -19.9...-21.67 (C H4): -43.1...-44.92 (C2H6): -31.1...-32.1 (C3H8): -23.5...-25.9 (CO2): -7.5...-8.9 (CH4): -69.6...-73.1 (C2H6): -51.37...-59.8
-
-
-
894
Chapter 40 - M. Gr6ning
pared as lithium isotope reference material by H. Svec, Iowa State University, USA. Because some reference materials were nearing exhaustion, three additional materials were introduced: IAEA-CO-1 (Carrara marble) and IAEA-CO-8 (calcite from the Kaiserstuhl, Germany), both prepared at the IAEA, and IAEA-CO-9 (BaCO3, prepared by C. Brenninkmeijer, IGNS, Lower Hutt, New Zealand). Due to the recent discovery of additional batches of NBS 18 and LSVEC from the original production time, the supply of those reference materials is secured for at least ten more years. NBS 16 and NBS 17, two pure CO2 materials, were prepared by T.B. Coplen and C. Kendall of the USGS in sealed glass ampoules (Coplen & Kendall, 1982). The two materials are exhausted. Several years ago, three new CO2 gas reference materials were prepared by M. Verkouteren, NIST, Gaithersburg, Maryland, USA (Verkouteren et al., 1998; Verkouteren, 1999), to provide a means for standardizing gas measurements without the necessity of any preparation by the carbonate acid reaction. These three materials span a broad isotopic range, covering values from those of atmospheric carbon dioxide to those of methane. These three gaseous reference materials were flame-sealed in glass tubes to ensure their isotopic stability on long time scales. Special care was taken for the filling procedure to avoid any isotope fractionation between individual tubes (see also Chapter 43 in this Volume). Three natural gases, consisting mainly of CH4, C2H6 and higher hydrocarbons, were collected and named NGS1, NGS2 and NGS3. They were stored in 150-ml pressurized steel cylinders (Hut, 1987). NGS1, being related to coal reservoirs, was collected by G. Hut, University of Groningen, Netherlands; the oil related NGS2 and the biogenic NGS3 were collected by T. Ricchiuto, AGIP, Milano, Italy. The three samples were originally available from IAEA and from M. Schoell, Chevron, La Habra, USA. New 50-ml sample cylinders for all three materials were filled recently at NIST, and those materials are now available from NIST. Isotopic values of all of these carbonate and CO2 reference materials versus VPDB are listed in Table 40.6, and reference materials still available are marked in bold. The ~}18OVPDB values may also be expressed versus VSMOW or VPDB-CO2 through the following two conversion formulae (Hut, 1987)" 618OvsMow = (1.0309 x 518OVPDB) + 30.9 618OVPDB_CO2 = (618OVPDB- 10.25) / 1.01025
[40.4] [40.5]
As a general rule, 6180 data of carbonates and of CO2 gas are reported versus VPDB, whereas for 6180 measurements of all other materials, the data should be reported versus VSMOW (Coplen et al., 1996). The author is of the opinion to avoid, whenever possible, reporting 6180 data relative to VPDB-CO2 due to the risk of confusing readers with data reporting relative to VPDB. Exceptions are only measurements on CO2 gas samples, as in atmospheric air, or on laboratory reference gases, which cases have to be clearly marked.
International StableIsotopeReferenceMaterials
895
No normalization is currently suggested for 613C,but it is good laboratory practice to state the isotopic composition of any reference material, which had been analysed with samples. That, in principle, allows any data user to normalize data later on, in case the need arises. A remark has to be given on the phosphoric acid digestion, which is used for preparation of CO2 gas from carbonate samples, as well as from the carbonate reference materials. The oxygen isotopic fractionation between carbonate and evolved CO2 depends on the acid reaction and associated parameters. It was determined experimentally more than three decades ago and later tabulated (Friedman & O'Neil, 1977) as common reference. It has been suggested that this oxygen isotope fractionation factor be re-determined (Gonfiantini, 1984) using normalised scales for the measurement results. This is not achieved yet. In fact several recipes exist for the preparation of the phosphoric acid, its storage and its use (e.g.McCrea, 1950; Urey et al., 1951; Coplen et al., 1983). With increasing acid concentration, its degree of polymerization and water release increases. Therefore, 100% phosphoric acid does not exist as such, and nominally even higher concentrations can be achieved (see also discussion in Volume II, Part 3, Chapter 4-5.1.5). The acid concentration can be controlled through density measurements. Because concentrated phosphoric acid is quite hygroscopic, each exposure of it to air slightly changes its concentration, and affects the oxygen isotopic fractionation of the acid reaction. Several chemicals added to the acid are thought to remove organic impurities (e.g. Cr203, H202, V205); however, no systematic assessments on possible effects of those additives on the resulting oxygen isotopic fractionation are available (see also discussion in Volume II, Part 3, Chapter 4-5.1.3 on the pretreatment of carbonate samples).
40.3.3.2 Organic carbon isotopic reference materials Quite early in the development of carbon stable isotope measurement methods, organic compounds were analysed to investigate the isotopic abundances and the different kinds of isotopic fractionation effects (Craig, 1953). The obvious principle to try to calibrate isotope ratio measurements of a certain compound by a reference material of similar chemical and physical properties (IT p r i n c i p l e - 'identical treatment') (Coplen et al., 1996; Werner & Brand, 2001) is especially relevant and applicable for organic compounds. There, the combustion techniques used to produce CO2 gas differ from the quite different preparation technique for usual carbonate reference materials (acid digestion). By the 1960s additional combustable reference materials were being produced, including NBS 21 (graphite) (Eckelmann et al., 1962) and NBS 22 (mineral oil, by S. Silverman, Chevron Oil Company, La Habra, USA) (Silverman, 1964). Graphite is included in the category organic reference materials due to its similar preparation as for organic materials, substantially different than for carbonate materials. Refinements of the initial measurements of those two materials were performed by interlaboratory studies (Schoell et al., 1983; Coplen et al., 1983). In the following years, additional materials were produced and calibrated, including polyethylene (PEF-1, now named IAEA-CH-7, by H. Gerstenberger and M. Herrmann, Zentralinstitut fuer Isotopen- und Strahlenforschung, Leipzig, Germany) (Ger-
896
Chapter 40- M. Gr6ning
stenberger & Herrmann, 1983), sucrose (Anu.Sucr., now IAEA-CH-6, by H. Polach, ANU, Canberra, Australia) (Hut, 1987) and graphite (USGS24, by T.B. Coplen, USGS, Reston, USA). All of those materials were produced more than two decades ago. Their calibration was performed by off-line combustion methods, using relative large amounts of substance. Newly developed on-line combustion techniques use amounts that are orders of magnitude lower. That imposes questions on the homogeneity of the existing reference materials at such low amount levels. No systematic tests have been reported so far. Only NBS 22 oil is an exception as liquid can be assumed to be homogeneous at all amount levels. Two sets of reference materials were produced 15 years ago consisting each of two materials enriched in 13C for applications in the medical and biological field (Parr & Clements, 1991). These include two 13C-labelled sodium bicarbonates prepared by I.I. Lesk, MSD Canada Ltd. (IAEA-303) and two 13C-labelled UL-D-glucoses, prepared by D. Halliday, Clinical Research Centre, Harrow, UK (IAEA-309). Two certified reference materials, BCR-656 ethanol from wine and BCR-660 hydro alcoholic solution 12% vol., were prepared by M. Lees, Eurofins Scientific, Nantes, France, one certified reference material, BCR-657 glucose powder, was prepared by C. Guillou and G. Remaud, EC Joint Research Centre, Ispra, Italy, in order to provide the means for proper analysis of wines and fruit juices in the European Community according to officially approved methods (Guillou et al., 2001). The parameters to be certified were the ~13C composition of the alcohol or sugar and the site specific deuterium content in the alcohol and/or water phase by deuterium- nuclear magnetic resonance. The isotopic composition of the available organic carbon isotopic reference materials is given in Table 40.7. A steadily spreading number of new applications of stable isotope measurements include more and more different organic compounds and at the same time require standardisation of such measurements using reference materials as similar as possible to the investigated compounds. It is certainly not possible at international level to provide dozens of stable isotopic reference materials for all possible applications, but attempts have been started to produce a limited number of suitable materials. The main criteria in the selection of substances are the stability of the compounds with time, ease of storage, ease of handling and preparation, combustibility, and suitability as reference material for a whole class of substances. Several additional materials are under investigations, including a batch of cellulose that was prepared by IAEA for stable carbon isotope analysis from a larger stock of cellulose by milling it down to a fine powder and homogenizing it. This material
897
International Stable Isotope Reference Materials
Table 40.7 - C a r b o n 6-values v e r s u s VPDB of the available organic carbon isotopic reference materials w i t h associated standard uncertainty at l o-level. References see in text. RM stands for "Reference material", CRM for "Certified reference material". Name
Material
NBS 21
graphite
Status
Distribution exhausted
613C [%o]
62H [%o]
-28.16+0.11
-
Hut (1987 NBS 22 USGS24
oil graphite
RM RM
IAEA, NIST IAEA, NIST
-29.74+0.12
-120+4
Gonfiantini et al. (1995)
Hut (1987)
-15.99+0.11
-
Gonfiantini et al. (1995) IAEA-CH-6
sucrose
RM
IAEA, NIST
-10.43+0.13
-
Gonfiantini et al. (1995) IAEA-CH-7
polyethylene
RM
IAEA, NIST
-31.83+0.11
-100.3+2.1
GonJi"antini et al. (1995) IAEA-303A
NaHCO3
RM
IAEA
93.3
-
Parr & Clements (1991) IAEA-303B
NaHCO3
RM
IAEA
466
-
Parr & Clements (1991) IAEA-309A
UL-D-glucose
RM
IAEA
93.9
-
Parr & Clements (1991) IAEA-309B
UL-D-glucose
RM
IAEA
535.3
-
Parr & Clements (1991) BCR-656
ethanol
CRM
BCR(IRMM)
-26.91+0.07
Guillou et at. (2001) BCR-657
sugar
CRM
BCR(IRMM)
d a t a see in: Guillou et al. (2001)
-10.76+0.04
Guillou et at. (2001) BCR-660
e t h a n o l in w a t e r
CRM
BCR(IRMM)
-26.72+0.09
Guiltou et al. (2001)
d a t a see in: Guillou et al. (2001)
was successfully tested for its 613C isotopic homogeneity (W. Stichler, GSF, Neuherberg, Germany; variability smaller than +0.02%o for samples amounts of about 1-2 mg). A batch of pure caffeine was recently produced by W. Brand & R. Werner, Max Planck Institute for Biogeochemistry, Jena, Germany. It is now at NIST for homogenization and bottling. One non-enriched benzoic acid (natural 6180 level) could potentially also be used as a carbon isotopic reference material. Two L-glutamic acids were prepared by H. Qi and T. Coplen, USGS, Reston, USA, named USGS40 and USGS41. One of them has natural C and N isotopic compositions, and the other is enriched in b o t h 13C and 15N by about 50 per mill. Eventually two batches of oxalic acid at the IAEA could also be milled and homogenized for use as stable carbon isotope reference materials. All of these new materials will need to be carefully calibrated relative to the VPDB scale by direct comparison to NBS 19 derived CO2. At the same time, the isotopic composition of the existing organic reference materials should be re-assessed to try to reduce the uncertainty of calibration values. This can be undertaken only in a carefully planned exercise involving several "high-precision" laboratories measuring the whole suite of organic reference materials and calibrating them versus NBS 19. It is hoped, that after bottling of the candidate materials, such an exercise can be organized
898
Chapter 40- M. Gr6ning
in the near future to improve the calibration of laboratories and their internal laboratory standards. While no normalization has been suggested for reporting of stable carbon isotope ratio data, it is recommended that authors report together with their data, also the values for reference material carbon isotope ratios that they did measure or would have measured had they analyzed them along with their samples. This provides a means to perform calculations on possible data normalization in the future.
40.3.4 Nitrogen stable isotopic reference materials The natural choice as the primary reference material for nitrogen relative isotope ratio measurements is atmospheric nitrogen gas, which seems to be isotopically homogeneous (Mariotti, 1983) with respect to the prevailing analytical precision of most laboratories. Several other reference materials with a wide range of isotope compositions exist and are used to calibrate nitrogen isotope measurements of different compounds (Coplen et al., 2002). These are especially useful for solid or liquid samples to test sample preparation methods and to avoid some problems associated with the purification of nitrogen from air. By the 1950s, G. Junk and H.J. Svec, Iowa State University, USA, prepared a nitrogen gas standard (Junk & Svec, 1958), which was later on split into aliquots in sealed glass tubes by C. Kendall, USGS (Kendall & Grim, 1990) and was named NSVEC. An additional nitrogen gas standard, NBS 14, prepared some decades ago has been exhausted since the 1960s. IAEA-N-1 and IAEA-N-2 are ammonium sulfates, prepared by E. Salati, Centro de Energia Nuclear na Agricultura, Brazil. IAEA-NO-3 is a potassium nitrate (formerly called IAEA-N3) and was prepared by A. Mariotti, Universit6 P. and M. Curie, Paris, France. USGS25, USGS26 (ammonium sulfates) and USGS32 (potassium nitrate) were prepared by J.K. B6hlke, USGS, Reston, USA (B6hlke et al., 1995; B6hlke & Coplen, 1995) by mixing commercially available 15Nenriched compounds with compounds of natural terrestrial abundance to achieve the desired 815N composition. The 815N value stated for USGS32 in Table 40.8 is the mean value derived from the whole data set in B6hlke & Coplen (1995). There is strong evidence that some of those USGS32 data were affected by systematic laboratory offsets, and the recommended value will most probably be adjusted to about +180%o in near future. A series of isotopic reference materials substantially enriched in 15N was prepared for the IAEA within an intercomparison study for medical and biological applications (Parr & Clements, 1991). Three 15N-enriched ammonium sulfates were prepared by E. Fern, Vevey, Switzerland (IAEA-305 set and IAEA-311). Two 15N-enriched urea reference materials (IAEA-310 set) were prepared by H. Faust, Leipzig, Germany. The values of the existing nitrogen isotopic reference materials are presented in Table 40.8.
899
International Stable Isotope Reference Materials
Table 40.8 - Nitrogen 6-values versus air-N2 for the existing nitrogen isotope reference materials and their associated standard uncertainties at l o-level. RM stands for "Reference material". Name
Material
Status
Distribution
615N [%o]
exhausted
-1.18 Kendall & Grim (1990) -2.77+0.05 B~hlke & Coplen (1995) +0.43+0.07 B~hlke & Coplen (1995) +20.32+0.09 B~hlke & Coplen (1995) +4.69+0.09 B~hlke & Coplen (1995) -30.25+0.38 B~hlke & Coplen (1995) +53.62+0.25 B~hlke & Coplen (1995) +179.2+1.3 B~hlke & Coplen (1995) + 180( no rmalized ) B~hlke & Coplen (1995) +39.8+0.25[46] Parr & Clements (1991) +375.3+1.2146] Parr & Clements (1991) +47.2+0.7[46] Parr & Clements (1991) +244.6+0.4[46] Parr & Clements (1991) +4693+29[46] Parr & Clements (1991) -1.8+0.1 B~hlke et al. (2003) +2.7+0.1 B~hlke et al. (2003)
NBS 14
nitrogen gas
NSVEC
nitrogen gas
RM
IAEA, NIST
IAEA-N-1
a m m o n i u m sulfate
RM
IAEA, NIST
IAEA-N-2
a m m o n i u m sulfate
RM
IAEA, NIST
IAEA-NO-3
potassium nitrate
RM
IAEA, NIST
USGS25
a m m o n i u m sulfate
RM
IAEA, NIST
USGS26
a m m o n i u m sulfate
RM
IAEA, NIST
USGS32
potassium nitrate
RM
IAEA, NIST
IAEA-305A
a m m o n i u m sulfate
RM
IAEA
IAEA-305B
a m m o n i u m sulfate
RM
IAEA
IAEA-310A
Urea
RM
IAEA
IAEA-310B
Urea
RM
IAEA
IAEA-311
a m m o n i u m sulfate
RM
IAEA
USGS34
potassium nitrate
IAEA, NIST
USGS35
sodium nitrate
IAEA, NIST
6180 [%o]
+25.6+0.2 B~hlkeet al. (2003)
+25.7+0.2 B~htkeet al. (2003)
-14.8+0.2 B~hlke et al. (2003) +51.5+0.3 B~htke et al. (2003)
Two new materials were recently prepared by J.K. B6hlke, T. Coplen, and S. Mroczkowski, USGS, Reston, USA. They include potassium nitrate USGS34 depleted in 170 and 180 (normal 1 7 0 / 1 8 0 ) and sodium nitrate USGS35 enriched in 170 and 180 with anomalous 170 (B6hlke et al., 2003). USGS40 and USGS41, prepared by T. Coplen and H. Qi, USGS, Reston, Virginia, USA, are two L-glutamic acids, one with a natural 615N level and one isotopically enriched in 15N by about 50 per mill. Additional materials are under discussion to provide a means for better calibration of organic nitrogen-bearing samples. They include thiourea, methionine and Nmethyl anthranilic ester (IAEA, 2001; see therein especially Table 2 in its annex and the report of working group B on biogeochemistry, food and ecology. Also found therein are results of discussions on an N20 gas reference material, mainly for atmo-
Chapter 40- M. Gr6ning
900 spheric and soil gas studies).
Past discussions on scale contractions observed in mass spectrometric measurements result in the recommendation to authors to report sample data together with the isotopic values of nitrogen isotopic reference materials had they been analyzed with the samples. When reporting values of 15N depleted or enriched reference materials used for the calibration of internal laboratory standards, one can provide information that can be efficiently used for re-assessing those data later on for any kind of normalization. 40.3.5 Sulfur stable isotopic reference materials The early choice of meteorite materials as reference materials for the calibration of sulfur stable isotope ratio measurements was certainly understandable because they represent the primordial sulfur composition as the average of terrestrial sulfur inventory. But this selection caused considerable problems due to chemical impurities, because of isotopic inhomogeneities, and due to the choice of different reference materials in different countries. By 1960, the Canyon Diablo Troilite CDT (FeS phase from a large octahedrite iron meteorite, Meteor Crater, Arizona, USA) had been adopted as the primary reference material (McNamara & Thode, 1950; Jensen & Nakai, 1962); nevertheless, Russian scientists continued to use the Shikote Alin meteorite (Robinson, 1995). The supply of CDT has been exhausted for more than a decade1, but this material was still used to define the internationally accepted stable sulfur isotope ratio scale (634SCDT) until recently.
After problems with the meteorite reference materials were recognized and discussed (Nielsen, 1984) (see also latest confirmation of this by SF6 measurements of CDT: Beaudoin et al. (1995)), a mineral sphalerite was introduced, but it was also found to be inhomogeneous. Several other approaches preparing additional natural sulfur-bearing materials were not quite successful. OGS, a raw precipitated BaSO4 from sea water (by Y. Horibe, University of Tokyo), caused quite early doubts on its homogeneity. The same occurred with Soufre de Lacq, an elemental sulfur that was derived from natural gas (later renamed to IAEA-S-4), provided by E. Roth, CEN, Saclay) France, which also caused doubts on its homogeneity. Distribution of all three reference materials was subsequently terminated by the IAEA. Distribution of NBS 122, a sphalerite supplied by S. Halas, University of Lublin, Poland, which was recognized to be inhomogeneous, was also terminated. Recent investigations on IAEA-S-4 (Soufre de Lacq) by the USGS (Qi & Coplen, 2003; Carmody & Seal, 1999) indicate that this material indeed is homogeneous and interlaboratory discrepancies may rather have been caused by laboratory offsets. Therefore the distribution of IAEA-S-4 by IAEA has started again. Two other materials that were subsequently prepared include NBS 123, another sphalerite, and NBS 127, a BaSO4 (ion exchanged sea water sulfate, prepared by J.R. O'Neil, USGS, Menlo Park, USA). This listing documents the prevailing unsatisfactory situation for sulfur isotope reference materials. 1. I was informed that one solid piece of the original CDT (approx. I kg) is stored at the Riksmuseum Stockholm, Sweden.
901
International Stable Isotope Reference Materials
Finally, the urgent need for pure chemical compounds as reference materials resulted in the subsequent production of three Ag2S reference materials with substantially different isotopic composition (IAEA-S-1, IAEA-S-2, IAEA-S-3) by B.W. Robinson, Institute for Geological and Nuclear Sciences, Lower Hutt, New Zealand. In view of the situation of an inhomogeneous and exhausted CDT reference, the change to a new scale was recommended (Gonfiantini et al., 1995), based on IAEA-S-1 as calibration material. The VCDT scale (Vienna-CDT) was established by agreement of CAWIA (IUPAC-CAWIA, 1997) by adopting a defined ~)34S value for IAEA-S-1 to keep the new scale as close to the CDT scale as possible. Five years ago, two additional BaSO4 materials named IAEA-SO-5 and IAEA-SO6, having different isotopic composition for both (534S and for 6180 values, were produced jointly by H.R. Krouse, University of Calgary, Canada and S. Halas, University of Lublin, Poland. Those two materials were prepared to be used in addition to NBS 127 to calibrate both 634S and 6180 values in sulfate samples. However, an initial interlaboratory comparison conducted in 1998 resulted in large value discrepancies and made it impossible to assign reliable isotope ratio values to the new materials. Only two years ago, results from SF6 measurements in three laboratories and improvements reported for SO2 measurement techniques resolved this dilemma. The isotopic data on the available sulfur reference materials and proposed reference mateTable 40.9- Sulfur 6-values versus VCDT for the available sulfur isotopic reference materials and their associated standard uncertainties at l o-level. 6-values marked with * are reported versus CDT. w provisional data from new unpublished measurements. CM stands for "Calibration material" and RM for "Reference material". Status
Distribution
634S [%o]
Troilite (FeS) BaSO4 S (elemental)
RM
exhausted discontinued IAEA, NIST
sphalerite sphalerite
RM
discontinued IAEA, NIST
NBS 127
BaSO4
RM
IAEA, NIST
IAEA-S-1 (NZ1) IAEA-S-2 (NZ2) IAEA-S-3 IAEA-SO-5
Ag2S AgRS Ag2S BaSO4
CM RM RM RM
IAEA, NIST IAEA, NIST IAEA, NIST IAEA
IAEA-SO-6
BaSO4
RM
IAEA
SF6
-
IRMM
0* +20.42+0.42" +16.90+0.12 Qi & Coplen (2003) +0.18+0.14" + 17.09+0.31" +17.44+0.10w +20.32+0.36* +21.17+0.12 Halas & Szaran (2001) -0.30 +22.66+0.13w -32.30+0.12w +0.49+0.09w +0.15+0.05 Halas & Szaran (2001) -34.18+0.07w -34.04+0.11 Halas & Szaran (2001) +17.33+0.22 e
Name CDT OGS IAEA-S-4 (Soufre de Lacq) NBS 122 NBS 123
IRMM-PIGS-2010
Material
e 634S value calculated from absolute isotope ratio measurements (see Table 40.1).
6180 [%o] +9.00+0.65
+9.34+0.32
902
Chapter 40- M. Gr6ning
rials were discussed during the last IAEA experts meeting on reference materials in the year 2000. Details of the discussions will be presented elsewhere. The isotope ratio values of the available sulfur reference materials are provided in Table 40.9. Before that IAEA meeting, an absolute ratio determination of the sulfur isotope abundance in IAEA-S-1 was undertaken using the SF6 technique at the Institute of Mineral Resources, CAGS, Beijing (Ding et al., 1999) and the Institute for Reference Materials and Measurements IRMM, European Commission, Geel, Belgium (T. Ding and S. Valkiers). The results compared quite favourably with relative 6-measurements using the SF6 technique performed by B.E. Taylor in Canada and T. Ding in China. This is a very successful case, which results in a direct link of the conventional relative 634S VCDT-scale to the S.I.-system of units. The differences of results produced by the SF6 and SO2 techniques are discussed elsewhere (Beaudoin & Taylor, 1995). But the establishment of these two sets of three reference materials each (in form of Ag2S and BaSO4), provides the means for the necessary routine normalization of ~534Smeasurements in laboratory operation. In addition, an SF6 gas, distributed by Messer-Griesheim GmbH, Krefeld, Germany, was assessed for its absolute sulfur isotope ratios by the Isotope Measurements Unit at IRMM in Geel and was described by them as Primary Isotopic Gas Standard (PIGS) (Taylor, 1998). Together with the calculated absolute isotopic abundance of the virtual VCDT, a 634S value for this SF6 gas can be calculated, which can be important for a direct calibration of secondary laboratory reference gases. In an advisory group meeting at the IAEA, the need for organic ~34S reference materials was expressed by the participants (IAEA, 2001). As for 615N measurements, the proposed materials included thiourea, methionine and N-methyl anthranilic acid ester. No results on the suitability of these materials are available so far. 40.3.6 Lithium stable isotopic reference materials For reporting of relative lithium stable isotope ratios, the primary reference material is LSVEC, a lithium carbonate prepared by H.J. Svec, Iowa State University, USA. It is assigned a 67Li value of 0%o by definition. For its Table 40.10 - Lithium 67Li-values versus LSVEC for the existing lithabsolute isotope ratios, see ium isotope reference materials. PRM stands for "Primary reference Table 40.1. Note that for material" and RM for "Reference material". lithium, the expression of Material Status Distribution 67Li [%o] relative isotope data as Name 66Li is still common. In LSVEC Li2CO3 PRM IAEA, NIST 0 accordance with IUPAC IRMM-016 Li2CO3 RM IRMM 0 (Coplen, 1996b), the IRMM-015 Li2CO3 RM IRMM -996 reporting as 67Li data is recommended.
International Stable Isotope ReferenceMaterials
903
In addition to LSVEC, two more lithium carbonate materials were prepared by the Institute for Reference Materials and Measurements IRMM in Geel, Belgium. IRMM016 (IRMM, 1997; Qi et al., 1997) is isotopically indistinguishable from LSVEC, and IRMM-015 (IRMM, 1993) is substantially enriched in 6Li. The isotope values of those reference materials are presented in Table 40.10. 40.3.7 Boron stable isotopic reference materials The primary reference material for boron isotope ratio measurements is boric acid SRM 951 (former name NBS 951) as supplied since the 1960s by NIST, Gaithersburg, USA (Catanzaro et al., 1970). Its 11B/10B ratio is given in Table 40.1; see also NIST Certificate of analysis (1999). A few other isotope ratio determinations on this material are discrepant for unknown reasons (Spivack & Edmond, 1986; Leeman et al., 1991). This material is used to define the zero-point of the 611B scale; so, by definition, 611BSRM951 is 0 %0. A boron acid material with a similar boron isotope ratio is provided as IRMM011 by the Institute for Reference Materials and Measurements (IRMM), Geel, Belgium (see Table 40.1) (De Bi6vre & Debus, 1969; IRMM Certificate, 2001). An overview on absolute ratio measurements of SRM 951 can be found in Deyhle (2001). In the past, only few additional materials were available for calibration of boron isotope ratio measurements: first, the JB-2 and JR-2 reference materials (island arc tholeiitic basalt and rhyolite distributed by the Geological Survey of Japan), which are used for boron isotope result normalization in silicates (Nakamura et al., 1992; Tonarini et al., 1997; Deyhle, 2001; Kasemann et al., 2001); second, the two NIST synthetic silicate glasses, SRM 610 and SRM 612, for which their boron isotopic composition was determined recently (Kasemann et al., 2001); third, ocean water, which has a uniform dissolved boron isotopic composition (Spivack & Edmond, 1987). Three other certified boron reference materials were prepared by J. Vogl et al. at Bundesanstalt ffir Materialforschung und-prfifung (BAM), Berlin, Germany with the primary purpose of producing boric acid reference materials for ICP-mass spectrometers used in nuclear reactors. These materials are available from BAM under the names BAM-I001, BAM-I002 and BAM-I003 (Vogl et al., 2002). Recently, the preparation of additional boron stable isotopic reference materials, covering different matrices, was initiated by R. Gonfiantini, Istituto di Geoscienze e Georisorse, Pisa, Italy. Eight different natural materials were prepared at that institute with assistance through an IAEA technical contract (two groundwater and a sea water sample, Elba tourmaline, Etna basalt, Lipari obsidian, Maiella limestone, Montelupo clay) (Tonarini et al., 2003). The provisional mean values were derived from the results of an extended interlaboratory comparison exercise with involvement of 15 laboratories engaged in boron isotopic measurements on natural matrices (Gonfiantini et al., 2003). Significant discrepancies exist in data reported from different laboratories, which can partly be attributed to differences in precision and systematic offsets, but also due to calibration problems (linear shifts of all data between laboratories), disturbing matrix effects and most probably to additional - so far, undetermined - systematic effects. The isotope values of existing boron reference materials are pro-
904
Chapter 40- M. Gr6ning
Table 40.11 - Boron 6-values versus SRM 951 for the existing boron isotopic reference materials and quality control materials with their associated standard uncertainty at l o-level. PRM stands for "Primary reference material", RM for "Reference material" and QCM for "Quality control material". Name
Material
Status
Distribution
6UB [%o]
SRM 951 IRMM-011 JB-2
boric acid boric acid basalt
PRM RM RM
NIST IRMM Jap.Geol.Surv.
0 -0.16 +7.09+0.08 Nakamura et al. (1992)
+7.23+0.24 Tonarini et al. (1997)
+7.13+0.34 Vogl et al. (2002)
JR-2
rhyolite
RM
Jap.Geol.Surv.
+2.71+0.8 Vogl et al. (2002)
SRM 610
silicate glass
RM
NIST
-1.05+0.8 Vogl et at. (2002)
SRM 612
silicate glass
RM
NIST
-1.07+0.8 Vogl et al. (2002)
IAEA-B-1
groundwater
QCM
IGG Pisa
+37.7+2.1 Gonfiantini et al. (2003)
IAEA-B-2
groundwater
QCM
IGG Pisa
+13.6+2.6 Gonfiantini et al. (2003)
IAEA-B-3
sea water
QCM
IGG Pisa
-21.3+0.9 Gonfiantini et al. (2003)
IAEA-B-4
tourmaline
QCM
IAEA
-10.3+2.9 Gonfiantini et al. (2003)
IAEA-B-5
basalt
QCM
IAEA
-4.2+2.7 Gonfiantini et al. (2003)
IAEA-B-6
obsidian
QCM
IAEA
-3.3+1.8 Gonfiantini et al. (2003)
IAEA-B-7
limestone
QCM
IAEA
+7.2+3.9 Gonfiantini et al. (2003)
IAEA-B-8
clay
QCM
IAEA
-5.4+1.2 Gonfiantini et al. (2003)
v i d e d in Table 40.11. A d d i t i o n a l efforts will be necessary to i m p r o v e the intercalibration of laboratories active in the field of b o r o n stable isotope ratio m e a s u r e m e n t s .
40.3.8 Chlorine stable isotopic reference materials For chlorine stable isotope m e a s u r e m e n t s , data are c o m m o n l y expressed relative to sea w a t e r chloride, w h i c h w a s t h o u g h t to be isotopically h o m o g e n e o u s for 637C1 w i t h i n a b o u t +0.15%o ( K a u f m a n n et al., 1988; Coplen, 2001b). Therefore, a S t a n d a r d M e a n O c e a n Chloride (SMOC) w a s p r o p o s e d as p r i m a r y reference material for chlorine stable isotope m e a s u r e m e n t s and the scale realized by use of i n d i v i d u a l s e a w a t e r s a m p l e s in laboratories. However, recently Y. Xiao of the Q i n g h a i Institute of Salt Lakes, Xining, China, has p u b l i s h e d sea w a t e r chlorine isotope ratios s h o w i n g sub-
905
International Stable Isotope Reference Materials
stantial variability (Xiao et al., 2002a). To eliminate the difficulty of using sea water as the chlorine isotopic reference material, Y. Xiao has collected sea water and purified about I kg of NaC1 for use as a relative chloride isotope ratio reference material called ISL 354, which can anchor the chlorine isotope ratio scale (Xiao et al., 2002b). Since the early 1960s, NIST SRM 975 sodium chloride (Shields et al., 1962) has been the basis for absolute isotope abundance measurements of chlorine. Since SRM 975 is exhausted, it has been replaced by SRM 975a sodium chloride (NIST Certificate of analysis, 2001). The isotope ratio values of the existing chlorine reference materials are presented in Table 40.12.
40.4 Concluding remarks In view of the preparation of new isotopic reference materials and the recent measurements on existing isotopic reference materials, a re-evaluation of recommended isotopic values for existing reference materials is ongoing and will result in a set of Reference Sheets issued by the IAEA according to the requirements as stated in ISOGuide 31 and its recent revision (ISO/REMCO, 1998). The uncertainties for all recommended isotope ratio values will be recalculated using a consistent approach for the whole dataset available. It has to be kept in mind, however, that many measurement data were provided long ago by laboratories that did not include sufficiently detailed information on their measurement uncertainty. Therefore, often the uncertainty for the reference values cannot be expressed according to international recommendations (ISO/BIPM, 1995; EURACHEM, 2000). The proper use of reference materials as described in ISO-Guide 33 (ISO, 1998) is discussed elsewhere (Gr6ning et al., 1999). The available reference materials are intended to calibrate local laboratory standards that are prepared by the individual laboratories. The reference materials are NOT intended to be used themselves for quality control purposes. For the distribution by IAEA of all stable isotope ratio reference materials, a rather strict rule applies" Each laboratory is entitled to order one unit of any reference material only once in a threeyear period. This limitation was set to preserve the availability of these valuable reference materials for the maximal possible time and therefore to ensure the comparability of results from laboratories for as many decades as possible. It is hoped that with Table 40.12 - Chlorine 6-values versus SMOC for the existing chlorine isotopic reference materials and their associated standard uncertainty at 1a-level. PRM stands for "Primary reference material" and RM for "Reference material" Name
Material
Status
Distribution
637C1 [%o]
SMOC SRM 975
NaC1
PRM RM
Exhausted
SRM 975a
NaC1
RM
NIST
ISL 354
NaC1
RM
IAEA
0 +0.43+0.04 Xiao et al. (2002) +0.2+1.5 Xiao et al. (2002) +0.05+0.02 Xiao et al. (2002)
906
Chapter 40 - M. Gr6ning
the preparation of additional reference materials, this limitation can be eased and more reference materials can be made available to laboratories more frequently for improved calibration of measurements. The following rules should be applied for an optimal calibration using available reference materials (Gr6ning et al., 2004), regardless of the element under consideration: The laboratory should prepare one or more local laboratory standards with chemical characteristics similar to those of the reference material used for the calibration. The amount of local laboratory standards to be prepared should be at least sufficient for all measurements performed within the timeframe of four calibration cycles (one cycle corresponds to three years at this moment for all internationally distributed stable isotopic reference materials supplied by IAEA and NIST). This enables the laboratory to perform several calibrations with the same local laboratory standards within the period of 12 years, therefore improving with time the accuracy of their calibrated isotopic values. The largest possible number of measurements should be performed during one analysis day, including analysis of relevant reference materials and local laboratory standards to minimise the effects of long-term performance drifts. The combination of individually assessed data from several measurement days should improve the reliability of the calculated average isotopic value for local laboratory standards. Additionally, the probability of any systematic offset, due to daily performance fluctuations, can be minimized with this strategy. A long-term average or floating mean should be applied for evaluation of the entire dataset for the internal laboratory standard. In most cases, weighted means should be applied to account for varying performance on individual calibration measurement days and to calculate a reasonable uncertainty for the mean value. This may serve as a proxy for the standard uncertainty achieved in the laboratory for routine measurements during that time.
Acknow legdements I wish to thank Tyler Coplen, Philip Taylor, Tiping Ding and Roberto Gonfiantini for many constructive comments to improve the manuscript. I am especially grateful to the careful review by Tyler Coplen with corrections on language, style and many historical details on reference materials.
Handbook of Stable IsotopeAnalyticalTechniques,Volume 1 P.A. de Groot (Editor) 9 2004 ElsevierB.V. All fights reserved.
CHAPTER 41 The Nature and Role of Primary Certified Isotopic Reference Materials" A Tool to Underpin Isotopic Measurements on a Global Scale P. D. P. Taylor1, P. De Bi6vre & S. Valkiers Institute for Reference Materials and Reference Measurements, JRC-European Commission, B-2440 Geel, Belgium e-mail: [email protected]
Abstract This chapter focuses on isotopic reference materials with certified values that are traceable in the International System of Measurements (the SI) and explains how they can be produced. An overview is given of cases where such materials have already been realised. Although for many end-users the problems of not having such reference materials (and only having materials with 'consensus-value-only' or'assigned' values ) are often hidden, the usefulness of this approach is elucidated in the light of increased pressures to improve comparability of isotopic measurements across the borders of space, time, and scientific disciplines.
41.1 Introduction As amply demonstrated in other chapters of this book, isotopic measurements are used as powerful investigative tools in many areas of science: medicine, biochemistry, geochemistry, climatic and environmental and nuclear chemistry, hydrology and archeology. In all these areas, the isotope ratios as such are either of interest or they are used in the isotope dilution process. The availability of instruments of various kinds has enabled a soaring flight of isotopic measurements. Mass spectrometry is predominantly used, although infrared absorption spectrometry (Roth, 1997) can be applied in particular cases (e.g. isotopes of hydrogen in water, carbon in CO2). In mass spectrometry, instruments with different ion sources (electron impact, thermal ionisation, ICP), various types of spectrometer configurations (quadrupoles, multiple collectors, ...) and interesting ancillary chemical reactors (modules for oxygen or carbon isotope ratio measurements producing CO2, CO, ...). Scientists developed suitable measurement procedures for these purposes, including the chemical sample preparation. Although in many cases measurements can be used primarily inside a 'closed community' (e.g. within a laboratory, a single hospital), in most cases the measurement scientists and even more so their customers want to compare the results 1. Correspondence should be adressed to this author.
908
Chapter 41 - P.D.P. Taylor, P. De Bibvre & S. Valkiers
obtained. How does a particular value compare with results from other laboratories, obtained at a different time? What trends can be observed in the carbon isotope ratios over the past 10 or 30 years ? Can I compare the laboratory test results for this patient coming from another hospital with what is produced in our laboratory? Are the oxygen isotope ratios for a sample of ethanol from wine identical when measured in country A or B? To increase the comparability of measurement results outside the perimeter of a single laboratory, several measures can be taken. Proper training of staff is of the utmost importance and the establishment of quality systems into the laboratory has become widespread. However probably the most important factor, recognised early on, is the use of a common standard or isotopic reference material. How does this help? Most routinely used isotopic measurement instruments today are comparator devices: a comparison is made between a particular measured isotope ratio of the sample with that of a reference sample (see Figure 41.1). The measurand is named delta of the sample versus the reference sample (expressed in %0):
6 x - (R~e f 1) 1000
[41.1]
and the reference sample therefore carries by definition the value (~Ref = 0. The delta was initially used by Urey and his collaborators (Sam Epstein primarily; Epstein & Mayeda, 1953) when developing the 180 thermometer. Being unable to measure a reliable absolute content of 180 of the samples (this was only achieved by Baertschi years
Figure 41.1 - Graphical presentation of possible traceability chains for measurements linked to 'artefact-only' IRMs (see text). The lines on the left (1,2,3) correspond to the 6 values, w h e n hidden uncertainties are taken into account. (It should not be concluded from this figure that delta values can be added linearly, as this only holds to a very rough approximation ; correct equation: 602 = 601 + 612 + 601" 612/1000)
The Nature and Role of Primary Certified Isotopic Reference Materials ...
909
later; Baertschi, 1976), a differential measurement technique was developed to establish the 180/160 value against the ocean value, assumed constant after preliminary investigations. The delta notation was then a very practical way of reporting results. Extension of the use of this notation together with materializing the reference to ocean water for oxygen and deuterium, and using other reference materials for other elements, followed. This procedure in principle enables laboratories, not interested in absolute calibration of their measurements, to compare their data with those of other laboratories. At first sight, it does not seem to be extremely important to know the absolute value Rref. Instead, the reference sample often only has an assigned (consensus) value. This approach is very practical, economical and straightforward and has helped the development of isotope ratio measurement sciences to a great extent. Reference samples such as PDB, CDT, SMOW all have been construed upon this principle (see Table 41.1 and Gonfiantini et al., 1993). Although such materials are regularly called 'primary' materials by practising isotope scientist because they were first used, from a metrological point of view they are not primary. Primary methods of measurements is a concept used (Quinn, 1997) in many areas of scientific measurements (temperature, mass, concentrations) to give 'direct access' to the international system of measurements (the SI). The SI being a fairly 'new' invention which was exactly conceived to improve comparability of measurements via better traceability. In this chapter, the limitations of relying on reference samples with 'consensus values only' will be explained as well as the ways and means to produce reference samples having 'true' ratios, i.e. with isotope ratios having a primary character, which can therefore solidly anchor the delta scales.
41.2 Commonly used reference samples A Reference Material (i.c. Isotopic RM or IRM), according to the International Vocabulary on Metrologyl, is "a material or substance one or more of whose property values are sufficiently homogeneous and well established to be used for the calibration of an apparatus, the assessment of a measurement method, or for assigning values to materials". The definition given for certified reference materials is "a reference material, accompanied by a certificate, one or more of whose property values are certified by a procedure which establishes traceability to an accurate realization of the unit in which the property values are expressed, and for which each certified value is accompanied by an uncertainty at a stated level of confidence". Many commonly used reference samples are of the first type, i.e. comparator samples with an assigned delta value (e.g. 8 - 0) that may be accompanied by an indicative value of the isotope ratio. Table 41.1 summarises most of the commonly used IRMs. Many of them started their life at the 'spur of the moment', in the 'heat of the 1. VIM (Vocabulaire International des Termes Fondamentaux et G6n6raux de M6trologie/ International Vocabulary of Basic and General Terms in Metrology), 2 nd Edition, ISO 1993.
910
Table 41.1 - Overview of 'comparator' IRMs commonly used in the differential isotopic measurement community. Nominal isotopic composition (in parts per 1000)
RM __
V SMOW-water Oa GISP-water -189.8 -428a SLAP-water -66.7 NBS30-biotite -118.5 NBS22-oil PEFI-polyethylene foil -100.3 USGS24-graphite Sucrose ANU-sucrose NBS18-carbonate NBS19-TS limestone LSVEC-lithium carbonate NBS28- silica sand (optical) IAEA-N1-ammonium sulfate IAEA-N2-ammonium sulfate IAEA-N3-potassium nitrate USGS25-ammonium sulfate USGS26-ammonium sulfate NSVEC- gaseous nitrogen Soufre de Lacq-elemental sulfur NZ1-silver sulfide NZ2-silver sulfide NBS123-sphalerite NBS127-barium sulfate USGS32-potassium nitrate
Oa
-24.85 -55.5a +5.1
-
0.0832a
-29.73 -31.77 -15.9 -10.47 -5.04 +1.95a -46.7
+7.16 +28.65 +3.0 +9.58
Oa
+0.4 +20.3 +2 to +4 -30.4 +53.5 -2.81
-
-
+9.4
-+16.0 -0.3 +21.0 -+17.0 +20.32
+179.9
~
Calibrated measurement. In the case of hydrogen, the absolute values of VSMOW and SLAP were determined by R. Hagemann, G. Nief and E. Roth. Tellus XXII, 712-715 (1970). In the case of oxygen, the absolute 1 8 0 / 1 6 0 ratio of VSMOW was determined by P. Baertschi, Earth and Planetary Science Letters 31,341-344 (1976). Table 41.1: IRMs issued by IAEA/NIST based on concencus values. a
Chapter 41 - P.D.P. Taylor, P. De Bibvre & S. Valkiers
8535 8536 8537 8538 8539 8540 8541 8542 8543 8544 8545 8546 8547 8548 8549 8550 8551 8552 8553 8554 8555 8556 8557 8558
The Nature and Role of Primary Certified Isotopic Reference Materials ...
911
battle' as it were, in response to a direct and immediate need. Nevertheless, for a material to be a reliable isotopic reference material (IRM), as for other reference materials, there are some basic requirements: 9 the material should be available in a stable form, so that the value of the property it carries will not change over time (i.e. the 6 value must remain zero!). For example, such a material should be made available in a stable chemical form, and stored and distributed in suitable, tight containers to avoid isotopic effects from evaportion, adsorption on the container wall, contamination deriving from the container material, etc. Else the value of the properties it carries (isotope ratios, elemental weight fraction) will change over time. This would imply a hidden shift of its 6 values, which therefore affects in a quantitatively unpredictable manner all measurements calibrated against it. This is graphically illustrated in Figure 41.1 for an unstable material, by the uncertainty given to the baseline. 9 the property to be measured should not vary across subsamples (homogeneity requirement). 9 the material should be available to the user community for a long period. 9 for CIRMs (Certified IRM), the certified value should have a sufficiently small combined uncertainty. Many of these requirements seem obvious, but nevertheless practice has shown that they have often not been met. Not only because the problems associated have been grossly underestimated, but also because guaranteeing these requirements can only be done at considerable expense, requiring resources often only available to specialised laboratories. There are a couple of well know examples, explained at length by Gonfiantini et al. (1993). The reference sample CDT previously used for sulfur isotopic measurements, was found to be inhomogeneous only after being in use for many years (Robinson, 1993). Similarly, SMOW originally used for oxygen and hydrogen isotopic measurements, purportedly corresponded to typical ocean water. In fact SMOW did not exist physically, but instead was only compared to values for NBS-1 a fresh water sample that was later found to be unstable over time. The reference material PDB used for carbon isotope ratios, also ran out of stock. The resulting 'instability' of the consensus reference materials is highly unsatisfactory, as traceability of such measurements is hampered. First of all, there is the considerable Babylonic confusion within the measurement community, which often perpetuates years after a new reference sample is introduced or a new zero point is assigned to the isotopic delta scale (e.g. VPDB instead of PDB, VCDT instead of CDT). Furthermore, when natural materials (especially for the light elements) are taken from the same location to approximate the original reference, it is found that the isotopic composition of the new sample is not identical to the predecessor. As the chance of finding a new reference sample with identical isotopic composition is rather dim, a 6 value relative to its predecessor is usually assigned to the replacement. For example for the sulfur delta scale, the material IAEA-S1 has been assigned a 634S value o f - 0.3 %0 on the new scale VCDT1. Note that a sample VCDT with delta value equal to zero 1. Reporting of relative sulfur isotope-ratiodata, Technical report IUPAC, 1997, Pure & Appl. Chem., 69, 293-295.
912
Chapter 41 - P.D.P. Taylor, P. De Bihvre & S. Valkiers
does not exist, i.e. the zero point of the VCDT scale only refers to a hypothetical or virtual material. This has been done to establish some comparability with previous values on the CDT scale, but it clearly demonstrates the weakness of a traceability scheme which only relies on artefacts and non-metrological values for RRef. A similar situation occurred in the case of PDB, where VPDB is a virtual reference material based on NBS-19, whose values vs. VPDB have been fixed by consensus 613C - +1.95 %o and 6180 = -2.20 %o. VPDB is used in all carbon isotopic studies, and in oxygen isotope studies of carbonates and its case was the first in which a virtual reference material has been adopted. Consequently, when scientists today state their measurements against SMOW, PDB, CDT (instead of VSMOW, VPDB, VCDT), this is probably erroneous. The letter 'V' (which stands for Vienna) refers to the measurement scale used. It implies also that the fractionation values recommended by ad-hoc IAEA working groups (to link references materials and measuring procedures), have been adopted (as for instance the fractionation between water and CO2, or between VPDB and VSMOW, etc.). Omitting the letter V may (unintentionally) imply that some other procedures have been adopted. The process of comparison of the 'old' versus the 'replacement' reference sample is usually performed by a set of selected laboratories (in a particular scientific discipline). The practicalities of this are not to be underestimated. The instruments used as comparators are nearly always all different (so what and how big are their specific biases?), the quality of the remaining predecessor sample available at some of the laboratories involved in this comparison might be questionable and the experience of the operators probably differs, as well as the measurement procedures. The establishment and maintenance of an international system to measure differences in isotope ratios is difficult, because of the difficulties to measure the uncertainty not only in the various comparison processes but also in the original zero line itself. As will be explained later on in this chapter (see e.g. Figure 41.6), this situation leads to floating delta scales, which are not anchored against the SI system. So apparently, there is some value in calibrating isotope ratios. Using a metaphor, the process of performing absolute ratio measurements could be compared to the process of measuring heights. Measuring the height of a house is much more easy to do if you are next to the house and using a ruler (with the zero reference point as the ground level). A more complicated approach would be to calculate this height by the difference of two heights relative to sea level (i.e. height of the roof minus height of the floor). Nevertheless, if you are concerned by global warming and want to know what the risk of flooding by the sea is for your house, its absolute height relative to sea level is probably important! Choosing the right reference system matters.
41.3 Comparability and traceability of the results of isotopic measurements In the past, isotopic measurements have often been confined to highly specialised research laboratories dealing with geo- or cosmochemistry. However, isotopic measurements are now even being incorporated into legislation1 and therefore much less 'non-committing'. Consequently, the comparability and traceability of the results of these measurements is even more important.
The Nature and Role of Primary Certified Isotopic Reference Materials ...
913
Traceability1 is not an aim in itself, but if a good traceability system is in place for a particular kind of measurement it will improve comparability. Although often 'willdo' solutions, 'quick fixes' work to a certain extent, at some stage if requirements for comparability increase, such traceability systems collapse. Figure 41.2 represents an 'artist impression' of what happens when isotope scientists continue to have 'artefactonly' based traceability systems. The absence of a good traceability system most often leads to hidden costs. Only in a limited number of cases does this become fully apparent. The 1999 crash of the Mars Pathfinder space craft failed as part of the mission control engineers where still calculating in Imperial Measures instead of using SI units, leading to a multi billion dollar loss.
Figure 41.2 - An 'artist impression' of what can happen if the measurement community does not use the concept of SI traceability (painting of Pieter Breughel the Elder (1520-1596; Museo Nazionale di Capodimonte, Napoli) 'The Blind', where one blind person is leading the others).
1. CEN European pre standards, ENV 12140 (13C/12C in sugars from fruit juice), ENV-xxx (in progress: comparison of 13C/12C in pulp and sugars), ENV 12141 (180/16 0 in water from fruit juice), EC / 822 / 97 (180 / 160 in water from wine). 1. Traceability is the property of the result of a measurement or the value of a standard whereby it can be related to stated references, usually national or international standards, through an unbroken chain of comparisons, all having stated uncertainties (see also Part 2, Chapter 42 on traceability).
914
Chapter 41 - P.D.P. Taylor, P. De Bi~vre & S. Valkiers
Figure 41.3 - Traceability of length measurements. In the past, a unit was realised by some physical artefact, this has been replaced by some measurement process today (e.g. in the SI, one metre is defined as the length of the path travelled by light in vacuum during a time interval of 1/299 792 458 of a second).
If measurement results are not traced back to a same common unit, problems are bound to arise sooner or later. As traceability is a property of the result of a measurement, it helps to clarify the nature of a measurement process: any measurement process consists of a series of comparisons. For a reliable measurement, the nature of these comparisons and their uncertainty should be known. The link thus established can be compared to a chain, whereby the thick lines in Figure 41.3 represent a comparison process with small uncertainty, whereas the opposite holds for the thin lines. Logically, better links (should) exist moving towards the top of the traceability chain. In the case of length measurements, the comparison process is traced back to a common unit. In the past, such units were anchored in some artefact (the King's foot). A 'novel' approach to improve comparability, was the introduction of the Convention of the Metre (1875), whereby several countries agreed to use the same artefact based unit, the realisation of which was based on a bar of Pt/Ir. This was the first truly internationally structured measurement system, which later evolved into what is now known as the International System of units. A remarkable evolution has taken place in the SI 9nearly all of the seven SI base units (except the Kg) have switched from being embedded in artefacts, to being embedded in measurement procedures (De Bi6vre & Taylor, 1997) making use of some fundamental properties of nature (e.g. one metre as a particular number of wavelengths of a 86Kr laser light source1). Thus, a solid anchor is provided, no longer subject to human arbitrariness. 1. strictly speaking, this is the old definition of the metre. In the new definition, the metre is defined as the length of the path travelled by light in vacuum during a time interval of 1/ 299 792 458 of a second.
The Nature and Role of Primary Certified Isotopic Reference Materials ...
915
Figure 41.4 - Comparability of delta measurements m a d e in different scientific disciplines suffers, as different communities use different working standards a n d / o r different measurement techniques.
So what is the situation for isotopic measurements? It can safely be stated that artefacts, such as the ones given in Table 41.1, are predominantly used to establish traceability of isotopic measurements. When values are compared over longer time periods, or when they circulate between different scientific disciplines using isotopic data, or between different countries, or have been generated using different measurement techniques (e.g. carbon isotopic measurements via infra red and mass spectrometry) the risk of incomparability increases (Figure 41.4). Traceability can only be realised if uncertainty statements are both reliable and realistic. For isotopic measurements, such statements are often unsatisfactory (or even non-existent). In many cases it is common practice just to quote reproducibility or even merely repeatability. Apart from reasons of prestige which may drive some investigators to assert exceedingly small uncertainties, the clear lack of training in matters of uncertainty also explains the current situation. The concept of evaluation of uncertainty in measurement, where error is not synonymous to uncertainty, is defined in the internationally agreed ISO document1. In this 1. ISO Guide on Uncertainty, Geneva 1993.
916
Chapter 41 - P.D.P. Taylor, P. De Bibvre & S. Valkiers
approach, a pragmatic stance is taken as for different kinds of uncertainty: they are different depending on the way they are evaluated. Uncertainty estimates resulting from repeated measurements are said to be of'type A" and all others are "type B'. Typical type B data are those derived from a certificate (e.g. of a working standard) or from a separate experiment (e.g. measuring the effect of a spectral interferent). For calculating a total uncertainty, both types are simply treated in an identical way. Why bother about uncertainty? Uncertainty is linked to quality. ISO-BIPM1 uncertainty can be seen as a way to make an objective statement about the quality of a measurement. Vague expressions such as 'inaccurate', 'fairly accurate' or 'highly accurate' can then be quantified. The different steps in the uncertainty estimation process are: 1) formulating the mathematical relationship between what is being measured (the measurand) and the experimental factors having an influence (expressed in the measurement equation), 2) assessing the magnitude of the various uncertainty contributions and 3) combining them to produce a combined uncertainty Uc either using the rules of uncertainty propagation or more simple approximations. If needed, an expanded uncertainty U is calculated, by multiplying with a coverage factor k (U = k.uc) depending on the level of confidence (k = 1, 2 ...). Specifically for uncertainties related to certified reference materials, readers are referred to a detailed description by Pauwels et al. (1999) explaining how to evaluate uncertainty contributions from inhomogeneity and instability. Finally, in isotopic measurements there is a strong tradition and tendency to use large numbers of replicate experiments (e.g. 100 replicate measurements on a sample aliquot in a 10 minute period) to reduce the standard uncertainty of the mean (= s/V'n) when carrying out type A uncertainty evaluations. Although it obviously leads to impressively small numerical values, this can be quite misleading. One must be aware of the other (often larger) sources of uncertainty when calculating combined uncertainties. The likelihood of obtaining comparable results is increased if such uncertainties are used. 41.4 Primary IRMs: CIRMs carrying SI traceable isotope ratios with the smallest achievable combined uncertainties The above explains why some laboratories have specialised in producing primary CIRMs. Such materials are also artefacts, but these artefacts only serve to carry the value. These values are far more important, as they are anchored in the SI system. Such Primary CIRMs can play a crucial role in the calibration process for isotopic measurements. In such a process, the output quantity (what one tries to measure, i.e. (6IRM)x, the delta value for the sample X in the IRM-scale) is linked to input quantities (what is actually measured, i.e. (6obs)x) for a particular system (Figure 41.5). By means of working standards and reference samples, the end user establishes the relationship between the measured value 6obs and the value carried by the reference samples with particular ~)IRMvalues. This operation is carried out at various intervals during a measurement session, depending on the required quality (in the example, a single calibra1. BIPM:International Bureau of Weights and Measures.
The Nature and Role of Primary Certified Isotopic Reference Materials ...
917
Figure 41.5 - Calibration (by the end user) of a differential isotope measurement using IRMs with certified 'delta' values (top). The calibration is shown for a single point. The bottom of this graph shows how the abscissa can be anchored to SI traceable values by the IRM supplier.. tion point is shown). The uncertainty on the values for 6IRM for the different calibration materials used, should be taken into account when the final value on the sample is calculated (see also Figure 41.1). If the values 6IRM of the used reference sample (or their uncertainties) are questionable or lacking in quality, this implies 'floating' points on the abscissa. The bottom part of Figure 41.5 then clarifies the role of Primary CIRMs when measuring differences in isotope ratios: their values (ratios of amounts of isotope) are 'locked' on the scale. This activity is the responsibility of the IRM supplier and means that the IRM isotope ratio(s) are not based on assumptions (e.g. mass dependent fractionation) or some arbitrary assigned value. Instead, it is possible to reproduce the value by a completely independent experiment for such a material thus reducing the problems arising when the IRMs run out of stock. Table 41.2 is an overview of presently existing CIRMs of this type. Using a real case, Figure 41.6 illustrates the value of 'anchoring' delta scales. After the many problems with the original CDT reference sample (see introduction), the alternative reference samples (e.g. IAEA-S1) carried an assigned value which was now (Ding et al., 1998) proven to be different to the SI-traceable value by about 20%o! So how are the 'absolute' ratios of amounts of isotopes, e.g. n(34S)/n(32S), obtained? A completely independent and reliable method is used to 'synthesise' these
918
Chapter 41 - P.D.P. Taylor, P. De Bi~vre & S. Valkiers
Table 41.2 - CIRMs issued by NIST (previously NBS) and IRMM (previously CBNM) based on calibrated isotope ratio mass spectrometry. The * indicates the sample is a 'spike' (i.e. can be used for isotope dilution). Also consult www.irmm.jrc.be and www.nist.gov for further information Element
Identification
Boron
IRMM 011 NIST SRM 951 NIST SRM 952 NIST SRM 977 IRMM 621-623 NIST SRM 975 IRMM 641 IRMM 642 NIST SRM 979 IRMM 012 IRMM 624 IRMM 625 NIST SRM 994 IRMM 632 IRMM 633 NIST SRM 994 IRMM 014 IRMM 620 IRMM 634 NIST SRM 981 NIST SRM 983 NIST SRM 991 NIST SRM 982 IRMM 016 IRMM 015 NIST SRM 980 IRMM 009 IRMM 637 IRMM 638 IRMM 639 IRMM 640 NIST SRM 986 IRMM 627 IRMM 628 IRMM 010 IRMM 630a IRMM 631 IRMM 290 A-G NIST SRM 985 NIST SRM 989 NIST SRM 984 NIST SRM 990 IRMM 017 IRMM 018 NIST SRM 978a
Boron-10 Bromine *Cadmium Chlorine *Chlorine *Chlorine Chromium Chromium *Chromium-50 *Chromium Copper *Copper-65 *Copper Gallium Iron *Iron-57 *Iron Lead normal Lead radiogen *Lead-206 Lead-206-238* Lithium Lithium-6 Mgnesium *Magnesium *Magnesium *Magnesium *Mercury *Mercury Nickel *Nitrogen-15 (nitrate) *Nitrogen-15 (nitrate) Platinum *Platinum-194 *Platinum Plutonium Potassium Rhenium Rubidium Silicon Silicon Silicon Silver
Year of Issue
Chemical Form
1969 1970 1969 1964 1997 1962 2000 2000 1966 1999 1999 1999 1964 1999 1999 1986 1992 1996 1999 1968 1968 1968 1968 1984 1986 1966 1999 1999 1999 1999 2000 1989 1998 1998 1999 1998 1998 1993 1975 1973 1969 1975 1989 1989 1962
H3BO3 H3BO3 H3BO3 NaBr sol in HNO3 NaC1 Aq. sol Aq. sol Cr(NO3)3 sol in HC1 sol in HC1 sol in HC1 Cu sol in HNO3 sol in HNO3 Ga Fe sol in HC1 sol in HC1 Pb Pb Pb Pb Li2CO3 Li2CO3 Mg sol in HNO3 sol in HNO3 sol in HNO3 sol in HC1 sol in HC1 Ni Aq. sol Aq. sol Metal sol in HC1 sol in HC1 Pu(NO3)4 KC1 Re RbC1 Si Si SiO2 AgNO3 (Table 41.2 continued >)
919
The Nature and Role of Primary Certified Isotopic Reference Materials ... (> Table 41.2 continued) Element
Identification
Strontium *Strontium Thallium Thorium Thorium Uranium Uranium Uranium
NIST SRM 987 IRMM 635 NIST SRM 997 IRMM 035 IRMM 036 NIST SRM U002 NIST SRM U970 IRMM 183-187
Year of Issue
Chemical Form
1982 1997 1980
SrCO3 sol in HC1 Th sol in HC1 sol in HC1 U308 sol in HNO3 UF
1966 1987 1988 / 1993
ratios, leading to a n Rtrue which can then be used to obtain a 'calibration' factor K for a particular ratio: K-
[41.2]
Rtrue
Robs It was A. Nier (Nier, 1950) w h o first reported on the use of such mixtures (for argon) to calibrate isotopic measurements. Later on, the use of such mixtures was further developed, especially in cases where traceability was very important, e.g. for w o r k on fundamental constants as well as for nuclear safeguards (for a review, consult De Bi6vre et al., 1993). For the latter type of measurements, research but especially field laboratories u n d e r s t o o d at a very early stage that the use of a relative measurement scale is not appropriate. As these scales are always based on assumptions or on some agreement, and therefore on m u t u a l trust between all parties, they were not viable in the context of nuclear/fissile material where such measurements are used to IAEA-S-3
IAEA-S-1
i I
.................
! I
23.394 -30 .... t ................
I
IAEA-S-2
...............................
22.644 -20 !
1
-10 . . . . . . . . .
0
I ...............
I
I
I
-30
-20
.............
,~ [n(32S) / n(34S)]
I ...... I
SI-traceable
22.143 +10
+20
+30
1
I
1
I
-10
....
t
0
........
6(34S)v-CDT
I
+10
SI-traceable +20
+30 &(34S)v-CDT Thode-traceable
Figure 41.6 - In the absence of SI-traceable values for the sulfur isotope ratios, the delta scale 'floats' relative to the SI traceable ratio scale.
920
Chapter 41 - P.D.P. Taylor, P. De Bibvre & S. Valkiers
keep track of the inventory of fissile material on a global scale. The smaller the uncertainty on such measurements, the smaller the uncertainty on the inventory, which is linked to the critical amount of U or Pu needed to make nuclear weapons. Questioning the appropriateness of differential measurement scales has just started in the area of chemical amount content measurements and isotope ratio measurements, for similar reasons: global trade, environmental policy issues and the growing interdisciplinary use of data (De Bi~vre, 1993; De Bi6vre et al., 1996). The value assignment process for CIRMs is based on two parts: they are underpinned by reliable and transparent measurement procedures to measure the isotope amount ratio (and not a difference in ratios) and a completely independent method (also called a primary method) is used to obtain the calibration K-factor. Both of these parts will be explained in detail.
41.5 Mixtures with gravimetrically prepared isotope ratios When a high purity substance of known chemical composition (e.g. Fe203 with molar mass M) is weighed on a balance (mass m), the chemical amount n(E) (e.g. n(Fe) is given by: n(E)
-
m(1-~im p )(1-r M
stoi
)
[41.3]
where ~imp and ~stoi are small correction terms for chemical impurity and stoichiometry (e.g. when using Fe203, is the amount ratio iron/oxygen really 2/3, or 2.001/3 ...). As the amount n(E) is not a function of other amounts (but function of molar mass and mass), such a method is called a primary method of measurement. These methods are perhaps not the fastest, most convenient or versatile, cheapest or most user friendly, but they have two important features which make them metrologically superior: the traceability of the results produced with such methods is easily understood and they yield small combined uncertainties. Gravimetry has been used in many cases to produce mixtures of isotopes with isotope ratios having small uncertainties. For this purpose, materials of high isotopic enrichment are used (e.g. typically 99% or more, see Figure 41.7) and the isotope ratio for a mixture X of such materials (A and B) is given by:
n (iE ,X)
n (iE,A) + n (iE,B)
n(JE,x)
n(JE,A) + n(JE,B)
f, A n(E,A) + f i, Bn(E, B) f j.An(E,A) + f j, Bn(E, B)
[41.4]
Where f denotes the isotope abundance (for isotope iE, jE), e.g. fi, A (= Ri,A/~Ri, A) the abundance of isotope i in sample A. The amounts n can be expressed using equation [41.3]. The starting materials A and B must be chemically purified and then transformed into a compound of well known stoichiometry. Their imperfection prior to mixing is then carefully assessed" purity, stoichiometry, degree of isotopic enrichment. The uncer-
ENRICHED Si ISOTOPE MATERIALS (RAW BASE MATERIAL)
PURE TEST MATERIAL (IRMM-018)
GRAVIMETRICALLYBLENDED si CRYSTAL DETERMINATION OF ISOTOPE MIXTURES AVOGADRO CONSTANT GIVEN MATERIALS
dissolve-in NaOH
INTERMEDIATE PRODUCTS
1
with H F and ion exchange convert to
1
con"ert to BaSiF6 then to SiF4 H2SiF6 further purify further Purify in gas phase in gas phase
+
+ convert to BaSiF6, then to SiF4, further purify in gas phase
t
convert to BaSiO6, then to SiF4, further purify in gas phase
with HF and ion exchange convert to H2SiFh convert to BaSiF6, then to SiF further pur&y in gas phase
The Nature and Role of Primary Certified Isotopic Reference Materials ...
GENERAL LAY-OUT OF PREPARATION AND MEASUREMENTS OF Si ISOTOPE MIXTURES
TARGET IONS FOR MS MEASUREMENT measure abundances
4
4
K 'higher' 3 K 'lower' 921
Figure 41.7 - Preparation and measurement scheme for gravimetric isotope mixtures of silicon, used in the context of the Avogadro project.
922
Chapter 41 - P.D.P. Taylor,P. De Bihvre & S. Valkiers
tainty of these three factors is measured and then combined with the uncertainty of the mass measurements. Figure 41.8 gives an example of the contribution of these different sources of uncertainty to the total combined uncertainty. Depending on the uncertainty required and on the amount of material available, solutions can either be prepared by starting from the enriched compounds and then mixed, or the solid compounds can be weighed, mixed and then brought into solution. The first approach reduces the amount of material required, but also often limits the combined uncertainty that can be reached. At IRMM, gaseous synthetic isotope mixtures have also been prepared (e.g. for Xe and Kr). Synthetic isotope mixtures have been prepared in the past by researchers at NBS (now NIST) and CBNM (now called IRMM) and also at CEA. Table 41.2 gives an over-
IRMM
Ins~tute for Reference M ~ e d a i s and Measummen~
CERTIFICA~ iliSOTO~C REFERE~E MATERii~ iRMM~72
~ e I ~ ~ i i c Refe~,~n~ M~dia.i ilRMM~72 ~s s u p p i ~ as a set w ~ c e M ~ ,as follows ....
~~r
ratios
Moi~ i ~ t o , ~ A i b : ~ , c , e Ra~o Code Numb~
,n
{i 0.03% ~ v~ue) ',ilRMM~7~t !IiRMM~72/2 LR',MM~T~3 ilRMM~7~S ~RiMM~7.~~6 !ilRMM~7~ ilRMM~'~8 iilR',MM~7~9 ~RMM~'7~!0 |RMM~,7..~I t', |iRMM~,7~13 |RMMR,~4 |iRMM~7~!.5
0o6~ 67 0o~87 0~1~ 014 0~,050 ~ t 0~,01~99 ~ 0~01',0 165 0,~2 ~1 2 0~ ~ 8 92 0 . ~ li~0! 82 0 o ~ 0'i9 9~.~
Figure 41.8- IRMM-072certificate
,,( 23:3u y.(:z~ u) {:t 0.03% ~ v:,~ue}
,(2~uy.(z~U) 0~1
0 , 4 ~ 91 0 , , ~ 3! 3
0,0t',9 ~:7 0,010 ~ 5 0o~ ~6 0 0,~ ~7 5 0~ ~ 2 34 O + ~ 49"7 0 . ~ 1',01 13 0 o ~ 0iS 8,~::
0,~2
0~ 0~3 .
.
.
.
.
.
.
.
.
.
.
.
.
0o~3 0,~ 0.993 0o~3 0.~3 0~3 0,~3
.
.
I',0 t7 2i 2"1 21 2t .~ 2i 21
The Nature and Role of Primary Certified Isotopic Reference Materials ...
923
view of CIRMs whose values have been obtained using this approach. The smallest combined uncertainties ever reported (De Bi6vre et al., 1995) using this approach has been for silicon isotope ratio mixtures yielding a relative combined uncertainties of 2.10-5 for ratios approximating those of silicon of natural isotopic composition. Figure 41.7 summarises the experimental procedures used to synthesize these samples. The compound Cs2SiF6 was selected as 'carrier molecule' as the uncertainty on the amounts of silicon weighed is reduced when Cs2SiF6 is used instead of Si. Other experimentally verified advantages of this molecule include constant and well known stoichiometry, very low hygroscopicity, chemical purity and ease of quantitative conversion into gaseous SiF4. We estimate the effort involved for the preparation of a set of 5 of such mixtures to be approximately 2 man years. Amongst the different sets of synthetic isotope mixtures which have been prepared for actinides measurements at IRMM, the IRMM-072 is probably the most exceptional one (Figure 41.8). It consists of a triple isotope mixture prepared by mixing U308 powder enriched in 233U, 235U, 238U, in such a way that the n(235U)/n(238U) is always close to unity while the n(233U)/n(238U) varies over 6 orders of magnitude. The expanded uncertainty however was 0.03% on n(233U) / n(238U), and 0.02% on n(235U) / n(238U). Such materials are exquisite tools to study instrumental bias (Taylor et al., 1995), and render visible anomalies otherwise unnoticeable. The built-in ratio close to unity can be used to correct mass discrimination effects (e.g. Rayleigh distillation effect in TIMS or the 'sampling' effects linked to jet expansion and ion beam transfer in the case of the ICP-MS). In this way, remaining ratio dependent effects which are made clearly visible" dead time effects for ion counting measurements, problems with non-linearity of high ohmic resistors of Faraday cups. Figure 41.9 shows the preparation scheme for a set of synthetic sulfur isotope mixtures (Ding et al., in press) used to calibrate for the first time ever the differential sulfur measurement scale (Figure 41.6). 41.6 The measurement procedure The measurement procedure is the other important ingredient for producing primary CIRMs. The method should be reliable and transparent. In this way, different sources of uncertainty can be easily identified in a step-by-step breakdown and, as a result, isotope ratios with small combined uncertainties can be obtained.
This approach leads to the selection of particular types of methods and instrumentation which are different from what researchers use in areas of differential isotopic measurements. This is not surprising, as the objectives are different. Contrary to differential measurements, the aim is initially to 'maximise' bias, so that it is visible, understandable and quantifiable. In a generic way, this can be described by"
Rtru e -
n(iE + ) n(JE+ )
=
I(iE+ ) I(iE+ )
.KToT -
I(iE+ ) I(iE+ )
K1.K2...K z
[41.5]
924
Chapter 41 - P.D.P. Taylor, P. De Bihvre & S. Valkiers
Enriched isotope starting materials
Gravimetrically blended isotope mixtures
S-33
r
Homogenization
HNO3+Br2 BaC12 S-32
HNO3+Br2 BaC12
BASO4-33
S-34
BaSO4-M-1
Ag2S-34
Ag2S-33
BaSO4-M-2
H3PO4+HCI+HI CdAc2 AgNO3
~ H3PO4+SnC12 AgNO3 Ag2S-32
Ag2S-M-2
L] Ag2S-M-1
Weigh
Ag2S-M-1
Ag2S-M-2
~r
~ SF6-32
BFF3
SF6-33
~ SF6-34
Assay Trace analysis ~
SF6-M-1
BFF3 SF6-M-2
Isotopic abundance ratio determination Figure 41.9 - Preparation scheme for sulfur synthetic isotope mixtures used to produce the
PIGS IRMM-2010.
This equation is an elaboration of equation [41.2]. Firstly, it clarifies that the ratio of amounts of isotopes is not directly accessible, but only via measurement of ratios of ion currents. Secondly, the component by component investigation will result in a series of different Ki factors (e.g. corresponding to bias in sampling, ion transfer, ...). Depending on the uncertainty target which is set, it might even be necessary to study the dependence of these K factors on the magnitude of the ratio. Some illustrations are given of how to choose the measurement strategy and equipment, when aiming at measurements of the highest metrological quality. Ion sources of very high stability, such as gas sources, are desirable. Additionally, the use of a molecular flow introduction system will results in a very predictable mass discrimination, primarily due to the effusion of the gas in the spectrometer, a property which has been exploited to the maximum extent in measurements carried out at IRMM in the framework of a redetermination of the Avogadro constant (Gonfiantini et al., 1997). A procedure was developed for measuring R values in near-ideal gas conditions, leading to a value for the Avogadro constant which is consistent in its relationship with other fundamental constants (Planck, Boltzmann, universal gas constant) to better than 10-6. Silicon molar masses were measured to an uncertainty of 10-7, which implies isotope ratios with an uncertainty of 2.10-5. A built-in quality control during such measurements (i.e. checking whether gas effuses as predicted by
The Nature and Role of Primary Certified Isotopic Reference Materials ... 925
Figure 41.10 -Experimentally observed change of the (29Si / W i ) isotope ratio over time, for the IRMM Avogadro spectrometer equipped with a molecular flow inlet system (effusing species SiF4). As ideal gas conditions are closely met, the decay of the ratio can be monitored against kinetic gas theory (i.e. the effusion process is dominant) during a normal measurement period (i.e. 2 hours).
926
Chapter 41 - P.D.P. Taylor, P. De Bihvre & S. Valkiers
kinetic gas theory) has proven to be a very powerful tool and enables to check whether isobaric interference occurs. Figure 41.10 shows how the experimentally observed change of isotope ratio over time can be monitored against a model taking into account effusion and small contributions from adsorption and desorption effects. Using these procedures, it is possible to identify and quantify most of the Ki factors of equation [41.5], just leaving a small 'residual' Kres which can be determined by comparison with the silicon synthetic mixtures. Identical procedures have been used at IRMM for the gas isotopic measurement of isotope ratios of gases as Xe, Kr and S and will be used for others in the future. In the absence of suitable gaseous chemical compounds, other ionisation sources are needed, such as thermal ionisation or inductively coupled plasmas. In those cases, a comparison of the metrological qualities is needed. Instrumentation equipped with an ICP sources often shows more severe mass bias than instrumentation equipped with a thermal ionisation source. For example, the mass bias for Li isotope ratios, using ICP-MS, is typically 15%, but is merely 0.2% when measured using thermal ionisation. On the other hand, significant progress has been made towards the understanding of the fundamental properties of ICP-MS equipment concerning mass bias (Mar6chal et al., 1999). As for the choice of detector systems, for calibrated measurements single collector faraday cups are preferred to multiple collectors, if the stability of the ion source allows this (e.g. thermal ionisation, gas source). With multiple cups, differences in gain and cup efficiency factors need to be carefully assessed (Fiedler & Donahue, 1988), which is an extremely difficult process and can only be carried out using special samples (e.g. IRMM-072). For single cup detector systems, though results obtained may have worse reproducibility, gain and cup bias factors do not exist at all and uncertainty assessment is far more easy. As for systems with electron multipliers, those operated in ion counting mode are to be preferred, as they are less prone to mass bias. Although larger combined uncertainties are expected from multipliers relative to faraday cups, their use is called for in those cases where maximal sensitivity is required. Their main source of uncertainty is detector dead time, the magnitude (and uncertainty) of which can accurately be assessed (Held & Taylor, 1999). 41.7 Conclusions
The importance of isotopic measurements reaches beyond the confines of pure research and has increased the demands on their quality and comparability. The isotopic measurement community needs to reflect how long it can afford to neglect the issue of traceability within the SI system. Primary isotopic reference materials, with values independent of human arbitrariness, can help to anchor the values of isotopic reference materials and therefore offer the basis for a truly internationally structured isotopic measurement system.
The Nature and Role of Primary Certified Isotopic Reference Materials ...
927
Acknowledgement The authors wish to thank the referees, viz. Dr. D. Wayne for his many valuable comments and especially Dr. E. Roth and Dr. R. Gonfiantini for providing the necessary historic perspective. The authors are indebted to J. Norgaard for his help in preparing the graphs.
Handbook of Stable Isotope AnalyticalTechniques, Volume 1 P.A. de Groot (Editor) 9 2004 Elsevier B.V. All fights reserved.
C H A P T E R 42 Traceability in Isotopic M e a s u r e m e n t s Heinrich Kipphardt Bundesanstalt ffir Materialforschung und -prfifung (BAM),D-12200 Berlin, Germany e-mail: [email protected]
42.1 Introduction Measurement results are not an end in itself but the basis for decision making in science, technology, commerce, health, environmental monitoring, food and agriculture, risk assessment, regulation and legal metrology, often with far reaching consequences. Comparability as condition for the establishment of the degree of equivalence of measurement results generated by different laboratories and at different times is very important to establish consistent data, which are required to gain insight, to avoid dispute and to save resources especially, when having globalisation in mind. Comparability of measurement results is established, if the values can be traced back to the same common basis. This is what traceability is about. The concept of traceability is not a 'new invention', but, in fact, formalised common sense of scientific work. Since comparability of measurement results is very useful, this aspect becomes increasingly important in laboratory accreditation as formulated in ISO 17025 (ISO, 1999). In this chapter, an attempt is made to explain the concept of traceability and its implications for isotopic measurements in a consistent way. Exemplified are the cases of amount of substance measurements by IDMS (Isotope Dilution Mass Spectrometry), isotope amount ratio measurements and differential measurements of so called 6values. They are illustrated in Figures 42.1 and 42.2. The related equations are given in the Annex. When dealing with the topic of traceability, it is important to keep in mind, that the discussion on metrology and the aspects of traceability is ongoing, and that there is no final 'textbook' solution available at this time.
42.2 The VIM definition of traceability A definition is a 'naming' and there is a large degree of freedom on how to define terms. Instead of 'wrong' or 'right', a definition can only be judged to be 'powerful/ useful' or 'meaningless/clumsy' and 'consistent' or 'inconsistent' with other existing systems. The term 'traceability' has just recently found its way into the field of isotopic and chemical measurements. However, it has a long tradition in physical measure-
Traceability in Isotopic Measurements
929
ments. The definition of traceability, as given in the 'International Vocabulary of Basic and General Terms in Metrology' (VIM, 1993), is widely accepted by the metrological community and, therefore, chosen to discuss the meaning of traceability for isotopic measurements: "6.10 traceability property of the result of a measurement or the value of a standard whereby it can be related to stated references, usually national or international standards, trough an unbroken chain of comparisons all having stated uncertainties" (VIM, 3993). 42.3 The role of values By definition, traceability is a property of values of a standard1 or of values resulting from a measurement. In isotopic measurements the values of interest are typically a-values, isotope amount ratios R and, when performing IDMS, an amount of substance n. In the traceability chains in Figures 42.1 and 42.2 values are framed by boxes. For each chain these boxes contain values of the same quantity class only i.e. n for amount of substance, R for isotope amount ratios and a-values for differential measurements.
Note, that traceability is not a property of a material nor a property of a measurement procedure, but a property of the value carried by a material or the value obtained by performing a measurement procedure. It is important to keep in mind that all values, except those which are defined, are only estimates and based on assuming a (scientific) model. This is also true for values of standards and for measurement results. Values are established on the basis of the model used and the care taken at the time of realisation/measurement. With additional insight and a refined model, values of standards or measurement results might be subject to change - or might be confirmed. 42.4 The role of stated references By definition traceability chains have their end at a stated reference. A stated reference is a 'known' value. The only ultimately known values are values which are defined. Therefore, the ultimate end of all traceability chains is a value defined in the definition of the unit of measurement (e.g. a SI unit, where SI is the Systame International d'Unitas). Abstract definitions of units are depicted in Figures 42.1 and 42.2 in the boxes of the last row. Preferably, the value defined is anchored unalterable in nature e.g. by means of a fundamental concept or a constant of nature. A value defined 'weakly' might have the drawback to be subject to variations against other unalterable values, as it might be the case, when values are defined by an artefact or
1. As explained in the next paragraph, a standard is understood to ' ... realize, conserve or reproduce a unit or value of a quantity ...' as defined by VIM (VIM, 1993). In isotopic measurements and chemistry, a standard is often a material. Note, that standards might differ in metrological quality and uncertainty (working standards; laboratory standards; secondary, primary and 'fundamental' standards).
930
Chapter 42 - H. Kipphardt
Sample
n(Cu, x)
R34 / 32(S, X)
amount of copper
isotope ratio
in a sample material X
in a sample material X
.,
I
Comparison
-~
R34/32(S,Y)
o..
ass s
Working standard
n(Cu, Y)
R34 / 32(S, Y)
amount of copper
isotope ratio
in an enriched spike material Y
in ref. material Y (e.g. Y = IAEA-S-1)
(s
Comparison
ID
Primary standard
n(Cu, z)
R34 / 32(S, Z)
amount of copper
isotope ratio
in a well characterised material Z
in synthetic isotope mixture Z
1 (best) Realisation
Abstract definition of unit
1 i
amount of substance: mole = as many entities as there are in12 g 12C.
---
...
(
isot
isotope ratio: = R2/1 = n(2E) / n(1E) [R] := In]/In] = 1
[n] :- 1 mol
Figure 42.1 - Illustration of traceability chains for amount of substance and isotope ratio values. Equations used are explained in the Annex.
931
Traceability in Isotopic Measurements
Sample
8X/V-CDT
R(x)
delta vs. V-CDT
isotope ratio
of a sample material
in a sample material
... (see eq. [42.13])
... (S
Comparison
ss sp
Working standard
~)Y/V-CDT
R (Y)
delta vs. V-CDT
isotope ratio
of a working standard Y
in a working standard Y
I R (Y)
... (S
(
Comparison
Primary standard
6Z/V-CDT
R (IAEA-S-1)
delta vs. V-CDT
isotope ratio
of a material Z
in IAEA-S-1
delta value vs. V-CDT
isotope ratio RRef
(best) Realisation
Abstract definition of unit
6X / V-CDT =
R (X) R (V-CDT)
-
1
:= R(V-CDT)
[81 := 1%
Figure 42.2 - Illustration of traceability chains fordelta values and for isotope ratios using delta measurements. Equations used are explained in the Annex.
932
Chapter 42 - H. Kipphardt
when values are defined by being produced by a certain institution. In a traceability chain the step before the abstract definition, is a practical realisation of the value which is defined. Such a realisation, is called (primary) standard, and is needed in order to have the value defined experimentally accessible. Other values can be compared to this standard experimentally. If possible, standard and sample should be very similar in value and composition, because this usually makes the experimental comparison easier to realise. In Figures 42.1 and 42.2 examples for realisation of the units concerned can be found in the boxes of the third row. In isotopic measurements, we have to look at the stated references for 6-values, isotope amount ratio values R and amount of substance values n. Starting with the latter the ultimate reference for amount of substance is the abstract definition of the mole: "The mole is the amount of substance of a system which contains as many elementary entities as there are atoms in 0.012 kilogram of carbon 12." In order to have this definition realised, it is not necessary to use exactly 12 g of 12C. A known multiple or fraction of 12 g of 12C would also be suitable. Moreover, it is not even necessary to realise the mole in 12C. A realisation of the mole in form of the element under investigation (e.g. Cu) is also feasible by using the concept of atomic weight. This is often advantageous, because chemical measurements are strongly dependent on the identity of element E under investigation. As depicted in the third box of the chain on the left hand side in Figure 42.1, the probably best amount of substance standard is a high purity substance well characterised for all impurities. Reading from bottom to top, the third step (see second box in Figure 42.1) in the traceability chain for amount of substance might then be a value carried by a spike, which has been 'calibrated' against this best realisation. For isotope amount ratios of an element E, as defined in equation [42.1], the unit of measurement is unity, since the dimensions of the individual amount of isotopes cancel. This is expressed in equation [42.2], where the square brackets represent the unit of the quantity. Note, that it is not even relevant, in which unit (mol, dozen or any other unit) the amount of the individual isotopes is measured, as long as the same is used for both amounts of isotopes.
R2/1 = n(2E) / n(1E) [R2/1] - [n(2E)] / [n(1E)l - ( m o l / m o l ) - 1
[42.1] [42.2]
Here, as explained above, there is no need to realise exactly a value of unity (= 1.000) as a primary standard. A known multiple or fraction of unity can also be used, and is especially advantageous, when the value realised is close to the size of the value, which is intended to be measured. In practice the best isotope amount ratio standard might be a synthetic isotope mixture prepared gravimetrically from highly enriched isotopes as depicted in the third box on the right hand side of Figure 42.1. The process of synthetic isotope mixing is briefly described in the Annex. Reading again from bottom to top, the third step in the traceability chain for isotope ratios might then be a value carried by an isotopic reference material, which has been 'cali-
Traceability in Isotopic Measurements
933
brated' against a synthetic isotope mixture (see second box on the right hand side of Figure 42.1). For delta measurements the same principle applies, however, the case is more complex, and therefore often found to be misunderstood. This has to do with the definition of a delta value. A delta value is defined to be the difference in isotopic composition between a sample material and a specified reference material (e.g. V-CDT in sulfur isotopic measurements) according to equation [42.3]. hSam./Ref. = ((RSam./ RRef.)-l)
[42.3]
For delta measurements the values appearing in the traceability chain are delta values as given in the chain on the left hand side of Figure 42.2. The comparison is here the comparison of two delta values, where a delta value is established by the experimental comparison of an isotope ratio in two materials. Since values from delta measurements are reported as multiples or fractions of 1%o, the unit of measurement for delta measurement is 1%o. The notation '%0' is nothing else than expressing 'multiply by 10-3', which was introduced - as any other u n i t - to keep the numbers in front convenient. As for any other kind of measurement, also for delta measurements the end of the traceability chain is the abstractly defined unit of measurement. As for amount of substance and isotope amount ratios explained above, there is no need to realise exactly a value of unity (= 1.000%o). A multiple or fraction of the value of unity would also be suitable, and even advantageous if close to the actual measurement problem. Formally, a known delta value carried by the isotopic difference of two materials would be required to realise the element of unity for delta measurements. In practice a different reasoning is common: Since the differences in the isotopic composition of the materials investigated are small, linearity is usually not the main problem in delta measurements. The primary standard artefact defining a delta scale, e.g. PDB, SMOW or CDT, is the best realisation of a 'known' value and therefore understood to be the best realisation of the unit of measurement. This is reasonable, although the delta value carried by these material is zero, and therefore not a realisation of the unit of measurement for delta measurements (The value of zero is due to the subtraction of I from the ratio of isotope ratios (see equation [42.3])). The role of the primary standard artefact in delta measurements is analogue to the role of a tara-setting on a balance, which is known to be linear and well calibrated for mass measurements. Traceability of the reference ratio used e.g. V-CDT, V-SMOW or V-PDB and traceability of delta values are two different pair of shoes, as the traceability of a tara-setting and traceability of mass measurements are. Confusion is often caused, since a 'reference' is involved in the equation for delta measurements. However, the specified reference is here used to define the measurand. This reference is often wrongly taken to be the end of the traceability chain for the value of delta measurements. This is as wrong as to state, that e.g. for sulfur iso-
934
Chapter 42 - H. Kipphardt
tope ratios Ri/32 with reference to the isotope 32S, the end of the traceability chain is the value of 32 coming from the 32S isotope. In both cases the end of the traceability chain is the unit value, which is 1%o (a sub-multiple of unity) for delta measurements and unity for R. A traceability chain for the reference ratio RRef used in delta measurements is illustrated in the chain on the right side of Figure 42.2. Here the values in the traceability chain are the isotope amount ratios, and the comparison consists of the comparison of these isotope amount ratios performed by delta measurements. The ultimate end of the traceability chain is the abstract definition of the isotope amount ratio, which is e.g. for sulfur the isotope ratio of V-CDT. A (best) realisation of this ratio is the value carried by the IAEA-S-1 sulfur standard, which currently defines the V-CDT sulfur scale. Reading the traceability chain from bottom to top, a third step might be a 'house standard' measured against IAEA-S-1 sulfur standard (see second box in chain on the right hand side of Figure 42.1). At first glance the traceability chains on the right hand side in Figures 42.1 and 42.2 do not seem to look very different - but they are. First, there is a difference in the comparison, which is in the first case done by isotope ratio mass spectrometry; probably over different orders of magnitude of R, and in the second case done by a delta measurements, which are usually valid only for small differences in R. The second and more important difference is, that the traceability chain on the right hand side of Figure 42.1 ends in the value of unity, whereas the traceability chain in Figure 42.2 ends in the value of V-CDT: Whatever the value of V-CDT is, it does not need to be known explicitly; The value is defined by an artefact. Even if formally consistent and useful, an artefact approach has a fundamental drawback. Comparability of data is only established, as long as the artefact is available and does not vary in time and space against other inalterable values - such as the values of the isotope amount ratio. However, when, as in the case of CDT and PDB, the reference material defining the delta scale is exhausted or suspected to be inhomogeneous problems concerning comparability of values arise. To avoid this, the role of the reference material in the measurement process must be inverted. The function of the reference material should not be to define a value, but to realise a (inalterable) value. For example, in delta measurements of sulfur isotopes this means, that IAEA-S-1, which currently defines the sulfur delta scale, must be seen as a carrier of an inalterable value of an isotope ratio. This value must be measured on 'absolute' scale with smaller or comparable uncertainty as coming from delta measurements. This has been done recently (Ding et al., 2001). For the practical work in the laboratory the inversion of the role of reference materials will make no difference" A sample is experimentally compared to a reference material. However, the difference is, that now the reference material is known for its unalterable values of isotopic composition.
Traceability in Isotopic Measurements
935
Note, that mass is at present the only remaining SI unit, which is also defined by an artefact- with all problems associated (Girard, 1994) -, since the mass of a kg is defined to be the mass of a certain Pt/Ir-Cylinder kept at the Bureau International des Poids et Mesures (BIPM). 42.5 The role of national and international standards For comparability in time and space, traceability of a measurement result to SI is very advantageous. Even if sometimes possible and done, a field laboratory does not need to establish the traceability chain for a measurement up to the abstract definition experimentally on its own. It is the task of the national and international metrology institutes (NMIs) to realise SI units and other units of measurement and to make them available to the measurement community in form of standards. For the standards disseminated, it is the responsibility of the NMI to establish and demonstrate a traceability chain to the abstract definition of the unit concerned. It is the responsibility of the field laboratory to establish correctly the traceability chain for their experimental comparisons down to the values carried by that standard.
Dealing with matter, for chemical and isotopic measurements these standards are mainly so called reference materials1. A reference material is carrying a 'known' value of a specified property. This might be formalised in a certificate issued by an organisation resulting in a certified reference material (CRM). In isotopic measurements standards distributed by NMIs are e.g. isotopic reference materials certified for their isotopic composition or for a delta value vs. a specified material and amount of substance standards (often called spike). The latter might be of natural isotopic composition or isotopically enriched. Dependent on the effort and care taken and the validity of the measurement model, reference materials can be quite different in quality with respect to uncertainty statement and metrological quality, and it is often difficult to judge the quality of a standard. Therefore a clear and transparent documentation of the certification process should always be available. 42.6 The role of chains By definition traceability runs along chains. This is depicted by the arrows in Figures 42.1 and 42.2. By nature, a traceability chain reads from the value of a measurement d o w n to the abstract definition of its corresponding unit. However, usually traceability chains are more easy to understand when starting with the abstract definition and then going up to the result of the measurement. This is indicated by the arrows hinting in two directions.
1. Especially in analytical chemistry, where there is a huge combination of analytes and matrices, and there is no method of measurement existing, which is independent of analyte and matrix, for many applications reference materials are not available. In such a case, often a complex problem is transformed into a more simple problem, e.g. by separation of the analyte to a matrix, for which a reference material is available.
936
Chapter 42 - H. Kipphardt
A chain is a one-dimensional relation. Note, that usually the links in the traceability chain, the comparisons - depicted as ovals in Figures 42.1 and 42.2 - involve other quantities, which often have their own traceability chains. The whole picture of a measurement process is therefore a (more-dimensional) net, and its degree of complexity depends on how detailed the process is formulated. If we look for example at the chain on the left hand side of Figure 42.1, in IDMS the unknown amount of substance n(E, X) of an element E in a sample X is compared to a known amount n(E, Y) of an element in the spike Y, done via isotope amount ratio measurements R(X), R(Y) and R(B) in sample, spike and blend B respectively. The main traceability chain relates n(E, X) to n(E, Y) by a comparison. This comparison is based on the measurement principle of IDMS as described by equation [42.4:]. The isotope amount ratios which are integral part of this comparison might have their own traceability chains e.g. via values established by isotopic reference materials. Therefore the chain on the right hand side of Figure 42.1 can be seen as a zoom-in of the chain on the left hand side of Figure 42.1.
42.7 The role of comparisons By definition traceability deals with the comparison of values. Comparisons are depicted in Figures 42.1 and 42.2 as ovals, with the exception, that the ovals at the bottom line of these figures represent realisations of abstract definitions. In the strict sense of the meaning, comparisons involve exactly two members of the same quantity (class). This forms a one-dimensional relation, a chain. Mass can only be compared to mass, an isotope ratio can only be compared to an isotope ratio. Occasionally traceability chains are drawn formally incorrect to include conversions from one quantity class to an other, such as a conversion from mass to amount of substance (via the concept of molar mass). Since a traceability chain shows only a spotlight of a complex measurement process, with many quantities involved this can be useful, especially when aiming at showing a measurement structure or when visualising the links having the major uncertainty contributions. However, this is wrong and confusing from the aspect of traceability. To establish the comparison is usually the very critical point in the traceability chain, and also most difficult to demonstrate, because the comparison is of experimental nature performed in a laboratory and there is no general recipe, on how to perform such a comparison correctly. Most often, lack of traceability is caused by lack of a correct comparison. In short, the comparison has to be done according to the state of the art and with the best knowledge available. Dealing with isotopic measurements, this book might be seen as a tool in order to help to achieve this. All problems of discrepancies between measurement model and physical reality i.e. sample preparation with losses and contamination, as well as the calibration and linearity of the measurement device come in at this stage.
Traceability in Isotopic Measurements
937
42.8 The role of uncertainty The VIM (VIM, 1993) defines uncertainty to be: "3.9 uncertainty of measurement parameter, associated with the result of a measurement, that characterizes the dispersion of the values that could reasonably be attributed to the measurand." The concept of uncertainty is explained in detail in the ISO/BIPM 'Guide to the expression of uncertainty in measurement' (GUM, 1995). More focused on chemical measurements there is the guide 'Quantifying Uncertainty in Analytical Measurement' (Eurachem, 2000) from EURACHEM/CITAC available. A detailed example for the uncertainty estimation for double IDMS for the measurement of lead in water using ICP-MS is given in the annex A7 of (Eurachem, 2000). In summary the uncertainty concept is based on taking into account, that in a measurement process the values of the influencing quantities are only known to a limited uncertainty. Therefore it is reasonable, that the measurement result derived from these quantities must also carry an uncertainty. All uncertainty contributions, either those, which can be determined from repeated observations (type A) or those, which can not be determined by repeated observations (type B) - such as estimates on experience or expert knowledge - are treated as standard uncertainties. Only the major contribution uncertainty components need to be taken into account. The individual contributions are combined to form the (combined) uncertainty of the measurement result. Already setting-up an uncertainty budget is very educational and often associated with more insight in the measurement process, resulting in improvements of the measurement procedure. Even if the uncertainty concept is powerful, one must keep in mind that it does not solve all problems: The uncertainty concept starts on the basis of an existing measurement model and can not account for incorrect measurement models - e.g. an influencing quantity has been overlooked - nor for a incorrect realisation of this model in the laboratory - e.g. a parameter thought to be under control, which is not under control. Therefore the uncertainty derived from this concept must be seen a minimum uncertainty. A traceability chain shows the values starting with the measurement result down to ultimately the abstract definition of the unit concerned and the links between these values. A traceability chain is therefore a prerequisite for an uncertainty estimate of a measurement result and not vice versa. In the traceability chain, uncertainty can be interpreted as the strength of a link in a traceability chain. The smaller the uncertainty, the stronger the link. 42.9 Conclusion The concept of traceability is a basis of scientific work. It is well established and has a long tradition in physical measurements. It can and must therefore be applied also to isotopic measurements. It is a powerful tool to establish comparability of mea-
938
Chapter 42 - H. Kipphardt
surement results in time and space. This is very desirable, because it avoids dispute and confusion, thus enhances development and saves resources. Without naming it explicitly, this concept or parts of it has been followed by some scientist in measurement practice, because it is in fact not more than formalised scientific common sense. As in any other field of measurement also in isotopic measurements sometimes lack of traceability is observed. Most often this is caused by lack of a correct comparison and less frequently by problems related to stated or varying references, wrong or missing uncertainty statements or broken chains.
Acknowledgements The scientific discussion with colleagues from JEPPIM (Joint European Programme for Primary Isotopic Measurements, a programme co-ordinated by the Institute for Reference Materials and Measurements (IRMM)), IRMM and BAM and especially with P. De Bi6vre is gratefully acknowledged. I am very grateful for the comments of the referees and especially for the very detailed and constructive comments from R.M. Verkouteren.
939
Traceability in Isotopic Measurements
ANNEX: Equations used in Figures 42.1 and 42.2 Equation [42.4] is the IDMS equation, it relates an amount of copper n(Cu, X) in a sample material X to the (known) amount of an isotopically different copper in a material Y (so called spike) by isotope amount ratio measurements on the copper isotopes in sample, spike and blend B. The equation given holds for an element with two isotopes. Equation [42.5] has the same structure as equation [42.4], however, other materials are involved. Here the amount of copper of a usually isotopically enriched spike Y is measured against a (known) amount of copper in a material Z (of well known purity). Equation [42.4] and equation [42.5] combine to what is called 'double IDMS'. IDMS is described in detail in this book by M. Berglund (Part 1, Chapter 37) and in De Bibvre (1994). - R(Cu,B) R(Cu,X) + 1 R ( C u , B ) - R(Cu,X) R(Cu,Y) + 1
[42.4]
n(Cu,Y) _ R ( C u , Z ) - R(Cu,B) R(Cu,Y) + 1 n(Cu,Z) R(Cu,B) - R(Cu,Y) R(Cu,Z) + 1
[42.5]
n(Cu,X)
n(Cu,Y)
_ R(Cu,Y)
o
Equation [42.6] shows the realisation of the abstract definition of the mole by a material Z. The mole is defined by 12C. The definition implies that pure 12C is meant. Thus the mass fraction w(12C, 12C) of 12C in pure 12C is unity. As described in section 42.4 'Role of stated references' the realisation of the mole does not need to be done in 12C but can be done using the element which is of interest for a specific application. This involves the ratio of the molar masses M. The molar mass M(12C, 12C) = 12 g/mol follows directly from the definition of the mole. The molar mass of Cu in material Z must be known, either estimated from what can be found in nature or, better, from a measurement with smaller uncertainty. When dealing with materials one must keep in mind, that ideal purity can never be reached. Therefore, the mass fraction w(Cu, Z) of the element (Cu) in the specific material Z must be measured. In practice this is often simply ignored and values for few metallic impurities given by the supplier of a material are taken synonymous for the total purity of the material, resulting in a too optimistic purity statement concerning value and uncertainty. At BAM there is a project running in close collaboration with Physikalisch Technische Bundesanstalt (PTB) to realise the SI definition of the mole with smallest uncertainty. Since in chemistry the method of measurement is usually strongly dependent on the element of interest, it is convenient to realise the SI definition for different elements. The target uncertainty for the mass fraction of the matrix element (in a high purity material) is 10-4 relative, which is about one order of magnitude lower than the uncertainty obtained in a typical measurement using isotope dilution mass spectrometry (IDMS). Since there is no method of measurement which can measure the mass fraction of an element in a high purity material with sufficient small uncertainty directly, the indirect approach is followed. This is to measure the mass fraction of all impurities, to sum them up and to subtract them from 100 % - the value for ideal purity. In this context, impurities are all elements of the periodic table not being the matrix element, including non-metals such as O, N, C, S, P and halogens, which are
Chapter42 - H. Kipphardt
940
usually difficult to measure. Since the total impurity content, composed of bulk and surface content, is of interest, a certain geometry and surface treatment by etching is prescribed. Copper has been the pilot element (Pattberg, 1999; Pattberg & Matschat 1999; Matschat et al., 2001, 2002; Kipphardt et al., 2002), work is going on Fe, Sn, Pb and others. n(Cu,Z) [n] - l m o l
m(Z) m(lkg)
~
w(Cu,Z)
w(12C12C) ~
M(12C,12C) M(Cu,Z)
[42.6]
Equation [42.7] describes the comparison of an isotope amount ratio R(S, X) of sulfur in a material X using the isotopes 34S and 32S against a (known) isotope amount ratio R(S, Y) of the same sulfur isotopes in a material Y. This is done by using the ion current ratio J produced by these isotopes in a mass spectrometer. The value of the ion current ratio J might be a good numerical approximation for the isotope amount ratio R, but isotope amount ratio and ioncurrent ratio should not be taken as synonymous, because both differ at least in uncertainty. The ioncurrent ratio is sometimes called 'observed','uncorrected' or'uncalibrated' isotope amount ratio, which is misleading, and should therefore be avoided. The factor K34/32(S,X) is the conversion factor between isotope amount ratio R and ion current ratio J for the corresponding isotopes. This factor is not only dependent on instrument and measurement procedure, but also often dependent on the size of the ratio. This is the reason, why the ratio of the K-factors does not cancel automatically. To obtain the K-factors is difficult and usually done by means of synthetic isotope mixtures. Equation [42.81has the same structure as equation [42.7], but is formulated with different materials.
R34/32(S,X) J34/32(S,X) K34/32(S,X) R34/32(S,Y)
J34/32(S,Y) K34/32(S,Y)
R34/32(S,Y) = J34/32(S,Y) . K34/32(S,Y) RB4/Ba(S,Z) /34/32(S,Z) KB4/g2(S,Z)
[42.7] [42.8]
Equation [42.9] is describing one part of isotope mixing. The idea behind isotope mixing is to achieve a material with a known value for the isotopic composition with small uncertainty, even if the input data carry a large uncertainties. It can be achieved when materials with high isotopic enrichment are blended. The technique is described in detail in De Bi~vre et al. (1995), Valkiers et al. (1998), Ding et al. (2000), Aregbe et al. (2001). For example: When two materials S-32 and S-34, with an isotopic purity of 99.9 % for 32S and 34S each, are blended to a 1"1 mixture Z, an uncertainty on the isotope amount ratio R(Z) < 0.001% relative can be achieved, even if the isotope amount ratios R(S-32) and R(S-34) of the starting materials carry an uncertainty of only 0.1% relative. This is because a large relative uncertainty is small in absolute when highly enriched materials are used. Practically the starting materials, which are usually available only in small amounts and limited chemical purity are purified and converted to stable chemical compounds. Since the values for impurities and stoichiometry (here expressed as mass fractions w(S)) are correlated due to the identical sam-
941
Traceability in Isotopic Measurements
ple preparation, they only need to be known within a larger uncertainty. A synthetic isotope mixture is then prepared gravimetrically, where the uncertainty on the mass ratio m(S-32)/m(S-34) is typically 10 -4 relative. Most often the isotopic composition of the mixture is close to the (natural) sample material. Preferably more than one mixture is prepared, however, this is most often restricted by the limited availability of the starting material. The next step is to establish the isotopic homogenisation of the blend. In case of solid materials the mixture is typically dissolved and the oxidation state of the element of interest is changed. For gases an equilibration using a catalyst might be necessary. The same procedure is applied to the starting material. Finally sub-samples from the blend and starting materials are measured by mass spectrometry, and the observed ion current ratios are recorded. Special care has to be taken to avoid (or to account for) memory effects and cross contamination. From the first observed ion currents ratios of the starting materials the isotopic composition of the blend is calculated. A calibration factor is calculated from the observed ion current ratio and the isotope amount ratio calculated for the blend. This calibration factor is applied iteratively to the input data (observed ion current ratios for the staring material). After a few iterations usually a consistent set of data is obtained. Alternatively, the analytical solution for the calculation of the calibration factor K from the observed ion current ratios J is given in equation [42.9b]. A prerequisite for the approach of isotopic mixing is a linear measurement capability over several orders of magnitude, which is a calibration factor independent of the size of the ratio. This can only be checked by mixtures of different isotopic composition. As for chemical work, a standard should be prepared from the same isotopes as involved in the sample, in order to make the experimental comparison between standard and sample easier. However, there are few cases, where synthetic mixing seems not to be necessary (Kipphardt et al., 2000). R34/32(S,Z)
n34 (S - 32) + n34(S - 34)
[R] "- 1 n32 (S - 32) + n32(S - 34) m(S - 32). w(S,S - 32) R(S - 32) + m ( S - 34). w(S,S - 34) R ( S - 34) M(S,S - 32) R(S - 32) + 1 M ( S , S - 34) R(S - 34) + 1 m ( S - 32) 9w ( S , S - 32) 1 1 + m ( S - 3 4 ) - w ( S , S - 34) M(S,S - 32) R(S - 32) + 1 M(S,S - 34) R(S - 34) + 1 [42.9a] R34/32(S,Z)
[R] " - 1
-- K34/32. J34/32(S,Z ) A+B
with K34/32 - C + D and the substitutions A = (J(S, Z ) - / ( S , S - 32)).
m ( S - 3 2 ) - w(S, S - 32) M(S, S - 32)
B - (/(S, Z ) - ](S, S - 34)). m(S - 34). w(S, S - 34) M(S, S - 34)
[42.9b]
942
Chapter 42 - H. Kipphardt
C - (J(S, S - 3 2 ) - J(S, z ) ) . J(S, S - 34).
m ( S - 32) 9w(S, S - 32) M(S, S - 32)
D - (J(S, S - 34) - J(S, Z ) ) . J(S, S - 32).
m(S-34)-w(S,S-34) M(S, S - 34)
Equation [42.10] describes the comparison of a delta value ~)X/V-CDTbetween a material X and the 'material' V-CDT against a (known) delta value ~Y/V-CDTbetween a material Y and the 'material' V-CDT. The comparison is directly established by comparing the corresponding isotope amount ratios. Since the differences in the isotope amount ratios R are usually small, the K-factor and its uncertainty, which are essential for Equations [42.7,8], are not relevant here. Equation [42.111 has the same structure as equation [42.10], but is formulated with different materials Y and Z. 8 X / V - CDT
[R(X)/R(V
~)Y/V- CDT
[R(Y)/R(V - CDT)] - 1
~)Y/V-CDT _ ~)Z/V- CDT
- CDT)]
- 1
[ R ( Y ) / R ( V - CDT)] - 1 [ R ( Z ) / R ( V - CDT)] - 1
[42.10] [42.11]
Equation [42.121 is a space holder for the realisation of the unit for delta measurements. This topic is treated in the in section 42.4 'Role of stated references'. For delta measurements the realisation is usually based on (reasonably) assuming linearity of a measurement device. ~)Z/V-CDT = 1000.
[6] "- 1%o
[R(Z)/R(V-CDT)]- 1
[42.12]
[8] "- 1%o
In Equation [42.13] isotope amount ratios of the same isotopes in different materials X and Y are compared. The comparison is performed via a delta measurement. This is analogous to equation [42.7] with the big difference, that the isotopic composition of X and Y is very very similar. Otherwise the K-factors need to be introduced. Equation [42.14] has the same structure as equation [42.13], but is formulated with different materials. R(X) = 8x + 1 R(Y) /Y
[42.13]
R(Y)
[42.14]
R ( I A E A - S - 1) = ~)Y/IAEA-S-1 -I- 1
Equation [42.15] stands again for a realisation of a unit. This topic is treated in the in section 42.4 'Role of stated references'. For isotope amount ratio measurement using delta values only, the realisation of unit and the definition fall together in an artefact.
Traceability in Isotopic Measurements
943
The delta value between CDT and IAEA-S-1 is defined by convention to be -0.30%o (As stated in Gonfiantini (1995)). R ( I A E A - S - 1) . = 1 - 0 . 3 0 %o R ( V - CDT)
[42.15]
Handbook of Stable Isotope AnalyticalTechniques, Volume 1 P.A. de Groot (Editor) 9 2004 Elsevier B.V. All fights reserved.
CHAPTER 43 Strategies and Practicalities in the Production and Use of Gas Isotope Standard Materials R. Michael Verkouteren Surface and Microanalysis Science Division, Chemical Science and Technology Laboratory, National Institute of Standards and Technology, Gaithersburg, MD 20899, U.S.A. e-mail: [email protected]
Abstract A practical guide is presented to the production and use of gas isotope Reference Materials (RMs) at the National Institute of Standards and Technology (NIST). These RMs are developed in partnership with the International Atomic Energy Agency (IAEA) and in accordance with guidelines of the International Organization for Standardization (ISO). Assessment of needs, gas selection criteria, safety considerations, measurement strategy, data normalization, and a Web-based interactive data reduction algorithm are discussed. Use of gas isotope RMs as a part of a total quality assurance system results in improved measurement reproducibility among laboratories. An example is presented using the carbon dioxide materials RM 8562, RM 8563, and RM 8564, where normalization of data improved interlaboratory reproducibility by a factor of three or more.
43.1 Introduction Isotope Reference Materials (RMs) are a critical component in the assurance of laboratory analytical quality, and have been applied to growing needs in environmental monitoring, medical and industrial applications, consumer protection, and in the traceability1 of measurements to primary standards (VIM, 1993; Garner & Rasberry, 1993; Peiser & De Bi6vre, 1999). Gas isotope RMs are especially critical for applications involving the light elements (H, C, N, O, S), which are among the most abundant and chemically diversified elements in nature and industry. The recent development of species-specific techniques for isotope determination (continuous-flow isotope ratio mass spectrometry (IRMS), or isotope ratio monitoring (Brand, 1996)), which combine the attributes of automated analysis and high sensitivity, have opened new applications having quality assurance needs addressable, in part, by gas isotope RMs. 1. The International Vocabulary of Basic and General Terms in Metrology (VIM) defines traceability as "the property of the result of a measurement or the value of a standard whereby it can be related to stated references, usually national or international standards, through an unbroken chain of comparisons all having stated uncertainties." Traceability is explained in Chapter 42 of Part 2 in this book volume.
Strategies and Practicalities in the Production and Use of Gas Isotope Standard Materials
945
N a t i o n a l m e a s u r e m e n t agencies, international a u t h o r i t a t i v e organizations, a n d suppliers of isotope s t a n d a r d s require i n f o r m a t i o n r e g a r d i n g the d e v e l o p m e n t of n e w gas isotope reference and i n t e r c o m p a r i s o n materials that will a d d r e s s the anticipated needs in the global inventory. This contribution relates successful strategies a n d practical aspects in the p r o d u c t i o n a n d use of gas isotope RMs at the N a t i o n a l Institute of S t a n d a r d s a n d Technology (NIST). Isotope m e a s u r e m e n t s are c o m m o n l y b a s e d on s t a n d a r d artifact scales. The value of an isotope ratio (R) in a s a m p l e relative to a s t a n d a r d is expressed as a d i m e n s i o n less quantity using the delta (6) n o t a t i o n (equation [43.1]),
m ~3standard =
[m
m
Rsample- Rstandard -~ . . . . . Rstandard
]
[43.1]
w h e r e the m t e r m designates the m i n o r isotope in the ratio considered (e.g., m - 13 for 13C/12C, or m - 18 for 180/160). C o m m o n l y , the delta value is r e p o r t e d as per mill (%o)(ISO, 1992). Table 43.1 lists the s t a n d a r d artifacts a n d their absolute isotope ratios as r e p o r t e d in the literature. These values are not certified by any a u t h o r i t a t i v e body, nor are the r e p o r t e d precisions of these values justified by their uncertainties. H o w ever, their precisions are "accepted" by the m e a s u r e m e n t c o m m u n i t y to p r o v i d e historical consistency for data r e d u c t i o n a l g o r i t h m s a n d h i g h precision in the i n t e r c o m p a r a b i l i t y of delta values s t a n d a r d i z e d onto the artifact scales. Few isotope Table 43.1 - Absolute Abundances of Artifacts* that Base Relative Isotopic Scales. Isotope Ratio
Historical Material
Primary Artifact
Recommended Scale
2H / 1H
SMOW
VSMOW
VSMOW***
13C / 12C
PDB-1
NBS-19
VPDB
15N / 14N
Atm-N2
Atm-N2
Air
180 / 160
SMOW
VSMOW
VSMOW*
34S / 32S
CDT
IAEA-S-1
VCDT
Absolute Isotope Ratio** SMOW: (155.76 + 0.05) x 10-6 [Hagemann et al., 1970] PDB-I: (11237+ 60) x 10-6 [Craig, 1957] NBS-19:(11202 + 28) x 10-6 [Zhang & Li, 1990] Atm-N2:, (3676+ 4)x 10-6 [Junk & Svec, 1958; Coplen et al., 1992] VSMOW: (2005.20 + 0.45) x 10-6 [Baertschi, 1976] CDT: (44360 + 40) x 10-6 [MacNamara & Thode, 1950]
* Acronyms: SMOW = Standard Mean Ocean Water; VSMOW = Vienna-SMOW (RM 8535); PDB = Pee Dee Belemnite (exhausted); VPDB = Vienna-PDB, defined through NBS-19 (RM 8544); CDT = Cation Diablo Troilite; VCDT = Vienna-CDT, defined through IAEA-S-1 ** Non-certified values, with reported expanded standard combined uncertainties (U = 2. Uc)[ISO/ BIPM, 1995]. ***Normalized to where Standard Light Antarctic Precipitation (SLAP, RM 8537) has 62H = - 428%o and 6180 = - 55.5%0 versus VSMOW [Coplen, 1996a].
946
C h a p t e r 43 - R.M. V e r k o u t e r e n
RMs exist with absolute ratios certified at the precision levels attained by relative (8) measurements, since proof of the validity of these absolute values is an extremely difficult task. Recent calibrated isotopic measurements at the Institute for Reference Materials and Measurements (IRMM) have determined precise absolute abundances of some fluorinated (SiF4, SF6) and noble gases (Aregbe et al., 1998; Valkiers et al., 1998, 1999). Work remains to expand this capability to other elements and relate precisely and accurately the current isotopic scales to the certified abundances. Quality assurance systems involve well-designed measurement methods, competent staff, adequate instrumentation, reliable RMs, laboratory quality assurance procedures, and proficiency testing. The scope of this contribution is limited to the constitution and use of "reliable RMs" specifically applied in gas IRMS. The International Organization for Standardization (ISO) Committee on Reference Materials (REMCO) has published guidelines on a variety of quality assurance aspects related to production (ISO 34, 1996), certification (ISO 31, 2000; ISO 35, 1989), and uses (ISO 33, 1999) of RMs. The practical compliance to these guidelines for isotopic standards will be described through two selected cases at the National Institute of Standards and Technology (NIST). These descriptions will include strategies in the selection of appropriate gases, production of RM units, characterization, use, and treatment of data.
43.2 Strategies and practical aspects 43.2.1 Needs Assessment
Developers of standard materials require documented needs from communities that can specify the critical requirements for matrix, chemical species, quality parameters, and the anticipated impacts and outcomes of having the particular material available. For many decades, the Hydrology Section of the International Atomic Energy Agency (IAEA) has provided the leadership to assess the global needs for light isotope RMs, and to help develop and characterize materials through interlaboratory comparison exercises. This effort was possible only through the significant and mostly pro bono efforts from many individuals who have served the isotope community by preparing the materials or by participating in IAEA intercomparisons. NIST, along with several other organizations, is now collaborating with the IAEA in the provision of isotopic RMs. At NIST, the decision to develop (or reissue) a particular Standard Reference Material (SRM| or RM is based upon several factors, including: 1) the timeframe of the industrial or regulatory need, 2) the nature and size of the anticipated impact, 3) technical feasibility, and 4) ability to respond with requisite quality and timeliness. These projects compete for very limited resources, so they must be accompanied by clear, documented justification. The SRM program collects requests through the Website at http://www.nist.gov/cgi-bin/ndassess.cgi, or by email at [email protected]. NIST has also recently begun to leverage its efforts through the NIST Traceable Reference Material (NTRM) program. This program relies on the private sector to develop
Strategies and Practicalities in the Production and Use of Gas Isotope Standard Materials
947
and provide suitable materials to the measurement communities, while quality of these materials is assured through representative analyses at NIST. The NTRM program has not yet been applied to isotopic RMs, but as the user community grows and demands on NIST resources increase, this may become a mechanism for satisfying the needs of the measurement community. 43.2.2 Gas Selection
For isotopic RMs, naturally occurring gases having the appropriate isotopic composition are preferred over synthetic gases and mixtures. This assures the complexity and similarity to samples, and avoids some pitfalls associated with blending, homogenization, and the assurance of stability. This preference must be balanced against matrix effects in natural materials that may influence the measurements and confound assignment of accurate values. Matrix effects are not an issue for chemically pure gaseous RMs, and pure synthetic materials may be used effectively in many cases. With some gases, however, blending to achieve target compositions should be avoided for two important reasons: 1) the resulting mixture may be difficult to homogenize below levels detectable by modern instrumentation; and 2) the mixture of gases may be isotopically unequilibrated even after complete physical blending. Data reduction algorithms used to convert molecular isotope ratio measurements (~45CO2 and ~46CO2) into reported delta values (513C and 5180) require that the distribution of the molecular isotopic masses in the measured CO2 population (44 amu through 49 amu) be predicated by random distribution of the constituent stable isotopes (12C, 13C, 160, 170, 180). In clean vessels at room temperature, however, the rate of isotopic exchange between two blended CO2 materials is expected to be extremely low, since trace water (if present) would offer an exchange mechanism. The result may be a drift, perhaps over years, in apparent ~13C and 5180 (as measured against a stable RM) as redistribution of the isotopes occurs within the equilibrating population of CO2 molecules (Verkouteren & Dorko, 1989; Gonfiantini et al., 1997). This may be conceptualized through the example in equation [43.2], Physical Blend [ 44amu, 49amuonly [ 16012C16 O + 18013C180 ~ 16012C18 O + 16013C16 O + 16013C18 O + 18012C18 O I
[43.2]
Equilibrated Blend (44 amu through 49 amu)
although natural CO2 distributions are more complex. As a practical example, Table 43.2 shows the predicted isotopic compositions observed in unequilibrated and equilibrated blends of three pure carbon dioxide gases. For blends of natural abundance CO2 materials, (e.g., mixtures 3 and 4, Table 43.2), the effect is insignificant. The use of 180-enhanced CO2 (i.e., tracers), however, creates mixtures that may exhibit a small yet significant isotopic drift (e.g., mixtures 1, 2, and 5, Table 43.2)1. The same effect would be expected in blends of any chemically pure material containing at least two different atoms, each having two or more stable isotopes (e.g., CO2, CO, SO2, N20).
948
Chapter 43 - R.M. Verkouteren
Table 43.2 - Comparison of Predictedt Isotopic Compositions of CO2 Materials, Non-Equilibrated (Physical Mixtures) and Isotopically Equilibrated Mixtures CO2 Blend (molar % of A:B:C)
Physical Mixture Only
Gas A (100:0:0) Gas B (0:100:0) Gas C (0:0:100) Mixture I (90:0:10) Mixture 2 (99:0:1) Mixture 3 (90:10:0) Mixture 4 (50:50:0) Mixture 5 (49:49:2)
-5.945 -4.199 -7.801 -23.005 -23.025
Observed 613CVPDB (%o) Isotopically Drift during Equilibrated Equilibration Mixture -4.000 -42.000 -20.000 -5.600 -4.160 -7.802 -23.004 -22.944
+0.345 +0.039 -0.001 +0.001 +0.081
Observed 618OVPDB (%o) Physical Isotopically Drift during Mixture Equilibrated Equilibration Only Mixture 72.541 0.052 -9.601 -16.002 0.291
-8.000 -24.000 +800.000 72.670 0.066 -9.600 -16.000 0.259
-0.129 -0.014 +0.001 +0.002 +0.032
Predicted values of 813C and 6180 from measurements of 645CO2 and846CO2 reduced using the standard algorithm, where the relationship among the stable oxygen isotopes for each gas is assumed to be: 6170 = a 6~80,where a = 0.50. The effects observed here are therefore not influenced by perturbations in this relationship, although variations in the value of a between the gases would add another layer of complexity to the calculations.
43.2.3 Preparation Preparation methods are generally RM specific, since each material offers unique challenges. While focused on chemical rather than isotopic RMs, background information and a broad array of methods may be found in the proceedings of the last seven symposia on Biological and Environmental Reference Materials (BERM)(Iyengar & Wolf, 1998). Preparation of some isotopic RMs may be found in publications of the IAEA (Hut, 1987; Gonfiantini et al., 1995). Few preparation procedures of gas isotope RMs have been published (Coplen & Kendall, 1982; Meijer, 1995; Valkiers et al., 1999; Verkouteren, 1999). Handling requires clean, high vacuum apparati and methods that minimize isotopic fractionation, a phenomenon that accompanies gas expansion into evacuated vessels. There are essentially two strategies for packaging: breakseals and cylinders, with the advantages and disadvantages of each listed in Table 43.3. The gas isotope program at NIST has utilized both strategies for application to different gases and community needs. Carbon dioxide isotopic RMs 8562, 8563, and 8564 in breakseals were recently prepared and characterized (Verkouteren, 1999), and the Natural Gas Standards (NGS-1, NGS-2, and NGS-3) have been prepared and distributed in pressurized cylinders since 1993.
1. In this example, we assume that the 180-enhanced CO2 follows the normal fractionation relationship among the 180, 170, and 160 isotopes. For CO2 that contains oxygen not following the normal relationship, the resulting mixtures will likely exhibit more pronounced drifts.
Strategies and Practicalities in the Production and Use of Gas Isotope Standard Materials
949
Table 43.3 - Pros and cons of two packaging strategies for gas isotope standards. Breakseal Strategy Advantages
Disadvantages
9 9 9 9 9 9
Simplicity of packaging High level of contamination control Thousands of independent units High statistical control of process High statistical control of process Batch certification
Cylinder Strategy 9 Ample amount of gas (may be pressurized but not liquefied) 9 Multiple uses (including conditioning) 9 Applicable to most gases 9 Convenient for users; rugged 9 Inexpensive (on per mass of gas basis)
Stability issues: requires determi9 One use design nation of compositional stability 9 Limited amount of gas (for w / r / t time, cylinder pressure, conditioning) temperature etc.: problematical, 9 Requires tube cracker may be cylinder dependent 9 Safety issues: applicable to noncombustible and non-toxic gases only 9 Contamination issues: possible back-streaming during use 9 Expensive (on per mass of gas basis) 9 Requires rigorous determination of 9 Certification issues: unit (rather than batch) certification probably sample-to-sample compositional heterogeneity and stability necessary
Examples
NSVEC ; [prepared by G. Junck, H.J. Svec & C. Kendall, unpublished] NBS-16,17; [Coplen & Kendall, 1982] RMs 8562,8563,8564 [Verkouteren, 1999]
H2 ring test; [Brand & Coplen, 2001] NGS-1,2,3; [Verkouteren, unpublished] GS-19,20; [Meijer, 1995] IMEP-8 [Norgaard et al., 2002]
Comments
Appropriate for some primary standards and reference materials; provides high level of traceability to primary scale
Appropriate for laboratory standards and intercomparison materials, and may be appropriate for some primary standards and RMs
43.2.3.1 Breakseal Strategy: C02 RM$ 8562-8564 A gas m a n i f o l d w a s d e s i g n e d to p r e p a r e over one t h o u s a n d RM units in t u b u l a r breakseals (Figure 43.1), while m i n i m i z i n g isotopic fractionation a n d protecting the gas from c o n t a m i n a t i o n over the t h r e e - w e e k d u r a t i o n of the p r o d u c t i o n stage. In the first week, the gas is i n t r o d u c e d into the entire s y s t e m a n d recirculated t h r o u g h a trap to d r y a n d h o m o g e n i z e the gas a n d equilibrate the system. P r e p a r a t i o n of the RM units occurs d u r i n g the following t w o weeks. The CO2 gas is circulated t h r o u g h either of t w o pathways" a h i g h v o l u m e p a t h w a y h o l d i n g u p to 20 L of CO2 in a stainless steel bellows, or a low v o l u m e p a t h w a y (250 mL) u s e d to cryogenically purify, h o m o g e nize, a n d distribute the C 0 2 into parallel glass tubes (9 m m outer diameter) that are each sealed into four RM units. P r e p a r e d in this way, the RM s a m p l e s are statistically indistinguishable across every nuisance factor of p r o d u c t i o n (day of production, lot n u m b e r , position on manifold, s p e e d of sealing process, etc). One i m p o r t a n t consideration is the fabrication, positioning, a n d t r e a t m e n t of the glass tubes. To allow easy
950
Chapter 43 - R.M. Verkouteren
Figure 43.1 - System schematic for production of gas isotopic RMs in breakseals. This system is appropriate only for non-flammable and non-toxic gases and mixtures. The gas is introduced into a 20 L stainless steel bellows and circulated through the entire system, including the trap chilled to - 78~ for 5 days at 400 mL/min. Glass tubes, each segmented into tubular breakseals, are filled with gas and sealed. See text and Verkouteren (1999) for further details. Other components include: (1) manual bellows valves; (2) pneumatic bellows valves; (3) diaphragm valve; (4) oil-free molecular drag vacuum system; (5) absolute pressure transducer. Electric and pneumatic air supplies are protected by backup systems.
sealing, each tube is pre-drawn on a lathe into four segments of about 25 cm each, with the tubes cut afterwards to a length of 121.9 cm + 0.3 cm. This length and precision is necessary for each tube to fit consistently into the manifold. While the ends of the glass tubes are sealed to the manifold with polytetrafluoroethylene compression fittings, Cajonl O-ring face seal fittings are also employed on one end to provide zeroclearance capability. The precision of length and the need to anneal the fragile tubes to remove surface water requires that this work be done on-site. After seven tubes are positioned in the manifold, they are leak checked by a helium detector, flushed with CO2 that remains in the low volume pathway from the production of the previous lot, evacuated, and opened to the bellows in the high volume pathway. This strategy minimizes effects from possible contamination by buffering and equilibrating each lot against the gas in the 20 L volume, and minimizes isotopic fractionation by filling every lot consistently at atmospheric pressure before recirculation and homogeniza1. Certain commercial equipment, instruments, and materials are identified in this paper to specify adequately the experimental procedure. Such identification does not imply recommendation or endorsement by the National Institute of Standards and Technology, nor does it imply that the equipment or materials identified are necessarily the best available for the purpose.
Strategies and Practicalities in the Production and Use of Gas Isotope Standard Materials
951
tion through a metal bellows pump. Further details of the production process are published elsewhere (Verkouteren, 1999).
43.2.3.2 Cylinder Strategy: The NGS Intercomparison Materials In 1984, a collaboration among the Chevron Oil Company, AGIP (Azienda Generale Italiana Petroli), the University of Groningen (The Netherlands), and the IAEA led to the development of three natural gas isotopic intercomparison materials (the so-called Natural Gas Standards" NGS-1, NGS-2, and NGS-3). These materials were comprised of light hydrocarbons (and associated gases) from distinct sources, and were needed by the petroleum and energy industries to test new measurement techniques and establish a means for international comparability of isotopic measurements. As these were the only isotopic standards that contained gaseous hydrocarbons, interest by other groups soon followed, including communities involved in the monitoring, speciation, and source apportionment of air toxics, atmospheric methane, and volatile organic compounds. Through an international comparison organized by the IAEA, chemical and isotopic values were reported (Hut, 1987), and since then, additional chemical and isotopic values for these materials have been published in the open literature (Dumke et al., 1989; Sohns et al., 1994). Distribution of the NGS gases was administered by the Chevron Oil Company until 1992, when under IAEA recommendation, the six 50 L high-pressure stock cylinders were transferred to NIST. Guidelines for filling small vessels were formulated by NIST and the IAEA as follows" 1) Use bakable all-welded stainless steel manifold, bellows valves, and oil free pump. 2) Use new 50 mL stainless steel cylinders, bellows valves, and associated hardware to construct vessels. Threaded connections and small cylinder size are required per U.S. Department of Transportation (DOT), International Air Transport Association (IATA), and International Civil Aviation Organization (ICAO) regulations regarding hazardous materials. 3) Assemble vessels using clean tools and working area, and mount onto manifold. 4) Leak check system using both high-pressure and vacuum helium tests. 5) Flush system with dry helium, and evacuate overnight using oil-free pump while heating at 100~ 6) Preheat stock cylinders to 80~ using progressive cylinder heating system (see section below on safety consideration), where bottom of cylinder is heated first to allow convective blending of the compressed hydrocarbon mixture. 7) Preheat all transfer lines and manifold hardware to 80~ and fill vessels with gas mixture at cylinder pressure (about 50 bar). 8) Uniquely label and record information regarding each unit, including position on manifold, filling pressure, and other variables. An exercise to assign isotopic values and uncertainties is currently in progress to upgrade these gases from intercomparison materials to RMs, after which they shall be designated RMs 8559, 8560, and 8561.
952
Chapter 43 - R.M. Verkouteren
Safety Considerations for Preparation of the NGS Materials Special procedures are followed when heating the 50 L pressurized steel stock cylinders to homogenize convectively the NGS hydrocarbon gases (C1 to C5 species) within each cylinder. Under no circumstances should this be attempted with any aluminum cylinder, or any steel cylinder that lacks a valid inspection for defects. Controlled heating is accomplished with three silicone rubber heater belts (208 volt, 1540 watts each) placed around the bottom, middle, and top of the cylinder, and controlled by a Watlow Series 965 microprocessor-based auto-tuning controller. Each heater belt has an embedded thermocouple (type J) and limit controller (Watlow series 140) that is set to cut power to all heaters should any temperature exceed ll0~ or if any of these thermocouple circuits is broken. Additionally, temperature is visually monitored on a computer from three thermocouples (type K) pinned underneath the heaters. Thirty-minute running histories of temperature at each position are displayed for the operator. Temperature control comes from a seventh thermocouple (type J) fastened to the cylinder about 5 cm above the middle heater. On our system, the control set point is 80~ with a 5 second cycle time, 20% power output, and high limit of 100~ Each heater is fitted with a manual on-off switch. Cylinder heating progresses in three stages over the course of a day: 1) At the start (T = 0 hrs), the bottom heater is activated; 2) at T - 2 hrs, the middle heater is activated (leaving the bottom heater on); and 3) at T - 4 hrs, the top heater is activated (leaving the other two on). The target temperature is reached by T - 5 hrs, and at T - 6 hrs the cylinder contents are adequately homogenized. The heated manifold is flushed with the gas mixture, vented and evacuated, then filled up to the pressure of the heated stock cylinder. 43.2.4 Characterization strategy The required characterizations of isotopic RMs include chemical composition or purity, sample-to-sample heterogeneity (Pauwels et al., 1998a), stability (Pauwels et al., 1998b), value assignment, and uncertainty estimation (ISO/BIPM, 1995). The implementation of these characterizations is frequently different between the breakseal and cylinder approaches. For RM breakseals, a representative sampling approach must be used, where a large number of units are randomly selected and sacrificed to obtain averages with valid statistical uncertainties. Cylinders are usually sampled rather than sacrificed for purposes of characterization. This sampling may involve a representative number of cylinders if the observed heterogeneity across cylinders is within the uncertainty needs of the customers, and longer-term observations indicate the isotopic compositions are stable within the stated shelf life. Otherwise, each cylinder may need to be sampled over time to individually correct for isotopic heterogeneity and drift. Value assignment, whether in breakseals or cylinders, involves measuring and relating the RM isotopic composition (and uncertainty) to the appropriate artifact scale or to fundamental SI units. While assignment of an absolute ratio is extremely challenging, the assignment of a standardized delta value is also deceptively difficult. Intercomparisons (Hut, 1987; B6hlke & Coplen, 1995; Robinson, 1995; Stichler, 1995; Kornexl et al., 1999b; Verkouteren, 1999; Norgaard et al., 2002; Verkouteren & Klinedinst, 2003) of RMs have invariably shown excellent measurement repeatability within laboratories while exhibiting generally poor measurement reproducibility across laboratories, indicating that individual laboratory procedures and
Strategies and Practicalitiesin the Productionand Use of Gas Isotope StandardMaterials
953
processes lead to systematic differences in realization of the standard isotopic scales. At NIST, several strategies are used for value assignment and uncertainty assessment (May et al., 2000). For isotopic materials, we currently employ intercomparisons coupled with NIST measurements and exhaustive knowledge about error sources and the state-of-the-art. New methods are needed to improve this process, and are being developed along several fronts worldwide, notably in the development of improved instrumentation, the development of primary methods for measuring the amount ratios of gas isotopes (Aregbe et al., 1998), the precise measurement of the relationship between 170 and 180 abundances in nature (Meijer & Li, 1998), the direct measurement and correction of cross contamination in ion sources (Meijer et al., 2000), the development of Web-based data processing systems (Verkouteren & Lee, 2001), and the use of extended isotopic measurements such as 847CO2 to provide information regarding the accuracy of customary isotopic measurements (845CO2, 846CO2)(Verkouteren, 1999). 43.2.5 Distribution
Once prepared and characterized, an RM is made available through any of the national or international standards agencies having mechanisms for inventory, marketing, and distribution. The importance to the community of valuing and maintaining an isotopic RM cannot be overstated, since they are fruits of tremendous efforts and may be difficult and costly to replace. By agreement between NIST and IAEA, only one unit of an isotope RM may be sold to any institution per 3-year period. This is to ensure that the RMs will be highly valued and used to assign values to secondary standards for routine use in laboratory quality assurance. At NIST, isotope RMs are accompanied by a Report of Investigation that describes the preparation process, the value assignments and methods of analyses, quality parameters, storage information, and recommended usage. 43.2.6 Measurement strategy
An IRMS instrument is needed, running in dual inlet or isotope ratio monitoring mode, with a clean ion source and using good vacuum practices. At least two equilibrated and dry laboratory reference gases in 2 L vessels or larger should be prepared that compositionally span across the working range of typical samples. We recommend that all RMs and laboratory reference gases be sampled and measured through the same capillary in a consistent, serial manner against a single machine reference gas entering through a matched capillary. Although leaks may be matched with respect to chemical leak rate and zero enrichment measurements at a given inlet pressure, variations may exist in fractionation behavior across a leak as a function of isotopic composition (Mook & Grootes, 1973). By relating RMs and laboratory reference gases by measurements through the same leak, this source of systematic error is reduced. 43.2.7 Data treatment
Delta measurements of molecular isotope ratios are reduced to the standardized delta values of the elemental isotope ratios. Standard algorithms have been published for carbon-13 and oxygen-18 in CO2 (Gonfiantini, 1981). These algorithms are usually
954
Chapter 43 - R.M. Verkouteren
embedded in the operating software of modern IRMS machines, but they should be tested since variations in outputs among algorithms- given the same input d a t a have been observed across instrument models and manufacturers (Allison & Francey, 1995). A set of standard test data (STD) for CO2 have been published (Allison et al., 1995) to test embedded algorithms, although these STD are useful only when an instrument manufacturer has allowed access to the embedded algorithm by user supplied input. NIST has posted a Web-based interactive standard reference algorithm (SRA) at http://www.nist.gov/widps-co2 that may be used to test embedded algorithms and associated data systems with any real measurement data (Verkouteren & Lee, 2001). This SRA, validated against the STD of Allison et al. (1995), will generate standardized and normalized results for any sample and RM(s). It may also be used to note the effects of changing fundamental assumptions that are inherent in the recommended algorithm. Proportional errors (Blattner & Hulston, 1978) in IRMS instruments are common problems that introduce systematic error into relative isotopic measurements, especially with samples having compositions quite different from the primary artifact. Normalization is one method used to correct for this error, and this technique has been recommended (Coplen, 1996a) by the International Union of Pure and Applied Chemistry (IUPAC) for improving the realization of the VSMOW scale for 180 and 2H. Similar recommendations for other scales have been more elusive since appropriate primary standards have not yet been developed. However, normalizing procedures may still be applied to any isotopic system having at least two reference materials with assigned values that bracket the working isotopic range. A normalized (norm) delta value (for CO2, m - 13, 18, 45, or 46) may be obtained through equation [43.3],
Im
m
m m (~Salnmeas -- t~RMl.meas ~Sarnnorm - ~RMl.assignd + - G - - - ~ - - --W-. . . . . ~RM2.meas -- ~RMl.measl
IIx
t~mM2.R m assignd - ~RMl.assignd
I [43.3]
where measured (meas) delta values for two RMs (RM1 and RM2) that bracket the composition of the sample (Sam) are entered, along with the value assignments (assignd) for the two RMs. When reporting measurement data that has been normalized in this manner, the assigned values of the RMs used to realize the isotopic scale should be reported. Improvement in realizing the VPDB scale by normalization is exemplified through data from a recent intercomparison (Table 43.4). Each of thirteen laboratories reported 613C and 6180 values versus VPDB for the CO2 RMs 8562, 8563, and 8564. For each laboratory, data from the bracketed RM (RM 8564 for ~13C; RM 8562 for 6180) was normalized through equation [43.3]. The resulting improvement realized in interlaboratory reproducibility was dramatic. The standard deviation of the mean decreased by a factor of 3.7 (from 0.078%o to 0.021%o) for 613C, and by a factor of 3.8 (from 0.204%o to 0.053%o) for 6180. By removing the statistical outlier (Lab ID #1), the ~5180interlaboratory variability decreased by a factor of 5.8.
-~
-__
Lab ID
~
~
_
RM 8562 (reported)
_
_
_
6 1 3 C V P D B (%o) ~ _ _ _
RM 8563 (reported)
RM 8564 (reported)
-3.75 -3.81 -3.81 -3.74 -3.78 -3.76 -3.67 -3.72 -3.79 -3.61 -3.76 -3.60 -3.77
1 2 3 4 5 6 7 8 9 10 11 12 13
_ _ _ _ _
~~
Mean S ~-
-41.27 -41.19 -41.64 -41.37 -41.51 -41.55 -41.22 -41.38 -41.67 -41.13 -41.39 -40.96 -41.11
-3.735 0.069
-41.319 0.219
~~
6180VPDB-C02 ~~
RM 8564 (normalized)+ ~~
~
~~
~~
('roo)
~~
RM 8563 (reported) __
RM 8564 (reported) ~
_
RM 8562 (reported) _
_
RM 8562 (normalized)*
_
_
-10.42 -10.44 -10.51 -10.40 -10.45 -10.45 -10.35 -10.40 -10.54 -10.29 -10.47 -10.27 -10.40
-10.48 -10.47 -10.46 -10.45 -10.44 -10.45 -10.48 -10.46 -10.49 -10.49 -10.50 -10.51 -10.47
-33.91 -33.13 -33.76 -33.54 -33.67 -33.71 -33.27 -33.24 -33.77 -32.90 -33.49 -32.88 -33.98
-10.05 -10.05 -10.08 -10.00 -10.03 -10.02 -9.94 -9.81 -10.10 -9.67 -10.01 -10.01 -10.37
-18.50 -18.38 -18.63 -7 8.48 -18.59 -18.52 -18.39 -18.29 -18.64 -18.04 -18.43 -18.20 -18.83
-18.34* -18.50 -18.50 -18.49 -18.53 -18.45 -18.54 -18.53 -18.50 -18.49 -18.45 -18.44 -18.44
-10.414 0.078
-10.474 0.021
-33.458 0.362
-10.010 0.160
-18.455 0.204
-18.488* 0.035*
~
~~~~~
_ _ _ _ _ _ _ _ _ _ _ _ _ ~
*Statistical outlier (Lab ID #1)removed before calculation of mean and standard deviation. Data normalized through equation [43.3],where assigned ~ Y ~ C V Pvalues D B of RM1 (RM 8562) = - 3.76%~ and RM2 (RM 8563) = - 41.56%. Data normalized through equation [43.3], where assigned ~ ~ * O V P D Bvalues - C O ~of RM1 (RM 8563) = - 33.63% and RM2 (RM 8564) = - 9.96%. IVerkouteren (1999).
*
Strategies and Practicalities in the Production and Use of Gas Isotope Standard Materials
Table 43.4 - Reported data of RMs 8562,8563, and 8564 from intercomparison] and effect of normalization.
955
956
Chapter 43 - R.M. Verkouteren
43.3 Conclusions
Gas isotope RMs can provide quality assurance for the increasing number of applications of isotopic measurement. Performance-based strategies and practical aspects regarding the selection, preparation, characterization, and use of gas isotope RMs have been described, especially in the context of two different material types currently being distributed at NIST: CO2 materials (RM 8562, RM 8563, and RM 8564); and the Natural Gas Standards (intercomparison materials NGS-1, NGS-2, and NGS-3). Assessment of needs, gas selection criteria, safety considerations, measurement strategy, data normalization, and a Web-based interactive data reduction algorithm were discussed. The practical aspects should not be considered prescriptive, but rather as examples that may be adapted to fit other specific projects. By utilizing gas isotope RMs and associated tools as a part of a total quality assurance system, improvements in reproducibility of measurement may be realized across laboratories, as exemplified by normalizing results of the recent intercomparison of three carbon dioxide RMs.
Handbook of Stable Isotope Analytical Techniques, Volume 1 P.A. de Groot (Editor) @ 2004 Elsevier B.V. All fights reserved.
