FATE and TRANSPORT of HEAVY METALS in the
VADOSE ZONE Edited by
H. Magdi Selim Iskandar K. Iskandar
Project Editor:...
132 downloads
1218 Views
14MB Size
Report
This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!
Report copyright / DMCA form
FATE and TRANSPORT of HEAVY METALS in the
VADOSE ZONE Edited by
H. Magdi Selim Iskandar K. Iskandar
Project Editor: Acquiring Editor Marketing Managers: Cover design: Manufacturing Manager:
Sylvia Wood Skip DeWall Bamara Glunn / Jane Stark Jonathan Pennell Carol Slatter
Library of Congress Cataloging-in-Publication Data Fate and transport of heavy metals in the vadose zone / edited by H.M. Selim, 1.K. Iskandar p. cm. Includes bibliographical references and index. ISBN 0-8493-4112-4 (alk. paper) 1. Soils-Heavy metal content. 2. Heavy metals-Environmental aspects 3. Zone of aeration. 1. Selim, Hussein Magdi Eldin, 1944- .II. Iskandar, 1.K. (Iskandar Karam), 1938- . S592.6.H43F37 1999 628.5'5-dc21
98-26915 CIP
This book contains information obtained from authentic and highly regarded sources. Reprinted material is quoted with permission, and sources are indicated. A wide variety of references are listed. Reasonable efforts have been made to publish reliable data and information, but the author and the publisher cannot assume responsibility for the validity of all materials or for the consequences of their use. Neither this book nor any part may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, microfilming, and recording, or by any information storage or retrieval system, without prior permission in writing from the publisher. All rights reserved. Authorization to photocopy items for internal or personal use, or the personal or internal use of specific clients, may be granted by CRC Press LLC, provided that $.50 per page photocopied is paid directly to Copyright Clearance Center, 222 Rosewood Drive, Danvers, MA 01923 USA. The fee code for users of the Transactional Reporting Service is ISBN 0-8493-94694/99/$0.00+$.50. The fee is subject to change without notice. For organizations that have been granted a photocopy license by the CCC, a separate system of payment has been arranged. The consent of CRC Press LLC does not extend to copying for general distribution, for promotion, for creating new works, or for resale. Specific permission must be obtained in writing from CRC Press LLC for such copying. Direct all inquiries to CRC Press LLC, 2000 Corporate Blvd., N.W., Boca Raton, Florida 33431. Trademark Notice: Product or corporate names may be trademarks or registered trademarks, and are used only for identification and explanation, without intent to infringe. © 1999 by CRC Press LLC
Lewis Publishers is an imprint of CRC Press No claim to original U.S. Government works International Standard Book Number 0-8493-4112-4 Library of Congress Card Number 99-26915 Printed in the United States of America 1 2 3 4 5 6 7 8 9 0 Printed on acid-free paper
THE EDITORS
H. Magdi Selim is Professor of Soil Physics at Louisiana State University, Baton Rouge. He received his M.S. and Ph.D. in Soil Physics from Iowa State University, Ames, in 1969 and 1971, respectively, and his B.S. in Soil Science from Alexandria University in Egypt, in 1964. Professor Selim has published numerous papers and book chapters, and is a coauthor of one book and several monographs. His research interests concern the modeling of the mobility of dissolved chemicals and their reactivity in soils and groundwaters. His research interests also include saturated and unsaturated water flow in multilayered soils. Professor Selim served as associate editor of Water Ruource.J Ruearch and the SoiL Science Society ofAmerica JournaL He is the recipient of several professional awards including the Phi Kappa Phi, Gamma Sigma Delta Award for Research, and the Doyle Chambers Career Achievements Award. Professor Selim is a Fellow of the American Society of Agronomy and the Soil Science Society of America.
Iskandar K. Iskandar received his Ph.D. degree in soil science and water chemistry at the University of Wisconsin-Madison, in 1972. He is a Research Physical Scientist at the Cold Regions Research and Engineering Laboratory (CRREL) and a Distinguished Research Professor at the University of Massachusetts, Lowell. He developed a major research program on land treatment of municipal wastewater, and coordinated a number of research areas including transformation and transport of nitrogen, phosphorus, and heavy metals. His recent research efforts focused on the fate and transformation of toxic chemicals, development of nondestructive methods for site assessments, and evaluation of in situ and on-site remediation alternatives. Dr. Iskandar has edited several books and published numerous technical papers. He organized several national and international conferences, workshops, and symposia. He received a number of awards including the Army Science Conference Award, and CRREL Research and Development Award. Dr. Iskandar is a Fellow of the Soil Science Society of America.
PREFACE
During the past decades, phenomenal progress has been made in several areas of biology, ecology, health, and environmental geochemistry of heavy metals in soils. Prior to the 1960s, research was focused on enhancing the plant uptake or availability of selected heavy metals or minor elements (also referred to as micro nutrients) from the soil. More recently, concerns regarding heavy metal contamination in the environment affecting all ecosystem components including aquatic and terrestrial systems have been identified with increasing efforts on limiting their bioavailability in the vadose zone. Moreover, several mathematical models for predicting the forms of metals in soils and the mechanisms of transformations and transport have been developed and evaluated. Because of the concerns regarding the role of heavy metals in the environment, a series of international conferences was held to explore the emerging issues of the biogeochemistry of heavy metals in the environment. In June 1997, the Fourth International Conference on the Biogeochemistry of Trace Elements was held in Berkeley, California. The contributions in this book were presented in part in the special symposium focusing on the fate and transport of heavy metals in the vadose zone as part of this international conference. The first four chapters of this book are devoted to sorption-desorption processes of selected heavy metals in the vadose zone. Kinetics of trace metal sorption-desorption with soil and soil components is the focus of Chapter 1. Importance of slow reactions and sorption mechanisms are also emphasized. In Chapter 2, adsorption of nickel by various soils and their isotherms are discussed. Moreover, a general isotherm approach based on intrinsic soil properties such as cation exchange characteristics and specific surface area is developed. Chapter 3 provides an overview of the sorption-desorption, precipitation, as well as complexation processes for cadmium reactions in soils. A discussion of sorption nonequilibrium during cadmium transport and reversibility of sorption processes are highlighted. Chapter 4 provides a comprehensive treatment of single and multiple retention mechanisms of the linear and nonlinear type which are commonly used to describe sorption-desorption of heavy several heavy metals in soils. Examples include hysteresis, reversibility and ion exchange retention kinetics during transport in soils. In the next three chapters, complexation and speciation processes and their influence on heavy metal mobility are discussed in detail. In Chapter 5, factors influencing complex formation of copper are emphasized. The effect of humic and fulvic acids on the retention of copper by soils and minerals is also presented. In Chapter 6, two sorption models that describe heavy metal binding of copper with solid and dissolved organic matter are presented. The applicability of such models to describe copper retention during transport is assessed. Also, the bioavailability (accumulation and excretion rates and toxicity) of copper for earthworms is discussed in Chapter 6. The effect of dissolved selenium species including metal selenium complexes and other dissolved organic carbon on selenium forms and their retention behavior in soils is presented in Chapter 7. Bioavailability and retention of heavy metals and their mobility in the vadose zone are presented in Chapters 8 through 11. The bioavailability and mobility of several divalent heavy metals as affected by pH and redox conditions are the focus of Chapter 8. Meth-
ods for quantif}ring and predicting the influence on mobility are also illustrated. In Chapter 9, the mobility of lead in calcareous mined soils is presented. The effect of various reagents on the mobility of lead in the vadose zone under different pH and redox conditions is also evaluated. In Chapter 10, an overview of modeling of heavy metal retention is given, along with factors influencing their mobilization/immobilization when organic residue/sewage sludge amendments are incorporated in the vadose zone. The use of a multiple reaction modeling approach is illustrated and the effect on retention parameters when organic waste is incorporated to the soil was also presented. The significance of the rhizosphere and its role on trace element interactions in the soil-plant system is the focus of Chapter 11. In addition, a conceptual model describing the dynamics of trace element processes between plant roots and the soil in the vadose zone is presented. In Chapter 12, plantavailable concentration levels for selected heavy metals in the vadose zone based on several extraction methods is discussed. The potential of various extraction methodologies is also evaluated. In Chapter 13, the authors discuss case studies of metal contamination from emission sources and old abandoned sites. The site investigations, monitoring, and alternative methods for remediation are given. We wish to thank the authors for their contributions to this book. Weare most grateful for their valuable time and effort in critiquing the various chapters and in keeping our focus on the main theme of our topic on heavy metals in the vadose zone. Special thanks are due to Drs. C. Hinz (University of Goettingen) and Giles Marion (U.S. Army CRREL) for their help in reviewing Chapters 3 and 11. Without the support of the Louisiana State University and the U.S. Army CRREL, this project could not have been achieved. Finally, we wish to express our appreciation to Ann Arbor Press for their help. H. Magdi Selim I.K. Iskandar
CONTRIBUTORS
K. Bajracharya Resource Sciences Center Department of Natural Resources Block C, Gate 2, 80 Meiers Road Indooroopilly Queensland 4068 Australia
D.A. Barry Department of Civil and Environmental Engineering University of Edinburgh Edinburgh EH9 3JN United Kingdom Klara Bujtas Research Institute for Soil Science and Agricultural Chemistry of the Hungarian Academy of Sciences (RISSAC) P.O. Box 35 H-1525 Budapest Hungary Philippe Cambier INRA Science du Sol Route de St Cyr F -78026, Versailles France Rayna Charlatchka INRA Science du Sol Route de St Cyr F -78026, Versailles France
Stephen Clegg Department of Ecology and Environmental Research Swedish University of Agricultural Sciences Box 7072 S-750 07 Uppsala Sweden Francois Courchesne Departement de Geographie Universite de Montreal C.P. 6128 Succursale Centre-Ville Montreal, Quebec H3C 3J7 Canada J. Csillag Research Institute for Soil Science and Agricultural Chemistry of the Hungarian Academy of Sciences (RISSAC) P.O. Box 35 H-1525 Budapest Hungary Carolina Garcia-Rizo University of Murcia Department of Agricultural Chemistry, Geology and Pedology Faculty of Chemistry E-30071 Murcia Spain George R. Gobran Department of Ecology and Environmental Research Swedish University of Agricultural Sciences Box 7072 S-750 07 Uppsala Sweden
Antonius A.F. Kettrup Institute of Ecological Chemistry GSF-National Research Center for Environment and Health N euherberg/Munich Postfach 1129 D-85764 Oberschleissheim Germany R.S. Kookana Cooperative Research Center for Soil and Land Management CSIRO Land and Water, PMB No.2 Glen Osmond SA 5064 Australia A. Lukacs Research Institute for Soil Science and Agricultural Chemistry of the Hungarian Academy of Sciences (RISSAC) P.O. Box 35 H-1525 Budapest Hungary
E.V. Mironenko Institute of Soil Science and Photosynthesis Academy of Sciences of Russia Pushchino, Moscow Region 142292 Russia R. Naidu Cooperative Research Center for Soil and Land Management CSIRO Land and Water, PMB No.2 Glen Osmond SA 5064 Australia T. Nemeth Research Institute for Soil Science and Agricultural Chemistry of the Hungarian Academy of Sciences (RISSAC) P.O. Box 35 H-1525 Budapest Hungary
Luis Madrid Instituto de Recursos Naturales y Agrobiologia (CSIC) Apartado 1052 E-41080 Seville Spain
G. Partay Research Institute for Soil Science and Agricultural Chemistry of the Hungarian Academy of Sciences (RISSAC) P.O. Box 35 H-1525 Budapest Hungary
Mari P.J.C. Marinussen Wageningen Agricultural University Sub-Department Soil Science and Plant Nutrition P.O. Box 8005 6700 Wageningen The Netherlands
Carmen Perez-Sirvent University of Murcia Department of Agricultural Chemistry, Geology and Pedology Faculty of Chemistry E-30071 Murcia Spain
Josefa Martinez-Sanchez University of Murcia Department of Agricultural Chemistry, Geology and Pedology Faculty of Chemistry E-30071 Murcia Spain
Alexander A. Ponizovsky Institute of Soil Science and Photosynthesis Academy of Sciences of Russia Pushchino, Moscow Region 142292 Russia
Katta J. Reddy Department of Renewable Resources P.O. Box 3354 University of Wyoming Laramie, WY S2071
Y.T. Tran Department of Environmental Engineering University of Western Australia Nedlands W A 6907 Australia
S. Schulte-Hostede Institute of Ecological Chemistry GSF-National Research Center for Environment and Health N euherberg/Munich Postfach 1129 D-S5764 Oberschleissheim Germany
Irena Twardowska Institute of Environmental Engineering Polish Academy of Sciences 34 M. Sklodowska-Curie Street 41-S19 Zabrze Poland
H. Magdi Selim Sturgis Hall Agronomy Department Louisiana State University Baton Rouge, LA 70S03
Sjoerd E.A.T.M. Van der Zee Wageningen Agricultural University Sub-Department of Soil Science and Plant Nutrition P.O. Box S005 6700 Wageningen The Netherlands
Donald L. Sparks Department of Plant and Soil Sciences University of Delaware Newark, DE 19717-1303 Daniel G. Strawn Department of Plant and Soil Sciences University of Delaware Newark, DE 19717-1303 T.A. Studenikina Institute of Soil Science and Photosynthesis Academy of Sciences of Russia Push chino, Moscow Region 142292 Russia Erwin J.M. Temminghoff Wageningen Agricultural University Sub-Department Soil Science and Plant Nutrition P.O. Box S005 6700 Wageningen The Netherlands
M. Th. van Genuchten U.S. Salinity Laboratory USDA, ARS 450 W. Big Springs Road Riverside, CA 92507 Walter W. Wenzel University of Agriculture Institute of Soil Science Gregor-Mendel Strasse 33 A-lISO Vienna Austria Franz Zehetner University of Agriculture Institute of Soil Science Gregor-Mendel Strasse 33 A-lISO Vienna Austria
CO.'\'TEl'\TS
Chapter 1. Sorption Kinetics of Trace Elements in Soils and Soil Materials ............... 1
DanieL G. Strawn and Dona{J L. SparlcJ Introduction ....................................................................................................................... 1 Evidence for Slow Sorption and Desorption Reactions ................................................. 3 Diffusion-Controlled Kinetic Reactions ........................................................................... 8 Kinetics and Mechanisms of Adsorption Processes .................................................. 10 Kinetics and Mechanisms of Surface Precipitation .................................................. 18 Summary .......................................................................................................................... 24 References ........................................................................................................................ 25
Chapter 2. Adsorption Isotherms of Nickel in Acid Forest Soils ................................ 29 Franz Zehetner and WaLter W. wenzeL
Introduction ..................................................................................................................... 29 Adsorption ....................................................................................................................... 29 Definition ..................................................................................................................... 29 The Diffuse Double-Layer ......................................................................................... 30 Adsorption Mechanisms ............................................................................................. 30 Adsorption Isotherms ...................................................................................................... 33 Classification ............................................................................................................... 33 The Langmuir Equation ............................................................................................. 34 The van Bemmelen-Freundlich Equation .................................................................. 39 Case Study ....................................................................................................................... 40 Adsorption versus Precipitation ................................................................................. 42 Langmuir and van Bemmelen-Freundlich Isotherms ............................................... 42 Effect of Soil:Solution Ratio on Quantity-Intensity Relationships ......................... 46 Fractionation of Adsorbed Nickel .............................................................................. 50 Adsorption Density and General Adsorption Density Isotherms ............................ 52 Summary .......................................................................................................................... 54 References ........................................................................................................................ 55
Chapter 3. Sorption-Desorption Equilibria and Dynamics of
Cadmium During Transport in Soil ......................................................................... 59 R.S. Kookana, R. NaiJu, D.A. Barry, Y.T. Tran, and K Bajracharya Introduction ..................................................................................................................... 59 Processes Governing Fate of Cadmium in the Soil Profile ........................................... 60 Sorption ....................................................................................................................... 60 Factors Affecting Cd Sorption in Soils ................................................................... 61 Precipitation ................................................................................................................ 69 Kinetics of Cd Sorption .............................................................................................. 70 Sorption Behavior of Cd During Transport Through Soil Columns ....................... 71 Batch versus Flow-Through Systems ..................................................................... 71
Evidence of Sorption Nonequilibrium During Cd Transport Through Soil ........... 74 Asymmetrical Breakthrough Curves ...................................................................... 74 Flow-Interruption as a Test for Sorption Nonequilibrium .................................... 75 Model Fitting ........................................................................................................... 76 Mass Balance Check for Complete BTCs .............................................................. 77 Causes of Sorption Nonequilibrium During Transport ............................................ 78 Cd Transport Under Field Conditions and Its Modeling ......................................... 78 Desorption and Reversibility of Cd Sorption ............................................................ 80 Desorption of Specifically Sorbed Cd .................................................................... 80 Partial Reversibility of Cd Sorption from Calcite and Calcareous Soils .............. 82 Cd Desorption Kinetics ........................................................................................... 82 Sorption Reversibility in Flow-Through Experiments .......................................... 83 Summary .......................................................................................................................... 83 References ........................................................................................................................ 85 Chapter 4. Modeling the Kinetics of Heavy Metals Reactivity in Soils ...................... 91 H. Magdi SeLim Introduction ..................................................................................................................... 91 Linear Retention .............................................................................................................. 92 Nonlinear Retention ........................................................................................................ 93 Langmuir or Second-Order Kinetics ............................................................................. 94 Hysteresis ......................................................................................................................... 96 Irreversible Reactions ..................................................................................................... 96 Specific Sorption ............................................................................................................. 96 Multiple Retention .......................................................................................................... 98 Ion Exchange Retention ................................................................................................ 100 Kinetic Ion Exchange ............................................................................................... 102 Case Study ................................................................................................................. 102 References ...................................................................................................................... 105 Chapter 5. Copper Retention as Mfected by Complex Formation with Tartaric and Fulvic Acids ....................................................................................... Alexander A. PonimIJcflcy, T.A. StUdenilcina, and E. V. Mironenlco Introduction ................................................................................................................... Copper(II) Retention by Soils, Oxides, and Clays ................................................. Solution Complex Formation and Cu(II) Adsorption ............................................ Complexes of Cu(II) with Fulvic Acids .................................................................. Influence of FA and Humic Acids (HA) on the Retention of Cu(II) by Solid Phases ......................................................................................... Copper Retention by Soil (A Case Study) ................................................................... Kinetics of Cu(II) Retention .................................................................................... Cu(II) Retention Isotherms and Cation Balance .................................................... Evaluation of Na2EDTA Ability to Extract Retained Copper ............................... Effect of Tartrate and Fulvic Acid on Cu(II) Retention Isotherms ....................... Modeling of Cu(II) Retention (Exchange) by Soil.. ...............................................
107 107 107 109 110 III III III 112 115 115 119
Summary ........................................................................................................................ 121 References ...................................................................................................................... 122
Chapter 6. Copper Mobility and Bioavailability in Relation with Chemical Speciation in Sandy Soil ........................................................................ E.J.M Temminghoffi MP.J.G. MarinLMden, and S.E.A.T.M Van der Zee Introduction ................................................................................................................... Sorption Models ............................................................................................................ Parameter Assessment Sorption Models ...................................................................... Copper Speciation in a Copper Contaminated SoiL ................................................... Mobility .......................................................................................................................... DOC Mobility Enhanced Copper Mobility ............................................................ Field Site Accumulation in Soil ................................................................................ Bioavailability ................................................................................................................ Bioavailability for Soil Organisms ........................................................................... Field Site Accumulation by Earthworms ................................................................ Summary ........................................................................................................................ References ......................................................................................................................
Chapter 7. Selenium Speciation in Soil Water: Experimental and Model Predictions .................................................................... Katta J. Reddy Introduction ................................................................................................................... Speciation of Dissolved Se ............................................................................................ Experimental and Model Predictions ........................................................................... Dissolved Se Speciation with CuO .......................................................................... Dissolved Se Speciation with GEOCHEM ............................................................ Comparison ............................................................................................................... Future Research ............................................................................................................ References ......................................................................................................................
Chapter 8. Influence of Reducing Conditions on the Mobility of Divalent Trace Metals in Soils ............................................................................... Philippe Cambier and Rayna Charlatchlca Introduction ................................................................................................................... Controversial Studies on Soil-Plant Systems .............................................................. Formation of Insoluble Sulfides and Other Solubility Equilibria .............................. Role of Fe and Mn Oxides as Trace Metal Sorbents .................................................. Reducing Processes Change pH ................................................................................... Role of Soluble Organic Ligands ................................................................................. Transformation of Insoluble Organics ......................................................................... Summary ........................................................................................................................ References ......................................................................................................................
127 127 128 129 130 133 133 136 136 136 139 143 145
147 147 148 149 149 149 150 156 156
159 159 160 161 162 164 168 170 171 172
Chapter 9. Lead Mobilization in Calcareous Agricultural Soils ................................ 177
Carmen Pirez-Sirvent, JOde/a Martinez-Sanchez, and Carolina Garda-Rizo Introduction ................................................................................................................... Soil Formation Factors ................................................................................................. Environmental Conditions ........................................................................................ Nature of the Materials ............................................................................................ Transport ........................................................................................................................ Dissolved Load .......................................................................................................... Particulate Forms: Suspended and Bed Loads ....................................................... Geochemical Processes ................................................................................................. Mobilization-Physical Weathering-Hydration Relations ....................................... Soluble Pb-Adsorbent Precursor Ratio ................................................................ Bicarbonated-Acidic Water Interaction ................................................................... Acid Water-Mineralized Particulate Material-0 2-C0 2 Interaction ................... Acid Water-Carbonated Particulate Material-0 2-C0 2 Interaction .................... Pb Sorption-Desorption ........................................................................................... Mobility .......................................................................................................................... Provoked Pb Mobility: Speciation Study ................................................................ Mobility in the Vadose Zone .................................................................................... Pb Assimilation by Plants ......................................................................................... Conclusion ..................................................................................................................... References ......................................................................................................................
177 178 178 178 180 182 186 186 187 187 188 189 189 190 191 191 193 195 196 197
Chapter 10. Metal Retention and Mobility as Influenced by Some Organic Residues Added to Soils: A Case Study ....................................... 201
LUM Madrw Introduction ................................................................................................................... 201 Soil as a Sink for Trace Metals ................................................................................. 201 Modeling Approaches for Retention of Metals by Soils ......................................... 202 Metal Concentrations in the Soil Solution ................................................................... 204 Factors Causing a Reversal of Immobilization ............................................................ 205 Interaction with Natural Organic Matter .................................................................... 207 Effect of Organic Residues on Metal Solubility .......................................................... 209 The Case of Sewage Sludge .......................................................................................... 210 A Mediterranean Concern: Olive Mill Wastewater .................................................... 211 Setting Up the Problem ............................................................................................ 211 Effect of OMW on Metal Retention Properties of Soils ........................................ 212 OMW in the Aqueous Phase as a Mobilizing Agent of Insoluble Metal Forms ......................................................................................... 215 Summary ........................................................................................................................ 218 References ...................................................................................................................... 219 Chapter 11. The Rhizosphere and Trace Element Acquisition in Soils .................... 225
George R. Gobran, Stephen Clegg, and FrancoM Courchune Introduction ................................................................................................................... 225 History ....................................................................................................................... 226
Rhizosphere - Defmitions ........................................................................................ 226 Methods of Rhizospheric Study ............................................................................... 226 Rhizodeposition ............................................................................................................. 228 Root Distribution and Longevity ............................................................................. 228 Belowground Carbon Flux ....................................................................................... 229 Exudates in the Rhizosphere .................................................................................... 229 Acid-Base Changes in the Rhizosphere ................................................................... 229 Rhizospheric Feedback Loops ...................................................................................... 232 Regulating Processes ................................................................................................ 232 Element Supply and Mobility in the Rhizosphere .................................................. 233 Microbial Activity and Element Accumulation in the Rhizosphere ....................... 234 Case Studies ................................................................................................................... 235 The Conceptual Model ............................................................................................. 236 Field Site and Treatments ......................................................................................... 236 Soil Fractionation ...................................................................................................... 236 Chemical Properties of the Soil Fractions ............................................................... 237 Weathering in Bulk and Rhizosphere Soil .............................................................. 239 Tree Growth and Rhizosphere Chemistry ............................................................... 242 Implications and Future Research ................................................................................ 242 References ...................................................................................................................... 245 Chapter 12. Distribution of Ecologically Significant Fractions of Selected Heavy Metals in the Soil Profile ............................................................. 251
T. Nemetb, K. Bujtd.1, J. CdUlag, G. Pdrtay, A. Lukdcd, and M. Tb. van Genucbten Introduction ................................................................................................................... 251 Sludge Application .................................................................................................... 252 Adsorption and Mobility .......................................................................................... 252 Extractions and Bioavailability ................................................................................ 253 Case Study ..................................................................................................................... 254 Nitric Acid Extraction ................................................................................................... 256 .-\AAc-EDTA Extraction .............................................................................................. 260 Concentrations in Soil Solution .................................................................................... 262 Movement ...................................................................................................................... 266 Summary ........................................................................................................................ 260 References ...................................................................................................................... 270 Chapter 13. Heavy Metal Contamination in Industrial Areas and Old Deserted Sites: Investigation, Monitoring, Evaluation, and Remedial Concepts .......................................................................................... 273
I
Irena Twardowdka, S. ScbuLte-Hodtede, and AntoniUd A.E Kettrup Introduction ................................................................................................................... Impact of Long-Term Stack Emission ......................................................................... Site Characteristics ................................................................................................... Site I: Nowa Huta n/Cracow, Area Adjacent to the Sendzimir Steelwork Complex, Poland ...........................................................
273 274 274 274
Site II: Irena Glasswork, Inowroclaw, Poland ..................................................... Soil Enrichment with Heavy Metals in the Areas Impacted by a Long-Term Stack Emission .............................................................................. Screening Survey and Methods ............................................................................ Metal Distribution in Soil vs. the Duration and Extent of Emission ................. Barrier Capacity of a Surface Soil Layer ............................................................. Heavy Metal Binding Strength and Mobility in Soils ......................................... Monitoring Program Requirements for Risk Assessment from Large-Area Soil Contamination by Trace Metals from Anthropogenic Sources ............... Evaluation of a Large-Area Deserted Industrial Site ................................................. Site Characteristics ................................................................................................... Sources of Heavy Metal Contamination in the Area ........................................... Monitoring Strategy ................................................................................................. Survey of Transfer Pathways and Risk Receptors .............................................. Human Risk Potential Assessment .......................................................................... Approach to Human Risk Potential Assessment ................................................. Applied Model: Quantitative Exposure Assessment (QEA) .............................. Remedial Concepts ........................................... ~ ............................................................ Summary ........................................................................................................................ References ......................................................................................................................
281 287 287 291 292 293 298 300 300 300 302 302 304 304 306 316 319 319
Index .............................................................................................................................. 323
CHAPTER I
Sorption Kinetics of Trace Elements in Soils and Soil Materials Daniel G. Strawn and Donald L. Sparks
INTRODUCTION Environmental contamination resulting from the extensive use of metals and semimetals in industry, agriculture, and in manufactured products has magnified the threat of toxicity for plants, animals, and society. Since soils and sediments have a large capacity for sorbing trace elements, an understanding of metal reaction mechanisms with natural materials is critical. Many studies have appeared in the literature on various aspects of metal sorption. Results from these studies have been used to develop government regulations, devise cleanup strategies, and develop models that predict the fate of trace elements in the environment. However, in conducting these studies researchers often overlook two important aspects: (1) the length of time soils are exposed to a contaminant (residence time) in the laboratory is relatively short compared with the much longer residence times that exist in field contaminated soils, and (2) the kinetics of metal sorption and desorption are often slow. These oversights lead to improper evaluation of contaminant behavior in the environment, resulting in regulations that may be improper, and models and remediation strategies that may be unsuccessful. This chapter will investigate the effects of residence time (aging) and slow kinetics on sorption and desorption reaction mechanisms of metals with soils and soil materials (e.g., clay minerals, metal oxides, and organic matter). Such information is important, and can be used in combination with transport models to predict the fate of trace metals through the vadose zone, and can provide information on metal bioavailability and speciation. Trace elements exist in the soil as either aqueous species, as structural elements in solids, or sorbed onto the surfaces of soil materials. While many of these trace elements are present naturally in the environment, their indigenous levels are usually nonthreatening. The buildup of these elements to dangerous levels is a result of commercial use and disposal practices. The following are a few examples of common sources of contamination: disposal of batteries that contain Pb, Cd, and Hg; exhaust from automobiles that
2
Fate and Transport of Heavy Metals in the Vadose Zone
burn gasoline with Pb additives; application of pesticides that contain Pb and As, e.g., Pb3 (As0 4)2; the use of Pb in paint; trace elements which are used in manufacturing that end up in waste disposal and the environment from either discarding the product or as a by-product of the manufacturing process; desiccation of agricultural runoff water in ponds which results in Se and As concentrating to dangerous levels; disposal of sewage which contains several trace elements, in particular heavy metals; and mine drainage which is often acidic and can increase the mobility of metals. Scientific studies have clearly shown that exposure to metal contaminants at higher than natural levels is toxic. As a result, many past uses and disposal practices of metals are now illegal, and trace element contamination of the environment is now regulated more closely. However, due to the relatively low solubility of many trace metals, and often strong sorption to soils, environmental contamination persists, and the threat from contaminants remains a problem that merits continued scientific investigation. While toxicity from trace elements, and their presence in the environment at dangerous levels are well-established facts, the questions remain: how does one remediate contaminated soils effectively, and how can significant risks be accurately evaluated? Finding effective answers to these questions hinges on a clear understanding of the behavior and interactions of trace elements with soils. In particular, an understanding of slow desorption and release kinetics from environmental settings which have been contaminated for long periods is critical. For example, Smith and Comans (1996) conducted sorption and desorption experiments on Cs contaminated sediments. They found that failure to include slow reactions in their model gave much lower estimates of the remobilization potential of the Cs. They concluded from model fits that sorption half-lives were between 50 and 125 days, and desorption half-lives were on the order of 10 years. Many studies rely on an equilibrium approach to predict the retention of contaminants on natural materials and subsequent migration through the vadose zone. Researchers often focus on determining parameters such as distribution coefficients, and the maximum amount of sorption possible. These studies are often based on the contaminantsolid interactions over a short period (24 hours or less) because it is assumed that the reaction has reached completion (Griffin et aI., 1986). However, field soils are seldom, if ever, at equilibrium, often laboratory studies are also far from equilibrium, and slow sorption may change the distribution between solid and solution over a period of time (Smith and Comans, 1996; Sparks, 1998). This is primarily due to slow metal sorption and desorption kinetics. The failure to account for the slow kinetics results in either underpredictions of the amount of contaminants retained by soils and minerals, or overpredictions of contaminant availability in the environment. A better approach is to base mobility estimates, remediation strategies, and risk assessments on the true availability of the contaminant, which is often controlled by a rate-limited sorption reaction. Most soils are heterogeneous media that contain a host of different minerals, solids, and organic materials. Thus, the interaction of trace elements with soils is a heterogeneous process. Several possible sorption mechanisms have been proposed (Figure 1.1): diffusion into micropores and solids followed by subsequent sorption onto interior surfaces; sorption to sites of variable reactivity, including sites which involve different bonding mechanisms, i.e., inner-sphere vs. outer-sphere and monodentate vs. bidentate; and surface precipitation (Fuller et aI., 1993; Loehr and Webster, 1996; Scheidegger and Sparks, 1996). Due to the heterogeneity of soil, these processes can occur simultaneously. A
Sorption Kinetics of Trace Elements in Soils and Soil Materials
3
Figure 1.1. Schematic of soil particles illustrating the different types of sorption that are possible. See text for definitions.
measured sorption or desorption rate often reflects a combination of all of the sorption mechanisms. However, it is possible that one mechanism may dominate at a particular time in the sorption reaction and the measured rate is primarily an expression of that reaction rate. For example, outer-sphere complexation can precede inner-sphere complexation, which can precede surface precipitation. The significance of this continuum in sorption is that while many sorption and desorption reactions may appear to have reached equilibrium, in fact the reaction can be continuous, and the slow process will not be measured if the experimentalist studies a short reaction time. In such cases, important secondary processes which are slower than the primary process may be completely overlooked. Thus, predictions on the fate of the contaminant may be inaccurate. This can cause increased threats of toxic exposure, improper evaluation of risks, and/or misappropriation of valuable cleanup and public safety funds. To protect human health and the environment from overexposure there must exist effective cleanup strategies, accurate risk assessment technologies, and models that correctly predict the fate of trace elements. For these tasks to be accomplished, time dependent reactions of trace elements with soils must be taken into consideration. Thus, the goals of this chapter are to discuss the kinetics of trace element interactions with soil and soil components, including the importance of slow reactions and possible sorption mechanisms.
EVIDENCE FOR SLOW SORPTION AND DESORPTION REACTIONS There are two separate phenomena associated with slow kinetic sorption processes: (1) a continuous slow removal of the sorptive from solution (sorption), and (2) a slow release of the sorbate from the sorbent (desorption). The second of these phenomena, desorption or release, may be influenced by the length of time in which the contaminant is in contact with the sorbent; i.e., there may be a decrease in the ability of the sorbate to be removed from the surface with increasing incubation or residence time. As mentioned above, several hypotheses for the cause of these two phenomena have been proposed (they are discussed in detail in later sections). An early report on the effect of incubation time on desorption reactions of metals from soils was given by McKenzie (1967). It was observed that manganese nodules present in Australian soils accumulated a large amount of Co. To account for this selective accumulation, a continuous sorption reaction was hypothesized. To test this, McKenzie (1967)
4
Fate and Transport of Heavy Metals in the Vadose Zone
determined both sorption and desorption kinetics of Co on manganese nodules isolated from soils. He found that removal of Co from solution slowed considerably after two days, but the extent of desorbability showed a continuous decrease with increasing aging periods. Thus, Co that was sorbed would become increasingly resistant to desorption from the nodule with time, resulting in an accumulation over time. Sorption processes commonly come to a state of quasi-equilibrium rapidly, and many researchers terminate their sorption experiments at relatively short times. However, it has been shown that sorption is a continuous process, and that the sorption mechanism can change over time, with little additional uptake. For example, Nyffeler et aI. (1984) found that the distribution coefficients for Be, Mn, Zn, Co, and Fe sorption on particulate matter from surface sediments and sediment traps increased over the entire time of observation, 108 days (Figure 1.2), suggesting that sorption is a slow process. Similarly Bruemmer et al. (1988) found that Ni, Zn, and Cd uptake by soils was continuous for times up to 42 days; e.g., Ni removal from solution at pH = 6 was 12% in two hours and 70% in 42 days. Bibak et al. (1995) studied the retention of Co by various goethite polymorphs and impure goethite. They found that Co sorption behavior varied between the different polymorphs and minerals, but in all samples the Co uptake increased with contact time (sorption kinetics measured from two hours to 504 hours). McBride (1982) found that sorption of Cu on noncrystalline aluminum oxide increased over periods of weeks, and proposed that different bonding mechanisms were responsible for the slow sorption process. McLaren et al. (1983) studied the desorption of Cu from humic acid, ferro-manganese concretions, and montmorillonite. In the desorption procedure the sorptive solution was replaced by the electrolyte solution (no metal), the suspension was allowed to incubate for four hours, and then, new electrolyte solution was added. The repeated washing of the soil removed little of the Cu, demonstrating that Cu sorption was strong. Young et al. (1987) compared Cu sorption and desorption reactions on river sediments with Cr and Zn. They observed that sorption of Zn was complete in four hours, Cr sorption was far from complete after 48 hours, and Cu sorption kinetics were intermediate. In addition, Young et al. (1987) concluded that desorption was not irreversible as McLaren et al. (1983) found, but that the observed irreversibility was a result of the slow kinetics involved. This slow desorption phenomenon was also observed for phosphate by Lookman et al. (1995). They found that slow phosphate desorption from soils continued for up to 1,600 hours, and showed no signs of reaching a plateau. In fact, using a rate constant derived from a first-order fit of the slow reaction, they predicted that 500 days would be required for desorption of 90% of the phosphate. Several researchers have noted that not only are trace elements strongly sorbed and exhibit slow desorption kinetics, but that the rate of desorption decreases with increasing residence times. Padmanabham (1983) conducted desorption experiments ofCu from goethite and concluded that Cu was sorbed in two different ways: a fraction was associated with low bonding energy and the rest was associated with high bonding energy. It was observed that a gradual interchange with increasing incubation time occurs between the readily desorbed fraction (low energy) and the less readily desorbed fraction (high energy). Similar results were found by Kuo and Mikkelsen (1980), Schultz et al. (1987), and Backes et al. (1995), who showed that the desorption rate of several transition metals (Zn, Co, and Cd) from soils and soil components decreased with increasing
Sorption Kinetics of Trace Elements in Soils and Soil Materials --- --------
5
---~-"'------
10000000 1000000 o Fe
~
100000
'ai
10000
0
a:
_____ ---------------------------.:e
:eMn oBe
aCo .Zn
c 0
.';=
::s .c ·c "Iii
is
1000 100 10
o
20
40
60
80
100
120
Incubation Time (Days)
Figure 1.2. Effect of incubation time on the distribution coefficient (1
aging time. Their findings further support the hypothesis that a slow process occurs between trace metals and soils that affects the availability of the metal. Smith and Comans (1996) observed an increase in the slowly desorbed fraction of sorbed Cs from sediments with increasing incubation time. Modeling of their data using a two-compartment model (Figure 1.3) suggests that there exists an exchangeable sorbed fraction and a "fixed" fraction. Slow transfer between the two fractions was responsible for slow kinetics of sorption and desorption. Another method for determining the reversibility of sorption, or the effects of aging, is to measure the exchangeability of the sorbate using isotopic exchange. This approach was used by McLaren et al. (1986) who studied the sorption and desorption behavior of Co from soil components. For humic acid, a large proportion of the Co was isotopically exchangeable, even for sorption incubation times as long as 50 days. However, for a soil oxide (ferro-manganese concretions) and montmorillonite, the fraction of nonisotopically exchangeable Co increased continuously as sorption time increased. Comans (1987) determined that the isotopic exchangeability of Cd on illite was 100%, but required an equilibration time of seven to eight weeks. These results suggest that with increasing incubation time the association between the sorbate and the sorbent changes to that of a more stable complex, i.e., less readily desorbed. Hysteresis, or nonsingularity, is a phenomenon in which the sorption and desorption isotherms do not coincide because of a shift in the equilibrium point. Pseudo-hysteresis is often observed in systems that have a slower desorption reaction than the sorption reaction (or vice versa), and is a result of not carrying the experiment out for a long enough period, i.e., the system has not reached equilibrium (McBride, 1994, p. 91). For example, Comans (1987) found that sorption and desorption isotherms of Co on illite are
(i
Fate and Transport of Heavy Metals in the Vadose Zone
IAqueous H Exchangeable I _ktI
.J
Fixed
I
Rate-Limited Process
Figure 1.3. Schematic illustrating a modeling approach used to describe the slow transition of (s from a mobile (exchangeable phase) to a fixed phase. From Smith and (omans (1996), with permission.
singular (nonhysteretic) only if the desorption reaction is allowed to come to equilibrium, which took 54 days. Many researchers have found that the magnitude of hysteresis increases with longer sorption incubation periods. Ainsworth et al. (1994) found that despite increasing the desorption times from 16 hours to nine weeks, hysteresis persisted for Co and Cd sorbed on hydrous ferric oxide (HFO). They also found that Cd and Co displayed increasing hysteretic behavior upon aging from two weeks to 16 weeks (Figure lA, Cd data not shown), while Pb sorption/desorption behavior was reversible (Figure 1.5). Oftentimes it is observed that the amount of sorbate able to be desorbed decreases with increasing incubation time. For example, the results of McKenzie (1980) suggest that Pb sorption on Fe-oxides is a slow process, and can affect the total amount of Pb able to be desorbed. McKenzie (1980) incubated Pb on hematite and goethite for periods from one day to 28 weeks, and then extracted the samples with 2.5% acetic acid until no additional Pb was removed. A comparison of the Pb removal from the different samples showed that increasing the incubation time from one day to 28 weeks increased Pb retention by 50% on hematite, and 100% on goethite. Similar experiments were conducted by Bibak et al. (1995) for Co sorbed on iron oxides; they used a strong acid extractant to measure the amount of Co released from iron oxides incubated for different periods. Their results indicated that the percentage of Co released decreased with increasing aging times. The results of McKenzie (1980) and Bibak et al. (1995) suggest that many adsorption reactions that appear to be at equilibrium are undergoing slow transformations that decrease the amount of sorbate able to be desorbed. Hysteresis has also been observed in cation exchange reactions, where the exchange of one sorbed cation with another is not completely reversible, i.e., the forward and reverse exchange reactions do not result in the same isotherms. The hysteretic behavior of cation exchange is abundantly reported in the literature; an excellent critical review of this literature was published by Verburg and Baveye (1994). From a survey of the literature they were able to categorize several elements into three categories (Table 1.1). The elements in each category were found to show hysteretic exchange between groups, but not within groups. Verburg and Baveye (1994) proposed that exchange reactions are most likely a multistage kinetic process in which the later rate-limiting processes are a result of physical transformation in the system; e.g., surface heterogeneity, swelling hysteresis, and formation of quasi-crystals, rather than simply a slow kinetic exchange process where there exists a unique thermodynamic relationship for forward and reverse reactions. While this may be true in some circumstances, an apparent (pseudo) hysteresis also can result from slow sorption and desorption reactions, i.e., lack of equilibrium.