CHAPTER 44 Data Corrections for Mass-Spectrometer Analysis of SO2 Max Coleman Postgraduate Research Institute for Sedimentology, University of Reading, UK Current address: Center for Life Detection, Jet Propulsion Laboratory, California Institute of Technology, 4800 Oak Grove Drive, MS 183-301, Pasadena, CA 91109-8099, USA e-mail: max.coleman~pl.nasa.gov
44.1 Introduction Analysis of stable isotope ratios of sulphur may be performed in an isotope ratio mass-spectrometer (IRMS) with the sample prepared either as SF6 (Rees, 1978) or as SO2 (Thode et al., 1961; Robinson & Kusakabe, 1975). Since fluorine is monoisotopic (19F only) ratios of m / z measured at 146 and 148 represent differences of 34S relative to 32S. However, SO2, which is used more frequently as the sulphur molecule to be analysed, contains oxygen which has three isotopes, 160, 170 and 180. The principle masses measured in a triple collector mass spectrometer are m / z - 64, 65, 66 which correspond predominantlyto 32S, 33S and 34S, however, neither of the latter two is simply the product of a heavier sulphur isotope. Thus, the purpose of the correction procedure is to calculate the ratios of the sulphur isotopes 34S/32S in terms of the data measured for the molecular species. The philosophy of the correction procedure is based on that of Thode et al. (1949) and Craig (1957) and in the approach presented here was first formulated for sulphur isotope corrections for a double collector mass-spectrometer (Coleman, 1980). The corrections below apply to a conventional dual inlet IRMS, where a sample of unknown composition and a reference gas are alternately introduced to the source of the mass spectrometer. This method may not be the most rapid but does produce the most precise results. The main problem then is accuracy. In this chapter, three aspects of data correction will be addressed" 9 Derivation of the algebraic expressions used to correct measured data 9 Evaluation of the sensitivity of the final result to uncertainties in the corrections 9 Practical routine application of the correction procedures in a laboratory Although sulphur isotope differences are usually measured in the mass spectrometer for the SO2 + ion, it is possible to use the SO + fragment, produced almost as abundantly in the ion source of most mass-spectrometers ( m / z - 48, 49, 50).
958
Chapter 44- M. Coleman
44.2 Isobaric interferences Table 44.1 - Contributions to the sulphur mass spectrum. The molecules contributContributing molecular species m/z ing to each relevant mass are given in Table 44.1. Thus, it 48 32S160 can be seen that there is a 49 33S160 + 32S170 potential need for compli50 34S160 + 33S170 + 32S180 cated data reduction becau32S16016O se masses 65 and 66 or 49 64 33S160160 + 32S170160 65 and 50 have contributions 34S160160 + 32S180160 + 32S170170 + 32S170160 66 from more than one oxygen isotope. The masses measured depend on the oxygen isotope composition of the SO2 as well as that of the sulphur. The aim of the correction is to remove the effects of oxygen isotope composition to reveal the true sulphur isotope values. 44.3 Notations used It should be noted that since there are two different oxygen atoms in the SO2 molecule, the molecular abundances of the trace oxygen isotopes, 170 and 180, are twice the atomic abundances. For example, that species at m / z = 66 consisting of 32S, 180 and 160 is written as 32S180160. Since there are two oxygen sites, the probability of 180 substituting for 160 is twice its atomic abundance. Therefore, pedantically, it should be written as 3 2 S 1 8 0 1 6 0 + 3 2 S 1 6 0 1 8 0 . Thus, the species is written only as 32S180160 but the abundances of trace oxygen isotopes in the SO2 molecule are not simply the atomic abundances. In evaluating such expressions, therefore where two oxygen atoms are present the atomic abundances multiplied by two are used. For the fragment SO +, the oxygen isotope abundances derive directly from the atomic abundances.
Table 44.2 - Notations used to describe mass ratios
Notation
Full expression
R17 R18
17/16 atomic abundance ratio 18/16 atomic abundance ratio
R33 R34
33/32 atomic abundance ratio 34/32 atomic abundance ratio
R49 R50
49 / 48 [m / z] / [m / z] ratio 50 / 48 [m / z] / [m / z] ratio
R65 R66
65 / 64 [m / z] / [m / z] ratio 66 / 64 [m / z] / [m / z] ratio
The notation used in the following calculations are given in Table 44.2. Note that m / z : 64 is abbreviated to 64 and similarly for all other masses measured and all atomic masses. Each of the ratios shown in Table 44.2 can apply to either the sample of u n k n o w n composition (x) or to the reference gas (r). Thus the mass spectrometer will alternately measure R66x and R66r. To abbreviate the algebraic expressions the values for the reference gas are shown without a suffix.
Data Corrections for Mass-SpectrometerAnalysis of SO2
959
Thus, R66 implies R66r. The 6 notation usually expresses the difference between sample and reference, relative to the reference value, in parts per thousand or per rail (%0). Thus, ~)34S =
((R34x - R34)/R34). 1000
[44.1]
However, for purposes of calculation it is easier to use the 6 value as a simple ratio to avoid multiplying many parts of the expression by 1000. Also, 634S is abbreviated to 634. So, 634 = (R34x - R34)/R34
[44.2]
This is more easily handled in the form of 634 = (R34x / R34)- 1
[44.3]
Similarly, 666 = (R66x / R66)- 1
[44.4]
and in the same way for 833, 849, 850 and 865 As outlined in section 44.2 (Table 44.1), in order to measure accurate isotopic compositions it is necessary to correct for oxygen isotope contributions to the SO2 mass spectrum. To achieve such corrections, the atomic ratios, R17, R18, R33 and R34 (Table 44.2) must be known. There are published R values for oxygen and sulphur in internationally-recognised reference materials: Standard Mean Ocean Water (SMOW) and Canyon Diablo Troilite (CDT); shown in Table 44.3. It can be seen that there are some discrepancies in the values but there are no clear criteria for best choice (Gonfiantini et al., 1995; and especially, Santrock et al., 1985). Therefore, the sensitivity of the corrections to these values needs to be ascertained too, and this matter is addressed below (section 44.6.2). We can now re-address the basic problem. The mass spectrometer measures 666 or 50 but we want to determine 834. 44.4 Relationship of 834 to ~66 and 865 866 is a function of R66x and R66 which, in turn, can be expressed in terms of the isotopic molecular entities involved (see Table 44.1). R66
-
(34S160160 + 328180160 + 32S170170 + 338170 160 ) 32S160160
[44.5]
960
Chapter 44- M. Coleman
Expressing the parts as simpler ratios by dividing each part of the n u m e r a t o r by the denominator (see Tables 44.1 and 44.2) R 6 6 - R34 + 2-R18 + R17. R17 + 2. R33. R17
[44.6]
Note that 2.R18 is used instead of R18 because, as explained in section 44.3, 32S180160 in fact represents 32S180160 and 32S160180 and therefore in this molecule there is twice the probability of 180 occurring than in the atomic oxygen; and similarly; for 2.R33.R17 from 33S170160 and 33S160170. R66x can be expressed similarly. These two values can be combined to give 666 (equation [44.4]). However, it is easier to rearrange equation [44.4] in the form of 1 + 866 = R66x / R66
[44.7]
1 + 666 - R34x + 2. R18x + R17x. R17x + 2. R33x. R17x R34 + 2 . R18 + R17. R17+ 2. R33. R17
[44.8]
But it is possible to express R34x in terms of a k n o w n quantity, R34, and the required value, 634. So by rearranging equation [44.3] R34x = (1 + 634). R34
[44.9]
Since, to a first approximation, any fractionation is linearly d e p e n d e n t on mass difference then any effect on the 33/32 ratio will be half that on the 34/32 ratio. So 633 - 634 / 2
[44.10]
and analogously with equation 440.9] it is possible to evaluate R33x. R 3 3 x - (1 + 634/2). R33
[44.11]
A similar relationship can Table 44.3 - Relative isotopic abundances of oxygen and sulphur. be formulaed for variations Ratio Value Source of 17/16x which can be related to 18/16x but a R17 0.0003720 Craig (1957) more precise factor has 0.000373 Santrock et al. (1985) been determined by Meijer 0.000402 Santrock et al. (1985) & Li (1998) 0.00037983 Li et al. (1988) R18 0.0019934 Craig (19 61) 0.0020052 Baertschi (1976) 617 = K. 61 [44.12] R33 0.0078795 Ding et al. (1998) 0.00789 MacNamarra & Thode (1950) where K - 0.528 R34 0.0441623 Ding et al. (1998) and, analogous to equa0.044994 Thode et al. (1961) tions [44.9] and [44.11] 0.045035 Zhang & Ding (1989)*
Data Corrections for Mass-Spectrometer Analysis of SO2
961
R18x = (1 + 518). R18
[44.13]
R17x = (1 + ~.. 618). R17
[44.14]
Substituting values from equations [44.9], [44.11], [44.12] and [44.14] in equation [44.8]. 1+ 566 =
(1 + 534)R34 + 2(1 + 618)R18 + (1 + ~.. 618)2. R172 + 2(1 + ~34/2)R33. (1 + ~. ~18)R17 R34 + 2. R18 + R172 + 2. R33. R17 [44.15] Equation [44.15] can be rearranged by cross multiplying; the products of any two 6 terms have been ignored since they are negligibly small. (R34 + 2. R18 + R172 + 2 - R 3 3 . R17) + 566(R34 + 2. R18 + R172 + 2. R33. R17) = (R34 + 2. R18 + R172 + 2. R33. R17) + 534(R34 + R33. R17) + 518(2. R18 + 2K. R172 + 2)~. R33. R17) [44.16] This reduces to an expression giving 634 in terms of 566 634 - 666- (R34 + 2. R18 + R172 + 2. R33. R17) (R34 + R33 + R17) - 6 1 8 . (2. R18 + 2. K. R172 + R 3 3 . 2 - K. R17) (R34 + R33 + R17)
[44.17]
Substituting R values (arbitrarily using the most-recently published data, rather than the others, Table 44.3) in equation [44.17] gives 634 = 1.089 {366 - 0.0891 {318
[44.18]
Clearly, there is not only a substantial correction needed for ~66, but also a dependence on the oxygen isotope difference of sample and reference gases. Evaluation of R65 can be m a d e in a similar way, but is m u c h simpler R65 -
33S 160 160
+
32S 170 160
32S16016 O
R65 = R33 + 2R17
[44.19]
[44.20]
Similarly, R65x = R33x + 2R17x
[44.21]
962
Chapter 44- M. Coleman
In the same way as before, the Rx values in equation [44.21] can be substituted by 5 values and R values 1 + 665 = R65x / R65
[44.22]
and from equations [44.20], [44.21] and [44.22]. 1 + 565 =
R33x + 2R17x R33 + 2R17
[44.23]
1 + 565 - (1 + 534/2)R33 + 2(1 + K. 518)R17 R33 + 2R17
[44.24]
this re-arranges to an expression for 634 in terms of 565 and 618 534 - 6 6 5 . 2 ( R 3 3 + 2R17)_ 6 1 8 . 2 . K. 2 - R 1 7 R33 R33
[44.25]
As before it is possible to evaluate this expression by involving the R values 634 = 2.193 665 -0.1017 518
[44.26]
Thus, in theory, 618 need not be known since measurement of both 665 and 666 gives two equations, [44.18] and [44.26], that can be solved simultaneously to eliminate 618. Since the 518 cannot be determined easily it might be thought valuable to solve the two simultaneous equations to eliminate the ~518terms, which gives 634 = 10.16 666 - 18.23 565
[44.27]
Clearly, the correction coefficients of the two measured values will impose a very large error multiplier on the uncertainty in the measured values of 666 and 665. However, there is a possible, alternative approach. This involves measurement of the SO + species.
44.5 Relationship of 634 to 650 and 649 Although the principle ion species produced from SO2 in the mass spectrometer source is SO2 +, the fragment SO + is produced almost as efficiently. If it is assumed that the ionisation processes which produce SO2 + and SO + do not give a significant isotopic fractionation, then measurement of the SO + masses should give comparable results. The equations, equivalent to those for SO2 +, [44.5] to [44.18], are [44.28] to [44.37] and are listed below, rather than described. R50 -
34S 160
+
32818 0 32816O
+
33S 170
[44.28]
963
Data Corrections for Mass-Spectrometer Analysis of SO 2
R50 = R34 + R18 + R33. R17
[44.29]
1 + ~50 = R50x / R50
[44.30]
1 + 650 - R34x + R18x + R33x. R17x R34 + R18 + R33. R17
[44.31]
1 + 650 - (1 + ~34)R34 + (1 + 1518)R18 + (1 + 6 3 4 / 2 ) R 3 3 . (1 + K. ~18). R17 R34 + R18 + R33. R17
[44.32]
(R34 + R18 + R33. R17)(1 + 650) = R34 + R18 + R33. R17 + ~34(R34 + R33. R17/2) + 618(R18 + K. R33. R17) [44.33] 634 - 650. (R34 + R18 + R33. R 1 7 ) _ 6 1 8 . (R18 + K . R33. R17) (R34 + R33. R 1 7 / 2 ) (R34 + R33- R 1 7 / 2 )
[44.341
Similarly for R49 R49 -
33S16 O
+
32S17 O
[44.35]
32S16 O
[44.36]
R49 ~ R33 + R17
Since this has the same form as equation [44.20] its final formulation in equation [44.37] will be similar to equation [44.25] but R17 is substituted for 2R17, because the values for oxygen isotope abundances for the SO + species are different from those for SO2 + 534 - 1549.2(R33 + R 1 7 ) _ 5 1 8 . 2 K - R 1 7 R33 R33
[44.37]
The numerical expressions of equations [44.34] and [44.37] are given below. [44.38]
534 = 1.045 ~50 - 0.0454 518 534 = 2.096 549 - 0.0509 518
[44.39] T a b l e 44.4 - C o m b i n a t i o n s
of s u l p h u r i s o t o p e
The two equations, [44.38] and [44.39] are measurements to e l i m i n a t e 618 similar in form to their equivalents for 534 = 650 x 2.09 - 666 x 1.09 m e a s u r e m e n t of the SO2 + ions a t m a s s e s 634 = 550 x 1.89- 665 x 1.77 64, 65 and 66, equations [44.18] and 6 3 4 = 6 4 9 x 4 . 7 7 - 6 6 6 x l . 3 9 [44.26], and do not immediately seem to 634 = 649 x 4 . 1 9 - 665 x 2.19 offer an opportunity to reduce uncertainty 634 = 666 x 9 . 7 4 - 665 x 17.4 by elimination of the u n k n o w n 518 t e r m . 634 = 650 x 10.2 - 649 x 18.2
[44.40] [44.41] [44.42] [44.431 [44.44] [44.451
964
Chapter 4 4 - M. Coleman
44.6 Minimising uncertainty in correction equations It is clear that the corrected value for 634 depends not only on the precision of measurement of the principal mass (66) in the sample and reference, but also on 618, R17 and R18 of the reference gas, amongst other factors. Thus, to use these equations it is necessary to calculate the relevant values for the specific reference gas. However, it is also important to estimate the uncertainties in the isotopic values of the reference gas and in the measured differences from those values. 44.6.1 Uncertainties in I~18 From equation [44.27] it is evident that it is possible to eliminate 618 from the correction procedure, but that the error magnification may be very large. However, there are other possible combinations of equations to eliminate the 618 term. All possible combinations of equations, [44.18], [44.26], [44.38] and [44.39] and the resultant correction coefficients are listed in Table 44.4, in order of increasing magnitude of the coefficients. The R values used to calculate the coefficients are again the most recently published ones (Table 44.3) as were used in equation [44.27].
It is now obvious that the overall uncertainty is dependent not only on the precision of the measured isotopic differences, but also on the magnitudes of the factors in the combined equations. Measurements that combine data from SO2 + (masses 64, 65, 66) and SO + (masses 48,49,50) give the best opportunity for minimising error magnification. However, there are also uncertainties in the absolute values of the isotopic abundances of oxygen and sulphur (see Table 44.3). This aspect is treated in the next section. The alternative solution to the 618 problem is to ensure that its value is zero, in which case that term in the correction equation disappears. How this may be achieved in certain circumstances is dealt with in section 44.7, below. 44.6.2 Uncertainties in absolute isotopic abundances To investigate the effects of the different estimates of R values, all possible combinations were calculated in a Monte Carlo simulation using Crystal Ball| software (reference to Table 44.3 shows that values of none of the variables are correlated). The results of these calculations are presented in Table 44.5, where mode values for each coefficient are shown [with the mean and pertinent range of values parenthesised].
Clearly, equations [44.50] and [44.51] have unacceptable error magnification because of the large coefficients so that they cannot be used realistically. It is also significant that the Table 44.5 - Sensitivity of correction equations to absolute isotopic a b u n d a n c e s distribution of 634 =650 634 =650 634 =649 ~)34 - 6 4 9 634 =666 634 =650
x x x x x x
2.09 1.84 4.99 4.20 6.99 6.69
[2.09, [1.84, [4.97, [4.20, [8.32, [7.98,
+0.00-0.00] +0.15-0.06] +0.05-0.39] +0.02 -0.01] +5.96-0.80] +5.70-0.76]
- 666 x - 665 x - 666 x - ~)65 x -665 x - 649 x
1.09 1.60 1.40 2.19 12.4 11.3
[1.09, [1.68, [1.49, [2.19, [14.6, [13.9,
+0.00 -0.00 +0.35 -0.25 +0.40 -0.30 +0.00 +0.00] +12.0 -9.85] +11.4 -9.88]
[44.46] [44.47] [44.48] [44.49] [44.50] [44.51]
coefficient values for these two equations are highly skewed as shown by the differ-
965
Data Corrections for Mass-Spectrometer Analysis of SO 2
ences between mode and mean (the latter is the first term in parentheses), which indicates that the coefficients are strongly dependent on the correct choice of R values. The best equations with regard to uncertainty in the R values are [44.46] and [44.49], since those solutions to the simultaneous equations are not sensitive to them. However, there are potential, practical problems associated with use of the parameters 549 and 565 in equation [44.49], which described in section 4:4.7, below. Therefore equation [44.46] will be shown to be the optimum choice.
44.7 Minimising uncertainties in practice 44.7.1 Internal precision of measurements Both 550 and 666 usually can be measured to high precision: 0.04%0 (2 x standard error of the mean) or less which is similar to the overall reproducibility of the method including sample preparation. However, in practice 565 cannot always be determined as precisely; and similarly for 54:9. When a sample gas runs well on the mass spectrometer the precisions are similar for all 5 values measured. However, it is more usual to find that the precision of 565 and 549 is a factor of two or three times as great as those for 566 and 550, respectively. In the laboratories at BP Research and at University of Reading, using a VG SIRA 9 mass spectrometer, typical average measurement precisions recorded are given in Table 4:4.6. Although the reasons for the variations in precision are not definitive, it is probable that fragments of other molecular species are collected at m / z 49 and 65 and may relate to imperfectly purified SO2. In a comparable, but not perfectly analogous system, mass spectrometer analysis of CO2 showed that trace contaminants could be collected with the minor isotope species giving spuriously larger measured 5 values (Weber et al., 1976). Analysis of many measurements indicates that 565 is more susceptible to this effect than 566, but can be used to advantage to act as a measure of analytical data quality. If there is no 618 difference between sample and reference, then equations [44.18] and [44.26] can be combined (with both 618 terms set to zero) to give 6 6 6 - 2.01 x 565
[44.52]
Therefore, if the 666 is less than twice as great as the 665, it seems probable that the sample is contaminated and that even the m/z:66 may have included a small amount of contaminant ions and Table 44.6 - Average internal give a more positive 566 value than it should. The s a m e precision of mass spectrometer reasoning applies to equations [44.38] and [44.39], which measurements if 518 - 0, yield 6 5 0 - 2.18 x 549
R
of the mean, %0
[44.53] 49
It is interesting to note that most analyses where the ratio of the two 6 values to each other is much greater than 2 are those where the internal precision of the 549 or 665 is
2 x standard error
5o 65 66
0.048 0.018 0.042 0.015
966
Chapter 44- M. Coleman
higher than average. This supports the concept of a contaminant effect. From the above it is obviously better to avoid measurement of 865 and 849 except for quality control purposes. Assuming that precise and valid measurements can be made of 850 and 866, equations [44.18], [44.38] and [44.46] show that if 818 is zero then the total error in a measurement will be reduced and either of equations [44.18] or [44.38] can be used to correct 866 or 850, respectively.
44.7.2 Minimising differences in 818 between sample and reference It is as important to reduce or eliminate the oxygen isotope difference between Samples and Reference as to determine the isotopic abundances of the reference gas precisely. Examples of the practical application of this approach are taken from the experience of routine analyses in the former BP Research and University of Reading Isotope Laboratories. Samples of sulphides are converted to sulphur dioxide using a stoichiometric excess of cuprous oxide as the oxidant (Robinson & Kusakabe, 1976). Reference gas is prepared by oxidising the laboratory standard (in this case a chalcopyrite), in the same way as samples: in this way oxygen isotope differences between Sample and Reference are reduced to a minimum. Although the exact oxygen isotope fractionation between cuprous oxide and sulphur dioxide cannot be established, precautions are taken to ensure that it is as small as possible. Equilibrium fractionation is at a minim u m at high temperatures and consequently the sample is introduced rapidly into the furnace (1070~ so that lower-temperature, and thus large, fractionations do not occur. This gives very reproducible results when making replicate measurements. However, to reduce the possibility of systematic error when measuring samples of different isotopic compositions they should all be prepared using the same batch of cuprous oxide. Ideally, the reference gas also should be prepared using the same oxidant but in practice a reference may have to serve more than two hundred analyses. This problem too can be overcome fairly readily. The relative differences between different samples in a batch will be maintained almost perfectly, even if the small 818 with respect to the reference gas is not known to high precision. However, all the samples will be subject to a systematic error after correction. This error can be eliminated by referring the batch of samples to the standards run with them. The procedure adopted is such that a number of samples, sufficient for one batch of analyses on the mass-spectrometer, is prepared together with at least two aliquots of the laboratory standard, but preferably more. This approach offers other benefits too. The overall reproducibility of the analyses can be ascertained since there are repeated determinations of the same material. Longer-term variations of isotopic composition of the reference gas can be monitored, but are automatically corrected. Equally, faults in the sample preparation system or mass-spectrometer are indicated immediately as measured isotopic differences between the laboratory standard run as a sample against the same material as reference gas.
Data Corrections for Mass-Spectrometer Analysis of SO 2
967
The remaining factor in the correction procedure, the absolute abundances of oxygen isotope in gas samples can be determined fairly readily too, even though the exact value is not critical. A sample of graphite is substituted for sulphide in the oxidation procedure and the resultant carbon dioxide is analysed isotopically against an oxygen isotope standard; e.g. NBS19 which is notionally calibrated against VPDB (Gonfiantini et al., 1995).
44.7.3 Isotopic measurements of sulphates
Although the problem of 618 differences can be managed for sulphide preparation, measurement and data correction, analysis of sulphates may pose a problem. A number of different methods are used to prepare sulphur dioxide from sulphates. In one the sulphate is reduced to sulphide before oxidation (Thode et al., 1961) which poses no obvious problems in terms of data correction. However, the direct reduction of sulphates (Holt & Engelkemeir, 1970; Coleman & Moore, 1978; Halas & Szaran~ 1999), even at high temperature could introduce oxygen with a different isotopic composition. However, during the extraction procedure described by Coleman & Moore (1978) the sulphur dioxide produced is kept for a few minutes in contact with a mixture of hot copper and cuprous oxide. This has two useful effects: any sulphur trioxide produced in the process is reduced to sulphur dioxide, and the oxygen in the gas equilibrates with that in the cuprous oxide. In this way all sulphur dioxide samples produced from a batch of sulphates should have similar oxygen isotope compositions. Similarly, the method of Halas & Szaran (1999), which involves reaction of BaSO4 with NaPO3 in the presence of copper, the oxygen in SO2 comes solely from the NaPO3. This ensures that all samples have identical oxygen isotope compositions. If the reference gas is also prepared from a sulphate standard then the 618 term in the correction equation will be zero. For ultimate precision the 634 of the sulphate standard can be determined with respect to sulphides (Canyon Diablo troilite or laboratory standard chalcopyrite) using the preparation method of Thode et al. (1961). Alternatively a portion of sulphide standard can be oxidized quantitatively to sulphate ion and then analyzed as BaSO4. Subsequently this standard sulphate can be included with batches of sulphates and their compositions can be referred to it, thus reducing the effect of possible 618 differences from the reference gas. This approach has been used for calibration of the standards SO-5, SO-6 and NBS-127 by Halas & Szaran (1999).
44.7.4 Choice of correction procedure in practice Analysis of sulphides can be performed in such a way that 618 differences can be minimised. Thus, the mass spectrometer uncertainty is limited to one measurement, that of 650 or 666: correction of data will use equation [44.38] or [44.18], respectively. However, where direct reduction of sulphates is used to prepare SO2, there is the possibility of variation in 618. In that case, the choice of correction procedure depends on the specific preparation method used and to what extent 618 values are uniform. It is impossible to make specific recommendations except for the need to experiment and determine the best approach.
968
Chapter 44- M. Coleman
The experience in the BP Research and University of Reading Laboratories indicates that overall uncertainty in the corrected 834 is lower when making a single 850 or 866 measurement rather than combining the errors from both measurements in equation [44.46], despite the possibility of variability of 818. 44.8 Correction procedure The calculations are performed on a Personal Computer or equivalent. They are accomplished most readily using Microsoft Excel| or another spreadsheet programme. The correction procedure is outlined below. The isotopic composition of the reference gas is not assumed and all standardisation is with respect to the standard material included with each batch of samples. a) After the completion of the analysis of a batch of samples (and standards run as samples) the measurements of the standard material with respect to the reference gas are averaged. The average value of the standard with respect to the reference gas is inverted to determine the 866 of the reference gas with respect to the standard.
[44.54]
866ref-std - (1 / (1 + 866std-ref/ 1000 - 1) x 1000
b) A correction equation is calculated using the isotopic abundances of the standard. These are derived from those for CDT and the 834 of the standard with respect to CDT. Appropriate oxygen values from oxidation of graphite, are included. This is applied to give a corrected value for the reference gas with respect to CDT. c) A new correction equation is calculated using the R values of the reference gas. d) The measured 866 values of the set of samples are corrected to produce 834 compositions with respect to the reference gas. Figure 44.1- Example of data correction programme output (for details see Table 44.7) PRISOTOPE LABS
Mode
No 818 Sample-Reference difference
S iso corrections
Int Std. 634S 6180
CP2 / 2 -4.56 -2.00
CORRECTIONS TO DATA 128/2/1990 ] [ Batch No.
Sample No. CP2/2 QHS-14 QHS-10 QHS-12 QHS-9 QHS-16 QHS-13 QHS-14D CP2/2
Description Chalcopyrite Std Ag2S Ppt Ag2S Ppt Ag2S Ppt Ag2S Ppt Ag2S Ppt Ag2S Ppt Ag2S Ppt Chalcopyrite Std
Extr. No. $6736" $6742 $6740 $6739 $6741 $6737 $6743 $6738 $6744"
CHALCOPYRITE
Material wrt CDT wrt SMOW [14
865
866
+0.067 +6.512 -0.775 -4.071 -1.328 -4.009 -5.007 +6.233 +0.101
+0.119 +12.270 -1.496 -8.218 -2.563 -8.299 -10.328 +12.129 +0.167
]
Name
634SCDT -4.586 +8.588 -6.337 -13.625 -7.494 -13.713 -15.913 +8.435 -4.534
[MaxColeman A65 +0.01 +0.38 -0.03 +0.04 -0.05 +0.14 +0.16 +0.17 +0.02
Comment
Contam.?
Duplicate
969
Data Corrections for Mass-Spectrometer Analysis of SO 2
e) The corrected values are adjusted by the ~)34 CDT value of the reference gas to relate all samples to CDT. f) The analyst can then relax with a clear conscience, satisfied that another set of accurate and precise data has been produced. A complete example of correction of a data set with input data and calculated results is shown as Figure 44.1. A65 is a possible measure of the extent of contamination as indicated by the enhancement of the 665 value (see equation [4:4.52]).
Table 44.7 - Details of correction p r o c e d u r e (Results p r e s e n t e d in Table 44.7) Action 1. Input values for standards run as samples Input Lab S t a n d a r d 634CDT I n p u t results from 8180 m e a s u r e m e n t of g r a p h i t e CO2 Calculate a v e r a g e m e a s u r e d 866 of stds (identified by * after s a m p no.) 2. Calc reference gas rel to lab. standards a) Calc atomic ratios for std Std R17 - 0.00037983 x (1+R18/2000) Std R18 = 0.0020052 x (1+ R18/1000) Std R33 = 0.007893 x (1+ R34 / 2000) Std R34 = 0.045035 x (1+ R34/1000) b) Calc correction coefficients (for lab. standard gas) C o e f f 6 6 s t d _ ( S t d R 3 4 + 2 x S t d R 1 8 + 4 x StdR172 + 2 9S t d R 3 3 x S t d R 1 7 ) (StdR34 + StdR33 x StdR17) c) Invert values of meas. stds (gives Ref. rel to Std.) Ref 666Std - (1 / (1 + Std 666Ref / 1000)-1), 1000 d) Correct Reference gas values relative to lab. Std. Ref 634Std = 666Ref-Std x Coeff 66Std e) Correct Reference gas values relative to CDT Ref 634CDT = ((1+634Std /1000),(1 + std 634CDT/1000)-1),1000 3. Calculate correction coefficients (for reference gas) a) Calculate atomic ratios for Reference gas Ref R17 = Std R17 Ref R18 = Std R18 Ref R33 - 0.007893-(1+ Ref 634CDT/2000) Ref R34 - 0.045035-(1+ Ref 634CDT/1000) b) Calculate correction coefficients (for reference gas) C o e f f 6 6 s t d = ( S t d R 3 4 + 2 x S t d R 1 8 + 4 x StdR172 + 2 9S t d R 3 3 x S t d R 1 7 ) (StdR34 + StdR33 x StdR17)
4. Correct all measured values a) rel to ref gas (using m e a s u r e d s t a n d a r d values as e x a m p l e & calculation check) Std 666Ref = M e a s u r e d Std66Ref Std 634Ref = Std 666Ref x coeff66Ref gas b) rel to C D T S a m p 634 = ((1 + Std 634Ref/1000)'(1 + Ref 634CDT/1000)-1),1000 c) Correct all measured values - see output in Figure 44.1 d) Calculate 666/665 values - see output in Figure 44.1
Value -4.56 -2.00 +0.143
0.00037945 0.0020012 0.007875 0.044830 1.08935
-0.143 -0.156 -4.715
0.00037945 0.0020012 0.007874 0.044823
1.08937
0.143 0.156 -4.560000001
970 A65 - 665 - 666 / 2
Chapter 44- M. Coleman [44.55]
The details of the successive stages in the calculation process are detailed in Table 44.7. After the correction coefficient has been calculated it is checked by application to the average of the measured values of the laboratory standard run as samples. This example shows a simple batch of samples of silver sulphide and two laboratory standards from which SO2 was prepared using the same batch of cuprous oxide. For other sample types and for mixed batches other correction procedures can be used. However, it must be stressed that this is but one example of the application of the generic set of correction equations presented in this chapter. It is essential for every laboratory to ensure that its own operating procedures produce both precise and accurate results.
Acknowledgements
Jill Banham is thanked for help with painstaking transcription of scrawl to text. Stan Halas and Kinga Revesz are gratefully acknowledged for careful and helpful reviews. This paper is PRIS Contribution no. 818.
Handbook of Stable IsotopeAnalyticalTechniques,Volume 1 P.A. de Groot (Editor) 9 2004 ElsevierB.V. All rights reserved.
CHAPTER 45 Oxygen Isotope Corrections in Continuous-Flow Measurements of SO2 Kristen Leckronel & Margaret Ricci2 1 Department of Chemistry, Roosevelt University, 430 S. MichiganAve., Chicago, IL 60605, U.S.A. 2 Department of BiologicalSciences, University of Idaho, Moscow, ID 83844,U.S.A. e-mail: 1 [email protected];[email protected]
45.1 Introduction and objectives Value of continuous-flow techniques. The emergence of continuous-flow techniques has significantly altered the practice of isotope ratio mass spectrometry (IRMS). In the continuous-flow approach, samples and standards are delivered to the ion source of an IRMS as pulses of analyte gas in a constant, low flow of helium. By increasing the fraction of sample admitted to the ion source, these methods decrease sample size requirements by 2-4 orders of magnitude relative to dual-inlet methods, thereby enabling isotopic analysis to be performed when sample size would otherwise be limiting (Brand, 1996). Continuous-flow IRMS instruments are faster, more easily automated, and less expensive than their dual-inlet counterparts. Perhaps most significantly, the continuous-flow approach is readily compatible with on-line techniques of sample preparation. This has enabled the development of a wide variety of valuable hybrid techniques including compound-specific isotopic analysis by gas chromatography - IRMS (Hayes et al., 1990), automated isotopic analysis of C, N and S in bulk samples by elemental analyzer - IRMS (Fry et al., 1992., 1996; Owens & Rees, 1989), spatially resolved isotope analysis by laser microprobe IRMS (Cerling & Sharp, 1996; Sharp & Cerling, 1996) and countless specialty inlets tailored to meet the demands of specific applications.
45.1.1 Continuous-flow sulfur isotope techniques While the earliest continuous-flow applications focused on carbon and nitrogen isotopes, methods have subsequently been developed for isotopic analysis of H, O and S. Although dual-inlet analysis of sulfur isotopes is conventionally performed using either SF6 or SO2 as the analyte, continuous-flow measurements exclusively use SO2 because it is easily generated by auxiliary sample preparation devices. The most widely-utilized continuous flow method for sulfur isotopic analysis is elemental analysis-isotope ratio mass spectrometery (EA-IRMS) (Giesemann et al., 1994), in which the SO2 generated by combustion of the sample in an elemental analyzer is channeled to an isotope ratio mass spectrometer. EA-IRMS has been successfully used to measure the 34S/32S ratio of small samples of sulfide (Volkert et al., 2000) and sulfate minerals (Alewell et al., 1999) as well as organosulfur compounds from biomass and
972
Chapter 45 - K. Leckrone & M. Ricci
sediments (MacAvoy et al., 2001; Passier et al., 1999). Another recent application of continuous-flow sulfur isotope techniques is the spatially-resolved, in-situ isotopic analysis of sulfide minerals by laser microprobe IRMS. In laser microprobe instruments, samples are placed in a chamber under an oxygen or fluorine atmosphere and spot-heated with a laser to generate SO2 or SF6 for IRMS analysis. The first generation of laser microprobe instruments utilized an infrared laser, a fluorine or oxygen atmosphere, and a dual-inlet IRMS to achieve a spatial resolution of ca 150-500 micron and isotopic precision of 0.3-0.5 permil (Crowe et al., 1990; Kelley & Fallick, 1990; Rumble et al., 1993). In the newest generation of instrument, spatial resolutions of 50 micron and isotopic precision of 0.3 permil have been achieved with an ultraviolet laser, an oxygen atmosphere, and continuous-flow IRMS analysis of the SO2 gas (Leckrone & Crowe, 1999).
45.1.2 Oxygen isotope corrections in continuous-flow sulfur isotope techniques Standard equations for the isotopic correction of SO2 (see e.g. Coleman, 1980; Rees, 1978) were developed at a time when preparation of samples by vacuum-line combustion, purification by cryogenic distillation, and isotopic analysis on a dualinlet instrument was the norm. Manual preparation of samples on a vacuum line, while slow, allows for very close control of variables such as reaction temperature, reaction rate, oxygen partial pressure, and oxygen source, all of which can affect both the sulfur and the oxygen isotopic composition of the product (Halas & Wolacewicz, 1981; Robinson & Kusakabe, 1975). However, as discussed in more detail below, the standard equation for isotopic correction of SO2 does not allow for independent calculation of both the oxygen and sulfur isotopic composition of a given SO2 sample. (This situation is in marked contrast to the isotopic correction of CO2 gas, in which the correction accurately calculates both 613C and 6180 even when carbon and oxygen isotopic composition of the CO2 sample vary freely relative to those of the standard). In light of the more variable combustion conditions inherent in continuous-flow sulfur isotope techniques, it is important to evaluate the assumptions underlying the standard isotope correction equations for SO2 and to ascertain under what conditions they are applicable in continuous-flow systems. Continuous-flow measurement of sulfur isotopes are measured based on the mass spectrum of the SO2+ molecular ion. Since both sulfur and oxygen are multi-isotopic, all measureable ion currents above the molecular ion (m/z 64) represent sums of multiple isotopic contributions (e.g. 3 4 S 1 6 0 1 6 0 + 32S180160 + 33S170160 + 32S170170 at m/z 66). Observed ratios of molecular species can be interpreted in terms of elemental sulfur isotope ratios only after correction for the oxygen isotopic contribution. A variety of solutions to the oxygen isotope correction of SO2 have been proposed (Coleman, 1980; Crowe & Vaughan, 1996; Holt & Engelkemeier, 1970; Rees, 1978). The most widely used approach takes the form i~34S = 1 . 0 8 9 ~)66 _ 0 . 0 8 9 1 ~)180
[45.1]
(see e.g. Halas & Wolacewicz, 1981 and Coleman, Part 2, Chapter 44), where the exact value of the coefficients depends on the absolute natural abundance assumed for each
Oxygen Isotope Corrections in Continuous-Flow Measurements of SO 2
973
isotope, as well as on the value of the standard relative to the CDT standard. The relative difference in sample (sa) and standard (st) sulfur isotopic ratio is expressed as a delta value in permil units
~34S _ (
34 s; t)
Rst Rs 1000
[45.2]
Similarl3r ~66 is the relative difference in measured 66/64 ion current ratios of sample and standard, and 6180 is the relative difference in oxygen isotope ratio of sample and standard, where 666 and ~180 are also expressed in permil units. The most serious shortcoming of the standard equation is that the oxygen isotopic composition of samples cannot actually be measured, so the latter term in equation [45.1] is in fact unknown. The equation is nevertheless useful because in vacuum-line combustions, samples and standards are prepared under such carefully controlled conditions that their oxygen isotopic composition are assumed to be identical and the second term of equation [45.1] reduces to zero. For samples such as sulfates, where oxygen isotopic variations between sample and standard raw materials cannot be avoided, the oxygen isotopic composition of the SO2 byproduct can be controlled by equilibration with a large excess of hot CuO (e.g. Halas & Wolacewicz, 1981). However, any random variations that do occur in a sample's oxygen isotopic composition will limit the precision of 634S calculated using equation [45.1],while systematic offsets in oxygen isotopic composition will affect the accuracy of the ~34S calculation. Several workers including Grassineau et al. (2001) and Fry et al. (2002) have emphasized the need to monitor and correct for changes in the oxygen isotopic composition of SO2 samples and standards in order to obtain acceptable sulfur isotope measurements when using continuous-flow methods. Control of oxygen isotopic variations in continuous-flow systems is more difficult than in conventional, vacuumline combustions because of the dynamic nature of the continuous-flow combustion, in which both the temperature and partial pressure of oxygen fluctuate rapidly during combustion. Additionally, multiple oxygen sources including metal oxide catalysts, pulses of oxygen gas, and oxygen within the sample matrix may contribute to oxygen isotopic variability from sample to sample and over the longer term. For analysis of sulfates and sulfides including sulfides in whole rock matrices, routine precision of 0.1 to 0.2 permil can be achieved through careful control of operating parameters including sample weight, V205 or other supplemental oxygen donors, the oxygen saturation of oxidation and reduction reactors, and frequency of internal standardization (Grassineau et al., 2001; Studley et al., 2002). The accuracy of 634S determinations for lowsulfur organic samples analyzed by elemental analysis- isotope ratio mass spectrometry (EA-IRMS) can be limited to >1-3 permil due to uncontrolled variability in the oxygen isotopic composition of the SO2 arising from exchangeable oxygen (e.g., organic oxygen and water) in the sample matrix. Such effects can be overcome through empirical correction factors or by isotopic equilibration of SO2 produced by the elemental analyzer in a 890~ SiO2 buffering furnace (Fry et al., 2002). While the effects of oxy-
974
Chapter 45 - K. Leckrone & M. Ricci
gen isotopic variability in continuous-flow sulfur isotope measurements can be overcome with sufficient experimental controls, an improved isotope correction algorithm capable of correcting for oxygen isotope variability is nevertheless desirable to reduce these experimental constraints and improve the robustness of continuous-flow sulfur isotope methods. 45.1.3 Objectives o f study Given the inherent limitations of the current isotope correction equation, the goal of this project has been to mathematically and experimentally evaluate two alternative methods for oxygen isotope correction which have the potential to correct for sample-to-sample variability in the oxygen isotopic composition of SO2. The first relies on simultaneous measurement of three adjacent ion currents (m/z 64, 65 and 66). This approach is similar to one suggested by Coleman (Coleman, 1980). Here, however, we use numerical techniques to solve exact expressions for S and O ratios assuming a triple collector instrument capable of simultaneous measurement of three distinct ion beams. This approach also assumes mass-dependent fractionations for both S and O isotopes. Accordingly, we first present a review of the sulfur and oxygen isotope literature to determine the most accurate values for these relationships. We then derive an exact expression for S and O isotope ratios in terms of measured mass spectrometric ion current ratios 66/64 and 65/64. Finally, we test the algorithm using a combination of computer-generated datasets and laboratory analyses of natural sulfide mineral samples and reference materials. Additionally, we have evaluated a previously proposed method for calculating 834S based on simultaneous collection of ion currents at m/z 48, 50, 64 and 66 (Holt & Engelkemeier, 1970; Coleman 1980). This approach does not assume mass-dependent fractionation of S or O isotopes. We derive an expression for the sulfur isotope ratio in terms of the measured mass spectrometric ion current ratios 66/64 and 50/48, and test the procedure using a variety of sulfide minerals and reference materials. 45.2 Corrections based on s i m u l t a n e o u s m e a s u r e m e n t of three adjacent ion currents 45.2.1 N o t a t i o n
Throughout this paper, we will use the letter R to indicate the ratio of a particular atomic or molecular isotopic species to the related isotopic species of lowest mass nR _ concentration of isotopic species of mass n concentration of base species related to n
[45.3]
For example, 34R refers to the isotopic abundance ratio 34S/32S of sulfur, while 66R refers to the measured ion current ratio m/z 66/m/z 64 of the sulfur dioxide molecule. The letter F will be used in a similar manner to represent fractional abundances nF = concentration of isotopic species of mass n concentration of all species related to n
[45.4]
For example, 13F refers to 13C/(12C + 13C). For molecular species, the denominator
Oxygen Isotope Corrections in Continuous-Flow Measurements of SO 2
975
refers to the sum of all species of SO2. The symbol cz refers to the fractionation factor, that is, the ratio of isotope ratios produced by a chemical or physical process
n
nR c~ -
products
[45.5]
nRreactants Classical calculations of 6 values for stable isotopes have often proceeded from the position that the mass spectrometer is measuring 6 values. On one level this treatment is correct, but on a deeper level, the mass spectrometer is actually measuring ion currents, which are then turned into ion current ratios, from which delta values are derived. Thus, in our calculations, ion current ratios are related directly to isotopic or molecular abundance ratios, and calculation of delta values is the last step rather than the first.
45.2.2 Mass spectrum of sulfur dioxide For SO2 samples with isotopic abundances near natural levels, ion currents large enough to allow rapid and precise measurements occur only at masses 64, 65 and 66. The fractional abundance of these SO2 molecular species are related to fractional abundance of individual isotopes according to simple probability functions 64F = 32F.16F.16F 65F = 33F.16F.16F + 2.32F.17F.16F 66F = 34F.16F.16 F + 2.32F.18F.16F + 2.33F.17F.16F + 32F.17F.17F
[45.6] [45.7] [45.8]
where factors of two accounts for the fact that there are two sites at which any given oxygen isotopic substitution can occur, hence the probability of oxygen isotopic substitution is twice its atomic abundance. It is convenient to divide equations [45.7] and [45.8] by equation [45.6] in order to express molecular abundance ratios (ion current ratios) in terms of atomic isotope ratios" 65R = 33R + 2.17R 66R = 34R + 2.18R + 2.33R17R + 17R2
[45.9] [45.10]
The molecular abundance ratios 65R and 66R are obtained from a triple collector instrument capable of simultaneous measurement of three distinct ion beams.
45.2.3 Mass-dependent relationships of oxygen and sulfur The fundamental difficulty underlying most methods for oxygen isotope correction of SO2 is that measurement of two (65R and 66R) ion-current ratios does not provide sufficient information to uniquely solve for the four contributing elemental ratios (33R, 34R, 17R and 18R). Functional relationships between 17R and 18R, and between 33R and 34R, are therefore required to provide the necessary additional equations. These relationships are based on the fact that nearly all physical and chemical processes fractionate multiple isotopes of the same element proportionally to differences in mass (Bigeleisen & Mayer, 1958). Consequently, the sulfur or oxygen isotope ratios
976
Chapter 45 - K. Leckrone & M. Ricci
of two pools (denoted by subscripts I and 2) linked by a single mass-dependent process are related quantitatively by 17R 1 / 17R 2 = (18R1/18R2)a
[45.11]
or 33R1 / 33R2 = (34R 1 / 34R2)a'
[45.12]
where a = In (17(I)/ln(18~) and a' = ln(33~)/ln(34~). However, fractionation factors are not fixed so that a or a' remain constant for all processes. Instead, values of a and a' depend on such variables as the molecular masses and reaction mechanisms of the species involved. In general, for both S and O, a and a' for specific processes take on values in the range of 0.50 to 0.53 (Hulston & Thode, 1965b; Matsuhisa et al., 1978). Furthurmore, equation [45.12] describes only the immediate isotopic relationship between reactant and product involved in a single fractionating process, rather than the overall composition of mixed isotopic pools. Thus, the relationships underlying the standard SO2 isotope correction equation
(18Rsal0"50 17Rst 18Rst)
17Rsa _
[45.13]
and
33asa _ (34Rsal0.50 33Rst 34Rst)
[45.14]
only approximately describe the isotopic relationships produced by the mixing of sulfur or oxygen pools of similar isotopic composition. The choices of a - a' - 0.5 represent minimal values of a and a' rather than the best possible approximations of the quantitative relationship between related isotope ratios in the terrestrial sulfur and oxygen pools. An optimal value of a can be obtained from paired measurements of ~170 and 6180 for a variety of terrestrial materials. The slope of the resulting 6170 vs. ~1180 crossplot is equivalent to a because the 6 values in such plots are identical to the first term in a series expansion of the log terms in
18Rsa In 17 "- a In 1 ~ Rst Rst 17R
sa
[45 15]
which is in turn derived from equation [45.11] (Santrock et al., 1985). An optimum
977
Oxygen Isotope Corrections in Continuous-Flow Measurements of SO 2
value for a' can be obtained from paired measurements of ~33S and ~34S by analogous logic. A more recent re-assessment (Robert et al., 1992 and references therein) of oxygen isotopes for a wide variety of terrestrial and lunar materials indicates ~170
-
(0.521 + 0.001) 915180
( n - 102)
[45.16]
where the stated uncertaintly is the l o standard deviation of the slope. The resulting value of a is intermediate between the recently reported value of 0.528 (Meijer & Lee, 1998) and the value of 0.516 suggested by Santrock et al. (1985). Paired measurements of ~133S a n d ~34S have been reported for a wide variety of sulfur-bearing terrestrial, lunar and meteorite samples prepared as SF6 (Beaudoin et al., 1994; Gao & Thiemens, 1991, 1993a, b; Heymann et al., 1998; Hulston & Thode, 1965a; Rees & Thode, 1977; Rumble et al., 1993; Thode & Rees, 1971). Our review of the SF6 literature indicates a relationship of 633S -
(0.515 + 0.001)-
634S
( n - 127)
[45.17]
This analysis excludes a small subset of iron meteorites which are known to have been selectively enriched in 33S by nucleogenic processes. In general it is important to note that the above relationships do not apply to materials affected by nucleosynthetic processes, admixture of separated isotopes, or mass-independent isotope effects. The isotope relationships expressed in equations [45.16] and [45.17] convey only the relative differences between the isotopic composition of unknown samples and an isotopic standard. In order to ascertain the absolute values of isotope ratios for use in equations [45.9] and [45.10], relative isotopic relationships need to be recast in the form (Santrock et al., 1985) 17R -
18Ra" K
[45.18]
34Ra'" K'
[45.19]
and 33R -
where K and K' are con- Table 45.1 - Absolute stants characteristic of dards the solar system's oxy- standard 17R gen and sulfur pools. Evaluation of K requires PDB 0.0003790 an absolute isotopic SMOW 0.0003730 abundance measure- Tank 02 0.0003775 ment of both 17R and 18R for at least one isoto- average St. dev. pic standard. As shown rsd in Table 45.1, there have
abundance measurements of oxygen isotope stan-
18R
reference
K
0.0020671 0.0020052 0.0020514
Mook & Grootes (1973) 0.0094915 Hoers (1997) 0.0094904 Santrock et al (1985) 0.0094916 0.0094912 0.0000007 0.007%
978
Chapter 45 - K. Leckrone & M. Ricci
been at least three independent analyses of the absolute 17R and 18R values in reference materials (Hoefs, 1980; Mook & Grootes, 1973; Santrock et al., 1985 and references therein), which we have used together with the a value of 0.521 to determine an average value of K - 0.0094912. Agreement among the three independent measurements is sufficient to calculate K with a relative standard deviation of 0.007%. Similarly, evaluation of K' requires an absolute isotopic abundance measurement of both 33R and 34R values for a single sulfur isotope standard. To date there have been absolute isotope abundance measurements of five separate sulfur standards (Ding & Zhang, 1989), resulting in an average value of K' - 0.03924 (0.3% relative standard deviation).
45.2.4 Calculation of sulfur and oxygen isotope ratios The exact expressions relating measured 65R and 66R ion current ratios to the isotopic abundances of their constituent elements (equations [45.9] and [45.10]) and expressions for mass-dependent relationships in the terrestrial oxygen and sulfur pools (equations [45.18] and [45.19]) can now be combined to yield a single master equation for 34R in terms of only measured quantities and the constants a, K, a' and K'. This method is analagous to the successful approach of Santrock et al. (1985) for carbon isotopic analysis of CO2, and apart from uncertainties in the values of the constants, the equation is exact. 66R - 3 4 R - 2( K'
65R/~
9
"
34Ra')1/a - K 965R- 34Ra 65R2 34R2a " K'2 ' '+ = 0 4
[45.20]
It is difficult to rearrange equation [45.20] to yield an explicit solution for 34R. Practically, the equation is most easily solved using numerical techniques. The overall procedure for SO2 isotope correction using this approach is as follows. Step 1" Raw ion current ratios are corrected for background currents, amplifier offsets,
mass discrimination effects, etc. by the usual method of differential measurement, in which m65.-.
65Rs a =
lKsa . t65Rs t _ 65 Q. t65 Rst, . Rst
m65
66
Rsa -
66
Q .
t66.-,
lKst
[45.21]
where m indicates a measured ion current ratio and t indicates the true isotope ratio of the standard gas as prepared. Working standards are determined by extension of this process to a primary standard. Step 2" The corrected ion current ratios are inserted into equation [54.20], which is
solved iteratively for 34R using Newton's method. The initial or "seed" value of 34R may be fixed (for example, set to equal 34Rst) or floating (for example, estimated as 34Rseed = 34Rst.66Rsa/66Rst). The former approach is adequate provided sample and standard differ by less than around 25%o, but the latter
Oxygen Isotope Correctionsin Continuous-FlowMeasurementsof SO 2
979
approach leads to more rapid convergence of equation [45.20] and avoids errors in cases of more widely differing samples and standards.
Step 3" Once 34R of the sample is known, 33R, 18R and 17R are calculated from equations [45.19], [45.9] and [45.18], respectively. Step 4" In the last step, 6 values are calculated from atomic isotope ratios according to ~)34S = 1000(34Rsa/34Rst - 1) and
6180 = 1000(18Rsa/18Rst-1)
[45.22]
As an alternative procedure, the four contributing equations can be combined to yield an exact expression for 18R in terms of 66R and 65R 66R 18R_ 1(65R-2KK'18Ra)1/a' 18Ra 3. K2 18R2a 2 2 - K-65R. + ~_ 9 - 0
[45.23]
in which case 17R, 34R and 33R are subsequently calculated by re-insertion in the primary contributing equations. However, we have not found any significant difference in the final isotopic values if 18R rather than 34S is calculted in Step 2. As with the standard SO2 isotope correction (equation [45.1]), corrections based on equation [45.20] require both ~)180 and ~)34Sof the reference gas to be known relative to primary isotopic standards. Unlike the standard correction, this approach does not require ~180 of the sample to be known or matched relative to the standard, and it accounts for all higher-order terms.
45.2.5 Results for model datasets The algorithm for oxygen and sulfur isotope calculations using the above four-step process was initially debugged and tested using two model datasets. In the first dataset, a series of 65R and 66R values were generated for hypothetical samples having 634SCDT ranging from -100 to 100 %o and ~)18OvsMow of 0 %o. In the second dataset, 65R and 66R values were generated for samples having input values of 6180 from-100 to 100%o and ~)34S of 0 %o. In each case, standards were assigned 65R and 66R equivalent to ~)34SCDT - 0%o and ~)18OvsMow of 0 %o. Thus, in the first dataset, samples varied relative to standards in ~)34S but not in 6180, and in the second set, samples and standards shared the same 634S differed in ~180 values. Sulfur isotopic ratios as 634S were calculated based on these spreadsheet-generated 65R and 66R values using three alternative isotope correction procedures. First, (~34S was calculated using the standard SO2 isotope correction equation (equation [45.1]). Second, ~)34S was calculated using an empirical two-point isotope correction of the form ~)34Ssa = m. ~)66Rsa+ b
[45.24]
in which m and b are determined by regression of measured ~)66R against known 634S for a pair of isotopic standards chosen to bracket the expected 634S of the sample (Crowe & Vaughan, 1996). This correction is particularly useful in laser-microprobe
980
Chapter 45 - K. Leckrone & M. Ricci
Figure 45.1 - Comparison of isotope corrections when oxygen isotopes of sample and standard are matched. IRMS because, although it assumes the oxygen isotopic composition of samples and standards are equivalent, knowledge of the actual value of 6180 is not required. Finally, both 634S and 6180 were calculated using the SO2cal algorithm based on equation [45.20]. Results are expressed as A values, where A34S is the error in the ~34S value calculated by the standard equation (equation [45.1]), empirical correction (equation [45.24]), or SO2cal algorithm, and A180 is the error in 6180 calculated with the SO2cal algorithm. Results for the dataset with constant 6180 but variable ~34S are summarized in Figure 45.1, and results for the dataset with constant 634S but variable 6180 are summarized in Figure 45.2. These modeling results illustrate several points. First, when the oxygen isotopic composition of sample and standard are matched, the standard correction displayed a slight systematic error of 0.0087%0 in the calculated 634S for every 1%o difference between the 634S of the sample and that of an arbitrarily chosen standard. These errors presumably arise due to the exclusion of higher-order
Oxygen Isotope Corrections in Continuous-FlowMeasurements of SO 2
981
Figure 45.2 - Comparison of isotope corrections when oxygen isotopes of sample and standard are mismatched. terms in the derivation of equation [45.1], and are negligible (<0.1%o) provided the sample and standard 634S are within about 10%o. Values of c534Scalculated using the empirical correction, and 634S and 6180 calculated using the SO2cal algorithm, displayed no systematic errors. When the oxygen isotopic composition of samples was allowed to vary relative to that of the standard, however, serious errors resulted if 634S was calculated using standard mathematical or empirical corrections. Each of these methods resulted in an approximately 0.09%0 error in calculated 634S for every 1%o mismatch between sample and standard 6180. In contrast, the SO2cal algorithm calculated accurate values for 634S and for 6180 despite wide variations in oxygen isotopic composition of samples relative to standards.
45.2.6 Results for sulfur isotope standards and sulfide mineral samples The three-mass algorithm was furthur tested on a series of sulfur isotope reference materials (IAEA-S-1 silver sulfide, IAEA-S-2 silver sulfide, NBS-123 sphalerite, and
982
Chapter 45 - K. Leckrone & M. Ricci
Soufre de Lacq elemental sulfur) for which 634S was independently known to range from -0.3 to +21.0%o. All materials were converted to SO2 by vacuum-line combustion with vanadium pentoxide. Sulfur and oxygen isotopes were calculated by the standard equation (equation [45.1]), empirical correction (equation [45.24]), or algorithm (based on equation [45.20]). As shown in Figure 45.3, sulfur isotopes calculated using the SO2cal algorithm covaried against known 834S values with a slope of 0.99 + 0.04 (lc,) and an intercept of 0.2 + 0.6%0 (lc~). When either the standard correction or an empirical correction were used, the correlation between calculated and known values was similar Figure 45.3 - 634S calculated for sulfur isotope standards using a but had better precision, standard correction, empirical calibration, or algorithm. with a slope of 0.10 + 0.01%o (lo) and an intercept of 0.0 + 0.2%0 (lo). The poorer precision of results calculated with the algorithm was also apparent in the greater scatter for replicate analyses of any given standard (lo deviations of + 0.9%0 using the algorithm and + 0.3%0 using a calibration curve). The precision of 6180 calculated by the algorithm was ca + 12%o, or about an order of magnitude worse than the precision of calculated &34S values. Statistical analyses excluded 3 points which had clearly anomalous 65R values (ie, 665R/666R ratios of 0.7-0.9, rather than the usual ~0.5). Elevated 665R/666R ratios are a strong indication of the presence of a low-level contaminant which imparts excess m/z 65 ion current. This raises the possibility that the remaining samples in the set contained a lower level of contamination insufficient to signal an obvious error in the 665R/666R ratio but still enough to contribute variability in 634S and 6180 calculated by the algorithm. Unfortunately, no mass scans of the samples were made to check for contamination. The experiment was repeated using a suite of natural, isotopically homogeneous sulfide minerals, several of which are routinely used at the University of Georgia as
Oxygen Isotope Corrections in Continuous-Flow Measurements of 502
983
in-house laser microprobe IRMS standards. No 665R/666R anomalies were observed for these data, and no contamination was indicated in mass scans of the sample gases. For these samples, 634S calculated using the SO2cal algorithm was compared directly with ~)34S calculated from a calibration curve. As shown in Figure 45.4, ~)34S values covaried with a slope of 1.03 + 0.01%o (lo) and an intercept of 0.4 + 0.3%o (lo). As with the previous set, the l o precision for replicate analyses was worse when 634S was calculated using the SO2cal algorithm (+ Figure 45.4 - 634S for laboratory analyses of sulfide samples using 1.2%o) compared with empirical calibration or SO2cal algorithm for isotope correction. results determined via a calibration curve (+ 0.3%o). The precision of 6180 was comparable to before (+ 14%o). The precision of the 665R and 666R measurements used as inputs for the SO2cal algorithm were on order of + 0.1 to 0.2%0 external precision for 3-7 replicate analyses of each standard or sample. The limited precision of isotope ratios calculated using the SO2cal algorithm was therefore disappointing and some-what surprising, since a fundamentally similar approach is used routinely and successfully in oxygen isotope corrections of CO2 and shows no evidence of excessive propagation of errors (Santrock et al., 1985). One possible source of error in the SO2cal algorithm could be incorrect assignment of values for the constants a, a', K and K'. However, we could detect no statistically significant offsets in either ~)34S or (~180. Neither was there any dependence or covariations in 634S and 6180 values calculated with the algorithm, as would be expected for incorrect assignment of the constants (Santrock et al., 1985)..It therefore seemed likely that propagation of errors in the measured 65R a n d / o r 66R ratios were the probable source for the limited experimental precision. 45.2.7 Propagation of errors in 65R and 66R ion current ratios A formal analysis of error propagation in equation [45.20] is difficult. Error propagation was instead analyzed using the model datasets described above. Randomly
984
Chapter 45 - K. Leckrone & M. Ricci
generated numbers we- Table 45.2 - Absolute abundance measurements of sulfur isotope stanre multiplied by appro- dards priate powers of ten to standard 33R 34R reference K add noise at various levels to either 65R or NBS 123 0.0079705 0.0458050 Zhang & Ding, 1989 0.0390046 66R model input values, V-CDT 0.0078795 0.0441623 " 0.0392914 " 0.0392954 and the effects on calcu- IAEA-S-1 0.0078791 0.0441491 " 0.0392935 lated output values of IAEA-S-2 0.0079713 0.0451620 IAEA-S-3 0.0077490 0.0427464 " 0.0392946 834S and 8 1 8 0 w e r e observed. Results are average 0.0392359 summarized in Table st dev 0.0001293 0.32% 45.2. Each entry sum- rsd marizes a model dataset comprising 20 or more points spanning a 60 permil range in input 834S values, where A65R and A66R indicate the root-mean-square error introduced into ion currrent ratio data, and A34S and A180 indicate the root-mean-square error in calculated elemental isotope ratios. Individual data sets are shown in Figures 45.5-8, with calculated 8 values plotted against the original, input 8 values to which noise in the corresponding 65R o r 66R was added.
Figure 45.5 - Effect of added noise in the 65/64 ion current ratio on calculated 834S.
Two trends in the data stand out. First, noise in 65R affects the calculated results about a factor of three more strongly than equivalent levels of noise in 66R, so that noise in the 65R is the most important factor limiting the precision of the final calculated result. This is unfortunate for several reasons. Be-cause of the low fractional abundance o f 33S and 1 7 0 , the magnitude of the m/z 65 ion current is only about 20% of the m/z 66 ion current. The 65R consequently has a poorer signal to noise ratio t h a n 66R. It also requires more amplification, causing furthur deterioration in
Oxygen Isotope Corrections in Continuous-FlowMeasurements of SO 2
985
signal to noise, and it appears to be more sensitive to trace contamination. A second clear trend in the data is that equivalent levels of noise added to either 65R or 66R result in roughly an order of magnitude poorer precision in calculated 6180 than in 634S. This was also clearly visible in the laboratory analyses of sulfide standards and samples. In many instances, freeing the 634S calculation from the constraint of constant 6180 is of primary importance, and obtaining the correct ~180 value itself is not required. For samples prepared by vacuum combustion, EAIRMS or laser microprobeIRMS, the oxygen isotopic composition of SO2 is set during combustion and carries no meaningful informa- Figure 45.6 - Effect of added noise in the 66/64 ion current ratio on sulfur isotope calculation. tion, except potentially as an indicator of the quality of combustion. However, the large errors in calculated ~180 values preclude use of the SO2cal algorithm for direct sulfur and oxygen isotopic analysis of natural SO2 samples. The propagation of error experiment provides a valuable insight into the measurement precision required to obtain a desired precision in calculated 6 values. Specifically, regression of obser-ved errors in 6 against known errors in input 65R values suggests that 65R must be measured with a precision of ca 6 x 10-6 in order to calculate ~34S to + 0.1%o and 6180 to + 1%o. Since the external precision for replicate analyses of 65R measured on our mass spectrometer is typically on the order of + 0.0001 (0.1%o), the + 1%o precision in ~}34S observed for laboratory analyses of sulfide minerals and reference materials was actually somewhat better than this simple propagation of error analysis suggests. Using the relationships in Table 45.3, however, we estimate conservatively that the precision of 65R measurements would need to improve by a factor of 5-15 in order to routinely calculate 634S to 0.3-0.1%o. In further experiments, however, increasing the integration time, idle time, and number of measurement cycles did not significantly improve the precision of the 65R or 66R measurements or of
986
Chapter 45 - K. Leckrone & M. Ricci
634S and 6180 calculated by this method. We can only speculate about why propagation of error so dominates the SO2 algorithm, while a fundamentally similar approach for CO2 is extremely robust. Since carbon has only one relevant higher isotope compared to two for sulfur, the CO2 algorithm requires only one mass-dependent relationship while the SO2 algorithm requires two. However, it seems unlikely that reducing the uncertainty in a, a' K or K' constants will do much to improve the robustness of the SO2 algorithm. Instead, differing fundamental patterns of isotopic abundance are most likely the root cause for the soundness Figure 45.7 - Effect of noise added in the 65/64 ratio on sulfur of the CO2 algorithm and isotope calculations. the relative weakness of the SO2 algorithm. Both sulfur and oxygen share a classic isotope distribution pattern, in which an odd-mass isotope one mass unit away from the base species (33S or 170) is much rarer than the evenmass isotopes two units away (34S and 180). At the same time, the natural abundance of both sulfur isotopes is higher than that of their oxygen isotope counterparts - that is, 33S is more abundant than 170, and 34S is more abundant than 180. As a consequence, substitutions involving higher sulfur isotopes dominate both the mass 65 ion current (91% 33S1602 vs. only 9% 32S170160) and the mass 66 ion current (92% 34S1602 vs. only 8% 32S180160). In contrast, the only higher carbon isotope has odd mass. Consequently, in the mass spectrum of CO2 the m/z 45 ion current is dominated by the higher carbon isotope (94% 13C1602)while the m/z 46 ion current is dominated by a higher oxygen isotope (>99% 12C180160). It therefore stands to reason that resolving the underlying sulfur and oxygen contributions to the SO2 mass spectrum based on adjacent ion currents is far more difficult than discerning the underlying carbon and oxygen contributions to the CO2 mass spectrum by a similar treatment. Another correction related to the one explored here could be made from simultaneous measurement of the three adjacent ion currents m/z 48, 49 and 50 of the SO + fragment ion. However, this approach seems to offer no real advantage, since it would not address
Oxygen Isotope Corrections in Continuous-Flow Measurements of SO 2
987
the fundamental problem related to parallel isotope distribution patterns of S and O, and would involve lower ion currents. If the isotope correction for S02 can in fact be improved, it seems best to turn to an altogether different approach. 45.3 Corrections based on simultaneous collection of parent-daughter fragments 45.3.1 S O + m a s s s p e c t r u m
As an alternative to the SO2cal algorithm, we have made a preliminary evaluation of the method first proposed by Holt & Engelkemier (Holt & Engelkemier, 1970) for isotope corrections based on the simultaneous collection of ion currents from the m/z 64 and 66 molecular ion as well as m/z 48 and 50 ion currents from the SO + daughter ion, which is produced at approximately 70% intensity relative to the parent ion by electron ionization at 70 eV. The parent-daughter correction method is particularly attractive because as outlined below, it does not employ functional relationships b e t w e e n 33R and 34R, o r between 170 and 180, and it therefore makes no assumption of mass-dependent fractionation. However, it implicitly assumes that any isotopic fractionation which may occur in the mass spectrometric measurement during fragmentation of the SO2 parent ion is adequately addressed through the standard isotopic practice of differential measurement. As with SO2, the fractional abundance of the relevant SO + fragment ions should be related to the fractional abundances of individual isotopes according to 48F = 32F16F
[45.25]
49F = 33F16F + 32F17F [45.26] 50F = 34F16F + 32F18F + 33F17F [45.27] In the case of the SO + fragment ion, only one possible site for oxygen isotopic substitution is present, so there are no multiplicative factors in the fractional abundance expressions. Dividing equations [45.26] and [45.27] by equation [45.25] yields expressions for ion current ratios in terms of elemental isotope ratios relationships in terms of ratios: 49R = 33R + 17R
Figure 45.8 - Effect of noise added in the 66/64 ratio on the [45.28] oxygen isotope calculation.
988
Chapter 45 - K. Leckrone & M. Ricci
50R = 34R + 18R + 33R17R
Table 45.3 - P r o p a g a t i o n of errors in 65R a n d 66R t h r o u g h the
[45.29] SORcala l g o r i t h m Inspection of equations [45.9], RMS noise a d d e d RMS error in RMS error in [45.10], [45.28] and [45.29] to ion c u r r e n t ratio calculated 634S in calculated 6180 show that they can be comA65R A34S A180 bined in various ways to yield ........................................................................................................................................................................................................................... expressions for elemental iso5.7 x 10-6 0.10 1.1 tope ratios in terms of mea5.6 x 10-5 0.95 11. 5.1 x 10 -4 10.0 113. sured ion current ratios. For example, equations [45.28] and A66R A34S A180 [45.9] can be combined to solve f o r 33R
[45.30]
2 . 4 9 R - 65R - 33R
5.3 x 10 -6 5.8 x 10-5 6.1 x 10 -4
0.03 0.31 3.3
0.3 3.1 33.
This approach is attractive in that it solves exactly for 33R, from which 34R can be obtained by equation [45.19]. However, it is based on collection of the least abundant SO2+ molecular ion and the least abundant SO + daughter ion, and so is experimentally undesirable due to its lower signal-to-noise ratio and potential sensitivity to contamination. Focusing on the more abundant ion, combination of equations [45.10] and [45.29] yields an exact expression for 34R in terms of measured ion currents and 17R 2.50R -
66R = 34R - 17R2
[45.31]
If the 17R2 term is ignored because of its low fractional abundance (0.007% of the m/z 50 ion current, compared to 95.7% 34R and 4.3% for 18R), (~34S c a n be determined directly from 50
66
6348 _ (2" ~Rsa - 66Rsa - 1 ) " 1 0 0 0 2 5~ Rst
[45.32]
for any S O 2 sample, regardless of whether or not its oxygen isotope composition matches that of the standard. Combination of equations [45.29] and [45.9], or of [45.28] and [45.10], are also possible, but are less desirable due to their dependence on at least one low intensity signal and the need to eliminate at least one higher order term.
45.3.2 Experimental evaluation of parent-daughter method The parent-daughter method using equation [45.32] was first tested under typical conditions using a set of sulfide and sulfate mineral samples which were converted to SO2 by vacuum-line combustion with V205 combustion. Samples ranged in ~)34S from 21.0%o to -32%o, and were assumed to have constant 6180. Due to limitations of the cup configuration on the mass spectrometer (a Finnigan MAT 252 with 7-cup
Oxygen Isotope Corrections in Continuous-Flow Measurements of SO 2
989
MEMCO configured for CO2, N2 and 0 2 ) , m/z 48 and 50 ion currents were sequentially measured using a peak-hopping procedure after measurement of the 66R w a s complete. As shown in Figure 45.9, values of ~34S calculated with the equation [45.32] covaried against those calculated with equation [45.1] with a slope of 1.01 + 0.01 and an intercept of 0.02 + 0.1 (R2 = 0.999; ls standard deviations). The external precision of replicate analyses was slightly poorer for the 4mass correction (+ 0.6%0) compared to when the standard correction (+ 0.4%o). However, precision of the parent-daughter method Figure 45.9 - Comparison of parent-daughter and standard may have been limited by the corrections when oxygen isotopes of samples and standards precision of the 50R measure- are matched. ment obtained the use of the peak-hopping procedure rather measurement of a true, simultaneous ion-current ratio. To furthur test equation [45.32], we analyzed a set of SO2 samples in which ~34S the sulfur isotopic composition was constant but the oxygen isotopic composition was varied by equilibrating half of the samples with Table 45.4 - Comparison of isotope corrections when the oxygen isotopes of sample and standard are mismatched water enriched in ~180 by a large evaporative reduc634S calculated by 634S calculated by SO2 sample tion. To promote recovery equation [45.1] equation [45.1] of the SO2, the water was acidified to pH 1 prior to unequilibrated 9.32 9.44 the experiment, and (original 6180) 9.66 9.54 equilibrated SO2 was 9.48 9.23 9.32 9.89 extracted and purified by average + sd (lo) 9.60 + 0.20 9.53 + 0.27 cryogenic vacuum distillation. As in the prior equilibrated 11.83 8.84 experiment, m/z 48 and 50 (altered 6180) 13.13 9.27 ion currents were sequen12.95 8.71 tially measured using a average + sd (lo) 12.64 + 0.70 8.94 + 0.30 peak-hopping procedure
990
Chapter 45 - K. Leckrone & M. Ricci
after measurement of the 66R for each sample was complete. The sulfur isotopic composition for each sample was then calculated using equation [45.1] or equation [45.32]. Initial results are shown in Table 45.4. For unequilibrated SO2 samples, ~34S calculated using equation [45.32] matched those calculated using equation [45.1], and both methods resulted in similar precision for replicate analyses. For SO2 samples equilibrated with the isotopically enriched water, the 634S values calculated with equation [45.1] were clearly in error, and appeared enriched relative to the unequilibrated control samples by +3.0%o. The 634S calculated using equation [4:5.32] were depleted b y 0.5%0. This difference in ~34S calculated for equilibrated and control samples is significant, but is not necessarily due to failure of equation [4:5.32]. Controlling the oxygen isotopic composition of SO2 by equilibration with water is experimentally difficult. Yields of SO2 for equilibrated samples were relatively low (ca 70%), presumably due the very large Henry's law constant for solubility of SO2 (aq) in water. An unknown amount of fractionation between the extracted SO2 and the residual aqueous SO2 may have occurred. In spite of this ambiguity, the results of this and the previous experiment suggest that equation [4:5.32] yields results which are comparable to standard calculations when oxygen isotopic composition of the samples and standards is matched, and result in more accurate results when the oxygen isotopes of samples vary relative to standards. 45.4 Conclusions and recommendations for future work
While the SO2cal algorithm developed here allows for independent calculation of 634S and 6180 from 65/64 and 66/64 ion current ratios which are routinely collected for SO2, the precision of calculated (~34S is currently limited to ca + 1%o, and the precision of calculated 6180 to over + 10%o. Thus, this approach is not recommended except in cases where large oxygen isotope variations between samples and standards limit the accuracy of other oxygen isotope correction procedures to > 1%o. The precision of ~)34Scalculated by the method described here appears to be limited primarily by the precision with which the 65R ion current ratio is measured. Readily made modifications in mass spectrometric conditions, such as increasing the measurement and idle times and the number of measurement cycles, did not significantly improve the precision of measured 65R and 66R ion current ratios, nor did they improve results from the SO2cal procedure. It seems likely that isotope corrections for SO2 using this approach may be fundamentally limited by the relatively minor contribution of the oxygen isotope ratio terms 17R and 18R to the m/z 65 and 66 ion current ratios respectively. An alternative oxygen isotope correction based on collection of parent ions at m/z 64 and 66, and their daughter ions at m/z 48 and 50, resulted in accurate and precise calculation of ~)34Sfor samples with constant oxygen isotopic compositions, and provided better accuracy and precision than the standard correction when oxygen isotopes of samples were variable. These preliminary results suggest that potential fractionations associated with fragmentation of the parent ion in the ion source of the mass spectrometer can be adequately addressed through the standard isotopic practise of relative measurements. Routine use of the parent-daughter correction in continuous-flow applications would require a mass spectrometer with a collector system capable of true simultaneous monitoring of m/z 48, 50, 64 and 66 ion beams. Although
Oxygen Isotope Corrections in Continuous-Flow Measurements of SO 2
991
simultaneous measurements were not possible on the Finnigan 252 used for this work, some commerically available instruments are already capable of the measurement.
Acknowledgements This work was supported by NSF-EAR-9811170 to D.C. and K.L., and by an NSF earth sciences postdoctoral fellowship to K.L.
Handbook of Stable Isotope Analytical Techniques, Volume 1 P.A. de Groot (Editor) 9 2004 Elsevier B.V. All fights reserved.
CHAPTER 46 Experimental Measurement of Isotopic Fractionation Factors and Rates and Mechanisms of Reaction Simon M. F. Sheppard Laboratoire de Sciences de la Terre, CNRS UMR5570, Ecole Normale Sup6rieure de Lyon, 46 All6e d'Italie, 69364 Lyon, Cedex 07, France e-mail: [email protected]
46.1 Introduction Laboratory and field experimental studies have played a fundamenal role in the development of both isotope chemistry, since its very beginnings in the 1930's (e.g., Weber et al., 1935) and, subsequently, in isotope geochemistry (e.g., Thode et al., 1949; Epstein et al., 1951). The stable isotope ratios of many light (H, Li, B, C, N, O, Mg, Si, S, C1,...) and some moderately heavy (Ca, Fe, Cu, Zn, Se, Sb,...) elements are measurably variable. Examples for all of these elements are given here with analyses being made by an isotope ratio mass spectrometer (IRMS), an ion micoprobe (SIMS) or a multiple collector inductively coupled plasma source mass spectrometer (MC-ICP-MS). Applications of NMR spectroscopy (170, 27A1, ...) to determine the kinetics of exchange reactions are not considered here (e.g., Casey & Phillips, 2001; Phillips et al., 2003). Interpretation of mass-dependent isotopic variations in natural phases invariably requires knowledge on how the isotopes are partitioned or fractionated among associated phases, and of the mechanisms and rates of exchange. All of these depend on experimental data. More recently, experimental and theoretical studies have also contributed to our understanding of mass-independent isotopic fractionations (Thiemens & Heidenreich, 1983; Thiemens, 1999, 2001; Gao & Marcus, 2001). Fractionation factors are determined by experimental, theoretical and / or empirical methods. All methods have their advantages and limitations. Although the theoretical methods are not discussed here (see Richet et al., 1977; Kieffer, 1982; O'Neil, 1986; Clayton & Kieffer, 1991; Polyakov, 1998 and references therein), it is worth recalling that (1) they require experimental spectroscopic data including, if possible, on the phase strongly enriched with the minor isotope, and (2) experimental fractionation data are essential to confirm or test the statistical mechanical model calculations (e.g., Urey & Rittenberg, 1933; Bigeleisen, 1965; Bigeleisen et al., 1973; Clayton & Kieffer, 1991). Thus, a detailed interplay between the results of theoretical studies and high precision experimental fractionation data is essential.
Experimental Measurementof Isotopic Fractionation Factors ...
993
In this presentation of experimental stable isotope geochemistry, emphasis is placed on the principles, methods and techniques of isotopic experiments, certain aspects of their design and some problems associated with the interpretation of the data. The examples given have been selected using a number of criteria including the pioneering application, and more recent studies to illustrate particular points. Exclusion of a paper in no way reflects on the quality of the work because there are more good studies than can be mentioned here. Such a chapter cannot review the experimentally derived data or end-results of a specific experiment. However, recent reviews of (1) equilibrium isotopic fractionation factors with a listing of systems studied and their calibration(s) has been given by Chacko et al. (2001), and (2) rates and mechanisms of isotopic exchange have been presented by Cole & Chakraborty (2001). Also, details of the equipment and its use, choice of materials for furnaces, thermocouples, etc.., and measurement of temperature, pressure, f02, pH, etc.., are not discussed because they would require a book to themselves. They are available in, for example, Edgar (1973), Ulmer & Barnes (1987), Holloway & Wood (1988) and references therein. Empirical methods of determining fractionation factors are also discussed. In general, these methods can be considered to be field experiments, in contrast to laboratory experiments. Such an approach has been an important source of fractionation data, particularly for biologically mediated reactions (e.g., aragonite or phosphate water systems, or sulphate reduction), and isotope exchange reactions with extremely slow exchange rates on a laboratory time-scale, often, but not invariably, at low temperatures (e.g., clay mineral systems). The experimental determination of fractionation factors is essential to provide constraints on the interpretations of isotopic data on natural systems and for understanding the principles underlying their behaviour. The temperature calibration of the isotopic geothermometers is one facet. Characterizing kinetic isotope effects, and mechanisms and rates of exchange are others. The general principles of isotopic exchange reactions and mechanisms of exchange as well as the notion of equilibrium are briefly discussed because these can influence the design of an experiment. Examples are inevitably dominated by the isotopic ratios of light elements measured by IRMS (H, C, N, O, Si, S, C1,...) because experimental studies have only really just started on some of the "new" elements (Mg, Ca, Fe, Cu, Zn, Se, Sb,...) that are measured by MC-ICP-MS (e.g., Mar6chal et al., 1999; Matthews et al., 2001; Zhu et al., 2001; Beard et al., 2003; Rouxel et al., 2003), or using new microtechniques such as SIMS (e.g., Valley et al., 1998) and laser probes (Rumble & Sharp, 1998; Shanks et al., 1998). 46.2 The isotopic fractionation factor To avoid any misunderstanding, the definition and implications of the isotopic fractionation factor between two substances A and B is: (~A-B = RA/RB
[46.1]
994
C h a p t e r 46 - S.M.F. S h e p p a r d
where R is the isotopic ratio of the less abundant and heavier isotope over the more abundant and lighter isotope of the element of interest (e.g, D / H , 180/160, 65Cu/ 6 3 C u , .... ), with two notable exceptions: 6 L i / 7 L i and 1 1 B / 1 0 B . Note that the definition is both quite general, being applicable to both mass-dependent and mass-independent isotopic fractionations, and practical; it is applicable to any two substances that contain the element and assumes no thermodynamic or other relationship between the two substances. For example, one can talk meaningfully about the c~'s between starting materials of an experiment before any isotopic exchange has occurred. The fractionation between any two substances can be derived directly from the 6-values of A and B that are measured relative to a common standard (O~A-B - [1000 + 6A]/[1000 + 6B]). Clearly, however, c~x-y is more interesting and useful if it is related to a welldefined kinetic or equilibrium process. But more information is required before c~ can be equated with a thermodynamic function such as the equilibrium constant, K. Although the temperature calibration of a fractionation factor is nearly always presented as a function of c~ (e.g. 1000 In (~A-B - f(T) rather that 1000 In K A - B - f(T) with T in Kelvin), it is too often implicitly assumed that c~ and K are interchangeable, without specific or unambiguous evidence for equilibrium, using the relation (~A-B = K1/nA_B where n is the number of exchangeable atoms in the reaction and assuming that the isotopes are randomly distributed over all possible sites. 46.3 Unidirectional and bidirectional reactions
Experiments to measure fractionation factors fall into two major families: unidirectional and bidirectional processes. The former are specifically for measuring kinetic effects since they are time and, therefore, path dependant; equilibrium may be attained but cannot be unambiguously demonstrated. The latter are for deriving equilibrium fractionations and kinetic information such as activation energies and rate constants for the exchange reaction. Basically it is necessary to devise an experimental setup where the uni- or bi-directional reaction can take place under strictly controlled physical, chemical, mineralogical and, if necessary, biological conditions. Although the experiment may appear to be rather straightforward in principle, in practice, a number of more or less surmountable problems may arise. For a unidirectional reaction, the fracfionation factor is strictly kinetic. Being path dependant, its value is not necessarily unique for any chosen overall reaction. Taking the reduction of sulphate to sulphide at 25~ and I bar as an example, experimentally measured fractionation factors vary from 0 to about 50%o, compared with the calculated equilibrium value of about 70%o (e.g., Kemp & Thode, 1968; Goldhaber & Kaplan, 1974; Canfield, 2001a). The magnitude of the kinetic fractionation depends on which step during sulphatereaction is rate limiting, the rate of reduction and whether the reduction is chemical or bacterial. For bacterial reduction, the fractionation varies with the strain of the bacteria, their degree of acclimatization to the various electron donors, nature of food and its supply, etc... Hidden or unidentified factors may play a decisive role during the course of the reaction. The situation can be quite complex. Isotopic exchange reactions that involve chemical reaction, and thus mineral synthesis, are unidirectional reactions. Such systems, which initially are not in chemical equilibrium, have been used to derive fractionation factors and to elucidate equilib-
Experimental Measurementof Isotopic FractionationFactors ...
995
rium fractionations (e.g., Clayton, 1959; O'Neil & Taylor, 1967, 1969; Matthews & Beckinsale, 1979; Bird et al., 1993; Vitali et al., 2001). Further aspects of this critical subject are discussed in section 46.5.
46.4 Equilibrium Measuring equilibrium fractionation factors has played a central role in experimental studies. Thus, the following remarks, concerning bidirectional exchange reactions that are designed to measure equilibrium fractionation factors, discuss criteria for complete exchange experiments and methods of extrapolation to the equilibrium value for partial exchange experiments. To determine equilibrium isotope fractionations, criteria must be built into the experimental setup to demonstrate that equilibrium was attained or, at least, more or less closely approached or bracketed (i.e., the reaction is reversible). Consider an exchange reaction between two substances where only the isotopes of a single element exchange between the two substances, that are neither created nor destroyed during the exchange reaction and experience no change in chemistry or texture" xA1 + yB2- xA2 + yB1
[46.2]
where A and B are the two substances, and I and 2 refer to the molecules totally substituted by the light and heavy isotope respectively. Thus, only the isotopic composition of the substances changes during the reaction. Equation [46.2] can refer to a homogeneous reaction taking place within a single physical phase, such as a fluid, or a heterogeneous reaction where the reaction occurs at the interface between the two phases (no production of a new phase or change in texture). For the latter case, exchange is considered to be purely diffusional and solution reprecipitation or modification of the surfaces is taken to be negligible (see section 46.5). It is assumed, for simplicity, that for substances with more than a single atom of the element of interest (e.g., oxygen in CaCO3), the several atoms are equivalent or at least indistinguishable during the exchange experiment. Although such an exchange reaction can be written between any two constituents, whether it actually works in the laboratory, or in nature, depends on the availability of a suitable mechanism of exchange and its rate. If equilibrium is attained (i.e., 100% exchange), the actual reaction path followed does not affect the value. Here, we assume that an efficient exchange mechanism exists. At any given time, t: CX(A-B)t- (1000 + ~At)/1000 + ~Bt)
[46.3]
but at equilibrium, eq: (Z(A-B)eq = (1000 + 6Aeq)/1000 + 6Beq) - K1/n(A-B)
[46.4]
where K is the equilibrium constant and n is the number of exchangable atoms with K - [(A2)x(B1)Y]/[(A1)x(B2)y] = (A2/A1) x/(B2/B1)y
[46.5]
996
Chapter 46 - S.M.F. Sheppard
Such an exchange reaction (equation [46.2]) implies that the equilibrium fractionation value can be approached from both sides of the value (equilibrium bracketing or reversing) without modifying the substances, except for their isotopic composition. That is, the reaction is perfectly reversible as the rate-controlling factors are independent of whether the isotopic ratio of the substances increases or decreases to arrive at the equilibrium value. The attainment of equilibrium can thus be unambiguously demonstrated. In practice, systems are not usually ideal and criteria for equilibrium are usually more complex, particularly for heterogeneous systems that contain a solid phase. Pattison (1994) gives an excellent discussion of the subtleties of this crucial point. After Pattison (1994), for most experimental isotopic systems between a solid phase and a fluid, what is usually referred to as an equilibrium fractionation factor is actually apparent isotopic compositional bracketing. Examples of the two-direction approach are given in Figure 46.1 for heterogeneous reactions. Figure 46.1a presents a simple, ideal model example from O'Neil (1986) for O-isotope exchange between exactly the same quartz sample (grain size, defect density, initial isotopic composition, etc...) and three isotopically different waters (W1, W2 and W3), that are the same except for their 180/160 ratios. The isotopic exchange rates for the forward and reverse reactions are taken to be identical. Note that isotopic exchange between three isotopically different quartz samples and a single water is not necessarily an equivalent system; the three quartz samples are highly unlikely to have exactly the same physical and chemical properties except for their isotopic compositions. At 500~ (Figure 46.1a), 50% exchange occurs per day in the three parallel experiments W1, W2 and W3 that have exactly the same P-T-t history and therefore identical rate-controlling factors. Thus, the degree of exchange in the three charges is proportional to the distance that each system is from isotopic equilibrium. After more than 5 days, > 97% exchange has taken place; the equilibrium fractionation of 3.0 can be determined with the necessary precision if the initial quartz-water fractionation in charge W2 was chosen to be close to the equilibrium (or final) value. Figure 46.1b, after Graham et al. (1980), presents the measured time dependence of the H-isotope fractionation between zoisite-water for two parallel runs where everything was identical except the D / H ratios of the waters, D and E. After about 10 days, essentially 100% exchange had been achieved. The close similarities between the nature of the curves of Figures 46.1a and 46.1b are evident. The form of the curves in Figure 46.1b 4 6 . 1 - Diagram illustrating the change in the fractionation factor with time for exchange reactions using the two-direction approach to the equilibrium fractionation, a Model O-isotope exchange between quartz-water with starting-material: 618Oquartz at + 10 and three waters at - 5 (A), + 5 (B), and + 15 (C), that proceeds at the rate of 50% exchange per day (after O'Neil, 1986). After about 5 days the fractionations for runs A, B, and C converge and bracket the equilibrium fractionation of 3%o.b - Measured H-isotope fractionations between zoisite and water with time for two companion runs with the same zoisite but isotopically different waters as starting-materials (after Graham, et al., 1980). The initial zoisite-water fractionations are both smaller and larger than the equilibrium value. After about 10 days, the two fractionations converge on the equilibrium value of about- 52%o (i.e.--- 100% exchange). Figure
Experimental Measurement of Isotopic Fractionation Factors ...
997
998
Chapter 46 - S.M.F. Sheppard
implies that the hypothesis that the rate of exchange, for both the forward and reverse reactions, is proportional to the distance that the run is from equilibrium is quite satisfactory for this system. Also, at the rather similar levels of concentration of the different isotopic species in these experiments, the rate constants for the different isotopic species are indistinguishable. For heterogeneous reactions, the rate of reaction will not necessarily be constant with time because the surface areas of the grains may change with time (see section 46.5). On Figure 46.1b, there are too few data points, particularly during the first few minutes and hours of the experiment, to determine how the rate of exchange changed with time. The process was probably not purely diffusional and without dissolution; it, therefore, probably was not constant. In reality, some exchange reactions do not attain, or even closely attain, equilibrium, because rates of exchange are just too low on the laboratory scale (usually less than a few months and very rarely over a year). Extrapolation techniques have been developed for partial exchange experiments.
46.4.1 Partial exchange: the Northrop-Clayton method Partial exchange techniques were developed by Northrop & Clayton (1966) for two phase systems, where each phase contains only a single site for the element of interest. In such systems, the rates of isotopic exchange for directly associated experiments are assumed to be equal but not necessarily constant. Associated experiments refer to 2 or more runs that are as identical as possible in (1) their preparation- identical solids (grain size and its distribution, isotopic composition, imperfections, mass,...), identical solutions (chemistry, mass,...) except for their isotopic compositions, identical reaction containers (volume, degree of filling, etc...), and (2) their P-T-t paths during their reaction history. Figure 46.1a illustrates the results of a model experiment where the rate of exchange is constant. Usually three runs are performed together. Two of the runs are chosen so that their initial isotopic compositions of the fluids are such that the isotopic differences with the solid, or salt to be dissolved, are both substantial and on both sides of the estimated equilibrium value, i.e. the initial a's are both smaller and larger than the equilibrium value of a. The evolution of these fractionations with time can thus bracket the equilibrium value (Figure 46.1). A third associated run is designed to have an initial fractionation that is as close as possible to the anticipated true equilibrium value. In this way (1) the degree of exchange, and (2) the extrapolated equilibrium fractionation value can be determined with the best precision for the degree of exchange. Northrop & Clayton (1966) derived the following equation [46.6] for the kinetics of partial isotopic exchange reactions in heterogeneous systems between phases A and B for all initial and final values of ~A-B (for simplicity just c~) close to unity (i.e., In a ~ 1 + c0 and assuming that the reaction rates in the forward and reverse directions are the same: In c t i - In CZeq- 1/F(ln ~f- In c~i)
[46.6]
where the subscripts i, f and eq refer to the initial, final and equilibrium fractionations,
999
Experimental Measurement of Isotopic Fractionation Factors ...
respectively, and F is the fractional approach to equilibrium. For a set of runs (usually 3), a plot of (In af- In cq) versus In ~i (Figure 46.2) gives a straight line with slope 1/F and an intercept with the y-axis of In aeq, if the assumptions are valid. For a single exchangeable atom, at equilibrium In af = In (~eq = In K and F - 1. The accuracy of the determination of the intercept or equilibrium value increases with increase in the percentage of exchange. Equation [46.6] has also been applied to homogeneous reactions. For hydrogen, the approximation In r ~ (1 + r is often not satisfactory. Therefore Suzuoki & Epstein (1976) used the following modified form of equation [46.6]" (ai
-
1) = (~eq - 1) - 1 / F(cxf - oti)
[46.7]
Equation [46.7] was used by Graham et al. (1980) for the analysis of the data presented on Figure 46.1b. Criss et al. (1987) and Criss (1999) reexamined the derivation of equations [46.6] and [46.7] and proposed a more complex quadratic equation [46.8] for a pair of exchange experiments, labelled by subscripts I and 2, based on intensive
Figure 46.2 - An example of the Northrop & Clayton (1966) partial exchange method where three companion runs, starting with the same mineral (dolomite) and three isotopically different waters, exchange for a given time and temperature (after Northrop & Clayton, 1966 and O'Neil, 1986). The intercept with the Y-axis gives the best estimate of the equilibrium fractionation for the set of runs, and the slope gives the degree of exchange (about 50%). With increase in percent of exchange, the line rotates anticlockwise, increasing the precision on "the best estimate of equilibrium fractionation".
1000
Chapter 46- S.M.F. Sheppard
variables only and ~ values rather than the logarithms of ~ values" ( 1 - H)(o~eq)2 + (Hcqi + HR2f- o~2i- C~lf)(Req + (O~lfC~2i-Ho~2fO~li)- 0
[46.8]
where H = (RBi/RBf)I (RBf/ RBi)2
[46.9]
using the same notation as above plus RB, the isotopic ratio of phase B, often water. Equation [46.8] is used to solve for (~eq. In the Criss derivation, the fractional approach to equilibrium, E is given by: 1- F - (RBf/RBi) [(af- ~eq)/(O~i- ~eq)]
[46.10]
and F is in fact a complex function of the relative mole fractions of A and B (Criss, 1999). All these equations [46.6,7 and 8] assume that the rate of isotopic exchange is proportional to the deviation of the isotopic ratio from the equilibrium ratio. This technique has been widely applied to the geothermometric calibration of heterogeneous single mineral-water (or aqueous solution) systems for H-, C- and O-isotopes (e.g., Northrop & Clayton, 1966; O'Neil & Taylor, 1967, 1969; Clayton et al., 1972; Suzuoki & Epstein, 1976; Matsuhisa et al., 1979; Graham et al., 1980, 1984:, 1987; Stoffregen et al., 1994a,1994b; Fortier et al., 1995; Guo & Qian, 1997; Cole & Ripley, 1998; Saccocia et al., 1998) and homogeneous reactions in water (e.g., Lloyd, 1968; Chiba & Sakai, 1985; L6cuyer et al., 1999; O'Neil et al., 2003). In order to avoid problems such as congruent or incongruent dissolution of the mineral in water, carbonate (e.g., Clayton et al., 1989, Chiba et al., 1989; Fortier et al., 1994), CO2 (e.g., Scheele & Hoefs, 1992; Matthews et al., 1994; Palin et al., 1996) and H2 (Vennemann & O'Neil, 1996) have been used as the exchange media. Based on a substantial number of experiments that have applied this technique and comparisons with other techniques, it has been observed that if the percentage of exchange was less than about 80 - 90% or so, then the intercept value for the incomplete exchange method may be larger than the "correct" equilibrium fractionation (Clayton et al., 1972; Matsuhisa et al., 1978; O'Neil, 1986). The reasons for this difference are not always understood but the reactions usually involve water or carbonate. Additionally, the basic hypothesis that the rates of isotopic exchange are identical for systems that only differ in their isotopic composition of the fluid reactant may not always hold. Recent experiments on exchange between (PO4)aq and H20 suggest that, under certain conditions of pH, the rate of exchange may depend on the distance from equilibrium (O'Neil et al., 2003). Further aspects of this observation need to be explored. Another way round this type of situation in systems with a solid reactant may be provided by application of the ion probe, if the individual crystals are sufficiently large, so that isotopic profiles can be measured (Fortier et al., 1995; Chacko et al., 1999). The Northrop-Clayton method may not be applicable to substances where the element of interest is present in more than a single crystallographic site or there are more
Experimental Measurementof Isotopic FractionationFactors ...
1001
than two phases (e.g., oxygen in micas, hydrogen and oxygen in chlorites, carbon and oxygen in calcite-dolomite-fluid, etc...)(O'Neil & Taylor, 1969; Sheppard & Schwarcz, 1970; Sheppard, 1980; Chacko et al., 1996; Cole & Ripley, 1998). For such systems, the exchange rates among the different sites may be similar or quite different, depending in large part on the mechanism of exchange (see section 46.5). Equations for exchange in three phase systems have been developed by Zheng et al. (1994). In systems with two solids and a fluid, the form of the curves representing the changes in the isotopic compositions of the solids and fluid with time - usually initially rapid followed by slower 8 changes - are much more complex than those given in Figure 46.1 because the kinetics are usually more complicated in systems with two solid phases (Zheng et a1.,1999).
46.4.2 Partial exchange: the three-isotope method The three-isotope method is an important modification of the Northrop-Clayton technique that was introduced by Matsuhisa et al. (1978), with later modifications by Matthews et al. (1983a). The isotope ratios 1 7 0 / 1 6 0 and 180/160 are measured on the same substances so that two fractionations before (or initial) and after (or final) are derived from each experimental capsule. The three-isotope plot, Figure 46.3, illustrates the method, with the assumption that the two isotopic ratios change at the same rate. The natural or synthetic starting mineral (Mo) plots on the primary mass-dependent fractionation line, PF, but the initial isotopic composition of the water sample (Wo) is chosen, by mixing with isotopically enriched water, so that it plots off the line PF in such a way that the 180/160 ratio is close to the presumed equilibrium value whilst the 170/160 ratio is far from the equilibrium value. In this way the change in the 170/160 fractionation sensitively monitors the extent of exchange while the 180/ 160 ratio closely brackets the equilibrium value, enabling accurate determination of its value. For 100% exchange, the final 8170 and 8180 mineral (Me) and water (We) equilibrium values plot on a secondary mass-dependent fractionation line that is parallel to PF and passes through the bulk isotopic composition of the mineral-water system. Isotopic measurements are made on pure oxygen gas. The combination of the three-isotope exchange method with the use of high pressures (e.g., piston cylinder apparatus) to accelerate the rate of exchange (see section 46.6.3) is thus an important technical innovation in experimental calibration. 46.5 Mechanisms and rates of isotopic exchange Many experimental studies designed to measure equilibrium or kinetic fractionation factors also yield information on the factors which affect the exchange reactions (e.g., Graham, 1981; Giletti, 1985; Cole & Ohmoto, 1986; Cole & Chakraborty, 2001). Some experiments were specifically designed to elucidate mechanisms and rates of exchange (e.g., Sakai & Dickson, 1978; Ohmoto & Lasaga, 1982; Giletti, 1985; Cole, 1992, 2000). Knowledge (or anticipation) of the mechanisms and extent or rates of exchange for both heterogeneous and homogeneous isotopic reactions can influence both the design of an experiment and the interpretation of the results. Isotopic exchange between a mineral and a fluid can occur by (1) chemical reaction or mineral synthesis, (2) dissolution-reprecipitation, or (3) diffusion. In reality, more than one
1002
Chapter 46 - S.M.F. Sheppard
Figure 46.3 - Schematic diagram of the three-isotope exchange method (after Matthews et al., 1983a). Natural samples plot on the primary mass fractionation line (PF) with slope of 0.52 (Matsuhisa et al., 1978). Initial isotopic compositions are mineral (Mo) on PF and water (Wo) whose fractionation with the mineral is well removed from the equilibrium value in 170/160 (by addition of labelled water; see Matthews et al., 1983a for details) but very close to equilibrium in 180/160. Complete isotopic exchange for the exchange reaction is defined by the secondary mass fractionation line (SF) parallel to PF and passing through the bulk isotopic composition of the total mineral plus water system. Isotopic compositions of partially equilibrated samples are Mf and Wf and completely equilibrated samples are Me and We. Values for Me and We can be determined by extrapolation from the measured values of Mo, Mr, Wo and Wf.
mechanism may operate in the system at the same time, but one mechanism often dominates. Chemical reactions involve the production of phases that were not present at the beginning of the experiment. They are unidirectional reactions which may attain equilibrium, but this cannot be unambiguously demonstrated. Although free energy changes associated with chemical reactions are usually one to three orders of magnitude larger than those for an isotope exchange reaction by itself, certain reactions can lead to metastable phases and/or disequilibrium fractionations, that are not necessarily readily detected (e.g., Matsuhisa et al., 1978). Although the subject of chemical versus isotopic equilibrium cannot be discussed here, it is noted that the attainment of chemical equilibrium does not imply that isotopic equilibrium has also been attained. Dissolution-reprecipitation processes group together a number of processes that occur at external surfaces or within the grains, and certain involve chemical reactions.
Experimental Measurement of Isotopic Fractionation Factors ...
1003
Experimental and theoretical studies have shown that, in general, the rate constants for exchange are related to surface area, fluid/solid ratio, the fraction of exchange and the time (Cole et al., 1983; Dubinina & Lakshtanov, 1997). A special case of dissolution-reprecipitation is the Ostwald ripening kinetic process (Ostwald, 1900; Baronnet, 1982; Eberl et al., 1990) which has been modelled by Chai (1974) and Stoffregen (1996). Because the starting solids invariably have a range of grain sizes and solubility varies with particle size (smaller grains have a higher surface energy), the larger particles grow at the expense of the smaller ones, without involving nucleation processes. The end products of an experiment can lead to a steady state condition where the grainsize distribution is independent of time and the physicochemical variables of the experiment, if the total number of particles is large enough to give a continuous size distribution. Exchange of O-isotopes between calcite and water during recrystallization of the former is consistent with an Ostwald ripening process (Anderson & Chai, 1974), and similarly for quartz-water (Matthews et al., 1983a), alunite-water (Stoffregen et al., 1994b) and chlorite-water (Cole & Ripley, 1998). Examples of changes in grain-size of run-products compared to their starting materials observed during experimental runs with quartz-water, anorthite-water and calcite-CO2 are given by Matsuhisa et a1.(1978), Matthews et al. (1983a) and Rosenbaum (1994). Note that the cores of the larger, growing grains may retain their initial isotopic composition because they are effectively armoured from isotopic exchange by the newly deposited layers. This leads to mineral-fluid isotopic disequilibrium without necessarily the intervention of a kinetic isotope effect. Chai's (1974) and Stoffregen's (1996) models enable one to calculate the fractional volume of the original material transferred through the fluid medium for a given increment of the average grain-size. For example, doubling the average grain-size implies that 75% of the initial solids have been transported through the solution (see Figure 8 in Chai, 1974). Importantly, isotopic overshooting of the true equilibrium value can occur by such a process. A dissolution-reprecipitation processes was proposed by O'Neil & Taylor (1967, 1969) for experiments where cation exchange took place to promote O-isotope exchange in systems such as albite-KC1-H20, sanidine-NaC1-H20 and muscoviteNaC1-H20. In particular, they showed (see Figures 5 & 6 in O'Neil & Taylor, 1967) that the gross morphology of the crystal was retained as the reaction front of both cation and isotopic exchange swept through the crystal, with the fluid within the crystal remaining in communication with the external fluid reservoir. There was no evidence for armouring of the original grains. Cation exchange dramatically increased the rate of isotopic exchange, but equilibrium cannot be demonstrated in such experiments. Dubinina & Lakshtanov (1997) have modelled these processes. The rate of change of solution-reprecipitation, textures and isotopic exchange in a charge and mineral-volatile fractionations can be influenced by the presence of "impurities". These may have been added during filling of the charge or produced within the capsule during the run and/or pressure-temperature device. For example, production of hydrogen from the dissociation of moisture in the pressure medium of a piston-cylinder apparatus, followed by its infiltration into the charge to react with CO2 to give CO and H20, has been discussed by Rosenbaum & Slagel (1995). Rosen-
1004
Chapter 46 - S.M.F. Sheppard
baum (1994) documented the role of minor quantities of water on these rates of change in the calcite-CO2 system. Importantly, he also demonstrated that, because the rate of exchange between CO2 and H20 was much faster than between either of these volatiles and the associated calcite, the final measured calcite-CO2 fractionation had been modified from its value attained at the temperature of the run (900~ thereby accounting for the scatter in the results of the runs of Chacko et al. (1991). For minerals like micas and chlorite with more than a single crystallographic site for oxygen, the rates of exchange of the different oxygens may be similar or different. If the mechanism of exchange is dominantly by recrystallization or dissolution-reprecipitation, then the rates may be very similar. On the other hand if pure diffusion is dominant, then the rates of exchange may be quite different. Diffusion, or volume diffusion, implies that each grain retains its dimensions and shape during the transport of isotopes across the surfaces. The magnitude of the diffusion coefficient of the element of interest depends on the nature of the diffusing or carrier species (often not well characterized) that in turn is related, but not necessarily directly) to the nature of the external reservoir (e.g., H20, CO2, 02, H2, CaCO3). In fact the experimental data can be used to make a choice among a number of possible carriers. For example, experiments that compare "dry" conditions with "wet" conditions can define the role, if any, of H-bearing species (e.g., Elphick et al., 1988). In many systems, isotopic exchange commences relatively rapidly with the surface layers and is followed by lower rates as the diffusing species penetrate into the bulk of the grain. Certain of the experimental data aimed at measuring diffusion coefficients have or may have been influenced by solution-reprecipitation processes. Some of the more recent studies measuring diffusion constants use isotopically enriched material and measure the isotopic depth profile in individual grains, often of known crystallographic orientation, using the ion microprobe (e.g., Giletti et al., 1978; Giletti & Yund, 1984; Elphick et al., 1986, 1988; Fortier & Giletti, 1991; Farver & Yund, 1995, 1999). In fact, equilibrium hydrogen isotope fractionation factors have been derived from depth profiles using the ion probe (Chacko et al., 1999). For further discussion of this topic see the review by Cole & Chakraborty (2001) and, for example, Graham (1981), Stolper & Epstein (1991), Zhang Y. et al. (1991a, b), and Palin et al. (1996). Another consequence of pure diffusion controlled exchange processes is that fractionation factors can evolve with depth of penetration, because surface properties are not identical to those of the bulk material (Hamza & Broecker, 1974). Application of suitable self-diffusion coefficient data are necessary to confirm that the length of the experimental run is long enough so that isotopic exchange has penetrated sufficiently deep (> ~ 0.3 mm) into the material to avoid the influence of surface fractionation factors (e.g., Stolper & Epstein, 1991; Matthews et al., 1994). Surface effects may also intervene, but be difficult to detect, if mineral to fluid ratios are very small and essentially complete exchange has only occurred between the surface layers and the fluid. Elphick et al. (1986) avoided grain boundary and pressure solution-reprecipitation processes from affecting their diffusion experiments by using a gel technique that hydrothermally deposited an isotopically labelled overgrowth on the quartz or feld-
Experimental Measurementof Isotopic Fractionation Factors ...
1005
spar starting material. Such experimental diffusion data are essential input to model quantitatively the effects of diffusion on stable isotope partitioning among minerals during prograde reaction histories and retrograde exchange processes, with applications to thermometry, open versus closed system behaviour, and cooling rates (Giletti, 1986; Eiler et al., 1992, 1993; Jenkin et al., 1994b; Kohn & Valley, 1998). However, these models can also be used to aid a field experiment aimed at calibrating an isotopic fractionation factor (J. W. Valley, pers. comm., 2001). For example, model results can aid the screening of samples by indicating the relative role of mineralogy, mode and grain size. The isotopic exchange reaction given in equation [46.2] implies that constituents A and B do not change chemically, i.e., they are in chemical equilibrium with each other. However, many experimental exchange studies start with an aqueous solution that initially is not in chemical equilibrium with the solids that are congruently or incongruently soluble in water. A certain amount of chemical reaction will inevitably take place and this can influence the results of the isotopic exchange reaction and quench products (see section 46.6.5). In incomplete exchange experiments which give extrapolated fractionation factors that are different from the true equilibrium value, water and carbonates are usually involved. Such problems appear to be particularly associated with the quartz-water and calcite-water systems (Matthews et al., 1994). Much effort to derive equilibrium fractionation factors has been invested in these important systems (quartz: O'Neil & Clayton (1964), Clayton et al. (1972, 1989), Matsuhisa et al. (1978, 1979), Matthews & Beckinsale (1979), Matthews et al. (1983a); calcite: Clayton (1961), O'Neil et al. (1969), Anderson & Chai (1974), Clayton et al. (1989), Cole (1992), Kim & O'Neil (1997)). On the other hand, dry gas-mineral exchange experiments, some of which may be controlled by pure diffusion, are often free of the above problems (Matthews et al., 1994; Rosenbaum, 1994). Future studies are required to characterize the order of the reactions and the mechanisms of exchange to advance our understanding of such systems.
46.6 Laboratory experimental methods Fractionation factors are so close to 1, usually between 1.05 and 0.95, that free energy changes of isotopic reactions are only a few calories. They are thus too small to be measured calorimetrically. Laboratory experiments therefore have to be designed so that the isotopic composition of the various substances before and after exchange can be measured directly or calculated from other measured parameters of the experiment. Experiments are carried out in a wide variety of containers made out of borosilicate or silica glass, plastic or noble metals, depending in large part on the selected experimental conditions. From an experimental point of view, it is convenient to divide laboratory experiments into two groups: (1) those conducted below, at or near atmospheric pressures, and (2) those contained in capsules that are placed within a high pressure-high temperature device. If the physical parameters cannot be con-
1006
Chapter 46 - S.M.F. Sheppard
trolled within the container then it is usually placed within an experimental P and/or T controlling device. A given type of device is usually associated with pre- and postexperimental characteristics such as the P-T-t heating up and quenching paths in addition to the designed P-T-t experimental conditions. Many of the P-T machines and techniques exploited by isotope geochemists have been adapted from those developed by the experimental petrology community. Table 46.1 summarises some of the characteristics of the principal devices that have been used for isotopic studies. Except for most of the near atmospheric pressure vessels, details of these devices and, in particular, their construction, calibration, operation, necessary safety precautions, basic principles of the preparation of sample materials in experimental geochemistry and original and key references are given in the books of Edgar (1973), Ulmer & Barnes (1987) and Holloway & Wood (1988). For experiments carried out above room temperature, the inevitable heating and/or quenching time-temperature trajectories can influence the end products, even if certain precautions are taken. However, quantitative or even semi-quantitative data on heating up and cooling histories are usually not given. Additionally; variations among different versions of the same general type of vessel can be quite considerable. Note that temperature-time quench paths follow an exponential curve so that the quench times given in Table 46.1 must only be taken as an extremely rough guide for the high temperature part of the curve.
46.6.1 Containers and capsules Containers can be basically an extremely simple test tube, flask or beaker in borosilicate or silica glass, or plastic, or a plastic box all at near atmospheric pressure, in the presence or absence of air or a controlled atmosphere. Others are much more complicated glass or precious metal capsule systems that require a device to at least tightly control the thermal regime. The principal purpose of the container is to hold the sample material in either a closed or controlled system during the run so that, for example, isotopic exchange can be interpreted in terms of the kinetic or equilibrium fractionation of interest. Under some conditions, usually at high temperatures, Fe loss to the metal capsule, loss of alkalis from a melt, or hydrogen diffusion can be so important that the chemistry of the phases changes substantially during the experiment, thus affecting the end results. From an isotopic point of view, a closed system requirement can readily be met for all elements except hydrogen, because glass and metals can act as semipermeable membranes for hydrogen. Sufficient quantities of hydrogen can diffuse through noble metal and silica glass capsules during a run to affect the interpretation of the end products of an experiment concerning H-isotopes unless special precautions are taken to reduce or eliminate such mass transfer processes (Graham et al., 1980; Vennemann & O'Neil, 1996). Note, as mentioned above in section 46.5, diffusion of hydrogen in piston-cylinder equipment can affect O-isotope fractionations.
46.6.2 Diffusion of hydrogen
The diffusion of hydrogen through metal capsules and its isotopic implications are discussed in some detail in the appendix to Graham et al. (1980). The permeability of precious metals to hydrogen have been measured at 2 kb and elevated temperatures by Chou (1986). Permeabilities increase in the sequence Au, Pt, Ag70Pd30. However,
Experimental Measurement of Isotopic Fractionation Factors ...
1007
Table 46.1 - Summary comparison of experimental devices for isotopic studies and their principal characteristics* Vacuum to Atmospheric Pressures#
Cold-seal Vessels
Range of T: Precision, T Accuracy, T
<3000~ 0.5-15~ 0.5-30~
<950~ <+10~ <+30~
Range of P: Precision, P Accuracy, P
0-400 bar 1% 1%
Sample container:
Internally Heated Vessels
PistonCylinder Devices
<600~ <+10~ +10~
<1500~ +2~ +5~
<1800~ +10~ +25~
<1-12 kbar <1% <1%
<2 kbar 0.02% 0.1%
<10 kbar 0.1% 0.1%
5-60 kbar <+3 <+5
glass, silica Pt, Au
Au, Pt Pd-Ag alloy
Au, Pt or Ti bag
Au, Pt, Pd-Ag alloy
Pt
Sample volume:
large
0.1-0.5 ml
250-103ml
0.6-4 ml
0.01-0.1 ml
Quench rate:
variable
variable, 150~ / min
slow; see text
water Ar, N2
water,
Pressure medium
Hydrothermal Rocking Bombs
variable, 60-600~ / min (100~ / sec)
very rapid, 100~ / sec
N2
halite, talc, barite P cells minutes to days
ar
Run times:
minutes to months
hours to months
weeks to months
hours to months
Remarks:
controlled atmosphere possible
rapid quench models
can sample liquid & vapour
controlled atmosphere possible
Example:
inorganic, mineralorganic and fluid biological H, C, O, S systems H, B, C, N, O, S
aqueous S systems, phase separation, H,O,S
mineral-, magmafluid H, C, O
mineralfluid H, C, O
Advantages:
relatively easy to operate
direct sampling of fluid phases
large volume
effect of P on kinetics
Disadvantages:
large T gradients, even with filler rods
start and end of run takes I to 3 hours
relatively complicated, often slow quench, safety
large T gradients, limited volume
Table 46.1 continued >
1008
Chapter 46 - S.M.F. Sheppard
> Table 46.1 continued * Data from Edgar (1973), Graham et al. (1980), Holloway and Wood (1988), Berndt et al. (1996). # Covers a very wide range of devices such as glass, silica and plastic flasks, tubes and boxes, controlled atmosphere furnaces, etc.., that operate from vacuum to slightly above atmospheric pressures (see Tables 46.2-5). w Note that as the temperature-time quenching curve is exponential, the rate given is no more that a very approximate estimated value for the high temperature part of the cooling curve.
their differences among these metals are quite sensitive to temperature. From a practical point of view, mass transfer of hydrogen, as protium (1H) or deuterium (D - 2H), is minimized by using thick walled (0.5 mm) as opposed to more normal walled (0.2-0.1 mm) noble metal (Au, Pt, Ag) or alloy (e.g. Ag70Pd30; Ag40Pd60) tubing. Wall thickness can be more critical than the specific metal used (see Figure 5 in Graham et al., 1980). If possible, the pressure medium for cold-seal vessels should be water with an isotopic composition close to that of the charge. This water, whose mass is substantial, can buffer the D / H ratio of the external system. However, some laboratories use argon, even though it is only essential for pressures above 8 kbar. (Argon or nitrogen is always employed in internally heated gas vessels; Table 46.1). Vessels using argon as the pressure medium can behave as if hydrogen moves around (in or out of the capsule) quite freely. For face centred cubic metals, that include the noble metals and their alloys amongst themselves, deuterium can diffuse more rapidly through them than hydrogen below a certain critical temperature (~ 500~ for Pd), i.e., there is major deviation from classical rate theory (V61kl & Alefeld, 1975). Whether hydrogen (H and/or D) diffuses into or out of the sealed capsule depends upon the run conditions, mass of hydrogen in the various reservoirs and the nature of the environments inside and outside of the capsule (hydrogen fugacity gradients, etc...). Free hydrogen in capsules can even be lost in minutes or hours, even at room temperature (V61kl & Alefeld, 1975). There is perhaps some confusion about hydrogen diffusion through noble metals. In part, this may have arisen from misunderstandings of the double capsule oxygen buffer technique using an assemblage of solids as developed by Eugster (1957). In its simplest form, the material of the charge is sealed in a thin-walled Pt capsule that is surrounded by the solid buffer assemblage and the whole is sealed within an Au capsule. In the presence of water, the f O 2 in the inner capsules is controlled by the buffer assemblage held between the capsules via diffusion of hydrogen through the semipermeable Pt. Hydrogen also diffuses through the outer Au capsule but typically at a slower rate, because the thickness of its walls is usually greater than that of the Pt. The length of run is chosen so that the buffer capacity of the outer capsule is not exhausted, giving the impression, on quick inspection, that the system as a whole is closed to hydrogen. The permeability of precious metals to hydrogen has also been discussed by Chou (1987). Hydrogen can infiltrate from the packing materials into the sample capsule during piston-cylinder experiments. Rosenbaum & Slagel (1995) have described glass and witherite packing materials that can maintain a low fH2 around the capsule, thus pre-
Experimental Measurement of Isotopic Fractionation Factors ...
venting the formation of CO and H20 from
CO2
1009
inside the capsule.
Hydrogen can also diffuse through non-metals. For example, Vennemann & O'Neil (1996) showed that diffusive loss of hydrogen prevented them using silica glass containers at temperatures above 300~
46.6.3 Effect of pressure
Based on theoretical and experimental evidence (Clayton et al., 1975), equilibrium O-isotope fractionation factors have usually been considered to be independent of pressure for crustal conditions (< 2 GPa) because molar volume changes during isotopic substitution are small compared with the analytical uncertainties. More recent theoretical and spectroscopic studies (Polyakov & Kharlashina, 1994; Gillet et al., 1996; Polyakov, 1998) suggest that pressure effects on isotopic fractionations may be nonnegligible at high pressures (< 3 GPa) and temperatures (< 1200K), particularly for highly anharmonic minerals like calcite and quartz. Further experimentation is called for at high pressures. For hydrogen, pressure dependence has been proposed under usually rather special conditions (Clayton et al., 1975; Sheppard, 1981; Blamart et al., 1986; Grinenko et al., 1987; Mineev & Grinenko, 1996; Driesner, 1997; Horita et al., 2002). Preliminary measurements of the H-isotope fractionation between albite melt and water at 3, 5 and 8 kb by Blamart et al. (1986) indicated that the fractionation varies with pressure. This is not too unexpected because the solubility of H20 increases significantly with increase in pressure (Burnham & Jahns, 1962), and Stolper (1982) and Silver & Stolper (1989) have shown that the proportions of the different species of hydrogen in the melt (molecular H20 and/or OH groups) change with pressure or concentration of hydrogen. Recently, both theoretical (Driesner, 1997) and experimental studies (Horita et al., 1999; Driesner & Seward, 2000) have demonstrated that hydrogen isotope fractionations for mineral-water systems are a function of pressure, particularly around the critical point of water. To give an idea of the importance of this pressure effect, Figure 46.4 summarizes the effect of pressure on the brucite-water D / H fractionation. For example, at 380~ the brucite-water fractionation changes from- 32%o at 15 MPa t o 20%0 at 800 MPa. The fractionation increases linearly with increase in the density of water from 0.070 to 1.035 gcm-3. Although the direction of the change measured experimentally agrees with the theoretical models, there are still a number of discrepancies that need to be resolved by future studies. Additionally, Clayton et al. (1975), Yund & Anderson (1978), Matsuhisa et al. (1979), Matthews et al. (1983a), and Cole & Chakraborty (2001) have shown experimentally that increasing pressure can increase the rate of isotopic exchange in some mineral-water systems (Ewald, 1985). Similarly, changing the chemical composition of the fluid can modify the effect of pressure on the rate of exchange in mineral-fluid systems and help to elucidate the reaction mechanism (Cole, 1992; Cole & Chakraborty, 2001). This increase in rate of exchange with pressure explains, in part, the importance of piston-cylinder techniques to experimental isotope geochemistry.
1010
Chapter 46 - S.M.F. Sheppard
Figure 46.4 - Experimental determination of the D / H fractionation between brucite and water as a function of pressure and temperature. Isochores (lines of constant density) with density values for pure water and the liquid-vapour boundary for water which ends at the critical point (CPt: 22.0 MPa and 374~ are given for reference. Modified after Horita et al. (1999)(see also Fig. 7 in Chacko et al. (2001)).
46.6.4 Starting-materials The selection and characterization of starting-materials are key elements of any experimental study and the outcome of the experiment can even be influenced by their choice. This section can only discuss a few selected aspects because the subject is so very large, reflecting the wide variety of both isotopic problems studied and experimental designs that have been used. Experimental petrologists and geochemists have discussed the importance of the selection, preparation and characterization of their starting-materials (e.g., Edgar, 1973; Holloway & Wood, 1988, and references therein), and this section assumes that the reader has a general familiarity with this subject. The above references also give 'standard' recipes for the preparation of gels, glasses, oxide mixes, etc., and these are not repeated here. 46.6.4.1 Inorganic systems The nature of the problem to be investigated and the type of equipment available dictates, in large part, the choice of starting-materials and mass of material required. Note that unfortunately, the mineralogy and/or chemistry of starting materials have occasionally not been characterized with the precision necessary for the problem or
Experimental Measurementof Isotopic Fractionation Factors ...
1011
the subsequent exploitation of the results. For solid starting-materials which are not in chemical equilibrium, it must be anticipated that isotopic exchange most probably will involve solution reprecipitation mechanisms that may pass by metastable phases (the production of metastable phases can be experimentally very reproducible). Similarly, high-entropy starting-materials, such as gels, glasses, amorphous silica or cristobalite within the stability field of quartz, etc..., are more reactive but usually lead to metastable phases that may never arrive at either isotopic or even stable phase equilibrium during convenient laboratory time scales. For example, Matsuhisa et al. (1978) showed that if cristobalite rather than quartz was used as the starting material in otherwise identical hydrothermal exchange experiments, final quartz-water fractionations at 250~ differed by about 3%0. Note also that laboratory runs are very rarely carried out for periods longer than a few months, and quite exceptionally for up to 10 years (Vitali et al., 2001).
46.6.4.2. Biological material The selection and cultivation of living material for experimental studies is too large and complex a subject to be discussed here. Only a few guidelines are given to help the reader appreciate certain problems that are specific to biological materials. Although isotopic experiments using bacteria have been practiced in the laboratory for a very long time (Thode et al., 1951; Table 46.2, N ~ 1), the identification, choice and cultivation of bacteria are specialized domains that are currently undergoing rapid development (e.g. Madigan et al., 2000). Recent reviews of the geochemical aspects of microbiology with implications for experimental studies include Nealson (1997), Banfield & Nealson (1997) and Canfield (2001a), and see papers in Lyons et al. (2003). Table 2 in An & Friedman (2000) lists common sources for providing microorganisms. A selection of more specific references is given in Table 46.2. A wide variety of invertebrates have been cultivated in aquaria under controlled conditions. Although certain biominerals apparently are in or are quite close to Cand/or O-isotope equilibrium with the free open sea or fresh water reservoir (e.g., brachiopods; Lowenstam, 1961; Carpenter & Lohmann, 1995) or some foraminifera (Emiliani, 1955; Shackleton, 1974; Erez & Luz, 1983)), in detail the situation may be quite complex based on laboratory and field experiments and analyses of fossils (e.g., Fairbanks et al., 1982; Erez & Luz, 1983; Grossman & Ku, 1986; Wefer & Berger, 1991; Garcia et a1.,1997; Bijma et al., 1998; Mulitza et al., 1998). This is perhaps not too surprising when one recalls that the biomineral is not precipitated directly from the nearby sea or fresh water reservoir but, for example for mollusca, from the extrapallial fluid that is physically isolated from the external environment by the periostracum, the shell itself, and the mantle epithelium (Wilbur & Saleuddin, 1983). Intraspecific variations, or variations among two or more adult individuals belonging to the same species, can be larger than the analytical precision. Interspecific variations, or variations among two or more different associated species can also occur. In addition, variations can occur during ontogeny, or growth of the shell, or between different layers of the same shell. Similarly, both interspecific and intraspecific hydrogen, carbon and oxygen isotope variations among plants have been observed (Ramesh et al., 1989;
Chapter 46 - S.M.F. Sheppard
1012
Table 46.2 - Summary of selected biological laboratory experiments carried out near atmospheric pressures Ref. #
System
Experimental set-up
Microbial mediated reactions ~
Reduction of sulphate and sulphite to sulphide.
Reduction of sulphate or sulphite solutions by Desulphovibrio desutphuricans from 11 -46~ in glass flask, under anaerobic conditions,
Reduction of sulphate to sulphide in organic-rich sediments,
Reduction of sulphate by natural populations of sulphate-reducing bacteria (cyanobacterial mats). Fractionations measured at different depths, sulphate concentrations and rates, and incubation temperatures (using gas-tight plastic bags). S~ also measured
(34S/ 32S)
~
(34S/32S) o
.
Reduction of thiosulfate and sulfite.
(34S/32S)
Fractionations measured during reduction and disproportionation of $2032- and SO32- by bacterial enrichments and pure cultures from marine and freshwater environments.
Disproportionation of elemental sulphur
Pure culture of Desulfocapsa thiozymogenes and enriched culture Kuhgraben, both freshwater, disproportionate S~ to sulphate and sulphide
(180 / 160, 34S/ 32S)
o
in closed bottles. FeCO3 and FeOOH used as scavengers for H2S.
Dissimilatory reduction of Se oxyanions. (80Se / 76Se)
Three Se oxyanion-respiring bacterial species grown under controlled closed system conditions using two different growth media. Double spike TIMS technique used.
Fermentation of methanol
Anaerobic conversion of methanol to methane using bacteria from Pacific Ocean mud at 30~ CH4 and CO2 fractions of evolved gases were separated before analysis.
Reduction of CO2 to CH4 in presence of H2 gas. (D / H, 13C/ 12C)
Cultures of Methanobacterium strain M.o.H., Methanosarcina barkeri in 200ml of medium solution within 500ml flask, incubated with H2CO2 gas mixture (80/20, v/v) bubbling through medium at 40~ Culture of M. formicicum at 34~ under 3 atm pressure of H2-CO2 using waters with different 8D values. Growth of M. thermoautotrophicum at 65~ in closed vessel under 3 atm of H2-CO2. Gases collected from vessel through serum caps to isotopically monitor depletion of substrate.
Reduction of CO2 to CH4 in presence of H2 gas and acetate.
Growth of Methanobacterium isolated from oil field with acetate in 12 litre Virtis fermentor, gassed at 1.0 to 1.5 1/min, 80:20 mixture of H2CO2 at 37 and 46~ CO2, CH4, lipids and total cell carbon analysed for 613C.
(13C/12C)
o
over range of metabolite concentrations, conditions of growth, electron donors, pH. Liberated H2S removed by N2 flow through solution and then trapped as CdS. See 1, Table 46.3 for equivalent chemical reduction of sulphate.
(13C/ 12C)
Table 46.2 continued >
Experimental Measurement of Isotopic Fractionation Factors ...
1013
> Table 46.2 continued Oxidation of CH4 to CO2. (D / H, 13C/ 12C)
Cultures of non-purified methane oxidizing bacteria grown in 500 ml mineral medium solution within 12 1 flask filled with CH4-air mixture (18-30% CH4; i.e., large gas/liquid ratio) at 11 and 26~ Pressure monitored manometrically. Gas sampled by syringe and monitored chemically (gas chromotrography) and isotopically (H and C). Different cultures gave different behaviour.
10. Aerobic CH4 oxidation and assimilation into biomass. (13C/12C)
Methanotropic bacteria Methylococcus capsulatus and Methylomonas methanica grown on CH4 using RuMP cycle for carbon assimilation. Intermolecular fractionation in individual lipids determined by GCC-IRMS.
11. Oxidation of CH4 in temperate soil. (13C/12C)
Soil subsamples were incubated in 16 1 mason jars. Consumption of CH4 in CH4-doped headspace 'air' above soil monitored under controlled conditions of moisture content, soil type, temperature. IRMS with on-line combustion.
12. Phosphate-water. (180/160)
Replacement of calcite by apatite during the growth of bacteria (mixed cultures) in medium containing either organophosphorus compounds or inorganic phosphate. Water-saturated CO2-free air bubbled through solution medium in glass or polyethylene flask reaction vessel to maintain constant O-isotope composition of growth medium.
13. Siderite-water. (180/160)
Geobacter metallireducens cultured anaerobically using acetate organic substrate and amorphous FeOOH as electron acceptor. Oxidation of acetate to CO2 with concurrent reduction of Fe and precipitation of siderite (18- 40)~
14. Oxidation and reduction of iron. (56Fe/54Fe)
Dissimilatory Fe-reducing bacteria were grown on ferrihydrate, and ferrous iron solutions were converted to magnetite by magnetotactic bacteria under controlled conditions between 4 ~ and 35~
15. Nitrate generated in acid forest soils by nitrification. (15N / 14N,180 / 160)
Incubation experiments carried out on forest floor material, in the dark, in open Nalgene percolation containers (free exchange with atmospheric 02). Irrigation waters monitored over 4 months.
Plants and algae 16. Photosynthetic fixation of CO2. (13C/12C)
Tomato plants grown in "box" under controlled temperature, CO2 concentrations and 613C value, and light conditions.
17. Photosynthetic fixation of CO2. (13C/12C)
Blue-green and green algae were cultured and harvested at various stages of growth on ASP-2 growth medium, at 39~ for the sea and fresh water ones, and < 70~ for the hot spring ones. Table 46.2 continued >
1014
Chapter 46 - S.M.F. Sheppard
> Table 46.2 continued 18. Cellulose-environment
Fractionation between bean cellulose and 6180 of CO2 and 02 of air, and 6180 of soil water, with control of relative humidity of air and temperature. Plants grown under glass bell jar.
19. Fractionation between cellulose-H20-CO2
Growth of wheat in plexiglass box under controlled environmental conditions and 6180 of H20 and CO2. All inlet and outlet flows of H20 and CO2 were monitored and analyzed isotopically.
20. Fractionation by root respiration
Roots were incubated in closed chambers. Fractionation measured by monitoring change in 6180 of air due to partial uptake by roots.
21. Fractionation in plant waters. (D / H, 180 / 160)
Plants grown in covered glass case. Source water, leaf and stem water, and water vapour isotopically monitored. Plant waters extracted by quantitative freeze drying.
22. Leaf transpiration efficiency.
Container grown plants (wheat) were cultured in glasshouse and in field, varying soil water status, vapour pressure deficit and nitrogen nutrition. C-isotope variations related to leaf transpiration efficiency.
23. Root water-leaf watertranspiration water (D / H, 180 / 160)
Wheat was grown in controlled climate boxes where 6D and ~180 of soil water, root water, leaf water and transpiration water vapour could be sampled under steady state conditions.
24. Land plants-environment
Vascular land plants were cultured under controlled environmental conditions (light intensity, temperature, humidity, Pc02, nutrient supply, ...) in growth chambers. 02/CO2 atmospheric ratio varied (21 and 35% 02, 330 ppm CO2).
(180/160)
(180/160)
(180/160)
(13C/12C)
(13C/12C)
25. Fractionations in aquatic plants,
Aquatic vascular plants grown in fresh-water aquarium under controlled lighting at constant temperature (15 ~ 20 ~ and 25~ and isoto(D/H, 13C/12C, 180/160) pic (H, C, O) compositions. Hydrogen refers to non-exchangeable carbon-bound hydrogen of cellulose.
26. Chlorophyll-total biomass (13C / 12C, 15N / 14N)
8 species of phytoplankton were cultured under controlled conditions (temperature, light, medium ..... ). Chlorophyll (purified pheophytin). separated from phytoplankton for sequential isotopic analysis of CO2 and N2.
27. Organic matter-DIC
Stock cultures of coccoliths (Emitiania huxleyi) acclimatized and then grown under varying CO2 concentrations. 613Cof particulate organic carbon and the fatty acids, phytol, sterols and alkenones analysed.
28. Organic matter-water (D/H)
Microalgae (18 strains, blue-green, green, diatoms) cultured under controlled conditions (light intensity and wavelength, temperature, nutrient availability water 6D) in test tubes in controlled water bath or glass water-jacketedgrowth chambers. 6D measured on total organic hydrogen, lipids, proteins or carbohydrates.
(13C/12C)
Table 46.2 continued >
Experimental Measurement of Isotopic Fractionation Factors ...
1015
> Table 46.2 continued 29. Diatomaceous silicawater. (180/160)
Freshwater diatoms (Stephanodiscus niagarae, Aulacoseira subarctica) were cultured at 3.6 - 20~ in waters with 2 different 6180 values in polystyrene bottles. Technique developed to dehydrate opaline frustules before isotopic analysis.
30. Marine diatom-water DIC (13C/12C)
Marine diatom (Phaeodactytum tricornutum) grown at 22~ in nitratelimited continuous culture growth chamber. Pco2 controlled. Concentration and 613C of DIC measured during experiment and 613C diatom at end.
31. Marine diatoms-sea water (dissolved silicic acid) (30Si / 28Si)
Marine diatoms (Skeletonema costatum, Thalassiosira weissflogii, Thalassiosira sp.) grown in batch culture at 12 - 22~ in model seawater medium (F/2 medium with extra Si(OH)4).
Animals 32. Fractionation between Animals raised in cages or water bath with controlled diet. Gut and animal (total body, chitin, digestive tracts purged or removed to avoid contamination. Respired collagen) and its diet. CO2 collected in glass line purged of atmospheric CO2. (13C / 12C, 15N / 14N) 33. Seawater-foraminifera (11B/10B)
Foraminifera (Orbulina univera) were cultured in seawater at different pH (7.7 to 9.0) contained in closed bottles to determine the pH dependence of the B-isotope fractionation between seawater and foraminifera. Experiment terminated before gametogenisis (i.e., before deposition of gametogenic calcite). 611B measured by NTIMS.
34. Seawater-carbonate (44Ca/40Ca)
Coccolithophorid, Emiliania huxleyi, cultured in seawater at 16~ 644Ca measured by SIMS
# References: 1. Harrison & Thode (1958), Kaplan & Rittenberg (1964), Kemp & Thode (1968), see also Goldhaber & Kaplan (1974); 2. Habicht & Canfield (1997), Canfield (2001b); 3. Habicht et al., 1998; 4. B6ttcher et al. (2001); 5. Herbel et al. (2000) ; 6. Rosenfeld & Silverman (1959); 7. Games et al. (1978), Balabane et al. (1987); 8. Belyaev et al. (1983); 9. Coleman et al. (1981); 10. Summons et al. (1994); 11. Tyler et al. (1994); 12. Blake et al. (1997); 13. Mortimer & Coleman (1997); 14. Beard et aI., (1999), Mandernack et al. (1999); 15. Mayer et al. (2001); 16. Park & Epstein (1960); 17. Pardue et al. (1976); 18. Ferhi & Letolle (1979); 19. DeNiro & Epstein (1979); 20. Angert & Luz (2001); 21. Cooper et al. (1991); 22. Condon et al. (1992); 23. Bariac et al. (1991), see also DeNiro & Cooper (1989); 24. Berner et al. (2000); 25. DeNiro & Epstein (1981a); 26. Sachs et al. (1999); 27. Riebesell et al. (2000); 28. Estep & Hoering (1980, 1981); 29. Brandriss et al. (1998); 30. Laws et al. (1995), Berner et al. (2000), see also Burkhardt et al. (1999); 31. De La Rocha et al. (1997); 32. DeNiro & Epstein (1978, 1981b); 33. Sanyal et al. (1996); 34. De La Rocha & DePaolo (2000).
Saurer et al., 1997b; Heaton, 1999; Tang et al., 2000). These remarks emphasize the importance of carrying out a number of tests in order to develop a viable sampling strategy.
1016 46.6.4.3
Chapter 46 - S.M.F. Sheppard
Preparation
Preparing gases or solutions of known chemistry and isotopic compositions usually does not raise major problems. Specific isotopic compositions can be derived by mixing either commercially available or natural (e.g., different meteoric waters) products. Two cautionary notes. (1) Having hydrogen gas with a known and constant 6D value can be more complicated. Hydrogen gas is usually composed of H2, HD and D2. Although the three molecules are related by H2 + D2 = 2HD
[46.11]
exchange equilibrium is not necessarily readily attained in the absence of a catalyst. For reaction [46.11], K is in the range of about 3.2 to 3.8 in the temperature range of 20 to 500~ (Richet et al., 1977). The IRMS measures the ratio mass 3/mass 2 or, after correction for H3+, the ratio H D / H H . Hence the measured ~SDdoes not take into account the presence of deuterium as D2. If the hydrogen gas contains some D2, even in trace quantities, the 6D value of the gas can change during the experiment simply through intermolecular hydrogen exchange reactions (reaction [46.11]). This effect was invoked by Vennemann & O'Neil (1996) to explain variations in the 15D value of their hydrogen gas. (2) Substances that do not contain the element of interest are sometimes added to water of known isotopic composition to increase the rate of exchange (e.g., NH4C1 or NaF; O'Neil et al., 1969; Clayton et al., 1972). In both these studies, the isotopic effect of the added solutes was shown to be negligible. Under certain conditions and concentrations, the isotopic behaviour of the aqueous solution is modified by the presence of such salts, the solute interaction effect (Taube, 1954; Sofer & Gat, 1972, 1975; Truesdell, 1974; Stewart & Friedman, 1975; Graham & Sheppard, 1980; O'Neil & Truesdell, 1991; Kakiuchi, 1994; Poulson & Schoonen, 1994; Horita et al., 1995; Shmulovich et al., 1999; Driesner & Seward, 2000; Chacko et al., 2001 and references therein). Water and aqueous solution are thus not necessarily isotopically equivalent for either hydrogen or oxygen, and this must be taken into consideration. The preparation of solids or minerals can be more complicated and time consuming. Ideally minerals should be both physically and chemically homogeneous, of known chemistry and mineralogy (e.g. aragonite versus calcite, or low-Mg calcite versus higher-Mg calcite) and well characterized with respect to their grain size or range of sizes, surface roughness, etc... In some high pressure experiments, the grain size may change as pressure is applied to the container during its collapse around the sample. The purity and grain size of clay minerals raise special problems (see Gilg et al., Part 1, Chapter 2). Scanning electron microscopy (SEM) is now widely used to examine textural and morphological features of both starting materials and run products; excellent examples of photomicrographs are given in Matsuhisa et al. (1978), Matthews et al. (1983a, b), Chacko et al. (1991) and Rosenbaum (1994). Surface areas of materials can be determined from SEM observations and krypton adsorption apparatus measurements. Note that Rosenbaum (1994) has shown that these two techniques give satisfactory agreement on synthetic calcite but the adsorption technique gave surface areas ~6 times larger than the geometric estimate from SEM because of the presence of unannealed fractures. Volatile containing minerals should be analyzed
Experimental Measurement of Isotopic Fractionation Factors ...
1017
before and after the reaction, and not assumed to be stoichiometric. For example, in certain hydrogen isotope experiments, the hydrogen content of the mineral has been shown to change significantly during the so called "simple" exchange reaction (Graham et al., 1984; Vennemann & O'Neil, 1996). Similarly for hydrogen isotope experiments, FeII and FelII contents should be determined before and after exchange. For minerals with complex structures, such as amphiboles, this is easier said than done because, for example, Mossbauer spectra of amphiboles cannot readily be interpreted in terms of FeII and FelII. Although mineral-water H-isotope fractionations are very sensitive to the Fell/Mg ratio of the mineral (Suzuoki & Epstein, 1976), relatively little is known about the effect of changing the FelI/FelII (or TiIII/TiTM etc...) ratio on the fractionation factor. Iron in smectite occurs principally as FeIII. For smectite-water fractionations, the fractionation increases with increase in the ratio octahedral FelII/ (FelII + Mg + A1)(see Figure. 5 in Sheppard & Gilg, 1996). These data do not follow the atomic mass/charge relationship of Suzuoki & Epstein (1976), that was derived principally from studies of di- and tri-octahedral micas where Fe is dominantly divalent. With the development of microanalytical techniques and small sample volumes of, for example, piston-cylinder capsules, high precision balances are needed. The working environment of such balances must be taken into consideration because they are significantly more sensitive to vibrations and small temperature variations.
46.6.5 Run-products Examination of run-products before isotopic analyses can provide crucial information on the experimental system and interpretation of the data. The relationship between the starting-materials and the run-products depend on several factors other than the design of the experiment. These include one or more of: evidence for chemical reaction, recrystallization, Ostwald ripening (see section 46.5), congruent or incongruent dissolution, presence or absence of quench phases that form during the relatively sudden changes of P-T conditions at the end of the experiment, chemical or isotopic zoning, change of grain size, loss or gain of mass, .... There is a range of techniques that can be applied such as optical microscopy, often in conjunction with refractive index oils, X-ray diffraction, and the several microbeam techniques - electron microprobe, transmission electron microscopy (TEM), scanning electron microscopy (SEM), micro-Raman, micro-infrared, ion microprobe, etc... The choice of technique(s) is dictated by the problem in hand and availability of or access to equipment. It is important to appreciate the sample requirments and limitations of the selected technique (e.g., X-ray diffraction methods do not detect phases that are present in less than about 5 per cent). For all experiments on mineral-fluid systems run under temperature conditions quite different from room temperature, the choice of fluid/mineral ratio can determine the importance of both dissolution processes, that are often incongruent, and quench products. If fluid/solid ratios are high, the chemistry of the starting solids may be substantially modified. Crystals formed during quenching may have the same mineralogy as the principal run products but be of different texture or size, usually smaller, and isotopic composition. Physical separation may be possible. However,
1018
Chapter 46 - S.M.F. Sheppard
Table 46.3 - Summary of selected inorganic experiments carried out in glass and plastic flasks at around atmospheric pressures. Ref .# ,
.
.
.
~
,
System
Experimental set-up
Reduction of sulphate solution to sulphide. (34S/ 32S)
Chemical reduction of sulphate solutions by a mixture of hydriodic (HI), hypophosphorus (H3PO2) and hydrochloric (HC1) acids at 18 50~ Oxygen-free nitrogen flushed through solution in glass reaction vessel to remove H2S that was then trapped as CdS.
Reduction of selenate to selenite. (82Se / 76Se or 80Se/ 76Se)
Chemical reduction of selenate (Se6+) to selenite (Se4+) in hot HC1. Selenite coprecipitated on to ferric hydroxide. Isotopic ratio 80Se/ 76Se measured by SIMS and 82Se/76Se by IRMS.
Oxidation of sulphide to sulphate and thiosulphate (34S/32S)
Oxidation of sodium sulphide solution carried out in closed bottle under slight positive pure 02 pressure at 22 - 25~ and pH 10.8 - 11. Sulphite also observed.
CO2(g)-CO2(aq)-HCO3(13C/12C)
Equilibration between dissolved bicarbonate and CO2 gas in 100ml flask at 5 - 25~ Some measurements to 125~
HCO3--CO3--CO2(g) (13C/12C)
Equilibrium fractionation between DIC (CO3 = + HCO3- + CO2(aq)) and CO2 gas in H20 (with Na2CO3, NaHCO3 or Na2CO3 + NaHCO3) and seawater at 4-80~ Experiments conducted in closed-system.
Air-water CO2 exchange (13C/12C)
Kinetic fractionation during the exchange of CO2 between air and sea-water determined from air to sea and sea to air in a simulator at 5, 15 and 21~
Diffusion of CO2 in (13C/12C)
Diffusion of CO2 carried out in a Stokes-type diaphragm cell at 25~ aqueous solution.
Solute-water interactions. (180/160)
Equilibration of CO2 with various aqueous solutions at 10 - 95~ *.