Sorption Kinetics of Trace Elements in Soils and Soil Materials
7
100 ~lll
~~
"0 (])
.0 .... 0
III
c
40
(])
~
a
1:1
c
0 ()
e e
e
odl
• ...
(])
Il..
0
~
2
~
4
Adsorption pH edge ~ aging time = 0 wk Desorption • aging time = 2 wk aging time = 9 wk C aging time = 16 wk
•
o
~.
~
h
8
6
12
10
pH Figure 1.4. Effect of sorption incubation time on the desorption of Co from Fe oxide. From Ainsworth et al. (1994), with permission.
1001------;:~[]T--r--____, "0 (])
.0
.... o
III
.0
-
a..
c::
(])
~
(])
a..
60 Adsorption pH edge ~ aging time = 0 wk Desorption aging time = 2 wk • aging time = 12 wk C aging time = 21 wk
o
12 pH Figure 1.5. Effect of sorption incubation time on the desorption of Pb from Fe oxide. From Ainsworth et al. (1994), with permission.
Regardless of the different reasons for hysteresis, we agree with the proposal of Verburg and Baveye (1994) that to be of practical value, kinetic models need to be complemented by detailed information on the mechanism (s) responsible for the slow kinetic reaction (s). The above discussion has presented many examples of slow kinetics for both sorption and desorption processes. Several of the researchers have speculated on the mechanism(s) that control slow reactions. However, these hypotheses are based primarily on macroscopic data, while sorption and desorption processes are microscopic phenomena. At
8
Fate and Transport of Heavy Metals in the Vadose Zone
Table 1.1. Classification of Cations into Three Groups in Such a Way that Hysteresis Has Been Documented in the Literature for Binary Reactions Involving Cations from Different Groupsa
a
Group 1
Group 2
Group 3
Na+ Li+
K+ Rb+ Cs+ NH/
Ca 2+ Ba 2+ sr2+ Mg2+ Mn 2+ Cu 2+ Ni 2+
From Verburg and Baveye (1994), with permission.
best, macroscopic investigations suggest a particular mechanism may be occurring; they provide little evidence that other mechanisms are not involved (Sposito, 1986; ChisholmBrause et aI., 1990). Despite this problem, reasonable predictions of sorption mechanisms based on macroscopic observations are often made. However, the uncertainty becomes clear when one observes the discrepancies in predictions of mechanisms between published reports on similar systems. For example, Ainsworth et al. (1994) predicted that the similarities of the ionic radii between Co and Fe suggested a coprecipitation mechanism as responsible for aging. This hypothesis was supported by their observation that as ionic radius decreased, i.e., Pb>Cd>Co, hysteresis increased. However, Bibak et al. (1995) predicted that the mechanism responsible for the slow reaction of Co on various iron oxides was diffusion. This prediction was based on a good fit of the data to a diffusion model. An important point to note about comparing these two systems is that in the experiments of Ainsworth et al. (1994) the initial Fe-oxide was amorphous and underwent recrystallization, while the Fe-oxides used in the experiments of Bibak et al. (1995) were crystalline and did not undergo a solid phase transformation. Such differences can have important consequences on sorption mechanisms. Despite this discrepancy, one can conclude from these studies that in order to better predict the mechanisms responsible for the slow kinetic processes, microscopic as well as macroscopic data are necessary.
DIFFUSION-CONTROLLED KINETIC REACTIONS Diffusion is an activated process driven by the necessity of a system to be at its lowest possible energy, i.e., uniformly distributed throughout space. Since soils are porous materials containing both macropores (>2 nm) and micropores «2 nm) (Pignatello and Xing, 1996), diffusion is a mechanism that can control the rate of sorption of trace elements on soils. These pores can be interparticle (between aggregates) or intraparticle (within an individual particle). Intraparticle pores can form during weathering, upon solid formation, or may be partially collapsed interlayer space between mineral sheets; i.e., vermiculite and montmorillonite. The rate of diffusion through a pore is dependent on pore size, particle size, tortuosity, chemical interactions, chemical flux through the soil, and whether the pore is continuous or discontinuous. Besides pore diffusion, solid-
Sorption Kinetics of Trace Elements in Soils and Soil Materials
9
phase diffusion is also a transport-limited process. Solid phase diffusion is dependent on the characteristics and interactions of the diffusant and the solid (Pignatello and Xing, 1996). Since there exists a range of diffusion rates in the soil, it follows that with increasing exposure time the fraction of contaminants in the more remote areas of soil particles (accessible via slow diffusion) will increase. This slow sorption phenomenon is often the explanation researchers use to account for the slow continuous sorption and desorption observed between metals and soil (Sparks, 1989; Burgos et aI., 1996). Bruemmer et aI. (1988) measured sorption and desorption of Cd, Zn, and Ni with goethite, a porous iron oxide known to have defects within the structure in which metals can be incorporated to satisfy charge imbalances. They found that the kinetics were described well with a solution to Fick's second law (a linear relation with the square root of time), and proposed that the uptake of the metal followed a three-step mechanism: "(i) adsorption of metals on external surfaces, (ii) solid-state diffusion of metals from external to internal sites,O and (iii) metal binding and fIxation at positions inside the goethite particle," suggesting that the second mechanism is responsible for the slow reaction (Bruemmer et aI., 1988). Similar observations on sorption of divalent metal ions were made by Coughlin and Stone (1995). They suggested that the slow sorption and desorption could be a result of slow diffusion that occurred because their synthetic goethite may have had an unusually high level of pores and cavities. Axe and Anderson (1997) also found that sorption of Cd and Sr could be characterized by a model which included two steps: a rapid reversible sorption step followed by a slow, rate-limiting process involving the diffusion of the cations through small pores existing along the surface. While the above examples have hypothesized that diffusion is the rate-limiting step based on good model fIts to data and some speculation, macroscopic sorption experiments are not defInitive proof of a mechanism (Sposito, 1989, p. 150). To give additional support to diffusion as a mechanism for sorption onto porous media, Papelis (1995) measured surface coverages of Cd and selenite on porous aluminum oxides using X-ray photoelectron spectroscopy (XPS). Papelis (1995) calculated the expected thickness of sorbed Cd and selenite from the total metal loss from solution using both external and internal surface areas. A good agreement was found between the calculated and the measured (using XPS) surface coverage thickness when the total surface area (i.e., internal and external surface area) was used. When the surface layer thickness was calculated without considering internal surface area, then the calculated thickness exceeds the thickness observed using XPS. Therefore, the most likely sorption mechanisms were sorption to external sites, diffusion of Cd into the internal structure, and subsequent sorption. While Papelis (1995) didn't measure the kinetics of the reaction, it seems probable that the sorption to the interior sites is slower than the exterior sites, and thus a slow kinetic sorption step would exist. Fuller et aI. (1993) combined kinetic sorption and desorption experiments with spectroscopic observations (Waychunas et aI., 1993) to conclude that the rate-limiting process in arsenate sorption by ferrihydrite is diffusion into the solid structure. Using X-ray
• Classical solid-state diffusion is a very slow process in crystalline structures, and usually only significant at very high temperatures (McBride, 1994, p. 28). In this case, solid state diffusion should be interpreted as diffusion processes through faults and micropores.
10
Fate and Transport of Heavy Metals in the Vadose Zone
0.12
60
0.1
__--------r---------r---A---~50
'0 0.08
40
CI>
1!i
'0.
e Il.
CI>
eo
o
III
~ 0.06
30 ~
...as
(5
::!!
CI>
u.
~
"C
.--------_......._.._..._...._....•..•....
0.04
_
............................
_ _ ...•....•...•.•..
~
-
....................... ....
20
:.!! 0
~/
0.02
/
/
./
10
0+----------.---------.----------.----------+0 o ~ 100 150 ~o Time (Hours)
Figure 1.6. Pore-space diffusion fit of As(V) adsorption density as a function of time for total (dark line), diffusion-limited (dotted line) and exterior surface components of adsorption (thin solid line). The solid triangles represent the adsorption data. Exterior sites are modeled based on equilibrium. From Fuller et al. (1993), with permission.
absorption fine structure (XAFS) spectroscopy, Waychunas et al. (1993) found that arsenate is sorbed predominantly as inner-sphere bidentate complexes, regardless of whether the arsenate was adsorbed post-mineralization of the ferrihydrite, or present during precipitation. Thus, at the pH of their study (8.00), arsenate surface precipitates were not formed. Slow sorption and desorption were explained as slow diffusion of the arsenate to or from interior surface complexation sites that exist within disordered aggregates of crystallites. The arsenate sorption and desorption kinetics (Figure 1.6) were explained well using a model which included two types of sorption sites: those easily accessible were described assuming equilibrium (thin solid line), while the sites which had limited accessibility (dotted line) were well represented by an equation which is based on Fick's second law of diffusion.
Kinetics and Mechanisms of Adsorption Processes Adsorption is a phenomenon in which matter accumulates at the interface between a solid phase and a solution phase; it is largely considered to be two-dimensional (Sposito, 1989, p. 132). Adsorption reactions are governed by the laws of thermodynamics: energy is conserved, and the entropy of a system increases to a maximum. These two concepts can be combined to create the Gibbs free energy (G) function. For a reaction to occur, the products must have a lower free energy than the reactants (~G < 0). This can occur by either a decrease in enthalpy, an increase in entropy, or both. It is important to note that a change in enthalpy can dominate the free energy function creating a negative ~G even when the entropy is decreased in the reaction, and vice versa. Therefore, an
Sorption Kinetics of Trace Elements in Soils and Soil Materials
11
adsorption process leads to an association between an ion and a surface, driven by the desire of the system to achieve an overall lower free energy. While thermodynamics can be used to determine if a reaction is favorable, it does not indicate the rate of the reaction, nor the pathways involved in arriving at the state with the lowest free energy. This information can be gained by measuring reaction kinetics. In real systems, such as soils and sediments where there exist several different types of sorption sites, reaction mechanisms and kinetics can be heterogeneous. In these systems kinetics plays an important role in the fate of trace elements since such systems are not at equilibrium, but are continuously undergoing chemical changes as they seek to produce the most stable species (Steinfield et al., 1989, p. 1). The change may be slow, resulting in the sorbate becoming less available with time (aging) (Koskinen and Harper, 1990), and can result in a change from one type of sorbed complex to another. This process is similar to the concept of the Ostwald-step rule: the first product in a precipitation reaction is that which has the highest solubility, followed by a slow continuous transformation to a more stable species (Stumm and Morgan, 1996, p. 807). An analogous process in adsorption would result in a multitude of adsorbed complexes, some of which may be in a metastable equilibrium state, undergoing continuous transformation to the most stable species. Evidence for this slow, continuous change to a more stable species is commonly observed for solid materials. Upon initial precipitation the solid is in an active form that has a disordered lattice (amorphous), and exists in a metastable equilibrium with the solution (Stumm and Morgan, 1996, p. 356). With time the solid slowly converts to the more stable inactive form. The inactive form is more crystalline-like, and has a lower solubility. This slow kinetic phenomenon may continue for geological time spans. An example is aragonite (a polymorph of calcite), which is found in rocks < 300 million years old. Aragonite is not thermodynamically stable, but forms under surficial temperatures and pressures, and slowly reverts to the more stable calcite (Blackburn and Dennen, 1994, p. 102). Waychunas et al. (1993), using XAFS data fitting, found that aging and continued polymerization of ferrihydrite resulted in a transformation of the number of linkages and interatomic distances to those suggesting a progression to the more ordered polymorph goethite. The slow transformation of a solid to a state with a lower free energy is often observed as an aging mechanism for precipitates, but transformations between sorption mechanisms is more difficult to distinguish, and little direct evidence exists for such processes. However, it seems reasonable to suggest that the energetics of sorption and desorption reaction processes are analogous to those of precipitation; i.e., kinetically limited by a transformation to the most stable sorption configuration (lowest ~G). Adsorption reactions occur via three different mechanisms: inner-sphere complexes, outer-sphere complexes, and diffuse ion (Figure 1.7, diffuse ion not indicated) (Sposito, 1989, p. 132). Outer-sphere bonds consist of a solvated ion that forms a complex with a charged functional group; the primary bonding force is electrostatic. An inner-sphere complex is partially dehydrated; the ion forms a direct ionic or covalent bond with the surface functional groups. A diffuse ion exists in the water layers near the surface, and is held by electrostatic attraction from permanent charges that exist in the solid structure. A major difference between the outer-sphere complex and the diffuse ion complex is in the strength of the electrostatic force, which is directly correlated to the proximity of the ion to the surface (McBride, 1994, p. 73). The type of sorption and bonding mechanism
12
Fate and Transport of Heavy Metals in the Vadose Zone
Metal
Oxygen H+
j
'H
a
Other Examples
aerD 0
"d Doa
Outer-Sphere Surface Complexes
Monodentate
Inner-Sphere Surface Complexes Bidentate
Figure 1.7. Schematic showing the different types of adsorption complexes that can occur on solid surfaces. See text for definitions. From Hayes (1987), with permission.
depends on several factors: (1) ionic radius, (2) electronegativity, (3) valence charge, (4) surface type, and (5) ionic strength of the sorptive solution. There are two major types of surface sites: variable charged sites, e.g., silanol and alumino!; and permanent charge sites that result from isomorphic substitution. To model surface complexation and understand the controlling mechanisms, scientists often assign a hypothetical bonding mechanism between an ion and a given surface. However, ions can bond to surfaces via several different mechanisms, and can undergo a continuous transition between adsorption mechanisms (Stumm and Morgan, 1995, p. 586). Waychunas et al. (1993) found that arsenate adsorbed onto ferrihydrate by both monodentate (30%) and bidentate bonding mechanisms. Bargar et al. (1996) used X-ray
Sorption Kinetics of Trace Elements in Soils and Soil Materials
13
absorption spectroscopy (XAS) to distinguish between outer- and inner-sphere sorbed Pb on CX-AI 20 3. They found that on the planar 0001 surface Pb-O-AI distances were consistent with an outer-sphere bond, while on the 1102 plane Pb was sorbed as an inner-sphere complex. Benjamin and Leckie (1981) conducted sorption experiments at several different loading levels and equilibrium pHs for Cd, Cu, Zn and Pb on amorphous iron oxyhydroxide. Their data suggested that there exist several types of bonding sites with variable bonding strengths, and that measured equilibrium constants are average values from these different types of sites. McBride (1982) found similar results on pure noncrystalline aluminum oxide using electron spin resonance (ESR) spectroscopy to study the change in Cu sorption mechanisms with time. He found that sorption involved sites of varying reactivity. The first reaction step was the rapid sorption of a low level of Cu; the second reaction occurred over several weeks and resulted in the uptake of a greater amount of Cu and ESR spectra distinct from the first reaction step. Such heterogeneity is enhanced in natural systems that contain materials with a variety of organic and inorganic surface sites. Adsorption reactions are often considered to form the most stable bond immediately, but commonly there are intermediates which can be metastable for long times. In fact, adsorption may consist of a series of chemical and physical reactions that may limit the overall reaction rate; i.e., ion and surface dehydration, breaking of a strong bond, bond formation, and surface diffusion (Stumm and Morgan, 1996, p. 761; McBride, 1994, p. 135). Hayes and Leckie (1986) and GrossI et al. (1994) used pressure-jump relaxation to measure the kinetics of Pb sorption on aluminum oxide and Cu(II) sorption on goethite, respectively. They found that the best fit to the data was obtained by fitting a kinetic model that included a transformation from outer-sphere to inner-sphere complexation. Their results also suggested that sorption behavior was biphasic, which they explained by suggesting that the slower reaction was a result of sites with lower affinities. This concept is similar to the high and low affinity site model proposed by Dzombak and Morel (1990, p. 92). While the kinetics of these reactions are quite rapid (reactions considered on a millisecond time scale), the demonstration of a multiple step adsorption mechanism rationalizes the hypothesis that in some systems one step may be slow enough to be responsible for the slow adsorption and desorption reactions often observed in soils (Sposito, 1989, p. 150). The kinetics of Pb sorption on y-AI203 are shown in Figure 1.8. These data show a fast initial reaction followed by a slow sorption reaction continuing for several hours. Such biphasic behavior is likely a result of sorption to sites of variable reactivity and/or diffusion limited sorption. Slow surface precipitation reactions can be ruled out because analysis of the radial structure function obtained using XAFS (Figure 1.9) does not exhibit any major features (e.g., second peaks indicative of second shell neighbors) beyond the primary Pb-O structural peak at -1.9 A (uncorrected for phase shifts) with long incubation times. Biphasic sorption reactions have also been observed in soils. An example is the result of Lehman and Harter (1984) who measured the kinetics of chelate-promoted Cu release from a soil to assess the strength of the bond formed. Their sorption/desorption data were biphasic, which they attributed to high and low energy bonding sites. They also found that with increased residence time, 30 minutes to 24 hours, there was a transition of the Cu from low energy sites to high energy sites (as evaluated by release kinet-
14
Fate and Transport of Heavy Metals in the Vadose Zone
--.
60
c:
50
. ~ 0
0 .;::;
::J
(5
C/)
E 0 ....
-"0 Q)
40 30
>
20
E Q)
c:
10
a..
0
0
• •
•
•
•
.n
0
2
4
6
8
188
190
192
Time (hours) Figure 1.S. Kinetics of Pb removal from solution by y-AI 2 0 3 • Ionic strength = 0.1, pH = 6.50, initial Pb concentration = 0.002 M.
"C
1.0 0.5
±::
0
Q)
:::I
c:
, -__~__~__________~70Days
g> ~
8 Days
E ....
~~------------
.E en c:
~
48 Hours
r-
24 Hours
o
2
3
4
5
R (A)
6
7
8
9
Figure 1.9. Radial distribution function (uncorrected for phase shifts) for Pb sorbed on y-A1 2 0 3 incubated for 24 hours to 70 days. Incubation conditions are the same as in Figure 1.8.
ics). Incubations for up to four days showed a continued uptake of Cu and a decrease in the fraction released within the first three minutes, which was referred to as the low energy adsorbed fraction. The results of Smith and Comans (1996), already mentioned, also showed that Cs sorption onto sediments is biphasic. They modeled exchange reactions assuming exchangeable and fixed fractions. The fixed fraction was assigned to Cs that was incorporated in the mineral lattice, i.e., predominantly specific exchange sites on illitic clay. The Cs adsorption mechanisms proposed by Smith and Comans (1996)
Sorption Kinetics of Trace Elements in Soils and Soil Materials
15
'were based on kinetic experiments, i.e., macroscopic observations. Kim et al. (1996) used nuclear magnetic resonance (NMR) spectroscopy to make microscopic observations of Cs sorption mechanisms on kaolinite, boehmite, silica gel, and illite. Their experiments coincide with those of Smith and Comans (1996), suggesting that Cs formed two distinct types of complexes on the surfaces of the minerals: inner-sphere and outersphere. The energy and stability of adsorbed species varies depending on the type of surface complex formed. It is generally accepted that surface complexes with more than one bond are more stable than complexes with a single bond (Stumm and Morgan, 1996, p. 276; McBride, 1994, p. 134), and likewise for inner-sphere vs. outer-sphere sorption (McBride, in Bolt, 1991, p. 168). One explanation for the increased stability of a multidentate bond over a monodentate bond may be the increased entropy gained from a more stable configuration (steric effect) (Steinfield et al., 1989; McBride, 1994, p. 80) . .w analogous phenomenon is the Chelate Effect; for example, the ~G for the ethylenediamine complex, a chelate ring with bidentate bonding to a cation, is lower than ~G of the diamine complex, which forms monodentate complexes with cations (Stumm and Morgan, 1996, p. 279, from Schwarzenbach, 1961). The lower ~G for the ethylenediamine complex means it is more stable. Since the enthalpies for the complexation of cations by the two chelates are similar, the lower ~G is a result of an increased entropy for the bidentate ring complex; as mentioned above, this phenomenon is often referred to as a steric effect or configurational entropy (Stein field et al., 1989, p. 250; McBride, 1994, p. 80). Since the reactive sites on minerals (silanol and aluminol sites) and organic matter (carboxyls and phenolic-OH) are often considered to be analogous to ligand functional groups, the steric effect is likely to be an important consideration when determining mechanisms of trace element adsorption reactions in soil. Thus, it is reasonable to conclude that if the coordination environment is appropriate, multidentatebonding will be favored (thermodynamically) over monodentate bonding. However, the formation of multiple bonds may have intermediate products that have a higher activation energy than a complex with only a single bond. As discussed below, an increase in the activation energy may limit the kinetics of complex formation. The formation of a surface complex, or conversion of an adsorbate from one bond type to another, may be thermodynamically favored but inhibited by an activation energy, which is the extra energy, beyond the difference in the free energy between the products and reactants (~GO), required to complete the reactions (Figure 1.10). The activation energy results from the energy required to form intermediate products not accounted for in the reaction stoichiometry (Noggle, 1989, p. 532). A large activation energy will result in slower adsorption and desorption kinetics compared to sorption processes which have a lower activation energy. Since the strength of adsorption varies depending on the surface and adsorptive being considered, the adsorbate availability (via desorption) and kinetics are variable (Pignatello and Xing, 1996). For many adsorbed ions it is found that the rate of adsorption is faster than desorption (McBride, 1994, p. 134; Swift and McLaren, in Bolt, 1991, p. 285). A possible reason for the slower rate of desorption is an increase in the activation energy required to break the adsorption bonds. The activation energy for desorption can be quantified as follows: ~G:j:desorption = ~G:j: adsorption + ~Go adsorption' where ~G:j: desorption = activation energy for desorption, ~G:j:adsorption = activation energy for adsorption (~O), and
16
Fate and Transport of Heavy Metals in the Vadose Zone
Activated Complex*
~G
Aque?us---------l-~~;-------
---
Species
--------------- ---------------- ------------------------------
Sorbed Complex
........E---------Desorption Sorption - - - - - - - -...
Reaction Coordinates Figure 1.10. Schematic diagram of G vs. reaction coordinate for sorption and desorption processes. Adapted from Sparks and Jardine (1981), with permission.
~Goadsorption
energy of adsorption, see Figure 1.10 (McBride, in Bolt, 1991, p. 168). This equation indicates that desorption of chemisorbed ions yields a larger activation energy than adsorption reactions, causing desorption to be a slower process. This may be the cause of the pseudo-hysteresis that is commonly observed in sorption and desorption experiments; i.e., the forward and reverse isotherms do not overlie when given the same reaction time. The experiments of McLaren et al. (1986) were discussed briefly in an earlier section; however, another look at their results is merited at this point to evaluate possible mechanisms. They found that Co sorbed by a soil oxide demonstrated a continuous decrease in isotopic exchangeability as sorption times increased (only 20% was exchangeable when sorption was carried out for 50 days) (Figure 1.11). For humic acid, the isotopic exchangeability of sorbed Co decreased only slightly with increased sorption incubation time (Figure 1.12) (the amount of Co that was isotopically exchangeable remained as high as 80% for 50 days of sorption incubation time). It is difficult to prescribe a particular mechanism as the cause for the aging observed in McLaren's studies; however, it is possible that a more stable complex is being formed on the oxide with increasing sorption incubation time, increasing the energy required for isotopic exchange. Eliminating diffusion as a slow exchange mechanism seems reasonable in this case since the humic acid fraction, a porous material, lacked a slow exchange portion. However, more detailed studies and measurements of the porosity of the two materials is needed for diffusion to be completely ruled out. Surface precipitation is difficult to eliminate; the authors =
Sorption Kinetics of Trace Elements in Soils and Soil Materials -
- ---- ---- -
---
17
-------
900 800
•
•
•
•
•
700 600
";"0)
~
:; 500 (I)
"§
400
0 (,)
300
III
-g
200 100
a a
10
20
30
40
50
Time (days) Figure 1.11. Isotopic exchangeability of Co sorbed by soil oxide: total Co sorbed (+), and isotopic exchangeable (.). The space in between the two lines indicates the nonisotopic exchangeable fraction. From McLaren et al. (1986), with permission.
20
•
•
18 -16 ~
";"0) 0)
.a.
14 12
"C (I)
~ 0
If)
a1 0
u
10 8 6 4 2 0 0
10
20
30
40
50
60
Time (days)
Figure 1.12. Isotopic exchangeability of Co sorbed by soil organic matter: total Co sorbed (.), and isotopic exchangeable (+). The space in between the two lines indicates the nonisotopic exchangeable fraction. From McLaren et al. (1986), with permission.
discounted it as the predominant sorption mechanism since surface coverages were low. For more conclusive evidence, microscopic measurements are necessary.
18
Fate and Transport of Heavy Metals in the Vadose Zone
In this section we have discussed adsorption and desorption kinetics and sorbate stability. The kinetics of sorption and stability of a surface complex is a factor of both entropy (steric factors) and enthalpy (bond energetics). However, the formation of the most stable adsorbed species may be limited by intermediate complexes. Thus, if a sorbate slowly converts from one sorption or bonding species to a more stable complex that has a lower free energy, the result is important toward controlling the rate of uptake and affecting the availability of trace metals.
Kinetics and Mechanisms of Surface Precipitation In contrast to adsorption, surface precipitation is a 3-dimensional growth phenomenon that occurs on surfaces. Classical solution chemistry defines aqueous systems in three states: undersaturated, saturated, and supersaturated, with respect to the solubility of inorganic precipitates. A system saturated or supersaturated has a negative ~G, indicating that the precipitation of a solid product is favored. Precipitation that occurs in a saturated system proceeds more slowly than a supersaturated system (Stumm and Morgan, 1996, p. 802). Surface precipitation during trace metal sorption has been observed in systems undersaturated with respect to the pure hydroxide, and below monolayer surface coverage (Fendorf et aI., 1992; Fendorf and Sparks, 1994; O'Day et al., 1994a,b; Scheidegger et aI., 1996). This means that the availability and transport of a cation or anion may be controlled by precipitation and dissolution mechanisms, as opposed to adsorption phenomena. Veith and Sposito (1977) showed that traditional sorption data are described equally well by both surface precipitate models and adsorption isotherms. In addition, it has been noted that solubility lines of many soil solutions (logarithm of metal activity plotted as a function of pH) reveal undersaturation with respect to common precipitates; however, they often have slopes paralleling those of pure precipitates (McBride, in Bolt et aI., 1991, p. 171, from Lindsay, 1979). Such examples display the ambiguity of macroscopic models in describing microscopic processes; i.e., surface precipitation and adsorption models seem to describe sorption data equally well. Since precipitation and dissolution reactions exhibit slower kinetics than adsorption and desorption (Farley et aI., 1985) they may be the mechanism responsible for aging and the slow kinetics of sorption and desorption often observed in experimental systems (Fendorf et aI., 1992). In this section we categorize surface precipitation into three different types that are commonly discussed in the literature. These include formation or sorption of metal polymers (dimers, trimers, etc.) on the surface (Chisholm-Brause et aI., 1990); a solid solution or coprecipitate that involves coions dissolved from the sorbent; and a homogeneous precipitate formed on the surface composed of ions from the bulk solution, or their hydrolysis products (Farley et aI., 1985). The continuum between surface precipitation and chemisorption is controlled by several factors, including (1) the ratio of the number of sites vs. the number of metal ions in solution, (2) the strength of the metal-oxide bond, and (3) the degree to which the bulk solution is undersaturated with respect to the metal hydroxide precipitate (McBride, in Bolt, 1991, p. 163). The different types of surface precipitation are explained in more detail below. Polymeric metal complexes that form at the surface, and/or the sorption of aqueous polymers, may be a mechanism that typifies surface precipitate-like complexes (Fendorf
Sorption Kinetics of Trace Elements in Soils and Soil Materials
19
et al., 1992). Chisholm-Brause et al. (1990) interpreted the presence of Pb in the second coordination shell of a sorbed Pb (determined using XAS) as small clusters or polynuclear structures that are analogous with hydroxy metal complexes formed in water solution. The formation of complete surface precipitates was ruled out because the number of Pb atoms in the second shell was small (0.3 to l.5). Bargar et al. (1997) observed similar polymer formation at high loading levels for Pb on AI-oxide surfaces. Fendorf and Sparks (1994) found that Cr polymerization, and eventually Cr-hydroxide surface clusters, began at surface coverages as low as 20%. It was proposed that when the structures of the sorbate and sorbent are dissimilar, epitaxial growth is energetically unfavorable and thus nucleation growth is away from the surface, i.e., surface clusters. The formation of a homogeneous solid on a surface can occur when a solution becomes saturated and the surface acts as a nucleation site, or from a chemisorption-precipitation continuum, i.e., when adsorption reaches monolayer coverage sorption continues on the newly created sites resulting in a precipitate on the surface (multilayer surface coverage) (McBride, in Bolt et al., 1991, p. 171; Farley et al., 1985). This phenomenon is analogous to the assumptions used in the classical Brunauer-Emmett-Teller (BET) isotherm model of gas sorption onto surfaces (Borg and Dienes, 1992, p. 400). The distinction between a surface precipitate and a sorbed metal complex can be subtle, and somewhat confusing, especially since polymer sorption can lead to, or preface, surface precipitation. Adding to the difficulty of distinguishing surface precipitation from sorbed metal complexes is the fact that methods for distinguishing between the two phenomena are at present in their early development, and few studies exist on this subject matter. The solid solution concept of surface precipitation was presented in detail by Farley et al. (1985); it is described as a process similar to homogeneous coprecipitation. The composition of the surface precipitate varies, "continuously between that of the original solid and a pure precipitate of the sorbing metal" (Farley et al., 1985). The solid solution concept differs from the multilayer precipitation concept in that it includes both desorption and/or dissolution of structural ions from the sorbent and the inclusion of ions from solution. The result of coprecipitation of the solution ions with ions dissolved from the surface is a solid with isomorphic substitution, or a stable mixture of two solids (Stumm and Morgan, 1996, p. 814). An important factor controlling which ions will form a solid solution is the ionic radius. For example, Ainsworth et al. (1994) found that the extent of reversibility with aging for Co, Cd, and Pb was inversely proportional to the ionic radius of the ions, where ionic radii increase in the order Co < Cd < Pb. Since the ionic radius of Co is the most similar to Fe, they concluded that the hysteresis was a result of the formation of a solid solution. Solid solution formation is probably limited by the rate of mineral dissolution, rather than a lack of thermodynamic favorability (McBride, 1994, p. 163; Scheidegger et al., 1998). O'Day et al. (1996) observed a small amount of Si backscattering from the XAFS spectra of Co sorption on quartz (a-Si0 2). They explained this by proposing that Co was coordinated in Si tetrahedra, which occurred by either diffusion to defect sites, or a small amount of quartz dissolution and reprecipitation of a mixed Co/Si phase (solid solution). Surface precipitation and dissolution are slower processes than adsorption and desorption. Farley et al. (1985) noted that the rate of Cd uptake by amorphous iron hydroxide was lower when the initial solution concentration exceeded that required for monolayer coverage. One possible reason for the slower precipitation reactions is that a
20
Fate an.d Transport of Heavy Metals in the Vadose Zone
precipitated ion must form several bonds, which requires more activation energy than adsorption complexes which have fewer bonds. Likewise, surface precipitates may be more stable than adsorbed species because of the formation of high energy bonds and increased coordination. Another factor which makes surface precipitates more stable is that only the surface of the precipitate is accessible to the solution for dissolution to occur (for a 3-dimensional object only the exposed surfaces are surrounded by solution). The formation of a surface precipitate involves several reactions, including (1) adsorption of the ion on the surface, (2) surface nucleation, and (3) crystal growth (Stumm and Morgan, 1996, p. 812). Each of these steps contains several independent reaction sequences, and the rate of precipitate formation is determined by the slowest reaction step. While the formation of surface precipitates is important for predicting the fate of trace elements in the environment, dissolution reactions are also important processes that may be the controlling mechanisms for trace element mobilization when a soil has been contaminated for long periods. For the dissolution of surface precipitates the reaction sequence is similar to the steps of dissolution of a pure solid: (1) transport of reactants from the bulk solution to the surface, (2) adsorption of solutes, (3) interlattice transfer of reacting species, (4) chemical reactions, (5) detachment of reactants from the surface, and (6) mass transport into the bulk solution (Stumm and W ollast, 1990). These steps can be summarized as transport and surface reaction mechanisms. The mechanism controlling the rate of dissolution is dependent on several factors; i.e., solution composition, pH, mixing, etc. The kinetics of surface precipitate formation and dissolution has not been extensively studied. In a recent study by Scheidegger and Sparks (1996) the rate of release of Ni from a pyrophyllite surface known to have Ni precipitates showed both a fast and slow reaction. The fast reaction was attributed to desorption of specifically sorbed Ni. The slow reaction was attributed to the slow dissolution of polynuclear Ni complexes, which were found to dissolve more slowly than pure Ni(OHh In another study, Scheidegger et al. (1998) monitored the kinetics of surface precipitate formation on pyrophyllite, montmorillonite, and gibbsite using XAFS. Surface precipitate formation was initially fast on pyrophyllite and gibbsite (within minutes), but did not occur until 48 hours on montmorillonite. Figure 1.13 shows that Ni uptake by pyrophyllite is initially rapid, with approximately 25% of the Ni being sorbed within the first 30 minutes. Then the reaction slowed considerably, but was continuous for times as long as 72 hours (97% of the Ni is removed from solution). Analysis of the radial structure function (Figure 1.14) derived from XAFS spectroscopic characterization of the samples after different sorption periods shows an increase in a second shell at -2.75 A (uncorrected for phase shifts). This suggests that the slow development of polynuclear Ni complexes is responsible for the slow sorption reaction. These complexes have been identified as mixed Ni-Al (takovitelike) hydroxide phases (Scheidegger et al., 1998). O'Day et al. (1996) used XAS and kinetic experiments to hypothesize the mechanisms of surface precipitation on two different minerals. Their hypotheses were strengthened by comparing and contrasting the spectroscopic and kinetics results for different mineral surfaces. XAFS results for Co on rutile (Ti02) showed an increase in the number of backscattering Co atoms for aging times of one day to 11 days, suggesting an increase in the size of multinuclear complexes formed on the surface. However, similar results were not seen for Co aging on quartz (a-Si0 2), which had Co(OH)2 surface
Sorption Kinetics of Trace Elements in Soils and Soil Materials
100 90
;e 80 e.. c:
o
5 en
'0 E
70 60
-e
50
' tJ
40
E CD
30
Z
20
~
•
•
•
•
21
•
•
• • • •
a:
10 O+-----.----.-----.----.-----.----.-----.-----r--~
o
24
48
72
96
120
144
168
192
216
Time (hours) Figure 1.13. Kinetics of Ni removal from solution by pyrophyllite. From Scheidegger et al. (1997b), with permission.
2.0
1.5 1.0 Q)
0.5
"'0
Months
:::J 0.0
~
c:
0)
rn ::E E
...0
-...
24 Hours
en rn c:
I-
o Figure 1.14. Radial distribution function (uncorrected for phase shifts) for Ni sorption on pyrophyllite incubated for 24 hours to 3 months. From Scheidegger et al. (1997b), with permission.