Liquid-vapour fractionations: water and salt solutions. (D/H, 180/160)
Aqueous liquid-vapour equilibrium and effect of various salts and their concentration from 0-100~ (for higher temperature experiments, see Table 46.4).
10. Hydrated saline mineralbrine. (D / H, 180 / 160)
Precipitation of hydrated minerals (MgC12.6H20; CaSO4.2H20; KMgC13.6H20; etc...) from supersaturated solutions at 10 - 40~
Table 46.3 continued >
1019
Experimental Measurement of Isotopic Fractionation Factors ... Table 46.3 continued 11. Goethites-water (D/H, 180/160)
Various methods of closed system synthesis of ferric oxides at 25 140~ Iron oxides precipitated from Fe(NO3)2 or FeC13 solutions (1) with NaOH or KOH, (2) with NaOH + HCO3-, and (3) hydrolysis. Exchange experiments using ferrihydrite gel. Importance of acid washing and vacuum drying of fine-grained precipitates and aging of initial precipitate.
12. Amino acid transamination (15N/14N)
Kinetic isotope effect measured during the transfer of amino nitrogen from glutamic acid to oxaloacetate to form aspartic acid. Transamination reactions catalyzed by glutamic oxalacetic transaminase. Fractionation measured for reverse reaction (transferring NH2 from aspartic acid to a-ketoglutarate). Percentage completion of reaction analyzed by monitoring appearance or disappearance of amino acids by high performance liquid chromatography.
13. Marine clay-seawater (liB/lOB)
Variation of adsorption constant and B-isotope fractionation measured as function of pH and temperature between natural sediments, prepared as a sediment slurry desorbed of B, and seawater contained in glass vials. ~11B measured by TIMS.
14. Halite-solution (lIB/lOB)
Evaporation of artificial and natural brines at 30~ examined.
15. Salt-solution
Chlorides of Na, K and Mg were precipitated from saturated solutions at 22~ by evaporation in glass beakers. Mineral-saturated solution fractionation measured.
(37C1/ 35C1)
Effect of Ca
16. Ion-exchange (65Cu / 63Cu, 68Zn / 66Zn) (40Ca/44Ca)
Solutions of Cu and Zn, in HC1 or HNO3 medium, were eluted on anion- exchange resin and isotopic ratios of eluted fractions measured by MC-ICP-MS. Ca isotope effects were measured by SIMS.
17. FelI-FeIII complexes in solution (57Fe / 54Fe, 56Fe / 54Fe)
Fractionation between tris(2,2'-bipyridine) iron-II ([FelI(bipy)3] 2+ and FeIII chloride complexes measured by MC-ICP-MS.
# 1. Harrison & Thode (1957), see also Grinenko et al. (1969); 2. Johnson et al. (1999), Johnson & Bullen (2003), Krouse & Thode (1962), Rees & Thode (1966); 3. Fry et al. (1988); 4. Mook et al. (1974); 5. Zhang J. et al. (1995), Halas et al. (1997, 2000); 6. Inoue & Sugimura (1985), Zhang J. et al. (1995); 7. O'Leary (1984); 8. Taube (1954), Truesdell (1974), Stewart & Friedman (1975), Kakiuchi (1994), O'Neil & Truesdell (1991); 9. Majoube (1971), Horita et al. (1993a, b), Driesner & Seward (2000), Bourg et al. (2001); 10. Horita (1989b); 11. Yapp (1987,1990), Bao & Koch (1999); 12. Macko et al. (1986); 13. Palmer et al. (1987), see also Schwarcz et al. (1969), Hemming & Hansen (1992), Xiao & Wang (2001); 14. Liu et al. (2000); 15. Eggenkamp et al. (1995); 16. Mar6chal & Albar6de (2002), Russel et al. (1978); 17. Matthews et al. (2001). * Note that the isotopic activity of H2180 that is measured is different from the concentration of H2180 in the solution (Sofer & Gat, 1972).
1020
Chapter 46 - S.M.F. Sheppard
Table 46.4 - Summary of selected sealed tube experimental systems carried out around or below atmospheric pressures. Ref. # .
o
~
~
o
.
System
Experimental set-upw
Dolomite-calcite-CO2 (13C / 12C, 180 / 160)
Dry CO2, with different 6-values, loaded in Ni vessels, equiped with glass stopcock, containing carbonate. CO2 oxygen amounted to only about 0.2% of total oxygen. Temperature range 350 - 610~
Self-diffusion of carbon and oxygen in calcite (13C / 12C, 180 / 160)
Exchange between limited reservoir of CO2 and annealed and nonannealed calcite of known grain size in Vycor reaction vessels at 650 850~ Progress of diffusion measured by change in composition of gas. Specific surface area and self-diffusion rate of C determined by exchange with 14C-labelled CO2 at 250 - 750~ Exchange between CO2 and Vycor measured.
CO2-carbonate surface (13C / 12C, 180 / 160)
Determination of the surface effect on the bulk CO2-carbonate fractionation. Coarse grains of calcite, dolomite or witherite in a nickel tube were powdered (N40 tLm) under vacuum with a steel ball. Aliquots of powdered carbonate were exchanged at 200~ with 3 isotopically different CO2 gases for various times.
Liquid CO2-water (180/160)
CO2 and H20 sealed in Pt tubes. Exchange at 25.3~ Complete exchange only achieved after about 3 days. Possible effect of CO2 clathrate hydrate formation on fractionation discussed.
CO2 ice / vapor (13C / 12C, 180 / 160)
Condensation and sublimation of CO2 at 130 - 170 K and 5 - 10 mbar. Influence of geometry of reservoir examined.
Quartz-CO2 (180/160)
Quartz sealed in Au capsules with dry CO2 (98% 180) and placed in silica tube loaded with dry Ar and sealed so that at run temperature of 888~ P in and outside of Au capsule was 0.6 bar. Study of diffusion using ion probe.
Albite-CO2 Albitic glass-CO2 Rhyolitic glass-CO2 (180/160)
Silicate powder and dry CO2 sealed in Pt capsule and placed in wirewound horizontal furnace (650 - 950~ Silicate-oxygen/CO2-oxygen ratio very high. Pressure 0.8 - 1.2 bars. Rates of oxygen isotope exchange measured to derive diffusion coefficient and mechanism of exchange.
Oxygen diffusion in sanidine (180 / 160)
Isotopically enriched 02 (75 atom % 180) exchanged with anhydrous sanidine in horizontally mounted quartz tube at I bar, 869 - 1053~ Oxygen isotope diffusion profiles measured in direction perpendicular to (001) with ion microprobe.
Zeolite-water vapour (180/160)
Framework oxygen of zeolites exchanged with water vapour at known PH20 in silica glass tubes at 23 < T < 500~ Table 46.4 continued >
Experimental Measurement of Isotopic Fractionation Factors ...
1021
> Table 46.4 continued 10. SO2-H2S
Sulphur and H 2 0 were sealed in a silica glass bulb under vacuum and heated to 500 - 1000~ to exploit reaction: 3S + 2H20 = 2H2S + SO2. After reaction, H20, H2S and SO2 separated by trap to trap vacuum distillation. At each temperature, one of two runs overshot final T by 100~ for 24 hr to check attainment of equilibrium.
11. Kaolinite-water (D/H)
Kaolinite sealed in glass tubes with isotopically different H20 (few % of total hydrogen) and run in furnace at 200 - 352~ for up to 4 months.
12. Silicate glasses-water vapour (D/H)
Rhyolitic and feldspathic glasses and melts were sealed in quartzvycor tubes with large excess of water and run at 530 - 850~ PH20 = 1.4 - 2.8 bars, < 54000 hr. Vapour-dissolved hydrogen fractionation factor determined. Reversal experiments indicate equilibrium was attained.
13. Hydrous mineral-H2 (D/H)
Exchange between molecular hydrogen and mineral (epidote, kaolinite, muscovite, biotite, hornblende) at 150-400~ in silica tubes. Run with either "infinite" reservoir of H2 or "infinite" reservoir of mineral. Small change in 8D of gases during run probably due to exchange with trace D2 present in H2-HD gas (see section 46.6.4.3). Diffusive loss of hydrogen through silica glass became significant for T > 300~
14. CH4-H2 (D/H)
CH4-H2 mixture at 770- 930 m m Hg, circulated in closed system over a Ni-Thoria catalyst at 200 - 500~ After sampling exchanged mixture (in-line sampling volume with no contribution of H2 adsorbed on catalyst) gas-chromatographic separation of H2 from CH4 before IRMS to determine equilibrium fractionation.
15. CH4-H2-H20 (D/H)
Gas mixtures (CD4-H2 and D20-H2) reacted in closed glass tube at 100 - 500~ Most experiments at P < i bar. Rates of reaction monitored from formation of HD after cryogenic separation of gas mixture including CD4-H2.
16. Fischer-Tropsch synthesis
Equimolar mixture of CO and H2 (0.8 1), initially at I atm, heated to 400~ in Vycor flask in presence of Co catalyst. CO2 removed continuously during reaction. Kinetic fractionation measured between CO2 and organics.
17. CO2-CH4 exchange
Exchange and equilibration between CO2-CH4 in pyrex tube along with varying amounts (10 - 50 mg) of catalytic material (iron oxides and commercial transition-metal (Ni, Pd-Pt, Rh) catalysts), at 200 600~
(180/160)
(13C/12C)
(13C/12C)
Table 46.4 continued >
1022
Chapter 46- S.M.F. Sheppard
> Table 46.4 continued 18. 03-02 (170 / 160, 180 / 160)
Synthesis of ozone from molecular oxygen using an electric discharge (Tesla coil) or microwave plasma. Effect of oxygen pressure measured. Ozone trapped cryogenically during its production. 6170 and 6180 measured to unravel mass-dependent and mass-independent fractionations.
# 1. O'Neil & Epstein (1966a); 2. Anderson (1968, 1969); 3. Hamza & Broecker (1974); 4. Rosenbaum (1993); 5. Eiler et al. (2000); 6. Sharp et al. (1991); 7. Matthews et al. (1994), Palin et al. (1996), see also Stolper & Epstein (1991); 8. Derdau et al. (1998), Elphick et al. (1988), Giletti et al. (1978), see also Freer & Dennis (1982), Freer et al. (1997); 9. Feng & Savin (1991, 1993a, 1993b), see also Karlsson & Clayton (1990); 10. Thode et al. (1971); 11. Liu & Epstein (1984); 12. Dobson et al. (1989); 13. Vennemann & O'Neil (1996); 14. Horibe & Craig (1995); 15. L6cluse & Robert (1994); 16. Lancet & Anders (1970); 17. Horita (2001); 18. Heidenreich & Thiemens (1986), Bains-Sahota & Thiemens (1987). w All isotopic measurements made by IRMS except where noted. some quench products may be precipitated as rims on existing crystals and difficult to identify; determination of their mass, which is not necessarily negligible, or their isotopic composition can be very difficult. If the fluid phase contains two constituents such as H 2 0 and CO2, O-isotope exchange can occur between them during and after the quench. In some synthesis experiments, gels and poorly-ordered very fine-grained precipitates may be the initial product. These may undergo isotopic exchange during subsequent aging processes or preparation of the sample for isotopic analysis. Such effects have been documented by Bao & Koch (1999) in ferric oxide-water systems (Table 46.3, N ~ 11).
46.7 The experiment For experiments carried out above room temperature, the working temperature should be approached from below, with no overshooting (the isotopes do not k n o w that they are not meant to start exchanging before arriving at the working temperature!). Note that in most fluid-mineral systems, the surface layer of the solid undergoes isotopic exchange much more rapidly than the bulk of the sample. The physical conditions within the container at any given m o m e n t are intimately related to the nature of the pressure-temperature device and its use. For these reasons, a brief summary of the principle types of equipment and laboratory setups that have been used is given as a complement to the information presented in Table 46.1. The purpose here is to help both the reader to appreciate the advantages and limitations of these techniques from an isotopic point of view, and the potential user to select or design the best system for his situation. This section, which is inevitably selective, will consider (1) experiments close to atmospheric pressures or under vacuum, (2) externally heated vessels - cold-seal and rocking bombs, and (3) internally heated vessels, including solid media apparati.
Experimental Measurement of Isotopic Fractionation Factors ...
1023
46.7.1 Experiments near atmospheric pressure or under vacuum
An extremely wide variety of experimental studies have been carried out below, at or slightly above atmospheric pressure in diverse types of containers. Many of these experiments were carried out near room temperatures, but some required the use of very high temperature furnaces (Table 46.1). Isotope ratios of most elements of interest have been studied. This section is therefore rather heterogeneous as there is no specific P-T device associated with these studies. The experiments can be divided into four groups" (1) systems with biological materials, (2) inorganic experiments carried Table 46.5 - Summary of selected open system experiments carried out at low pressures (< 10MPa, < 100 bar) or under vacuum. Ref. #
System
Experimental set-up
1. Silicate-melt or mineral-CO2 or 02 (180/160)
Silicate sample suspended in 1 atmosphere double-ceramic tube quenching furnace through which CO2 or 02 of constant isotopic composition was passed continuously at 1250- 1525~ Calculated diffusion coefficient and equilibrium fractionation factors
2. Calcite-bicarbonate solution (13C/12C)
Inorganic precipitation of calcite or aragonite in temperature controlled (10 - 40~ water-jacketed reaction vessel by bubbling CO2 N2 gas of known chemical and isotopic composition through Na-CaC1 solutions using synthetic calcite or aragonite seeds. Precipitation rates varied from 102.6 to 104.8tcmolm-2h-1.
3. Evaporation of basalt and chondrites (30Si/28Si)
Basalt and two carbonaceous chondrites were evaporated using a solar furnace.
4. Evaporation of forsterite (26Mg / 24Mg)
Single crystals of forsterite were evaporated at 1500 - 1800~ in a vacuum furnace. Measured solid-gas kinetic fractionation of Mg and derived diffusion coefficients.
5. Evaporation of synthetic "solar" material (26Mg/24Mg, 170/160, 180 / 160, 30Si / 28Si)
A FeO-MgO-SiO2-CaO-A1203-TiO2 rare earth element melt system was evaporated in vacuum furnace at 1800 - 2000~ for few seconds to 5 h. Chemical fractionations monitored along with kinetic isotope effects.
6. Evaporation of solid silica (180 / 160, 170 / 160)
Partial evaporation of silica at 1600 - 1700~ under vacuum (10-9 bars) and in H2 or N2 (> 10-5 bars) atmosphere. Degree of 180 and 170 enrichment of solid controlled by kinetic fractionations at ablating surface, rate of sublimation and efficiency of O-self diffusion in solid.
7. Fractionation of CH4 during diffusion (13C/12C)
Methane diffused at 90~ and 9 MPa through shales with varying organic carbon contents in triaxial flow cell. Fractionation during molecular transport (diffusion) measured.
# 1. Muehlenbachs & Kushiro (1974), see also Appora et al. (2003); 2. Romanek et al. (1992), see also McCrea (1950), Turner (1982); 3. Molini-Velsko et al. (1987); 4. Wang et al. (1999); 5. Wang et al. (2001), see also Richter et al. (2002); 6. Young et al. (1998c); 7. Zhang T. & Krooss (2001).
1024
Chapter 46- S.M.F. Sheppard
out in glass or plastic flasks where the contents in the flask may or may not strictly be a closed system, (3) experiments run in sealed containers, and (4) open system experiments. The division between (2) and (3) is sometimes rather arbitrary. A selection of such experimental set-ups is summarized in Tables 46.2 to 46.5. Because of their large diversity, details of the individual experiments cannot be presented; the references give the necessary entry to their descriptions. A few general observations and remarks can be made.
46.7.1.1 Biological systems (Table 46.2) Most of these experiments were designed to measure kinetic isotope fractionations. Many of the experiments using biological material require a certain specialized knowledge of the biological elements under study and access to particular equipment that is not brought out directly in Table 46.2. The reader is referred to the references for such details on, for example, the methods of bacterial culture, etc... Similarly, isotopic analysis of certain of the starting materials and end products require special techniques (see references, Table 46.2 and Lyons et al., 2003). Much attention has been directed to growing plants under various controlled conditions. The systems are quite complex because, among several things, (1) the element of interest is usually present in different sites and substances with different degrees of exchangeability, (2) tissues or organs from the same plant may exhibit isotopic variations substantially larger than the analytical precision, and (3) isotopic compositions can be a function of a wide variety of variables including temperature, humidity, water availability, light intensity and wavelength, CO2/02 ratio, nutrient supply, etc... (e. g., Farquhar et al., 1989a; Heaton, 1999; Tang et al., 2000).
46.7.1.2 Inorganic systems in more or less closed flasks (Table 46.3) Such experiments in glass or plastic flasks were designed to measure either equilibrium or kinetic fractionations at temperatures that were usually below 100~ The contents of a flask in certain experiments were essentially a closed system. An extremely wide variety of systems have been explored for most elements including B, C1 and Se. Certain, N ~ 1 and 2 of Table 46.3, are chemical reduction reactions which can be considered as the chemical equivalent of biological reduction reactions (Table 46.2, N ~ 1, 2 and 5). In addition, to measure the thermal diffusion constants for 29N2/ 28N2 in air at temperatures from -60 to O~ Grachev & Severinghaus (2003) developed a two-bulb apparatus that fits into a dual temperature bath. 46.7.1.3 Sealed tubes (Table 46.4) These experiments have yielded kinetic, including mass-independent, and equilibrium fractionation data and diffusion data. Tube and fluid volumes were selected so that internal pressures were less than atmospheric to a few atmospheres at the working temperature, to avoid the use of more complicated pressure controlling devices. Quenching of the tube can usually be carried out in seconds to a few minutes, depending on the mass of the tube system. Alternatively, the substance of interest (e.g., 03; Thiemens & Heidenreich, 1983) can be cryogenically removed during the reaction. Note that silica glass becomes inconveniently permeable to H2 gas at tern-
Experimental Measurementof Isotopic FractionationFactors ...
1025
peratures above 300~ (Vennemann & O'Neil, 1996).
46.7.1.4 Open system reactions (Table 46.5) These systems are relatively limited in number but wide in application and aimed at deriving either kinetic or equilibrium fractionations. They include studies on silicate melts at about atmospheric pressures (N ~ 1) or under high vacuum, the latter directed at fractionations of Si, Mg and O under simulated space conditions (N ~ 3, 4, 5 and 6), precipitation of carbonates at variable rates (N ~ 2), and diffusion of gas through water-saturated shales using a flow cell (N ~ 7). 46.7.2 Externally heated pressure-vessels Externally heated pressure-vessels cover a wide variety of steel and nickel-chromium-cobalt-based alloy vessels, with or without pressure control, that are placed within an electric furnace. Two types of autoclaves have principally been used for isotopic studies: cold-seal vessels and hydrothermal rocking bombs. The fugacity of hydrogen within the capsule or container is controlled by the vessel and its pressure medium unless special precautions are taken (see section 46.6.2). The experimental data are principally for H, C, O and S. 46.7.2.1 Cold-seal vessels These refer to Tuttle-type cold-seal vessels (Tuttle, 1949), or a modified form. Some are fitted with a water cooled jacket near the cold-seal end so that by removing the furnace and turning the bomb, attached to its flexible capillary or pressure tubing, into a vertical position, the capsules inside can drop down into the cold zone for very rapid quenching. Cold-seal vessels of one sort or another have been very widely used for isotopic calibration studies of minerals for all elements since the late 50's (Clayton, 1959). Three capsules are often placed in the same bomb so that P-T-t conditions are identical for the set. The Northrop & Clayton (1966) method (see section 46.4.1) can then be applied to derive the best estimated equilibrium fractionation factor and degree of exchange in incomplete exchange runs. Quite large noble metal capsules have been used (e.g., 5 mm diameter and 40 - 70 mm long by Clayton et al., 1972, and 8 mm diameter and 100 mm long by Czamanske & Rye, 1974). Experimental configurations using big capsules may have large temperature variations along their length (~ 50~ for Clayton et al. system) or apparently relatively small estimated variations (3 - 5~ for Czamanske & Rye, 1974). Note that the presence of a vertical temperature gradient was part of the experimental design by Czamanske & Rye to aid dissolution of reactant sulphides (galena and sphalerite) at the hotter lower end of the tube and, after transport of metals and sulphur in solution, co-precipitation of sulphides at the cooler upper end. Although the temperature may be monitored to + 5~ or better, the actual capsule temperature may not be known to better than + 25~ (Table 46.1). The heating-up times to the chosen temperature may be tens of minutes to an hour and the temperature should be approached from below. Quench rates to 100~ or below are highly variable but can be as long as 10 minutes or more for large conventional coldseal bombs; the rapid quench models, which can quench within a minute or so, have rarely been used for isotopic studies. Some examples are (1) for oxygen in mineralwater systems: Clayton (1959), O'Neil & Taylor (1967, 1969), O'Neil et al. (1969), Clay-
1026
Chapter 46- S.M.F. Sheppard
ton et al. (1972), (2) for oxygen and carbon in carbonates: Northrop & Clayton (1966), Sheppard & Schwarcz (1970), (3) for sulphur in sulphides: Czamanski & Rye (1974), and (4) for hydrogen in mineral-water systems: Suzuoki & Epstein (1976), Graham et al. (1980, 1984).
46.7.2.2 Hydrothermal rocking bombs These autoclaves were developed by Barnes (1963) and Dickson et al. (1963), with modifications by Seyfried et al. (1987), to study aqueous sulphur systems, hydrothermal alteration of rocks, mineral solubilities, speciation and kinetic studies where substantial sample volumes were necessary. The working volume is very large by other autoclave standards. The rocking action increases reaction rates. Parameters such as pH can be varied. Note that aliquots of the fluid phase (e.g., 10 ml of solution), liquid and/or gas phase, can be sampled under working conditions. Although the temperature quenching of such a fluid takes place in a few seconds, the fluid may be such a highly reactive, multicomponent solution that chemical separation must be carried out immediately on sampling (e.g., removing sulphide from the solution to reduce back reaction between H2S and SO4 2- or loss of S2- as H2S which takes a few minutes; Sakai & Dickson, 1978). Any solids in the run can only be sampled after quenching the autoclave which can take an hour or so. This is an important technique for sulphur isotope studies. Some examples for sulphur are Robinson (1973), Sakai & Dickson (1978), and see Ohmoto & Lasaga (1982) for evaluation of the data, and for oxygen, Chiba & Sakai (1985). For sulphur isotopes in the dissolved sulphide-sulphate system, fractionation and rate data as a function of pH were obtained by Igumnov (1976) and Igumnov et al. (1977) using large volume, but non-rocking, titanium alloy autoclaves. 46.7.3 Internally heated vessels Internally heated vessels have the furnace within the pressure device. Their main advantage is that they can operate up to both higher pressures and temperatures than externally heated autoclaves (Table 46.1). The experimental data are principally for H, C and O. There are two principal types of apparati used by experimental geochemists: gas vessels and piston-cylinders. 46.7.3.1 Internally heated gas vessel Such a vessel is relatively complicated because both the furnace and thermocouples are inside the thick-walled pressurized vessel whose outer walls are water-cooled to retain its strength. The internal working volume at constant temperature is relatively large as the furnace is usually double or triple so that internal tempertaure gradients can be reduced. The sample, sealed in a metal capsule, can be as large as 150 250 mg. Gases such as CO2 are introduced as oxalic acid (anhydrous, dihydrate) or silver oxalate (Holloway & Wood, 1988) which decompose to CO2, CO and H20. Because of redox reactions in the charge and diffusion of hydrogen through the Pt capsule walls, the CO is largely converted to COa. The pressure medium is argon up to about 12 kbar (argon solidifies at about 13 kbar or 1.3 GPa), or nitrogen. Because the volume of argon or nitrogen under high pressure and high temperature can be substantial and the compressibility of argon or nitrogen is relatively high, an operating vessel is literally a big bomb, so extremely strict safety precautions must be in force
Experimental Measurementof Isotopic Fractionation Factors ...
1027
during the operation of such autoclaves. Large vessels can take several hours to heat up, but smaller versions need only 15 minutes or so. Quench time is often several minutes (Table 46.1) but very rapid quench versions exist (~ 100~ Such devices have been used for mineral-water H- and O-isotope systems (Graham et al., 1980; Karlsson & Clayton, 1990), and albite melt-H20, CO2-basaltic and H20-basaltic magma fractionations (Javoy et al., 1978; Richet et al., 1986; Pineau et al., 1998).
46.7.3.2 Piston-cylinder apparatus A solid medium piston-cylinder apparatus consists of a piston that presses into a cylinder containing the solid materials, capsule and furnace (often graphite) assembly (Boyd & England, 1960). Filling the small capsule with microquantities of substances, including fluids, followed by sealing (arc welding, etc.) requires a certain expertise. Pressure is transmitted to the charge by the solid material(s) such as salt, talc or barium carbonate and it may not be perfectly uniform. Considerable care has to be exercised to derive the accurate value of the pressure from the geometry of the apparatus and corrections made for friction and other effects. Although the size of the capsule is small (Table 46.1), temperature gradients can be important. Corrections have to be made for the effect of P on thermocouples. Since operation up to 30 kbar and 1700~ is relatively straightforward, and Clayton et al. (1975) have demonstrated that increasing pressure increases the rate of O-isotope exchange and recrystallization without modifying the equilibrium fractionation factor, piston-cylinder apparati are now often used for stable isotope studies. Because the capsule volumes are small, two phase mixtures are usually chosen to contain roughly equal proportions of the element of interest and all end products are analyzed isotopically. There may only be sufficient material for a single analysis of each phase with the volatile phase being analyzed by puncturing the capsule with a steel needle under vacuum. Despite the short length of the capsule (few mm), CO2 is added either by freezing in a known mass of CO2 or as oxalic acid dihydrate (giving CO2 and H20 at low fH2 or silver oxalate (giving CO2 only). Since the sample material is contained in a sealed Au or Pt tube that is permeable to hydrogen under the run conditions, the nature of the packing materials around the capsule can influence the end products. Rosenbaum & Slagel (1995) have examined this problem in detail emphasizing the importance of eliminating all traces of water and creating an environment with low hydrogen activity around the capsule to improve experimental reproducibility. Some examples are: (1) for oxygen in mineralH20 systems: Matsuhisa et al. (1978, 1979), Matthews et al. (1983b), (2) for mineralcalcite or BaCO3: Clayton et al. (1989), Chiba et al. (1989), Rosenbaum et al. (1994), Rosenbaum & Mattey (1995), (3) for carbon and oxygen between CO2 and calcite: Chacko et al. (1991), Rosenbaum (1994), (4) for carbon between CO2 and basaltic melt: Mattey (1991), and (5) for hydrogen in mineral-H20 systems" Horita et al. (1999).
46.8 Field experiments Field experiments refer to those conducted under natural or field environmental conditions that are more or less well defined. Such experiments have and will continue to yield crucial fractionation data that often cannot be obtained during laboratory experiments. Although these experiments are not ideal, comparison of such empirical calibrations of fractionation factors with laboratory calibrations of equilib-
1028
Chapter 46 - S.M.F. Sheppard
rium fractionations can, for example, give much insight into certain biological processes. The limitations of this approach are usually rather different from those associated with laboratory experiments and they can be difficult to access. The major weakness is the general lack of unambiguous criteria for equilibrium. A number of field calibrations of fractionation factors give such highly systematic temperature dependent fractionations that they have been interpreted as equilibrium values (see section 46.8.2). It is also possible that some so-called non-equilibrium fractionation factors represent equilibrium for a reaction that has not been clearly identified. For example, certain biogenic carbonate fractionations only consider the fractionation between the biogenic mineral and open sea or fresh water rather than the body fluids that were actively involved in the precipitation process. If the pH of the body fluids is not identical to that of the external water reservoir, then different fractionations may be found because carbonate-water or-DIC fractionations are pH sensitive (speciation is pH sensitive). Important aspects of the role of changing speciation in the solution on fractionation factors for elements such as B, C, O, and S are discussed in Fry et al. (1986), Gessler & v. Gehlen (1986), Usdowski et al. (1991), Hemming & Hanson (1992), Usdowski & Hoefs (1993), Zeebe (1999) and O'Neil et al. (2003). Table 46.6 summarizes a number of experimental field calibrations of isotopic fractionations for H-, C-, N-, O-, and S-isotopes as a function of environmental parameters as well as a few examples for Li, B, Si, C1 and Ca. A large number of systems have been selected to emphasize the wide variety of field experiments that have been carried out. Four principal types of systems dominate Table 46.6: (1) biominerals and plants, (2) weathering environments, (3) hydrothermal and magmatic systems, and (4) between major earth reservoirs. Such a table can be neither exhaustive nor comment on the quality of the results. Some of the temperatures are not known with any precision (e.g., surface temperatures for weathering reactions could cover the range 0~ to 50~ or so). Alternatively, the temperature today, measured with precision, in a geothermal system is not necessarily the same as that during the formation of the mineral. Such things must be borne in mind when using Table 46.6. Also, the remarks column can only include one or two key words about the system. Four specific systems are discussed in a little more detail.
46.8.1 Carbonate-water systems One of the very first field calibrations was of the carbonate (CaCO3)-water system by Epstein et al. (1951, 1953) where the O-isotope fractionation was directly measured between a living mollusk and its environmental seawater at a measured temperature or small range of temperatures. Importantly, at least two series of experiments were conducted in controlled temperature baths of seawater by analysing the regenerated carbonate infilling a notch or drill hole that had been made in the shell at the beginning of the experiment (i.e., Table 46.2 type laboratory experiments). The results were identical and also agreed satisfactorily with one of the relationships for inorganically precipitated CaCO3 determined experimentally in the same laboratory by McCrea (1950). The similarity of the organic and inorganic temperature scales is highly suggestive, but not proof, that equilibrium was attained. Although this interpretation is supported by subsequent experimental studies (O'Neil et al., 1969; Kim & O'Neil,
System ~
Element
Temperature
Remarks
References ~
~~~~~~
_
_
_
_
A. Direct temperature measurement
1. Aragonite-H20DIC
2-22°C
Living and modern corals, forams, mollusks.
Grossman & Ku (1986), Patterson et al. (1993), Leder et al. (1996), Swart et al. (1996), McConnaughey et al. (1997), Bohm et al. (2000)
2. Calcite-HzO
7-30°C
Marine mollusks and associated seawater (sampling strategy, metabolic effects, salinity, growth rate).
Epstein et al. (1951, 1953); Klein et al. (1996)
3. Carbonate-sea water
8-11°C
Living foraminifera and coccolithophorids and sea water.
De La Rocha & DePaolo (2000)
4. Epidote-H20
90-320°C
Geothermal wells: Reykjanes, Salton Sea, Wairakei.
Graham & Sheppard (1980), Bird et al. (1988)
5. Kaolinite-H20
80°C 65-280°C
Geothermal mudpot, Yellowstone Geothermal wells: Valles Caldera; 6D water inferred.
Sheppard et al. (1969), Lambert & Epstein (1980), Gilg & Sheppard (1996)
6. Magnetite-H20
9°C Chiton "teeth" and associated seawater 112". 175°C Geothermal well: Wairakei.
O'Neil & Clayton (1964) Blattner et al. (1983)
7. Opaline-H20
11-18°C
Opaline phytoliths from stems and leaves of wheat relative to stem and leaf water.
Shahack-Gross et al. (1996), see also Webb & Longstaffe (2000; 2002)
8. Phosphate-H20
3-40°C
Biogenic shell, bone or tooth apatite and associated or inferred water; T measured or calculated from carbonate using Epstein et al. (1953)equation.
Longinelli & Nuti (1973a), Kolodny et al. (1983); D'Angela & Longinelli (1990), Lecuyer et al. (1996a);Iacumin et al. (1996) 1029
Table 46.6 continued >
Experimental Measurement of Isotopic Fractionation Factors ...
Table 46.6 - Some experimental field calibrations of isotopic fractionations
1030
Table 46.6 continued > 9. Phosphatecarbonate
0
37°C
Biogenic apatite (bone, teeth) of modern mammals. Carbonate of phosphate.
Iacumin et al. (1996), Bryant et al. (1996)
10. Quartz-adulariacalcite
0 0
275 k2O"C 275 +20"C
Geothermal well: Broadlands. Geothermal well: Broadlands.
Blattner (1975) Blattner (1975)
11. Quartz-illite
0
160-270°C
Geothermal well: Broadlands.
Eslinger & Savin (1973)
12. Silica-H20
0
-1.5-29°C
Living marine and freshwater diatoms and sponges, and associated waters.
Labeyrie (1974), Knauth & Epstein (1975); Schmidt et al. (1997), Shemesh et al. (1992a); Clayton (1992), Shemesh (1992)
13. Sphalerite-galena
S
273°C
Geothermal well: Broadlands; fluid inclusion T.
Czamanske & Rye (1974)
14. C02-CH4
C
20-600°C
Geothermal systems: kinetic and equilibrium fractionations.
Sheppard (1981), Giggenbach (1982,1997)
15. Plant-CO2
C
surface
Photosynthesis: C3 and C4 plants.
Park & Epstein (1960), Smith & Epstein (1971)
16. Cellulose-H20
H, 0
15-30°C
Plant and animal cellulose: terrestrial, marine and freshwater.
DeNiro & Epstein (l979,1981a), Yapp & Epstein (1982), White et al. (1994) Tang et al. (2000), Aucour et al. (2002)
17. Calcite-graphite
C
600-800°C 400-680°C
Kfeld-plagio and magnetite-ilmenite. Calcite-dolomite solvus thermometer.
Valley & O'Neil (1981) Wada & Suzuki (1983)
18. Dolomite-calcite
C, 0
200-650°C
Mg-calcite-dolomite solvus thermometer.
Sheppard & Schwarcz (1970) Table 46.6 continued >
Chapter 4 6 - S.M.F. Sheppard
B. Chemical and isotopic thermometers
Experimental
Table 46.6 continued > ~~
~~
~
~~~~~~
~~
~
~
~
~
_
_
_ ~~
_ ~~
_
~~~~~~~~~~~
~
200-300°C
Internal mineral fractionation. [OH] oxygen of illite measured by vacuum dehydration and/or low-temperature fluorination.
Bechtel & Hoernes (1990)
20. Silicates and magnetite
0
500-800°C
Various selected isotopic thermometers and coexisting mineral systematics.
Bottinga & Javoy (1973,1975)
21. Pyroxene-olivine
0
855-1300°C
Various chemical thermometers.
Kyser et al., (1981), Gregory & Taylor (1986)
22. Calcite-diopside
0
800°C
Diffusion rates of oxygen in diopside estimated using both infinite and finite reservoir model, mineral modal abundance and 8180 variations.
Sharp & Jenkin (1994)
23. Quartz-calcite
0
80-730°C
Various chemical, mineralogic and isotopic geothermometers.
Sharp & Kirschner ( 3 994)
24. Serpentinemagnetite
0
50-500°C
Quartz-muscovite 1 8 0 thermometer, chlorite-Fe-Ti oxide fractionations.
Wenner and Taylor (1971)
25. Tourmaline-H20quartz
H
200-600°C
Quartz, muscovite, illite, chlorite, biotite isotopic thermometers.
Kotzer et al. (1993)
26. Zircon-titanite Quartz-titanite
0
650°C
Various experimental and empirical isotopic thermometers. Comparison with semi-empirical increment method. Closure temperature taken into consideration.
King et al. (2001)
surface
Farm, zoo and wild eggs. Fractionation between calcite, total organic C or individual acids (9) and diet, local water, climate.
Johnson et al. (1998)
0
o f I s o t o p i c F r a c t i o n a t i o n F a c t o r s ...
0
Measurement
19. Illite (whole)illite [OH]
C. Estimated or inferred temperatures 27. Calcite (Ostrich eggshell)-diet
C, N, 0
1031
Table 46.6 continued >
1032
Table 46.6 continued > seawater surface
H- and 0-isotope systematics in cherts. T calibration from oceanic chert.
Knauth & Epstein (1975,1976), Kolodny & Epstein (1976)
29. Gibbsite-Hz0
surface
Weathering zones, bauxites.
Lawrence & Taylor (1971), Bernard et al. (1976), Chen et al. (1988, 1990), Bird et al. (1989, 1990), Vitali et al. (2001)
30. Goethite-H20
surface
Weathering and supergene zones, isotopic comp. water inferred.
Yapp & Pedley (1985),Yapp (1987,1990,2000), Bao et al. (2000a)
31. Kaolinite-H20
surface
Weathering zones.
Savin & Epstein (1970), Gilg & Sheppard (1996), Sheppard & Gilg (1996)
32. Opaline-seawater
seawater
Fractionation with dissolved silicic acid.
De La Rocha et al. (2000)
33. Smectite-H20
surface surface
Weathering and oceanic clays. Weathering; function of octahedral Fe3+/(Fe3f + Mg + Al). Interlayer water-water.
Savin & Epstein (1970),Lawrence & Taylor (1971), Gilg & Sheppard (1995), Sheppard & Gilg (1996) France-Lanord & Sheppard (1992)
surface
Hydrated obsidians and perlites; H20 >-2wt.%.
Friedman & Smith (1958), Taylor (1968), Cerling et al. (1985),Gilg & Sheppard (1999)
35. Zeolite-Hz0
surface
Analcime, natrolite: channel water (chabazite, clinoptilolite, laumontite: rapid exchange with ambient water vapour).
Karlsson & Clayton (1990)
36. Bone collagen and phosphate-water
surface
Deer bone from across N. America related to environment (meteoric water, humidity).
Cormie et al. (1994)
37. Plant leafwater transpiration
surface
Leaves of astomatal plants sampled under measured air T and humidity conditions.
Cooper et al. (1991)
seafloor 34. Volcanic glass-
Hz0
Table 46.6 continued >
Chapter 46 - S.M.F. Sheppard
28. Chert-water
38. C3 plantsenvironment
C
surface
Spatial, species and temporal variations in leaves, needles and wood.
Heaton & Crossley (1998), Heaton (1999)
39. Tree ringenvironment
H, c , 0
surface
Isotopic variations of cellulose nitrate of tree rings with climatic variables (temperature, humidity, precipitation, topography, soil, within site and between species,. . .).
Edwars & Fritz (1986), Feng & Epstein (1995a, 1995b), White et al. (1994), Tang et al. (2000), Aucour et al. (2002).
40. Nitrification
N
surface
Fractionation during nitrification and algal assimilation of NH4+ estimated from measurement of dissolved NH4+ and N03in river water.
Cifuentes et al. (1989), see also Mariotti et al. (1981)
41. Bacterial sulphate S reduction in organicsediments
surface
Fractionation during sulphate reduction in cyano-bacterial microbial mats and sediments at known rates of sulfate reduction. Part of experiment carried out in laboratory
Habicht & Canfield (1997)
42. Microbial sulphate reduction
0, S
sea floor
S- and 0-isotope fractionations at oil and gas seeps. Rates of sulphate reduction estimated. Seep sediments compared with reference sediments.
Aharon & Fu (2000)
43. Magma (granitic)-
H
700°C
Rayleigh distillation model with several assumptions.
Nabelek et al., (1983)
44. Magma (rhyo1itic)H20
H
magmatic
From obsidians in tephra flows, Rayleigh model with assumptions.
Taylor et al. (1983)
45. Altered oceanic crust-seawater
Li
crustal
Chan et al. (1992) Fractionation estimated between altered basalt and seawater from 66Li measurements of fresh and altered oceanic basalts.
H20
1033
Table 46.6 continued >
Experimental Measurement of Isotopic Fractionation Factors ...
Table 46.6 continued >
1034
Table 46.6 continued > 46. Continental crustseawater
B
crustal
Model fractionation derived from 611B values of continental tourmalines with time.
Chaussidon & AlbarPde (1992)
47. Evaporite mineralbrine
B
surface
Fractionations associated with precipitation of borates and adsorption onto clays in natural evaporitic environments.
Xiao et al. (1992b), Vengosh et al. (1995)
48. Diffusion coefficient in water
C1
surface
Fractionation associated with diffusion of C1-derived from S37C1 in groundwatersaline water mixing zone, in absence of advection.
Desaulniers et al. (1986)
49. Surface-mantle reservoirs
c1
crust to mantle
Fractionation (37C1/35C1) estimated from fresh MORB glass, high and low T altered oceanic crust.
Magenheim et al. (1995), see also Eggenkamp & Koster van Groos (1997)
Chapter 4 6 - S.M.F. Sheppard
Experimental Measurementof Isotopic FractionationFactors ...
1035
1997), it is well documented that the equation does not apply to all low temperature carbonate-water systems. Certain biogenic and inorganic carbonates are not precipitated in equilibrium with their associated fresh or saline waters (see sections 46.6.4.2 and 46.8) and the nature of the biogenic carbonate (aragonite, low- or high-Mg calcite) has to be considered. 46.8.2 Phosphate-water systems The phosphate-water fractionation expression was initially derived by Longinelli and his coworkers (Longinelli, 1965, 1966; Longinelli & Nuti, 1973a) from the analysis of biologically precipitated (biogenic) invertebrate phosphate and its associated biogenic calcium carbonate (high- and low-Mg calcite, aragonite) from the same shell, plus application of the Epstein et al. (1953) equation for calcium carbonate-water to determine the average growth temperature. Subsequently, Kolodny et al. (1983) measured directly the biogenic fish phosphate-water fractionation, in both sea and fresh waters, at measured temperatures. They also demonstrated that the isotopic composition of fish bone phosphate is only influenced by the isotopic composition and temperature of the water, and is not inherited from either the food or dissolved phosphate oxygen. L6cuyer et al. (1996a) combined measured isotopic compositions of living inarticulate brachiopods (lingulides) with inferred isotopic composition, salinity and temperature of seawater data to confirm that a single equation describes the phosphate-water fractionation of many marine vertebrate and invertebrate organisms, except mammals. All these temperature dependent empirical fractionations have usually been considered to represent equilibrium rather than kinetic effects, although direct evidence for equilibrium was lacking. Recent laboratory experimental determination of the equilibrium isotopic fractionation between the dissolved inorganic orthophosphate ion, H2PO4-, and water by L6cuyer et al. (1999) indicates that biogenic apatites are systematically 1sO-depleted (~ 8%o at 20~ However, this ~ 8%o difference represents the equilibrium fractionation between phosphate solutions with seawater type pHs, where H P O 4 2- is the dominant ion, and lower pH solutions, where H2PO4- is the dominant ion (O'Neil et al., 2003). This example emphasizes the importance of speciation on mineral-solution fractionations. 46.8.3 Plants Isotopic compositions and their variations of plants, their organs (e.g., leaf, wood, roots) or constituents (e.g., cuticles, phytoliths,...) are increasingly being used as proxies for environmental conditions in palaeoclimatic studies. Although a certain number of controlled laboratory experiments have been conducted (see section 46.7.1.1 and Table 46.2), field experiments have also played their part (N ~ 7, 15, 16, 36-39, Table 46.6). In particular, such studies have demonstrated that variations substantially larger than the analytical precision commonly occur in the same substance from different parts of the same plant as well as between plants of the same species from the same site, not to mention differences between different species from the same site (for references see above examples from Table 46.6). Because the isotopic composition is usually a function of several environmental variables- temperature, soil water, humidity, insolation, etc...- unscrambling the isotopic signal(s) in terms of a particular parameter becomes quite complex. Additionally, because the analytical techniques are
1036
Chapter 46 - S.M.F. Sheppard
often quite labour intensive, isotopic analyses of, for example tree rings, are rarely taken from more than a single tree per site (Tang et al., 2000; Aucour et al., 2002). Moreover, an isotopic calibration carried out on a currently living species may not be applicable, in detail, to fossil material that is similar, but not identical, to the so-called calibrated reference living material.
46.8.4 Between reservoir fractionations Examples N ~ 1 to 38 (Table 46.6) measured fractionations between substances minerals, plants, magma, liquid or gaseous species. In contrast to these more classical examples, N ~ 3 and 44 to 48 are for the more recently studied isotopic ratios of Li, B, Si, C1 and Ca. The empirical experiments were principally designed to measure fractionations among major earth reservoirs - seawater, oceanic crust, mantle, etc...- in order to understand the large-scale geochemical cycle of these elements. 46.9 Conclusions
Stable isotope geochemistry has come a long way over the past 50 years or so. The results of carefully designed laboratory and field experimental isotopic studies, many of which have been extremely time-consuming, have contributed fundamentally to this progress. On quick inspection, many experimental laboratory studies seem to be quite straightforward, but, as partially summarized in Tables 46.1 to 46.5, much time, ingenuity and a wealth of ideas have been invested behind the results. The development of new analytical techniques has played its role, of which the ion microprobe and multiple collector inductively coupled plasma source mass spectrometer are two examples. Future laboratory and experimental field studies will be equally challenging. Interactions among the three approaches to characterize isotopic fractionation fact o r s - experimental, empirical and theoretical- are fundamental to improving our understanding of isotope effects, because each method suffers from certain difficulties and limitations. Laboratory experiments, however, hold a key position in this trinity because they can be repeated or redesigned to increase their credibility in a way that is often less evident for the others. It is recalled that the theoretical modelling of isotopic fractionations requires experimental spectroscopic data as basic input. Surprisingly few minerals, however, have been synthesised experimentally with strong enrichment in the usually minor occurring isotope(s) so that the effects of isotopic substitution on the vibrational spectra (infrared, Raman,...) can be measured (e.g., Sato & McMillan, 1987; Gillet et al., 1996; B6ttcher et al., 1997c) rather than calculated using some model. Comparison of experimental fractionation data with theoretical models is essential to assess whether a harmonic approximation is satisfactory or not. Although emphasis has been placed on how isotopic fractionation factors have been measured experimentally both in the laboratory and in the field, much detailed information and understanding have also been derived on mechanisms and rates of reactions. These have many applications. In fact there is a chicken and egg situation because understanding the mechanism of the reaction can aid both the design of the experiment and the interpretation of the results.
Experimental Measurement of Isotopic Fractionation Factors ...
Acknowledgments
1037
Since my introduction to experimental isotope geochemistry at McMaster University, Canada with H. P. Schwarcz, D. M. Shaw and H. G. Thode, I have particularly benefited from my collaborations and/or interactions with S. Epstein, C. M. Graham, J. J. Hemley, M. Pichavant and H. R Taylor, Jr., all of whom are gratefully acknowledged. This chapter has benefited from the helpful comments of A. Matthews and J. W. Valley.
Handbook of Stable Isotope AnalyticalTechniques, Volume 1 P.A. de Groot (Editor) 9 2004 Elsevier B.V. All fights reserved.
CHAPTER 47 Laboratory Set-Up for GC-MS and Continuous-Flow IRMS Wolfram Meier-Augenstein Queen's University Belfast, Environmental Engineering Research Centre, David Keir Building, Belfast, BT9 5AG, UK e-mail: [email protected]
47.1 The Ideal MS / IRMS Laboratory It is probably fair to say that every scientist dreams about access to unlimited resources to design and build the ideal laboratory for their work. It seems to be equally true that those scientists involved in mass spectrometry (MS) and isotope ratio mass spectrometry (IRMS) are more likely than others to have their wish granted. In reality though, this is mainly due to administrative considerations concerning the protection of a major cash investment in the instruments rather than assuring high quality analytical work. However, if you should find yourself in the fortunate position of being asked to submit a wish list for your IRMS laboratory, take advantage of this situation as best as you can. The performance of your instruments and the quality of the results they produce not only depend on the amount of time and effort you put into sampling, sample preparation, sample separation and instrument maintenance but also on the quality of the environment your instruments operate in. In the following paragraphs, consideration is given to all the points that affect instrument installation and its performance. This listing is the result of foresight and hindsight in conjunction with experiences made or rather suffered when setting up two MS / IRMS facilities at two universities. Since individual circumstances and precise requirements vary, a broad and general outline of bare necessities is given that can easily be adapted to and incorporated into the situation at hand. 47.2 Location While planning a new or completely refurbished laboratory, usually the first step is to decide where this new facility should be located. The choice of locale should be dictated by four considerations" (1) Access; (2) Floor space; (3) Floor stability and (4) Climate. Of course, there are other things to consider as well such as noise level, the lab's location within the exisiting power grid (section 47.4), gas supply (section 47.5) and connections to telephone and IT services. These and similar points are secondary to identifying the premises that will accomodate the future MS / IRMS laboratory and
Laboratory Set-Up for GC-MS and Continuous-Flow IRMS
1039
will be discussed separately. The first three points almost always lead to a choice of ground floor or basement level accommodation. The new lab should have level access or access via a gently sloped ramp. The doors must be wide enough to manoeuvre the instrument(s) through with ease. The floor must be rock solid, i.e. free of vibrations, and the floor space generous enough to allow easy access to the instrument or instruments from all sides, e.g. instrument floor space plus 1 metre on each side. In terms of overall floor space, one must also consider the spatial requirements of ancillary instruments, future equipment, distance of computers and monitors from the IRMS magnets and the installation of a fume cupboard. In line with local safety regulations, think about space required for gas cylinder storage (see section 47.5). Lastly, the largest outer wall area should face North (Northern Hemisphere locations) or South (Southern Hemisphere locations) since it's easier and cheaper to keep a place warm than transport excess heat out.
47.3 Temperature Speaking of heat, all your analytical instruments will benefit from an environment with constant ambient temperature, preferably around 23 ~ with a temperature stability of + 1 ~ i.e. drift in temperature should not exceed I ~ per hour. So, air conditioning is a must on the check-list for the new laboratory. If money is no object, one should chose an air conditioning system that provides humidity control as well. The size and capacity of the air conditioner depends on the number of instruments in your laboratory and their respective heat output at peak. This includes GCs, vacuum pumps, drying ovens, personnel and spare capacity for future equipment.
47.4 Power Supply When planning your new laboratory, make sure it will have its own dedicated power supply line from the nearest local sub-station. In other words, insist that your laboratory will not be the last in line of an already several times over extended campus grid serving several other major consumers. All analytical instruments are sensitive to fluctuation in the power supply and IRMS instruments are no exception. Apart from maintaining a constant voltage, your power supply should be free of surges, spikes and brown-outs. If that cannot be guaranteed, make sure one or more UPS units (uninterrupted power supply) are installed and connected between wall socket and instrument. An UPS will enable you to safe your data and shut down your IRMS in a controlled fashion should there be a power cut. In addition, modern UPS units act as surge filters and provide a constant voltage output even if your mains power supply does not. An UPS unit providing up to 2 hours backup supply is probably the best compromise in terms of expense and space requirement in a situation where you are not served by an emergency power generator. Should your mains power supply be backed up by a emergency power generator, a smaller UPS unit will suffice, e.g. 15 to 30 minutes power backup, but is still essential since its short reaction time will bridge the time before the emergency power generator kicks in.
1040
Chapter 47 - W. M e i e r - A u g e n s t e i n
Another point worth mentioning is the number of sockets in the new laboratory. In a variation of Murphy's Law there never seem to be enough sockets in a lab and if there are, they are never where you need them. This inevitably leads to the intricate and intertwined system of tripwires otherwise known as extension leads and extension sockets. As a general rule, for an IRMS lab you will require at least one dedicated three-phase, five-connector power line that must be fused separately. In addition, each wall should contain one dedicated, separately fused circuit feeding four blocks of four sockets each, two at floor level and two at bench-top level. In addition, consider the possibility of one or two socket cubes suspended from the ceiling in the centre of the lab.
47.5 Gas Supply In the context of an IRMS (and / or MS) facility, gas supply means the supply of gases required to run your instruments, i.e. gases such as Helium, Oxygen, Hydrogen, Nitrogen, Synthetic Air, Carbon Dioxide and Carbon Monoxide as well as a supply of oil-free compressed air. The overriding rule for gas supply in instrumental analysis in general and IRMS in particular is "the cleaner the better". This maxim can only be achieved if every step in the gas train is set up in such a way as to exclude leaks, atmospheric break-in and contamination. To this end, all materials used for gas supply should be made of stainless steel. This includes pressure regulators that should contain a stainless steel diaphragm, tubing, connectors, valves and ferrules. Regulator diaphragms made from other materials will deteriorate over time and start leaking. Please note, using stainless steel tubing and then connecting it with brass unions and brass ferrules is false economy. Brass ferrules are softer than stainless steel and will therefore not cut into the stainless steel tubing and, hence, not afford a leak tight seal. By the way, the stainless steel tubing must be clean on the inside, i.e. free of any contaminants associated with its manufacture. Gases should be of highest purity, i.e. 99.998% (= 4.8 in European notation) or 99.999% (= 5.0). Ideally, the carrier gas should be of even higher quality, i.e. 99.9999% (= 6.0). However, Helium of this quality is rather expensive and the same quality can be achieved by using 4.8 or 5.0 Helium in conjunction with a high-capacity gas purifier. The best gas purifier on the market is Supelco's thermo-chemical absorption system, the "High Capacity Gas Purifier" that consists of a purifier tube and a small oven. To monitor the performance of this system as well as providing a back-up once the purifier tube is exhausted, it is recommended to install an additional self-indicating purifier tube (e.g. Supelco's OMI-2) in line behind the thermo-chemical trap. The two major reasons for the high purity demand on the carrier gas He are the capillary columns in the GC and the IRMS itself. Stationary phases of medium to high polarity are extremely sensitive to oxygen being present in the carrier gas, which leads to column deterioration and excessive column bleed. To achieve the best possible accuracy and precision of IRMS measurements, the carrier gas must also be free of organic contaminants and moisture.
Laboratory Set-Up for GC-MS and Continuous-Flow IRMS
1041
With the possible exception of the smaller reference gas cylinders, ideally all the gases should be stored outside the lab in two safety cabinets, accommodating four cylinders each. A typical set-up would be N2 and H2 in one cabinet, and He, 02 and synthetic air in the other, thus keeping combustable and flamable gases separate. To safeguard against running out of He when one can afford it least, two He cylinders should be connected simultaneously to a change-over regulator with a pressure dependent toggle switch. Unfortunately, these change-over regulators come with a high price tag and, hence, in most cases must be regarded as a luxury albeit a desirable one. From the central gas storage point, 1/4" stainless steel tubing should be run into the lab and along the walls at head height. From these 1/4" supply lines 1/8" stainless steel tubing tee off at or near the instruments. A pressure gauge and a shut-off valve should be placed between the T-union and the instrument supply tubing. Prior to its first use, such a gas supply line set-up should be flushed with clean, dry nitrogen to remove air and moisture. Similarly, compressed air should be supplied from an external compressor and brought into the lab via a dedicated line that incorporates an oil mist filter. Also, consider a natural gas supply line for your Bunsen burners or, alternatively, use gas cartridge operated burners. Finally, an exhaust line for the vacuum pumps should be installed, either plumbed into the exhaust of the fume cupboard or leading directly outside. A brief note on gas safety. Even with gas cylinders stored outside the laboratory, it is a good idea to have a H2 monitor with alarm in the laboratory. In addition, due to the increased interest in on-line IRMS measurement of 1 8 0 / 1 6 0 isotope ratios of organic compounds, consideration should be given to the installation of a CO monitor (with alarm), because 5180 measurement by on-line thermal conversion (pyrolysis) of organic compounds requires the use of CO as reference gas. For this application provision must be also made for an extraction line positioned at the conversion interface of the GC-IRMS, either branched into a fume hood or connected to a separate extractor fan, to remove CO emanating from the reference gas inlet and potential leaks in the interface.
47.6 Finishing Touches When choosing the floor cover, practicality should overrule appearances. A carpetlike floor cover might look nice but is prone to induce static electricity discharges and is difficult to clean. A chemically inert surface is better suited, especially if the lab floor has been fitted with a drainage point at the lowest point of the floor. Make provisions for ample work surface / bench-top area with underneath storage facilities. Ideal are mobile cupboard, drawer and shelf units on castors with a small work surface of their own. Ask for a solvent storage cabinet that is earthed and connected to the exhaust of your fume cupboard.
1042
Chapter 47 - W. Meier-Augenstein
Good, i.e. sufficient and evenly spaced lighting is also important. Soft fluorescent light tubes should be suspended from the ceiling at a height of approx. 2.5 metres. Equally important are telephone points and sockets to connect your computer to your local server and / or local intranet. From a practical point of view, a sink with running hot and cold water is desirable as long as it is not directly adjacent to the mass spectrometer. Similarly~ a supply line of de-ionised water is a valuable thing to have. Mass spectrometer laboratories are notoriously noisy, chiefly due to the high vacuum systems. Accommodating the rotary pumps in ventilated yet noise insulated cupboards will go a long way to reduce the noise level in the lab. However, this still leaves to high-pitched noise of the turbo-molecular pumps. Again, if money is no object a sound engineer should be consulted to suggest appropriate sound proofing measures tailor-made to the situation at hand. Lastly, if you want to see all your planning and requests come to fruition, do not sit back and wait for the hand-over day of your new lab. Be involved and get involved during the entire process of your new lab being designed, built and fitted. Be prepared for specifications not being met, the drainage point not being at the lowest point of the floor and doors not being wide enough despite clear instructions and specifications. By the same token, be prepared to discover that the architect, the engineers, the builders, plumbers, electricians etc. will all work on the assumption that you, the MS / IRMS specialist and end-user of this lab will have no idea of what you are talking about and what you really need. For that reason they will make changes to plans, dimensions and specifications without telling you about it. So, get involved armed with a tape measure, a calibre, a volt meter (or even better a voltage monitor) and the pre-installation requirement booklet of your IRMS and fight your corner.
Isogeochem List
1043
APPENDIX A I s o g e o c h e m List ISOGEOCHEM is an e-mail discussion list in stable isotope geochemistry. The list currently counts 1900 subscribers but is still increasing in size. As a result, there are active discussions on all different topics involving stable isotopes. The list represents a useful tool for obtaining information on stable isotope analysis and a range of related subjects. The list-owner, Andrea Lini, is acknowledged for his allowance to present details of the list, including procedures to subscribe or unsubscribe, in this book. Although there always is a risk that this information will become invalid because of changes and improvements in the fast developing internet technology, it is considered highly valuable to include current information here for those not yet aware of the list. The objectives of the ISOGEOCHEM list are to promote the exchange of news and information among those with an interest in stable isotope geochemistry, and to provide new contacts and enhance collaboration among researchers from different disciplines (e.g. geology, biology, chemistry). The list is intended not only as a discussion forum for isotope geochemists but also as a source of information and help for researchers from other fields interested in applying stable isotopes as an additional tool in their own studies.
Procedures for (un)subscribing to the ISOGEOCHEM list In order to gain the benefit of this discussion list, you must subscribe to it. Simply send an electronic mail message to this address: [email protected] Leave the subject field blank. Construct a subscribe command as follows and place it as the first line of your message: sub ISOGEOCHEM fname lname where fname is your first name and lname is your last name. To signoff from the list, email to [email protected] with the following request: signoff ISOGEOcHEM or unsubscribe ISOGEOcHEM
1044
Appendix A
If you have any questions regarding the ISOGEOCHEM discussion list, feel free to contact the "list-owner", Dr. Andrea Lini ([email protected]). I S O G E O C H E M and related WEB-Sites ISOGEOCHEM also has a web-site:
http: / / www.isogeochem.com This site provides information about conferences, vacant positions, IRMS manufacturers and suppliers of isotopic materials. It also contains links to numerous stable isotope laboratories worldwide. Archive All postings to the ISOGEOCHEM list are recorded in ISOGEOCHEM's Archive. The archive has a full search engine for quick retrieval of anything ever posted on the list. This archive can be found on WEB-Site:
http" / / list.uvm.edu / archives / isogeochem.html or by connecting to the ISOGEOCHEM WEB-Site given above, and selecting the link to the Archive. Address list and E-mail address list An address list of subscribers to the list can be found on WEB-Site:
http://geology.uvm.edu / address.html This site can also be accessed from the ISOGEOCHEM WEB-Site. Note on the addresslist: adresses of only about 10% of the subscribers is only given; only those who provided a complete address are included in that list.
The Web Stable Isotope Fractionation Calculator
1045
APPENDIX B The Web Stable Isotope Fractionation Calculator Georges Beaudoinl & Pierre Therrien D6partement de G6ologie et de G6nie G6ologique, Universit6 Laval, Qu6bec, Canada GIK 7P4 e-mail: i [email protected]
The fractionation of stable isotopes between two isotopic species is an inverse function of temperature. It is convinient to transform the isotope fractionation (a) into a polynominal inverse function of temperature such as: ( 10181 ( 10121 ( 109/ (~26) 1 )+F 10001nc~ - A T6 ) + B T4 ) + C - ~ j +D +E( 3T
[B.1]
where A, B, C, D, E, and F are variables and T is the temperature in K. The advantage of the 1000 In (Z(a-b)logarithmic transformation is that its numerical value is similar to the difference (Aa-b) in g-values of each isotopic species a and b (Hoefs, 1997) that can be determined experimentally in the laboratory" [B.2]
Aa-b = 6a-~lb
where the h-value is the per mil difference in isotope ratios relative to a standard: 180/16Oi 6i -
180/16OvsMow
-1) 91000
[B.3I
where VSMOW is a standard, here Vienna-Standard Mean Ocean Water. The fractionation factor O~a-bis related to the 5-value by the following relation: 1000 4- ~}a
[B.4]
~ a - b = 1000 + 5 b
Considering the uncertainty of 5-values, which is commonly up to 0.2%0 for 513C, 6180, ~34S, and of 1-5%o for 5D, propagation of errors yields uncertainties on Aa-bvalues of about 0.3%0 and up to 7%o, respectively. These experimental uncertainties are larger than the difference between the Aa-b and 1000 In Or(a-b) for Aa-b values below
1046
Appendix B - G. Beaudoin & P. Therrien
20%o and 6-val- Table B1 - Evaluation of the 1000 In a(a-b) logarithmic transformation. ues below about 6a 6b O~(a-b) Aa-b 1000 In O~(a-b) 20%o for (~13C, 6180, (~34S (Table 0 1.0100 10 9.95 B1) For hydro10 10 1.0099 10 9.85 9 20 gen isotope data, 20 0 1.0200 20 19.80 higher 6Da and 30 10 1.0198 20 19.61 30 1.0194 20 19.23 Aa-b have higher 50 -30 1.0619 60 60.02 numerical values 30 -20 0.9388 -60 -63.18 and uncertainties -80 such that use of the 1000 In O~(a-b)logarithmic transformation remains a good approximation in most situations. A large number of experimentally and theoretically derived 1000 In C~(a-b)equations exist in the literature and they have been compiled periodically (Friedman & O'Neil, 1977; O'Neil, 1986; Kyser, 1987b). Traditional hardcopy compilation become
Figure B1 - Main w i n d o w of the WEB S t a b l e I s o t o p e Fractionation Calculator that permits input of the temperature and selection of the equation of interest.
The Web Stable Isotope Fractionation Calculator
1047
Figure B2 - Result of the computation of the stable isotope fractionation at the temperature and for the equation selected. The coefficients of the equation are given with the reference (s) for the equation. quickly obsolete, however, and may contain typographic errors that are not easily corrected (Savin & Lee, 1988). The Internet offers an alternative application that is widely accessible and which can be conveniently updated to account for new data or to correct errors. The Internet Stable Isotope Fractionation Calculator:
http" / / www.ggl.ulaval.ca / cgi-bin / isotope / generisotope.cgi is based on a regularly updated database listing a large number (currently more than 450) of fractionation equations between isotopic species of hydrogen, carbon, oxygen and sulfur. The user inputs a temperature (in C) and selects an equation between two isotope species (Figure B1). The program returns the 1000 In ~(a-b) value for the input temperature, the values of coefficients A to F used for the computation, and the reference(s) for the equation (Figure B2). It thus offers a simple yet powerful tool to obtain a list of fractionation equations and their coefficients to interpret stable isotope data.
1048
Appendix C
APPENDIX C S u p p l i e r s of R e f e r e n c e Materials
C1 Adresses for reference and stable isotope materials International Atomic Energy Agency (IAEA) Analytical Quality Control Services Agency's Laboratories Seibersdorf P.O.Box 100 A-1400 Vienna Austria Stable isotope reference materials at environmental level: M. Gr6ning Phone: +43-1-2600-21740 Fax: +43-1-26007 E-mail: [email protected] General Phone: Fax: E-mail:
information on all reference materials: Mr. Radecki +43-1-2600-28226 +43-1-2600-28222 [email protected]
Institute for Reference Materials and Measurements Joint Research Centre of the European Commission B-2440 Geel Belgium For any information or for ordering: E-mail: [email protected] Andr6 Verbruggen Coordinator Isotope Reference Materials Phone: +32 14 571 617 Fax: +32 14 591 978 E-mail: [email protected] Information also can be obtained from the IRMM Web site: http"//www.irmm.jrc.be / mrm.html
Suppliers of Reference Materials
1049
National Institute of Standards and Technology (NIST) Standard Reference Materials Program Room 204, Building 202 Gaithersburg, Maryland 20899-0001 USA
Phone: +1-301-975-6776 Fax: +1-301-948-3730 E-mail: [email protected] Orders for NIST must be accompanied by an order number of a selected reference material (see: Coplen, 1996, or contact NIST for order numbers). CEA: distributes stable isotopes through its daughter company" EUROISO-TOP Parc des Algorithmes (Bat. Hombre) F-91194 St. Aubin France http://www.eurisotop.fr / Messer Griesheim (MG) Futingsweg 34 47805 Krefeld Germany Phone: +49-203-6002-388 Fax: +49-203-6002-460
A list of suppliers of stable isotope compounds and stable isotope labelled compounds can be found on the WEB-site: http: / / www.uvm.edu / --geology / geowww / suppliers.html
1050
Appendix C
C2 Indiana Zinc
The Biogeochemical Laboratories at Indiana University use several methods to reduce water to elemental hydrogen for subsequent determination of D / H ratios. For quantitative production of hydrogen from microliter-quantities of water by use of a simple and inexpensive batch process, we recommend the use of zinc metal (Coleman et al., 1982). Pure zinc is unsuitable because it will not react with water quantitatively. Zinc with suitable characteristics had been available from BDH. As BDH zinc is not produced specifically for hydrogen isotopic analysis its characteristics in the application are unreliable. During the 1980's, privately funded research at Indiana University resulted in a proprietary method to prepare zinc turnings for rapid quantitative reduction of water to H2. Our "optimally contaminated" zinc, which we are calling "Indiana Zinc", is used in dozens of laboratories worldwide. It has a long-standing track record of providing reproducible hydrogen stable isotope ratios (for example, see Schimmelmann & DeNiro, 1993). Since 1999, two types of Indiana Zinc are available. The first type is the traditional formula with its long track record of proven reliability for laboratories that can keep a constant zinc" water ratio. A newly developed zinc formula is available upon request that has a reduced hydrogen blank ("amount effect"; see: Dem6ny, 1995) and is especially suited for analyses where the zinc" water ratio cannot be controlled within a narrow range. For more information on how to obtain "Indiana Zinc" and how to convert water to elemental hydrogen by using our zinc, send e-mail to: [email protected], or fax (812) 855-7961, or send letter to: Arndt Schimmelmann, Indiana University, Department of Geological Sciences, Biogeochemical Laboratories, 1001 East Tenth Street, Bloomington, IN 47405-1405, USA. Indiana Zinc is a research material offered for use without guarantees and without acceptance of any responsibilities for damages arising from its use or possible failure in any application. It is distributed as a service to those engaged in stable isotope research. For additional information, see: http" / / www.indiana.edu / ~geosci / research / biogeochem / biogeochem.html
Suppliers of ReferenceMaterials
1051
C3 H and C stable isotope standards for organic compound-specific investigations Compound-specific hydrogen isotope ratios for organic hydrogen are now analytically accessible by a combination of GC, high-temperature pyrolysis to elemental hydrogen, and subsequent on-line irm-MS (Sessions et al., 1999; Scrimgeour et al., 1999; Hilkert et al., 1999; Tobias & Brenna, 1997). Complete irmGCMS-systems with interfaces are commercially available, but there has been no isotopically defined set of organic compounds for routine calibration and D / H quality control. Standards that reliably establish or confirm isotopic calibrations must be in the same form as the unknown analytes. In contrast, the use of intermittent spikes of introduced elemental "standard" hydrogen gas is fraught with potential problems because it does not take into account D / H fractionations that may occur in the analytical train between injection in the GC and the exit of the pyrolysis reactor. The Biogeochemical Laboratories at Indiana University, in collaboration with Woods Hole Oceanographic Institution (A. Sessions, J.M. Hayes), first established the purity of n-alkanes (range from C-12 to C-50) and n-alkanoic acid methyl esters (C-10, C-20, C-30 FAMEs) by GC-MS, followed by the measurement of D / H and 12C/13C ratios for each compound. Up to five replicate analyses for each compound were performed off-line, via conventional combustion of milligram-amounts of individual compounds in quartz ampules and cryogenic purification of combustion gases in a vacuum line. Water was converted to elemental hydrogen in contact with uranium, followed by collection of hydrogen gas using a Toepler pump. Gas yields and elemental H / C ratios were routinely monitored manometrically for quality control. Hydrogen and carbon isotopic ratios were determined using MAT252 mass-spectrometers at Indiana University. The hydrogen isotopic calibration employed the conventional normalization to VSMOW (zero per mil) and SLAP (-428 per mil), according to Coplen (1996). Our standards are typically requested by the research community in form of solutions in hexane that are sealed under argon in glass ampules. We offer individual solutions of n-alkanes (C-12 to C-50) and fatty acid methyl esters (C-10, C-20, C-30). They are useful for co-injection as internal isotopic standards. One particular'mixture B' should be mentioned here because it is designed to test the accuracy of H3 + correction in hydrogen-isotope-IRMS (see Sessions et al., 1999). Mixture B can be used to measure the H3 + factor under conditions closely matching those experienced by analytes. It contains fifteen n-alkanes (C-16 to C-30) containing a 5-fold range of concentrations (arranged in three pentads with rising concentrations), from 20 nmol H2 to 100 nmol H2 per compound per microliter of solution (see Figure C1). Custom mixtures may be available upon request.
1052
Appendix C
Figure C1 - irm-GCMS chromatogram showing ion currents (m/z 2 and 3) of n-alkane mixture "B" (courtesy of Alex Sessions, Woods Hole Oceanographic Institution).
Although intended primarily for D / H research, our standards are equally useful for irmGCMS work on carbon isotope ratios. For more information, please send email to: [email protected], fax (812) 855-7961,or send letter to: Arndt Schimmelmann, Indiana University, Department of Geological Sciences, Biogeochemical Laboratories, 1001 East Tenth Street, Bloomington, IN 47405-1405, USA. Our isotope standards are research materials offered for use without guarantees and without acceptance of any responsibilities for damages arising from its use or possible failure in any application. As a public university, we supply this reference standard material as a service to those engaged in stable isotope research, rather than as a commercial product. For additional information, see: http: / / www.indiana.edu / ~geosci / research / biogeochem / biogeochem.html
REFERENCES
This Page Intentionally Left Blank
References
1053
--
(1983) International vocabulary of basic and general terms in metrology. International Organization for Standardization, Geneva, Switzerland, 2nd edition. [40]1 -(1990)Commission regulation (EEC) No 2676/90 of 17 September 1990 determining community methods for the analysis of wines. Off. J. Eur. Commun., L 272:33 p. [5] -(1995) Reference and intercomparison materials for stable isotopes of light elements. IAEATECDOC-825, IAEA, Vienna: 165 p. [5] -(1997) Element sector ICPMS (1997) and the VG-MicroTrace 9000 sector GDMS (ca. 1995). Promotional publication, Finnigan-MAT [35] -(2000) Commission regulation (EC) No 2729/2000 of 14 December 2000 laying down detailed implementing rules on controls in the wine sector. Off. J. Eur. Commun., L 316:16 p. [5] Abbas M. M., Guo J., Carli B., Mencaraglia F., Carlotti M. & Nolt I. G. (1987) Heavy ozone distribution in the stratosphere from far-infrared observations. J. Geophys. Res., 92: 13231-13239. [18] Abelson P. H. & Hoering T. C. (1961) Carbon isotope fractionation in formation of amino acids by photosynthetic organisms. Proc. Natl. Acad. Sci. U.S.A., 47: 623-632. [9] Abragam A. (1961) The principle of nuclear magnetism. Clarendon Press, Oxford. [5] Abrajano T. A. Jr., Murphy D. E., Fang J., Comet P. & Brooks J. M. (1994) 13C/12C ratios in individual fatty acids of marine mytilids with and without bacterial symbionts. Org. Geochem., 21: 611-617. [9] Adams L. H. (1949-1950) Ratio of oxygen isotopes in quartz of contrasted origin. In: Carnegie Inst. Of Washington Year Book, Vol. 49: 36-37. [20] Adams N. G., Smith D. & Henchman M. J. (1982) Isotope exchange in the reactions H30 + + D20, NH4 + + ND3, CH5 + + CD4 and their mirror isotope reactions at thermal energies. Int. J. Mass Spectrom. Ion Phys., 42: 11-23. [4] Adams R. M., Hamilton S. A. & McCarl B. A. (1985) An assessment of the economic effects of ozone on U.S. agriculture. J. Air Pollut. Cotrol Assoc., 35: 938-943. [18] Adamsen E J. & Reeder J. D. (1983) A comparison of diffusion and distillation methods for preparing samples for 15N analyses. Soil Sci. Soc. Am. J., 47: 612-613. [15] Ader M., Boudou J.-P., Javoy M., Goffe B. & Daniels E. (1998) Isotope study on organic nitrogen of Westphgalian anthracites from the Western Middle field of Pennsylvania (U.S.A.) and from Bramsche Massif (Germany). Org. Geochem., 29: 315-323. [16] Ader M., Coleman M. L., Doyle S. P., Stroud M. & Wakelin D. (2001) Methods for the stable isotopic analysis of chlorine in chlorate and perchlorate compounds. Anal. Chem., 73: 4946-4950. [28] Aggarwal J. K. & Palmer M. R. (1995) Boron isotope studies of natural materials: A review of analytical methods. Analyst, 120: 1301-1307. [7] Aggarwal J. K., Rehktimper M. & Mezger K. (1999) Boron isotope measurements by multiple-collector ICP-MS. EOS Trans. AGU, 80: Fl118. [31] Aggarwal P. K. & Dillon M. A. (1998) Stable isotope composition of molecular oxygen in soil gas and groundwater: A potentially robust tracer for diffusion and oxygen consumption processes. Geochim. Cosmochim. Acta, 62(4): 577-584. [15] Aggett J., Burton C. A., Lewis T. A., Llewellyn D. R., O'Connor C. & Odell A. L. (1965) The isotopic analysis of oxygen in organic compounds and in co-ordination compounds containing organic ligands. Int. J. Appl. Radiat. Isot., 16: 165-170. [23] Aguilera R., Becchi M., Casabianca H., Hatton, C. K., Catlin D. H., Starcevic B. & Pope H. G. (1996) Improved method of detection of testosterone abuse by gas-chromatography combustion isotope ratio mass-spectrometry analysis of urinary steroids. J. Mass Spectrom., 31: 169-176. [8] Agyei E. K. & McMullen C. C. (1968) A study of the isotopic abundance of boron from various sources. Can. J. Earth Sci., 5: 921-927. [7] Ahoron P. & Fu B. (2000) Microbial sulfate reduction rates and sulfur and oxygen isotope fractionations at oil and gas seeps in deepwater Gulf of Mexico. Geochim. Cosmochim. Acta, 64: 233-246. [46] Aiken G., Thurman E. M., Malcolm R. L. & Walton H. E (1979) Comparison of XAD macroporous resins for the concentration of fulvic acid from aqueous solution. Anal. Chem., 51: 335-366. [15] 1. The number(s) between the square hooks at the end of each reference relate to the chapter(s) where that reference is cited.
1054
References
Akagi T., Franchi I. A. & Pillinger T. (1993) Oxygen isotope analysis of quartz by neodymium-yttrium aluminium garnet laser-fluorination. Analyst, 118: 1507-1510. [20] Akagi T., Franchi I. A. & Pillinger T. (1995) Isotope analysis of oxygen in minerals using Nd:YAG laserfluorination: the use of stainless-steel wool trap as a fluorine getter. Geochem. J., 29: 115-122. [20] A1-Aasm I. S., Taylor B. E. & South B. (1990) Stable isotope analysis of multiple carbonate samples using selective acid extraction. Chem. Geol. (Isot. Geosci. Sect.), 80, 119-125. [20] Alderton D. H. M. & Harmon R. S. (1991) Fluid inclusion and stable isotope evidence for the origin of mineralizing fluids in south-west England. Mineral. Magazine, 55: 605-611. [3] Aleu J., Fronza G., Fuganti C., Serra S., Fauhl C., Guillou C. & Reniero F. (2002) Differentiation of natural and synthetic phenylacetic acids by 2H NMR of the derived benzoic acids. Eur. Food Res. Technol., 214: 63-66. [5] Alewell C., Michell M. J. & Likens G. E. (1999) Sources of stream sulfate at the Hubbard Brook experimental forest: Long-term analyses using stable isotopes. Biogeochemistry, 44(3): 281-300. [45] Alibert C., Vicet A., Yarekha D. A., P6rona A., Rouillard Y., Chevrier F., Glastre G. & Baranov A. N. (2001) Gas analysis using room temperature diode lasers in the 2 - 2.5 micron range. Proceedings of the 3rd International Conference on Tunable Diode Laser Spectroscopy, Zermatt, Switserland, July 8-12, 2001. [34] Alisuaskas R. T., Klaas E. E., Hobson K. A. & Ankney C. D. (1998) Stable-carbon isotopes support use of adventitious color to discern winter origins of lesser snow geese. J. Field Ornith., 69: 262-268. [9] Allen R. E. & Parkes R. J. (1995) Digestion procedures for determining reduced sulfur species in bacterial cultures and in ancient and recent sediments. Am. Chem. Soc. Symp. Ser., 612: 243-257. [27] Allenby R. J. (1954) Determination of the isotopic ratios of silicon in rocks. Geochim. Cosmochim. Acta, 5: 40-48. [20, 25] Allison C. E. & Francey R. J. (1995) High precision stable isotope measurements of atmospheric trace gases. In: Reference and intercomparison materials for stable isotopes of light elements. IAEATECDOC-825, IAEA, Vienna: 131-153. [43] Allison C. E., Francey R. J. & Meijer H. A. J. (1995) Recommendations for the reporting of stable isotope measurements of carbon and oxygen in CO2 gas. In: Reference and intercomparison materials for stable isotopes of light elements. IAEA-TECDOC-825, IAEA, Vienna: 155-162. [14, 43] Amaral J. A., Hesslein R. H., Rudd J. W. M. & Fox D. E. (1989) Loss of total sulfur and changes in sulfur isotopic ratios due to drying of lacustrine sediments. Limnol. Oceanogr., 34: 1351-1358. [26] Amaral J. A., Kelly C. A. & Flett R. J. (1993) Measurement of organic-35S and organic-S in lake sediments: Methodological considerations. Biogeochemistry, 23: 61-78. [26] Amberger A. & Schmidt H.-L. (1987) Nati~rliche Isotopengehalte van Nitrat als Indicatoren f~ir dessen Herkunft. Geochim. Cosmochim. Acta, 51: 2699-2705. [15] Ambrose S. H. (1990) Preparation and characterization of bone and tooth collagen for isotopic analysis. J. Archaeol. Sci., 17: 431-451. [9] Ambrose S. H. (1991) Effects of diet, climate and physiology on nitrogen isotope abundances in terrestrial foodwebs. J. Archaeol. Sci., 18: 293-317. [9] Ammons J. M. (1980) Preconcentration methods for the determination of gaseous sulfur compounds in air. PhD Thesis, University of South Florida. [26] An Y. H. & Friedman R. J. (2000) Handbook of Bacterial Adhesion. Humana Press, Totowa, NJ. [46] Anbar A. D., Papanastassiou D. A., Creaser R. A. & Wasserburg G. J. (1997) The determination of iridium in natural waters by clean chemical preconcentration and negative thermal ionization mass spectrometry. Anal. Chem., 69: 2444-2450. [29] Anbar A. D., Zhang C., Barling J. & Roe J. E. (1999) Iron isotope fractionation studies using MC-ICPMS. Ninth Annual V.M. Goldschmidt Conference, Abstract #7612. [31] Anbar A. D., Roe J. E., Barling J. & Nealson K. (2000) Nonbiological fractionation of iron isotopes. Science, 288: 126-128. [29, 31] Anbar A. D., Knab K. A. & Barling J. (2001) Precise determination of mass dependant variations in the isotopic composition of molybdenum using MC-ICPMS. Anal. Chem., 73: 1425-1431. [31] Anderson D. Z., Frisch J. C. & Masser C. S. (1984) Mirror reflectometer based on optical cavity decay time. Appl. Opt., 23: 1238-1245. [34]
References
1055
Anderson S. & Mauersberger K. (1981) Calibration of a mass spectrometer experiment for ozone. Rev. Sci. Instrum., 52: 1025-1028. [18] Anderson S. M. & Kaye J. A. (1987) Comment on "Heavy ozone in the stratosphere". Geophys. Res. Lett., 14: 91-93. [18] Anderson S. M., Morton J. & Mauersberger K. (1989) Laboratory measurements of ozone isotopomers by tunable diode laser absorption spectroscopy. Chem. Phys. Lett., 156: 175-180. [18, 34] Anderson S. M., Mauersberger K., Morton J. & Schueler B. (1992) Heavy ozone anomaly, evidence for a mysterious mechanism. In: Isotope effects in gas-phase chemistry. J. A. Kaye (ed.). Am. Chem. Soc., Washington DC: 155-166. [18] Anderson S. M., Hupalo P. & Mauersberger K. (1993) Rotational structure in the near-infrared absorption spectrum of ozone. J. Chem. Phys., 99: 737-739. [18] Anderson S. M. & Mauersberger K. (1995) Ozone absorption spectroscopy in search of low-lying electric states. J. Geophys. Res., 100: 3033-3048. [18] Anderson S. M., Mauersberger K. & Morton J. (1995) The ozone molecule: Isotope effects and electronic structure. In: Progress and problems in atmospheric chemistry. J. R. Barker (ed.). World Scientific, Signapure: 473-497. [18] Anderson S. M., Hiilsebusch D. & Mauersberger K. (1997) Surprising rate coefficients for four isotopic variants of O + 02 + M. J. Chem. Phys., 107: 5385-5392. [18] Andersen T., Burke E. A. J. & Austrheim H. (1989) Nitrogen-bearing, aqueous fluid inclusions in some eclogites from Western Gneiss Region of the Norwegian Caledonides. Contrib. Mineral. Petrol., 103: 153-165.[16] Andersen T., Burke E. A. J. & Neumann E.-R. (1995) Nitrogen-rich fluid in the upper mantle: fluid inclusions in spinel dunite from Lanzarote, Canary Islands. Contrib. Mineral. Petrol., 120: 20-28. [16] Anderson T. F. (1968) Surface area measurement in calcite grains by isotopic exchange with C14-1abeled carbon dioxide. Geochim. Cosmochim. Acta, 32: 1177-1186. [46] Anderson T. F. (1969) Self-diffusion of carbon and oxygen in calcite by isotope exchange with carbon dioxide. J. Geophys. Res., 74: 3918-3932. [46] Anderson T. F. & Chai B., H.-T. (1974) Oxygen isotope exchange between calcite and water under hydrothermal conditions. In: Geochemical transport and kinetics. A. W. Hofmann, B. J. Giletti, H. S. Jr. Yoder & R. A. Yund (eds.). Carnegie Inst. Washington Publ., Vol. 634:219-227. [46] Anderson T. F., Bopp B. N., Williams A. C., Ho L. Z. & Hudson J. D. (1994) The stable isotopic record of fossils from the Peterborough member, Oxford clay formation (Jurassic), U.K.: Palaeoenvironmental implications. J. Geol. Soc. London, 151: 125-138. [22] Anderson W. B. & Polls G. A. (1998) Marine subsides of island communities in the Gulf of California: evidence from stable carbon and nitrogen isotopes. Oikos, 81: 75-80. [9] Andrawes F. F. & Gibson E. K. Jr. (1979) Release and analysis of gases from geological samples. Am. Mineralogist, 64: 453-463. [3] Andreae M. O. & Raemdonck H. (1983) Dimethylsulfide in the surface ocean and the marine atmosphere: A gobal view. Science, 221: 744-747. [26] Angerosa F., Camera L., Cumitini S., Gleixner G. & Reniero F. (1997) Carbon stable isotopes and olive oil adulteration with pomace oil. J. Agric. Food Chem., 45: 3044-3048. [8] Angert A. & Luz B. (2001) Fractionation of oxygen isotopes by root respiration: implications for the isotopic composition of atmospheric 02. Geochim. Cosmochim. Acta, 65: 1695-1701. [46] Annyas J., Bicanic D. & Schouten F. (1999) Novel instrumental approach to the measurement of total body water: optothermal detection of heavy water in the blood serum. Appl. Spectrosc., 53: 339-343. [34] Anonymous (2000) Porewater extraction from argillaceous rocks for geochemical characterization: methods and interpretation. OECD Publications, Paris: 188 p. [2] Anovitz L. M. & Grew E. S. (1996) Mineralogy, petrology and geochemistry of boron: an introduction. In: Boron: Mineralogy, petrology and geochemistry. E. S. Grew & L. M. Anaovitz (eds.). Mineralogical Society of America, Reviews in Mineralogy, Vol. 33: 1-40. [7] Appelblad P. K. & Baxter D. C. (2000) A model for calculating dead time and mass discrimination correction factors from inductively coupled plasma mass spectrometry calibration curves. J. Anal. At. Spectrom., 15: 557-560. [37].
1056
References
Appora I., Eiler J. M., Matthews A. & Stolper E. M. (2003) Experimental determination of oxygen isotope fractionation between CO2 vapor and soda-melilite melt. Geochim. Cosmochim. Acta, 67: 459471. [46] Aragulis-Aragu~s L., Rozanski K., Gonfiantini R. & Louvat D. (1995) Isotope effects accompanying vacuum extraction of soil water for stable isotope analyses. J. Hydrol., 168: 159-171. [1, 2] Aravena R., Schiff S. L., Trumbore S. E., Dillon P. J. & Elgood R. (1992) Evaluating dissolved inorganic carbon cycling in a forested lake watershed using carbon isotopes. Radiocarbon, 34: 636-645. [10] Aravena R., Evans M. L. & Cherry J. A. (1993) Stable isotopes of oxygen and nitrogen in source identification of nitrate from septic systems. Ground Water, 31: 180-186. [15] Arden J. W. (1983a) Distribution of lead and thallium in the matrix of the Allende meteorite and the extent of terrestrial lead contamination in chrondites. Earth Planet. Sci. Lett., 62: 395-406. [31] Arden J. W. (1983b) Electrochemical separation and isotopic determination of thallium at the nanogram level by surface ionization mass spectrometry. Anal. Chim. Acta, 148: 211-223. [31] Aregbe Y., Valkiers S., Mayer K., De Bi6vre P., Wessel R. M. & Alink A. (1998) Measuring amount ratios of gas isotopes by two primary methods. Metrologia, 35: 7-16. [43] Aregbe Y., Valkiers S., Poths J., Norgaard J., Kipphardt H., De Bi6vre P. & Taylor P. D. P. (2001) A primary isotopic gas standard for krypton with values for isotopic composition and molar mass traceable to the Syst6me International d'Unit6s. Int. J. Mass Spectrom., 206: 129-136. [42] Arlinghaus H. F., Spaar M. T., Tanigaki T., McMahon A. W. & Holloway P. H. (1994) Sputter-initiated Resonance Ionization Spectroscopy - An analytical technique for quantitative and sensitive measurement of impurities and ultra-shallow doping profiles in semiconductors. J. Vac. Sci. Technol., B 12: 263-268. [33] Arlinghaus H. F., Guo X. Q., Whitaker T. J. & Kwoka M. N. (1996) Fish & Chips: Analytical applications of resonance ionization mass spectrometry. Resonance Ionization Spectroscopy 1996, Eight International Symposium. N. Winograd & J.E. Parks (eds.). AIP Press, Woodbury, NY, AIP Conference Proceedings 388: 139-144. [33] Ash R. D., Arden J. W., Grady M. M., Wright I. P. & Pillinger C. T. (1990) Recondite interstellar carbon components in the Allende meteorite revealed by preparative precombustion. Geochim. Cosmochim. Acta, 54: 455-468. [13] Asghar R. B., Diskin A. M., Spanel P., Smith D. & Davies S. J. (2003) Measuring transport of water across the peritoneal membrane. Kidney Int., in press. [4] Asprey L. B. (1976) The preparation of very pure fluorine gas. J. Fluorine Chem., 7: 359-361. [20] ASTM (1958) Tentative method for sulfur in petroleum oil (quartz-tube method). In: Annual Book of ASTM Standards. American Society for Testing and Materials, Philadelphia: 828-832. [26] ASTM (1964) Standard method of test for sulfur in petroleum products by the bomb method. In: Annual Book of ASTM Standards. American Society for Testing and Materials,, Philadelphia. Vol. D129: 76-80. [26] ASTM (1993) Standard test methods for total sulfur in the analysis sample of coal and coke. In: Annual Book of ASTM Standards. American Society for Testing and Materials, Philadelphia. Vol. 05.05, D317789: 333-336. [26] Aston F. W. (1919) The constitution of the elements. Nature, 104: 393. [28] Aston F. W. (1920) The mass-spectra of chemical elements (Part 2). Phil. Magazine, 40: 628-634. [28] Aston F. W. (1942) Mass-spectra and isotopes. Longmans, Green & Co, NY, 2nd ed.: 276 p. [28, 35, 38] Assonov S. S. & Brenninkmeijer C. A. M. (2003) A redetermination of absolute values for 17RVPDB-CO2 and 17RVSMOW.Rapid Commun. Mass Spectrom., 17: 1017-~029. [40] Atekwana A. E. & Krishnamurthy R. V. (1998) Seasonal variations of dissolved inorganic carbon and 613C of surface waters: Application of a modified gas evolution technique. J. Hydrol., 205: 265-278. [10] Atkins P. W. (1988) Molecular quantum mechanics. Oxford University Press, Oxford: 199 p. [33] Atkins P. W. (1990) Physical chemistry. Oxford University Press, Oxford. [4] Attar A. (1979) Sulfur groups in coal and their determination. In: Analytical Methods for Coal and Coal Products. C. Karr (ed.). Academic Press, New York NY, USA, Vol III: 585-624. [26] Attar A. & Dupuis F. (1979) Data on the distribution of organic functional groups in coal. Amer. Chem. Soc. Div., Fuel Chem., 24: 166-177. [26]
References
1057
Aucour A.-M., Tao F.-X., Sheppard S. M. F., Huang N.-W. & Liu C. Q. (2002) Climatic and monsoon isotopic signals (6D, 613C) of northeastern China tree rings. J. Geophys. Res., 107(D9: Atmospheres): 1.1-1.9. [46] Aufderheide A. C., Kelley M. A., Rivera M., Gray L., Tieszen L. L., Iversen E., Krouse H. R. & Carevic A. (1994) Contributions to chemical dietary reconstruction to the assessment of adaption by ancient highland immigrants (also Ramirez) to coastal conditions at Pisagua, North Chile. J. Archaeol. Sci., 21: 515-524. [26] Aursand M., Mabon F. & Martin G. J. (1997) High-resolution 1H and 2H NMR spectroscopy of pure essential fatty acids for plants and animals. Magn. Reson. Chem., 35: $91-$100. [5] Aursand M., Mabon F. & Martin G. J. (2000) Characterization of farmed and wild salmon (Salmo salar) by a combined use of compositional and isotopic analysis. J.A.O.C.S., 77(6): 659-666. [5] Ayliffe L. K., Lister A. M. & Chivas A. R. (1992) The preservation of glacial-interglacial climatic signatures in the oxygen isotopes of elephant skeletal phosphate. Palaeogeogr., Palaeoclimat., Palaeoecol., 99: 179-191. [22] Baer D., Gupta M., O'Keefe A. & Paul J. (2002) Novel ultra-sensitive instrumentation for trace gas measurements. Presentation at the 16th Int. Forum Process Anal. Chem., San Diego. [34] Baertschi P. (1950) Isotopic composition of the oxygen in silicate rocks. Nature, 166: 112-113. [20] Baertschi P. & Silverman S. R. (1951) The determination of relative abundances of oxygen isotopes in silicate rocks. Geochim. Cosmochim. Acta, 1: 317-328. [20, 25] Baertschi P. (1953) Uber die relativen Unterschiede in H2180-Gehalt nati~rlicher Wasser. Helv. Chim. Acta, 36: 1352-1369. [1] Baertschi P. (1976) Absolute 180 content of standard mean ocean water. Earth Planet. Sci. Lett., 31: 341344. [4, 40, 43, 44] Bailey S. A. & Smith J. w. (1972) Improved method for the preparation of sulphur dioxide from barium sulphate for isotope ratio studies. Anal. Chem., 44: 1542-1543. [26] Bailey K., Chen C. Y., Du X., Li Y. M., Lu Z.-T., O'Connor T. P. & Young L. (2000) ATTA - A new method of ultrasensitive isotope trace analysis. Nucl. Instrum. Methods Phys. Res., B 172: 224-227. [34] Bains-Sahota S. K. & Thiemens M. H. (1987) Mass-independent oxygen isotope fractionation in a microwave-plasma. J. Phys. Chem., 91: 4370-4374. [18, 46] Bains-Sahota S. K. & Thiemens M. H. (1989) A mass-independent sulfur isotope effect in the nonthermal formation of $2F10. J. Chem. Phys., 90: 6099-6109. [18] Bakel A. J., Ostrom P. H. & Ostrom N. E. (1994) Carbon isotopic analysis of individual n-alkanes: evaluation of accuracy and application to marine particulate organic material. Org. Geochem., 21: 595-602. [8] Baker L., Franchi I. A., Maynard J., Wright I. P. & Pillinger C. T. (2002) A technique for the determination of 180/160 and 170/160 isotopic ratios in water from small liquid and solid samples. Anal. Chem., 74: 1665-1673. [1] Balabane M., Galimov E., Hermann M. & L6tolle R. (1987) Hydrogen and carbon isotope fractionation during experimental production of bacterial methane. Geochim. Cosmochim. Acta, 51: 115-119. [46] Ball J. D., Crowley S. F. & Steele D. F. (1996) Carbon and oxygen isotope ratio analysis of small carbonate samples by conventional phosphoric acid digestion: sample preparation and calibration. Rapid. Commun. Mass Spectrom., 10(8): 987-995. [12] Ball J. W. (1996) Thermodynamic and isotopic systematics of chromium chemistry. PhD Thesis, University of Arizona, Tucson, Arizona, USA. [29] Ballentine D. C., Macko S. A., Turekian V. C., Gilhooly W. P. & Matincingh B. (1996) Transport of biomass burning products through compound specific stable isotope analysis. Org. Geochem., 25: 97104. [9] Balzar A., Gleixner G., Grupe G., Scmidt H.-L., Schramm S. & Turban-Just S. (1997) In vitrio decomposition of bone collagen by soil bacteria: the implications for stable isotope analysis in archaeometry. Archaeometry, 39: 415-429. [9] Bandura D. R., Baranov V. I. & Tanner S. D. (2001) Reaction chemistry and collisional processes in multiple devices for resolving isobaric interferences in ICP-MS. Fresenius J. Anal. Chem., 370: 454-470. [31]
1058
References
Banfield J. F. & Nealson K. H. (1997) Geomicrobiology: Interactions between Microbes and Minerals. Mineralogical Society of America, Reviews in Mineralogy, Vol. 35:448 p. [46] Bao H. & Koch P. L. (1999) Oxygen isotope fractionation in ferric oxide-water systems: Low temperature synthesis. Geochim. Cosmochim. Acta, 63(5): 599-613. [46] Bao H. & Thiemens M. H. (2000) Generation of 02 from BaSO4 using a CO2-1aser fluorination system for simultaneous analysis of 8180 and 8170. Anal. Chem., 72(17): 4029-4032. [20] Bao H., Koch P. L. & Thiemens M. H. (2000a) Oxygen isotopic composition of ferric oxides from recent soil, hydrologic, and marine environments. Geochim. Cosmochim. Acta, 64: 2221-2231. [46] Bao H., Thiemens M. H., Farquhar J., Campbell D. A., Lee C. C.-W., Heine K. & Loope D. B. (2000b) Anomalous 170 compositions in massive sulphate deposits on the Earth. Nature, 406: 176-178. [18] Barbeni M., Cisero M. & Fuganti C. (1997) Natural abundance 2H nuclear magnetic resonance study of the origin of (Z)-3-hexenol. ]. Agric. Food Chem., 45: 237-241. [5] Barbieri B., Beverini N. & Sasso A. (1990) Optogalvanic spectroscopy. Rev. Mod. Phys., 62: 603-644. [34] Bariac T., Jusserand C. & L6tolle R. (1982) A routine method for 180 analysis of water vapor. Int. ]. AppI. Radiat. Isot., 33: 1365-1368. [1] Bariac T., Deleens E., Gerbaud A. & Mariotti A. (1991) La composition isotopique (180, 2H) de la vapeur d'eau transpir6e: 6tude en conditions asservies. Geochim. Cosmochim. Acta, 55: 3391-3402. [46] Bariac T., Gonzalez-Dunia J., Tardieu F., Tessier D. & Mariotti A. (1994) Variabilit6 spatiale de la composition isotopique del'eau (180, 2H) au sein des organes des plantes a6riennes: 1. Approche en conditions contr616es. Chem Geol., 115: 307-315. [21] Barker C. & Torkelson B. E. (1975) Gas adsorption on crushed quartz and basalt. Geochim. Cosmochim. Acta, 39: 212-218. [3] Barnes H. L. (1963) Ore solution chemistry. I. Experimental determination of mineral solubilities. Econ. Geol., 58: 1054-1060. [46] Barnes I. L., Garner E. L., Gramlich J. W., Moore L. J., Murphy T. J., Machlan L. A., Shileds W. R., Tatsumoto M. & Knight R. J. (1973) Determination of lead, uranium, thorium and thallium in silicate glass standard materials by isotope dilution mass spectrometry. Anal. Chem., 45: 880-885. [35] Barnes I. L., Moore L. J., Machlan L. A., Murphy T. J. & Shields W. R. (1975) Absolute isotopic abundance ratios and atomic weight of a reference sample of silicon. ]. Research Nat. Bureau Stand., 79A: 727-735. [25] Barnes J. H. IV, Sperline R., Denton M. B., Barinaga C. J., Koppenaal D. W., Young E. T. & Hieftje G. M. (2002) Characterization of a Focal Plane Camera Fitted to a Mattauch-Herzog Geometry Mass Spectrograph: 1. Use with a Glow-Discharge Source. Anal. Chem., 74: 5327-5332. [35] Barnes L., Downes C. T. & Hulston R. J. (1978) Warm Springs, South Island, New Zealand, and their potential to yield laumontite. Am. J. Sci., 278: 1412-1427. [10] Barnes R. M. (1998) Plasma source mass spectrometry in experimental nutrition. In: Mathematical modelling in experimental nutrition. A. J. Clifford & H.-G. M/iller (eds.),Vol. 46, Plenum Press: 379396. [31] Barnhart J. (1997) Chromium chemistry and implications for environmental fate and toxicity. In: Chromium in soil; perspectives in chemistry, health, and environmental regulation. D. M. Proctor, B. L. Finley, D. J. Harris, D. J. Paustenbach & D. Rabbe (eds.). Lewis Publishers: 561-568. [29] Baronnet A. (1982) Ostwald ripening in solution. The case of calcite and mica. Estudios Geol. (Madrid), 38: 185-198. [46] Barrer R. M. (1978) Zeolites and clay minerals as sorbents and molecular sieves. Academic Press: 497 p. [36]. Barrie A. & Coward W. A. (1985) A rapid analytical technique for the determination of enery expenditure by the doubly labelled water method. Biomed. Mass Spectrom., 12: 535-541. [1] Barrie A. (1991) New methodologies in stable isotope analysis. In: Stable isotopes in plant nutrition, soil fertility and environmental studies. IAEA-SM-313 / 106, Proceedings Series, IAEA, Vienna: 3-25. [15, 24] Barrie A. & Prosser S. J. (1996) Automated analysis of light-element stable isotopes by isotope ratio mass spectrometry. In: Mass spectrometry of soils. T. W. Boutton & S. Y. Yamasaki (eds.). Marcel Dekker Inc, NY: 1-46. [1, 9, 15]
References
1059
Barshick C. M., Smith D. H., Wade J. W. & Bayne C. K. (1994) Isotope dilution glow discharge mass spectrometry as applied to the analysis of lead in waste oil samples. J. Anal. At. Spectrom., 9: 83-87. [351 Barth S. (1993) Boron isotope variations in nature: a synthesis. Geol. Rundschau, 82: 640-651. [7] Barth S. (1997) Boron isotopic analysis of natural fresh and saline waters by negative thermal ionization mass spectrometry. Chem Geol., 143: 255-261. [7] Barth S. (1998) Application of boron isotopes for tracing sources of anthropogenic contamination in groundwater. Water Res., 32: 685-690. [7] Bartholomew R. M., Brown F. & Lounsbury M. (1954) Chlorine isotope effect in reactions of tert-butyl chloride. Can. J. Chem., 32: 979-983. [28] Bartoli F., Burtin G. & Herbillon A. J. (1991) Disaggregation and clay dispersion of oxisols: Na resin, a recommended methodology. Geoderma, 49: 301-317. [2] Bates A. L., Spiker E. C., Orem W. H. & Burnett W. C. (1993) Speciation and isotopic composition of sulfur in sediments from Jellyfish Lake, Palau. Chem. Geol., 106: 63-76. [26] Bates D. R. (1986) Heavy ozone in the stratosphere. Geophys. Res. Lett., 13: 664-666. [18] Bates D. R. (1988) Suggested explanation of heavy ozone. Geophys. Res. Lett., 15: 13-16. [18] Bates D. R. (1990) Isotopic fractionation in the formation of ozone and carbon dioxide. J. Chem. Phys., 93: 8739-8744. [18] Battagliarin M., Sentimenti E. & Scattolin R. (1995) An innovative sample preparation procedure for trace and ultra-trace analysis on non-conducting powders by direct current glow discharge mass spectrometry. Spectrochim. Acta, B 50: 13-25. [35] Baylis S. A., Hall K. & Jumeau E. J. (1994) The analysis of the C1-C5 components of natural-gas samples using gas-chromatography combustion isotope ratio mass-spectrometry. Org. Geochem., 21: 777-785.