22
Fate and Transport of Heavy Metals in the Vadose Zone
precipitates present. They hypothesized that the reason for the observed slow change in the multinuclear surface precipitate on the rutile and not on the quartz was a result of the similar radii between Co and Ti, while the radius of Si is larger than Co. As a result of this difference in atomic radii, the Co hydroxide-like precipitate formed on quartz was attached only to corners of select Si tetrahedra on the surface. However, Co sorption on the rutile was consistent with the formation of a precipitate that had similar lattice dimensions as the surface, effectively extending the lattice structure of the bulk solid; i.e., an epitaxial growth. The resulting Co surface precipitate was structurally strained due to charge imbalances and distortion of the Co0 6 octahedra. Thus, they proposed that the change with increasing equilibration times was due to the formation of an anatase-like structure (conclusion made based on similar octahedra bond distances between the anatase and the surface precipitate). The anatase structure, a Ti0 2 polymorph, has favorable lattice dimensions for Co because of a more open structure which better accommodates the slightly larger Co ion. There are several thermodynamic reasons for the formation of surface precipitates in unsaturated systems. For example, the solid surface may lower the energy of nucleation by providing sterically similar sites (McBride, in Bolt, 1991, p. 171), the activity of the surface precipitate is less than unity (Sposito, 1986), and the solubility of the surface precipitate is lowered because the dielectric constant of the solution near the surface is less than that of the bulk solution (O'Day et al., 1994a). It has not been established which one of these mechanisms predominates; however, it is possible that the three phenomena simultaneously influence precipitation reactions on surfaces. To ensure that precipitation is truly a surface-induced phenomenon, experimental systems should be run at conditions undersaturated with respect to known phases. However, the solubility products of many possible phases are unknown, making it difficult to determine if such phases will precipitate in a given system. This is particularly true for mixed cation hydroxides and coprecipitation reactions on surfaces (d'Espinose de la Caillerie et al., 1995; Towle et al., 1997; Scheidegger et al., 1998) The theories on the enhancement of surface precipitation by the three mechanisms mentioned above are discussed in more detail below. As discussed above, Farley et al. (1985) presented a solid solution model to explain the continuum between precipitation and chemisorption onto solid surfaces. The model suggests that sorption is a process that includes solid dissolution, and then reprecipitation onto the surface. Thus, the formation of a surface complex involves the coprecipitation of both the ions dissolved from the sorbent, and the ions present in the bulk solution. Therefore, assuming the solid phase is a pure crystal and has unit activity (relative to the pure macro crystal) is an inappropriate assumption, and invalidates solubility determinations based on the law of mass action, and ion activity products that do not account for surface activity. In addition, the resulting surface complex may not be compositionally homogeneous and completely free from inclusions, causing the activity of the solid phase to be even lower (Sposito, 1986). Sposito (1986) illustrated this idea by considering the dissolution of CdC03 : (1) Ion Activity Product (lAP)
=
[Cd][C0 3]
= K.o
[CdC0 3 (s)]
(2)
Sorption Kinetics of Trace Elements in Soils and Soil Materials
23
where the brackets indicate ion activity. The activity of a pure solid is often considered to be one; however, if Cd forms a mixed precipitate (coprecipitate) with another ion in solution, such as Ca, the result is not [CdC03 (s)] = 1, but Cdx C
(3) where gj is the activity coefficient of solid i, ~ is the mole fraction of solid i and K; so is the solubility product of the pure mineral i (Van Riemsdijk and Van der Zee, in Bolt, 1991, p. 251). From this equation one sees that a coprecipitate with one constituent present in minor amounts has a decreased solubility product with respect to the pure mineral. The enhancement of precipitate formation on the surface may also be due to the reduction of the energy barrier necessary for nucleation processes to occur in an aqueous solution. This is a factor of the lattice dimensions of the solid, and those of the precipitate to be formed, i.e., a steric interaction (McBride, in Bolt et al., 1991, p. 171). The result of surface nucleation sites is that the extent of supersaturation required for precipitation is decreased. However, there may be other important factors contributing to surface nucleation interactions. For example, Fendorf et al. (1992) observed AI surface precipitates on Mn0 2 using high resolution transmission electron micrography (HRTEM), but not on Ti0 2 under the same conditions (undersaturated with respect to the most likely AI hydroxide precipitates, and equivalent surface coverages). If promotion of surface precipitation below saturation was a result of the presence of nucleation sites, then one would expect to see precipitates on both surfaces. Thus, another factor inhibited surface precipitate formation on the Ti0 2, i.e., steric hindrances between the two surfaces. The dielectric constant of the solution at the interface of a solid is much less than it would be in the bulk solution (Hiemenz, 1986, p. 725). This is a result of the surface charge aligning the dipoles of the water layers nearest the surface. This phenomenon is called dielectric saturation, and results in a dielectric constant an order of magnitude or less in the first few angstroms of the surface (McBride, 1994, p. 296). The activity of individual ions in solution is inversely proportional to the dielectric constant of water. Consequently, near the surface the lowered dielectric constant of the water causes an increased ion activity, and the lAP near the surface will exceed that of the bulk solution. O'Day et al. (1994a) used a more direct approach to this concept that calculates the change in the free energy near the surface as a function of the dielectric constant using the Born charging equation for a spherical ion. However, calculations for sorption of Co on kaolinite revealed that the average surface dielectric constant is not low enough to account for surface precipitation of a pure hydroxide phase as the predominant mechanism of sorption. Thus, if surface precipitation was the mechanism of sorption, then either the value of the dielectric constant near the surface was incorrect (possible, since dielectric constants near surfaces are difficult to determine), or precipitation was enhanced by one of the aforementioned mechanisms (O'Day et al., 1994a). Regardless of the mechanism, surface precipitation is an important process affecting trace metal reactions with natural materials. Since surface precipitation kinetics can be
24
Fate and Transport of Heavy Metals in the Vadose Zone
slow, the extent of precipitation and subsequent dissolution of surface precipitates are affected by residence time; thus, they are important slow kinetic mechanisms which can control the fate of trace elements in the environment.
SUMMARY In this chapter we have presented evidence that slow kinetics are important when estimating the extent and reversibility of trace metal sorption on soils and soil materials. We have also discussed three possible mechanisms for such slow kinetics: diffusion, rate-limited adsorption processes, and precipitation reactions on surfaces. While evidence exists that suggests all three mechanisms occur, the slow sorption mechanism occurring in a particular system is highly dependent on the prevailing environmental conditions; e.g., solution pH, sorbent characteristics, ionic strength, trace metal physicochemical characteristics, dissolution rate of the solid, and the microporosity of the solid. In addition to a review of relevant studies in the literature, two examples from our own research were given that suggested different mechanisms as the rate controlling processes responsible for the slow sorption of metals on model soil components. In one case Pb sorption on y-AI203 resulted in little change in the local chemical environment (determined using XAFS) with increased incubation time. In another example, Ni sorption on the clay mineral pyrophyllite resulted in increased polynuclear surface complexation with increasing reaction time. While it is difficult to make direct comparisons of the two metals since the surfaces present are different, we think it is justifiable since additional experiments (data not shown) for Pb on pyrophyllite (AIcacio, 1997) and Ni on gibbsite (a form of AI hydroxide) (Scheidegger et al., 1997a) showed similar sorption behavior to the systems being compared above. We propose that one reason for the different apparent bonding mechanisms is the difference in the ionic radius of the two metals (1.20 Afor Pb and 0.69 Afor Ni). The sorbents studied in these two cases have AI present as a structural component. Since AI can dissolve and has a similar radius as Ni they can form a coprecipitate, while the Pb ion is too large to form such a coprecipitate. Ainsworth et al. (1994) observed that the extent of sorption reversibility was positively correlated to the ionic radius of the sorbing metals. Coughlin and Stone (1995) also suggested that coprecipitation of metal ions with Fe is directly dependent on ionic radius. While these studies do not provide direct evidence on the formation of a coprecipitate, they agree well with our data; Ni (the smaller ion with an ionic radius similar to AI) forms a mixed precipitate, while Pb does not. This information can be used to improve predictions on the fate of these metals in the environment, and will allow for better simulations in the laboratory. When predicting the transport of trace elements in the vadose zone, researchers must know the kinetics of sorption and desorption behavior. If slow kinetics are controlling these mechanisms then reaction-transport models should include such chemical processes. This will result in more accurate predictions and improved management of existing and potential risks. In addition, if the mechanism responsible for slow sorption or desorption is known, researchers can design remediation strategies more efficiently. This may include mobilizing or immobilizing contaminants based on the pH of the soil solution, treating the soil with chelating ligands, or creating treatments for specific exposure times based on the kinetics of the reactions.
Sorption Kinetics of Trace Elements in Soils and Soil Materials
25
To obtain complete sorption and desorption kinetic behavior, researchers should conduct experiments in the lab for extended time periods. When possible, mechanisms of slow kinetics should be determined to better predict the fate of trace elements in the environment. To ascertain the mechanisms, both macroscopic and microscopic experiments should be used. In this chapter we have presented several methods that have been used for the determination of mechanisms. With the rapid advancement of technology the future should bring an even better understanding of soil sorption mechanisms and kinetic processes. It is critical that researchers combine their efforts with those in related fields so that the most contemporary and valid models can be developed and employed to predict the fate of trace elements in the environment.
REFERENCES Ainsworth, C.C., J.L. Pilon, P.L. Gassman, and W.G. Van der Sluys. Cobalt, cadmium, and lead sorption to hydrous iron oxide: Residence time effect. Soil Sci. Soc. Am. J. 58, pp. 1615-1623, 1994. Alcacio, T.E. An XAFS Survey of Pb Complexes at the Gibbsite Solid/Liquid Interface. Masters Thesis, University of Delaware, 1997. Alexander, M. How toxic are chemicals in soils? Environ. Sci. Tec/moL. 29(11), pp. 2713-2717, 1995. Axe, L. and P.R Anderson. Experimental and theoretical diffusivities of Cd and Sr in hydrous ferric oxide. J. Collow Interface Sci. 185(2), pp. 436-448, 1997. Backes, C.A., RG. McLaren, A.W. Rate, and RS. Swift. Kinetics of cadmium and cobalt desorption from iron and manganese oxides. Soil Sci. Soc. Am. J. 59, pp. 778-785, 1995. Bargar, J.R, G.E. Brown, Jr., and G.A. Parks. Surface complexation of Pb(II) at oxide-water interfaces: I. XAFS and bond-valence determination of mono- and polynuclear Pb(II) sorption products on Al-oxides. Geochimica et CO.JmochimicaActa, 61(13), pp. 2617-2637, 1997. Bargar, J.R, S.N. Towle, G.E. Brown, Jr., and G.A. Parks. Outer-sphere Pb(II) adsorbed at specific surface sites on single crystal a-alumina. Geochimica et CO.Jmochimica Acta 60(18), pp. 3541-3547, 1996. Benjamin, M.M. and J.O. Leckie. Multiple-site adsorption of Cd, Cu, Zn, and Pb on amorphous iron oxyhydroxide. J. CoLww Interface Sci. 79(1), pp. 209-221, 1981. Bibak, A., J. Gerth, and O.K Borggaard. Retention of cobalt by pure and foreign-element associated goethites. Clay.J Clay Minerau 43(2), pp. 141-149, 1995. Blackburn, W.H. and W.H. Dennen.PrincipluofMinerawgy. Wm. C. Brown Publishers, Dubuque, lA, 1994. Borg, RJ. and G.J. Dienes. The Pby.JicaL Chemutry of SoLw.J. Academic Press, Inc., London, 1992. Bruemmer, G.W., J. Gerth, and KG. Tiller. Reaction kinetics of the adsorption and desorption of nickel, zinc and cadmium by goethite. I. Adsorption and diffusion of metals. J. SoiL Sci. 39, pp. 37-52, 1988. Burgos, W.D., J.T. Novak, and D.P. Berry. Reversible sorption and irreversible binding of naphthalene and a-Naphthol to soil: Elucidation of processes. Environ. Sci. TechnoL. 30(4), pp. 1205-1211, 1996. Chisholm-Brause, C.J., KF. Hayes, A.L. Roe, G.E. Brown, Jr., G.A. Parks, and J.O. Leckie. Spectroscopic investigation of Pb(II) complexes at the y-Al 20iwater interface. Geochimica et CO.JmochimicaActa 54, pp. 1897-1909, 1990. Comans, RN.J. Adsorption, desorption and isotopic exchange of cadmium on illite: Evidence for complete reversibility. Water RM. 21 (12), pp. 1573-1576, 1987.
26
Fate and Transport of Heavy Metals in the Vadose Zone
Coughlin, B.R and A.T. Stone. Nonreversible adsorption of divalent metal ions (Mn", Co", Ni", Cu", and Pb") onto goethite: Effects of acidification, Fe" addition, and picolinic acid addition. Environ. Sci. Technol. 29(9), pp. 2445-2455, 1995. d'Espinose de la Caillerie, J.B., M. Kermarec. and O. Clause. Impregnation of y-alumina with Ni(II) and Co(II) ions at neutral pH: Hydrotalcite-type coprecipitate formation and characterization. J. Am. Chem. Soc. 117, pp. 11471-11481. 1995. Dzombak, D.A. and F.M.M. Morel. Sur/ace Complexation MOdeling, HYdrOLM Ferric Oxide. John Wiley & Sons, New York, 1990. Farley KJ., D.A. Dzombak, and F.M.M. Morel. A surface precipitation model for the sorption of cations on metal oxides. J. Colloid Inter/ace Sci. 106(1), pp. 226-242, 1985. Fendorf, S.E., M. Fendorf. D.L. Sparks, and R Gronsky. Inhibitory mechanisms of Cr(I1l) oxidation by 8-Mn0 2 • J. Colloid Inter/ace Sci. 153(1), pp. 37-54, 1992. Fendorf. S.E., G.M. LambIe, M.G. Stapleton, M.J. Kelley, and D.L. Sparks. Mechanisms of chromium(III) sorption on silica. 1. Cr(III) surface structure derived by extended X-ray absorption fine structure spectroscopy. Environ. Sci. Technol. 28(2), pp. 284-289. 1994. Fendorf, S.E. and D.L. Sparks. Mechanisms of Chromium(III) sorption on silica. 2. Effect of reaction conditions. Environ. Sci. Technol. 28(2), pp. 290-297, 1994. Fendorf, S.E., D.L. Sparks, and M. Fendorf. Surface precipitation reactions on oxide surfaces. J. Colloid Inter/ace Sci. 148(1), pp. 295-298, 1992. Fuller, C.C., J.A. Davis, and G.A. Waychunas. Surface chemistry offerrihydrite: Part 2. Kinetics of arsenate adsorption and coprecipitation. Geochimica et COdmochimica Acta 57, pp. 2271-2282, 1993. Griffin. RA., W.A. Sack, W.R Roy, C.C. Ainsworth, and LG. Krapac. Batch Type 24-h Distribution Ratio for Contaminant Adsorption by Soil Materials. In HazardOLM anO IndLMtrial Solid Wadte Tedting and DiJpodal, Vol. 6, D. Lorenzen, RA. Conway, L.P. Jackson, A. Hamza, C.L. Perket. and W.J. Lacy, Eds., American Society for Testing and Materials. Philadelphia, 1986. pp. 390-408. GrossL P.R. D.L. Sparks, and C.C. Ainsworth. Rapid kinetics of Cu(II) adsorption/desorption on goethite. Environ. Sci. Techno!. 28(8), pp. 1422-1429. 1994. Hayes, KF. and J.O. Leckie. Mechanism of Lead Ion Adsorption at the Goethite-Water Interface. In Geochemical PrOCedded at Mineral Sur/aced, J.A. Davis and K.F. Hayes. Eds .• American Chemical Society, Washington, DC, 1986, pp. 115-141. Hayes. KF. Equilibrium, Spectroscopic, and Kinetic Studies of Ion Adsorption at the Oxide/ Aqueous Interface. Ph.D. Dissertation, Stanford University, 1987. Hiemenz, P.C. Principled of Colloid and Sur/ace ChemiJtry. Marcel Dekker, Inc., New York. 1986. Kim, Y., RJ. Kirkpatrick, and RT. Cygan. J33Cs NMR study of cesium on the surfaces of kaolinite and illite. Geochimica et C(},jmochimicaActa 60(21), pp. 4059-4074. 1996. Koskinen. W.C. and S.S. Harper. The Retention Process: Mechanisms. In Pedticided in the Soil Environment: PrOCedded, Impactd, and MOdeling, H.H. Cheng, Ed., Soil Science Society of America, Madison. WI, 1990, pp. 51-77. Kuo, S. and D.S. Mikkelsen. Kinetics of zinc desorption from soils. Plant Soil 56, pp. 355-364, 1980. Lehmann, RG. and RD. Harter. Assessment of copper-soil bond strength by desorption kinetics. Soil Sci. Soc. Am. J. 48. pp. 769-772, 1984. Loehr, R.C. and M.T. Webster. Behavior of fresh vs. aged chemicals in soils. J. Soil Contam. 5(4), pp. 361-393, 1996. Lookman, R, D. Freese. R. Merckx. K. Vlassak, and W.H. Van Riemsdijk. Long-term kinetics of phosphate release from soil. Environ. Sci. Technol. 29(6). pp. 1569-1575, 1995. McBride. M.B. Cu 2+ characteristics of aluminum hydroxide and oxyhydroxides. Clayd and Clay Minerau 30(1), pp. 21-28. 1982.
Sorption Kinetics of Trace Elements in Soils and Soil Materials
27
McBride, M.M. Processes of Heavy and Transition Metal Sorption by Soil Minerals. In InteractWIM at the Soil Col!.JlJ-Soil Solution Interface, Vol. 190, G.H. Bolt, M.F.D. Boodt, M.H.B. Hayes, and M.B. McBride, Eels., Kluwer Academic Publishers, 1991, pp. 149-176. McBride, M.M. Environmental Chemiltry of SoiU. Oxford University Press, New York, 1994. McKenzie, RM. The sorption of cobalt by manganese minerals in soils. AliA. J. Soil Ru. 5, pp. 235-246, 1967. McKenzie, RM. The adsorption oflead and other heavy metals on oxides of manganese and iron. AlMt. J. Soil Ru. 18, pp. 61-71, 1980. McLaren, RG., D.M. Lawson, and RS. Swift. Sorption and desorption of cobalt by soils and soil components. J. Soil Sci. 37, pp. 413-426, 1986. McLaren, RG., J.G. Williams, and R.S. Swift. Some observations on the desorption and distribution behavior of copper with soil components. J. Soil Sci. 34, pp. 325-331, 1983. Noggle, J.H. PhYdical Chemiltry. Harper Collins, New York, 1989. Nyffeler, U.P., H.Y. Li, and P.H. Santschi. A kinetic approach to describe trace-element distribution between particles and solution in natural aquatic systems. Geochimica et COdmochimica Acta 48, pp. 1513-1522, 1984. O'Day, P., G.E. Brown, Jr., and G.A. Parks. X-Ray absorption spectroscopy of cobalt (II) multinuclear surface complexes and surface precipitates on kaolinite. J. CofWlJ Interface Sci. 165, pp. 269-289, 1994a. O'Day, P., G.A. Parks, and G.E. Brown, Jr. Molecular structure and binding sites of cobalt (II) surface complexes on kaolinite from X-ray absorption spectroscopy. CIaYd and Clay Minerau 42(3), pp. 337-355, 1994b. O'Day, P.A., C.J. Chisholm-Brause, S.N. Towle, G.A. Parks, and G.E. Brown. X-ray absorption spectroscopy of Co(II) sorption complexes on quartz (u-Si0 2) and rutile (Ti0 2). Geochimica et COdmochimica Acta 60(14), p. 2515, 1996. Padmanabham, M. Adsorption-desorption behavior of copper (II) at the goethite-solution interface. AlMt. J. Soil Ru. 21, pp. 209-320, 1983. Papelis, C. X-Ray Photoelectron spectroscopic studies of cadmium and selenite adsorption on aluminum oxides. Environ. Sci. Technol. 29, pp. 1526-1533, 1995. Pignatello, J.J. and B. Xing. Mechanisms of slow sorption of organic chemicals to natural particles. Environ. Sci. Technol. 30(1), pp. 1-11, 1996. Scheidegger, A.M., G.M. Lambie, and D.L. Sparks. Investigation ofNi sorption on pyrophyllite: An XAFS study. Environ. Sci. TechnoL 30(2), pp. 548-554, 1996. Scheidegger, A.M. and D.L. Sparks. Kinetics of the formation and the dissolution of nickel surface precipitates on pyrophyllite. Chem. GeoL 132(1-4), p. 157, 1996. Scheidegger, A.M., G.M. Lambie, and D.L. Sparks. Spectroscopic evidence for the formation of mixed-cation hydroxide phases upon metal sorption on clays and aluminum oxides. J. Col!.JlJ Interface Sci. 186, pp. 118-128, 1997a. Scheidegger, A.M., G.M. Lambie, and D.L. Sparks. The kinetics of nickel sorption on pyrophyllite as monitored by X-ray absorption fine structure (XAFS) spectroscopy. Journal de PhYdique IV, 7(C2-773), 1997b. Scheidegger, A.M., D.G. Strawn, G.M. Lambie, and D.L. Sparks. The kinetics of mixed Ni-Al hydroxide formation on clays and aluminum oxides: A time-resolved XAFS study. Geochimica COdmochimica Acta, in press, 1998. Schultz, M.F., M.M. Benjamin, and J.F. Ferguson. Adsorption and desorption of metals on ferrihydrite: Reversibility of the reaction and sorption properties of the regenerated solid. Environ. Sci. TechnoL 21(9), pp. 863-869, 1987. Smith, J. T. and RN.J. Comans. Modeling the diffusive transport and remobilization of CS 137 in sediments: The effects of sorption kinetics and reversibility. Geochimica et COdmochimica Acta 60(6), pp. 995-1004, 1996.
28
Fate and Transport of Heavy Metals in the Vadose Zone
Sparks, D.L. and P.M. Jardine. Thermodynamics of potassium exchange in soil using a kinetic approach. Soil Sci. Soc. Am. 1. 45, pp. 1094-lO99, 1981. Sparks, D.L. Kinetic.J of SoiL Chemical Proce.JJe.J. Academic Press, San Diego, 1989. Sparks, D.L. EnvironmentaL SoiL ChemiJtry. Academic Press, San Diego, 1995. Sparks, D.L. Kinetics of Soil Chemical Phenomena: Future Directions. In Future ProJpectJ for SoiL ChemiJtry. Soil Science Society of America Special Publication, P.M. Huang et al., Eds., Soil Science Society of America, Madison, WI, in press, 1998. Sposito, G. Distinguishing Adsorption from Surface Precipitation. In GeochemicaL Proce.JJe.J at MineraL Surfaced, J.A. Davis and K.F. Hayes, Eds., American Chemical Society, Washington, DC, 1986, pp. 217-228. Sposito, G. The ChemiJtry of Soil!. Oxford University Press, New York, 1989. Steinfeld, J.I., J.S. Francisco, and W.L. Hase. ChemicaL Kinetic.J and DymanicJ. Prentice Hall, Englewood Cliffs, NJ, 1989. Stumm, W. and J.J. Morgan. Aquatic ChemiJtry, ChemicaL EquiLihria and Rate.J in NaturaL WaterJ. John Wiley & Sons, New York, 1996. Stumm, W. and R. W ollast. Coordination chemistry of weathering: Kinetics of the surfacecontrolled dissolution of oxide minerals. Rev. GeophYJ. 28(1), pp. 53-69, 1990. Swift, R.S. and R.G. McLaren. Micronutrient Adsorption by Soils and Soil Colloids. In InteractionJ atthe SoiL CoL"'w-SoiL SoLution Inter./ace, Vol. 190, G.H. Bolt, M.F.D. Boodt, M.H.B. Hayes, and M.B. McBride, Eds., Kluwer Academic Publishers, Dordrecht, The Netherlands, 1991, pp. 257-292. Towle, S.N., J.R. Bargar, G.E. Brown, Jr., and G.A. Parks. Surface precipitation of cobalt on Al 2 0 3 .1. CoL"'w Interface Sci. 187, pp. 62-82, 1997. Van Riemsdijk, W.H. and S.E.A.T.M. Van der Zee. Comparison of Models for Adsorption, Solid Solution and Surface Precipitation. In InteractionJ at the SoiL CoL"'w-SoiL SoLution Interface, Vol. 190, C.H. Bolt, M.F.D. Boodt, M.H.B. Hayes, and M.B. McBride, Eds., Kluwer Academic Publishers, The Netherlands, 1991, pp. 241-256. Veith, J.A. and G. Sposito. On the use of the Langmuir equation in the interpretation of "adsorption" phenomena. Soil Sci. Soc. Am. 1. 41, pp. 697-702, 1977. Verburg, K. and P. Baveye. Hysteresis in the binary exchange of cations on 2: 1 clay minerals: A critical review. ClayJ and CLay Mineral! 42(2), pp. 207-220, 1994. Waychunas, G.A., B.A. Rea, C.C. Fuller, and J.A. Davis. Surface chemistry of ferrihydrite: Part 1. EXAFS studies of the geometry of coprecipitated and adsorbed arsenate. Geochimica et COJmochimica Acta 57, pp. 2251-2269, 1993. Young, T.C., J.V. DePinto, and T.W. Kipp. Adsorption and desorption of Zn, Cu, and Cr by sediments from the Raisin River Valley. J. Great LaIee.J &d. 13(3), pp. 353-366, 1987.
CHAPTER 2
Adsorption Isotherms of Nickel Acid Forest Soils
•
In
Franz Zehetner and Walter W. Wenzel
INTRODUCTION Recent studies reveal that the deposition of Pb, As, Co, Cr, and Cd in Europe has decreased since the early eighties while the level of Ni deposition has not changed significantly (Schulte and Gehrmann, 1996; Schulte et al., 1996). Ni concentrations in soil solutions may be controlled by either adsorption/desorption or by precipitation/dissolution processes. According to Ma and Lindsay (1995), adsorption/desorption may be important in uncontaminated soils oflow pH whereas precipitation/dissolution may operate in soils of high pH and/or high Ni levels. Ni adsorption is influenced by a number of soil factors, including CEC, pH, texture, CaC03 content, organic matter, sesquioxides, and chelating agents (Adriano, 1986). The most important adsorbents in soils are organic matter, layer silicates, hydrous oxides of Si, Al, Fe, and Mn, and carbonates. The range of different (external and internal) surfaces available for Ni adsorption in soils results in a continuum of adsorption sites with various affinities for the species of Ni. The most strongly adsorbed species are the cationic species Ni 2 + and NiOH+ (Uren, 1992). In this contribution adsorption isotherms of Ni are presented for individual acid forest soils. Individual mass-based isotherms are only of limited value for modeling Ni mobility in soils since determination for each individual soil is required. Therefore, the potential of a general surface-based adsorption isotherm was tested using specific surface area (SSA) instead of soil mass as the reference quantity. Since typically, cation exchange capacity (CEC) rather than SSA is available from soil databases, a chargebased approach was developed by further substituting SSA for CEC.
ADSORPTION
Definition Adsorption, as defined by Everett (1972), is the process producing net accumulation of a substance at the common boundary of two contiguous phases. In soil/soil solution .,0
30
Fate and Transport of Heavy Metals in the Vadose Zone
systems the term adsorption refers to the accumulation of molecules and ions at the interface between the soil solid phases and the soil solution forming two-dimensional arrangements. Adsorption does not include three-dimensional processes such as (surface) precipitation and diffusion into crystal structures (Sposito, 1989; Scheidegger and Sparks, 1996a). The adsorbed substance is termed an addorbate, the solid surface on which it accumulates is the addorbent. A molecule or an ion in the soil solution that potentially can be adsorbed is termed an ad(JOrptive.
The Diffuse Double-layer Adsorption phenomena can be described by means of molecular adsorption models, which are based on hypotheses about the interactions between an adsorptive and an adsorbent resulting in a particular arrangement of an adsorbate on a surface. The diffLMe double-layer model is the oldest molecular adsorption model. Gouy (1910) and Chapman (1913) derived an equation describing the ionic distribution in the diffuse layer formed on a uniformly charged plane surface. The classical approach to understanding the distribution of cations in the solution close to a negatively charged surface is the application of diffuse double-layer theory, which provides a mathematical description of the decrease in electrical potential with increasing distance from a charged plane surface. The typical distribution curve for cations away from a negatively charged surface (Figure 2.1) can be considered as a balance between electrical attraction and a diffusion away from the surface due to the concentration gradient that is established as a result of the attraction. In the diffuse double-layer theory, charge is the exclusively important property whereas ion size, as well as the effects of hydration are not accounted for. Thus, the main use of the diffuse double-layer approach is the description of effects at some distance from the surface in a region beyond that in which surface complexes are formed. Stern (1924) modified the model proposed by Gouy and Chapman accounting for the special nature of the first layers (within about 0.5 nm from the surface) of counterions against the charged surface compared to the truly diffuse ions beyond. This generally improves the application of diffuse double-layer theory but still gives little explanation to the real situation close to the surface. In summary, a diffuse layer of ions may extend from a charged surface, but only after the ions in solution have formed any specific interaction with the surface.
Adsorption Mechanisms According to Sposito (1989), adsorption on soil particle surfaces can involve three mechanisms, i.e., in the order of decreasing interaction strength, inner-dphere dur/ace complexation, outer-dphere dur/ace complexation, and diffLMe ion addoclation (Figure 2.2). In innersphere surface complexes, the molecule or ion is directly bound to a surface functional group, involving ionic, as well as covalent bonding. Inner-sphere surface complexation can be considered as the molecular basis for the term dpecific addorption, since covalent bonding is significantly influenced by the particular electron configurations of the atoms involved. In the less stable outer-sphere surface complexes at least one water molecule is interposed between the surface functional group and the bound molecule or ion, e.g., by
Adsorption Isotherms of Nickel in Acid Forest Soils ~""-
-~"
31
-~"
I Stern layer Diffuse layer -.I •.
solution ••.External -------------
--------------~----------+
(/)
c:
-. -.Q o c: o
e!
c:
~ c: o
,....
()
/'
- --
Anions
Distance from surface
Figure 2.1. Distribution of cations and anions away from a negatively charged surface (after Russell, 1988).
the solvation shell of an ion. If a solvated ion does not form a complex with a charged surface functional group, but balances charge in a delocalized sense, it is adsorbed in the diffuse ion swarm. These ions are fully dissociated from surface functional groups and are free to move about. The outer-sphere surface complexation and the diffuse ion association involve electrostatic bonding and can be considered as the molecular basis for the term nOfMpecijic ad
32
Fate and Transport of Heavy Metals in the Vadose Zone
Stern layer
s i
a
Diffuse layer d
w »zn2+« _
Diffuse ion swarm
::~ Outer-sphere surface complexes
Inner-sphere surface complexes
o
Metal
o Oxygen cIo Water molecule Surface precipitate
Figure 2.2. Schematic representation of possible reactions at a mineral surface (after Sposito, 1984; Russell, 1988; Scheidegger and Sparks, 1996a). The s plane is that of surface hydroxyl groups, the i and 0 planes are associated with inner-sphere and outer-sphere surface complexes, respectively, and the d plane indicates the beginning of the diffuse layer.
Adsorption Isotherms of Nickel in Acid Forest Soils
Q
Q
33
L-curve isotherm
S-curve isotherm
c
c
C-curve isotherm
Q
H-curve isotherm
c
Q
c
Figure 2.3. The four main classes of adsorption isotherms (after Sposito, 1984, 1989).
ADSORPTION ISOTHERMS
If the adsorbed amount (Q) of a substance and the corresponding concentration in solution (C) are known at fIxed temperature (ilotherm) and applied pressure, the pairs of Q and C can be plotted against one another with Q as the dependent variable, obtaining an adsorption isotherm.
Classification Adsorption isotherms for solutes in dilute solution can be classifIed according to initial slope (Giles et aL, 1960, 1974a, 1974b). The four main classes, which are shown in Figure 2.3, are briefly described in Sposito (1984, 1989). The S-curve isotherm is characterized by an initially small slope that increases with adsorptive concentration. This indicates that at low concentration the affInity of the soil solid phases for the adsorbate is less than that of the soil solution for the adsorptive. Metal adsorption in soil may follow an S-shaped isotherm if dissolved organic compounds form strong, nonadsorbing complexes with the metaL When the metal concentration exceeds the complexing capacity of these ligands, adsorption on the solid soil particles increases and the isotherm takes on its characteristic S-shape. The S-curve isotherm occurs also as a result of interactions among adsorbed organic molecules. These interactions (e.g., surface polymerization) cause the adsorbate to become stabilized on the solid surface and lead to enhanced affInity of the surface for the adsorbate with increasing amounts adsorbed.
34
Fate and Transport of Heavy Metals in the Vadose Zone
The L-curve isotherm, which typically is concave to the concentration axis, shows an initial slope that does not increase with adsorptive concentration. This is the result of high affinity of the soil solid phases for the adsorbate at low surface coverage combined with a decreasing amount of adsorbing surface with progressing adsorption. The H-curve isotherm represents an extreme version of the L-curve isotherm. Its characteristic large initial slope indicates very high affinity of the soil solid phases for the adsorbate, which is caused either by very specific interactions between the solid phases and the adsorbate (inner-sphere surface complexation) or by significant van der Waals interactions in the adsorption process. The C-curve isotherm shows an initial slope that remains independent of adsorptive concentration until the maximum possible adsorption. This type of isotherm is produced either by a constant partitioning of a substance between the interfacial region and the soil solution or by a proportionate increase in the amount of adsorbing surface as the amount adsorbed increases. The adsorption of ions by soil particles usually follows an L-curve isotherm. The mathematical description of this isotherm almost invariably involves either the Langmuir equation or the van Bemmelen-Freundlich equation. The development of these two equations to describe solute adsorption from aqueous solutions was reviewed from a historical point of view by Forrester and Giles (1972).
The Langmuir Equation By means of kinetic theory, Langmuir (1918) developed an equation describing gaseous adsorption on plane solid surfaces (glass, mica, platinum). Langmuir assumed that the surface of a solid possesses a finite number of adsorption sites. If a gas molecule strikes an unoccupied site, it is adsorbed whereas if it strikes an occupied site, it is reflected back into the gas phase. This model leads to the concept of an upper limit of adsorption which occurs when the surface of the solid is covered with a closely packed mono-molecular layer of gas molecules. According to Langmuir, adsorption can be regarded as a chemical reaction, given by
A
+
S
~
AS
(1)
and described as a dynamic equilibrium. A denotes some gaseous substance that is adsorbed on a plane solid surface (S). At equilibrium, the rate of adsorption equals the rate of desorption, i.e.,
(2) where kads and kdes are the rate coefficients in the forward and reverse direction, respectively. The rate of adsorption is proportional to the pressure of the gas (p) -i.e., the amount of gas molecules per unit of volume, and to the fraction of the uncovered surface (1 - 8). The desorption rate is proportional to the fraction of the surface covered (8), that is obtained as
Adsorption Isotherms of Nickel in Acid Forest Soils
35
(3)
By introducing a constant quantity (adsorption coefficient K) for the ratio of the rate coefficients (kad/kdeJ, Equation 3 can be rearranged to
8= Kp I+Kp
(4)
Since the adsorption coefficient (K) is defined as the ratio of the forward to the reverse rate coefficient, it can be regarded as equilibrium constant of the adsorption reaction (Eq. 1) and directly related to the standard free energy change of the reaction (~GO) by
(5) where R is the gas constant [8.3143 J K- I mol-I] and T is absolute temperature. The fraction of the surface covered (8) can be expressed by the ratio of the adsorbed gas volume (y) to the value of y that is approached asymptotically as p becomes arbitrarily large; i.e., the maximum adsorption capacity (Yma,). Thus,
8=--L. Ymax
(6)
Equations 4 and 6 can be combined, obtaining the nonlinear form of the traditional Langmuir equation, given by
y=
YmaxKp I+Kp
(7)
that can be linearized to y -=y max K-Ky
p
(8)
by multiplying both sides by (lip + K) and solving for yip. At vel}' small values of p, K p is negligibly small and Equation 7 can be rearranged to
y=YmaxKp
(9)
36
Fate and Transport of Heavy Metals in the Vadose Zone
The adsorbed gas volume (y) increases linearly with increasing pressure (p). At high pressure, K p is much greater than 1 and Equation 7 can be simplified to
Y =Ymax
(10)
Under these conditions the maximum adsorption capacity is obtained. For the description of ion adsorption reactions in soil/soil solution systems, the adsorbed gas volume (y) is replaced by the adsorbed amount of the studied ion per unit soil (Q) and the equilibrium pressure (p) by the equilibrium soil solution concentration of the ion (C). Parameter b is introduced as a measure of the adsorption maximum instead of Ymax' and the Langmuir equation is obtained as
Q= bKC I+KC
(11)
and in its linear form as
Q
-=bK-KQ
C
(12)
Quotient Q/C is known as the distribution coefficient (Kd), and parameter K determines the magnitude of the initial slope of the isotherm and is therefore a measure of adsorption energy. The advantages of Equation 12 over other linear forms of the Langmuir equation are discussed in Veith and Sposito (1977). A graph of Q/C against Q should be a straight line with the slope -K and an x-intercept equal to b if the Langmuir equation is applicable to the studied system. The typical shape of the nonlinear Langmuir isotherm is shown in Figure 2.4. Harter and Smith (1981) summarized the assumptions on which the theoretical model of Langmuir is based: • • • • • •
The plane surfaces have a fixed number of only one kind of elementary space. Each space is able to hold only one adsorbed molecule. The surface is covered with a monolayer only. The adsorption reaction is reversible. Adsorbed molecules are not free to move laterally on the surface. Adsorption energy is the same for all sites and is not dependent on surface coverage. • There is no interaction between adsorbate molecules. • The probability of a molecule condensing on an unoccupied site or dissociating from an occupied site is not affected by coverage of adjoining sites.
Many of these assumptions are not valid in soil/soil solution systems, where different types of adsorption sites exist, and the surfaces can be covered by more than one layer. Irreversible reactions are involved, adsorbed molecules can move laterally on the surface showing interactions between each other, and the adsorption energy, as well as the
Adsorption Isotherms of Nickel in Acid Forest Soils
37
b· ................................................................................................. Q
c Agure 2.4. The Langmuir isotherm (b
= adsorption maximum).
probability of adsorption or desorption are dependent upon surface coverage. The application of the Langmuir equation for describing adsorption of solutes in dilute solution requires further assumptions (Harter and Smith, 1981) which are seldom attained in soil/soil solution systems: • There is no specific adsorption of the solvent by the surface. • There is no interaction between solvent and solute (Henry's law obeyed). • Adsorption sites are empty "holes" having an activity coefficient of 1, as well as entropy and adsorption energy.
o
Nevertheless, the Langmuir equation has been widely used in the literature, as reviewed by Travis and Etnier (1981). Adsorption data of ions in soils often show the shape of Langmuir isotherms and can be accurately described with the Langmuir equation. This may be explained according to Brunauer et al. (1967), who pointed out that the derivation of the Langmuir equation does not require all of the usual assumptions if it is developed, as originally by Langmuir (1918), on the basis of kinetic arguments instead of statistical thermodynamics, which has imposed further restrictions on the original assumptions. Moreover, Brunauer et al. (1967) noted that for adsorption on energetically heterogeneous surfaces, the free energy of adsorption would decrease with increasing surface coverage whereas the lateral interaction energies between adsorbed molecules would raise the free energy of adsorption with increasing surface coverage. In certain cases the two opposing effects would compensate for each other, making the free energy of adsorption approximately constant for a particular adsorption isotherm. Veith and Sposito (1977) found the Langmuir equation to apply, on the basis of statistical analysis, even when the concentration of the studied ions was much larger than the threshold value required to initiate precipitation. Thus, the Langmuir equation cannot be used statistically to determine whether adsorption or precipitation is occurring and the Langmuir parameters cannot be interpreted in terms of surface reactions unless there is independent evidence that only adsorption is involved. Furthermore, the calculation of the maximum adsorption capacity can involve errors of 50% and more (Harter, 1984)
38
Fate and Transport of Heavy Metals in the Vadose Zone
if the isotherm does not have the correct Langmuir shape and only low concentration data are used for the calculation. Although, according to Equation 12, a linear relationship is expected between QlC and Q, a curvilinear relationship is frequently reported, which has been attributed to nonuniformity of soil adsorption sites with respect to adsorption energy (Holford et al., 1974; Holford, 1978) or to the effect of desorbed ions in the equilibrium solution on the competition for adsorption sites (Griffin and Au, 1977; Harter and Baker, 1977, 1978). Holford et al. (1974) used the equation proposed by Langmuir (1918) for gaseous adsorption on surfaces with more than one type of elementary space of different adsorption energies, given by
(13)
and described o-phosphate adsorption by 41 soils of widely varying physical and mineralogical properties almost perfectly with the equation
Q= b1K1C I+K 1C
+
b 2K 2C I+K 2C
(14)
where subscripts indicate different discrete energy adsorption sites. The traditional Langmuir equation (Eq. 11) fitted none of the adsorption data as well as the "two-surface" Langmuir equation (Eq. 14). However, according to Posner and Bowden (1980) and Sposito (1982), the adjustable parameters in Equation 14 (b l , K I , b 2, and K 2) cannot be interpreted in terms of surface reactions unless there is independent evidence for two types of surface sites involved in the adsorption reaction. For the description of Zn adsorption by soils, Harter and Baker (1977) used the equation proposed by Boyd et al. (1947), who modified the Langmuir equation to describe simultaneous competitive adsorption of two equally charged cations, obtaining
(15)
where subscripts 1 and 2 refer, respectively, to the adsorbate ion and the ion originally on the adsorbent surface. The calculated adsorption maxima (b) were similar to those obtained from the traditional Langmuir equation (Eq. 11). Thus, the use of the traditional Langmuir equation seems satisfactory for the calculation of adsorption maximum values that are subsequently used for correlation with chemical and physical soil properties. However, according to Harter and Baker (1977), adsorption energy (K) is significantly affected by desorbed ions and should therefore be calculated by Equation 15. Following suggestions from Holford (1978), Harter and Baker (1978) combined Equations 15 and 14, accounting for heterogeneity of adsorption energy over the entire surface. The resulting equation is given by
Adsorption Isotherms of Nickel in Acid Forest Soils
39
(16)
where subscripts 1 and 2 refer to the adsorbate ion and the ion originally on the adsorbent surface, respectively, and the single and double primes indicate different discrete energy adsorption sites.
The van Bemmelen-Freundlich Equation In the works of van Bemmelen (1888) and Freundlich (1909), an empirical equation was developed which applies to the adsorption of trace amounts of molecules or ions on solid surfaces. For gaseous adsorption, the van Bemmelen-Freundlich equation has the form
y= A p
lin
(17)
where A and lin are positive-valued empirical constants that decrease with increasing temperature, with n being greater than unity in most cases of practical interest. For soil/soil solution systems, Equation 17 can be rearranged to
Q=AC13
(18)
with 13 constrained to lie between 0 and 1. By logarithmizing Equation 18, the linear form of the van Bemmelen-Freundlich equation is obtained as log Q =log A + 13 log C
(19)
Thus, a plot of log Q against log C should be a straight line, with the slope 13 and a y-intercept equal to log A if the van Bemmelen-Freundlich equation is applicable to the studied system. The typical shape of the nonlinear van Bemmelen-Freundlich isotherm is shown in Figure 2.5. Although the van Bemmelen-Freundlich equation is empirical, it was tried to establish a theoretical derivation. Actually the equation may be obtained from Langmuir's theory of monolayer adsorption by assuming that the free energy of adsorption decreases logarithmically with increasing surface coverage (Halsey and Taylor, 1947), which may be caused by particle interactions or by surface heterogeneity. The van BemmelenFreundlich equation does not provide direct information on adsorption capacity and adsorption energy, however, according to Slejko (1985), the constants A and 13 may be regarded as relative indicators of capacity and affinity, respectively. The adsorption of ions on soil particle surfaces can often be described accurately by the van Bemmelen-Freundlich equation, which has been extensively used in the literature, as reviewed by Travis and Etnier (1981). However, the same authors noted that the flexibility of the two constants allows for easy curve fitting, but does not guarantee accuracy if the data are extrapolated beyond the experimental points. Furthermore, Hemwall (1957) observed the van Bemmelen-Freundlich isotherm to apply also to pre-
40
Fate and Transport of Heavy Metals in the Vadose Zone ----
....
---.--.~
Q
c Figure 2.5. The van Bemmelen-Freundlich isotherm.
cipitation reactions. Thus, it cannot be used to differentiate between adsorption and precipitation. In the range of low adsorption, the van Bemmelen-Freundlich isotherm is similar to the Langmuir isotherm, and with proper choice of constants both isotherms can almost be brought to coincidence. Schulte (1988) reported that both equations accurately described the adsorption of heavy metals in soils. However, neither of the two equations provides information on the adsorption mechanisms involved or even whether adsorption or precipitation has occurred. Thus, both may be regarded as curve-fitting models without particular molecular significance, but with predictive capability under limited conditions (Sposito, 1989).
CASE STUDY Five forest sites in the Giinser Mountain region, eastern Austria, were selected for the study. The mean annual precipitation in the region is 1000 mm and the mean annual temperature is 8°C. The parent material of the soils is phyllite. Soil samples (n = 20) were collected according to genetic horizons, air-dried, and passed through a 2-mm sieve. The screened soils were mixed to obtain homogeneous samples and stored in plastic bags until analysis. The studied soils represent acid forest soils [pH(CaCI 2) between 2.9 and 4.7] showing high variation in organic carbon content, specific surface area, and cation exchange capacity. Soil texture, however, displays low variation. The characteristic soil properties are given in Zehetner (1997). Particle size analysis of the fraction < 2 mm was carried out by a combined sieve and pipette method (Blum et al., 1996). The specific surface area (SSA) was determined by means of a single point in the water adsorption isotherm according to Puri and Murari (1964), who found for various soils and clays that a mono-molecular water layer was completed in equilibrium with a relative water vapor pressure (p/po) of 0.S273. Three grams of air-dried soil were equilibrated with p/po = 0.S2 maintained by a saturated Mg(N0 3h solution at 24.SoC. The samples were weighed, dried to constant weight at lOsoC, and weighed again. The time required for equilibration (seven days at p/po = 0.S2 and 48 hours at lOsoC, respectively) was determined in preliminary experiments. The specific surface area [m 2 g-l] was calculated by the equation
Adsorption Isotherms of Nickel in Acid Forest Soils
41
(20)
where X [g g-I] is the amount of water adsorbed per mass soil in equilibrium with p/po = 0.52, N is Avogadro's number [6.02205 X 1023 mol-I], A is the area occupied per molecule of water [1.08 X 10- 19 m 2], and M(H 2 0) is the molecular weight of water [18.015 g mol-I]. For the maintenance of a constant relative water vapor pressure, a saturated salt solution is preferable to conc. H 2S04 since the concentration of the latter changes due to evaporation and condensation during the tests. The experiments were conducted without vacuumizing (Ponizovskiy et al., 1993) and by applying a desorption rather than an adsorption method (Farrar, 1963) in order to avoid irreversible changes in soil structure. Soil pH was measured in a 1:2.5 soil:O.Ol M CaC12 extract after two hours of equilibration using a combination pH electrode (Blum et al., 1996). The cation exchange capacity (CEC) at soil pH was calculated as the sum of Al 3 +, Ca2 +, Fe 3 +, H+, K+, Mg2+, Mn2+, and Na+ extracted by 0.1 M BaCl2 at 1:20 (Blum et al., 1996), and corrected for H+ due to Al hydrolysis (Meiwes et al., 1984). The sum of mono- and divalent cations in the BaCl2 extract was calculated and termed CEC(2+)' Soil organic carbon (OC) corresponds to the total carbon content measured with a CNS analyzer since the studied soils are carbonate-free. Oxide-bound Al, Fe, and Mn were extracted with the bicarbonatecitrate-dithionite method according to Mehra and Jackson (1960). For each soil sample the quantitative soil:solution ratio necessary to obtain a soil paste at the Atterberg "upper plastic limit" was determined with the liquid limit apparatus described in Hillel (1980). It varied between 1:0.4 and 1:0.7 for the studied soils, corresponding to their specific surface area. The determined soil:solution ratio of each soil was used for a saturated water extract (equilibrium soil solution) and for the adsorption experiments. Aqua bidest. and Ni solutions, respectively, were added to 100 g of the airdried and screened soil. The soil slurry was repeatedly stirred during the 48-hour-equilibration period at 20°C, which was determined in preliminary tests. The solution was separated from the soil by centrifugation and membrane filtration (cellulose acetate; 0.45 pm). In the equilibrium soil solution, major cations were measured by ICP-AES, Ni by GF-AAS, and anions by ion chromatography. pH and electrolytic conductivity (EC) were determined, and dissolved organic carbon (DOC) was estimated by correlation between DOC and absorbance at 254 nm measured with a UV-VIS spectrophotometer (Brandstetter et al., 1996). The properties of the equilibrium soil solutions are given in Zehetner (1997). For each soil a Ni adsorption isotherm with 13 points was obtained by applying NiC1 2 solutions in increasing concentrations (between 2 X 10-7 and 10-2 mol L- I). Depending on the quantitative soil:solution ratio and on the concentration of Ni in the added solution, between 0.08 and 7000 }lmol Ni per kg soil were added. The soluble Ni was measured with Flame- and GF -AAS. The total amount ofNi retained by the soil was calculated by the difference between the initial and the final concentration in each solution. The horizons of a spodosal (soils no. 8-12) were used for adsorption experiments at soil:solution ratio of 1:5. NiCl 2 solutions of 10-2, 10-4, and 2 X 10-6 mol L- I, and a blank
42
Fate and Transport of Heavy Metals in the Vadose Zone
solution (aqua bidest.) were added to 8 g of the air-dried and screened soil. The obtained soil suspensions were shaken for 48 hours at 20°C, centrifuged and passed through membrane filters (cellulose acetate; 0.45 pm). For the calculation of adsorption isotherms, the initial amounts of potentially mobile Ni on the soil solid surfaces have to be taken into account. They were estimated by the fraction extractable with 0.05 M Na2EDTA at 1:10 according to Blum et al. (1996). Since EDTA does not attack the silicate structures (Brummer et aI., 1986; Konig et aI., 1986), it may be considered as an extractant for potentially mobile heavy metals (Schulte, 1988, 1994a, 1994b; Schulte and Beese, 1994a, 1994b). The horizons of a spodosal (soils no. 8-12) were used for fractionating adsorbed Ni. Subsequently to the adsorption experiments (saturation extracts with NiCl2 solutions of 10-2, 10-4, and 2 X 10--6 mol L- i , and aqua bidest. as blank solution) an aliquot of each soil was washed chloride-free, air-dried and sequentially extracted by a seven-step sequential extraction procedure (Table 2.1) developed by Zeien and Brummer (1989). The residual fraction (fraction 7) was obtained by microwave digestion with H 20 2IHN0 3 instead of the procedure suggested by Zeien and Brummer (1989) using HCIO/HN03 or HF/HCI0 4 •
Adsorption versus Precipitation Sadiq and Zaidi (1981) suggested the precipitation of nickel ferrite (NiFe204) to be the major retention process of Ni in 27 studied soils (pH range from 5.5 to 8.6), and Sadiq and Enfield (1984a, 1984b) concluded from theoretical and experimental studies on the solid phase formation ofNi that NiFe204 was the most stable solution controlling solid species of Ni in soils. Other solid phases of Ni including carbonate, phosphate, sulfates, halides, oxides and hydroxides, sulfides, as well as nickel silicate (Ni2Si04) and nickel aluminate (NiAl 20 4) , were metastable with respect to nickel ferrite under the conditions most commonly encountered in soils. However, according to Ma and Lindsay (1995), NiFe204 has not yet been observed in soils and the conditions necessary for its formation have not yet become clear. Thus, the validity of NiFe204 as a mineral controlling Ni 2+ activities in soil solutions needs further examination. In the presented study the soil solutions were undersaturated with respect to NiFe204 even at the highest applied initial Ni concentration (10 mmol L- i ), assuming Fe3 + activity in equilibrium with soil-Fe according to Lindsay (1979). Thus, adsorption including ion exchange is likely to be the process that controlled solution concentrations of Ni under the studied conditions.
langmuir and van Bemmelen-Freundlich Isotherms The traditional, as well as the "two-surface" Langmuir equations (Eqs. 11 and 14, respectively) were fitted to the experimental data of individual soils by means of an iterative nonlinear x2-technique. The results of the fitting procedure are presented in Zehetner (1997). Nonlinear fitting of Equation 11 was preferred over linear regression according to Equation 12 since data variability is reduced in a graph of Q/C against Q. The quality of fit was improved, especially in the low concentration range, by applying the "two-surface" instead of the traditional Langmuir equation. This is evident from obtained X2 values, which are lower by an order of magnitude for the "two-surface"
Table 2.1. Sequential Extraction Procedure Used for the Fractionation of Adsorbed Nia Fraction 1
2
3
4
5
6
7 a
Mobility and Binding Form Mobile fraction: Water-soluble and exchangeable (Le., nonspecifically adsorbed) metals as well as easily soluble metal-organic complexes Easily mobilizable fraction: More specifically adsorbed metals, close to the surface occluded and carbonate bound metals as well as less stable metal-organic complexes Metals bound to Mn-oxides
Organically bound metals
Metals bound to amorphous Fe-oxides Metals bound to crystalline Fe-oxides Residual fraction
Adapted from Zeien and Briimmer, 1989.
Extractant
Extraction Conditions
Soil:Solution Ratio 0
1 M NH4N03 (unbuffered)
24 hr shaking at 20
(
1 :25
1 M NH40Ac (pH 6.0)
24 hr shaking at 200 (
1 :25
Wash Step
1 M NH 4N0 3 (unbuffered) at 1 :12.5; 10 min shaking at 200 (
;x:. Q. VI
0 .... "t:I
.... o· ::
0.1 M NH 2 OH-H(1 + 1 M NH 40Ac (pH 6.0) 0.025 M NH4EDTA (pH 4.6) 0.2 M NH 4-oxalate (pH 3.25) 0.1 M ascorbic acid + 0.2 M NH 4-oxalate (pH 3.25) 65% HN0 3 + 30% H2 0 2
30 min shaking at 200 (
90 min shaking at 20 0 (
4 hr shaking at 20 0 (
1 :25
1 :25
1 :25
in the dark
30 min shaking in a water basin at 96 ± 3°( in the
1 :25
light
Microwave digestion
1 :20
2 times 1 M NH 4 0Ac (pH 6.0) at 1 :12.5; 10 min shaking at 200 ( 1 M NH4 0Ac (pH 4.6) at 1 :12.5; 10 min shaking at 20 0 ( 0.2 M NH4-oxalate (pH 3.25) at 1 :12.5; 10 min shaking at 20 0 ( in the dark 0.2 M NH4-oxalate (pH 3.25) at 1 :12.5; 10 min shaking at 20 0 ( in the dark
iii
0 ....::r
fb ....
3
VI
0
-+.
z ;::;. ;s:;
!:E.
:i"
,.,
;x:.
0: ."
..,
0 fb
.... VI
Vi
9.
iii
J:;;.
w
44
Fate and Transport of Heavy Metals in the Vadose Zone
10
AEh (soil no. 8)
....-. ......,
en
.::s:.
0.1
0
E
E .........
0.01
0 - - - - - Langmuir - - "two-surface" Langmuir - - - van Bemmelen - Freundlich
0.001
0.001
0.01
0.1
10
C [mmol L."1] Figure 2.6. Adsorption isotherms fitted to the experimental data of soil no. 8.
10
Bw (soil no. 11)
....-. ......,
en
.::s:.
0.1
0
E E 0.01 ......... 0 - - - - - Langmuir - - "two-surface" Langmuir - - - van Bemmelen - Freundlich
0.001
0.001
0.01
0.1
10
C [mmol L- 1] Figure 2.7. Adsorption isotherms fitted to the experimental data of soil no. 11.
Langmuir equation compared to the traditional form (Zehetner, 1997), and can be seen in Figures 2.6 and 2.7 for a topsoil and a subsoil horizon (soils no. 8 and 11). In the "twosurface" Langmuir equation, subscript 1 indicates surfaces with low adsorption maximum and high adsorption energy, and subscript 2 indicates surfaces with high adsorption maximum and low adsorption energy (Zehetner, 1997). The linear form of the van Bemmelen-Freundlich equation (Eq. 19) was fitted to the experimental data of individual soils by linear regression, obtaining correlation coefficients (R) ~ 0.988, and significance at the 0.001 level (Zehetner, 1997). For soils no. 8
Adsorption Isotherms of Nickel in Acid Forest Soils --_._--
---------------
----
.---~,.
--
- - - - - - ..
45
- - ---- - - - - - - - - ----- ---..-
50 0
0 0
40
0
0
0
,........,
..-, 30 Ol .::t:-
0
AEh (soil no. 8)
•
Bw (soil no. 11)
0
o
0
~
"0
~
20 10
0 0
• • ••
0 0.0001
0.001
•• 0.01
0.1
1
10
Initial Ni concentration [mmol L- 1] Figure 2.8. Effect of applied initial Ni concentration on distribution coefficient values (~) for soils no. 8 and 11.
and 11, the resulting van Bemmelen-Freundlich isotherms are shown in Figures 2.6 and 2.7, respectively. Correlation coefficients of calculated adsorption isotherm parameters and the soil properties SSA, CEC, and CEC(2+) are presented in Zehetner (1997). Except for b l , the R values were lowest for correlation with CEC, intermediate with SSA, and highest with CEC(2+). Thus, ion exchange against mono- and divalent cations was possibly a major mechanism of Ni adsorption under the studied conditions whereas sites occupied by trivalent cations like Al 3 + and Fe 3 + were apparently less involved. Partial hydrolysis of trivalent cations may result in particularly strong adsorption due to decreased hydration and increased charge intensity. The adsorption maximum of the high energy sites (b l ) displayed the tightest correlation with SSA, followed by CEC and CEC(2+). This may be due to specific adsorption on organic surfaces, of which SSA is a better measure than CEC and CEC(2+)' as indicated by correlation with OC (R = 0.957***, 0.841 ***, and 0.781 ***, respectively). The adsorption energy parameters (K, Kl' and K 2), which may be considered as pHdependent (Schulte, 1988; ~chulte and Beese, 1994a, 1994b), did not correlate with pH(CaCI 2) nor with pH(H 20). This may be due to the relatively narrow pH range of the studied soils. EDTA-extractable Ni correlated with KI (R = 0.907~'**) but not with K2 and K. This indicates that initially adsorbed Ni is predominantly bound to high energy sites, which may be caused by specific adsorption on organic surfaces. As shown in Figure 2.8 for soils no. 8 and 11, distribution coefficients (~) were higher for topsoils than for subsoils and decreased with increasing applied initial Ni concentration. This is consistent with the results of Basta and Tabatabai (1992), and may be attributed to affinity of Ni for highly selective sites (specific adsorption) at low concentration and to adsorption on less selective sites (nonspecific adsorption) at larger amounts of Ni added. In topsoil horizons (OC > 15 g kg-I) organic carbon content (OC) was the soil constituent mainly affecting adsorption capacity parameters whereas in subsoil horizons
46
Fate and Transport of Heavy Metals in the Vadose Zone 10
•
AEh (soil no. 8)
0 0
......
.,....
•
I
Q)
.::t:. 0
E E ........
a
0.1
0 0
0 0
0.01
0.001 0.0001
.~o
0.001
saturation extract
•
1 : 5 extract
0.1 [mmol L-1]
0.01
C
0
10
Figure 2.9. Quantity-intensity relationships of Ni obtained in saturation extracts and at 1 :5 for soil no. 8.
(oe < 15 g kg-I) dithionite-extractable Mn (Mnd) was the major influencing constituent. Correlation coefficients are presented in Zehetner (1997). In the topsoils, oe displayed tighter correlation with b l than with b 2 whereas in the subsoils, Mnd showed the opposite trend. Thus, organic matter may be strongly involved in (specific) adsorption of Ni on high energy sites. Neither in topsoils nor in subsoils did clay content show significant correlations with adsorption capacity parameters. This may be due to the relatively low variation in the clay content of the studied soils and the old age of the soil material, indicated by low feldspar and smectite contents and high kaolinite contents, which was determined for selected soils by X-ray diffraction. Effect of Soil:Solution Ratio on Quantity-Intensity Relationships The adsorption experiments were conducted in saturation extracts at soil:solution ratios between 1:0.4 and 1:0.7, which are closer to field conditions than tighter soil:solution ratios most commonly applied in adsorption experiments in the literature. For the horizons of a spodosal (soils no. 8-12), the adsorption isotherms obtained in saturation extracts were compared with quantity-intensity relationships at a soil:solution ratio of 1:5 (Figures 2.9 to 2.13). In the low concentration range (Ni solution concentrations in the order of 10-4 to 10-.3 mmol L -I), the quantity-intensity relationships were hardly affected by the soil:solution ratio. At applied initial Ni concentrations of 0.1 and 10 mmol L -I, however, the amounts of Ni adsorbed per mass soil were significantly higher in the 1:5 extracts than in the saturation extracts, and despite a slighter decrease of Ni solution concentrations due to adsorption in the 1:5 extracts, the obtained datapoints were significantly above the adsorption isotherms obtained in saturation extracts. Thus, in the low concentration range quantity-intensity relationships may be satisfactorily described by using tighter soil:solution ratios whereas at higher concentrations these would cause false assessment of adsorption. At a given solution concentration the amount adsorbed per mass soil would
Adsorption Isotherms of Nickel in Acid Forest Soils 10
•
Bhs (soil no. 9)
o
o
.......... ::t:.
o E E .........
a
o
•o
'0>
0.1
o
47
o
~o
0.01
o saturation extract •
1: 5 extract
0.001 +-...,.......,...,........,.".---..-...........rTTTTr-....-r-rTTTnr-.....-.........n-nr--,-.......,..........." 0.1 10 0.001 0.01 0.0001
C [mmol L- 1] Figure 2.1 O. Quantity-intensity relationships of Ni obtained in saturation extracts and at 1 :S for soil no. 9.
10
•
Bs (soil no. 10)
.......... ,
0
0>
::t:. 0
E E .........
a
·0
0
0
0.1 0 0
0.01
0.001 0.0001
~o 0.001
saturation extract
•
1 : 5 extract
0.1 [mmol L- 1]
0.01
C
0
10
Figure 2.11. Quantity-intensity relationships of Ni obtained in saturation extracts and at l:S for soil no. 10.
be overestimated, and at a given adsorbate concentration the corresponding equilibrium solution concentration would be underestimated. Thus, soil:solution ratios of 1:5 and tighter, used in the majority of (Ni) adsorption studies in the literature, may result in quantity-intensity relationships that strongly overestimate adsorption, especially at higher concentrations, and are therefore of limited value for modeling metal mobility in natural systems. Specific adsorption may be responsible for the observed low influence of the soil:solution ratio on quantity-intensity relationships in the low concentration range. Involving covalent and ionic bonding, specific adsorption can be regarded as a chemical
48
Fate and Transport of Heavy Metals in the Vadose Zone 10
Bw (soil no. 11)
•
....... ......
~ ~
• 0
o
0.1
E ........
a
o
o
o 0.Q1
o
o
0.0001
0.001
o saturation extract • 1: 5 extract 0.01
10
0.1
C [mmol L- 1] Figure 2.12. Quantity-intensity relationships of Ni obtained in saturation extracts and at 1:5 for soil no. 11. 10
•
Cw (soil no. 12) o
.......... ...... '0)
•
~
~
0
o 0.1
o
E ........
a
o
o 0.01
o
o saturation extract • 1: 5 extract
o.
0.001
0.1 [mmol L- 1]
0.01
C
10
Figure 2.13. Quantity-intensity relationships of Ni obtained in saturation extracts and at 1:5 for soil no. 12.
equilibrium reaction, establishing a certain solution concentration at a given adsorbate concentration regardless of the soil:solution ratio. However, at higher concentrations, where nonspecific coulombic interactions tend to dominate, adsorption was enhanced due to larger amounts of adsorptive per mass soil at 1:5 compared to saturation extracts. Applied initial Ni concentration of 10 mmol L- i provided 4 to 7 mmol Ni per kg soil in saturation extracts depending on the respective soil:solution ratio, but as much as 50 mmol Ni per kg soil at 1:5. Differences in ionic strength and pH may contribute to enhanced Ni adsorption at 1:5 compared to saturation extracts. In 1:5 extracts ionic strength was lower than in saturation extracts, as indicated by electrolytic conductivity (Zehetner, 1997). This may cause
Adsorption Isotherms of Nickel in Acid Forest Soils 100 -
AEh (soil no. 8) r -
c
(])
r-
-
-
(]) L-
r-
r*,
~ ()
-
49
50-
C. .........
c
a
.~
c.
L-
aC/)
0.0005 0.001 0.002 0.005
0.02
0.05
0.1
0.5
5
10
Initial Ni concentration [mmol L- 1]
'0
ro ~ :.;::::; ro
0.01
-50 -
D
(])
0:::
*
saturation extract
1: 5 extract
-100- ' -
Figure 2.14. Effect of applied initial Ni concentration and soil:solution ratio on relative adsorption for soil no. 8.
lower ion competition and greater width of diffuse layers. Ni adsorption did not affect solution pH at applied initial Ni concentrations < 1 mmol L-I, however, it resulted in a significant decrease of pH at 10 mmol L-i, which was more strongly pronounced in saturation extracts than in l:S extracts (Zehetner, 1997). Relative adsorption (Arel) [percent] was calculated as
(21)
where Co and CI are the Ni concentrations in the applied solution before and after equilibration, respectively. For a topsoil and a subsoil horizon (soils no. 8 and 11) the influence of applied initial Ni concentration and soil:solution ratio on relative adsorption is shown in Figures 2.14 and 2.1S. In the low concentration range, desorption played a significant role leading to a negative value of 1\.el at the lowest applied initial Ni concentration. Relative adsorption then increased with applied initial Ni concentration until a maximum was reached at around O.OOS to 0.01 mmol L-I, and decreased again above 0.1 mmol L- 1 in the presented subsoil horizon and above 1 mmol L -I in the presented topsoil horizon. This difference may be explained by adsorption capacities and affinities higher in topsoils than in subsoils (Zehetner, 1997). Although the amount of Ni adsorbed per mass soil was significantly higher in l:S extracts than in saturation extracts, relative adsorption was significantly lower at the highest applied initial Ni concentration in the topsoil horizon and at all studied concentrations in the subsoil horizon. The similar 1\.el at applied initial Ni concentrations of 0.002 and 0.1 mmol L -I in the topsoil may again result from higher adsorption capacities and affinities compared to the subsoil (Zehetner, 1997).
50
Fate and Transport of Heavy Metals in the Vadose Zone 100 -
Bw (soil no. 11) -
........ +-' C
~
I-
Q)
-
50 -
-
c.. .......... c
o :;::::;
c.. oen "'0 ro
o
0.0002
I-
n 0.0005 0.001 0.002 0.005
0.01
0.02
0.05
0.1
0.5
1
Initial Ni concentration [mmol
L- 1]
5
10
Q)
> ro
:;::::;
-50 -
Q)
c::
-
o *
saturation extract
1: 5 extract
-100 -
Figure 2.15. Effect of applied initial Ni concentration and soll:solution ratio on relative adsorption for soil no. 11.
Fractionation of Adsorbed Nickel The h izons of a spodosal (soils no. 8-12) were used to study the influence of applied imtla i concentration on the fractionation of adsorbed Ni (Figures 2.16 to 2.20). e blank treatment displays the distribution of initially present Ni. In all the studied rizons, the dominant initially present Ni fractions were, in the order of decreasing Ni concentration, residual Ni, Ni bound to crystalline Fe-oxides, and Ni bound to amorphous Fe-oxides (fractions 7, 6, and 5). In the Bs, Bw, and Cw horizons (soils no. 10, 11, and 12) these three fractions almost exclusively contributed to the total Ni contents, which were higher than in the horizons above. In the Bhs horizon (soil no. 9), organically bound Ni (fraction 4) also significantly contributed to the total Ni content, and notable Ni contents could be identified, in the order of decreasing Ni concentration, in the mobile, the Mn-oxide bound, and the easily mobilizable fractions (fractions 1, 3, and 2). In the AEh horizon (soil no. 8) besides fractions 7, 6, and 5, only organically bound and mobile Ni were identified in substantial concentrations. According to Tu (1996), previous studies on the distribution of Ni in soils identified 50-80% and 20-30% of native Ni in the residual and in the Fe- and Mn-oxide bound fractions, respectively, and usually less than 2% in the soluble plus exchangeable fractions. Generally, this was observed in the presented study in which, however, the ratio of residual to oxide-bound Ni was shifted toward the oxide-bound fractions. Applied initial Ni concentrations of 0.002 and 0.1 mmol L- 1 were too small to significantly affect total Ni contents and to identify clear trends in the fractionation of adsorbed Ni, ho'Wever, especially in the uppermost horizons AEh and Bhs, the mobile Ni fraction clearly increased with applied initial Ni concentration. At 10 mmol L -I, total Ni contents were significantly raised in all the studied horizons, and the major part of adsorbed Ni was in the mobile fraction that became the dominant fraction in the AEh and
Adsorption Isotherms of Nickel in Acid Forest Soils
51
6
AEh (soil no. 8) ~
5
~
~
Fraction:
o 4 E
.sZ
=
7 56
~ _
3
4 3 2 1
EEEE!I
= = =
"0 Q)
t5 2 ~
x UJ o
0.002
0.1
10
Initial Ni concentration [mmol L-
1
]
Figure 2.16. Distribution of Ni among the seven sequentially extracted fractions at different applied initial Ni concentrations for soil no. 8. 6
Bhs (soil no. 9) ~
5
~ (5 4
Fraction'
E
=
7 5 6
EITEll
4 3 2 1
.sz
= _
3
=
"0
= =
Q)
t5
~
2
x
UJ
o
0.002
0.1
10
Initial Ni concentration [mmol L-
1
]
Figure 2.17. Distribution of Ni among the seven sequentially extracted fractions at different applied initial Ni concentrations for soil no. 9.
Bhs horizons. The easily mobilizable fraction was also increased in all the studied horizons whereas Mn-oxide bound and organically bound Ni were substantially increased only in the AEh and Bhs horizons and to a smaller extent in the Bs horizon. An increase in the fraction bound to amorphous Fe-oxides was identified in the AEh horizon, as well as in the Bw and in the Cw horizons. Fractions 6 and 7 were apparently not affected by adsorption at applied initial Ni concentrations of lO mmol L- 1 and below. The fractionation of adsorbed Ni was conducted immediately after the adsorption experiments, which may explain the strong dominance of mobile Ni in all the studied horizons. Redistribution of adsorbed Ni toward more stable forms may occur with time.
52
Fate and Transport of Heavy Metals in the Vadose Zone 6
Bs (soil no. 10) ~
5
'0)
.::<:
Fraction'
(5 4
= = = =
E
.sz
3
-
"C (])
=
tl 2
=
~
>< w
o
7 6 5 4 3 2 1
o 0.002 0.1 10 Initial Ni concentration [mmol r1]
Figure 2.18. Distribution of Ni among the seven sequentially extracted fractions at different applied initial Ni concentrations for soil no. 1O. 6
Bw (soil no. 11) 5
~ (5
E
Fraction:
4
= =
7
3
-
5
2
= =
1
~
Z
= =
"C
~ ~
x UJ o
o
0.002
0.1
6
4 3 2
10
Initial Ni concentration [mmol L- 1] Figure 2.19. Distribution of Ni among the seven sequentially extracted fractions at different applied initial Ni concentrations for soil no. 11.
Adsorption Density and General Adsorption Density Isotherms For different acid forest soils, Schulte (1988, 1994a, 1994b), as well as Schulte and Beese (1994a, 1994b) were able to approximately describe quantity-intensity relationships of various heavy metals by a single adsorption isotherm for each, by plotting the equilibrium solution concentration (C) against the adsorption density based on the specific surface area (AD ssA) [ions m-2], defined as
AD SSA
_ (SEDTA + S) N SSA 106
(22)
Adsorption Isotherms of Nickel in Acid Forest Soils
53
6
Cw (soil no. 12) ~
5
~ o 4 E
=
7
z
= -
5 4 3 2
Fraction·
E.
3
= = = =
"0 Q)
13 2 ~
x w o
0.002
0.1
6
1
10
Initial Ni concentration [mmol r1]
Figure 2.20. Distribution of Ni among the seven sequentially extracted fractions at different applied initial Ni concentrations for soil no. 12.
where SEDTA [mmol kg-I] is the EDT A-extractable amount of Ni initially adsorbed, S [mmol kg-I] is the amount adsorbed during the test, N [6.02205 X 1023 ions mol-I] is Avogadro's number, and SSA [m 2 g-I] is the specific surface area. Adsorption processes, especially nonspecific adsorption through ion exchange, are influenced by the intrinsic surface charge, which CEC is a measure of. By dividing ADsSA through the intrinsic surface charge density (CEC/SSA) [mole m-2], the adsorption density based on exchange sites (AD cEC) [ions mole-I] was obtained as AD
_ (SEDTA + S) N CEC
CEC -
(23)
where CEC [mmole kg-I] is the cation exchange capacity. Since it was shown that adsorption isotherm parameters were closely related to the exchange sites occupied by mono- and divalent cations (Zehetner, 1997), CEC in Equation 23 was further substituted for CEC(2+)' obtaining AD cE C(2+). Surface based and charge based adsorption densities were calculated for each point of the isotherm of each soil, and the linear form of the van Bemmelen-Freundlich equation (Eq. 19) was fitted by linear regression to the pairs of log ADsSA (log AD cEC and log AD cEC(2+)' respectively) and log C, obtaining general adsorption density isotherms by combining all the studied soils (Figures 2.21 to 2.23). Similar fits were obtained by using SSA and CEC as reference quantities, however, the quality of fit was improved when only the proportion of CEC occupied by mono- and divalent cations (CEC(2+» was used as the reference. By means of the proposed general adsorption density isotherms, quantity-intensity relationships of native Ni can be estimated and the behavior of deposited Ni can be assessed in acid soils of different composition if the initially adsorbed amount (Q) or the corresponding solution concentration (C) and either SSA, CEC, or CEC(2+) are available. Use of CEC as the reference quantity, which is usually available in soil databases, yields similar accuracy as SSA, how-
54
Fate and Transport of Heavy Metals in the Vadose Zone
........ 1016
,
N
E
IIIPO
II)
c:
g
1015
«
II) II)
Cl
«
10 14
log ADSSA = 16.1948 + 0.5887 log C R = 0.930***, n = 260 1013+-~~~r-~~~~~~~~~~~~~~
0.0001
0.001
0.01
10
0.1
C [mmol r1] Figure 2.21. General adsorption density isotherm of Ni with SSA as the reference quantity.
1023
..;() 1022
-0
E II)
§
1021
;=.. ()
w
()
Cl «
1020
log AD cEC = 22.3098 + 0.5937 log C
R =0.928***, n =260
1019 -I-~"""""""""r-~--""'TTTTl-""""'~-.nr-'-..........r-rn-nr--r-"""""TTT1"T1 0.0001
0.001
0.01
0.1 1
C [mmol L-
10
]
Figure 2.22. General adsorption density isotherm of Ni with CEC as the reference quantity.
ever, if CEC(2+) is available, the accuracy can be strongly improved. As indicated by the correlation of adsorption isotherm parameters with SSA, CEC, and CEC(2+) (Zehetner, 1997), the better generalizibility of adsorption density isotherms with CEC(2+) as the reference shows that ion exchange against mono- and divalent cations was probably the principal mechanism of Ni adsorption under the studied conditions.
SUMMARY Nickel adsorption was studied in acid forest soils. The traditional and the "two-surface" Langmuir equations as well as the van Bemmelen-Freundlich equation were fitted to the experimental data. At low concentration, specific adsorption on organic surfaces
Adsorption Isotherms of Nickel in Acid Forest Soils
-
..--,
55
1023
u
0
E 1022 IIJ
c::
g £' 10 o W
21
()
o
«
log AD CEC {2+) = 22.8322 + 0.5948 log C
1020
0.0001
R
0.001
= 0.969***, n = 260 0.01
C
0.1 [mmoll- 1]
10
Figure 2.23. General adsorption density isotherm of Ni with CEC(2+l as the reference quantity.
may occur to a certain degree, however, exchange against mono- and divalent cations is considered as the primary mechanism of Ni adsorption in the studied soils. Organic matter and Mn-oxides may be the most effective adsorbents in topsoils and subsoils, respectively. By means of sequential extraction, adsorbed Ni was predominantly found in the mobile fraction, involving water-soluble and exchangeable Ni, as well as easily soluble metal-organic complexes. In order to obtain close-to-field conditions, adsorption experiments were conducted in saturation extracts. Comparison to adsorption at 1:5 showed that, especially at higher concentrations, adsorption would be strongly overestimated if tighter soil:solution ratios were applied. Using the van Bemmelen-Freundlich equation, general adsorption density isotherms were developed for the studied soils. Similar fits were obtained by using specific surface area (SSA) and cation exchange capacity (CEC) as reference quantities, however, the quality of fit was improved when only the proportion of CEC occupied by mono- and divalent cations (CEC(2+) was used as the reference.
REFERENCES Adriano, D.C. Trace Element.J in the TerrutriaL Environment. Springer-Verlag, New York, 1986. Basta, N.T. and M.A. Tabatabai. Effect of cropping systems on adsorption of metals by soils: II. Effect of pH. Soil Sci., 153, pp. 195-204, 1992. Blum, W.E.H., H. Spiegel, and W.W. Wenzel. Boden.wAand.Jinventur. Konzeptwn, Durchfuhrung und Bewertung. Empfehlungenmr VereinheitLichung der Vorgang.JweiJe in (j.Jterreich. 2. uberarbeitete Auflage, Bundesministerium fur Land- und Forstwirtschaft, Bundesministerium fur Wissenschaft, Verkehr und Kunst, Wien, 1996. Boyd, G.E., J. Schubert, and A.W. Adamson. The exchange adsorption of ions from aqueous solutions by organic zeolites. 1. Ion exchange equilibria. J. Am. Chern. Soc., 69, pp. 2818-2829, 1947. Brandstetter, A., R.S. Sletten, A. Mentler, and W.W. Wenzel. Estimating dissolved organic carbon in natural waters by UVabsorbance (254 nm). Zeit.Jchrijt fur Pjlanzenerniihrung und BOden!cunde, 159, pp. 605-607, 1996.