[8]
Beard B. L. & Johnson C. M. (1999) High-precision iron isotope measurements of terrestrial and lunar materials. Geochim. Cosmochim. Acta, 63: 1653-1660. [29, 31] Beard B. L., Johnson C. M., Cox L., Sun H., Nealson K. & Aguilar C. (1999) Iron isotope biosignatures. Science, 285: 1889-1892. [29, 31, 46] Beard B. L., Johnson C. M., Skulan J. L., Nealson K. H., Cox L. & Sun H. (2003) Aplication of Fe isotopes to tracing the geochemical and biological cycling of Fe. Chem. Geol., 195: 87-117. [46] Bearhop S., Thompson D. R. & Waldron S. (1999) Stable isotopes indicate the extent of freshwater feeding by cormorants Phalacrocorax carbo shot at inland fisheries in England. J. Appl. Ecology, 36: 75-84. [91 Bearhop S., Teece M. A., Waldron S. & Waldron S. F. (2000) The influence of lipid and uric acid upon 613C and h15N values of avain blood: Implications for trophic studies. The Auk, 117: 504-507. [9] Beaudoin G. & Taylor B. E. (1993) MILES laser microprobe. Part 2: Preliminary assessment of precision and accuracy of sulphur isotope analysis. Geol. Surv. Canada, Current Res., Part D, Paper 93-ID: 199204. [20] Beaudoin G. & Taylor B. E. (1994) High precision and spatial resolution sulfur isotope analysis using MILES laser microprobe. Geochim. Cosmochim. Acta, 58(22): 5055-5063. [20] Beaudoin G., Taylor B. E., Rumble D. III& Thiemens M. (1994) Variations in the sulfur isotope composition of troilite from Canon Diablo iron meteorite. Geochim. Cosmochim. Acta, 58: 4253-4255. [26, 45] Beaudoin G. & Taylor B. E. (1995) Comparison of the environmental-SO2 and the laser-SF6 methods: implications for the sulfur isotope scale. In: Reference and intercomparison materials for stable isotopes of light elements. Proceedings of a Consultants Meeting held in Vienna, 1-3 Dec., 1993. IAEA, Vienna. IAEA-TECDOC-825: 49-50. [40] Beaudoin G., Taylor B. E., Rumble D. III & Thiemens M. (1995) Variations in the sulfur isotope composition of CDT. In: Reference and intercomparison materials for stable isotopes of light elements. Proceedings of a Consultants Meeting held in Vienna, 1-3 Dec., 1993. IAEA, Vienna. IAEATECDOC-825: 47-48. [40] Beaufort D., Dudoignon P., Proust D., Parneix J. C. & Meunier A. (1983) Microdrilling in thin section: a useful method for the identification of clay minerals in situ. Clay Minerals, 18: 219-222. [2] Beaumont V. & Robert F. (1999) Nitrogen isotope ratios of kerogens in Precambrian cherts: a record of the evolution of atmospheric chemistry? Precambrian Res., 96: 63-82. [13]
1060
References
Bebout G. E. & Fogel M. L. (1990) Techniques for extraction of nitrogen from silicate minerals and whole-rock samples for isotopic analysis. Ann. Rept. Dir. Geophys. Lab. 1989-1990, No. 2200. Carnegie Institution, Washington: 123-128. [36] Bebout G. E. & Fogel M. L. (1992) Nitrogen-isotope compositions of metasedimentary rocks in the Catalina Schist, California: implications for metamorphic devolatilization history. Geochim. Cosmochim. Acta, 56: 2839-2849. [16] Bebout G. E. & Barton M. D. (1993) Metasomatism during subduction: Products and possible paths in the Catalina Schist, California. Chem Geol., 108: 61-92. [16] Bebout G. E. (1995) Impact of subduction-zone metamorphism on mantle-ocean chemical cycling. Chem. Geol., 126: 191-218. [16] Bebout G. E. (1997) Nitrogen-isotope tracers of high-temperature fluid-rock interactions: case study of the Catalina Schist, California. Earth Planet. Sci. Lett., 151: 77-95. [16] Bebout G. E., Cooper D. C., Bradley A. D. & Sadofsky S. J. (1999a) Nitrogen-isotope study of fluid-rock interactions in the Skiddaw Aureole and Granite, English Lake District. Am. Mineralogist, 84: 14951505. [16] Bebout G. E., Ryan J. G., Leeman W. P. & Bebout A. E. (1999b) Fractionation of trace elements by subduction-zone metamorphism: Effect of convergent margin thermal evolution. Earth Planet. Sci. Lett., 171: 63-81. [16] Becchi M., Aguilera R., Farizon Y., Flament M., Casabianca H. & James P. (1994) Gas-chromatography combustion isotope ratio mass-spectrometry analysis of urinary steroids to detect misuse of testosterone in sport. Rapid Commun. Mass Spectrom., 8: 304-308. [8] Bechtel A. & Hoernes S. (1985) Sauerstoffisotopenuntersuchungen an OH-haltigen Silikaten: Fraktionierung zwischen OH- und Silikat-Sauerstoff. Fortschr. Mineral., 63(Suppl. 1): 18. [20] Bechtel A. & Hoernes S. (1990) Oxygen isotope fractionation between oxygen of different sites in illite minerals: a potential single-mineral thermometer. Contrib. Mineral. Petrol., 104: 463-470. [2, 20, 46] Bechtel A. & Hoernes S. (1993) Stable isotopic variations of clay minerals: A key to the understanding of Kupferschiefer-type mineralization, Germany. Geochim. Cosmochim. Acta, 57: 1799-1816. [2] Becker J. F., Sauke T. B. & Loewenstein M. (1992) Stable isotope analysis using diode laser spectroscopy. Appl. Opt., 31: 1921-1927. [34] Becker J. S., Breuer U., Westheide J., Saprykin A. I., Holzbrecher H., Nickel H. & Dietze H.-J. (1996) Trace and surface analysis of ceramic layers of solid oxide fuel cells by mass spectrometry. Fresenius J. Anal. Chem., 335: 626-632. [35] Becker J. S. & Dietze H.-J. (1998) Ultratrace and precise isotope analysis by double-focusing sector field inductively coupled plasma mass spectrometry. J. Anal. At. Spectrom., 13: 1057-1063. [35] Becker J. S. & Dietze H.-J. (2003) State-of-the-art in inorganic mass spectrometry for analysis of highpurity materials. Int. J. Mass Spectrom., 228: 127-150. [35] Becker R. H. & Clayton R. N. (1975) Nitrogen abundances and isotopic compositions in lunar samples. 6th Lunar Sci. Conf., Proceedings: 2131-2149. [17] Becker R. H. & Clayton R. N. (1977) Nitrogen isotopes in igneous rocks. EOS Trans. AGU, 58:536 (Abstract). [16] Becker R. H. & Epstein S. (1982) Carbon, hydrogen and nitrogen isotopes in solvent extractable organic matter from carbonaceous chondrites. Geochim. Cosmochim. Acta, 34: 257-278. [11] Begley I. S. & Sharp B. L. (1994) Occurence and reduction of noise in inductively coupled plasma mass spectrometry for enhanced precision in isotope ratio measurement. J. Anal. At. Spectrom., 9: 171-176. [32] Begley I. S. & Scrimgeour C. M. (1996) On-line reduction of H20 for 62H and 6180 measurement by continuous-flow isotope ratio mass spectrometry. Rapid Comm. Mass Spectrom., 10: 969-973. [1, 8] Begley I. S. & Scrimgeour C. M. (1997) High-precision 62H and 6180 measurement for water and volatile organic compounds by continuous flow pyrolysis isotope ratio mass spectrometry. Anal. Chem., 69: 1530-1535. [1, 4, 8, 23] Begley I. S. & Sharp B. L. (1997) Characterisation and correction of instrumental bias in inductively coupled plasma quadrupole mass spectrometry for accurate measurement of lead isotope ratios. J. Anal. At. Spectrom., 12: 395-402. [32]
References
1061
Begun G. M. & Landau L. (1961) Mass spectra and metastable transitions in isotopic nitrous oxide. J. Chem. Phys., 35: 547-551. [19] Behne W. (1953) Untersuchungen zur Geochemie des Chlor und Brom. Geochim. Cosmochim. Acta, 3: 186214. [28] Belshaw N. S., O'Nions R. K., Martel D. J. & Burton K. W. (1994) High resolution SIMS analysis of common lead. Chem. Geol., 112: 57-70. [35] Belshaw N. S., Freedman P. A., O'Nions R. K., Frank M. & Guo Y. (1998) A new variable dispersion double-focusing plasma mass spectrometer with performance illustrated for Pb-isotopes. Int. J. Mass Spectrom., 181: 51-58. [31] Belshaw N. S., Zhu X. K., Guo Y. & O'Nions R. K. (2000) High precision measurement of iron isotopes by plasma source mass spectrometry. Int. J. Mass Spectrom., 197: 191-195. [31] Belsky T. & Kaplan I. R. (1971) Light hydrocarbon gases, 13C and origin of organic matter in carbonaceous chondrites. Geochim. Cosmochim. Acta, 34: 257-278. [11] Belyaev S. S., Wolkin R., Kenealy W. R., DeNiro M. J., Epstein S. & Zeikus J. G. (1983) Methanogenic bacteria from the Bondyuzhskoe oil field: general characterization and analysis of stable-carbon isotopic fractionation. Applied Environ. Microbiol., 45: 691-697. [46] Ben-David M., Flynn R. W. & Schell D. M. (1997) Annual and seasonal changes in diets of martens: evidence from stable isotope analysis. Oecologia, 111: 280-291. [9] Bender M., Labeyrie L. D., Raynaud D. & Lorius C. (1985) Isotopic composition of atmospheric 02 in ice linked with deglaciation and global oprimary productivity. Nature, 318: 349-352. [21] Bender M., Sowers T. & Labeyrie L. (1994) The Dole effect and its variation during the last 130,000 years as measured in the Vostok ice core. Global Biogeochem. Cycles, 8(3): 363-376. [21] Benner R., Fogel M. L., Sprague E. K. & Hodson R. E. (1987) Depletion of 13C in lignin and its implications for stable isotope studies. Nature, 329: 708-710. [9] Benninghoven A., Riidenauer F. G. & Werner H. W. (1987) Secondary ion mass spectrometry: Basic concepts, instrumental aspects, applications and trends. Chemical Analysis Monograph, Vol. 86. Wiley & Sons, NY: 1216 p. [30, 39] Berden G., Peeters P. & Meijer G. (2000) Cavity ring-down spectroscopy: Experimental schemes and applications. Int. Rev. Chem. Phys., 19: 565-607. [34] Bergamaschi P., Schupp M. & Harris G. W. (1994) High-precision direct measurements of 13CH4/ 12CH4 and 12CH3D/12CH4 ratios in atmospheric methane sources by means of a long-path tunable diode laser absorption spectrometer. Applied Optics, 33(33): 7704-7716. [34] Bernard C., Bocquier G. & Javoy M. (1976) D6termination des rapports isotopiques 180/160 de la gibbsite et de la boehmite pr6sentes dans diff6rentes bauxites. C. R. Acad. Sci. Paris, 282: 1089-1092. [46] Berndt M. E., Seal R. R. II, Shanks W. C. III & Seyfried W. E. Jr. (1996) Hydrogen isotope systematics of phase separation in submarine hydrothermal systems: experimental calibration and theoretical models. Geochim. Cosmochim. Acta, 60: 1595-1604. [46] Berner R. A., Petsch S. T., Lake J. A., Beerling D. J., Popp B. N., Lane R. S., Laws E. A., Westley M. B., Cassar N., Woodward F. I. & Quick W. P. (2000) Isotopic fractionation and atmospheric oxygen: implications for Phanerozoic 02 evolution. Science, 287: 1630-1633. [46] Berner U. & Faber E. (1996) Empirical carbon isotope/maturity relationships for gases from algal kerogens and terrigenous organic matter, based on dry, open-system pyrolysis. Org. Geochem., 24: 947-955. [8] Bert D., Leavitt S. W. & Dupouey J.-L. (1997) Variations of wood ~13C and water-use efficiency of Abies alba during the last century. Ecology, 78: 1588-1596. [9] Bettany J. R., Stewart J. w. B. & Saggar S. (1979) The nature and forms of sulfur in organic matter fractions of soils selected along an environmental gradient. Soil Sci. Soc. Am. J., 43: 981-985. [26] Betti M., Giannarelli S., Hiernaut T., Rasmussen G. & Koch L. (1996a) Detection of trace radioisotopes in soil, sediment, and vegetation by glow-discharge mass spectrometry. Fresenius J. Anal. Chem., 355: 642-646. [35] Betti M., Rasmussen G. & Koch L. (1996b) Isotopic abundance measurements on solid nuclear-type samples by glow discharge mass spectrometry. Fresenius J. Anal. Chem., 355: 808-812. [35] Betz G. (1980) Surface sputtering. Surface Sci., 92: 283-309. [35]
1062
References
Betz G. & Wien K. (1994) Energy and angular distributions of sputtered particles. Int. J. Mass Spectrom. Ion Proc., 140: 1-110. [35] Bhattacharya S. K. & Thiemens M. H. (1989a) New evidence for symmetry dependent isotope effects: O + CO reaction. Z. Naturforsch., 44a: 435-444. [18] Bhattacharya S. K. & Thiemens M. H. (1989b) Effect of isotopic exchange upon symmetry dependent fractionation in the O + CO ~ CO2 reaction. Z. Naturforsch., 44a: 811-813. [18] Bickle M. J., Chapman H. J. & You C. F. (2000) Measurement of lithium isotopic ratios as lithium tetraborate ions. Int. J. Mass Spectrom., 202: 273-282. [6] Bidigare R. R., Fluegge A., Freeman K. H., Hanson K. L., Hayes J. M., Hollander D., Jasper J. P., King L. L., Laws E. A., Milder J., Millero F. J., Pancrost R., Popp B. N., Steinberg P. A. & Wakeham S. G. (1997) Consistent fractionation of 13C in nature and in the laboratory: Growth-rate effects in some haptophyte algae. Global Biogeochem. Cycles, 11: 279-292. [9] Biederbeck V. O. (1978) Soil organic sulfur and fertility. In: Soil Organic Matter. M. Schnitzer & S. U. Khan (eds.). Elsevier, New York: 273-310. [26] Bigeleisen J., Perlman M. L. & Prosser H. C. (1952) Conversion of hydrogenic materials to hydrogen for isotopic analysis. Anal. Chem., 24: 1356-1357. [1, 2, 3, 21] Bigeleisen J. & Wolfsberg M. (1958) Theoretical and experimental aspects of isotope effects in chemical kinetics. Adv. Chem. Phys., 1: 15-76. [18, 45] Bigeleisen J. (1965) Chemistry of isotopes. Science, 147: 463-471. [46] Bigeleisen J., Lee M. W. & Mandel F. (1973) Equilibrium isotope effects. Ann. Rev. Phys. Chem., 24: 407440. [46] Bijma J., Hemleben C., Huber B. T., Erlenkeuser H. & Kroon D. (1998) Experimental determination of the ontogenetic stable isotope variability in two morphotypes of Globigerinella siphonifera (d'Orbigny). Marine Micropaleo., 35: 141-160. [46] Birck J. L. & All6gre C. J. (1988) Manganese-chromium isotope systematics and the development of the early solar system. Nature, 331(No. 6157): 579-584. [29] Birck J. L. & Lugmair G. W. (1988) Nickel and chromium isotopes in Allende inclusions. Earth Planet. Sci. Lett., 90(2): 131-143. [29] Bird D. K., Cho M., Janik C. J., Liou J. G. & Caruso L. J. (1988) Compositional, order/disorder, and stable isotope characteristics of A1-Fe epidote, state 2-14 drill hole, Salton Sea geothermal system. J. Geophys. Res., 93(Bll): 13135-13144. [46] Bird M. I., Chivas A. R. & Andrew A. S. (1989) A stable-isotope study of lateritic bauxites. Geochim. Cosmchim. Acta, 53: 1411-1420. [2, 46] Bird M. I., Chivas A. R. & Andrews A. S. (1990) Reply to comments by C.-H. Chen, Liu K.-K. and Shieh Y.-N. on "A stable-isotope study of lateritic bauxites". Geochim. Cosmochim. Acta, 54: 1485-1486. [46] Bird M. I., Longstaffe F. J., Fyfe W. S. & Bildgen P. (1992) Oxygen-isotope systematics in a multiphase weathering system in Haiti. Geochim. Cosmochim. Acta, 56: 2831-2838. [2] Bird M. I., Longstaffe F. J. & Fyfe W. S. (1993) Oxygen-isotope fractionation in titanium-oxide minerals at low temperature. Geochim. Cosmochim. Acta, 57: 3083-3091. [2, 46] Bird M. I., Summons R. E., Gagan M. K., Roksandic Z., Dowling L., Head J., Fifield L. K., Cresswell R. G. & Johnson D. P. (1995) Terrestrial vegetation change inferred from n-alkane 613C analysis in the marine environment. Geochim. Cosmochim. Acta, 59: 2853-2857. [9] Bishop P. K. (1990) Precipitation of dissolved carbonate species from natural waters for 613C analysis A critical appraisal. Chem. Geol. (Isot. Geosci. Sect.), 80: 251-259. [10] Bizarro M., Simoneitti A., Stevenson R. K. & Kurszlaukis S. (2003) In situ 87Sr/86Sr investigation of igneous apatites and carbonates using laser-ablation MC-ICP-MS. Geochim. Cosmochim. Acta, 67: 289302. [20] Bjoroy M., Hall K., Gillyon P. & Jumeau J. (1991) Carbon isotope variations in n-alkanes and isoprenoids of whole oils. Chem. Geol., 93: 13-20. [9] Bjoroy M., Hall K. & Moe R. P. (1994) Stable carbon-isotope variation of n-alkanes in central graben oils. Org. Geochem., 22: 355-381. [8] Blair N., Leu A., Munoz E., Olsen J., Kwong E. & des Marais D. (1985) Carbon isotopic fractionation in heterotrophic microbial metabolism. Appl. Environm. Microbiol., 50: 996-1001. [9]
References
1063
Blake R. E., O'Neil J. R. & Garcia G. A. (1997) Oxygen isotope systematics of biologically mediated reactions of phosphate: I. Microbial degradation of organophosphoris compounds. Geochim. Cosmochim. Acta, 61: 4411-4422. [22, 46] Blake R. E., O'Neil J. R. & Garcia G. A. (1998) Effects of microbial activity on the 6180 of dissolved inorganic phosphate and textural features of synthetic apatites. Am. Mineralogist, 83: 1516-1531. [22] Blamart D., Graham C. M. & Sheppard S. M. F. (1986) The system NaA1Si308 (melt)-H20-seawater: solubility of water and hydrogen isotope fractionation factor at 3, 5 and 8 kb. Intern. Symp. Experimental Mineral. Geochem., Nancy, France: 23-24 (Abstract). [46] Blanchar R. W. (1986) Measurement of sulfur in soils and plants. In: Sulfur in Agriculture. M. A. Tabatabai (ed.). American Society of Agronomy - Crop Science Society of America - Soil Science Society of America, Madison, Wisconsin, USA. Monograph No. 27: 455-490. [26] Blatt H. & Brown V. M. (1974) Prophylactic separation of heavy minerals. J. Sedim. Petrol., 44: 260-261. [44] Blatt H. (1987) Oxygen isotopes and the origin of quartz. J. Sediment. Petrol., 57: 373-377. [2] Blattner P. (1973) Oxygen from liquids for isotopic analysis and a new determination of RCO2-H20 at 25~ Geochim. Cosmochim. Acta, 37: 2691-2693. [1] Blattner P. (1975) Oxygen isotope composition of fissure-grown quartz, adularia and calcite from Broadlands geothermal fields, New Zealand with an appendix on quartz-K-feldspar-calcitemuscovite oxygen isotope geothermometers. Am. J. Sci., 275: 785-800. [46] Blattner P. & Hulston J. R. (1978) Proportional variations of geochemical 6180 scales - an interlaboratory comparison. Geochim. Cosmochim. Acta, 42: 59-62. [43] Blattner P., Braithwaite R. B. & Glover R. B. (1983) New evidence on magnetite oxygen isotope geothermometers at 175~ and 112~ in Waikarei steam pipelines (New Zealand). Isot. Geosc., 1: 195-204. [29, 46] Blau K. & Halket J. (1993) Handbook for derivatives of chromatography. John Wiley & Sons, West Sussex, England. [9] Blaum K., Bushaw B. A., Geppert C., Miiller P., N/~rtersh~iuser W., Schmitt A., Trautmann N. & Wendt K. (1998) Diode-laser-based resonance ionization mass spectrometry of gadoliunium. Resonance Ionization Spectroscopy, Ninth International Symposium. J.C. Vickerman, I. Lyon, N.P. Lockyer & J.E. Parks (eds.). AlP Press Woodbury, NY. AIP Conference Proceedings 454:275 p. [33] Bligh E. G. & Dyer W. J. (1959) A rapid method of total lipid extraction and purification. Can. J. Biochem. Physiol., 37: 911-917. [9] Blum J. S., Bindi A. B., Buzzelli J., Stolz J. & Oremland R. S. (1998) Bacillus arsenicoselenatis, sp. nov., and Bacillus selenitireducens, sp. nov.: two haloalkaliphiles from Mono Lake, California that respire oxyanions of selenium and arsenic. Arch. Microbiol., 171: 19-30. [29] Blunier T., Chappellaz J. A., Schwander J., Barnola J. M., Desperts T., Stauffer B. & Raynaud D. (1993) Atmospheric methane, record from a Greenland ice core over the last 1000 year. Geophys. Res. Lett., 20(20): 2219-2222. [14] Bogaerts A. & Gijbels R. (1996) Relative sensitivity factors in glow discharge mass spectrometry. The role of charge transfer ionization. J. Anal. At. Spectrom., 11: 841-847. [35] Bogaerts A. & Gijbels R. (1998) Fundamental aspects and applications of glow discharge spectrometric techniques. Spectrochim. Acta, B 53: 1-42. [35] Bogaerts A. & Gijbels R. (2000) Behavior of the sputtered copper atoms, ions and excited species in a radio-frequency and direct current glow discharge. Spectrochim. Acta, B 55: 279-297. [35] Bogaerts A., Chen Z. & Gijbels R. (2003) Glow discharge modelling: from basic understanding towards applications. Surf. Interface Anal., 35: 593-603. [35] Bohn H. L., Barrow N. J., Rajan S. S. S. & Parfitt R. L. (1986) Reactions of inorganic sulfur in soils. In: Sulfur in Agriculture. M. A. Tabatabai (ed.). American Society of Agronomy - Crop Science Society of America - Soil Science Society of America, Madison, Wisconsin, USA. Monograph No. 27: 233-249. [26] Bol R. A. & Harkness D. D. (1995) The use of zeolite molecular sieves for trapping low concentrations of CO2 from environmental atmospheres. Radiocarbon, 37: 643-647. [36] Borella S. (1994) Jahrringstudien von 13C und 180 in Hinsicht auf Klimarekonstruktionen. Diploma Thesis, Universit~it Bern, Switserland. [24]
1064
References
Borella S. (1998) Stable isotope measurements of oxygen and carbon in tree rings: Method improvements and synergy of both parameters for ecological and climatological studies. PhD Thesis, University of Bern, Switserland. [24] Borella S., Leuenberger M., Saurer M. & Siegwolf R. (1998) Reducing uncertainties in ~)13Canalysis of tree rings: Pooling, milling, and cellulose extraction. J. Geophys. Res., 103(D16): 19519-19526. [24] Borella S., Leuenberger M. & Saurer M. (1999) Analysis of 6180 in tree rings: Wood-cellulose comparison and method dependent sensitivity. J. Geophys. Res., 104(D16): 19519-19526. [23, 24] Borggaard O. K. (1982) Selective extraction of amorphous iron oxides by EDTA from selective silicates and mixtures of amorphous and crystalline iron oxides. Clay Minerals, 17: 365-368. [2] Borthwick J. & Harmon R. S. (1982) A note regarding C1F3 as an alternative to BrF5 for oxygen isotope analysis. Geochim. Cosmochim. Acta, 46: 1665-1668. [2, 20] Boschetti A., Bassi D., Iacob E., Ianetti S., Ricci L. & Scotoni M. (2002) Resonant photoacoustic simultaneous detection of methane and ethylene by means of a 1.63 ~m diode laser. Appl. Phys., B 74: 273278. [34] Bottinga Y. (1969) Calculated fractionation factors for carbon and hydrogen isotope exchange in the system calcite-carbon dioxide-graphite-methane-water vapour. Geochim. Cosmochim. Acta, 33: 49-64. [13] Bottinga Y. & Craig H. (1969) Oxygen isotope fractionation between CO2 and water, and the isotopic composition of marine atmospheric CO2. Earth Planet. Sci. Lett., 5: 285-295. [1] Bottinga Y. & Javoy M. (1973) Comments on oxygen isotope geothermometry. Earth Planet. Sci. Lett., 20: 250-265. [46] Bottinga Y. & Javoy M. (1975) Oxygen isotope partitioning among the minerals in igneous and metamorphic rocks. Rev. Geophys. Space Physics, 13: 401-418. [46] Bottomley D. J., Veizer J., Nielsen H. & Moczydlowska M. (1992) Isotopic composition of disseminated sulfur in Precambrian sedimentary rocks. Geochim. Cosmochim. Acta, 56: 3311-3322. [26] Bottomley D. J., Katz A., Chan L. H., Starinsky A., Douglas M., Clark I. D. & Raven K. G. (1999) The origin and evolution of Canadian Shield brines: evaporation or freezing of seawater: New lithium isotope and geochemical evidence from the Slave craton. Chem. Geol., 155: 295-320. [6] Bottomley D. J., Chan L. H., Katz A, Starinsky A. & Clark I. D (2003) Lithium isotope geochemistry and origin of Canadian shield brines. Ground Water, in press. [6] Bottrell S. H., Carr L. P. & Dubessy J. (1988) A nitrogen-rich metamorphic fluid and coexisting minerals in slates from North Wales. Mineral. Magazine, 52: 451-457. [16] Bottrell S. H., Louie P. K. K., Timpe R. C. & Hawthorne S. B. (1994) The use of stable isotope ratio analysis to access selectivity of chemical analyses and extractions of forms of sulphur in coal. Fuel, 73: 1578-1582. [26] Bottrell S. H. & Novak M. (1997) Sulphur isotopic study of two pristine Shagnum bogs in the western British Isles. J. Ecology, 85: 125-132. [26] Bouman C., Vroon P. Z., Elliott T. R., Schwieters J. B. & Hamester M. (2002) Determination of lithium isotope compositions by MC-ICPMS (Thermo Finnigan MAT Neptune). Geochim. Cosmochim. Acta, 66:A97 (Abstract). [6] Bourg C., Stievenard M. & Jouzel J. (2001) Hydrogen and oxygen isotopic composition of aqueous salt solutions by gas-water equilibration method. Chem. Geol., 173: 331-337. [46] Boutton T. W. (1991) Stable carbon isotope ratios of natural materials 1. Sample preparation and mass spectrometer analysis. In: Carbon isotope techniques. D. C. Coleman& B. Fry (eds.). Academic Press, San Diego, USA: 155-171. [9] Bowling D. R., Sargent S. D., Tanner B. D. & Ehleringer, J. R. (2003) Tunable diode laser absorption spectroscopy for stable isotope studies of ecosystem-atmosphere CO2 exchange. Agric. Forest Meteorol., 118: 1-19. [34] Boyd J. R. & England J. L. (1960) Apparatus for phase-equilibrium measurements at pressures up to 50 kilobars and temperatures up to 1750~ J. Geophys. Res., 65: 741-748. [46] Boyd S. R., Mattey D. P., Pillinger C. T., Milledge H. J., Mendelssohn M. & Seal M. (1987) Multiple growth events during diamond genesis: an integrated study of carbon and nitrogen aggregation state in coated stones. Earth Planet. Sci. Lett., 86: 341-353. [16]
References
1065
Boyd S. R., Wright I. P., Franchi I. A. & Pillinger C. T. (1988) Preparation of sub-nanomole quantities of nitrogen gas for stable isotopic analysis. J. Phys. E: Sci. Instrum., 21: 876-885. [13, 17, 36] Boyd S. R. & Pillinger C. T. (1990) Determination of the abundance and isotopic composition of nitrogen within organic compounds: a sealed tube technique for use with static vacuum mass spectrometers. Meas. Sci. Technol., 1: 1176-1183. [13, 16] Boyd S. R. & Pillinger C. T. (1991) Rubidium sulfate - ammonium sulfate solid solution: a standard for use during the determination of the abundance and isotopic composition of nitrogen, at the ppm level, by static vacuum mass spectrometry. Anal. Chem., 63: 1332-1335. [13] Boyd S. R., Hall A. & Pillinger C. T. (1993a) The measurement of 615N in crustal rocks by static vacuum mass spectrometry: application to the origin of the ammonium in the Cornubian batholith, southwest England. Geochim. Cosmochim. Acta, 57: 1339-1347. [13, 16] Boyd S. R., Wright I. P. & Pillinger C. T. (1993b) Accurate determination of nitrogen concentrations by static vacuum mass spectrometry. Meas. Sci. Technol., 4: 1000-1005. [13, 17] Boyd S. R., Rejou-Michel A. & Javoy M. (1994) Non-cryogenic purification of nanomole quantities of nitrogen gas for isotopic analysis. Anal. Chem., 66: 1396-1402. [13] Boyd R. S., Rejou-Michel A. & Javoy M. (1995) Improved techniques for the extraction, purification and quantification of nanomole quantities of nitrogen gas: the nitrogen content of diamond. Meas. Sci. Technol., 6: 297-305. [13] Boyd S. R. (1997) Determination of the ammonium content of potassic rocks and minerals by capacitance manometry: a prelude to the calibration of FTIR microscopes. Chem. Geol., 137: 57-66. [13, 16] Boyd S., Wright I. P. & Pillinger C. T. (1997) Stepped-heating of carbonates and carbon-bearing quartz grains. Chem. Geol., 134: 303-310. [13] Boyd S. R. & Philippot P. (1998) Precambrian ammonium biogeochemistry: a study of the Moine metasediments, Scotland. Chem. Geol., 144: 257-268. [13, 16] Boyd S. R., Wright I. P., Alexander C. M. O. D. & Pillinger C. T. (1998) High resolution steppedcombustion mass spectrometry: Application to the detection and analysis of fine-grained diamond in meteorites and rocks. Geostandards Newsl., 22: 71-83. [13] Boyer P. D., Graves D. J., Suelter C. H. & Dempsey M. E. (1961) Simple procedure for conversion of oxygen of orthophosphate or water to carbon dioxide for oxygen-18 determination. Anal. Chem., 33: 1906-1909. [1] Biihlke J. K., Gwinn C. J. & Coplen T. B. (1993) New reference materials for nitrogen-isotope-ratio measurements. Geostand. Newsl., 17(1): 159-164. [15, 40] Biihlke J. K. & Coplen T. B. (1995) Interlaboratory comparison of reference materials for nitrogenisotope-ratio measurements. In: Reference and intercomparison materials for stable isotopes of light elements. Proceedings of a Consultants Meeting held in Vienna, 1-3 Dec 1993. IAEA, Vienna. IAEATECDOC-825: 51-66. [15, 40, 43] Biihlke J. K. & Denver J. M. (1995) Combined use of ground-water dating, chemical, and isotopic analyses to resolve the history and fate of nitrate contamination in two agricultural watersheds, atlantic coastal plain, Maryland. Water Resour. Res., 31: 2319-2339. [15] Biihlke J. K., Mroczkowski S. J. & Coplen T. B. (2003) Oxygen isotopes in nitrate: new reference materials for 180:170:160 measurements and observations on nitrate-water equilibration. Rapid Commun. Mass Spectrom., 17(16): 1835-1846. [15, 40] Bi3hm F., Joachimski M. M., Dullo W.-C., Eisenhauer A., Lehnert H., Reitner J. & WiJrheide G. (2000) Oxygen isotope fractionation in marine aragonite of coralline sponges. Geochim. Cosmochim. Acta, 64: 1695-1703. [46] Bi~ttcher J., Strebel O., Voerkelius S. & Schmidt H.-L. (1990) Using isotope fractionation of nitratenitrogen and nitrate-oxygen for evaluation of microbial denitrification in a sandy aquifer. J. Hydrol., 114(3-4): 413-424. [15, 29] B/Sttcher G., Brumsack H.-J., Heinrichs H. & Pohlmann M. (1997a) A new high-pressure squeezing technique for pore fluid extraction from terrestrial soils. Water, Air and Soil Pollution, 94: 289-296. [2] B/Sttcher M. E., Rusch A., HiJpner T. & Brumsack H.-J. (1997b) Stable sulfur isotope effects related to local intense sulfate reduction in a tidal sandflat (Southern North Sea): Results from loading experiments. Isot. Environ. Health Stud., 33: 109-129. [27]
1066
References
B/~ttcher M. E., Gehlken P.-L., Fernandez-Gonzalez A. & Prieto M. (1997c) Characterization of syn-
thetic BaCO3-SrCO3 (witherite-strontianite) solid-solutions by Fourier transform infrared spectoscopy. Eur. J. Mineral., 9: 519-528. [46] B/ittcher M. E., Brumsack H.-J. & de Lange G. J. (1998a) Sulfate reduction and related stable isotope (34S, 180) variations in interstitial waters of the eastern Mediterranean. In: Proceedings of the Ocean Drilling Program (ODP). A. H. F. Robertson, K.-C. Emeis & A. Camerlenghi (eds.). Sci. Res., Vol. 160: 365-373. [27] Bi~ttcher M. E., Oelschltiger B., Hi~pner T., Brumsack H.-J. & Rullkiitter J. (1998b) Sulfate reduction related to the early diagenetic degradation of organic matter and "black spot" formation in tidal sandflats of the German Wadden Sea: Stable isotope (13C, 34S, 180) and other geochemical results. Org. Geochem., 29: 1517-1530. [27] Bi~ttcher M. E., Smock A. M. & Cypionka H. (1998c) Sulfur isotope fractionation during experimental precipitation of iron(II) and manganese(II) sulfide at room temperature. Chem. Geol., 146: 127-134. [27] Bi~ttcher M. E. & Lepland A. (2000) Biogeochemistry of sulfur in a sediment core from the west-central Baltic Sea: Evidence from stable isotopes and pyrite textures. J. Mar. Sys., 25: 299-312. [27] Bi~ttcher M. E., Thamdrup B. & Vennemann T. W. (2001) Oxygen and sulphur isotope fractionation during anaerobic bacterial disproportionation of elemental sulfur. Geochim. Cosmochim. Acta, 65: 1573-1581. [27, 46] Brand W. A., Tegtmeyer A. R. & Hilkert A. (1994) Compound-specific isotope analysis: extending toward 15N/14N and 180/160. Org. Geochem., 21: 585-594. [1, 8, 23] Brand W. A. (1995a) Precon: A fully automated interface for the pre-GC concentration of trace gases in air for isotopic analysis. Isot. Environ. Health Stud., 31: 277-284. [8] Brand W. (1995b) High precision isotope ratio monitoring techniques in mass spectrometry. J. Mass Spectrom., 31: 225-235. [16] Brand W. A., Avak H., Seedorf R., Hofmann D. & Conradi T. (1996) New methods for fully automated isotope ratio determination from hydrogen at the natural abundance level. Isot. Environ. Health Stud., 32: 263-273. [1] Brand W. A. (1996) High precision isotope ratio monitoring techniques in mass spectrometry. J. Mass Spectrom., 31: 225-235. [8, 38, 43, 45] Brand W. A. (1998) Isotope ratio mass spectrometry: Precision from transient signals. In: Adv. Mass Spectrom., Vol. 14. E. J. Karjalainen, A. E. Hesso, J. E. Jalonen & U. P. Karjalainen(eds.): 655-680. [38] Brand W. A. & Coplen T. B. (2001) An interlaboratory study to test instrument performance of hydrogen dual-inlet isotope-ratio mass spectrometers. Fresenius J. Anal. Chem., 370: 358-362. [38, 43] Brandes J. A. & Devol A. H. (1997) Isotopic fractionation of oxygen and nitrogen in coastal marine sediments. Geochim. Cosmochim. Acta, 61: 1793-1801. [9] Brandriss M. E., O'Neil J. R., Edlund M. B. & Stoermer E. F. (1998) Oxygen isotope fractionation between diatomaceous silica and water. Geochim. Cosmochim. Acta, 62: 1119-1125. [2, 46] Brantley S. L., Liermann L. & Bullen T. D. (2001) Fractionation of Fe isotopes by soil microbes and organic acids. Geology, 29: 535-538. [29] Bray A. M. (2001) The geochemistry of boron and lithium in mid-ocean ridge hydrothermal vent fluids. PhD Thesis, University of New Hampshire, Durham: 125 p. [6] Br~iuer K. & Strauch G. (2000) An alternative procedure for the 180 measurement of nitrate oxygen. Chem. Geol., 168: 283-290. [15] Breck D. H. (1974) Zeolite molecular sieves: Structure, chemistry, and use. Wiley, NY. [36] Breit G. N., Simmons E. C. & Goldhaber M. B. (1985) Dissolution of barite for the analysis of strontium isotopes and other chemical and isotopic variations using aqueous sodium carbonate. Chem. Geol., 52: 333-336. [26] Bremner J. M. (1965a) Organic forms of nitrogen, n: Methods of soil analysis, Part 2. Chemical and microbiological properties. C. A. Black, D. D. Evans, L. E. Ensminger, J. L. White & F. E. Clark (eds.). American Society of Agronomy Inc., Madison, Wisconsin. Agronomy, Vol. 9: 1238-1255. [15] Bremner J. M. (1965b) Inorganic forms of nitrogen. In: Methods of soil analysis, Part 2. Chemical and microbiological properties. C. A. Black, D. D. Evans, L. E. Ensminger, J. L. White & F. E. Clark (eds.). American Society of Agronomy Inc., Madison, Wisconsin. Agronomy, Vol. 9: 1179-1237. [15, 28]
References
1067
Bremner J. M. (1965c) Isotope-ratio analysis of nitrogen in nitrogen-15 tracer investigations. In: Methods of soil analysis, Part 2. Chemical and microbiological properties. C. A. Black, D. D. Evans, L. E. Ensminger, J. L. White & F. E. Clark (eds.). American Society of Agronomy Inc., Madison, Wisconsin. Agronomy, Vol. 9 1256-1286. [15] Bremner J. M. (1965d) Total nitrogen. In: Methods of soil analysis, Part 2. Chemical and microbiological properties. C. A. Black, D. D. Evans, L. E. Ensminger, J. L. White & F. E. Clark (eds.). American Society of Agronomy Inc., Madison, Wisconsin. Agronomy, Vol. 9: 1149-1178. [15] Bremner J. M. & Edwards A. P. (1965) Determination and isotope-ratio analysis of different forms of nitrogen in soils: I. Apparatus and procedure for distillation and determination of ammonium. Soil Sci. Soc. Am. Proc., 29: 504-507. [15] Bremner J. M. & Keeney D. R. (1965) Steam distillation methods for determination of ammonium, nitrate and nitrite. Anal. Chim. Acta, 32: 485-495. [15] Bremner J. M. & Mulvaney C. S. (1982) Nitrogen - Total. In: Methods of soil analysis, Part 2. Chemical and microbiological properties. A. L. Page, R. H. Miller & D. R. Keeney (eds.). American Society of Agronomy Inc. - Soil Science Society of America Inc., Madison, Wisconsin. Agronomy Series, 2nd edition, No. 9: 595-624. [15] Bremner J. M. (1996) Nitrogen-Total. In: Methods of soil analysis. Part 3. Chemical methods. D. L. Sparks et. al. (eds.). Soil Science Society of America and American society of Agronomy, Madison, Wisconsin: 1085-1121. [15] Brendel O., Iannetta P. P. M. & Steward D. (2000) A rapid and simple method to isolate pure alphacellulose. Phytochem. Anal., 11: 7-10. [24] Brenna J. T. & Morrison G. H. (1984) Low temperature ashing for elemental sensitivity enhancement in secondary ion mass spectrometry. Anal. Chem., 56: 2791-2797. [2] Brenna J. T. (1994) High-precision gas isotope ratio mass spectrometry: Recent advances in instrumentation and biomediacal applications. Acc. Chem. Res., 27: 340-346. [8] Brenna J. T., Corso T. N., Tobias H. J. & Caimi R. J. (1997) High-precision continuous-flow isotope ratio mass spectrometry. Mass Spectrom. Rev., 16: 227-258. [8] Brenninkmeijer C. A. M. & Mook W. G. (1981) A batch process for direct conversion of organic oxygen and water to CO2 for 180/160 analysis. Int. J. Appl. Rad. Isot., 32: 137-141. [1] Brenninkmeijer C. A. M. (1983) Deuterium, oxygen-18 and carbon-13 in tree rings and peat deposits in relation to climate. PhD Thesis, University of Groningen, Groningen, the Netherlands: 146 p. [23, 24] Brenninkmeijer C. A. M. & Morrison P. D. (1987) An automated system for isotopic equilibration of CO2 and H20 for 180 analysis of water. Chem. Geol. (Isot. Geosci. Sect.), 66: 21-26. [1] Brenninkmeijer C. A. M. & Ri~ckmann T. (1999) Mass spectrometry of the intramolecular nitrogen isotope distribution of environmental nitrous oxide using fragment-ion analysis. Rapid Commun. Mass Spectrom., 13: 2028-2033. [19] Brewer A. K. (1936) The isotope effect in the evaporation of lithium, potassium and rubidium ions. J. Chem. Phys., 4: 350-353. [6] Br6as O., Guillou C., Reniero F., Sada E. & Tanet G. (1996) Deuterium/hydrogen measurements in wines and fruit juices by platinum catalysed equilibration method. Rapid Commun. Mass Spectrom., 10:246-249 [1] Br4as O., Guillou C., Reniero F., Sada E. & Angerosa F. (1998) Oxygen-18 measurement by continuous flow pyrolysis / isotope ratio mass spectrometer of vegetable oils. Rapid Commun. Mass Spectrom., 12: 188-192. [1, 23] Brilliant D. R., Franchi I. A. & Pillinger C. T. (1994) Nitrogen components in lunar 12023: Complex grains are not the carrier of isotopically light nitrogen. Meteoritics, 29:718-723 [13] Brimmer S. P., Fawcett W. R. & Kulhavy K. A. (1987) Quantitative reduction of selenate ion to selenite in aqueous samples. Anal. Chem., 59: 1470-1471. [29] Brindle I. D., Alarabi H., Karshman S., Le X. & Zheng S. (1992) Combined generator / separator for continuous hydride generation: Application to on-line pre-reduction of arsenic(V) and determination of arsenic in water by atomic emission spectrometry. Analyst, 117: 407-411. [29] Brindley G. W. & Lemaitre J. (1987) Thermal, oxidation and reduction reacions of clay minerals. In: Chemistry of clays amd clay minerals. A. C. D. Newman (ed.). Mineralogical Society, London: 319370. [2]
1068
References
Bronk D. A. & Glibert P. M. (1991) A 15N tracer method for the measurement of dissolved organic nitrogen release by phytoplankton. Mar. Ecol. Prog. Ser., 77: 171-182. [15] Brooks J. R., Flanagan L. B., Buchmann N. & Ehleringer J. R. (1997) Carbon isotope composition of boreal plants: functional grouping of life forms. Oecologia, 110: 301-311. [9] Brooks P. D., Stark J. M., McInteer B. B. & Preston T. (1989) Diffusion method to prepare soil extracts for automated nitrogen-15 analysis. Soil Sci. Soc. Am. J., 53: 1707-1711. [15] Brooks P. D. & Atkins G. J. (1992) Rapid isotopic analysis of trace gases at atmospheric levels. Isotopenpraxis, 28: 106-109. [8] Brown F., Gillies A. & Stevens W. H. (1953) A note on the preparation of chlorine gas containing 36C1. Can. J. Chem., 31: 768. [28] Brown K. A. (1986) Formation of organic sulphur in anaerobic peat. Soil Biol. Biochem., 18: 131-140. [26] Brown L., Rajan R. S., Roberts R. B., Tera R. & Whitford D. J. (1978) A new method for determining the isotopic composition of lithium. Nucl. Inst. Method, 156: 541-546. [6] Brown M. (1993) Deduction of emissions of source gases using an objective inversion algorithm and a chemical-transport model. J. Geophys. Res., 98(D7): 12639-12660. [14] Brumsack H.-J. (1980) Geochemistry of Cretaceous black shales from the Atlantic ocean (DSDP Legs 11, 14, 36 and 41). Chem. Geol., 31: 1-25. [27] Brumsack H. & Zuleger E. (1992) Boron and boron isotopes in pore waters from ODP Leg 127, Sea of Japan. Earth Planet. Sci. Lett., 113: 427-433. [7] Brunn4e C. (1987) The ideal mass analyzer: fact or fiction? Int. J. Mass Spectrom. Ion Proc., 76: 121-237. [32] Burns D. A. & Kendall C. (2002) Analysis of sources of 15N and 180 to differentiate NO3- sources in runoff in two watersheds in the Catskills Mountain of New York. Water Resour. Res., 38(5): doi 10.1029 / 2001WR000292,2002. [15] Briichert V. & Pratt L. M. (1996) Contemporaneous early diagenetic formation of organic and inorganic sulfur in estuarine sediments from St. Andrew Bay, Florida, USA. Geochim. Cosmochim. Acta, 60(13): 2325-2332. [29] Bryant J. D., Koch P. L., Froelich P. N., Showers W. J. & Genna B. J. (1996) Oxygen isotope partitioning between phosphate and carbonate in mammalian apatite. Geochim. Cosmochim. Acta, 60: 5145-5148. [46] Buchmann N., Brooks J. R., Rapp K. D. & Ehleringer J. R. (1996) Carbon isotope composition of C4 grasses is influenced by light and water supply. Plant, Cell Environm., 19: 392-402. [9] Buchmann N., Guehl J.-M., Barigah T. S. & Ehleringer J. R. (1997) Interseasonal comparison of CO2 concentrations, isotopic composition, and carbon dynamics in an Amazonian rainforest (French Guiana). Oecology, 110: 120-131. [9] Budd P., Lythgoe P., McGill R. A. R., Pollard A. M. & Scaife B. (1999) Zinc isotope fractionation in liquid brass (Cu-Zn) alloy: potential environmental and archeological applications. In: Geoarcheology: exploration, environments, resources. A. M. Pollard (ed.). Geological Society, Geological Society Special Publications, Vol. 165: 147-153. [31] Buhay W. M., Wolfe B. B., Elgood R. J. & Edwards T. W. D. (1995) A report on the 6180 measurements of the IAEA cellulose intercomparison material IAEA-C3. In: Reference and intercomparison materials for stable isotopes of light elements. IAEA-TECDOC-825, IAEA, Vienna: 85-86. [23] Bullen T. D., White A. F., Childs C. W., Vivit D. V. & Schulz M. S. (2001) A demonstration of significant abiotic Fe isotope fractionation in nature. Geology, 29: 699-702. [29] Bunch T. E. & Chang S. (1980) Carbonaceous chondrites II: Carbonaceous chondrites phyllosilicates and light element geochemsitry as indicators of parent body processes and surface conditions. Geochim. Cosmochim. Acta, 44: 1543-1557. [11] Bunn S. E., Davies P. M. & Kellaway D. M. (1997) Contributions of sugar cane and invasive pasture grass to the aquatic food web of tropical lowland stream. Mar. Freshwater Res., 48: 173-179. [9] Burdett J. W., Arthur M. A. & Richardson M. (1989) A Neogene seawater sulfur isotope age curve from calcareous pelagic microfossils. Earth Planet. Sci. Lett., 94: 189-198. [26] Burgoyne T. W., Hieftje G. M. & Hites R. A. (1997) Space charge evaluation in a plasma-source mass spectrograph. Anal. Chem., 69: 485-489. [32] Burgoyne T. W. & Hayes J. M. (1998) Quantitative production of H2 by pyrolysis of gas chromatographic effluents. Anal. Chem., 70: 5136-5141. [3, 14, 38]
References
1069
Burk R. L. & Stuiver M. (1981) Oxygen isotope ratios in tree cellulose reflect mean annual temperature and humidity. Science, 211: 1417-1419. [23] Burkhardt S., Riebesell U. & Zondervan I. (1999) Effects of growth rate, CO2 concentration, and cell size on the stable carbon isotope fractionation in marine phytoplankton. Geochim. Cosmochim. Acta, 63: 3729-3741. [46] Burnham C. W. & Jahns R. H. (1962) A method for determining the solubility of water in silicate melts. Am. J. Sci., 260: 721-745. [46] Burns D. A. & Kendall C. (2002) Analysis of sources of 15N and 180 to differentiate NO3- sources in runoff in two watersheds in the Catskills mountain of NewYork. Water Resour. Res., 38(5): 10.1029/ 2001WR000292. [15] Bushaw B. A., Juston F., Ni~rdh/iuser W., Trautmann N., Voss-de Haan P. & Wendt K. (1996) Multiple resonance RIMS, measurements of calcium isotopes using diode lasers. Resonance Ionization Spectroscopy 1996, Eight International Symposium. N. Winograd & J.E. Parks (eds.). AIP Press Woodbury, NY, AIP Conference Proceedings 388: 115-118. [33] Bushaw B. A. & Cannon B. D. (1998) Diode-laser-based RIMS measurements of 90St. Resonance Ionization Spectroscopy, Ninth International Symposium. J.C. Vickerman, I. Lyon, N.P. Lockyer & J.E. Parks (eds.). AlP Press Woodbury, NY, AIP Conference Proceedings 454: 177-182. [33] Buzek F. (1983) A rapid procedure for preparing oxygen-18 determination in water samples. Isotopenpraxis, 19: 70-72. [1] Cai W. J., Wang Y.C., Krest J. & Moore W. S. (2003) The geochemistry of dissolved inorganic carbon in a surficial groundwater aquifer in North Inlet, South Carolina, and the carbon fluxes to the coastal ocean. Geoch. Cosmochim. Acta, 67: 631-639. [10] Caimi R. J., Huoghton L. A. & Brenna J. T. (1994) Condensed-phase carbon isotopic standards for compound-specific isotope analysis. Anal. Chem., 66: 2989-2991. [8] Caimi R. J. & Brenna J. T. (1996) Direct analysis of carbon isotope variability in albumins by liquid flowinjection isotope ratio mass spectrometry. J. Am. Soc. Mass Spectrom., 7: 605-610. [8] Caimi R. J. & Brenna J. T. (1997) Quantitative evaluation of carbon isotopic fractionation during reversed-phase high-performance liquid chromatography. J. Chromatogr., 757A: 307-310. [8] Calhoun J. A. (1990) Chemical and isotopic methods for understanding the natural marine sulfur cycle. PhD Thesis, University of Washington: 86 p. [26] Calhoun J. A., Bates T. S. & Charlson R. J. (1991) Sulfur isotope measurements of submicrometer sulfate aerosol particles over the Pacific Ocean. Geophys. Res. Lett., 18(10): 1877-1880. [26] Callis E. L. & Abernathey R. M. (1991) High-precision isotopic analyses of uranium and plutonium by total sample volatilization and signal integration. Int. J. Mass Spec. Ion Proc., 103: 93-105. [35] Calvert S. E., Thode H. G., Yeung D. & Karlin R. E. (1996) A stable isotope study of pyrite formation in Late Pleistocene and Holocene sediments of the Black Sea. Geochim. Cosmochim. Acta, 60: 1261-1270. [27] Cameron A. E. (1955) Variation in the natural abundance of the lithium isotopes. J. Am. Chem. Soc., 77: 2731-2733. [6] Cameron A. E. & Lippert E. L. Jr. (1955) Isotopic composition of bromine in nature. Science, 121: 136-137. [281 Cameron A. E., Herr W., Herzog W. & Lund6n A. (1956) Isotopen-Anreicherung beim Brom durch electrolytische 0berftihrung in geschmolzenem Bleibromid. Z. Naturforsch., 11a: 203-205. [28] Campbell D. J. & Kaufmann R. S. (1984) Fractionation of chlorine isotopes by flow through semipermeable membranes. EOS Trans. AGU, 65: 882. [28] Campbell D. J., Kendall C., Chang C. C. Y., Silva S. R. & Tonnesen K. A. (2002) Pathways for nitrate release from an alpine watershed: determination using 6180 and 615N. Water Resour. Res., 38(5): doi 10.1029 / 2001WR000294. [15] Cane G. & Clark I. D. (1998) Tracing ground water recharge in an agricultural watershed with isotopes. Ground Water, 27: 133-139. [10] Canfield D. E., Raiswell R., Westrich J. T., Reaves C. M. & Berner R. A. (1986) The use of chromium reduction in the analysis of reduced inorganic sulphur in sediments and shales. Chem. Geol., 54: 149155. [26]
1070
References
Canfield D. E. & Thamdrup B. (1994) The production of 34S depleted sulfide during bacterial disproportionation of elemental sulfur. Science, 266: 1973-1975. [26] Canfield D. E. (2001a) Biogeochemistry of sulfur isotopes. Rev. Mineral. Geochem., 43: 607-636. [46] Canfield D. E. (2001b) Isotope fractionation by natural populations of sulphate-reducing bacteria. Geochim. Cosmochim. Acta, 65: 1117-1124. [46] Cantrell C. A., Shetter R. E., McDaniel A. H., Calvert J. A., Davidson J. A., Lowe D. C., Tyler S. C., Cicerone R. J. & Greenberg J. P. (1990) Carbon kinetic isotope effect in the oxidation by the hydroxyl radical. J. Geophys. Res., 95(D13): 22455-22462. [14] Canuel E. A., Freeman K. H. & Wakeham S. G. (1997) Isotopic compositions of lipid biomarker compounds in estuarine plants and surface sediments. Limnol. Oceanogr., 42: 1570-1583. [9] Caraco N. E & Cole J. J. (1999) Human impact on nitrate export: an analysis using major wolrd rivers. Ambio, 28: 167-170. [15] Carle R., Beyer J., Cheminat A. & Krempp E. (1992) 2H NMR determination of site-specific natural isotope fractionation in (-)-c~-bisabolols. Phytochem., 31(1): 171-174. [5] Carmody R., Plummer L. N., Busenberg E. & Coplen T. B. (1998) Methods for collection of dissolved sulfate and sulfide and analysis of their sulfur isotopic composition. U. S. Geol. Surv., Reston, Virginia. Open-File Report 97-234: 91p. [26] Carmody R. W. & Seal R. R. II (1999) Evaluation of the sulfur isotopic composition and homogeneity of the Soufre de Lacq reference material. Chem. Geol., 153: 289-295. [40] Carpenter S. J. & Lohmann K. C. (1995) 8180 and 813C values of modern brachiopod shells. Geochim. Cosmochim. Acta, 59: 3749-3764. [46] Carr R. H., Wright I. P., Joines A. W. & Pillinger C. T. (1986) Measurement of carbon stable isotopes at the nanomole level: A static mass spectrometer and sample preparation technique. J. Phys. E: Sci. Instrum., 19: 798-808. [3, 13] Casabianca H., Graff J. B., Jame P., Perrucchietti C. & Chastrette M. (1995) Application of hyphenated techniques to the chromatographic authentication of flavours in food-products and perfumes. HRCJ. High Resolut. Chromatogr., 18: 279-285. [8] Casciotti K. L., Sigman D. M., Galanter Hastings M., Biihlke J. K. & Hilkert A. (2002) Measurement of the oxygen isotopic composition of nitrate in seawater and freshwater using the denitrifier method. Anal. Chem., 74: 4905-4912. [15] Casey W. H. & Phillips B. L. (2001) Kinetics of oxygen exchange between sites in the GaO4A112(OH)24(H20)127+(aq) molecule and aqueous solution. Geochim. Cosmochim. Acta, 65: 705714. [46] Cassagnab6re A. (1998) Caract6risation et interpr6tation de la transition kaolinite-dickite dans les r6servoirs a hydrocarbures de Froy et Rind (Mer du Nord, Norv6ge). Th6se d'Universit6, Universit6 de Poitiers, Poitiers, France: 238 p. [2] Catanzaro E. J., Murphy T. J., Garner E. L. & Shields W. R. (1964) Absolute isotopic abundance ratio and the atomic weight of bromine. J. Res. Nat. Bur. Stand., 68A: 593-599. [28] Catanzaro E. J. & Murphy T. J. (1966) Magnesium isotope ratios in natural samples. J. Geophys. Res., 71: 1271-1274. [31] Catanzaro E. J., Murphy T. J., Garner E. L. & Shields W. R. (1966) Absolute isotopic abundance ratios and atomic weights of magnesium. J. Res. Nat. Bureau Stand., 70A: 453-458. [31] Catanzaro E. J., Champion C. E., Garner E. L., Marineneko G., Sappenfield K. M. & Shields W. R. (1970) Standard Reference Material: Boric Acid; Isotopic, and assay of standard reference materials. Natl. Bur. Stand. (US), Spec. Publ. (Febr. 1970), 260-17:70 p. [7, 40] Cerling T. E., Brown F. H. & Bowman J. R. (1985) Low-temperature alteration of volcanic glass: hydration, Na, K, 180 and Ar mobility. Chem. Geol., 52: 281-293. [46] Ceding T. E. (1992) Use of carbon isotopes in paleosols as an indicator of the p(CO2) of the paleoatmosphere. Global Biogeochem. Cycles, 6: 307-314. [9] Cerling T. E. & Sharp Z. D. (1996) Stable carbon and oxygen isotope analysis of fossil tooth enamel using laser ablation. Palaeogeogr. Palaeoclimatol. Palaeoecol., 126: 173-186. [20, 45] Chacko T., Mayeda T. K., Clayton R. N. & Goldsmith J. R. (1991) Oxygen and carbon isotope fractionations between CO2 and calcite. Geochim. Cosmochim. Acta, 55: 2867-2882. [46]
References
1071
Chacko T., Hu X., Mayeda T. K., Clayton R. N. & Goldsmith J. R. (1996) Oxygen isotope fractionations in muscovite, phlogopite, and rutile. Geochim. Cosmochim. Acta, 60(14): 2595-2608. [46] Chacko T., Riciputi L. R., Cole D. R. & Horita J. (1999) A new technique for determining equilibrium hydrogen isotope fractionation factors using the ion microprobe: Application to the epidote-water system. Geochim. Cosmochim. Acta, 63(1): 1-10. [46] Chacko T., Cole D. R. & Horita J. (2001) Equilibrium oxygen, hydrogen and carbon isotope fractionation factors applicable to geological systems. Rev. Mineral. Geochem., 43: 1-81. [46] Chae Y. M. & Krouse H. R. (1986) Alteration of sulfur-34 natural abundance in soil by application of feedlot manure. Soil Sci. Soc. Am. J., 50: 1425-1430. [26] Chai B. H. T. (1974) Mass transfer of calcite during hydrothermal recrystallization. In: Geochemical Transport and Kinetics. A. W. Hoffmann, B. J. Giletti, H. S. Jr. Yoder & R. A. Yund (eds.). Carnegie Inst. Wash. Pubt., Vol. 634: 205-218. [46] Chamberlain C. P., Blum J. D., Holmes R. T., Feng X., Sherry T. W. & Graves G. R. (1997) The use of isotope tracers for identifying populations of migratory birds. Oecologica, 109: 132-141. [9] Chan L. H. (1987) Lithium isotope analysis by thermal ionization mass spectrometry of lithium tetraborate. Anal. Chem., 59: 2662-2665. [6, 7, 31] Chan L.-H. & Edmond J. M. (1988) Variation of lithium isotope composition in the marine environment: a preliminary report. Geochim. Cosmochim. Acta, 52: 1711-1717. [6] Chan L. H., Edmond J. M., Thompson G. & Gillis K. (1992) Lithium isotopic composition of submarine basalts: Implications for the lithium cycle in the oceans. Earth Planet. Sci. Lett., 108: 151-160. [6, 46] Chan L. H., Edmond J. M. & Thompson G. (1993) A lithium isotope study of hot springs and metabasalts from mid-ocean ridge hydrothermal systems. J. Geophys. Res., 98: 9653-9659. [6] Chan L. H., Gieskes J. M., You C. F. & Edmond J. M. (1994) Lithium isotope geochemistry of sediments and hydrothermal fluids of the Guaymas Basin, Gulf of California. Geochim. Cosmochim. Acta, 58: 4443-4454. [6] Chan L. H., Leeman W. P. & You C. F. (1999) Lithium isotopic composition of Central American Volcanic Arc lavas: Implications for modification of subarc mantle by slab-derived fluids. Chem. Geol., 160: 255-280. [6] Chan L. H. & Kastner M. (2000) Lithium isotopic compositions of pore fluids and sediments in the Costa Rica subduction zone: implications for fluid processes and sediment contribution to the arc volcanoes. Earth Planet. Sci. Lett., 183: 275-290. [6] Chan L. H., Leeman W. P. & Tonarini S. (2001) Lithium isotope compositions of South Sandwich Arc and Southwest Washington Cascades: A comparative study of arc processes. EOS Trans. AGU, 82: F1293 (Abstract). [6] Chan L. H., Alt J. C. & Teagle D. A. H. (2002a) Lithium and lithium isotope profiles through the upper oceanic crust: a study of seawater-basalt exchange at ODP Sites 504B and 896A. Earth Planet. Sci. Lett., 201: 187-201. [6] Chan L. H., Leeman W. P., Tonarini S. & Singer B. (2002b) Lithium and boron isotopes in Aleutian Islands: Contribution of marine sediments to island arc magmas. EOS Trans. AGU, 83, F1428 (Abstract). [6] Chan L. H., Leeman W. P. & You C. F. (2002c) Lithium isotopic composition of Central American Volcanic Arc lavas: Implications for modification of subarc mantle by slab-derived fluids, correction. Chem. Geol., 182: 293-300. [6] Chan L. H., Starinsky A. & Katz A. (2002d) The bahavior of lithium and its isotopes in oilfield brines; Evidence from the Heletz-Kokhav field, Israel. Geochim. Cosmochim. Acta, 66: 615-623. [6] Chan L. H. & Frey F. A. (2003) Lithium isotope geochemistry of the Hawaiian Plume: Results from the Hawaii Scientific Drilling Project and Koolau Volcano. Geochem. Geophys. Geosyst., 4(3), 8707, doi 10.1029 / 2002GC000365. [6] Chang C. C. Y., Langston J., Riggs M., Campbell D. H., Silva S. R. & Kendall C. (1999) A method for nitrate collection for 615N and 6180 analysis from waters with low nitrate concentrations. Can. J. Fish Aquat. Sci., 56: 1856-1864. [9, 15] Chang C. C. Y., Kendall C., Silva S. R., Battalgin W. A. & Campbell D. H. (2002) Nitrate stable isotopes: tools for determining nitrate sources among different land uses in the Mississippi River Basin. Can. J. Fish. Aq. Sci., 59: 1874-1885. [15]
1072
References
Chanton J. P. & Martens C. S. (1985) The effects of heat and stannous chloride addition on the active
distillation of acid volatile sulfide from pyrite-rich marine sediment samples. Biochem., 1: 375-383. [26] Chanton J. P., Martens C. S. & Goldhaber M. B. (1987) Biogeochemical cycling in an organic-rich coastal marine basin. 8. A sulfur isotope budget balanced by differential diffusion across the sediment-water interface. Geochim. Cosmochim. Acta, 51: 1201-1208. [27] Chapelle F. H. & Knobel L. L. (1985) Stable carbon isotopes of HCO3- in the Aquiua aquifer, Maryland: evidence for an isotopically heavy source of CO2. Ground Water, 23: 592-599. [10] Chapelle F. H., Morris J. T., McMahon P. B. & Zelibor J. L. (1988) Bacterial metabolism and the 613C composition of groundwater, Floridan aquifer system, South Carolina. Geology, 16: 117-121. [10] Chapelle F. H. & McMahon P. B. (1991) Geochemistry of dissolved inorganic carbon in a Coastal Plain aquifer. 1. Sulfate from confining beds as an oxidant in microbial CO2 production. J. Hydrol., 127: 85108.[10] Chapman B. (1980) Glow discharge processes: Sputtering and plasma etching. Wiley, NY: 406 p. [35] Chapman R. (1972) Versatile wide range electron current regulator. Sci. Instrum., 43: 1536-1538. [39] Charef A. & Sheppard S. M. F. (1984) Carbon and oxygen isotope analysis of calcite or dolomite associated with organic matter. Isotope Geosci., 2: 325-333. [2] Chartier F. & Tabarant M. (1997) Comparison of the performance of a laboratory-built high resolution glow discharge mass spectrometer with that of a quadrupole inductively coupled plasma (glow discharge) mass spectrometer for boron and gadolinium isotopic analysis. J. Anal. At. Spectrom., 12: 1187-1193. [35] Chau Y. K. & Riley J. P. (1965) The determination of selenium in sea water, silicates, and marine organisms. Anal. Chim. Acta, 33: 36-49. [29] Chaussidon M. & Lorand J.-P. (1990) Sulphur isotope composition of orogenic spinel lherzolite massifs from Ari6ge (North-Eastern Pyr6n6es, France): an ion microprobe study. Geochim. Cosmochim. Acta, 54: 2835-2846. [20] Chaussidon M. & Albar6de F. (1992) Secular boron isotope variations in the continental crust: an ion microprobe study. Earth Planet. Sci. Lett., 108: 229-241. [46] Chaussidon M. & Jambon A. (1994) Boron content and isotopic composition of oceanic basalts: Geochemical and cosmochemical implications. Earth Planet. Sci. Lett., 121: 277-291. [30] Chaussidon M. & Uiterdijk Appel P. W. (1997) Boron isotopic composition of tourmalines from the 3.8Ga-old Isua supracrustals, West Greenland: implications on the 611B value of early Archean seawater. Chem. Geol. (Isot. Geosci. Sect.), 136: 171-180. [30] Chaussidon M., Robert F., Mangin D., Hanon P. & Rose E. F. (1997) Analytical procedures for the measurement of boron isotope compositions by ion microprobe in meteorites and mantle rocks. Geostand. Newslett., 21: 7-16. [30] Chaussidon M. & Robert F. (1998)7Li/6Li and 11B/10B variations in chondrules from the Semarkona unequilibrated chondrite. Earth Planet. Sci. Lett., 164: 577-589. [6, 30] Chaussidon M. & Robert F. (1999) Lithium nucleosynthesis in the Sun inferred from the solar wind 7Li/ 6Li ratio. Nature, 402: 270-273. [30] Chaux R. & Lavorel B. (2000) Relative line intensity measurement in absorption spectra using tunable diode laser at 1.6 ~m: application to the determination of 13CO2/12CO2isotope ratio. Appl. Phys., B 72: 237-240. [34] Chazot G. (1997) t~tude de la composition isotopique de l'oxyg6ne des surfaces cuirass6es du bouclier guyanais. BRGM, Report R39465:33 p. [2] Chen C.-H., Liu K.-K. & Shieh Y.-N. (1988) Chemical and isotopic studies of bauxitization in the Tatun Volcanic area, northern Taiwan. Chem. Geol., 68: 41-56. [46] Chen C.-H., Liu K.-K. & Shieh Y.-N. (1990) Comments on "A stable-isotope study of lateritic bauxites" by M. I. Bird, A. R. Chivas and A. S. Andrews. Geochim. Cosmochim. Acta, 54:. 1483-1484. [46] Chen C. Y., Li Y. M., Bailey K., O'Connor T. P., Young L. & Lu Z.-T. (1999) Ultrasensitive isotope trace analysis with a magneto-optical trap. Science, 286: 1139-1141. [34]
References
1073
Cheng H. H. & Bremner J. M. (1965) Gaseous forms of nitrogen. In: Methods of soil analysis, Part 2. Chemical and microbiological properties. C. A. Black, D. D. Evans, L. E. Ensminger, J. L. White & F. E. Clark (eds.). American Society of Agronomy Inc., Madison, Wisconbsin, USA. Vol 9: 1287-1323. [36] Chen X. & Houk R. S. (1996) Spatially-resolved measurements of ion density behind the skimmer of an inductively coupled plasma mass spectrometer. Spectrochim. Acta, B 51: 41-54. [35] Chen Huang T. S. C. & Chin H.-C. (1973) Infrared methods for analysis of very low and very high concentration of deuterium oxide in water. Internal Report INER-62-D-0117. Inst. Nucl. Energy Res., Atomic Energy Council, Taiwan. [7] Chiba H. & Sakai H. (1985) Oxygen isotope exchange rate between dissolved sulfate and water at hydrothermal temperatures. Geochim. Cosmochim. Acta, 49: 993-1000. [26, 46] Chiba H., Sakai H. & Yasutake M. (1985) Automatic measurement of oxygen isotope ratio of water samples. Hot Spring Institute, Okayama University, Okayama, Japan, Report No. 56:27-34 (In Japanese with English abstract). [1] Chiba H., Chadko T., Clayton R. N. & Goldsmith J. R. (1989) Oxygen isotope fractionations involving diopside, forsterite, magnetite, and calcite: Application to geothermometry. Geochim. Cosmochim. Acta, 53: 2985-2995. [46] Chihara K., Suzuki M. & Kawazoe K. (1978) Adsorption rate on molecular sieving carbon by chromatography. AIChE Journal, 24: 237-246. [36] Chou I. M. (1986) Permeability of precious metals to hydrogen at 2 kb total pressure and elevated temperatures. Am. J. Sci., 286: 638-658. [46] Chou I. M. (1987) Oxygen buffer and hydrogen sensor techniques at elevated pressures and temperatures. In: Hydrothermal experimental techniques. G. C. Ulmer & H. L. Barnes (eds.). WileyInterscience, NY: 61-99. [46] Chu X. L., Zhao R., Zang W. X., Lei J. J. & Tang u G. (1994) Extraction of various sulfurs in coal and sedimentary rock and preparation of samples for sulfur isotopic analysis. Chin. Sci. Bull., 39: 140145. [26] Chukhrov F. V., Ermilova L. P., Churikov V. S. & Nosik L. P. (1980) The isotopic composition of plant sulfur. Org. Geochem., 2: 69-75. [26] Ciais P., Tans P., White J. W., Francey R., Trolier M., Berry J., Randall D., Sellers P., Collatz J. & Schimrnel D. (1995) The global carbon budget inferred from a new worldwide data set of 613C measurements in atmospheric CO2. J. Geophys. Res., 100(D3): 5051-5070. [14] Ciais P., Denning A. S., Tans P. P., Berry J. A., Randall D. A., Collatz G. J., Sellers P. J., White J. W. C., Trolier M., Meijer H. A. J., Francey R. J., Monfray P. & Heimann M. (1997) A three-dimensional synthesis study of 6180 in atmospheric CO2. 1. Surface fluxes. J. Geophys. Res., 102: 5857-5872. [14] Cicerone R. J. & McCrumb J. L. (1980) Photodissociation of isotopically heavy 02 as a source of atmospheric 03. Geophys. Res. Lett., 7: 251-254. [18] Cicerone R. J. & Oremland R. S. (1988) Biogeochemical aspects of atmospheric methane. Global Biogeochem. Cycles, 2: 299-327. [14] Cifuentes L. A., Sharp J. H. & Fogel M. L. (1988) Stable carbon and nitrogen isotope biogeochemistry in the Delaware estuary. Limnol. Oceanogr., 33: 1102-1115. [9] Cifuentes L. A., Fogel M. L., Pennock J. R. & Sharp J. H. (1989) Biogeochemical factors that influence the stable nitrogen isotope ratio of dissolved ammonium in the Delaware Estuary. Geochim. Cosmochim. Acta, 53: 2713-2721. [9, 46] Clayton D. D., Obradovich M., Guha M. & Brown L. E. (1991) Silicon and titanium isotopes in SiC from AGB stars. Lunar Planet. Sci., XXII: 221-222 (Abstract). [25] Clayton R. N. & Epstein S. (1958) The relationship between 180/16 0 ratios in coexisting quartz, carbonate, and iron oxides from various geological deposits. J. Geol., 66: 352-373. [20] Clayton R. N. (1959) Oxygen isotope fractionation in the system calcium carbonate - water. J. Chem. Phys., 30: 1246-1250. [46] Clayton R. N. (1961) Oxygen isotope fractionation between calcium carbonate and water. J. Chem. Phys., 34: 724-726. [46] Clayton R. N. & Mayeda T. K. (1963) The use of bromine pentafluoride in the extraction of oxygen from oxides and silicates for isotopic analysis. Geochim. Cosmochim. Acta, 27: 43-52. [2, 20, 25, 36]
1074
References
Clayton R. N., O'Neil J. R. & Mayeda T. K. (1972) Oxygen isotope exchange between quartz and water. J. Geophys. Res., 77: 3057-3067. [46]
Clayton R. N., Mayeda T. K. & Hurd J. M. (1974) Loss of oxygen, silicon, sulfur and potassium from the lunar regolith. Proc. Fifth Lunar Sci. Conf.: 1801-1809. [25]
Clayton R. N., Goldsmith J. R., Karel K. J., Mayeda T. K. & Newton R. C. (1975) Limits on the effect of pressure on isotopic fractionation. Geochim. Cosmochim. Acta, 39: 1197-1201. [46]
Clayton R. N., Mayeda T. K. & Epstein S. (1978) Isotopic fractionation of Si in Allende inclusions. Proceedings of the 9th Lunar Sci. Conf., Houston. Vol. 1: 129-130. [20, 25] Clayton R. N. & Mayeda T. K. (1983) Oxygen isotopes in eucrites, shergottites, nakhlites, and chassignites. Earth Planet. Sci. Lett., 62: 1-6. [36] Clayton R. N. (1986) High temperature isotope effects in the early solar system. In: Stable isotopes in high temperature geological processes. J. W. Valley, H. P. Jr. Taylor & J. R. O'Neil (eds.). Mineral. Soc. America, Reviews in Mineralogy, Vol. 16: 129-140. [25] Clayton R. N., Goldsmith J. R. & Mayeda T. K. (1989) Oxygen isotope fractionation in quartz, albite, anorthite and calcite. Geochim. Cosmochim. Acta, 53: 725-733. [46] Clayton R. N. & Kieffer S. W. (1991) Oxygen isotope thermometer calibrations. In: Stable Isotope Geochemistry: A tribute to Samuel Epstein. H. P. Jr. Taylor, J. R. O'Neil & I. R. Kaplan (eds.). The Geochemical Society, Spec. Publ. 3: 3-10. [46] Clayton R. N. (1992) Silica-carbonate isotopic temperature calibration. Science, 258: 1162-1163. [46] Clement S. W. & Compston W. (1989) SIMS at high sensitivity and high mass resolution. Proc. of the 7th International Conference on Secondary Ion Mass-Spectroscopy. J. Wiley & Sons. [25] Cliff S. S. & Thiemens M. H. (1997) The 180/160 and 170/160 ratios in atmospheric nitrous oxide: A mass-independent anomaly. Science, 278: 1774-1776. [18, 19] Cliff S. S., Thiemens M. H. & Brenninkmeijer C. A. M. (1999) First measurement of the 180/160 and 170/160 ratios in stratospheric nitrous oxide: A mass independent anomaly. J. Geophys. Res., 104: 16171-16175.[19] Cline J. D. & Kaplan I. R. (1975) Isotopic fractionation of dissolved nitrate during denitrification in the eastern tropical North Pacific ocean. Marine Chem., 3: 271-299. [15] Close K. J. & Yarwood J. (1972) A precision electron emission regulator. Vacuum, 22: 45-46. [39] Coburn J. W. & Kay E. (1971) A new technique for the elemental analysis of thin surface layers of solids. Phys. Lett., 19: 350-352. [35] Coburn J. W. & Harrison W. W. (1981) Plasma sources in analytical mass spectrometry. Appl. Spec. Rev., 17: 95-164. [35] Cohn M. & Urey H. C. (1938) Oxygen exchange reactions of organic compounds and water. J. Am. Chem. Soc., 60: 679-682. [1] Cole D. R., Ohmoto H. & Lasaga A. C. (1983) Isotopic exchange in mineral-fluid systems. I. Theoretical evaluation of oxygen isotopic exchange accompanying surface reactions and diffusion. Geochim. Cosmochim. Acta, 47: 1681-1693. [46] Cole D. R. & Ohmoto H. (1986) Kinetics of isotopic exchange at elevated temperatures and pressures. In: Stable isotopes in high temperature geological processes. J. W. Valley, H. P. Jr. Taylor & J. R. O'Neil (eds.). Mineral. Soc. America, Reviews in Mineralogy, Vol. 16: 41-90. [46] Cole D. R. (1992) Influence of solution composition and pressure on the rates of oxygen isotope exchange in the system: calcite-H20-NaC1 at elevated temperatures. Chem. Geol. (Isot. Geosc. Sect.), 102: 199216. [46] Cole D. R. & Ripley E. M. (1998) Oxygen isotope fractionation between chlorite and water from 170 to 350~ a preliminary assessment based on partial exchange and fluid/rock experiments. Geochim. Cosmochim. Acta, 63: 449-457. [46] Cole D. R. (2000) Isotopic exchange in mineral-fluid systems. IV. The crystal chemical controls on oxygen isotope exchange rates in carbonate-H20 and layer silicate-H20 systems. Geochim. Cosmochim. Acta, 64: 921-931. [46] Cole D. R. & Chakraborthy S. (2001) Rates and mechanisms of isotopic exchange. Rev. Mineral. Geochem., 43: 83-223. [46] Coleman D. C. & Fry B. (1991) Carbon isotope techniques. Isotopic techniques in plant, soil, and aquatic biology series. E. A. Paul & J. M. Melillo (series eds.). Academic Press, San Diego: 274 p. [9]
References
1075
Coleman D. D., Risatti J. B. & Schoell M. (1981) Fractionation of carbon and hydrogen isotopes by methane-oxidizing bacteria. Geochim. Cosmochim. Acta, 45: 1033-1037. [46] Coleman J. J., Xu x., Thiemens M. H. & Trogler W. C. (1996) Photopolymerization and mass independent fractionations in carbon disulfide. Science, 273: 774-776. [18] Coleman M. L. & Moore M. P. (1978) Direct reduction of sulfates to sulfur dioxide for isotopic analysis. Anal. Chem., 50: 1594-1595. [26, 44] Coleman M. L. (1980) Corrections for mass spectrometer analysis of SO2.Stable Isotope Report, Inst. Geol. Sci., London: 12 p. [44, 45] Coleman M. L., Shepherd T. J., Durham J. J., Rouse J. E. & Moore G. R. (1982) Reduction of water with zinc for hydrogen isotope analysis. Anal. Chem., 54: 993-995. [1, 3, 21, App C] Coleman R. G. & Delevaux M. H. (1957) Occurrence of selenium in sulfides from some sedimentary rocks of the western United States. Econ. Geol., 52(5): 499-527. [29] Collister J. W., Rieley G., Stern B., Eglinton G. & Fry B. (1994) Compound-specific 812C analyses of leaf lipids from plants with differing carbon dioxide metabolisms. Org. Geochem., 21: 619-627. [9] Colman A. S., Karl D. M., Fogel M. L. & Blake R. E. (2000) A new technique for the measurement of phosphate oxygen isotopes of dissolved inorganic phosphate in natural waters. EOS Trans. AGU, 81(48, Fall Meeting Suppl.): 176 (Abstract). [22] Compston W. (1996) SHRIMP: origins, impact and continuing evolution. J. Royal Soc. West. Aust., 79: 109-117. [35] Condon A. G., Richards R. A. & Farquhar G. D. (1992) The effect of variations in soil water availability, vapour pressure deficit and nitrogen nutrition on carbon isotope discrimination in wheat. Aust. J. Agric. Res., 43: 935-947. [46] Connie J. M. & Currie V. A. (1996) The isotopic characterization of methane, non-methane hydrocarbons and formaldehyde in the troposphere. Atmosph. Environ., 30(4): 621-638. [14] Conway C. M. & Taylor H. P. Jr. (1969) O18/O16 and C 13/C 12 ratios of coexisting minerals in the Oka and Magnet Cove carbonatite bodies. J. Geol., 77: 618-626. [20] Conway E. J. (1947) Microdiffusion analysis and volumetric error. Crosby, Lockwood, London. [15] Conway T. J., Tans P. P., Waterman L. S., Thoning K. W., Kitzis D. R., Massarie K. A. & Zhang N. (1994) Evidence for interannual variability of the carbon cycle for the National Oceanic and Atmospheric / Climate monitoring diagnostics Laboratory Global Network. J. Geophys. Res., 99(Dll): 22831-22855. [14] Cooper D. E., Martinelli U., Carlisle C. B., Riris H., Bour D. B. & Menna R. J. (1993) Measurement of 12CO2:13CO2 ratios for medical diagnostics with 1.6-~m distributed-feedback semiconductor diode lasers. Appl. Opt., 32: 6727-6731. [34] Cooper G. W., Thiemens M. H., Jackson T. L. & Chang S. (1997) Sulfur and hydrogen isotope anomalies in meteorite sulfonic acids. Science, 277: 1072-1074. [11, 18] Cooper L. W., DeNiro M. J. & Keeley J. E. (1991) The relationship between stable oxygen and hydrogen isotope ratios of water in astomatal plants. In: Stable Isotope Geochemistry: A tribute to Samuel Epstein. H. P. Jr. Taylor, J. R. O'Neil & I. R. Kaplan (eds.). The Geochemical Society, Spec. Publ. 3: 247255. [46] Cooper W. C. & Glover J. R. (1974) The toxicology of selenium and its compounds. In: Selenium. R. A. Zingaro & W. C. Cooper (eds). Van Nostrand Reinhold Co: 654-674. [29] Coplen T. B. & Hanshaw B. B. (1973) Ultrafiltration by a compacted clay membrane - 1. Oxygen and hydrogen isotopic fractionation. Geochim. Cosmochim. Acta, 37: 2295-2310. [2] Coplen T. B. & Kendall C. (1982) Preparation and stable isotope determination of NBS-16 and NBS-17 carbon dioxide reference samples. Anal. Chem., 54: 2611-2612. [1, 40, 43] Coplen T. B., Kendall C. & Hopple J. (1983) Comparison of stable isotope reference samples. Nature, 302: 236-238. [12, 14, 40] Coplen T. B. (1988) Normalization of oxygen and hydrogen isotope data. Chem. Geol. (Isot. Geosc. Sect.), 72: 293-297. [1, 38] Coplen T. B., Wildman J. D. & C h e n J. (1991) Improvements in the gaseous hydrogen-water equilibration technique for hydrogen isotope ratio analysis. Anal. Chem., 63: 910-912. [1, 21] Coplen T. B., Krouse H. R. & Biihlke J. K. (1992) Reporting of nitrogen-isotope abundances. Pure Appl. Chem., 64(6): 907-908. [43]
1076
References
Coplen T. B. (1994) Reporting of stable hydrogen, carbon, and oxygen isotopic abundances. Pure & Appl. Chem., 66(2): 273-276. [21] Coplen T. B. & Harper I. T. (1994) An improved technique for the 2H/1H analysis of urines from diabetic volunteers. Biol. Mass Spectrom., 23: 437-439. [1] Coplen T. B. (1996a) New guidelines for reporting stable hydrogen, carbon, and oxygen isotope-ratio data. Geochim. Cosmochim. Acta, 60(17): 3359-3360. [14, 43] Coplen T. B. (1996b) Atomic weights of the elements 1995. Pure Appl. Chem., 68: 2339-2359. [40] Coplen T. B. (1996c) Editorial: More uncertainty than necessary. Paleoceanography, 11(4): 369-370 [40] Coplen T. B., De Bii~vre P., Krouse H. R., Vocke R. D. J., Griining M. & Rozanski K. (1996) Ratios for light-element isotopes standardized for better interlaboratory comparison. EOS Trans. AGU, 77(27): 255-256. [40] Coplen T. B. & Krouse H. R. (1998) Sulphur isotope data consistency improved. Nature, 392: 32. [26, 27] Coplen T. B. (2000) Laboratory information management system (LIMNS) for light stable isotopes. U.S. Geological Survey. Open-File Report 00-345:121 p. [15] Coplen T. B. (2001a) Manual for operation of an isotope hydrology laboratory. IAEA Guidelines, IAEA, Vienna. [1] Coplen T. B. (2001b) Atomic weights of the elements 1999. Pure Appl. Chem., 73: 667-683. [40] Coplen T. B., Hopple J. A., Bi~hlke J. K., Peiser H. S., Rieder S. E., Krouse H. R., Rosman K. J. R., Ding T., Vocke R. D. Jr., R4v6sz K. M., Lamberty A., Taylor P. D. P. & De Bi6vre P. (2002) Compilation of minimum and maximum isotope ratios of selected elements in naturally occurring terrestrial materials and reagents. U.S. Department of the Interior- U.S. Geological Survey, Reston, Virginia. WaterResources Investigations Report 01-4222:98 p. [26, 40] Cormie A. R., Luz B. & Schwarcz H. P. (1994) Relationship between the hydrogen and oxygen isotopes of deer bone and their use in the estimation of relative humidity. Geochim. Cosmochim. Acta, 58: 34393449. [46] Cornwell J. C. & Morse J. W. (1987) The characterization of iron sulfide minerals in anoxic marine sediments. Marine Chem., 22: 193-206. [26] Corso N. T. & Brenna J. T. (1997) High-precision position-specific isotope analysis. Proc. Natl. Acad. Sci., U.S.A., 94: 1049-1053. [8] Cortecci G. & Longinelli A. (1970) Isotopic composition of sulfate in rain water, Pisa, Italy. Earth Planet. Sci. Lett., 8: 36-40. [26] Council N. R. (1981) Prudent practices for handling hazardous chemicals in laboratories. National Academic Press, Washington, DC. [9] Cowie G. L. & Hedges J. I. (1984) Determination of neutral sugars in plankton, sediments, and wood by capillary gas chromatography of equilibrated isomeric mixtures. Anal. Chem., 56: 497-504. [9] Craig H. (1953) The geochemistry of the stable carbon isotopes. Geochim. Cosmochim. Acta, 3: 53-92. [40] Craig H. (1957) Isotopic standards for carbon and oxygen and correction factors for mass-spectrometric analysis of carbon dioxide. Geochim. Cosmochim. Acta, 12: 133-149. [1, 14, 15, 19, 23, 38, 40, 43, 44] Craig H. (1961) Standard for reporting concentrations of deuterium and oxygen-18 in natural waters. Science, 133: 1833-1834. [40, 44] Craig H. & Gordon L. I. (1965) Deuterium and oxygen-18 variations in the ocean and the marine atmosphere. In: Stable isotopes in oceanographic studies and palaeotemperatures. E. Tongiorgio (ed.). Spoleto Conference, Italy. Laboratorio di Geologia Nucleare, Pisa: 9-130. [1, 21] Craig H., Marti K. & Wiens R. (1993) A static mass spectrometer with triple collection for nitrogen and neon isotopes. Scripps Inst. Oceanogr., San Diego. Techn. Rep. No93-11. [17] Creaser R. A., Papanastassiou D. A. & Wasserburg G. J. (1991) Negative thermal ion mass spectrometry of osmium, rhenium, and iridium. Geochim. Cosmochim. Acta, 55(1): 397-401. [29] Cremonini M. A., Tacconi D., Clementi V. & Luchinat C. (1998) Accurate determination of deuterium/ hydrogen ratios in natural organic compounds through a nuclear magnetic resonance time-domain reference convolution method: application to ethanol from three botanical sources and critical analysis of systematic inaccuracies in previous method. J. Agric. Food Chem., 46: 3943-3946. [5] Criss R. E., Gregory R. T. & Taylor H. P. Jr. (1987) Kinetic theory of oxygen isotope exchange between minerals and water. Geochim. Cosmochim. Acta, 51: 1099-1108. [46] Criss R. E. (1999) Principles of Stable Isotope Distribution. Oxford University Press, Oxford, UK. [46]
References
1077
Croce de Cobos A. E. & Troe J. (1984) High-pressure range of the recombination O + 02 --+ 03. Int. J. Chem. Kinet., 16: 1519-1530. [18] Cronin J. R., Pizzarello S. & Cruikshank D. P. (1988) Organic matter in carbonaceous chondrites, planetary satelites, asteroids and comets. In: Meteorites and the early solar system. J. F. Kerridge & M. S. Mathews (eds.). Univ. Arizona Press: 819-857. [11] Cronin J. R. & Pizzarella S. (1990) Aliphatic hydrocarbons of the Murchison meteorite. Geochim. Cosmochim. Acta, 54: 2859-2868. [11] Cronin J. R. & Chang S. (1993) Organic matter in meteorites: Molecular and isotopic analyses of the Murchison meteorite. In: Chemistry of life's origins. J. M. Greenburg & V. Pirronello (eds.). Kluwer: 209-258. [11] Cronin J. R., Pizzarello S., Epstein S. & Krishnamurthy R. V. (1993) Molecular and isotopic analyses of the hydroxy acids, dicarboxylic acids and hydroxylicarboxylic acids of the Murchison meteorite. Geochim. Cosmochim. Acta, 57: 4745-4752. [11] Cross W. G. (1951) Two-directional focussing of charged particles with a sector-shaped, uniform magnetic field. Rev. Sci. Instrum., 22: 717-722. [38] Crosson E. R., Ricci K. N., Richman B. A., Chilese F. C., Owano T. G., Provencal R. A., Todd,M. W., Glasser J., Kachanov A. A. & Paldus B. A. (2002) Stable Isotope Ratios Using Cavity Ring-Down Spectroscopy: Determination of 13C/12C for Carbon Dioxide in Human Breath. Anat. Chem., 74: 2003-2007. [34] Crowe D. E., Valley J. W. & Baker K. L. (1990) Micro-analysis of sulfur-isotope ratios and zonation by laser microprobe. Geochim. Cosmochim. Acta, 54: 2075-2092. [20, 26, 45] Crowe D. E. & Vaughan R. G. (1996) Characterization and use of isotopically homogeneous standards for in situ laser microprobe analysis of 34S/32S ratios. Am. Mineralagist, 81: 187-193. [45] Crowson R. A., Showers W. J., Wright E. K. & Hoering T. C. (1991) Preparation of phosphate samples for oxygen isotope analysis. Anal. Chem., 63: 2397-2400. [20, 22] Cuevas J., Villar M. V., Fern~ndes A. M., G6mez P. & Martin P. L. (1997) Pore waters extracted from compacted bentonite subjected to simultaneous heating and hydration. Appl. Geochem., 12: 473-481.
[2]
Cumrning G. L. (1973) Propagation of experimental errors in lead isotope ratio measurements using the double spike method. Chem. Geol., 11(3): 157-165. [29] Curie I. (1921) Sur le poids atomique du chlore dans quelques mineraux. C. R. Acad. Sci. Paris, 172: 10251028. [28] Czamanske G. K. & Rye R. O. (1974) Experimentally determined sulfur isotope fractionations between sphalerite and galena in the temperature range 600~ to 275~ Econ. Geol., 69:17-25 [46]. D'Allessandro M., Bidoglio G., Mousty F., Sala Benito J. V. & Yllera De Llano A. (1997) Laboratory and in situ investigation of the isotopic ratio 79Br/81Br as a tracer of water migration in crystalline rocks. J. Hydrol., 193: 351-362. [28] D'Angela D. & Longinelli A. (1990) Oxygen isotopes in living mammal's bone phosphate: Further results. Chem. Geol. (Isot. Geosci. Sect.), 86: 75-82. [22, 46] Dahnke H., Kleine D., Urban W., GHering P. & Miirtz M. (2001) Isotopic ratio measurement of methane in ambient air using mid-infrared cavity leak-out spectroscopy. Appl. Phys., B 72: 121-125. [34] Dai S., Ren D., Tang Y. G., Shao L. & Li S. (2002) Distribution, isotopic variation and origin of sulfur in coals in the Wuda coalfield, Inner Mongolia, China. Int. J. Coal Geology, 51: 237-250. [26] Dallwitz M. J. (1970) Fractionation corrections in lead isotopic analysis. Chem. Geol., 6(4): 311-314. [29] Damon P. E. & Kulp J. L. (1958) Excess helium and argon in beryl and other minerals. Am. Mineralogist, 43: 433-459. [16] Danho D., Naulet N. & Martin G. J. (1992) Deuterium, carbon and nitrogen analysis of natural and synthetic caffeines. Authentication of coffee and coffee extracts. Analusis, 20: 179-184. [5] Dansgaard W. (1953) The abundabce of O18 in atmospheric water vapour. Tellus, 5: 461-469. [1] Darimont A., Burke E. & Touret J. (1988) Nitrogen-rich metamorphic fluids in Devonian metasediments from Bastogne, Belgium. Bull. Mineralogie, 111: 321-330. [16]
1078
References
Datta B. P., Khodade P. S., Parab A. R., Goyal A. H. & Chitambar S. A. (1992) Thermal ionization mass spectrometry of Li2BO2+ ions: determination of the isotopic abundance ratio of lithium. Int. J. Mass Spectrom. Ion Proc., 116: 87-114. [6] Dauphas N. & Marty B. (1999) Heavy nitrogen in carbonatites of the Kola peninsula: A possible signature of hte deep mantle. Science, 286: 2488-2490. [17] Dauphas N., Reisberg L. & Marry B. (2001) Solvent extraction, ion chromatography, and mass spectrometry of molybdenum isotopes. Anal. Chem., 73: 2613-2616. [31] David M. B., Mitchell M. J., Aldcorn D. & Harrison R. B. (1989) Analysis of sulfur in soil, plant and sediment material: sample handling and use of an automated analyzer. Soil Biol. Biochem., 21: 119124. [26] Davies S., Spanel P. & Smith D. (1997) Quantitative analysis of ammonia on the breath of patients in end-stage renal failure. Kidney Int., 52: 223-228. [4] Davies S., Spanel P. & Smith D. (2001a) A new 'online' method to measure increased exhaled isoprene in end-stage renal failure. Nephrol. Dial. Transplant., 16: 836-839. [4] Davies S., Spanel P. & Smith D. (2001b) Rapid measurement of deuterium content of breath following oral ingestion to determine body water. Physiol. Meas., 22: 651-659. [4] Davis A. & Olsen R. L. (1995) The geochemistry of chromium migration and remediation in the subsurface. Ground Water, 33: 759-768. [29] Dawson P.H. (1995) Quadrupole mass spectrometry and its application. American Institute of Physiscs, New York, USA: p. 143. [32] De Bi6vre P. J. & Debus G. H. (1969) Absolute isotope ratio determination of a natural boron standard. Int. J. Mass Spectrom. Ion Phys., 2: 15-23. [40] De Bi6vre P. & Barnes I. L. (1985) Table of the isotopic composition of the elements as determined by mass spectrometry. Int. J. Mass Spectrom. Ion Proc., 65: 211-230. [35] De Bi6vre P. (1990) Isotope dilution mass spectrometry: what can it contribute to accuracy in trace analysis? Fresenius J. Anal. Chem., 337: 766-771. [37] De Bi6vre P. (1994) Isotope dilution mass spectrometry. In: Trace element analysis in biological specimens. R. F. M. Herber & M. Stoeppler (eds.). Elsevier, Vol. 15: 169-183. [32, 37, 42] De Bi6vre P. & Valkiers S. (1994) Molar mass determination for NA (Avogadro) projects. Present status and potential for the future. Metrologia, 31: 245-249. [25] De Bi6vre P., Valkiers S. & Peiser H. S. (1994) New values of silicon reference materials, certified for isotope abundance ratios. J. Res. Nat. Ist. Stand. Techn., 99(2): 201-202. [25] De Bi6vre P., Lenaers G., Murphy T. J., Peiser H. S. & Valkiers S. (1995) The chemical preparation and characterization of specimens for "absolute" measurements of the molar mass of an element on silicon, for redeterminations of the Avogadro constant. Metrologia, 32: 103-110. [25, 42] De Bolle W., Alonso A., Verbruggen A. & Wellum R. (1999) The blending of UF6 for the preparation of isotopic reference materials with specified 234U, 235U and 236U abundances. In: Symposium on safeguards and nuclear materials management. C. Foggi & E. Petraglia (EC-JRC-Ispra)(eds.). Proc. of the 21st ESARDA meeting. EC-JRC publication EUR-18963 EN: 585-587. [32] De Endredy A. S. (1963) Estimation of free iron oxides in soils and clays by photolytic method. Clay Minerals Bull., 5: 209-217. [2] De Gendt S., Schelles W., Van Grieken R. & Muller V. (1995a) Quantitative analysis of iron-rich and other oxide-based samples by means of glow discharge mass spectrometry. J. Anat. At. Spectrom., 10: 681-687. [35] De Gendt S., Van Grieken R., Hang W. & Harrison W. W. (1995b) Comparison between direct current and radiofrequency glow discharge mass spectrometry for the analysis of oxide-based samples. J. Anat. At. Spectrom., 10: 689-695. [35] DeLa Rocha C. L., Brzezinski M. A. & DeNiro M. J. (1996) Purification, recovery, and laser-driven fluorination of silicon from dissolved and particulate silica for the measurement of natural stable isotope abundances. Anal. Chem., 68(21): 3746-3750. [20, 25] De La Rocha C. L., Brzezinski M. A. & DeNiro M. J. (1997) Fractionation of silicon isotopes by marine diatoms during biogenic silica formation. Geochim. Cosmochim. Acta, 61: 5051-5056. [46] De La Rocha C. L., Brzezinski M. A. & DeNiro M. J. (2000) A first look at the distribution of the stable isotopes of silicon in natural waters. Geochim. Cosmochim. Acta, 64(14): 2467-2477. [20, 46]
References
1079
De La Rocha C. L. & DePaolo D. J. (2000) Isotopic evidence for variations in the marine calcium cycle over the Cenozoic. Science, 289: 1176-1178. [46] De La Rocha C. L. (2003) Silicon isotope fractionation by marine sponges and the reconstruction of the silicon isotope composition of ancient deep water. Geology, 31: 423-426. [20] De Laeter J. R. & Kennedy A. K. (1998) A double focussing mass spectrometer for geocghronology. Int. J. Mass Spectrom. Ion Proc., 178: 43-50. [35] De Laeter J. R. (2001) Applications of inorganic mass spectrometry. John Wiley & Sons, NY. [39] De Wit J. C., Van der Straaten C. M. & Mook W. G. (1980) Determination of the absolute hydrogen isotopic ratio of V-SMOW and SLAP. Geostand. Newslett., 4(1): 33-36. [40] de Wit M. J., Hart R., Martine A. & Abbott P. (1982) Archean abiogenic and probable biogenic structures associated with mineralised hydrothermal vent systems and regional metasomatism, with implications for greenstone belt studies. Econ. Geol., 77: 1783-1803. [13] Debeer E., Delange C. A. & Westwood N. P. C. (1992) Resonance-enhanced multiphoton-ionization photoelectron-spectroscopy of np and nf Rydberg states of atomic nitrogen. Phys. Rev., A 46: 56535665. [33] Dechan M. (1886) Detection and estimation of iodine, bromine, and chlorine. J. Chem. Soc., 49: 682-685.
[28]
Decitre S., Deloule E., Reisberg L., James R., Agrinier P. & Mevel C. (2002) Behavior of Li and its isotopes during serpentinization of oceanic peridotites. Geochem. Geophys. Geosyst., 3(1): doi 10.1029/ 2001GC000178. [6, 30] Dedina J. & Tsalev D. L. (1995) Hydride generation atomic absorption spectrometry. Wiley. [29] Deiana M., Corongiu F. P., Dessi M. A., Scano P., Casu M. & Lai A. (2001) NMR determination of sitespecific deuterium distribution (SNIF-NMR) in squalene from different sources. Magn. Reson. Chem., 39: 29-32. [5] Deines P. D., Langmuir D. & Harmon R. S. (1974) Stable carbon isotope ratios and the existence of a gas phase in the evolution of carbonate ground waters. Geochim. Cosmochim. Acta, 38: 1147-1164. [10] DeLong W. A. & Lowe L. E. (1961) Note on carbon-bonded sulphur in soil. Can. J. Soil Sci., 42: 223. [26] Deloule E., All6gre C. J. & Doe B. (1986) Lead and sulphur isotope microstratigraphy in galena crystals from Mississippi-valley type deposits. Econ. Geol., 81: 1307-1321. [26] Deloule E., France-Lanord C. & Albar6de F. (1991) D / H analysis of minerals by ion probe. In: Stable Isotope Geochemistry: A tribute to Samuel Epstein. H. P. Jr. Taylor, J. R. O'Neil & I. R. Kaplan (eds.). The Geochemical Society, Spec. Publ. 3: 53-62. [30] Dem6ny A. (1995) H isotope fractionation due to hydrogen-zinc reactions and its implications on D / H analysis of water samples. Chem. Geol., 121: 19-25. [1, App C] Demir I. & Harvey R. D. (1991) Variation of organic sulfur in macerals of selected Illinois Basin coals. Org. Geochem., 17: 525-533. [26] Demmelmair J. & Schmidt H.-L. (1993) Precise 613C-determination in the range of natural abundance of amino acids from protein hydrolysates by gas chromatography-mass spectrometry. Isot. Environ. Health Stud., 29: 237-250. [8] Dempster A. J. (1935) New methods in mass spectroscopy. Proceed. Am. Phil. Soc., 75: 755-767. [35, 38] Demtri3der W. (1982) Laser spectroscopy. Basic concepts and instrumentation. 2nd edition. Springer Verlag, Berlin. [34] DeNiro M. J. & Epstein S. (1978) Influence of diet on the distribution of stable isotopes in animals. Geochim. Cosmochim. Acta, 42: 495-506. [46] DeNiro M. J. & Epstein S. (1979) Relationship between the oxygen isotope ratios of terrestrial plant cellulose, carbon dioxide and water. Science, 204: 51-53. [21, 46] DeNiro M. J. (1981) The effects of different methods of preparing cellulose nitrate on the determination of the D / H ratios of non-exchangeable hydrogen of cellulose. Earth Planet. Sci. Lett., 54: 177-185. [24] DeNiro M. J. & Epstein S. (1981a) Isotopic composition of cellulose from aquatic organisms. Geochim. Cosmochim. Acta, 45: 1885-1894. [46] DeNiro M. J. & Epstein S. (1981b) Influence of diet on the distribution of nitrogen isotopes in animals. Geochim. Cosmochim. Acta, 45: 341-351. [46]
1080
References
DeNiro M. J. & Cooper L. W. (1989) Post-photosynthetic modification of oxygen isotope ratios of carbohydrates in the potato: implications for paleoclimatic reconstruction based upon isotopic analysis of wood cellulose. Geochim. Cosmochim. Acta, 53: 2573-2580. [46] Dennis P. F., Rowe P. J. & Atkinson T. C. (1998) Stable isotope composition of palaeprecipitation and palaeogroundwaters from speleothem fluid inclusions. In: Isotope techniques in the study of environmental change. IAEA, Vienna: 663-671. [3] Dennis P. F., Rowe P. J. & Atkinson T. C. (2001) The recovery and isotopic measurement of water from fluid inclusions in speleothem. Geochim. Cosmochim. Acta, 65(6): 871-884. [3] Deqiu J. & Zhengxin W. (1985) Determination of the deuterium content in water with Mg-Pt reduction method. Acta Scient. Natur. Univers. Pekin., 1:42-46 (in Chinese; English translation by R. Krouse). [1] Dequasie H. L. & Grey D. C. (1970) Stable isotopes applied to pollution studies. Am. Lab., 2: 19-27. [26] Derdau D., Freer R. & Wright K. (1998) Oxygen diffusion in anhydrous sanidine feldspar. Contrib. Mineral. Petrol., 133: 199-204. [46] Derwish G. A. W., Galli A., Giardini-Guidoni A. & Volpi G. G. (1964) High-pressure mass spectrum of N20. J. Chem. Phys., 40: 3450-3451. [19] Desage M., Guilluy R., Brazier J. L., Chaudron H., Girard J., Cherpin H. & Jumeau J. (1991) Gaschromatography with mass-spectrometry or isotope-ratio mass-spectrometry in studying the geographical origin of heroin. Anal. Chim. Acta, 247: 249-254. [8] Desaulniers D. E., Kaufmann R. S., Cherry J. A. & Bentley H. W. (1986) 37C1-35C1 variations in a diffusion-controlled groundwater system. Geochim. Cosmochim. Acta, 50: 1757-1764. [28, 46] DesMarais D. J. (1978a) Carbon, nitrogen, and sulfur in Apollo 15, 16, and 17 rocks. Proceedings of the 9th Lunar Planet. Sci. Conf.: 1451-2467. [13] DesMarais D. J. (1978b) Variable-temperature cryogenic trap for the separation of gas mixtures. Anal. Chem., 50: 1405-1406. [20] DesMarais D. J., Mitchell J. M., Meinschein W. G. & Hayes J. M. (1980) The carbon isotope biogeochemistry of the individual hydrocarbons in bat guano and the ecology of the insectivorous bats in the egion of Carlsbad, New Mexico. Geochim. Cosmochim. Acta, 44: 2075-2086. [9] DesMarais D. J. & Moore J. G. (1984) Carbon and its isotopes in mid-oceanic basaltic glasses. Earth Planet. Sci. Lett., 69: 43-57. [3] Dewar W. A. & McDonald P. (1961) Determination of dry matter in silage by distillation with toluane. J. Sci. Food Agric., 12: 790-795. [21] Dews J. R. (1966) The isotopic composition of lithium in chondrules. J. Geophys. Res., 71: 4011-4020. [6] Deyhle A. (2001) Improvements of boron isotope analysis by positive thermal ionization mass spectrometry using static multicollection of Cs2BO2+ ions. Int. J. Mass Spectrom., 206: 79-89. [40] Dhamala B. R. & Mitchell M. J. (1992) Utilization of 35S in analyzing sulfur transformations and fluxes in forest soils. Proceedings of NHRI Symposium on Sulphur Transformations in Soil Ecosystems. Saskatoon (Canada): 297-314. [26] Dhamenlincourt P., Beny J. M., Dubessy J. & Poty B. (1979) Analysis of fluid inclusions with the MOLE Raman microprobe. Bull. Mindralogie, 102: 600-610. [3] Dickman M. & Thode H. (1990) Sulphur bacteria and sulphur isotope fractionation in a meromictic lake near Toronto, Canada. In: Facets of modern biogeochemistry. V. Ittekot, S. Kempe, W. Michaelis & A. Spitzy (eds.). Springer Verlag: 225-241. [29] Dickson F. W., Blount C. W. & Tunell G. (1963) Use of hydrothermal solution equipment to determine the solubility of anhydrite in water from 100~ to 275~ and from I bar to 1000 bars pressure. Am. J. Sci., 261: 61-78. [46] Dickson J. A. D. (1991) Disequilibrium carbon and oxygen isotope variations in natural calcite. Nature, 353: 842-844. [22] Diemer J., Qu4tel C. R. & Taylor P. D. P. (2002) Comparison of the performance of different ICP-MS instruments on the measurement of Cu in a water sample by ICP-IDMS. J. Anal. At. Spectrom., 17: 1137-1142. [32] Dimeff, J. (1972) Nondispersive gas analyzing method and apparatus wherein radiation is serially passed through a reference and unknown gas. U.S. Patent 3,679,899. [34]
References
1081
Ding T.-P., Li-Hong, Zhang-Guobing, Li-u & Fan-Tianyi (1985) Sulphur isotope determination by using the SF6 method. Chinese Academy of Geological Sciences, Beijing, China. Annual Rept, p.21: 163164. [20] Ding T.-P., Li-Hong, Zhang-Guobing, Li-Yanhe & Li-Jincheng (1987) A study on determining 33S/32S, 34S/ 32S and 36S/ 32S by SF6 method. Mineral Deposits, 6: 81-93. [20] Ding T., Wan D., Li J., Song H., Li Y. & Liu Z. (1988) The analytical method of silicon isotopes and its geological application. Mineral Deposits, 7(4): 90-95 (in Chinese). [20, 25] Ding T.-P. & Zhang Q.-L. (1989) Analysis of the reference material NBS-123 and the atomic weight of sulfur. Chin. Sci. Bull., 34(13): 1086-1089. [45] Ding T., Wan D., Jiang S. & Liu Z. (1990) A study on silicon compositions of some rocks and minerals. ICOG-7, Canberra: 27 (Abstract). [25] Ding T. (1991) A study on analytic method of silicon isotope and its geological application. 2nd ChinaJapan joint symposium on geochronology, cosmochronology and isotope geology, Guangzhou, China: 11 (Abstract). [25] Ding T., Jiang S., Wan D. & Li Y. (1994) The silicon isotope variations in geochemical circulation. ICOG8, Berkeley. U.S. Geol. Surv. Circular 1107:82 (Abstract). [25] Ding T., Jiang S., Was D., Li Y., Li J., Song H., Liu Z. & Yao X. (1996) Silicon isotope geochemistry. Geological Publishing House, Beijing, China: 125 p. [20, 25] Ding T., Wang C., Zhang F. & Wan D. (1998) Silicon isotope fractionation in some surface processes (A study on river water and plants). ICOG-9, Beijing, Chin. Sci. Bull., 43 (Supplem.): 33 (Abstract). [25] Ding T., Bai R., Li Y., Wan D., Zou X. & Zhang Q. (1999) Determination of the absolute 32S/34S ratio of IAEA-S-1 reference material and V-CDT sulfur isotope standard. Sci. China, Ser. D (Earth Sciences), 42: 45-51. [40, 44] Ding T., Valkiers S., Kipphardt H., Qu6tel C., De Bi6vre P., Taylor P. D. P. & Gonfiantini R. (2000) Preparation of two synthetic mixtures for the calibration of isotope amount ratio measurements of sulfur. Int. J. Mass Spectrom., 197: 131-137. [42] Ding T., Valkiers S., Kipphardt H., De Bi6vre P., Taylor P. D. P., Gonfiantini R. & Krouse R. (2001) Calibrated sulfur isotope abundance ratios of three IAEA sulfur isotope reference materials and VCDT with a reassessment of the atomic weight of sulfur. Geochim. Cosmochim. Acta, 65(15): 2433-2437. [26, 27, 40, 42] Ding T., Wan D., Wang C. & Zhang F. (2002) Si, O, H isotopic compositions of water and suspended materials in the Yangtze River. Geochim. Cosmochim. Acta, 66:A185 (Abstract). [20] Ding T., Wan D., Wang C. & Zhang F. (2003) Silicon isotope compositions of dissolved silicon and suspended matter in the Yangtze River, China. Geochim. Cosmochim. Acta, 67: in press. [20] Dittel A. I., Epifanio C. E., Cifuentes L. A. & Kirchman D. L. (1997) Carbon and nitrogen sources for shrimp postlarvae fed natural diets from a tropical mangrove system. Est. Coast. Shelf Sci., 45: 629637. [9] Divis A. F. & Clark J. R. (1978) Exploration for blind ore deposits and geothermal reservoirs by lithium isotope thermometry- atomic absorption "mass spectrometry". Assoc. Explor. Geochem. Proceedings of the Seventh International Geochemical Exploration Symposium, Rexdale, Ontario, Canada: 233241. [6] Dixon P. R., Perrin R. E., Rokop D. J., Maeck R., Janecky D. R. & Banar J. P. (1993) Measurements of iron isotopes (54Fe, 56Fe, 57Fe, 58Fe) in sub microgram quatities of iron. Anal. Chem., 65: 2125-2130. [29] Dlugokencky E. J., Masarie K. A., Lang P. M. & Tans P. P. (1998) Continuing decline in the growth rate of the atmospheric methane burden. Nature, 393(No. 6684): 447-450. [14] Dobson P. F., Epstein S. & Stolper E. M. (1989) Hydrogen isotope fractionation between coexisting vapor and silicate glasses and melts at low pressure. Geochim. Cosmochim. Acta, 53: 2723-2730. [46] Dodson M. H. (1970) Simplified equations for double-spiked isotopic analyses. Geochim. Cosmochim. Acta, 34(11): 1241-1244. [29] Doering W. v. E. & Dorfman E. (1953) Mechanism of the peracid ketone-ester conversion. Analysis of organic compounds for oxygen-18. J. Am. Chem. Soc., 75: 5595-5598. [23] Dongmann G., Niirnberg H. W., Fi~rstel H. & Wagener K. (1974) On the enrichment of H2180 in the leaves of transpiring plants. Radiat. Environm. Biophys., 11: 41-52. [21]
1082
References
D o n o h u e D. L. & Harrison W. W. (1975) Radiofrequency cavity ion source in solids mass spectrometry.
Anal. Chem., 47: 1528-1531. [35] Donohue D. L. & Petek M. (1991) Isotopic measurement of palladium metal containing protium and deuterium by glow discharge mass spectrometry. Anal. Chem., 63: 740-744. [35] Dontsova E. I. (1959) Method of determination of oxygen isotope ratios in igneous rocks and minerals. Geokhimiya, 8:669-678 (in Russian). [20] Dore J. E., Popp B. N., Karl D. M. & Sansone F. J. (1998) A large source of atmospheric nitrous oxide from subtropical North Pacific surface waters. Nature, 396: 63-66. [19] Dostrovsky I. & Klein F. S. (1952) Mass spectrometric determination of oxygen in water samples. Anal. Chem., 24: 414-415. [1] Doucett R. R., Hooper W. & Power G. (1999) Identification of anadromous and nonamadromous adult brook trout and their progeny in the Tabusintac River, New Brunswick, by means of multiple stable isotope analysis. Trans. Am. Fish. Soc., 128: 278-288. [9] Douglas D. J. & Tanner S. D. (1998) Fundamental considerations in ICPMS. In: Inductively coupled plasma mass spectrometry. A. Montaser (ed.). Wiley-VCH: 615-679. [31] Dourson M., Stern B., Griffin S. & Bailey K. (1991) Impact of risk-related concerns on the U.S. Environmental Protection Agency Programs. In: Nitrate contamination:Exposure, consequence, and control. I. Bogardi & R. D. Kuzelka (eds.). Springer Vertag, Berlin. NATO SI Ser. G: Ecological Sciences, Vol. 30: 477-487. [15] Douthitt C. B. (1982) The geochemistry of the stable isotopes of silicon. Geochim. Cosmochim. Acta, 46: 1449-1458. [20, 25] Dowdle P. R. & Oremland R. S. (1998) Microbial oxidation of elemental selenium in soil slurries and bacterial cultures. Environ. Sci. Technol., 32: 3749-3755. [29] Dowling L. M., Boreham C. J., Hope J. M., Murray A. P. & Summons R. E. (1995) Carbon isotopic composition of hydrocarbons in ocean-transported bitumens from the coastline of Australia. Org. Geochem., 23: 729-737. [8] Downey S. W., Emerson A. B., Kopf R. F. & Schubert E. F. (1992) Resonance ionization massspectrometry of device materials. Inst. Phys. Conf. Series, 128: 255-258. [33] Downs M. R., Michener R. H., Fry B. & Nadelhoffer K. J. (1999) Routine measurement of dissolved inorganic 15N in streamwater. Environ. Monitoring Assessment, 55: 211-220. [15] Dreibus G., Spettel B. & W~inke H. (1979) Halogens in meteorites and their primordial abundances. Origin and distribution of the elements. Proceedings of the second Symposium, Paris 1977. Pergamon Press: 33-38. [28] Drever J. I. (1982) The Geochemistry of Natural Waters. Prentice-Hall, Englewood Cliffs, NJ: 388 p. [26] Driesner T. (1997) The effect of pressure on deuterium-hydrogen fractionation in high-temperature water. Science, 277: 791-794. [46] Driesner T. & Seward T. M. (2000) Experimental and simulation study of salt effects and pressure/ density effects on oxygen and hydrogen stable isotope liquid-vapor fractionation for 4-5 molal aqueous NaC1 and KC1 solutions to 400~ Geochim. Cosmochim. Acta, 64(10): 1773-1784. [46] Du Y.-C., Nose A., Wasano K. & Uchida Y. (1998) Responses to water stress of enzyme activities and metabolite levels in relation to sucrose and starch synthesis, the Calvin cycle and the C4 pathway in sugarcane (Saccharumsp.) leaves. Aust. J. Plant Physiol., 25: 253-260. [9] Duan W.-M., Coleman M. L. & Pye K. (1997) Determination of reduced sulphur species in sediments an evaluation and modified technique. Chem. Geol., 141: 185-194. [26] Dubinina E. O. & Lakshtanov L. Z. (1997) A kinetic model of isotopic exchange in dissolution-precipitation processes. Geochim. Cosmochim. Acta, 61: 2265-2273. [46] Dubois A. (1985) An accurate zinc charcoal reduction system for D / H measurements of water and cellulose. Bull. Belg. Veren. Geol., 94: 113-115. [1] DuChateau N. L., Verbruggen A., Hendrickx F. & De Bi6vre P. (1986) Rapid accurate isotopic measurements on boron in boric acid and boron carbide. Talanta, 33: 291-294. [7] Duckworth D. C., Barshick C. M., Bostick D. A. & Smith D. H. (1993) Direct measurement of uranium isotopic ratios in soils by glow discharge mass spectrometry. Appt. Spectrosc., 47: 243-245. [35]
References
1083
Dugan J. P., Borthwick J., Harmon R. S., Gagnier M. A., Glahn J. E., Kinsel E. P., McLeod S., Viglino J. A. & Hess J. W. (1985) Guadinine hydrochloride method for determination of water oxygen isotope ratios and the oxygen-18 fractionation between carbon dioxide and water at 25~ Anal. Chem., 57: 1734-1736. [1, 3, 20] Dugan J. P. & Borthwick J. (1986) Carbon dioxide-water oxygen isotope fractionation factor using chlorine trifluoride and guanidine hydrochloride techniques. Anat. Chem., 58: 3052-3054. [1] Duit W., Jansen J. B. H., van Bremen A. & Bos A. (1986) Ammonium micas in metamorphic rocks as exemplified by Dome de L'Agout (France). Am. J. Sci., 286: 702-732. [16] Dumke I., Faber E. & Poggenburg J. (1989) Determination of stable carbon and hydrogen isotopes of light hydrocarbons. Anal. Chem., 61(19): 2149-2154. [43] Dungan R. S. & Frankenberger W. T. Jr. (1998) Reduction of selenite to elemental selenium by Enterobacter cloacae SLDla-1. J. Env. Qual., 27: 1301-1306. [29] Dunstan L. P., Gramlich J. W., Barnes I. L. & Purdy W. C. (1980) Absolute isotopic abundance and the atomic weight of a reference sample of thallium. J. Res. Natl. Bur. Stand., 85: 1-10. [31] Durakiewicz T. & Halas S. (1995) A novel electronic circuit for a Pirani gauge. Vacuum, 46: 101-103. [39] Durakiewicz T. (1996) Electron emission controller with pulsed heating of filament. Int. J. Mass Spectrom. Ion Proc., 156: 31-40. [39] Durand B., Espitali6 J., Nuaise G. & Combaz A. (1972) ]~tude de la mati6re organique insoluble (kerogene) des argiles du Joacien du bassin de Paris. Premiere partie: ]~tude par les proced6s optiques, analyse elementaire, 6tude en microscopie et diffraction 61ectroniques. Rev. L'Inst. Franfais du Petrol, 27: 865-884. [26] Durand B. & Nicaise G. (1980) Procedures for kerogen isolation. In: Kerogen- Insoluable organic matter from sedimentary rocks. G. Durand (ed.). Edition Technip, Paris: 113-142. [2] Durka W., Schulze E. D., Gebauer G & Voerkelius S. (1994) Effects of forest decline on uptake and leaching of deposited nitrate determined from 15N and 180 measurements. Nature, 372: 765-767. [15] Durrant S. F. (1999) Laser ablation inductively coupled plasma mass spectrometry: achievements, problems and prospects. J. Anal. At. Spectrom., 14: 1385-1403. [20] Eakin P. A., Fallick A. E. & Gerc J. (1992) Some instrumental effects in the determination of stable carbon isotope ratios by gas chromatography - isotope ratio mass spectrometry. Chem. Geol. (Isot. Geosc. Sect.), 101: 71-79. [8, 9] Eastoe C. J., Guilbert J. M. & Kaufmann R. S. (1989) Preliminary evidence for fractionation of stable chlorine isotopes in ore forming hydrothermal systems. Geology, 17: 285-288. [28] Eastoe C. J. & Guilbert J. M. (1992) Stable chlorine isotopes in hydrothermal systems. Geochim. Cosmochim. Acta, 56: 4247-4255. [28] Eberl D. D., Srodon J., Kralik M., Taylor B. E. & Peterman Z. E. (1990) Ostwald ripening of clays and metamorphic minerals. Science, 248: 474-477. [46] Eckelmann W. R., Broecker W. S., Whitlock D. W. & Allsup J. R. (1962) Implications of carbon isotopic composition of total organic carbon of some recent sediments and ancient oils. AAPG Bull., 46: 699704. [40] Ecker K. H. & Pritzkow W. (1994) Measurement of isotope ratios in solids by glow discharge mass spectrometry. Fresenius J. Anal. Chem., 349: 207-208. [35] Edgar A. D. (1973) Experimental Petrology: Basic Principles and Techniques. Oxford University Press, Oxford, UK: 217 p. [46] Edwards T. W. D. & Fritz P. (1986) Assessing meteoric water composition and relative humidity from 180 and 2H in wood cellulose: paleoclimatic implications for southern Ontario, Canada. Appl. Geochem., 1: 715-723. [46] Edwards T. W. D., Buhay W. M., Elgood R. J. & Jiang H. B. (1994) An improved nickel-tube pyrolysis method for oxygen isotope analysis of organic matter and water. Chem. Geol. (Isot. Geosci. Sect.), 114: 179-183. [1, 23] Eggenkamp H. G. M. (1994) 637C1. The geochemistry of chlorine isotopes. PhD Thesis, University of Utrecht, Utrecht, the Netherlands. Geologica Ultraiectina, No. 116:149 p. [28] Eggenkamp H. G. M., Kreulen R. & Koster van Groos A. F. (1995) Chlorine stable isotope fractionation in evaporites. Geochim. Cosmochim. Acta, 59: 5169-5175. [28, 46]
1084
References
Eggenkamp H. G. M. & Schuiling R. D. (1995) 637C1 variations in selected minerals: a possible tool for exploration. J. Geochem. Explor., 55: 249-255. [28] Eggenkamp H. G. M. & Koster van Groos A. F. (1997) Chlorine stable isotopes in carbonatites: evidence for isotopic heterogeneity in the mantle. Chem. Geol., 140: 137-143. [28, 46] Eggenkamp H. G. M. & Coleman M. L. (2000) The application of classical methods for the measurement of stable bromine isotopes in natural samples. Chem. Geol., 167: 393-402. [28] Ehhalt D. H. (1974) Vertical profiles of HTO, HDO, and H20 in the troposphere. NCAR Technical Note, NCAR-TN / STR-100. [1] Ehleringer J. R. & Osmond C. B. (1989) Stable isotopes. In: Plant physiological ecology. R. W. Pearcy, J. R. Ehleringer, H. A. Mooney & P. W. Rundel (eds.). Chapman and Hall, London: 281-299. [21] Ehleringer J. R. (1991) 13C/12C fractionation and its utility in terrestrial plant studies. In: Carbon isotope techniques. D. C. Coleman & B. Fry (eds.). Academic Press: 187-200. [9] Ehleringer J. R. & Dawson T. E. (1992) Water uptake by plants: perspective from stable isotope composition. Plant, Cell Environ., 15: 1073-1082. [21] Ehleringer J. R., Hall A. E. & Farquhar G. D. (1993) Stable isotopes and plant carbon-water relations. Academic Press, San Diego: 555p. [9] Ehleringer J. R. & Cook C. S. (1998) Carbon and oxygen isotope ratios of ecosystem respiuration along an Oregon conifer transect: preliminary observations based on small-flask sampling. Tree Physiol., 18: 513-519. [8] Ehleringer J. R., Casale J. F., Loft M. J. & V.L. Ford V. L. (2000) Tracing the geographical origin of cocaine. Nature, 408: 311-312, 2000. [8] Ehrlich H. L. (1996) Geomicrobiology. Marcel Dekker. [29] Eiler J. M., Baumgartner L. P. & Valley J. W. (1992) Intercrystalline stable isotope diffusion: a fast grain boundary model. Contrib. Mineral. Petrol., 112: 543-557. [46] Eiler J. M., Valley J. W. & Baumgartner L. P. (1993) A new look at stable isotope thermometry. Geochim. Cosmochim. Acta, 57: 2571-2583. [46] Eiler J. M., Graham C. W. & Valley J. W. (1997) SIMS analysis of oxygen isotopes: matrix effects in complex minerals and glasses. Chem. Geol., 138: 221-244. [30] Eiler J. M., Kitchen N. & Rahn T. A. (2000) Experimental constraints on the stable-isotope systematics of CO2 ice / vapor systems and relevance to the study of Mars. Geochim. Cosmochim. Acta, 64: 733-746. [46] Eiler J. M. & Kitchen M. (2001) Hydrogen-isotope analysis of nanomole (picomole) quantities of H20. Geochim. Cosmochim. Acta, 65: 4467-4479. [1] Eldridge C. S., Compston W., Williams I. S., Walshe J. L. & Both R. A. (1987) In situ microanalysis for 34S/ 32S ratios using the ion microprobe SHRIMP. Int. J. Mass Spectrom. Ion Proc., 76: 65-83. [20, 26, 30] Eldridge C. S., Compston W., Williams I. S., Both R. A., Walshe J. L. & Ohmoto H. (1988) Sulfur isotope variability in sediment hosted massive sulfide deposits as determined using the ion microprobe SHRIMP. I. An example from the Rammelsberg ore body. Econ. Geol., 83: 443-449. [26] Eldridge C. S., Compston W., Williams I. S. & Walshe J. L. (1989) Sulfur isotopic analyses on the SHRIMP ion microprobe. In: New frontiers in stable isotopic research: Laser probes, ion probes, and small-sample analysis. W. C. Shanks III & R. E. Criss (eds.). U.S. Geol. Surv., Bull. 1890: 163-174. [25] Ellis A. S., Johnson T. M. & Bullen T. D. (2002) Cr isotopes and the fate of hexavalent chromium in the environment. Science, 195: 2060-2062. [29] Ellis A. S., Johnson T. M., Bullen T. D. & Herbel M. J. (2003) Stable isotope fractionation of selenium by natural microbial consortia. Chem. Geol., 195: 119-129. [29] Ellis L. & Fincannon A. L. (1998) Analytical improvements in irm-GC/MS analyses: Advanced techniques in tube furnace design and sample preparation. Org. Geochem., 29: 1101-1117. [8] Elphick S. C., Dennis P. F. & Graham C. M. (1986) An experimental study of the diffusion of oxygen in quartz and albite using an overgrowth technique. Contrib. Mineral. Petrol., 92: 322-330. [46] Elphick S. C., Graham C. M. & Dennis P. F. (1988) An ion microprobe study of anhydrous oxygen diffusion in anorthite: a comparison with hydrothermal data and some geological implications. Contrib. Mineral. Petrol., 100: 490-495. [46] Elrashidi M. A., Adriano D. C., Workman S. M. & Lindsay W. L. (1987) Chemical equilibria of selenium in soils; a theoretical development. Soil Sci., 144(2): 141-152. [29]
References
1085
Elsenheimer D. & Valley J. W. (1992) In situ oxygen isotope analysis of feldspar and quartz by Nd:YAG laser micro probe. Chem. Geol. (Isot. Geosci. Sect.), 101: 21-42. [20] Elsenheimer D. & Valley J. W. (1993) Sub-millimeter scale zonation of 8180 in quartz and feldspar, Isle of Skye, Scotland. Geochim. Cosmochim. Acta, 57: 3669-3676. [20] Emiliani C. (1955) Pleistocene temperatures. J. Geol., 63: 538-578. [46] Engel M. H., Macko S. A. & Silfer J. A. (1990) Carbon isotope composition of individual amino acids in the Murchison meteorite. Nature, 348: 47-49. [9, 11] Engel M. H. & Macko S. A. (1997) Isotopic evidence for extraterrestrial non-racemic amino acids in the Murchison meteorite. Nature, 389: 265-268. [9, 11] Engeln R., yon Helden G., Berden G. & Meijer G. (1996) Phase shift cavity ring down absorption spectrometry. Chem. Phys. Lett., 262: 105-109. [34] Engeln R., Berden G., Peeters P. & Meijer G. (1998) Cavity enhanced absorption and cavity magnetic rotation apectroscopy. Rev. Sci. Instrum., 69: 3763-3769. [34] England J. G., Zindler A., Reisberg L. C., Salters V., Marcantonio F., Bourdon B., Brueckner H., Turner P. J., Weaver S. & Read P. (1992) The Lamont-Doherty Geological Observatory Isolab 54 isotope ratio mass spectrometer. Int. J. Mass Spectrom. Ion Proc., 121: 201-240. [30] Engrand C., McKeegan K. D. & Leshin L. A. (1999) Oxygen isotopic compositions of individual minerals in Antarctic micrometeorites: Further links to carbonaceous chondrites. Geochim. Cosmochim. Acta, 63(17): 2623-2636. [30] Ensminger L. E. (1954) Some factors affecting the adsorprtion of sulfate by Alabama soils. Soil Sci. Soc. Am. Proc., 18: 259-264. [26] Entwistle D. C. & Reeder S. (1993) New apparatus for pore fluid extraction from mudrocks for geochemical analysis. In: Geochemistry of clay- pore fluid interaction. D. A. C. Manning, P. L. Hall & C. R. Hughes (eds.). The Mineralogical Society Series 4, Chapman & Hall, London: 365-388. [2] Epstein S., Buchsbaum R., Lowenstam H. & Urey H. C. (1951) Carbonate-water isotopic temperature scale. Geol. Soc. Am. Bull., 62: 417-425. [46] Epstein S., Buchsbaum R., Lowenstam H. A. & Urey H. C. (1953) Revised carbonate-water isotopic temperature scale. Geol. Soc. Am. Bull., 64: 1315-1325. [40, 46] Epstein S. & Mayeda T. K. (1953) Variations of 180 of waters from natural sources. Geochim. Cosmochim. Acta, 4: 213-224. [1, 2, 20, 40] Epstein S. & Taylor H. P. Jr. (1970) Stable isotopes, rare gases, solar wind, and spallation products. 180 / 160, 30Si/28Si, D/H, 13C/12C studies of lunar rocks and minerals. Science, 167: 533-535. [25] Epstein S. & Taylor H. P. Jr. (1971) 180/160, 30Si/28Si, D / H and 13C/12C ratios in lunar samples. Proc. Second Lunar Sci. Conf., 2: 1421-1441. [20] Epstein S., Yapp C. J. & Hall J. H. (1976) The determination of the D / H ratio of non-exchangeable hydrogen in cellulose extracted from aquatic and land plants. Earth Planet. Sci. Lett., 30: 241-251. [24] Epstein S., Krishnamurthy R. V., Cronin J. R., Pizzarello S. & Yuen G. U. (1987) Unusual stable isotope ratios in amino acid and carboxylic extract from the Murchison meteorite. Nature, 326: 477-479. [11] Erd61yi M., Richter D. & Tittel F. K. (2002) 13CO2/12CO2 isotopic ratio measurements using a difference-frequency-based sensor operating at 4.35 ~m. Appl. Phys., B 75: 289-295. [34] Erdmann N., Griining C., Trautmann N., Waldek A., Huber G., Kunz P., Nennemann M. & Passler G. (1998) Trace analysis of plutonium by resonance ionization mass spectrometry. J.C. Vickerman, I. Lyon, N.P. Lokyer & J.E. Parks (eds.). Resonance Ionization Spectroscopy, Ninth International Symposium. AIP Press Woodbury, NY, AlP Conference Proceedings 454: 279-282. [33] Erez J. & Luz B. (1983) Experimental paleotemperature equation for plaktonic foraminifera. Geochim. Cosmochim. Acta, 47: 1025-1031. [46] Eriksen T. E. (1972) Sulfur isotope effects I. The isotopic exchange coefficient for the sulfur isotopes 34S32S in the system SO2-g-HSO3-aq at 25, 35, and 45~ Acta Chem. Scandinavia, 26: 573-580. [26] Esat T. M. (1984) A 61 cm radius multi-detector mass spectrometer at the Australian National University. Nucl. Instrum. Meth. Phys. Res., B5: 545-553. [30] Esat T. M. (1995) Charge collection thermal ionization mass spectrometry of thorium. Int. J. Mass Spectrom. Ion Proc., 148: 159-170. [30]
1086
References
Escande M. (1983) Echangeabilit6 et fractionnement isotopique de l'oxig6ne des smectites magnesiennes de synth6ses. Etablissement d'un g6othermom6tre. Th6se, Univ. Paris XI, Paris, France: 150 p. [2] Escande M., Decarreau A. & Labeyrie L. (1984) t~tude exp6rimentale de l'6changeabilit6 des isotopes de l'oxyg6ne des smectites. C. R. Acad. Sc. Paris, 299(series II, No. 11): 707-710. [2] Esler M. B., Griffith D. W. T., Turatti F., Wilson S. R., Rahn T. A. & Zhang H. (2000a) N20 concentration and flux measurements and complete isotopic analysis by FTIR spectroscopy. ChemosphereGlobal Change Sci., 2: 445-454. [19, 34] Esler M. B., Griffith D. W. T., Wilson S. R. & Steele L. P. (2000b) Precision trace gas analysis by FT-IR spectroscopy. 2. The 13C/12Cisotope ratio of CO2. Anal. Chem., 72: 216-221. [34] Eslinger E. V. (1971) Mineralogy and oxygen isotope ratios of hydrothermal and low grade metamorphic argillaceous rocks. PhD Thesis, Case Western Reserve University, Cleveland, Ohio, USA: 205 p. [2] Eslinger E. V. & Savin S. (1973) Mineralogy and oxygen isotope geochemistry of the hydrothermally altered rocks of the Ohaki-Broadlands, New Zealand geothermal area. Am. J. Sci., 273: 240-267. [46] Estep M. L. F. & Hoering T. C. (1980) Biogeochemistry of the stable hydrogen isotopes. Geochim. Cosmochim. Acta, 44: 1197-1206. [46] Estep M. L. F. & Hoering T. C. (1981) Stable hydrogen isotope fractionations during autotrophic and mixotrophic growth of microalgae. Plant Physiol., 67: 474-477. [46] Etheridge D. M., Pearman G. I. & Fraser P. J. (1992) Changes in tropospheric methane between 1841 and 1978 from a high accumulation-rate Antarctic ice core. Tellus, 44(Ser. B; 4): 282-294. [14] Etheridge D. M., Steele L. P., Francey R. J. & Langenfelds R. L. (1998) Atmospheric methane between 1000 AD and present: Evidence of anthropogenic emissions and climatic variability. J. Geophys. Res., 103(D13): 15979-15993. [14] Eugster H. P. (1957) Heterogeneous reactions involving oxidation and reduction at high pressures and temperatures. J. Chem. Phys., 26: 1760-1761. [46] EURACI-IEM (2000) Quantifying Uncertainty in Analytical Measurement. EURACHEM / CITAC Guide, 2nd edition: 120 p. [40, 42] Eugster O., Tera F. & Wasserburg G. J. (1969) Isotopic analyses of barium in meteorites and in terrestrial samples. J. Geophys. Res., 74(15): 3897-3908. [29] Evershed R. P., Arnot K. I., Collister J., Eglinton G. & Charters S. (1994) Application of isotope ratio monitoring gas-chromatography mass-spectrometry to the analysis of organic residues of archaeological origin. Analyst, 119: 909-914. [8] Evershed R. P., Stott A. W., Raven A., Dudd S. N., Charters S. & Leyden A. (1995) Formation of longchain ketones in ancient-pottery vessels by pyrolysis of acyl lipids. Tetrahedron Lett., 36: 8875-8878.
[81 Evershed R. P., Mottram H. R., Dudd S. N., Charters S., Stott A. W., Lawrence G. J., Gibson A. M., Conner A., Blinkhorn P. W. & Reeves V. (1997) New criteria for the identification of animal fats preserved in archaeological pottery. Naturwissensch., 84: 402-406. [8] Ewald A. H. (1985) The effect of pressure on oxygen isotope exchange in silicates. Chem. Geol., 49: 179185. [46] Exley R. A., Mattey D. P., Clague D. A. & Pillinger C. T. (1986) Carbon isotope systematics of a mantle "hotspot": a comparison of Loihi Seamount and MORB glasses. Earth Planet. Sci. Lett., 78: 189-199. [3] Fagan R. & Longstaffe F. J. (1997) Hydrogen-isotope exchange in smectite? 11th International Clay Conference, Ottawa, Canada (Abstract). [2] Fahey A. J. (1998) Measurements of dead time and characterization of ion counting systems for mass spectrometry. Rev. Sci. Instrum., 69: 1282-1288. [4] Fairbank W. M. Jr., Spaar M. T., Parks J. E. & Hutchinson J. M. R. (1989) Measurement of the odd-even effect in the resonance ionization of tin as a function of laser intensity. Phys. Rev., A40: 2195-2198. [33] Fairbank W. M. Jr. (1996) A seat-of-the-pants approach to RIS theory. N. Winograd & J.E. Parks (eds.). Resonance Ionization Symposium 1996, Eighth International Symposium. AIP Press Woodbury, NY, AIP Conference Proceedings 388: 457-464. [33] Fairbanks R. G. (1982) The origin of continental shelf and slope water in the New York Bight and Gulf of Maine: evidence from H2180/H2160 ratio measurements. J. Geophys. Res., 87: 5796-5808. [1]
References
1087
Fairbanks R. G., Sverdlove M., Free R., Wiebe P. H. & B6 A.W.H. (1982) Vertical distribution and isotopic fractionation of living planktonic foraminifera from the Panama Basin. Nature, 298: 841-844. [46] Fang D. & Marcus R. K. (1993) Fundamental plasma processes. In: Glow discharge spectroscopies. R. K. Marcus (ed.). Plenum Press, NY: 17-66. [35] Fang J., Barcelona M. J., Krishnamurthy R. V. & Atekwana E. A. (2000) Stable carbon biogeochemistry of a shallow sand aquifer contaminated with fuel hydrocarbons. Appl. Geochem., 15: 157-169. [10] Fantle M. S., Dittel A. I., Schwalm S. M., Epifano C. E. & Fogel M. L. (1999) A food web analysis of the juvenile blue crab, Callinectus, sapidus, using stable isotopes in whole animals and individual amino acids. Oecologia, 120: 416-426. [9] Farmer V. C. & Mitchell B. D. (1963) Occurence of oxalates in soil clays following hydrogen peroxide treatment. Soil Science, 96(4): 221-229. [2] Farquhar G. D., Ehleringer J. R. & Hubick K. T. (1989a) Carbon isotope discrimination and photosynthesis. Annu. Rev. Plant Physiol. Plant Mol. Biol., 40: 503-537. [9, 46] Farquhar G. D., Hubick K. T., Condon A. G. & Richards R. A. (1989b) Carbon isotope fractionation and plant water-use efficiency. In: Stable isotopes in ecological research. P. W. Rundel, J. R. Ehleringer & K. A. Nagy (eds.). Springer Verlag, NY, Vol. 68: 21-40. [21] Farquhar G. D. & Lloyd J. (1993) Carbon and oxygen isotope effects in the exchange of carbon dioxide between terrestrial plants and the atmosphere. In: Stable isotopes and plant carbon-water relations. J. R. Ehleringer, A. E. Hall & G. D. Farquhar (eds.). Academic Press: 47-70. [21]
Farquhar G. D., Lloyd J., Taylor J. A., Flanagan L. B., Syvertsen J. P., Hubrick K. T., Wong S. C. & Ehleringer J. R. (1993) Vegetation effects on the isotope composition of oxygen in atmospheric CO2. Nature, 363: 439-443. [14, 21]
Farquhar G. D., Henry B. K. & Styles J. M. (1997) A rapid on-line technique for determination of oxygen isotope composition of nitrogen-containing organic matter and water. Rapid Commun. Mass Spectrom., 11: 1554-1560. [1, 8, 15, 21, 23] Farquhar G. D., Barbour M. M. & Henry B. K. (1998) Interpretation of oxygen isotope composition of leaf material. In: Stable isotopes. H. Griffiths (ed.). BIOS Scinetific, Oxford: 27-62. [21] Farquhar J. & Rumble D. III (1998) Comparison of oxygen isotope data obtained by laser fluorination of olivine with KrF excimer laser and CO2 laser. Geochim. Cosmochim. Acta, 62(18): 3141-3149. [20] Farquhar J. & Thiemens M. H. (2000) A 33S enrichment in ureilite meteorites; evidence for a nebular sulfur component. Geochim. Cosmochim. Acta, 64: 1819-1825. [20] Farquhar J., Bao H. & Thiemens M. H. (2000a) Atmospheric influence of Earth's earliest sulfur cycle. Science, 289: 756-758. [20, 26] Farquhar J., Savarino J., Jackson T. & Thiemens M. (2000b) Evidence of atmospheric sulphur in the martian regolith from sulphur isotopes in meteorites. Nature, 404: 50-52. [18] Farquhar J., Savarino J., Airieau S. & Thiemens M. H. (2001) Observation of wavelength-sensitive mass-independent sulfur isotope effects during SO2 photolysis: Implications for the early atmosphere. J. Geophys. Res., 106: 32829-32839. [26] Farver J. R. & Yund R. A. (1995) Interphase boundary diffusion of oxygen and potassium in K-feldspar / quartz aggregates. Geochim. Cosmochim. Acta, 59: 3697-3705. [46] Farver J. R. & Yund R. A. (1999) Oxygen bulk diffusion measurements and TEM characterization of a natural ultramylonite: implications for fluid transport in mica-bearing rocks. J. Metamorphic Geol., 17: 669-683. [46] Fassett J. D. & Paulsen P. J. (1989) Isotope dilution mass spectrometry for accurate elemental analysis. Anal. Chem., 61(10): 643A-649A. [37] Fauhl C. & Wittkowski R. (1996) On-line 1H-NMR to facilitate tube preparation in SNIF-NMR analysis. Z. Lebensm. Unters. Forsch., 203: 541-545. [5] Faure G. (1986) Principles of isotope geology. Wiley & Sons, 2nd edition: 589 p. [29] Fayek M., Harrison T. M., Grove M., McKeegan K. D., Coath C. D. & Boles J. R. (2001) In situ isotopic evidence for protracted and complex carbonate cementation in a petroleum reservoir, North Coles Leveem, San Joaquin Basin, California, U.S.A.J. Sedim. Res., 71(3): 444-458. [30]
1088
References
Feldmann I., Jakubowski N. & Stuewer D. (1999a) Application of a hexapole collision and reaction cell in ICP-MS; Part I: Instrumental aspects and operational optimization. Fresenius J. Anal. Chem., 365: 415-421. [311 Feldmann I., Jakubowski N., Thomas C. & Stuewer D. (1999b) Application of a hexapole collision and reaction cell in ICP-MS; Part II: Analytical figures of merit and first applications. Fresenius J. Anal. Chem., 365: 422-428. [31] Felgett P. B. & Usher M. J. (1980) Fluctuation phenomena in instrument science. J. Phys. E., Sci. Instrum., 13: 1041-1046. [39] Feng F. & Horlick G. (1994) Analysis of aluminum alloys using inductively coupled plasma and glow discharge mass spectrometry. J. Anal. At. Spectrom., 9: 823-831. [35] Feng P., Bu X. & Stucky G. D. (1997) Hydrothermal syntheses and structural characterization of zeolite analogue compounds based on cobalt phosphate. Nature, 388: 735-740. [36] Feng X. & Savin S. M. (1991) Oxygen isotope studies of zeolites: Stilbite, analcime, heulandite, and clinoptilolite- I. Analytical technique. In: Stable Isotope Geochemistry: A tribute to Samuel Epstein. H.P. Jr. Taylor, J. R. O'Neil & I. R. Kaplan (eds.). The Geochemical Society, Spec. Publ. 3: 271- 283. [46] Feng X., Krishnamurthy R. V. & Epstein S. (1993) Determination of D / H ratios of nonexchangeable hydrogen in cellulose: A method based on the cellulose-water exchange reaction. Geochim. Cosmochim. Acta, 57: 4249-4256. [24] Feng X. & Savin S. M. (1993a) Oxygen isotope studies of zeolites- stilbite, analcime, heulandite, and clinoptilolite: II Kinetics and mechanisms of isotopic exchange between zeolites and water vapour. Geochim. Cosmochim. Acta, 57: 4219-4238. [46] Feng X. & Savin S. M. (1993b) Oxygen isotope studies of zeolites: stilbite, analcime, heulandite, and clinoptilolite; III, Oxygen isotope fractionation between stilbite and water or water vapor. Geochim. Cosmochim. Acta, 57(17): 4239-4247. [46] Feng X. & Epstein S. (1995a) Carbon isotopes of trees from arid environments and implications for reconstructing atmospheric CO2 concentrations. Geochim. Cosmochim. Acta, 59: 2599-2608. [46] Feng X. & Epstein S. (1995b) Climatic temperature records in 6D data from tree rings. Geochim. Cosmochim. Acta, 59: 3029-3037. [46] Ferhi A. & L4tolle R. (1979) Relation entre le milieu climatique et les teneurs en oxyg6ne-18 de la cellulose des plantes terrestres. Physiol. Vdg., 17: 107-117. [46] Ferhi A., Bariac T., Jusserand C. & L6tolle R. (1983) An integrated method for isotopic analysis of oxygen from organic compounds, air water vapor and leaf water. Int. J. Appl. Radiat. Isot., 34: 14511457.[1] Ferretti R. J., Lowe D. C., Martin R. J. & Brailsford G. B. (2000) A new gas chromatograph-isotope ratio mass spectrometry technique for high precision, N20-free analysis of 613C and 6180 in atmospheric CO2 from small air samples. J. Geophys. Res., 105(D5): 6709-6718. [14] Fessenden J. E., Cook C. S., Loft M. J. & Ehleringer J. R. (2002) Rapid 180 analysis of small water and CO2 samples using a continuous-flow isotope ratio mass spectrometer. Rapid Commun Mass Spectr., 16: 1257-1260. [1] Fetter C. W. (1994) Applied Hydrogeology. Prentice Hall. [29] Feuerstein T. P., Ostrom P. H. & Ostrom N. E. (1997) Isotopic biogeochemistry of dissolved organic nitrogen: a new technique and application. Org. Geochem., 27: 363-370. [15] Fiedler R. & Proksch G. (1972) The determination of the nitrogen content and nitrogen-15 abundance by means of the nitrogen gas generated from inorganic and organic materials. Anal. Chim. Acta, 60: 277-285. [15] Field E. M., Switsur V. R. & Waterhouse J. S. (1994) An improved technique for the determination of oxygen stable isotope ratios in wood cellulose. Int. J. Appl. Radiat. Isot., 45: 177-181. [23] Filer C. N. (1999) Isotopic fractionation of organic compounds in chromatography. J. Labelled Compd. Radiopharm., 42: 169-197. [8] Finlay J. C. (2003) Controls of streamwater dissolved inorganic carbon dynamics in a forested watershed. Biogeochemistry, 62: 231-252. [10] Firsching F. H. (1961) Precipitation of silver phosphate from homogeneous solution. Anal. Chem., 33: 873-878. [20, 22]
References
1089
Fischer T. P., Giggenbach W. F., Sano Y. & Williams S. N. (1998) Fluxes and sources of volatiles dis-
charged from Kudryavy, a subduction zone volcano, Kurile Islands. Earth Planet. Sci. Lett., 160: 8196. [16] Fitzsimons I. C. W., Harte B. & Clark R. M. (2000) SIMS stable isotope measurement counting statistics and analytical precision. Mineral. Magazine, 64: 59-83. [30] Flesch G. D., Anderson A. R. Jr. & Svec H. J. (1973) A secondary isotopic standard for 6Li/7Li determination. Intl. J. Mass Spectrom. Ion Phys., 12: 265-272. [6, 40] Flintrop C., Hohlmann B., Jasper T., Korte C., Podlaha O. G., Scheele S. & Veizer J. (1996) Anatomy of pollution: rivers of North-Westphalia, Germany. Am. J. Sci., 296: 58-98. [10] Florkowsky T. (1985) Sample preparation for hydrogen isotope analysis by mass spectrometry. Int. J. Appl. Radiat. Isot., 36(12): 991-992. [1] Focken U. & Becker K. (1998) Metabolic fractionation of stable carbon isotopes: implications of different proximate compositions for studies of the aquatic food webs using 613C data. Oecologia, 115: 337-343.
[9]
Fogel M. L., Aguilar C., Cuhel R., Hollander D. J., Willey J. D. & Paerl H. W. (1999) Biological and isotopic changes in coastal waters induced by Hurricane Gordon. Limnol. Oceanogr., 44: 1359-1369.
[9]
Fogel M. L. & Tuross N. (1999) Transformations of plant biochemicals to geological macromolecules during early diagenesis. Oecologia, 120: 336-346. [9] Forrest J. & Newman L. (1973) Sampling and analysis of atmospheric sulfur compounds for isotope ratio studies. Atmos. Environm., 7: 561-573. [26] Fortier S. M. & Giletti B. J. (1991) Volume self-diffusion of oxygen in biotite, muscovite, and phlogopite micas. Geochim. Cosmochim. Acta, 55: 1319-1330. [46] Fortier S. M. (1994) An on-line experimental/analytical method for measuring the kinetics of oxygen isotope exchange between CO2 and saline/hypersaline salt solutions at low (25-50~ temperatures. Chem. Geol. (Isot. Geosci. Sect.), 116: 155-162. [1] Fortier S. M., Liittge A., Satir M. & Metz P. (1994) Oxygen isotope fractionation between fluorphlogopite and calcite: an experimental investigation of temperature dependence and F-/OH- effects. Eur. J. Mineral., 6: 53-65. [2, 46] Fortier S. M., Cole D. R., Wesolowski D. J., Riciputi L. R., Paterson B. A., Valley J. W. & Horita J. (1995) Determination of the magmatic-water equilibrium oxygen isotope fractionation factor at 350~ A comparison of ion microprobe and laser fluorination techniques. Geochim. Cosmochim. Acta, 59(18): 3871-3875. [46] Fortier S. M. & Liittge A. (1995) An experimental calibration of the temperature dependence of oxygen isotope fractionation between apatite and calcite at high temperatures (350-800~ Chem. Geol. (Isot Geosci. Sect.), 125: 281-290. [20, 22] Fossing H. & Jorgensen B. B. (1989) Measurement of bacterial sulfate reduction in sediments: Evaluation of a single-step chromium reduction method. Biogeochem., 8: 205-222. [26] Fossing H. (1995) 35S radiolabeling to probe biogeochemical cycling of sulfur. In: Geochemical transformations of sedimentary sulphur. M. A. Vairavamurthy, M. A. A. Schoonen, T. I. Eglinton, G. W. Luther, and B. Manowitz (eds.). American Chemical Society, Washington. ACS Symposium Series 612: 349-364. [26] Fouillac C., Fouillac A. M. & Criaud A. (1990) Sulphur and oxygen isotopes of dissolved sulphur species in formation water from the Dogger geothermal aquifer, Paris Basin, France. Appl. Geochem., 5: 415-429. [26] Fouillac A.-M. & Girard J.-P. (1996) Laser oxygen isotope analysis of silicate/oxide grain separates: Evidence for a grain effect? Chem. Geol., 130: 31-54. [20] France-Lanord C. & Sheppard S. M. F. (1992) Hydrogen isotope composition of pore waters and interlayer water in sediments from the Central Western Pacific, LEG 129. Proceed. ODP, Sci. Results, 129: 295-302. [2, 46] France-Lanord C. & Derry L. A. (1994) 613C of organic carbon in the Bengal Fan: Source evolution and transport of C3 and C4 plant carbon to marine sediments. Geochim. Cosmochim. Acta, 58: 4809-4814.
[9]
1090
References
Francey R. J. (1986) Carbon isotope measurements in baseline air, forest canopy air, and plants. In: The changing carbon cycle. J. R. Trabalka & D. E. Reichle (eds.). Springer Verlag: 160-174. [24] Francey R. J. & Tans P. P. (1987) The latitudinal variation in oxygen-18 of atmospheric CO2. Nature, 327: 495-497. [141 Francey R. J., Allison C. E., Steele L. P., Langenfelds R. L., Welch E. D., White J. W. C., Trollier M., Tans P. P. & Massarie K. A. (1994) Intercomparison of stable isotope measurements of CO2. In: Climate monitoring and diagnostics laboratory (CMDL). J. T. Peterson & R. M. Rosson (eds.). US Department of Commerce, Boulder, Colorado, USA. Summary report 1993, No. 22: 106-110. [14] Francey R. J., Tans P., Allison C. E., Enting I. G., White J. W. C & Trolier M. (1995) Changes in oceanic and terrestrial carbon uptake since 1982. Nature, 373: 326-330. [14] Franchi I. A., Wright I. P., Gibson E. K. Jr. & Pillinger C. T. (1986) The laser microprobe: A technique for extracting carbon, nitrogen and oxygen from solid samples for isotopic measurements. J. Geophys. Res., Supplem. 91 (B4; Proceedings of the sixteenth lunar and planetary science conference, Part. 2): D514-D524. [17, 20] Franchi I. A., Akagi T. & Pillinger C. T. (1993a) Laser fluorination of silicates for 6170 and 6180 using 02 gas. TERRA Nova, 5 (Abstr. Suppl. No.l): 370. [20] Franchi I. A., Wright I. P. & Pillinger C. T. (1993b) Constraints on the formation conditions of iron meteo-rites based on concentration and isotopic compositions of nitrogen. Geochim. Cosmochim. Acta, 57: 3105-3121. [13] Franconi J. M., Naulet N. & Martin G. J. (1989) Natural factors of site-specific isotope fractionation of hydrogen in the fermentation products (ethanol and water) of barley. J. Cereal Sci., 10: 69-80. [5] Frazer J. w. & Crawford R. (1963) Modifications in the simultaneous determination of carbon, hydrogen, and nitrogen. Mikrochim. Acta, 1963(3): 561-566. [15] Freeman K. H., Hayes J. M., Trendel J. M. & Albrecht P. (1990) Evidence from carbon isotope measurements for diverse origins of sedimentary hydrocarbons. Nature, 343: 254-256. [9] Freeman K. H., Wakeham S. G. & Hayes J. M. (1994) Predictive isotopic biogeochemistry: Hydrocarbons from anoxic marine basins. Org. Geochem., 21: 629-644. [9] Freer R. & Dennis P. F. (1982) Oxygen diffusion studies. I. A preliminary ion microprobe investigation of oxygen diffusion in some rock-forming minerals. Mineral. Magazine, 45: 179-192. [46] Freer R., Wright K., Knoll H. & Gi~ttlicher J. (1997) Oxygen diffusion in sanidine feldspars and a critical appraisal of oxygen isotope-mass-effect measurements in non-cubic materials. Phil. Magazine, A 75: 485-503. [46] Freney J. R. (1961) Some observations on the nature of organic sulphur compounds in soil. Aust. J. Agric. Res., 12: 424-432. [26] Freney J. R., Melville G. E. & Williams C. H. (1970) The determination of carbon bonded sulfur in soil. Soil. Sci., 109: 310-318. [26] Freney J. R. (1986) Forms and reactions of organic sulfur compounds in soils. In: Sulfur in Agriculture. M. A. Tabatabai (ed.). American Society of Agronomy - Crop Science Society of America - Soil Science Society of America, Madison, Wisconsin, USA. Monograph No. 27: 207-232. [26] Frezzotti M. L., Dallai L. & Sharp Z. D. (2000) Fluid-inclusion and stable-isotope evidence for fluid infiltration and veining during metamorphism in marbles and metapelites. Eur. J. Mineral., 12(1): 231-246. [3] Frick U. & Pepin R. O. (1981) Microanalysis of nitrogen isotope abundances: Association of nitrogen with noble gas carriers in Allende. Earth Planet. Sci. Lett., 56: 64-81. [13, 17] Fricke H. C., O'Neil J. R. & Lynnerup N. (1995) Oxygen isotope composition of human tooth enamel from medieval Greenland: Linking climate and society. Geology, 23: 869-872. [22] Fricke H. C. & O'Neil J. R. (1996) Inter- and intra-tooth variation in the oxygen isotope composition of mammalian tooth enamel phosphate: implications for palaeoclimatological and palaeobiological research. Palaeogeogr., Palaeoclim., Palaeoecol., 126: 91-99. [22] Fricke H. C., Clyde W. C., O'Neil J. R. & Gingerich P. D. (1998a) Evidence for rapid climate change in North America during the latest Paleocene thermal maximum: oxygen isotope compositions of biogenic phosphate from the Bighorn Basin (Wyoming). Earth Planet. Sci. Lett., 160: 193-208. [22]
References
1091
Fricke H. C., Clyde W. C., O'Neil J. R. & Gingerich P. D. (1998b) Intra-tooth variations in ~)180 (PO4) of mammalian tooth enamel as a record of seasonal variations in continental climate variables. Geochim. Cosmochim. Acta, 62: 1839-1850. [22] Friedheim C. & Meyer R. J. (1892) Ober die quantitative Trennung und Bestimmung von Chlor, Brom und Jod. Z. anorg. Chemie, 1: 407-422. [28] Friedman I. (1953) Deuterium content of natural waters and other substances. Geochim. Cosmochim. Acta, 4:89-103.[1] Friedman I. & Smith R. L. (1958) The deuterium content of water in some volcanic glasses. Geochim. Cosmochim. Acta, 15: 218-228. [3, 46] Friedman I. (1970) Some investigations of the deposition of travertine from hot springs, I. Geochim. Cosmochim. Acta, 34: 1303-1315. [10] Friedman I. & Hardcastle K. (1970) A new technique for pumping hydrogen gas. Geochim. Cosmochim. Acta, 34: 125-126. [1] Friedman I. & O'Neil J. R. (1977) Compilation of stable isotope fractionation factors of geochemical interest. U. S. Geol. Survey Prof. Paper, 440-KK. [40, App B] Friedman I., O'Neil J. & Cebula G. (1982) Two new carbonate stable isotope standards. Geostand. Newslett. 6, 11-12. [40] Friedman L. & Bigeleisen J. (1950) Oxygen and nitrogen isotope effects in the decomposition of ammonium nitrate. J. Chem Phys., 18: 1325-1331. [19] Fritz P. & Fontes J. C. (1966) Fractionnement isotopique pendant l'attaque acide des carbonates naturels: r61e de la granulom6trie. C. R. Acad. Sc. Paris, t. 263: 1345-1348. [12] Fritz P., Drimmie R. J. & Nowicki V. K. (1974) Preparation of sulfur dioxide for mass spectrometer analysis by combustion of sulfides with copper oxide. Anal. Chem., 46: 164-166. [26] Fritz P., Frape S. K., Drimmie R. J. & Heemskerk A. R. (1986) Reply to comments by Grobczak et al. on "Water-rock interaction and chemistry of groundwaters from the Canadian Shield". Geochim. Cosmochim. Acta, 50: 1561-1563. [1] Fronza G., Fuganti C., Grasselli P., Barbeni M. & Cisero M. (1993) Natural abundance 2H nuclear magnetic resonance study of the origin of (R)-h-decanolide. J. Agric. Food Chem., 41: 235-239. [5] Fronza G., Fuganti C., Grasselli P., Servi P., Zucchi G., Barbeni M. & Cisero M. (1995) Natural abundance 2H nuclear magnetic resonance study of the origin of 2-phenylethanol and 2-phenylethyl acetate. J. Agric. Food Chem., 43: 439-443. [5] Fronza G., Fuganti C., Guillou C., Reniero F. & Joulain D. (1998) Natural abundance 2H nuclear magnetic resonance study of the origin of raspberry ketone. J. Agric. Food Chem., 46(1): 248-254. [5] Fry B. & Sherr E. B. (1984) 813C measurements as indicators of carbon flow in marine and freshwater ecosystems. Contrib. Mar. Sci., 27: 13-47. [9] Fry B. (1986) Stable sulfur isotopic distributions and sulfate reduction in lake sediments of the Adirondack Mountains, New York. Biogeochemistry, 2: 329-343. [26] Fry B., Gest H. & Hayes J. M. (1986) Sulfur isotope effects associated with protonation of HS- and volatilization of H2S. Chem. Geol. (Isotope Geosci. Sec.), 58: 253-258. [46] Fry B., Ruf W., Gest H. & Hayes J. M. (1988) Sulfur isotope effects associated with oxidation of sulphide by 02 in aqueous solution. Chem. Geol., 73: 205-210. [46] Fry B. (1991) Stable isotope diagrams of freshwater food webs. Ecology, 72: 2293-2297. [9] Fry B., Brand W., Mersch F. J., Tholke K. & Garritt R. (1992) Automated analysis system for coupled 813C and 815N measurement. Anal. Chem., 64: 288-291. [9, 45] Fry B., Garritt R., Tholke K., Neill C., Michener R. H., Mersch F. J. & Brand W. (1996) Cryoflow: cryofocussing nanomole amounts of CO2, N2 and SO2 from an elemental analyzer for stable isotopic analysis. Rapid Commun. Mass Spectrom., 10: 953-958. [15, 45] Fry B., Silva S. R., Kendall C. & Anderson R. K. (2002) Oxygen isotope corrections for online 834S analysis. Rapid Commun. Mass Spectrom., 16: 854-859. [26, 45] Fung I., John J., Lerner E., Matthews E., Prather M., Steele L. P. & Fraser P. J. (1991) Dimensional model synthesis of the global methane cycle. J. Geophys. Res., 96(D7): 13033-13065. [14] Fusch C. (1985) Quantitative Analyse von D20-Konzentrationen im Serum mit der Infrarot-Spektroskopie. Inaugural-Dissertation (Inauguration Thesis), UniversiNit T/,ibingen. [34]
1092
References
Fusch C., Spririg N. & Moeller H. (1993) Fourier transform infrared spectroscopy measures 1H/2H
ratios of native water with a precision comparable to that of isotope ratio mass spectrometry. Eur. J. Clin. Chem. Clin. Biochem., 31: 639-644. [34] Gaillardet J. & All6gre C. J. (1995) Boron isotopic compositions of corals: seawater or diagenesis record? Earth Planet. Sci. Lett., 136: 665-676. [7] Gale N. H. (1970) A solution in closed form for lead isotopic analysis using a double spike. Chem. Geol., 6: 305-310. [29] Gale N. H., Woodhead A. P., Stos-Gale Z. A., Walder A. & Bowen I. (1999) Natural variations detected in the isotopic composition of copper: possible applications to archaeology and geochemistry. Int. J. Mass Spectrom., 184: 1-9. [31] Galy A., Young E. D., Ash R. D. & O'Nions R. K. (2000) The formation of chondrules at high gas pressure in the solar nebula. Nature, 290: 1751-1753. [31] Galy A., Belshaw N. S., I-Ialicz L. & O'Nions R. K. (2001) High-precision measurement of magnesium isotopes by multiple-collector inductively coupled plasma mass spectrometry. Int. J. Mass Spectrom., 208: 89-98. [31] Games L. M. & Hayes J. M. (1976) Isotopic and quantitative analysis of the major carbon fractions in natural water samples. Anal. Chem., 48: 130-135. [10] Games L. M. & Hayes J. M. (1977) Carbon isotopic study of the landfill leachate in groundwater. J. Water Poll. Control Fed., 49: 668-677. [10] Games L. M., Hayes J. M. & Gunsalus R. P. (1978) Methane-producing bacteria: natural fractionations of the stable carbon isotopes. Geochim. Cosmochim. Acta, 42: 1295-1297. [46] Ganther H. E. (1974) The biochemistry of selenium. In: Selenium. R. A. Zingaro& W. C. Cooper (eds.). Van Nostrand Reinhold Co: 546-614. [29] Gao X. & Thiemens M. H. (1991) Systematic study of sulfur isotopic composition in iron meteorites and the occurence of excess 33S a n d 36S. Geochim. Cosmichim. Acta, 55: 2671-2679. [20, 45] Gao X. & Thiemens M. H. (1993a) Isotopic composition and concentration of sulfur in carbonaceous chondrites. Geochim. Cosmochim. Acta, 57: 3159-3169. [45] Gao X. & Thiemens M. H. (1993b) Variations of the isotopic composition of sulfur in enstatite and ordinary chondrites. Geochim. Cosmochim. Acta, 57: 3171-3176. [45] Gao Y. Q. & Marcus R. A. (2001) Strange and unconventional isotope effects in ozone formation. Science, 293: 259-263. [18, 46] Gao Y. Q. & Marcus R. A. (2002) On the theory of the strange and unconventional isotope effects in ozone formation. J. Chem. Phys., 116(1): 137-154. [18] Garcia J.-P., Sheppard S. M. F., Picard S., L4cuyer C. & Ernig C. (1997) Carbon and oxygen isotopic signals in living brachiopods: prerequisites for the record of oceanic secular variations. Terra Nova, 9 (Abstracts suppl. No. 1): 271 (Abstract). [46] Garlick G. D. & Epstein S. (1967) Oxygen isotope ratios in coexisting minerals of regionally metamorphosed rocks. Geochim. Cosmochim. Acta, 31: 181-214. [20] Garner E. L. & Raspberry S. D. (1993) What's new in traceability. J. Test. Eval., 21: 505-509. [43] Garrels R. M. & Christ C. L. (1965) Solutions, minerals and equilibria. Harper and Row, NY. [10] Garten C. T. Jr. (1992) Nitrogen isotope composition of ammonium and nitrate in bulk precipitation and forest throughfall. Intern. J. Environ. Anal. Chem., 47: 33-45. [15] Gaudette H. E. (1990) Chlorine and boron isotopic analysis of Antarctic ice and snow: Indicators of marine and volcanic atmospheric inputs. GSA Ann. Mtg. Abstr. Prog.: 173. [28] Gaunt J. (1956) The analysis of heavy water by infra-red spectrometry. Spectrochim. Acta, 8: 57-65. [34] Gazi D., Spiro B. & Greenwood P. B. (1993) Experiments on laser fluorination of silicate and oxide minerals. Terra Nova, 5 (Abstr. Suppl. No.l): 371. [20] Gearing J. N. (1991) The study of diet and trophic relationships through natural abundance 13C. In: Carbon isotope techniques. D. C. Coleman & B. Fry (eds.). Academic Press Inc, San Diego: 201-218. [9] Gehre M., I-Ioefling R. & Kowski P. (1996a) Methodical studies for D/H-isotope analysis - A new technique for the direct coupling of sample preparation to an IRMS. Isot. Environ. Health Stud., 32: 335-340. [1] Gehre M., Hoefling R., Kowski P. & Strauch G. (1996b) Sample preparation device for quantitative hydrogen isotope analysis using chromium metal. Anal. Chem., 68(24): 4414-4417. [1]
References
1093
Gellene G. I. (1992) Symmetry restrictions in diatom/diatom reactions. I. Group theoretical analysis. J. Chem. Phys., 96: 4387-4402. [18] Gellene G. I. (1993) Symmetry-dependent isotope fractionation in the formation of He2 +. J. Chem. Phys., 97: 34-39. [18] Gellene G. I. (1996) An explanation for symmetry-induced isotopic fractionation in ozone. Science, 274: 1344-1346. [18] Gemery P. A., Trolier M. & White J. W. C. (1996) Oxygen isotope exchange between carbon dioxide and water following atmospheric sampling using glass flasks. J. Geophys. Res., 101: 14415-14420. [14] Genoni L., Iacumin P., Nikolaev V., Gribchenko Y. & Longinelli A. (1998) Oxygen isotope measurements of mammoth and reindeer skeletal remains: an archive of late Pleistocene environmental conditions in Eurasian Arctic. Earth Planet. Sci. Lett., 160: 587-592. [22] Gensho R. & Honda M. (1971) Measurements of the isotopic ratio of boron. Shitsuryo Bunseki (Mass Spectrom.), 19:131-143 (in Japanese, with English abstract). [7] Genty D., Plagnes V., Causse C., Cattani O. & Stievenard M. (2002) Fossil water in large stalagmite voids as a tool for paleoprecipitation stable isotope composition reconstitution and paleotemperature calculation. Chem. Geol., 184: 83-95. [3] Gerstenberger H. & Herrmann M. (1983) Report on the intercomparison for the isotope standards Limestone KH2 and Polyethylene Foil PEF 1. Zfl-Mitteilungen, Zentralinstitut f/Jr Isotopen- und Strahlenforschung. Leipzig, Germany. Vo160: 67-83. [40] Gessler R. & v. Gehlen K. (1986) Investigation of sulfur isotope fractionation between H2S gas and aqueous solutions. Fresenius Z. Anal. Chem., 324: 130-136. [46] Gianfrani L., Gagliardi G., van Burgel M. & Kerstel E. (2003) Isotope analysis of water by means of near-infrared dual-wavelength diode laser spectroscopy. Opt. Express, 11:1566-1576 (http:// www.opticsexpress.org / abstract.cfm?URI=OPEX-11-13-1566). [34] Giesemann A., J/iger H.-J., Norman A. L., Krouse H. R. & Brand W. A. (1994) On-line sulfur-isotope determination using an elemental analyzer coupled to a mass spectrometer. Anal. Chem., 66: 28162819. [26, 27, 45] Gifford S., Bentley H. W. & Graham D. L. (1985) Chlorine isotopes as environmental tracers in Columbia River basalt groundwaters. Proc. 17th Int. Congr. I.A.H.. Hydrogeology of rocks of low permeability. Vol. 17: 417-429. [28] Giggenbach W. F. (1982) Carbon-13 exchange between CO2 and CH4 under geothermal conditions. Geochim. Cosmochim. Acta, 46:159-165 [46]. Giggenbach W. F. (1997) Relative importance of thermodynamic and kinetic processes in governing the chemical and isotopic composition of carbon gases in high-heat flow sedimentary basins. Geochim. Cosmochim. Acta, 61: 3763-3785. [46] Giletti B. J., Semet M. P. & Yund R. A. (1978) Studies in diffusion. III. Oxygen in feldspar: an ion microprobe determination. Geochim. Cosmochim. Acta, 42: 45-57. [46] Giletti B. J. & Yund R. A. (1984) Oxygen diffusion in quartz. J. Geophys. Res., 89: 4039-4046. [46] Giletti B. J. (1985) The nature of oxygen transport within minerals in the presence of hydrothermal water and the role of diffusion. Chem. Geol., 53: 197-206. [46] Giletti B. J. (1986) Diffusion effects on oxygen isotope temperatures of slowly cooled igneous and metamorphic rocks. Earth Planet. Sci. Lett., 77: 218-228. [46] Gilg H. A. & Sheppard S. M. F. (1995) Hydrogen isotope fractionation between smectite and water. Terra Nova, Abstracts suppl. No. 1:329 (Abstract.) [46] Gilg H. A. & Sheppard S. M. F. (1996) Hydrogen isotope fractionation between kaolinite and water revisited. Geochim. Cosmochim. Acta, 60: 529-533. [46] Gilg H. A. & Sheppard S. M. F. (1999) Stability of H and O isotope ratios in natural hydrated silicic glasses: estimation of diffusion coefficient for water at ambient temperatures. EUG 10, Conference Abstracts, 4:531 (Abstract). [46] Gillet P., McMillan P., Schott J., Badro J. & Grzechnik A. (1996) Thermodynamic properties and isotopic fractionation of calcite from vibrational spectroscopy of 180-substituted calcite. Geochim. Cosmochim. Acta, 60: 3471-3485. [46]
1094
References
Gillson G. R., Douglas D. J., Fulford J. E., Halligan K. W. & Tanner S. D. (1988) Nonspectroscopic interelement interferences in inductively coupled mass spectrometry. Anal. Chem., 60: 1472-1474. [6,
35]
Gilmour I. & Pillinger C. T. (1994) Isotopic compositions of individual polycyclic aromatic hydrocarbons from the Murchison meteorite. Mon. Not. R. Astron. Soc., 269: 235-240. [11] Gilmour I., Johnston M. A., Pillinger C. T., Pond C. M., Mattacks C. A. & Prestrud P. (1995a) The carbon isotopic composition of individual fatty acids as indicators of dietary history in arctic foxes on Svalbard. Phil. Trans. R. Soc. Lond., 349(B): 135-142. [9] Gilmour J. D., Hewett S. M., Lyon I. C., Stringer M. & Turner G. (1991) A resonance ionization mass spectrometer for xenon. Meas. Sci. Technol., 2: 1-7. [33] Gilmour J. D., Ash R. D., Lyon I. C. & Turner G. (1994a) I-Xe dating using the RELAX mass spectrometer. Resonance Ionization Spectroscopy 1994 (RIS 94). AIP Press Woodbury, NY, AIP Conference Proceedings 329: 233-236. [33] Gilmour J. D., Johnston W. A., Lyon I. C. & Turner G. (1994b) RELAX-l: An ultrasensitive resonance ionization mass spectrometer for xenon. Rev. Sci. Instrum., 65(3): 617-625. [33] Gilmour J. D., Ash R. D., Lyon I. C. & Turner G. (1995b) Iodine-xenon dating of Bjub61e and Parnallee using RELAX. Meteoritics, 30: 405-411. [33] Gim6nez-Miralles J. E., Salazar D. M. &Solana I. (1999) Regional origin assignment of red wines from Valencia (Spain) by 2H-NMR and 13C IRMS stable isotope analysis of fermentative ethanol. J. Agric. Food Chem., 47(7): 2645-2652. [5] Giral S. (1994) Variations des rapports isotopiques 180/160 des kaolinites de deux profils lat6ritiques amazoniens: signification pour la p6dologie et la pal6oclimatologie. Thbse Universit6, Universit6 Aix-Marseille III, Aix en Province-Marseille, France: 198 p. [2] Giral-Kacmarcik S., Savin S. M., Nahon D., Girard J.-P., Lucas Y. & Abel L. (1998) Oxygen isotope geochemistry of kaolinite in laterite-forming processes, Manaus, Amazonas, Brazil. Geochim. Cosmochim. Acta, 62: 1865-1879. [2] Girard G. (1994) The third periodic verification of national prototypes of the kilogram (1988-1992). Metrologia, 31: 317-336. [42] Girard J.-P., Deynaux M. & Nahon D. (1989) Diagenesis of the upper Proterozoic siliciclastic sediments of the Taoudeni Basin (West Africa) and relation to diabase emplacement. J. Sediment. Petrol., 59: 233248. [2] Girard J.-P. & Fouillac A.-M. (1995) G6ochimie isotopique de l'oxyg6ne et de l'hydrog6ne des argiles: examples d'application aux domaines diag6n6tique et g6othermique. Bull. Centres Rech. Explor.-Prod. ElfAquit., 19(1): 167-195. [2] Girard J.-P. & Savin S. M. (1996) Intracrystalline fractionation of oxygen isotopes between hydroxyl and non-hydroxyl sites in kaolinite measured by thermal dehydroxylation and partial fluorination. Geochim. Cosmochim. Acta, 60: 469-487. [2, 20] Girard J.-P., Freyssinet P. & Chazot G. (2000) Unraveling climatic changes from intra-profile variation in oxygen and hydrogen isotopic composition of goethite and kaolinite: An integrated study from Yaou, French Guyana. Geochim. Cosmochim. Acta, 64: 409-426. [2] Giraudon S., Danzart M. & Merle H. (2000) Deuterium nuclear magnetic resonance spectroscopy and stable carbon isotope ratio analysis/mass spectrometry of certain monofloral honeys. J. AOAC Int., 83(6): 1401-1409. [5] Giuliani G., France-Lanord C., Zimmermann J. L., Arboleda C., Charoy B., Coget P., Fontan F. & Giard D. (1997) Fluid composition, 6D of channel H20 and 6180 of lattice oxygen in beryls: genetic implications for Brazilian, Columbian, and Afghanistani emerald deposits. Intern. Geol. Rev., 39: 400424. [16] Giussani A., Roth P., Werner E., Schramel P., Wendler I. & Nusslin F. (1997) Thermal ionization and inductively coupled plasma mass spectrometry: Potential for the application to studies on the biokinetics of molybdenum in humans. Isot. Environ. Health Stud., 33: 207-215. [31] Gleason J. D., Friedman I. & Hanshaw B. B. (1969) Ectraction of dissolved carbonate species from natural water for carbon-isotope analysis. U.S. Geol. Surv. Prof. Paper, 650-D: D248-D250. [10] Gleditsch E. & Sandahl B. (1922) Radioactivit6 sur le poids atomique de chlore dans un mineral ancien, l'apatede Balme. C. R. Acad Sci. Paris, 174: 746-748. [28]
References
1095
Gleit C. E. & Holland W. D. (1962) Use of electrically excited oxygen for the low temperature decomposition of organic substances. Anal. Chem., 34: 1454-1457. [2] Glibert P. M., Lipschultz E, McCarthy J. J. & Altabet M. A. (1982) Isotope dilution models of uptake and remineralization of ammonium by marine plankton. Limnol. Oceanogr., 27: 639-650. [15] Godfrey J. D. (1962) The deuterium content of hydrous minerals from the East-Central Sierra Nevada and Yosemite National Park. Geochim. Cosmochim. Acta, 26: 1215-1245. [1] Goerges T. & Dittert K. (1998) Improved diffusion technique for 15N:14N analysis of ammonium and nitrate from aqueous samples by stable isotope spectrometry. Comm. Soil Sci. Plant Anal., 29(3&4): 361-368. [15] Goldhaber M. B. & Kaplan I. R. (1974) The sulphur cycle. In: The Sea. E. D. Goldberg (ed.). Interscience Wiley, NY, Vol. 5: 569-655. [46] Goldhaber M. B. & Kaplan I. R. (1975) Controls and consequences of sulfate reduction rates in recent marine sediments. Soil Sci., 119: 42-55. [27] Goldman A., Murcray F. J., Murcray D. G., Kosters J. J., Rinsland C. P., Flaud J. M., Camy-Peyret C. & Barbe A. (1989) Isotopic abundances of stratospheric ozone from balloon-borne high resolution infrared solar spectra. J. Geophys. Res., 94: 8467-8473. [18] Gonfiantini R. (1977) Consultants' meeting on stable isotope standards and intercalibration in hydrology and geochemistry. 8-10 Sept., 1976. IAEA, Vienna. Internal Report, March 1977. [40] Gonfiantini R. (1978) Standards for stable isotope measurements in natural compounds. Nature, 271(No. 5645): 534-536. [1, 40] Gonfiantini R. (1981) The h-notation and the mass-spectrometric measurement techniques. In: Stable isotope hydrology, deuterium and oxygen-18 in the water cycle. J. R. Gat & R. Gonfiantini (eds.). IAEA, Vienna. Technical Report Series, No. 210: 35-84. [1, 14, 15, 40, 43] Gonfiantini R. (1984) Advisory group meeting on stable isotope reference samples for geochemical and hydrological investigations. 19-21 Sept, 1983. Report to Director General, IAEA, Vienna: 77 p. [21, 40] Gonfiantini R., Stichler W. & Rozanski K. (1995) Standards and intercomparison materials distributed by the International Atomic Energy Agency for stable isotope measurements. In: Reference and intercomparison materials for stable isotopes of light elements. Proceedings of a Consultants Meeting held in Vienna, 1-3 Dec 1993. IAEA-TECDOC-825, IAEA, Vienna: 13-29. [40, 42, 43, 44] Gonfiantini R., Valkiers S., Taylor P. D. P. & De Bi6vre P. (1997) Using isotopic disequilibrium of CO2 to model gas adsorption in mass spectrometric measurements. Int. J. Mass Spectrom. Ion Proc., 161: 15-26. [43] Gonfiantini R., Tonarini S., Griining M., Adorni-Braccesi A., A1-Ammar A. S., Astner M., Btichler S., Barnes R. M., Bassett R. L., Cocherie A., Deyhle A., Dini A., Ferrara G., Gaillardet J., Grimm J., Guerrot C., Kr/ihenbiihl U., Layne G., Lemarchand D., Meixner A., Northington D. J., Pennisi M., Reitznerov~i E., Rodushkin I., Sugiura N., Surberg R., Tonn S., Wiedenbeck M., Wunderli S., Xiao Y. & Zack T. (2003) Intercomparison of boron isotope and concentration measurements. Part Ih Evaluation of the results obtained. Geostand. Newslett., 27(1): 41-57. [40] Gonzalez J., Jamin E., Remaud G., Martin Y. L., Martin G. G. & Martin M. L. (1998) Authentication of lemon juices and concentrates by a combined multi-isotope approach using SNIF-NMR and IRMS. J. Agric. Food Chem., 46: 2200-2205. [5] Gofii M. A. & Eglinton T. I. (1996) Stable carbon isotopic analyses of lignin-derived CuO oxidation products by isotope ratio monitoring-gas chromatography-mass-spectrometer (irm-GC-MS). Org. Geochem., 24: 601-615. [9] Goodman K. J. & Brenna J. T. (1994) Curve fitting for restoration od accuracy for overlapping peaks in gas chromatography/combustion isotope ratio mass spectrometry. Anal. Chem., 66: 1294-1301. [8] Goodman K. J. & Brenna J. T. (1995) High-precision gas-chromatography combustion isotope ratio mass-spectrometry at low signal levels. J. Chromatogr., A 689: 63-68. [8] Goodman K. J. (1998) Hardware modifications to an isotope ratio mass spectrometer continuous-flow interface yielding improved signal, resolution, and maintenance. Anal. Chem., 70: 833-837. [8] Grabczak J., Rozanski K. & Zuber A. (1986) Comments on "Water-rock interaction and chemistry of groundwaters from the Canadian Shield" by Frape, Fritz and McNutt. Geochim. Cosmochim. Acta, 50: 1559-1560. [1]
1096
References
Graber E. R. & Aharon P. (1991) An improved microextraction technique for measuring dissolved inorganic carbon (DIC), 613CDIC and 618OH20 from milliliter-size water samples. Chem. Geol., 94(2): 137144. [1, 10] Grachev A. M. & Severinghaus J. P. (2003) Laboratory determination of thermal diffusion constants for 29N2/28N2 in air at temperatures from -60 to 0~ for reconstruction of magnitudes of abrupt climate changes using the ice core fossil-air paleothermometer. Geochim. Cosmochim. Acta, 67: 345-360. [46] Gradsztajn E., Salome M., Yaniv A. & Bernas R. (1967) Isotopic analysis of lithium in the Holbrook meteorite and in terrestrial samples with a sputtering ion mass spectrometer. Earth Planet. Sci. Lett., 3: 387-393. [6] Gradsztajn E. & Guez M. (1969) Mass discrimination introduced by the optical properties of a sputtering ion source during the isotopic analysis of lithium in stony meteorites. Int. J. Mass Spectrom. Ion Phys., 4: 305-313. [6] Grady M. M., Verchovsky A., Wright I. P. & Pillinger C. T. (1996) Fully automated stepped combustion analysis of C, N, and Ar in QUE 94201: Comparison with other shergottites. Meteoritics, 31: A52-A53. [13] Graff J. & Rittenberg D. (1952) Microdetermination of deuterium in organic compounds. Anal. Chem., 24: 878-881. [1] Graham C. M. & Sheppard S. M. F. (1980) Experimental hydrogen isotope studies, II. Fractionations in the system epidote - NaC1 - H20, epidote - CaC12- H20 and epidote - seawater, and the hydrogen isotope composition of natural epidotes. Earth Planet. Sci. Lett., 49: 237-251. [46] Graham C. M., Sheppard S. M. F. & Heaton T. H. E. (1980) Experimental hydrogen isotope studies - I. Systematics of hydrogen isotope fractionation in the system epidote - H20, zoisite - H20 and A10(OH) - H20. Geochim. Cosmochim. Acta, 44: 353-364. [46] Graham C. M. (1981) Experimental hydrogen isotope studies: III. Diffusion of hydrogen in hydrous minerals, and stable isotope exchange in metamorphic rocks. Contrib. Mineral. Petrol., 76: 216-228. [46] Graham C. M., Harmon R. S. & Sheppard S. M. F. (1984) Experimental hydrogen isotope studies: hydrogen isotope exchange between amphibole and water. Am. Mineralogist, 69: 128-138. [46] Graham C. M., Viglino J. A. & Harmon R. S. (1987) Experimental study of hydrogen-isotope exchange between aluminous chlorite and water and of hydrogen diffusion in chlorite. Am. Mineralogist, 72: 566-579. [46] Graham C. M., Valley J. W. & Winter B. L. (1996) Ion microprobe analysis of 180 / 160 in authigenic and detrital quartz in the St. Peter Sandstone, Michigan Basin and Wisconsin Arch, USA: Contrasting diagenetic histories. Geochim. Cosmochim. Acta, 60(24): 5101-5116. [30] Graham R. P., MacNamara J., Crocker I. H. & MacFarlene R. B. (1951) The isotopic constitution of germanium. Can. J. Chem., 29: 89-102. [28] Grassineau N. V., Mattey D. P. & Lowry D. (2001) Sulfur isotope analysis of sulfide and sulfate minerals by continuous flow-isotope ratio mass spectrometry. Anal. Chem., 73(2)" 220-225. [26, 45] Gray J. & Tompson P. (1976) Climatic information from 180 / 160 ratios of cellulose in tree rings. Nature, 262: 481-482. [23] Gray J. & Thompson P. (1977) Climatic information from 180/160 analysis of cellulose, lignin and whole wood from tree rings. Nature, 270: 708-709. [24] Gray J., Yonge C. & Thompson P. (1984) Simultaneous determination of 180/160 and D / H in compounds containing C, H and O. Int. J. Appl. Radiat. Isot., 35(6)" 525-530. [1] Green J. W. (1963) Wood cellulose. In: Methods of carbohydrate chemistry. R. L. Whistler (ed.). Academic Press, NY: 9-21. [24] Green L. W., Leppinen J. J. & Elliot N. L. (1988) Isotopic analysis of lithium as thermal dilithium fluoride ions. Anal. Chem., 60: 34-37. [6] Green M. D., Rosman K. J. R. & De Laeter J. R. (1986) The isotopic composition of germanium in terrestrial samples. Int. J. Mass Spectrom Ion Proc., 68: 15-24. [31] Greenwood J. P., Mojzsis S. J. & Coath C. D. (2000) Sulfur isotopic compositions of individual sulfides in martian meteorites ALH84001 and Nakhla: Implications for crust-regolith exchange on Mars. Earth Planet. Sci. Lett., 184: 23-35. [30] Greenwood N. N. & Earnshaw A. (1997) Chemistry of the Elements. Butterworth & Heinemann. [26]
References
1097
Gregory R. T. & Taylor H. P. Jr. (1986) Possible non-equilibrium oxygen isotope effects in mantle nodules an alternative to the Kyser-O'Neil-Carmichael 180/16 0 geothermometer. Contrib. Mineral. Petrol., 93: 114-119. [46] Grey D. C. & Jensen M. L. (1972) Bacteriogenic sulfur in air pollution. Science, 177: 1099-1100. [26] Gr6goire D. C., Acheson B. M. & Taylor R. P. (1996) Measurement of lithium isotope ratios by inductively coupled plasma mass spectrometry: Application to geological materials. J. Anat. Atom. Spectrom., 11: 765-772. [6, 31] Griffith K. S. & Gellene G. I. (1992) Symmetry restrictions in diatom/diatom reactions. II. Nonmassdependent isotope effects in the formation of 04 +. J. Chem. Phys., 96: 4403-4411. [18] Griffiths H. (1998) Stable isotopes: Integration of biological ecological and geochemical processes. Bios Scientific Publishers, Oxford, UK: 438 p. [9, 24] Grimm W. (1968) Eine neue Glimmentladungslampe ftir die Optische Emissionspsektralanalyse. Spectrochim. Acta, B 23: 443-450. [35] Grinenko V. A., Grinenko L. N. & Zagryazhskaya G. D. (1969) Kinetic isotope effect in high temperature reduction of sulfate. Geokhimiya, 4: 484-491. [46] Grinenko V. A., Mineyev S. D., Devirts A. L. & Laguntina Y. P. (1987) Hydrogen isotope fractionation in the lizardite-water system at 100~ and I atm. Geokhimiya, 4: 574-579. [46] Grinsted M. J., Wilson A. T. & Ferguson C. W. (1979) 13C / 12C ratio variations in Pinus Iongaeva (Bristlecone pine) cellulose during the last millenium. Earth Planet. Sci. Lett., 42: 251-253. [24] Grob K. (1986) Making and manipulating capillary columns for gas chromatography. Dr. Alfred Huethig Verlag, Heidelberg, Germany. [8] Grossman E. L. & Ku T.-L. (1986) Oxygen and carbon isotope fractionation in biogenic aragonite: Temperature effects. Chem. Geol., 59: 59-74. [46] Grove M. (1993) Thermal histories of southern California basement terranes. PhD Thesis, University of California, Los Angeles, California, USA. [16] Grove M. & Bebout G. E. (1995) Cretaceous tectonic evolution of coastal southern California: insights from the Catalina Schist. Tectonics, 14: 1290-1308. [16] GriJning M., Fr/Jhlich K., De Regge P. P. & Danesi P. R. (1999) Intended use of the IAEA reference materials Part II: Examples on reference materials for stable isotope composition. In: The use of matrix reference materials in environmental analytical processes. A. Fajgelj & M. Parkany (eds.). Royal Society of Chemistry, Cambridge. Spec. Pub. No. 238: 81-92. [40] Gr/Jning M., van Duren M., Andreescu L., Tanweer A., Jaklitsch M. & Gattin M. (2004) Quality assurance for stable isotope analysis of water samples by mass spectrometry. Accr. Qual. Assur., (in press). [40] Guenther J., Krankowski D. & Mauersberger K. (2000) Third-body dependence of rate coefficients for ozone formation in 160-180 mixtures. Chem. Phys. Lett., 324: 31-36. [18] Guillou C. (1986) Application de la mesure du fractionnement isotopique de l'hydrogbne a la d6termination de l'enrichissement des vins. PhD Thesis, University of Nantes, Nantes, France. [5] Guillou C., Trierweiler M. & Martin G. J. (1988) Repeatability and reproducibility of site-specific isotope ratios in quantitative 2H NMR. Magn. Reson. Chem., 26: 491-496. [5] Guillou C. (1991) L'analyse isotopique des vins. Analusis Magazine, 19(10): M29-M34. [5] Guillou C. & Martin G. J. (1993) Characterization of reference tetramethylurea (TMU) for the determination of H / D ratio in alcohols by SNIF-NMR. BCR information - chemical analysis. Report CDNA-14396-EN-C. Commission of the European Communities, Directorate XIII, Luxembourg: 43 p. [5] Guillou C., Remaud G. & Lees M. (2001) EUR 20064- The certification of the carbon-13 and deuterium isotopic ratio content and the alcoholic grade of ethanol from wine, BCR-656 (96% vol). The carbon13 isotopic content of sugar, BCR-657. The oxygen-18 isotopic ratio content of water from wine, BCR-658 (7% vol) BCR-659 (12% vol). The carbon-13 and deuterium isotopic ratio content and the alcoholic grade of ethanol from wine BCR-660 (12% vol). European Commission, BCR information series, Office for Official Publications of the European Communities. Luxembourg: 94 p. (also available under URL: http://www.irmm.jrc.be/mrm.html). [40] Gulens J., Olmstead W. J., Denovan A. S., Gale K. L. & Rolston J. H. (1985) Determination of oxygen isotopes in water by direct oxidation and oxygen mass spectrometry. Int. J. Appl. Radiat. Isot., 36: 663667. [1]
1098
References
GUM (1995) Guide to the expression of uncertainty in measurement. IFCC, ISO, IUPAC, IUPAP, OIML. International Organization for Standardization, Geneva: 101 p. ISBN 92-67-10188-9. [32, 42] Gunter B. D. & Gleason J. D. (1971) Isotope fractionation during gas chromatographic separations. J. Chromat. Sci., 9: 191-192. [38] Guo J. & Qian Y. (1997) Hydrogen isotope fractionation and hydrogen diffusion in the tourmalinewater system. Geochim. Cosmochim. Acta, 61: 4679-4688. [46] Guzowski J. P. Jr., Broekaert J. A. C. & Hieftje G. M. (2000) Electrothermal vaporization for sample introduction into a gas sampling glow discharge time-of-flight mass spectrometer. Spectrochim. Acta, B 55: 1295-1314. [35] Giinther D., Frischknecht R. & Heinrich C. (1977) Capabilities of a 193 nm Arf excimer laser for LAICP-MS micro analysis of geological materials. J. Anal. At. Spectrom., 12: 939-944. [20] Giinther D. & Heinrich C. A. (1999) Comparison of the ablation behaviour of 266 nm Nd:YAG and 193 nm ArF excimer lasers for LA-ICP-MS analysis. J. Anal. At. Spectrom., 14: 1369-1374. [20] Gwiazda R., Woolard D. & Smith D. (1998) Improved lead isotope ratio measurements in environmental and biological samples with a double focusing magnetic sector inductively coupled plasma mass spectrometer. J. Anal. At. Spectrom., 13: 1233-1238. [35] Gygli A. (1993) Microdetermination of oxygen in organic compounds using a glassy carbon pyrolysis tube and nondispersive infrared detection. Mikrochim. Acta, 111: 37-43. [23] Haberhauer G. & Blochberger K. (1999) A simple cleanup method for the isolation of nitrate from natural water samples for O isotope analysis. Anat. Chem., 71: 3587-3590. [15] I-Iabfast K. (1960) Praezisionsbestimmung von Isotopenhaufigkeiten mit dem Massenspektrometer. Z. Instrumentenk., 68: 82-86. [39] I-Iabfast K. (1997) Advanced isotope ratio mass spectrometry I: Magnetic isotope ratio mass spectrometers. In: Modern isotope ratio mass spectrometry. I. T. Platzner (ed.). John Wiley & Sons, Chichester, UK: 11-82 [31, 38] I-Iabfast K. (1998) Fractionation correction and multiple collectors in thermal ionization isotope ratio mass spectrometry. Int. J. Mass Spectrom., 176: 133-148. [31] Habicht K. S. & Canfield D. E. (1997) Sulfur isotope fractionation during bacterial sulfate reduction in organic-rich sediments. Geochim. Cosmochim. Acta, 61: 5351-5361. [46] Habicht K. S., Canfield D. E. & Rethmeier J. (1998) Sulfur isotope fractionation during bacterial sulfate reduction and disproportionation of thiosulfate and sulfite. Geochim. Cosmochim. Acta, 62: 2585-2595. [46] Haendel D., Muhle K., Nitzsche H.-M., Stiehl G. & Wand U. (1986) Isotopic variations of the fixed nitrogen in metamorphic rocks. Geochim. Cosmochim. Acta, 50: 749-758. [16] Hagedorn M. L. (1992) Differentiation of natural and synthetic benzaldehydes by 2H nuclear magnetic resonance. J. Agric. Food Chem., 40: 634-637. [5] Hagemann R., Nief G. & Roth E. (1970) Absolute isotopic scale for deuterium analysis of natural waters. Absolute D / H ratio for SMOW. Tellus, 22: 712-715. [40, 43] Hagemann R. & Lohez P. (1978) Twin mass spectrometers for simultaneous isotopic analysis of hydrogen and oxygen in water. Adv. Mass Spectrom., 7A: 504-508. [1] Hagiwara Y. (2000) Selenium isotope ratios in marine sediments and algae - A reconaissance study. MSc Thesis, Univ. of Illinois at Urbana-Champaign, Urbana, IL, USA: 62 p. [29] Haglund R. F. Jr. & Itoh N. (1994) Electronic processes in laser ablation of semiconductors and insulators. In: Laser ablation. J. C. Miller (ed.). Springer-Verlag. Springer series in material sciences, Vol. 28: 11-52. [20] Haimson M. & Knauth L. P. (1983) Stepwise fluorination- a useful approach for the isotopic analysis of hydrous minerals. Geochim. Cosmochim. Acta, 47: 1589-1595. [20] Haisch M., Hering P., Fabinski W. & Z/~chbauer M. (1996) Isotopenselektive Kozentrationsmessungen an Atemgasen mit einem NDIR-Spektrometer. Technisches Messen, 63: 322-328. [34] Halas S. (1979) An automatic inlet system with pneumatic changeover valves for isotope ratio mass spectrometer. J. Phys. E: Sci. Instrum., 12: 418-420. [26] Halas S. & Skorzynski Z. (1980) An inexpensive device for digital measurements of isotopic ratios. J. Phys., E: Sci. Instrum., 13: 346-349. [39]
References
1099
Halas S. & Wolacewicz W. P. (1981) Direct extraction of sulfur dioxide from sulfates for isotopic analysis. Anal. Chem., 53: 686-689. [26, 45] Halas S. (1987) On bias in 34S/34S data obtained using SO2 gas in mass spectrometry. Advisory Group Meeting on Studies on Sulphur Isotope Variations in Nature. IAEA, Vienna: 105-111 [26]. Halas S. & Sikora J. (1987) Precise m / e monitoring in a magnetically scanned mass spectrometer by using two Hall probes. J. Phys. E., Sci. Instrum., 20: 662-664. [39] I-Ialas S. & Sikora J. (1990) Electron emission stabiliser with double negative feedback loop. Meas. Sci. Technol., 1: 980-982. [39] Halas S., Kaminski A. & Durakiewicz T. (1993) A bridge-based temperature controller with a resister as temperature sensor and heater. Meas. Sci. Technol., 4: 1208-1212. [39] Halas S. & Durakiewicz T. (1995) Improved adsorbant trap for gases. Chem. Geol. (Isot. Geosci. Sect.), 122: 271-273. [36] Halas S. & Kaminski A. (1995) An instrumental bridge. Patent description PL 1667134 B1, (in Polish). [39] Halas S. & Jasinska B. (1996) Platinized magnesium as reducing agent for hydrogen isotope analysis. Isot. Environm. Health Stud., 32(1): 105-109. [1] Halas S., Szaran J. & Niezgoda H. (1997) Experimental determination of carbon isotope equilibrium fractionation between dissolved carbonate and carbon dioxide. Geochim. Cosmochim. Acta, 61: 26912695. [46] Halas S. & Durakiewicz T. (1998a) Temperature distribution along a metal filament heated in vacuum by DC current. Vacuum, 49: 331-336. [39] Halas S. & Durakiewicz T. (1998b) Filament temperature stabilizer for a thermal ionization mass spectrometer. Int. J. Mass Spectrom. Ion. Proc., 181: 167-171. [39] Halas S. & Szaran J. (1999) Low-temperature thermal decomposition of sulfates to SO2 for on-line 34S/ 32S analysis. Anal. Chem., 71: 3254-3257. [26, 44] I-lalas S., Szaran J. & Niezgoda H. (2000) Reply to the comment by M. P. Verma on "Experimental determination of carbon isotope equilibrium fractionation between dissolved carbonate and carbon dioxide". Geochim. Cosmochim. Acta, 64: 575-576. [46] Halas S. & Szaran J. (2001) Improved thermal decomposition of sulfates to SO2 and mass spectrometric determination of 634S of IAEA SO-5, IAEA SO-6 and NBS-127 sulfate standards. Rapid Commun. Mass Spectrom., 15: 1618-1620. [26, 40] Halas S., W6jtowicz A., Nowak J. & Durakiewicz T. (2001) Temperature controller for thermal ionization mass spectrometry. Rapid Commun. Mass Spectrom., 15: 1-4. [39] I-Ialicz L., Galy A., Belshaw N. S. & O'Nions, R. K. (1999) High-precision measurement of calcium isotopes in carbonates and related materials by multiple collector inductively coupled plasma mass spectrometry (MC-ICP-MS). J. Anal. Atom. Spectrom., 14: 1835-1838. [31] Hall A. (1999) Ammonium in granites and its petrogenitic significance. Earth Sci. Revs., 45: 145-165. [16] Hall G. E. M., Pelchat J.-C. & Loop J. (1988) Separation and recovery of various sulphur species in sedimentary rocks for stable isotopic determination. Chem. Geol., 67: 35-45. [26] Hall J. A., Barth J. A. C. & Kalin R. M. (1999) Routine analysis by high precision gas chromatography mass selective detector isotope ratio mass spectrometry to 0.1 parts per mil. Rapid Commun. Mass Spectrom., 13: 1231-1236. [8] Halliday A. N., Lee D.-C., Christensen J. N., Walder A. J., Freedman P. A., Jones C. E., Hall C. M., Yi W. & Teagle D. (1995) Recent developments in inductively coupled plasma magnetic sector multiple collector mass spectrometry. Intl. J. Mass Spec. Ion Proc., 146 / 147: 21-33. [31, 35] Halliday A. N., Lee D.-C., Christensen J. N., Rehk~imper M., Yi W., Luo X., Hall C. M., Ballentine C. J., Pettke T. & Stifling C. (1998) Applications of multiple collector-ICPMS to cosmochemistry, geochemistry, and paleoceanography. Geochim. Cosmochim. Acta, 62(6): 919-940. [31] Halliday A. N., Christensen J. N., Lee D.-C., Rehktimper M., Hall C. M. & Luo X. (2000a) Multiplecollector inductively coupled plasma mass spectrometry. In: Inorganic mass spectrometry. C. M. Barshick, D. C. Duckworth & D. H. Smith (eds.). Marcel Dekker: 291-328. [31] Halliday A. N., Rehk/imper M., Schi~nb/ichler M., Oberli F., Freedman P. A., Frank M., Miiller W., Teutsch N., Baur H. & Wiechert U. (2000b) MC-ICPMS - the good, the small and the massive. J. Conf. Abstr., 5:474 (Abstract). [31]
11 O0
References
Halsted R. E. & Nier A. O. (1950) Gas flow through the mass spectrometer viscous leak. Rev. Sci.