56
Fate and Transport of Heavy Metals in the Vadose Zone
Brummer, G.W., J. Gerth, and U. Herms. Heavy metal species, mobility and availability in soils. Zeitdchriftfur Pflanunerniihrung und Bodenkunde, 149, pp. 382-398, 1986. Brunauer, S., L.E. Copeland, and D.L. Kantro. The Langmuir and BET Theories. In The SoLidGad Interface, Volume 1, pp. 77-103, E.A. Flood, Ed., Marcel Dekker, New York, 1967. Chapman, D.L. A Contribution to the Theory of Electrocapillarity. The London, Edinburgh, and DubLin PhilfJdophicaL Magazine and JournaL of Science, 6th series, 25, pp. 475-481, 1913. Everett, D.H. ManuaL of Sym6014 and Terminology for PbydicochemicaL QpantitieJ and UnitJ. Appendix II: Dejinilwnd, Terminology and Sym6014 in CoLLoid and Swface ChemiJtry. Butterworths, London, 1972. Farrar, D.M. The use of vapour-pressure and moisture-content measurements to deduce the internal and external surface area of soil particles. J. SoiL Sci., 14, pp. 303-321, 1963. Fendorf, S.E., G.M. Lambie, M.G. Stapleton, M.J. Kelley, and D.L. Sparks. Mechanisms of chromium(III) sorption on silica: 1. Cr(lll) surface structure derived by extended X-ray absorption fine structure spectroscopy. Environ. Sci. TechnoL., 28, pp. 284-289, 1994. Forrester, S.D. and C.H. Giles. From manure heaps to monolayers. One hundred years of solutesolvent adsorption isotherm studies. Chem. Ind., pp. 318-325, 1972. Freundlich, H. KapiLlarchemie. Akademische Verlagsgesellschaft, Leipzig, 1909. Giles, C.H., T.H. MacEwan, S.N. Nakhwa, and D. Smith. Studies on adsorption. Part XI: A system of classification of solution adsorption isotherms, and its use in diagnosis of adsorption mechanisms and in measurement of specific surface areas of solids. J. Chem. Soc., London, pp. 3973-3993, 1960. Giles, C.H., D. Smith, and A. Huitson. A general treatment and classification of the solute adsorption isotherm. I: Theoretical. J. CoLloid Interface Sci., 47, pp. 755-765, 1974a. Giles, C.H., A.P. D'Silva, and LA. Easton. A general treatment and classification of the solute adsorption isotherm. Part II: Experimental interpretation. J. CoLloid Interface Sci., 47, pp. 766778, 1974b. Gouy, M. Sur la constitution de la charge electrique a la surface d'un electrolyte. JournaL de PhYdi<JUC Theori<Juc et AppLi<Juee, 4" serie, 9, pp. 457-468, 1910. Griffin, RA. and A.K. Au. Lead adsorption by montmorillonite using a competitive Langmuir equation. Soil Sci. Soc. Am. J., 41, pp. 880-882, 1977. Halsey, G. and H.S. Taylor. The adsorption of hydrogen on tungsten powders. J. Chem. PhYd., 15, pp. 624-630, 1947. Harter, RD. Curve-fit errors in Langmuir adsorption maxima. SoiL Sci. Soc. Am. J., 48, pp. 749752, 1984. Harter, R.D. and D.E. Baker. Applications and misapplications of the Langmuir equation to soil adsorption phenomena. SoiL Sci. Soc. Am. J., 41, pp. 1077-1080, 1977. Harter, RD. and D.E. Baker. Further reflections on the use of the Langmuir equation in soils research. SoiL Sci. Soc. Am. J., 42, pp. 987-988, 1978. Harter, RD. and G. Smith. Langmuir Equation and Alternate Methods of Studying "Adsorption" Reactions in Soils. In ChemiJtry in the Soil Environment, ASA Special Publication Number 40, pp. 167-182, R.H. Dowdy et aI., Eds., American Society of Agronomy, Soil Science Society of America, Madison, WI, 1981. Hemwall, J.B. The fixation of phosphorus by soils. Adv. Agron., 9, pp. 95-112, 1957. HilleL 0. Fundamenta14 of SoiL Pbyd0. Academic Press, London, 1980. Holford, I.C.R. Soil adsorption phenomena and the Langmuir equation. SoiL Sci. Soc. Am. J., 42, pp. 986-987, 1978. Holford, I.C.R., RW.M. Wedderburn, and G.E.G. Mattingly. A Langmuir two-surface equation as a model for phosphate adsorption by soils. J. SoiL Sci., 25, pp. 242-255, 1974. Konig, N., P. Baccini, and B. Ulrich. Der EinfluE der naturlichen organischen Substanzen auf die Metallverteilung zwischen Boden und Bodenlosung in einem sauren Waldboden. Zeitdchrift fur Pflanunerniihrung und Bodenkunde, 149, pp. 68-82, 1986.
Adsorption Isotherms of Nickel in Acid Forest Soils
57
Langmuir, I. The adsorption of gases on plane surfaces of glass, mica, and platinum. J. Am. Chem. Soc., 40, pp. 1361-1403, 1918. Lindsay, W.L. ChemicaL equilibria in Soiu. John Wiley & Sons, New York, 1979. Ma, Q. Y. and W.L. Lindsay. Estimation of Cd 2+ and Ni 2+ activities in soils by chelation. Geoderma, 68, pp. 123-133, 1995. McBride, M.B. Reactions Controlling Heavy Metal Solubility in Soils. In AdvanCed in Soil Science, Volume 10, B.A. Stewart, Ed., Springer-Verlag, New York, 1989, pp. 1-56. Mehra, a.p. and M.L. Jackson. Iron oxide removal from soils and clays by a dithionite:citrate system buffered with sodium bicarbonate. ClayJ and Clay Minerau, 7, pp. 317-327, 1960. Meiwes, K-J., N. Konig, P.K Khanna, J. Prenzel, and B. Ulrich. Chemische Untersuchungsverfahren fur Mineralboden, Auflagehumus und Wurzeln zur Charakterisierung und Bewertung der Versauerung in Waldboden. In Berichte deJ ForJchungJzentrumJ Wa[Jo"lcOJYJtemeIWauJJterben, 7, Gottingen, 1984, pp. 1-67. a'Day, P.A., G.E. Brown, Jr., and G.A. Parks. X-ray absorption spectroscopy of cobalt(II) multinuclear surface complexes and surface precipitates on kaolinite. J. CoLlolJ Inter/ace Sci, 165, pp. 269-289, 1994a. a'Day, P.A., G.A. Parks, and G.E. Brown, Jr. Molecular structure and binding sites of cobalt(II) surface complexes on kaolinite from X-ray absorption spectroscopy. ClayJ and Clay Minerau, 42, pp. 337-355, 1994b. Ponizovskiy, A.A., L.P. Korsunskaya, T.A. Polubesov, a.A. Salimgareyeva, Ye.S. Aleksane, and Ya.A. Pachepskiy. Methods for determining the specific surface area of soil from water-vapor adsorption. EuraJian SoiL Sci, 25, pp. 12-29, 1993. Posner, A.M. and J.W. Bowden. Adsorption isotherms: Should they be split? J. SoiL Sci., 31, pp. 1-10, 1980. Puri, B.R. and K Murari. Studies in surface area measurements of soils: 2. Surface area from a single point on the water isotherm. SoiL Sci., 97, pp. 341-343, 1964. Russell, E.W. RMJeLL'.J SoiL ConditionJ and Plant Growth. 11th ed., A. Wild, Ed., Longman Scientific & Technical, Essex, 1988. Sadiq, M. and T.H. Zaidi. The adsorption characteristics of soils and removal of cadmium and nickel from wastewaters. Water Air Soil PolLut. 16, pp. 293-299, 1981. Sadiq, M. and C.G. Enfield. Solid phase formation and solution chemistry of nickel in soils: l. Theoretical. Soil Sci, 138, pp. 262-270, 1984a. Sadiq, M. and C.G. Enfield. Solid phase formation and solution chemistry of nickel in soils: 2. Experimental. SoiL Sci, 138, pp. 335-340, 1984b. Scheidegger, A.M. and D.L. Sparks. A critical assessment of sorption-desorption mechanisms at the soil mineral/water interface. SoiL Sci, 161, pp. 813-831, 1996a. Scheidegger, A.M. and D.L. Sparks. Kinetics of the formation and the dissolution of nickel surface precipitates on pyrophyllite. Chem. GeoL., 132, pp. 157-164, 1996b. Scheidegger, A.M., G.M. Lambie, and D.L. Sparks. Investigation ofNi sorption on pyrophyllite: An XAFS study. Environ. Sci. Techno!., 30, pp. 548-554, 1996a. Scheidegger, A.M., M. Fendorf, and D.L. Sparks. Mechanisms of nickel sorption on pyrophyllite: Macroscopic and microscopic approaches. SoiL Sci. Soc. Am. J., 60, pp. 1763-1772, 1996b. Schulte, A. AdJorption von SchwermetaLlen in repriMentativen Boden IJraeu und NordwedtdeutJchlandJ in Abhangiglceit von derJpezifiJchen Ober/liiche. Berichte des Forschungszentrums Wald6kosysteme, Reihe A, 46, Gottingen, 1988. Schulte, A. Adsorption density, mobility and limit values of Cd, Zn, Cu and Pb in acid forest soils. Archil' fur Aclcer- und Pjlanzenbau und Bodenlcunde, 38, pp. 71-81, 1994a. Schulte, A. The Adsorption of Cd, Zn, Cu and Pb in Acid Forest Soils. In BiogeochemiJtry of Trace ElementJ, D.C. Adriano et aI., Eds., Science and Technology Letters, Northwood, 1994b, pp. 525-535.
58
Fate and Transport of Heavy Metals in the Vadose Zone
Schulte, A. and F. Beese. Adsorptionsdichte-Isothermen von Schwermetallen und ihre okologische Bedeutung. Zeitdchrijt fur Pflanzenerniihrung und BOdenkunde, 157, pp. 295-303, 1994a. Schulte, A. and F. Beese. Isotherms of cadmium sorption density. J. Environ. QuaL., 23, pp. 712718, 1994b. Schulte, A. and J. Gehrmann. Entwicklung der Niederschlags-Deposition von Schwermetallen in Westdeutschland. 2. Arsen, Chrom, Kobalt und Nickel. Zeitdchrijt fur Pjlanzenerniihrung und Bodenkunde, 159, pp. 385-389, 1996. Schulte, A., A. Balazs, J. Block, and J. Gehrmann. Entwicklung der Niederschlags-Deposition von Schwermetallen in Westdeutschland. 1. Blei und Cadmium. Zeitdchrijt fur Pjlanzenerniihrung und Bodenkunde, 159, pp. 377-383, 1996. Slejko, F.L. Addorption Technology. Marcel Dekker, New York, 1985. Sparks, D.L. EnvironmentaL Soil ChemiJlry. Academic Press, San Diego, 1995. Sposito, G. On the use of the Langmuir equation in the interpretation of" adsorption" phenomena: II. The "two-surface" Langmuir equation. Soil Sci. Soc. Am. J., 46, pp. 1147-1152, 1982. Sposito, G. The Surface Chemi.Jtry of SoiU. Oxford University Press, New York, 1984. Sposito, G. The Chemi.Jtry of Soiu. Oxford University Press, New York, 1989. Stern, O. Zur Theorie der elektrolytischen Doppelschicht. Zeitdchrijt fur Elektrochemie und Angewandte PhYdikaLi.Jche Chemie, 30, pp. 508-516, 1924. Travis, C.C. and E.L. Etnier. A survey of sorption relationships for reactive solutes in soil. J. Environ. QuaL., 10, pp. 8-17, 1981. Tu, C. Distribution and transformation of native and added Ni fractions in purple soils from Sichuan Province. Pedodphere, 6, pp. 183-192, 1996. U ren, N. C. Forms, reactions, and availability of nickel in soils. Adv. Agron., 48, pp. 141-203, 1992. van Bemmelen, J .M. Die Absorptionsverbindungen und das Absorptionsvermogen der Ackererde. Die Landwirtdcha/tLichen Verdllchd-Stationen, 35, pp. 69-136, 1888. Veith, J.A. and G. Sposito. On the use of the Langmuir equation in the interpretation of "adsorption" phenomena. SoiL Sci. Soc. Am. J., 41, pp. 697-702, 1977. Zehetner, F. Addorption IdothermJ of NickeL in Acid Fore.Jt Soiu. Diplomarbeit, Institut fur Bodenforschung, Universitat fur Bodenkultur, Wien, 1997. Zeien, H. and G.W. Brummer. Chemische Extraktionen zur Bestimmung von Schwermetallbindungsformen in Boden. Mitteilungen der Deutdchen BOdenkundLichen Ge.JeLucha/t, 59, pp. 505509, 1989.
CHAPTER ;)
Sorption-Desorption Equilibria and Dynamics of Cadmium During Transport in Soil R.S. Kookana, R. Naidu, D.A. Barry, Y.T. Tran, and K. Bajracharya
INTRODUCTION Cadmium (Cd) is a soil contaminant that is toxic to plants and animals. It is a major human health hazard due to its potential accumulation in kidneys leading to kidney dysfunction (Alloway, 1990). Both natural and anthropogenic activities are important sources of Cd to the soil environment. Sources include atmospheric fallout from metallurgical industries, application of sewage sludge or other waste containing Cd on land, and use of phosphatic fertilizers in agricultural areas. The result is substantial Cd accumulation in soil, providing the source for plant uptake, as well as leaching of Cd under certain soil and environmental conditions (Alloway, 1990; Naidu et al., 1997). The mean concentrations of Cd in agricultural soil are generally less than 0.4 mg/kg (Alloway, 1990), with a maximum up to 3 mg/kg. Highly contaminated soils can, however, contain Cd concentrations up to 10 mg/kg or even more (e.g., Hornburg and Brummer, 1993). Elevated concentrations of Cd in soil can also originate from geologic sources. For example, soils developed from black shales can contain Cd concentrations> 20 mg/kg (Alloway, 1990). In soil solutions, however, Cd concentrations rarely exceed 10 /J-g/L under agricultural conditions. McLaughlin et al. (1997) reported that in solutions extracted at moisture contents equivalent to -5 kPa from 50 Australian agricultural (saline/sodic) soils, the total concentration of Cd ranged from 2.7 to 222.4 nM (0.3024.9 /J-g/L). Natural Cd concentrations up to 800 /J-g/L (far exceeding USEPA drinking water quality standard -10 /J-g/L) have been observed in filtered samples of stream waters and shallow groundwaters associated with unmined ore deposits in some areas (Runnells et aI., 1992). Among anthropogenic activities, the disposal of sewage sludge and/or metal rich industrial wastes on soils can lead to unacceptably high concentrations of Cd in both soils and associated water resources. 59
60
Fate and Transport of Heavy Metals in the Vadose Zone
The fate of Cd from anthropogenic sources is largely determined by its retention and mobility in soils (Naidu et al., 1997). The chemodynamics of Cd in the soil environment are influenced by its interactions with both the solid and aqueous phases of soil. In addition, the soil solution and solid phase concentrations of Cd can have significant bearing on the nature of the interactions. Leaching of Cd through soil profiles has implications for both Cd buildup in the root zone and the potential for contamination of surface and groundwater associated with soils showing low sorption capacity. Some soils can have an inherently low sorption capacity for Cd due to either their low surface negative charge density or their sandy nature and high permeability. Cd mobility in such soils can be rather high (Boekhold and Van der Zee, 1991; Kookana et aI., 1994). Groundwater monitoring studies (Lieber et aI., 1964) as well as simulations of long-term leaching behavior of Cd (Boekhold and Van der Zee, 1991) indicate that Cd leaching in some instances, e.g., in sandy soils receiving high input (> 50 glha.yr), can potentially exceed the acceptable levels in groundwater. The objective of this report is to provide an overview of important processes which largely determine the fate of Cd in soil environment. The processes covered here include sorption, desorption, precipitation, complexation - all factors affecting transport or mobility. The influences of soil solid as well as the solution phase compositions, and of the soil physical conditions, e.g., static and flow conditions, on sorption, desorption, and mobility of Cd in soils are discussed.
PROCESSES GOVERNING FATE OF CADMIUM IN THE SOil PROFilE The processes of Cd retention in a soil include adsorption/desorption, ion exchange, and precipitation/dissolution reactions, their occurrence depending on the nature of the solid phase, the concentration of Cd present in soil solution, and soil solution composition. Cd mobility and transport through the soil profile are directly linked to retention and release processes. Precipitation and dissolution reactions of metals including Cd may involve discrete solid phases or solid phases adsorbed to soil component surfaces and usually occur either in the presence of specific anions or at high concentration of metal ions (Naidu et aI., 1997). While ion exchange causes an exchange between ionic species in solution phase and those held by the soil surface, adsorption reactions can involve ionic as well as molecular species of metals (Amacher et al., 1986). Since it is often not possible to make a clear distinction between various retention processes of metals in soils, the term sorption is more appropriate than adsorption. These key processes (sorption, desorption, precipitation, complexation, and transport) regulate the mobility and availability of Cd in soil solution (Naidu et al., 1997), and are discussed in the following sections, particularly in relation to the soil and solution phase composition.
Sorption Sorption is one of the most important processes that govern the bioavailability and mobility of Cd in soil. Sorption of metals has generally been related to metal ion hydrolysis (Hodgson et aI., 1964; Forbes et al., 1976) and therefore to their pKa values. On the basis of pKa values (in parentheses), metal sorption is expected to decrease in the following sequence: Hg (3.4) » Pd (7.7) > eu (7.7) » Zn (9.0) > Co (9.7) > Ni (9.9) > Cd (l0.1). Based on studies carried out on soil clay isolates from a range of soils, Tiller et al.
Sorption-Desorption Equilibria and Dynamics of Cadmium During Transport in Soil
61
(l984a) concluded that both specific and nonspecific sorption mechanisms contribute to Cd retention in soils. Cd can form inner-sphere complexes with amorphous Fe and Ai oxyhydroxides (Hayes and Leckie, 1987). Silicate minerals exhibit a range of affinities for Cd (Garcia-Miragaya and Page, 1978; Tiller et al., 1979) and the nature of sorption mechanisms also varies with the type of mineral. As Zachara and Smith (1994) stated, Cd can form at least two types of complexes with layer silicates, including outer-sphere ion exchange complexes on the basal plane, and coordination complexes with SiOH or AlOH groups exposed at the crystalline edges. However, they found that on soil smectite edge complexation reactions were more important. Sorption of Cd on kaolinite is mainly ascribed to ion exchange and possibly some inner-sphere complexation at pH > 6.5 (Schindler et al., 1987). While Cd sorption through metal ion hydrolysis and specific adsorption in the presence of different electrolytes has been documented (e.g., Tiller et al., 1979; Brummer et al., 1988; Tiller 1989), the nature of mechanisms operative in the presence of specifically adsorbed ligand ions such as sulfate, phosphate, and organic ligands remains unclear (Naidu et al., 1997). Sorption of Cd on amorphous metallic oxides has been postulated to be a continuum between surface reactions and precipitation (Farley et al., 1985). Farley et al. (1985) extended the well-established surface complexation approach (coordination reactions between solute and functional groups on the surface) through a "surface precipitation" model which is based on a series of adsorption followed by precipitation reactions on the oxide surfaces. The precipitation on the solid phase was described by the formation of a solid-solution. The model satisfactorily described the sorption behavior of Cd on iron hydroxide surfaces (Farley et al., 1985; Dzombak and Morel, 1986). Silicate clay minerals, amorphous oxides of Fe and Mn, and particulate organic C present in soil all influence the sorption potential of metals in soils. However, the relative contribution of these soil fractions to Cd sorption varies depending on soil type and its solution composition. For example, in the presence of Ca concentrations high enough to suppress ion exchange mechanisms, layer silicates have a relatively lesser role than organic carbon and Fe and Mn oxides, whereas layer silicates can contribute significantly to Cd sorption on Na-saturated clay-sized isolates from soils (e.g., Zachara et al., 1992). Obviously the contribution will vary with the type of layer silicates. Often, layer silicates act as substrates on which amorphous Fe and Al oxides precipitate. Therefore, fixedcharge sites on soil clays with a potential to bind Cd can be blocked by oxides (Zachara et al., 1992). The presence of mineral-bound organic material can enhance Cd sorption (Haas and Horowitz, 1986) or have no effect (Davis, 1984; Harter and Naidu, 1995). Not only the nature of the sorbent but also a range of other soil properties and factors such as soil pH, ionic strength of the soil solution, the presence of inorganic and organic ligands, competing ions, etc., affect Cd sorption. The influence of some of these factors is discussed in the following sections.
Factors Affecting Cd Sorption il1 Soils Cd interactions with the soil solid phase include ion adsorption at surface sites, ion exchange with clay minerals, binding by organically coated particulate matter or organic colloidal material, or adsorption of metal ligand complexes (Naidu et al., 1997). All of these interactions are influenced by soil solution composition and characteristics such as
62
Fate and Transport of Heavy Metals in the Vadose Zone
pH and ionic strength, nature of the metal species, dominant cation, and ligands (inorganic and organic) present in the soil solution.
pH A large body of literature based on both pure mineral systems and soils has established that soil solution pH has the most critical influence on sorption of Cd (e.g., Tiller et al., 1984b; Haas and Horowitz, 1986; Briimmer et al., 1988; Biirgisser et al., 1991; Gerth et al., 1993; Naidu et al., 1994a; Naidu et al., 1997). These studies show that over a narrow pH range, the sorption of Cd increases very rapidly leading to a so-called adsorption edge (Figure 3.1). Recently, Tran et al. (1998a) performed Cd sorption experiments at a range of pH values in a sandy soil. They found that for every increase of 0.5 unit of pH for the range of 5.5 to 6.5, twice as much sorption of Cd was observed. Such increases in sorption presumably occur because of the rapid increase in the concentration of the metal-hydroxy species, believed to be the active component adsorbed by soils (Hodgson et al., 1964; Davis and Leckie, 1978; Tiller et al., 1979). However, speciation calculations show that the concentration of CdOH+ is negligible relative to Cd2+ at the pH of the adsorption edge (Naidu et al., 1994a). Therefore, for the metalhydroxy species to be adsorbed, it must have a very high affinity for the soil surface which will then drive the metal hydrolysis reaction (Eqs. 1 and 2) to the right, maintaining Le Chatelier's principle of equilibrium (Naidu et al., 1994a): (1) MOH+
+
Soil
¢:::}
Soil - MOH
(2)
On surfaces exhibiting variable charge, in addition to the metal species, the surface charge density of the adsorbent is strongly influenced by pH. Increasing solution pH leads to a rapid increase in net negative surface charge which may explain the enhanced affinity for metal ions in such systems, e.g., Fe and Al oxides, organic matter (Garcia-Miragaya and Page, 1978; Naidu et al., 1997).
Ionic Strength Soil properties such as pH, charge density distribution, thickness of the diffuse double layer, and the activity of Cd present in solution are all influenced by the composition of soil solution (Harter and Naidu, 1995). Several studies (e.g., Homann and Zasoski, 1987; Boekhold et al., 1993; Zachara et al., 1993; Naidu et al., 1994a,b) have demonstrated the importance of soil solution composition on the nature and extent of Cd sorption by soils and their constituents. Increasing ionic strength (1) generally reduces metal sorption due to its influence on both sorbate and sorbent properties (Naidu et al., 1997). Depending on the nature of the surface properties of the sorbent, the effect of ionic strength on the sorption of Cd can be marked. On minerals with permanent charge density, a substantial reduction in Cd sorption with increasing ionic strength of NaCI04, NaCl, or Na2S04 was noted by several workers (e.g., Garcia-Mirgaya and Page, 1976, 1977; Zhu and Alva, 1993). However, when Cd forms inner-sphere complexes with sorbents (e.g., with amorphous Fe and Al
Sorption-Desorption Equilibria and Dynamics of Cadmium During Transport in Soil ---~,-'""'------""'.,,----~-
---
----
63
"--------
--,----""'"
Xeralf
-
100
100
~
~ "C Q)
.0 '-
0
en "C
U
0
c:: 0
t
0
61lmal 151lmal 30 Ilmal 150llmal
Cl.
0
'-
a.. 20
0
2
4
L'10 L'l ... L"1 I:J. L'l •
6
1.51lmal L'l ... 3 Ilmal r 1 I:J. 61lmal L'l •
8
Oxisol
0
2
4
6
8
Oxisol
0
c::
0
Maianda
t
0
0.75 1lmal 1.5llmal 31lmol 61lmal
Cl.
0 '-
a.. 20
0
2
4
6
L'l ... L'l I:J. C1 • L'10
4
6
8
Figure 3.1. Effect of pH on sorption of Cd in four soils at different initial concentrations (Naidu et aI., 1994, with permission).
oxyhydroxide, smectite), sorption is little affected by ionic strength (Hayes and Leckie, 1987). For example, Naidu et al. (1994a) observed that doubling of I did not influence Cd sorption in a smectite-dominated Vertisol, whereas this increase in I caused a 50-fold drop in Cd sorption on an Inceptisol (Figure 3.2), On variable charge surfaces, the effect of ionic strength depends on pH, In an Oxisol, Cd sorption was found to increase with ionic strength at pH values below the point of zero net charge (PZNC) but the reverse was observed at pH values above the PZNC (Naidu et al., 1994a). The effect of ionic strength on metal sorption via its effect on electrostatic potential in the plane of adsorption is likely to be most marked in variable charge soils (Barrow, 1987). Complexation with Ligands Complexation of heavy metal cations with a variety of ligands in soils (both inorganic and organic) has been long been recognized (Harter and Naidu, 1995), Indeed, several
64
Fate and Transport of Heavy Metals in the Vadose Zone 4.4.. '""=+==-=--~='fl Xeralf
4.0
4.0
~
'0 en
b
3.6
~ H20
(5
0.01 0.030.150.300.75 1.5
3.6 ~""""'----''--'''''''''---I._-L---'--I H20 0.01 0.030.150.300.75 1.5
E
.6
Oxisol
c::
5.04
E
.03
.01 O~~~~--~~~~
H20 0.01 0.030.150.300.75 1.5
H20 0.010.030.150.300.75 1.5
Ionic strength NaN0 3 (mol L-1) Figure 3.2. Effect of ionic strength on Cd sorption in six soils (Naidu et aI., 1994, with permission).
computer models including GEOCHEM-PC (Parker et al., 1995) and MINTEQA2/ PRODEFA2 (Allison and Brown, 1995), which are recent improvements on previous versions of these models, are available to calculate various ionic speciations of the metals in soil solution. Model simulation as well as experimental results show that complex formation can significantly affect the activity of Cd in soil solution and therefore its sorption and leaching potential through a soil profile (Garcia-Miragaya and Page, 1976; Naidu et al., 1994b; Temminghoff et al., 1995: Bolton et al., 1996). Cd+ 2 makes stable complexes with CI-, SO/-; principally CdCI+ and CdS04, respectively. However, model calculations show that at the same concentration of Cd 2 +, Cl-, and SO/-, the fraction complexed with CI- is higher than that with Chlorocomplexes of Cd can significantly reduce the sorption of Cd in soils, depending on chloride concentration in soil solution (Boekhold et al., 1993). Formation of metal-ligand complexes can have a variable impact on Cd sorption. It has been observed, for example by Homann and Zasoski (1987), that the presence ofeland S04 2- had either no effect or reduced the Cd sorption on some forest soils. In calcareous soils with pH > 7.5 (Typic Torrifluvents, Ustollic Caciorthids, and Petro calcic Paleustolls), O'Connor et al. (1984) observed that Cd sorption decreased in the presence of CaCl2 but increased in the presence of CaS04' the latter being the result of a decreased Ca activity and reduced competition with Cd for sorption. Cd-ligand com-
sol-.
Sorption-Desorption Equilibria and Dynamics of Cadmium During Transport in Soil
65
plexes can have important consequences in terms of Cd bioavailability and food quality. Studies on the effect of chloride salinity on Cd bioavailability in soil solution and plant uptake (Bingham et aI., 1986; McLaughlin et aI., 1994) show that Cd availability is enhanced in the presence of CI ions in soil solution. Organic ligands through their complexation with Cd can either enhance or reduce Cd sorption in soil, depending on the functional groups of the organic molecules and the charge characteristics of the soil (Harter and Naidu, 1995). For example, Elliott and Denneny (1982) studied sorption of Cd in the presence of acetate, oxalate, nitrilotriacetate, and EDTA in three soils. They observed that in two of the three soils, oxalate increased the sorption of Cd at pH values below 5.0, whereas the other ligands caused a reduction in sorption. Neal and Sposito (1986) observed that, in soils with permanent charge, Cd sorption was reduced by the presence of organic matter from sewage sludge. Bolton et al. (1996) studied the Cd complexation properties of a humic acid fraction extracted from a soil and concluded that it forms relatively strong complexes with soil humic acid at pH values relevant to natural environments. Low molecular weight organic acids such as acetic, citric, fumaric, oxalic, and succinic acids have been shown to enhance the release behavior of Cd present in soil and, therefore, its bioavailability (Krishnamurti et al., 1997). Complexation of organic and inorganic ligands can result in enhanced transport of Cd through the soil profile, as discussed in a later section.
Presence Of Other Metals and Cations and Competition for Sorption Sites Competition by other metal ions (Cu, Ni, Pb, Cr, etc.) has been shown to cause a reduction in Cd sorption by several workers (e.g., Garcia-Mirgaya and Page, 1976; Homann and Zasoski, 1987; Christensen, 1987). Christensen (1987), while studying Cd sorption on 12 Danish soils, found that in the presence of mixtures of heavy metals such as Ni, Co, and Zn, the sorption coefficient for Cd was up to 14 times lower than those observed for Cd alone. The reduction was ascribed mostly to Zn due to its relatively higher concentration than other competing metals. Other cations such as Ca, Mg, and Na in soil solution can also compete with Cd for sorption sites. However, Ca has been reported to be a much stronger competitor than Na for Cd sorption on soils and pure systems (Christensen, 1984a; Zhu and Alva, 1993; Zachara et aI., 1993; Boekhold et aI., 1993; Naidu et al., 1994a; Kookana and Naidu, 1998). Boekhold et aI., (1993) reported that due to competition between Cd and Ca, sorption of Cd in soil was reduced by 80% in the Ca-electrolytes as compared to the Na-electrolytes. Similarly, Naidu et aI. (1994a) found that, in Australian soils, even when the ionic strengths were kept constant, Ca2 + caused much greater reduction in Cd sorption than Na+. Such reduction in sorption can have a major influence on Cd mobility and transport behavior in soils, as discussed below.
Effect Of pH and Soil Solution Composition on Cd Transport As mentioned already, soil pH, through its effect on both surface charge density and the formation of hydroxy metal species, exerts a strong influence on Cd sorption and hence transport in soils. The influence of pH may, however, vary with soil type. In soils with high surface charge densities, limited movement of heavy metals has been observed at pH values above 6.0. For example, in a Typic Hapludoll from Minnesota (pH = 6.4),
66
Fate and Transport of Heavy Metals in the Vadose Zone
Cd did not leach beyond 0.4 m depth over a period of three years even after the incorporation of sludge in the top 0.2 m at a loading of 25 kg Cd ha-1 (Dowdy and Yolk, 1983). In contrast, these authors noted that in a Typic Hapludult at pH 5, despite the high OM content (10.6%) in the topsoil, 4-7% of the applied Cd leached beyond a depth of 1.2 m. On the basis of a field study, Streck and Richter (1997a) reported that following 29 yr of wastewater (Cd concentration ranging from 2-40 /lg/L during 1980-1990) application, only about 5% of Cd and Zn were found below 0.7 and 0.9 m depths, respectively, in a soil (current pH = 5.2-5.4 in 1:2.5, 0.01 M CaCI 2). In flow-through experiments, Bajracharya et al. (1996) found that sorption was reduced drastically when the pH is reduced from 6 to 4.3. This resulted in an early breakthrough of Cd and a reduction in the value of the partition or Freundlich coefficients by an order of magnitude, consistent with Al-Soufi (1994). A reduction in Cd sorption due to increase in ionic strength or the presence of competing species is expected to influence its transport through soils. Recently, Kookana and Naidu (1998) studied the Cd transport behavior in laboratory columns of an Oxisol and an Alfisol, as influenced by varying ionic strengths and index cations. They observed that when the concentration of NaN0 3 was increased in the background solutions from 0.03 M to 0.30 M, the breakthrough of Cd through the Oxisol soil column occurred 3-4 times faster (Figure 3.3A). While for breakthrough only approximately 30 pore volumes were needed at 0.30 M NaN0 3, about 110 pore volumes were needed in the presence of 0.03 M NaN0 3 • At the same ionic strength, Cd eluted much earlier in the presence of 0.01 M Ca(N03)2 than in the presence of 0.03 M NaN0 3 • An increase in Ca(N03)2 concentration, however, had relatively less impact than NaN0 3, mainly due to the significant competition for sorption sites in the presence of even relatively low (0.01 M) concentration of Ca(N03h. In contrast, even at a very high ionic strength of NaN03 (0.075 M), the Cd breakthrough in the Alfisol took more than 100 pore volumes, whereas only 10 pore volumes were needed in the presence of 0.05 M Ca[N03h (Figure 3.3B). An increase in ionic strength of Ca[N03h from 0.05 M to 0.25 M enhanced the Cd transport by a factor of 2 in the Alfisol. For both soils, Cd movement at constant ionic strength was an order of magnitude faster in the presence of Ca(N03)2 solution compared to NaN0 3 • The presence of inorganic and organic ligands in soil solution can markedly affect the mobility of Cd through soil. Chloride forms stable complexes with Cd and given the natural abundance of CI- in soils, it is particularly significant to consider the role of CI- in Cd mobility. Doner (1978) studied the mobility of Cd, Cu, and Ni through a Typic Xerorthent with pHsat of 6.6, demonstrating that in the presence of 0.5 M NaCI solution, Cd moved up to 4 times more rapidly than in the presence of NaCI0 4- solutions. It is worth noting that very high Cd concentrations (10 IlglmL) were used in this study. The role of chlorocomplexes in enhancement of Cd mobility was similarly observed by Kookana et al. (unpublished data) during miscible displacement studies on Oxisol soil columns. These experiments were carried out under conditions similar to those in Kookana et al. (1994) but at a higher effluent pH. They found that the Cd BTC (breakthrough curve) in the presence of O.OlM CaCl2 was twice as fast as in O.OlM Ca [N0 3]2 (Figure 3.4). This was because Cd forms complexes with CI- but not Nol-. This is consistent with calculations which show that at I = 0.03, in 0.02 M NaCI solutions, only 48% of dissolved Cd is present as Cd2 + (Boekhold et aI., 1993).
,
Sorption-Desorption Equilibria and Dynamics of Cadmium During Transport in Soil 1
A
0.9
;
0.8
~
0.7
/ -
~
0
0.5 0.4 0.3 ~
0.2
,
I
1-
0.6
~
./...,.,.,:
./
Oxisol
~ 0.01
M Ca[N0312 [J 0.03 M NaN03 .0.3 M NaN03
0.1
I
67
I
o~"""memmmDmmmmEDmma.. . . .~--------~ 150 100 50 o
,...,
1
B•
I
0.9
0.8
-
-·----·---_~-I
Alfisol
,~
o
~
o
B
o
0.6 0
~PQcAII1'
o
•
0.7
0 0
----_._---
00.75 M NaN0 3
o 0.4
• 0.25 M Ca(N0 3)2 00.05 M Ca(N0 3)2
0.3
•
0.2
I I
0
o o o
•
0.5
I I
[J
o [J
o
o
a
[J
0.1 0 0
20
40
60
80
100
120
140
160
Pore Volumes
Figure 3.3. Ionic strength and index cation effect on the mobility of Cd in an (A) Oxisol and (B) Alfisol (after Kookana and Naidu, 1998, with permission).
Characterization of the Combined Effects of Ionic Strength, pH, and Cd Concentration The variation in ionic strength often causes a change in soil solution pH and, therefore, the results are often a reflection of combined effects of pH and ionic strength. However, some workers in recent years have tried to quantify the individual contributions of pH and ionic strength. To discern the effect of ionic strength, pH, and [Ca], a simple theoretical relationship was developed by Temminghoff et al. (1995). The rela-
68
Fate and Transport of Heavy Metals in the Vadose Zone
0.9 0.8
• 0.01 M CaCI 2 o 0.01 M Ca[N0312
0.7
•
0.6
•
•
•• •
••••• o
~
0.2 0.1
o
•
•
0
o
u 0.4 0.3
o
o
•
o 0.5
o o
o o
o +---~~~~~Yr=-~~b-=----r------+------+------1-----~ 80 70 30 50 60 10 20 40 o Pore Volumes
Figure 3.4. Cd breakthrough curves as influenced by complexation with chloride in an Oxisol [data for 0.01 M Ca(N0 3h from Kookana et aI., 1994, with permission). Flow rate 13.5 cm/h (Cn and 12.3 cm/h (N0 3-); pH 4.8-4.9 (Cn, 4.9-5.0 (N0 3-). Other experimental conditions were same as described by Kookana et al. (1994).
tionship describes the pH-dependency of Cd ion binding and also the effect of competition by Ca. Since [Ca2+] » [Cd2+] in most cases, this simple relation can be expressed as:
(3) where ~ is the adsorbed quantity of Cd, K' is a modified Freundlich sorption coeffi. ned an d nCa are Cd - an d C a-speCl'filC nonl'd eal'lty parameters, an d m IS . a parameter Clent, based on ned and nCa representing the relative replacement ratio of H+ by Cd 2+ (for details see Temminghoff et aI., 1995). In Equation 3, activities instead of concentrations have to be used. According to Equation 3, the Cd sorption data corresponding to different pH, [Cd], and [Ca] obtained from batch and transport experiments (Kookana and Naidu, 1998) follows a straight line, as shown in Figure 3.5. In Equation 3, the exponents for pH and Ca are indicative of the sensitivity of Cd sorption to these two parameters. These exponents have been reported to vary with soil type. For example, for a clayey soil (Oxisol) the value of the exponent m was found to be -1.3 by Naidu et al. (l994a), in contrast to a value of -0.77 for a sandy soil reported by Boekhold et aI. (1993). The marked differences in m for these soils may signifY the different nature of the soil particle surfaces and electrolytes used during sorption studies. Similarly, the sensitivity of sorption to [Ca2 +] has also been found to vary between soils. Kookana and Naidu (1998) found the nCa values for an Oxisol to be -0.21 and -0.61 for an AlfisoI. Temminghoff et aI. (1995) reported a value of nCa of -0.34 for a sandy soil, whereas Chardon (1984) reported an average value of -OA1± 0.07 for six different soil
Sorption-Desorption Equilibria and Dynamics of Cadmium During Transport in Soil
69
0.04 - . - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - ,
•
0.03
•
OJ
.:.: (5
E
E
;-0.02
•
QI
.c ....
o
UI
"C
o
0.01
'"•••'"•
0 0
'" •
'"
0.0005
•
0.001
0.0015
0.002
0.0025
Figure 3.5. Cd sorption as a function of [WI, [Cd 2 +[ concentration and [Ca 2 +) activities in soil solution. column data, .A. • batch data (Kookana and Naidu, 1998, with permission).
types. The low nCa for the Oxisol may be due to its low inherent affinity for Cd. In this soil the pH is close to the PZC (point of zero charge), and so the effect of ionic strength on Cd sorption is much less pronounced.
Precipitation The role of precipitation and dissolution reactions in determining solubilities of trace elements in soil solution is generally only important in cases where either the concentration of metal is very high or when the conditions in soils are such that certain anions or ligands are present. For example, under reducing conditions sulfide salts of metals such as Cd, Zn, Hg, Fe, etc., may be important (Tiller, 1996). Interactions of Cd with surfaces of calcite-the most common carbonate in soilshave been described by various authors as chemisorbed complexes, surface precipitates, etc. (see Zachara et al., 1991). However, initial rapid sorption of Cd on calcite surfaces may be followed by a slower phase of dehydration and coprecipitation (e.g., Davis et al., 1987; Zachara et al., 1991). There is a possibility that following sorption, Cd may migrate into the solid phase, forming a solid solution, as demonstrated by Stipp et al' (1992). Stipp et al. (1992), on the basis of studies on Cd uptake by calcite using near-surface sensitive techniques, provided experimental evidence for the "formation of a thin, surficial, solid solution precipitate during initial contact with solution," which disappeared on the order of weeks due to the diffusion of the trace metal into the solid phase. They also verified by surface analysis, the precipitation of a nearly pure, crystalline otavite (CdC03)
70
Fate and Transport of Heavy Metals in the Vadose Zone
overlayer on calcite. Therefore it is possible that at elevated levels of Cd in the carbonate-rich systems, otavite can playa role in controlling concentrations of Cd in terrestrial and aquatic environments. However, as studies by Holm et al. (1996) showed, even in carbonate-rich aerobic soils the Cd concentration in solution was not governed by otavite, presumably due to the presence of inhibitors of precipitation such as dissolved organic matter in soil solution. The precipitation that follows the sorption reaction of Cd on calcite and the probable incorporation of the sorbate into calcite can result in its limited desorption (Zachara et aI., 1991). Therefore, in calcareous soils and groundwaters, calcite can act as an important sink for Cd and some other metals. The kinetics of calcite recrystallization (Zachara et aI., 1991) and Cd diffusion into the solid phase (Stipp et aI., 1992) may contribute to a nonequilibrium sorption behavior of Cd during transport in soil and groundwater.
Kinetics of Cd Sorption In well-mixed systems involving soils and other sorbents, the sorption of Cd as well as other ion exchange reactions has generally been found to be fast (Christensen, 1984a; Hayes and Leckie, 1987; Hachiya et al., 1979). With reaction half-lives on the order of minutes or less, they are often complete by the time solid and liquid phases can be separated (e.g., Zasoski and Burau, 1978; Harter and Lehmann, 1983; Jardine and Sparks, 1984). Some of these studies have been carried out at very high solution to soil ratios, and involve continuous agitation facilitating the accessibility of sorption sites to sorbent. The sorption reaction of Cd in soils has been found to be generally complete within hours. For example, Christensen (1984a), using low concentrations of cadmium, reported that more than 95% sorption in soils (pH 6.0 to 6.5, background salt 0.001 M CaCI 2) studied was complete within 10 minutes and the equilibrium was achieved within 1 hour with no further increase in sorption for 67 weeks (Figure 3.6). Chardon (1984) studied sorption of Cd after 23 hand 46 h on 12 soils [pH 3.3 to 7.6; background salt 0.0015 M Ca(N03hJ and reported that for one soil only a slightly higher sorption (but significant) was noted after 46 h. Kookana (1997; unpublished data) observed that cadmium sorption on four soils (Alfisols from South Australia, pH 5.5 to 7.3) was essentially complete within 3 h of shaking; more than 95% of maximum observed sorption occurred in the first 30 minutes. On synthetic minerals and natural soil sorbents, biphasic sorption reactions have been reported for Cd. On synthetic goethite, for example, Brummer et ai. (1988) noted a rapid initial sorption of the metals (Cd, Zn, and Ni), followed by a much slower reaction akin to a diffusion-controlled penetration into goethite. However, among the three metals studied, the smallest increase in the magnitude of sorption with time at any given pH was with Cd. A batch kinetic study on Cd sorption in five soils by Selim (1989) also showed a fast sorption reaction followed by a slower reaction, more pronounced in some soils than others. Sorption of trace elements showing a rapid initial sorption followed by a much slower reaction has been reported by other workers also (Gerth et aI., 1993). Similarly, Fuller and Davis (1987) noted that Cd sorption on a calcareous aquifer sand followed two reaction steps, the second slower step continuing for 7 days. In this case, the sorption might have been followed by a precipitation reaction.
Sorption-Desorption Equilibria and Dynamics of Cadmium During Transport in Soil
c;;E
71
8·~-----------------------------------, 66 ...
~ 50C)
t::.. sandy loam
::i.
-;- 40 o
r-
• loamy sand
~
cQ): c.:>
A
c:
A
A
o
"C
U
0
o
1
1
10 20
1 ..
1
-
1..>..
-
c.:>
1
30 ·60 180
..
I
A I..>.
1
1
1200 1400 1680
Contact time (minutes) Figure 3.6. Solution concentration of Cd as a function of time in a batch sorption experiment in two soils at low concentrations (after Christenson, 1984a, with permission).
The rate of sorption reaction can significantly influence the transport behavior of Cd through soil profiles or aquifer systems. The observation that in calcareous soils or aquifers the sorption may be followed by precipitation is particularly noteworthy, due to its potential effect on reaction kinetics and thereby on transport of Cd through the systems. Several studies on Cd transport are available in the literature (e.g., Dowdy et al. 1991; Dunnivant et aI., 1992; Boekhold and Van der Zee, 1992; Kookana et aI., 1994; Bajracharya et aI., 1996; Streck and Richter, 1997); however, most have been carried out in soil columns under laboratory conditions. While a detailed account of Cd transport is beyond the scope of this chapter, studies on sorption of Cd during transport have been discussed below.
Sorption Behavior of Cd During Transport Through Soil Columns
Batch versus Flow-Through Systems At a constant pH and for the same adsorbing media, it is often assumed that the amount of Cd sorbed by soil under batch and flow-through systems should remain the same. However, a comparison of data from the two systems has shown that sorption coefficients obtained by the two methods may not always be in good agreement (e.g., Boekhold and Van der Zee, 1992; Burgisser et aI., 1991; Grolimund et al., 1995; Bajracharya et al., 1996). Klamberg et al. (1989) found that the maximum sorption capacity for copper with humic acids was greater in column than in batch experiments. There may be several reasons for these differences (Grolimund et al., 1995), as described below.
Effect of Solid-Solution Ratio A major difference between batch and flow-through systems (particularly using packed soil columns) is the soil:solution ratio. Past research indicates that the solid-to-solution ratio affects both the rate and extent of sorption (Tan and Teo, 1987; Boesten and Van der Pas, 1988). However, different solid-to-solution ratios have been used in various Cd sorption studies published in the literature. Bajracharya et al. (1996) compared Cd sorption under batch and flow conditions and found that the batch-determined Cd partition
72
Fate and Transport of Heavy Metals in the Vadose Zone
coefficients (linear and Freundlich) were around 60-80% higher than that determined by any of the column experiments conducted at various flow rates. The solid-to-solution ratio in their column experiments was about 5 g: 1 mL, in contrast to a ratio of 1:50 in the batch experiments. Tan and Teo (1987) observed the influence of solid-to-solution ratio on equilibrium solid phase concentration and noted that for the same initial concentration, the equilibrium solid phase concentration was lower at a higher solid-to-solution ratio. Hence, the higher linear and Freundlich coefficient values observed in batch experiments by Bajracharya et al. (1996) could be due to the solid-to-solution ratio effect. This was also apparent in their batch experiments with different solid-to-liquid ratios (Figure 3.7). As shown in the figure, the partition coefficient decreases with increasing solid-to-liquid ratio. However, for the sand used in these experiments, as the ratio becomes smaller, very little effect is observed beyond the ratio of 0.5 g: 1 mL. The partition coefficient at a 1 g: 1 mL solid-to-liquid ratio (8.65), i.e., similar to that in a flow-through experiment, was less than the column-determined values (between 10 and 12), suggesting that the dynamic effect of flow could influence Cd sorption. Thus, even though the solid-to-solution ratio does affect the partition coefficient, it does not explain the difference in ~ values between batch and column experiments. Akratanakul et al. (1983) suggested that even when the soil-to-solution ratio is the same, the sorption in flow-through experiments should be greater than that in corresponding batch experiments. They reasoned that since batch systems are closed, the amount of ions in the liquid phase decreases with time as more and more ions are adsorbed onto soil particles. In flow systems, desorbed ions are carried away by the flowing solution, replacing desorbed ions with a higher concentration of Cd ions. This effectively exposes the soil particles to a continuous source of ions. They also suggested that the amount of ions that are exposed to soil particles in a batch system is equal to the concentration of the solution times the volume. In a dynamic system, the product of concentration, flow rate, and exposure time gives the amount of ions exposed to soil particles. This implies that there are more Cd ions available for sorption in the flow-through system and that greater exposure of Cd ions increases sorption of Cd to soil particles. A schematic representation of the two systems has been provided in Figure 3.8. In batch and column experiments on copper sorption in soils, Grolimund et al. (1995) observed a significant dependence on sorbed amount of the solid concentration. However, once the sorbate was prewashed before sorption was measured, this effect disappeared and the results from different techniques were in good agreement. Therefore, it was concluded that the particle concentration effect was due to incomplete removal of preadsorbed ions (e.g., AI, K, Mg) or the presence of complexing agents (e.g., dissolved organic matter). Often, column experiments involve preconditioning of soil with a background electrolyte solution before sorption is measured, which is usually not done in batch experiments. The disagreement between batch and column studies, therefore, can be attributed to the physicochemical differences resulting in a multicomponent effect on sorption equilibria.
Flow-Affected Sorption The amount of sorption is expected to be same at different flow velocities as long as sufficient time of contact with the sorbent is allowed. However, there are conflicting
Sorption-Desorption Equilibria and Dynamics of Cadmium During Transport in Soil
73
20
•
,-.,
~ ....;;l
•
•
'-'
..... s:: Q) ...... C,) ......
•
1t Q) 10
•
0
C,)
s:: 0
'p
.~
p..
0 0.0
0.5 Solid-to-liquid ratio
1.0
Figure 3.7. Effect of sOlid-to-liquid ratio on adsorption coefficients in batch reactors (Bajracharya et aI., 1996, with permission).
Solution
Sorbed
• t Sorbed Batch system
Time Solution Sorbed
§ ~
§ U
Flow-through system
Sorbed
.~
Solution Time
Figure 3.8. A schematic representation of concentration changes in sorbed and solution phases with time as an equilibrium between the two phases is approached for the batch and flowthrough systems.
reports in the literature on this phenomenon. For example, Akratanakul et al. (1983) conducted flow-through experiments at three different flow velocities (1.3, 2.3, and 2.7 cmlh) for Pope Ridge soil developed from volcanic deposits and found the rate of sorption of Cd increased with flow velocity. Kookana et al. (1994) conducted flow-through experiments on Cd at two different flow velocities for an Oxisol soil. From their two breakthrough curves, it is evident that the curve at the higher velocity emerged later, indicating greater sorption at the higher velocity. However, the higher pH at higher flow velocity, albeit by only 0.2 unit, would have contributed to the increased sorption. The linear sorption coefficients evaluated by fitting a two-site model to two other column experimental breakthrough curves in an Oxisol and a Spodosol were found to be larger than the corresponding batch-determined ones.
74
Fate and Transport of Heavy Metals in the Vadose Zone
There are other reports in which the pore water velocity has the opposite effect on solute sorption. For example, Bajracharya (1989) conducted Cd sorption experiments in river sand at flow velocities of 0.83 cmlh and 9.17 cmlh and found that at the higher velocity, sand exhibited a lower sorption of Cd. Similar velocity effects on sorption have also been reported for other less strongly adsorbed solutes (Miller et aI., 1989; Shimojima and Sharma, 1995). Retardation factors for tritium and bromide have been observed to depend strongly on pore water velocity, with the retardation factor decreasing with increasing velocity (Schulin et aI., 1987). There are, again, reports where batch- and flow-determined sorption parameters were within the same range. For Cd, Theis et al. (1988) reported that the total reactive surface site density from column studies agreed well with the values obtained from batch studies. Boekhold and Van der Zee (1992) reported that batch-determined Freundlich sorption coefficients of Cd for a soil adequately described the observed Cd breakthrough curves from column experiments. Similarly, Burgisser et al. (1991) in their experiments on Cd sorption on sand (particle size, 125 to 250 I.lm) found that the batch sorption data was in good agreement with the isotherm calculated from the BTCs, obtained at two different flow velocities. They showed that kinetic effects were absent in their experiments, as the BTCs were unaffected by flow velocities. The factors affecting heavy metal sorption are difficult to isolate from experiments conducted on natural soils. Characteristics of natural soils are very difficult to ascertain precisely. Soil heterogeneity makes replicate experiments subject to uncertainty. The effects of pore water velocity on Cd sorption are difficult to establish even in relatively homogenous porous media, as shown by Bajracharya et al. (1996). They carried out 13 flow-through Cd sorption experiments on a uniform graded silica sand, mostly at a constant pH of6, but using a wide velocity range (5-214 cmlh). The partition coefficient, Kd [L 3/M] , obtained by fitting solutions of the advection-diffusion equation to concentration breakthrough curves from their experiments, is plotted against pore water velocity in Figure 3.9. Even in this artificial homogeneous sorbing medium, no conclusive trend in ~ with pore water velocity could be established. From the above discussion it is unreasonable to draw any definite conclusion in terms of similarity in results from batch and the flow-through systems. Not only are the two systems very different from each other, the sorption of Cd is influenced by several physical and chemical factors in the two systems, as discussed above. However, it is clear that, in addition to the differences in factors such as soil solution composition and pH, it is important to establish that kinetic effects are absent under the experimental conditions employed in flow-through systems.
Evidence of Sorption Nonequilibrium During Cd Transport Through Soil Asymmetrical Breakthrough Curves The linear equilibrium sorption during transport of a solute normally results in symmetrical breakthrough curves. On the other hand, a nonequilibrium process is said to occur when a solute undergoes any time-dependent reaction in addition to the standard advection and dispersion transport processes. Early and asymmetric breakthrough curves are characteristic of nonequilibrium processes, and can be caused by various physical and chemical processes. On the other hand, asymmetrical BTCs can result from nonlin-
Sorption-Desorption Equilibria and Dynamics of Cadmium During Transport in Soil
-
14
"""' btl
~ .....l
....I::
'-"
75
13
Q)
.(3 1+= 4-<
Q)
12
•
0
u I::
... 11 0 .;:1
~
•
~
10
1
• •
• •
• • •
•
100 10 Pore water velocity (cmlh)
1000
Figure 3.9. Variation of linear Cd sorption coefficient with pore water velocity (Bajracharya et aI., 1996, with permission).
ear equilibrium sorption from the liquid to solid phase. In principle, a nonreactive chemical would exhibit no asymmetry in the breakthrough curve when passed through the same homogeneous soil in which a reactive chemical shows a time-dependent sorption reaction. Tailing in the breakthrough data indicating sorption nonequilibrium during Cd transport has been reported in miscible displacement studies carried out on soils and clay-humic acid mixtures by several authors (Morisawa and Inoue, 1985; Campbell et al., 1987; Selim et al., 1992; Boekhold and Van der Zee, 1992; Kookana et al., 1994).
Flow-Interruption as a Test for Sorption Nonequilibrium Another technique to check for the occurrence of nonequilibrium processes is the flow-interruption method. The presence of a measurable depression in the rising limb of the BTC immediately after the flow is resumed signifies a nonequilibrium process method (Murali and Aylmore, 1980; Brusseau et al., 1989; Kookana et al., 1994). The purpose of stopping the flow is to enable sufficient time for solute to be adsorbed onto the sorption sites. Using the flow-interruption technique, Kookana et al. (1994) noted that nonequilibrium conditions existed during their flow experiments involving Cd. Recently, Tran et al. (1998a) further investigated Cd sorption nonequilibria through a series of flow-interruption experiments. In these experiments, a steady flow in homogeneous sand was established, after which Cd was introduced at constant pH. A flow-interruption period of 24 hr was applied once a complete breakthrough curve had been obtained; i.e., when the effluent concentrations were close to the influent concentration. Immediately after this no-flow period, the effluent solution was monitored for both Cd concentration and pH. As already noted, drop in effluent concentration immediately after the resumption of flow indicates nonequilibrium behavior. This behavior, however, was not observed. Figure 3.10 shows a plot reported by Tran et al. (1998a) in which an increase in effluent concentration was observed immediately after resumption of flow. Clearly this increase of effluent concentration was the result of desorption of Cd from the solid phase during
76
Fate and Transport of Heavy Metals in the Vadose Zone
•,,,,
1.5
,,
",...
•
• Observed data MCMFIT fitted - - - Two-site model simulated
u 0.5
0.0 . ._"'-----l'-----"_--'-_......._ - ' - _........_""---_"'----' o 20 40 60 80 100
Time (h) Figure 3.10. Cd flow-interruption experiment reported by Tran et al. (1998a, with permission). Circles are experimental data, and lines are fits of the reaction-advection-diffusion transport model.
the no-flow period causing an increase in the Cd concentration of the interstitial pore water. The temperature and pH changes were small and insufficient to explain the observed desorption. However, in some other studies on Cd, increases in effluent concentrations have been observed at the start of desorption in flow-through experiments (e.g., Dunnivant et al., 1992; Selim et al., 1992). This has generally been termed the "snowplow effect," which is not related to the nonequilibrium effect and can be caused by factors such as changes in ionic strength of influent solutions. Tran et al. (1998a) fitted transport model solutions to the experimental data; some results are shown in Figure 3.10. Note that the post-interruption limb of the data is fitted well by the two-site, nonequilibrium transport model, which suggested that either non equilibrium sorption was occurring, or that a nonlinear, equilibrium isotherm was active.
ModeJ Fitting Sorption equilibrium or nonequilibrium during transport of a solute can be tested by checking the fit of various equilibrium or nonequilibrium models to the data. It has commonly been observed that for a range of inorganic and organic solutes, a two-site nonequilibrium model successfully described the observed BTCs (e.g., Valocchi, 1985; Dunnivant et al., 1992; Gaber et al., 1992; Kookana et al., 1993, 1994). Similar sorption behavior has been observed for Cd. Dunnivant et al. (1992), for example, noted that a two-site nonequilibrium approach was necessary to describe Cd BTCs in aquifer columns, particularly in the presence of dissolved organic carbon. Kookana et al. (1994) reported that the two-site non equilibrium model fitted their observed breakthrough data better than the equilibrium model in two soils. From experimental verification with nonreactive solutes, they suggested that in an Oxisol, the observed nonequilibrium behavior of Cd might have been due to sorption-related nonequilibrium, whereas physical nonequilibrium, perhaps due to the presence of immobile water, was more likely in a
Sorption-Desorption Equilibria and Dynamics of Cadmium During Transport in Soil
77
Spada sol. Hinz and Selim (1994) studied Cd movement in two soils and used the observed BTCs to test equilibrium sorption models based on eight different sorption isotherms. They observed that, while in one soil (Typic Udipsamment) transport models based on equilibrium sorption isotherms were adequate to describe Cd and Zn BTCs, none of the models could describe the BTCs in the other soil (Aquic Fragiudalf). The authors suggested that possibly the finer texture of the latter soil may have caused the nonequilibrium conditions. In other studies, equilibrium sorption models have been found to be sufficient to describe the observed Cd sorption in column experiments (e.g., Gerritse, 1996; Bajracharya et aI., 1996). Gerritse (1996) in his study, however, noted that with the equilibrium sorption model, the dispersivity for Cd was greater than that for Cl, a nonreactive solute. The increased dispersivity was ascribed to the heterogeneous distribution of sorption sites in soil. Gerritse (1996) also showed that a BTC from a Spodosol, which showed substantial tailing and fitted a two-site nonequilibrium model (Kookana et aI., 1994), could be described by allowing a higher dispersivity in the equilibrium model. The increase in dispersivity could result from both nonequilibrium as well as equilibrium sorption conditions. For example, the dispersion due to solute exchange between mobile and immobile regions (physical nonequilibrium) can be lumped into a dispersivity term (e.g., Passioura, 1971; van Genuchten and Dalton, 1986). Similarly, pore scale heterogeneity could lead to different dispersion coefficients for reactive and nonreactive solutes even when local equilibrium conditions are maintained during flow (e.g., Sugita and Gillham, 1995). A clear discrimination between equilibrium and nonequilibrium processes should be possible only through experiments in which Cd transport is studied by combining varying flow rates with flow interruption tests.
Mass Balance Check for Complete BTCs In cases where complete breakthrough is achieved (i.e., the concentration in the breakthrough curve is the same as that in the feed solution), the concept of column holdup (Huber and Gerritse, 1971; van Genuchten and Parker, 1984; Barry and Sposito, 1988; Barry and Bajracharya, 1995; Bajracharya and Barry, 1997a), i.e., the mass of Cd retained in column, can be used to check the results of model fitting. The holdup or the total mass of Cd in the column at the instant when influent concentration equals that of the effluent is (Barry and Bajracharya, 1995):
(4)
where So is the solid phase concentration existing in equilibrium with the influent liquid phase concentration, Co, and Vo is the pore volume in the sand column. Note that this expression is valid for either linear or nonlinear, and for both equilibrium and none quilibrium sorption. Holdup can be calculated in two ways, either directly from the breakthrough curves or from the above equation using the fitted isotherm. Bajracharya et al. (1996) calculated H using both approaches, and found the results agreed very well except for the single experiment where the model fits indicated nonequilibrium sorption to be active.
78
Fate and Transport of Heavy Metals in the Vadose Zone
In this one case, the holdup estimates based on the nonequilibrium model differed from the experimental value by 20%. Their results show that some care is needed in interpreting results of model fits to data, even in cases where good fits are obtained.
Causes of Sorption Nonequilibrium During Transport Sorption nonequilibrium may arise due to the sorption reaction itself being slow (chemical kinetics), or if the soil solution is not well mixed (physical nonequilibrium), or both. Skopp (1986) provided a critical analysis of time-dependent flow processes in soils. Given that chemical kinetics of Cd are relatively fast and that the occurrence of sorption nonequilibrium during transport tends to depend on soil type (as discussed above), it is likely that in most cases reported above, the time-dependency of Cd sorption represents physical nonequilibrium, such as incomplete mixing of solution between mobile and immobile domains in soil. Time-dependence of Cd sorption during transport might be due to film or intraparticle diffusion, or diffusive transfer between mobile and immobile portions of the flow domain (Barry and Li, 1994). In both cases, the nonequilibrium is caused by diffusive transport within some stagnant liquid in the soil and is characterized by a transfer rate parameter, a and /3, the fraction of pore space that is mobile. The parameters, a and /3, are known as nonequilibrium parameters, and can arise in modeling either physical or chemical nonequilibrium solute transport processes. The rate parameter, a, is best considered as an apparent parameter. It has been shown to vary under variations in flow velocity (Schulin et al., 1987; Griffioen et al., 1998). The parameter, /3, in general does not vary much with water content or pore water velocity (Schulin et al., 1987; Bajracharya and Barry, 1997b; Griffioen et al., 1998), although there are some reports indicating velocity dependence (e.g., Nkedi-Kizza et al., 1983; Brusseau et al., 1994). The main physical transport processes are characterized by the solute advection rate, V [LT- 1], and the dispersion coefficient, D [L2 T- 1]. Solutes that undergo time-dependent reactions with soil minimally require two more parameters; viz., the equilibrium partition coefficient, K [0 M- 1], and the (in this case) chemical reaction rate, a [T- 1]. Both K and a are usually determined from batch experiments, while V and D are determined using a nonreactive tracer in a flow experiment. This separation of physical and chemical parameters is not always useful. Gerritse (1996), for example, found that the dispersivity for transport of Cd was much greater than that for CI-, which is considered to be a tracer. Other, similar, findings have been reported by Boekhold and Van der Zee (1992), Kookana et al. (1994) and Hinz and Seiim (1994).
Cd Transport Under Field Conditions and Its Modeling There are relatively few studies on the transport behavior of Cd under field conditions and its model simulation (Sidle et al., 1977; Dowdy et al., 1991; Streck and Richter, 1997a). Slow displacement of heavy metal cations and therefore the need for long-term experimentation, and the environmental concerns associated with any input of heavy metals for experimental purposes may be some of the reasons for small-scale rather than large-scale studies (Streck and Richter, 1997a). However, continuous intentional or unintentional applications of metal-rich sewage sludge or wastewater to land or the expo-
Sorption-Desorption Equilibria and Dynamics of Cadmium During Transport in Soil
79
sure of sites to industrial wastes has provided a situation under which heavy metal transport has been studied. Generally, the downward movement of heavy metals has been found to be very slow, restricting them within a few centimeters of the incorporation depth. However, in some cases significant leaching of Cd and Zn has been observed (Dowdy et al., 1991). In the study by Dowdy et al. (1991), however, massive amounts of sludge were applied every year for 14 years on highly structured soils. Lund et al. (1976) reported elevated concentrations of Cd, and other heavy metals up to a depth of 3 m in coarse-textured soils under sludge drying ponds in use for >20 years. In another study involving a sewage sludge disposal site, Sidle and Kardos (1977) found that 6.6% of the applied Cd in the form of sludge was recovered in the percolate at 120 em depth in the soil profile. In this study, simulations with a simple transport model based on Freundlich adsorption isotherm of cationic form of Cd predicted virtually no movement of Cd in the soil (Sidle et al., 1977). They concluded that the complexed form of Cd as well as its movement through preferential pathways should be taken into account for any predictive simulation. Recently, in a study by Streck and Richter (1997a), transport of Cd and Zn was investigated at field scale following application of wastewater for 29 years on a sandy soil. They found that Cd and Zn were partly displaced to a depth of 0.9 and 0.7 m, respectively. The simulations of Cd leaching were carried out with various types of modeling approaches employing parallel soil column (PSC) and convective-dispersive (CD E) approaches using either a grid model (points in the field) or Monte-Carlo model (Streck and Richter, 1997b). Sorption was described by an extended Freundlich equation capable of taking the spatially variable organic carbon content and pH at the site into account. The observed profile of Cd agreed well with that predicted with the PSC model, both with grid and Monte-Carlo simulations. Therefore, it was concluded that the spatial variability of sorption could adequately describe the field-scale dispersion of Cd. It is noteworthy that the sorption data used in the simulations were measured in the presence of an electrolyte matching the mean ionic strength of the wastewater used for irrigation and included CI anion (0.0025 M CaCI 2). The simulations with the CDE modeL however, could agree with measured data only when the dispersivity parameter was adjusted to 0.29 m-a value considered to be on the higher side of those commonly reported. As discussed in the earlier section, increased dispersivity may result from the physicochemical heterogeneity or nonequilibrium conditions during sorption. The results from this study highlight that spatial variability of key soil properties affecting Cd sorption, such as pH and organic carbon content, has a greater bearing on transport behavior of Cd than perhaps the nonidealilty of microscale processes, such as sorption nonequilibrium. The importance of heterogeneity of soil properties and spatial variability in determining Cd transport through soil was also demonstrated through simulations with a simple root zone model (using stochastic theory) by Boekhold and Van der Zee (1991) for a sandy soil. In the example they considered (representing a site exposed to combined agricultural and industrial activities with 50 glhafyr rate of Cd input), simulations showed that after 40 years, the average Cd concentration in mean water flux can reach the Dutch reference value for groundwater (1.5 J.lglL). They recommended that when groundwater quality is of major concern, accurate knowledge of sorption parameters and input rates of Cd are crucial for reliable results, because leaching rates are very sensitive to Cd input rate and to flow and sorption parameters. They observed that large areas in the field may
80
Fate and Transport of Heavy Metals in the Vadose Zone
have high leaching rates, which may remain undetected by simulations with the average behavior of Cd. Therefore, soil heterogeneity of both soil physical and chemical properties must be taken into account in an assessment of Cd leaching through soil profIles.
Desorption and Reversibility of Cd Sorption Reports of desorption of Cd from soils are relatively fewer in the literature than those of sorption, except those carried out using specifIc extractants to establish the solid phase speciation of Cd (Tiller, 1996). In sorption-desorption experiments on Cd in soils, both complete and partial reversibility of sorption have been reported in the literature. For example, Christensen (1984b) studied desorption of Cd at low Cd concentrations (0.1 to 6 mg/g in soil) in two Danish soils (loamy sand and sandy loam) at pH 6.0. They observed a full reversibility of Cd sorption in the loamy sand but only partial in the sandy loam. Mayer (1978) also noted similar full reversibility of Cd sorption in an acid subsurface soil over a wide range of solution concentrations of Cd (1-10000 ~/L). Complete reversibility of sorbed Cd from poorly crystalline kaolinite was also reported by PuIs et al. (1991). Similarly, Cd sorption was found to be completely reversible in both column and batch experiments in an Australian Oxisol, whereas hysteresis was observed in AlfIsol (Kookana et al., 1994; Naidu et al., 1997). The ambient pH of the Oxisol was closer to the point of net zero charge and therefore the soil had a very low cation exchange capacity. In contrast, Amacher et al. (1986), while studying the desorption behavior of Cd in fIve different soils also found incomplete reversibility of several metals, including Cd. Mter allowing the sorption reaction between metals and soils to proceed for 336 hr, they carried out desorption in 0.005 M Ca(N03h A signifIcant fraction of the sorbed metals did not desorb from soils (Table 3.1) even after the long desorption period. In this study the Ca was present in suffIciently high concentration as Ca(N03)2 to replace the Cd on exchange sites. It was suggested that the poor reversibility in these studies may have been due to specifically sorbed Cd on metal oxides or organic matter, or due to formation of insoluble compounds or coprecipitated Cd. Recently, Kookana (unpublished data) observed sorption desorption hysteresis in two AlfIsols from South Australia, as shown by the data presented in Figure 3.llA,B. Tran et al. (1998b) reported Cd desorption experiments carried out on a homogeneous sand medium at constant-pH of 5.5 and 6.5. They compared sorption/desorption isotherms and found signifIcant hysteretic behavior (Figure 3.11 C) at a pH of 6.5. The partial reversibility of Cd is likely to be linked to the mechanism of sorption in soils, as discussed below.
Desorption of Specifically Sorbed Cd Tiller et al. (1984b) in their desorption studies on clay-sized isolates from several soils found that at a soil pH of 5, up to 85% of sorbed Cd was easily desorbed rapidly in 0.01 M Ca(N03h However, at pH 7, they noted that the easily desorbed fraction (they called it nonspecifically sorbed) was much lower, particularly from the clay-sized fraction from Oxisol. The lower proportion of desorbed Cd at higher pH may be because the surface may be highly undersaturated relative to the number of sorption sites available for binding (Naidu et al., 1997). Thus, it appears that the reversibility of sorbed Cd is a function of the nature of sorbent and the soil conditions determining the affinity of Cd for soil (Kookana et al., 1997). As discussed earlier, Cd can form high affinity inner-
Sorption-Desorption Equilibria and Dynamics of Cadmium During Transport in Soil
81
Table 3.1. Fraction of Sorbed Cd Released in Solution after 336 hr Desorptiona Cd Released
Freundlich K
pH
CEC cmol(+)/kg
Fe 2 0 3
Soil Type
(%)
(%)
(n)
Typic Hapludults Typic Udifluvents Aquic Fragiudalfs Vertic Haplaquepts Typic Udipsamments
5.1 7.4 6.4 5.4 5.4
3.72 6.20 8.31 31.3 1.20
10.2 0.44 1.14 0.94 2.20
50.3-74.9 5.4-34.1 2.2-15.6 3.6-9.4 24.7-51.5
a
8.0 71.5 147.2 189.2 21.8
(0.89) (0.81) (0.85) (0.92) (0.84)
After Amacher et aI., 1986. 8.0,------------------,
A
6.0
f-
OJ
I
I
0.0 r- B 30.0 I-
I
~e
0-----
.o-~i----
~ 20.0 I~ E /e C 10.01- . /
~o
C/)
~~ ~IL_ ~I_~IL__~I
0.0 ...
__
__
c
~
cp
4.0 -
//
3.5 -
I
3.0 -
__
cp I
10
,.
/'
e
I _ __'__ I _ _I'___--' L__---'-_ _~
2.5 0.0
0.1 0.2 0.3 0.4 Solution concentration (rng/L)
0.5
Figure 3.11. Cd adsorption and desorption batch isotherms for (A) an Alfisol, solution pH 6.1 ± 0.1 , (B) an Alfisol pH 6.6 ± 0.2 (Kookana, unpublished data), and (C) a homogeneous sand; solution pH 6.5 (Tran et aI., 1997b, with permission). The ordinate represents the Cd solution phase concentration, and the abscissa is the solid phase concentration.
sphere complexes as well as coprecipitates with certain minerals and from such sorbent the desorption of Cd is unlikely to be fully reversible. On certain sorbents, the desorption of Cd and other metals has been found to be influenced by the time of metal-sorbent contact (Brummer et al., 1988; Backes et al., 1995). Backes et al. (1995) studied the kinetics of Cd and Co desorption from synthetic Fe and Mn oxides (at pH - 6) by both batch and flow methods. They concluded that not only the oxide sorbed large amounts of Cd and Co but substantial proportions of sorbed metals could not readily be des orbed in soil solution, especially from Mn oxides. The
82
rate and Transport of Heavy Metals in the Vadose Zone
rate of desorption from goethite became progressively slower with contact time between sorbate and the sorbent. While, as the authors state, several mechanisms may cause such effects, it is not clear if the physical or chemical changes were responsible for reduced desorption with contact time.
Partial Reversibility of Cd Sorption from Calcite and Calcareous Soils In batch experiments on calcite surfaces, sorption of Cd has been found to be only partially reversible (Zachara et al., 1991). Similarly, in flow-through experiments on a calcareous soil, Buchter et al. (1996) found that 35% of the applied Cd did not elute from the column, indicating sorption hysteresis. They also reported that, in a batch study on the same soil, pronounced hysteresis in Cd sorption-desorption was observed. Batch desorption experiments on a calcareous soil from South Australia (Kookana, unpublished data in Figure 3.11) also show the sorption hysteresis of Cd. The partial reversibility from calcite may be due to dehydration of sorbed Cd and coprecipitation as suggested by Davis et al. (1987). Cd reversibility may be time dependent and may be so slow that during its transport in natural systems such as groundwater, nonequilibrium behavior could become evident (Zachara et al., 1991).
((1 Desorption Kinetics Desorption kinetics of Cd in soils is relatively little understood. However, from published studies on natural soils and synthetic minerals it appears that desorption kinetics of Cd depend on the sorbent properties as well as experimental conditions. In batch experiments involving shaking, often the desorption equilibrium is achieved within hours. For example, Tiller et al. (1984b) noted that, at pH 5, up to 85% of Cd sorbed on claysize fractions from soils was desorbed rapidly in one quick wash (5 min) with 0.01 M Ca(N03h Kookana et al. (1997) reported that, in two Australian soils, solution concentration of Cd during desorption did not significantly change after 2 hours of shaking. Similarly, Tran et al. (1998b) carried out a series of batch kinetic desorption experiments at pH 6 which showed no significant difference in Cd solution phase concentration between 1 day and 10 days equilibration, indicating that desorption of Cd was not time dependent over that time scale. In contrast, however, Amacher et al. (1986), in their batch kinetic studies on Cd desorption in five soils, noted a rapid initial phase of desorption followed by a slower phase. However, they noted that although the overall retention/release reaction was not in equilibrium, the metal and soil reaction was almost instantaneous. On synthetic Fe and Mn oxides the desorption of Cd has been found to continue for several days and may be diffusion controlled. Using synthetic goethite, Gerth et al. (1993) found that extraction of metals (added at 10-6 M and sorbed during a reaction period of 21 d at 35°C) with 0.7 M HN0 3 for 14 days at 35°C released 89,72, and 60% Ni, Cd, and Zn, respectively. This supported the observations of Brummer et al. (1988) that, during the sorption process, metals become immobilized, possibly by diffusion into highly specific binding sites in goethite micropores, which protect them against acid attack. Recently, Backes et al. (1995) studied the desorption behavior of Cd and Co on Fe and Mn oxide surfaces at 20°C and compared the amounts desorbed from soils in contact with metals for 1- and 15-week periods. In this study, the desorption of metals was in-
Sorption-Desorption Equilibria and Dynamics of Cadmium During Transport in Soil
83
duced by continuous pumping of 0.0 1M Ca(N0 3)2 through oxides at a flow rate of 2.53 mLimin. The results showed a rapid and slower phase of desorption, the latter being the dominant in terms of total desorption. The amount of Cd associated with the slower phase increased with increased contact time between Cd and sorbent during sorption. When sorption was allowed over 15 weeks, the amount of Cd desorbed (within 5 hr) was found to be almost half of that desorbed when the sorption reaction time was only 1 week. Therefore, it appears that both the rate and extent of Cd release is dependent on the nature of the sorbent and mechanism of sorption of Cd.
Sorption Reversibility in Flow-Through Experiments A comparison of the sorption and desorption flanks of the breakthrough curves (BTC), can provide clues about the reversibility of solute sorption during transport. For a solute showing a linear sorption isotherm and symmetrical breakthrough curve, the desorption flank of BTC when inverted and superimposed on the sorption flank should match if the sorption is reversible. However, nonlinearity of sorption isotherm and kinetics of sorption-desorption reactions can influence the shape of sorption and desorption flanks of a BTC and, hence, deviations on superimposition may be seen. In such cases it is important is to assess the mass balance for the solute entering and eluting out of the column. KookanCl. et al. (1994) conducted Cd transport experiments on an Oxisol to obtain both sorption and desorption fronts of a BTC. When the sorption and desorption flanks of the Cd BTCs were inverted and superimposed, some deviation between the two flanks of the Cd BTC was noted, but the mass balance (the areas on the left of the two flanks of the BTCs) were essentially the same during sorption and desorption phases (Naidu et al., 1997). These results show that nearly all the Cd that was introduced into the soil was recovered during desorption. Campbell et al. (1987) similarly carried out Cd sorptiondesorption studies on montmorillonite-humic acid mixtures using the miscible-displacement technique. An examination of their BTCs and mass balance also shows that most of bound Cd did elute from the columns. Partial reversibility of sorption can influence the desorption flank of a BTC markedly, as shown by Tran et al. (1998b). Column experiments were reported by Tran et al. (1998b), in which a 1 mg/L Cd solution was introduced into short (approximately 5 cm) columns containing homogeneous sand, which showed sorption desorption hysteresis for Cd in batch experiments. After the sorption phase was complete (effluent concentration at 1 mg/L), the influent was switched to a Cd-free solution. The resultant breakthrough curve is shown in Figure 3.12. Clearly the symmetry of the influent pulse is not maintained in the breakthrough curve, the asymmetry reflecting the nonsingular or hysteretic sorption/desorption isotherm of Cd. From the above discussion it is clear that Cd sorption in soils is not always reversible. The reversibility is likely to be influenced by the nature of sorbent, pH, and composition of soil solution. Currently the desorption process of Cd is poorly understood and warrants further research.
SUMMARY Cd is sorbed by both specific and nonspecific interactions with soils, depending upon the nature of mineral matter present in soil and soil solution composition. Similarly the
84
Fate and Transport of Heavy Metals in the Vadose Zone
1.0
;J'
bb
5 0 .5 U
o.o~__. . .
o
20
40
60
80
100
Time (h) Figure 3.12. Column effluent concentrations (e) showing adsorption then desorption of a Cd pulse (Tran et aI., 1997b, with permission) at pH 6. The line is a model fit to the experimental data.
extent of Cd sorption is also influenced by the nature of the sorbent as well the composition of soil solution in terms of ionic strength, nature of competing and complexing ions. Ca ions are particularly effective in competing with Cd for sorption sites even at lower ionic strengths. It is therefore important to adequately account for such competition in assessing the mobility of Cd in soils. Similarly, the presence of ligands, especially Cl, in soil solution can markedly influence Cd sorption and mobility. Further research to improve the understanding of the effects of organic ligands on Cd behavior is warranted. In a heterogeneous and dynamic system such as the soil environment, several factors together determine the nature and extent of retention reactions of Cd. While it is difficult to isolate the individual effects of various factors influencing Cd, some workers have been able to develop relations which can quantitate the individual influence of Cd concentration, pH, and ionic strength of Ca, during sorption. Clearly, more efforts are needed in this area to develop a better understanding of influences of various factors in multivariate systems. Most research on Cd sorption desorption equilibria and kinetics has been carried out in well-mixed batch systems. While such systems are easy, quick, and suitable for establishing the fundamental retention reactions, they do not always represent the realistic conditions under which Cd mobility needs to be assessed; e.g., under flow conditions. Cd sorption behavior has been found to be different in flow-through systems as compared to batch systems. However, from the current work it is not possible to conclude whether batch systems over- or underestimate the sorption of Cd. Clearly the sensitivity of Cd sorption to several factors, and the fundamental differences between the techniques, together with varying conditions in different experiments even with the same technique, makes it very difficult to draw any meaningful conclusions. Desorption of sorbed Cd from soil is probably more relevant for the assessment of its mobility and potential adverse impact on the environment. However, it remains a poorly
Sorption-Desorption Equilibria and Dynamics of Cadmium During Transport in Soil
85
understood phenomenon. From the limited work available in literature, it is concluded that sorption of Cd is not always reversible. Indeed, the reversibility of Cd sorption depends upon the nature of the sorbent as well as the desorbing solution. Cd interactions with calcite, which is the essential component of calcareous soils and some aquifers, can have significant implications for Cd mobility in the environment. It has been shown that sorption of Cd followed by coprecipitation and dehydration on calcite can result in partial reversibility and possible nonequilibrium conditions under flow conditions. Also, the high pH in calcareous soils favors Cd retention and therefore its mobility is likely to be limited in such soils. Ease of Cd desorption is likely to be linked to the binding affinity of Cd to the sorbent and, therefore, under conditions where Cd shows feeble binding, its sorption is likely to be reversible. Under such conditions, both sorption as well as desorption processes will favor greater mobility of Cd. Highly weathered acidic soils (e.g., some Australian Oxisols), or acidic sandy soils with inherently low cation exchange capacities, are therefore likely to have greater availability of Cd in soil solution, which may have implications for its plant uptake of leaching through the soil profile.