Instrum., 21: 1019-1021. [38] Hamelin B., Manhes G., Albar6de F. & All6gre C. J. (1985) Precise lead isotope measurements by the double spike technique; a reconsideration. Geochim. Cosmochim. Acta, 49(1): 173-182. [29] Hamester M., Wiederin D., Wills J., Kerl W. & Douthitt C. B. (1999) Strategies for isotope ratio measurements with a double focusing sector field ICP-MS. Fresenius J. Anal. Chem., 364: 495-497. [31] Hamza M. S. & Broecker W. S. (1974) Surface effect on the isotopic fractionation between CO2 and some carbonate minerals. Geochim. Cosmochim. Acta, 38: 669-681. [46] Hamza M. & Epstein S. (1980) Oxygen isotopic fractionation between different oxygen sites in hydrous silicate minerals. Geochim. Cosmochim. Acta, 44: 173-182. [2, 20] Hang W., Walden W. O. & Harrison W. W. (1996) Microsecond pulsed glow discharge as an analytical spectroscopic source. Anal. Chem., 68: 1148-1152. [35] Hang W., Yan X., Wayne D. M., Olivares J. A., Harrison W. W. & Majidi V. (1999) Glow discharge source interfacing to mass analyzers: Theoretical and practical considerations. Anal. Chem., 71: 32313237. [35] Hanneguelle S., Thibault J. N., Naulet N. & Martin G. J. (1992) Authentication of essential oils containing linalool and linalyl acetate by isotopic methods. J. Agric. Food Chem., 40: 81-87. [5] Hanschmann G. (1981) Berechnung von isotopeneffekten auf quantenchemischer Grundlage am Beispiel stickstoffhaltiger Molekule. Zitl.-Mitt., 41: 19-39. [16] Hansen C. S., Calaway W. F., Pellin M. J., Wiens R. C. & Burnett D. S. (1996) Three-color resonance ionization spectrometry of Zn in Si. Resonance Ionization Spectroscopy 1996, Eighth International Symposium N. Winograd & J.E. Parks (eds.). AIP Press Woodbury, NY, AIP Conference Proceedings 388: 215-218. [33] Hanson D. & Mauersberger K. (1986) The vapor pressure of solid and liquid ozone. J. Chem. Phys., 85: 4669-4672. [18] Hansson S., Hobbie J. E., Elmgren R., Larsson U., Fry B. & Johansson S. (1997) The stable nitrogen isotope ratio as a marker of food-web interactions and fish migration. Ecology, 78: 2249-2257. [9] Hardcastle K. G. & Friedmann I. (1974) A method for oxygen isotope analysis of organic material. Geophys. Res. Lett., 1: 165-167. [23] Hare P. E., Fogel M. L., Stafford T. W., Mitchell A. D. & Hoering T. C. (1991) The isotopic composition of carbon and nitrogen in individual amino aciuds isolated from modern and fossil proteins. J. Archaeol. Sci., 18: 277-292. [9] Harkins W. D. & Stone S. B. (1926) The isotopic composition and atomic weight of chlorine from meteorites and from minerals of non-marine origin. J. Am. Chem. Soc., 48: 938-949. [28] Harmon R. S., Schwarcz H. P & O'Neil J. R. (1979) D / H ratios in speleothem fluid inclusions; a guide to variations in the isotopic composition of meteoric precipitation. Earth Planet. Sci. Lett., 42: 254-266. [3] Harmon R. S. & I-Iinton R. W. (1992) Frontiers in isotope geosciences, part II. Chem. Geol., 102: 169-199. [30] Harren F. & Reuss J. (1997) Photoacoustic spectroscopy. In: Encyclopedia of applied physics. G. L. Trigg (ed.).Weinheim, Vol. 19: 413-435. [34] Harren F. J. M., Cotti G., Oomens J. & te Lintel Hekkert S. (2000) Photoacoustic spectroscopy in trace gas monitoring. In: Encyclopedia of analytical chemistry. R. A. Meyers (ed.). John Wiley Ltd, Chichester: 2203-2226. [34] Harrington R. R., Kennedy B. P., Chamberlain C. P., Blum J. D. & FoR C. L. (1998) 15N enrichment in agricultural catchments: field patterns and applications to tracking Atlantic salmon (Salmo salar). Chem. Geol., 147: 281-294. [15] Harris C., Compton J. S. & Bevington S. A. (1999) Oxygen and hydrogen isotope composition of kaolinite deposits, Cape Peninsula, South Africa: Low-temperature, meteoric origin. Econ. Geol., 94: 13531366. [2] Harris N. B. W., Jackson D. H., Mattey D. P., Santosh M. & Bartlett J. (1993) Carbon-isotope constraints on fluid advection during contrasting examples of incipient charnockite formation. J. Metam. Geol., 11: 833-843. [3]
References
1101
Harrison A. G. & Thode H. G. (1957) The kinetic isotope effect in chemical reduction of sulphate. Trans. Faraday Soc., 53: 1648-1651. [46] Harrison A. G. & Thode H. G. (1958) Mechanism of the bacterial reduction of sulphate from isotope fractionation studies. Trans. Faraday Soc., 54: 84-92. [46] Harrison W. G. (1978) Experimental measurements of nitrogen remineralization in coastal waters. Limnol. Oceanogr., 23: 684-694. [15] Harrison W. W. & Magee C. W. (1974) Hollow cathode ion source for solids mass spectrometry. Anal. Chem., 46: 461-464. [35] Harrison W. W., Hess K. R., Marcus R. K. & King F. L. (1986) Glow discharge mass spectrometry. Anat. Chem., 58: 341A-356A. [35] Harrison W. W. (1988) Glow discharge mass spectrometry. In: Chemical analysis. F. Adams, R. Gijbels & R. Van Grieken (eds.). Wiley-Interscience, NY, Vol. 95: 85-123. [35] Harrison W. W. & Bentz B. L. (1988) Glow discharge mass spectrometry. Prog. Analyt. Spectrosc., 11: 53110.[35] Harrison W. W. (1998) Pulsed glow discharge as a solid analysis source. J. Anat. At. Spectrom., 13: 10511056. [35] Hart S. R. & Zindler A. (1989) Isotope fractionation laws: a test using calcium. Int. J. Mass Spectrom. Ion Proc., 89: 287-301. [31] Harte B., Fitzsimons I. C. W., Harris J. W. & Otter M. L. (1999) Carbon isotope ratios and nitrogen abundances in relation to cathodoluminescence characteristics for some diamonds from the Kaapvaal Province, S. Africa. Mineral. Magazine, 63(6): 829-856. [30] Hartley P. E. (1980) Mass spectrometric determination of deuterium in water at natural levels. Anat. Chem., 52: 2232-2234. [1] Hartmann M. & Nielsen H. (1969) 634S-Werte in rezenten Meeressedimenten und ihre Deutung am Beispiel einiger Sedimentprofile aus der westlichen Ostsee. Geol. Rundschau, 58: 621-655. [27] Hashimoto I. & Jackson M. L. (1960) Rapid dissolution of allophane and kaolinite-halloysite after dehydration. Clays Clay Minerals, 7: 102-113. [2] Hashizume K. & Sugiura N. (1990) Precise measurement of nitrogen isotopic composition using a quadrupole mass spectrometer. Mass Spectroscopy, 38: 269-286. [13, 17] Hashizume K. & Sugiura N. (1992) Measurement of cosmogenic nitrogen using a static massspectrometry system and its implication. Geochim. Cosmochim. Acta, 56: 1625-1631. [17] Hashizume K. & Sugiura N. (1995) Nitrogen isotopes in bulk ordinary chondrites. Geochim. Cosmochim. Acta, 59: 4057-4069. [16] Hashizume K., Chaussidon M., Marty B. & Robert F. (2000) Solar wind record on the moon: Deciphering presolar from planetary nitrogen. Science, 290: 1142-1145. [17] Hashizume K., Marty B. & Weiler R. (2002) Analyses of nitrogen and argon in single lunar grains: towards a quantification of the asteroidal contribution to planetary surfaces. Earth Planet. Sci. Lett., 202: 201-216. [17] Hassan A. A. (1982) Methodologies for extraction of dissolved inorganic carbon for stable carbon isotope studies evaluation and alternatives. U. S. Geol. Survey, Water Resour. Invest. Report No. 82-6: 51p. [10] Hathorn B. C. & Marcus R. A. (1999) An intramolecular theory of the mass-independent isotope effect for ozone. I. J. Chem. Phys., 111: 4087-4100. [15, 18] Hathorn B. C. & Marcus R. A. (2000) An intramolecular theory of the mass-independent isotope effect for ozone. II. Numerical implementation at low pressure using a loose transition state. J. Chem. Phys., 113(21): 9497-9509. [18] Hattori K. & Sakai H. (1979) D / H ratios, Origins, and Evolution of the Ore-Forming Fluids for the Neogene Veins and Kuroko Deposits of Japan. Econ. Geol., 74: 535-555. [3] I-Iauck R. D. & Bremner J. M. (1976) Use of tracers for soil and fertilizer nitrogen research. Adv. Agron., 28: 219-266. [15] Hauck R. D. (1982) Nitrogen-isotope-ratio analysis. In: Methods of soil analysis, Part 2. Chemical and microbiological properties. A. L. Page, R. H. Miller & D. R. Keeney (eds.). American Society of Agronomy Inc. - Soil Science Society of America Inc., Madison, Wisconsin. Agronomy Series, 2nd edition, No. 9: 735-779. [15]
1102
References
Haur A., Hladikov~i J. & Smejkal V. (1973) Procedure of direct conversion of sulfates into SO2 for mass spectrometric analysis of sulfur. Isotopenpraxis, 9: 329-331. [26] Hayatsu R. & Anders E. (1981) Organic compounds in meteorites and their origins. Top. Curr. Chem., 99: 1-37.[11] Hayes J. M. & Schoeller D. A. (1977) High precision pulse counting: Limitations and optimal conditions. Anal. Chem., 49(2): 305-311. [30] Hayes J. M. & Baker K. L. (1986) Reagent and procedure for preparation of H2 for hydrogen-isotopic analysis of water. Indiana University, Bloomington. Intern. Rpt. (unpublished). [1] Hayes J. M. & Johnson M. W. (1988) Reagent and procedure for preparation of H2 for hydrogen-isotopic analysis of water. Indiana University, Bloomington. Intern. Rpt. (unpublished). [1] Hayes J. M., Freeman K. H., Popp B. N. & Hoham C. H. (1990) Compound-specific isotopic analyses, a novel tool for reconstruction of ancient biogeochemical processes. Org. Geochem., 16: 1115-1128. [9, 26, 45] Hays P. A., Remaud G. S., Jamin E. & Martin Y. L. (2000) Geographic origin determination of heroin and cocaine using site-specific isotopic ratio deuterium NMR. J. Forensic Sci., 45(3): 552-562. [5] Heathcote J. A. (1981) Hydrochemical aspects of the Gipping Chalk salinity investigation. PhD Thesis, University of Birmingham. [10] Heathcote J. A. (1985) Carbonate chemistry of recent Chalk groundwater in part of East Anglia. J. Hydrol., 78: 215-227. [10] I-Ieaton T. H. E. & Collett G. M. (1985) The analysis of 15N / 14N ratios in natural samples, with emphasis on nitrate and ammonium in precipitation. Conc. Sci. Ind. Res., Pretoria. Res. Report: 624. [15] Heaton T. H. E. (1999) Spatial, species, and temporal variations in the 13C / 12C ratios of C3 plants: implications for palaeodiet studies. J. Archaeological Soc., 26: 637-649. [46] Heidenreich J. E. III & Thiemens M. H. (1986) A non-mass-dependent oxygen isotope effect in the production of ozone from molecular oxygen: The role of molecular symmetry in isotope chemistry. J. Chem. Phys., 84: 2129-2136. [18, 46] Heier K. S. & Billings G. K. (1972) Lithium. In: Handbook of Geochemistry. K. H. Wedepohl (ed.). Springer-Verlag, Vol. II-I, Chap. 3. [6] Hein R., Crutzen P. J. & Heimann M. (1997) An inverse modeling approach to investigate the global atmospheric methane cycle. Global Biogeochem. Cycles, 11(1): 43-76. [14] I-Ieinrichs H. & Herrmann A. G. (1990) Praktikum der analytischen Geochemie. Springer-Verlag. [27] Helle J. F., Hillaire-Marcel C. & Rondeau B. (2002) Seasonal changes in the sources and fluxes of dissolved inorganic carbon through the St. Lawrence River- isotopic and chemical constraint. Chem. Geol., 186: 117-138. [10] Helliker B. R. & Ehleringer J. R. (2000) Establishing a grassland signature in veins: 180 in the leaf water of C3 and C4 grasses. Proceed. Nat. Acad. Sci., 97(14): 7894-7898. [21] Helliker B. R., Roden J. S., Cook, C. & Ehleringer J. R. (2002) A rapid and precise method for sampling and determining the oxygen isotope ratio of atmospheric water vapor. Rapid Commun Mass Spectr., 16: 929-932. [1] Hemming N. G. & Hanson G. N. (1992) Boron isotopic composition and concentration in modern marine carbonates. Geochim. Cosmochim. Acta, 56: 537-543. [7, 46] Hemming N. G. & Hanson G. N. (1994) A procedure for the isotopic analysis of boron by negative thermal ionization mass spectrometry. Chem. Geol. (Isot. Geosci. Sect.), 114: 147-156. [7, 29] Henchman M. J., Smith D. & Adams N. G. (1982) Estimation of the enthalpy changes in several ionmolecule reactions involving H / D exchange from zero-point-energy considerations. Int. J. Mass Spectrom. Ion Phys., 42: 25-32. [4] Henderson J. H., Clayton R. N., Jackson M. L., Syers J. K., Rex R. W., Brown J. L. & Sachs I. B. (1972) Cristobalite and quartz isolation from soils and sediments by hydrofluosilicic acid treatment and heavy liquid separation. Soil Sci. Soc. Am. Proc., 36: 830-835. [2] Hener U., Brand W. A., I-Iilkert A. W., Juchelka D., Mosandl A. & Podebrad F. (1998) Simultaneous online analysis of 180/16 0 and 13C/12C ratios of organic compounds using GC-pyrolysis-IRMS. Food Res. Technol. Z. Lebensm. Unters. Forsch., A206(3): 230-232. [8, 23] Herbel M. J. (1997) Stable selenium isotopic ratios of saline environmental samples from the western San Joaquin Valley of California. PhD Thesis, University of California, Davis. [29]
References
1103
Herbel M. J., Johnson T. M., Oremland R. S. & Bullen T. D. (1998) Selenium stable isotope fractionation during bacterial dissimilatory reduction of selenium oxyanions. Geochim. Cosmochim. Acta, 64: 37013709. [291 Herbel M. J., Johnson T. M., Oremland R. S. & Bullen T. D. (2000) Fractionation of selenium isotopes during bacterial respiratory reduction of selenium oxyanions. Geochim. Cosmochim. Acta, 64: 37013709. [29, 46] Herbelin J. M., McKay J. A., Kwok M. A., EUenten R. H., Urevig D. S., Spencer D. J. & Bernard D. J. (1980) Sensitive measurement of photon lifetime and true reflectances in an optical cavity by a phase-shift method. Appl. Opt., 19: 144-147. [34] Herbert B. K. (1976) A circuit for stabilizing the electron current to the anode of a hot-filament device. Vacuum, 26: 363-369. [39] Herczeg A. L. (1987) A stable isotopic study of dissolved inorganic carbon cycling in a softwater lake. Biogeochemistry, 4: 231-263. [10] Hermann A. (1999) Determination of site-specific D / H isotope ratios of glycerol from different sources by 2H-NMR spectroscopy. Z. Lebensm. Unters. Forsch., A 208: 194-197. [5] Hermann A. (2001) Determination of D / H isotope ratio in acetic acid from vinegars and pickled products by 2H-NMR-spectroscopy. Eur. Food Res. Technol., 212: 683-686. [5] Herriott D. R., Kogelnik H. & Kompfner R. (1964) Off-axis paths in spherical interferometers. Appl. Opt., 3: 523-526. [34] Hersh C. K. (1961) Molecular sieves. Reinhold, NY: 129 p. [36] Herut B., Spiro B., Starinsky A. & Katz A. (1995) Sources of sulfur in rainwater as indicated by isotopic h34S data and chemical composition, Israel. Atmos. Environ., 29: 851-857. [26] Hervig R. L. (1992) Oxygen isotope analysis using extreme energy filtering. Chem. Geol. (Isot. Geosci. Sect.), 101: 185-186. [30] Hervig R. L., Williams P., Thomas R. N., Schauer S. N. & Steele I. M. (1992) Microanalysis of oxygen isotopes in insulators by secondary ion mass spectrometry. Int. J. Mass Spec. Ion Proc., 120: 45-63. [30] Herzog G. F., Xue S., Hall G. S., Nyquist L. E., Shih C.-Y., Wiesmann H. & Brownlee D. E. (1999) Isotopic and elemental composition of iron, nickel, and chromium in type I deep-sea spherules: Implications for origin and composition of the parent micrometeoroids. Geochim. Cosmochim. Acta, 63(9): 1443-1457. [29] Herzog R. (1934) Ionic- and electro-optical cylindrical lenses. Z. Phys., 89: 447-. [38] Herzog W. & Di~rnenburg E. (1958) Die Verwendung von Methylchlorid zur massenspektrometrischen Isotopenanlyse von Chlor. Z. Naturforsch., 13a: 51-52. [28] Herzog W. & Klemm A. (1958) Die Temperaturabh~ingigkeit des Isotopie-Effektes bei der elektrolytischen Wanderungen der Chlorionen in geschmolzenem Thallium(I)-chlorid. Z. Naturforsch., 13a: 7-16. [28] Hesslein R. H., Capel M. J., Fox D. E. & Hallard K. A. (1991) Stable isotopes of sulfur, carbon, and nitrogen as indicators of trophic level and fish migration in the lower Mackenzie River Basin, Canada. Can. J. Fish Aquat. Sci., 48: 2258-2265. [26] Hesslein R. H., Hallard K. A. & Ramlal P. (1993) Replacement of sulfur, carbon, and nitrogen in tissue of growing broad whitefish (Coregonus nasus) in response to a change in diet traced by h34S, 613C, and 615N. Can. J. Fish. Aquat. Sci., 50: 2071-2076. [9] Heumann K. G. & Zeininger H. (1985) Boron trace determination in metals and alloys by isotope dilution mass spectrometry with negative thermal ionization. Int. J. Mass Spectrom. Ion Proc., 67: 237252. [7] Heumann K. G. & R~idlein N. (1989) Negative thermal ionization mass spectrometry of selenium, Part 3. Selenium trace determination in food samples. Fresenius Z. Anal. Chem., 335: 751-754. [29] Heumann K. G. & Wachsmann M. (1989) Negative thermal ionization mass spectrometry of selenium, Part 4. Selenium trace determination in sediments and related samples. Fresenius Z. Anal. Chem., 335: 755-758. [29] Heumann K. G., Eisenhut S., Gallus S., Hebeda E. H., Nusko R., Vengosh A. & Walczyk T. (1995) Recent developments in thermal ionization mass spectrometric techniques for isotope analysis: a review. Analyst, 120: 1291-1299. [7]
1104
References
Heumann K. G., Gallus S. M., Radlinger G. & Vogl J. (1998) Precision and accuracy in isotope ratio measurements by plasma source mass spectrometry. J. Anal. At. Spectrom., 13: 1001-1008. [26, 32, 35] Heuser E. (1944) Cellulose chemistry. J. Wiley & Sons. [24] Heymann D., Yancey T. E., Wolbach W. S., Thiemens M. H., Johnson E. A., Roach D. & Moecker S. (1998) Geochemical markers of the Cretaceous-Tertiary boundary event at Brazos River, Texas, USA. Geochim. Cosmochim. Acta, 62(1): 173-181. [45] Hildebrand P. & Beglinger C. (1997) Nondispersive infrared spectrometry. A new method for the detection of Heliobacter pylori infection with the 13C-urea breath test. Clin. Infect. Dis., 25: 1003-1005. [34] Hilkert A. W., Douthitt C. B., Schliiter H. J. & Brand W. A. (1999) Isotope ratio monitoring gas chromatography/mass spectrometry of D / H by high temperature conversion isotope ratio mass spectrometry. Rapid Commun. Mass Spectrom., 13: 1226-1230. [1, 14, 38] Hill J. W. & Fry A. (1962) Chlorine isotope effects in the reactions of benzyl and substituted benzyl chlorides with various nucleophiles. J. Am. Chem. Soc., 84: 2763-2769. [28] Hillier S., Fallick A. E. & Matter A. (1996) Origin of pore-lining chlorite in the aeolian Rotliegend of northern Germany. Clay Minerals, 31: 153-171. [2] Hinners T. A., Hughes R., Outridge P. M., Davis W. J., Simon K. & Woolard D. R. (1998) Interlaboratory comparison of mass spectrometric methods for lead isotopes and trace elements in NIST SRM 1400 Bone Ash. J. Anal. At. Spectrom., 13: 963-970. [35] Hinton R. W., Long J. V. P., Fallick A. E. & Pillinger C. T. (1983) Ion probe measurements of D / H ratios in meteorites. Lunar Planet. Sci., 14: 313-314. [30] Hinton R. W. (1995) Ion microprobe analysis in geology. In: Microbe techniques in Earth sciences. P. J. Potts, J. F. W. Bowles, S. J. B. Reed & M. R. Cave (eds.). Chapman & Hall, London: 235-289. [30] Hintze C. (1915) Handbuch der mineralogie. Verlag von Veit & Comp., Leipzig. [28] I-Iippler H., Rahn R. & Troe J. (1990) Temperature and pressure dependence of ozone formation rates in the range 1-1000 bar and 90-370 K. J. Chem. Phys., 93: 6560-6569. [18] Hirata T. (1996) Lead isotopic analyses of NIST standard reference materials using multiple collector inductively coupled plasma mass spectrometry coupled with a modified external correction method for mass discrimination effect. Analyst, 121: 1407-1411. [31] I-Iirata T. (1997) Isotopic variations of germanium for iron and stony iron meteorites. Geochim. Cosmochim. Acta, 61: 4439-4448. [31] Hirata T. & Ohno T. (2001) In-situ isotopic ratio analysis of iron using laser ablation-multiple collectorinductively coupled plasma mass spectrometry (LA-MC-ICP-MS). J. Anal. At. Spectrom., 16: 487-491. [31] Hirata T. (2003) Chemically assisted laser ablation ICPO mass spectrometry. Anal. Chem., 75: 228-233. [20] Hitchcock D. R. & Black M. S. (1984) 34S/32Sevidence of biogenic sulfur oxides in a salt marsh atmosphere. Atmos. Environ., 18: 1-17. [26] Hitchon B. (1974) Aplication of geochemistry to the search for crude oil and natural gas. In: Introduction to Exploration Geochemistry. A. A. Levinson (ed.). Applied Publishing Ltd.: 509-545 [26] Hobson K. A. & Clark R. G. (1992) Assessing avian diets using stable isotopes II: factors influencing diet-tissue fractionation. The Condor, 94: 189-197. [9] Hobson K. A., Hughes K. D. & Ewins P. J. (1997) Using stable isotope analysis to identify endogenous and exogenous sources of nutrients in eggs of migratory birds: applications to Great Lakes contaminants research. The Auk, 114: 467-478. [26] I-Iobson K. A. & Schell D. M. (1998) Stable carbon and nitrogen isotope patterns in baleen from eastern Artic bowhead whales (Balaena mysticetus). Can. J. Fish. Aquat. Sci., 55: 2601-2607. [9] Hobson K. A. (1999) Tracing origins and migration of wildlife using stable isotopes: a review. Oecologia, 120: 314-326. [26] Hobson K. A., Drever M. C. & Kaiser G. W. (1999) Norway rats as predators of burrow-nesting seabirds: insights from stable isotope analyses. J. Wildl. Manage., 63: 14-25. [26] Hockey J. A. & Pethica B. A. (1961) Surface hydration of silicas. Trans. Faraday Soc., 57: 2247-2262. [3] Hoefs J. & Stalder H. A. (1977) Die C-Isotopenzusammensetzung von CO2-haltigen Flfissigkeitseinschlfisse in Kluftquartzen der Zentralalpen. Schweiz. Miner. Petrol. Mitt., 57: 329-347. [3]
References
1105
Hoefs J. & Morteani G. (1979) The carbon isotopic composition of fluid inclusions in Alpine fissure quartzes from the western Tauern Window (Tyrol, Austria). Neues Jhb. Mineral., 3: 123-134. [3] Hoefs J. (1980) Stable Isotope Geochemistry. Springer-Verlag. 2nd edition: 203 p. [45] Hoefs J. (1997) Stable Isotope Geochemistry. Springer-Verlag. 4th completely revised, updated, and enlarged edition: 201 p. [App B] Hoefs J. & Sywall M. (1997) Lithium isotope composition of Quaternary and Tertiary biogene carbonates and a global lithium isotope balance. Geochim. Cosmochim. Acta, 61: 2679-2690. [6] Hoekstra H. R. & Katz J. J. (1953) Direct determination of oxygen in less familiar metal oxides. Anal. Chem., 25: 1608-1612. [20] Hoering T. (1955) Variations of nitrogen fifteen abundance in naturally occuring substances. Science, 122: 1233-1234. [16] I-Ioering T. (1957) The isotopic composition of ammonia and nitrate ion in rain. Geochim. Cosmochim. Acta, 12: 97-102. [15] Hoering T. C. & Parker P. L. (1961) The geochemistry of the stable isotopes of chlorine. Geochim. Cosmochim. Acta, 23: 186-199. [28] Hoering T. C. (1990) Development of the sulfur hexafluoride method for sulfur isotope analysis. Carnegie Institution, Washington. Ann. Rept. Dir. Geophys. Lab. 1989-1990, No. 2200: 128-131. [20] Hofmann A. (1971) Fractionation corrections for mixed-isotope spikes of Sr, K, and Pb. Earth Planet. Sci. Lett., 10(4): 397-402. [29] Hofmann D., Jung K., Segschneider H.-J., Gehre M. & Schiiiirmann G. (1995) 15N/14N analysis of amino acids with GC-C-IRMS - methodical investigations and ecotoxicological applications. Isot. Environ. Health Stud., 31: 367-375. [8] Hogg A. J. C., Pearson M. J. & Fallick A. E. (1993) Pretreatment of Fithian illite for oxygen isotope analysis. Clay Minerals, 18: 149-152. [2] Holloway J. M. & Dahlgren R. A. (2002) Nitrogen in rock: occurrences and biogeochemical implications. Global Biogeochem. Cycles, 14(4): doi 10.1029/2002GB001862,2002. [15] Holloway J. R. & Wood B. J. (1988) Simulating the Earth: Experimental Geochemistry. Unwin Hyman, Boston, USA: 196 p. [46] Holmden C., Papanastassiou D. A. & Wasserburg G. J. (1997) Negative thermal ion mass spectrometry of oxygen in phosphates. Geochim. Cosmochim. Acta, 61: 2253-2264. [22] Holmes R. M., McLelland J. W., Sigman D. M., Fry B. & Peterson B. J. (1998) Measuring 15N-NH4 + in marine, estuarine and fresh waters: An adaption of the ammonia diffusion method for samples with low ammonium concentrations. Marine Chem., 60: 235-243. [9, 15] Holmes R. M., Aminot A., Kerouel R., Hooker B. A. & Peterson B. J. (1999) A simple and precise method for measuring ammonium in marine and freshwater ecosystems. Can. J. Fish. Aquat. Sci., 56: 1801-1808. [9] Holt B. D. & Engelkemeir A. G. (1970) Thermal decomposition of barium sulfate to sulfur dioxide for mass spectrometric analysis. Anal. Chem., 42: 1451-1453. [26, 44, 45] Holt B. D., Engelkemeir A. G. & Venters A. (1972) Variations of sulfur isotope ratios in samples of water and air near Chicago. Environ. Sci. Tech., 6: 338-341. [26] Holt B. D. (1975) Determination of stable isotope ratios in the environment. Prog. Nucl. Energy Series, Anal. Chem., 12: 11-26. [26] Holt B. D., Sturchio N. C., Arehart G. B. & Bakel A. J. (1995) Ultrasonic vacuum extraction of gases from water for chemical and isotopic analysis. Chem. Geol. (Isot. Geosci. Sect.), 122: 275-284. [10] Holt B. D., Sturchio N. C., Abrajano T. A. & Heraty L. J. (1997) Conversion of chlorinated volatile organic compounds to carbon dioxide and methyl chloride for isotopic analysis of carbon and chlorine. Anal. Chem., 69: 2727-2733. [28] Honig R. E. (1945) Gas flow in the mass spectrometer. J. Appl. Phys., 16: 646-654. [38] Honma H. & Itihara Y. (1981) Distribution of ammonium in minerals of metamorphic and granitic rocks. Geochim. Cosmochim. Acta, 45: 983-988. [16] Hopple J. A., Hannon J. E. & Coplen T. B. (1998) Comparison of two stable hydrogen isotope-ratio measurement techniques on Antarctic surface-water and ice samples. Chem. Geol., 152: 321-323. [1] Horibe Y. & Kobayakawa M. (1960) Deuterium abundance of natural waters. Geochim. Cosmochim. Acta, 20: 273-283. [1, 40]
1106
References
Horibe Y. & Craig H. (1995) D / H fractionation in the system methane-hydrogen-water. Geochim. Cosmochim. Acta, 59(24): 5209-5217. [46] Horita J. & Matsuo S. (1986) Extraction and isotopic analysis of fluid inclusions in halites. Geochem. J., 20: 261-272. [1, 3] Horita J. (1988) Hydrogen isotope analysis of natural waters using an H2-water equilibration method: A special implication to brines. Chem. Geol. (Isot. Geosci. Sect.), 72: 89-94. [1, 3, 38] I-Iorita J. & Gat J. R. (1988) Procedure for the hydrogen isotope analysis of water from concentrated brines. Chem. Geol. (Isot. Geosci. Sect.), 72: 85-88. [1] Horita J. & Gat J. R. (1989) Deuterium in the Dead Sea: Remeasurement and implications for the isotopic activity correction in brines. Geochim. Cosmochim. Acta, 53: 131-133. [1] Horita J., Ueda A., Mizukami K. & Takatori I. (1989) Automatic 6D and 6180 analyses of multi-water samples using H2- and CO2-water equilibration methods with a common equilibration set-up. Appl. Radiot. Isot., 40(9): 801-805. [1, 3] Horita J. (1989a) Analytical aspects of stable isotopes in brines. Chem. Geol. (Isot. Geosci. Sect.), 79: 107112.[1] Horita J. (1989b) Stable isotope fractionation factors of water in hydrated saline mineral-brine systems. Earth Planet. Sci. Lett., 95: 173-179. [46] Horita J. (1990) Stable isotope paleoclimatology of brine inclusions in halite; modeling and application to Searles Lake, California. Geochim. Cosmochim. Acta, 54: 2059-2073. [3] Horita J., Cole D. R. & Wesolowski D. J. (1993a) The activity-composition relationship of oxygen and hydrogen isotopes in aqueous salt solutions: II. Vapor-liquid water equilibration of mixed salt solutions from 50 to 100~ and geochemical implications. Geochim. Cosmochim. Acta, 57: 4703-4711. [1, 46] Horita J., Wesolowski D. J. & Cole D. R. (1993b) The activity-composition relationship of oxygen and hydrogen isotopes in aqueous salt solutions. I. Vapor-liquid equilibration of single salt solutions from 50 to 100~ Geochim. Cosmochim. Acta, 57: 2797-2817. [1, 46] I-Iorita J. & Wesolowski D. J. (1994) Liquid-vapor fractionation of oxygen and hydrogen isotopes of water from the freezing to the critical temperature. Geochim. Cosmochim. Acta, 58: 3425-3437. [1] Horita J., Cole D. R. & Wesolowski D. J. (1995) The activity-composition relationship of oxygen and hydrogen isotopes in aqueous salt solutions: III. Vapor-liquid water equilibration of NaC1 solutions to 350~ Geochim. Cosmochim. Acta, 59(6): 1139-1151. [46] I-Iorita J., Dreisner T. & Cole D. R. (1999) Pressure effect on hydrogen isotope fractionation between brucite and water at elevated temperatures. Science, 286: 1545-1547. [46] Horita J. (2001) Carbon isotope exchange in the system CO2-CH4 at elevated temperatures. Geochim. Cosmochim. Acta, 65: 1907-1919. [46] Horita J., Cole D. R., Polyakov V. B. & Driesner T. (2002) Experimental and theoretical study of pressure effects on hydrogen isotope fractionation in the system brucite-water at elevated temperatures. Geochim. Cosmochim. Acta, 66: 3769-3788. [46] I-Iorlick G. & Montaser A. (1998) Analytical characteristics of ICPMS. In: Inductively coupled plasma mass spectrometry. A. Montaser (ed.).Wiley-VCH: 503-613. [31] Hornibrook E. R. C. & Longstaffe F. J. (1996) Berthierine from the Lower Cretaceous Clearwater formation, Alberta, Canada. Clays Clay Minerals, 44: 1-21. [2] I-Iouerou G., Kelly S. D. & Dennis M. J. (1999) Determination of the oxygen-18/oxygen-16 isotope ratios of sugar, citric acid and water from single strength orange juice. Rapid Commun. Mass Spectrom., 13: 1257-1262. [1] Hough R. M., Gilmour I., Pillinger C. T., Arden J. W., Gilkes K. W. R., Yuan J. & Milledge H. J. (1995) Diamonds and silicon carbide in impact melt rock from the Ries impact crater. Nature, 378: 41-44. [13] Houston P. L., Suits A. G. & Toumi R. (1996) Isotopic enrichment of heavy ozone in the stratosphere. J. Geophys. Res., 101: 18829-18834. [18] Houweling S., Kaminski T., Dentener F., Lelieveld J. & Heimann M. (1999) Inverse modeling of methane sources and sinks using the adjoint of a global transport model. J. Geophys. Res., 104(D21): 26173-26160.[14] Howald R. A. (1960) Ion pairs. I. Isotope effects shown by chloride solutions in glacial acetic acid. J. Am. Chem. Soc., 82: 20-24. [28]
References
1107
H s i e h J. C. C. (1997) An oxygen isotopic study of soil water and pedogenic clays in Hawaii. PhD Thesis,
California Institute of Technology, Pasadena, California, USA: 181 p. [2] Hsieh J. C. C., Savin S. M., Kelley E. F. & Chadwick O. A. (1998) Measurement of soil-water 6180 values by direct equilibration with CO2. Geoderma, 82: 255-268. [1, 2] Hsieh J. C. C. & Yapp C. J. (1999) Hydrogen isotope exchange in halloysite. Clays Clay Minerals, 47: 811816. [2] Hughes W. B., Holba A. G. & Dzou L. I. P. (1995) The ratios of dibenzothiophene to phenanthrene and pristane to phytane as indicators of depositional environment and lithology of petroleum source rocks. Geochim. Cosmochim. Acta, 59: 3581-3598. [8] Huh Y., Chan L. H., Zhang L. & Edmond J. M. (1998) Lithium and its isotopes in major world rivers: implications for weathering and the oceanic budget. Geochim. Cosmochim. Acta, 62: 2039-2052. [6] Huh Y., Chan L. H. & Edmond J. M. (2001) Lithium isotopes as a probe of weathering process: Orinoco River. Earth Planet. Sci. Lett., 194: 189-199. [6] Hulston J. R. & Thode H. G. (1965a) Cosmic-ray-produced 36S and 33S in the metallic phase of iron meteorites. J. Geophys. Res., 70(18): 4435-4442. [45] Hulston J. R. & Thode H. G. (1965b) Variations in the S33, S 34, and S36 contents of meteorites and their relation to chemical and nuclear effects. J. Geophys. Res., 70: 3475-3484. [20, 26, 45] Humbert F., Libourel G., France-Lanord C., Zimmermann L. &Marty B. (2000) CO2-1aser extractionstatic mass spectrometry analysis of ultra-low concentrations of nitrogen in silicates. Geostand. Newslett., 24: 255-260. [16, 17] Huneke J. C., Armstrong J. T. & Wasserburg G. J. (1983) Fun with Panurge: High mass resolution ion microprobe measurements of Mg in Allende inclusions. Geochim. Cosmochim. Acta, 47: 1635-1650. [25] Hunter K. L. & Gray J. W. (1993) ETH Technical Publication, TP007. [37] Hunkeler D., Hohener P., Bernasconi S. & Zeyer J. (1999) Engineered in situ bioremediation of a petroleum hydrocarbon-contaminated aquifer: assessment of mineralization based on alkalinity, inorganic carbon and stable carbon isotope balances. J. Cont. Hydrolgeol. 37: 201-223. [10] Hurai V., Simon D. & Bezik V. (1997) Contrasting chemistry and H, O, C isotope composition of greenschist-facies, Hercynian and Alpine metamorphic fluids (Western Carpathians). Chem. Geol., 136: 281-293. [3] Hurst G. S. & Payne M. G. (1988) Principles and applications of resonance ionization spectroscopy. Adam Hilger, Boston & Philadelphia. [33] Hurtgen T. M., Arthur M. A., Suits N. S. & Kaufman A. J. (2002) The sulfur isotopic composition of Neoproterozoic seawater sulfate: implications for a snowball Earth? Earth Planet. Sci. Lett., 203: 413429. [26] Husek P. (1995) Simultaneous profile analysis of plasma amino and organic-acids by capillary gaschromatography. J. Chromatogr. Biomed. Appl., 669: 352-357. [8] Husek P. & Matucha P. (1997) Simple approach for analysis of plasma oxo-, hydroxy- and dicarboxylic acids. J. Chromaogr., B 693: 499-502. [8] Husek P. (1998) Chloroformates in gas chromatography as general purpose derivatizing agents. J. Chromatogr., B 717: 57-91. [8] Hut G. (1987) Consultants' group meeting on stable isotope reference samples for geochemical and hydrological investigations. Scientific Meeting 16-18 Sept., 1985. IAEA, Vienna. Internal Report: 42 p. [1, 40, 43] Hutchinson D. A. (1954) Natural abundance of the lithium isotopes. Phys. Rev., 96: 1018-1021. [6] Huygen C. (1963) The sampling of sulfur dioxide in air with impregnated filter paper. Anal. Chim. Acta, 28: 349-360. [26] Hyeong, K. & Capuano, R. M. (2000) The effect of organic matter and the H202 organic-matter-removal method on the 6D of smectite-rich samples. Geochim. Cosmochim. Acta, 64: 3829-3837. [2] IAEA (2001) New approaches for stable isotope ratio measurements. Proceedings of an Advisory Group Meeting, held in Vienna, 20-23 Sept., 1999. IAEA, Vienna. TECDOC-1247:151 p. [34, 40] Iacumin P., Bocherens H., Mariotti A. & Longinelli A. (1996) Oxygen isotope analyses of co-existing carbonate and phosphate in biogenic apatite: a way to monitor diagenetic alteration of bone phosphate? Earth Planet. Sci. Lett., 142: 1-6. [46]
1108
References
Igumnov S. A. (1976) Sulfur isotope exchange between sulfide and sulfate in hydrothermal solutions. Geokhimiya, 4: 497-503. [46] Igumnov S. A., Grinenko V. A. & Porter N. B. (1977) Temperature dependence of 634S for H2S and dissolved sulfate over the range 260-400~ Geokhimiya, 7: 1085-1087. [46] Ihinger P. D., Hervig R. L. & McMillan P. F. (1994) Analytical methods for volatiles in glasses. In: Volatiles in magma. M. R. Carrol & J. R. Holloway (eds.). MSA Rev. Mineral., 30: 67-121. [1] Ikezoe Y., Matsuoka S., Takebe M. & Viggiano A. (1987) Gas phase ion-molecule reaction rateconstants through 1986. Maruzen, Japan. [4] Ingraham N. L. & Shadel C. (1992) A comparison of the toluene distillation and vacuum/heat methods for extracting soil water for stable isotopic analysis. J. Hydrol., 140: 371-387. [1, 2, 21] Inoue H. & Sugimura Y. (1985) Carbon isotopic fractionation during the CO2 exchange process between air and sea water under equilibrium and kinetic conditions. Geochim. Cosmochim. Acta, 49: 2453-2460. [46] IPCC (Intergovernmental Panel on Climate Change) (1990) Scientific assessment of climate change. Rept. Working Group h J.T. Houghton, G.J. Jenkins & J.J. Ephraums. Cambridge University Press, Cambridge, UK: 365 p. [14] IPCC (Intergovernmental Panel on Climate Change) (1995) Climate change 1995: The scientific basis contribution of Working Group I to the Third Assessment Report of the IPCC. J. T. Houghton, L. G. Meiro Filho, B. A. Callender, N. B. W. Harris, A. Kattenberg & K. Maskell (eds.). Cambridge University Press, Cambridge, UK: 572 p. [14] IPCC (Intergovernmental Panel on Climate Change) (2001) Climate change 2001: The scientific basis contribution of Working Group I to the Third Assessment Report of the IPCC. J. T. Houghton, Y. Ding, D. J. Griggs, M. Noguer, P. J. van der Linden & D. Xiaosu (eds.). Cambridge University Press, Cambridge, UK: 944 p. [14], 241-287. [19] Ireland T. R. (1986) Isotope systematics of refractory inclusions from carbonaceous chondrites. PhD Thesis, The Australian National University (unpublished). [30] Ireland T. R. (1995) Ion microprobe mass spectrometry: Techniques and applications in cosmochemistry, geochemistry, and geochronology. In: Advances in analytical geochemistry. M. Hyman & M. Rowe (eds.). JAI Press, Vol. 2: 1-118. [30] Irion F. W., Gunson M. R., Rinsland C. P., Yung Y. L., Abrams M. C., Chang A. Y. & Goldman A. (1996) Heavy ozone enrichments from ATMOS infrared solar spectra. Geophys. Res. Lett., 23: 2377-2380. [18] IRMM (1993) IRMM Certificate: Isotopic reference material IRMM-015. Institute for Reference Materials and Measurements, Geel, Belgium: 2 p. [40] IRMM (1997) IRMM Certificate: Isotopic reference material IRMM-016. Institute for Reference Materials and Measurements, Geel Belgium: 2 p. [40] IRMM (2001) Certificate: Isotopic reference material IRMM-011. Institute for Reference Materials and Measurements, Geel, Belgium: 2 p. [40] Irving C. S., Klein P. D., Navratil P. R. & Boutton T. W. (1986) Measurement of 13CO2/12CO2abundance by nondispersive heterodyne ratiometry as an alternative to gas isotope mass spectrometry. Anat. Chem., 58: 2172-2178. [33, 34] Ishikawa T. & Nakamura E. (1990) Suppression of boron volatilization from a hydrofluoric acid solution using a boron-mannitol complex. Anal. Chem., 62: 2612-2616. [7] Ishikawa T. & Nakamura E. (1993) Boron isotope systematics of marine sediments. Earth Planet. Sci. Lett., 117: 567-580. [7] Ishikawa T. & Nakamura E. (1994) Origin of the slab component in arc lavas from across-arc variation of B and Pb isotopes. Nature, 370: 205-208. [7] ISO Guide 35 (1989) Certification of reference materials - General and statistical principles. International Organization for Standardization, Geneva: 32 p. [43] ISO (1992) Quantities and units, Part 0: General principles, subclause 2.3.3, ISO 31-0. International Organization for Standardization, Geneva: 21 p. [43] ISO Guide 33 (1998) Uses of Certified Reference Materials. Revision of ISO Guide 33 (1989). International Organisation for Standardisation, Geneva. [40] ISO (1999) General requirements for the competence of testing and calibration laboratories. International Organization for Standardization, Geneva. IEC 17025. [42]
References
1109
ISO Guide 31 (2000a) Reference materials - Contents of certificates and labels. International Organization for Standardization, Geneva: 7 p. [43] ISO Guide 33 (2000b) Uses of certified reference materials. International Organization for Standardization, Geneva: 23 p. [43] ISO Guide 34 (2000c) General requirements for the competence of reference material producers. International Organization of Standardization, Geneva: 22 p. [40, 43] ISO/BIPM (1995) Guide to the Expression of Uncertainty in Measurement. International Organisation for Standardisation, Geneva, Switzerland: 110 p. [40, 43] ISO/REMCO (1998) Reference Materials - Contents of Certificates and Labels. ISO/REMCO No. 488, Apr. 1998, Revision of ISO Guide 31. International Organisation for Standardisation, Geneva. [40] Itoh S. & Hasegawa R. (1998) Measurements of isotope ratios of boron and silicon by glow-discharge mass-spectrometry using a small quantity of sample. J. Japan Inst. Metals, 62: 289-292. [35] IUPAC-CAWIA (1997) Reporting of relative sulfur isotope-ratio data. Pure Appl. Chem., 69: 293-295. [40] Iyengar V. & Wolf W. (1998) Global activities in hte world of reference materials including the needs of developing countries. Fresenius J. Anal. Chem., 360: 282-286. [43] Jackson C. H. N. (1933) On the true density of tsetse flies. J. Animal Ecol., 2: 204-209. [37] Jackson D. H., Mattey D. P. & Harris N. B. W. (1988) Carbon isotope compositions of fluid inclusions in charnockites from southern India. Nature, 333: 167-170. [3] Jackson M. C. & Young W. A. P. (1973) A capacitive integration system for the precise measurement of isotopic ratios in a mass spectrometer. Rev. Sci. Instrum., 44: 32-34. [39] Jackson M. L., Savin M. & Clayton R. N. (1976) Hexafluorosilicic acid reagent modification for quartz isolation. Soil Sci. Soc. Am. J., 40: 958-960. [2] Jackson M. L. (1979) Soil chemical analysis- advanced course. Madison, Wisconsin. Published by the author. 2nd edition. [2] Jackson S. E., Griffin W. L. & Pearson N. J. (1999b) Determination of high-precision isotopic ratios by laser ablation MC-ICP-MS. Ninth Annual V.M. Goldschmidt Conference, Abstr. #7361. [31] Jackson S. E. & Ryde N. (2001) The application of Nd:YAG lasers in LA-ICP-MS. In: Laser-AblationICPMS in the Earth Sciences. P. Sylvester (ed.). Mineral. Assoc. of Canada. Short Course Vol. Series, Vol. 29, Chapter 3: 29-45. [20] Jackson S. E. & Giinther, D. (2003) The nature and sources of laser induced isotopic fractionation in laser ablation-multicollector-inductively coupled plasma-masspectrometry. J. Anal. At. Spectrom., 18: 205-212. [20] Jackson S. M., Morgan G. H., Morse A. D., Butterworth A. L. & Pillinger C. T. (1999a) The use of static mass spectrometry to determine the combined stable isotopic composition of small samples of atmospheric methane. Rapid Commun. Mass Spectrom., 13: 1329-1333. [36] Jackson T. A. (2001) Variations in isotope composition of mercury in a freshwater sediment sequence and food web. Can. J. Fish. Aquat. Sci., 58: 185-196. [31] Jacob E. & Christie K. O. (1977) A simple method for the purification of fluorine. J. Fluorine Chem., 10: 169-172. [20] James B. R. (1994) Hexavalent chromium solubility and reduction in alakaline soils enriched with chromite ore processing residue. J. Environ. Qual., 23: 227-233. [29] James R. H., Rudnicki M. D. & Palmer M. R. (1999) The alkali element and boron geochemistry of the Escanaba trough sediment-hosted hydrothermal system. Earth Planet. Sci. Lett., 171: 157-169. [6] James R. H. & Palmer M. R. (2000) The lithium isotope composition of international rock standards. Chem. Geol., 166: 319-326. [6] Jameson R. F. (1959) The composition of the "strong" phosphoric acids. J. Chem. Soc., 1: 752-759. [12] Jamin E., Naulet N. & Martin G. J. (1997) Multi-element and multi-site isotopic analysis of nicotine from tobacco leaves. Plant Celt Environ., 20: 589-599. [5] Jancso G. & Van Hook W. A. (1974) Condensed phase isotope effects. Chem. Rev., 74: 689-710. [4] Janghorbani M., Davis T. A. & Ting B. T. G. (1988) Measurement of stable isotopes of bromine in biological fluids with inductively coupled plasma mass spectrometry. Analyst, 113: 405-411. [28] Janssen C., Guenther J., Krankowski D. & Mauersberger K. (1999) Relative formation rates of 5003 and 5203 in 160-180 mixtures. J. Chem. Phys., 111: 7179-7182. [18]
1110
References
Jaques A. L., Hall A. E., Sheraton J. W., Smith C. B., Sun S.-S., Drew R. M., Foudoulis C. & Ellingsen K. (1989) Comparison of crystalline inclusions and C-isotopic composition of Argyle and Ellendale diamonds. In: Kimberlites and related rocks. J. Ross (ed.). Blackwell (Geol. Soc. Austr. Spec. Publ., 14): 966-989. [13] Jarman W. M., Hilkert A., Bacon C. E., Collister J. W., Ballschmiter K. & Risebrough R. W. (1998) Compound-specific carbon isotopic analysis of arochlors, clophens, kaneclors, and phenoclors. Environm. Sci. Technol., 32: 833-836. [9] Jarvis K. E., Gray A. L. & Houk R. S. (1992) Handbook of inductively coupled plasma mass spectrometry. Blackie and Son - Chapman and Hall (USA): 380 p. [31] Jasper J. P. (2001) Quantitative estimates of precision for molecular isotopic measurements. Rapid Commun. Mass Spectrom., 15: 1554-1557. [34] Javoy M. & Fourcade S. (1973) 160/180 ratios in lunar fines and rocks. Earth Planet. Sci. Lett., 211: 377382. [20] Javoy M., Marette J. & Pineau F. (1973) 160/180 ratios in Luna 16 fines. Geochim. Cosmochim. Acta, 37: 2017-2019. [20] Javoy M., Pineau F. & Iiyama I. (1978) Experimental determination of the isotopic fractionation between gaseous CO2 and carbon dissolved in tholeiitic magma: apreliminary study. Contrib. Mineral. Petrol., 67: 35-39. [46] Javoy M., Pineau F. & Delorme H. (1986) Carbon and nitrogen isotopes in the mantle. Chem. Geol., 57: 41-62. [16] Jenden P. D., Kaplan I. R., Poreda R. J. & Craig H. (1988) Origin of nitrogen-rich natural gases in the California Great Valley. Evidence from helium, carbon and nitrogen isotope ratios. Geochim. Cosmochim. Acta, 52: 851-861. [16] Jendrzejewski N., Eggenkamp H. M. G. & Coleman M. L. (1997) Sequential determination of chlorine and carbon isotopic composition in single microlitre samples of chlorinates solvent. Anal. Chem., 69: 4259-4266. [28] Jenkin G. R. T., Fallick A. E. & Leake B. E. (1992) A stable isotope study of retrograde alteration in SW Connemara, Ireland. Contrib. Mineral. Petrol., 110: 269-288. [3] Jenkin G. R. T., Craw D. & Fallick A. E. (1994a) Stable isotopic and fluid inclusion evidence for meteoric fluid penetration into an active mountain belt; Alpine Schist, New Zealand. J. Metam. Geol., 12: 429444. [3] Jenkin G. R. T., Farrow C. M., Fallick A. E. & I-Iiggins D. (1994b) Oxygen isotope exchange and closure temperatures in cooling rocks. J. Metamorphic Geol., 12: 221-235. [46] Jensen E. S. (1991) Evaluation of automated analysis of 15N and total N in plant material and soil. Plant Soil, 133: 83-92. [15] Jensen M. L. & Nakai N. (1962) Sulfur isotope meteorite standards results and recommendations. In: Biogeochemistry of sulfur isotopes. M. L. Jensen (ed.). Yale University, New Haven, Connecticut. Nat. Sci. Found. Symp. Vol.: 30-35. [26, 40] Jeong D.-Y., Rhee Y.-J., Lee J. & Rhee B. K. (1998) CW laser-initiated resonance ionization mass spectrometry of Gd atoms. Resonance Ionization Spectroscopy, Ninth International Symposium. J.C. Vickerman, I. Lyon, N.P. Lockyer & J.E. Parks (eds.). AIP Press Woodbury, NY. AIP Conference Proceedings 454: 289-292. [33] Jia Y. J. & Kerrich R. (1999) Nitrogen isotope systematics of mesothermal lode gold deposits: metamorphic, granitic, meteoric water, or mantle origin? Geology, 27: 1051-1054. [16] Jiang S., Ding T. & Wan D. (1992) A study on silicon and oxygen isotope geochemistry of some leadzinc ore depositis in China. Chinese Sci. Bull., 37(12): 1022-1027. [25] Jiang S., Ding T., Wan D. & Li Y. (1993) Silicon isotopic composiitons of Archean banded Si-Fe formation (BIF) in the Gongchangling ore deposit, Liaoning, China. Chinese Sci. Bull., 36: 482-489.
[25]
Jiang S.-Y., Palmer M. R., Ding T.-P. & Wan, D.-F. (1994) Silicon isotope geochemistry of the Sullivan Pb-Zn deposit, Canada: A preliminary study. Econ. Geol., 89: 1623-1629. [20] Johnson B. J., Fogel M. L. & Miller G. H. (1998) Stable isotopes in modern ostrich eggshell: a calibration for paleoenvironmental applications in semi-arid regions of southern Africa. Geochim. Cosmochim. Acta, 62: 2451-2461. [9, 46]
References
1111
Johnson C. M. & Nishita H. (1952) Microestimation of sulfur in plant materials, soils, and irrigation waters. Anal. Chem., 24: 736-742. [26] Johnson C. M. & Ulrich A. (1959) Analytical methods for use in plant analysis. Calif. Agric. Exp. Stn. Bull., 766: 25-78. [26] Johnson C. M. & Beard B. L. (1999) Correction of instrumentally produced mass fractionation during isotopic analysis of Fe by thermal ionization mass spectrometry. Int. J. Mass Spectrom., 193(1): 87-99. [29, 31] Johnson D. G., Jucks K. W., Traub W. A. & Chance K. V. (2000a) Isotopic composition of stratospheric ozone. J. Geophys. Res., 105(D7): 9025-9031. [18] Johnson R. C. & Tieszen L. L. (1994) Variation for water-use efficiency in alfalfa germplasm. Crop Sci., 34: 452-458. [9] Johnson T. C., Bonrud P. A., Dosch T. L., Kilness A. W., Senger K. A., Busch D. C. & Meyer M. R. (1987) Fatal outcome of methemoglobinemia in an infant. J. Am. Med. Assoc., 257: 2796-2797. [15] Johnson T. M., Herbel M. J., Bullen T. D. & Zawislanski P. T. (1999) Selenium isotope ratios as indicators of selenium sources and oxyanion reduction. Geochim. Cosmochim. Acta, 63(18): 2775-2783. [20, 29, 30, 46] Johnson T. M., Bullen T. D. & Zawislanski P. T. (2000b) Selenium stable isotope ratios as indicators of sources and cycling of selenium: Results from the Northern Reach of San Francisco. Environ. Sci. Technol., 34: 2075-2079. [30] Johnson T. M. & Bullen T. D. (2003) Selenium isotope fractionation during reduction by Fe(II)-Fe(III) hydroxide-sulfate (green rust). Geochim. Cosmochim. Acta, 67: 413-419. [46] Johnson T. M. (in press) A review of mass-dependent fractionation of selenium isotopes and implications for other "heavy stable isotopes". Chem. Geol. [29] Johnston A. M., Scrimgeour C. M., Henry M. O. & Handley L. L. (1999) Isolation of NO3--N as 1-phenylazo-2-naphthol (Sudan-I) for measurement of 615N. Rapid Commun. Mass Spectrom., 13: 15311534. [15] Johnston J. C. & Thiemens M. H. (1997) The isotopic composition of troposheric ozone in three environments. J. Geophys. Res., 102(D21): 25395-25404. [18] Johnston J. C., Ri~ckmann T. & Brenninkmeijer C. A. M. (2000) CO2 + O(1D) isotopic exchange: Laboratory and modeling studies. J. Geophys. Res., 105: 15213-15229. [18] Johnston W. H. & Arnold J. R. (1953) The existence of multiple charged molecular ions of HBr, HC1, and DC1. J. Chem. Phys., 21: 1499-1502. [28] Jones A. M., Iacumin P. & Young E. D. (1999) High-resolution 6180 analysis tooth enamel phosphate of by isotope ratio monitoring gas chromatography mass spectrometry and ultraviolet laser fluorination. Chem. Geol., 153: 241-248. [20, 22] Jones B. J. J. & Steyn w. J. A. (1973) Sampling, handling, and analyzing plant and tissue samples. In: Soil testing and plant analysis. L. W. Walsh & J. D. Beaton (eds.). Soil Science Society of America, Madison, Wisconsin, USA: 249-270. [26] Jones D. M., Carter J. F., Eglinton G., Jumeau E. J. & Fenwick C. S. (1991) Determination of 613C values of sedimentary straight chain and cyclic alcohols by gas chromatography/isotope mass spectrometry. Biol. Mass Spectrom., 20: 641-646. [9] Jones R. I., Grey J., Sleep D. & Quarmby C. (1998) An essessment, using stable isotopes, of the importance of allochthonous organic carbon sources to the pelagic food web in Loch Ness. Proc. R. Soc. London, B 265: 105-111. [9] Jones R. L. & Pyle J. A. (1984) Observations of CH4 and N20 by the Nimbus-7 SAMS: a comparison with in situ data and two-dimensional numerical model calculations. J. Geophys. Res., 89(D4): 5263-5279. [14] Jovanovic S. & Reed G.W. Jr. (1976) 196Hg and 202Hg isotopic ratios in chondrites: revisited. Earth Planet. Sci. Lett., 31: 95-100. [31] Juchelka D., Beck T., Hener U., Dettmar F. & Mosandl A. (1998) Multidimensional gas chromatography coupled on-line with isotope ratio mass spectrometry (MDGC-IRMS): Progress in the analytical authentication of genuine flavor components. HRC-J. High Resolut. Chromatogr., 21: 145151. [8]
1112
References
Juillet-Leclerc A. & Labeyrie L. D. (1987) Temperature dependence of the oxygen isotopic fractionation between diatom silica and water. Earth Planet. Sci. Lett., 84: 69-74. [2] Junge F., Seltmann R. & Stiehl G. (1989) Nitrogen isotope characteristics of breccias, granitoids, and greisens from eastern Erzgebirge tin ore deposits, Sadisdorf: Altenberg, GDR. Proc. 5th Working Meeting, Isotopes in Nature: 321-332. [16] Junk G. & Svec H. J. (1958) The absolute abundance of the nitrogen isotopes in the atmosphere and compressed gas from various sources. Geochim. Cosmochim. Acta, 14: 234-243. 15,[40, 43] Jusserand C. (1980) Extraction de l'eau interstitielle des s6diments et des sols - Comparison des valeurs de l'oxyg6ne 18 par diff6rentes m6thodes, premiers r6sultats. Catena, 7: 87-96. [2] Kakihana I-I., Kotaka M., Satoh S., Nomura M. & Okamoto M. (1977) Fundamental studies on the ionexchange separation of boron isotopes. Bull. Chem. Soc. Jap., 50: 158-163. [7] Kakiuchi M. (1994) Temperature dependence of fractionation of hydrogen isotopes in aqueous sodium chloride solutions. J. Sol. Chem., 23: 1073-1087. [46] Kampschulte A., Bruckschen P. & Strauss H. (2001) The sulphur isotopic composition of trace sulphates in Carboniferous brachiopods: implications for coeval seawater, correlation with other geochemical cycles and isotope stratigraphy. Chem. Geol., 175: 165-189. [26] Kaplan I. R., Emery K. O. & Rittenberg S. C. (1963) The distribution and isotopic abundance of sulphur in recent marine sediments off Southern California. Geochim. Cosmochim. Acta, 27: 297-331. [29] Kaplan I. R. & Rittenberg S. C. (1964) Microbiological fractionation of sulfur isotopes. J. Gen. Microbiol., 34: 195-212. [46] Kaplan I. R. (1983) Stable isotopes of sulfur, nitrogen and deuterium in recent marine environments. In Stable isotopes in sedimentary geology. M. A. Arthur, T. F. Anderson, I. R. Kaplan, J. Veizer & L. S. Land (eds.). Society of Sedimentary Geology. Vol. 10: 2.1-2.108. [29] Karasek F. W. & Clement R. E. (1988) Basic gas chromatography - mass spectrometry. Elsevier, Amsterdam. [4] Karhu J. & Epstein S. (1986) The implication of the oxygen isotope record in coexisting cherts and phosphates. Geochim. Cosmochim. Acta, 50: 1745-1756. [20, 22] Karhu J. A. (1997) Catalytic reduction of water to hydrogen for isotopic analysis using zinc containing traces of sodium. Anat. Chem., 69: 4728-4730. [1] Karlsson H. R. & Clayton R. N. (1990) Oxygen isotope fractionation between analcime and water: An experimental study. Geochim. Cosmochim. Acta, 54: 1359-1368. [36, 46] Karlsson H. R. (1995) Application of oxygen and hydrogen isotopes top zeolites. In: Natural zeolites '93: Occurence, properties, use. D. W. Ming & F. A. Mumpton (eds.). Int. Comm. Nat. Zeol., NY: 125-140. [36] Kasemann S., Meixner A., Rocholl A., Vennemann T. W., Rosner M., Schmitt A. K. & Wiedenbeck M. (2001) Boron and oxygen isotopic composition of certified reference materials NIST SRM-610/612 and reference materials JB-2 and JR-2. Geostand. Newslett., 25: 405-416. [40] Kator H., Morris L. J., Wetzel R. L. & Koepfler E. T. (1992) A rapid chromatographic method for recovery of 15NO2- and NO3- produced by nitrification in aqueous samples. Limnol. Oceanogr., 39(4): 900907. [15] Katz S. A. & Salem H. (1994) The biological and environmental chemistry of chromium. VCH Publishers. [29] Katzenberg M. A. & Krouse H. R. (1989) Application of stable isotopes in tissues to problems in identification. Cdn. Soc. Forensic Sci. J., 22: 7-19. [26] Kaufmann R. S., Long A., Bentley H. & Davis S. (1983) Application of chloride stable isotope analyses to hydrogeology. Proc. 1983 meet. Amer. Water Res., Ariz. & Ariz.-Nev. Acad. Sci., Hydr. sect. Hydr. Water Res. Ariz. Southwest.: 85-90. [28] Kaufmann R. S. (1984) Chlorine in groundwater: Stable isotope distribution. PhD Thesis, University of Arizona, Tucson, Arizona, USA: 137 p. [28] Kaufmann R. S. & Long A. (1984) Stable chlorine ratio as a ground-water tracer. EOS Trans. AGU, 65: 886. [28] Kaufmann R. S., Long A., Davis S. & Bentley H. (1984a) Natural variations of chlorine stable isotopes. Geol. Soc. Am. Abstr. w. Prog., 16: 556. [28]
References
1113
Kaufmann R. S., Long A., Bentley H. & Davis S. (1984b) Natural chlorine isotope variations. Nature, 309: 338-340. [28] Kaufmann R. S. & Arn6rsson S. (1986) 37C1/35C1 ratios in Icelandic geothermal waters. Proc. 5th International Symposium on Water-Rock Interaction: 325-327. [28] Kaufmann R. S., Frape S., Fritz P. & Bentley H. (1987) Chlorine stable isotope composition of Canadian Shield brines. In: Saline water and gases in crystalline rocks. P. Fritz & S. K. Frape (eds.). Geol. Ass. Can. Spec. Paper, Vol. 33: 89-93. [28] Kaufmann R. S., Long A. & Campbell D. J. (1988) Chlorine isotope distribution in formation waters, Texas and Louisiana. AAPG Bull., 72: 839-844. [28, 40] Kaufmann R. S. (1989) Equilibrium exchange models for chlorine stable isotope fractionation in high temperature environments. Proc. 6th International Symposium on Water-Rock Interaction: 365-368. [281 Kaufmann R. S., Frape S. K., McNutt R. & Eastoe C. J. (1992) Chlorine stable isotope distribution of Michigan Basin Shield formation waters. Proc. 7th International Symposium on Water-Rock Interaction: 943-946. [28] Kaufmann R. S., Frape S. K., McNutt R. & Eastoe C. J. (1993) Chlorine stable isotope distribution of Michigan basin formation waters. Appl. Geochem., 8: 403-407. [28] Kaye J. A. & Strobel D. F. (1983) Enhancement of heavy ozone in the earth's atmosphere? J. Geophys. Res., 88: 8447-8452. [18] Kaye J. A. (1986) Theoretical analysis of isotope effects on ozone formation in oxygen photochemistry. J. Geophys. Res., 91: 7865-7874. [18] Kazahaya K. & Matsuo S. (1985) A new ball-milling method for extraction and fluid inclusions from minerals. Geochem. J., 19: 45-54. [1, 3] K/irger J. & Ruthven D. M. (1992) Diffusion in zeolites and other microporous solids. Wiley, NY: 605 p. [36] Keeling C. D., Bacastow R. B., Carter A. F., Piper S. C., Whorf T. P., Heimann M., Mook W. G. & Roeloffzen H. (1989) A three-dimensional model of atmospheric CO2 transport based on observed winds: 1. Analysis of observational data. In: Aspects of climate variability in the Pacific and the Western Americas. D. H. Peterson (ed.). Am. Geophys. Union, Geophys. Monogr. 55: 165-236. [14] Keeling C. D., Whorf T. P., Wahlen M. & van der Plicht J. (1995) Interannual extremes in the rate of rise of atmospheric carbon dioxide since 1980. Nature, 375: 666-670. [14] Keene W. C., Pszenny A. A. P., Galloway J. N. & Hawley M. E. (1986) Sea-salt corrections and interpretation of constituent ratios in marine precipitation. J. Geophys. Res., D91: 6647-6658. [26] Keeney D. R. & Nelson D. W. (1982) Nitrogen- Inorganic forms. In: Methods of soil analysis, Part 2. Chemical and microbiological proper. A. L. Page, R. H. Miller & D. R. Keeney (eds.). American Society of Agronomy Inc. - Soil Science Society of America Inc., Madison, Wisconsin. Agronomy Series, 2nd edition, No. 9: 643-698. [15] Keller J. & Krafft M. (1990) Effusive natrocarbonatite activity of Oldoinyo Lengai, June 1988. Bull. Volcanol., 52: 629-645. [28] Kelley D. B. & Friih-Green G. L. (2001) Volatile lines of descent in submarine plutonic environments: Insights from stable isotope and fluid inclusion analyses. Geochim. Cosmochim. Acta, 65(19): 33253346. [3] Kelley K. R., Ditsch D. C. & Alley M. M. (1991) Diffusion and automated nitrogen-15 analysis of lowmass ammonium samples. Soil Sci. Soc. Am. J., 55: 1016-1020. [15] Kelley S. P. & Fallick A. E. (1990) High precision spatially resolved analysis of 634S in sulphides using a laser extraction technique. Geochim. Cosmochim. Acta, 54: 883-888. [20, 26, 45] Kelly J. F. (1999) Stable isotopes of carbon and nitrogen in the study of avian and mammalian ecology. Can. J. Zoology, 78: 1-27. [9] Kelly S. D., Parker I. G., Sharman M. & Dennis M. J. (1998) On-line quantitative determination of 2H/ 1H isotope ratios in organic and water samples using an elemental analyser coupled to an isotope ratio mass spectrometer. J. Mass Spectrom., 33: 735-738. [1] Kemp A. L. W. & Thode H. G. (1968) The mechanism of the bacterial reduction of sulphate and of sulphite from isotope fractionation studies. Geochim. Cosmochim. Acta, 32: 71-91. [46]
1114
References
Kendall C. & Coplen T. B. (1985) Multisample conversion of water to hydrogen by zinc for stable isotope determination. Anal. Chem., 57: 1437-1440. [1, 21] Kendall C. & Grim E. (1990) Combustion tube method for measurement of nitrogen isotope ratios using calcium oxide for total removal of carbon dioxide and water. Anal. Chem., 62: 526-529. [15, 16, 40] Kendall C., Campbell D. H., Burns D. A., Shanley J. B., Silva S. R. & Chang C. C. (1995a) Tracing sources of nitrate in snowmelt runoff using the oxygen and nitrogen isotopic compositions of nitrate. Proceedings of a Boulder Symposium, IAHS Publ. No. 228. [15] Kendall C., Sklash M. G. & Bullen T. D. (1995b) Isotope tracers of water and solute sources in catchments. In: Solute modelling in catchment systems. S. T. Trudgill (ed.). John Wiley & Sons Ltd, NY: 261-303. [1 ] Kendall C. (1998) Tracing nitrogen sources and cycling in catchments. In: Isotope tracers in catchment hydrology. C. Kendall & J. J. McDonnell (eds.). Elsevier Science B.V, Amsterdam: 534-569. [15] Kerrich R., Beckinsale R. D. & Shackleton N. J. (1978) The physical and hydrothermal regime of tectonic vein systems; evidence from stable isotope and fluid inclusion studies. N. Jhb. Miner. Abhand., 131: 225-239. [3] Kerridge J. F. (1985) Carbon, hydrogen, and nitrogen in carbonaceous chondrites: Abundances and isotopic compositions in bulk samples. Geochim. Cosmochim. Acta, 49: 1707-1714. [16] Kerridge J. F. (1993) Long-term compositional variation in solar corpuscular radiation: Evidence from nitrogen isotopes in the lunar regolith. Rev. Geophys., 31: 423-437. [17] Kerstel E. R. T., van Trigt R., Dam N., Reuss J. & Meijer H. A. J. (1999) Simultaneous determination of the 2H/1H, 170/160, and 180/160 isotope abundance ratios in water by means of laser spectrometry. Anal. Chem., 71: 5297-5303. [34] Kerstel E. R. T., van Trigt R., Dam N., Reuss J. & Meijer H. A. J. (2001) Laser spectrometry applied to the simultaneous determination of the 62H, 6170, and 6180 isotope abundances in water. In: New approaches for stable isotope ratio measurements. Proceedings of an Advisory Group meeting held in Vienna, 20-23 September 1999. IAEA, Vienna. IAEA-TECDOC-1247: 7-13. [34, 40] Kerstel E. R. T. (2001) O-17 analysis of Ontario Hydro heavy water. Center for Isotope Research, University of Groningen, Groningen, the Netherlands. Customer Report No. 2001-32. [34] Kerstel E. R. T., Gagliardi G., Gianfrani L., Meijer H. A. J., van Trigt R. & Ramaker R. (2002) Determination of the 2H/1H, 170/160, and 180/160 isotope ratios in water by means of tunable diode laser spectroscopy at 1.39 ~m. Spectrochim. Acta, A 58: 2389-2396. [34] Kester C. L., Rye R. O., Johnson C. A., Schwartz C. & Holmes C. (2001) On-line sulfur isotope analysis of organic material by direct combustion: preliminary results and potential applications. Isot. Environ. Health Stud., 37: 53-65. Ketterer M. E., Peters M. J. & Tisdale P. J. (1991) Verification of a correction procedure for measurement of lead isotope ratios by inductively coupled plasma mass spectrometry. J. Anal. At. Spectrom., 6: 439443. [35] Khalfallah Y., Normand S., Tissot S., Pachiaudi C., Beylot M. & Riou J. P. (1993) Determination of 613C values of plasma lactate by gas chromatography/isotope ratio mass-spectrometry. Biol. Mass Spectrom., 22: 707-711. [8] Khan S. A., Mulvaney R. L. & Mulvaney C. S. (1997) Accelerated diffusion methods of inorganicnitrogen analysis of soil extracts and water. Soil Sci. Soc. Am. J., 61: 936-942. [15] Khan S. A., Mulvaney R. L. & Brooks P. D. (1998) Diffusion methods for automated nitrogen-15 analysis using acidified disks. Soil Sci. Soc. Am. J., 62: 406-412. [15] Khan S. A., Mulvaney R. L. & Hoeft R. G. (2000) Direct-diffusion methods for inorganic-nitrogen analysis of soil. Soil Sci. Soc. Am. J., 64: 1083-1089. [15] Kiba T., Tagaki T., Yoshimura Y. & Kishi I. (1955) Tin (II)-strong phosphoric acid. A new reagent for the determination of sulfate by reduction to hydrogen sulfide. Bull. Chem. Soc. Japan, 28: 641-644. [26, 27] Kiba T., Akaza I. & Sugishita N. (1957) Rapid determination of inorganic sulfur in various forms, particularly in sulfide ores, by the tin(II)-strong phosphoric acid reduction method. Bull. Chem. Soc. Japan, 30: 972-975. [26]
References
1115
Kieffer S. W. (1982) Thermodynamics and lattice vibrations of minerals: 5. Applications to phase equilibria, isotopic fractionation, and high-pressure thermodynamic properties. Rev. Geophys. Space Phys., 20: 827-849. [46] Kilmer V. J. & Nearpass D. C. (1960) The determination of available sulfur in soils. Soil Sci. Soc. Am. Proc., 24: 337-340. [26] Kim C.-K., Seki R., Morita S., Yamasaki S., Tsumura A., Takaku Y., Igarashi Y. & Yamamoto M. (1991) Application of a high resolution inductively coupled plasma mass spectrometer to the measurement of long-lived radionuclides. J. Anal. At. Spectrom., 6: 205-209. [35] Kim K.-R. & Craig H. (1990) Two-isotope characterization of N20 in the Pacific Ocean and constraints on its origin in deep water. Nature, 347: 58-61. [19] Kim K.-R. & Craig H. (1993) Nitrogen-15 and oxygen-18 characteristics of nitrous oxide: A global perspective. Science, 262: 1855-1857. [19] Kim S. A. & O'Neil J. R. (1997) Equilibrium and nonequilibrium oxygen isotope effects in synthetic carbonates. Geochim. Cosmochim. Acta, 61: 3461-3475. [46] Kindness A. & Marr I. L. (1997) Improved infrared spectroscopic method for the measurement of 13C:12C ratios. Appl. Spectrosc., 51(1): 17-21. [34] King E. M., Valley J. W., Davis D. W. & Kowallis B. J. (2001) Empirical determination of oxygen isotope fractionation factors for titanite with respect to zircon and quartz. Geochim. Cosmochim. Acta, 65: 3165-3175. [46] King F. L. & Harrison W. W. (1990) Glow discharge mass spectrometry. An introduction to the technique and its utility. Mass Spec. Rev., 9: 285-317. [35] King F. L. & Harrison W. W. (1993) Glow discharge mass spectrometry. In" Glow discharge spectroscopies. R. K. Marcus (ed.). Plenum Press, NY: 175-214. [35] King S. L., Quay P. D. & Lansdown J. M. (1989) The C-13/C-12 kinetic isotope effect for soil oxidation of methane at ambient atmospheric concentrations. J. Geophys. Res., 94(D15): 18273-18277. [14] Kipphardt H., Valkiers S., Taylor P. D. P. & De Bi6vre P. (2000) "Calibration" in isotopic measurements. Int. J. Mass Spectrom., 198: 71-76. [42] Kipphardt H., Dudzus T., Meier K. A., Recknagel S., Hedrich M. & Matschat R. (2002) Measurement of oxygen and nitrogen in high purity metals used as national standards for element analysis in Germany by classical carrier gas hot extraction (HE) and HE after activation with photons. Mater. Trans. (JIM), 43: 98-100. [42] Kirschner D. L. & Sharp Z. D. (1997) Oxygen isotope analyses of fine-grained minerals and rocks using the laser-extraction technique. Chem. Geol., 137: 109-115. [20] Kirshenbaum I., Smith J. S., Crowell T., Graft J. & McKee R. (1947) Separation of the nitrogen isotopes by the exchange reaction between ammonia and solutions of ammonium nitrate. J. Chem. Phys., 15: 440-446. [15] Kirshenbaum I. (1951) Isotopic analysis of heavy water by the mass spectrometer. In: Physical properties and analysis of heavy water. McGraw-Hill Book Comp. Inc. [1] Kiser R. W. (1965) Introduction to mass spectrometry and its applications. Prentice Hall, NY. [38] Kishima N. & Sakai H. (1980) Oxygen-18 and Deuterium determination on a single water sample of a few milligrams. Anal. Chem., 52: 356-358. [1, 3, 20] Kita I. (1981) A new type of ball mill made of Pyrex glass. Geochem. J., 15: 289-291. [3] Klaue B. & Blum J. D. (2000) Mercury isotopic analyses by single- and multi-collector magnetic sector inductively coupled plasma mass spectrometry. J. Conf. Abstr., 5:591 (Abstract). [31] Klein R. T., Lohmann K. C. & Thayer C. W. (1996) Sr / Ca and 13C / 12C ratios in skeletal calcite of Mytilus trossulus: covariation with metabolic rate, salinity, and carbon isotopic composition of seawater. Geochim. Cosmochim. Acta, 60: 4207-4221. [46] Klemm A. & Lund6n A. (1955) Isotopenanreicherung beim Chlor durch electrolytische Uberffihrung in geschmolzenem Bleichlorid. Z. Naturforsch., 10a: 282-284. [28] Klossa B., Pierre A. & Minster J. F. (1981) Mesures de la composition isotopique du lithium dans les inclusions. Earth Planet. Sci. Lett., 52: 25-30. [6] Kli~tzli U. S. (1992) Negative thermal ionization mass spectrometry: a new approach to boron isotope geochemistry. Chem. geol. (Isot. Geosci. Sect.), 101: 111-122. [7]
1116
References
Knauth L. P. & Epstein S. (1975) Hydrogen and oxygen isotope ratios in silica from the Joides deep drilling project. Earth Planet. Sci. Lett., 25: 1-10. [46] Knauth L. P. & Epstein S. (1976) Hydrogen and oxygen isotope ratios in nodular and bedded cherts. Geochim. Cosmochim. Acta, 40: 1095-1108. [46] Knauth L. P. & Beeunas M. A. (1986) Isotope geochemistry of fluid inclusions in Permian halite with implications for the isotopic history of ocean water and the origin of saline formation waters. Geochim. Cosmochim. Acta, 50: 419-433. [3] Knoll G. F. (2000) Chapter 4: General properties of Radiation detectors. In: Radiation detection and measurement. Wiley & Sons, New York. 3rd edition: 802 p. [32] Knowles R. & Blackburn H. T. (1993) Nitrogen isotope techniques. Academic Press Inc., San Diego: 311 p. [9, 15] Koch P. L., Fogel M. L. & Tuross N. (1994) Tracing the diets of fossil animals using stable isotopes. In: Stable isotopes in ecology and environmental science, methods in ecology. K. Lajtha & R.H. Michener (eds.). Blackwell Scientific Publications, Oxford, UK: 63-92. [9] Koch P. L., Heisinger J., Moss C., Carlson R. W., Fogel M. L. & Behrensmeyer A. K. (1995) Isotopic tracking of change in diet and habitat use in African elephants. Science, 267: 1340-1343. [9] Koehler G. D., Chipley D. & Kyser T. K. (1991) Measurement of the hydrogen and oxygen isotopic compositions of concentrated chloride brines from fluid inclusions in halite. Chem. Geol. (Isot. Geosci. Sect.), 94: 45-54. [1] Koehler G., Wassenaar L. I. & Hendry M. J. (2000) An automated technique for measuring 6D and 6180 values of porewater by direct CO2 and H2 equilibration. Anal. Chem., 72: 5659-5664. [1, 2] Kohlmuller D. & Kochen W. (1993) Is n-pentane really an index of lipid-peroxidation in humans and animals - a methodological reevaluation. Anal. Biochem., 210: 268-276. [8] Kohn M. J., Valley J. W., Elsenheimer D. & Spicuzza M. J. (1993) O isotope zoning in garnet and staurolite. Evidence for closed-system mineral growth during metamorphism. Am. Mineral., 78: 9881001. [20] Kohn M. J. (1996) Predicting animal 6180: accounting for diet and physiological adaptation. Geochim. Cosmochim. Acta, 60: 4811-4829. [20] Kohn M. J., Schoeninger M. J. & Valley J. W. (1998) Variability in oxygen isotope compositions of herbivore teeth: reflections of seasonality or development physiology? Chem. Geol., 152: 97-112. [20, 22] Kohn M. J. & Valley J. W. (1998) Obtaining equilibrium oxygen isotope fractionations from rocks: theory and examples. Contrib. Mineral. Petrol., 132: 209-224. [46] Kohnen M. E. L. (1988) Stabiele chloorisotopen onderzoek. University of Utrecht, Internal Report: 17 p.
[28]
Koirtyohann S. R. (1994) Precise determination of isotopic ratios for some biologically significant elements by inductively coupled plasma mass spectroscopy. Spectrochim. Acta, B 49: 1305-1311. [32] Koletzko S., Haisch M., Seeboth I., Braden B., Hengels K., Koletzko B. & Hering P. (1995) Isotopeselective non-dispersive infrared spectrometry for detection of Helicobacter pylori infection with 13Curea breath test. Lancet, 345: 961-962. [34] Kolodny Y. & Epstein S. (1976) Stable isotope geochemistry of deep sea cherts. Geochim. Cosmochim. Acta, 40: 1195-1209. [46] Kolodny Y., Luz B. & Navon O. (1983) Oxygen isotope variation in phosphate of biogenic apatites. 1. Fish bone apatite - rechecking the rules of the game. Earth Planet. Sci. Lett., 64: 298-404. [20, 22, 46] Kolodny Y. & Raab M. (1988) Oxygen isotopes in phosphatic fish remains from Israel: paleothermometry of tropical Cretaceous and Tertiary shelf waters. Palaeogeogr., Palaeoclimatol., Palaeoecol., 64: 5967. [22] Kolodny Y. & Luz B. (1992) Isotope signatures in phosphate deposits: formation and diagenetic history. In: Isotopic signatures and sedimentary records. S. Chauduri & N. Clauer (eds.). Springer: 69-122. [22] Kolodny Y., Luz B., Sander M. & Clemens W. A. (1996) Dinosaur bones: Fossils or pseudomorphs? The pitfalls of physiology reconstruction from apatitic fossils. Palaeogeogr., Palaeoclimatol., Palaeoecol., 126: 161-171. [22]
References
1117
Komor S. C. & Magner J. A. (1996) Nitrate in groundwater and water sources used by riparian trees in an agricultural watershed: a chemical and isotopic investigation in southern Minnesota. Water Resour. Res., 32: 1039-1050. [15] Konstantinov B. P. & Bakulin E. A. (1965) Separation of chloride isotopes in aqueous solutions of lithium chloride, sodium chloride, and hydrochloric acid. Russ. J. Phys. Chem., 39: 315-318. [28] Kornexl B. E., Gehre M., Hi, fling R. & Werner R. A. (1999a) On-line 6180 measurements of organic and inorganic substances. Rapid Commun. Mass Spectrom., 13: 1685-1693. [1, 9, 15, 20, 22, 23] Kornexl B. E., Werner R. A. & Gehre M. (1999b) Standardization for oxygen isotope ratio measurement - Still an unsolved problem. Rapid Commun. Mass Spectrom., 13: 1248-1251. [1, 23, 43] Kosir I. J., Kocjancic M., Ogrinc N. & Kidric J. (2001) Use of SNIF-NMR and IRMS in combination with chemometric methods for the determination of chaptalisation and geographical origin of wines (the example of Slovenian wines). Anal. Chim. Acta, 429: 195-206. [5] Kosler J., Kucera M. & Sylvester P. (2001) Precise measurement of Li isotopes in planktonic foraminiferal tests by quadrupole ICPMS. Chem. Geol., 181: 169-179. [6] Kosterev A. A., Curl R. F., Tittel F. K., Gmachl C., Capasso F., Sivco D. L., Baillargeon J. N., Hutchinson A. L. & Cho A. Y. (1999) Methane concentration and isotopic composition measure-ments with a mid-infrared quantum-cascade laser. Opt. Lett., 24: 1762-1764. [34] Kotaka M. (1973) Chromatographic separation of boron and nitrogen isotopes using pure water as an eluant. PhD Thesis, Tokyo Institute of Technology, Tokyo, Japan. [7] Kotzer T. G., Kyser T. K., King R. W. & Kerrich R. (1993) An empirical oxygen- and hydrogen-isotope geothermometer for quartz-tourmaline and tourmaline-water. Geochim. Cosmochim. Acta, 57: 34213426. [46] Koumenis I. L., Vestal M. L., Yergey A. L., Abrams S., Deming S. N. & Hutchens T. W. (1995) Quantitation of metal isotope ratios by laser-desorption time-of-flight mass-spectrometry. Anal. Chem., 67: 4557-4564. [33] Koziet J., Rossmann A., Martin G. J. & Johnson P. (1995) Determination of the oxygen-18 and deuterium content of fruit and vegetable juice water. A European inter-laboratory study. Anal. Chim. Acta, 302: 29-37. [1] Koziet J. (1997) Isotope ratio mass spectrometric method for the on-line determinartion of oxygen-18 in organic matter. J. Mass Spectrom., 32: 103-108. [1, 15, 23] Kragten J. (1994) Calculating standard deviations and confidence intervals with a universally applicable spreadsheet technique. Analyst, 119: 2161-2165. [32, 37] Krankowsky D., Bartecki F., Klees G. G., Mauersberger K., Schellenbach K. & Stehr J. (1995) Measurement of heavy isotope enrichment in tropospheric ozone. Geophys. Res. Lett., 22(13): 1713-1716. [18] Krankowsky D., L~immerzahl P. & Mauersberger K. (2000) Isotope measurement of stratospheric ozone. Geophys. Res. Lett., 27(17): 2593-2595. [18] Kretz R. (1983) Symbols for rock forming minerals. Am. Mineralogist, 68(1-2): 277-279. [3] Kreulen R. (1980) CO2-Rich fluids during regional metamorphism on Naxos (Greece): carbon isotopes and fluid inclusions. Am. J. Sci., 280: 745-771. [3] Kreulen R. & Schuiling R. D. (1982) N2-CH4-CO2 fluids during formation of the Dome de L'Agout, France. Geochim. Cosmochim. Acta, 46: 193-203. [3, 16] Kreulen R., Van Breemen A. & Duit W. (1982) Nitrogen and carbon isotopes in metamorphic fluids from the Dome de L'Agout, France. Proc. Fifth Int. Conf. Geochronol. Cosmochronol. Isot. Geol.: 191. [16] Krishnamurthy R. V., Epstein S., Cronin J. R., Pizzarello S. & Yuen G. (1992) Isotopic and molecular analyses of hydrocarbons and monocarboxylic acids of the Murchison meteorite. Geochim. Cosmochim. Acta, 56: 4045-4058. [11] Krishnamurthy R. V., Atekwana E. A. & Guha H. (1997) A simple, inexpensive carbonate - phosphoric acid reaction method for the analysis of carbon and oxygen isotopes of carbonates. Anal. Chem., 69: 4256-4258. [10] Krohn M. D., Kendall C., Evans J. R. & Fries T. L. (1993) Relations of ammonium minerals at several hydrothermal systems in the western U.S.J. Volcanol. Geotherm. Res., 56: 401-413. [16]
1118
References
Krouse H. R. & Thode H. G. (1962) Thermodynamic properties and geochemistry of isotopic compounds of selenium. Can. J. Chem., 40: 367-375. [20, 29, 46] Krouse H. R. & Modzeleski V. E. (1970) 13C/12C abundances in components of carbonaceous chondrites and terrestrial samples. Geochim. Cosmochim. Acta, 34: 459-474. [11] Krouse H. R. (1977) Sulphur isotope studies and their role in petroleum exploration. J. Geochem. Explor., 7: 189-211. [26] Krouse H. R. (1980) Sulphur isotopes in our environment. In: The terrestrial environment, A. P. Fritz & J. C. Fontes (eds.). Elsevier. Handbook of environemental isotope geochemistry, Vol. 1: 435-471. [26] Krouse H. R. & Tabatabai M. A. (1986) Stable sulfur isotopes. In: Sulfur in agriculture. M. A. Tabatabai (ed.). Amer. Soc. Agronomy, Crop Sci. Soc. Amer., Soil Sci. Soc. Amer.: 169-205. [26] Krouse H. R. & Ueda A. (1987) Sulphur isotope analyses of trace sulphide and sulphate in various materials using Kiba reagent. In: Studies on sulphur isotope variations in nature. IAEA, Vienna. Vol. STI/ PUB/747: 113-121. [26] Krouse H. R., Levinson A. A., Piggott D. & Ueda A. (1987a) Further stable isotope investigations of human urinary stones: comparison with other body components. Appl. Geochem., 2: 205-211. [26] Krouse H. R., Ritchie R. G. G. & Roche R. S. (1987b) Sulphur isotope composition of H2S evolved during nonisothermal pyrolysis of sulphur containing materials. J. Anal. Appl. Pyrol., 12: 19-29. [26] Krouse H. R. (1991) Tracing of the fate of emissions from sour gas processing in Alberta, Canada. In: Stable Isotopes in the Assessment of Natural and Anthropogenic Sulphur. H. R. Krouse & V. A. Grinenko (eds.). John Wiley & Sons: 307-326. [26] Krouse H. R., Mayer B. & Schoenau J. J. (1996) Applications of stable isotope techniques to soil sulfur cycling. In: Mass spectrometry of soils. T. W. Boutton & S. Yamasaki (eds.). Marcel Dekker, Inc., New York: 247-284. [26] Krylov D. P., Zagnitko V. N., Hoernes S., Lugovaja I. P. & Hoffbauer R. (2002) Oxygen isotope fractionation between zircon and water; experimental determination and comparison with quartzzircon calibrations. Eur. J. Mineral., 14: 849-853. [20] Kung C. C. & Clayton R. N. (1978) Nitrogen abundances and isotopic compositions in stony meteorites. Earth Planet. Sci. Lett., 38: 421-435. [17] Kuroda P. K. & Sandell E. B. (1953) Chlorine in igneous rocks. Geol. Soc. Am. Bull., 64: 879-896. [28] Kusakabe M., Rafter T. A., Stout J. D. & Collie T. W. (1976) Sulfur isotopic variations in nature. 12. Isotopic ratios of sulfur extracted from some plants, soils and related materials. N. Z. J. Sci., 19: 433440. [26] Kusakabe M. (2001) A simple method for sampling total dissolved carbonate in carbonate-rich natural waters and CO2 preparation for 613C determination. Geochem. J., 35: 459-464. [10] Kush P., Hustralid A. & Tate J. T. (1937) Dissociation processes produced in SbC13, AsC13 and PC13 by electron impact. Phys. Rev., 52: 840-842. [28] Kvenvolden K., Lawless J., Pering K., Peterson E., Flores J., Ponnamperuma C., Kaplan I. & Moore C. (1970) Evidence for extra-terrestrial amino acids and hydrocarbons in the Murchison meteorite. Nature, 228: 926-928. [11] Kwak T. J. & Zedler J. B. (1997) Food web analysis of southern California coastal wetlands using multiple stable isotopes. Oecologia, 110: 262-277. [26] Kyser K., Klassen K. & Mitchell R. (2000) Analysis of N isotopes in micas and amphiboles using EACF-IRMS: problems, solutions and implications for the origin of N in hydrous phases from the mantle. Proc. GeoCanada 2000, Calgary, Canada. [16] Kyser T. K., Carmichael I. S. E. & O'Neil J. R. (1981) Oxygen isotope thermometry of basic lavas and mantle nodules. Contrib. Mineral. Petrol., 81: 88-102. [20, 46] Kyser T. K. (1995) Micro-analytical techniques in stable isotope geochemistry. Can. Mineral., 33: 261-278. [20] Labeyrie L. (1974) New approach to surface seawater paleotemperatures using 180/16 0 ratios in silica diatom frustules. Nature, 248: 40-42. [2, 46] Labeyrie L. D. Jr. & Juillet A. (1982) Oxygen isotopic exchange-ability of diatom valve silica; interpretation and consequences for paleoclimatic studies. Geochim. Cosmochim. Acta, 46: 976-975. [2] Lai A., Casu M., Saba G., Corongiu F. P. & Dessi M. A. (1995) NMR investigation of the intramolecular distribution of deuterium in natural triacylglycerols. Magn. Reson. Chem., 33: 163-166. [5]
References
1119
Lajtha K. & Michener R. H. (1994) Stable isotopes in ecology and environmental science, methods in ecology. Blackwell Scientific Publications, Oxford, UK: 316 p. [9] Lal D., Arnold J. R. & Somayajulu B. L. K. (1964) A method for the extraction of trace elements from seawater. Geochim. Cosmochim. Acta, 7: 1111. [22] Lambert S. J. & Epstein S. (1980) Stable isotope investigations of an active geothermal system in valles Caldera, Jemez Mountains, New Mexico. J. Volcan. Geotherm. Res., 8: 111-129. [46] Lamberty A., Michiels E. & De Bi6vre P. (1987) On the atomic weight of lithium. Int. J. Mass Spectrom. Ion Phys., 79: 311-333. [6] Lancet M. S. & Anders E. (1970) Carbon isotope fractionation in the Fischer-Tropsch synthesis and in meteorites. Science, 170: 980-982. [46] Lancoursi6re J., Meyer M., Nahon L., Morin P. & Larzilli6re M. (1994) Resonance-enhanced twophoton ionization of helium using an Ar+ mode-locked laser synchronized with VUV synchrotron radiation pulses. H.-J. Kluge, J.E. Parks & K. Wendt (eds.). Resonance Ionization Spectroscopy 1994, Seventh International Symposium. AIP Press Woodbury, NY. AIP Conference Proceddings 329: 423426. [33] Landmeyer J. E. & Stone P. A. (1995) Radiocarbon and 513C values related to groundwater recharge and mixing. Ground Water, 33: 227-234. [10] Landon M. K., Delin G. N., Komor S. C. & Regan C. P. (1999) Comparison of the stable-isotopic composition of soil water collected from suction lysimeters wick samples, and cores in a sandy unsaturated zone. J. Hydrot., 224: 45-54. [1] Lang P., Steele L. P. & Martin R. C. (1990) Atmospheric methane data for the period 1986-1988 from the NOAA/CMDL global cooperative flask sampling network. NOAA Climate Monitoring and Diagnostics Laboratory, Boulder, Colorado, USA: 108 p. [14] Lange J. & Brumsack H.-J. (1977) Total sulfur analysis in geological and biological materials by coulometric titration following combustion. Fresenius Z. Anal. Chem., 286: 361-366. [27] Langvad T. (1954) Separation of chlorine isotopes by ion exchange chromatography. Acta Chem. Scandinavia, 8: 526-527. [28] Larras-Regard E. & Mony M.-C. (1997) Improvements in biomedical imaging by SIMS (IMS 3F, 4F, and nanoSIMS). In: SIMS XI. G. Gillen, R. Lareau, J. Bennett & F. Stevie (eds.): 119-122. [30] Larsen N. W., Pedersen T. & Sehested J. (1991) Isotopic study of the mechanism of ozone formation. Int. J. Chem. Kinet., 23: 331-343. [18] Lashof D. A. & Ahuja D. R. (1990) Relative contributions of greenhouse gas emissions to global warming. Nature, 344: 529-531. [14] Lasorsa B. & Casas A. (1996) A comparison of sample handling and analytical methods for determination of acid volatile sulfides in sediment. Marine Chem., 52: 211-220. [26] Lassen J., Benck E. C. & Schuessler H. A. (1994) Ultra-sensitive detection of krypton atoms utilizing step-wise-laser excitation and field ionization. H.-J. Kluge, J.E. Parks & K. Wendt (eds.). Resonance Ionization Spectroscopy 1994, Seventh International Symposium. AIP Press Woodbury, NY. AlP Conference Proceddings 329: 327-330. [33] Lauretta D. S., Bevouard B. & Buseck P. R. (1999) The cosmochemical behavior of mercury. Earth Planet. Sci. Lett., 171: 35-47. [31] Lauretta D. S., Klaue B., Blum J. D. & Buseck P. R. (2001) Mercury abundances and isotopic compositions in the Murchison (CM) and Allende (CV) carbonaceous chondrites. Geochim. Cosmochim. Acta, 65: 2807-2818. [31] Lawrence J. R. (1970) O18/O16 and D / H ratios of soils, weathering zones and clay deposits. PhD Thesis, California Institute of Technology, Pasadena, California, USA: 263 p. [2] Lawrence J. R. & Taylor H. P. Jr. (1971) Deuterium and oxygen-18 correlation: clay minerals and hydroxides in Quaternary soils compared to meteoric waters. Geochim. Cosmochim. Acta, 35: 993-1003. [46] Laws E. A., Popp B. N., Bidigare R. R., Kennicutt M. C. & Macko S. A. (1995) Dependence of phytoplankton carbon isotopic composition on the growth rate and [CO2]aq: theoretical considerations and experimental results. Geochim. Cosmochim. Acta, 59: 1131-1138. [46] Lazar B. & Holland H. D. (1988) The analysis of fluid inclusions in halite. Geochim. Cosmochim. Acta, 52: 485-490. [3]
1120
References
Leaney F. W., Smettem K. R. J. & Chittleborough D. J. (1993) Estimating the contribution of preferential flow to subsurface runoff from a hillslope using deuterium and chloride. J. Hydrol., 147: 83-103. [2] Leavitt S. W. & Long A. (1984) Sampling strategy for stable carbon isotope analysis of tree rings in pine. Nature, 311: 145-147. [24] Leavitt S. W. & Long A. (1991) Seasonal stable isotope variability in tree rings: Possible paleoenvironmental signals. Chem. Geol., 87: 59-70. [24] Leavitt S. W. & Danzer S. R. (1993) Method for batch processing small wood samples to holocellulose for stable-carbon isotope analysis. Anal. Chem., 65: 87-89. [24] Leavitt S. W. (1993a) Environmental information from 13C/12C ratios of Wood. In: Climate change in continental isotopic records. Geophysical Monograph, Vol. 78: 325-331. [24] Leavitt S. W. (1993b) Seasonal 13C/12C changes in tree rings: Species and site coherence, and a possible drought influence. Can. J. Forest Res., 23: 210-218. [24] Leavitt S. W. & Lara A. (1994) South American tree rings show declining 613C trend. Tellus, 46(B): 152157. [24] Leckrone K. J. & Hayes J. M. (1997) Efficiency and temperature dependence of water removal by membrane dryers. Anal. Chem., 69: 911-918. [14] Leckrone K. J. & Hayes J. M. (1998) Water-induced errors in continuous-flow carbon isotope ratio mass spectrometry. Anal. Chem., 70: 2737-2744. [8, 14, 38] Leckrone K. J. & Crowe D. E. (1999) In-situ isotopic analysis of sulfide minerals by UV laser ablationgas chromatography - isotope ratio mass spectrometry. EOS Trans. AGU, 89(17): 51-V. [45] Leder J. J., Swart P. K., Szmant A. M. & Dodge R. E. (1996) The origin of variations in the isotopic record of scleractinian corals: I. Oxygen. Geochim. Cosmochim. Acta, 60: 2857-2870. [46] Lee D.-C. & Halliday A. N. (1995) Precise determinations of the isotopic compositions and atomic weights of molybdenum, tellurium, tin, and tungsten using ICP magnetic sector multiple collector mass spectrometry. Int. J. Mass Spectrom. Ion Proc., 146/147: 35-46. [31, 35] Lee P. S. & Majkowski R. F. (1986) High resolution infrared diode laser spectroscopy for isotope analysis - measurement of isotopic carbon monoxide. Appl. Phys. Lett., 48: 619-621. [34] Lee T., Mayeda T. K. & Clayton R. N. (1980) Oxygen isotopic anomalies in Allende inclusions HAL. Geophys. Res. Lett., 7: 493-496. [20] Lee T. & Tera F. (1986) The meteoritic chromium isotopic composition and limits for radioactive (super 53) Mn in the early solar system. Geochim. Cosmochim. Acta, 50(2): 199-206. [29] Lee-Thorp J. A. & van der Merwe N. J. (1991) Aspects of the chemistry of modern and fossil biological apatites. J. Archaeol. Sci., 18: 343-354. [9] Leeman W. P., Vocke R. D. Jr., Beary E. S. & Paulsen P. J. (1991) Precise boron isotopic analysis of aqueous samples: Ion exchange extraction and mass spectrometry. Geochim. Cosmochim. Acta, 55: 3901-3907. [40] Leeman W. P. & Sisson V. P. (1996) Geochemistry of boron and its implications for crustal and mantle processes. In: Boron: Mineralogy, petrology and geochemistry. E. S. Grew & L. M. Anovitz (eds.). Mineralogical Society of America, Reviews in Mineralogy, Vol. 33: 645-707. [7] Legge A. H., Corbin J., Bogner J. C., Strosher M., Krouse H. R., Laishley E. J., Bryant R. D., Cavey M. J., Prescott C. E., Nosal M., Schellhase H. U., Weidensaul T. C. & Mayo J. (1988) Acidification in a temperate forest ecosystem: the role of sulfur gas emissions and sulfur dust. Phoenix Press Inc., Calgary, Canada: 399 p. [26] Lehmann B., Wahlen M., Zumbrunn R. & Oescher H. (1977) Isotope analysis by infrared laser absorption spectroscopy. Appl. Phys., B 13: 153-158. [34] Lehmann K. K. (1996) Ring-down cavity spectroscopy cell using continuous wave excitation for trace species detection. U.S. Patent 5,528,040. [34] Lehmann M. F., Bernasconi S. M. & McKenzie J. A. (2001) A method for the extraction of ammonium from freshwaters for nitrogen isotope analysis. Anal. Chem., 73: 4717-4721. [15] Lei J. J., Ren D. u Tang Y. G., Chu X. L. & Zhao R. (1994) Sulfur-accumulating model of super-high organosulfur coal from Guiding, China. Chinese Sci. Bull., 39: 1817-1821. [26] Leshin L. A., Rubin A. E. & McKeegan K. D. (1997) The oxygen isotopic composition of olive and pyroxene from CI chondrites. Geochim. Cosmochim. Acta, 61: 835-845. [30]
References
1121
Leshin L. A., McKeegan K. D., Carpenter P. K. & Harvey R. P. (1998) Oxygen isotopic constraints on the genesis of carbonates from martian meteorite ALH84001. Geochim. Cosmochim. Acta, 62: 3-13. [30]
Leshin L. A. (2000) Implications for the origin and history of water on Mars from analyses of martian meteorite QUE94201. Geophys. Res. Lett., 27: 2017-2020. [30]
Leuenberger M., Borella S., Stocker T., Saurer M., Siegwolf R., Schweingruber F. & Matyssek R. (1998) Stable isotopes in tree rings as climate and stress indicators, vdf Hochschulverlag AG an der ETH Zfirich, Zfirch, Switserland. [24] Leuenberger M. & Huber C. (2002) On-line determination of oxygen isotope ratios of water or ice by mass spectrometry. Anal. Chem., 74: 4611-4617. [1] Lewis R. S., Tang M., Wacker J. F., Anders E. & Steel E. (1987) Interstellar diamonds in meteorites. Nature, 326: 160-162. [13] L6cluse C. & Robert F. (1994) Hydrogen isotope exchange reaction rates: origin of water in the inner solar system. Geochim. Cosmochim. Acta, 58: 2927-2939. [46] L6cuyer C., Grandjean P., O'Neil J. R., Cappette H. & Martineau F. (1993) Thermal excursions in the ocean at the Cretaceous-Tertiary boundary (northern Morocco): 6180 record of phosphatic fish debris. Palaeogeog. Palaeoclimatol. Palaeoecol., 105: 235-243. [20, 22] L6cuyer C. & O.Neil J. R. (1994) Stable isotope compositions of fluid inclusions in biogenic carbonates. Geochim. Cosmochim. Acta, 58: 353-363. [3] L6cuyer C., Grandjean P. & Emig C. C. (1996a) Determination of oxygen isotope fractionation between water and phosphate from living lingulids: potential application to palaeoenvironmental studies. Paleogeogr., Palaeoclim., Palaeoecol., 126: 101-108. [22, 46] L6cuyer C., Grandjean P., Paris F., Robardet M. & Robineau D. (1996b) Deciphering "temperature" and "salinity" from biogenic phosphates: the 6180 of coexisting fishes and mammals of the Middle Miocene sea of western France. Palaeogeogr., Palaeoclim., Palaeoecol., 126: 61-74. [22] L6cuyer C., Grandjean P., Barrat J.-A., Nolvak J., Emig C., Paris F. & Robardet M. (1998) 6180 and REE contents of phosphatic brachiopods: A comparison between modern and lower Paleozoic populations. Geochim. Cosmochim. Acta, 62(14): 1429-2436. [22] L6cuyer C., Grandjean P. & Sheppard S. M. F. (1999) Oxygen isotope exchange between dissolved phosphate and water at temperatures M35~ Inorganic versus biological fractionations. Geochim. Cosmochim. Acta, 63(6): 855-862. [22, 46] L6cuyer C., Grandjean P., Reynard B., Albar6de F. & Telouk P. (2000) 11B/ 10B analysis of geochemical materials by ICP-MS Plasma 54: methodology and paleo-pH of seawater. EOS Trans. AGU, 81: F700. [31] Li B. & Shi J. (1983) Continuous determination of stable isotopes of hydrogen-gases and salts in fluid inclusions. Scientia Sinica, Series B: Chemical, Biological, Agricultural, Medical and Earth Sciences (Chung Kuo K'o Hsueh Yuan, Chu Pan), 26(1): 93-102. [3] Li G. Q., Duan Y. X. & Hieftje G. M. (1995) Space-charge effects and ion distribution in plasma source mass spectrometry. J. Mass Spectrom., 30: 841-848. [35] Li W., Ni B., Jin D. & Zhang Q. (1988) Measurement of the absolute abundance of oxygen-17 in VSMOW. Kexue Tongbao (Chinese Sci. Bull.), 33: 1610-1613. [4, 40, 44] Li Y. M. (1996) Sulfur isotope composition of SO2 prepared by laser assited oxidation of sulfur. PhD Thesis, University of Calgary, Calgary, Canada: 149 p. [26] Libby L. M., Pandolfi L. J., Payton P. H., Marshall J., Becker B. & Giertz-Siebenlist V. (1976) Isotope tree thermometers. Nature, 261: 284-288. [23] Lichtfouse 1~., Elbisser B., Balesdent J., Mariotti A. & Bardoux G. (1995) Isotope and molecular evidence for direct input of maize leaf wax n-alkanes into crop soils. Org. Geochem., 22: 349-351. [9] Lichtfouse l~., Bardoux G., Mariotti A., Balesdent J., Ballentine D. C. & Macko S. A. (1997) Molecular, 13C, and 14C evidence for the allochthonous and ancient origin of C16 - C18 n-alkanes in modern soils. Geochim. Cosmochim. Acta, 61: 1891-1898. [9] Lide R. A. (1994) Handbook of chemistry and physics. CRC-Press, Ann Arbor, USA. 75th edition. [36] Liebl H. (1983) Ion gun systems for submicron SIMS. In: SIMS. J. Okano, R. Shimizu & H. W. Werner (eds.). Springer. Vol. IV: 114-118. [30] Liewig N., Caluer N. & Sommer F. (1987) Rb-Sr and K-Ar dating of clay diagenesis in Jurassic sandstone oil reservoir, north Sea. Am. Assoc. Petrol. Geol. Bull., 71: 1467-1474. [2]
1122
References
Lindars E. S., Grimes S. T., Mattey D. P., Collinson M. E., Hooker J. J. & Jones T. P. (2001) Phosphate ~)180 determination of modern rodent teeth by direct laser fluorination: An appraisal of methodology and potential application to palaeoclimate reconstruction. Geochim. Cosmochim. Acta, 65(15): 2535-2548. [20] Linder P., Bermann E. & Gamarnik B. (1996) Characterization of citrus honey by deuterium NMR. J. Agric. Food Chem., 44: 139-140. [5] Lipschultz F., Wofsy S. C. & Fox L. E. (1986) Nitrogen metabolism of the eutrophic Delaware River ecosystem. Limnol. Oceanogr., 31: 701-716. [15] Litaor M. I. (1988) Review of soil solution samplers. Water Resour. Res., 24: 727-733. [2] Liu K. K. & Epstein S. (1984) The hydrogen isotope fractionation between kaolinite and water. Isotope Geosci., 2: 335-350. [46] Liu K. K., Su M. J., I-Isueh C. R. & Gong G. C. (1996) The nitrogen isotopic composition of nitrate in the Kuroshio Water northeast of Taiwan: evidence for nitrogen fixation as a source of isotopically light nitrate. Marine Chem., 54: 273-292. [9] Liu S. C. & Trainer M. (1988) Responses of the tropospheric ozone and odd hydrogen radicals to column ozone change. J. Atmos. Chem., 6: 221-234. [18] Liu W. G., Xiao Y. K., Peng Z. C., An Z. S. & He X. X. (2000) Boron concentration and isotopic composition of halite from experiments and salt lakes in the Qaidam Basin. Geochim. Cosmochim. Acta, 64: 2177-2183. [46] Liu u P. & Mulvaney R. L. (1992) Use of diffusion for automated nitrogen-15 analysis of soil extracts. Commun. Soil. Sci. Plant Anal., 23: 613-629. [15] Livingston S. W. & Spittlehouse D. L. (1996) Carbon isotope fractionation in tree ring early and late wood in relation to intragrowing season water balance. Plant Cell Environm., 19: 768-774. [24] Lloyd R. M. (1968) Oxygen isotope behavior in the sulfate-water system. J. Geophys. Res., 73: 6099-6110. [46] Loader N. J., Robertson I., Barker A. C., Switsur V. R. & Waterhouse J. S. (1997) An improved technique for the batch processing of small wholewood samples to c~-cellulose. Chem. Geol., 136: 313-317. [24] Loader N. J.& Buhay W. M. (1999) Rapid catalytic oxidation of CO to CO2 - On the development of a new approach to on-line oxygen isotope analysis of organic matter. Rapid Commun. Mass Spectrom., 13: 1828-1832. [1] Lober R. W., Reeder J. D. & Porter L. K. (1987) New and conventional evaporitive systems in concentrating nitrogen samples prior to isotope-ratio analysis. Soil. Sci. Soc. Am. J., 51: 951-958. [15] Lockheart M. J., Van Bergen P. F. & Evershed R. P. (1997) Variations in the stable carbon isotope compositions of individual lipids from the leaves of modern angiosperms: implications for the study of higher land plant-derived sedimentary organic matter. Org. Geochem., 26: 137-153. [9] Long A., Eastoe C. J., Kaufmann R. S., Martin J. G., Wirt L. & Finley J. B. (1993) High-precision measurement of chlorine stable isotope ratios. Geochim. Cosmochim. Acta, 57: 2907-2912. [28] Longerich H. P., Fryer B. J. & Strong D. F. (1987) Determination of lead isotope ratios by inductively coupled plasma-mass spectrometry (ICP-MS). Spectrochim. Acta, B 42: 39-48. [31, 32, 35] Longinelli A. (1965) Oxygen isotope composition of orthophosphate from shells of living marine organisms. Nature, 207: 716. [20, 22, 46] Longinelli A. (1966) Ratios of oxygen-18:oxygen-16 in phosphate and carbonate from living and fossil marine organisms. Nature, 211: 923-927. [22, 46] Longinelli A. & Nuti S. (1968a) Oxygen isotope composition of phosphates from marine formations. Earth Planet. Sci. Lett., 5: 13-17. [22] Longinelli A. & Nuti S. (1968b) Oxygen-isotope ratios in phosphate from marine organisms. Science, 160: 879-882. [22] Longinelli A. & Nuti S. (1973a) Revised phosphate-water isotopic temperature scale. Earth Planet. Sci. Lett., 19: 373-376. [20, 22, 46] Longinelli A. & Nuti S. (1973b) Oxygen isotope measurements of phosphate from fish teeth and bones. Earth Planet. Sci. Lett., 20: 337-340. [22] Longinelli A., Bartelloni M. & Cortecci G. (1976) The isotopic cycle of oceanic phosphate, I. Earth Planet. Sci. Lett., 32: 389-392. [22]
References
1123
Longinelli A. (1984) Oxygen isotopes in mammal bone phosphate: A new tool for paleohydrological and paleoclimatological research? Geochim. Cosmochim. Acta, 48: 385-390. [20, 22] Longstaffe F. J. (1986) Oxygen isotope studies of diagenesis in the basal Belly River sandstone, Pembina I-pool, Alberta. J. Sediment. Petrol., 56: 78-88. [2] Longstaffe F. J. & Mizota C. (1999) Rapid changes in the oxygen- and hydrogen-isotope composition of halloysite during laboratory wetting at 23~ Geol. Soc. Am. Abstr. w. Prog., (Abstract). [2] Lory J. & Russelle M. P. (1994) Evaluation of diffusion method for preparing low-nitrogen samples for 15N analysis. Soil Sci. Soc. Am. J., 58: 1400-1404. [15] Lovley D. R., Stolz J. F., Nord G. L. & Phillips E. J. P. (1987) Anaerobic production of magnetite by dissimilatory iron-reducing microorganism. Nature, 330: 252-254. [29] Lovley D. R. & Phillips E. J. P. (1994) Reduction of chromate by Desulfovibrio viulgaris and its C-3 cytochrome. Appl. Env. Microbiol., 60: 726-728. [29] Lowe D. C., Brenninkmeijer C. A. M., Tyler S. C. & Dlugokencky E. J. (1991) Determination of the isotopic composition of atmospheric methane and its application in the Antarctic. J. Geophys. Res., 96(D8): 15455-15467. [14] Lowe D. C., Brenninkmeijer C. A. M., Brailsford G. W., Lassey K. R., Gomez A. J. & Nisbet E. G. (1994) Concentration and C-13 records of atmospheric methane in New-Zealand and Antarctica- Evidence for changes in methane sources. J. Geophys. Res., 99(D8): 16913-16925. [14] Lowe D. C., Manning M. R., Bromley T. B., Allen W., Brailsford G., Ferretti D., Gomez A., Knobben R., Martin R., Mei Z., Moss R., Koshy K. & Maata M. (1999) Shipboard determinations of the distribution of 13C in atmospheric methane in the Pacific. J. Geophys. Res., 104(D21): 26125-26135. [14] Lowe L. E. (1965) Sulphur fractions of selected Alberta soil profiles of the chernozemic and podzolic orders. Can. J. Soil Sci., 45: 297-303. [26] Lowenstam H. A. (1961) Mineralogy, 180/16 0 ratios, and strontium and magnesium contents of recent and fossil brachiopods and their bearing on the history of the oceans. J. Geol., 69: 241-260. [46] Lu Z.-T. & Wendt K. D. A., (2003) Laser-based methods for ultrasensitive trace-isotope analyses. Rev. Sci. Instr., 74: 1169-1407. [34] Lund6n A. & Herzog W. (1956) Isotopenanreicherung bei Chlor durch electrolytische Oberfiihrung in geschmolzenem Zinkchlorid. Z. Naturforsch., 11a: 520. [28] Lund6n A. & Lodding A. (1960) Isotopenanreicherung bei Brom durch electrolytische Oberfiihrung in geschmolzenem Zinkbromid. Z. Naturforsch., 15a: 320-322. [28] Luo Y. H. & Sternberg L. (1992) Spatial D / H heterogeneity of leaf water. Plant Physiol., 99: 348-350. [21] Lusk J. & Krouse H. R. (1997) Comparative stable isotope and temperature investigation of minerals and associates fluids in two regionally metamorphosed (Kuroko-type) volcanogenic massive sulfide deposits. Chem. Geol., 143: 231-253. [3] Luz B., Kolodny Y. & Kovach J. (1984) Oxygen isotope variations in phosphate of biogenic apatites. III. Conodonts. Earth Planet. Sci. Lett., 69: 255-262. [20, 22] Luz B. & Kolodny Y. (1985) Oxygen isotope variation in phosphate of biogenic apatites, IV. Mammal teeth and bones. Earth Planet. Sci. Lett., 75: 29-36. [22] Luz B., Barkan E., Bender M. L., Thiemens M. H. & Boering K. A. (1999) Triple-isotope composition of atmospheric oxygen as a tracer for biospheric productivity. Nature, 400: 547-550. [18] Lyon G. L. & Cox M. A. (1978) Stable isotope analysis of hydrogen and methane mixtures. Inst. Nucl. Sci., DSIR, Lower Hutt, New Zealand. Int. Rept. INS-R--282. [33] Lyon G. L. & Cox M. A. (1980) The reduction of water to hydrogen for D / H ratio analysis, using zinc in a matrix of sand. Inst. Nucl. Sci., DSIR, Lower Hutt, New Zealand. Int. Rept. INS-R-282, November, 1980: 7p. [1] Lyon I. C., Saxton J. M., McKeever P. J., Chatzitheodoridis E. & Van Lierde P. (1995) Precision and reproducibility on quartz obtained using an Isolab 54 ion microprobe. Int. J. Mass Spectrom. Ion Phys., 9(8): 1-16. [30] Lyon T. D. B. & Fell G. S. (1990) Isotopic composition of copper in serum by inductively coupled plasma mass spectrometry. J. Anal. Atom. Spectrom., 5: 135-137. [31] Lyon T. D. B., Fletcher S., Fell G. S. & Patriarca M. (1996) Measurement and application of stable copper isotopes to investigations of human metabolism. Microchem. J., 54: 236-245. [31]
1124
References
Lyons J. R. (2001) Transfer of mass-independent fractionation in ozone to other oxygen-containing radicals in the atmosphere. J. Geophys. Res., 28(17): 3231-3234. [18] Lyons T. W., Zhang C. L. & Romanek C. S. (2003) Isotopic records of microbially mediated processes. Chem. Geol., 195(edited issue): 1-274. [46] MacAvoy S. E., Macko S. A. & Garman G. C. (1998) Tracing marine biomass into tidal freshwater ecosystems using stable sulfur isotopes. Naturwissensch., 85: 544-546. [26] MacAvoy S. E., Macko S. A., McInnich S. P. & Garman G. C. (2000) Marine nutrient contributions to freshwater apex predators. Oecologia, 122: 568-573. [26] MacAvoy S., Macko S. A. & Garman G. (2001) Isotopic turnover in aquatic predators: quantifying the explotation of migratory prey. Can. J. Fish. Aq. Sci., 58: 923-932. [45] MacDaniel C. V. & Maher P. K. (1976) Zeolite stability and ultrastable zeolites. In: Zeolite chemistry and catalysis. J. A. Rabo (ed.). American Chemical Society, Washington. ACS Monograph 171, Chapter 4: 285-331. [36] MacElroy J. M. D., Freidman S. P. & Seaton N. A. (1999) On the origin of transport resistences within carbon molecular sieves. Chem. Eng. Sci., 54: 1015-1027. [36] MacFarlane A. W. & Shimizu N. (1991) SIMS measurement of 634S in sulfide minerals from adjacent vein and stratabound ores. Geochim. Cosmochim. Acta, 55: 525-541. [26, 30] Macko S. (1981) Stable nitrogen isotope ratios as tracers of organic geochemical processes. PhD Thesis, University of Texas, Austin, Texas, USA: 181 p. [15, 16, 36] Macko S. A., Fogel Estep M. L., Engel M. H. & Hare P. E. (1986) Kinetic fractionation of stable nitrogen isotopes during amino acid transamination. Geochim. Cosmochim. Acta, 50: 2143-2146. [46] Macko S. A., Estep M. L. F., Hare P. E. & Hoering T. C. (1987) Isotopic fractionation of nitrogen and carbon in the synthesis of amino acids by microorganisms. Chem Geol. (Isot. Geosc.), 65: 79-92. [9] Macko S. A., Helleur R., Hartley G. & Jackman P. (1990) Diagenesis of organic matter - A study using stable isotopes of individual carbohydrates. Org. Geochem., 16: 1129-1137. [9] Macko S. A., Engel M. H., Hartley G., Hatcher P., Helleur R., Jackman P. & Silfer J. A. (1991) Isotopic compositions of individual carbohydrates as indicators of early diagenesis of organic matter in peat. Chem. Geol., 93: 147-161. [9] MacNamara J. & Thode H. G. (1950) Comparison of the isotopic constitution of terrestrial and meteoritic sulfur. Phys. Rev., 78: 307-308. [26, 40, 43, 44] Macpherson C., Hilton D. R., Newman S. & Mattey D. P. (1999) CO2, 13C/12C and H20 variability in natural basaltic glasses: A study comparing stepped heating and FTIR spectroscopic techniques. Geochim. Cosmochim. Acta, 63: 1805-1813. [36] Madigan M. T., Martinko J. M. & Parker J. (2000) Brock Biology of Microorganisms. Prentice Hall Int. 9th edition. [46] Madorsky S. L. & Strauss S. (1948) Concentration of isotopes of chlorine by the counter-current electromigration method. J. Res. Nat. Bur. Stand., 38: 185-189. [28] Madu B. E., Nesbitt B. E. & Muehlenbachs K. (1990) A mesothermal gold-stibnite-quartz vein occurrence in the Canadian Cordillera. Econ. Geol., 85(6): 1260-1268. [3] Magaritz M. & Taylor H. P. Jr. (1976) 180/160 and D / H studies along a 500 km traverse across the Coast Range batholith and its country rocks, central British Columbia. Can. J. Earth Sci., 13: 1514-1536. [20] Magenheim A. J., Spivack A. J., Volpe C. & Ranson B. (1994) Precise determination of stable chlorine isotope ratios in low-concentration natural samples. Geochim. Cosmochim. Acta, 58: 3117-3121. [28] Magenheim A. J., Spivack A. J., Michael P. J. & Gieskes J. M. (1995) Chlorine stable isotope composition of the oceanic crust: implications for Earth's distribution of chlorine. Earth Planet. Sci. Lett., 131: 427-432. [46] Mahon K. I., Harrison T. M. & McKeegan K. D. (1998) The thermal and cementation histories of a sandstone petroleum reservoir, Elk Hills, California. Part 2: In situ oxygen and carbon isotopic results. Chem. Geol., 152: 257-271. [30] Majoube M. (1971) Fractionnement en oxyg6ne 18 et en deut6rium entre l'eau et sa vapeur. J. Chim. Phys. Physicochim. Biol., 68: 1423-1436. [46] Majzoub M. A. (1966) Une methode d'analyse isotopique de l'oxyg6ne sur des microquantit6s d'eau; determination des coefficients de partage a l'6quilibre de l'oxig6ne 18 entre H20 et CO2; D20 et CO2. J. chim. Phys., 63(4): 565-568. [1, 20]
References
1125
Majzoub M. & Nief G. (1968) A mass spectrometer for the isotopic oxygen analysis of water. Adv. Mass Spectrom., 4: 511-520. [1] Maliva R. G., Dickson J. A. D. & Fallick A. E. (1999) Kaolin cements in limestones: Potential indicators of organic-rich pore waters during diagenesis. J. Sedim. Petrol., 69: 158-163. [2] Mandernack K. W., Bazylinski D. A., Shanks W. C. III& Bullen T. D. (1999) Oxygen and iron isotope studies of magnetite produced by magnetotactic bacteria. Science, 285: 1892-1896. [29, 31, 46] Mandernack K. W., Lynch L., Krouse H. R. & Morgan M. D. (2000) Sulfur cycling in wetland peat of the New Jersey Pinelands and its effect on stream water chemistry. Geochim. Cosmochim. Acta, 64: 3949-3964. [26] Manheim F. T. & Sayles F. L. (1974) Composition and origin of interstitial waters of marine sediments based on deep sea drill cores. In: The Sea. E. D. Goldberg (ed.). Wiley, NY. Vol. 5 : 527-568. [2] Manian S. H., Urey H. C & Bleakney W. (1934) An investigation of the relative abundance of the oxygen isotopes O16:O 18 in stone meteorites. J. Am. Chem. Soc., 56: 2601-2609. [20] Manowitz B., Krouse H. R., Barker C. & Premuzic E. T. (1990) Sulfur isotope data analysis of crude oils from Bolivar Coastal fields (Venezuela). In: Geochemistry of sulfur in fossil fuels. W. L. Orr & C. M. White (eds.). American Chemical Society. Vol. 429: 592-612. [26] Mansuy L., Philp R. P. & Allen J. (1997) Source identification of oil spills based on the isotopic composition of individual components in weathered oil samples. Environm. Sc. & Techn., 31: 3417-3425. [8] Manzano B. K., Fowler M. G. & Machel H. G. (1997) The influence of thermochemical sulphate reduction on hydrocarbon composition in Nisku reservoirs, Brazeau river area, Alberta, Canada. Org. Geochem., 27: 507-521. [26] Marcus R. K. (1993) Operation principles and design considerations for radiofrequency powered glow discharge devices. J. Anal. At. Spectrom., 8: 935-943. [35] Marcus R. K., Harville T. R., Mei Y. & Schick C. R. Jr. (1994) Rf-powdered glow discharges: Elemental analysis across the solids spectrum. Anal. Chem., 66: 902A-911A. [35] Mar6chal C. N., T61ouk P. & Albar6de F. (1999) Precise analysis of copper and zinc isotopic compositions by plasma-source mass spectrometry. Chem. Geol., 156: 251-273. [31, 32, 46] Mar6chal C. N., Nicolas E., Douchet C. & Albar6de F. (2000) Abundance of zinc isotopes as a marine biogeochemical tracer. Geochem. Geophys. Geosys., 1: 1999GC000029. [31] Mar6chal C. & Albari~de F. (2002) Ion-exchange fractionation of copper and zinc isotopes. Geochim. Cosmochim. Acta, 66:1499-1509 (and see GCA, 66: 3675). [46] Mariotti A., German J. C., Hubert P., Kaiser P., Letolle R., Tradieux A. & Tradieux P. (1981) Experimental determination of nitrogen kinetic isotope fractionation: some principles; Illustration for the denitrification and nitrification processes. Plant and Soil, 62: 413-430. [46] Mariotti A. (1983) Atmospheric nitrogen is a reliable standard for natural 15N abundance measurements. Nature, 303: 680-683. [40] Marra P. P., Hobson K. A. & Holmes R. T. (1998) Linking winter and summer events in a migratory bird by using stable-carbon isotopes. Science, 282: 1884-1886. [9] Marsh K. L., Mulvaney R. L. & Khan S. A. (2003) Use of diffusion for enzymatic determination of urea-nitrogen in soil extracts. Commun. Soil Sci. Plant Anal., in press. [15] Marshall D., Pfeiffer H. R. & Sharp Z. D. (1994) A re-evaluation of Raman as a tool for the determination of 12C and 13C in geological fluid inclusions. Analusis, 22(10): 38-41. [3] Martin G. J. & Martin M. L. (1981) Deuterium labelling at the natural abundance level as studied by high field quantitative 2H-NMR. Tetrahedron Lett., 22(36): 3525-3528. [5] Martin G. J., Martin M. L., Mabon F. & Bricout J. (1982a) A new method for the identification of the origin of natural products. Quantitative 2H NMR at the natural abundance level applied to the characterization of anetholes. J. Am. Chem. Soc., 104: 2658-2659. [5] Martin G. J., Martin M. L., Mabon F. & Michon M. J. (1982b) Identification of the origin of natural alcohols by natural abundance hydrogen-2 nuclear magnetic resonance. Anal. Chem., 54(13): 2380-2382.