REFERENCES AkratanakuL S., L. Boersma, and G.O. Klock. Sorption process in soils as influenced by pore water velocity: II. Experimental results. SoiL Sci. 135, pp. 331-341, 1983. Allison, J.D. and D.S. Brown. MINTEQA2/PRODEFA2-A Geochemical Speciation Model and Interactive Preprocessor, in ChemicaL Equilibrium and Reaction Modeu, SSSA Special Publication 42, pp. 241-252. (Soil Science Society of America, Madison, WI), 1995. Alloway, B.J. Cadmium, in Heavy Metau in Soiu. pp. 100-124. B.J. Alloway, Ed., John Wiley & Sons, Inc., New York, 1990. Al-Soufi, R.W. A method for simulating cadmium transport in soil: Model development and experimental evaluation. J. Hydro!' 163, pp. 233-247, 1994. Amacher, M.C., J. Kotuby-Amacher, H.M. Selim, and I.K Iskandar. Retention and release of metals by soils-Evaluation of several models. Geoderma. 38, pp. 131-154, 1986. Backes, C.A., R.G. McLaren, A.W. Rate, and R.S. Swift. Kinetics of cadmium and cobalt desorption from iron and manganese oxides. SoiL Sci. Soc. Am. J. 59, pp. 778-785, 1995. Bajracharya, K Transport of Cadmium in Soil. D. Eng. Thesis. Asian Institute of Technology, Bangkok, Thailand, 1989. Bajracharya, K and D.A. Barry. Accuracy criteria for linearised diffusion wave flood routing. J. Hydro!' 195, pp. 200-217, 1997a. Bajracharya, K and D.A. Barry. Nonequilibrium solute transport parameters and their physical significance: Numerical and experimental results. J. Contam. HydroL. 24, pp. 185-204, 1997b. Bajracharya, K, Y.T. Tran, and D.A. Barry. Cadmium adsorption at different pore water velocities. Geoderma 73, pp. 197-216, 1996. Barrow, N.J. Reactions with variable-charge soils. FertiLizer &d. 14, pp. 1-100, 1987. Barry, D.A. and K Bajracharya. Optimised Muskingum-Cunge solution method for solute transport with equilibrium Freundlich reaction. J. Contam. HydroL. 18, pp. 221-238, 1995. Barry, D.A. and L. Li. Physical Basis of Nonequilibrium Solute Transport in Soil, in 15th InternationaL CongrNJ of SoiL Science TranJactionJ, AcapuLco, Mexico, JuLy 10-16. International Society of Soil Science & Mexico Society of Soil Science, 2a, pp. 86-105, 1994. Barry, D.A. and G. Sposito. Application of the convection-dispersion model to solute transport in finite soil columns. SoiL Sci. Soc. Am. J. 52, pp. 3-9, 1988.
86
Fate and Transport of Heavy Metals in the Vadose Zone
Bingham, F.T., G. Sposito, and J.E. Strong. The effect of sulfate on the availability of cadmium. SoiL Sci. 141, pp. 172-177, 1986 Boekhold, A.E. and S.E.A.T.M. Van der Zee. Long-term effects of soil heterogeneity on cadmium behaviour in soil. J. Contam. HydroL. 7, pp. 371-390, 1991. Boekhold, A.E. and S.E.A. T.M. Van der Zee. A scaled sorption model validated at the column scale to predict cadmium contents in a spatially variable field soil. Soil Sci. 154, pp. 105-112, 1992. Boekhold, A.E., E.J.M. Temminghoff, and S.E.A.T.M. Van der Zee. Influence of electrolyte composition and pH on cadmium sorption by an acid soil. J. SoiL Sci. 44, pp. 85-96, 1993. Boesten, J.J. T.1. and L.J. T. Van der Pas. Modeling adsorption/desorption kinetics of pesticides in a soil suspension. SoiL Sci. 146, pp. 221-231, 1988. Bolton, K.A., S. Sjoberg, and L.J. Evans. Proton binding and cadmium complexation constants for a soil humic acid using a quasi-particle model. SoiL Sci. Soc. Am. J. 60, pp. 1064-1072, 1996. Brummer, G., J. Gerth, and K.G. Tiller. Reaction kinetics of the adsorption and desorption of Ni, Zn and Cd by goethite. I. Adsorption and diffusion of metals. J. SoiL Sci. 39, pp. 35-52, 1988. Brusseau, M.L., Z. GerstL D. Augustijn, and P.S.C. Rao. Simulating solute transport in an aggregated soil with the dual-porosity model: Measured and optimised parameter values. J. HydroL. 163, pp. 187-193, 1994. Brusseau, M.L., P.S.C. Rao, R.E. Jessup, and J.M. Davidson. Flow interruption: A method for investigating sorption non-equilibrium. J. Contam. HydroL. 4, pp. 223-240, 1989. Brusseau, M.L. and P.S.C. Rao. Sorption nonideality during organic contaminant transport in porous media. CRC Crit. Rev. Environ. ControL 19, pp. 33-99, 1989. Buchter, B., C. Hinz, M. Gfeller, and H. Fluhler. Cadmium transport in an unsaturated stony subsoil monolith. Soil Sci. Soc. Am. J. 60, pp. 716-721, 1996. Burgisser, C., A. Scheidegger, M. Borkovec, and H. Sticher. Transport and adsorption of cadmium in columns. Deat. BOden!.:. GeddeLL. MilteiL. 66, pp. 283-286, 1991. CampbelL G.D., H.F. Galcia, and P.W. Schindler. Binding of cadmium by montmorillonitehumic acid mixtures: Miscible displacement experiments. AuAraL. J. SoiL Red. 25, pp. 391-403, 1987. Chardon, W. Mobility of Cadmium in Soil (in Dutch). PhD Thesis. Agricultural University, Wageningen, The Netherlands, 1984. Christensen, T.H. Cadmium soil sorption at low concentrations: Effect of time, cadmium loading, pH and calcium. Water Air SoiL PoLLut. 21, pp. 105-114, 1984a. Christensen, T.H. Cadmium soil sorption at low concentrations: II. Reversibility, effect of changes in solute composition, and effect of soil aging (die.). Water Air SoiL PoLLut. 21, pp. 115125, 1984b. Christensen, T.H. Cadmium soil sorption at low concentrations: V. Evidence of competition by other heavy metals. Water Air SoiL PoLLut. 34, pp. 293-303, 1987. Davis, J.A. and J.O. Leckie. Surface ionization and complexation at the oxide/water interface. J. CoLwiJ Interface Sci. 67, pp. 91-107, 1978. Davis, J.A. Complexation of trace metals by adsorbed natural organic material. Geochimca et COdmochimca Acta. 48, pp. 677-691, 1984. Davis, J.A., C.C. Fuller, and A.D. Cook. A model for trace metal sorption process at the calcite surface: Adsorption of Cd+ 2 and subsequent solid solution formation. Geochimiea et COdmochimiea Acta, 51, pp. 1477-1490, 1987. Doner, H.E. Chloride as a factor in mobilities of Ni(I1), Cu(II) and Cd(II) in soil. SoiL Sci. Soc. Am. J. 42, pp. 882--885, 1978. Dowdy, R.H., J.J. LattereiL T.D. Hinesly, R.B. Grossman, and D.L. Sullivan. Trace metal movement in an aeric Ochraqualf following 14 years of annual sludge applications. J. Environ. QULlL 20, pp. 119-123, 1991.
Sorption-Desorption Equilibria and Dynamics of Cadmium During Transport in Soil
87
Dowdy, RH. and V.V. Volk. Movement of Heavy Metals in Soils, in Chemical MObility and Reactivity in Soil SY
88
Fate and Transport of Heavy Metals in the Vadose Zone
Holm, P.E., B.B.H. Anderson, and T.H. Christensen. Cadmium solubility in aerobic soils. Soil Sci. Soc. Am. J. 60, pp. 775-780, 1996. Homann, P.S. and RJ. Zasoski. Solution composition effects on cadmium sorption by forest soil profiles. J. Environ. QuaL. 16, pp. 429-433, 1987. Hornburg, V. and G.W. Brummer. Verhalten von Schwermetallen in Boden. 1. Untersuchungen zur Schwermetallmobilitat. Zeitdchrift fur Pflanzenernahrung und BOdenkunde 156, pp. 467-477,
1993. Huber, J.F.K and RG. Gerritse. Evaluation of dynamic gas chromatographic methods for the determination of adsorption and solution isotherms. J. Chromatography 58, pp. 137-158, 1971. Jardine, P.M. and D.L. Sparks. Potassium-calcium exchange in a multireactive soil systems. 1. Kinetics. SoiL Sci. Soc. Am. J. 48, pp. 39-45, 1984. Klamberg, H., G. Matthess, and A. Pekdeger. Organo-Metal Complexes as Mobility-Determining Factors of Inorganic Toxic Elements in Porous Media, in Inorganic Contaminantd in the VadOde Zone, Ecological Studies, Vol. 74, B. Bar-Yosef, N.J. Barrow, and J. Goldsmith, Eds., Springer-Verlag, Berlin, pp. 3-17, 1989. Kookana, RS. and R Naidu. Effect of Soil Solution Composition on Cadmium Transport Through Variable Charge Soils, in Proceedingd Firdt InternationaL Conference Contaminantd and the SoiL Environment-Feu. '96, AdelaUJe, AlldlraLia, pp. 29-30, 1996. Kookana, RS. and R Naidu. Effect of soil solution composition on cadmium transport through variable charge soils. Geoderma, 84, pp. 235-248, 1998. Kookana, RS., R Naidu, and KG. Tiller. Sorption non-equilibrium during cadmium transport through soils. Alldtrai. J. SoiL Ru. 32, pp. 635-651, 1994. Kookana, RS., R Naidu, and KG. Tiller. Desorption of Cadmium is Determined by Its Adsorption Affinity to Soils, in Proceedingd Fourth InternationaL Conference on the Biogeochemi.Jtry of TraceElementd. Berkeley, CaLifornia, I.K Iskandar, S.E. Hardy, A.C. Chang, and G.M. Pierzynski, Eds., 1997, pp. 399-400. Kookana, RS., RD. Schuller, and L.A.G. Aylmore. Simulation of simazine transport through soil columns using time-dependent sorption data measured under flow conditions. J. Contam. HydroL. 14, pp. 93-115, 1993. Krishnamurti, G.S.R, G. Cieslinski, P.M. Huang, and KC.J. Van Rees. Kinetics of cadmium release from soils as influenced by organic acids: Implication in cadmium availability. J. Environ. QuaL. 26, pp. 271-277, 1997. Lieber, M., N.M. Perlmutter, and H.L. Frauenthal. Cadmium and hexavalent chromium in Nassau County ground water. J. Am. Water Workd A
1013-1018, 1994. McLaughlin, M.J., KG. Tiller, and M.K Smart. Speciation of cadmium in soil solutions of saline/ sodic soils and relationship with cadmium concentrations in potato tubers (Solanum tuuerodum L.). Alldtrai. J. SoiL Ru. 35, pp. 183-198, 1997. Miller, D.M., M.E. Sumner, and W.P. Miller. A comparison of batch- and flow-generated anion adsorption isotherms. SoiL Sci. Soc. Am. J. 53, pp. 373-380, 1989. Morisawa, S. and Y. Inoue. Mathematical models for estimating dynamic performance in the saturated zone: Discussions on the mechanisms of groundwater pollution by cadmium. Proc. Environ. Sanitation Eng. Ru. 21, pp. 43-55, 1985.
Sorption-Desorption Equilibria and Dynamics of Cadmium During Transport in Soil
89
Murali. V. and L.A.G. Aylmore. No-flow equilibrium and adsorption dynamics during ionic transport in soils. Nature 283. pp. 467-469. 1980. Naidu. R .• N.J. deLacy. N.S. Bolan. RS. Kookana. and KG. Tiller. Effect of Inorganic Ligands on Adsorption of Cadmium by Soils. in ProceedingJ of the 15th InternationaL Soil Science Society Congrc.:l.:l, JuLy 1994. pp. 190-191. 1994b. Naidu. R. N.S. Bolan. RS. Kookana. and KG. Tiller. Ionic strength and pH effects on surface charge and Cd sorption characteristics of soils. J. SoiL Sci. 45. pp. 419-429. 1994a. Naidu. R. RS. Kookana. M.E. Sumner. R.D. Harter. and KG. Tiller. Cadmium sorption and transport in variable charge soils: A review. J. Environ. QuaL. 26. pp. 602-617. 1997. Neal. RH. and G. Sposito. Effects of soluble organic matter and sewage sludge amendments on cadmium sorption by soils at low cadmium concentrations. SoiL Sci. 142. pp. 164-172. 1986. Nkedi-Kizza. P .• J.W. Biggar. H.M. Selim. M.Th. van Genuchten. P.J. Wierenga. J.M. Davidson. and D.R. Nielsen. On the equivalence of two conceptual models for describing ion exchange during transport through an aggregated oxisol. Water ReJour. ReJ. 20. pp. 1123-1130. 1984. Nkedi-Kizza. P .. J.W. Bigger. M.Th. van Genuchten. P.J. Wierenga. H.M. Selim. J.M. Davidson. and D.R Nielsen. Modeling tritium and chloride 36 transport through an aggregated Oxisol. Water ReJour. ReJ. 19. pp. 691-700. 1983. O·Connor. G.A .. C. O·Connor. and G.R Cline. Sorption of cadmium by calcareous soils: Influence of solution composition. SoiL Sci. Soc. Am. J. 48. pp. 1244-1247. 1984. Parker. D.R. W.A. Norvell. and RL. Chaney. GEOCHEM-PC-A Chemical Speciation Program for IBM and Compatible Personal Computers. in ChemicaL Equilibrium and Reaction Modeu. SSSA Special Publication 42. Soil Science Society of America. Madison. WI. 1995. pp. 253-269. Passioura. J.B. Hydrodynamic dispersion in aggregated media. Soil Sci. Ill, pp. 339-344. 1971. PuIs. RW .. RM. Powell. D. Clark. and C.J. Eldred. Effect of pH. solid/solution ratio. ionic strength. and organic acids on Pb anJ Cd sorption on kaolinite. Water Air SoiL PoLLut. 57-58. pp. 423-430. 1991. Runnells. D.D.. T.A. Shepherd. and E.E. Angino. Metals in water: Determining natural background concentrations in mineralized areas. Environ. Sci. TechnoL. 26. pp. 2316-2323. 1992. Schindler. P.W.• P. Leichti. and J.C. Westall. Adsorption of copper. cadmium. and lead from aqueous solutions to the kaolinite/water interface. NetherlandJ J. Agric. Sci. 35. pp. 219-230. 1987. Schulin. R. P.J. Wierenga. H. Fliihler. and J. Leuenberger. Solute transport through a stony soil. SoiL Sci. Soc. Am. J. 51. pp. 36-42. 1987. Selim. H.M.• B. Buchter. C. Hinz. and L. Ma. Modeling the transport and retention of cadmium in soils: Multireaction and multicomponent approaches. SoiL Sci. Soc. Am. J. 56. pp. lO04-1015. 1992. Selim. H.M. Prediction of contaminant retention and transport in soils using multireaction models. Environ. HeaLth PerJpect. 39. pp. 69-75. 1989. Shimojima. E. and M.L. Sharma. The influence of pore water velocity on transport of sorptive and non-sorptive chemicals through an unsaturated sand. J. HydroL. 164. pp. 239-261. 1995. Sidle. R.C. and L.T. Kardos. Transport of heavy metals in a sludge-treated forested area. J. Environ. QuaL. 6. pp. 431-437. 1977. Sidle. RC .• L.T. Kardos. and M.Th. van Genuchten. Heavy metals transport model in a sludgetreated soil. J. Environ. QuaL. 6. pp. 438-442. 1977. Skopp. J. Analysis of time-dependent chemical processes in soils. J. Environ. QuaL. 15. pp. 205213. 1986. Stipp. S.L.. M.F. Hochella. G.A. Parks. and J.O. Leckie. Cd2+ uptake by calcite. solid-state diffusion. and the formation of solid-solution: Interface processes observed with near-surface sensitive techniques (XPS. LEED. and AES). Geochimica et COJmochimca Acta. 56. pp. 19411954. 1992.
90
Fate and Transport of Heavy Metals in the Vadose Zone
Streck. T. and J. Richter. Heavy metal displacement and in a sandy soil at the field scale: 1. Measurements and parameterization of sorption. J. Environ. QuaL. 26. pp. 49-56. 1997a. Streck. T. and J. Richter. Heavy metal displacement and in a sandy soil at the field scale: II. Modeling. J. Environ. QuaL. 26: pp. 56-62. 1997b. Sugita. F. and RW. Gillham. Pore scale variation in retardation factor as a cause of nonideal reactive breakthrough curves. 3. Column investigations. Water Redour. Red. 31. pp. 121-128. 1995. Tan. T.C. and W.K Teo. Combined effect of carbon dosage and initial adsorbate concentration on the adsorption isotherm of heavy metals on activated carbon. Water Red. 21. pp. 1183-1188. 1987. Temminghoff. E.J.M.• S.E.A.T.M. Van der Zee. and F.A.M. De Haan. Speciation and calcium competition effects on cadmium sorption by sandy soil at various pHs. European J. SoiL Sci. 46. pp. 649-655. 1995. Theis. T.L.. R Iyer. and L.W. Kaul. Kinetic studies of cadmium and ferricyanide adsorption on goethite. Environ. Sci. Tec/moL. 22. pp. 1013-1017. 1988. Tiller. KG. Heavy metals in soils and their environmental significance. Adv. SoiL Sci. 9. pp. 113142. 1989. Tiller. KG. Soil Contamination Issues: Past. Present and Future. A Personal Perspective. in ContaminantdandtheSoiLEnvironmentintheAuAralMiaPacificRegion. R Naidu. RS. Kookana. D.P. Oliver. S. Rogers. and M.J. McLaughlin. Eds .. Kluwer Academic Publishers. 1996. pp. 1-28. Tiller. KG .• J. Gerth. and G. Brummer. The sorption of Cd. Zn. and Ni by a soil clay fraction: Procedures for partition of bound forms and their interpretation. Geoderma. 34. pp. 1-16. 1984a. Tiller. KG .. J. Gerth and G. Brummer. The relative affinities of Cd. Ni and Zn for different clay fractions and goethite. Geoderma. 34. pp. 17-35. 1984b. Tiller. KG .• V.K Nayyar. and P.M. Clayton. Specific and non-specific sorption of cadmium by soil clays as influenced by zinc and calcium. AwtraL. J. SoiL Red. 17. pp. 17-28. 1979. Tran. Y.T .• K Bajracharya. and D.A. Barry. Anomalous cadmium sorption in flow interruption experiments. Geoderma, 84. pp. 169-184. 1998a. Tran. Y.T .• D.A. Barry. and K Bajracharya. Cadmium desorption in sand. Environ. SCt: TechnoL. (in review). 1998b. Valocchi A.J. Validity of the local equilibrium assumption for modelling sorbing solute through homogenous soils. Water Redoar. Red. 21. pp. 808--820. 1985. van Genuchten. M.Th. and F.N. Dalton. Models for simulating salt movement in aggregated field soils. Geoderma 38. pp. 165-183. 1986. van Genuchten. M.Th. and J.C. Parker. Boundary conditions for displacement experiments through short laboratory soil columns. SoiL Sci. Soc. Am. J. 48. pp. 703-708. 1984. Zachara. J.M. and S.C. Smith. Edge complexation reaction of cadmium on specimen and soilderived smectite. SoiL Sci. Soc. Am. J. 58. pp. 762-769. 1994. Zachara. J.M .• C.E. Cowan. and C.T. Resch. Sorption of divalent metals on calcite. Geochimica et COdmochimica Acta. 55. pp. 1549-1562. 1991. Zachara. J.M.• S.C. Smith. C.T. Resch. and C.E. Cowan. Cadmium sorption to soil separates containing layer silicates and iron and aluminum oxides. SoiL Set: Soc. Am. J. 56. pp. 1075-1084. 1992. Zachara. J.M.. S.C. Smith. J.P. McKinley. and C.T. Resch. Cadmium sorption on specimen and soil smectites in sodium and calcium electrolytes. Soil Sci. Soc. Am. J. 57. pp. 1491-1501. 1993. Zasoski. RG. and RG. Burau. A technique for studying the kinetics of adsorption in suspensions. Soil Sci. Soc. Am. J. 42. pp. 372-374. 1978. Zhu. B. and A.K Alva. Differential adsorption of trace metals by soils as influenced by exchangeable cations and ionic strength. SoiL Sci. 155. pp. 61-66. 1993.
CHAPTER
~
Modeling the Kinetics of Heavy Metals Reactivity in Soils H. Magdi Selim
INTRODUCTION The reactivity, toxicity, and mobility of heavy metal in the soil system is of major environmental concern. Thus, understanding of the complex interactions of heavy metals in the environment is a prerequisite in the effort to predict their behavior in the vadose zone. To predict the fate and transport of heavy metals in the soil, models that include retention and release reactions of solutes with the soil matrix are needed. Retention and release reactions in soils include ion exchange, adsorption/desorption, precipitation/dissolution, and other mechanisms such as chemical or biological transformations. Adsorption is the process where solutes bind or adhere to soil matrix surfaces to form outer- or inner-sphere solute surface-site complexes. Over the last 30 years, several models describing the retention behavior of heavy metals in the soil environment have been developed (see Selim et aI., 1990, and Selim, 1992 for a review). Table 4.1 provides a listing of commonly used equilibrium and kinetic retention approaches. A disadvantage of several of such empirical approaches lies in the basic assumption of local equilibrium of the governing reactions. Alternatives to equilibrium-based approaches are the multisite or multireaction models which deal with the multiple interactions of one solute species in the soil environment. Multiple interaction approaches assume that a fraction of the total sites is time dependent; i.e., kinetic in nature whereas the remaining fraction interacts rapidly or instantaneously with that in the soil solution. Nonlinear equilibrium (Freundlich) and first- or nth-order kinetic reactions are the associated processes (for a review see Selim, 1992). Another two-site approach was proposed by Theis et aI. (1988) for Cd mobility and adsorption on goethite. They assumed that the nature of reactions for both sites was governed by secondorder kinetic reactions. The reactions were assumed to be consecutive where the second reaction was irreversible in nature. Amacher et aI. (1988) developed a multireaction model that includes concurrent and concurrent-consecutive processes of the nonlinear 91
92
Fate and Transport of Heavy Metals in the Vadose Zone
Table 4.1. Selected Equilibrium and Kinetic Type Models for Heavy Metal Retention in Soils Model
Formulationa
Equilibrium Type Linear Freundlich General Freundlich Rothmund-Kornfeld ion exchange Langmuir General Langmuir-Freundlich Langmuir with sigmoidicity
S=K.JC S = K.J Cb S/S max = [w CI(l + wC))~ Sj/ST = ~K (C/CT)n S/S max = w CI[ 1 + we] S/S max = (w C)~ 1[1 + (wC)~) S/S max = w CI[l + wC + oK)
Kinetic Type First-order nth-order Irreversible (sink/source) Second-order irreversible Langmuir kinetic Elovich Power Mass transfer
aS/at aS/at aS/at aS/at aS/at aS/at aS/at aS/at
= = = = = = = =
k f (Sip) C - kb S k f (Sip) Cn - kb S ks (Sip) (C - Cpl ks (Sip) C (Smax - S) k f (Sip) C (Smax - S) - kb S A exp(-BS) K (Sip) Cn Sm K (Sip) (C - CO)
a A, B, b, C*, CP' K, K.J, ~K' kb' kf' k" n, m, Smax, W, /3, and (J are adjustable model parameters, p is bulk density, S is volumetric soil water content, CT is total solute concentration and ~ is total amount sorbed of all competing species.
kinetic type. The model was capable of describing the retention behavior of Cd and Cr(VI) with time for several soils. Selim et al. (1992) developed a multicomponent approach which accounts for an equilibrium ion exchange and a specific sorption process based on a second-order (Langmuir) kinetic reaction. The multicomponent model adequately predicted the observed breakthrough results where a pronounced snowplow or chromatographic effect was observed. Eff1uent peak concentrations were 3-5 fold that of the input Cd pulse. In this chapter, we focus on major features of kinetic retention processes and modeling approaches of heavy metals in the vadose zone. Retention approaches of the linear and nonlinear reversible and irreversible kinetic types are first discussed. Models of the multiple-reaction type, including the two-site equilibrium-kinetic models, and ion-exchange multicomponent models are also presented. Illustrative examples of Cu and Cd isotherms and breakthrough results (BTCs) are discussed.
LINEAR RETENTION For several heavy metals (e.g, Cr, Cu, Zn, Cd, and Hg), retention and release reactions in the soil solution have been observed to be strongly time-dependent. Recent studies on the kinetic behavior of the fate of several heavy metals include Aringhieri et al. (1985), and Amacher et al. (1986) among others. A number of empirical models have been proposed to describe kinetic retention and release reactions of solutes in the solution phase. The first-order kinetic approach is perhaps the earliest single (linear) form of reaction used to describe time-dependent sorption which may be expressed as
Modeling the Kinetics of Heavy Metals Reactivity in Soils
P
as
a;- = kf e c - kb P S
93
(1)
where C is the heavy metal concentration in solution (llg/mL), S is the amount of heavy metal sorbed or retained by the soil (Ilg/g soil), P is the soil bulk density (g/cm 3) , is the volumetric soil-water content (cm 3/cm 3 ), and t is time (h). Reaction given in Equation 1 is fully reversible between the heavy metal species present in solution and that sorbed by soil matrix surfaces where k f and kb represent the forward and backward rate coefficients (h- 1). The first-order reaction was first incorporated into the classical (convective-dispersive) transport equation by Lapidus and Amundson (1952) to describe solute retention during transport under steady water flow conditions. Integration of the first-order reaction (Eq. 1) subject to initial conditions of C = C i ' and S = 0 at t = 0, yields a system of linear sorption isotherms. That is, for any reaction time t, a linear relation between Sand C is obtained. However, linear isotherms are not often encountered except for selected cations and heavy metals at low concentrations (see Selim et aI., 1992). In contrast, nonlinear retention behavior is commonly observed for several heavy metals as depicted by the nonlinear isotherms for CU for Cecil and McLaren soils shown in Figure 4.1.
e
NONLINEAR RETENTION As a consequence of nonlinear retention behavior such as depicted in Figure 4.1, the single reaction given by Equation 1 has been extended to include the nonlinear kinetic form,
(2)
where m is a dimensionless parameter commonly less than unity and represents the order of the nonlinear reaction. For both single kinetic forms (Eqs. 1 and 2), the magnitude of the rate coefficients dictates the extent of the kinetic behavior of the reaction. For small values of k f and kb' the rate of retention is slow, and strong kinetic dependence is anticipated. In contrast, for large values of k f and K b, the retention reaction is a rapid one and should approach quasi-equilibrium in a relatively short time. In fact, at large times (t ~ 00) equilibrium is attained and the rate of retention (aSia t) approaches zero and the above equation yields
where
(3)
which is an analogous form to the Freundlich equilibrium equation where Kd is the distribution coefficient (cm 3/g). Therefore, for linear or Freundlich isotherms, one may regard Kd as the ratio of the rate coefficient for sorption (forward reaction) to that for desorption or release (backward reaction).
94
Fate and Transport of Heavy Metals in the Vadose Zone -------
500
o
·0 400 (/)
~
6
2 hours
•
4
•
12
.," 0"
~
Ol
--------~------
300
"0
.... 200
.0 0 (/)
"0
«
::J
()
100 0
t
0
Cu Adsorption
McLaren Soil 20
10
30
40
50
60
70
250 o
2 hours
• 12
~
·0 200 (/)
..
~
Ol
=t .......-
150
"0
.... 100 (/)
.0 0
«"0 ::J
()
50
Cu Adsorption
Cecil Soil 0
0
20
40
60
80
100
Cu in Solution (llg/mL) Figure 4.1. Cu adsorption isotherms at several reaction times in Cecil and McLaren soils.
LANGMUIR OR SECOND-ORDER KINETICS Based on this approach, heavy metal retention mechanisms are assumed to follow a second-order kinetic type reaction where the forward process is controlled by the product of the solution concentration C (mg L- 1) and the amount of unoccupied or unfilled sites (<1» (Selim and Amacher, 1988). Specifically, the reaction may be expressed by the following reversible process:
(4)
Therefore, the kinetic rate equations for heavy metal retention may be expressed by the following kinetic Langmuir (or second-order) equation:
Modeling the Kinetics of Heavy Metals Reactivity in Soils
95
where k f and kb (h- 1) are forward and backward rate coefficients for the retention sites. If is unity, the above equation yields a first-order kinetic retention reaction (Lapidus and Amundson, 1952). However, a major disadvantage of first-order kinetic reactions is that as the concentration in solution increases, a maximum heavy metal sorption is not attained, which implies that there is an infinite heavy metal retention capacity of the soil or infinite amount of exchange sites on matrix surfaces. In contrast, the second-order approach achieves maximum sorption when all unfilled sites become occupied (i.e., ~ 0). We recognize that the unfilled or vacant sites (
(6)
where w represents the equilibrium constant for the reaction associated with the retention sites. Moreover, this form (Eq. 6) is analogous to expressions for homovalent ionexchange equilibrium reactions. In this sense, the equilibrium constant w resembles the selectivity coefficients for exchange reactions of soil matrix surfaces (Sposito, 1984). However, a major difference between ion-exchange and the proposed second-order approach is that no consideration of other competing ions in solution or on matrix surfaces is incorporated into the second-order rate equations. In a strict thermodynamic sense, the above formulation should be expressed in terms of activities rather than concentrations. However, we use the implicit assumption that solution phase ion activity coefficients are constant in a constant ionic strength medium. Further rearrangement of Equation 6 yields the following expression at t ~ 00,
S wC ------
(7)
which is the widely recognized Langmuir isotherm equation. This formulation is the oldest and most commonly encountered in soils. It was developed to describe the adsorption of gases by solids where a finite number of adsorption sites in the surface is assumed. The terms wand Smax are adjustable parameters. Here W (mL Ilg- 1) is a measure of the bond strength of molecules on the matrix surface and Smax (Ilg g-l of soil) is the maximum sorption capacity or total amount of available sites per unit soil mass. In an
96
Fate and Transport of Heavy Metals in the Vadose Zone
attempt to classifY the various shapes of sorption isotherms, it was recognized that the Langmuir isotherm is the most commonly used and is referred to as the L-curve isotherm (Sposito, 1984). Moreover, Langmuir isotherms were used successfully to describe Cd, Cu, Pb, and Zn retention in soils. Figure 4.2 shows experimental and fitted isotherm examples of the use of the Langmuir equation to describe Cu retention in Cecil and McLaren soil after 192 h of reaction.
HYSTERESIS Adsorption-desorption hysteretic behavior has been observed by several scientists. Examples of hysteretic behavior for Cu adsorption-desorption for a McLaren soil is shown in Figure 4.3. Cu desorption shows significant hysteresis or nonsingularity behavior which was apparent at high initial concentrations. Based on the hysteresis behavior for several initial concentrations (not shown), desorption results suggest that part of the adsorbed Cu was not easily desorbed or becomes nondesorbable by forming strong interaction with the soil matrix. Selim et al. (1976) showed that singularity or hysteresis may be partly the result from failure to achieve equilibrium during adsorption and desorption. If adsorption as well as desorption were carried out for times sufficient for equilibrium to be attained, or if the kinetic rate coefficients were sufficiently large, such hysteretic behavior would be minimized.
IRREVERSIBLE REACTIONS Irreversible processes of various heavy metals in the soil account for various (sink! source) reactions, including precipitation/dissolutions, mineralization, immobilization, biological transformations, volatilization and radioactive decay, among others. First-order kinetic reactions have been utilized to quantifY the irreversible retention processes by several authors. Models that account for first-order kinetic and sequential first-order (irreversible) decay reactions include Rasmuson and Neretienks (1981) and Amacher et al. (1988). The first-order irreversible retention form is
(8) where k lrr is the irreversible rate of reaction (h-'). This mechanism has been used to describe various heavy metals adsorption/precipitation, and radio-nuclide decay. Description of precipitation reactions that involve secondary nucleation is not an easy task, and it is often difficult to distinguish between precipitation and adsorption.
SPECIFIC SORPTION It has been postulated that specific sorption on high affinity sites may be regarded as an irreversible process. A second-order approach was modified to describe irreversible or weakly reversible retention when the rate of desorption or release is small, i.e., when the backward rate coefficient kb approaches zero (see Selim et al., 1992). Therefore, a second-order irreversible process can be expressed as
Modeling the Kinetics of Heavy Metals Reactivity in Soils 500
.---------------------------_~-~-~
&---
~
97
--- ---
--McLaren
400
-E
Cl 300
w CD a: 200
•
o
Cecil
(/)
:J
o
100
40
20
60
80
Cu CONCENTRATION (mg/L)
Figure 4.2. Retention isotherms for Cu after 8 days of reactions for Cecil and McLaren soils. Solid curves are calculated isotherms using equilibrium Langmuir model.
500.---------------------~
:=o
400
Desorption
(J)
-~ "0
,
Q)
.0
(; 200 (J)
~ :J
o
r~------------
300
100
I // ,.' ••
o
-'
Adsorption
,
Cu - McLaren Soil 10
20
30
40
50
60
Cu in Solution (J.Lg/mL)
Figure 4.3. Cu adsorption and desorption isotherms for McLaren soil.
as
pa;-=kf
e (ST -S)C
(9)
where only two parameters, ST and kf' are required to account for irreversible retention. The term ST represents the total amount of specific sites (J.Lg g-l of soil). For several metal ions including Cd, Ni, Co, and Zn, specific sorption has been shown to be dependent on time of reaction. Therefore, the use of a kinetic rather than an equilibrium sorption mechanism is recommended. A major advantage of the formulation of irreversible reaction (Equation 7) is that a sorption maximum is achieved when all unfilled sites become occupied (i.e., S ~ ST)' In contrast, for the first-order type Equation 6, as metal ion concentration increases, an irreversible sorption maximum is not attained.
98
Fate and Transport of Heavy Metals in the Vadose Zone
MULTIPLE RETENTION Several studies showed that retention of heavy metals in several soils was not adequately described by use of a single reaction of the equilibrium or kinetic types. The inadequacy of single reactions is not surprising since they only describe the behavior of one species with no consideration to the simultaneous reactions of others in the soil system. Multicomponent models consider a number of processes governing several species in the soil solution including ion exchange, complexation, dissolution/precipitation, and competitive adsorption. However, multicomponent models often consider the local equilibrium assumption (LEA) to be valid. On the other hand, multisite or multireaction models deal with the multiple interactions of one species in the soil environment. Such models are empirical in nature and based on the assumption that a fraction of the total sites are highly kinetic, whereas the remaining fraction of sites interact slowly or instantaneously with that in the soil solution (Selim et al., 1976; Jardine et al., 1985). Nonlinear equilibrium (Freundlich) and first- or nth-order kinetic reactions were the associated processes. The two-site approach proved successful in describing observed extensive tailing of breakthrough results and has been used by several scientists including Jardine et al. (1985), Parker and Jardine (1986), among others. The model proved successful in describing the retention and transport of several dissolved chemicals including aluminum, phosphorus, potassium, cadmium, chromium, and methyl bromide. However, there are several inherent disadvantages of the two-site model. First, the reaction mechanisms are restricted to those that are fully reversible. Moreover, the model does not account for possible consecutive type solute interactions in the soil system. Another two-site approach was proposed by Theis et al. (1988) for Cd mobility and adsorption on goethite. They assumed that the nature of reactions for both sites to be governed by second-order kinetic reactions. The reactions were assumed to be consecutive where the second reaction was irreversible in nature. Other empirical approaches include the consecutive (equilibrium-kinetic) model of Barrow and Shaw (1979). Here an adsorbed (surface) phase was assumed to be in direct equilibrium with that in the solution phase and slowly (and reversibly) with an absorbed (internal) phase within the soil matrix. Another approach is that of van der Zee and van Riemsdijk (1986), in which a reversible reaction for P sorption-desorption was governed according to Langmuir kinetic model (Eq. 5). In addition, precipitation reaction was accounted for as a (kinetic) diffusion process of P through a thin layer of metal phosphate coating that surrounds metal oxides. Metal oxide is assumed to be converted to metal phosphate in the reaction zone by a precipitation-like reaction. Amacher et al. (1988) and Selim et al. (1989) proposed a simplified model which accounts for multiple reactions of solutes during transport in soils. In addition to the soil solution phase (C) of a solute in the soil, four other phases representing solute retained by the soil matrix (Se' S\, S2' S3' and Sire> were also considered. A schematic of the multireaction model is shown in Figure 4.4. We assume Se to be governed by an equilibrium Freundlich reaction, whereas S\ and S2 were governed by nonlinear kinetic reactions,
(10)
Modeling the Kinetics of Heavy Metals Reactivity in Soils
99
Figure 4.4. Schematic diagram of the multi reaction model.
aSl
at
_
-
k 1 -e C n p
-
k 2S1
(11)
(12) where kl to k6 are the associated rates coefficients (h- 1). These two phases (Sl and S2) may be regarded as the amounts sorbed on surfaces of soil particles and chemically bound to Ai and Fe oxide surfaces or other types of surfaces. Amacher et al. (1988) pointed out that it is not necessary to have a prior knowledge of the exact retention mechanisms for these reactions to be applicable. Moreover, these phases may be characterized by their kinetic sorption and release behavior to the soil solution and thus are susceptible to leaching in the soil. In addition, the primary difference between these two phases not only lies in the difference in their kinetic behavior but also on the degree of nonlinearity as indicated by the parameters nand m. The consecutive reaction between S2 and S3 represents slow reaction as a result of further rearrangements of solute retained on matrix surfaces. Incorporation of S3 in the model allows the description of the frequently observed very slow release of solute from the soil. As a result, this strongly retained phase was represented by
(13)
The multireaction model also considers irreversible solute removal via a retention sink term Q in order to account for irreversible reactions such as precipitation/dissolution, mineralization, and immobilization, among others. We expressed the sink term as a firstorder kinetic process in a similar fashion to that given by Equation 6, where k irr is the associated rate coefficient (h- 1). The multireaction model (Eqs. 6 to 10) was incorporated into the classical convectivedispersion transport equation (CDE) which can be expressed as (Selim et al., 1989)
100
Fate and Transport of Heavy Metals in the Vadose Zone
(14)
where D is the hydrodynamic dispersion coefficient (cm 2 h- I), v is Darcy's water flux I density (cm h- ), and z is soil depth (cm). The term R is dimensionless, referred to here as the Freundlich retardation coefficient which is dependent on C,
(15)
For the linear case where the exponent b in Equation 7 is unity, the well-known retardation factor R is obtained
(16)
which is constant. The multireaction model was capable of describing the kinetic retention behavior of P, Cd, Cr, Zn, and Hg based on batch data sets for several soils (see Amacher et al., 1988; Selim et aI., 1990).