[51
Martin G. J., Martin M. L., Mabon F. & Michon M. J. (1983a) A new method for the identification of the origin of ethanols in grain and fruit spirits: high field quantitative deuterium nuclear magnetic resonance at the natural abundance level. J. Agric. Food Chem., 31: 311-315. [5]
1126
References
Martin G. J., Martin M. L., Mabon F. & Bricout J. (1983b) Natural Abundance Deuterium NMR, a new method of identification of the origin of food products applied to the recognition of anetholes and estragoles. Sci. Aliments, 3: 147-155. [5] Martin G. J., Benbernou M. & Lantier F. (1985a) Application of site-specific (SNIF NMR) natural isotope fractionation of hydrogen to the characterization of european beers. J. Inst. Brew., 91: 242-249.
[51
Martin G. J., Janvier P. & Mabon F. (1985b) Caract6risation des an6tholes par RMN du deuterium. Analyse factorielle des rapports isotopiques sp6cifiques. Analusis, 13(6): 267-274. [5] Martin G. J., Zhang L. L., Naulet N. & Martin M. L. (1986a) Deuterium transfer in the bioconversion of glucose to ethanol studied by specific isotope labeling at the natural abundance level. J. Am. Chem. Soc., 108(17): 5116-5122. [5] Martin G. J., Guillou C., Naulet N., Brun S., Tep Y., Cabanis J. C., Cabanis M. T. & Sudraud P. (1986b) Contr61e de l'origine et de l'enrichissement des vins par analyse isotopique sp6cifique. Sci. Aliments, 6: 385-405. [5] Martin G. J., Janvier P., Akoka S., Mabon F. & Jurczak J. (1986c) A relation between the site-specific natural deuterium contents in R-pinenes and their optical activity. Tetrahedron Lett., 27(25): 28552858. [51 Martin G. J., Guillou C., Martin M. L., Cabanis M. T., Tep Y. & Aerny J. (1988) Natural factors of isotope fractionation and the characterization of wines. J. Agric. Food Chem., 36: 316-322. [5] Martin G. J. & Naulet N. (1988) Precision, accuracy and referencing of isotope ratios determined by nuclear magnetic resonance. Fresenius Z. Anal. Chem., 332: 648-651. [5] Martin G. J., Danho D. & Vallet C. (1991a) Natural isotope fractionation in the discrimination of sugar origins. J. Sci. Food. Agric., 56: 419-434. [5] Martin G. J., Martin M. L. & Zhang B. L. (1992) Site-specific natural isotope fractionation of hydrogen in plant products studied by nuclear magnetic resonance. Plant Cell Environ., 15: 1037-1050. [5] Martin G., Remaud G. S. & Martin G. J. (1993a) Isotopic methods for control of natural flavours authenticity. Flavour Frag. J., 8: 97-107. [5] Martin G. J., Hanneguelle S. & Remaud G. (1993b) Application of site-specific natural isotope fractionation studied by Nuclear Magnetic Resonance (SNIF-NMR) to the detection of flavour and fragrance adulteration. Ital. J. Food Sci., 3: 191-213. [5] Martin G. J., Trierweiler M., Ristow R., Hermann A. & Belliardo J. J. (1994) The certification of the three reference ethanols for SNIF-NMR. BCR certified materials, EUR 15347 EN. [5] Martin G. J., Koziet J., Rossmann A. & Dennis J. (1996b) Site-specific natural isotope fractionation in fruit juices determined by deuterium NMR. An european inter-laboratory comparison study. Anal. Chim. Acta, 321: 137-146. [5] Martin G. G., Guillou C. & Caisso M. (1991c) Detection of beet sugar addition in fruit juices and concentrates. An overview of isotopic techniques including new development made possible by 2H-NMR. Flfdssiges Obst, 1(2): 18-21. [5] Martin G. G., Hanote V., Lees M. & Martin Y. L. (1996a) Interpretation of combined 2H SNIF/NMR and 13C SIRA/MS analyses of fruit juices to detect added sugar. J. AOAC Int., 79(1): 62-72. [5] Martin G. G., Martin Y. L., Naulet N. & McManus H. J. D. (1996c) Application of 2H SNIF-NMR and 13C SIRA-MS analyses to maple syrup: detection of added sugars. J. Agric. Food Chem., 44: 3206-3213.
[51
Martin M. L. & Martin G. J. (1990) Deuterium NMR in the study of site-specific natural isotope fractionation (SNIF-NMR). In: NMR basic principles and progress. P. Dietl, E. Fluck, H. Gfinter, R. Kosfeld& J. Seelig (eds.). Springer Verlag, Berlin. Vol. 23: 1-61. [5] Martin M. L., Martin G. J. & Guillou C. (1991b) A site-specific and multi-element isotopic approach to origin interference of sugars in foods and beverages. Mikrochim. Acta [Wien], Ih 81-91. [5] Martin Y. L. (1994) A global approach to accurate and automatic quantitative analysis of NMR spectra by complex least-squares curve fitting. J. Magn. Reson., Ser. A, 111: 1-10. [5] Martinelli R. U. (1996) Mid-infrared wavelengths enhance trace-gas sensing. Laser Focus World, 3 (March): 7781-7783. [34] Marry B., Lenoble M. & Vassard N. (1995) Nitrogen, helium and argon in basalt: a static mass spectrometry study. Chem. Geol., 120: 183-185. [13, 17]
References
1127
Marty B. & Humbert F. (1997) Nitrogen and argon isotopes in oceanic basalts. Earth Planet. Sci. Lett., 152: 101-112. [17] Marumo K., Longstaffe F. J. & Matsubaya O. (1995) Stable isotope geochemistry of clay minerals from fossil and active hydrothermal systems, southwestern Hokkaido, Japan. Geochim. Cosmochim. Acta, 59: 2545-2559. [2] Mas F., Beemsterboer B., Veltkamp A. C. & Verweij A. M. A. (1995) Determination of common-batch members in a set of confiscated 3,4-(Methylendioxy)Methylamphetamine samples by measuring the natural isotope abundances - a preliminary-study. Forensic Sci. Int., 71: 225-231. [8] Masarie K. A., Langenfelds R. L., Allison C. E., Conway T. J., Dlugokencky E. J., Francey R. J., Novelli P. C., Steele L. P., Tans P. P., Vaughn B. & White J. W. C. (2001) The NOAA / CSIRO Flask-Air Intercomparison Program: A strategy for directly assessing consistency among atmospheric measurements derived from independent laboratories. J. Geophys. Res., 105(D17): 20445-20464. [14] Mason P. R. D., de Hoog C.-J. & Meffan-Main S. (1999a) A new method for in-situ high precision sulfur isotope ratio measurmeents by laser ablation multiple collector ICP-MS. Ninth Annual V.M. Goldschmidt Conference, Abst. #7393. [31] Mason P., Kaspers K. & Van Bergen M. J. (1999b) Determination of sulphur isotope ratios and concentrations using ICP-MS incorporating hexapole ion optics. J. Anal. At. Spectrom., 14: 1067-1084. [26, 31] Matheney R. K. & Knauth L. P. (1989) Oxygen-isotope fractionation between marine biogenic silica and seawater. Geochim. Cosmochim. Acta, 53: 3207-3214. [2] Mathieu H. J. & Leonard D. (1998) Use of post-ionization techniques to complement SIMS analysis. A review with practical aspects. High Temp. Materials Proc., 17: 29-44. [33] Matschat R., Czerwensky M., Pattberg S. & Heinrich H.-J. (2001) The certification of high purity metals as primary referencve materials - A challenge for multi-element trace analysis. Phys. Stat. Solidi, A 189: 107-122. [42] Matschat R., Czerwensky M., Pattberg S., Heinrich H.-J. & Tutschku S. (2002) High purity metals as primary calibration materials for element analysis - Their importance and their certification. Mater. Trans. (JIM), 43: 90-97. [42] Matsuhisa Y., Goldsmith J. R. & Clayton R. N. (1978) Mechanisms of hydrothermal crystallization of quartz at 250~ and 15 kbar. Geochim. Cosmochim. Acta, 42: 173-182. [45, 46] Matsuhisa Y., Goldsmith J. R. & Clayton R. N. (1979) Oxygen isotopic fractionation in the system quartz-albite-anorthite-water. Geochim. Cosmochim. Acta, 43: 1131-1140. [46] Matsumi Y., Kishigami M., Tanaka N., Kawasaki M. & Inoue G. (1998) Isotope 180 / 160 ratio measurements of water vapor by use of photoacoustic spectroscopy. Appl. Opt., 37: 6558-6562. [34] Matsuo S. (1991) Technical problems related to the precise measurements of isotope ratios of H20 and CO2 extracted from fluid inclusions. International Institute for Geothermal Research, Pisa, Italy: 383-408.
[3]
Mattauch J. & Herzog R. (1934) Uber einen neuen Massenspektographen. Z. Physik, 89: 786-795. [38] Mattey D. P., Carr R. H., Wright I. P. & Pillinger C. T. (1984) Carbon isotopes in submarine basalts. Earth Planet. Sci. Lett., 70: 196-206. [3] Mattey D. P., Exley R. A. & Pillinger C. T. (1989) Isotopic composition of CO2 and dissolved carbon species in basalt glass. Geochim. Cosmochim. Acta, 53: 2377-2386. [3] Mattey D. P. (1991) Carbon dioxide solubility and carbon isotope fractionation in basaltic melt. Geochim. Cosmochim. Acta, 55: 3467-3473. [46] Mattey D. & Macpherson C. (1993) High-precision oxygen isotope microanalysis of ferromagnesian minerals by laser-fluorination. Chem. Geol. (Isot. Geosci. Sect.), 105: 305-318. [20] Mattey D., Jackson D. H., Harris N. B. W. & Kelley S. (1994) Isotopic constraints on fluid infiltration from an eclogite facies shear zone, Holsenoy, Norway. Metamorph. Geol., 12: 311-325. [16] Matthews D. E. & Hayes J. M. (1978) Isotope-ratio-monitoring gas chromatography-mass spectrometry. Anal. Chem., 50: 1465-1473. [11, 20, 22, 38] Matthews A. & Beckinsale R. D. (1979) Oxygen isotope equilibration systematics between quartz and water. Am. Mineralogist, 64: 232-240. [46] Matthews A., Goldsmith J. R. & Clayton R. N. (1983a) On the mechanisms and kinetics of oxygen isotope exchange in quartz and feldspar at elevated temperatures and pressures. Geol. Soc. Am. Bull., 94: 396-412. [46]
1128
References
Matthews A., Goldsmith J. R. & Clayton R. N. (1983b) Oxygen isotope fractionatios involving pyroxenes: the calibration of mineral-pair geothermometers. Geochim. Cosmochim. Acta, 47: 631-644. [46] Matthews A., Palin J. M., Epstein S. & Stolper E. M. (1994) Experimental study of 180 / 160 partitioning between crystalline albite, albitic glass, and CO2 gas. Geochim. Cosmochim. Acta, 58(23): 5255-5266. [46] Matthews A., Ayalon A. & Bar-Matthews M. (2000) D / H ratios of fluid inclusions of Soreq cave (Israel) speleothems as a guide to the Eastern Mediterranean Meteoric Line relationships in the last 120 ky. Chem. Geol., 166: 183-191. [3] Matthews A., Zhu X.-K. & O'Nions K. (2001) Kinetic iron stable isotope fractionation between iron (-II) and (-III) complexes in solution. Earth Planet. Sci. Lett., 192: 81-92. [46] Matucha M., Jockisch W., Verner P. & Anders G. (1991) Isotope effect in gas-liquid-chromatography of labelled compounds. J. Chromatogr., 588: 251-258. [8] Matucha M. (1995). Isotope effects (IEs) in gas chromatography (GC) of labelled compounds (LCs). In: Synthesis and applications of isotopically labelled compounds. J. Allen (ed.). John Wiley & Sons Ltd.: 489-494. [8] Maubert C., Gu6rin C., Mabon F. & Martin G. J. (1988) D6termination de l'origine de la vanilline par analyse multidimensionnelle du fractionnement isotopique naturel sp6cifique de l'hydrog6ne. Analusis, 16(8): 434-439. [5] Mauersberger K. (1977) Mass spectrometer beam system for applications in the stratosphere. Rev. Sci. Instrum., 48: 1169-1173. [18] Mauersberger K. & Finstad R. (1979) Further laboratory studies and stratospheric flight of a mass spectrometer beam system. Rev. Sci. Instrum., 50: 1612-1617. [18] Mauersberger K. (1981) Measurement of heavy ozone in the stratosphere. Geophys. Res. Lett., 8(8): 935937.[18] Mauersberger K. (1987) Ozone isotope measurements in the stratosphere. Geophys. Res. Lett., 14(1): 8083.[18] Mauersberger K., Morton J., Schueler B., Stehr J. & Anderson S. M. (1993) Multi-isotope study of ozone: Implications for the heavy ozone anomaly. Geophys. Res. Lett., 20: 1031-1034. [18] Mauersberger K., Erbacher B., Krankowski D., Giinther J. & Nickel R. (1999) Ozone isotope enrichment: Isotopomer-specific rate coefficients. Science, 283: 370-372. [18] Mauersberger K., L~immerzahl P. & Krankowski D. (2001) Stratospheric ozone isotope enrichments revisited. Geophys. Res. Lett., 28(16): 3155-3158. [18] May W., Parris R., Beck C., Fassett J., Greenberg R., Guenther F., Kramer G., Wise S., Gills T., Colbert J., Gettings R. & MacDonald B. (2000) Definitions of terms and modes used at NIST for value-assignment of reference materials for chemical measurements. NIST, Special Publ. 260-136:16 p. [43] Mayer B. (1993) Untersuchungen zur Isotopengeochemie des Schwefels in Waldb6den und neu gebildetem Grundwasser unter Wald. GSF-Forschungszentrum ffir Umwelt und Gesundheit, Neuherberg, Germany. GSF-Bericht Rep. 2/93. [26] Mayer B., Feger K. H., Giesemann A. & J~iger H.-J. (1995) Interpretation of sulfur cycling in two catchments in the Black Forest (Germany) using stable sulfur and oxygen isotope data. Biogeochemistry, 30: 31-58. [26] Mayer B. & Krouse H. R. (1996) Prospects and limitations of an isotope tracer technique for understanding sulfur cycling in forested and agro-ecosystems. Isotopes Environ. Health Stud., 32: 191-201. [26] Mayer B., Bollwerk S. M., Mansfeldt T., Hiitter B. & Veizer J. (2001) The oxygen isotope composition of nitrate generated by nitrification in acid forest floors. Geochim. Cosmochim. Acta, 65: 2743-2756. [46] Mayer B., Boyer E., Goodale C., Jawoarski N., van Bremen N., Howarth R. W., Seitzinger S., Billen G., Lajtha K., Nadelhoffer K., Van Dam D., Hetling L., Nosil M., Paustian K. & Alexander R. (2002) Sources of nitrate in rivers draining 16 major watersheds in the northeastern US: Isotopic Constraints. Biogeochem., 57/58: 86-90. [15] Maynard D. G. & Addison P. A. (1985) Extraction and colorimetric determination of elemental sulfur in organic horizons of forest soils. Can. J. Soil Sci., 65: 811-813. [26] Mayne Y. (1957) Natural variation in the nitrogen isotope ratio in igneous rocks. Geochim. Cosmochim. Acta, 12: 185-189. [16]
References
1129
Mazany T., Lerman J. C. & Long A. (1980) Carbon-13 in tree-ring cellulose as an indicator of past climates. Nature, 287: 432-435. [24] McArdle N. C. & Liss P. S. (1995) Isotopes and atmospheric sulfur. Atmos. Environ., 29(18): 2553-2556. [26] McArthur J. M., Coleman M. L. & Bremner J. M. (1980) Carbon and oxygen isotopic composition of structural carbonate in sedimentary francolite. J. Geol. Soc. London, 137: 669-673. [22] McArthur J. M. & Herczeg A. (1990) Diagenetic stability of the isotopic composition of phosphateoxygen: palaeoenvironmental implications. In: Phosphorite research and development. A. J. G. Notholt & I. Jarvis (eds.). Geol. Soc. Am., Vol. 52: 119-124. [22]. McBain J. W. (1932) The sorption of gases and vapors by solids. Rutledge and Sons, London. [36] McCarroll D. & Pawellek F. (1998) Stable carbon isotope ratios of latewood cellulose in Pinus sylvestris from northern Finland: variability and signal strength. The Holocene, 8: 693-702. [24] McCarthy M. D. B., Newton R. J. & Bottrell S. H. (1998) Oxygen isotopic compositions of sulphate from coals: implications for primary sulphate sources and secondary weathering processes. Fuel, 77: 677682. [26] McConnaughey T. A., Burdett J., Whelan J. F. & Paull C. K. (1997) Carbon isotopes in biological carbonates: Respiration and photosynthesis. Geochim. Cosmochim. Acta, 61: 611-622. [46] McConville C., Kalin R. M. & Flood D. (1999) Direct equilibration of soil water for 6180 analysis and its application to tracer studies. Rapid Commun. Mass Spectrom., 13: 1339-1345. [1, 2] McCord B. R. & Taylor J. W. (1986) Computer-controlled detection system for high precision isotope ratio measurements. Anal. Chem., 58: 2589-2591. [39] McCrea J. M. (1950) On the isotopic chemistry of carbonates and a paleotemperature scale. J. Chem. Phys., 18: 849-857. [10, 12, 40, 46] McCready R. G. L., Grinenko V. A. & Krouse H. R. (1980) Sulfur isotope fractionation by Proteus vulgaris and Salmonella heidelberg during the reduction of thiosulfate. Can. J. Microbiol., 26: 1173-1177. [26] McIlvane T. C. (1921) A buffer solution for calorimetric comparison. J. Biol. Chem., 49: 183-186. [28] McInteer B. B. & Montoya J. G. (1981) Automation of a mass spectrometer for nitrogen isotope analysis. In: Recent developments in mass spectrometry in biochemistry, medicine and environmental research. A. Frigerio (ed.). Elsevier Scientific PubI., Amsterdam, the Netherlands. Vol. 7: 343-349. [15] McKeegan K. D., Walker R. M. & Zinner E. (1985) Ion microprobe isotopic measurements of individual interplanetary dust particles. Geochim. Cosmochim. Acta, 49: 1971-1987. [25] McKeegan K. D. (1987) Ion microprobe measurements of H, C, O, Mg, and Si isotopic abundances in individual interplanetary dust particles. PhD Thesis (unpublished), Washington University, St. Louis, Missouri, USA: 187 p. [30] McKeegan K. D., Leshin L. A., Russell S. S. & MacPherson G. J. (1998) Oxygen isotopic abundances in calcium-aluminum-rich inclusions from ordinary chondrites: Implications for nebular heterogeneity. Science, 280: 414-418. [30] McKeegan K. D., Chaussidon M. & Robert F. (2000) Incorporation of short-lived 10Be in a calciumaluminum-rich inclusion from the Allende meteorite. Nature, 289: 1334-1337. [30] McKibben M. A. & Eldridge C. S. (1989) Sulfur isotopic variations among minerals and aqueous species in the Salton Sea geothermal system: A SHRIMP ion microprobe and conventional study of active ore genesis in a sediment-hosted environment. Am. J. Sci., 289: 661-707. [26] McKibben M. A. & Eldridge C. S. (1995) Microscopic sulfur isotope variations in ore minerals from the Viburnum Trend, Southeast Missouri: A SHRIMP study. Econ. Geol., 90(2): 228-245. [26] McKibben M. A. & Riciputi L. R. (1998) Sulfur isotopic ratios determined by ion microprobe. In: Applications of microanalytical techniques to understanding mineralizing processes. M. A. McKibben, W. C. Shanks III & W. I. Ridley (eds.). Society of Economic Geologists, Littleton, CO, USA. Reviews in Econ. Geol., Vol. 7: 121-139. [20, 26] McKibben M. A.& Shanks W. C. (1998) Applications of microanalytical techniques to understanding mineralizing processes. Reviews in economic geology. C. M. Lesher (series ed). Pubco, Vol. 7:263 p.
[30]
1130
References
McKinney C. R., McCrea J. M., Epstein S., Allen H. A. & Urey H. C. (1950) Improvements in mass spectrometers for the measurement of small differences in isotopic abundance ratios. Rev. Sci. Instrum., 21: 724-730. [28, 38, 40] McLelland J. W., Valiela I. & Michener R. H. (1997) Nitrogen-stable isotope signatures in estuarine food webs: A record of increasing urbanization in coastal watersheds. Limnol. Oceanogr., 42: 930-937. [9] McMahon P. B. & Chapelle F. H. (1991) Geochemistry of dissolved inorganic carbon in a Coastal Plain aquifer. 2. Modelling carbon sources, sinks and 613C evolution. J. Hydrol., 127: 109-135. [10] McMahon P. B. & Bohlke J. K. (1996) Denitrification and mixing in a stream-aquifer system; effects on nitrate loading to surface water. J. Hydrol., 186(1-4): 105-128. [29] McMahon A. W., Gilmour J. D., Hernandez M. B. & Rateitzak M. (1998) A RIS-TOF instrument for the measurement of ultratrace quantities of uranium and plutonium. Resonance Ionization Spectroscopy, Ninth International Symposium. J.C. Vickerman, I. Lyon, N.P. Lockyer & J.E. Parks (eds.). AIP Press Woodbury, AIP Conference Proceedings 454: 269-274. [33] McManus J. B., Zahniser M. S., Nelson D. D., Williams L. R. & Kolb C. E. (2002) Infrared laser spectrometer with balanced absorption for measurement of isotopic ratios of carbon gases. Spectrochim. Acta, A58: 2465-2480. [34] McMullen C. C., Cragg C. B. & Thode H. G. (1961) Absolute ratios of 11B/10B in Searles Lake borax. Geochim. Cosmochim. Acta, 23: 147-149. [7] McNeal J. M. & Balistrieri L. S. (1989) Geochemistry and occurence of selenium; an overview. In: Selenium in agriculture and the environment. L. W. Jacobs (ed.). Soil Science Society of America: 1-13. [29] Mehra O. P. & Jackson M. L. (1960) Iron oxide removal from soils and clays by a dithionite-citrate system buffered by sodium bicarbonate. 7th Natl. Conf. on Clays and Clay Minerals. A. Swineford (ed.). Pergamon Press, Washington DC: 317-327. [2] Mei Y. & Harrison W. W. (1991) Getters as plasma reagents in glow discharge mass spectrometry. Spectrochim. Acta, B 46: 175-182. [35] Meier A. L. (1982) Determination of lithium isotopes at natural abundance levels by atomic absorption spectrometry. Anal. Chem., 54: 2158-2161. [6] Meier A. & Notholt J. (1996) Determination of the isotopic abundances of heavy 03 as observed in arctic ground-based FTIR-spectra. Geophys. Res. Lett., 23: 551-554. [18] Meier L. P. & Menegatti A. P. (1997) A new, effcient, one-step method for the removal of organic matter from clay-containing sediments. Clay Minerals, 32: 557-563. [2] Meier-Augenstein W., Brand W., Hoffmann G. F. & Rating D. (1994a) Bridging the information gap between isotope ratio mass spectrometry and conventional mass spectrometry. Biol. Mass Spectrom., 23: 376-378. [8] Meier-Augenstein W., Hess A., Hoffmann G. F. & Rating D. (1994b) Evaluation of water removal and memory effect in 13CO2 breath tests by isotope ratio mass-spectrometry. Isot. Environ. Health Stud., 30: 349-358. [36] Meier-Augenstein W. (1995) Online recording of 13C / 12C ratios and mass-spectra in one gas-chromatographic analysis. HRC-J. High Resolut. Chromatogr., 18: 28-32. [8] Meier-Augenstein W., Watt P. W. & Langhans C. D. (1996) Influence of gas-chromatographic parameters on measurment of 13C / 12C isotope ratios by gas-liquid-chromatography combustion isotope ratio mass-spectrometry. I. J. Chromatogr., A 752: 233-241. [8] Meier-Augenstein W. (1997a) The chromatographic side of isotope ratio mass spectrometry: Pitfalls and answers. LC/GC, 15: 244-252. [8] Meier-Augenstein W. (1997b) A reference gas inlet module for internal isotopic calibration in high precision gas chromatography/combustion-isotope ratio mass spectrometry. Rapid Comm. Mass Spectrom., 11: 1775-1780. [8] Meier-Augenstein W. (1999) Applied gas chromatography coupled to isotope ratio mass spectrometry. J. Chromatography, A. 842: 351-371. [8] Meier-Augenstein W. (2002) Stable isotope analysis of fatty acids by gas chromatography-isotope ratio mass spectrometry. Anal. Chem. Acta, 465: 63-79. [8]
References
1131
Meijer H. A. J. (1995) The isotopic composition of the Groningen GS-19 and GS-20 pure CO2 standards. In: Reference and intercomparison materials for stable isotopes of light elements. IAEA-TECDOC825, IAEA, Vienna: 81-83. [43] Meijer H. A. J. & Li W. J. (1998) The use of electrolysis for accurate 6170 and 6180 isotope measurements in water. Isot. Environ. Health Stud., 34(1): 349-369. [1, 43, 44, 45] Meijer H. A. J., Neubert R. E. M. & Visser G. H. (2000) Cross contamination in dual inlet isotope ratio mass spectrometers. Int. J. Mass Spectrom., 198: 45-61. [26, 34, 43] Melander L. & Saunders W. (1980) Reaction rates of isotopic molecules. J. Wiley-Interscience, New York.
[8]
Menegatti A. P., Friih-Green G. L. & Stille P. (1999) Removal of organic matter by disodium peroxodisulphate: effects on mineral structure, chemical composition and physicochemical properties of some clay minerals. Clay Minerals, 34: 247-257. [2] Merritt D. A., Brand W. A. & Hayes J. M. (1994) Isotope-ratio-monitoring gas chromatography-mass spectrometry: methods for isotopic calibration. Org. Geochem., 21: 573-583. [8, 38] Merritt D. A. & Hayes J. M. (1994) Factors controlling precision and accuracy in isotope-ratio-monitoring mass spectrometry. Anal. Chem., 66(14): 2336-2347. [8, 14, 20, 38] Merritt D. A., Freeman K. H., Ricci M. P., Studley S. A. & Hayes J. M. (1995a) Performance and optimization of a combustion interface for isotope ratio monitoring gas chromatography/mass spectrometry. Anal. Chem., 67: 2461-2473. [8] Merritt D. A., Hayes J. M. & DesMarais D. J. (1995b) Carbon isotopic analysis of atmospheric methane by isotope-ratio-monitoring gas-chromatography mass-spectrometry. J. Geophys. Res., 100(D1): 1317-1326. [8, 14] Mestres M., Busto O. & Guasch J. (2000) Analysis of organic sulfur compounds in wine aroma. J. Chromatogr., A 881: 569-581. [26] Metges C. C., Petzke K.-J. & Hennig J. (1996) Gas chromatography/combustion/isotope ratio mass spectrometric comparison of N-acetyl- and N-pivaloyl amino acid esters to measure 15N isotopic abundances in physiological samples: a pilot study on amino acid synthesis in the upper gastrointestinal tract of minipigs. J. Mass Spetrom., 31: 367-376. [8, 9] Metges C. C. & Petzke K. J. (1997) Measurement of 15N / 14N isotopic composition in individual plasma free amiono acids of human adults at natural abundance by gas chromatography-combustion isotope ratio mass spectrometry. Anal. Biochem., 247: 158-164. [8] Metges C. C. & Petzke K. J. (1999) The use of GC-C-IRMS for the analysis of stable isotope enrichment in nitrogenous compounds. In: Methods for investigation of amino acid and protein metabolism. A. E. E1-Khoury (ed.). CRC Press, Boca Raton, Florida: 121-134. [8] Meyer M., Lancoursi6re J., Nahon L., Gisselbrecht M., Morin P. & Larzilli6re M. (1996) Resonant 2photon-ionization of Xe. Resonance Ionization Spectroscopy 1996, Eighth International Symposium. N Winograd & J.E. Parks (eds.). AIP Press Woodbury, NY, AIP Conference Proceedings 388: 357-360. [33] Michalski G. M. & Thiemens M. (2000) Mass independent fractionation in nitrate aerosols. EOS, Trans. AGU, 81:F120 (Abstract). [15] Michalski G., Savarino J., BShlke J. K. & Thiemens M. (2002) Determination of the total oxygen isotopic composition of nitrate and the calibration of a A170 nitrate reference material. Anal. Chem., 74: 4989-4993. [15] Michener R. H. & Schell D. M. (1994) Stable isotope ratios as tracers in marine aquatic food webs. In: Stable isotopes in ecology and environmental science. K. Lajtha & R. H. Michener (ed.). Blackwetl Scientic Publications, Oxford: 138-157. [9] Michiels E. & De Bi6vre P. (1983) Absolute isotopic composition and the atomic weight of a natural sample of lithium. Int. J. Mass Spectrom. Ion Phys., 49: 265-274. [6] Milatz J. M., Kluyver J. C. & Hardebol J. (1951) Determination of isotope ratios, e.g. in tracer work, by infrared absorption method. J. Chem. Phys., 19: 887-888. [34] Miller R. L., Suits A. G., Houston P. L., Toumi R., Mack J. A. & Wodtke A. M. (1994) The "ozone decifit" problem: 02 (X, ve 26) + O(3P) from 226-nm ozone photodissociation. Science, 265: 1831-1838. [18]
1132
References
Miller M. & Pillinger C. T. (1997) An appraisal of stepped heating release of fluid inclusion CO2 for isotopic analysis: A preliminary to 613C characterisation of carbonaceous vesicles at the nanomole level. Geochim. Cosmochim. Acta, 61: 193-205. [3, 13] Miller M. F., Franchi I. A., Sexton A. S. & Pillinger C. T. (1999) High precision 6170 isotope measurements of oxygen from silicates and other oxides: Method and applications. Rapid Commun. Mass Spectrom., 13: 1211-1217. [20, 36] Miller J. B., Mack K. A., Dissly R., White J. W. C., Dlugokencky E. J. & Tans P. P. (2002) Development of analytical methods and measurements of 13C/12C in atmospheric CH4 from the NOAA-CMDL global air sampling network. J. Geophys. Res., 107: doi 10.1029/2001JD000630. [14] Mills G. A. & Urey H. C. (1939) Oxygen exchange between carbon dioxide, bicarbonate ion, carbonate ion and water. J. Am. Chem. Soc., 61: 534. [1] Mills A. & Urey H. C. (1940) The kinetics of isotopic exchange between carbon dioxide, bicarbonate ion, carbonate ion, and water. J. Am. Chem. Soc., 62: 1019-1026. [1] Milton D. M. P. & Hutton R. C. (1993) Investigations into the suitability of using a secondary cathode to analyse glass using glow discharge mass spectrometry. Spectrochim. Acta, B 48: 39-52. [35] Minagawa M., Winter D. A. & Kaplan I. R. (1984) Comparison of Kjeldahl and conmbustion method for measurement of nitrogen isotope ratios in organic matter. Anal. Chem., 56: 1859-1861. [16] Mineev S. D. & Grinenko V. A. (1996) The pressure influence on hydrogen isotopes fractionation in the serpentine-water system. J. Conf. Abstr., 1:404 (Abstract). [46] Mitchell B. D. & Smith B. F. L. (1974) The removal of organic matter from soil extracts by bromine oxidation. J. Soil Sci., 25: 239-241. [2] Mitchell S. (1996) Biological Interactions of Sulfur Compounds. Taylor & Francis. [26] Miyajima T., Yamada Y. & Hanba Y. T. (1995) Determinating the stable isotope ratio of total dissolved inorganic carbon in lake water by GC/C/IRMS. Limnol. Oceanogr., 40: 994-1000. [10] Miyake Y., Shimomura K., Mills A. P. Jr., Marangos J. P. & Nagamine K. (1998) Ultra-sensitive detection of hydrogen isotopes by the Lyman-a RIS. Resonance Ionization Spectroscopy, Ninth International Symposium. J.C. Vickerman, I. Lyon, N.P. Lockyer & J.E. Parks (eds.). AIP Press Woodbury, NY. AIP Conference Proceedings 454: 47-52. [33] Miyazaki T., Wada E. & Hattori A. (1973) Capacities of shallow waters of Sagami Bay for oxidation and reduction of inorganic nitrogen. Deep-Sea Res., 20: 571-577. [15] Mizota C., Domon Y. & Yoshida N. (1992) Oxygen isotope composition of natural phosphates from volcanic ash soils of the Great Rift Valley of Africa and east Java, Indonesia. Geoderma, 5: 111-123. [2] Mizota C. (1996) Pedogeochemistry of hydrogen and oxygen isotope ratios. In: Mass spectrometry of soils. T. W. Boutton & S.-I. Yamasaki (eds.). Marcel Dekker Inc.: 155-176. [2] Mizutani Y. & Rafter T. A. (1969) Oxygen isotopic composition of sulphates - Part 5. Isotopic composition of sulphate in rain water, Gracefield, New Zealand. N. Z. Jl. Sci., 12: 69-80. [26] Moers M. E. C., Jones D. M., Eakin P. A., Fallick A. E., Griffths H. & Larter S. R. (1993) Carbohydrate diagenesis in hypersaline environments: application of GC-IRMS to the stable isotope analysis of derivatives saccharides from surficial and buried sediments. Org. Geochem., 20: 927-933. [9] Moerth C.-M. & Torssander P. (1995) Sulfur and oxygen isotope ratios in sulfate during an acidification reversal study at Lake Garsjon, Western Sweden. Water Air Soil Pollut., 79: 261-278. [26] Mohler F. L. (1960) Isotopic abundance ratios reported for reference samples stocked at the National Bureau of Standards. National Bureau of Standards, Washington. Tech. Note No. 51:8 p. [40] Mojzsis S. J., Arrhenius G. A., McKeegan K. D., Harrison T. M., Nutman A. P. & Friend R. L. (1996) Evidence of life on Earth before 3,800 million years ago. Nature, 384: 55-59. [30] Mojzsis S. J., Harrison T. M. & McKeegan K. D. (2001) Oxygen isotope evidence from ancient zircons for liquid water at the Earth's surface 4,300 Myr ago. Nature, 409: 178-181. [30] Mojzsis S. J., Coath C. D., Greenwood J. P., McKeegan K. D. & Harrison T. M. (2003) Mass-independent isotope effects in Archean (2.5 - 3.8 Ga) sedimentary sulfides determined by ion microprobe analysis. Geochim. Cosmochim. Acta, 67: 1635-1658. [30] Molini-Velsko C. A. (1983) Isotopic composition of silicon in meteorites. Ph.D. Thesis, University of Chicago, Chicago, USA: 174p. [36]
References
113 3
Molini-Velsko C., Mayeda T. K. & Clayton R. N. (1986) Isotopic composition of silicon in meteorites. Geochim. Cosmochim. Acta 50: 2719-2726. [25] Molini-Velsko C., Mayeda T. K. & Clayton R. N. (1987) Silicon isotope systematics during distillation. Lunar Planet. Sci., 18: 657-658. [46] Monaghan J. M., Scrimgeour C. M., Stein W. M., Zhao F. J. & Evans E. J. (1999) Sulphur accumulation and re-distribution in wheat (Triticum aestivum): a study using stable sulphur isotope ratios as a tracer system. Plant Cell Environ., 22, 831-839. [26] Moncaster S. J. & Bottrell S. H. (1991) Extraction of low-level sulphide from groundwaters for sulphur isotope analysis. Chem. Geol. (Isot. Geosci. Sect.), 94: 79-82. [26] Monetti A., Reniero F. & Versini G. (1994) Interregional and interannual variability of ethanol sitespecific deuterium content in Italian wines. Z. Lebensm. Unters. Forsch., 199: 311-316. [5] Monetti A., Versini G., Dalpiaz G. & Reniero F. (1996) Sugar adulterations control in concentrated rectified grape musts by finite mixture distribution analysis of the myo- and scyllo-inositol content and the D / H methyl ratio of fermentative ethanol. J. Agric. Food Chem., 44: 2194-2201. [5] Monin J. C., Pelet R. & Fevrier A. (1978) Geochemical analysis of organic matter extracted from sedimentary rocks. Part 4. Extraction of small quantities of rock. Revue Inst. Franr Pdtrole, 33: 223-240.
[2]
Monster J. (1972) Homogeneity of sulfur and carbon isotope ratios S34 /S 32 and C 13 / C 12 in petroleum. AAPG Bull., 54: 941-949. [26] Montaser A., Minnich M. G., McLean J. A., Liu H., Caruso J. A. & McLeod C. W. (1998) Sample introduction in ICPMS. In: Inductively coupled plasma mass spectrometry. A. Montaser (ed.). WileyVCH: 83-264 [31] Mook W. G. (1968) Geochemistry of the stable carbon and oxygen isotopes of natural waters in the Netherlands. PhD Thesis, University of Groningen. [10] Mook W. G. & Grootes P. M. (1973) The measuring procedure and corrections for the high precision mass spectrometric analysis of isotopic abundance ratios, especially referring to carbon, oxygen and nitrogen. Int. J. Mass Spectrom. Ion Phys., 12: 273-298. [23, 43, 45] Mook W. G., Bommerson J. C. & Staverman W. H. (1974) Carbon isotope fractionation between dissolved bicarbonate and gaseous carbon dioxide. Earth Planet. Sci. Lett., 22: 169-176. [46] Mook W. G. & Jongsma J. (1987) Measurement of the N20 correction for 13C / 12C ratios of atmospheric CO2 by removal of N20. Tellus, 39B: 96-99. [14] Mook W. G. (2000) Environmental isotopes in the hydrological cycle: Principles and applications, Volume I: Introduction: Theory, Methods, Review. International Hydrological Programme IHP-V. UNESCO, Paris. Technical Documents in Hydrology, No.39, Vol.l: 49-51. [40] Mooney-Slater R. C. L. (1962) Polymorphic forms of bismuth phosphate. Z. Kristallogr., 117: 371-385. [22] Moore G. S. M. & Rose H. E. (1973) The structure of powdered quartz. Nature, 242: 187-190. [3] Moore H. (1977) The isotopic composition of ammonia, nitrogen dioxide and nitrate in the atmosphere. Atmos. Environ., 21: 1239-1243. [15] Moore P. B. & Shen J. (1983) An X-ray structural study of cacoxenite, a mineral phosphate. Nature, 306: 356-358. [36] Moran K. K., Mulvaney R. L. & Khan S. A. (2002) A Technique to facilitate diffusions for nitrogen-isotope analysis by direct combustion. Soil Sci. Soc. Am. J., 66: 1008-1011. [15] Moreau-Le Golvan Y., Michelot J.-L. & Boisson J.-Y. (1997) Stable isotope contents of porewater in a claystone formation (Tournemire, France): assessment of the extraction technique and preliminary results. Appl. Geochem., 12: 739-745. [2] Morgan M. D. & Mandernack K. W. (1996) Biogeochemistry of sulfur in wetland peat following 3.5 y of artificial acidification (HUMEX). Environ. Int., 22: 605-610. [26] Moriguti T. & Nakamura E. (1993) Precise lithium isotope analysis by thermal ionization mass spectrometry using lithium phosphate as an ion source. Proc. Japan Acad. Sci.: 69B: 123-128. [6] Moriguti T. & Nakamura E. (1998a) High-yield lithium separation and precise isotopic analysis for natural rock and aqueous samples. Chem. Geol., 145: 91-104. [6, 31] Moriguti T. & Nakamura E. (1998b) Across-arc variation of Li isotopes in lavas and implications for crust/mantle recycling at subduction zones. Earth Planet. Sci. Lett., 163: 167-174. [6]
1134
References
Morrison J., Brockwell T., Merren T., Fourel F. & Phillips A. M. (2001) On-line high-precision stable hydrogen isotopic analyses on nanolitre water samples. Anal. Chem., 73: 3570-3575. [1] Morse A. D., Wright I. P. & Pillinger C. T. (1993) An investigation into the cause of memory effects associated with the conversion of H20 to H2 for D / H measurement. Chem. Geol. (Isot. Geosc. Sect.), 107: 147-158. [1] Mortimer R. J. G. & Coleman M. L. (1997) Microbial influence on the oxygen isotopic composition of diagenetic siderite. Geochim. Cosmochim. Acta, 61: 1705-1711. [46] Morton J., Schueler B. & Mauersberger K. (1989) Oxygen fractionation of ozone isotopes 4803 through 5403. Chem. Phys. Lett., 154: 143-145. [18] Morton J., Barnes J., Schueler B. & Mauersberger K. (1990) Laboratory studies of heavy ozone. J. Geophys. Res., 95: 901-907. [18] Morton R. D. & Catanzaro E. J. (1964) Stable chlorine isotope abundances in apatites from Odegardens Verk, Norway. Norsk Geol. Tiddsk., 44: 307-313. [28] Morville J., Chenevier M., Kachanov A. A. & Romanini D. (2001) Trace gas detection with DFB lasers and cavity ring-down spectroscopy. Proceedings SPIE 4485 (Optical spectroscopic techniques, remote sensing, and instrumentation for atmospheric and space research IV): 236-243. [34] Morville J., Romanini D., Kachanov A. A. & Chenevier, M. (2003) Two schemes for trace gas detection using cavity ring down spectroscopy. Appl. Phys., B (submitted). [34] Mosandl A. (1995) Enantioselective capillary gas-chromatography and stable-isotope ratio massspectrometry in the authenticity control of flavors and essential oils. Food Rev. Int., 11: 597-664. [8] Mosandl A., Braunsdorf R., Bruche G., Dietrich A., Hener U., Karl V., Kopke T., Kreis P., Lohmann D. & Maas B. (1995) New methods to assess authenticity of natural flavors and essential oils. ACS Symp. Ser., 596: 94-112. [8] Mosandl A. (1997) Progress in the authenticity assessment of wines and spirits. Analusis, 25: M31-M38.
[81
Motz J. E., Edwards T. W. D. & Buhay W. M. (1997) Use of nickel-tube pyrolysis for hydrogen-isotope analysis of water and other compounds. Chem. Geol., 140: 145-149. [1] Moum J. & Rosenqvist I. T. (1958) Hydrogen (protium)-deuterium exchange in clays. Geochim. Cosmochim. Acta, 14: 250-252. [2] Moussa I., Naulet N., Martin M. L. & Martin G. J. (1990) A site-specific and multielement approach to the determination of liquid-vapor isotope fractionation parameters. The case of alcohols. J. Phys. Chem., 94(21): 8303-8309. [5] Mroz E. J. (1993) Deuteromethanes: potential fingerprints of the sources of atmospheric methane. Chemosphere, 26: 45-53. [34] Muehlenbachs K. & Kushiro I. (1974) Oxygen isotope exchange and equilibrium of silicates with CO2 and 02. Carnegie Instn. Washington, Year Book, 74: 232-240. [46] Mulitza S., Wolff T., P~itzold J., Hale W. & Wefer G. (1998) Temperature sensitivity of planktic foraminifera and its influence on the oxygen isotope record. Marine Micropaleo., 33: 223-240. [46] Mullane M. V., Waterhouse J. S. & Switsur v. R. (1988) On the development of a novel method for the determination of stable oxygen isotope ratios in cellulose. Int. J. Appl. Radiat. Isot., 39: 1029-1035. [23] Mullis J., Dubessy J., Poty B. & O'Neil J. R. (1994) Fluid regimes during late stages of a continental collision: physical, chemical and stable isotope measurements of fluid inclusions in fissure quartz from a geotraverse through the Central Alps, Switzerland. Geochim. Cosmochim. Acta, 58: 2239-2267.
[3]
Mulvaney R. L. (1986) Comparison of procedures for reducing cross-contamination during steam distillations in nitrogen-15 tracer research. Soil Sci. Am. J., 50: 92-96. [15] Mulvaney R. L., Fohringer C. L., Bojan V. J., Michlik M. M. & Herzog L. F. (1990) A commercial system for automated nitrogen isotope-ratio analysis by the Rittenberg technique. Rev. Sci. Instrum., 61: 897-903. [15] Mulvaney R. L. & Liu Y. P. (1991) Refinement and evaluation of an automated mass spectrometer for nitrogen isotope analysis by the Rittenberg technique. J. Autom. Chem., 13: 273-280. [15] Mulvaney R. L. (1993) Mass spectrometry. In: Nitrogen isotope techniques. R. Knowles & T. H. Blackburn (eds.). Academic Press, San Diego, CA: 11-57. [15]
References
113 5
Mulvaney R. L., Ashraf M., Azam F., Brooks P. D. & Herman D.J. (1994) Evaluation of methods for nitrogen-15 analysis of inorganic nitrogen in soil extracts: I. Steam-distillation methods. Commun. Soil Sci. Plant Anal., 25: 2187-2200. [15] Mulvaney R. L. (1996) Nitrogen- Inorganic forms. In: Methods of soil analysis. Part3. Chemical Methods. D. L. Sparks et al. (eds.). Soil Science Society of America and American society of Agronomy, Madison, Wisconsin: 1123-1184. [15] Mulvaney R. L., Khan S. A., Stevens W. B. & Mulvaney C. S. (1997) Improved diffusion methods for determination of inorganic nitrogen in soil extracts and water. Biol. Fertil. Soils, 24: 413-420. [15] Mulvaney R. L. & Khan S. A. (1999) Use of diffusion to determine inorganic nitrogen in a complex organic matrix. Soil Sci. Soc. Am. J., 63: 240-246. [15] Mulvaney R. L. & Khan S. A. (2001) Diffusion methods to determine different forms of nitrogen in soil hydrolysates. Soil Sci. Soc. Am. J., 65: 1284-1292. [15] Mulvaney R. L., Khan S. A., Hoeft R. G. & Brown H. M. (2001) A soil organic nitrogen fraction that reduces the need for nitrogen fertilization. Soil Sci. Soc. Am. J., 65: 1164-1172. [15] Murnick D. E. & Peer B. J. (1994) Laser-based analysis of carbon isotope ratios. Science, 263: 945-947. [34] Murnick D. E., Colgan M. J. & Stoneback D. N. (1998) Optogalvanic effect isotope ratio analysis in carbon dioxide. In: Synthesis and applications of isotopes and isotopically labeled compounds. J. R. Heys & D. G. Melillo (eds.). John Wiley and Sons, Chichester: 111-115. [34] Murnick D. E. (2001) Laser assisted ratio analysis - An alternative to GC/IRMS for CO2. In: New approaches for stable isotope ratio measurements. Proceedings of an Advisory Group meeting held in Vienna, 20-23 September 1999. IAEA-TECDOC-1247, IAEA, Vienna: 15-22. [34, 40] Murphy B. F. (1947) The temperature variation of thermal diffusion factors for binary mixtures of H, D and He. Phys. Rev., 72: 834-837. [38] Murty S. V. S. & Goswami J. N. (1992) Nitrogen, noble gases and nuclear tracks in lunar meteorites MAC 88104/105. Proc. Lunar Planet. Sci., 22: 225-237. [17] Musashi M., Markl G. & Kreulen R. (1998) Stable chlorine-isotope analysis of rock samples: new aspects of chlorine extraction. Anal. Chem. Acta, 362: 261-269. [28] Miiller P., Bushaw B. A., Blaum K., Ntirtersh~iuser W., Trautmann N. & Wendt K. (1998) Ultratrace determination of hte long-lived isotope 41Ca by narrowband CW-RIMS. Resonance Ionization Spectroscopy, Ninth International Symposium. J.C. Vickerman, I. Lyon, N.P. Lockyer & J.E. Parks (eds.). AIP Press Woodbury, NY. AIP Conference Proceedings 454: 73-78. [33] Myers W. R. & DeLong W. B. (1984) Fluorine corrosion: High-temperature attack on metals by fluorine and hydrogen fluoride. Chem. Eng. Prog., 44: 359-362. [20] Myhre G., Highwood E. J., Shine K. P. & Stordal F. (1998) New estimates of radioactive forcing due to well-mixed greenhouse gases. Geophys. Res. Lett., 25: 2715-2718. [14] Nabelek P. I., O'Neil J. R. & Papike J. J. (1983) Vapor phase exsolution as a controlling factor in hydrogen isotope variation in granitic magmas: the Notch Peak granitic stock, Utah. Earth Planet. Sci. Lett., 66: 137-150. [46] Nadeau S., Pineau F., Javoy M. & Francis-Donald M. (1990) Carbon concentrations and isotopic ratios in fluid-inclusion-bearing upper-mantle xenoliths along the northwestern margin of North America. Chem. Geol., 81: 271-297. [3] Naden J., Kilias S. P., Leng M. J., Cheliotis I. & Sheperd T. J. (2003) Do fluid inclusions preserve 6180 values of hydrothermal fluids in epithermal systems over geological time? Evidence from paleo- and modern geothermal systems, Milos island, Aegean Sea. Chem. Geol., 197: 143-159. [3] Nakai N. & Jensen M. L. (1967) Sources of atmospheric sulfur compounds. Geochem. J., 1: 199-210. [26] Nakamura E., Ishikawa T., Birck J.-L. & All6gre C. J. (1992) Precise boron isotopic analysis of natural rock samples using a boron-mannitol complex. Chem. Geol. (Isot. Geosci. Sect.), 94: 193-204. [40] Nakamura T., Kojima S., Ohta H., Oda H., Ikeda A., Okuno M., Yokota K., Mizutani Y. & Kretschmer W. (1998) Isotopic analyses and cycling of dissloved inorganic carbon at Lake Biwa, central Japan. Radiocarbon, 40: 933-944. [10] Nakano T. & Nakamura E. (2001) Boron isotope geochemistry of metasedimentary rocks and tourmalines in a subduction zone metamorphic suite. Phys. Earth Planet. Int., 127: 233-252. [30]
1136
References
Nakazawa T., Morimoto S., Aoki S. & Tanaka M. (1997) Temporal and spatial variations of the carbon isotopic ratio of atmospheric carbon dioxide in the western Pacific region. J. Geophys. Res., 102: 12711285. [141 Nascimento C., Atekwana E. A. & Krishnamurthy R. V. (1997) Concentrations and isotope ratios of dissolved inorganic carbon in groundwaters of denitrifying environments. Geophys. Res. Lett., 24: 1511-1514. [10] Navon O. & Wasserburg G. J. (1985) Self shielding in 02 - A possible explanation for oxygen isotopic anomalies in meteorites? Earth Planet. Sci. Lett., 73: 1-16. [18] Neal R. H. & Spositom G. (1989) Selenate adsorption on alluvial soils. Soil Sci. Soc. Am. J., 53(1): 70-74. [29] Nealson K. H. (1997) Sediment bacteria: who's there, what are they doing, and what's new? Ann. Rev. Phys. Chem., 25: 403-434. [46] Nehring N. L., Bowen P. A. & Truesdell A. H. (1977) Techniques for the conversion to carbon dioxide of oxygen from dissolved sulfate in thermal waters. Geothermics, 5: 63-66. [20] Nelms S. M., Qu6tel C., Prohaska T., Vogl J. & Taylor P. D. P. (2001) Evaluation of detector dead time calculation models for ICP-MS. J. Anal. At. Spectrom., 16: 333-338. [32, 37] Nelson D. E., Angerbjorn A., Liden K. & Turk I. (1998) Stable isotopes and the metabolism of the European cave bear. Oecologia, 116: 177-181. [9] Nelson S.T. & Dettman D. (2001) Improving hydrogen isotope ratio measurements for on-line chromium reduction systems. Rapid Commun Mass Spectr., 15: 2301-2306. [1] Nesbitt B. E., Muehlenbachs K. & Murowchick J. B. (1989) Genetic implications of stable isotope characteristics of mesothermal Au deposits and related Sb and Hg deposits in the Canadian Cordillera. Econ. Geol., 84: 1489-1506. [3] Nesbitt B. E. & Muehlenbachs K. (1995) Geochemical studies of the origins and effects of synorogenic crustal fluids in the southern Omineca Belt of British Columbia, Canada. Geol. Soc. Am. Bull., 107: 1033-1050. [3] Nesbitt B. E. & Muehlenbachs K. (1997) Paleo-hydrogeology of late proterozoic units of southeastern Canadian Cordillera. Am. J. Sci., 297: 359-392. [3] N e w m a n L., Forrest J. & Manowitz B. (1975) The application of an isotopic ratio technique to a study of the atmospheric oxidation of sulphur dioxide in the plume from coal-fired power plant. Atmos. Environ., 9: 959-968. [26] Newton R. J., Bottrell S. H., Dean S. P., Hatfield D. & Raiswell R. (1995) An evaluation of the use of the chromous chloride reduction method for isotopic analysis of pyrite in rocks and sediment. Chem. Geol. (Isot. Geosci. Sect.), 125: 317-320. [26] Nicolussi G. K., Pellin M. J., Lykke K. R., Trevor J. L., Mencer D. E. & Davis A. M. (1996) Surface analysis by SNMS: Femtosecond laser postionization of sputtered and laser desorbed atoms. Surf. Interface Anal., 24: 363-370. [33] Nicolussi G. K., Davis A. M., Pellin M. J., Lewis R. S., Clayton R. N. & Amari S. (1997a) The isotopic composition of molybdenum in single presolar carbide grains. Lunar and Planetary Science Conference, Ext. Abs. XXVIIh 1021-1022 (Abstract). [33] Nicolussi G. K., Davis A. M., Pellin M. J., Lewis R. S., Clayton R. N. & Amari S. (1997b) S-process zirconium in individual presolar silicon carbide grains. Lunar and Planetary Science Conference, Ext. Abs. XXVIII: 1023-1024 (Abstract). [33] Nicolussi G. K., Pellin M. J., Lewis R. S., Davis A. M., Amari S. & Clayton R. B. (1998) Molybdenum isotopic composiiton of individual presolar silicon carbide grains from the Murchison meteorite. Geochim. Cosmichim. Acta, 62: 1093-1104. [31] Nief G. & Botter R. (1959) Mass spectrometric analysis of simple hydrogen compounds. Adv. Mass Spectrom., 1: 515-525. [1] Nielsen H. (1984) The standard situation for sulfur isotopes. In: Advisory Group Meeting on stable isotope reference samples for geochemical and hydrological investigations, 19-21 Sept., 1983. R. Gonfiantini (ed.). IAEA, Vienna: 73-77. [40] Nier A. O. & Hanson E. E. (1936) A mass-spectrographic analysis of the ions produced in HC1 under electron impact. Phys. Rev., 50: 722-726. [28]
References
1137
Nier A. O. (1940) A mass spectrometer for routine isotope abundance measurements. Rev. Sci. Instrum., 11: 212-216. [38] Nier A. O. (1947) A mass spectrometer for isotope and gas analysis. Rev. Sci. Instrum., 18: 398-411. [28,
38, 40] Nier A. O., Ney E. P. & Inghram M. G. (1947) A null method for the comparison of two ion currents in a mass spectrometer. Rev. Sci. Instrum., 18: 294. [28] Nier A. O. (1950) A redetermination of the relative abundances of the isotopes of carbon, nitrogen, oxygen, argon and potassium. Phys. Rev., 77: 789-793. [40] Nier A. O. (1955) Determination of isotopic masses and abundances by mass spectrometry. Science, 121: 737-744. [28] Niewenhuize J., Maas Y. E. M. & Middelburg J. J. (1994) Rapid analysis of organic carbon and nitrogen in particulate materials. Marine Chem., 45: 217-224. [9] Nirel P. M. V. & Lutz T. M. (1991) Plasma-source mass spectrometry (ICP-MS) application to multielement analysis in environmental samples, a practical evaluation. J. Trace Microprobe Techn., 9: 95106. [28] Nishio Y. & Nakai S. (2002) Accurate and precise lithium isotopic determinations of igneous rock samples using multi-collector inductively coupled plasma mass spectrometry. Anal. Chim. Acta, 456: 271-281. [6] Nishizawa K., Maeda Y., Kawashiro F., Fujii T. & Yamamoto T. (1998) Contribution of nuclear size and shape, nuclear mass, and nuclear spin to enrichment factors of zinc isotopes in a chemical exchange reaction. Sep. Sci. Technol., 33: 2102-2112. [31] NIST Certificate of analysis (1999) Standard reference material 951, boric acid standard. National Institute of Standards and Technology, Gaithersburg, USA: 2 p. [40] NIST Certificate of analysis (2001) Standard reference material 975a, isotopic standard for chlorine. National Institute of Standards and Technology, Gaithersburg, USA: 2 p. [40] Nitz S., Weinreich B. & Drawert F. (1992) Multidimensional gas-chromatography- isotope ratio massspectrometry (Mdgc-irms). A system description and technical requirements. HRC-J. High Resolut. Chromatogr., 15: 387-391. [8] Niu H. & Houk R. S. (1996) Fundamental aspects of ion extraction in inductively coupled plasma mass spectrometry. Spectrochim. Acta, B 51: 779-815. [35] Nomura M., Okamoto M. & Kakihana H. (1973) Determination of boron isotope ratio by the surface ionization method. Shitsuryo Bunzeki, 21:277-281 (in Japanese). [7] Norman A. L. & Krouse H. R. (1992) Stable isotope studies of atmospheric sulphur: comparison of Calgary, Canada and Bermuda. Papers from the 9th World Clean Air Congress: Towards the Year 2000. IU-8.01. [26] Norman D. I. & Sawkins F. J. (1987) Analysis of volatiles in fluid inclusions by mass spectrometry. Chem. Geol., 61: 1-10. [3] Northrop D. A. & Clayton R. N. (1966) Oxygen-isotope fractionations in systems containing dolomite. J. Geology, 74: 174-196. [46] Noto M. & Kusakabe M. (1995) The reactivity of zinc used for preparation of water samples for D / H isotope analysis. Okayama University. Tech Report, Inst. Study Earth's Interior, Ser. B, No. 14 (in Japanese). [1] Novak M., Adamova M. & Milicic J. (2003a) Sulfur metabolism in polluted shagnum peat bogs: a combined 34S-35S-210Pb study. Water Air Soil Pollut., 3: 181-200. [26] Novak M., Buzek F., Harrison A. F., Prechova E., Jackova I. & Fottova D. (2003b) Similarity between C, N and S stable isotope profiles in European spruce forest soils: implications for the use of ~34S as a tracer. Appl. Geochem., 18: 765-779. [26] Ni~rtersh~iuser W., Blaum K., Bushaw B. A., Miiller P., Trautmann N. & Wendt K. (1998) Lineshapes and optical selectivity in double-resonance RIMS measurements. Resonance Ionization Spectroscopy, Ninth International Symposium. J.C. Vickerman, I. Lyon, N.P. Lockyer & J.E. Parks (eds.). AIP Press Woodbury, NY. AIP Conference Proceedings 454: 171-176. [33] Norgaard J. V., Valkiers S., Van Nevel L., Vendelbo E., Papadakis I., Br6as O. & Taylor P. D. P. (2002) The international measurement evaluation programme, IMEP-8: carbon and oxygen isotope ratios in C02. Anal. Bioanal. Chem., 374: 1147-1154. [43]
1138
References
Nriagu J. O. & Coker R. D. (1983) Sulphur in sediments chronicles past changes of lake acidification. Nature, 303: 692-694. [26] Nriagu J. O. & Soon u K. (1985) Distribution and isotopic composition of sulfur in lake sediments of Northern Ontario. Geochim. Cosmochim. Acta, 49: 823-834. [26] Nriagu J. O., Holdway D. A. & Coker R. D. (1987) Biogenic sulfur and the acidity of rainfall in remote areas of Canada. Science, 237: 1189-1192. [26] Nriagu J. O., Coker R. D. & Barrie L. A. (1991) Origin of sulphur in Canadian Artic haze from isotope measurements. Nature, 349: 142-145. [26] O'Keefe A. & Deacon D. A. G. (1988) Cavity ring-down optical spectrometer for absorption measurements using pulsed laser sources. Rev. Sci. Instrum., 59: 2544-2551. [34] O'Leary M. H. (1984) Measurement of the isotope fractionation associated with diffusion of carbon dioxide in aqueous solution. J. Phys. Chem., 88: 823-825. [46] O'Malley V. P., Abrajano T. A. Jr. & I-Iellou J. (1994) Determination of the 13C/12C ratios of individual PAH from environmental samples: can PAH sources be apportioned? Org. Geochem., 21: 809-822. [9] O'Neil J. R. & Clayton R. N. (1964) Oxygen isotope thermometry. In: Isotopic and cosmic chemistry. H. Craig, G. J. Miller, and G. J. Wasserburg (eds.). North Holland: 157-168. [29, 46] O'Neil J. R. & Epstein S. (1966a) Oxygen isotope fractionation in the system dolomite-calcite-carbon dioxide. Science, 152: 198-201. [1, 46] O'Neil J. R. & Epstein S. (1966b) A method for oxygen isotope analysis of milligram quantities of water and some of its applications. J. Geophys. Res., 71: 4956-4961. [3, 20] O'Neil J. R. & Taylor H. P. Jr. (1967) The oxygen isotope and cation exchange chemistry of feldspars. Am. Mineralogist, 52: 1414-1437. [46] O'Neil J. R., Clayton R. N & Mayeda T. K. (1969) Oxygen isotope fractionation in divalent metal carbonates. J. Chem. Phys., 51: 5547-5558. [46] O'Neil J. R. & Taylor H. P. Jr. (1969) Oxygen isotope equilibrium between muscovite and water. J. Geophys. Res., 74: 6012-6022. [46] O'Neil J. R. (1986) Theoretical and experimental aspects of isotopic fractionation. In: Stable isotopes in high temperature geological processes. J. W. Valley, H. P. Jr. Taylor & J. R. O'Neil (eds.). Mineral. Soc. America, Reviews in Mineralogy, Vol. 16: 1-40. [46, App B] O'Neil J. R. & Truesdell A. H. (1991) Oxygen isotope fractionation studies of solute-water interactions. In: Stable isotope geochemistry: A tribute to Samuel Epstein. H. P. Jr. Taylor, J. R. O'Neil & I. R. Kaplan (eds.). The Geochemical Society, Spec. Publ. 3: 17-25. [46] O'Neil J. R., Roe L. J., Reinhard E. & Blake R. E. (1994) A rapid and precise method of oxygen isotope analysis of biogenic phosphate. Isr. J. Earth Sci., 43: 203-212. [20, 22] O'Neil J. R., Vennemann T. W. & McKenzie W. F. (2003) Effects of speciation on equilibrium fractionations and rates of oxygen isotope exchange between (PO4)aq and H20. Geochim. Cosmochim. Acta, 67 (Clayton issue): 3135-3144. [46] O'Reilly C., Jenkin G. R. T., Feely M., Alderton D. H. M. & Fallick A. E. (1997) A fluid inclusion and stable isotope study of 200 Ma of fluid evolution in the Galway Granite, Connemara, Ireland. Contrib. Mineral. Petrol., 129: 120-142. [3] Oana S. & Ishikawa H. (1966) Sulfur isotopic fractionation between sulfur and sulfuric acid in the hydrothermal solution of sulfur dioxide. Geochem. J., 1: 45-50. [26] Ohashi S. (1955) Rapid determination of sulfur in organic compounds by the tin (II)-strong phosphoric acid reduction method. Bull. Chem. Soc. Japan, 28(9): 645-649. [26] Ohba T. (1987) 180/160 and D / H determinations for small amounts of water. Geochem. J., 21: 183-186. [1,20] Ohba T., Kazahaya K. & Matsuo S. (1995) Diffusiona1180 loss from inclusion water in a natural hydrothermal quartz from the Kaneuchi tungsten deposit, Japan. Geochim. Cosmochim. Acta, 59: 3039-3047.
[3]
Ohba T. & Hirabayashi J. (1996) Handling of Pt catalyst in H2-H20 equilibration method for D / H measurement of water. Geochem. J., 30: 373-377. [1] Ohba T. & Matsuo S. (1998) Precise determination of hydrogen and oxygen isotope ratios of water in fluid inclusions of quartz and halite. Geochem. J., 22(2): 55-68. [3]
References
1139
Ohkouchi N., Kawamura K., Kajiwara Y., Wada E., Okada M., Kanamatsu T. & Taira A. (1999) Sulfur isotope records around Livello Bonarelli (northern Apennines Italy) black shale at the CenomanianTuronian boundary. Geology, 27: 535-538. [26] Ohmoto H. & Lasaga A. C. (1982) Kinetics of reactions between aqueous sulfates and sulfides in hydrothermal systems. Geochim. Cosmochim. Acta, 46: 1725-1745. [46] Ohmoto H. (1992) Biogeochemistry of sulphur and the mechanisms of sulfide-sulfate minralization in Archean oceans. In: Early organic evolution. M. Schidlowski, S. Golubic, M. M. Kimberley, D. M. McKirdy & P. A. Trudinger (eds.). Springer Verlag, Berlin-Heidelberg: 378-397. [27] Ohmoto H. & Goldhaber M. B. (1997) Sulfur and carbon isotopes. In: Geochemistry of hydrothermal ore deposits. H. L. Barnes (ed.). Wiley & Sons: 570-611. [27] Ohorodnik S. K. & Harrison W. W. (1993) Cryogenic coil for glow discharge sources. Anal. Chem., 65: 2542-2544. [35] Ohsumi T. & Fujino H. (1986) Isotopic exchange technique for preparation of hydrogen gas in mass spectrometric D /H analysis of natural waters. Anal. Sci., 2: 489-490. [1] Oi T., Kawada K., Hosoe M. & Kakihana H. (1991) Fractionation of lithium isotopes in cation-exchange chromatography. Sep. Sci. Tech., 26: 1353-1375. [6] Oi T., Odagiri T. & Nomura M. (1997) Extraction of lithium from GSJ rock reference samples and determination of their lithium isotopic compositions. Anal. Chim. Acta, 340: 221-225. [6] Olive K. A. & Schramm D. N. (1992) Astrophysical 7Li as a product of Big Bang nucleosynthesis and galactic cosmic-ray spallation. Nature, 360: 439-442. [6] Oremland R. S., HoUibaugh J. T., Maest A. S., Presser T. S., Miller L. G. & Culbertson C. W. (1989) Selenate reduction to elemental selenium by anaerobic bacteria in sediments and culture: Biogeochemical significance of a novel, sulfate-independent respiration. Appt. Env. Microbiol., 55(9): 2333-2343. [29] Orlando T. M., Burns A. R., Stechel E. B. & Jennison D. R. (1991) Resonance ionization detection of neutral O(3PJ) produced from stimulated surface processes. J. Vac. Sci. Technol. A - Vac Surf. Films, 9: 1769-1773. [33] Orr W. L. (1974) Changes in sulfur content and isotopic ratios of sulfur during petroleum maturation study of Big Horn Basin Paleozoic oils. Am. Assoc. Pet. Geol. Bull., 58: 2295-2318. [26] Osberghaus O. (1950) Die Isotopenh/iufigkeit des Bors. Massenspektrometrische untersuchung der Elektronenstossprodukte von BF3 und BC13. Z. Physik, 128: 366-377. [28] Ostrom P. H., Colunga-Garcia M. & Gage S. H. (1997) Establishing pathways of energyfow for insect predators using stable isotope ratios: field and laboratory evidence. Oecotogia, 109: 108-113. [9] Ostwald W. (1900) Ober die vermeintliche Isomerie des roten und gelben Quecksilberoxyds und die Oberfl/ichenspannung fester K6rper. Z. Phys. Chem. Stoechiom. Verwandschaftslehre, 34: 495-503. [46] Otsuki A., Ino Y. & Fujii T. (1983) Simultaneous measurements and determinations of stable carbon and nitrogen isotope ratios, and organic carbon and nitrogen contents in biological samples by coupling of a small quadrupole mass spectrometer and modified carbon-nitrogen elemental analyzer. Int. J. Mass Spectrom. Ion Phys., 48: 343-346. [15] Oversby V. M. (1969) Determination of lead isotope compositions in iron meteorites by the double spike technique. Meteoritics, 4(3): 201. [29] Owen H. R. & Schaeffer O. A. (1954) The isotope abundance of chlorine from various sources. J. Am. Chem. Soc., 77: 898-899. [28] Owens N. J. P. & Rees A. P. (1989) Determination of nitrogen-15 at sub-microgram levels of nitrogen using automated continuous-flow isotope ratio mass spectrometry. Analyst, 114: 1655-1657. [45] Pajo L., Tamborini G., Rasmussen G., Mayer K. & Koch L. (2001) A novel isotope analysis of oxygen in uranium oxides: comparison of secondary ion mass spectrometry, glow discharge mass spectrometry, and thermal ionization mass spectrometry. Spectrochim. Acta, B 56: 541-549. [35] Paldus B. A., Harb C. C., Spence T. G., Wilke B., Xie J., Harris J. S. & Zare R. N. (1998) Cavity-locked ring-down spectroscopy. J. Appl. Phys., 83: 3991-3997. [34] Paldus B. A., Spence T. G., Zare R. N., Oomens J., Harren F. J. M., Parker D. H., Gmachl C., Cappasso F., Sivco D. L., Baillargeon J. N., Hutchinson A. L. & Cho A. Y. (1999) Photoacoustic spectroscopy using quantum-cascade lasers. Opt. Lett., 24: 178-180. [34]
1140
References
Palin J. M., Epstein S. & Stolper E. M. (1996) Oxygen isotope partitioning between rhyolitic glass/melt
and CO2: an experimental study at 550-950~ and I bar. Geochim. Cosmochim. Acta, 60: 1963-1973. [46] Palmer D. A. S. & Williams-Jones A. E. (1996) Genesis of the carbonatite-hosted fluorite deposit at Amba Dongar, India; evidence from fluid inclusions, stability isotopes, and whole rock-mineral geochemistry. Econ. Geol., 91: 934-950. [3] Palmer G. H. (1958) The thermal-emission ion source in solid source mass spectrometry. J. Nuclear Energy, 7: 1-12. [7] Palmer M. R., Spivack A. J. & Edmond J. M. (1987) Temperature and pH controls over isotopic fractionation during adsorption of boron on marine clay. Geochim. Cosmochim. Acta, 51: 2319-2323. [46] Palmer M. R. (1991) Boron isotope systematics of Halmahera arc (Indonesia) lavas: Evidence for involvement of the subducted slab. Geology, 19: 215-217. [7] Palmer M. R. & Swihart G. H. (1996) Boron isotope geochemistry: An overview. In: Boron. Mineralogy, petrology and geochemistry. E. S. Grew & L. M. Anovitz (eds.). Mineral. Soc. of America, Reviews in Mineralogy, Vol. 33: 709-744. [7] Palmer M. R., Pearson P. N. & Cobb S. J. (1998) Reconstructing past ocean pH-depth profiles. Science, 282: 1468-1471. [7] Papadakis I., Poulsen E., Taylor P. D. P., Van Nevel L. & De Bi6vre P. (1999) IMEP-9 trace elements in water. IRRM, Report to participants, EUR 18724 or Internat. Rept. (IRRM)GE / R / IM / 15 / 1999. [37] Papadakis I., Taylor P. D. P. & De Bi6vre P. (2001) CCQM-K2 key comparison: cadmium and lead content in natural water. Metrologia, 38: 543-547. [37] Papic P. B., Ciric M. M. & Zmbov K. F. (1979) Isotopic analysis of boron by negative thermal ionization. Glas. Hem. Drus. Beograd, 44: 195-201. [7] Pardue J. W., Scalan R. S., Van Baalen C. & Parker P. L. (1976) Maximum carbon isotope fractionation in photosynthesis by blue-green algae and a green algae. Geochim. Cosmochim. Acta, 40: 309-312. [46] Park H., Yi J.-H., Han J.-M., Rhee Y.-J. & Lee J. (1998) Studies on autoionization states of Sm by 3-step resonance photoionization. Resonance Ionization Spectroscopy, Ninth International Symposium. J.C. Vickerman, I. Lyon, N.P. Lockyer & J.E. Parks (eds.). AIP Press Woodbury, NY. AIP Conference Proceedings 454: 257-260. [33] Park R. & Epstein S. (1960) Carbon isotope fractionation during photosynthesis. Geochim. Cosmochim. Acta, 21: 110-126. [46] Parr R. M. & Clements S. A. (1991) Intercomparison of enriched stable isotope reference materials for medical and biological studies. IAEA, Vienna. IAEA Report NAHRES-5:31 p. [40] Passier H. F., Biittcher M. E. & De Lange G. J. (1999) Sulphur enrichment in organic matter of eastern Mediterranean sapropels: A study of sulfur isotope partitioning. Aq. Geochem., 5(1): 99-119. [45] Patris N., Delmas R. J. & Jouzel J. (2000a) Isotopic signatures of sulfur in shallow Antarctic ice cores. J. Geophys. Res., 105: 7071-7078. [26] Patris N., Mihalopoulos N., Baboukas E. D. & Jouzel J. (2000b) Isotopic composition of sulfur in sizeresolved marine aerosols above the Atlantic Ocean. J. Geophys. Res., 105: 14449-14457. [26] Pattberg S. (1999) Multielement-Spurenbestimmung in Reinstkupfer mit HR-ICP-MS. PhD Thesis, Humboldt-University, Berlin, Germany. [42] Pattberg S. & Matschat R. (1999) Determination of trace impurities in high purity copper using sectorfield ICP-MS: continuous nebulization, flow injection analysis and laser ablation. Fresenius J. Anal. Chem., 364: 410-416. [42] Patterson K. Y., Veillon C., Hill A. D., Moser-Veillon P. B. & O'Haver T. C. (1999) Measurement of calcium stable isotope tracers using cool plasma ICP-MS. J. Anal. Atmos. Spectrom., 14: 1673-1677. [31] Patterson W. P., Smith G. R. & Lohmann K. C. (1993) Continental paleothermometry and seasonality using the isotopic composition of aragonitic otoliths of freshwater fishes. In: Climate Change in Continental Isotopic Records. P. K. Swart, K. C. Lohmann, J. McKenzie & S. Savin (eds.). Am. Geophys. Union. Monograph 78: 191-202. [46] Pattison D. R. M. (1994) Are reversed Fe-Mg exchange and solid solution experiments really reversed? Am. Mineralogist, 79: 938-950. [46] Patton C. J. (1995) U.S. Geological Survey nutrient preservation experiment - Nutrient concentration data for surface-, ground-, and municipal-supply water samples and quality-assurance samples. U.S. Geological Survey. Open-File Report 95-141:140 p. [15]
References
1141
Patton C. J., Fischer A. E., Campbell W. H. & Campbell E. R. (2002) Corn leaf nitrate reductace - a nontoxic alternative to cadmium for photometric nitrate determinations in water samples by air-segmented continous-flow analysis. Environ. Sci. Technol., 36(4): 729-735. [15] Pauling L. (1964) College chemistry. W.H. Freeman and Co., San Francisco, 3rd edition: 832 p. [20] Paulsen P. J. & Kelly W. (1984) Determination of sulfur as arsenic monosulfide by isotope dilution thermal mass spectrometry. Anal. Chem., 56: 708-713. [26, 29] Pauwels J., Lamberty A. & Schimmel H. (1998a) Homogeneity testing of reference materials. Accred. Qual. Assur., 3: 51-55. [43] Pauwels J., Lamberty A. & Schimmel H. (1998b) Quantification of the expected shelf-life of certified reference materials. Fresenius J. Anal. Chem., 361: 395-399. [43] Pawellek F. & Veizer J. (1994) Carbon cycle in the upper Danube and its tributaries: 13CDICconstraints. Isr. J. Earth Sci., 43: 187-194. [10] Paytan A., Luz B. & Kolodny Y. (1990) 6180 in dissolved phosphate. Geol. Soc. Am., ICOG 7, Vol. 27:77 (Abstract.) [22] Pearson F. J. Jr., Fisher D. W. & Plummer L. N. (1978) Correction of ground-water chemistry and carbon isotopic composition for effects of CO2 outgassing. Geochim. Cosmochim. Acta, 42: 1799-1807. [10] Peele G. L. & Brent D. (1977) Cleaning of mass spectrometer ion sources by electropolishing. Anal. Chem., 49: 674. [39] Peiser H. S. & De Bi6vre P. (1999) Traceability in measurement - time for an update? Accred. Qual. Assur., 4: 154-155. [43] Penman A. D. & Wright I. A. (1987) Determination of deuterium level in biological fluids by isotope ratio mass spectrometry. Biomed. Environ. Mass Spectrom., 14: 339-342. [1] Penzias A. A. (1980) Nuclear processing and isotopes in the galaxy. Science, 208: 663-669. [11] Perera I. K., Lyon I. C., Turner G. & Van Lierde P. (1994) Continuous wave laser resonance ionization mass spectrometric analysis of strontium. Int. J. Mass Spetrom. Ion Process., 132: 225-238. [33] Perera I. K., Lyon I. C & Turner G. (1995) Isotope ratio measurement in strontium using two-photon two-colour resonance ionization mass spectrometry. J. Anal. Atom. Spectrosc., 10: 273-280. [33] Perlmutter N. M., Lieber M. & Frauenthal H. L. (1963) Movement of waterborne cadmium and hexavalent chromium wastes in South Farmingdale, Nassau County, Long Island. U.S. Geological Survey Professional Paper, Vol. 475C: C170-C184. [29] Petersen B. J., Howarth R. W. & Garritt R. H. (1986) Sulfur and carbon isotopes as tracers of salt-marsh organic matter flow. Ecology, 67: 865-874. [26] Peterson B., Wollheim W., Mulholland P., Webster J., Meyer J., Tank J., Marti E., Bowden W., Valett M., Hershey A., McDowell W., Dodds W., Hamilton S., Gregory S. & Morral D. (2001) Control of nitrogen export from watersheds by headwater streams. Science, 292: 86-90. [15] Petterson D. W. & Hayes J. M. (1978) Contemporary topics in analytical and clinical chemistry. Plenum Press. New York: 217-252. [14] Phillips B. L., Lee A. & Casey W. H. (2003) Rates of oxygen exchange between the A1208A128(OH)56 (H20)2618+(aq)(A130) molecule and aqueous solution. Geochim. Cosmochim. Acta, 67: 2725-2733. [46] Picard S., Garcia J.-P., L6cuyer C., Sheppard S. M. F., Cappetta H. & Emig C. C. (1998) 5180 values of coexisting brachiopods and fish: Temperature differences and estimates of paleo-water depths. Geology, 26(11): 975-978. [22] Pichlmayer F. & Blochberger K. (1988) Isotopenh~iufigkeitsanalyse von Kohlenstoff, Stickstoff and Schwefel mittels Ger~itekopplung Elementaranalysator-Massenspectrometer. Fresenius Z. Anal. Chem., 331: 196-201. [26, 27] Pichlmayer F., Lamprecht G. & Rossmann A. (1998) Sulfur isotope measurements in organic substances. Isotopes Environ. Health Stud., 34: 227-335. [26] Pickthorn W. J. & O'Neil J. R. (1985) 180 relations in alunite minerals: a potential single-mineral thermometer. Geol. Soc. Am. Abstr. w. Progr., 17:689 (Abstract). [20] Pillinger C. T. (1992) New technologies for small sample stable isotope measurement: static vacuum gas source mass spectrometry, laser probes, ion probes and gas chromatography-isotope ratio mass spectrometry. Int. J. Mass Spectrom. Ion Proc., 118/119: 477-501. [20] Pimminger M., Grasserbauer M., Schroll E. & Cerny I. (1984) Microanalysis in galena by secondary ion mass spectrometry for determination of sulfur isotopes. Anal. Chem., 56: 407-411. [30]
1142
References
Pineau F. & Javoy M. (1983) Carbon isotopes and concentrations in mid-oceanic ridge basalts. Earth Planet. Sci. Lett., 62: 239-257. [3] Pineau F., Shilobreeva S., Kadik A. & Javoy M. (1998) Water solubility and D / H fractionation in the system basaltic andesite-H20 at 1250~ and between 0.5 and 3 kbars. Chem. Geol., 147: 173-184. [46] Piperov N. B. & Penchev N. P. (1973) A study on gas inclusions in minerals. Analysis of the gases from micro-inclusions in allanite. Geochim. Cosmochim. Acta, 37: 2075-2097. [3] Piperov N. B. & Penchev N. P. (1982) Deuterium content of the inclusion water from hydrothermal galenas, Madan, Bulgaria; preliminary investigation. Econ. Geol., 77(1): 195-197. [3] Pisonero J., Costa J. M., Pereiro R., Bordela N. & Sanz-Medel A. (2001) A simple glow discharge ion source for direct solid analysis by on-axis time-of-flight mass spectrometry. J. Anal. At. Spectrom., 16: 1253-1258. [35] Pizzarello S., Krishnamurthy R. V., Epstein S. & Cronin J. R. (1991) Isotopic analyses of amino acids from the Murchison meteorite. Geochim. Cosmochim. Acta, 55: 905-910. [11] Pizzarello S., Feng X., Epstein S. & Cronin J. (1994) Isotopic analyses of nitrogenous compounds from the Murchison meteorite: Ammonia, amines, amino acids and polar hydrocarbons. Geochim. Cosmochim. Acta, 58: 5579-5587. [11] Pizzarello S. & Cronin J. (1998) Alanine enantiomers in the Murchison meteorite. Nature, 394: 236. [11] Plas A. & Friih-Green G. L. (1997) Intracrystalline fractionation of oxygen isotopes in serpentine minerals determined by thermal dehydroxylation. Terra Nova, 9(Abs. Suppl. No. 1): 445 (Abstract). [2] Platzner I. T., Habfast K., Walder A. J. & Goetz A. (1997) Modern isotope ratio mass spectrometry. John Wiley & Sons, Chichester, West Sussex, England: 514 p. [1, 32, 38] Podosek F. A., Prombo C. A., Grossman L. & Zinner E. K. (1991) Chromium isotopic compositions of individual spinel crystals from the Murchison meteorite. 54th meeting of the Meteoritical Society. Arizona State University, Center for Meteorite Studies, Vol. 26:385-386 (Abstract). [29] Polyakov V. B. & Kharlashina N. N. (1994) Effect of pressure on equilibrium isotopic fractionation. Geochim. Cosmochim. Acta, 58: 4739-4750. [46] Polyakov V. B. (1998) On anharmonic and pressure corrections to the equilibrium isotopic constants for minerals. Geochim. Cosmochim. Acta, 62: 3077-3085. [46] Pombo S. A., Pelz O., Schroth M. H. & Zeyer J. (2002) Microbial influence in natural and polluted groundwater: Field-scale C-13-1abeling of phospholipid fatty acids (PLFA) and dissolved inorganic carbon: tracing acetate assimilation and mineralization in a petroleum hydrocarbon-contaminated aquifer. Ferns Microbiot. Ecol., 41: 259-267. [10] Pond D. W., Bell M. V., Dixon D. R., Fallick A. E., Segonzac M. & Sargent J. R. (1998) Stable-carbon isotope composition of fatty acids in hydrothermal vent mussels containing methanotrophic thiotrophic bacterial endosymbionts. Appt. Environm. Microbiol., 64(1): 370-375. [9] Porter L. K. & O'Deen W. A. (1977) Apparatus for preparing nitrogen ammonium chloride for 15Ndeterminations. Anal. Chem., 49: 514-516. [15] Poschenrieder W. P., Herzog R. F. & Barrington A. E. (1965) The relative abundance of the lithium isotopes in the Holbrook meteorite. Geochim. Cosmochim. Acta, 29: 1193-1195. [6] Poulson S. R. & Schoonen M. A. A. (1994) Variations of the oxygen isotope fractionation between NaCO3- and water due to the presence of NaC1 at 100-300~ Chem. Geol. (Isot. Geosci. Sect.), 116: 305315. [46] Powell M. D. & Kyser T. K. (1991) Analysis of 613C and 6180 in calcite, dolomite, rhodochrosite and siderite using a laser extraction system. Chem. Geol. (Isot. Geosci. Sect.), 94: 55-66. [22] Presser T. S. (1994a) Geological origin and pathways of selenium from the California Coastal Ranges to the west-central San Joaquin Valley. In: Selenium in the environment.W.T. Jr. Frankenberger & S. Benson (eds.). Marcel Dekker: 139-155. [29] Presser T. S. (1994b) The Kesteron effect. Environ. Manage., 18: 437-454. [29] Preston T. & Owens N. J. P. (1983) Interfacing an automatic elemental analyser with an isotope ratio mass spectrometer: the potential for fully automated total nitrogen and nitrogen-15 analysis. Analyst, 108: 971-977. [15] Preston T. (1992) The measurement of stable isotope natural abundance variations. Plant Cell Environ., 15: 1091-1097. [8]
References
1143
Preston T. & Slater C. (1994) Mass spectrometric analysis of stable-isotope-labelled acid tracers. Proc. Nutr. Soc., 53: 363-372. [8] Previs S. F., Hazey J. W., Diraison F., Beylot M., David F. & Brunengraber H. (1996) Assay of the deuterium enrichment of water via acetylene. J. Mass Spectrom., 31(6): 639-642. [1] Price G. D., McKenzie J. E., Pilcher J. R. & Hoper W. T. (1997) Carbon-isotope variation in Sphagnum from hummock-hollow complexes: implications for Holocene climate reconstruction. The Holocene, 7: 229-233. [24] Prombo C. A. & Clayton R. N. (1993) Nitrogen isotopic compositions of iron meteorites. Geochim. Cosmochim. Acta, 57: 3749-3761. [16] Prosser S. J., Wright I. P. & Pillinger C. T. (1990) A preliminary investigation into the isotopic measurement of carbon at the picomole level using static vacuum mass spectrometry. Chem. Geol., 83: 71-88. [13] Prosser S. J. & Scrimgeour C. M. (1995) High-precision determination of 2H / 1H in H2 and H20 by continuous-flow isotope ratio mass spectrometry. Anal. Chem., 67: 1992-1997. [1, 3, 8, 38] Pruden G. & Bloomfield C. (1968) The determination of iron(II) sulphide in soil in the presence of iron(III) oxide. Analyst, 93: 532-534. [26] Puchelt H., Sabels B. R. & Hoering T. C. (1971) Preparation of sulfur hexafluoride for isotope geochemical analysis. Geochim. Cosmochim. Acta, 35: 625-628. [20, 26, 36] Pugh L. A. & Rao K. N. (1976) Intensities from infrared spectra. In: Molecular spectroscopy: Modern research. K. N. Rao (ed.). Academic Press, New York Vol. II. [34] Pupin A. M., Dennis M. J., Parker I., Kelly S., Bigwood T. & Toledo M. C. F. (1998) Use of isotopic analyses to determine the authenticity of brazilian orange juice (Citrus sinensis). J. Agric. Food Chem., 46: 1369-1373. [5] Purnell A. L. & Doolan K. J. (1983) Rapid determination of forms of sulfur in coal. Fuel, 62: 1107-1110. [26] Qi H. P., Taylor P. D. P., Berglund M. & De Bi6vre P. (1997) Calibrated measurement of the isotopic composition and atomic weight of the natural Li isotopic reference material IRMM-016. Int. J. Mass Spectrom. Ion Proc., 171: 263-268. [6] Qi H. P. & Coplen T. B. (2003) Evaluation of the 34S / 32S ratio of Soufre de Lacq elemental sulfur isotopic reference material by continuous flow isotope-ratio mass spectrometry. Chem. Geol., 199: 183-187. [40] Qi-Lu & Masuda A. (1994) The isotopic composition and atomic weight of molybdenum. Int. J. Mass Spectrom. Ion Proc., 130: 65-72. [31] Quay P. D., Emerson S. R., Quay B. M. & Devol A. H. (1986) The carbon cycle for Lake WashingtonA stable isotope study. Limnol. Oceanogr., 31: 596-611. [10] Quay P. D., King S., Wilbur D. O. & Wofsy S. (1989) 13C/12C of atmospheric CO2 in the Amazon basin: forest and river sources. J. Geophys. Res., 94: 18327-18336. [9] Quay P. D., King S. L., Stutsman J., Wilbur D. O., Steele L. P., Fung I., Gammon R. H., Brown T. A., Farwell G. W., Grootes P. M. & Schmidt F. H. (1991) Carbon isotopic composition of atmospheric CH4: Fossil and biomass burning source strengths. Global Biogeochem. Cycles, 5: 25-47. [14] Quay P. D., Wilbur D. O., Stuiver M., Hedges J. I., Devol A. H. & Richey J. E. (1992) Carbon cycling in the Amazon River: Implications from the 13C composition of particulate and dissolved carbon. Limnol. Oceanogr., 37: 857-870. [9] Quay P., Stutsman J., Wilbur D., Snorer A., Dlugokencky E. & Brown T. (1999) The isotopic composition of atmospheric methane. Global Biogeochem. Cycles, 13(2): 445-461. [14] Quemerais B., Mabon F., Naulet N. & Martin G. J. (1995) Site-specific isotope fractionation of hydrogen in the biosynthesis of plant fatty acids. Plant Cell Environ., 18: 989-998. [5] Qu6tel C. R., Thomas B., Donard O. F. X. & Grousset F. E. (1997) Factorial optimization of data acquisition factors for lead isotope ratio determination by inductively coupled plasma mass spectrometry. Spectrochim. Acta, B 52: 177-187. [32] Qu6tel C. R., Prohaska T., Hamester M., Kerl W. & Taylor P. D. P. (2000a) Examination of the performance exhibited by a single detector double focusing magnetic sector ICP-MS instrument for uranium isotope abundance ratio measurements over almost three orders of magnitude and down to pg g-1 concentration levels. J. Anal. At. Spectrom., 15: 353-358. [32]
1144
References
Qu6tel C. R., Vogl J., Prohaska T., Nelms S., Taylor P. D. P. & De Bi6vre P. (2000b) Comparative performance study of ICP mass spectrometers by means of U "isotopic measurements". Fresenius J. Anal. Chem., 368: 148-155. [32] Qu6tel C., Prohaska T., Nelms S., Diemer J & Taylor P. D. P. (2001) ICP-MS applied to isotope abundance ratio measurements: performance study and development of a method for combining uncertainty contributions from measurement correction factors. In: Plasma source mass spectrometry: the new millennium. G. Holland & S. D. Tanner (eds.). The RSC Publishers, UK.Proc of the 7th Durham Conf.: 257-269. [32] Quinn P. K. & Bates T. S. (1989) Collection efficiencies of a tandem sampling system for atmospheric aerosol particles and gaseous ammonia and sulfur dioxide. Environ. Sci. Technol., 23: 736-739. [26] Rabalais N. N., Turner R. E., Justic D., Dortch Q. & Wiseman W. (1999) Characterization of hypoxia: topic I report for the integrated assessment on hypoxia in the Gulf of Mexico. NOAA Coastal Ocean Program, NOAA Coastal Ocean Program Decision Analysis, Silver Spring, Md., Series No. 15:167 p. [15] Rabo J. A. (1976) Zeolite chemistry and catalysis. American Chemical Society, Washington DC, ACS Monograph 171:796 p. [36] Rahn T. A. & Wahlen M. (1997) Stable isotope enrichment in stratospheric nitrous oxide. Science, 278: 1776-1778. [19] Rahn T. A. & Wahlen M. (2000) A reassesment of the global isotopic budget of atmospheric nitrous oxide. Glob. Biogeochem. Cycles, 14: 537-543. [34] Rajan R. S., Brown L., Tera F. & Whitford D. J. (1980) Lithium isotopic composition in some stone meteorites. Earth Planet. Sci. Lett., 51: 41-44. [6] Ramakumar K. L., Parab A. R., Khodade P. S., Almaula A. I., Chitambar S. A. & Jain H. C. (1985) Determination of isotopic composition of boron. J. Radioanal. Nucl. Chem., 94: 53-62. [7] Rameb~ick H., Berglund M., Vendelbo D., Wellum R. & Taylor P. D. P. (2001) On the determination of the true dead-time of a pulse-counting system in isotope ratio mass spectrometry. J. Anal. At. Spectrom., 16: 1271-1274. [37] Ramesh R., Bhattacharya S. K. & Pant G. B. (1989) Climatic significance of 5D variations in a tropical tree species from India. Nature, 337: 149-150. [46] Rashid K., Krouse H. R. & McCready R. G. L. (1978) Selenium isotope fractionation during bacterial selenite reduction. In: Short papers of the fourth international conference, geochronology, cosmochronology, isotope geology, 1978. R. E. Zartman (ed.). U. S. Geological Survey: 347-348. [29] Rashid K. & Krouse H. R. (1985) Selenium isotopic fractionation during SeO3 reduction to Se0 and H2Se. Can. J. Chem. 63: 3195-3199. [29] Rau G. (1978) Carbon-13 depletion in a subalpine lake: carbon flow implications. Science, 210: 901-902. [10] Rautenschlein M., Habfast K. & Brand W. (1990). In: Stable isotopes in pediatric nutritional and metabolic research. T. E. Chapman, R. Berger, D. J. Reijngoud & A. Okken (eds.). Intercept Ltd, Andover, MA: 133-148. [8] Raven A. M., van Bergen P. F., Stott A. W., Dudd S. N. & Evershed R. P. (1997) Formation of long-chain ketones in archaeological pottery vessels by pyrolysis of acyl lipids. J. Anal. Appl. Pyrol., 40-1: 267-
285. [8]
Rayleigh Lord R.J. (1939) Nitrogen, argon and neon in the earth's crust with applications to cosmology. Proc. Roy. Soc. Lond., Ser. A 170: 451. [16] Read J. J., Johnson D. A., Asay K. H. & Tieszen L. L. (1992) Carbon isotope discrimination: Relationship to yield, gas exchange, and water-use efficiency in field-grown crested wheatgrass. Crop Science, 32: 168-175. [9] Reardon E. J. & Fritz P. (1978) Computer modeling of groundwater 13C and 14C isotope composition. J. Hydrol., 36: 201-224. [10] Reardon E. J., Allison G. B. & Fritz P. (1979) Seasonal chemical and isotopic variations of soil CO2 at Trout Creek, Ontario. J. Hydrol., 43: 355-371. [10] Rees C. E. & Thode H. G. (1966) Selenium isotope effects in the reduction of sodium selenite and of sodium selenate. Can. J. Chem., 44: 419-427. [29, 46]
References
1145
Rees C. E. & Thode H. G. (1977) A 33S anomaly in the Allende meteorite. Geochim. Cosmochim. Acta, 41: 1679-1682. [20, 26, 45] Rees C. E. (1978) Sulphur isotope measurements using SO2 and SF6. Geochim. Cosmochim. Acta, 42: 383389. [20, 26, 44, 45] Rees C. E.& Holt H. D. (1991) The isotopic analysis of sulphur and oxygen. In: Stable isotopes: Natural and anthropogenic sulphur in the environment; Scope 43. H. R. Krouse & V. A. Grinenko (eds.). John Wiley & Sons, Chichester: 43-64. [26] Rehk~imper M. & Halliday A. N. (1998) Accuracy and long-term reproducibility of lead isotopic measurements by multiple-collector inductively coupled plasma mass spectrometry using an external method for the correction of mass discrimination. Int. J. Mass Spectrom., 181: 123-133. [31] Rehktimper M. & Halliday A. N. (1999) The precise measurement of T1 isotopic compositions by MCICPMS: Application to the analysis of geological materials and meteorites. Geochim. Cosmochim. Acta, 63: 935-944. [31] Rehk~imper M. & Mezger K. (2000) Investigation of matrix effects for Pb isotope ratio measurements by multiple collector ICP-MS: verification and application of optimized analytical protocols. J. Anal. At. Spectrom., 15: 1451-1460. [31] Rehk~imper M., Frank M., Halliday A. N. & Hein J. R. (2001a) Thallium isotope variations in ferromanganese crusts. J. Conf. Abstr., 6:689 (Abstract). [31] Rehk~imper M., SchSnbachler M. & Stirling C. H. (2001b) Multiple collector ICP-MS: Introduction to instrumentation, measurement techniques and analytical capabilities. Geostand. Newslett., 25: 23-40. [31] Rein J. E. & Abernathey R. M. (1972) Rapid mass-spectrometric determination of boron isotopic distribution in boron carbide. Talanta, 19: 857-862. [7] Remaud G., Guillou C., Vallet C. & Martin G. J. (1992) A coupled NMR and MS isotopic method for the authentification of natural vinegars. Fresenius J. Anal. Chem., 342: 457-461. [5] Remaud G., Debon A. A., Martin Y. L., Martin G. G. & Martin G. J. (1997a) Authentication of bitter almond oil and cinnamon oil: application of the SNIF-NMR method to benzaldehyde. J. Agric. Food Chem., 45: 4042-4048. [5] Remaud G. S., Martin Y. L., Martin G. G. & Martin G. J. (1997b) Detection of sophisticated adulterations of natural vanilla flavours and extracts: application of the SNIF-NMR method to vanillin and p-hydroxybenzaldehyde. J. Agric. Food Chem., 45: 859-866. [5] Remaud G. S., Martin Y. L., Martin G. G., Naulet N. & Martin G. J. (1997c) Authentication of mustard oils by combined stable isotope analysis (SNIF-NMR and IRMS). J. Agric. Food Chem., 45: 1844-1848. [26, 5] Rendigs R. R. & Commeau J. A. (1987) Effects of disaggregation on a fine-grained marine mud by two ultrasonic devices. J. Sediment. Petrol., 57: 786-787. [2] Rennie M. J., Meier-Augenstein W., Watt P. W., Patel A., Begley I. S. & Scrimgeour C. M. (1996) Use of continuous-flow combustion MS in studies of human metabolism. Biochem. Soc. Trans., 28: 927932. [8] Reuter J. H., Epstein S. & Taylor H. P. Jr. (1965) O18 / O16 ratios of some chondritic meteorites and terrestrial ultramafic rocks. Geochim. Cosmochim. Acta, 29: 481-488. [20] Reynolds J. H. & Verhoogen J. (1953) Natural variations in the isotopic constitution of silicon. Geochim. Cosmochim. Acta, 3: 224-234. [20, 25] R6v6sz K. & Woods P. H. (1990) A method to extract soil water for isotopic analysis. J. Hydrol., 115: 397406. [1, 2, 21] R6v6sz K., BShlke J. K. & Yoshinari T. (1997) Determination of 5180 and 615N in nitrate. Anal. Chem., 69: 4375-4380. [15] R6v6sz K. M., BShlke J. K., Smith R. L. & Yoshinari T. (1999a) Stable isotope composition of dissolved 02 undergoing respiration in a ground-water contamination gradient. In: U. S. Geological Survey Toxic Substances Hydrology Program; proceedings of the technical meeting. D. W. Morganwalp & H. T. Buxton (eds.). U. S. Geological Survey. Water-Resources Investigations, Report WRI99-4018-C: 323-328. [15]
1146
References
R6v6sz K. M., B/Shlke J. K., Smith R. L. & Yoshinari T. (1999b) Delta 18-O composition of dissolved 02 undergoing respiration in a contaminated ground water. IAEA, Vienna. IAEA International Symposium on isotope techniques in water resources development and management, Vienna, Austria. 10-14 May, 1999. (on CD-ROM). [15] R4v4sz K. & BiShlke J. K. (2002) Comparison of 6180 measurements in nitrate by different combustion techniques. Anal. Chem., 74: 5410-5413. [15] Rhee S. K., Reed R. G. & Brenna J. T. (1997) Fatty acid carbon isotope ratios in humans on controlled diets. Lipids, 32: 1257-1263. [9] Rhee Y.-J., Yi J.-H., Kim J. T., Park H., Han J.-M., Lee J. & Baik M.-G. (1998) Monitoring of Gd photoionization process by detection of fluorescence charateristics. Resoannce Ionization Spectroscopy, Ninth International Symposium J.C. Vickerman, I. Lyon, N.P. Lockyer & J.E. Parks (eds.). AIP Press Woodbury, NY. AIP Conference Proceedings 454: 261-265. [33] Rhine E. D., Sims G. K., Mulvaney R. L. & Pratt E. J. (1998) Improving the Berthelot reaction for determining ammonium in soil extracts and water. Soil Sci. Soc. Am. J., 62: 473-480. [15] Rhodes A. L. & Oreskes N. (1999) Oxygen isotope composition of magnetite deposits at E1Laco, Chile: Evidence of formation from isotopically heavy fluids. In: Geology and ore deposits of the central Andes. B. J. Skinner (ed.). Soc. Economic Geologists Inc., Spec. Publ. No. 7, Chap. 11: 333-351. [20] Ricci M. P. & Douthitt C. B. (1998) 180/16 0 in terrestrial material: Consistent and exact calculation of the 170 "ion correction" for O-containing molecules used in gas isotope ratio mass spectrometry. 5th Canadian Continuous-Flow Isotope Ratio Mass Spectrometry, Ottawa, Canada. Poster Presentation (Unpublished). [23] Rice C. A., Tuttle M. L. & Reynolds R. L. (1993) The analysis of forms of sulphur in ancient and sedimentary rocks: comments and cautions. Chem. Geol., 107: 83-95. [26] Richard L. (2001) Calculation of the standard molal thermodynamic properties as a function of temperature and pressure of some geochemically important organic sulfur compounds. Geochim. Cosmochim. Acta, 65(21): 3827-3877. [26] Richer P., Bottinga Y. & Javoy M. (1977) A review of hydrogen, carbon, nitrogen, oxygen, sulphur, and chlorine stable isotope fractionation among gaseous molecules. Ann. Rev. Earth Planet. Sci., 5: 65-110. [16,46] Richet P., Roux J. & Pineau F. (1986) Hydrogen isotope fractionation in the system H20-liquid NaA1Si308: new data and comments on D / H fractionation in hydrothermal experiments. Earth Planet. Sci. Lett., 78: 115-120. [46] Richter F. M., Davis A. M., Ebel D. S. & Hashimoto A. (2002) Elemental and isotopic fractionation of Type B calcium-, aluminum-rich inclusions: experiments, theoretical consideration, and constraints on their thermal evolution. Geochim. Cosmochim. Acta, 66: 521-540. [46] Richter R. & Hoerness S. (1988) The application of the increment method in comparison with experimentally derived and calculated O-isotope fractionations. Chemie der Erde, 48: 1-18. [20] Riciputi L. R. & Paterson B. A. (1994) High spatial-resolution measurements of O isotope ratios in silicates and carbonates by ion microprobe. Am. Mineral., 79: 1227-1230. [20] Riciputi L. R., Duckworth D. C., Barshick C. M. & Smith D. H. (1995) Isotope ratio measurements using glow discharge mass spectrometry. Int. J. Mass Spectrom. Ion Proc., 146/147: 55-64. [35] Riciputi L. R. (1996) A comparison of extreme energy filtering and high mass esolution techniques for measurement of 34S/32S ratios by ion microprobe. Rapid Commun. Mass Spectrom., 10: 282-286. [30] Riciputi L. R., Cole D. R. & Machel H. G. (1996) Sulfide formation in reservoir carbonates of the Devonian Nisku Formation, Alberta, Canada: An ion microprobe study. Geochim. Cosmochim. Acta, 60: 325-336. [26] Riciputi L. R. & Greenwood J. P. (1998) Analysis of sulfur and carbon isotope ratios in mixed matrices by secondary ion mass spectrometry: Implications for mass bias corrections. Int. J. Mass Spectrom., 178: 65-71. [30] Riebesell U., Revill A. T., Holdsworth D. G. & Volkman J. K. (2000) The effects of varying CO2 concentration on lipid composition and carbon isotope fractionation in Emiliania huxleyi. Geochim. Cosmochim. Acta, 64: 4179-4192. [46]
References
1147
Rieley G., Collier R. J., Jones D. M., Eglington G., Eakin P. A. & Fallick A. E. (1991) Sources of sedimentary lipids deduced from stable carbon-isotope analyses of individual compounds. Nature, 342: 425-427. [9] Rieley G. (1994) Derivatization of organic compounds prior to gas chromatographic - combustion isotope ratio mass spectrometric analysis: identification of isotope fractionation processes. Analyst, 119: 915-919. [8] Rieley G., Van Dover C. L., Hedrick D. B. & Eglinton G. (1999) Trophic ecology of Rimicaris exoculota: a combined lipid abundance/stable isotope approach. Marine Biol., 133: 495-499. [9] Rigby D. & Batts B. D. (1986) The isotopic composition of nitrogen in Australian coals and oil shales. Chem Geol., 58: 273-282. [16] Righi D. & Jadault P. (1988) Improving soil clay minerals studies by high-gradient magnetic separation. Clay Minerals, 23: 225-232. [2] Riley J. T., Ruba G. M. & Lee C. C. (1990) Direct determination of total organic sulfur in coal. In: Geochemistry of Sulphur in Fossil Fuels. W. L. Orr & C. M. White (eds.). American Chemical Society, Washington. ACS Symposium Series 429:429-438. [26] Rinsland C. P., Malathy Devi V., Flaud J.-M., Camy-Peyret C., Smith M. A. & Stokes G. M. (1985) Identification of 180-isotope lines of ozone in infrared ground-based solar absorption spectra. J. Geophys. Res., 90" 10719-10725. [18] Rittenberg D., Keston A. S., Rosebury E & Schoenheimer R. (1939) Studies in protein metabolism. II. The determination of nitrogen isotopes in organic compounds. J. Biol. Chem., 127: 291-299. [15] Rittenberg D. (1948) The preparation of gas samples for mass spectrographic isotope analysis. In: Preparation and measurement of isotopic tracers. D. W. Wilson, A. O. C. Nier & S. P. Reimann (eds.). J.W. Edwards, Ann Arbor, Michigan: 31-42. [15] Rittenberg D. & Ponticorvo L. (1956) A method for the determination of the 0 TM concentration of the oxygen of organic compounds. Int. J. Appl. Radiat. Isot., 1" 208-214. [23] Robert F. & Epstein S. (1982) The concentration and isotopic composition of hydrogen, carbon and nitrogen in carbonaceous meteorites. Geochim. Cosmochim. Acta, 46: 81-95. [11] Robert F. & Halbout J. (1990) A numerical model for isotopic patterns from thermal-extraction experiments. Meteoritics, 25: 291-299. [13] Robert F., Rejou-Michel A. & Javoy M. (1992) Oxygen isotopic homogeneity of the earth: New evidence. Earth Planet. Sci. Lett., 108: 1-9. [45] Roberts T. L. & Bettany J. R. (1985) The influence of topography on the nature and distribution of soil sulfur across a narrow environmental gradient. Can. J. Soil. Sci., 65: 419-434. [26] Robinson B. W. (1973) Sulphur isotope equilibrium during sulphur hydrolysis at high temperatures. Earth Planet. Sci. Lett., 18: 443-450. [46] Robinson B. W. (1974) The origin of mineralization at the Tui Mine, Te Aroha, New Zealand, in the light of stable isotope studies. Econ. Geol., 69: 910-925. [3] Robinson B. W. & Kusakabe M. (1975) Quantitative preparation of sulfur dioxide, for 34S/32S analysis, from sulfides by combustion with cuprous oxide. Anal. Chem., 47: 1179-1181. [26, 44, 45] Robinson B. W. (1995) Sulphur isotope standards. In: Reference and intercomparison materials for stable isotopes of light elements. Proceedings of a Consultants Meeting held in Vienna, 1-3 Dec 1993. IAEA-TECDOC-825, IAEA, Vienna: 39-45. [40, 43] Roedder E., Ingram B. & Hall W. E. (1963) Studies of fluid inclusions, III: extraction and quantitative analysis of inclusions in the milligram range. Econ. Geol., 58: 353-374. [1, 3] Roehl R. & Gomez J. (1995) Correction of mass bias drift in inductively coupled plasma mass spectrometry measurements of zinc isotope ratios using gallium as an isotope ratio internal standard. J. Anal. Atom. Spectrom., 10: 15-23. [31] Roether W. (1970) Water-CO2 exchange set-up for the routine oxygen-18 assay of natural waters. Int. J. Appl. Radiat. Isot., 21" 379-387. [1] Roller C. B., Weidmann D., Kosterev A., Tittel E K. & Uehara K. (2003) Quantum cascade laser spectrometer for carbonyl sulfide isotopomer analysis. 2nd International Symposium on Isotopomers, 4-6 November, Stresa, Italy. [34] Rolston J. H., den Hartog J. & Butler J. P. (1976) The deuterium isotope separation factor between hydrogen and liquid water. J. Phys. Chem., 80: 1064-1067. [1]
1148
References
Romanek C. S., Grossman E. L. & Morse J. W. (1992a) Carbon isotopic fractionation in synthetic
aragonite and calcite: Effects of temperature and precipitation rate. Geochim. Cosmochim. Acta, 56: 419-430. [46] Romanini D., Kachanov A. A., Sedeghi N. & Stoeckel F. (1997a) CW cavity ringdown spectroscopy. Chem. Phys. Lett., 264: 316-322. [34] Romanini D., Kachanov A. A. & Stoeckel F. (1997b) Diode laser cavity ringdown spectroscopy. Chem. Phys. Lett., 270: 538-545. [34] Romanini D., Kachanov A. A. & Stoeckel F. (1997c) Cavity ringdown spectroscopy: broad band absolute absorption measurements. Chem. Phys. Lett., 270: 546-550. [34] Romanini D., Kachanov A. A., Morville J. & Chenevier M. (1999) Measurement of trace gases by diode laser cavity ringdown spectroscopy. Proceedings SPIE 3821 (Environmental Sensing and Applications): 94-104. [34] Romanini D. (2003) Cavity ring down and cavity enhanced absorption spectroscopy, and trace detection, with diode lasers. Proceedings of the 4 th International Conference on Tunable Diode Laser Spectroscopy, July 14-18, 2003, Zermatt, Switzerland. [34] Rose S. & Long A. (1989) Dissolved inorganic carbon in the Tucson Basin aquifer. Ground Water, 27: 4349.[10] Rosenbaum J. M. (1993) Room temperature oxygen isotope exchange between liquid CO2 and H20. Geochim. Cosmochim. Acta, 57: 3195-3198. [46] Rosenbaum J. M. (1994) Stable isotope fractionation between dioxide and calcite at 900~ Geochim. Cosmochim. Acta, 58(17): 3747-3753. [46] Rosenbaum J., Kyser T. K. & Walker D. (1994) High temperature oxygen isotope fractionation in the enstatite-olivine-BaCO3 system. Geochim. Cosmochim. Acta, 58: 2653-2660. [46] Rosenbaum J. & Mattey D. (1995) Equilibrium garnet-calcite oxygen isotope fractionation. Geochim. Cosmochim. Acta, 59: 2839-2842. [46] Rosenbaum J. & Slagel M. M. (1995) C-O-H speciation in piston-cylinder experiments. Am. Mineralogist 80: 109-114. [46] Rosenbaum J. M., Cliff R. A. & Coleman M. L. (2000) Chlorine stable isotopes: A comparison of dual inlet and thermal ionization mass spectrometric measurements. Anal. Chem., 72: 2261-2264. [28] Rosenfeld W. D. & Silverman S. R. (1959) Carbon isotope fractionation in bacterial production of methane. Science, 130: 1658-1659. [46] Rosman K. J. R. (1972) A survey of the isotopic and elemental abundance of zinc. Geochim. Cosmochim. Acta, 36: 801-819. [31] Rosman K. J. R., De Laeter J. R. & Gorton M. P. (1980) Cadmium isotope fractionation in fractions of two H3 chondrites. Earth Planet. Sci. Lett., 48: 166-170. [31] Rosman K. J. R., Lycke W., Damen R., Werz R., Hendrickx F., Traas L. & De Bi6vre P. (1987) The preparation and use of synthetic isotope mixtures for testing mass spectrometers. Int. J. Mass Spectrom. Ion Proc., 79: 61-71. [32] Rosman K. J. R. & De Laeter J. R. (1988) Cadmium mass fractionation in unequilibrated ordinary chondrites. Earth Planet. Sci. Lett., 89: 163-169. [31] Rosman K. J. R. & Taylor P. D. P. (1998) Isotopic compositions of the elements 1997. Pure & Appl. Chem., 70(1): 217-235. [28, 32, 37] Ross P. J. & Martin E. A. (1970) A rapid procedure for preparing gas samples for nitrogen-15 determination. Analyst, 95: 817-822. [15] Ross G. J. (1969) Acid dissolution of chlorites: release of magnesium, iron and aluminium and mode of acid attack. Clays Clay Minerals, 17: 347-354. [2] Rotaru M., Birck J. L. & All6gre C. J. (1992) Clues to early solar system history from chromium isotopes in carbonaceous chondrites. Nature, 385(No. 6386): 465-470. [29] Roth E. (1997) Critical evaluation of the use and analysis of stable isotopes. Pure & Appl. Chem., 69(8): 1753-1828. [37, 40]
References
1149
Rothman L. S., Rinsland C. P., Goldman A., Massie S. T., Edwards D. P., Flaud J.-M., Perrin A., CamyPeyret C., Dana V., Mandin J.-Y., Schroeder J., McCann A., Gamache R. R., Wattson R. B., Yoshino K., Chance K. V., Jucks K. W., Brown L. R., Nemtchinov V. & Varanasi P. (1998) The HITRAN molecular spectroscopic database and HAWKS (HITRAN atmospheric Workstation): 1996 edition. J. Quant. Spectrosc. Rad. Trans., 60: 665-710. [34] Rouxel O., Ludden J., Fouquet Y. & Carignan J. (2000) Natural variations of selenium isotopes determined by multicollector plasma source mass spectrometry: Application to seafloor hydrothermal systems. J. Conf. Abstr. (Goldschmidt 2000 Conference), 5: 858-859. [31] Rouxel O., Fouquet Y. & Ludden J. (2001) Copper, selenium and sulphur isotope systematics of hydrothermal systems. J. Conf. Abstr., 6: 689. [31] Rouxel O., Ludden J., Carginan, J., Martin, L. & Fouquet Y. (2002) Natural variations of Se isotopic composition determined by hydride generation multiple collector inductively coupled plasma mass spectrometry. Geochim. Cosmochim. Acta, 66(18): 3191-3199. [20, 29] Rouxel O., Ludden J. & Fouquet Y. (2003) Antimony isotope variations in natural systems and implications for their use as geochemical tracers. Chem. Geol., 200: 25-40. [46] Rozanski K., Dulinski M. & Rzepska J. (1987) D and 180 analysis of milligram water samples. Appl. Radioat. Isot., 38: 713-716. [1] Rozanski K. & Rzepka J. (1991) Modification of D and O-18 content of water samples due to evaporation from leaky containers. Appl. Radiat. Isot., 42: 129-133. [1] RiJckmann T., Brenninkmeijer C. A. M., Neeb P. & Crutzen P. J. (1998a) Ozonolysis of non-methane hydrocarbons as a source of the observed mass independent oxygen isotope enrichment in tropospheric CO. J. Geophys. Res., 103: 1463-1470. [18] RiJckmann T., Brenninkmeijer C. A. M., Saueressig G., Bergamaschi P., Crowley J. N., Fisher H. & Crutzen P. J. (1998b) Mass-independent oxygen isotope fractionation in atmospheric CO as a result of the reaction CO + OH. Science, 281: 544-546. [18] R6ckmann T., Kaiser J., Brenninkmeijer C. A. M., Crowley J. N., Borchers R., Brand W. A. & Crutzen P. J. (2001) The isotopic enrichment of nitrous oxide (15N14NO, 14N15NO and 14N14N180) in the stratosphere and in the laboratory. J. Geophys. Res., 106: 10403-10410. [19, 34] Rudolph J., Lowe D. C., Martin R. J. & Clarkson T. S. (1997) A novel method for compound specific determination of 613C in volatile organic compounds at ppt levels in ambient air. Geophys. Res. Lett., 24: 659-662. [8] Rumble D., Palin J. M. & Hoering T. C. (1991) Laser fluorination of sulfide minerals with F2 gas. Carnegie Institution of Washington, Geophysical Laboratory. Annual Report of the Director, 1990-1991, No. 2250: 30-34. [20] Rumble D., Hoering T. C. & Palin J. M. (1993) Preparation of SF6 for sulfur isotope analysis by laser heating sulfide minerals in the presence of F2 gas. Geochim. Cosmochim. Acta, 57: 4499-4512. [20, 26, 45] Rumble D. III& Hoering T. C. (1994) Analysis of oxygen and sulfur isotope ratios in oxide and sulfide minerals by spot heating with a carbon dioxide laser in a fluorine atmosphere. Acc. Chem. Res., 27: 237-241. [20] Rumble D. III, Farquhar J., Young E. D. & Christensen C. P. (1997) In situ oxygen isotope analysis with an excimer laser using F2 and BrF5 reagents and 02 gas as analyte. Geochim. Cosmochim. Acta, 61(19): 4229-4234. [20] Rumble D. & Sharp Z. D. (1998) Laser microanalysis of silicates for 180/170 /160 and of carbonates for 180/160 and 13C/12C ratios. In" Applications of microanalytical techniques to understanding mineralizing processes. M. A. McKibben, W. C. III Shanks & W. I. Ridley (eds.). Society of Economic Geologists, Littleton, CO, USA. Vol. 7: 99-119. [20, 46] Rundel P. W., Ehleringer J. R. & Nagy K. A. (1989) Stable isotopes in ecological research. In: Ecological studies: Analysis and Synthesis, 68. W. D. Billings, F. Golley, O. L. Lange, J. Olson & H. Remmert (eds.). Springer Verlag, New York. [9] Rusak D. A., Anderson J. E., Keller R. A., Wilkerson C. W. Jr. & Kunhardt E. E. (1999) Signal dependence on light modulation frequency in optogalvanic spectroscopy. Appl. Spectrosc., 53: 1651-1653. [34]
1150
References
Russ G. P. & Bazan J. M. (1987) Isotopic ratio measurements with an inductively coupled plasma source
mass spectrometer. Spectrochim. Acta, 42B: 49-62. [31, 35] Russell J. D., Birnie A. & Fraser A. R. (1984) High-gradient magnetic separation (HGMS) in soil clay
mineral studies. Clay Minerals, 19: 771-778. [2] Russell R. D. (1971) The systematics of double spiking. J. Geophys. Res., 76(20): 4949-4955. [29] Russell R. D. (1972) Reply to discussion of "The systematics of double spiking". ]. Geophys. Res., 77(2):
381. [29]
Russell S. S., Arden J. W. & Pillinger C. T. (1991) Evidence for multiple sources of diamond from primi-
tive chondrites. Science, 54: 1188-1191. [13] Russell S. S. (1992) A carbon and nitrogen isotope study of chondritic diamond and silicon carbide. PhD
Thesis, Open University, Milton Keynes, England. [13] Russell W. A. & Papanastassiou D. A. (1978) Calcium isotope fractionation in ion-exchange chromato-
graphy. Anal. Chem., 50: 1151-1154. [29] Russell W. A., Papanastassiou D. A. & Tombrello T. A. (1978) Ca isotope fractionation on the Earth and
other solar system materials. Geochim. Cosmochim. Acta, 42: 1075-1090. [29, 31, 46] Russo R. E., Mao X. & Mao, S. S. (2002) The physics of laser ablation in microchemical analysis. Anal.
Chem., 74: 71A-77A. [20] Ryan J. N. & Gschwend P. M. (1991) Extraction of iron oxides from sediments using reductive dissolution by titanium(III). Clays Clay Minerals, 39: 509-518. [2] Ryan K. R. (1972) Ionic collision processes in gaseous N20. J. Chem. Phys., 57: 271-276. [19] Ryan P. C., Conrad M. E., Brown K., Chamberlain C. P. & Reynolds R. C. Jr. (1998) Oxygen isotope compositions of mixed-layer serpentine-chlorite and illite-smectite in hte Tuscaloosa formation (US Gulf Coast): Implications for pore fluids and mineralogic reactions. Clays Clay Minerals, 46: 357-368. [2] Rye R. O. & O'Neil J. R. (1968) The 018 content of water in primary fluid inclusions from Providencia, North-Central Mexico. Econ. Geol., 63: 232-238. [3, 20] Rye R. O., Schuiling R. D., Rye D. M. & Jansen J. B. H. (1976) Carbon, hydrogen, and oxygen isotope studies of the regional metamorphic complex at Naxos. Geochim. Cosmochim. Acta, 40: 1031-1049. [3] Rye R. O., Back W., Hanshaw B. B., Rightmare C. T. & Pearson E J. Jr. (1981) The origin and isotopic composition of dissolved sulphide in groundwaters from carbonate aquifers in Florida and Texas. Geochim. Cosmochim. Acta, 45: 1841-1950. [26] Rye R. O. (1993) The evolution of magmatic fluids in the epithermal environment. Econ. Geol., 63: 232238. [20] Saccocia P. J., Seewald J. S. & Shanks W. C. III (1998) Hydrogen and oxygen isotope fractionation between brucite and aqueous NaC1 solutions from 250 to 450~ Geochim. Cosmochim. Acta, 62: 485492. [46] Sachs J. P., Repeta D. J. & Goericke R. (1999) Nitrogen and carbon isotopic ratios of chlorophyll from marine phytoplankton. Geochim. Cosmochim. Acta, 63: 1431-1441. [46] Sadofsky S. J. & Bebout G. E. (2000) Ammonium partitioning and nitrogen isotope fractionation among coexisting micas during high temperature fluid-rock interactions: Examples from Vermont and Maine, U.S.A. Geochim. Cosmochim. Acta, 64: 2835-2849. [16] Sahoo S. K. & Masuda A. (1995) Simultaneous measurement of lithium and boron isotopes as lithium tetraborate ion by thermal ionization mass spectrometry. Analyst, 120: 335-339. [6, 7] Saito M. (1995) A contribution to the study of matrix effects in the analysis of solid samples by DC glow discharge mass spectrometry. Spectrochim. Acta, B 50: 171-178. [35] Sajjad M. I. & Tasneem M. A. (1983) An efficient hydrogen preparation system. The Nucleus, 20: 31-34.
[1]
Sakai H. & Honma H. (1966) Isotopic analysis of oxygen in rocks and minerals. I. On the conversion of
oxygen gas into carbon dioxide. Mass Spectroscopy, 14:183-186 (in Japanese; with English abstract).
[20]
Sakai H. & Yamamoto M. (1966) Fractionation of sulfur isotopes in the preparation of sulfur dioxide.
An improved technique for the precision analysis of stable sulfur isotopes. Geochem. J., 1: 35-42. [26] Sakai H., Smith J. W., Kaplan I. R. & Petrowski C. (1976) Micro-determinations of C, N, S, H, He, metallic Fe, 613C, 615N and ~)34S in geologic samples. Geochem. J., 10: 85-96. [13]
References
1151
Sakai H. & Dickson F. W. (1978) Experimental determination of the rate and disequilibrium fractionation factors of sulfur isotope exchange between sulfate and sulfide in slightly acid solutions at 300~ and 1000 bars. Earth Planet. Sci. Lett., 39: 151-161. [46] Sakai H., DesMarias D. J., Ueda A. & Moore J. G. (1984) Concentrations and isotope ratios of carbon, nitrogen and sulfur in ocean-floor basalts. Geochim. Cosmochim. Acta, 48: 2433-2441. [3] Salata G. G., Roelke L. A. & Cifuentes L. A. (2000) A rapid and precise method for measuring stable carbon isotope ratios of dissolved inorganic carbon. Marine Chem., 69: 153-161. [10] Saloman E. B. (1990) RIS data tables. Spectrochim. Acta, 45B: 37-83. [33] Saloman E. B. (1991) RIS data tables. Spectrochim. Acta, 46B: 319-378. [33] Saloman E. B. (1992) RIS data tables. Spectrochim. Acta, 47B: 517-543. [33] Saloman E. B. (1993) RIS data tables. Spectrochim. Acta, 48B: 1139-1203. [33] Saltzman E. S., Brass G. W. & Price D. A. (1983) The mechanism of sulfate aerosol formation: chemical and sulfur isotopic evidence. Geophys. Res. Lett., 10: 513-516. [26] Sanchez Chillon B., Alberdi M. T., Leone G., Bonadonna F., Stenni B. & Longinelli A. (1994) Oxygen isotopic composition of fossil equid tooth and bone phosphate: an archive of difficult interpretation. Palaeogeogr., Palaeoclimatol., Palaeoecol., 107: 317-328. [22] Sano Y., Takahata N., Nishio Y. & Marty B. (1998) Nitrogen recycling in subduction zones. Geophys. Res. Lett., 25: 2289-2292. [16] Sansone F. J., Popp B. N. & Rust T. M. (1997) Stable carbon isotope analysis of low-level methane in water and gas. Anal. Chem., 69: 40-44. [14] Santos R. V. & Clayton R. N. (1995) Variations of oxygen and carbon isotopes in carbonatites: A study of Brazilian alkaline complexes. Geochim. Cosmochim. Acta, 59: 1339-1352. [22] Santrock J., Studley S. A. & Hayes J. M. (1985) Isotopic analysis based on the mass spectrum of carbon dioxide. Anal. Chem., 57: 1444-1448. [1, 15, 44, 45] Santrock J. & Hayes J. M. (1987) Adaptation of the Unterzaucher procedure for determination of oxygen-18 in organic substances. Anal. Chem., 59: 119-127. [23] Sanyal A., Hemming N. G., Hanson G. N. & Broecker W. S. (1995) Evidence for a higher pH in the glacial ocean from boron isotopes in foraminifera. Nature, 373: 234-236. [7] Sanyal A., Hemming N. G., Broecker W. S., Lea D. W., Spero H. J. & Hanson G. N. (1996) Oceanic pH control on the boron isotopic composition of foraminifera: evidence from culture experiments. Paleoceanography, 11: 513-517. [46] Sasaki A., Arikawa Y. & Folinsbee R. E. (1979) Kiba reagent method of sulphur extraction applied to isotopic work. Bull. Geol. Surv. Japan, 30: 241-245. [26, 27] Sato R. K. & McMillan P. F. (1987) An infrared and Raman study of the isotopic species of a-quartz. J. Phys. Chem., 91: 3494-3498. [46] Sauer P. E. & Sternberg L. d. S. L. (1994) Improved method for the determination of the oxygen isotopic composition of cellulose. Anal. Chem., 66: 2409-2411. [23] Sauke T. B. & Becker J. F. (1998) Stable isotope laser spectrometer for exploration of Mars. Planet. Space Sci., 46: 805-812. [34] Saurer M., Borella S. & Leuenberger M. (1997a) 6180 of tree rings of beech (Fagus sylvatica) as a record of 6180 of growing season precipitation. Tellus, 49B: 80-92. [23] Saurer M., Aellen K. & Siegwolf R. (1997b) Correlating 613C and 6180 in cellulose of trees. Plant Cell Environm., 20: 1543-1550. [46] Saurer M., Robertson I., Siegwolf R. & Leuenberger M. (1998) Oxygen isotope analysis of cellulose: An interlaboratory comparison. Anal. Chem., 70(10): 2074-2080. [23] Savarino J. & Thiemens M. H. (1999a) Analytical procedure to determine both 6180 and 6170 of H202 in natural water and first measurements. Atmos. Environ., 33: 3683-3690. [18] Savarino J. & Thiemens M. H. (1999b) Mass independent oxygen isotope (160, 170, 180) fractionation found in the H+O2+M reaction. J. Chem. Phys., A 103: 9221-9229. [18] Savin S. M. (1967) Oxygen and hydrogen isotope ratios in sedimentary rocks and minerals. PhD Thesis (unpublished), California Institute of Technology. [2, 20] Savin S. M. & Epstein S. (1970) The oxygen and hydrogen isotope geochemistry of clay minerals. Geochim. Cosmochim. Acta, 34: 25-42. [2, 46]
1152
References
Saxton J. M., Lyon I. C., Chatzitheodoridis E., Perera I. K., Van Lierde P., Freedman P. & Turner G. (1996) The Manchester Isolab 54 ion microprobe. Int. J. Mass Spectrom. Ion Proc., 154: 99-131. [30] Scalan R. S. (1958) The isotopic composition, concentration, and chemical state of the nitrogen in igneous rocks. PhD Thesis, The University of Arkansas: 79 p. [16] Schadewaldt P., Schommartz B., Wienrich G., BriJsicke H., Piolot R. & Ziegler D. (1997) Application of isotope-selective nondispersive infrared spectroscopy (IRIS) for evaluation of [13C]octanoic acid gastric-emptying breath tests: comparison with isotope ratio-mass spectrometry (IRMS). Ctin. Chem., 43(3): 518-522. [34] Scheele N. & Hoers J. (1992) Carbon isotope fractionation between calcite, graphite and CO2: an experimental study. Contrib. Mineral. Petrol., 112: 35-45. [46] Schell D. M. (1978) Chemical and isotopic methods in nitrification studies. In: Microbiology-1978. D. Schlesinger (ed.). Am. Soc. Microbiol., Washington: 292-295. [15] Schelles W., De Gendt S., Muller V. & Van Grieken R. (1995) Evaluation of secondary cathodes for glow discharge mass spectrometry analysis of different nonconducting sample types. Appl. Spectrosc., 49: 939-944. [35] Schelles W., De Gendt S. & Van Grieken R. (1996) Optimization of secondary cathode thickness for direct current glow discharge mass spectrometric analysis of glass. J. Anal. At. Spectrom., 11: 937-941.
[35]
Schelles W. & Van Grieken R. (1996) Direct current glow discharge mass spectrometric analysis of a Macor ceramic using a secondary cathode. Anal. Chem., 68: 3570-3574. [35] Scherer J. J., Paul J. B., O'Keefe A. & Saykally R. J. (1997) Cavity ringdown laser absorption spectroscopy. History, development, and application to molecular beams. Chem. Rev., 97: 25-51. [34] Schidlowski M., Hayes J. M. & Kaplan I. R. (1983) Isotopic interferences of ancient biochemistry of carbon, sulfur, hydrogen and nitrogen. In: Earth's earliest biosphere: Its origin and evolution. J. W. Schopf (ed.). Princeton University Press, Princeton: 149-186. [27] Schiegl W. E. & Vogel J. C. (1970) Deuterium content of organic matter. Earth Planet. Sci. Lett., 7: 307313. [1] Schiffries C. M. & Rye D. M. (1990) Stable isotopic systematics of the Bushveld complex: II. Constraints on hydrothermal processes in layered intrusions. Am. J. Sci., 290: 209-245. [3] Schimmelmann A. & DeNiro M. J. (1985) Determination of oxygen stable isotope ratios in organic matter containing carbon, hydrogen, oxygen, and nitrogen. Anat. Chem., 57: 2644-2646. [23] Schimmelmann A. (1991) Determination of the concentration and stable isotopic composition of nonexchangeable hydrogen in organic matter. Anal. Chem., 63: 2456-2459. [24] Schimmelmann A. & DeNiro M. J. (1993) Preparation of organic and water hydrogen for stable isotope analysis: Effects due to reaction vessels and zinc reagents. Anal. Chem., 65: 789-792. [1, App C] Schleser G. H. (1990) Investigations of the 613C pattern in leaves of Fagus sylvatica L. J. Experim. Botanic, 41: 565-572. [24] Schlesinger W. H. (1997) Biogeochemistry - An analysis of global change. Academic, San Diego, California: 588 p. [15] Schmidt H.-L., Kempe K., Demmelmair J. & Medina R. (1992) On-line gas chromatography and isotope ratio mass spectrometry of organic compounds in the range of natural abundances. Kyoto '92 Int. Conf. Biol. Mass Spectrom. Osaka University, Japan, Proceedings: 382-383. [23] Schmidt M., Botz R., Stoffers P., Anders T. & Bohrmann G. (1997) Oxygen isotopes in marine diatoms: a comparative study of analytical techniques and new results on the isotope composition of recent marine diatoms. Geochim. Cosmochim. Acta, 61: 2275-2280. [46] Schmutz J. A. & Hobson K. A. (1998) Geographic, temporal, and age-specific variation in diets of glaucous gulls in western Alaska. The Condor, 100: 119-130. [9] Schoell M., Faber E. & Coleman M. (1983) Carbon and hydrogen isotopic compositions of the NBS-22 and NBS-21 stable isotope reference materials: An Inter-laboratory comparison. Org. Geochem., 5: 36. [40] Schoenau J. J. & Bettany J. R. (1988) A method for determining sulfur-34 abundance in soil and plant sulfur fractions. Soil Sci. Soc. Am. J., 52: 297-300. [26] Schoenau J. J. & Bettany J. R. (1989) 34S natural abundance variations in prairie and boreal forest soils. J. Soil Sci., 40: 397-413. [26]
References
115 3
Schoeninger M. J., Iwaniec U. T. & Nash L. T. (1998) Ecological attributes recorded in stable isotope ratios of aboreal prosimian hair. Oecology, 113: 222-230. [9] Schopf J. W. (1983) Earth's earliest biosphere: Its origins and evolution. Princeton University Press, Princeton: 543 p. [13]
Schopf J. W. & Klein C. (1992) The Proterozoic biosphere. Cambridge University Press, Cambridge, England: 1348 p. [13]
Schueler B., Morton J. & Mauersberger K. (1990) Measurements of isotopic abundances in collected stratospheric ozone samples. Geophys. Res. Lett., 17: 1295-1298. [18] Schumacher K., Turgeon H. & Mosandl A. (1995) Sample preparation for gas-chromatography isotope ratio mass-spectrometry - an investigation with volatile components from strawberries. Phytochem. Anal., 6: 258-261. [8] Schupp M., Bergamaschi P., Harris G. W. & Crutzen P. J. (1993) Development of a tunable diode laser absorption spectrometer for measurements of the 13C/12C ratio in methane. J. Chemosphere, 26: 1322. [34] Schiitze M. (1939) Ein neues Verfahren zur direkten Sauerstoffbestimmung in organischen Sub-stanzen. Fresenius Z. Anal. Chem., 118: 245-258. [23] Schwander H. (1953) Bestimmung des relativen Sauerstoffisotopen-Verhatnisses in Silikatgesteinen und Mineralen. Geochim. Cosmochim. Acta, 4: 261-291. [20] Schwartz H. P., Agyei F. K. & McMullen C. C. (1969) Boron isotopic fractionation during clay absorption from sea-water. Earth Planet. Sci. Lett., 6: 1-5. [7, 46] Schwarcz H. P., Harmon R. S., Thompson P. & Ford D. C. (1976) Stable isotope studies of fluid inclusions in speleothems and their paleoclimatic significance. Geochim. Cosmochim. Acta, 40: 657-665. [3] Schwarcz H. P., Melbye J., Katzenberg M. A. & Knyf M. (1985) Stable isotopes in human skeletons of southern Ontario: reconstructing palaeodiet. J. Archaeol. Sci., 12: 187-206. [22] Schwarz G., Ohm K., Yamanouchi T., Furukawa T., Kowski P. & Gernandt H. (1996) Stable isotopic composition of Antarctic air moisture and precipitation. Nankyoku Shiry6 (Antarctic Record), 40: 169178. [36] Schwarz G., Kowski P. & Gernandt H. (1998a) Saisonale Variation von Deuterium und Sauerstoff-18 in Luftfeuchte und Niederschlag tiber Antarktika. Isot. Environ. Health Stud., 34: 159-168. [36] Schwarz G., Kowski P., Kaneto S. & Gernandt H. (1998b) Meridional distributions of deuterium in atmospheric water vapour between tropical and southern polar latitudes. Mem. Natl. Inst. Polar Res., 52 (Special Issue): 102-110. [36] Scrimgeour C. M., Rollo M. M., Mudambo S. M. K. T., Handley L. L. & Prosser S. J. (1993) A simplified method for deuterium/hydrogen isotope ratio measurements on water samples of biological origin. J. Biol. Mass Spectrom., 22: 383-387. [1] Scrimgeour C. M. (1995) Measurement of plant and soil water isotope composition by direct equilibration methods. J. Hydrol., 172: 261-274. [1, 21] Scrimgeour C. M., Begley I. S. & Thomason M. L. (1999) Measurement of deuterium incorporation into fatty acids by gas chromatography isotope ratio mass spectrometry. Rapid Commun. Mass Spectrom., 13: 271-274. [8] Seaton N. A., Friedman S. P., MacElroy J. M. D. & Murphy A. (1997) The molecular sieving mechanism in carbon molecular sieves: a molecular dynamics and critical path analysis. Langmuir, 13:1199-1204. [36] Seboldt W., Linger S., Hoernes S., Grimmeisen W., Lekies R., Herkelmann R. & Burr D. (1993) Lunar oxygen extraction using fluorine. In: Resources of near-Earth space. J. S. Lewis, M. S. Matthews & M. L. Guerrieri (eds.). Univ. Arizona Press, Tucson: 129-147. [20] Segschneider H.-J., Hofmann D., Schmidt G. & Russow R. (1995) Incorporation of 15NO2 nitrogen into individual amino acids by suflowers using GC-C-IRM. Isot. Environ. Health Stud., 31: 315-325. [8] Seiler R. L. (1998) Prediction of lands susceptible to irrigation-induced selenium contamination of water. In: Environmental chemistry of selenium. W. T. Jr. Frankenberger & R. A. Engberg (eds.). Marcel Dekker, Inc., Vo|. 64: 397-418. [29]
1154
References
Selby D. & Nesbitt B. E. (1996) British Columbia; implications for P-V-T-X history of ore and post ore fluids. Biennial Pan-American Conference on Research on Fluid Inclusions, 6. PACROFI VI; Sixth biennial Pan-American Conference on Research on Fluid Inclusions; Program and Abstracts: 116117. [3] Sennaroglu A., Konca A. O. & Pollock C. R. (2000) Continuous-wave power performance of a 2.47 ~m Cr2+:ZnSe laser: Experiment and modeling. IEEE J. Quantum El., 36: 1199-1205. [34] Sephton M. A., Pillinger C. T. & Gilmour I. (1998) 613C of free and macromolecular aromatic structures in the Murchison meteorite. Geochim. Cosmochim. Acta, 62: 1821-1828. [11] Sephton M. A., Pillinger C. T. & Gilmour I. (1999) Small-scale hydrous pyrolysis of macromolecular material in meteorites. Planet. Space Sci., 47: 181-187. [11] Sephton M. A., Pillinger C. T. & Gilmour I. (2001) Normal alkanes in meteorites: molecular 613C values indicate an origin by terrestrial contamination. Precambrian Res., 106: 45-56. [11] Sephton M. A. & Gilmour I. (2001) Pyrolysis-gas chromatography-isotope ratio monitoring-mass spectrometry of macromolecular material in meteorites. Planet. Space Sci., 49: 465-471. [11] Seyfried W. E. Jr., Janecky D. R. & Brendt M. E. (1987) Rocking autoclaves for hydrothermal experiments II: the flexible cell system. In: Experimental Hydrothermal Techniques. G. Ulmer & H. L. Barnes (eds.). Wiley-Interscience, New York: 216-240. [46] Shackleton C. H. L., Roitman E., Phillips A. & Chang T. (1997) Androstanediol and 5-androstenediol profiling for detecting exogenously administered dihydrotestosterone, epitestosterone, and dehydroepiandrosterone: Potential use in gas chromatography isotope ratio mass spectrometry. Steroids, 62: 665-673. [8] Shackleton N. J. (1974) Attainment of isotopic equilibrium between ocean water and the benthonic foraminifera genus Uvigerina: Isotopic changes in the ocean during the last glacial. In: Variation du climat au cours du Pleistoc6ne. J. Labeyrie (ed.). Colloq. Intern. CNRS, Paris, Vol. 219: 203-209. [46] Shahack-Gross R., Shemesh A., Yakir D. & Weiner S. (1996) Oxygen isotopic composition of opaline phytoliths: potential for terrestrial climatic reconstruction. Geochim. Cosmochim. Acta, 60: 3949-3953. [46] Shanks W. C. III, Boehlke J. K. & Seal R. R. II (1995) Stable isotopes in mid-ocean ridge hydrothermal systems: Interaction between fluids, minerals and organisms. In: Seafloor hydrothermal systems: Physical. chemical, biological, and geological interactions. S. E. Humphris. et al. AGU, Monograph 91: 194-221. [1] Shanks W. C., Crowe D. E. & Johnson C. (1998) Sulfur isotope analyses using the laser microprobe. In: Applications of microanalytical techniques to understanding mineralizing processes. M. A. McKibben, W. C. III Shanks & W. I. Ridley (eds.). Society of Economic Geologists, Littleton, CO, USA, Vol. 7: 141-153. [46] Sharp Z. D. & O'Neil J. R. (1988) A laser-based carbon reduction technique for oxygen isotope analysis of microsamples of silicates and oxides. Geol. Soc. Am. Abstr. w. Prog., 20:A280 (Abstract). [20] Sharp Z. D. & O'Neil J. R. (1989) A laser-based carbon reduction technique for oxygen isotope analysis of silicates and oxides. In: Annual Report of the Director, Geophysical Laboratory: 1988-1989. Carnegie Institution, Washington, Vol. 89: 72-78. [20] Sharp Z. D. (1990) A laser-based microanalytical method for the in situ determination of oxygen isotope ratios of silicates and oxides. Geochim. Cosmochim. Acta, 54: 1353-1357. [2, 20] Sharp Z. D., Giletti B. J. & Yoder H. S. Jr. (1991) Oxygen diffusion rates in quartz exchanged with CO2. Earth Planet. Sci. Lett., 107: 339-348. [46] Sharp Z. D. (1992) In situ laser microprobe techniques for stable isotope analysis. Chem. Geol. (Isot. Geosci. Sect.), 101: 3-19. [26, 22] Sharp Z. D. & Jenkin G. R. T. (1994) An empirical estimate of the diffusion rate of oxygen in diopside. J. Metamorphic Geol., 12: 89-97. [46] Sharp Z. D. & Kirschner D. L. (1994) Quartz-calcite oxygen isotope thermometry: A calibration based on natural isotopic variations. Geochim. Cosmochim. Acta, 58: 4491-4501. [46] Sharp Z. D. & Moecher D. (1994) O-isotope variations in a porphyroclastic metaanorthosite: Diffusion effects and false isotherms. Am. Mineralogist, 79: 951-959. [20] Sharp Z. D. (1995) Oxygen isotope geochemistry of the A12SiO5 polymorphs. Am. J. Sci., 295: 1058-1076. [20]
References
1155
Sharp Z. D. & Cerling T. E. (1996) A laser GC-IRMS technique for in situ stable isotope analyses of carbonates and phosphates. Geochim. Cosmochim. Acta, 60(15): 2909-2916. [22, 45] Sharp Z. D. & Cerling T. E. (1998) Fossil isotope records of seasonal climate and ecology. Straight from the horse's mouth. Geology, 26: 219-222. [22] Sharp Z. D., Atudorei V. & Furrer H. (2000) The effect of diagenesis on oxygen isotope ratios of biogenic phosphates. Am. J. Sci., 300: 222-237. [22] Sharp Z. D., Atudorei V. & Durakiewicz T. (2001) A rapid method for the determination of hydrogen and oxygen isotope ratios of water from water and solid hydrous substances. Chem. Geol., 178: 197210. [1, 3] Shaw S. Y. & Lue J. T. (1980) An integrated circuit based vacuum ionization gauge meter. J. Phys. E.: Sci. Instrum., 13: 1150-1152. [39] Shearer G. & Kohl H. (1993) Natural abundance of 15N: Fractional Contribution of two sources to a common sink and use of isotope discrimination. In: Nitrogen isotope Techniques. R. Knowles & T. H. Blackburn (eds.). Academic Press Inc.: 89-125. [15] Sheets R. W., Nesbitt B. E. & Muehlenbachs K. (1996) Meteoritic water component in magmatic fluids from porphyry copper mineralization, Babine Lake area, British Columbia. Geology, 24: 1091-1094.
[3]
Shelton K. L. (1983) Composition and origin of ore-forming fluids in a carbonate-hosted porphyry copper and skarn deposit: a fluid inclusion and stable isotope study of Mines Gasp6, Quebec. Econ. Geol., 78: 387-421. [3] Shelton K. L., Taylor R. P. & So C. S. (1987) Stable isotope studies of the Dae Hwa Tungsten-Molybdenum Mine, Republic of Korea: evidence of progressive meteoric water interaction in Tungstenbearing hydrothermal system. Econ. Geol., 82: 471-481. [3] Shemesh A., Kolodny Y. & Luz B. (1988) Isotope geochemistry of oxygen and carbon in phosphate and carbonate of phosphorite francolite. Geochim. Cosmochim. Acta, 52: 2565-2572. [22] Shemesh A. (1990) Crystallinity and diagenesis of sedimentary apatites. Geochim. Cosmochim. Acta, 54: 2433-2438. [22] Shemesh A. (1992) Response to R. N. Clayton on "Silica-carbonate isotopic temperature calibration". Science, 258: 1163. [46] Shemesh A., Charles C. D. & Fairbanks R. G. (1992a) Oxygen isotopes in biogenic silica: global changes in ocean temperature and isotopic composition. Science, 256: 1434-1436. [46] Shemesh A., Ron H., Erel Y., Kolodny Y. & Nur A. (1992b) Isotopic composition of vein calcite and its fluid inclusions: implication to paleohydrologica systems, tectonic events and vein formation processes. Chem. Geol., 94: 307-314. [3] Sheppard S. M. F., Nielsen R. L. & Taylor H. P. Jr. (1969) Oxygen and hydrogen isotope ratio of clay minerals from porphyry copper deposits. Econ. Geol., 64: 755-777. [2, 46] Sheppard S. M. F. & Schwarcz H. P. (1970) Fractionation of carbon and oxygen isotopes and magnesium between coexisting metamorphic calcite and dolomite. Contrib. Mineral. Petrol., 26: 161-198. [46] Sheppard S. M. F., Nielsen R. L. & Taylor H. P. Jr. (1971) Hydrogen and oxygen isotope ratios in minerals from porphyry copper deposits. Econ. Geol., 66: 515-542. [2] Sheppard S. M. F. (1980) Isotopic evidence for the origins of water during metamorphic processes in oceanic crust and ophiolite complexes. In: Associations mafiques ultramafiques dans les orog6nes, CNRS, Colloq. Intern. No. 272: 135-147. [46] Sheppard S. M. F. (1981) Stable isotope geochemistry of fluids. In: Chemistry and geochemistry of solutions at high temperatures and pressures. D. T. Rickard & F. E. Wickman (eds.). Physics and Chemistry of the Earth, Pergamon Press, Vol. 13 & 14: 419-445. [46] Sheppard S. M. F. & Gilg H. A. (1996) Stable isotope geochemistry of clay minerals. Clay Minerals, 31: 1-24. [2, 46] Sherwood B., Fritz P., Frape S. K., Macko S. A., Weise S. M. & Whelhan J. A. (1988) Methane occurences in the Canadian Shield. Chem. Geol., 71: 223-236. [16] Sheu D. D. & Chiu C. H. (1995) Evaluation of cellulose extraction procedures for stable carbon isotope measurement in tree ring research. Int. J. Environ. Anal. Chem., 59: 59-67. [24] Shields W. R., Murphy T. J., Garner E. L. & Dibeter V. H. (1962) Absolute isotopic abundance ratios and the isotopic weight of chlorine. J. Am. Chem. Soc., 84: 1519-1522. [28, 40]
1156
References
Shields W. R., Goldich S. S., Garner E. L. & Murphy T. J. (1965) Natural variations in the abundance ratio and the atomic weight of copper. J. Geophys. Res., 70: 479-491. [31] Shields W. R., Murphy T. J., Catanzaro E. J. & Garner E. L. (1966) Absolute isotopic abundance ratios and the atomic weight of a reference sample of chromium. J. Res. Nat. Bur. Stand., 70A: 193-197. [29] Shima M. (1962) Boron in meteorites. J. Geophys. Res., 67: 4521-4523. [7] Shima M. & Honda M. (1963) Isotopic abundance of meteoritic lithium. J. Geophys. Res., 68: 2849-2854. [6] Shimamura T., Takahashi T., Honda M. & Nagai H. (1993) Multi-element and isotopic analyses of iron meteorites using a glow discharge mass spectrometer. J. Anal. At. Spectrom., 8: 453-460. [35] Shimizu N. & Hart S. R. (1982a) Applications of the ion microprobe to geochemistry and cosmochemistry. Ann. Rev. Earth Planet. Sci., 10: 483-526. [6, 30] Shimizu N. & Hart S. R. (1982b) Isotope fractionation in secondary ion mass spectrometry. J Appl. Phys., 53(3): 1303-1311. [30] Shmulovich K., Landwehr D., Simon K. & Heinrich W. (1999) Stable isotope fractionation between liquid and vapour in water-salt systems up to 600~ Chem. Geol., 157: 343-354. [46] Shouakar-Stash O., Drimmie R., Morrison J., Frape S. K., Heemskerk A. R. & Mark W. A. (2000) Online D / H analysis for water, natural gas and organic solvents by manganese reduction. Anal. Chem., 72: 2664-2666. [1] Shukolyukov A. & Lugmair G. W. (1998) Isotopic evidence for the Cretaceous-Tertiary impactor and its type. Science, 282(No. 5390): 927-929. [29] Sickman J. O., Leydecker A., Chang C. C. Y., Kendall C., Melack J. M., Lucero D. M. & Schimel J. (2003) Mechanisms underlying export of N from high-elevation catachments during seasonal transitions. Biogeochem., 64: 1-24. [15] Siebert C., N/igler T. F. & Kramers J. D. (2001) Determination of molybdenum isotope fractionation by double-spike multicollector inductively coupled plasma mass spectrometry. Geochem. Geophys. Geosys., 2, 2000GC000124. [31] Siegfriedt R. K., Wiberley J. S. & Moore R. W. (1951) Determination of sulfur after combustion in a small oxygen bomb. Rapid titrimetric method. Anal. Chem., 23: 1008-1011. [26] Sigman D. M., Altabet M. A., Michener R., McCorckle D. C., Fry B. & Holmes R. M. (1997) Natural abundance-level measurement of the nitrogen isotopic composition of oceanic nitrate: an adaptation of the ammonia diffusion method. Marine Chem., 57: 227-242. [15] Sigman D. M., Casciotti K. L., Andreani M., Barford C., Galanter M. & Bi~hlke J. K. (2001) A bacterial method for the nitrogen isotopic analysis of nitrate in seawater and freshwater. Anal. Chem., 73: 41454153.[15] Silfer J. A., Engel M. H., Macko S. A. & Jumeau E. J. (1991) Stable carbon isotope analysis of amino acid enantiomers by conventional isotope ratio mass spectrometry and combined gas chromatography/isotope ratio mass spectrometry. Anal. Chem., 63: 370-374. [9] Silva S. R., Kendall C., Wilkison D. H., Ziegler A. C., Chang C. C. Y. & Avanzino R. J. (2000) A new method for collection of nitrate from fresh water and the analysis of nitrogen and oxygen isotope ratios. J. Hydrol., 228: 22-36. [15] Silver J. A. (1992) Frequency-modulation spectroscopy for trace gas detection: theory and comparison among experimental methods. Appl. Opt., 31: 707-717. [34] Silver L. A. & Stolper E. M. (1989) Water in albitic glasses. J. Petrol., 30: 667-709. [46] Silverman S. R. (1951) The isotope geology of oxygen. Geochim. Cosmochim. Acta, 2: 26-42. [20] Silverman S. R. (1964) Investigations of petroleum origin and evolution mechanisms by carbon isotope studies. In: Isotopic cosmic chemistry. H. Craig, G. J. Miller & G. J. Wasserburg (eds.). North Holland: 92-102. [40] Simon K. (2001) Does 6D from fluid inclusion in quartz reflect the original hydrothermal fluid? Chem. Geol., 177: 483-495. [3] Simpson I. A., Evershed R. P., Bull I. D. & Dockrill S. J. (1997) Identification of manuring practices associated with early anthropogenic soil formation in Orkney, Scotland. Abstr. of Papers Am. Chem. Soc., 213: 54-GEOC. [8]
References
1157
Skorupa J. P. (1998) Selenium poisening of fish and wildlife in nature: Lessons from twelve real-worldexamples. In: Environmental chemistry of selenium. W. T. Jr. Frankenberger & R. A. Engberg (eds.). Marcel Dekker Inc., Vol. 64: 315-354. [29] Skulan J., DePaolo D. J. & Owens T. L. (1997) Biological control of calcium isotopic abundances in the global calcium cycle. Geochim. Cosmochim. Acta, 61(12): 2505-2510. [29] Slodzian G., Chaintreau M. & Dennebouy R. (1986) The emission objective lens working as an electron mirror: self regulated potential at the surface of an insulating sample. In: Secondary ion mass spectrometry. A. Benninghoven, D. S. Colton, D. S. Simons & H. W. Werner (eds.). Springer, SIMS V: 158-160. [30] Slodzian G., Chaintreau M., Dennebuoy R. & Rousse A. (2001) Precise in situ measurements of isotopic abundances with pulse counting of sputtered ions. Eur. Phys. J. Appl. Phys., 14: 199-231. [30] Smith B. F. L. (1994) Characterisation of poorly ordered minerals by selective chemical methods. In: Clay mineralogy: Spectroscopy and chemical determinative methods. M. J. Wilson (ed.). Chapman & Hall, London: 333-357. [2] Smith B. N. & Epstein S. (1971) Two categories of 13C/12C ratios for higher plants. Plant Physiol., 47: 380-384. [46] Smith D., Adams N. G. & Henchman M. J. (1980) Studies of the binary reactions of H30+.(H20)0,1,2 ions and their deuterated analogues with D20, H20 and NH3. J. Chem. Phys., 72: 4951-4957. [4] Smith D. & Adams N. G. (1988) The selected ion flow tube (SIFT). Studies of ion-neutral reactions. Adv. Atom. Mol. Phys., 24: 1-49. [4] Smith D. & Spanel P. (1996a) Application of ion chemistry and the SIFT technique to the quantitative analysis of trace gases in air and on breath. Int. Reviews Phys. Chem., 15: 231-271. [4] Smith D. & Spanel P. (1996b) The novel selected-ion flow tube approach to trace gas analysis of air and breath. Rapid Commun. Mass Spectrom., 10:1183-1198 [4]. Smith D., Wang T. S. & Spanel P. (2002a) Kinetics and isotope patterns of ethanol and acetaldehyde emissions from yeast fermentations of glucose and glucose-6,6-d2 using selected ion flow tube mass spectrometry: a case study. Rapid Commun. Mass Spectrom., 16: 69-76. [4] Smith D., Wang T. S. & Spanel P. (2002b) On-line, simultaneous quantification of ethanol, some metabolites and water vapour in breath following the ingestion of alcohol. Physiol. Meas., 23, 477-489. [4] Smith J. W., Young N. B. & Lawlor D. L. (1964) Direct determination of sulphur forms in Green River oil shale. Anal. Chem., 36:618-622. [26] Smith J. W. & Kaplan I. R. (1970) Endogenous carbon in carbonaceous meteorites. Science, 167: 13671370.[11] Smith J. W. &Batts B. D. (1974) The distribution and isotopic composition of sulphur in coal. Geochim. Cosmochim. Acta, 38: 121-133. [26] Smith J. W. & Croxford R. J. W. (1975) An isotopic investigation of the environment of deposition of the McArthur mineralisation. Miner. Deposita, 10: 269-276. [2] Smith M.-A. H., Rinsland C. P., Fridovich B. & Rao K. N. (1985) Intensities and collision broadening parameters from infrared spectra. In: Molecular spectroscopy: Modern research. K. N. Rao (ed.). Academic Press, Orlando, Vol. III. [34] Smith M. P. & Yardley B. W. D. (1996) The boron isotopic composition of tourmaline as a guide to fluid processes in the southwestern England orefield: An ion microprobe study. Geochim. Cosmochim. Acta, 60: 1415-1427. [30] Smithwick R. W., Lynch D. W. & Franklin J. C. (1993) Relative ion yields measured with a high-resolution glow discharge mass spectrometer operated with an Argon/Hydrogen mixture. J. Am. Soc. Mass Spectrom., 4: 278-285. [35] Smykatz-Kloss W. (1974) Differential thermal analysis. Springer, Berlin: 185 p. [13, 16] So C.-S., Rye D. M. & Shelton K. L. (1983) Carbon, hydrogen, oxygen, and sulphur isotope and fluid inclusion study of the Weolog tungsten-molybdenum deposit, republic of Korea: Fluid histories of metamorphic and ore-forming events. Econ. Geol., 78: 1551-1573. [3] Socki R. A., Karlsson H. R. & Gibson E. K. Jr. (1992) Extraction technique for the determination of oxygen-18 in water using preevacuated glass vials. Anal. Chem., 64: 829-831. [1] Socki R. A., Romanek C. S. & Gibson E. K. Jr. (1999) On-line technique for measuring stable oxygen and hydrogen isotopes from microliter quantities of water. Anal. Chem., 71: 2250-2253. [1, 20]
1158
References
Sofer Z. & Gat J. R. (1972) Activities and concentrations of oxygen-18 in concentrated aqueous salt solutions: Analytical and geophysical implications. Earth Planet. Sci. Lett., 15: 232-238. [46] Sofer Z. & Gat J. R. (1975) The isotope composition of evaporating brines: Effect of the isotopic activity ratio in saline solutions. Earth Planet. Sci. Lett., 26: 179-186. [46] Sohns E., Gerling P. & Faber E. (1994) Improved stable nitrogen isotope ratio measurements of natural gases. Anal. Chem., 66: 2614-2620. [8, 43] Solarzano L. (1969) Determination of NH4 + in natural water by the phenolhypochlorite method. Limnol. Oceanogr., 14: 799-801. [9] Song T. R. & Ding T. P. (1990) A new probe on application of silicon isotopes composition (630Si) in siliceous rocks to sedimentary facies analysis. Chinese Sci. Bull., 35: 761-766. [25] Sorokina I. T., Sorokin E., DiLieto A., Tonelli M., Page R. H. & Schaffers K. I. (2001) Efficient broadly tunable continuous-wave Cr2+:ZnSe laser. J. Opt. Soc. Am., B 18: 926-930. [34] Sorensen P. & Jensen E. S. (1991) Sequential diffusion of ammonium and nitrate from soil extracts to a polytetrafluoroethylene trap for 15N determination. Anal. Chim. Acta, 252: 201-203. [15] Si~derholm S., Meinander N. & Kauppinen J. (1998) The 13C/12C isotopic ratio determined by FTIR spectroscopy. AIP Conf. Proc., No. 430: 223-226. [34] Si~derlund R. & Svensson B. H. (1976) The global nitrogen cycle. Ecol. Geol. (Stockholm), 22(SCOPE Report 7), 23073. [15] Spanel P. & Smith D. (1996) The selected ion flow tube, a novel technique for the quantitative trace gases analysis of air and human breath. Med. Biol. Eng. Comp., 34: 409-419. [4] Spanel P., Davies S. & Smith D. (1998) Quantification of ammonia in human breath by the selected ion flow tube analytical method using H30 + and 02 + precursor ions. Rapid Commun. Mass Spectrom., 12: 763-766. [4] Spanel P., Smith D., Holland T. A., A1 Singary W. & Elder J. B. (1999) Analysis of formaldehyde in the headspace of urine from bladder and prostate cancer patients using selected ion flow tube mass spectrometry. Rapid Commun. Mass Spectrom., 13: 1354-1359. [4] Spanel P. & Smith D. (2000) Selected flow tube mass spectrometry analyses of stable isotopes in water: Isotopic composition of H30 + and H30+(H20)3 ions in exchange reactions with water vapor. J. Am. Soc. Mass Spectrom., 11: 866-875. [4] Spanel P. & Smith D. (2001) Accuracy and precision of flowing afterglow mass spectrometry for the determination of the deuterium abundance in hte headspace of aqueous liquids and exhaled breath water. Rapid Commun. Mass Spectrom., 15: 867-872. [4] Sparks J. P. & Ehleringer J. R. (1997) Leaf carbon isotope discrimination and nitrogen content for riparian trees along elevational transects. Oecologia, 109: 362-367. [9] Spicuzza M. J., Valley J. w., Kohn M. J., Girard J. P. & Fouillac A. M. (1998a) The rapid heating, defocused beam technique: a CO2-1aser-based method for highly precise and accurate determination of 6180 values of quartz. Chem. Geol., 144: 195-203. [20] Spicuzza M. J., Valley J. W. & McConnell V. S. (1998b) Oxygen isotope analysis of whole rock via laser fluorination: An air-lock approach. Geol. Soc. Am. Abstr. w. Prog., 30(7): 80 (Abstract). [2, 20] Spivack A. J. & Edmond J. M. (1986) Determination of boron isotope ratios by thermal ionization mass spectrometry of the dicesium metaborate cation. Anal. Chem., 58: 31-35. [7, 40] Spivack A. J. & Edmond J. M. (1987) Boron isotope exchange between seawater and the oceanic crust. Geochim. Cosmochim. Acta, 51: 1033-1043. [40] Spivack A. J., Palmer M. R. & Edmond J. M. (1987) The sedimentary cycle of the boron isotopes. Geochim. Cosmochim. Acta, 51: 1939-1949. [7] Spivack A. J., You C.-F. & Smith H. J. (1993) Foraminiferal boron isotope ratios as a proxy for surface ocean pH over the past 21 Myr. Nature, 363: 149-151. [7] Spivack A. J. & You C.-F. (1997) Boron isotopic geochemistry of carbonates and pore waters, Ocean Drilling Program Site 851. Earth Planet. Sci. Lett., 152: 113-122. [7] Spooner N., Rieley G., Collister J. W., Lander M., Cranwell P. A. & Maxwell J. R. (1994) Stable carbon isotopic correlation of individual biolipids in aquatic organisms and a lake bottom sediment. Org. Geochem., 21: 823-827. [9]
References
1159
Spratt H. G. & Morgan M. D. (1990) Sulfur cycling over an annual cycle in a cedar dominated freshwater wetland. Limnol. Oceanogr., 35: 1586-1593. [26] Sprinson D. B. & Rittenberg D. (1948) Preparation of gas samples for mass spectrometric analysis of isotope abundance. U. S. Naval Med. Bull., Supplement 115 (March-April): 82-93. [15] Sprinson D. B. & Rittenberg D. (1949) The rate of utilization of ammonia for protein synthesis. J. Biol. Chem., 180: 707-714. [15] Sridhar K., Jackson M. L. & Clayton R. N. (1975) Quartz oxygen isotopic stability in relation to isolation from sediments and diversity of source. Soil Sci. Soc. Am. Proc., 39: 1209-1213. [2] Stahl W. J. (1977) Carbon and nitrogen isotopes in hydrocarbon research and exploration. Chem. Geol., 20: 121-149. [26] Stanley K., Lyon G. L. & Stewart M. K. (1984) Hot shot reduction of water to hydrogen for isotopic analysis. Institute of Nuclear Sciences, DSIR, Lower Hutt, New Zealand. Rept. INS-R--322, February, 1984: llp. [1] Stapp P., Polis G. A. & Pinero S. (1999) Stable isotopes reveal strong marine and E1 Nifio effects on island food webs. Nature 401, 467-469 [9]. Stark J. M. & Hart S. C. (1996) Diffusion technique for preparing salt solutions, Kjeldahl digests, and persulfate digests for nitrogen-15 analysis. Soil Sci. Soc. Am. J., 60: 1846-1855. [15] Staudt W. J. & Schoonen M. A. A. (1994) Sulfate incorporation into sedimentary carbonates. In: Geochemical transformations of sedimentary sulfur. M. A. Vairavamurthy, M. A. A. Schoonen, T. I. Eglinton, G. W. Luther, and B. Manowitz (eds.). American Chemical Society, Washington. ACS Symposium Series 612: 332-345. [26] Steele L. P., Dlugokencky E. J., Lang P. M., Tans P. P., Martin R. C. & Masarie K. A. (1992) Slowing down of the global accumulation of atmospheric methane during the 1980's. Nature, 358(6384): 313316. [14] Stehr J., Krankowsky D. & Mauersberger K. (1996) Collection and analysis of atmospheric ozone samples. J. Atmos. Chem., 24: 317-325. [18] Steiner R. E., Lewis C. L. & King F. L. (1997) Time-of-flight mass spectrometry with a pulsed glow discharge ionization source. Anal. Chem., 69: 1715-1721. [35] Steinfeld J. I., Adler-Golden S. M. & Gallagher J. W. (1987) Chem. Ref. Data, 16(4): 911. [18] Stephan E. (2000) Oxygen isotope analysis of animal bone phosphate: method refinement, influence of consolidants, and reconstruction of paleotemperatures for Holocene sites. J. Archaeol. Sci., 27: 523535. [22] Stern L. A., Chamberlain C. P., Reynolds R. C. & Johnson G. D. (1997) Oxygen isotope evidence of climate change from pedogenic clay minerals in the Himalayan molasse. Geochim. Cosmochim. Acta, 61: 731-744. [2] Sternberg L. d. S. L., DeNiro M. J. & Savidge R. A. (1986) Oxygen isotope exchange between metabolites and water during biochemical reactions leading to cellulose synthesis. Plant Physiol., 82: 423427. [23] Stevens C. M. & Rust F. E. (1982) The carbon isotopic composition of atmospheric methane. J. Geophys. Res., 87(C7): 4879-4882. [14] Stevens C. M. (1995) Carbon-13 isotopic abundance and concentration of atmospheric methane for background air in the southern and northern hemispheres from 1978 to 1989. Environmental Science Division, U. S. Dept. of Energy, Publ. 4388. [14] Stevens W. B., Mulvaney R. L., Khan S. A. & Hoeft R. G. (2000) Improved diffusions methods for nitrogen and 15 nitrogen analysis of kjeldahl digests. J. Assoc. Off. Anal. Chem. Int., 38: 1039-1046. [15] Stevens W. H. (1971) Catalyst for hydrogen-deuterium exchange. Ger. Often. Rept. 2104019: 22p. [1] Stevenson E J. (1982) Nitrogen - Organic forms. In: Methods of soil analysis, Part 2. Chemical and microbiological properties. A. L. Page, R. H. Miller & D. R. Keeney (eds.). Am. Assoc. of Agronomy Inc- Soil Sci. Soc. of Amer. Inc., Madison, Wisconsin. Agronomy Series, No.9, 2nd edition: 625-641. [15] Stevenson F. J. (1996) Nitrogen - Organic forms. In: Methods of soil analysis. D. L. Sparks (ed.). Soil Sci. Soc. of Amer. Inc., Madison, Wisconsin. Ser. 5, Part 3: 1185-1200. [15] Stewart M. K. & Friedman I. (1975) Deuterium fractionation between aqueous salt solutions and water vapor. J. Geophys. Res., 80: 3812-3818. [46]
1160
References
Stewart M. K. & James M. R. (1981) Conversion of water to hydrogen for deuterium assay using uranium at 700~ Inst. Nucl. Sci., DSIR, Lower Hutt, New Zealand. Rept. INS-R-292. [1, 33] Stewart M. K. (1981) 180 and D enrichment by evaporation from sample containers. Int. J. Appl. Radiat. Isot., 32: 159-163. [1] Stichler W. (1995) Interlaboratory comparison of new materials for carbon and oxygen isotope measurements. In: Refrence and intercomparison materials for stable isotopes of light elements. IAEA-TECDOC-825, IAEA, Vienna: 67-74. [43] Stickrod R. D. & Marshall J. D. (2000) On-line nitrate-615N extracted from groundwater determined by continuous-flow elemental analyzer/isotope ratio mass spectrometry. Rapid Commun. Mass Spectrom., 14: 1266-1268. [15] St-Jean G. (2003) Automated quantitative and isotopic (C-13) analysis of dissolved inorganic carbon and dissolved organic carbon in continuous-flow using a total organic carbon analyzer. Rapid Commun. Mass. Spectrom., 17: 419-428. [10] Stoffregen R. E., Rye R. O. & Wasserman M. D. (1994a) Experimental studies of alunite: I. 180-160 and D-H fractionation factors between alunite and water at 250-450~ Geochim. Cosmochim. Acta, 58(2): 903-916. [46] Stoffregen R. E., Rye R. O. & Wasserman M. D. (1994b) Experimental studies of alunite: II. Rates of alunite-water alkali and isotopic exchange. Geochim. Cosmochim. Acta, 58: 917-929. [46] Stoffregen R. (1996) Numerical simulation of mineral-water isotope exchange via Ostwald ripening. Am. J. Sci., 296: 908-931. [46] Stoffyn-Egli P. & Mackenzie F. T. (1984) Mass balance of dissolved lithium in the oceans. Geochim. Cosmochim. Acta, 48: 859-872. [6] Stolper E. (1982) The speciation of water in silicate melts. Geochim. Cosmochim. Acta, 46: 2609-2620. [46] Stolper E. & Epstein S. (1991) An experimental study of oxygen isotope partitioning between silica glass and CO2 vapor. In: Stable isotope geochemistry: A tribute to Samuel Epstein. H. P. Jr. Taylor, J. R. O'Neil & I. R. Kaplan (eds.). The Geochemical Society, Special Publication No. 3: 35-51. [46] Stone J., Hutcheon I. D., Epstein S. & Wasserburg G. J. (1991) Silicon, carbon and nitrogen isotope studies of silicon carbide in carbonaceous ad enstatite chondrites. In: Stable isotope geochemistry: A tribute to Samuel Epstein. H. P. Jr. Taylor, J. R. O'Neil & I. R. Kaplan (eds.). The Geochemical Society, Special Publication No. 3: 487-504. [25] Stott A. W., Davies E., Evershed R. P. & Tuross N. (1997a) Monitoring the routing of dietary and biosynthesised lipids through compound-specific stable isotope (613C) measurements at natural abundance. Naturwissenschaften, 84: 82-86. [9] Stott A. W., Evershed R. P. & Tuross N. (1997b) Compound-specific approach to the ~)13C analysis of cholesterol in fossil bones. Org. Geochim., 26: 99-103. [9] Strauss H. (1997) The isotopic composition of sedimentary sulfur through time. Palaeogeogr., Palaeo-clim., Palaeoecol., 132: 97-118. [27] Strebel O., BSttcher J. & Fritz P. (1990) Use of isotope fractionation of sulfate-sulfur and sulfate-oxygen to assess bacterial desulfurication in a sandy aquifer. J. Hydrol., 121(1-4): 155-172. [29] Strelow, F. W. E. (1973) Application of ion exchange to element separation and analysis. In: Ion exchange and solvent extraction. J. A. Marinsky & Y. Marcus (eds.). Marcel Dekker. [29] Strelow F. W. E., Weinert C. H. S. W. & van der Walt T. N. (1974) Separation of lithium from sodium, beryllium and other elements by cation-exchange chromatography in nitric acid-methanol. Anal. Chim. Acta, 71: 123-132. [6] Strickland J. D. H. & Parsons T. R. (1972) A practical handbook of seawater analysis. Bull. Fish. Res. Bd. Can., 167: 1-130. [9] Stuart-Williams H. L. Q. & Schwarcz H. P. (1995) Oxygen isotopic analysis of silver orthophosphate using a reaction with bromine. Geochim. Cosmochim. Acta, 58: 3837-3841. [12, 22] Stuart-Williams H. L. Q. & Schwarcz H. P. (1997) Oxygen isotopic determination of climatic variation using phosphate from beaver bone, tooth enamel, and dentine. Geochim. Cosmochim. Acta, 61: 25392550. [22] Studley S. A., Ripley E. M., Elswick E. R., Dorais M. J., Fong J., Finkelstein D. & Pratt L. M. (2002) Analysis of sulfides in whole rock matrices by elemental analyzer - continuous flow isotope ratio mass spectrometry. Chem. Geol., 192(1-2): 141-148. [26, 45]
References
1161
Stuiver M. & Braziunas T. F. (1987) Tree cellulose 13C/12C isotope ratios and climatic change. Nature, 328: 58-60. [9] Stuiver M., Grootes P. M. & Braziunas T. F. (1995) The GISP2 6180 climate record of the past 16,500 years and the role of the sun, ocean, and volcanoes. Quater. Research, 44: 341-354. [1] Stull D. R. (1947) Vapour pressure of pure substances. Inorganic compounds. Ind. Eng. Chem., 39: 540-
550. [28]
Sturchio N. C. & C h a n L. H. (2003) Lithium isotope geochemistry of the Yellowstone hydrothermal system. Soc. Econ. Geol., Spec. Publ., 10: 171-180. [6] Stiirup S., Hansen M. & Molgaard C. (1997) Measurement of 44Ca:43Ca and 42Ca:43Ca isotopic ratios in urine using high resolution inductively coupled plasma mass spectrometry. J. Anal. Atom. Spectrom., 12: 919-923. [31] Sudzuki N. (1987) A water conversion method for D / H ratio analyes and its accuracy. Geochem. J., 21: 29-33. [1] Sugiura N., Hashizume K., Miura Y., Kiyota K. & Zashu S. (1995) Cosmogenic nitrogen isotopic disequilibrium observed in meteorites. Meteoritics, 30: 646-653. [17] Summons R. E., Jahnke L. L. & Roksandic Z. (1994) Carbon isotopic fractionation in lipids from methanotrophic bacteria: relevance for interpretation of the geochemical record of biomarkers. Geochim. Cosmochim. Acta, 58: 2853-2863. [46] Summons R. E., Franzmann P. D. & Nichols P. D. (1998) Carbon isotopic fractionation associated with methylotrophic methanogenesis. Org. Geochem., 28: 465-475. [9] Supplee J. M., Whittaker E. A. & Lenth W. (1994) Theoretical description of frequency modulation and wavelength modulation spectroscopy. Appl. Opt., 33: 6294-6302. [34] Suryanarayana M. V., Sankari M. & Gangadhara S. (1998) Determination of Li-6/Li-7 isotope ratio using two photon resonance three photon resonance ionization mass spectrometry. Int. J. Mass Spectrom., 173: 177-189. [33] Sutka R. L., Ostrom N. E., Ostrom P. H., Gandhi H. & Breznak J. A. (2003) Nitrogen isotopomer site preference of N20 produced by Nitrosomonas europaea and Methylococcus capsulatus Bath. Rapid Commun. Mass Spectrom., 17: 738-745. [19] Suvorova V. A. & Dubinina Y. O. (1996) Determinating oxygen isotope compositions on small amounts of water. Geochem. Int., 33: 50-54. [1] Suzuoki T. & Epstein S. (1976) Hydrogen isotope fractionation between OH-bearing minerals and water. Geochim. Cosmochim. Acta, 40: 1229-1240. [46] Svec H. J. & Anderson A. R. Jr. (1965) The absolute abundance of the lithium isotopes in natural sources. Geochim. Cosmochim. Acta, 29: 633-641. [6] Swart P. K. (1983) Carbon and oxygen isotope fractionation in scleractinian corals: A review. Earth Sci. Rev., 19: 51-80. [3] Swart P. K., Burns S. J. & Leder J. J. (1991.) Fractionation of the stable isotopes of oxygen and carbon in carbon dioxide during the reaction of calcite with phosphoric acid as a function of temperature and technique. Chem. Geol., 86: 89-96. [12] Swart P. K., Leder J. J., Szmant A. M. & Dodge R. E. (1996) The origin of variations in the isotopic record of scleractinian corals: II. Carbon. Geochim. Cosmochim. Acta, 60: 2871-2885. [46] Swihart G. H. (1996) Instrumental techniques for boron isotope analysis. In: Boron. Mineralogy, petrology and geochemistry. E. S. Grew & L. M. Anovitz (eds.). Mineralogical Society of America, Reviews in Mineralogy, Vol. 33: 845-862. [7] Switsur R., Waterhouse J. S., Field E. M., Carter T. & Loader N. (1995) Stable isotope studies in tree rings from o a k - technique and some preliminary results. In: Problems of stable isotopes in treerings, lake sediments and peat-bogs as climatic evidence for the Holocene. B. Frenzel (ed.). Gustav Fischer Verlag, Stuttgart, Vol. 15: 129-140. [24] Syers J. K., Chapman S. L., Jackson M. L., Rex R. W. & Clayton R. N. (1968) Quartz isolation from rocks, sediments and soils for determination of oxygen isotopes composition. Geochim. Cosmochim. Acta, 32: 1022-1025. [2] Szaran J. (1996) Experimental investigation of sulphur isotopic fractionation between dissolved and gaseous H2S. Chem. Geol., 127: 223-228. [26]
1162
References
Szostak R. (1989) Molecular sieves: Principles of synthesis and identification. Van Nostrand Reinhold, New York: 524 p. [36] Tabatabai M. A. & Bremner J. M. (1970) An alkaline oxidation method for determination of total sulfur in soils. Soil Sci. Soc. Am. Proc., 34: 62-65. [26] Tabatabai M. A. (1982) Sulfur. In: Methods of soil analysis, part 2. Chemical and microbiological properties. A. L. Page, R. H. Miller & D. R. Keeney (eds.). American Association of Agronomy Inc. - Soil Science Society of America Inc., Agronomy Series, No. 9: 501-538. [26] Tabatabai M. A. (1992) Methods of measuring of sulphur in soils, plant materials, and waters. In: Sulphur cycling on the continents: Wetlands, terrestrial ecosystems and associated water bodies. R. W. Howarth, J. W. B. Stewart & M. V. Ivanov (eds.). John Wiley & Sons, New York: 307-344. [26] Taieb R. (1990) Les isotopes de l'hydrogene, du carbone et de l'oxygene dans les s6diments argileux et les eaux de formation. Th6se, Institut national Polytechnique de Lorraine, Nancy, France: 141 p. [2] Takahata N., Nishio Y., Yoshida N. & Sano Y. (1998) Precise isotopic measurements of nitrogen at the sub-nanomole level. Anal. Sci., 14: 485-491. [17] Takala K., Olkkonen H. & Krouse H. R. (1991) Sulphur isotope composition of epiphytic and terricolous lichens and pine bark in Finland. Environ. Pollut., 69: 337-348. [26] Tan T. C., Pearson G. J. & Walker R. W. (1973) Sampling extraction and 13C/12C analysis of total dissolved CO2 in marine environments. Bedford Institute of Oceanography, Canada. Report Series, BI-R-73-16:17 p. [10] Tanaka N. & Rye D. M. (1991) Chlorine in the stratosphere. Nature, 353: 707. [28] Tanaka N., Rye D. M., Xiao u & Lasaga A. C. (1994) Use of stable sulfur isotope systematics for evaluating oxidation reaction pathways and in-cloud-scavenging of sulfur dioxide in the atmosphere. Geophys. Res. Lett., 21(14): 1519-1522. [26] Tanaka N., Rye D. M., Rye R., Avak H. & Yoshinari T. (1995) High precision mass spectrometric analysis of isotopic abundance ratios in nitrous oxide by direct injection of N20. Int. J. Mass Spectrom. Ion Proc., 142: 163-175. [19] Tang D., Yang Q., Zhou C., Kang X., Liu D. & Huang W. (2001) Genetic relationships between swamp microenvironment and sulfur distribution of the late Paleozoic coals in North China. Science in China (Series D), 44: 555-565. [26] Tang K., Feng X. & Ettl G. J. (2000) The variation in dD of tree rings and the implications for climatic reconstruction. Geochim. Cosmochim. Acta, 64: 1663-1673. [46] Tanner S. D. (1992) Space charge in ICP-MS: Calculation and implications. Spectrochim. Acta, 47B: 809823. [6] Tans P. P., de Jong A. F. M. & Mook W. G. (1978) Chemical pretreatment and radial flow of 14C in tree rings. Nature, 271: 234-235. [24] Tans P. P., Fung I. Y. & Takahashi T. (1990) Observational constraints on the global atmospheric carbon dioxide budget. Science, 247: 1431-1438. [14] Tans P. P., Berry J. A. & Keeling R. F. (1993) Oceanic 13C/12Cobservations: A new windowq on ocean CO2 uptake. Global Biogeochem. Cycles, 7: 353-368. [14] Tanweer A., Hut G. & Burgman J. O. (1988) Optimal conditions for the reduction of water to hydrogen by zinc for mass spectrometric analysis of deuterium content. Chem. Geol. (Isot. Geosc. Sect.), 73(2): 199-203. [1] Tanweer A. (1990) Importance of clean metallic zinc for hydrogen isotope analysis. Anal. Chem., 62: 2158-2160.[1] Tanweer A. (1993a) Investigation of BDH zinc for hydrogen isotope analysis by mass spectrometry. Z. Naturforsch., 48a: 739-740. [1] Tanweer A. (1993b) Determination of deuterium in brines and in hypersaline aqueous solutions by mass spectrometry using zinc as reducing agent. Analyst, 118: 835-838. [1] Tanweer A. & H a n L.-F. (1996) Reduction of microliter amounts of water with manganese for D / H isotope ratio measurement by mass spectrometry. Isot. Environ. Health Stud., 31: 1-7. [1] Tanzer D. & Heumann K. G. (1991) Determination of dissolved selenium species in environmental water samples using isotope dilution mass spectrometry. Anal. Chem., 63: 1984-1989. [29] Taube H. (1954) Use of oxygen isotopes effects in the study of hydration of ions. J. Phys. Chem., 58: 523528. [46]
References
1163
Taylor B. E. & O'Neil J. R. (1977) Stable isotope studies of metasomatic Ca-Fe-A1-Si skarns and associated metamorphic and igneous rocks, Osgood Mountains, Nevada. Contrib. Mineral. Petrol., 63: 1-49. [20] Taylor B. E., Foord E. E. & Friedrichsen H. (1979) Stable isotope and fluid inclusion studies of gembearing granitic pegmatite-aplite dikes, San Diego Co., California. Contrib. Mineral. Petrol., 68: 187205. [3] Taylor B. E., Eichelberger J. C. & Westrich H. R. (1983) Hydrogen isotopic evidence of rhyolitic magma degassing during shallow intrusion and eruption. Nature, 306: 541-545. [46] Taylor B. E. & Beaudoin G. (1993) MILES laser microprobe. Part I: System description. Geol. Surv. Canada. Current Research, Part D, Paper 1993-ID: 191-198. [20, 26] Taylor B. E. & Beaudoin G. (2000) Sulphur stratigraphy of the Sullivan Pb-Zn-Ag deposit, B.C: Evidence for hydrothermal sulphur, and bacterial and thermochemical sulphate reduction. In: The Sullivan deposit and its geological environment. J. W. Lydon, T. H6y, J. F. Slack & M. Knapp (eds.). Mineral Deposits Division of the Geological Association of Canada. Spec. Publ. No. 1: 696-719. [20] Taylor B. E., Ding T., Halas S., Br6as O. & Robinson B. W. (2001) Accurate calibration of the V-CDT scale: Proposed ~134Svalues for calibration reference materials and methods of correction for SO2based analyses. AGU Mtgn: $426 (Abstract). [20] Taylor B. E. (2004) Biogenic and thermogenic sulfate reduction in the Sullivan Pb-Zn-Ag deposit, B.C.: Evidence from micro-isotopic analysis of carbonate and sulfide in bedded ores. Chem. Geol.: in press [20] Taylor B. E. (in press) Laser-assisted micro-analysis of the sulfur isotope reference materials using the SF6-method: Calibration of the sulfur isotope reference scale and consistent discrepancies with SO2based measurements. IAEA, Vienna. TECDOC. [20, 26] Taylor B. E., Ding T., Halas S., Br6as O. & Robinson B. W. (in press) Accurate calibration of the V-CDT scale: proposed 634S values for calibration and reference materials and methods of correction for SO2 based analyses. IAEA, Vienna. [20, 26] Taylor C. B. (1973) Measurement of oxygen-18 ratios in environmental waters using the EpsteinMayeda technique. Part I: Theory and experimental details of the equilibration technique. Inst. Nucl. Sci., DSIR, Lower Hutt, New Zealand, Rept N. 556 (INS-LN-37), April, 1973:24 p. [1] Taylor C. B. & Fox V. J. (1996) An isotopic study of dissolved inorganic carbon in the catchment of the Waimakariri River and deep ground water of North Canterbury plains, New Zealand. J. Hydrol., 186: 161-190. [10] Taylor J. W. & Grimsrud E. P. (1969) Chlorine isotope ratios by negative ion mass spectrometry. Anal. Chem., 41: 805-810. [28] Taylor J. W. & Chen I.-J. (1970) Variables in oxygen-18 isotopic analysis by mass spectrometry. Anal. Chem., 42: 224-228. [23] Taylor H. E., Huff R. A. & Montaser A. (1998) Novel applications of ICPMS. In: Inductively coupled plasma mass spectrometry. A. Montaser (ed.). Wiley-VCH: 681-807. [31] Taylor H. P. Jr. & Epstein S. (1962) Relationship between O18/O16 ratios in coexisting minerals of igneous and metamorphic rocks. Part I: Principles and experimental results. Geol. Soc. Am. Bull., 73: 461-480. [2, 20, 25] Taylor H. P. Jr. (1968) The oxygen isotope geochemistry of igneous rocks. Contrib. Mineral. Petrol., 19: 171. [46] Taylor H. P. Jr. & Epstein S. (1970) Oxygen and silicon isotope ratios of lunar rocks. Earth Planet. Sci. Lett., 9: 298-310. [20] Taylor H. P. Jr. & Epstein S. (1973) 180/160 and 30Si/28Si studies of some Apollo 15, 16 and 17 samples. 4th Lunar Science Conference, Proceedings, Vol. 2: 1657-1679. [25] Taylor P. D. P., De Bi6vre P., Walder A. J. & Entwistle A. (1995) Validation of the analytical linearity and mass discrimination correction model exhibited by a multiple collector inductively coupled plasma mass spectrometer by means of a set of synthetic uranium isotope mixtures. J. Anal. At. Spectrom., 10: 395-398. [31, 32, 35] Taylor P. D. P. (1998) IRMM Certificate: Sulfur primary isotopic gas standard IRMM-PIGS-2010. Institute for Reference Materials and Measurements, Geel, Belgium: 2 p. [40]
1164
References
Taylor T. I. 6: Urey H. C. (1938) Fractionation of the lithium and potassium isotopes by chemical exchange with zeolites. J. Chem. Phys., 6: 429-438. [6] Tebo B. M. & Obraztova A. Y. (1998) Sulfate reducing bacterium grows with Cr(VI), U(VI), Mn(IV), and Fe(III) as electron acceptors. FEMS Microbiol. Lett., 162: 193-198. [29] Teece M. A., Fogel M. L., Dollhopf M. E. & Nealson K. H. (1999) Isotopic fractionation associated with biosynthesis of fatty acids by marine bacterium under oxic and anoxic conditions. Org. Geochim., 30: 1571-1579. [9] Tellier K. E., Hluchy M. M., Walker J. R. & Reynolds R. C. Jr. (1988) Application of high gradient magnetic separation (HGMS) to structural and compositional studies of clay mineral mixtures. J. Sedim. Petrol., 58: 761-763. [2] Teng J., Barshick C. M., Duckworth D. C., Morton S. J., Smith D. H. & King F. L. (1995) Factors influencing the quantitative determination of trace elements in soils by glow discharge mass spectrometry. Appl. Spectrosc., 49: 1361-1366. [35] Tennie A., Hoffbauer R. & Hoernes S. (1998) The oxygen isotope fractionation of kyanite in experiment and nature. Contrib. Mineral. Petrol., 133: 346-355. [20] Thielecke F. & Noack R. (1997) Evaluation of an automated equilibration technique for deuterium/hydrogen isotope ratio measurements with respect to assessing total energy expenditure by the doubly labelled water method. J. Mass Spectrom., 32: 323-327. [1] Thielecke F., Brand W. & Noak R. (1998) Hydrogen isotope determination for small-size water samples using ban equilibration technique. J. Mass Spectrom., 33: 342-345. [1] Thiemens M. H. & Heidenreich J. E. III (1983) The mass-independent fractionation of oxygen: A novel isotope effect and its possible cosmochemical implications. Science, 219: 1073-1075. [15, 18, 46] Thiemens M. H. & Jackson T. (1987) Production of isotopically heavy ozone by ultra-violet light photolysis of 02. Geophys. Res. Lett., 14: 624-627. [18] Thiemens M. H. & Jackson T. (1988) New experimental evidence for the mechanism for production of isotopically heavy 03. Geophys. Res. Lett., 15: 639-643. [18] Thiemens M. H. & Jackson T. L. (1990) Pressure dependency for heavy isotope enhancement in ozone formation. Geophys. Res. Lett., 17: 717-719. [18] Thiemens M. H. & Trogler W. C. (1991) Nylon production: An unknown source of atmospheric nitrous oxide. Science, 251: 932-934. [19] Thiemens M. H. (1992) Mass-independent isotopic fractionation and their applications. In: Isotope effects in gas-phase chemistry. J. A. Kaye (ed.). American Chemical Society, Washington DC: 138-154.
[18]
Thiemens M. H. (1999) Mass-independent isotope effects in planetary atmospheres and the early solar system. Science, 283: 341-345. [18, 46] Thiemens M. H. (2001) The mass-independent ozone isotope effect. Science, 283: 341-345. [46] Thiemens M. H., Savarino J., Farquhar J. & Bao H. (2001) Mass-independent isotopic compositions in terrestrial and extraterrestrial solids and their applications. Acc. Chem. Res., 34: 645-652. [15, 18] Thirlwall M. (2001) Inappropriate tail corrections can cause large inaccuracy in isotope ratio determination by MC-ICP-MS. J. Anal. At. Spectrom., 16: 1121-1125. [32] Thode H. G., MacNamara J., Lossing F. P. & Collis C. B. (1948) Natural variations in isotopic content of boron and its chemical atomic weight. J. Am. Chem. Soc., 70: 3008-3011. [7] Thode H. G., MacNamara J. & Collins C. B. (1949) Natural variations in the isotopic content of sulfur and their significance. Can. J. Res., 27B: 361-373. [44, 46] Thode H. G., Kleerekoper H. & McElchern D. (1951) Isotopic fractionation in the bacterial reduction of sulphate. Research (London), 4: 581-582. [46] Thode H. G., Monster J. & Dunford H. B. (1958) Sulphur isotope abundances in petroleum and associated minerals. AAPG Bull., 42: 2619-2641. [26] Thode H. G., Monster J. & Dunford H. B. (1961) Sulphur isotope geochemistry. Geochim. Cosmochim. Acta, 25: 159-174. [26, 40, 44] Thode H. G. & Monster J. (1970) Sulfur isotope abundances and genetic relations of oil accumulations in Middle East basin. AAPG Bull., 54: 627-637. [26] Thode H. G., Cragg C. B., Hulston J. R. & Rees C. E. (1971) Sulphur isotope exchange between sulphur dioxide and hydrogen sulphide. Geochim. Cosmochim. Acta, 35: 35-45. [46]
References
1165
Thode H. G. & Rees C. E. (1971) Measurement of sulphur concentrations and the isotope ratios 33S/32S, 34S / 32S and 36S / 32S in Apollo samples. Earth Planet. Sci. Lett., 12: 434-438. [20, 26, 45] Thode H. G. & Rees C. E. (1979) Sulphur isotopes in lunar and meteorite samples. Proc. Lunar Planet. Sci. Conf. (10th): 1629-1636. [26] Thode H. G. (1981) Sulfur isotope ratios in petroleum research and exploration: Williston Basin. AAPG Bull., 65: 1527-1535. [26] Thompson A. M. (1992) The oxidizing capacity of the earth's atmosphere: probable past and future changes. Science, 256: 1157-1165. [14] Thompson P. & Gray J. (1977) Determination of 180/160 ratios in compounds containing C, H and O. Int. J. Appl. Rad. Isot., 28: 411-415. [23] Thomson J. J. (1910) Rays of positive electricity. Philosoph. Mag., 20: 752-767. [38] Thomson J. J. (1913) Rays of positive electricity and their application to chemical analyses. Longman, Green and Co., London. [38] Thonnard N. & Lehmann B. E. (1994) The institute for rare isotope measurements: An interdisciplinary facility for RIS-based applications. H.-J. Kluge, J.E. Parks & K. Wendt (eds.). Resonance Ionization Spectroscopy, Seventh International Symposium. AIP Press Woodbury, New York, AIP Conference Proceedings 329: 335-338. [33] Thornton V. & Condon F. E. (1950) Infrared spectrometric determination of deuterium oxide in water. Anal. Chem., 22: 680-681. [34] Thurman E. M. & Malcolm R. L. (1981) Preparative isolation of aquatic humic substances. Env. Sci. Technol., 15: 463-466. [15] Thurston W. M. (1970) Automatic mass spectrometric analysis of the deuterium to hydrogen ratio in natural water. Rev. Sci. Instrum., 41(7): 963-966. [1] Thurston W. M. (1971) Steam film sampling of water for mass spectrometric analysis of the deuterium content. Rev. Sci. Instrum., 42: 700-703. [1] Thurston W. M. & James M. W. D. (1984) Absolute determination of deuterium content of heavy water standards by distillation and mass spectrometry. Anal. Chem., 56: 386-389. [1] Tieszen L. L., Boutton T. W., Tesdahl K. G. & Slade N. A. (1983) Fractionation and turnover of stable carbon isotopes in animal tissues: Implications for 613C analysis of diet. Oecologia, 57: 32-37. [9] Tilles D. (1961a) Variations of silicon isotope ratios in a zoned pegmatite. J. Geophys. Res., 66: 3015-3020. [20] Tilles D. (1961b) Natural variations in isotopic abundances of silicon. J. Geophys. Res., 66: 3003-3014. [20] Tittel E K., Richter D. & Fried A. (2003) Mid-infrared laser applications in spectroscopy. In: Solid-state mid-infrared laser sources. I. T. Sorokina & K. L. Vodopyanov (eds.). Springer Verlag, Berlin. Topics Appl. Phys., 89: 445-516. [34] Tobias H. J., Goodman K. J., Blacken C. E. & Brenna J. T. (1995) High precision D / H measurement from hydrogen gas and water by continuous flow isotope ratio mass spectrometry. Anal. Chem., 67(14): 2486-2492. [1] Tobias H. J. & Brenna J. T. (1996) High-precision D / H measurement from organic mixtures by gas chromatography continuous flow isotope ratio mass spectrometry using a palladium filter. Anal. Chem., 68: 3002-3007. [8] Tobias H. J. & Brenna J. T. (1997) On-line pyrolysis as a limitless reduction source for high-precision isotopic analysis of organic-derived hydrogen. Anal. Chem., 69: 3148-3152. [8] Todd B. J. (1955) Outgassing of glass. J. Appl. Phys., 26: 1238-1243. [1] Todt W., Cliff R. A., Hanser A. & Hofmann A. W. (1996) Validation of the analytical linearity and mass discrimination correction model exhibited by a multiple collector inductively coupled plasma mass spectrometer by means of a set of synthetic uranium isotope mixtures. J. Anal. At. Spectrom., 10: 395398. [31] Tokunaga T. K., Zawislanski P. T., Johannis P. W., Lipton D. S. & Benson S. (1994) Field investigations of selenium speciation, transformation, and transport in soils from Kestrson Reservoir and Lahontan Valley. In: Selenium in the environment. W. T. Jr. Frankenberger & S. Benson (eds.). Marcel Dekker: 119-138. [29] Tomascak P. B., Krogstad E. J., Prestegaard K. L. & Walker R. J. (1995) Li isotope variations in groundwaters from eastern Maryland. EOS Trans. AGU, 76:$112 (Abstract). [6]
1166
References
Tomascak P. B., Carlson R. W. & Shiney S. B. (1999a) Accurate and precise determinations of Li isotopic composition by multi-collector sector ICP-MS. Chem. Geol., 158: 145-154. [6, 31] Tomascak P. B., Hauri E. H. & Lassiter J. C. (1999b) Lithium isotope constraints on Hawaiian plume components. EOS Trans. AGU, 80: $354. [31] Tomascak P. B. & Langmuir C. H. (1999) Lithium isotope variability in MORB. EOS Trans. AGU, 80: F1086-1087. [6, 31] Tomascak P. B., Tera F., Helz R. T. & Walker R. J. (1999c) The absence of lithium isotope fractionation during basalt differentiation: New measurements by multicollector sector ICP-MS. Geochim. Cosmochim. Acta, 63(6): 907-910. [6, 31] Tomascak P. B., Ryan J. G. & Defant M. J. (2000) Lithium isotope evidence for light element decoupling in the Panama subarc mantle. Geology, 28: 507-510. [6] Tomascak P. B., Widom E., Benton L. D., Goldstein S. L. & Ryan J. G. (2002) The control of lithium budgets in island arcs. Earth Planet. Sci. Lett., 196: 227-238. [6] Tomascak P. B., H e m m i n g N. G. & H u m m i n g S. R. (2003) The lithium isotope composition of waters of the Mono Basin, California. Geochim. Cosmochim. Acta, 67: 601-611. [6] Tonarini S., Pennisi M. & Leeman W. P. (1997) Precise boron isotopic analysis of complex silicate (rock) samples using alkali carbonate fusion and ion-exchange separation. Chem. Geol., 142: 129-137. [7, 40] Tonarini S., Pennisi M., Adorni-Braccesi A., Dini A., Ferrara G., Gonfiantini R., Wiedenbeck M. & Grinning M. (2003) Intercomparison of boron isotope and concentration measurements. Part I: Selecion, preparation and homogeneity tests of the intercomparison materials. Geostand. Newslett., 27(1): 21-39. [40] Tong S. L. & Harrison W. W. (1993) Glow discharge mass spectrometric analysis of non-conducting materials. Spectrochim. Acta, B 48: 1237-1245. [35] Toth R. A. (1994a) Extensive measurements of H2160 line frequencies and strengths: 5750 to 7965 cm-1. Appl. Opt., 33: 4851-4867. [34] Toth R. A. (1994b) Transition frequencies and strengths of H2170 and H2180:6600 to 7640 cm -1. Appl. Opt., 33: 4868-4879. [34] Toulemonde B., Horman I., Egli H. & Derbesy M. (1983) Food-related applications of high-resolution NMR. Part II. Differentiation between natural and synthetic vanillin samples using 2H-NMR. Helv. Chim. Acta, 66: 2342-2345. [5] Toyoda S. & Yoshida N. (1999) Determination of nitrogen isotopomers of nitrous oxide on a modified isotope ratio mass spectrometer. Anal. Chem., 71: 4711-4718. [19] Toyoda S., Yoshida N., Urabe T., Aoki S., Nakazawa T., Sugawara S. & Honda H. (2001) Fractionation of N20 isotopomers in the stratosphere. J. Geophys. Res., 106: 7515-7522. [19] Toyoda S., Yoshida N., Miwa T., Mastsui Y., Yamagishi H., Tsunogai U., Nojiri Y. & Tsurushima N. (2002) Production mechanism and global budget of N20 inferred from its isotopomers in the western North Pacific. Geophys. Res. Lett., 29(3): 7-1 - 7-4. [19] Trautmann N. (1994) Ultratrace analysis of long-lived radioisotopes in the environment. H.-J. Kluge, J.E. Parks & K. Wendt (eds.). Resonance Ionization Spectroscopy, Seventh International Symposium. AIP Press Woodbury, New York, AIP Conference Proceedings 329: 243-250. [33] Truesdell A. H. (1974) Oxygen isotope activities and concentrations in aqueous salt solutions at elevated temperatures: consequences for isotope geochemistry. Earth Planet. Sci. Lett., 23: 387-396. [46] Trust B. A. & Fry B. (1992) Stable sulphur isotopes in plants: a review. Plant Cell Environ., 15: 1105-1110. [26] Trust Hammer B. A., Fogel M. L. & Hoering T. C. (1998) Stable carbon isotope ratios of fatty acids in seagrass and redhead ducks. Chem. Geol., 152: 29-41. [9] Tse R. S., Wong S. C. & Yuen C. P. (1980) Determination of deuterium/hydrogen ratios in natural waters by Fourier transform nuclear magnetic resonance spectrometry. Anal. Chem., 52: 2445-2448. [40] Tucker M. E. (1988) Techniques in sedimentology. Blackwell Scientific Publications, Oxford. [2] Tudge A. P. (1960) A method of analysis of oxygen isotopes in orthophosphate - its use in the measurement of paleotemperatures. Geochim. Cosmochim. Acta, 18: 81-93. [2, 20, 22]
References
1167
Turatti F., Griffith D. W. T., Wilson S. R., Esler M. B., Rahn T. A., Zhang H. & Blake G. A. (2000) Positionally dependent 15N fractionation factors in the UV photolysis of N20 determined by high resolution FTIR spectroscopy. Geophys. Res. Lett., 27: 2489-2492. [34] Turekian V. C., Macko S. A. & Keene W. C. (2001) Application of stable sulfur isotopes to differentiate sources of size-resolved particulate sulfate in polluted marine air at Bermuda during spring. Geophys. Res. Lett., 28(8): 1491-1494. [26] Turner J. V. (1982) Kinetic fractionation of carbon-13 during calcium carbonate precipitation. Geochim. Cosmochim. Acta, 46: 1183-1191. [46] Turner J. V. & Gailitis V. (1988) Single-step method for hydrogen isotope ratio measurement of water in porous media. Anal. Chem., 60: 1244-1246. [1] Turner P. J., Mills D. J., SchrSder E., Lapitajas G., Jung G., Iacone L. A., Haydar D. A. & Montaser A. (1998) Instrumentation for low- and high-resolution ICPMS. In: Inductively coupled plasma mass spectrometry. A. Montaser (ed.). Wiley-VCH: 421-501. [31] Tuttle M. L., Goldhaber M. B. & Williamson D. L. (1986) An analytical scheme for determining forms of sulfur in oil shales and associated rocks. Talanta, 33: 953-961. [26] Tuttle O. F. (1949) Two pressure vessels for silicate-water studies. Geol. Soc. Am. Bull., 60: 1727-1729. [46] Tyler S. C. (1986) Stable carbon isotope ratios in atmospheric methane and some of its sources. J. Geophys. Res., 91(D12): 13232-13238. [14] Tyler S. C., Crill P. M. & Brailsford G. W. (1994) 13c/12 C fractionation of methane during oxidation in a temperate forested soil. Geochim. Cosmochim. Acta, 58: 1625-1633. [46] Ueda A. & Sakai H. (1983) Simultaneous determination of the concentration and isotope ratio of sulfate and sulfide sulfur and carbonate carbon in geological samples. Geochem. J., 17: 185-196. [26] Ueda A. & Krouse H. R. (1986) Direct conversion of sulphide and sulphate minerals to SO2 for isotope analysis. Geochem. J., 20: 209-212. [26] Uehara K., Yamamoto K., Kikugawa T. & Yoshida N. (2001) Isotope analysis of environmental substances by a new laser-spectroscopic method utilizing different pathlengths. Sensors and Actuators, B 74: 173-178. [34] Uehara K., Yamamoto K., Kikugawa T. & Yoshida N. (2002) Site-selective nitrogen isotopic ratio measurement of nitrous oxide using 2-/2m diode lasers. Spectrochim. Acta, Part A, 59: 957-962. [19] Ulmer G. C. & Barnes H. L. (1987) Hydrothermal experimental techniques. John Wiley & Sons: 523 p. [46] UnsSld A. (1968) Physik der Sternatmosph/iren. Springer Verlag, Berlin: 90-91. [39] Unterzaucher J. (1940) Die mikroanalytische Bestimmung des Sauerstoffes. Chem. Berichte, 73B: 391-404. [23] Urey H. C. & Rittenberg D. (1933) Some thermodynamic properties of the HIH2, H2H 2 molecules and compounds containing the H2 atom. J. Chem. Phys., 1(2): 137-143. [46] Urey H. C. (1947) The thermodynamic properties of isotopic substances. J. Chem. Soc., 69: 562-581. [15, 18, 20, 28] Urey H. C., Lowenstam H., Epstein S. & McKinney C. R. (1951) Measurement of paleotemperatures and temperatures of the Upper Cretaceous of England, Denmark, and the southeastern United States. Geol. Soc. Am. Bull., 62: 399-416. [40] Usdowski E. & Hoefs J. (1993) Oxygen isotope exchange between carbonic acid, bicarbonat, and water: A re-examination of the data of McCrea (1950) and an expression for the overall partitioning of oxygen isotopes between the carbonate species and water. Geochim. Cosmochim. Acta, 57: 3815-3818. [46] Usdowski E., Michaelis J., BSttcher M. E. & Hoefs. J. (1993) Factors for the oxygen isotope equilibrium fractionation between aqueous and gaseous CO2, carbonic acid, bicarbonate, carbonate, and water (19~ Z. Phys. Chem., 170: 237-249. [46] Uzaki M., Yamada K. & Ishiwatari R. (1993) Carbon-isotope evidence for oil-pollution in long-chain normal alkanes in Tokyo bay sediments. Geochem. J., 27: 385-389. [8] Valentini J. J. (1987) Mass-independent isotope fractionation in nonadiabatic molecular collisions. J. Chem. Phys., 86: 6757-6765. [18] Valentini J. J., Gerrity D. P., Phillips D. L., Nieh J. C. & Tabor K. D. (1987) CARS spectroscopy of O2(1Dg) from the Hartleyband photodissociation of 03: Dynamics of the dissociation. J. Chem. Phys., 86: 6745-6756. [18]
1168
References
Valkiers S., Aregbe Y., Taylor P. D. P. & De Bi6vre P. (1998) A primary xenon isotopic gas standard with traceable values for isotopic composition and molar mass. Int. J. Mass Spectrom., 173: 55-63. [42, 43] Valkiers S., Kipphardt H., Ding T., Damen R., De Bi6vre P. & Taylor P. D. P. (1999) A primary isotopic gas standard for sulfur in the form of SF6, with Systeme International d'Unites traceable values for isotopic composition and molar mass. Int. J. Mass Spectrom., 193: 1-8. [43] Vallet C., Arendt M. & Martin G. J. (1988) Site specific isotope fractionation of hydrogen in the oxidation of ethanol into acetic acid. Application to vinegars. Biotechnol. techniq., 2(2): 83-88. [5] Vallet C., Arendt M., Mabon F., Naulet N. & Martin G. J. (1991) Combination of mass spectrometry and site-specific NMR isotope analyses in the characterization of amino acids. J. Sci. Food. Agric., 56: 167-185. [5] Valley J. M., Chiarenzelli J. R. & McLelland J. M. (1994) Oxygen isotope geochemistry of zircon. Earth Planet. Sci. Lett., 126: 187-206. [20] Valley J. W. & O'Neil J. R. (1981) 13C/12C exchange bewteen calcite and graphite: a possible thermometer in grenville marbles. Geochim. Cosmochim. Acta, 45: 411-419. [46] Valley J. W., Kitchen N., Kohn N. J., Niendorf C. R. & Spicuzza M. J. (1995) UWG-2, a garnet standard for oxygen isotope ratios: Strategies for high precision and accuracy with laser heating. Geochim. Cosmochim. Acta, 59(24): 5223-5231. [16, 20] Valley J. W. & Graham C. M. (1996) Ion microprobe analysis of oxygen isotope ratios in quartz from Skye granite: Healed microcracks, fluid flow, and hydrothermal exchange. Contrib. Mineral. Petrol., 124: 225-234. [20] Valley J. W., Graham C. M., Harte B., Eiler J. M. & Kinny P. D. (1998) Ion microprobe analysis of oxygen, carbon and hydrogen isotope ratios. In: Applicants of microanalytical techniques to understanding mineralizing processes. M. A. McKibben, W. C. III Shanks & W. I. Ridley (eds.). Society of Economic Geologists, Rev. in Econ. Geol., Vol. 7: 73-98. [20, 46] Valyi L. (1977) Atom and ion sources. Akademiai Kiado, Budapest. [38] Van der Kamp G., Van Stempvoort D. R. & Wassenaar L. I. (1996) The radial diffusion method 1. Using intact cores to determine isotopic composition, chemistry and effective porosity for groundwater in aquitards. Water Resource Res., 32: 1815-1822. [2] Van der Meer M. T. J., Schouten S. & Sinninghe Damste J. S. (1998) The effect of the reversed tricarboxylic acid cycle on the 13C contents of bacteriallipids. Org. Geochem., 28: 527-533. [9] Van Everdingen R. O., Shakur M. A. & Krouse H. R. (1982) Isotope geochemistry of dissolved, precipitated, airborne and fallout sulphur species associated with springs near Paige Mountain, Normal Range, N.W.T. Can. J. Earth Sci., 19: 1395-1407. [26] Van Hook W. A. (1972) Vapor pressure isotope effect in aqueous systems. III. The Vapor pressure of HOD (-60 to 200~ J. Phys. Chem., 76: 3040-3043. [4] Van Kreel B. K., Van der Vegt F., Meers M., Wagenmakers T., Westerterp K. & Coward A. (1996) Determination of total-body water by a simple and rapid mass-spectrometric method. J. Mass Spectrom., 31(1): 108-111. [1] van Stempvoort D. R., Reardon E. J. & Fritz P. (1990) Fractionation of sulfur and oxygen isotopes in sulfate by soil sorption. Geochim. Cosmochim. Acta, 54: 2817-2826. [26] van Trigt R., Kerstel E. R. T., Visser G. H. & Meijer H. A. J. (2001) Stable isotope ratio measurements on highly enriched water samples by means of laser spectrometry. Anal. Chem., 73: 2445-2452. [34] van Trigt R., Meijer H. A. J., Sveinbjornsdottir A. E., Johnsen S. J. & Kerstel E. R. T. (2002a) Measuring stable isotopes of hydrogen and oxygen in ice by means of laser spectrometry: the Bolling transition in the Dye-3 (South Greenland) ice core. Ann. Glaciology, 35: 125-130. [34] van Trigt R., Kerstel E. R. T., Meijer H. A. J., McLean M. & Visser G. H. (2002b) Validation of the DLW method in Japanese quail at various water fluxes using laser and IRMS. J. Am. Phys., 93: 2147-2154. [34] van Warmerdam E. M., Frape S. K., Aravena R., Drimmie R. J., Flatt H. & Cherry J. A. (1995) Stable chlorine and carbon isotope measurement of selected chlorinated organic solvents. Appl. Geochem., 10: 547-552. [28] Vanhaecke F., Moens L., Dams R. & Taylor P. (1996) Precise measurement of isotope ratios with a double-focusing magnetic sector ICP mass spectrometer. Anal. Chem., 68: 567-569. [31, 35]
References
1169
Vanhaecke F., Moens L., Dams R., Papadakis I. & Taylor P. (1997) Applicability of high-resolution ICP-
mass spectrometry for isotope ratio measurements. Anal. Chem., 69: 268-273. [31] Vanhaecke F., Moens L., Dams R., Allen L. & Georgitis S. (1999a) Evaluation of the isotope ratio per-
formance of an axial time-of-flight ICP mass spectrometer. Anal. Chem., 71: 3297-3303. [31] Vanhaecke F., Moens L. & Taylor P. (1999b) Use of ICP-MS for isotope ratio measurements. In: Induc-
tively coupled plasma spectrometry and its applications. S. J. Hill (ed.). Sheffield Academic Press/CRC Press: 145-207. [37] Vaughn B. H., White J. W. C., Delmotte M., Trolier M., Cattani O. & Stievenard M. (1998) An automated system for hydrogen isotope analysis of water. Chem. Geol., 152: 309-319. [1] Veblen D. R. & Ferry J. M. (1983) A TEM study of the biotite-chlorite reaction and comparison with petrologic observations. Am. Mineralogist, 68: 1160-1168. [16] Velinsky D. J., Pennock J. R., Sharp J. H., Cifuentes L. A. & Fogel M. L. (1989) Determination of the isotopic composition of ammonium-nitrogen at the natural abundance level from estuarine waters. Mar. Chem., 26: 351-361. [9, 15] Vengosh A., Chivas A. R. & McCulloch M. (1989) Direct determination of boron and chlorine isotopic compositions in geological materials by negative thermal-ionization mass spectrometry. Chem. Geol., 79: 333-343. [7, 28] Vengosh A., Kolodny Y., Starinsky A., Chivas A. R. & McCulloch M. T. (1991) Coprecipitation and isotopic fractionation of boron in modern biogenic carbonates. Geochim. Cosmochim. Acta, 55: 29012910. [7] Vengosh A., Chivas A. R., Starinsky A., Kolodny Y., Zhang B. & Zhang P. (1995) Chemical and boron isotope compositions of non-marine brines from the Qaidam Basin, Qinghai, China. Chem. Geol., 120: 135-154. [46] Venkitachalem T. V. & Rao A. S. (1991) Resonance-enhanced multiple-photon ionization spectroscopy of excited sulphur atoms. Appl. Phys. B - Photophys. Laser Chem., 52: 102-107. [33] Vennemann T. W. & Smith H. S. (1990) The rate and temperature of reaction of C1F3 with silicate minerals, and their relevance to oxygen isotope analysis. Chem. Geol. (Isot. Geosci. Sect.), 86: 83-88. [20] V e n n e m a n n T. W., Kesler S. E. & O'Neil J. R. (1992) Stable isotope compositions of quartz pebbles and their fluid inclusions as tracers of sediment provenance; implications for gold- and uranium-bearing quartz pebble conglomerates. Geology, 20: 837-840. [3] V e n n e m a n n T. W. & O'Neil J. R. (1993) A simple and inexpensive method of hydrogen isotope and water analyses of minerals and rocks based on zinc reagent. Chem. Geol. (Isot. Geosci. Sect.), 103: 227234. [1, 2] V e n n e m a n n T. W. & O'Neil J. R. (1996) Hydrogen isotope exchange reactions between hydrous minerals and molecular hydrogen: I. A new approach for the determination of hydrogen isotope fractionation at moderate temperatures. Geochim. Cosmochim. Acta, 60(13): 2437-2451. [46] V e n n e m a n n T. W. & Hegner E. (1998) Oxygen, strontium, and neodynium isotope composition of fossil shark teeth as a proxy for the palaeoceanography and palaeoclimatology of the Miocene northern Alpine Paratethys. Palaeogeogr., Pataeoclimatol., Palaeoecol., 142: 107-121. [22] Vennemann T. W., Fricke H. C., Blake R. E., O'Neil J. R. & Colman A. (2002) Oxygen isotope analysis of phosphates: a comparison of techniques for analysis of Ag3PO4. Chem. Geol., 185: 321-336. [20] Verchovsky A. B., Fisenko A. V., Semjonova L. F. & Pillinger C. T. (1997) Heterogeneous distribution of Xenon-HL within presolar diamonds. Meteoritics, 32: A131-A132. [13] Verkouteren R. M. & Dorko W. D. (1989) High-accuracy gas analysis via isotope dilution mass spectrometry: carbon dioxide in air. Anat. Chem., 61: 2416-2422. [43] Verkouteren R. M., Klouda G. A. & Currie L. A. (1995) The carbon dioxide isotopic measurement process: Progress at NIST on measurements, reduction algorithms and standards. In: Reference and intercomparison materials for stable isotopes of light elements. IAEA-TECDOC-825, IAEA, Vienna: 111-129. [1, 15] Verkouteren R. M., Lagergren E. S. & Velapoldi R. A. (1998) Report of Analysis. NIST, Gaithersburg, USA. Internal Report 837-6-98. [40] Verkouteren R. M. (1999) Preparation, characterization, and value assignment of carbon dioxide isotopic reference materials: RMs 8562, 8563, and 8564. Anal. Chem., 71: 4740-4746. [40, 43]
1170
References
Verkouteren R. M. & Lee J. N. (2001) Web-based interactive data processing: Application to stable isotope metrology. Fresenius J. Anal. Chem., 370: 803-810. [43] Verkouteren R. M. & Klinedinst D. B. (2003) Value assignment and uncertainty estimation of selected light stable isotope reference materials: RMs 8543-8545, RMs 8562-8564, and RM 8566. NIST, Special Publication 260-149. [43] Vet5 I., Het6nyi M., Dem6ny A. & Hertelendi E. (1994) Hydrogen index as reflecting intensity of sulphidic diagenesis in non-bioturbated, shaly sediments. Org. Geochem., 22: 299-310. [27] Vieth W. & Huneke J. C. (1991) Relative sensitivity factors in glow discharge mass spectrometry. Spectrochim. Acta, B 46: 137-153. [35] Vilagura R. A., Alexander R. B., Castro M. S., Tilden P. M., Pearl H. W., Stacey P. E. & Turner R. E. (2001) Nitrogen loading in coastal water bodies: an atmospheric perspective. American Geophysical Union, Washington DC. Coastal and Estuarine studies, Vol. 57. [15] VIM (1993) International vocabulary of basic and general terms in metrology. BIPM, IEC, IFCC, ISO, IUPAC, IUPAP, OIML, International Organization for Standardization. ISBN 92-67-01075-1. [42, 43] Vinogradov A. P. & Dontsova E. I. (1947) Izotopnyi Sostav Kisloroda Alyuminosilikatnykh Gornykh Porod (Isotopic composition of the alumino-silicate rocks). Dokl. Akad. Nauk SSSr, 58:83-84 (in Russian). [20] Vistousek P. M., Aber J. D., Howarth R. W., Likens G. E., Matson P. A., Schindler D. W., Schlesinger W. D. & Tilman D. G. (1977) Human alterations of the global nitrogen cycle: sources and consequencaes. Ecol. Appl., 7: 737-750. [15] Vitali F., Longstaffe F. J., Bird, M. I., Gage K. L. & Caldwell W. G. E. (2001) Hydrogen-isotope fractionation in aluminium hydroxides: Synthesis products versus natural samples from bauxites. Geochim. Cosmochim. Acta, 65: 1391-1398. [46] Vityk M., Krouse H. R. & Demihov Y. (1993) Preservation of 6180 values of fluid inclusion water in quartz over geological time in an epithermal environment: Beregovo Deposit, Transcarpathia, Ukraine. Earth Planet. Sci. Lett., 119(4): 561-568. [3] Voerkelius S. (1990) Isotopendiskriminierungen bei der Nitrifikation und Denitrifikation: Grundlagen und Anwendungen der Herkunfts-Zuordnungen von Nitrat und Distickstoffmonoxid. PhD Thesis, Technical University of Mfinich, Mfinich, Germany: 119 p. [15] Vogel A. I. (1951) A textbook of quantitative inorganic analysis, theory and practice. Longmans, Green and Co., London: 918 p. [28] Vogel J. C., Talma A. S. & Heaton T. H. E. (1981) Gaseous nitrogen as evidence for denitrification in groundwater. J. Hydrol., 50: 191-200. [15] Vogl J., Becker D., Birke W., Janisch N., Klingbeil P., Ludwig H. P., Riebe W., Pritzkow W. & Wermann G. (2002) Zertifizierungsbericht, Herstellung und Zertifizierung von 3 Bor-Isotopenreferenzmaterialien mit unterschiedlicher Bor-Isotopenzusammensetzung. Bundesanstalt fiir Materialforschung und -pr~ifung, Berlin, Germany. [40] Vogler E. A. & Hayes J. M. (1980) Carbon isotopic compositions of carboxyl groups of biosynthesized fatty acids. In: Advances in organic geochemistry, 1979. A. G. Douglas & J. R. Maxwell (eds.). Pergamon Press, Oxford, Vol. 12: 697-704. [9] Volkert R., Johnson C. M. & Tamashausky A. (2000) Mesoproterozic graphite depositits, New Jersey Highlands: geologic and stable isotopic evidence for possible algal origins. Can. J. Earth Sci., 37: 16651675. [45] yon Kallman H. & Lasareff W. (1932) l]-ber die Isotopenuntersuchungen (Sauerstoff, Neon und Chlor). Z. Physik, 80: 333-343. [28] Voncken J. H. L., Van Roermund H. L. M., van der Eerden A. M. J., Jansen J. B. H. & Erd R. C. (1993) Holotype buddingtonite; an ammonium feldspar without zeolitic H20. Am. Mineralogist, 78: 204209. [16] Viilkening J. & Papanastassiou D. A. (1989) Iron isotope anomalies. Astrophys. J., 347: L43-L46. [29, 31] VSlkering J., Koppe M. & Heumann K. G. (1991) Tungsten isotope ratio determinations by negative thermal ionization mass spectrometry. Int. J. Mass Spectrom. Ion Proc., 107: 361-368. [35] VSlkl J. & Alefeld G. (1975) Hydrogen diffusion in metals. In: Diffusion in Solids: Recent Developments. A. S. Nowick & J. J. Burton (eds.). Academic Press: 231-302. [46]
References
1171
Vreca P. (2003) Carbon cycling at the sediment-water interface in a eutrophic mountain lake (Jezero na
Planini pri Jezeru, Slovenia). Org. Geochem., 34: 671-680. [10] Wachniew R. & Rozanski K. (1997) Carbon budget of a mid-latitude, groundwater-controlled lake: Isotopic evidence for the importance of dissolved inorganic carbon recycling. Geochim. Cosmochim. Acta, 61: 2453-2465. [10] Wachsmann M. & Heumann K. G. (1991) Negative thermal ionization mass spectrometry of main group elements, Part 1.5 th group: phosphorus, arsenic and antimony. Int. J. Mass Spectr. Ion Proc., 108: 75-86. [39] Wachsmann M. & Heumann K. G. (1992) Negative thermal ionization mass spectrometry of main group elements. Part 2. 6th group: sulfur, selenium, and tellurium. Int. J. Mass Spectrom. Ion. Proc., 114: 209-220. [20, 26, 29, 39] Wada H. & Siuzuki K. (1983) Carbon isotopic thermometry calibrated by dolomite-calcite solvus temperatures. Geochim. Cosmochim. Acta, 47: 697-706. [46] Wada H. (1988) Microscale isotopic zoning in calcite and graphite crystals in marble. Nature, 331: 61-63. [12] Wadleigh M. A., Schwarcz H. P. & Kramer J. R. (1994) Sulphur isotope tests of seasalt correction factors in precipitation: Nova Scotia, Canada. Water Air Soil Pollut., 77: 1-16. [26] Wadleigh M. A. & Blake D. M. (1999) Tracing sources of atmospheric sulphur using epiphytic lichens. Environ. Pollut., 106: 265-271. [26] Wahlen M. & Yoshinari T. (1985) Oxygen isotope ratios in N20 from different environments. Nature, 313: 780-782. [19, 34] Wainwright S. C., Haney J. C., Kerr C., Golovkin A. N. & Flint M. V. (1998) Utilization of nitrogen derived from seabird guano by terrestrial and marine plants at St. Paul, Pribilof Islands, Bering Sea, Alaska. Marine Biol., 131: 63-71. [9] Wakeham S. G. & Volkman J. K. (1991) Sampling and analysis of lipids in marine particulate matter. Geophys. Monographs, 63: 171-179. [9] Wakshal E. & Nielsen H. (1982) Variations of ~)34S(SO4), ~)180(H20) and C1/SO4 ratio in rainwater over northern Israel, from the Mediterranean coast to Jordan Rift Valley and Golan Heights. Earth Planet. Sci. Lett., 61: 272-282. [26] Walder A. J. & Freedman P. A. (1992) Isotopic ratio measurement using a double focusing magnetic sector mass analyzer with an inductively coupled plasma as an ion source. J. Anal. At. Spectrom., 7: 571-575. [31, 35] Walder A. J. & Furuta N. (1993) High-precision lead isotope ratio measurement by inductively coupled plasma multiple collector mass spectrometry. Anal. Sci., 9: 675-680. [31] Walder A. J., Abell I. A., Platzner I. & Freedman P. A. (1993a) Lead isotope ratio measurement of NIST 610 glass by laser ablation inductively coupled plasma mass spectrometry. Spectrochim. Acta, B 48: 397-402. [35] Walder A. J., Koller D., Reed N. M., HuRon R. C. & Freedman P. A. (1993b) Isotope ratio measurement by inductively coupled plasma multiple collector mass spectrometry incorporating a high efficiency nebulization system. J. Anal. At. Spectrom., 8: 1037-1041. [31] Walder A. J., Platzner I. & Freedman P. A. (1993c) Isotope ratio measurement of lead, neodymium and neodymium - samarium mixtures, hafnium and hafnium - lutetium mixtures with a double focusing multiple collector inductively coupled plasma mass spectrometer. J. Anal. At. Spectrom., 8: 19-23. [31, 35] Walder A. J. (1997) Advanced isotope ratio mass spectrometry. II: isotope ratio measurement by multiple collector inductively coupled plasma spectrometry. In: Modern isotope ratio mass spectrometry. I. T. Platzner (ed.). John Wiley & Sons. Chemical analysis, a series of monographs on analytical chermistry and its applications, Vol. 145: 83-108. [31] Walker G. R., Woods P. H. & Allison G. B. (1994) Interlaboratory comparison of methods to determine the stable isotope composition of soil water. Chem. Geol. (Isot. Geosci. Sect.), 111: 297-306. [1, 2] Waiters L. J. Jr., Claypool G. E. & Choquette P. W. (1972) Reaction rates and 6180 variation for the carbonate-phosphoric acid preparation method. Geochim. Cosmochim. Acta, 36: 129-140. [12] Wan D. F., Li Y. H., Song H. B. & Liu Z. J. (1997) Preparation and calibration of reference material for silicon isotope. Acta Geoscientia Sinica, 18(3): 330-335 (in Chinese, with English abstract). [25]
1172
References
Wang C.-H. & Yeh H.-W. (1985) Oxygen isotopic compositions of DSDP Site 480 diatoms: Implications and applications. Geochim. Cosmochim. Acta, 49: 1469-1478. [2] Wang J., Davis A. M., Clayton R. N. & Hashimoto A. (1999) Evaporation of single crystal forsterite: evaporation kinetics, magnesium isotope fractionation, and implications of mass-dependent isotopic fractionation of a diffusion-controlled reservoir. Geochim. Cosmochim. Acta, 63: 953-966. [46] Wang J., Davis A. M., Clayton R. N., Mayeda T. K. & Hashimoto A. (2001) Chemical and isotopic fractionation during the evaporation of the FeO-MgO-SiO2-CaO-A1203-TiO2 rare earth element system. Geochim. Cosmochim. Acta, 65: 479-494. [46] Wang X. F. & Yakir D. (1995) Temporal and spatial variations in the oxygen-18 content of leaf water in different plant species. Plant, Cell Environm., 18(12): 1377-1385. [21] Wang Y., Huntington T. G., Osher L. J., Wassenaar L. I., Trumbone S. E., Amundson R. G., Harden J. W., McKnight D. M., Schiff S. L., Aiken G. R., Lyons W. B., Aravena R. O. & Baron J. S. (1998) Carbon Cycling in Terrestrial Environments. I. In: Isotope tracers in catchment hydrology. C. Kendall & J. J. McDonnel (eds.). Elsevier Science B.V., New York: 577-610. [26] Wankel S. D., Kendall C., Silva S. R. & Bernis B. E. (2002) Estimating atmospheric nitrate deposition using nitrate 6170, 6180 and 615N. EOS, Trans. AGU. Annual Fall Meeting, San Francisco, CA. December 6-10. [15] Ward M. H., Mark S. D., Cantor K. P., Weisenburger D. D. & Correa-Villasenor A. (1996) Drinking water nitrate and the risk of non-Hodgkin's lymphomia. Epidemiology, 7: 465-471. [15] Ward S., Scantlebury M., Krol E., Thompson P. J., Sparling C. & Speakman J. R. (2000) Preparation of hydrogen from water by reduction with lithium aluminum hydride for the analysis of 62H by isotope ratio mass spectrometry. Rapid Commun. Mass Spectrom., 14: 450-453. [1] Ware R. K. & Pirooz P. P. (1967) Gas adsorption on freshly broken glass surfaces- a source of error in analysis of bubbles in glass. Glass Technol., 8: 86-87. [3] Wassenaar L. I. (1995) Evaluation of the origin and fate of nitrate in the Abbotsford Aquifer using the isotopes of 15N and 180 in NO3-. Appl. Geochem., 10: 391-405. [15] Wassenaar L. I. & Koehler G. (1999) An on-line technique for the determination of the 6180 and 6170 of gaseous and dissolved oxygen. Anal. Chem., 71: 4965-4968. [15, 36] Wasserburg G. J., Lee T. & Papanastassiou D. A. (1977) Correlated O and Mg isotopic anomalies in Allende inclusions. II: Magnesium. Geophys. Res. Lett., 4: 299-302. [31] Wasserburg G. J., Jacobsen S. B., DePaolo D. J., McCulloch M. T. & Wen T. (1981) Precise determination of S m / N d ratios, Sm and Nd isotopic abundances in standard solutions. Geochim. Cosmochim. Acta, 45: 2311-2323. [31] Watanabe S., Kazunobu M. & Matsuo S. (1983) Isotopic ratios of carbonaceous materials incorporated in olivine crystals from the Hualalai Volcano, Hawaii- An approach to mantle carbon. Geochem. J., 17: 95-104. [3] Watson L. L., Hutcheon I. D., Epstein S. & Stolper E. M. (1994) Water on Mars: Clues from deuterium/ hydrogen and water contents of hydrous phases in SNC meteorites. Science, 265: 86-90. [30] Wayne D. M., Yoshida T. M. & Vance D. E. (1996) Analysis of trace impurities in palladium metal powders by glow discharge mass spectrometry. J. Anal. At. Spectrom., 11" 861-869. [35] Wayne D. M. (1997) Direct determination of trace noble metals (palladium, platinum, and rhodium) in automobile catalysts by glow discharge mass spectrometry. J. Anal. At. Spectrom., 12: 1195-1202. [35] Wayne D. M., Schulze R. K., Maggiore C., Cooke D. W. & Havrilla G. (1999) Characterization of tantalum films on analytical surfaces: Insight into sputtering of non-conductors in a dc-glow discharge using secondary cathodes. Appl. Spectrosc., 53: 266-277. [35] Webb E. A. & Longstaffe F. J. (2000) The oxygen isotopic compositions of silica phytoliths and plant water in grasses: implications for the study of paleoclimate. Geochim. Cosmochim. Acta, 64: 767-780. [46] Webb E. A. & Longstaffe F. J. (2002) Climatic influences on the oxygen isotopic composition of biogenic silica in prairie grass. Geochim. Cosmochim. Acta, 66, 1891-1904. [46] Weber D., Kexel H. & Schmidt H.-L. (1997) 13C-pattern of natural glycerol: Origin and practicle importance. J. Agric. Food Chem., 45: 2042-2046. [8]
References
1173
Weber J. N., Deines P., Weber P. H. & Baker P. A. (1976) Depth related changes in the 13C/12C ratio of skeletal carbonate deposited by the Caribbean reefframe building coral Montastrea annularis: further implications of a model for stable isotope fractionation by scleractinian corals. Geochim. Cosmochim. Acta, 40: 31-39. [44] Weber P. K., Hutcheon I. D., McKeegan K. D. & Ingram B. L. (2002) Otolith sulfur isotope method to reconstruct salmon (Oncorhynchus tshawytscha) life history. Can. J. Fish Aquat. Sci., 59: 587-591. [26] Webster C. L. (1972) Selenium isotope analysis and geochemical applications. PhD Thesis, Colorado State University. [29] Webster C. L. Jr. & Warren C. G. (1981) Selenium isotope analysis. U. S. Geological Survey, Open-File Rept., 81-992:20 p. [20] Webster C. R. & May R. D. (1992) In situ stratospheric measurements of CH4, 13CH4, N20 and OC180 using the BLISS tunable diode laser spectrometer. Geophys. Res. Lett., 19: 45-48. [34] Webster C. R., May R. D., Owen T., Herman R. L., Rossman G. R. & Yung Y. L. (1999) Tunable laser spectrometers for in-situ gas, particulate, and mineralogical studies of Mars. Centre National d'Etudes Spatiales, International Conference. Mars Exploration Program and Sample Return Missions, P8 $3(64). [33] Webster L. A., Wahl M. H. & Urey H. C. (1935) The fractionation of the oxygen isotopes in an exchange reaction. J. Chem. Phys., 3: 129. [46] Wedepohl K. H. (1995) The composition of the continental crust. Geochim. Cosmochim. Acta, 59: 12171232. [15] Wefer G. & Berger W. H. (1991) Isotope paleontology: growth and composition of extant calcareous species. Marine Geol., 100: 207-248. [46] Weidmann D., Roller C., Tittel E K., Curl R. F., Uehara K., Oppenheimer C. & DeNatale P. (2003) Development of a 4.3 ~m quantum cascade laser based 13CO2/12CO2 isotope ratio sensor. 2nd International Symposium on Isotopomers, 4-6 November, Stresa, Italy. [34] Weilacher T., Gleixner G. & Schmidt H.-L. (1996) Carbon-isotope pattern in purine alkaloids a key to isotope discriminations in C-1 compounds. Phytochem., 41: 1073-1077. [8] Weiler R. R. & Nriagu J. O. (1978) Isotopic composition of dissolved inorganic carbon in the Great lakes. J. Fisheries Res. Board of Canada, 35: 422-430. [10] Welch S. A., Beard B. L., Johnson C. M. & Braterman P. S. (2003) Kinetic and equilibrium Fe isotope fractionation between aqueous Fe(II) and Fe(III). Geochim. Cosmochim. Acta, 67: in press. [46] Wendt K., Bushaw B. A., Bhowmick G., Bystrow V. A., Kotovski N., Kratz J. V., Lantzsch J., Miiller P., NSrtersh/iuser W., Often E. W., Seibert A., Trautmann N., Waldek A. & Yushkevich Y. (1996) 89,90Sr-determination on various environmental samples by collinear resonance ionization spectroscopy. Resonance Ionization Spectroscopy 1996, Eighth International Symposium. N. Winograd & J.E. Parks (eds.). AIP Press Woodbury, New York, AIP Conference Proceddings 388: 361-364. [33] Wendt K. (1998) Ultratrace determination of long-lived radioactive isotopes. Resonance Ionization Spectroscopy, Ninth International Symposium. J.C. Vickerman, I. Lyon, N.P. Lockyer & J.E. Parks (eds.). AIP Press Woodbury, NY. AIP Conference Proceedings 454: 39-46. [33] Wendt K., Bushaw B. A., Gruning C., Horn R., Huber G., Katz J. V., Kunz P., Miiller P., NiJrtersh~iuser W., Nunnemann M., Passler G., Schmitt A., Trautmann N. & Waldek, A. (1999) Recent developments in and applications of resonance ionization mass spectrometry. Fresenius J. Anal. Chem., 364: 471-477. [33] Wenner D. B. & Taylor H. P. Jr. (1971) Temperatures of serpentinization of ultramafic rocks based on 180 / 160 fractionation between coexisting serpentine and magnetite. Contrib. Mineral. Petrol., 32: 165185. [46] Wenzel B., L6cuyer C. & Joachimski M. M. (2000) Comparing oxygen isotope records of silirian calcite and phosphate - 6180 compositions of brachiopods and conodonts. Geochim. Cosmochim. Acta, 64: 1859-1872. [22] Werle P. (1999) Laseroptical sensors for in-situ gas analysis. Rec. Res. Developm. Opt. Eng., 2: 143-160. [34] Werne J. P., Lyons T. W., Hollander D. J., Formolo M. J & Damste J. S. S. (2003) Reduced sulfur in euxinic sediments of the Cariaco Basin: sulfur isotope constraints on organic sulfur formation. Chem. Geol., 195: 159-179. [26]
1174
References
Werner R. A., Kornexl B. E., Rossmann A. & Schmidt H.-L. (1996) On-line determination of 6180 values
of organic substances. Anal. Chim. Acta, 319: 159-164. [1, 23] Werner R. A. & Brand W. A. (2001) Referencing strategies and techniques in stable isotope ratio analysis.
Rapid Commun. Mass Spectrom., 15: 501-519. [40] Werner R. A., Br~iunlich S. & Brand W. A. (2001) On-line determination of 5180 in silicates, carbonates
and other inorganic substances: First results. Abstr. Papers Am. Chem. Soc., 221 (98-GEOC, Part I, Apr. 1, 2001). [201 Westaway K. C., Koerner T., Fang Y.-R., Rudzinski J. & Paneth P. (1998) A new method of determining chlorine kinetic isotope effects. Anal. Chem., 70: 3548-3552. [28] Westgate L. M. & Anderson T. E (1982) Extraction of various forms of sulphur from coal and shale for stable sulphur isotope analysis. Anal. Chem., 54: 2136-2139. [26] Weston R. E. (1999) Anomalous or mass-independent isotope effects. Chem. Rev., 99: 2115-2136. [18] Wetherill G. W. (1964) Isotopic composition and concentration of molybdenum in iron meteorites. 1. Geophys. Res., 69(20): 4403-4408. [29] Wetherill G. W. & Chapman C. R. (1998) Asteroids and meteorites. In: Meteorites and the early solar system. J. F. Kerridge & M. S. Mathews (eds.). Univ. Arizona Press: 35-67. [11] Wheat J. A. (1971) Isotopic analysis of lithium by atomic absorption spectrophotometry. Appl. Spectrosc., 25: 328-330. [6] Wheeler M. D., Newman S. M., Orr-Ewing A. J. & Ashfold M. N. R. (1998) Cavity ring-down spectroscopy. J. Chem. Soc. Farad. Trans., 94: 337-351. [34] White A. F. & Chuma N. J. (1987) Carbon and isotopic mass balance models of Oasis Valley-Fortymile Canyon groundwater basin, southern Nevada. Water Resour. Res., 23: 571-582. [10] White J. U. (1976) Very long optical paths in air. J. Opt. Soc. Am., 66: 411-416. [34] White J. w. & Gedzelman S. D. (1984) The isotopic composition of atmospheriic water vapor and the concurrent meteorological conditions. J. Geophys. Res., 89: 4937-4939. [1] White J. w. C., Lawrence J. R. & Broecker W. S. (1994) Modeling and interpreting D / H ratios in tree rings: a test case of White Pine in the northeastern United States. Geochim. Cosmochim. Acta, 58: 851862. [46] Whittaker M., Pollard S. J. T. & Fallick A. E. (1995) Characterization of refractory wastes at heavy oilcontaminated sites - a review of conventional and novel analytical methods. Environ. Technol., 16: 1009-1033. [8] Whittaker M., Pollard S. J. T., Fallick A. E. & Preston T. (1996) Characterization of refractory wastes at hydrocarbon-contaminated sites. 2. Screening of reference oils by stable carbon isotope fingerprinting. Environ. Pollut., 94: 195-203. [8] Whittaker P. G., Barrett J. F. R. & Williams J. G. (1992) Precise determination of iron isotope ratios in whole blood using inductively coupled plasma mass spectrometry. J. Anal. Atom. Spectrom., 7: 109113. [31] Wiechert U. & Hoefs J. (1995) An excimer laser-based micro analytical preparation technique for in-situ oxygen isotope analysis of silicate and oxide minerals. Geochim. Cosmochim. Acta, 59: 4093-4101. [20] Wiechert U., Fiebig J., Przybilla R., Xiao Y. & Hoefs J. (2002) Excimer laser isotope-ratio-monitoring mass spectrometry for in situ oxygen isotope analysis. Chem. Geol., 182: 179-194. [20] Wieder R. K., Lang G. E. & Granus V. A. (1985) An evaluation of wet chemical methods for quantifying sulfur fractions in freshwater wetland peat. Limnol. Oceanogr., 30: 1109-1115. [26] Wieder R. K. & Lang G. E. (1988) Cycling of inorganic and organic sulfur in peat from Big Run Bog, West Virginia. Biogeochem., 5: 221-242. [26] Wieser M. E. (1997) Stable isotope ratio mass spectrometry of nanigram quatities of boron and sulfur. PhD Thesis, University of Calgary, Calgary, Canada: 127 p. [26] Wigley T. M. L., Plumer L. N. & Pearson F. J. (1978) Mass transfer and carbon isotope evolution in natural waters. Geochim. Cosmochim. Acta, 42: 1117-1139. [10] Wijbrands J. R. & McDougall I. (1986) 40Ar/39Ar dating of white micas from an Alpine high-pressure metamorphic belt on Naxos (Greece): the resetting of the argon isotopic system. Contrib. Mineral. Petrol., 93: 187-194. [16] Wilbur K. M. & Saleuddin A. S. M. (1983) Shell formation. In: The Mollusca. A. S. M. Saleuddin & K. M. Wilbur (eds.). Academic Press: 236-279. [46]
References
1175
Wilde S. A., Valley J. W., Peck W. H. & Graham C. M. (2001) Evidence from detrital zircons for the existence of continental crust and oceans on the Earth 4.4 Gyr ago. Nature, 409: 175-178. [30] Wiley K. F., Arlinhaus H. F. & Whitaker T. J. (1999) Routine analysis at sub-micron resolution through the use of sputter initiated resonance ionization spectroscopy. Appl. Surf. Sci., 145: 36-40. [33] Wilkinson J. J., Jenkin G. R. T., Fallick A. E. & Foster R. P. (1995) Oxygen and hydrogen isotopic evolution of Variscan crustal fluids, south Cornwall, U.K. Chem. Geol., 123: 239-254. [3] Willey J. F. & Taylor J. W. (1978) Capacitive integration to produce high-precision isotope ratio measurements on methyl chloride and methyl bromide samples. Anal. Chem., 50: 1930-1933. [28] Williams J. D. H., Syers J. K. & Walker T. W. (1967) Fractionation of soil inorganic phosphate by a modification of Chang and Jackson's procedure. Soil Sci. Soc. Am. Proc., 31: 736-739. [2] Williams J. M., Jegou H., Langlois V., Zhang B. L. & Martin M. L. (1993) The search for D-glucose derivates suitable for the study of natural hydrogen isotope fractionation. Carbohydr. Res., 245: 57-64. [5] Williams L. B., Ferrell R. E., Hutchean I., Bakel A. J., Walsh M. M. & Krouse H. R. (1995) Nitrogen isotope geochemistry of organic matter and mineral during diagenesis and hydrocarbon migration. Geochim. Cosmochim. Acta, 59: 765-779. [16] Williams L. B. & Hervig R. L. (1997) Potential for in-situ measurements of isotopic ratios in clay minerals using secondary ion mass spectrometry. 11th International Clay Conference, Ottawa, Canada (Abstract). [2] Williams L. B., Hervig R. L., Holloway J. R. & Hutcheon I. (2001a) Boron isotope geochemistry during diagenesis. Part I. Experimental determination of fractionation during illitization of smectite. Geochim. Cosmochim. Acta, 65: 1769-1782. [2] Williams L. B., Hervig R. L., Wieser M. E. & Hutcheon I. (2001b) The influence of organic matter on the boron isotope geochemistry of the gulf coast sedimentary basin, USA. Chem. Geol., 174: 445-461. [30] Williams P. (1979) The sputtering process and sputtered ion emission. Surf Sci., 90: 588-634. [30] Williams P. (1982) On mechanisms of sputtered ion emission. Appl. Surface Sci., 13: 241-259. [30] Wilson A. T. & Grinsted M. J. (1975) Palaeotemperatures from tree rings and the D / H ratio of cellulose as a biochemical thermometer. Nature, 257: 387-388. [24] Wilson A. T. & Grinsted M. J. (1977) 13C/12 C in cellulose and lignin as palaeothermometers. Nature, 265: 133-135. [24] Wilson B. E. Jr. & Wells A. J. (1946) The experimental determination of the intensities of infra-red absorption bands. I. The theory of the method. J. Chem. Phys., 14: 578-565. [34] Wilson B. E. Jr., Decius J. C. & Cross P. C. (1980) Molecular vibrations: The theory of infrared and raman vibrational spectra. Dover Publications, New York. [34] Wilson P. F., Freeman C. G., McEwan M. J., LMiligan D. B., Allardyce R. A. & Shaw G. M. (2001) Alcohol in breath and blood: a selected ion flow tube mass spectrometric study. Rapid Commun. Mass Spectrom., 15: 413-417. [4] Wilson R. G. (1995) SIMS quantification in Si, GaAs, and diamond - an update. Int. J. Mass Spectrom. Ion Proc., 143: 43-49. [30] Winchester M. R., Lazik C. & Marcus R. K. (1991a) Characterization of a radio frequency glow discharge emission source. Spectrochim. Acta, B 46: 483-499. [35] Winchester M. R., Hayes S. M. & Marcus R. K. (1991b) Determination of platinum and rhodium in alumina catalysts by glow discharge atomic absorption spectrophotometry. Spectrochim. Acta, B 46: 615627. [35] Winchester M. R., Duckworth D. C. & Marcus R. K. (1993) Analysis of nonconducting sample types. In: Glow discharge spectroscopies. R. K. Marcus (ed.). Plenum Press, New York: 263-328. [35] Winteringham E P. (1980) Nitrogen balance and related studies: a global review. In: Soil Nitrogen as Fertilizer or Pollutant. IAEA, Vienna: 307-344. [15] Wise L. E., Murphy M. & D'Addieco A. (1946) Chlorite holocellulose, its fractionation and bearing on summative wood analysis and on studies in the hemicellulose. Paper Trade J., 122: 35. [24] w i t t G. B., Moll E. J., Beeton R. J. S. & Murray P. J. (1998) Isotopes, wool, and rangeland monitoring: Let the sheep do the sampling. Environ. Management, 22: 145-152. [9] Witter A. E. & Jones A. D. (1999) Chemical characterization of organic constituents from sulfide-rich produced water using gas chromatography/mass spectrometry. Environ. Toxicol. Chem., 18: 19201926. [26]
1176
References
Wolfe R. R. (1992) Radioactive and stable isotope tracers in biomedicine. Wiley-Liss Inc., New York. [8] Wombacher F., Rehk~imper M., Miinker C. & Mezger K. (2000) Cadmium isotope analysis by MCICPMS. EOS Trans. AGU, 81: 1362. [31] Wombacher F., Rehk~imper M., Mezger K., Miinker C. & Bischoff A. (2001) Mass-dependent cadmium isotope fractionation in meteorites and experiments. J. Conf. Abstr., 6: 689. [31] Wong W. W., Caberra M. P. & Klein P. D. (1984) Evaluation of a dual mass spectrometer system for rapid simultaneous determination of hydrogen-2/hydrogen-1 and oxygen-18/oxygen-16 ratios in aqueous samples. Anal. Chem., 56: 1852-1858. [1] Wong W. W. (1985) Comparison of infrared and mass-spectrometric measurements of carbon-13/carbon-12 ratios. Int. J. Appl. Radiat. Isot., 36: 997-999. [34] Wong W. W. & Klein P. D. (1986) A review of techniques for the preparation of biological samples for mass-spectrometric measurements of hydrogen-2/hydrogen-1 and oxygen-18/oxygen-16 isotope ratios. Mass Spectrom. Rev., 5: 313-342. [1, 23] Wong W. W., Lee L. S. & Klein P. D. (1987a) Oxygen isotope ratio measurements on carbon dioxide generated by reaction of microliter quantities of biological fluids with guanidine hydrochloride. Anal. Chem., 59: 690-693. [1] Wong W. W., Lee L. S. & Klein P. D. (1987b) Deuterium and oxygen-18 measurements on microliter samples of urine, plasma, saliva and human milk. Am. J. Clin. Nutr., 45: 905-913. [1] Wong W. W., Clarke L. L., Llaurador M. & Klein P. D. (1992a) A new zinc product for the reduction of water in physiological fluids to hydrogen gas for 2H/1H isotope ratio measurements. Eur. J. Clin. Nutr., 46: 69-71. [1] Wong W. W., Clarke L. L., Johnson G. A., Llaurador M. & Klein P. D. (1992b) Comparison of two elemental-analyzer gas-isotope-ratio mass spectrometer system in the simultaneous measurement of 13C/12C ratios and carbon content in organic samples. Anal. Chem., 64: 354-358. [9] Wong W. W., Klein P. D., Parr R. M. & Clements S. A. (1993) Interlaboratory analysis of reference water samples enriched in deuterium and oxygen-18. Int. J. Appl. Radiat. Isot., 44: 561-566. [1] Woo J. C., Jakubowski N. & Stuewer D. (1993) Analysis of aluminium oxide powder by glow discharge mass spectrometry with low mass resolution. J. Anal. At. Spectrom., 8: 881-889. [35] Woodbury S. E., Evershed R. P., Rossell J. B., Griffith R. E. & Farnell P. (1995) Detection of vegetable oil adulteration using gas-chromatography combustion isotope ratio mass-spectrometry. Anal. Chem., 67: 2685-2690. [8] Woodbury S. E., Evershed R. P. & Rossell J. B. (1998a) 613C analyses of vegetable oil fatty acid components, determined by gas chromatography combustion isotope ratio mass spectrometry, after saponification or regiospecific hydrolysis. J. Chromatogr.,, A 805: 249-257. [8] Woodbury S. E., Evershed R. P. & Rossell J. B. (1998b) Purity assessments of major vegetable oils based on 613C values of individual fatty acids. J. Am. Oil Chem. Soc., 75: 371-379. [8] Wortel T. M. (1979) Some applications of zeolites in organic chemistry. Delft University Press, Delft, Netherlands: 127 p. [36] Woutersen S., Milan J. B., Buma W. J. & deLange C. A. (1997) Resonance-enhanced multiphotonionization photoelectron spectroscopy of even-parity autoionizing Rydberg states of atomic sulphur. J. Chem. Phys., 106: 6831-6838. [33] Wright I. P., Boyd S. R., Franchi I. A. & Pillinger C. T. (1988) High-precision determination of nitrogen stable isotope ratios at the sub-nanomole level. J. Phys. E: Sci. Instrum., 21: 865-875. [13, 17] Wright I. P. & Pillinger C. T. (1989) Carbon isotopic analysis of small samples by use of stepped-heating extraction and static mass spectrometry. In: New frontiers in stable isotopic research; Laser probes, ion probes, and small-sample analysis. W. C. III Shanks & R. E. Criss (eds.). U.S. Geol. Surv., Bull. 1890: 9-34. [13] Wu S., Ding T., Meng X. & Bai L. (1997) Determination and geological implication of O-Si isotope of the sediment core in the CC area, the Pacific Ocean. Chinese Sci. Bull., 42(17): 1462-1465. [25] Xenakis D., Faucher O., Charalambidis D. & Fotakis C. (1996) Two-XUV-photon ionization of argon using a harmonic radiation source. N. Winograd & J.E. Parks (eds.). Resonance Ionization Symposium 1996, Eighth International Symposium. AIP Press Woodbury, NY, AIP Conference Proceedings 388: 365-368. [33]
References
1177
Xiao Y.-K., Beary E. S. & Fassett J. D. (1988) An improved method for the high-precision isotopic measurement of boron by thermal ionization mass spectrometry. Int. J. Mass Spectrom. Ion Proc., 85: 203213. [7] Xiao Y.-K. & Beary E. S. (1989) High-precision isotopic measurement of lithium by thermal ionization mass spectrometry. Int. J. Mass Spectrom. Ion Proc., 94: 101-114. [6] Xiao Y.-K., Jin L. & Qi H.-P. (1991) Investigation of thermal ion emission characteristics of graphite. Int. J. Mass Spectrom. Ion Proc., 107: 205-213. [7] Xiao Y.-K. & Zhang C.-G. (1992) High precision isotopic measurement of chlorine by thermal ionization mass spectrometry of the Cs2C1+ ion. Int. J. Mass Spectrom. Ion Proc., 116: 183-192. [28] Xiao Y. K., Qi H. P., Wang Y. H. & Jin L. (1992a) The comparison of loading forms in isotopic measurement of lithium by thermal ionization mass spectrometry. Chin. Sci. Bull., 37: 469-472. [6] Xiao Y., Sun Y. K., Wang Y. H., Oi H. P. & Jin L. (1992b) Boron isotope composition of brine, sediments and source water in Da Qaidam Lake, Qinghai, China. Geochim. Cosmochim. Acta, 56: 1561-1568. [46] Xiao Y.-K., Liu W.-G., Qi H.-P. & Zhang C.-G. (1993) A new method for the high-precision isotopic measurement of bromine by thermal ionization mass-spectrometry. Int. J. Mass Spectrom. Ion Proc., 123: 117-123. [28] Xiao Y. & Wang L. (2001) The effect of pH and temperature on the isotopic fractionation of boron between saline brine and sediment. Chem. Geol., 171: 253-261. [46] Xiao Y., Zhou Y., Wang Q., Wei H., Liu W. G. & Eastoe C. J. (2002a) A secondary isotopic reference material of chlorine from selected seawater. Chem. Geol., 182: 655-661. [40] Xiao Y. K., Zhou Y. M., Liu W. G., Hong A. S., Wang Q. Z., Wei H. Z. & Shirodkar P. V. (2002b) The character of chlorine isotopic compositions in ocean water. Geol. Review, 48: in press (in Chinese). [40] Xue S., Yang Y.-L., Hall G. S. & Herzog G. F. (1997) Germanium isotopic compositions in Canyon Diablo spheroids. Geochim. Cosmochim. Acta, 61: 651-655. [31] Yakir D. (1992) Variations in the natural abundance of oxygen-18 and deuterium in plant carbohydrates. Plant, Cell Environ., 15: 1005-1020. [23] Yakir D. & Wang X. F. (1996) Fluxes of CO2 and water between terrestrial vegetation and the atmosphere estimated from isotope measurements. Nature, 380: 515-517. [21] Yamamuro M. & Kayanne H. (1995) Rapid direct determination of organic carbon and nitrogen in carbonate-bearing sediments with a Yanaco MT-5 CHN analyzer. Limnol. Oceanogr., 40: 1001-1005.
[9]
Yarnamoto T., Hashizume K., Matsuda K. & Kase T. (1998) Multiple nitrogen isotopic components coexisting in ureilites. Meteorit. Planet. Sci., 33: 857-870. [17] Yamulki S., Toyoda S., Yoshida N., Veldkamp E., Grant B. & Bol R. (2001) Diurnal fluxes and the isotopomer ratios of N20 in a temperate grassland following urine amendment. Rapid Commun. Mass Spectrom., 15: 1263-1269. [19] Yang J. & Epstein S. (1983) Interstellar organic matter in meteorites. Geochim. Cosmochim. Acta, 47: 21992216. [11] Yang J. & Epstein S. (1987) The effect of the isotopic composition of oxygen on the non-mass-dependent isotopic fractionation in the formation of ozone by discharge of 02*. Geochim. Cosmochim. Acta, 51: 2011-2017. [18] Yang W., Spencer R. J. & Krouse H. R. (1995) Stable isotope and major element compositions of fluid inclusions in Devonian and Cambrian dolomite cements, western Canada. Geochim. Cosmochim. Acta, 59: 3159-3172. [3] Yang C., Telmer K. & Veizer J. (1996a) Chemical dynamics of the "St. Lawrence" riverine system: 6DH20, 618OH20, 613CDIC, 634Ssulfide, and dissolved 87Sr/86Sr. Geochim. Cosmochim. Acta, 60: 851-866.
[lOl
Yang W., Krouse H. R. & Spencer R. J. (1996b) Improved techniques for stable isotope analyses of microlitre quantities of water from fluid inclusions in halite and concentrated brines. Chem. Geol., 130: 139-145. [1, 3] Yanagisawa E & Sakai H. (1983) Thermal decomposition of barium sulfate - vanadium pentaoxide silica glass mixtures for preparation of sulfur dioxide in sulfur isotope ratio measurements. Anat. Chem., 55: 985-987. [26]
117 8
References
Yapp C. J. & Epstein S. (1982) A reexamination of cellulose carbon-bound hydrogen 6D measurements and some factors affecting the plant-water D /H relationships. Geochim. Cosmochim. Acta, 46" 955-965. [46] Yapp C. J. & Pedley M. D. (1985) Stable hydrogen isotopes in iron oxides - II. D / H variations among natural goethites. Geochim. Cosmochim. Acta, 49: 487-495. [2, 46] Yapp C. J. (1987) Oxygen and hydrogen isotope variations among goethites (R-FeOOH) and the determination of paleotemperatures. Geochim. Cosmochim. Acta, 51" 355-364. [2, 46] Yapp C. J. (1990) Oxygen isotopes in iron (III) oxides 1. Mineral-water fractionation factors. Chem. Geol., 85: 329-335. [461 Yapp C. J. (1991) Oxygen isotopes in an oolitic ironstone and the determination of goethite 6180 values by selective dissolution of impurities: The 5 M NaOH method. Geochim. Cosmochim. Acta, 55: 26272634. [2] Yapp C. J. & Poths H. (1991) 13C/12Cratios of the Fe(III) carbonate component in natural goethites. In" Stable isotope geochemistry: In: Stable Isotope Geochemistry: A tribute to Samuel Epstein. H. P. Jr. Taylor, J. R. O'Neil & I. R. Kaplan (eds.). The Geochemical Society, Spec. Publ. 3: 257-270. [2] Yapp C. J. & Poths H. (1993) The carbon isotope geochemistry of goethite (cz-FeOOH) in ironstone of the Upper Ordovician Neda Formation, Wisconsin, USA: Implications for early Paleozoic continental environments. Geochim. Cosmochim. Acta, 57: 2599-2611. [2] Yapp C. J. & Poths H. (1995) Stable hydrogen isotopes in iron oxides: III. Nonstoichiometric hydrogen in goethite. Geochim. Cosmochim. Acta, 59(16): 3405-3412. [2] Yapp C. J. (1998) Paleoenvironmental interpretation of oxygen isotope ratios in oolitic ironstones. Geochim. Cosmochim. Acta, 62: 2409-2420. [2] Yapp C. J. (2000) Climatic implications of surface domains in arrays of 6D and 6180 from hydroxyl minerals: goethite as an example. Geochim. Cosmochim. Acta, 64: 2009-2025. [46] Yeh H. W. (1974) Oxygen isotope studies of ocean sediments during sedimentation and burial diagenesis. PhD Thesis, Case Western Reserve University: 136 p. [2] Yeh H.-W. & Epstein S. (1978) 29Si/28Si and 30Si/28Si of meteorites and Allende inclusions. Lunar Planet. Sci., IX: 1289-1291 (Abstract). [25] Yeh H. W. (1980) D / H ratios and late stage dehydration of shales during burial. Geochim. Cosmochim. Acta, 44: 341-352. [2] Yi W., Halliday A. N., Lee D.-C. & Christensen J. N. (1995) Indium and tin in basalts, sulfides, and the mantle. Geochim. Cosmochim. Acta, 59: 5081-5090. [31] Yi W., Budd P., McGill R. A. R., Young S. M. M., Halliday A. N., Haggerty R., Scaife B. & Pollard A. M. (1999) Tin isotope studies of experimental and prehistoric bronzes. In: The beginnings of metallurgy. A. Hauptmann, E. Pernicka, T. Reheren, and U. Yalcin (eds.). Deutsches Bergbau-Museum: 285290. [31] Yildiz O. (1995) Ist der Isotopengehalt des Holzes aus Jahrringen von B~iumen als Klimaindikator geeignet? Diploma Thesis, Fachhochschule Aachen. [24] Yong J. W. H., Wong S. C., Letham D. S., Hocart C. H. & Farquhar G. D. (2000) Effects of elevated [CO2] and nitrogen on cytokinins in he xylem sap and leaves of cotton. Plant Physiol., 124: 767-779. [21] Yonge C. J. (1982) Stable isotope studies of water extracted from speleothems. PhD Thesis, McMaster University: 298 p. [3] Yoo R. K. & Gellene G. I. (1995) Symmetry induced kinetic isotope effects in the formation of (CO2)2+. J. Chem. Phys., 102: 3227-3237. [18] Yorozu M., Okada Y. & Endo A. (1999) Detection of hydrogen isotope atoms by mass spectrometry combined with resonant ionization. Appl. Phys. B - Lasers and Optics, 69" 129-132. [33] Yoshida N. & Matsuo S. (1983) Nitrogen isotope ratio of atmospheric N20 as a key to the global cycle of N20. Geochem. J., 17: 231-239. [19] Yoshida N. & Mizutani Y. (1986) Preparation of carbon dioxide for oxygen-18 determination of water by use of a plastic syringe. Anal. Chem., 58: 1273-1275. [1] Yoshida N. & Toyoda S. (2000) Constraining the atmospheric N20 budget from intramolecular site preference in N20 isotopomers. Nature, 405- 330-334. [15, 19, 34]
References
1179
Yoshinari T. (1990) Emissions of N20 from various environments The use of stable isotope composition of N20 as tracer for the studies of N20 biogeochemical cycling. In: Denitrification in soil and sediment. N. P. Revsbech & J. Sorensen (eds.). Plenum Press, New York: 129-150. [19] Yoshinari T., Altabet M. A., Naqvi S. W. A., Cosdispoti L., Jayakumar A., Kuhland M. & Devol A. (1997) Nitrogen and oxygen isotopic composition of N20 from suboxic waters of the eastern tropical North Pacific and the Arabian Sea - measurement by continuous-flow isotope-ratio monitoring. Marine Chem., 56: 253-264. [19] You C. F., Spivack A. J., Smith J. H. & Gieskes J. M. (1993a) Mobilization of boron in convergent margins: Implications for boron geochemical cycle. Geology, 21: 207-210. [7] You C.-F., Gieskes J. M., Chen R. & Spivack A. J. (1993b) I, Br, dissolved organic carbon, B, 611B and Mn in pore waters of Site 808, Nankai Accretionary Prosm. ODP, Scientific Results, 131: 165-174. [7] You C.-F., Butterfield D., Spivack A. J., Gieskes J. M., Gamo T. & Campbell A. (1994) Boron and halide systematics in submarine hydrothermal vent fluids: effects of phase separation and sediment contribution. Earth Planet. Sci. Lett., 123: 227-238. [7] You C.-F., Chan L.-H., Spivack A. J. & Gieskes J. M. (1995a) Lithium, boron and their isotopes in sediments and pore waters of Ocean Drilling Program Site 808, Nankai Trough: Implications for fluid expulsion in accretionary prisms. Geology, 23: 37-40. [6, 7] You C.-F., Spivack A. J., Gieskes J. M., Rosenbauer R. & Bischoff J. M. (1995b) Experimental study of boron geochemistry implication for fluid processes in convergent margin. Geochim. Cosmochim. Acta, 59: 2435-2442. [7] You C. F. & Chan L. H. (1996) Precise determination of lithium isotopic composition in low concentration natural samples. Geochim. Cosmochim. Acta, 60(5): 909-915. [6, 31, 40] You C.-F., Spivack A. J., Gieskes J. M., Martin J. B. & Davisson L. (1996a) Boron contents and isotopic compositions in pore waters: new approach to evaluate the temperature squeezing artifacts-geochemical implications. Marine Geol., 129: 351-361. [7] You C.-F., Castillo P., Gieskes J. M., Chan L. & Spivack A. J. (1996b) The role of slab derived fluids in arc magmatism. Earth Planet. Sci. Lett., 140: 41-52. [7] You C.-F., Shen J.-S. & Lee T. (1999) Precise boron isotopic determination by negative thermal ionization mass spectrometry. J. Nat. Cheng Kung Univ., Sci. Eng. & Med. Sect., 34: 1-11. [7] Young E. D., Fogel M. L., Rumble D. I. & Hoering T. C. (1998a) Isotope-ratio-monitoring of 02 for microanalysis of 180/160 and 170/160 in geological materials. Geochim. Cosmochim. Acta, 62(18): 3087-3094. [20] Young E. D., Coutts D. W. & Kapitan D. (1998b) UV laser ablation and irm-GCMS microanalysis of 180 / 160 and 170/160 with applicatiopn to a calcium-aluminium-rich inclusion from the Allende meteorite. Geochim. Cosmochim. Acta, 62(18): 3161-3168. [20] Young E. D., Nagahara H., Mysen B. O. & Audet D. M. (1998c) Non-Rayleigh oxygen isotope fractionation by mineral evaporation: Theory and experiments in the system SiO2. Geochim. Cosmochim. Acta, 62: 3109-3116. [46] Young J. P., Barshick C. M., Shaw R. W. & Ramsey J. M. (1994) Application of diode lasers to the isotopically selective determination of uranium in oxides by optogalvanic spectroscopy. H.-J. Kluge, J.E. Parks & K. Wendt (eds.). Resonance Ionization Spectroscopy 1994, Seventh International Symposium. AIP Press Woodbury, NY. AIP Conference Proceddings 329: 111-114. [33] Yu F. J. (1991) The measurement of oxygen isotope composition of exceeding micro-water; capillary CO2-H20 room temperature equilibrium method. Kexue Tongboa (Chinese Sci. Bull.), 36(2): 140-144.
[3]
Yuen G., Blair N., DesMarais D. J. & Chang S. (1984) Carbon isotope composition of low molecular weight hydrocarbons and monocarboxylic acids from the Murchison meteorite. Nature, 307: 252-254.
[11]
Yun M., Wadleigh M. A. & Pye A. (in press) Direct measurement of sulphur isotopic composition in lichens by continuous flow isotope ratio mass spectrometry (CF-IRMS). Chem. Geol. [26] Yund R. A. & Anderson T. F. (1978) The effect of pressure on oxygen isotope exchange between feldspar and water. Geochim. Cosmochim. Acta, 42: 235-239. [46]
1180
References
Yung Y. L., Lee A. Y. T., Irion F. W., DeMore W. B. & Wen J. (1997) Carbon dioxide in the atmosphere: Isotope exchange with ozone and its use as a tracer in the middle atmosphere. J. Geophys. Res., 102: 10857-10866. [18] Yung Y. L. & Miller C. E. (1997) Isotopic fractionation of stratospheric nitrous oxide. Science, 278: 17781780. [19] Zaback D. A. & Pratt L. M. (1992) Isotopic composition and speciation of sulfur in the Miocene Monterey Formation; reevaluation of sulfur reactions during early diagenesis in marine environments. Geochim. Cosmochim. Acta, 56(2): 763-774. [26, 29] Zaback D. A., Pratt L. M. & Hayes J. M. (1993) Transport and reduction of sulfate and immobilization of sulfide in marine black shales. Geology, 21(2): 141-144. [29] Zadnik M. G., Specht S. & Begemann F. (1989) Revised isotopic composition of terrestrial mercury. Int. J. Mass Spectrom. Ion Proc., 89: 103-110. [31] Zaluski G., Nesbitt B. & Muehlenbachs K. (1994) Hydrothermal alteration and stable isotope systematics of the Babine porphyry Cu deposits, British Columbia: implications for fluid evolution of porphyry system. Econ. Geol., 89: 1518-1541. [3] Zawislanski P. T. & Zavarin M. (1996) Nature and rates of selenium transformations; a laboratory study of Kesteron Reservoir soils. Soil Sci. Soc. Amer. J., 60(3): 791-800. [29] Zeebe R. E. (1999) An explanation of the effect of seawater carbonate concentration on foraminiferal oxygen isotopes. Geochim. Cosmochim. Acta, 63: 2001-2007. [46] Zehr J. P. & Oremland R. S. (1987) Reduction of selenate to selenide by sulfate-respiring bacteria: Experiments with cell suspensions and estuarine sediments. Appl. Env. Microbiol., 53(6): 1365-1369. [29] Zeng Y., Mukai H., Bandow H. & Nojiri Y. (1994) Application of gas chromatography-combustionisotope ratio mass spectrometry to carbon isotopic analysis of methane and carbon monoxide in environmental samples. Anal. Chim. Acta, 289: 195-204. [14] Zhabina N. N. & Volkov I. I. (1978) A method of determination of various sulfur compounds in sea sediments and rocks. In: Environmental biogeochemistry and geomicrobiology. W. E. Krumbein (ed.). Ann Arbor Science, Ann Arbor, MI: 735-746. [26, 27] Zhai M., Nakamura E., Shaw D. M. & Nakano T. (1996) Boron isotope ratios in meteorites and lunar rocks. Geochim. Cosmochim. Acta, 60: 4877-4881. [7] Zhang B. L., Vallet C., Blanchard F., Martin G. J. & Combaut G. (1994a) Charaterization of galactose from Carrageenans by natural abundance site-specific isotope parameters. Phytochem. Analysis, 5: 291-296. [5] Zhang B. L., Quemerais B., Martin M. L., Martin G. J. & Williams J. M. (1994b) Determination of the natural deuterium distribution in glucose from plants having different photosynthetic pathways. Phytochem. Analysis, 5: 105-110. [5] Zhang B. L., Buddrus S., Trierweiler M. & Martin G. J. (1998a) Charaterization of glycerol from different origins by 2H- and 13C-NMR studies of site-specific natural isotope fractionation. J. Agric. Food Chem., 46(4): 1374-1380. [5] Zhang D. (1988) Nitrogen concentrations and isotopic compositions of some terrestrial rocks. PhD Thesis, The University of Chicago: 157 p. [16] Zhang J., Quay P. D. & Wilbur D. O. (1995) Carbon isotope fractionation during gas-water exchange and dissolution of CO2. Geochim. Cosmochim. Acta, 59(1): 107-114. [46] Zhang L., Chan L. H. & Gieskes J. M. (1998b) Lithium isotope geochemistry of pore waters from Ocean Drilling Program Sites 918/919, Irminger Basin. Geochim. Cosmochim. Acta, 62: 2437-2450. [6] Zhang Q. & Ding T. (1989) Analysis of the reference material NBS-123 and the atomic weight of sulphur. Kexue Tongbao (Chinese Sci, Bull.), 34: 1086-1089. [44] Zhang Q.-L. & Li W.-J. (1990) A calibrated measurement of the atomic weight of carbon. Chinese Sci. Bull., 35(4): 290-296. [40, 43] Zhang T. & Krooss B. H. (2001) Experimental investigation on the carbon isotope fractionation of methane during gas migration by diffusion through sedimentary rocks at elevated temperatures and pressure. Geochim. Cosmochim. Acta, 65: 2723-2742. [46] Zhang X., Nesbitt B. & Muehlenbachs K. (1989) Gold mineralization in the Okanangan Valley, southern British Columbia: fluid inclusion and stable isotope studies. Econ. Geol., 84: 410-424. [3]
References
1181
Zhang Y., Stolper E. & Wasserburg G. J. (1991a) Diffusion of a multi-species component and its role in oxygen and water transport in silicates. Earth Planet. Sci. Lett., 103: 228-240. [46] Zhang Y., Stolper E. & Wasserburg G. J. (1991b) Diffusion of water in rhyolitic glasses. Geochim. Cosmochim. Acla, 55: 441-456. [46] Zheng H. W. (1993) Calculation of oxygen isotope fractionation in hydroxyl-bearing silicates. Earth Planet. Sci. Left., 120: 247-263. [2] Zheng Y.-F., Metz P., Satir M. & Sharp Z. D. (1994) An experimental calibration of oxygen isotope fractionation between calcite and forsterite in the presence of a CO2-H20 fluid. Chem. Geol., 116: 1727. [46] Zheng Y.-F., Satir M., Metz P. & Sharp Z. D. (1999) Oxygen isotope exchange processes and disequilibrium between calcite and forsterite in an experimental C-O-H fluid. Geochim. Cosmochim. Acta, 63: 1781-1786. [46] Zhu X. K., O'Nions R. K., Guo Y., Belshaw N. S. & Rickard D. (2000a) Determination of natural Cuisotope variation by plasma-source mass spectrometry: implications for use as geochemical tracers. Chem. Geol., 163: 139-149. [31] Zhu X.-K., O'Nions R. K., Guo Y. & Reynolds B. C. (2000b) Secular variations of iron isotopes in North Atlantic deep water. Science, 287: 2000-2002. [31] Zhu X.-K., Guo Y., O'Nions R. K., Young E. D. & Ash R. D. (2001) Isotopic homogeneity of iron in the early solar system. Nature, 412: 311-313. [31] Zhu X. K., Guo Y., Williams R. K., O'Nions R. K., Matthews A., Belshaw N. S., Canters G. W., de Waal E. C., Weser U., Burgess B. K. & Salvato B. (2002) Mass fractionation processes of transition metal isotopes. Earth Planet. Sci. Lett., 200: 47-62. [46] Zinner E., McKeegan K. D. & Walker R. M. (1983) Laboratory measurements of D / H ratios in interplanetary dust. Nature, 305: 119-121. [30] Zinner E. & Crozaz G. (1986) A method for the quantitative measurement of rare earth elements in the ion microprobe. Int. ]. Mass Spectrom. Ion Proc., 69: 17-38. [30] Zinner E., Tang M. & Anders E. (1987) Large isotopic anomalies of Si, C, N and noble gases in interstellar silicon carbide from the Murray meteorite. Nature, 330: 730-732. [20, 25, 30] Zinner E. (1989) Isotopic measurements with the ion microprobe. In: New frontiers in stable isotopic research: Laser probes, ion probes, and small-sample analysis. W. C. III Shanks & R. E. Criss (eds.). U.S. Geol. Surv., Bull. 1890: 145-162. [30] Zinner E. (1998) Stellar nucleosynthesis and the isotopic composition of presolar grains from primitive meteorites. Annu. Rev. Earth Planet. Sci., 26: 147-188. [17] Zuleger E., Gieskes J. M. & You C.-F. (1996) Interstitial water chemistry of sediments of the Costa Rica Accretionary Complex off the Nicoya Peninsula. Geophys. Res. Lett., 23: 899-902. [7]
This Page Intentionally Left Blank
SUBJECT INDEX
This Page Intentionally Left Blank
1183
Subject Index
A absence of air absolute calibration absolute isotope abundance absolute isotope composition absolute isotope ratio
1006 506, 777, 890, 909 136, 902, 905, 964, 977, 978 706
165, 526, 533, 828, 874878,885,901,902,945 absolute measurement 706, 776 absolute sensitivity 853 absorption 13,15,104,156,282,378, 381, 405, 433, 446, 455471, 748-758, 760-783, 812,1040 absorption (on resin) 336 absorption lines 378 absorption spectrometry 47, 123, 907 absorption spectroscopy 380, 396, 399, 768-783 absorption spectrum 466, 776-779,880 abundance of artifacts 945 abundance sensitivity 96, 666, 694, 695, 697, 728,742,743,851,854 accelerating beams 642,655,657,664 accelerating voltage 123, 534, 656, 659, 727, 734, 735, 798, 799, 844, 847,860,862,864-866 acceleration of electrons 667, 791, 792, 859, 860, 864-866 acceleration of ions 661, 694, 791, 839, 841 acceleration potential 661, 685, 709, 804, 839, 843, 844 accelerator mass spec- 204, 609 trometer (AMS) acetic acid 40, 43, 48, 120, 125, 201, 231, 320, 512, 514, 517 acetic anhydride 165 acetone 16, 117, 339, 365, 427, 448, 451, 457, 458, 528, 529, 531, 572, 576, 577, 589, 613, 616 acid chemistry 240 acid digestion 147, 318, 319, 427, 450, 487, 571, 720, 895 acid environment 249 acid lake water 11 acid matrix 636, 637, 708 acid soil 590, 592, 1013 acid treatment 43, 56, 257, 266 acid volatile sulfur 570, 571, 574-577, 580, 603 acrylic fibers 490 acrylic matter 490
activated clay activation energies active fluorine / fluorination compounds active submarine chimney acetylene adiabatic energies adjacent peaks adjacent standard adsorbed air adsorbed moisture adsorbed water
581 581, 994 451, 526, 531 723
4,17 387 168, 848 133, 703 365 461 8, 9, 13, 47, 49, 53, 59, 72, 74, 298, 444, 493, 502 adsorption 4, 6, 8, 12, 14, 36, 72, 74, 123, 158, 189, 190, 214, 215, 318, 327-330, 338, 378, 430, 465, 471, 559, 564, 585, 591, 649, 809814, 819, 911, 926, 1016, 1019, 1034 adsorption (on resin) 330, 338 adsorption peak 811 aerated soil 586, 587, 590, 592 aerodynamic sizing 561 aerosol 132, 560, 561 aerosol (dry) 720, 725 aerosol nitrate 336, 376 aerosol particles 550, 561, 562 aerosol sulfate 376, 560, 561 Ag + exchange membrane337 Ag3PO4 crystal 482, 486-494 agar 200 AgPd alloy 1007, 1008 agricultural 2-5, 17, 19, 25, 29, 32, 180, 181, 305, 313, 320, 340 agriculture 813 air 3, 10, 18, 19, 25, 45, 95, 157, 174, 183, 189, 214223, 227, 274-280, 283, 287, 292-303, 307, 317, 341, 354, 369, 371, 380, 383, 458, 465, 475, 486, 488, 502, 557-561, 565, 566, 570, 611, 613, 616, 657, 815, 817, 858, 863, 870, 879, 895, 898, 899, 910, 945, 1013, 1014, 1018, 1024 air actuated pistons 284 air bath 20, 22, 25 air bubbling 207, 557 air conditioning 1039
1184 air extraction air flow rate air intake air leak / leaking air moisture air of breath air pollution air pressure air purification system air reference air sample air sample (modern) air sampler air standard air stream air temperature air toxics airborne emission airlock module airtight conditions airtight vial / container airway reactivity alcohol aldehydes algae algae strains algal assimilation algal culture algal population algal proteins algal species algal uptake algorithm alignment of magnet alignment of primary beam alignment / centering check aliphatic hydrocarbons alkali basalt alkali salts alkaline earth elements alkaline oxidation allanite crystal Allende chondrite altered basalt alumina tube aluminium foil boat
Subject Index 275 558,815 213 341,534 502,815,892 89,9Z 98,158 191 100,293 8O8 279,293,294,300,372 179, 272-297, 302, 303, 761,815,818 269 558 274, 280, 289, 302, 363, 364,367-372 281 473 951 381 50 215 215,220,224,328,347 378 4, 19, 25, 32, 159, 160, 163-166, 195, 196, 198, 221,858,896 159,163 650,1013 1014 1033 649 179 200 179 649 115, 168, 169, 439, 504, 634, 646, 779, 944-956, 974,979-990 851 675 458 233,234 85 559 31,867 586,588,593 66 721 1033 175 242
aluminum boat aluminosilicates alunite Amberjet (ion exch. resin) Amberlite (Na resin) Amberlyst ambient air amino acid
amino acid standard amino (acid) sugar ammonium carbamate ammonium citrate reagent ammonium contamination amorphous silica amphibole amplifier gain drift analyte matrix analytical artifact analytical chemistry analytical contamination analytical science analytical sensitivity analytical setup analytical statistics ancient (early) atmosphere ancient materials ancient organic matter ancient sediments Anderson pyrite andesite angular momentum anhydrite anhydrous environment anhydrous phosphoric acid anhydrous process aniline animal animal animal animal animal animal animal
cellulose feeding glue protein teeth tissue
190 40, 43 444, 445, 572, 1003 486-488, 492, 493 41, 333, 486-489, 493, 564 499 9Z 159,211,293,300 120, 158, 161-167, 173, 189-201, 230-236, 555, 565, 585, 588, 592, 595, 1019 199 320,344,346 24 485 190,320 43,56,1011 570,687,1017 634 708,723 123,687 154,155,639,794,935 274 820 130,139,149 204, 684 288 269,448 229 185 597 687 127 103, 748-750, 752, 754 553,568,572,573,599 255 27 246 338,339 179-183, 200, 378, 594, 595,1015 1030 177 529 200 183 182,186,191
Subject Index animal waste anion exchange chemistry anisotopic standard annual plants anode dark space Antarctic
1185 305 715
136 50L 516 791, 792 7, 80, 84, 292, 305, 310, 311, 376, 476, 681, 775, 815,885,889,945 564 Antarctic firn cores Antarctic nitrate aerosol 336 377 anthropogenic hydrocarbons anthropogenic sources 556, 557, 559 58, 68, 407, 446, 483-486, apatite 495, 496, 605, 619, 687, 688, 1013 apparent isotopic frac- 447 tionation 1 applied science 16,457 aqua regia 643 Aquadag paste 187 aquatic ecology 178,179,181,190,329 aquatic ecosystem 1014 aquatic plants 743 aquisition cycle Ar ion/atom interaction 792, 793 46 aragonite 269 Archean age 268,269 Archean chert Archean sedimentary 685,687 sulfides 716, 724 archeological artifact archeological samples 716, 724 716 archeological studies archeology 907 231,233-235,581 aromatic compounds aromatic deuterium 120 nuclei aromatic entities 235 aromatic formulation 118 (alkyl) aromatic frag233 ments aromatic hydrocarbons 163,198 aromatic positions 118 aromatic ring 118, 120 aromatic signals 117 aromatics 581,582,584 Arrhenius diagram 263 Arrhenius equation 266 Arrhenius expression 263 arsenopyrite 569 945,952 artifact scale Ascarite trap 334
Ascarite II ascorbic acid aspartic acid asphaltenes asteroid asymmetric stretching band Atlantic seawater atmosphere (the)
213,223,293 649 161,191,231,232,1019 581,582,584 229,236 779
137, 712 49, 54, 177, 180, 220, 257, 307, 309, 375-379, 389, 390, 417, 501, 538, 541, 552-559, 596, 607, 889 atmosphere monitoring 273, 290 atmospheric air 307, 364, 372, 501, 782, 894 272-304, 775, 778, 818, atmospheric CH4 951-956 atmospheric chemistry 390 376 atmospheric CO 180, 204, 213, 215, 21Z atmospheric CO2 220, 224, 272-304, 376, 779,894 atmospheric components 378, 383 349 atmospheric contamination 38O atmospheric data 189,379,380, 794 atmospheric gases 376 atmospheric H202 305 atmospheric load 291,775,778,951 atmospheric methane 9,10,416,457,502 atmospheric moisture 199, 357, 361, 392, 877, atmospheric N2 879,898 376,394,775 atmospheric N20 390 atmospheric N20 lifetime 309,329,336,343,376 atmospheric nitrate 335, 336, 376, 380, 455, atmospheric 02 1013,1015 309,376-381,384,388 atmospheric ozone 46, 98, 694, 789, 950, atmospheric pressure 1005-1008, 1012, 10181025 375,384 atmospheric species 343,380,899 atmospheric studies 3,25,47 atmospheric vapor 3,203,443,491,496,688 atmospheric water 106 atomic environment 703, 706,718,835 atomic weight 1008,1020 Au capsule 178 Australian tropical stream 1025-1027 autoclaves automated combustion 322,341
1186 automated device automated instrumentation automated Rittenberg analysis (ARA-MS) autosampler
Subject Index 178, 237, 238, 838 339 341
6, 7, 36, 165, 321, 341, 503, 570 Avogadro constant 924 921 Avogadro project Avogadro spectrometer 925 azeotropic distillation 3, 4, 31, 54,473-475, 479, 480 R-cellulose 507-521
B Ba(OH)2 reagent 211, 213 backbone procedure 419 back-diffusion 805 background concentra- 290 tion background condition 852 background contamina- 286 tion background contribution 537 background correction 743, 744, 852 background currents 669, 978 background determina- 852, 854 tion background drift 299 background emission 536 background estimation 743 background impurities 861 background information 314, 948 background interference 101 background level 287, 393, 503 background measurement852 background monitoring 470 background nitrogen 312 background noise 287, 677 background peaks 612 background pressure 873 background problem 500 background reduction 468 background samples 161 background scan 611 background science 89 background signal 742, 778 background water 287 backup filter 561 BaC12-NaOH reagent 211 bacterial experiment 650 bacterial methane 291 bacterial strains 335, 650, 994, 1012
bactericide bacteriological reduction bacterioplankton BaF2 window Balmat pyrite barite basalt basalt glass basalt standard basaltic magma basaltic melt baseline baseline baseline baseline baseline baseline baseline
conditions correction counts oscilations peaks separation
206, 211, 562 994
184 468,532 686,687 66, 43Z 465, 552, 570, 572, 573, 576, 600,602,1007 68, 125-127, 137, 350, 648, 710, 712, 904,1023 70,443,689 133,134,648 1027 1027 158, 161, 299, 614, 852,911 300 115 646 777 168,169 161, 168, 170, 173, 193 773 pro- 715
568, 579, 139, 903,
769,
174,
baseline slope baseline subtraction cedure baseline values 173, 181, 184 basic isotopic fractiona- 624 tion systematics batch reactor 647 batch reduction 9, 11, 22, 35, 36 beam defining slits 860, 861 Beer's law 768, 771, 774, 776, 780 beet root sugar 116 beet sucrose standard 476 bellows pump 457, 951 bellows reservoir (IRMS)242, 276, 297, 299, 300, 350, 357, 358, 838, 851, 879 275,410,416,950,951 bellows valves 120,231 benzaldehyde 17, 19, 44, 117, 118, 231, benzene 516,572,582,584 118,120,121,891,897 benzoic acid 315 Berthelot reaction 40, 43, 48, 61, 564, 896, bicarbonate 1018,1023 108, 161, 165, 175, 177, biochemical 181-192 103,820,907 biochemistry 202 biochemists 407,437,496,1030,1035 biogenic apatite
Subject Index biogenic mineral biogenic opal biogenic pyrite biogenic shell biogeochemistry biological activity Biological and Environmental Reference Materials (BERM) biological (re)cycling biological fluid biological fossils biological harmful radiation biological marker biological method biological process
1187 1028 52 578 1029 179,358,891,89G 899 684,717 948 306, 496, 628 3-5, 10, 17, 19, 25, 29, 32, 481 684
377 187 118 177, 181, 183, 450, 482, 625, 653, 1028 biological sample/ 31, 177, 179, 185, 186, matter 191, 194, 197, 201, 306, 313, 341, 344, 541, 555, 593, 594, 716, 1011, 1023, 1024 biological science 340 biological studies 474 biological tissue 188 biological tracer 2 biology 1, 1043 biomedical applications 756, 891, 898 biomedical fluid 11 biomedical science 122, 172 biomedical study 6 biomedicine 772 biominerals 1011, 1028 BioRad 328, 329, 331, 564, 640, 642 biosphere 177, 200, 272, 538, 552, 596 biosynthesis 116, 175, 182, 192, 196, 198-200 biosynthetic 108 biotite 348, 351-354, 357-359, 429, 441, 442, 445, 570, 687, 884, 890, 910, 1021, 1031 BiPO4 crystal 485, 486, 491 birch leaves 519 bird contamination 313 bird droppings 312 birds 172, 178, 181, 200, 312, 313, 594 bitumen 173, 578-583 bituminous organic sulfur572
black shale bladder stones blank
45, 46, 597-603 595 8, 9, 59, 127, 246, 251, 324, 328, 336, 348-350, 356-374, 410, 423-440, 452, 453, 459-461, 465, 471, 496, 497, 500, 501, 505, 550, 558, 569, 578, 640, 647, 648, 709, 729, 731, 736, 743, 748, 756, 800, 817, 818, 833, 852, 1050 blank contribution 328, 373, 374, 469, 822 blank correction 8 blank run 303 blank solution 165 bleaching 512,516,517 blend samples 729, 736, 738-740 block heater 11 blood (serum) 4, 29, 88, 98, 102, 182, 191, 197, 206, 594, 595 Boltzmann constant 105, 669, 867, 924 Boltzmann equilibrium 105, 114 Boltzmann populations 770 Boltzmann relation 105 bond stretching 385 bone 180, 186, 188, 1029, 1030 bone collagen 199, 1032 bone-dry 299 boric acid 125, 127, 128, 201, 314, 317, 322, 327, 345, 641, 645, 876, 903, 904 boric acid standard 142, 147, 148, 150 boron reagent 128 boron spike 146 borosilicate 12, 14, 15, 237, 240, 242, 532, 533, 612, 615, 805, 806 botanical origin 112, 116, 117, 121 botanical source 115 Boudouard equilibrium 498, 499 Bouguer-Lamber-Beer 771 law brachiopods 1011,1035 brackish water 639 breath (human) 2, 88, 89, 93, 95-99, 101, 102, 158, 159, 761-763, 778, 779,816 brines 4, 5, 7, 16, 1Z 2Z 29-33, 74, 130, 139-141, 269, 452,1018,1019,1034 broad peak 268,839 broad-leaved trees 515,520 bromination 490,491
1188
Subject Index
bromine pentafluoride/ 24, 30, 49, 50, 57, 400BrF5 472, 482, 483, 490-493, 524, 526-531, 542, 546, 818 bromine trifluoride/ 408, 409, 412, 417, 424, BrF3 442, 446, 447, 451, 490, 528, 531, 542, 545, 546, 924 bromoform 42 bromomethane 606, 619-621 bronze alloy 724 bronze artifact 724 bulk air 295 bulk meteorite 371 bulk tissue 190, 201 butane 231
C caffeine calcite lattice calcium carbide calcsilicate calculated equilibrium calculated isotope fractionation calibration
334, 897 573 17 78, 83 62 359
22, 30, 34, 111, 132, 274, 279, 280, 298, 300, 311, 337, 357, 398, 425, 439, 471, 476, 478, 498, 503, 599, 675, 681, 687, 745, 765, 779, 780, 784, 795, 796, 874-906, 909, 916920, 993, 994, 1000, 1006, 1028, 1032 calibration (one point) 336 calibration (two point) 980,981 calibration (mult. point) 336 478,599,795,982,983 calibration curve 906 calibration cycle 358 calibration line 881, 889, 890, 892, 893, calibration material 901, 917 35,392,505,506,794 calibration method 398,425 calibration standard 200 Calvin cycle 116 cane sugar 183 canopy air(forest) 864,865 capacitance-resistance filter 11, 18, 20, 25, 28, 63, 87, capillary (-ies) 97, 98,153, 155, 157, 168173, 276-299, 451, 502, 543, 546, 598, 611, 836,
838,853,953,1025,1040 342 290 273,274,304 59,174, 180, 248, 272, 321, 499, 602, 843, 894, 944, 947, 948, 956, 96Z 1040 carbon isotope labeled 896,1020 162,377,499,1040 carbon monoxide 489 carbon oxidation 8,29,36,343 carbon pyrolysis 7, 23, 7Z 342, 401-404, carbon reduction 408,429 carbon rod reactor 410 carbonaceous chondrite 229, 233, 720, 1023 carbonaceous meteorite 256, 260 carbonate 27, 31, 40-48, 56, 58-61, 64, 69, 74, 81, 125, 129, 132, 133, 136, 139, 180, 183, 188, 203, 211-223, 237-255, 258, 261-263, 280, 402, 405, 427, 444, 461, 483-485, 488, 495, 496, 497, 524, 529, 563, 572, 573, 579, 586, 590, 591, 610, 614, 618, 619, 625, 657, 681, 684, 707, 713, 723, 838, 852, 876883, 892-895, 902, 903, 910, 912, 1000, 1005, 1015, 1020, 1025-1030, 1035 carbonatite 64, 83, 241, 243, 618, 619, 876, 884, 892 carbonic anhydrase 26 carboniferous chomdrite 536 503 carbosorb CO2 trap 161, 164, 166, 230, 231, carboxylic acid 233, 234 514, 582, 584, 625 carcinogen 178 carniforous fish 255, 882, 894 Carrara marble 8, 9, 18, 29, 45, 46, 89, 92carrier gas 102, 132, 157-175, 225, 286, 296, 297, 303, 334, 341, 358, 419, 430, 433, 459-471, 501, 503, 546, 550, 551, 569, 571, 612, 618, 637, 638, 725, 782, 790-794, 817, 818, 853, 854, 879, 882, 1040 560, 561 cascade impactor
capsule crimping carbon cycle (the) Carbon Cycle Group carbon dioxide
Subject Index catalyst (a, the)
18, 19, 20, 88, 160, 298, 318, 319, 402, 501, 552, 598, 805, 941, 1016, 1021 catalyst poisoning 19 catalytic activity/action 17, 19, 22, 430 catalyze(d)/catalysis 7, 19, 26, 29, 75, 164, 298, 335, 344, 417, 499, 625, 1019 catalyzed graphite 333, 334, 340, 410 / carbon cathode dark space 791, 792 cathodic sputtering 789, 792 cathodoluminescence 674 causal law 699 cavity enhanced absorp- 777 tion spectroscopy (CEAS) cavity ring down 776 spectroscopy (CRDS) celestite 599 cell membrane 20O cellulose 200, 201, 474-480, 497522,883,896,1014,1030 cellulose filters 556,558,560,562 cellulose nitrate / nitrat. 508,509,1033 cellulose oxidation 499 cellulose standard 503,511,520, centrifugation 3, 4, 40-42, 46, 53, 56, 201, 335, 338, 486, 590, 617, 639 centrifugation-agitation 98 cycle centrifuge tube 188, 318, 550 ceramic frit 582 ceramic insulator 858, 862 ceramic material 500, 803 ceramic parts 863, 864 ceramic rods 864 ceramic spacers 858, 864 ceramic tube 165, 298, 334, 501, 1023 ceramics 788, 794, 859 certified isotopic compo- 648, 706 sition certified reference mate- 112, 828, 830, 881, 890, rial 891, 896, 897, 903, 907927, 935 certified site specific 112 ratio certified spike 820,827,829 certified values 733, 734, 740-742, 907, 945 chalcopyrite 548,599,716,966-968 changeover valve 836-838,843,851-853 channeltrons 670,695,697
1189 charcoal charcoal (activated) charge neutralization charge-balanced lattice charging cycle chemical bonding environment chemical conversion chemical environment chemical impurity chemical lifetime chemical memory chamical oxidation chemical reactivity chemical reactor chemical reagent chemical separation resin chemical thermometer chemical tracer chemical treated filters chemically homogene-
6, 23, 185, 338, 363 19, 338, 439, 462, 819 654, 664, 675 813 872 601 338, 541, 598 106 920 379 843 117, 197, 586, 593 547 907 4, 257 339. 342, 493 1031 557 557,558, 562, 565 1016
OUS
chemically scrubbed air 578 chemisorption 366, 813, 814, 818 chemistry 1, 152, 202, 464, 564, 673, 748, 808, 882, 929, 939, 995, 998, 1010, 1016, 1017, 1043 chemistry laboratory 496 chemistry of the phases 1006 chert 268, 269, 406, 1032 Chilean perchlorate 604 chitin 200, 1015 chlorinated solvent 7, 615 chlorine radical 291 chlorine trifluoride 24, 30, 49, 400, 404, 408C1F3 412, 422, 423, 431, 433436, 442, 446, 452, 455, 467, 468, 471 chlorite 41-43, 48, 49, 51, 351, 429, 445, 514, 1001, 1003, 1004, 1031 chloroform 44, 165, 197, 517 chloromethane 607-610, 613, 615, 616, 620, 621 chlorophyll 200, 710, 712 cholesterol 186, 191, 196-198 chondritic meteorite 724 chromatographic 74 analysis chromatographic column193, 195, 284, 494, 495, 546, 598, 602, 818, 854, 891
1190 chromatographic chromatographic ference chromatographic tope effect chromatographic
Subject Index filter inter-
416 855
iso-
168,171
peak
156, 168, 169, 297, 853, 854 171
chromatographic peak distortion chromatographic performance chromatographic separation chromatographic technique chromatography chromium
169 156, 195, 293, 335, 547, 557,621,1021 113, 125, 54Z 853, 856, 880 160, 169, 230, 232, 287, 836,853,854,856 Z 15, 623, 625, 627, 640, 644, 648, 650, 783, 829, 918 648 783 626,650
chromium chloride chromium doped... chromium isotopic fractionation chromium powder chromium reduced sulfur570-580 chromium reduction 571, 572, 576, 579, 588 Chromosorb 157, 441, 564 cinnabar 447, 720 circulating/pumping air 3, 275, 288, 815 citric acid 120, 610, 611, 618 classical fluorination 407 classical isotope frac309 tionation clay (mineral/sedim.) 15, 38-61, 422, 461, 590, 795, 813, 904, 993, 1016, 1019, 1032, 1034 clay fraction 39, 41 clay membrane 53 clay rich 22, 38, 53-56, 60, 139 clay separation 38-40 clay size 41, 42 clay standard 45, 47, 50 clean air site 274, 277 cleaning procedure 13, 320, 365, 709, 863 clear glass window 467 climate change 180, 183, 306 climate conditions 36, 497 climate monitoring 273, 274, 304 climate reconstruction 497 climate sensitive ar507 chive climate, affect on.. 272 climate, disrupt the.. 378
climate, influence on.. clinical applications clinical diagnosis clinical environment clinical investigation clinical setting clinical tracer study clinopyroxene Co catalyst CO gas laser CO standard CO2 contaminant CO2 fluorination technique CO2 laser
377 758, 784 88,98 88,779 98 779 6 68 1021 778,780 503,504 205,220,22Z 286 337
361, 362, 364, 436, 458, 459, 463, 464, 466-468, 472, 495, 532, 760, 764, 780 CO2-CH4 equilibration 1021 211-213,218,221,223 CO2-free atmosphere CO2-dissolved bicarbo- 1018 nate equilibration CO2-plant water equi- 473-481 libration CO2-water equilibration 3, 4, 17-35, 55, 66, 226, 408, 451, 476, 1018 coal 291, 559, 560, 568, 578583, 894 coal carbon 46 coal-fired power plant 559 coca leaves 172 cocaine 172 collector arrangment 534 collector array 697 collector assembly 858, 870 collector block 695, 697 collector ceramics 858 collector feed throughs 858 collector plate 847 collector slits 393, 664, 665, 675, 857, 859, 861 collector system 390, 393, 857, 859, 990 collision cell 695, 706, 722 collision gas 695, 708, 722 collisional interaction 793 colloidal graphite 643 colloidal Pt powder 494 colloidal silica 645, 646 colloidal solution 617 colloidal suspension 640, 645 color center crystals 782 color center laser 764, 772, 782 combined uncertainty 731, 733, 734, 738, 740, 745, 911, 916, 922
Subject Index 154, 157, 162, 165, 170, 182, 189, 192, 193, 225, 233, 256, 258, 293, 296, 298, 332, 333, 338-340, 342, 350, 356, 357, 366, 403, 505, 544, 545, 559, 597, 598, 602, 615, 817, 971-973, 982, 985, 988, 1051 combustion / oxidation 593 combustion / pyrolysis 258, 817, 818 combustion apparatus 294 578 combustion boat 162, 166 combustion catalyst 162 combustion chamber 182, 298 combustion efficiency combustion experiment 261, 266, 366 combustion furnace 224, 225, 294, 297-299 combustion gases 168, 365, 366 combustion interface 154, 165, 171 combustion mode 362, 366 combustion of sample on 326, 327 resin combustion procedure 601,818 171 combustion process 168, 174, 321, 341, 578, combustion products 615 162,165,173 combustion reactor 330 combustion reagent 162,302,358,544 combustion system 544,547,578,817,895 combustion technique / method combustion tube/vessel 70, 165, 298, 324, 330, 332, 336, 340, 342, 346, 544, 593, 618 common air 380 complete electrolysis 23 (in)complete fluorination 420, 443, 450, 461 complete oxidation 44 complex ecosystem 202 complex interactions 178, 180, 184 complex matrix 112, 154, 197 complex organic mixtures232, 233 compound peaks 17, 126, 129, 144, 146, 149, 152, 156, 168, 171, 284-286, 288, 290, 297299, 303, 341, 367-369, 371, 537, 593, 612, 621, 639, 641, 657, 799 compound specific 153, 156, 179, 182, 185188, 191-193, 202, 229236, 547, 596 242,1040, 1041 compressed air 492 computed statistics
combustion
1191 computerized environ- 121 ment condenser 189, 317, 467 conducting matrices 788 conduction heating 366 conodont 483, 495 conservation of momen- 749 tum constant combustion con-544 ditions 836 constant sensitivity 55, 215, 216 constant temperature (thermostated) water bath 944 consumer protection 24, 38, 58, 235, 236, 257, contaminants 285, 286, 306, 343, 373, 410, 412, 447, 448, 450, 465, 469, 496, 529, 531, 543, 546, 657, 743, 794, 817, 818, 863, 966, 982, 1040 3, 4, 8-10, 25, 28, 29, 36, contamination 47, 113, 185, 186, 190199, 211, 242, 250, 256, 257, 268, 278, 285-287, 303, 305, 312, 315, 320, 327, 331, 365, 412, 443, 450, 456-503, 524, 558, 565, 612, 625, 626, 649, 660, 707, 737, 743, 746, 779, 780, 794, 832, 847, 911, 936, 949, 950, 965, 969, 982, 983, 988, 1015, 1040, 1050 continental crust 307, 1034 continental rocks 139, 140 continental scale 183 continental tourmalines 1034 continental USA 292 continents 556 continuous flow MS (CF) 1, 7-9, 17, 22, 24, 29, 36, 168, 287, 335, 341, 343, 474, 497, 543, 544, 546, 564, 578, 595, 817, 944, 882, 972 399 CF analysis 637 CF apparatus 990 CF applications 505 CF approaches 972, 974 CF autosampler technique 973 CF combustion CF hydride generation 637
1192 CF loop CF measurement CF method CF mode CF preparation procedure CF purification method CF pyrolysis technique CF spectrophotometry CF sulfur isotope technology CF system
Subject Index 42 882, 971, 972 637, 971, 973 8, 17, 494, 501, 878 403 637, 638 476 333, 334 972, 974
676, 168, 190, 498, 593, 64Z883,891,972,973 77, 338, 390, 391, 394, CF technique / ology 459,544,626,879 CF ultracentrifugation 41 66 CF water reduction 333,468 CF with He 182,460 CF-GC-IRMS continuous fluorination 533 technique continuous isotopic spec-510 trum continuous monitoring 107, 192, 872 continuous sputtering 674,799 controlled climate boxes 1014 controlled environment 1014 conventional fluorination50, 410, 412, 413, 433, 443, 446, 447, 453, 458, 459, 461, 466, 681 conventional methodo- 471 logy 5, 8, 24, 30, 33, 120, 161, conversion (method) 165, 221, 223, 297, 313, 321, 329, 334, 335, 337, 340, 392, 410, 411, 417419, 423, 427, 430, 433, 436, 442, 446, 455, 474, 498-501, 508, 524, 541, 547, 558, 561, 556, 562, 563, 569, 570, 573, 574, 586, 590, 591, 595, 614, 775, 838, 847, 936, 1012 335, 339, 430, 501, 506, conversion (complete) 525, 586 conversion (incomplete) 9, 571, 847 conversion dynode/ 668, 679 channel plates conversion efficiency 303, 419 conversion equation 855, 894 (formula) conversion factor 940 conversion interface 1041 conversion relationship 136
conversion system 5 conversion temperature 13 cool plasma 713 Cooperative Air Samp- 292 ling Network 49, 69, 257, 258 copper oxide furnace 344 corn leaves 861 correction plate 706, 715, 717 correction technique cosmochemical fractio- 718 nation process cosmochemical implica- 122 tions cosmochemical signifi- 268 cance cosmochemistry 268, 709, 725, 912 cosmogenic mass-depen- 632 dent fractionation 391 cosmogenic nitrogen 448 cosmogenic studies 475 cotton plants 597-600 coulometric method 600 coulometric measurement 599-601 coulometry 677 counting artifact 373, 374, 658, 659, 667, counting statistics 671, 731, 734 7 Cr reactor 7, 23, 34, 36, 571, 572, Cr reduction 575, 576, 579, 588, 626, 650 181 crabs 33, 311 Craig correction 578 Cretaceous 879, 892 Cretaceous belemnite 482 critical nutrient 181 crop plants 50, 320, 322, 407, 408, cross contamination 439, 461, 539, 647, 784, 941, 953 503 cross-calibration 173, 578, 579, 584 crude oil 349 crust-mantle cycling 348 crust-mantle transfer 293, 294, 296, 297, 325, cryo-focusing 430, 433, 462, 465, 469, 470 427, 447, 452, 972 cryogenic distillation 444 crystal face 485 crystal growth 458, 484, 485 crystal lattice 349 crystal rocks environment
Subject Index crystal structure crystalline KBr crystalline ozone crystalline pyrite crystalline pyrrhotite crystalline residue crystalline rock crystalline solids crystallinity crystallization crystallographic orientation crystallographic site crystallographic structure Cu/CuO reagent Cu20 cubic pyrite CuO/copper oxide
CuO/NiO system CuO/NiO/Pt'catalyst' CuO / Pt' catalyst' CuO-Cu20 mixture CuO oxidation CuO oxygen reservoir CuO wires cw laser cyclic compound cyclic hydrocarbons cyclic structure cyclic sulfide cyclone pre-separator cyclosilicates cysteine cystine
1193 68, 70, 259, 805, 806 418 379, 383 582, 600 571, 574 165 140, 268, 349 259, 805 47, 51, 445 349, 1003 1004 810, 817, 1000, 1004 442 352, 358 75, 544, 545 549 58, 59, 68, 70, 175, 190, 258, 321, 339, 350, 352, 362, 363, 366, 401, 544, 616, 818, 973 162 162 162 59, 257, 407 494 59 170, 258, 350, 615 777 553 636 553 554 560 358 554, 555, 585, 595 554, 555, 571, 585, 595
D Daly collector/ion coun- 130, 144, 627, 695 ter data reduction 33, 36, 150, 169, 262, 311, 343, 548, 628-634, 644, 854, 944, 945, 947, 956, 958 data normalization 944, 956 dead biological material 188 Dead Sea 27, 31-33 dead time 95, 364, 667-669, 677, 679, 681, 728-743, 832, 923, 926
dead volume decripitation experiments deep earth fluids deep-sea spherule deer bones degassing degradative enzymes dehydration dehydroxylation dendrochronology dendroclimatology dendroecology denitrification denitrifier method denitrifying bacteria densimetry derivatization derivatization agents derivatization method derivatization procedure derivatization reactions derivatization scheme derivatization step derivatization technique derivatized derivatized compounds desiccator desolvating nebulizer desorption effects detectable contamination detection limit detrital deuterium labeled glucose Devarda's alloy Devonian diagenesis diagenetic diagentic opal
3, 9, 208, 275, 285, 290, 313, 455, 843 66, 358 142 632 1032 12-14, 23, 28, 32, 47, 49, 53, 57, 60, 204, 207, 349, 357,491,565,566,870 186 56,57,239,420,491,806, 1031 46, 49, 51-53, 422, 423, 445 508 508 507 308,335,336,344 306, 333, 335, 336, 343, 344 335 40,42,48,61 155, 160-166, 193, 19519G 233 15~ 160 162,164,166,165 197 160 194 198, 201 198 196, 199 160, 162, 163, 195, 199 215, 220, 328, 345, 485 698, 708-710, 713, 714, 716, 719-722 926 287 11, 65, 155, 755, 780 39, 187 101 314-324, 327, 328, 330, 332,346,347 354,579 139,140,199,201,271 38, 39, 42, 56, 484, 488, 602,689 56
1194 diamond
Subject Index 30, 257, 259, 265-267, 270, 271, 349, 350, 433, 467, 684 27O
diamondiferous diatreme diatary protein 200 dietary sources 181,198 diatom 534,535,1014,1015 diatom opal 56 diatom silica 52 diatomaceous silica 1015 diatomic species 608 dichloromethane 197-199,201,572,576 diethyl ether 158,618 differences in reactivity 444 different matrices 903 differential sputtering 659,793 diffusion 3, 16, 51, 189, 190, 264, 306, 314, 320328, 330, 345, 346, 375, 605, 809, 812, 813, 821, 828, 995, 998, 1001, 1004-1006, 1008, 1018, 1020, 1023, 1025,1026,1034 diffusion coefficient 604, 1004, 1020, 1023, 1034 diffusion constants 1004 diffusion data 1005,1024 diffusion disks 322 diffusion experiments 1004 diffusion method 189,190,330,344 diffusion packets 346 diffusion period 323,346 diffusion process 324 diffusion rates 810,812,1020,1031 diffusion technique 318,324,330,332,342 diffusional loss 72 digestion catalyst 319 digestion decomposition 124 digestion problem 832 digestion procedure 831 digestion rate 319 digestion time 319 digital conversion 115 diode laser 386, 388, 761, 772-775, 77Z 780,782,783 777 diode laser light injection direct combustion 230, 232, 322 direct diffusion 322 direct injection 11, 25, 650 direct sputtering 794 direct-sampling ion 789 source
dis-(non-)equilibrium fractionation dispersed graphite dispersion of ions dispersion range (IRMS) disposible syringe dissociation / ed
1002,1028
403 842 848, 849 213 26,45,362,369,378,384, 385, 389, 397, 408, 499, 553,860,1003 dissociation constants 158,239 dissociation process 370,384,463 dissociation reaction 26 dissolved inorganic car- 203-228 bon (DIC) dissolved organic car565 bon (DOC) dissolved organic mat- 23, 31, 314, 331 ter (DOM) dissolved salt 4, 5, 22, 23, 32, 33, 342 distillation 113, 116, 156, 189, 190, 313-327, 330, 332, 340, 345, 349, 356, 412, 414, 451, 455, 499, 531, 619, 620 distillation apparatus 316, 345, 567, 570-575, 586, 588, 591, 619 distillation equipment 316, 317, 620 distillation flask 189, 316, 317, 345 distillation method 54, 315, 320, 323, 328 distillation procedure 316 distillation processes 53, 54, 430 distillation products 116 distillation technique 54, 55, 60, 583 distillation time 315, 317, 320, 321 distillation unit 316,319 distillation yield 55, 116 divinylbenzene 125, 327 Dole effect 481 double filament 130, 138, 550, 643 double focussing 132, 549, 550, 677, 726, technique 738, 740-742, 788, 798, 804, 844 double focussing ion op- 697 tical system double spike calibration 626, 633, 650 double(isotope)spike 450, 623-651, 701, 709, 711, 713, 716, 719, 720, 724, 1012 dried atmosphere 49 Drierite 322, 345 drift correction 171, 300, 301 drift line 302 drinking water 305, 729, 736 dry (atmospheric) air 278-281,283, 287
1195
Subject Index 49,408,423 20, 25, 55, 75, 185, 208, 221, 448, 427, 451, 528, 529, 531, 533, 609, 613, 616,811,814,816 dry leaf samples 19Z 476,477,480 dry weight 200 drying agent 157,158,410 drying of air 280 drying (biol.) tissue 186,188 627 dual collector TIMS dual inlet (unspecified) 1, 6, 272, 282, 290, 300, 348,350 dual inlet (MS/IRMS) 1,4,21,22,157,168,259, 272-274, 276, 282, 299, 334, 441, 460, 539, 543, 544, 551, 564, 572, 578, 590, 875, 882, 883, 957, 972 dual inlet analysis 282, 288, 391, 971 dual inlet configuration 879 dual inlet instrument 299, 972 dual inlet measurement 273, 288, 882 dual inlet method 282,971 dual inlet mode 283,595,878,953 dual inlet system 272, 390, 391, 396, 498, 543,836-838,853,854 dual inlet technique 283,290 dual inlet valves 543 dual path gas cell 772 Dumas combustion 318,319,321,339 Dumas method 321 Dumas process 330,340 duoplasmatron 656,657,659,680 dye laser 758, 764,776 dynamic conversion 5 dynamic extraction 208 dynamic headspace 158 dynamic mass spectro- 1,262,270,348,349,362, metry 468 dynamic methods 5 dynamic measurements 697 dynamic mode 16,350 dynamic pumping 495 dynamic range 95, 363, 669, 670, 782, 795 dynamic reduction 5,16,35 dynamic systems 8,36 dynamic vacuum 528 dynamic zoom lens 69Z 727 system dry box dry ice
E Earth (the) earth materials earth science earth scientists earth's atmosphere earth's crust earth's early atmosphere earth's early history earth's surface
1,122,139,141,142,229, 235,875 626 122, 142, 152, 349, 471, 476,496,874 40O 375-377 307,625 448 540 38, 143, 236, 37Z 378, 384,625,648 85 188 17L 183 17L 183,189,191,306
East Pacific rise ecological laboratories ecological processes ecological samples / material ecological studies / 177-179, 183, 191, 198 research / investigation ecological systems 175, 178 ecologists 177, 179, 202 ecology / ecological 172, 202, 181, 596, 899 science economic geology 684 ecosystem 177-184,201,202,307 ecosystem analysis 180 ecosystem components 183 ecosystem cycles 329 ecosystem dynamics 183 ecosystem location 184 ecosystem processes 202 ecosystem seasonal va- 184 riation ecosystem studies 177, 182, 184 ecosystem survey 184 ecosystem temporal va- 183 riation ectotherms 483 effective filtering area 560 egg shell 1031 eigenstate 749 Elba tourmaline 903 electron emission current 841, 859, 864-866 electron gun 661-663, 675, 678-680 electron shell 654 electron trap 394, 398, 859, 860, 866 electronic environment 106 electrostatic interac729 tion
1196 elemental analyzer (EA) 29, 190, 202, 340, 341, 40Z 429, 542, 546, 54Z 569, 593, 595, 598, 602, 817,891,971,973 342 EA (automated) 341 EA (automated C/N) EA (high temperature) 8,340 305,319,341,971 EA-IRMS EA-CF-IRMS 358 476 elemental standard 24 electrolysis method 21,23,30 electrolysis of water 611,863 electrolyte electrolytical recoverage 401 electron bombardment 123 electron collector 840 electron impact frag396 mentation electron impact ion 390, 839, 840, 859, 907 source electron injection scheme654 electron multipliercol- 95, 132, 364, 367-369, lector/ion counter 378, 549, 550, 667-670, 677, 681, 684, 695-697, 832, 926 electronegative element 652, 653, 679 654 electronegative ions electronegative species 655 675 electronics drift electropositive element 653,654 electropositive species 655 7,695,727,728 electrostatic filter 538,597 element cycles 125, 127, 195, 284, 287, elution peak 299,854,855 859,865 emission controller 982,983,1027 empirical calibration 703, 704, 716, 717, 719, empirical correction 973 980-982 empirical derived law 699 empirical derived con- 504 stants 1036 empirical experiments 629 empirical fit empirical fractionation 484, 1035, 1036 empirical fractionation 671 law empirical increment me- 1031 thod empirical isotope correc- 979 tion empirical isotope ratio 150 empirical isotope ther- 1031 mometer
Subject Index empirical method empirical normalization empirical procedure energy dispersion energy filter
992, 993 706, 723 702 661, 694, 697, 742, 844 17, 662, 666, 676, 677, 679, 682, 742-744, 854 energy filtering 662, 666, 676, 677, 679, 682 energy window 662, 666, 677 enstatite chondrite 536 environment 38, 52, 143, 177, 186, 198, 383, 391, 541, 597, 605, 650, 746, 1008, 1027, 1032, 1038, 1039 environmental chemistry 907 environmental condi196, 625, 1014, 1035 tions environmental extremes 412 environmental factors 418 environmental geoche- 685 mistry environmental health 558 viewpoint environmental level 879, 880, 1048 environmental molecule 759, 779, 782 environmental monito- 928, 944 ring environmental N20 390 environmental para191, 202, 1028 meters environmental policy 920 issues environmental questions 175 environmental recon507 struction environmental research 343, 497, 782 environmental samples/ 149, 738, 739, 757 sampling environmental science 154,239 environmental seawater 1028 environmental study 178, 507, 778 environmental tracing 757 environmental variables 1035 environmental variation 508 environmental water 189 enzyme 26, 185, 595, 625 enzyme-mediated reac- 196, 199 tion equal size peaks 664 equilibration conditions 408 equilibration effect 504 equilibration method 5, 31, 55, 60 equilibration rate 25 equilibration system 17
Subject Index equilibration time equilibration vessels equilibrium bracketing equilibrium conditions equilibrium constant (K) equilibrium distribution equilibrium fractionation factor (c,) equilibrium isotope / ic fractionation equilibrium isotope effects equilibrium (chemical) processes equilibrium thermodynamics equilibrium time error propagation Eschka method Eschka mixture essential nutrient essential oils estuarine water ethane ethanol
1197 27 18 996 675, 867 91, 543, 994, 995 92 28, 34, 850, 993-1009, 1023,1025,1027 34, 375, 993-1000, 1004, 1006, 1009, 1018, 1021, 1024-1030,1035 2,375 375, 994 224 21,27 633,634,673,833,983 563, 565, 574-57G 587, 593,594 568,569,578,586 624 172 316-318 4,231 44, 55, 75, 101, 108, 109, 112-116, 127, 185, 189, 195, 275, 277, 320, 345, 412, 451, 457, 458, 516, 517, 529, 571, 582, 614, 816,891,896,897,908 16 175, 231 163-166
ether ethylbenzene ethyl-chloroformates (ECF) Etna basalt 903 Europa mission 758 eutrophication 305 evacuating air 20, 25, 206, 216, 217, 221 evaporation filament 131 evaporitic environment 1034 exchange equilibrium 1016 excimer laser 458,459,466,469 exhalation-inhalation 98 cycle exothermic dissolution 514 exothermic reaction 44,413 expansion bellows 7,949,950 experimental artifacts 27Z 385 experimental background237 experimental blank 337 experimental calibration 1001 experimental capsule 1001
experimental comparison932, 934, 935 experimental conditions 47, 121, 398, 491, 727, 738, 769, 1005, 1006 experimental data 16, 263, 264, 266, 267, 270,992,993,1004,1025, 1026,1036 experimental design 184,658,1010,1025 experimental details 780 experimental determi- 671,744,1010,1035 nations experimental evidence 464, 1009 experimental fractiona- 484 tion equations experimental geoche1006 mistry experimental isotope 992-1037 fractionation measurements experimental method 309, 337, 392, 992, 1005 experimental observa- 385, 386 tions experimental parameters 395 experimental petrology 1006, 1010 experimental procedure 20, 112, 923, 950 experimental results 266, 267, 727, 729, 741 experimental section 245 experimental set-up 668, 774,778, 994, 995, 1012, 1018, 1020, 1023, 1024 experimental settings 729 experimental study 32, 47, 52, 261, 309, 350, 756, 796, 992, 1001, 1003, 1009-1011, 1023, 1028, 1036 experimental system 193, 1017, 1020 experimental technique 95, 391 experimental time 114, 115 experimental work 376 experiments 13, 14, 52, 55, 74, 107, 208, 211, 218, 233, 237, 238, 240, 263, 264, 271, 322, 324, 353, 356, 358, 384, 385, 393, 398, 491, 511, 526, 568, 582, 604, 624, 625, 650, 708, 727, 731, 744, 816, 985, 9921037 explosive reagent 30 exponential equation 733, 743 exponential law 629, 631, 671, 699-704, 706, 716, 733, 788, 796, 799-802 exponential model 629, 730, 733, 735, 738, 743
1198 external external external external
Subject Index calibration contamination environment normalization
731, 735 320 1011 701, 702, 704, 706, 708, 711, 715-725 external standard 134, 138, 279, 598, 671, 701-703, 708-717, 723 extracting plate 860, 861 extraction efficiency ex- 217 periment extraction line 52, 66, 67, 70, 71, 76, 208, 357, 358, 452, 453, 491, 526, 1041 extraction plate voltage 395 extrapolation 995, 998, 1002 extraterrestrial bodies 747 extraterrestrial environ- 229, 230 ment extraterrestrial feature 235 extraterrestrial material 230, 361, 362, 374, 406, 535, 653 extraterrestrial O-iso680 tope analysis extraterrestrial organic 230, 235 matter extraterrestrial property 235 extraterrestrial samples 257, 267, 361, 371, 374 extraterrestrial source 230 region extraterrestrial water 25 extremely poisonous Cd 566 acetate
Faraday collector / cup
fat-free tissue fatty acid fatty acid standard Fe chemistry Fe isotope standard Fe oxidation feathers Fe-Cr-Ni alloy
71, 129, 132, 133, 276, 284, 364, 367, 369, 393, 550, 614, 626, 644, 645, 667, 669, 670, 679, 681, 685, 68Z 690, 694, 695, 69Z 698, 726, 728, 729, 739, 743, 832, 841, 847849, 853, 854, 857, 858, 870 191 120, 162-166, 170, 172, 175, 193, 195, 196, 198, 1014,1051 196,199 625 627,648 650,1013 181,186,594 795
feldspar feldspar crystal ferromanganese crusts fertilizer field environment field experiments filament filament filament filament
(metal) (Ni) (Pt) (Re)
filament (Ta) filament (thoriated Ir) filament (W) filament amendment proportions filament assembly filament centre filament characteristics filament conditioning filament contamination filament corrosion filament current filament electron density filament graphite technique filament length filament loading filament material filament protection filament resistance stabilizer filament safity filament stabilization filament status filament surface filament temperature filament voltage filament work function filter capsule filter discs filter glass tube filter material filter packs
56, 264, 265, 349, 356, 358, 403, 406, 429, 442, 444,445,461,530 256 714, 721 181, 305, 312, 328, 329, 344 1027 540,992,993,1005,1011, 1027,1028,1035,1036 148 621 494,868,870 130, 131, 549, 641, 643646,839,868 129,614,868 839 363,839,868 627 131 867 869 843 647 22 124, 129-131, 363, 614, 645, 646, 864 643 643 867 130, 149, 550, 561, 614, 62Z 639, 640, 643, 645, 650 147,646 864 868 866 868,869,871 870 645,843 130, 131, 368, 62Z 628, 647,857,859,867-870 865,869,871 494 331 346 521 186 346
Subject Index filter paper
215, 322, 476, 529, 558562,565-567,578 filter sheets 560 filter strips 327,328 filter system 214,560 fine-grained matrices 465 fish 178, 181, 200, 305, 483, 625,1035 fish bones 482,1035 fish flesh 490 fish otoliths 594 fish skin 594 fish tooth 484 flame sealing 215 flash combustion 601 flask air 293 flat-topped peaks 694,697,706,799 Floridian aquifer 567 flow-through reactor 421 fluid inclusion 2, 4, 8, 36, 62-8Z 269, 349,358,359,452,1030 fluor-apatite 425 fluoroscence techniques 747, 756 fluorescent instruments 202 fluoric acid 365 fluoride 124, 162, 442, 444, 454, 456, 457, 464, 467, 468, 614,617 fluoride(-rich) atmo461, 972 sphere fluoride compound 451, 453, 455, 471 fluoride reagent 420 fluorinated gases 946 fluorinated plastic 412 fluorinated polymer 168 fluorinated substance 107 fluorinating compounds 546 reactivity fluorination / ed 49, 58, 66, 400-472, 482, 484, 488, 490-492, 524534,626,816,1031 fluorination / ing agent 49,58,409,418,456,546 fluorination chemistry 471 fluorination cycles 413 fluorination line 412, 416, 418, 443, 453, 455,462 fluorination reaction 49, 405, 413, 418-421, 443, 453, 455, 460, 464, 468,471,491 fluorination reagent 403-409, 414-416, 419, 420, 423, 450, 457, 459, 461, 466, 468, 471, 524, 526 fluorination stripping 443,453
1199 fluorine
58, 107, 162, 400, 408, 411-413, 414, 418, 420, 421, 443, 445, 446, 452454, 461, 524, 532, 534, 545, 957, 972 400, 409, 421, 448
fluorine(-bearing) compounds fluorine frequency 114 fluorine gas 404, 409, 626 fluorine generator 413, 532 fluorine molecules 199 fluorine nuclei 113 fluorine oxidation 49, 401 fluorine pump 413, 418, 452 fluorocarbon contami448 nant fluorosilicic acid 445, 525 fly dilution 820-822 focus position drift 147 foil boat (small/tiny) 242, 337, 341 Fomblin 46 food quality control 172 food science 1, 103, 121 foraminifera 130, 1011, 1015, 1029 foraminifera shell 132, 141, 152 foraminiferal tests 144 formaldehyde 195 formic acid 588, 591, 638 forrest ecosystem 183 forrest soil 1013 forsterite crystal 1023 fossil fuel combustion 557, 559, 562 fossil hydrocarbons 235 fossil sediments 601 Fourier transform IR 779 spectroscopy Fourier transformation 47, 105, 114, 115, 380, (FTIR) 382, 763-765, 779, 782 fractional approach to 1000 equilibrium fractionated distillation 157 fractionation standard 325 fractionation factor 19, 20, 28, 34, 52, 55, 133, 141, 147, 243, 246, 247, 359, 360, 396, 408, 673, 796, 800, 802, 850, 975, 976, 992-1037 fragment ions 154, 390-397, 957, 958, 962, 986, 987 fragment pattern 396 fragment ratio 394, 395 framboidal pyrite 549 freeze drying 187, 188, 194, 314, 330, 339, 576, 1014
1200 frequency modulation spectroscopy (FMS) frequency window fresh leaf fresh water
Subject Index 773
114 474-481 1Z 33,314,324,331,335, 345,346,911,1013,1028, 1030,1035 1012 freshwater bacteria 1015,1030 freshwater diatoms freshwater ecosystem 306 freshwater environment 1012 freshwater fish 181 fruit juice 116, 481, 896, 913 fulvic acid 563, 588 fundamental standard 929 fusel oil 118
G 66, 441, 447, 548, 549, 552,568,570,1025,1030 359,429,441-443,466 garnet 171 gas calibration 8, 74, 76, 77, 88, 153, gas chromatograph(y) 154, 160, 161, 168, 169, (GC) 173, 182, 190-199, 223225, 233, 272, 282-284, 292, 321, 333, 350, 395, 430, 447, 448, 469, 471, 474, 476, 478, 494, 495, 508, 544, 546, 557, 609, 612, 613, 620, 621, 818, 853,854,880,891,971 GC technology 153-176 GC-C-IRMS 153-155, 157, 160-166, 168, 170-173, 175, 182, 192-194, 196, 197, 199, 201,350 gas equilibration 1Z 25,32,250,941,948 gas evolution technique 205, 206, 208, 210, 220, 226 29 gas permeable membrane 417, 418, 455, 81Z 1039, gas storage 1041 62-87 gaseous inclusion 55,224,225,615 gas-tight syringe 510 Gaussian distribution 168 Gaussian function 284,301,467 Gaussian peak Gaussian shaped signals 168 Gaussian source of noise 681 gelatinous mass 525 gelatinous silica 525 galena
genesis geochemical cycle geochemical tracer geochemistry
geographical origin geographical latitude geographical site geological analysis geological application geological community geological context geological environment geological fluid geological implications geological investigation geological material geological origin geological phenomenon geological processes geological recycling geological sample
geological settings geological studies geological time geology geoscience geostandard shale geothermal mudpot geothermal system geothermal well geothermal water geothermometer/metry
38,56,140,141,445 1036 152 103, 141, 154, 400, 496, 523, 596, 604, 605, 619, 622, 680, 685, 788, 789, 803, 805, 907, 992, 993, 1009,1036,103Z 1043 103,108,115,116,121 108 172 654 123,138,139,496 789 199 607 2 143 65 122, 123, 127, 132, 138, 139, 569, 570, 573, 614, 652, 664, 692, 717, 794 892 690 122,141,653 224 4, 8, 122, 128, 130, 132134, 139, 141, 256, 568, 569, 572-574, 609, 619, 707,709,713-725 78 124,141 62,820,1043 1 652, 788 597 1029 1028,1030 1029,1030 2 51, 52, 406, 444, 445, 993, 1000, 1031
geothermometric calibration glass ampoule glass boat glass fiber filter glass vial glass standard glass wool glass-ceramic standard
1000 3,7,35,894 215,216 186, 187, 189, 190, 214, 322, 323, 327, 341, 558, 560-562,610,617 1Z 26,161,198,214,215, 32G 1019 688, 788,798,802-804 293,294,564 80O
Subject Index glassy carbon Global Meteoric Water Line (GMWL) globalisation glove box glow discharge ion source glow discharge mass spectrometry (GDMS) glucose glutamic acid glycine glycolipids gold wire Gorleben aquifer grain fragments grain size
grain size distribution grain size fraction grain size range grain size segregation grain size separation granite graphite
graphite capsule graphite container graphite crucible graphite powder graphite reagent graphite reduction graphite resistance furnace graphite rod graphite slurry graphite standard grass-like plants gravimetric calibration
1201 76, 77, 334, 340, 494, 501, 503 15, 65 928 214, 215, 218, 223, 558, 568 789, 790, 791, 798, 803, 8O4 143, 788-804 101,200,201,896,897 232,235,884,89Z 1019 199,200,231,232 2O0 559 565 464 11-13, 23, 39, 54, 55, 68, 72, 240, 245, 253, 262, 265-267, 373, 418, 421, 434, 442, 463, 812, 996, 998, 1003, 1005, 1016, 1017,1020 39,41,42 68 41 41 39,41,42 39, 81-83, 85, 86, 264, 265, 395, 402, 431, 530, 568,569 24, 25, 30, 49, 271, 334, 340, 401-403, 430, 482, 483, 488, 489, 492, 493, 495, 529, 614, 636, 643, 64G 789, 808, 876, 883, 884, 895-89G 910, 967969,102G 1030 402 402 529 402,492,493 492,493 401,482,489 401 417, 423, 457 147, 614 494 517 631, 633, 635
gravimetric determination gravimetric (isotope) mixtures gravimetric preparation gravimetric techniques greenhouse effect greenhouse gas greenhouse gas lifetime greenhouse warming Greenland Ice Sheet Precipitation (GISP) Grimm-type discharge cell Grimm-type ion source groundwater
566, 567, 571, 574, 827 539, 540, 889, 920, 921, 924, 932, 941 627 570 291 272, 280, 377, 390 378 291 476, 887 798
789, 791 2, 36, 134, 142, 180, 181, 203, 204, 306, 312, 541, 567, 624-626, 649, 650, 757, 903, 904, 1034 groundwater chemistry 55 guanidine (hydro24, 30, 31, 33, 66, 76, 333, chloride) 334, 340, 408 gypsum 43, 552, 568, 572, 573 H H3+ contribution H3+ correction H3+ effect H3+ factor H3 + formation H3+ ion H3+ production H3+ signal H3BO3 reagent H202 oxidation H2S evolution peaks H2-water equilibration Haarhof / Van-derLinde function habitat hair halite crystal Hall probe halloysite halogen fluoride
21, 22, 33, 850, 853 842, 851, 1016 33, 34 848, 850 847, 848 841, 843, 850 22 21 345 565 581-583 3, 4, 7, 17-35, 55, 889 168
108, 177 182, 594 65 364, 861, 862 47-49, 53, 60 410-412, 418-420, 446, 451, 453, 455, 461, 471 halogen waste products 448 halogen 341, 410, 412, 604, 619, 621, 622, 655, 939 hard coal 46 hardwood 514 Harriott-type gas cell 772
1202 Hay(e)sep H-bearing contaminants headspace air headspace equilibration health and waste issues heavily contaminated heavy liquid heavy oil heavy water Heliobacter pylori hemicellulose hemoglobin herbaceous plants herbivore teeth herbivores heterogeneous pathway heterogeneous reaction heterogeneous sulfide heterogeneous system heterogeneous target heteronuclear 02 molecules hexane hexanediethyl ether hexapole collision cell Hg catalyst HgC12 poisoning HgC12 pyrolysis high detection sensitivity high dispersion IRMS high(er) mass resolution
Subject Index 157, 294 38 1013 226 319 313 42, 570 173, 581, 583 1,17 779 200, 201, 512-517 625 517 437 180 557 995, 996, 998, 1001 660 996,998, 1000 652 387 54, 191, 198, 502, 1051 198 696, 723 499 2 497, 499, 505 782 175 364, 551, 666, 676, 677, 693, 697, 790 334, 349
high temperature combustion high temperature oxi405, 529, 600 dation high-sensitivity analy- 795 tical technique high-sensitivity MS 259, 362 high-voltage discharge 499 reactor HNO3-Br2 digestion 574-576 HNO3-Br2 oxidation 594 Hokko beads 19, 22 holocellulose 508-520 homeotherms 483 homogeneous compound881 homogeneous crystal 690 homogeneous elemental 877 reservoir homogeneous gas 5 homogeneous magnetic 840, 844 field
homogeneous natural material homogeneous pathway homogeneous reaction homogeneous reference material homogeneous sample homogeneous seawater homogeneous solution homogeneous standard homogeneous surface homogeneously heating homogenized resin homonuclear diatomics horizontal momentum host lattice hostile environment HPLC HPLC peak HR-ICP-MS human activity human arbitrariness human error human hair human hands human health human inhalation human milk human nose human origin human population human samples human subjects human waste humans humic acid humid air humid locations humidity humidity conditions humidity control humidity levels humidity of air hydration hydration limited pro-
691 557 995, 999-1001 881, 882, 896, 900, 909 508, 512, 519, 586, 725 136 481 259, 511, 690 12 364 328 387 664 813 756 6, 7, 153, 1019 158 692, 693, 697, 710, 713, 715, 718, 720 390 914, 926 17 594 194 378 561 2, 10 19 2 291 11 173, 816 305 175, 177, 594 330, 563, 565 89, 280, 502 280, 281 185, 186, 280, 456, 473, 675, 1014, 1024, 1032, 1033, 1035 9, 1032 1039 172 3, 502, 558, 1014, 1032 26, 27, 31, 491, 806 26
cess
hydration reaction hydride contribution hydride formation hydride generation hydride interferences
26 681 743 636-639,647, 651, 744 536
1203
Subject Index hydride process hydride technique hydrides hydriodic acid hydrobromic acid hydrocarbon
hydrocarbon contaminants hydrocarbon oil hydrocarbon skeletons hydrochloric acid hydrofluorocarbons (HFCs) hydrogen atmosphere hydrogen background hydrogen diffusion hydrogen gas hydrogen peroxide
637,638 637,638 657,666,707 569,570,588,591 450 4, 44, 46, 54, 58, 59, 158, 159, 185, 193, 195-198, 230, 231, 361, 366-369, 401, 657, 794, 805, 814, 894,951,952 245 46 233, 236 195, 317, 332, 1018 377 559,1023 856 1006,1008 4,9,68,71,819,843,847, 850,1016,1051 44, 45, 48, 484, 489, 490, 559,579,618 236,582 307 204
hydrogenation hydrological data hydrological evolution of water 2,3,25 hydrological fluid hydrological processes 142 142 hydrological reservoir 35, 36, 313 hydrological sample 203 hydrological systems hydrological tracer study 2 hydrologists 306 1, 103, 189, 907 hydrology 35 hydrology laboratory 17, 19, 21 hydrophobic (Pt) catalyst 323 hydrophobic tape 538, 552, 562, 569, 683 hydrosphere 141 hydrothermal activity 38, 83, 443, 444, 1026 hydrothermal alteration hydrothermal apatite 407 hydrothermal aragonite 46 hydrothermal deposits 349, 1004 hydrothermal exchange 1011 hydrothermal fields 723 130, 139, 142 hydrothermal fluid 719 hydrothermal molybdenite hydrothermal quartz 68 hydrothermal reaction 140
hydrothermal rocking bombs hydrothermal system hydrothermal water /fluid hydrous iron oxides hydrous mineral hydrous phases hydrous pyrolysis hydrous silicates hydrous whole-rock samples hydroxybenzalacetone hydroxyl (OH) hydroxyl group hydroxyl hydrogen hydroxyl ions hydroxyl oxygen hydroxyl radical hydroxyl site hydroxyl water hydroxylamine hygroscopic hygroscopic fluoride hygroscopic salts hypersaline brine hypersaline water hypobromite oxidation hypochlorite hypothetical contamination
1007, 1025, 1026 56, 1028 32, 130, 139, 142, 149 56,625 2, 4, 8, 11, 53, 69, 687, 1020,1021 69 233 423,444 453 117 164, 377, 625 47, 116, 160, 265 46, 508 265 5O 291, 377, 484 112, 495 49 428, 429, 450 413, 425, 442, 444, 457, 491, 502, 506, 895, 923 451 631 32 11 317, 319, 321, 340 44, 45, 573 735
IAEA (International 34, 35, 111, 280, 300, 305 Atomic andEnergy 307, 310, 311, 504-506, Agency) 539, 599, 739-741, 772, 784, 853, 874-906, 910912, 917, 930, 931, 934, 943-946, 948, 951, 953, 981, 984, 1048 IAEA Advisory Group 892 Meeting IAEA Consultants' 886 Meeting IAEA experts meeting 883, 902 IAEA intercomparison 946 885 IAEA interlaboratory comparison IAEA laboratory 13 IAEA panel group 887 IAEA Panel Meeting 885
1204 IAEA reports IAEA technical contract ice (frozen water) ice bath ice core ice crystals ice slush ice trap Icelandic basalt ice-water bath ICP ion source ICP technique ICP-MS
ICP-MS experiment ICP-MS standard ICP-TOF-MS ideal standard igneous rock illite impermeable glass impregnated filters impurity interference impurity ions impurity isotopes
impurity of'.
BrF3 lignin (in cellulose) oxygen SeF4 sodium transition metals 80Se (in double spike) incomplete combustion incomplete diffusion incremental heating incubated soil incubation experiments incubation time incubation temperature independent isotope fractionation Indiana zinc industrial standard inert gas infrared (IR) IR (near) IR (mid)
Subject Index 887 903 3,186,187,313,75Z 875 621 29,290,336,772 275,277 208,215,216 207 648 615 132, 651, 692, 726, 803, 907 47 123, 127, 132, 138, 551, 621, 692-745, 788, 789, 794-800, 804, 820, 828, 923,926,937 729 648 692,693 259 39,64,404,648,716,721 39-51, 60, 80, 429, 445, 1030,1031 250 558 138 102 633,634 409 511 461 429 128 549 630 161, 171 325 349, 356, 870 649 1013 325, 346 1012 389 11-13 168 74, 157, 188, 410, 414, 455, 503, 545, 778 104, 753, 759-787, 915, 1036 396, 466, 783 766, 767, 780, 783
IR (far) IR absorption
382 156, 396, 399, 767, 880, 907 IR beam 756 IR detector 772, 778 IR emission 380 IR lamp 609 IRlaser 458-460, 463, 466, 467, 471, 772, 972 IR photon 766 IR radiation 377, 390, 458, 756, 779, 780, 783 IR spectrometer / try 380, 382, 759-787, 907 IR spectroscopy/ic 156, 391, 396, 399 IR spectrum 760, 769, 880 IR wavelength 458 inhomogeneous dust par-555 ticles inhomogeneous reference538, 900, 901, 911 material inhomogeneous sample 519 series in-house standard 598, 599, 934, 982 initial experiment 356, 384 injection current 773 injection flow rate 288 injection nebulizers 709, 722 injection valve 6, 7, 293 in-line combustion 333 in-line configuration 456 in-line gas purifier 286 in-line reaction traps 433, 461 in-line reactor 410 in-line sampling 1021 inner leaf section 481 inorganic acid 164 inorganic carbon 27, 189, 602, 892, 893 inorganic carbonates 1023, 1035 inorganic cation com812 plexes inorganic cation ex125 change inorganic chemistry 831 inorganic components 307, 579 inorganic compound 313, 315, 538, 541, 552, 553, 563, 568, 570, 574, 575, 578, 585-587, 590, 592, 595 inorganic conditions 496 inorganic dust particles 555 inorganic experiments 1018, 1023 inorganic forms 315, 585 inorganic fraction 321, 576, 592
1205
Subject Index inorganic isotopic reservoir inorganic material inorganic matter inorganic minerals inorganic nitrogen
186
189 8,177 579 179, 305, 312-316, 322, 323,325,330 inorganic phosphate 490,1013,1035 inorganic processes 482,624 inorganic residue 576 inorganic samples 605 inorganic sequestration 314 inorganic species 2Z 329,562,565 inorganic standard 547 inorganic substances 334,501 inorganic sulfates 547, 581, 585, 588, 590592,594 inorganic sulfides 547, 581 inorganic sulfur 552, 553, 562, 563, 565, 568, 570, 573-576, 579, 581, 585-588, 590, 592, 595 inorganic system 1007,1010,1024 insectivorous bats 198 insects 186,555 in-situ analysis 123, 405, 421, 444, 459, 462-466, 469, 471, 495, 535, 537, 548, 652, 661, 684, 688-692, 715, 725, 747,972 in-situ analytical capa- 652,691 bility in-situ chemical destruc- 378 tion in-situ chemistry 748 in-situ CO source 291 in-situ fluorination 464,465 in-situ laser sampling 459,482,483 in-situ measurement 50,433,652,723 in-situ method / techni- 1,380,483, 754 que / system in-situ micro sampling 467 in-situ photochemical 378 production in-situ Rb / Sr dating 757 in-situ reaction 465 in-situ reaction volume 405 in-situ sampling 56,466 in-situ stable isotope 725 studies in-situ 40K decay 269 instrument / ation 96-98, 107, 110, 115, 135, 154, 165, 175, 182, 188, 192, 202, 273, 274, 288,
instrumental instrumental instrumental instrumental instrumental
341, 376, 380, 383, 391, 395, 398, 450, 453, 535, 548-550, 594, 596, 614, 626, 627, 642, 644-646, 650, 653, 65G 663, 664, 674-679, 692-699, 706, 708, 710, 712, 714, 725740, 743-745, 747, 757, 758, 766-768, 772, 775784, 788-790, 794-798, 814, 824, 830, 835, 838, 843, 844, 847, 851, 856, 863, 870-873, 878, 880882, 907, 908, 912, 923, 926, 940, 946, 947, 950, 953, 954, 971, 972-975, 991,1038-1041 analysis 122,143,196 aspects 654,853 artifact 850,852 background726, 731, 736, 740, 743 (mass)bias 698, 701, 706, 708, 718, 725, 923 blank 709 calibration 779 characte- 728
instrumental instrumental instrumental ristics instrumental conditions 108, 123, 129, 397, 709, 710 instrumental configura- 93 tion instrumental corrections 33, 311 instrumental design 663, 803, 856 instrumental detection 65 limits instrumental (mass)dis-623-635, 644, 647, 650, crimination 651, 698, 701-703, 708, 716, 720, 721, 724 instrumental drift 124,284, 778,784,838 instrumental effects 675,833,848 instrumental factors 123,133,670 instrumental features 678 instrumental (isotope / 50, 141, 149, 152, 536, mass) fractionation 549, 652, 653, 655, 658, 659, 662, 666, 670, 672,677, 680, 687, 690, 699, 701, 706, 709-713 instrumental installa1038 tion instrumental knowledge 835 instrumental limitations 687 instrumental magnetic 108 field
1206 instrumental manufacturers instrumental memory instrumental methods instrumental offset instrumental parameters instrumental performance instrumental procedures instrumental resolution instrumental response instrumental sensitivity instrumental setting instrumental setup instrumental stability instrumental step instrumental sources instrumental techniques insulated ice chest insulating materials insulators (samples)
Subject Index 155, 165, 954 709 404, 450 427 671, 674, 776 697 123 771, 779 842 725, 729, 824 738, 832 674 778, 784 743 726 141, 143, 152, 460 3, 313 789, 794, 795 654, 675
interaction between:
adsorbent-nitrogen 158 analyte-stationay 171 phase environment-orga177 nisms fluid-mineral 687 fluid-rock 38, 349 gas-filament surface 843 ions-He or H2 651 laser-mineral 459 laser-sample 433, 436, 458, 461, 464 light-atoms 749 organisms 177, 191 plasma-organic matter 46 rotation-vibration 770 seawater-ocean crust 139, 1033 sediment-water 141 solute-water 31, 1018 solute-stationary 160, 168 phase surface water-ground- 204 water various carbon species 203 water-rock 142 interaction with high 863 speed ions intercalibration 132, 273, 290, 723, 904 intercomparison 280, 952-956 intercomparison material506, 539, 853, 881, 945, 949, 951, 956 intercomparison program280, 303 intercomparison results 505
intercomparison studies 290, 850, 898 intercrystalline fractio- 50, 51 nation interhalogen fluorides 400, 404, 409, 410, 412, 415, 442, 454, 455 interhalogens 401 inter-laboratory cali34, 273, 290, 311 bration inter-laboratory stan646 dard interlayer water 47, 49, 50, 52, 53, 55, 60 internal laboratory 408, 671, 731, 900, 1027 calibration internal normalization 149, 629, 699-703, 706, 718 157, 193, 195, 210, 241, internal standard 278, 408, 476, 478, 481, 512, 646, 794, 796, 882, 898, 900, 906, 1051 international standard 34, 138, 200, 259, 311, 360, 476, 478, 488, 599, 605, 606, 647, 759, 772, 784, 850, 855, 880, 903, 929, 935, 944, 953 international system of 907, 909, 912 measurement international system of 914 units 1043 internet technology interplanetary dust par- 661 ticles interstellar environment 229, 236 interstellar space 230, 236 interstitial water 53, 55, 139, 140 intracellular structure 682 intra-crystalline 423, 443 intra-mineral isotope / ic 444, 445 fractionation inverse linear relation 325 iodine pentoxide 500 iodomethane 609, 611, 613, 616, 618, 621 ion beam dispersion 550 ion beam double focus- 726 sing ion beam drift 857 ion beam monitoring 990 ion chemistry 91, 97, 856 ion (exchange)chromato-125, 127, 129, 158, 195, graphy 198, 337, 339, 552, 614, 716, 720, 721, 724 ion collection statistics 288 ion exchange resin 125, 129, 155, 313, 317, 325-338, 482-492, 499,
1207
Subject Index
ion microscope ion ion ion ion ion ion ion
optica crossover optical elements optica properties optica system peak repeller source
ion source compound ion source conditions ion source design ion source geometry ion source material ion source memory ion source parameters ion source plates ion source pressure ion source schematics ion source setting ion source technology ion source vacuum ionic contaminants ionic strength ionization cage ionization chamber ionization efficiency
ionization filament ionization sources ion-laser pump irradiation of protons irreproducible values isobaric contamination isobaric interference
564, 565, 614, 617, 637641, 646, 714, 1019 652, 656, 657, 663, 664, 675, 677, 678, 680 664 657 844, 851 694, 697 96 860, 964 95, 123, 127, 135, 137, 156, 162, 165, 168, 171, 175, 199, 283, 287, 289, 299, 302, 312, 344, 364, 368, 371, 374, 378, 390, 391, 394, 395, 398, 407, 466, 501, 539, 543, 546, 547, 593, 613, 642, 643, 657, 694, 766, 790, 796, 798, 803, 836, 838-843, 847, 848, 851, 857, 858865, 924, 926, 953, 957, 971, 990 124, 137 22, 368, 374, 395, 843, 859 842 873 125, 127, 131, 138 408, 446 173, 174, 395, 398 642 798, 841, 854 839 361, 873 789 636 487, 965 226, 323, 326, 336, 346, 610,617 859, 860, 863 22, 393-395, 397, 398 131, 138, 144, 146-149, 152, 303, 494, 535, 643, 645, 650, 654, 692, 694, 725, 746-748, 859, 860 130,131, 550, 643, 861 144, 926 782 114 123 149 101, 145-148, 156, 162, 168, 337, 343, 446, 537,
isobaric peaks Isogeochem (list) iso-propanol isothermal equilibrium isotope change monitoring isotope compositional bracketing isotope dilution / IDMS
545, 561, 635, 644, 664, 666, 695, 707, 710-719, 723, 731, 798, 800, 822, 832, 847, 853, 926, 958 536 19, 34, 1043, 1044 199 216 1014
645, 708, 743, 841,
996
88, 125, 540, 632-634, 637, 727, 729, 736, 738740, 745, 820-834, 907, 918, 928-930, 936, 937, 939 isotope (ratio) drift 133, 300, 478, 647, 947, 948, 952 isotope exchange kinetics496 isotope equilibration 20, 21, 27, 973 isotope/ic fractionation 2, 25, 34, 51, 54, 70, 74, 108, 122, 124, 128-130, 138, 141, 146, 147, 156158, 160, 171, 182, 189191, 193, 196, 199-202, 245, 249, 307, 311, 317, 320, 327, 329, 375, 384, 396, 402, 406, 407, 430, 436, 444, 447, 460, 463466, 471, 473, 483, 484, 497, 515, 525, 540, 541, 546, 555, 566, 586, 595, 603, 611, 6 1 3 , 624629,634, 649, 650, 655, 672, 701, 706, 716, 718, 719, 722, 724, 772, 813, 815, 889, 894, 895, 948950, 966, 996, 998, 1006, 1009, 1015, 1017, 1019, 1028, 1029, 1033, 1039, 1045, 1047 isotope fractionation 1045-1047 calculator isotope/ic fractionation 113, 375, 454, 781, 895 effect isotope/ic fractionation 28, 359, 992, 993, 1004, factors 1005, 1036 isotope hydrology group 622 isotope memory 455 isotope monitoring 707 isotope ratio measure909 ment science
1208 isotope ratio monitoring 350, 441, 460, 468, 495, 597, 598, 713, 854, 856, 944, 953 isotope salt effect 4, 31-33 isotopic calibration 34, 124, 154, 155, 171, 311, 503, 506, 526, 874906, 1025, 1029, 1036, 1051 isotopic diffusion 836 isotopic disequilibrium 1003 isotopic enriched tracer 190 isotopic equilibrium 21, 22, 29, 31, 62, 89, 371, 372, 471, 541, 559, 996, 1002, 1011, 1035 isotopic fingerprint 108, 121, 329, 723 isotopic fractionation 150, 708 artifact isotopic fractionation 1036 interaction isotopic fractionation 376 processes isotopic geothermome- 406,993,1031 try isotopic labeled drugs 682 isotopic labeling 540 isotopic lifetime 379 isotopic mass balance 445,477 isotopic mass fractiona- 677 tion drift isotopic peaks 77,670,742 isotopic standard 35, 123, 124, 132, 133136, 155, 380,476, 533, 604, 606, 627, 701, 708, 710, 714, 935, 945, 946, 949, 951, 967, 977-984, 1051,1052 isotopic thermometer 1030,1031 isotopic tracer 1,140,142,152,161,182, 306, 314, 316, 338, 341, 343, 507, 540, 621, 623, 649,856,947 isotopically depleted 2, 5, 51, 56,64, 68, 72, 158, 191, 199, 200, 245, 249, 251, 254, 273, 280, 387, 389, 540, 803, 889, 891, 892, 899, 900, 990, 1035 isotopically enriched 2, 6, 17, 35, 49, 72, 92, 107, 125, 128, 133, 158, 172, 178, 183, 188, 199, 201, 241, 243, 249, 251254, 260, 264, 308, 323, 328, 380, 382, 387, 397, 473, 474, 477, 481, 491,
Subject Index 540, 543, 566, 579, 621, 631, 634, 655, 714, 719, 723, 729, 75Z 780, 820834, 880, 888, 889, 891, 896-900, 903, 921-924, 930, 932, 935, 939, 940, 977, 989, 990, 992, 1001, 1004,1012,1020 182
isotopically heterogeneous isotopically homogene- 549,648,898,904,982 OUS
isotopically inhomogeneous isotopically labeled isotopomer
isotopomer bands isotopomer calibration technique isotopomer detector isotopomer determination isotopomer distribution isotopomer ion mass spectra isotopomer overlap isotopomer ratio isotopomer triplets
566 52, 682, 1004, 1049 89-95, 108, 110, 111, 168, 343, 344, 380, 387, 391394, 398, 399, 759, 766, 768-770, 772, 775, 778780, 782, 784, 786, 854 779 392 778 396 386 95 92 392, 396-399 849
jarosite 572 Johnson-Nishita reduct. 573, 586, 588, 591, 593, 594 juices 2, 4, 29, 35
K Kalrez kaolinite KBr KBr coarse scrap KBr furnace KBr reactor Kel-F kerogen
468 15, 39-51, 56, 57, 60, 61, 423, 429, 443-445, 461, 1021, 1029, 1031 410, 414, 418, 419, 421, 423, 435, 436, 439, 461, 470, 620 418 456 414, 418, 419, 456 411-416, 455, 462 58, 574, 579-581
Subject Index kerogen sulfur kerosene K-factor
579, 580 54 730, 740, 828, 829, 919, 920, 924, 940-942 K-factor drift 832 Kiba (reducing)reagent 581, 594, 595, 597, 598, 600 573,576,579 Kiba reduction 595 kidney stones kinetic analysis of data 258 810,814 kinetic diameter 239,264,813,994,1035 kinetic effect 550, 661,676, 681, 792, kinetic energy 793,844,859 925,926 kinetic gas theory 994 kinetic information 2, 94, 156, 160, 175, 234, kinetic isotope effect 235, 243, 247, 252, 339, 375, 387, 430, 450, 541, 548, 993, 1003, 1019, 1023 kinetic isotope / ic frac- 51, 244, 291, 309, 375, 649, 994, 1001, 1018, tionation 1021-1025,1030 337,994,1003 kinetic process 403 kinetic reaction 388,1026 kinetic studies 248 kinetically induced fractionation 249 kinetically related effects 316 Kjeldahl apparatus Kjeldahl determination 350 306, 315, 316, 318, 319, Kjeldahl digestion 340 Kjeldahl distillation 306,314-316,318,350 306,321,339 Kjeldahl method 319,321 Kjeldahl procedure Kjeldahl technique 315,350 659,674 Kohler illumination 659,674 Kohler method 20O Krebs cycle
L labeled compound 20, 103, 168, 178, 1049 labeled fatty acids 175 labeled food products 115 labeled gas 393, 398 labeled inorganic nitro- 322 gen labeled position 392, 396, 397 labeled storage room 418 labeled water 1002
1209 labeled working standard labeling studies laboratory atmosphere laboratory environment laboratory experiment laboratory intercomparison laboratory standard
laboratory tissue lamproite landfills large-scale ecosystem laser ablation
laser analysis laser applications laser bandwidth laser based fluid inclusion extraction laser beam
laser laser laser laser laser
chamber characteristics desorption determination diffraction particle sizer laser drilling laser emission laser extraction laser energy laser firings laser fluorination laser frequency laser heating laser laser laser laser
intensity interaction ionization light
393 540 285 779 376, 389, 647 290, 850 22, 28, 34, 243, 311, 357, 408, 476, 478, 648, 719, 880, 882, 898, 900, 905, 906, 929, 949, 966, 967970 457 269-271 209,210,212,775 179 1, 8, 350, 408, 409, 466, 495, 542, 547, 551, 570, 692, 711, 715, 723, 725, 756,788, 789 50,464,482,495 459,468 749, 752, 753 76 364, 374, 405, 459, 461, 462, 466, 467, 532, 533, 548, 715, 748, 751, 752, 755,772, 789 362,366,367,373 458,466 757 5O 42 76 362, 780 361, 364, 453, 458, 459, 532 458, 459, 463, 464, 466, 468,471 548 50, 405, 407, 412, 447, 458-461, 464, 465, 471, 495,537 777 50, 350, 358, 362, 400, 403, 407, 421, 495, 532, 972 772,777 436,458,459,461 747, 755, 793 754,772,776,777,914
1210
Subject Index
laser methods laser methodologies laser microanalysis laser noise laser peak laser photoacoustic spectroscopy laser pit laser power laser (micro-)probe laser pulse laser pyrolysis laser radiation laser laser laser laser laser laser laser laser
reaction craters resonance sampling selection source spectrometry spectroscopy system
laser technique laser tuning laser wavelength lateritic soil lattice atoms lattice binding energy leading peak lead-salt (diode) laser leaf/leaves leaf discs leaf length leaf litter leaf margin leaf material leaf mesophyll water leaf petiole leaf punch leaf pyrolysis leaf samples leaf segment leaf surface leaf temperature
44Z 461,537 403 430, 453, 461, 465, 469, 471 772,773,776 467 780 496 414, 465, 495, 748, 753, 780 359, 532, 782, 971, 972, 979,982,985,993 495,752 408 76, 405, 458, 467, 776, 777,780 464 746, 747,756 40Z 459 465,471 772,773,782,783 784 772,780,782 50,7Z 364,386,458,459, 466, 469, 495, 496, 725, 783 50, 361, 400, 403, 458, 483,495 775 458, 459, 463, 466, 467, 748,773,775 42,43 792 792, 793 168,169 760-763, 765, 772, 774, 783 180, 182, 183, 473-475, 478, 479, 514, 515, 518, 1029,1032,1033,1035 475, 479,481 481 188 481 514 474 474 474, 475, 479, 481 479-481 474, 476-478 475 481 473
182,183,473 1014 273, 473-481, 497, 1014, 1029,1032 164 Lewis acid 712 Li isotope standard 129,133,136 Li2CO3 standard LiBO2-V205 flux mixture 350 496 life science 169 lifetime of GC column 124, 132, 654, 693, 695, light elements 698, 702, 72Z 853, 911, 944,992,993 508, 509, 511, 512, 514, lignin 516,517,520,521 196,197 lignin-phenols 241,876,884,893,904 limestone 771 linear absorption 107 linear amplifiers 629 linear approximation 842 linear behavior linear/non-linear beha- 549 vior linear calibration curve 795 647 linear change 294 linear column 15,628,699,851 linear correction 363, 368, 478, 518, 703, linear correlation 704,707 671,868,960 linear dependence 795 linear dynamic range 629,731 linear equations 628,629 linear fractionation 369,848 linear function 299,300,302 linear interpolation 671, 699, 700, 702, 730, linear law 733 941 linear measurements 734, 735 linear model linear molecule density 780 344,385,390 linear molecules 479,515,518,539 linear regression 264, 396, 420, 702, 842, linear relationship 851 665 linear representation 782,853 linear response 903 linear shift 582 linear temperature increase 22 linear time-correction 302, 395, 398, 551, 669, linearity 774, 843, 851, 854, 873, 875, 931, 933, 936, 942 leaf tissue leaf transpiration leaf water
1211
Subject Index linearity limits of the IRMS Lipari obsidian lipids lipopolysaccharide liquid air liquid chromatograph liquid chromatography liquid reagent liquid-solid chromatography lithium aluminium hydride (LAH) LAH reduction lithogenic material lithogenic S compounds lithology lithosphere living material living organism living tissue liver tissue local ecosystem local standard long-term performance drift long-term storage Lorentzian rule low ionization sensitivity low-pass filter low-temperature interaction low-temperature environment lunar basalt lunar fines lunar grains lunar materials lunar nitrogen lunar regolith lunar rocks lunar samples lunar soil lunar spectra Luer ground joint Luer lock
477 903 182, 189-198, 201, 511, 517,1012-1014 200 402,421,429,450 193 153,199,201,1019 412 584 16, 572 23, 572 568 569 62 538,552, 568, 596 1011,1036 1 197 594 306 784,905,906 906 187,211,224,227,228 844 144 774 683 236 625 443 361,366,373,426 977 373 361 526,648 526,977 267,688 382 206,208,215 207
M Madan galena magnesia reagent magnetic interactions
65 485 104, 106
magneto-optical trap Maiella limestone main peaks mais germ oil major earth reservoirs major element chemistry major peak mammals Manhattan Project mannitol mantle (Earth) mantle epithelium mantle materials mantle petrologic work mantle (derived) rock mantle source marble standard Mariana trough marine air marine diatom marine ecosystem marine limestone Mars (planet) Mars mission Mars Pathfinder Mars Polar Lander Martian atmosphere Martian crustal fluid mass balance mass balance approach mass balance calculations mass balance consideration mass balance correction mass balance equation mass balance studies mass bias drift mass bias correction mass centered peak mass discrimination
mass dispersion mass filter mass filtered ions mass fragmentation analysis mass interference
759 903 608 172 1028,1036 795 199,583 200,483,1030,1035 1 147,200,722 140,141,683,1034,1036 1011 349 349 140,349,648,682,688 183 241 85,137 561 1015 172 892 688 758 913 758 688 687 224,335,432,477,574 325,474 53,56,196 55, 59 852 8, 371, 484 624 702, 710, 713, 723 699, 701, 704, 709, 715 636 94, 96, 122, 123, 132, 133, 135, 156, 158, 160, 171, 551, 698-704, 706, 708, 710, 713, 714, 716-718, 720, 721, 723-725, 737, 804, 923, 924, 978 132, 175, 664, 665, 679, 697, 851 97, 657, 658 95 394, 399 38, 58, 101, 158, 335, 361, 367, 409, 450, 508
1212 mass peak
90, 128, 149, 150, 644, 645, 728 mass resolution 367, 657, 664-666, 675, 677, 679-682, 685, 693, 697, 706, 727, 739, 798, 800 mass resolution capabi- 679 lity mass resolution require- 667, 675 ment 109, 367, 536, 664-666, mass resolving power 766, 771, 788 832 mass spectral interferences 795, 804, 835 mass spectrograph 34 mass spectrometer tuning 89, 95, 98, 146, 173, 607, mass spectrum / tra 636, 639, 641, 646, 664, 735, 839, 847, 873, 958, 972, 975, 986, 987 540, 541 mass-(in)dependent conversion mass-dependent isotope 25, 148, 623, 624, 710 fractionation mass-independent iso- 394, 534, 540, 992, 994 tope fractionation 378, 383, 804, 844 Mattauch-Herzog instrument (MS) Mattauch-Herzog geo- 679 metry 115 mathematical window functions 2, 36, 197, 261, 343, 492, matrix (a, the) 655, 673, 716, 721, 724, 729, 935, 946 681 matrix correction 788, 793, 828 matrix dependence 831 matrix difference 125, 1 3 8 , 395, 5 3 5 , matrix effect 570,657,672, 673, 6 8 8 , 690, 708, 710, 723, 726, 793, 794, 903, 947 133, 768, 939 matrix element 158 matrix interaction 789 matrix-matched standard 795 matrix matching 17 matrix materials 556 maximum conversion 465 measurable isotopic fractionation 644, 647, 743, 744, 985, measurement cycle 990
Subject Index measurement statistics measurement transparency medical applications medical field medical research medical science medical study medicine membrane desolvation membrane filtration membrane valves memory (effect)
memory issue memory signal memory size mercaptan mercuric cyanide mercury (native) mercury(II) chloride mercury pistons mesh size of exchange resin mesquite leaves metabasalt metabolism metal boat metal capsule metal catalyst metal fluoride metal hydrides metal oxide catalyst metal reactor metal reagent metal-fluoride melts metasediments metasomatic alteration metasomatic products metastable species lifetime metastable state lifetime metavolcanic meteoric sources meteoric water
690 822,833 944 896 176 1,122 98,176, 716 103,907 698, 713 556,563-56Z 572,590 836 4-9, 22, 23, 25, 29, 98, 14Z 159, 285, 286, 288, 294, 300, 313, 321, 452, 497, 498, 500-505, 543, 561, 585, 598, 60Z 608, 709, 722, 729, 743, 772, 784,816,842,852,941 114 852 114 553,554,585 333 319,421,461,499,720 312, 330, 497, 499, 500, 504,505 838 332 191 79,351 177,191,200 190 1006,1026 24 416,420,442 798 973 8 4,5 464 78-81, 84, 256, 269, 271, 351,354,356,357,360 348,359 351 386 386, 387 79, 80 63 15, 36, 682, 1016, 1032
Subject Index meteorite
meteorite (Allende) meteorite (Brownfield) meteorite (Cation Diablo) meteorite (iron)
1213 2,122-124,229,233,235, 236, 261, 266, 349, 362, 378, 404, 429, 443, 523, 526, 534, 625-628, 632, 688, 713-715, 718, 720, 723, 724, 788, 789, 803, 816,818,900,977 422,540,710,721 724 538,717
26~ 268, 350, 715, 717, 718,879,900,977 685,687,688 meteorite (Martian) meteorite (Murchison) 229,230-235 724 meteorite (Orgueil) 362 meteorite (primitive) meteorite (Semarkona) 724 meteorite (Sikhote Alin) 538, 900 meteorite impact 632 meteoritic amino acids 236 meteoritic components 632 meteoritic compounds 230,232 meteoritic diamond 266 meteoritic organic fea- 235 ture meteoritic organic frac- 233 tion meteoritic organic matter 229, 230, 236 meteoritic refractory 653, 689 inclusion meteoritic samples 123, 361, 816 meteoritic SiC 261 meteoritic sulfur 879 meteorological inviron- 508 mental variations 384, 508 meteorological parameters 4, 7, 67, 75, 76, 169, 231, methane 272, 291-303, 377, 766, 774, 775, 818, 847, 892, 894,1012,1023 methane combustion 298, 300 methane oxidation 1013 methanol 44, 127, 164, 166, 197, 198, 201, 223, 317, 318, 322, 323, 338, 516, 561, 584,609,1012 methanosulfonate (MSA)559, 561, 562 methanotropic bacteria 1013 methemoglobinemia 305 methionine 554, 555, 585, 595, 899, 902
149, 193, 195, 418, 445, 502, 541, 661, 668, 671, 680,687 537 methodology study 608,614,615 methyl chloride 163,166 methylates 164,166 methylation 822,928,929,935 metrology 133 Mg contaminants Mg isotope standard 710 Mg perchlorate 503,593,598 16 Mg-Pt reagent 57, 261, 264, 265, 348, mica 349, 351-354, 356, 357, 429,431,530,1001,1004, 1017 351 mica crystal mica sheets 240 351,356 micaceous rock 306,314 micro diffusion Micro Isotope Laser Ex- 462 traction System (MILES) micro reduction appa550 ratus micro scale fluorination 458, 460 microalgae 1014 microbe cultures 650 microbeam technique 459, 1017 microbes 291, 308, 344, 649 microbial conversion 591 microbial denitrifica330 tion microbial denitrifier 306 method (non-)microbial reduc- 343 tase microbial respiration 211 1011 microbiology microdiamond 270, 271 micro-IR 1017 micrometeorite 681 microporosity 843 microwave plasma 1022 Mid Atlantic 85 migration of air 410 milk 2,31 milk casein 595 millipore filter 215, 488 mineral composition 38, 407, 408, 690 mineral contamination 38, 43 mineral deposition 62 mineral grains 361, 422, 461, 462 mineral horizon 592
methodology
1214 mineral identification technique mineral oil mineral phases mineral separates mineral specimen mineral standard mineral surface mineral synthesis mineralization mineralogy
Subject Index 44
895 39,62,68,72,535,673 351-353,681 524,548,551 673 4,458,541,549 994 185 46,673,1005,1010,1016, 1017 miniaturization 451,458,461,46Z 471 minimal contamination 555 minor contamination 252, 657 minor peak 96, 144, 583 mixed matrices 673 Mn reactor 16 Mn reduction 7, 23, 36 model seawater 1015 moisture 20, 28, 206-208, 281, 457, 461, 474, 502, 685, 815, 816, 1003, 1040, 1041 moisture content 53, 1013 moisture traces 170 molar mass 526, 920, 924, 936, 939 molecular beam 378, 388, 657 molecular flow 838, 924, 925 molecular fragmentation 390, 393, 394, 655, 860, 987, 990 molecular interference 666, 714 molecular ion beam 393, 657 molecular radical 839 molecular sieve 3, 76, 158, 165, 174, 175, 189, 296, 297, 318, 337, 357, 383, 403, 430, 441, 465, 469, 470, 501, 503, 557, 559, 805-819 mollusk/ca 1011, 1028, 1029 molybdenite 719, 720 momentum transfer 655 Monel alloy 414, 429, 453, 462 monochromatic radiation758 monoisotopic 123, 128, 957 monosaccharide 201 Monte Carlo simulation 964 Montelupo clay 903 MORB 85,139,1034 mosses 507,516,517 mudstone 43 multi(ple)-collector 1,129,132,134,344,393, (MC) 450, 533, 546, 550, 645, 649, 650, 670, 674, 679, 681, 68Z 690, 692-725,
MC-ICP-MS multiple interactions multiple-pass gas cell multiple-reflection gas cell muscle muscle tissue muscovite mustard plants m / z peaks
733, 745, 800, 802-804, 90Z 926,992,1036 132-139, 450, 649-651, 692-725,992,993,1019 182 774, 775 774 191,594 181, 182, 191, 19Z 198, 594 351, 353, 356, 429, 445, 1003,1021,1031 594 97,29Z 299,358,598
N N contamination N20 contamination N20 cycle Na2CO3 standard Nation Nation dryer / drying Nation membrane Nation tube nails Nakhla pyrrhotite nano-diamond narcotic drugs narrow band-pass filter narrow peaks National Metrology Institution (NMI) national standard natural abundance
318,320,330 282 390 209 168,28Z 469 283,284,28Z 593 287,469 168,469 186,594 685 256,260,266 172 764, 778 853 822,824,935
935 1, 6, 103, 104, 108, 109, 115, 121, 155, 161, 172, 175, 178, 180, 306, 312, 314-316, 320, 323, 328, 378, 391, 392, 49Z 499501, 503, 550, 606, 632, 639, 754, 760, 761, 763, 769, 772, 775, 84Z 852, 898,947,972,986 natural abundance stan- 328 dard natural air 760-763,778 natural brines 32,1019 natural compound 108, 175, 597, 599, 881, 892 natural concentrations 323 natural contribution 632 natural cycling 273 natural distribution 947
Subject Index natural natural natural natural natural natural
ecosystem 200 element 628 environment 399, 523, 1027, 1034 exchange process 650 fluids 149 gas 7,157,173,159,169,291, 349, 579, 585, 894, 900, 951,1041 natural gas standards 948,951,956 natural groundwater 204 natural grain size 41 natural hydrocarbon gas 884, 893 natural isotope compo- 629-632, 633, 635, 648, sition 825, 829-831, 833, 889, 897, 923, 935 natural isotope / ic 122, 624, 631, 701, 702, fractionation 708, 718 natural isotopic ratios 16, 629, 630, 634, 635, 647 natural levels 975 natural material 122, 139, 147, 523, 538, 604, 605, 648, 691, 715, 717, 720, 875, 877, 880, 883, 900, 903, 911, 941, 947 natural matrices 903 natural mixture 123 natural origin 103, 112, 118, 120, 121, 128 natural phases 992 natural precursors 118, 120 natural products 118, 1016 natural profile 267 natural reference 137, 538 natural samples 36, 44, 119, 128, 135, 143, 152, 159, 193, 206, 209, 211, 222, 226, 266, 267, 315, 319, 320, 323, 328, 373, 376, 392, 406, 450, 585, 598, 600-602, 604, 606, 621, 625-628, 646, 709, 720, 721, 723, 756, 767, 772, 941, 985, 1002 natural sources 119, 121, 122, 556, 557 natural species 389 natural stable isotopes 112, 142, 146, 378, 387, 609, 623, 631, 635, 715, 717, 718, 720, 724, 835, 897, 899 natural standard 646, 647 natural state 233 natural settings 624, 626 natural substances 540 natural system 52, 993
1215 natural natural natural natural
uranium values variability variation
natural water
natural wetlands naturally occuring processes NBS22 oil NBS120c phosphorite standard Nd glass laser Nd-YAG laser nebulization negative peak negative TIMS
735, 738, 739 632 116, 209, 211, 713 122, 406, 549, 604, 605, 621, 623, 627, 630, 693, 713, 717, 718, 725, 831, 875 2, 4, 5, 17, 19, 21, 25, 36, 125, 133, 142, 203, 204, 206, 209, 211, 212, 219, 222, 224, 227, 312-315, 320, 329, 482, 483, 639, 767, 772, 780, 885, 888 775 391 896, 884, 897, 910 488-493
403 403, 465-467, 544, 548 138, 550, 711 859 143-145, 148, 149, 151, 152, 450, 451, 494, 550, 637, 642, 643, 747, 1015 neutralization 332, 337, 415, 418, 419, 421, 442, 451, 455, 488, 655, 661 neutralization processes 655 neutralization reactor 421 NH4+-free environment 320 Ni alloy wire 578 Ni bomb/tube pyrolysis 8, 23, 30, 497, 498 Ni catalyst 502, 1021 Ni-Cr-Co alloy 1025 Ni-Thoria catalyst 1021 nickelized graphite 334, 404 nickelized graphite 342 reactor nicotinic acid 319 Nier Johnson geometry 697, 727 Nier type ion source 364, 859, 860, 872 ninhydrin oxidation 320 NIST (National Insti34, 129, 133, 142, 305, 310, 311, 487, 533, 646, tute of Standards 706, 707, 712, 715, 716, and Technology) 719, 721, 750, 798-803, 874-906, 910, 918, 919, 922, 944948, 951, 953, 954, 956, 1049 nitrate contamination 329 nitrate radical 377
1216 nitrate reduction
320, 324, 330, 343, 344, 624, 649 nitrate standard 309, 335 nitric acid 10, 12, 13, 127, 320, 365, 366, 572, 579, 610, 611, 617-619, 637, 643, 739 nitrification 312, 1013, 1033 nitrogen atmosphere 215, 218, 570, 574, 579, 586, 595, 877, 1023 nitrogen isotope labeled 320, 345, 392, 396, 398 nitrogen standard 259, 357, 360, 368, 369, 371, 898 181 nitrogenous nutrient nitrous oxide reductase 335 899, 902 N-methyl anthranilic ester 835 Nobel Prize 1008 noble metal alloys noble metal ion-exchan- 813 ged zeolite non-absolute (isotope) 828 amount ratio non-adiabatic collision 385 non-aqueous environ240 ment non-baseline conditions 278 non-condensable gas 55, 59,76, 208, 216, 223, 379, 412, 451, 533, 818 non-dispersive technique 778 non-electrolytes 32 non-fluorination techni- 468 que non-homogeneous leaf 481 water non-hygroscopic 425, 486, 633 non-linear discrimina- 647 tion drift non-linear equations 629 non-linear expression 629 non-linear fractionation 687 non-linear scale contrac- 784 tion non-linearity 284, 288, 302, 312, 370, 647, 825, 838, 842, 848, 851, 852, 868, 875, 923 non-methane hydrocar- 377 bons (NMHC) non-quantitative con160, 162 version non-radiogenic isotope 623 non-reactive window 468 non-seasalt sulfate 557, 561, 562 non-silicate fraction 41 non-silicate standard 358
Subject Index 514 517 20, 150, 152, 699, 782, 784, 886, 887, 891, 895, 898, 900, 903, 954, 955, 1051 normalization correction 15, 150, 151 normalization laws 671 normalization procedure 144, 149, 151 normalization process 668 North Atlantic 82, 562, 714 North Pacific 137 Northrop-Clayton 998, 1000, 1001 method n-pentane 582, 584, 613 NSS aerosol 559 nuclear activities 738, 739 nuclear application 122 nuclear charges 110 nuclear chemistry 907 nuclear contamination 738 nuclear magnetic reso- 103-121,896 nance (NMR) NMR interactions 104, 105, 110 NMR magnetic inter104, 106 actions NMR power decoupling 109, 114 nuclear magnetogyric 104, 108 ratio nuclear mass spectro831 metry nuclear material 919 nuclear Overhauser en- 106, 110 hancement (NOE) 230 nuclear processing 124,540 nuclear reactions 1,122,903 nuclear reactor 729,738,919 nuclear safeguards 122 nuclear science 104, 749,753, 754 nuclear spin 124 nuclear technique nuclear test ban treaties 273 nuclear weapons 920 nucleation process 1003 nuclei 103-114, 118, 120, 766, 835 nucleopore filter 187 nucleosynthesis 2 nucleosynthetic effects 626 nucleosynthetic process 977 nuclidic mass 376 nutrient autoanalyzer 202 nutrient concentration 179, 181, 184, 187 nutrient distribution 179
non-wood sample non-woody plants normalization
1217
Subject Index nutrient loading nutrient source nutrient supply nutrients nutritional tracer study
179 179 1014, 1024 180, 181, 189, 306 6
O 180 thermometer 908, 1031 02 contaminant 430 02 oxidation 464 02 / CO2 reactor 430 observed isotope / ic 158, 375 fractionation oceanic basalt 688, 1033 oceanic chert 1032 oceanic rock 130 off-line (method/system)6, 157, 204, 239, 240, 255, 335-337, 350, 498-500, 502, 505, 511, 520, 598600, 602, 896, 1051 off-line attachments 192 off-line batch mode 16 off-line combustion 199, 201, 333, 336, 337, 340, 891, 896 off-line determination 603 off-line preparation 7, 8, 430, 432, 542, 595, 597-600, 883 off-line process 239 off-line pyrolysis 233, 497, 498 off-line studies 337 off-line technique 157, 191, 300, 336, 497, 498, 511, 519, 520, 891 oil 3, 4, 46, 172, 578, 579, 581-584, 883, 894 oil field 1012 oil free compressed air 1040 oil free pump 951 oil mist filter 1041 oil polution 173 oil sand 583 oil sand bitumen 581 oil shale 575, 599 oil seep 1033 oil spills 173 oil wells 159 oil-fired power plant 559, 562 olive oil 172 olivine 68, 70, 125, 402, 406, 429, 441-443, 464, 466, 471, 530, 673, 1031 olivine crystal 70 one-line sample combus- 164 tion
on-line (method/system)2, 6, 7, 11, 17, 23, 25, 29, 30, 74, 88, 96, 102, 157, 173-175, 159, 186, 202, 333, 335, 336, 403, 429, 497, 498, 501-504, 506, 508, 510, 511, 520, 547, 593, 598, 817 on-line analysis 156, 391, 505 on-line collection 418 on-line combustion 162, 319, 333, 334, 543, 891, 896 on-line coupling 174 on-line determination 603 on-line equilibration 18, 25 on-line furnace 258 on-line GC 853 on-line heating 350 on-line inlet 433 on-line IRMS 156, 1041, 1051 on-line measurement 8, 782, 851, 853, 1041 on-line neutralization 418, 419 on-line preparation 6, 290, 319, 542, 599, 879 on-line pyrolysis 174, 195, 233, 497, 500, 501, 847, 1041 on-line reactor 401 on-line reduction 7, 25 on-line technique 6, 497, 500, 511, 519, 520, 547, 598, 879, 891, 896, 971 on-line thermal conver- 1041 sion opal 56, 420 opaline frustules 1015 optical absorption tech- 282 nique optical components 467, 777, 842 optical detection 759, 766 optical excitation 780 optical feedback 777 optical geometry 735 optical grade window 421 optical isolator 774 optical isotope labora782 tory optical isotope ratio 784 instrumentation optical isotope ratio 760, 766, 780, 782 measurement optical microscopy 39, 659, 679, 1017 optical parametric 783 oscillation optical path length 768, 769, 771, 774-776 optical prism 844 optical properties 123
1218 optical pumping optical pyrometer optical set-up optical shielding optical spectroscopy optical spectrum optical system optical technique / method optically active 03 optogalvanic effect optogalvanic signal optogalvanic spectroscopy ore composition ore samples organ tissue organic acid organic assemblage organic carbon
Subject Index 754, 755 402 777 384 125,380,803 766 467 756, 759, 766, 779, 781, 782 377 780,880 780 780
716 707, 716 1024 163,164,366 236 47, 184, 185, 188, 200, 225, 345, 484, 529, 54Z 565, 600, 602, 895-89Z 1014,1023,1031 346 organic carryover organic cation exchanger 125 578 organic coatings 30Z 579 organic components 4, 7, 46, 112, 153, 155, organic compound 156, 158, 168, 172, 174, 175, 184, 189, 191, 194, 230, 132-234, 278, 313, 315, 334, 338, 343, 344, 362, 373, 407, 450, 490, 496, 506, 538,541, 546, 54Z 554, 569, 571, 574, 575, 578, 581, 588, 590, 594-596, 605, 615, 624, 635, 637, 639, 646, 895, 896,1041,1051 229,232,478 organic constituents organic contamination 70, 257, 270, 336, 338, 365, 489, 492, 565, 639, 646,1040 44,149 organic contents organic cosmochemist 229 245 organic dust 19Z 198 organic extract(ion) 235 organic feature 546 organic fluoride compounds 233 organic fraction organic fragment peaks 641 organic hydride peaks 639 organic hydrogen 46,1014, 1051
organic impurities organic interference organic isotope ratio organic inventory organic mass spectrometry organic matter/material
101, 895 315, 346, 636, 646, 647 173 230 154, 175, 844
3, 8, 11, 22, 24, 40,43-46, 58-60, 64, 69, 125, 127, 138, 149, 177, 178, 181, 182, 184, 186, 189-191, 200, 201, 203, 229, 230, 235, 236, 256, 25Z 268, 269, 271, 319-321, 338, 365, 366, 448, 473, 476, 482, 484, 489, 490, 498, 501, 506, 508, 54Z 573, 595, 613, 624, 682, 891, 895,1014 organic molecules 103, 339, 489, 624, 635, 798 organic network 233 organic nitrogen 256, 268, 305, 306, 312321, 330, 339, 344, 346, 899 organic oxygen 973 organic phosphate 490 organic precursors 236 organic reducing agents 582 organic reference mate- 895, 89Z 902 rial 173,642 organic residue 30Z 310, 341, 342, 49Z organic samples 507, 515, 546, 580, 615, 81G 899,973 322,345 organic solutions 7, 44, 195, 196, 208, 229, organic solvents 516,517,555,572,581 329,639 organic species 498,505 organic standard organic structure 553 45,334,501 organic substances 1013 organic substrate 552-554, 562, 565, 568organic sulfur 57G 579, 581,, 585, 586, 588,590-596,603 349,1007 organic systems 44,59,191,360,579,580, organic-rich 1012 41, 43, 44, 46, 53, 57, 58, organics 565, 636, 639-641, 646, 1021 organisms environment 196 organochlorine 615, 616 origin of life 229
1219
Subject Index 346 1003,1017 243-245, 249, 250, 278, 300, 515, 519, 644, 681, 685 883 outlier exclusion 277, 278 outlier rejection 883 outlier test 335 overnight incubation 897,899,1026,1027 oxalic acid oxic surface environment 329 59, 64, 75, 118, 188, 203, oxidation 262, 265, 267, 407, 427, 430, 490, 552, 565, 567, 571, 572, 586, 594, 595, 637, 639, 649, 967, 968 162 oxidation catalyst 199, 303, 342 oxidation furnace 307 oxidation numbers 967 oxidation procedure 319 oxidation promotion 75 oxidation rate 45, 46, 430, 624 oxidation reaction 973 oxidation reactor 377, 552, 553, 601, 624, oxidation state 625, 641, 941 637, 638 oxidation trap oxidation-reduction be- 619 havior oxidative pyrolysis 68, 70 oxidizing environment 625 oxidizing reagent 489 oxyfluorides 617 oxygen atmosphere 59, 256, 258, 265, 362, 366, 403, 489, 548, 972 oxygen sink 162 oxygen-free atmosphere 500 ozone/03 291, 309, 375-390, 1022, 1024 ozone chemistry 376, 382 ozone cycling 385 osmosis Ostwald ripening outlier (results)
P Pacific ocean Pacific seawater palaeoenvironment paleoclimate paleo-dietary habits paleohumidity palladium foil parabola spectrograph
137, 292, 550, 559, 561, 1012 137, 712 198, 482, 483, 597 38, 483 173 482 8 835
para-hydroxybenzaldehyde parent-daughter comparison parent-daughter correction parent-daughter fragments parent-daughter method parrafin wax Parr bomb Parr bomb combustion Parr bomb oxidation Parr bomb technique partial electrolysis partial exchange experiments partial exchange method partial exchange reactions partial combustion partial fluorination partition function passive fluorination Pd filter Pd foil membrane Pd-Pt catalyst peak area peak broadening peak data peak definition peak detection peak distortion peak flanks peak height peak hopping peak integration peak intensity peak interference peak maximum peak overlap peak profile peak reproducibility peak resolution peak schematics peak separation peak shape
117, 120 989 987, 990 987 988, 989 4, 224 578-580, 582, 584, 594, 615 574, 579 593 581, 595, 615 30 995,998 998, 999, 1001 64 515 51, 52, 420, 423, 427, 443-445, 464, 471 384, 770 5O 8, 175 175 1021 115, 162, 226, 297, 302, 303, 503, 840 169, 170, 855 168 170 168 170, 171 843 115, 297, 364, 366-368, 372, 664, 799, 832 634, 646, 989 168, 286 96, 113, 115, 768, 769, 773, 795 113, 794 664 113, 115, 156, 161, 168170 367 285 664, 665 665 666 96, 170, 665, 680, 693, 694, 844, 845, 851, 873
1220 peak shift peak signal peak size peak slope peak symmetry peak tailing peak to tail signal peak width peak-jump(ing) peat pectin pedosphere PeeDee belemnite PeeDee formation pegmatite perchlorates perchloric acid periodic table permeability permeable tape permeable to H2 permeable to liquid permeable to water persulfate oxidation petroleum petrologic persuits petrologic problems petrologic study petrologic system phenyl group phenylacetaldehyde phenylacetic acid phenylethanol phloem phosphate chemistry phosphate contamination phosphate radical phosphatic tissue phospholipids phosphoric acid
Subject Index 125,126 743 199,312 665 115 169,171,299,742,744 666 226, 296, 664, 69Z 799, 855 29,76,465,549,728,847 568,585-591 200 638,541,552,596 879 879,892 351,406 124,125,60G 605 124,426,524 653, 694, 789, 790, 793, 835,939 41, 53, 378, 418, 752, 1006,1008 323 1024,1027 513 469 6O8 44, 185, 235, 578, 579, 583,951 358 349,358 349,359 360 553 118 119,120 118-120 473,481 239 644
482,484-486 483,490 196 24, 58, 12Z 206-209, 211, 215, 216, 218, 220-223, 237-239, 243, 245, 248, 250, 253-255, 569, 588, 591,645,895 496 phosphorus cycle photoacoustic detection 767, 778, 779 photoacoustic spectro- 762-764,780 scopy photochemical control 384
photochemical equili384 brium model photochemical forma- 377 tion photochemical lifetime 379 photochemical produc- 378 tion photochemical smog 378 photodissociation 377, 385 photolysis lifetime 379 photon flux 750, 752 photon stop 727 photorespiratory process 200 photosynthesis 177, 200, 206, 211, 272, 1030 photosynthetic fixation 183, 200, 1013 photosynthetic fluxes 290 photosynthetic metabo- 200 lism photosynthetic pathway 116 photosynthetic process 177 photosynthetic produc- 481 tivity photosynthetic uptake 272 phototoxic effect 378 phyllosilicates 56 physical-chemical 267 background physiological changes 183, 191 physiological effects 108 physiological invariant 182 source physiological processes 182, 507 physiological questions 175 physiological variation 182 phytoplankton 178, 179, 182, 184, 187, 200, 1014 pioneering application 993 pioneering technique 138 pioneering work 349, 622 piston-cylinder capsule 1017 plagioclase 68, 351, 352 Planck's constant 752, 924 planetary cores 625 planetary differentia632 tion planetary system 229 planet-forming processes 2 229, 361 planets plant appearance 185 473, 1033 plant environment 174, 191, 197, 201, 473plant matter / material 475, 479, 497, 507, 517, 593, 594
Subject Index plant mesophyll plant parts plant physiology plant press plant samples plant plant plant plant plant plant plant plant
sap sources species structure tissue transpiration rate type water
plant wax esters plant-derived hydrocarbons plants
plasma (blood) plasma (ionized gas) plasma amino acids plasma analysis plasma ashing plasma conditions plasma discharge plasma flicker plasma furnace plasma interference plasma ionization plasma lactate plasma power plasma source plasma structure plasma torch plastic syringe platinized platinum black Pleistocene Pliocene pneumatic air supply pneumatic control pneumatic cylinder pneumatic valves
1221 182 473 497,507 185 3,181,185,197,474,476, 593,626 474 192 517 518 180,197,592,593 473 517 2, 473, 474, 481, 1014, 1032 196 196 3, 22, 29, 108, 11Z 17Z 179, 180, 181, 185, 195, 197-200, 272, 378, 473, 497, 516, 592-594, 649, 1011, 1013, 1014, 1024, 1028, 1030, 1032, 1033, 1035,1036 2,10,31,32 45, 132, 458, 489, 550, 650, 651, 656, 694, 70Z 720, 791-794,796,926 173 139 40,44,45,48,58 713 748 694, 726 45,46,489 696,698 693,698,700,701,706 161 46 138, 139, 551, 651, 692, 694, 697-699, 701, 709, 713, 725,74Z 796 790, 792 550 26,20Z 212-214,225 5,9,23,423 16 495 495 950 474,476 475 284,462,950
Poisson statistics 667, 855 polar hydrocarbons 230 polar ice 564 polar organic compounds236 Polar Science Program 889 polarization 749, 753, 754 polarized light 749 pollen 555 polyatomic interference 651 polyatomic ion 642, 651, 666, 707, 715 polycarbonate 186, 506 polycyclic sulfide 554 polyethylene 883, 884, 895, 897 polyethylene beaker 525 polyethylene bottle 3, 212, 213, 313, 618 polyethylene flask 1013 polyethylene foil 910 polyethylene glycol 163, 170 polyethylene screen 326 polymerization 239, 895 polythionate 553 poorly defined matrices 673 Porapak 285, 286, 612 Poraplot 296 pore water 15, 22, 38, 53-56, 60, 132, 204, 572, 590, 603 porous membrane 54 porous resin beads 19 porous Teflon frit 640 porphyrins 200 possible interaction 792 possible isotope/ic frac- 36, 343, 350 tionation positive peak 859 positive TIMS 143-150, 152, 614, 549 post-depositional envi- 483 ronment potential contamination 41 potential health hazard 578 potential isotopic frac- 5, 34, 311 tionation potential mass fractio- 144, 146 nation artifact potential sensitivity 988 powdered wood grain 507, 521 size power density 463,548,751 power generator (UPS) 1039 power law 671, 699, 700, 702, 704, 706,707,730,788,796 Precambrian 268 Precambrian cherts 268 precious / noble metal 1006,1025 capsule
1222 precipitation (rain, ..) precipitation/rainfall collectors precipitation procedure precipitation process precipitation reactions precipitation reagent precipitation technique
Subject Index 2, 36, 306, 497, 557, 564, 1033 312, 556
212 221 203 212 211-214, 217-220, 222, 228 pre-combustion 25Z 365,366 pre-concentration 293-29Z 303,556,565 pre-concentration device 295, 303 predominant sink 377 prefluorination 50, 57, 432, 435, 438, 440, 456, 467, 496 preliminary experiment 50, 819 presolar grains 362 pre-sputtered 798 pressure-equilibration 283, 285 (-volume) pressure-filtered 590 pretreatment 4, Z 16,1Z 22,23,38,44, 45, 318, 319, 422, 424, 426-428, 432, 435, 438, 440, 556, 563, 568, 759, 766,775,781,895 primary artifact 945,954 primary beam 123, 535, 536, 549, 652, 655, 656, 658, 660-662, 66Z 670, 674, 675, 679, 680,685,688 primary ion 535,655,658,668 primary ion beam 53Z 652, 655, 65Z 659, 674,680,682 primary ion delivery 658 primary ion spot 660 primary reference mate- 892, 898, 900, 902-905 rial primary sink 291 primary standard 133, 772, 784, 881, 929933, 944, 949, 954, 978, 979 primates 181 principles of statistics 872 pristane / phytane ratio 578 probability functions 975 pronounced isotopic 123 fractionation propanol 4, 165, 533 protein 4, 32, 161, 182, 195, 198200, 489, 592, 624, 1014 Pt boat 617,618 Pt capsule 1008, 1020, 1026
Pt catalyst
17, 19, 20, 21, 55, 403, 473, 850 Pt-A1203 catalyst 21, 474 PTFE membrane 323, 572 pulse counting 364, 667-669, 739, 740 pulsed laser 777 pump oil 46, 456, 457, 850 purines 200 Pyrex 10, 13, 14, 16, 24, 28, 65, 72, 208, 215, 220-222, 274, 275, 77, 287, 337, 410, 414, 427, 467, 495, 525, 528, 583, 611, 1021 pyridineacetic anhydride 201 pyrite 441, 447, 547-549, 552, 568-576, 579, 581, 582, 599, 602, 603, 686, 723 pyrite fraction 574, 576 pyrite grains 465, 685 pyrohydrolysis method 617 Pyrolysis/pyrolyzed 22, 65, 70, 87, 174, 175, 233, 256, 258, 262, 265, 268, 334, 342, 362, 366, 373, 401, 404, 408, 474, 474-481, 497-506, 582, 818, 847, 1041, 1051 pyrolysis fragments 175 pyrolysis interface 174, 175 pyrolysis reactor 8, 494, 501, 1051 pyrolysis system 174, 175, 342, 476, 478 pyrolysis technique 24, 476, 479, 481, 494, 497-506 pyrolysis temperature 8, 499 pyrolytic products 175, 497, 500 pyrolytic reaction 174 pyroxene 70, 431, 433, 441, 442, 530, 570 pyroxene glass 443 pyrrhotite 441, 447, 465, 568, 570
Q Q (quadrupole)-ICP-MS 132, 692, 693, 698, 702, 708-717, 720, 727, 734 Q-switched laser 753 quadrupole field disper- 734 sion quadrupole mass filter 93, 95 quality control 2, 36, 290, 293, 301, 343, 739, 1051 quality control material 881, 882, 904 quality control purpose 905 quality control tank 278, 279 quantitative conversion 4, 5, 923
Subject Index quantitative peak collection quantum cascade (QC) laser quantum interaction quartz (mineral)
1223
158 760,765,783
748 46, 56, 5~ 66, 68, 69, 72, 74, 359, 401-403, 406, 420, 421, 429, 433, 436, 442-445, 451, 457, 458, 461, 463, 465-467, 530, 532-534, 658, 878, 996, 1003-1005, 1009, 1011, 1020,1030,1031 quartz arenite 465 quartz furnace 858 quartz insulator 791 quartz phenocryst 443,444 quartz sand 533-535 quartz siltite 465 quartz standard 453 quartz vein 71,358,359 quartz wool 5,59 quartz wool filter 559 quartz xenolith 68 quartzite 535 quasi simultaneous arri- 668 val (QSA) effect quaternary science 154
R Rabi frequency RC filter radial diffusion cell method radiation radiation intensity radiation wavelength radiative forcing radio frequency radio frequency heating radio frequency induction furnace radio frequency radiation radiogenic radiogenic contents radiogenic decay radiogenic effects radiogenic isotope rain rainfall
750, 752 859 55,60
rainstorms rainwater rainy season Raman Raman microprobe Raman spectroscopy random interleaved sampling Raney-Nickel alloy rapid irradiation rare bird species rare earth elements raspberry ketone Rayleigh distillation reacted sample storage reaction bottle incubation reaction equilibrium reaction kinetics reaction pathways reaction products reaction products reactivity reactive compound reactivity of zinc reactor (unspecified) reactor (high temperature) reactor tube reactor surface reactor temperature reactor walls reagent
366,777,783 790 458 272,291 45,114,115,384,789 404 402 650 523,632 632 671 632 623, 624, 627, 628, 632, 693,699, 701, 715 3, 36, 308, 317, 325-327, 331,332,556 108,116,312
reagent blank reagent contribution reagent cost reagent cylinder reagent mixture reagent neutralization reagent poisoning reagent pressure reagent reactivity
306 108,313,559 172 101Z 1036 77 65,68,77 777 588 463 820 666,806,1023 114,117-119 430,923,1033 14 346 91,92 2Z 91,998 203,309,653,839 39, 252, 356, 402, 419, 421,451 442 1
11 5, 7, 23, 165, 418, 421, 500, 502, 583 578 583 5OO 494 5O0 15, 23-25, 30, 36, 50, 51, 138, 148, 160, 164, 165, 184, 185, 217, 220, 227, 238, 241, 250, 251, 318, 320, 330, 338-340, 343, 352-356, 404, 408-413, 417, 418, 420, 422, 424, 426, 428, 431-437, 440, 441, 443, 451 ,452, 455, 456, 516, 543, 572, 578, 637, 638, 644, 729, 743 127, 452 23 412 412, 415, 417 321 416, 418 4 466 16
1224 220, 410, 415-419, 455, 462,527,528,895 415 reagent transfer 213 reagent water sample 387 recombination theory 1004,101Z 1027 recrystallization 30Z 625 redox chemistry 624 redox conditions 624 redox indicator 624,625 redox process 30Z 493,624,1026 redox reaction 625 redox transformations reduced inorganic sulfur 571, 587, 588, 592 reducing agent 7, 333, 483, 553, 581, 582, 650 reducing environment 562,625 reducing reagent 5 550,650 reductants injection 7 reduction / pyrolysis 55O reduction flask 973 reduction reactor 342 reduction system 52,239,240,250 re-equilibration reference gas injection 284,285,880,882 4Z 112, 131, 133, 136, reference material 241, 243, 261, 26Z 271, 30Z 310, 311, 33Z 341, 392, 506, 526, 533, 538, 539, 544, 546, 551, 596, 598, 599, 606, 648, 706, 729, 735, 736, 738, 765, 768, 769, 772, 774, 778, 784, 786, 828, 833, 855, 874-927, 932-936, 944, 949, 954, 959, 974, 978, 981,985,1048,1049 reference material sto- 888,896,953 rage 171,297,855 reference peaks 15,33,135,311,336,604, reference standard 605, 648, 798, 803, 882, 1052 413 re-fluorination 1017 refractive index oil 70Z 708 refractory oxides 173 refractory waste 339,500,703-707 regression line 475,502,1014 relative humidity 104, 109, 494, 670, 779, relative sensitivity 795 498 reliable standard 690 reliable statistics 298 repeated oxidation 916 replicate experiment reagent storage
Subject Index reproducible analysis / measurement reproducible apparant mass discrimination reproducible data reproducible DIC concentration reproducible experiment reproducible extraction technique reproducible filament loading reproducible isotope mass-discrimination reproducible isotope values
495, 507, 675, 688, 690, 726 734 69, 548, 690 209 1011 508, 516, 520 627 675
45, 193, 195, 209, 217, 310, 311, 358, 447, 488, 495, 507, 546, 726, 1050 74 reproducible method reproducible procedure 222 reproducible reaction 243 reproducible repeata729 bility reproducible results 209, 219, 674, 966 reproducible yield 488 residual artifact 673 resin/diffusion method 328 resin exchange capacity 327, 639 resin exchange column 125, 323, 326, 328, 331, 338, 564, 640, 642 resin from bitumen 582 resin volume (in column) 125, 127, 326, 564 resonance excitation 755 resonance frequency 104, 105, 107, 114 resonance ionization 746-756 resonance signal 106 respiration 206 respiratory fluxes 273, 290 respiratory releases 272 retardation lens 695, 696 reverse electrolysis 863 Rh catalyst 1021 rhyolite 127, 137, 140, 903, 904 risk assessment 928 Rittenberg analysis 322 Rittenberg method 340, 341 Rittenberg oxidation 321 Rittenberg technique 321 Rittenberg Y-tube 340 river water 132, 139, 140, 1033 rock matrix 570, 647 rock matrix standard 647 rock standard 132, 137, 138, 647, 795 rodent teeth 437 root respiration 203, 1014
Subject Index rose oil rotary evaporator/ion rotational constant rotation-vibration bands rotation-vibration motions rotation-vibration interaction rotation-vibration transition routine chemistry routine normalization rubber septum RuMP cycle
34S enriched spike salicilic acid saline brine saline fluid saline groundwater saline water
1225 118 197,201, 564 770 766 759 770 770 487 902 55, 221, 224, 571 1013
540 318 130, 1018 6 134 11, 22, 324, 639, 1034, 1035 saliva 2, 10, 31, 88 salt 2, 27, 32, 33, 40, 43, 146, 185, 310, 318, 319, 330, 337, 342, 346, 410, 421, 490, 553, 610, 620, 998, 1016, 1018, 1027 salt lakes 140 salt solution 31, 32, 1019 Salton Sea 1029 sample blend 821 sample capsule 615, 1008 sample combustion 298, 341 sample digestion proto- 314 col sample filament 550,643,645 sample impurity 138 sample incubation 188,345 sample injection 6,7,9,156,162,169,171, 206, 211, 283, 285, 295, 297.470 sample introduction 285, 293, 458, 468, 708710, 713-715, 717, 725 sample loop 6,283-285,288,293,296, 298,302,303,470 sample loop equilibra- 285 tion sample matrix / ces 178, 188, 450, 596, 635, 637, 651, 706-708, 731, 735, 738, 796, 973 sample peak 168, 289, 299
sample purity sample storage
38, 138, 710 2, 3, 157, 179, 184-187, 193, 200, 205, 206, 211, 212, 214, 215, 220, 228, 280, 312, 475, 479, 562, 593 sampling protocol 178, 184 sandstone 39, 42, 79 satellite observations 385 scale contraction 539, 784, 900 scale conversion 887, 888 scale shrinkage 7, 14, 15, 34, 35, 311 scale stretching 34, 35, 311 Schumann-Runge bands 378 Schtitze-Unterzaucher 500,501 method science 928 Se contamination 643, 647 Se oxidation 649 Se reduction 626, 636, 637, 649, 1018 sea-air interface 559 seafloor basalt 139 sealed-tube 12, 14, 59, 189, 191, 201, 333, 353, 404, 406, 611, 616, 1024 sealed-tube combustion 326, 327, 332, 339, 340, 351 sealed-tube decripitation 358 sealed-tube extraction 348,358 sealed-tube experiment 348, 1020 seasalt sulfate 556, 557 seasalt tracer 557 seawater 32, 125, 132-142, 316, 323, 335, 336, 490, 606, 614, 710, 712, 721, 722, 900, 904, 905, 1015, 1018, 1019, 1028, 1029, 1032, 1034, 1036 seawater chloride 904 seawater compartments 878 seawater derived 559 seawater matrix 149 seawater pH 1035 seawater samples 25, 129, 132, 324, 723, 903,904 seawater source 63 seawater sulfate 556,900 seawater temperature 1035 seawater-rich oceanic 689 crust secondary ion 1, 535-537, 549, 652-691 secondary ion beam 653, 655-657, 659, 661664, 669, 675, 677-679
1226 secondary ion characte- 658 ristics secondary ion composi- 755 tion secondary ion collection 668 secondary ion current 549 secondary ion emission 658 secondary ion extraction 656, 658, 674, 675, 679 secondary ion generation 549 secondary ion mass spec-123, 405, 444, 523, 535, trometry/SIMS 537, 543, 548, 549, 551, 652-691, 788, 789, 799, 800, 802, 992, 993, 1015, 1018, 1019 secondary ion optical 663, 679 axes secondary ion signal 659, 670 secondary ion speciation 657 secondary ion spectra 664 secondary ion system 674 secondary ion transmis- 536 sion secondary ion tuning 660 conditions secondary ion yield 658 secondary ionization 654, 655, 746-748 sedimentary apatites 486 sedimentary basin 140, 349 sedimentary deposits 483, 813 sedimentary phosphate 483 sedimentary processes 38 sedimentary rock 43, 81, 88, 404, 568, 574, 575 sedimentary samples 547 sedimentary sulfides 685, 687 sedimentary sulfur cycle 572 sedimentary system 547 sediments 38, 39, 41, 45, 56, 62, 64, 124, 132, 139, 140, 141, 180, 184, 185, 188, 191, 196-198, 204, 227, 228, 270, 271, 336, 450, 568, 572, 576-579, 597, 602, 603, 632, 639, 649, 650, 689, 715-720, 972, 1012, 1019, 1033 seeds 555, 978, 1023 seismic monitoring 273 selected ion monitoring 153, 154 selective fluorination 404 seleniferous shale 624 self-diffusion 1004, 1020, 1023 self-neutralization 661 Semarkona chondrite 688
Subject Index semi-conductor laser semipermeable membrane semipermeable Pt semisynthetic sensitivity
775 168, 1006
1008 121 103, 108, 109, 130, 132, 144-146, 155, 15Z 284, 29Z 362, 364, 375, 395, 398, 399, 401, 453, 460, 468, 500, 508, 535, 550, 632, 634, 658, 666, 679, 680, 691, 69Z 698, 713, 723, 725, 74Z 773, 782, 832, 838, 839, 841, 84Z 848, 851, 854, 880, 926, 944,957,959,964 sensitivity estimates 482 sensitivity limitation 108, 109 sensitivity optimization 154 sensitivity specification 842 septum 11,98, 208, 216, 217, 225, 550, 616 septum coring 206 septum tubes 206-209, 211, 215-217 septum-sealed 17, 29, 98 sequential diffusion 324 Se-rich environment 626 seriously contaminated 281 serpentine 52 serpentinized peridotite 123, 889 224-226 serum bottles shaking cycles 26 shale 39, 43-46, 53, 54, 56, 58, 140, 354, 580, 597, 624, 1023,1025 shell 6,69,1011,1028,1035 22,50Z 516,517 shrubs Si isotope standard 533 Si metal matrix 654 594 Siberian mammoth side filament 861 signal dispersion (NMR) 120 significant contamination49, 320, 566 significant isotope / ic 591, 625, 636 fractionation 637 silane coating 3,186,188,322,357,430, silica gel 433, 439, 469, 549, 550, 617,643-645,809 silica gel matrix 643 silica glass capsule 1006 silica window 76 silicate carrier 223 silicate detritus 125
Subject Index silicate materials silicate matrix silicate melt silicate rock silicate standard silicate studies silicate system silicates (minerals) silicic acid silver capsule silver cyanide silver filters single collector silicon tetrafluoride silver boat single filament ion source site specific characterization site specific deuterium content site specific isotope ratio site specific nitrogen isotope analysis site specific parameters skimmer (cone) smectite SMOB SMOC Sn catalyst snow
snowmelt SO2 gas lifetime SO3 radical sodium-spiked zinc software (computer)
softwood soil
soil amino acid
1227 350 465 253 124,348,349 35Z 360 350 349,358,359 24,39,41,50,52,56,124, 348,350,359 1015,1032 332,503 333,334,340 312 139, 427, 645, 684, 692, 693, 697, 788, 790, 798, 800,803,804,926 405 337,341,342 643,644,867
soil carbonates soil digest soil extracts soil fraction soil gas soil horizon soil hydrolysates soil material soil microbes soil phosphate soil properties soil samples soil scientists soil solutions soil sulfate soil sulfer compounds soil water
103 896 110, 111 390-399 113 133, 550, 551, 728, 730, 796 41, 44-49, 52, 53, 60, 1017,1032 606 606,609,877,904,905 493 3,308,331,536,775 306,885 552 238 14 33, 115, 168, 169, 284, 286, 294, 299, 311, 599, 731, 798, 833, 852, 856, 954,964 514-516 3,10,15,38-61,172,173, 179-181, 184, 188, 197, 203, 312, 315, 322, 329, 341, 344, 428, 450, 559, 585, 586, 588, 590, 591, 649, 745, 788, 789, 803, 1013,1033 173
soil zone soil-respired CO2 solar absorption solar furnace solar material solar nebula solar neutrino experiment solar radiation solar system solar system (early) solar system environment solar system oxygen solar wind Solenhofen limestone solid hydrous substance solid standard soluable nitrate reductace (NaR) solute interaction effects solvent solvent extraction / extracts solvent impurities solvent storage solvent system solvent-insoluable sonification Soufre de Lacq
180,183 636,638 314, 315, 320-323, 344, 345 588 55,817,900 541,586,590 344 586,590 291,649 58 59O 3, 10, 15, 22, 41, 44, 54, 55, 344, 586-591, 626, 795,1013 42,306,315,341 322,345 590,591 586,590 2, 15, 22, 29, 54, 55, 273, 49Z 1014,1035 312 183 38O 1023 1023 229 772 377 236, 361, 628 229, 627, 632, 689, 710, 715 229 977 688 892 481 209, 218, 476, 478 344 1016 107, 109, 110, 113, 117, 165, 185, 188, 191, 19Z 198,572,576,70Z 708 113, 191, 230, 572, 576, 579,583,584,717 155 1041 118 233 41 884,900,901,910,981
1228 source environment source filament source tuning sources and sinks South Pacific Soxhlet apparatus Soxhlet distillation Soxhlet extractor / tion Soxhlet system space shuttle mission space technology spark discharge conversion spark source of MS spatial resolution
spectral interference spectrographic spectroscopic interference sphalerite
Subject Index 232 368 279 179, 272, 273, 375, 378, 379, 390, 399 559 44, 513, 517, 572 513 191, 513, 516, 517, 579, 580 197 382 866 404
sputtering fractionation sputtering ion source sputtering procedures sputtering process sputtering region square peak SrC12-NH4OH reagent stable isotope methodology stainless steel filter standard
790, 795 8, 405, 407, 439, 444, 453, 459, 464, 471, 482, 483, 495 ,496, 547, 551, 652, 659, 682, 691, 747, 754, 755, 758, 788, 972 706-708, 717, 727, 832 11 726
441, 447, 465, 466, 549, 568, 570, 599, 884, 900, 901, 910, 981, 1025, 1030 spike compounds 738 spike material 736, 738, 821, 830, 930 spike solution 628, 633, 634, 640, 641, 736 spiked standard 634, 636 spin-lattice (NMR) 105, 113 splitless injection 171, 159 spring water 142, 209, 649 sputter cleaning 688, 794 sputter rate 535, 536, 655, 673, 793 sputter site 652, 657, 661 sputter yield 793 sputtered area 789 sputtered atoms 535, 536, 672, 790, 792795 sputtered material 655 sputtered neutrals 793 sputtered pit / crater 549, 799 sputtering / sputtered 405, 532, 537, 549, 652, 653, 655, 664, 671, 675, 678, 680, 754, 755, 792794, 843, 862 sputtering behavior 793 sputtering characteristics 795 sputtering effects 848 sputtering event 674
standard standard standard standard standard standard standard standard
standard
670, 674 123, 789 664 670 661 299 211-214,218, 220-222 178, 324
54 6, 15, 16, 21, 34, 35, 44, 132, 133, 136-138, 155, 157, 171, 191, 196, 204, 209, 218, 240, 241, 255, 274, 279, 282, 297, 301, 303, 308, 311, 312, 325, 335, 392, 502-506, 523, 547, 599, 604, 606, 623, 627, 631-638, 646-653, 659, 668-676, 681, 682, 685, 687, 690, 702-723, 768, 772, 776, 798, 799, 830, 832, 842, 850, 873906, 908, 913, 929, 932, 935, 941, 944, 945, 953, 966-990, 994, 1045, 1051, 1052 additions 125, 325 analysis 300, 647, 681 artifacts 945 compound 193, 195, 201 conversion 337 data 681, 954 definition of 6 307 deviation 7, 45, 50, 54, 129, 130, 132, 141, 145, 146, 174, 175, 199, 207, 219, 234, 243, 278, 279, 298, 290, 301, 302, 324, 334, 337, 383, 384, 396, 478, 488, 493, 494, 506, 511, 515, 520, 534, 601, 615, 621, 646, 685, 726, 735, 738, 765, 799, 801, 882, 887, 890, 954, 955, 977, 978, 989 error 100, 303, 325, 328, 480, 481, 505, 520, 648, 744, 889, 965 fluorination 413
standard line standard gas
35, 165, 300-302, 310, 311, 364, 392, 393, 408, 501, 503, 543, 855, 898,
1229
Subject Index
standard heterogeneity standard isotope abundance standard isotopic scale standard manifold standard material
978, 1051 672 101
953 274 259, 261, 311, 392, 547, 648, 681, 682, 759, 768, 784, 795, 798, 803, 944, 946,968 651 standard matrix / ces standard measurement 202, 364, 702, 716, 720, 724,873 5, 15, 70, 164, 31G 328, standard method 348,615 101 standard mixture 60, 319, 357, 364, 374, standard procedure 524,610,646,852 Standard Reference Ma- 946 terial (SRM) 44, 96, 710, 712 standard samples 208, 219, 223, 317, 324, standard solutions 633, 634, 648, 704, 708721 556 standard state conditions 5O standard system 275,280,294,302,368 standard tank 35, 39, 49, 53, 351, 357, standard technique 703,716,852,853 154 standard tool 729, 733, 734, 744, 834, standard uncertainty 876, 883, 887, 893, 89Z 899, 901, 904-906, 916, 937,945 149,491,969 standard value 6, 20-22, 28, 34, 35, 92, standard water 280, 408, 476-478, 481, 850,885 2, 20, 32, 35, 112, 310, standardization 476, 668, 701-703, 708716, 723, 854, 879, 880, 886,896,968,973 16, 133, 243, 249, 282, standardized 309, 610, 632, 894, 945, 952-954 standard-sample brack- 284, 288, 702, 703, 710, eting 711, 713-716, 723 starch 200 stars 229, 755 static batch method 5, 9 static charge 549, 845 static dilution 392
static electricity discharges static gas cell static mass discrimination effect static mass spectrometer static measurement static mode static multicollection static reduction static vacuum stationary phase stationary phase wax statistical normalization statistical outlier steam distillation steel SRB1098 standard steel SRM368 standard steel standard stepped combustion stepped heating stepped pyrolysis step-wise fluorination steroids (a)stigmatic focussing storage conditions stratosphere stratospheric 03 stretching motions stretching transitions strong acid resin strong base resin structural anhydrous feldspar styrene divinyl benzene sublimation sublimation rate subsequent conversion substantial isotope fractionation Sudan-1 sugar sugar fermentation sugar-beet suite of standards sulfanilamide
1041 770 699 1, 70, 259, 265, 267, 268, 348-350, 358, 362, 460, 468,817,818 697 363,652 847 9 528 155, 157, 160-163, 168171 163 309 954,955 189, 312, 315, 316, 318320,326,327 350 259 795,804 257, 258, 260, 266-271, 362,817 49,256-271,349,350 262,264 56,443,445,453 163,170 664,846,847 208,228 291,377-386,390,399 377, 378, 380, 382, 385, 386,389 766,767 780 639 489,617,641 265 17, 19, 125, 327 12, 13, 548, 1020 1023 575 359 338,339 4, 32, 116, 332, 883, 896, 897,913 117 112 279 599
1230 591 585, 559, 569, 573, 575, 585, 96Z 993, 994, 1012, 1018,1033 sulfate standard 967 568,619 sulfide ores 723,985 sulfide standard 125, 231, 232, 234, 236, sulfonic acid 32Z 553,554 450 sulfosalt minerals 552 sulfur chemistry sulfur compound oxida- 543, 559, 568, 570, 574, tion 595, 1018 sulfur cycle 538, 572, 596 sulfur hexafluoride 545 sulfur isotope methodo- 540 logy sulfur isotope standard 978, 981, 984 sulfur oxyfluoride 546 sulfur trioxide 967 sulfuric acid 32, 305, 312, 314, 317, 318, 342, 346, 347, 553, 607-609 sulfuric acid catalyst 609 sulfurous acid 553 sulphamic acid 237-255, 315, 318, 330 sulphamic acid (dry) 237 summary of experiments 345 suprasil window 433 surface contaminants 593, 688, 793 surface water 2, 32, 156, 203, 204, 306, 541, 678, 680, 950 surficial environment 180, 625 (a)symmetric isotopomer 380, 382, 387-389 synthetic 121, 599 synthetic Ag2S 461 synthetic air 1040, 1041 synthetic carbonates 241, 253-255, 1016, 1023 synthetic compound 881 synthetic data 633, 634 synthetic fluid 2 synthetic gases 947 synthetic glass 903 synthetic goethite 56, 57 synthetic material 120, 506, 880, 947, 1023 synthetic mineral 1001 synthetic mixed matrices 673 synthetic (isotope)mix- 526, 922-924, 926, 930, ture 932, 933, 940, 941, 947 synthetic molecular 806, 808 sieves synthetic nitrates 336 synthetic origin 103, 116, 118 synthetic polymer 469 sulfate desorption sulfate reduction
Subject Index synthetic precursors 118 synthetic process 235 synthetic product 175 synthetic pyrite 571 synthetic reagent 599 synthetic sample 119, 721 synthetic SiC 260 synthetic solids 805 synthetic solution 721 synthetic source 119 synthetic spectra 779 synthetic standard 123, 135, 599, 647 synthetic sulfanilamide 599 synthetic sulfate 599,601 synthetic sulfur com59Z 599 pounds 173 synthetic testosterone 891 synthetic wines 805,806,810,813 synthetic zeolite synthetically produced 883 substances 9-11, 26, 54, 55, 206, syringe 212-214, 221, 225, 571, 1013 213 syringe caps 207 syringe filter 213,214,21Z 219 syringe precipitation technique T table sugar tail contribution tandem filters tank air tap water tap water vapour taphonomy tar tarry products target chemistry TCA cycle technology Teflon Teflon bag Teflon cap Teflon ferrules Teflon filters Teflon holders Teflon liners Teflon matrix material Teflon membrane Teflon O-rings
332 33 562 276,279,300 34,90,454 95 483 4,813 5OO 690 200 928 323,345,469 19,21 207 295, 414, 415, 448, 454, 457 556,560-562 513,514,522 3,313 17 346,347 293
1231
Subject Index Teflon plug Teflon reaction tube Teflon seal Teflon sheets Teflon tubing Teflon-lined cap tektite temperate soil temperature drift temperature monitoring terrestrial terrestrial apatite standard terrestrial carbon terrestrial contamination terrestrial ecosystem terrestrial environment terrestrial standard terrestrial / land plants Tertiary Tesla coil testing environment tetramethyl urea tetrathionate thawing-pumpingfreezing cycles theoretical peak therapeutic drug thermal conversion thermal conversion EA (TC/EA)) thermal decomposition
10 446 54,454 543 316,550,583 19L 198 534,535 1013 1039 418,448 2,25,123,139 687 178 235,236,362 180,290 235 68G 720,724 180,200,273,1014 578 1022 747 111, 112, 117 553 55 270 122 1041 334
25G 33G 343, 398, 542547 thermal desorption 559 thermal diffusion 548,836,1024 thermal ion source 130,657,857,907 thermal ionization 123, 124, 128, 138, 147, 626, 629, 635, 712, 71Z 720, 747,789,793,857 thermalionizationmass 123, 12Z 129, 133, 135, spectrometry/TIMS 13G 138, 139, 141-152, 450, 540, 543, 549-551, 55Z 559, 561, 562, 614, 623-651, 669, 689, 692694, 700-702, 709, 710, 712, 713, 715-718, 724, 725, 739, 741, 795, 796, 798-800, 802, 820, 828, 839, 861, 867, 923, 1012, 1019 TIMS analyses 701, 796 TIMS data 699,700,712,802
TIMS filaments TIMS instrument TIMS measurement TIMS reference values TIMS results TIMS source TIMS studies TIMS values time-of-flight (TOF) Titan mission thermodynamic (TD) data TD fractionation TD framework TD function TD isotope effect TD principles TD properties TD relationship thermodynamically stable thermogenic methane thermogravimetric analysis thermometric applications thermometry thermostatic bath thin layer chromatography thiourea Thode reduction Ti-alloy autoclaves time window tin boat tin capsule tissue tissue water Toepler pump toluene toluol-ethanol tomato plants tooth / teeth tooth apatite tooth enamel topaz
623 626, 627, 645, 646, 650, 692 133, 138, 698, 709, 710, 713,716, 720,739 701 740 644,645 713 133,802 664,692,693,789,804 758 402 309 444 994 156,158 203 379, 401 603,994 566 291 49 484 1005 485 195 899,902 560,573 1026 168 328,339,341,347 318, 321, 328, 332, 341, 342, 346, 476, 493, 503, 593,598 181, 185, 186, 18Z 191, 199,594,595 474 5, 6, 23, 49, 430, 814, 1051 44,54,231,475,516,582, 584 512-517 1013 180-183, 186, 439, 482, 595,1029,1030 180,183,1029 407,43Z 483,495,496 467,530
1232 total impurity content total inorganic carbon (TIC) total organic load total system drift tourmaline trace contaminants traceability
Subject Index 940 600
331 300 441,688,904,1031 11,12,965,984 833, 883, 908, 909, 911915, 919, 920, 926, 928944,949 875,914,929-937 traceability chain 540 tracer compounds 160,161 tracer dilution 6,161,316,338,540,621, tracer study 649,852 168,171 trailing peak transbasalt 85 882 transfer standard transient sputtering loci 674 1021 transition metal catalyst 813 transition metal ionexchanged zeolite 312 treated waste 528,531 treatment of waste 500,539 tree ring cellulose 497 tree ring record 516 tree ring research 180,508 tree ring study 180, 512, 517, 838, 1033, tree rings 1036 triethanolamine (TEA) 558 199 trifluoroacetic anhydride 195 trimethylsilyl ether 424, 429, 447, 538, 539, troilite 599, 959, 967 200 trophic interactions 177, 180, 181, 191, 200, trophic level 201 181, 187 trophic relationship 191 trophic structure 191 trophic web analysis 186 tropical regions 43 tropical soil 291, 377-383 troposphere 183 tropospheric CO2 377, 378, 382, 383, 389 tropospheric 03 390, 394, 399 tropospheric N20 892, 910 TS limestone 820-822 tsetse flies 24, 59, 241, 612, 616, 949 tube cracker tube cracker (automated) 321 tuna oil 175
tunable IR laser tunable laser tunable radiation Tuttle-type vessel two stage reactor 1,1,1 trichloroethane
767 386, 388, 751, 758, 760, 761, 763, 765, 774-776, 780, 782 778 1025 175 615
U ultrasonic agitator / ion ultrasonic bath ultrasonic cleaning ultrasonic probe ultrasonic tank ultrasonication ultraviolet (UV) UV laser UV light UV photodissociation UV photolysis UV (ir)radiation UV spectrum UV/Vis spectrophtometer uncertainty propagation undisclosed contamination unidirectional conversion universal gas constant universe unpolarized light unstable reagent uranium furnace uranium hydride uranium oxide uranium reactor uranium reduction uranium turnings urban environment urea uric acid urine US Environmental Protection Agency
208, 209, 221, 558, 572 206-208, 516, 518, 560 432 41 197 197 76, 104, 377, 431, 433, 436, 437, 753 77, 433, 436, 458, 459, 466, 469, 495, 496, 715, 723, 725, 972 125, 384, 390, 466, 496, 496, 560, 565 377 377 236, 377, 639, 641, 646 466, 471 202 916 12 541 924 2 749, 754 340 5,59 6 5 5 5, 6, 13, 22-25, 29, 33-36, 66, 766 55 383 32, 191, 329, 344, 898, 899 191 2, 10, 19, 31, 88, 98, 102, 130, 174, 322, 595 305, 378
Subject Index
1233
V
W
Vacutainer
waste Br waste disposal waste line waste products water bath
16, 17, 206, 208-210, 215, 216, 226, 228 vacuum distillation 3, 4, 29, 54, 410, 412, 474 vacuum extraction line 74, 217, 350, 458, 493 vacuum fluorination ap-406, 409, 415, 416, 453, paratus / line 455, 456 valve mixing 33, 311 Van der Waals disper168 sion forces vanadium pentoxide/ 407, 544, 545, 593, 895, V205 973, 982 V205 combustion 988 V205 reagent 598 vapor phase equilibra- 205, 223-228 tion variable combustion con-972 ditions variable dispersion optics729 (non-)vascular plants 197, 198, 200, 516, 517, 518, 1014 Venus mission 758 vibrational energy 458, 766 virtual material 912 viscous flow 348-350, 836, 838, 853 Viton 76, 274, 275, 436, 463, 468 vitreous carbon 342, 502 Voigt profile 769 volatile fluoride 420 volatile hydrides 637 volatile meteoritic com- 233 pound volatile organic com153, 157, 159, 502, 615, pound 951 volcanic air samples 761 volcanic arcs 139 volcanic ash 140, 376 volcanic gases 349 volcanic glass 70 volcanic rock 85, 132, 133 voltage window 680 volume diffusion 261, 1004 VUV synchrotron radia- 757 tion Vycor (glass) 14, 1020 Vycor flask 1021 Vycor tube/vessel 10, 12, 14, 330, 333, 336, 1020, 1021
411 528 451, 456, 543 418, 419, 442, 455-457 20, 22, 25, 44, 207, 208, 513, 514, 529-531, 578, 615, 1014, 1015 27, 326 water chemistry 680 water contaminant 36, 68, 74, 496 water desorption 53, 74, 232, 475 water extraction 4, 6, 10, 13, 15, 66, 76, water reduction 850 2, 3, 28, 36, 52, 88-102, water vapor 165, 189, 2391 255, 275, 281, 291, 293, 296, 300, 323, 330, 340, 341, 525, 559, 794, 1014, 1020 619 water-soluable salt water-soluable solvent 4 wavelength modulation 773 spectroscopy (WMS) wax paper funnel 216 waxes 163, 195, 475, 517 weathered basalt 139 weathering 38, 39, 42, 140, 141, 445, 1028, 1032 weathering environment 1028 wet air 281 wet chemistry 482-489, 525 wetland 209, 291, 559 wetland runoff 210,212 whalebones 198 whales 181 wheat plants 593, 1014 white-mica 348, 351-357 whole ecosystem 192 whole rock matrices 973 whole tissue 201 Wien filter 657, 658 wildlife 594, 625 window materials 463 wine 2, 19, 112, 116, 117 Wilson-Wells method 771 wood 183, 188, 197, 457, 502, 507-521, 1033, 1035 wood density 510 woody plant tissue 201 woody plants 516 working environment 415, 863, 1017 working lifetime 199
1234
working standard
Subject Index
21, 28, 32-35, 117, 118, 311, 396, 398, 534, 879, 882, 908, 929-931, 978
111, 113, 392, 393, 543, 855, 916, 915,
X X-ray fluorescence xylem
46 473, 474
Y yeast cells
101
Z 53, 125, 189, 314, 318, 342, 379, 805-819, 1020, 1032 817 zeolite crystal 780 zero background 32, 33, 310, 311, 952 zero-enrichment zinc artifact 13 12 zinc hydroxide 11, 13 zinc oxidation zinc reactor 5 11, 14 zinc reagent 8-10, 22, 23, 29, 31-36, zinc reduction 66, 76, 889 zinc reduction (dynamic) 8 9,10 zinc shot 409, 441, 466, 535, 681zircon 683,787 463,467,495 ZnSe window 996,997 zoisite 697 zoom lens dispersion 187,200,201 zooplankton zeolite