ION EXCHANGE RETENTION This mechanism is commonly considered as a rapid reaction of the nonspecific sorption type. The mechanism is a fully reversible reaction between ions in the soil solution and those retained on charged matrix surfaces. The exchange reaction for two competing ions i and j, having valencies Vi and Vj' respectively, may be written as T
K·IJ
(17)
where TK;j denotes the thermodynamic equilibrium constant and a and a" (omitting the subscripts) are the ion activity in soil solution and on the exchanger surfaces, respectively. For the case of a binary homovalent ion, a generic selectivity coefficient K;j (Rubin and James, 1973) or a separation factor for the affinity of ions on exchange sites is often used. Examples of calculated and measured homovalent ion exchange isotherms are illustrated in Figure 4.5 for Cd-Ca for two soils (Selim et aI., 1992). For :KcJCa > 1, sorption of ion Cd is preferred and the isotherms are convex, whereas for K0ICa < 1, sorption affinity is apposite and the isotherms are concave. The capability of the ion exchange approach in describing multiple pulse applications is illustrated in Figure 4.6. Here three input pulses of Cd having a concentration (Co) of 10 mg L -I were consecutively applied to a Windsor soil column (Selim et aI., 1992). The ion exchange model well predicted the position of the BTC peaks. Moreover, the assumption of equilibrium ion exchange with the selectivity coefficient (:KcJCa) based on Figure 4.5 adequately predicts the observed
Modeling the Kinetics of Heavy Metals Reactivity in Soils
-><
1.0
101
,....----------.-----'::;00 Cd-Ca ISOTHERM
c
w
al
a: 0.8
oen
"C
o
LL
0.6
t;
0.4
o z o
« a:
- Kc.tc.= 1 - - Kc.tc. = 0.5 _.- !
LL
w 0.2 ...J
o
:E
•
WINDSOR SOIL
o
EUSTIS SOIL
........ 0.8
0L-~--L.~~...1._~....J._~
o
0.2
0.4
0.6
~---'
1.0
Cd CONCENTRATION (C/C o) Figure 4.5. Cadmium-calcium exchange isotherm for Windsor and Eustis soils. Solid and dashed curves are simulations using different selectivities (I
8 Cd - WINDSOR (Co = 10 mg/L)
I
6
t
2
o ~~..... 20
70
: io
• '(- .......... 120 VNo
j 170
··· ·r~
t-..... 220
Figure 4.6. Measured (closed circles) and predicted breakthrough curves in Windsor soil column for three Cd pulses of Co = 10 mg L-1. The solid curve is prediction using equilibrium ion exchange.
chromatographic effect where the effluent concentrations exceed that of the input pulse solution (C/Co > 1). Selim et al. (1992) obtained equally good predictions for multiple input pulse applications with Co of 100 mg L- I . The assumption of equilibrium ion exchange reaction has been employed to describe sorption of heavy metals in soils by several investigators (Abd-Elfattah and Wada, 1981; Harmsen, 1977; Bittel and Miller, 1974; Selim et al., 1992; Hinz and Selim, 1994). In general, the affinity of heavy metals increase with decreasing heavy metal fraction on exchanger surfaces. Using an empirical selectivity coefficient, it was shown that Zn affinity increased up to two orders of magnitude for low Zn surface coverage in a Ca background solution (Abd-Elfattah and Wada, 1981). The Rothmund-Kornfeld approach incorporates variable selectivity based on the amount adsorbed (s) or exchanger com-
102
Fate and Transport of Heavy Metals in the Vadose Zone
position. The approach is empirical and provides a simple equation that incorporated the characteristic shape of binary exchange isotherms as a function of Sj as well as the total solution concentration in solution (Cr). Harmsen (1977) and Bond and Phillips (1990) expressed the Rothmund-Kornfeld as
(Sir:
= R
(sY'
K. 'J
J
[(Cir:]n (C.)Vl
(18)
J
where n is a dimensionless empirical parameter associated with the ion pair i-j, and R~j is the Rothmund-Kornfeld selectivity coefficient. The above equation is best known as a simple form of the Freundlich equation which applies to ion exchange processes. As pointed out by Harmsen (1977), the Freundlich equation may be considered as an approximation of the Rothmund-Kornfeld equation valid for Sj « Sj and Cj « Cj' where
s. ,
=
RK.(c.)n 'J'
(19)
Based on best-fit predictions, Hinz and Selim (1994) showed strong Zn and Cd affinity (compared to Ca) at low Zn and Cd concentrations based on parameter estimates of ionexchange isotherms using the Rothmund-Kornfeld approach.
Kinetic Ion Exchange Recently, Sparks (1989) compiled an extensive list of cations (and anions) that exhibited kinetic ion exchange behavior in soils; e.g., AI, NH 4, K, and several heavy metal cations. According to Ogwada and Sparks (1986a,b,c), kinetic ion exchange behavior was probably due to mass transfer (or diffusion) and chemical kinetic processes. The proposed approach was analogous to mass transfer or diffusion between the solid and solution phase such that, for ion species i,
as· at
-' =
0
a(s· - s,.)
'
(20)
where at any time t, the symbol Sj denotes the amount sorbed, where SjG is the amount sorbed at equilibrium, and a is an apparent rate coefficient Cd-I) for the kinetic-type sites. In Equation 17, the amount sorbed at equilibrium SjG is calculated using the respective isotherm relations similar to Equation 14. Expressions similar to Equation 17 have been used to describe mass transfer between mobile and immobile water as well as chemical kinetics (Parker and Jardine, 1986). For large a, Sj approaches SjG in a relatively short time and equilibrium is rapidly achieved, whereas for small a, kinetic behavior should be dominant for extended period of time.
Case Study We investigated Cu retention by monitoring its concentration in the soil solution with time for a wide range of input concentrations. Cu transport was studied using miscible
Modeling the Kinetics of Heavy Metals Reactivity in Soils
103
displacement methods. Two different soils were studied. Cecil soil was chosen as a benchmark and McLaren soil was obtained from a site near an abandoned Cu mine on Fisher Mountain, Montana. Isotherms were measured using standard ion exchange methods where ten-gram samples of soil were equilibrated with Cu and Mg at varying ratios. The samples were shaken for 24 h on a reciprocal shaker with 30 mL of various proportions of CUS04 and MgS04 solutions. The solutions were then centrifuged and decanted. For the first two steps (24 h each step) total concentration was 0.5 N followed by four time steps at 0.01 N. Triplicate samples were used for each solution ratio. Adsorbed cations were removed by three extractions with 1 N NaOAc and corrections were made for the entrained solution. Cu and Mg in solution and extractant solution were analyzed by ICP. Based on these Cu-Mg exchange isotherm experiments, selectivity for Cecil and McLaren soils were obtained. The transport of Cu in the Cecil and McLaren soils were investigated using the miscible displacement technique. Plexiglas columns (6.4 cm i.d. X 10 cm) were uniformly packed with air-dry soil and were slowly water-saturated. Upon 1 saturation, the fluxes were adjusted to the desired flow rates. A Cu pulse of 100 mg Lwas introduced into each column after it was totally saturated with 0.01 N MgS04 or Mg(CI04h as the background solution. Perchlorate as the background solution was used to minimize ion pair formation. The Cu pulse was eluted subsequently with 0.005 M MgS04 solution. The ionic strength was maintained nearly constant throughout the experiment. In other soil column experiments, similar conditions were used except that no background solution was used in the Cu pulse input solution. This resulted in a condition of variable total ionic concentrations or ionic strength during input pulse application and the subsequent leaching solution. A fraction collector was used to collect column effluent. Figures 4.7 and 4.8 show the effect of total concentration or ionic strength of the input pulse solution on Cu breakthrough results. When Cu was introduced in Mg background solution with minimum change in ionic strength, Cu breakthrough curves (BTCs) appear symmetrical in shape with considerable tailing and peak concentration of 40 mg L- 1• Mg BTC shows an initial increase in concentration due to slight increase in ionic strength followed by a continued decrease during leaching. When Cu was introduced in the absence of a background solution, the total concentration considerably decreased from 0.005 to 0.0015 M. As shown in Figure 4.8, Cu BTC showed a sharp increase in concentration due to chromatographic (or snowplow) effect (Selim et al., 1992). The Cu peak concentration was 94 mg L- 1 and the corresponding Mg concentration in the effluent decreased due to depletion of Mg during the introduction of Cu. Mg concentration increased, thereafter, to a steady state level during subsequent leaching, however. This snowplow effect is a strong indication of competitive ion exchange between Mg and Cu cations. The amount of Cu recovered in the effluent was 53% of that applied in the presence of MgS04 as the background solution, whereas only 38% was recovered when no background solution was used. For McLaren soil (Figure 4.9), the snowplow effect was pronounced as shown in Figures 4.7 and 4.8 due to changes in total concentration of input solutions with a recovery of 60% of that applied. Therefore, miscible displacement or transport experiments indicated that there was strong ion exchange between Cu and Mg cations which was also affected by the counter ion used. Effluent peak concentrations were 3-5 fold that of the input Cd pulse, which is indicative of pronounced chromatographic effect.
104
Fate and Transport of Heavy Metals in the Vadose Zone ----- ----
--
-
-~-
---~-.----
-----
--
200
Cecil Soil
...J
Column I Sulfate
Ol
-
E C 0
150
:;::
...c:
Mg
~ ....
Q)
-~----~--'-
100
0
c: 0
...c:
U
Q)
50
:J
:::: w
Cu 0 0
10
20
30
40
50
60
Pore Volume Figure 4.7. BTCs of Cu and Mg for Cecil soil (Column I).
--
200
c
150
...J
en
-E
Column II Sulfate
Cecil Soil
0 +:i
-
Mg
...cu C
(I)
u
100
c
0
()
c
50
(I)
::::I
tt:
w
Cu
0 0
10
20
30
40
50
60
Pore Volume Figure 4.8. BTCs of Cu and Mg for Cecil soil (Column /I).
In summary, we presented an overview of several models which are used for the description of the retention of heavy metals in soils. Single reactions models were classified into equilibrium and kinetic types. A general purpose multireaction kinetic and transport model was also presented. Major features of multi reaction models are that they are fiexible and are not restricted by the number of solute species present in the soil system
Modeling the Kinetics of Heavy Metals Reactivity in Soils
--
200
c o
150
. .J
McLaren Soil
C)
-E
105
Column V Sulfate Mg
~
~
,..
+"
C
Q) (.)
100
C
o
•
(.)
+"
C
50
Q)
::::J
e W
Cu 0 0
10
20
30
40
50
60
Pore Volume Figure 4.9. BTCs of cu and Mg for McLaren soil (Column V).
nor the governing retention reaction mechanisms. This includes reversible and irreversible reactions of the linear and nonlinear kinetic types. Moreover, these models can incorporate concurrent as well as consecutive type retention reactions which may be equilibrium or kinetic in nature. Ion exchange mechanisms of the instantaneous and kinetic types were also presented. Case studies of Cd and Cu isotherms as well as transport in soil columns provided illustrations of model applications.
REFERENCES Abd-Elfattah, A. and K. Wada. Adsorption oflead, copper. zinc. cobalt. and calcium by soils that differ in cation-exchange materials. J. SoiL Sci. 32. pp. 271-283. 1981. Amacher. M.C.• H.M. Selim. and I.K. Iskandar. Kinetics of chromium (VI) and cadmium retention in soils; A nonlinear multireaction model. Soil Sci. Soc. Am. J. 52. pp. 398-408. 1988. Amacher. M.C.• J. Kotuby-Amacher. H.M. Selim. and I.K. Iskandar. Retention and release of metals by soils-evaluation of several models. Geooerma. 38. pp. 131-154. 1986. Aringhieri. R.. P. Carrai. and G. Petruzzelli. Kinetics of Cu and Cd adsorption by and Italian soil. SoiL Sci. 139. pp. 197-204. 1985. Barrow. N.J. and T.C. Shaw. Effects of solution and vigor of shaking on the rate of phosphate adsorption by soil. J. Soil Sci. 30. pp. 67-76. 1979. BitteL J.E. and R.J. Miller. Lead. cadmium and calcium selectivity coefficients of a montmorillonite. illite and kaolinite. J. Environ. QuaL. 3. pp. 250-253. 1974. Harmsen. K. Behavwr of Heavy Metau in Soiu. Centre for Agriculture Publishing and Documentation. Wageningen. The Netherlands. 1977. Hinz. C. and H.M. Selim. Transport of Zn and Cd in soils: Experimental evidence and modelling approaches. SoiL Sci. Soc. Am. J. 58. pp. 1316-1327. 1994.
106
Fate and Transport of Heavy Metals in the Vadose Zone
Jardine, P.M., J .C. Parker, and L.W. Zelazny. Kinetics and mechanisms of aluminum adsorption on kaolinite using a two-site nonequilibrium transport model. Soil Sci. Soc. Am. J. 49, pp. 867873, 1985. Lapidus, L. and N.L. Amundson. Mathematics for adsorption in beds. VI. The effect of longitudinal diffusion in ion exchange and chromatographic column. J. PhYd. Chem. 56, pp. 984-988, 1952. Ogwada, RA. and D.L. Sparks. A critical evaluation on the use of kinetics for determining thermodynamics of ion exchange in soils. Soil Sci. Soc. Am. J. 50, pp. 300-305, 1986a. Ogwada, R.A. and D.L. Sparks. Kinetics of ion exchange on clay minerals and soil. I. Evaluation of methods. Soil Sci. Soc. Am. J. 50, pp. 1158-1162, 1986b. Ogwada, RA. and D.L. Sparks. Kinetics of ion exchange on clay minerals and soil. II. Elucidation of rate-limiting steps. Soil Set: Soc. Am. J. 50, pp. 1162-1166, 1986c. Parker, J.C. and P.M. Jardine. Effect of heterogeneous adsorption behavior on ion transport. Water Ruour. Ru. 22, pp. 1334-1340, 1986. Rasmuson, A. and I. Neretienks. Migration of radionuclides in fissured rock. The influence of micropore diffusion and longitudinal dispersion. J. GeopPyd. Ru. 86, pp. 3749-3758, 1981. Rubin, J. and RV. James. Dispersion-affected transport of reacting solution in saturated porous media, Galerkin method applied to equilibrium-controlled exchange in unidirectional steady water flow. Water Ruour. Ru. 9, pp. 1332-1356, 1973. Selim, H.M. Prediction of contaminant retention and transport in soils using kinetic multireaction models. Environ. HeaLth Perdpec. 83, pp. 69-75, 1989. Selim, H.M. Modeling the transport and retention of inorganics in soils. Adv. Agron., 47, pp. 331384, 1992. Selim, H.M. and M.C. Amacher. A second-order kinetic approach for modeling solute retention and transport in soils. Water Ruourc. Ru. 24, pp. 2061-2075, 1988. Selim, H.M., B. Buchter, C. Hinz, and L. Ma. Modeling the transport and retention of cadmium in soil: Multireaction and multicomponent approaches. Soil Sci. Soc. Am. J. 56, pp. 1004-1015, 1992. Selim, H.M., J.M. Davidson, and R.S. Mansell. Evaluation of a 2-site Adsorption-Desorption Model for Describing Solute Transport in Soils. Proc. Summer Computer Simulation Con! Wadhington, DC, 12-I4Ju!y, 1976, La Jo!la, CA, Simulation Councils Inc., La Jolla, CA. 1976, pp. 444448. Selim, H.M., M.C. Amacher, and I.K. Iskandar. Modeling the Transport of Heavy Metals in Soils. CRREL Monograph 90-2, U.S. Army Corps of Engineers, 1990, p. 158. Sparks, D.L. Kinet0 of Soil Chemica! ProceddU. Academic Press, San Diego, CA, 1989. Sposito, G. The Surface Chemidtry of Soil!. Oxford University Press, New York, 1984. Theis, T.L., R Iyer, and L.W. Kaul. Kinetic studies of cadmium and ferricyanide adsorption on goethite. Environ. Sci. Techno/. 22, pp. 1032-1017, 1988. van der Zee, S.E.A. T.M. and W.H. van Riemsdijk. Sorption kinetics and transport of phosphate in a sandy soil. Geoderma 38, pp. 293-309, 1986.
CHAPTER S
Copper Retention as Affected by Complex Formation with· Tartaric and Fulvic Acids Alexander A. Ponizovsky, T.A. Studenikina, and E.V. Mironenko
INTRODUCTION Soil solutions from humic horizons and surface waters contain inorganic and organic compounds capable of forming complexes with heavy metals (HM). Complexation can influence mobility of metals and their availability to living organisms. Many experimental studies of the HM retention in soils have been conducted without taking into account complexing ligands. Copper is a trace metal commonly present in many surface waters and soil solutions in detectable amounts. The influence of complex formation on copper behavior in soils and waters is the subject of this chapter.
Copper(lI) Retention by Soils, Oxides, and Clays Copper retention in soils is a rather complex process even without any substances able to associate with the metal ions in solutions. McLaren and Crawford (1973a, 1973b) suggested separating the following pools of natively available copper in soils: • • • • •
present in soil solution and exchangeable Cu (extractable by 0.05 M CaCI2) specifically bound by inorganic sites (soluble in 2.5% acetic acid) specifically bound with organic matter (soluble in K-pyrophosphate) occluded by free oxides (soluble in acid solution of ammonium oxalate) residual (bound in the lattices of minerals, soluble in HF).
These pools are not strictly related to types of chemical bounds. The total amount of copper retained in soil may not be in equilibrium with the Cu(II) in the soil solution, and may not be exchangeable and extractable with salt solutions. Zhang and Sparks (1996) reported that the quantity of Cu extracted by Na2EDTA from the montmorillonite used 10"'7
108
Fate and Transport of Heavy Metals in the Vadose Zone
in Cu-retention experiments was only 1.0 to 1.2% of the Cu-exchange capacity. Copper may be adsorbed by soil colloids in amounts in excess of their conventional exchange capacities, assuming adsorption as the Cu2+ ion (e.g., McLaren and Crawford, 1973a, 1973b). This phenomenon has been termed "specific adsorption" and is shown to occur on clays (Bingham et aI., 1964), organic matter (DeMumbrum and Jackson, 1956), and free oxides (Grimme, 1968). Copper sorption has been actively studied on soil components, such as Mo, Fe, AI oxides, or bentonite. Murray (1975) studied Cu 2+ retention on manganese oxide and found that adsorption of one Cu 2+ion leads to the release of 1.2 H+ ions. An inequivalence of proton released to retained metal was also found for Ca2+, Mg2+, Zn 2+and some other metals. This was explained by the replacement of a proton on the surface by a divalent 2 metal ion. Shindler et al. (1976) found that the number of protons released per Cu + adsorbed increases from 1 to 2 with the enhancement of pH. They attributed such a phenomenon to the retention of [CuOHr complexes instead of free Cu 2+ ions. Infrared spectroscopic studies by Parfitt and Russel (1977) showed that sorption of Cu 2 + on goethite at pH 7 did not lead to direct interaction between copper ions and surface OH groups. According to Rodda et al. (1996), experimental data on adsorption on goethite can be fitted by a model in which monomeric Cu(OH)+ and dimeric CU2(OH)22+ complexes compete for surface sites, with the dimer more strongly adsorbed to the surface. The number of protons released per Cu(II) ion adsorbed at pH 5.0 increased from 1.76 at 25 D C to 1.92 at 70 D C. Bower and Truog (1941) reported enhanced retention of Cu(II) by bentonite from chloride background medium. This was explained by the CuCI+ and [CuOHr ion pair involvement in the ion exchange and/or copper hydroxide precipitation. Sposito et al. (1981) also found that copper sorption by Na-bentonite in chloride solutions exceeds that in perchlorate solution. They suggested that monovalent CuCI+ complex was responsible for this effect, though this ion pair is known to be unstable and its concentration in solutions low. In these experiments, the solution pH was not stabilized and decreased with the increase in exchangeable copper content. Such results were not confirmed in the study of Zhang and Sparks (1996). They obtained values of Cu-exchange capacity for Na-montmorillonite similar for CI-, (CI0 4)-, (N0 3)-, and (S04)2- background at 0.25 M Cu(II) solution concentration and pH from 4.31 to 4.54. This concentration level was rather high, and they suggested that some other relations could be observed at a much lower Cu concentration, when specific adsorption occurs mainly at nonexchangeable sites on clay edges. H-montmorillonite was shown to retain copper from the chloride and acetate solutions at pHd in amounts equal to the standard cation exchange capacity (CEC) (Bingham et aI., 1964). In acetate solutions at pH>3, the amount of Cu 2+ retained exceeded the CEC. Acetate was sorbed by clay, but the sorbed amount was not equal to the excess of Cu2+ retained. Soils selectively adsorb Cu2+. Harter (1992) studied this phenomenon and revealed that the amount of Cu 2+ adsorbed exceeded that of the Ca2+ desorbed. He hypothesized that this discrepancy could be attributed either to nonexchange sorption processes or to the adsorpt~on of both free ITletal ~ons and their charged complexes, e.g., [CuOHr which results in maximum metal retention greater than CEC. McLaren and Crawford (1973b) suggested that specific and nonspecific adsorption of copper by soils and montmorillo-
Copper Retention as Affected by Complex Formation
109
nite take place simultaneously and independently. The specific adsorption exceeds nonspecific adsorption at low copper concentrations, while the relation can be opposite as the concentration increases. They assumed that nonspecific adsorption isotherms can be described by the Vanselow equation, whereas the Langmuir adsorption equation was found suitable for the "specific adsorption" isotherms within solution concentration ranges of 1-5 mg mL- 1 (0.015 to 0.075 mmol L- 1). Sposito et al. (1981) described sodium-copper exchange isotherm on bentonite in the perchlorate solution by the Vanselow equation and obtained the selectivity coefficient independent of exchangeable Na/Cu ratio. The same equation was applied to describe Cu-Na exchange on montmorillonite at pH from 5.17 to 6.50 and total concentration of metal ions in the solution of 0.02 mole L- 1 (Zhang and Sparks, 1996).
Solution Complex Formation and Cu(1I) Adsorption Copper(II) forms complexes with a number of anions commonly found in natural waters (Yatsimirsky and Vasilyev, 1959). In water solutions free of other ligands, Cu(II) is present mainly as a complex [Cu(H 2 0)6J 2+, a Lewis acid with pK=6.8, and [Cu(OH) (H 20hr, a Lewis base. Simulations of the Cu(II) speciation in fresh waters indicate a possible presence of a variety of complexes with different charges, depending on the solution pH (Vuceta and Morgan, 1978). The discrimination between copper complexes in natural waters remains a challenging analytical problem. A number of direct electrochemical techniques have been used in an attempt to distinguish between chemical forms of metals. Some of the existing methods are based on the separation of the different species. Dialysis, ultrafiltration, and centrifugation have been used to separate "free" metals in natural waters from metals associated with colloid material. Extraction with chloroform and chelating with ion exchange resins have been applied to separate "bound" copper from the free ions. Several studies were carried out on sea water (e.g., Florence and Batley, 1977) and sewage effluents. In sewage effluents Bender et al. (1970) using gel filtration chromatography (Sephadex G-50) found no free Cu 2+ ions. About 13% of the copper in the sample was associated with molecular weights of 104 or greater, while the remaining molecular weights ranged from 500 to 1000. A ligand that can be sorbed on a surface can enhance or inhibit retention of Cu (Davies and Leckie, 1978). Negatively charged Cu 2+ complexes with organic acids can be only slightly sorbed on the surface of layer silicate minerals (Bloomfield et al., 1976). However, glutamic acid increases retention of Cu(II), the influence of salicylic acid is negligible, and picolinic acids inhibits copper retention by hydrous oxides (Davies and Leckie, 1978). Pampura (1993) found that from 1 to 3 mmol L- 1 citric, malic, and salicylic acids decreased the Cu(II) sorption by soil. Such an effect can be explained by the formation of complex ions of various charge and with different abilities to be adsorbed on charged surfaces. Janvion et al. (1995) found that the concentration of acetate affected the distribution coefficient of Cu(II) between the acetate buffer in the mobile and stationary phases having both cation- and anion-exchange groups. This was explained by suggesting that copper retention involves not only Cu 2+ cation exchange, but also anion exchange of copper anionic complexes, e.g., with 2,3-pyrazinedicarboxilic acid (2,3-PDCA), [Cu(PDCA)2f-.
ilO
Fate and Transport of Heavy Metals in the Vadose Zone
Elliot and Huang (1979) reported on a survey of a number of studies on the influence of several chelating agents on the adsorption of various metal ions by some oxides. They presented results on adsorption of Cu(II) complexes with glycine, nitrilotriacetic, and aspartic acids on y-Al203. These authors also studied adsorption of metal-amino acid complexes by y-Al203' which is a relatively hydrophilic and hydrophobic activated carbon, and adsorption of Cu(II) complexes with various ligands by several alumosilicates with different surface charges and ion-exchange properties (Elliot and Huang, 1980, 1981). Their results indicated that several adsorption mechanisms were affecting the retention. Besides hydrophylic complexes, hydrophobic complexes probably could be sorbed on the hydrophobic parts of particle surfaces. This was proved by Baffi et al. (1994), who studied Cu(II) adsorption on inorganic fractions of marine sediments in the presence of aminoacids. Stumm et al. (1976) assumed that the interaction of the surface functional groups with the metal ions to be similar with the complexation with ligands in the solution. That is probably the reason why the modification of the surface of clay minerals, e.g., by polyphosphates, enhances their ability to sorb trace metals (Klimova and Tarasevitch, 1992). Cu (II) adsorption by hydrous oxides was found to decrease in the presence of a complexing ligand (citric acid and EDT A) in a manner that suggests competition between the ligand and oxide surface for complexation of the metal ion (Davies and Leckie, 1978). They studied the influence of several complexing ligands, capable of adsorbing on the surface of amorphous iron oxide on the trace metal uptake by oxide surface. The adsorption edge for Cu(II) was found in the pH range 5 to 6. Adsorption of metal ions on various oxide surfaces increased abruptly in this pH range, where hydrolysis products became a significant fraction of the dissolved metal. The authors suggested that adsorbed CuOH+ was the main form of copper(II) on the surface. The location of the adsorption edge for Cu(II) was influenced by ionic strength and total Cu(II) concentration. Cu(II) binding by amorphous Fe(OH)3 was not affected by the addition of salicylic and protocatecholic acids in the experiments of Davies and Leckie (1978). Adsorption was enhanced by glutamic acid and 2,3-PDCA. Picolinic acid effectively prevented copper(II) removal from the solution. The authors attributed these effects to the adsorption of glutamic acid and 2,3-PDCA on Fe(OHh This adsorption should increase the binding strength of the surface for Cu(II). Adsorbed picolinate could not function as a complexing ligand for metals due to the fact that coordinating groups were bound with the surface and unavailable to the metal ion.
Complexes of CI.I(II) with Fulvic Acids Fulvic acids (FA) are important components of surface waters and soil solutions. The complexing ability of FA is well known but rather difficult to be described quantitatively since FA are high molecular weight substances of complex composition with a great number of functional groups. Complex formation between Cu(II) and FA was studied by different techniques. A set of the methods was applied to calculate stability constants of soil organic matter with Cu(II) (see, e.g., Young et aI., 1982; Hirose et al., 1982; Perdue and Lytle, 1983; Ephraim and Marinsky, 1990; and a review by Bizri et aI., 1984). These methods provide some conditional values, related to definite solution ionic strength, copper and FA concentrations, etc. For example, stability constants calculated
Copper Retention as Affected by Complex Formation
111
by Ruzie and Scatchard plots, obtained by ion selective electrodes and potentiometric stripping analysis, decreased continuously with the degree of site occupation (Soares and Vasconcelos, 1994).
Influence of FA and Humic Acids (HA) on the Retention of Cu(lI) by Solid Phases Humic substances can be sorbed on mineral surfaces containing hydroxylated Al, Fe, or Mn sites (Parfitt et al., 1977; Murphy et al., 1990), thereby changing the mineral surface properties. Surface bound humic substances increase the adsorption of some metal cations on single mineral solids by contributing additional potentially high affinity complexation sites. For copper which forms strong complexes with humic substances, these substances enhance metal binding to Fe and AI oxides at lower pH (e.g., < 6) by cosorption and decrease metal ion retention at higher pH by formation of nonadsorbing
aqueous complexes (Tipping et al" 1983; Allard et al., 1989). For other combinations of metals and sorbents, mineral-bound humic substances exhibit variable effects on metal binding (Xu et al., 1989). It depends on the complexation ability of humic substance for the metaL the distribution of the humic substance between the solution and solid, and the variation of both of them with respect to pH and ionic strength (Zachara et al., 1994). It has not been estimated whether the mineral-bound humic substance exhibits comparable complexation properties for metals to that in solution. Both Davies (1984) and Laxen (1985) concluded that metal-humic substance interactions were stronger on surfaces than in solution. Tipping et al. (1983) suggested that when humic substances bind to goethite, new high affinity metal complexation sites are formed. Dissolved FA decreased Cu(II) retention by montmorillonite, though the effect was less than the influence of citrate and EDTA (Vuceta and Morgan, 1978). Clay in Cuform adsorbs FA and HA (Theng and ScharpenseL 1975). The authors attributed this phenomenon to the precipitation of slightly soluble copper salts of FA and HA due to release of exchangeable Cu(II). Chakrabarti et al. (1984) mentioned that dissolved FA decrease the velocity of the dissolved copper retention by the solid ion-exchange phase. This was perhaps caused by the low velocity of dissociation of the Cu-FA complex. Wershaw et al. (1983) suggested that Cu(II), forming charge transfer complexes with F A, in some conditions coordinates not one, but two or more FA molecules, and this leads to the aggregation of molecules. Such a point of view coincides with the existing conception of humic substances aggregation by metal ions (Gamble et al., 1984). It was found that, while extracting bound copper from bentonite, the amount of Cu(II) released was dependent on the concentration of ligand and the stability constant of the metal complex. The ratio of metal release to complexing sites decreased in the order: EDTA> humic acid> tannic acid (Guy and Chakrabarti, 1976).
COPPER RETENTION BY SOil (A CASE STUDY) Kinetics of Cu(lI) Retention A case study of the Cu(II) retention was carried out on a silty clay Typic Haplustoll soil (leached chernozem) sampled in the Tula region, Russia, from the A horizon (0-20 cm depth). Soil organic C content was 27.5 g kg-I, and exchangeable cation contents,
112
Fate and Transport of Heavy Metals in the Vadose Zone
determined by modified Pfeffer method (Khitrov, 1982) (extraction by 0.1 M NH4 CI solution in 70% alcohol), were 27.4, 1.7,0.13, and 0.70 cmole kg-I ofCa2+, Mg2+, Na+, and K+, respectively, with total cation exchange capacity 29.9 cmole kg-I. pH measured in water extract with 1: 1 soil: water ratio was 5.77. Each of these values is an average of three subsamples. The soil sample was air dried and ground to pass through a 2-mm sieve. To remove carbonates that can be present in the leached chernozem in trace amounts, the sample was leached with 0.1 M HCI until the leachate pH was about 3. Then sediment was washed out with water to remove the excess HCl. Then sample was saturated with Ca2+ by treatments with 0.1 M CaCI2, to diminish to a trace level an exchangeable H+ content, estimated by 1 M KCI extraction with subsequent KOH titration. To remove extra CaCl 2 the sediment was washed with water, until the final Ca2+ concentration in leachates became 2.5 mM (further decrease leads to peptization of the sediment). Then, the solution was decanted, the sample was air dried, ground in a mortar and sieved to pass a 1-mm mesh screen. Soil samples were placed into the flasks and suspended in 3 mM Ca(N03)2 solution as a background. Cu(N03h solution was added and pH value was adjusted by adding HN0 3 or KOH. The flasks were shaken at 25±1 °C for 90 days. The supernatant solutions were decanted, filtered through 0.2 pm membrane filters and copper concentration was measured with atomic adsorption spectroscopy (AAS). The amount retained was taken as the difference between the amount added and the amount recovered in the equilibrium solution. Equilibrium of Cu(n) retention by soil was found to be obtained for the period from 4 to 24 hours (Figure 5.1). Sorption velocity was highest at pH 6 (about 95% of the maximal copper retention was observed already in 1 hour) and lowest at pH 4. Mter 24 h no increase in soil copper content was observed.
Cu(1I) Retention Isotherms and Cation Balance Soil samples were suspended in 3 mM Ca(N0 3h Then some amounts of the 0.1 1\ 1 CU(N0 3)2 solution were added, and pH value was adjusted, adding HN0 3 or KOB The isotherms were obtained for pH 4,5, and 6. The flasks were shaken at 25±1 °C for I day. In some intervals the pH value was corrected by titration with HN0 3 or KOH. The suspensions were centrifuged and supernatant solutions were analyzed. Retained copper was calculated as above. Amount of H+ displaced from soil by Cu 2+ was evaluatec from the amount of HN0 3 or KOH used to adjust the pH value. Isotherms of Cu(n) displacement by Ca2+ from the contaminated soil were obtainec by treatment of the soil residue in the flasks by 3 mM Ca(N03h with the pH adjustmen: as mentioned above. The suspensions were shaken for 1 day, centrifuged, and analyzec Cu(n) retention isotherms at different pH are presented on Figure 5.2. Increase i:pH leads to the enhanced copper sorption. At pH 6 the sorption is much higher, and the' shape of the isotherm is different from those at pH 4 and 5. Cu(n) sorption in all cases was accompanied by the release of both Ca2+ and H+ ions from soil to solution. To maintain pH level, KOH was added to the suspension and Kwas partly retained by soil.
Copper Retention as Affected by Complex Formation 60
---
50
""'"
~-
40
r
";
CI
-.
r-
-.
..w: "5
______ pH = 4 _____ pH = 5
E 30 E
I--
--A-- pH = 6
:, 0
en
20
10
o
o
20
40
60
80
100
120
140
160
180
time, h
Figure 5.1. Kinetics of Cu(11) retention by soil.
350 _ pH=4, Cu retention ___ pH=5, Cu retention ~ pH=6, Cu retention -cr- pH=4, Cu displacement - 0 - pH=5, Cu displacement
300
250
~ a
200
:,
150
E E
0
en
100
50
0 0
1
2
3
meu' mM
Figure 5.2. Isotherms of Cu(1I) retention by soil without ligands at pH 4, 5, and 6.
113
114
Fate and Transport of Heavy Metals in the Vadose Zone 250 ,---------------------------------------,
";"0
200
~
_I>
o E E
150
~ Q)
c:: ·iii
i!....
100
o
"5lVI
m
~
50
VI
c::
.2
B
0
80
100 120 140 160 180 200 220 240 260 280 300 Cu sorbed, mmolc kg·1
Figure 5.3. Cation balance at Cu(1I) retention, pH=5.
It was found that the relations of released and retained ions amounts, e.g., at pH 5 (Figure 5.3) in the studied range can be expressed as 2
Ca \el.
=
0.765 CUsorb.
H+ reI.
=
0.292 CUsorb.
K\orb.
=
0.0332 CUsorb.
Here Ca2+re l.' H+rel.' K\orb., and CUsorb. are amounts of Ca2+, H+ released and K+, Cu(II) retained (mole kg- 1 soil), respectively. Thus total cation balance for copper retention is
Taking into account measurements errors, it could be concluded that amount of iom released is equal to the amount retained. Balance between Cu(II) retained and Ca2+and H+ released at pH 4, 5, and 6 is presented in Table 5.1. Impact of K+ retention or displacement on cation balance is rather small. In all studied cases Cu(II) displaced from soil both Ca2+ and H+. At pH 4 and 6 (Ca2+re l. + H+rel)/Cu2+sorb. ratio was 0.89 and 0.98, respectively, i.e., not so much different from the value of 1.05 at pH 5. Variation of thi;; ratio, caused probably by the measurements errors, doesn't allow rejecting the hypothesis on the exchange equivalence. The "reverse" isotherm of copper displacement by Ca2+ at pH 4 does not diverge significantly from the direct one (Figure 5.2). The increased distinction at pH 5 can be
Copper Retention as Affected by Complex Formation
115
Table 5.1. Cation Balance at Cu(I1) Retention by Soil (mole per mOle Cu(I1)) NL - no Ligands, (1) Cu(1I) Retention, (2) - Cu(I1) Displacement pH=4 (2)
TA (1)
FA (1 )
NL
(1)
(1 )
pH=5 TA (1)
0.64 0.25 0.89
0.64 0.32 0.96
0.54 0.29 0.83
0.63 0.18 0.81
0.76 0.29 1.05
0.43 0.34 0.77
NL
Ca 2 + H+ (Ca 2 ++W)
pH=6 FA (1)
NL (1 )
TA (1 )
FA (1)
0.39 0.19 0.58
0.21 0.77 0.98
0.16 0.76 0.92
0.19 0.77 0.96
attributed to (i) displacement of Cu(II) mainly by Ca2+ and not by H+ ions; (ii) the differences in the solution Ca2+ concentrations in direct and reverse runs; (iii) slower velocity of the reverse process comparatively with the direct one. Thus the Cu2+_(Ca2++ H+) exchange seems to be a reversible process, though it could be difficult to displace all the copper retained due to the higher soil selectivity with respect to Cu (II).
Evaluation of Na 2 EDTA Ability to Extract Retained Copper To extract retained copper, contaminated soil was 3 times extracted with portions of 0.02 mole L- 1 Na2EDTA. To prevent peptization and to improve copper displacement, lO mM Ca(N03)2 was added to the Na2EDT A solution. As it is shown on Figure 5.4, copper was quantitatively revealed from the contaminated soil by 0.02 mole L -I Na2EDTA extraction. Though CuEDTA is a very stable complex (pK = 18.9) (Sillen and Martell, 1970), first treatment revealed only about 80% of the total copper retained amount, and 3 subsequent extractions by Na2EDTA solution were necessary for complete copper displacement. Ca(N03)2 + Na2EDTA solution was shown to extract Cu(II) more efficiently and did not peptisize soil as it was observed for Na2EDTA.
Effect of Tartrate and Fulvic Acid on Cu(lI) Retention Isotherms Tartaric acid (TA) and FA were taken as samples of soluble organic compounds of soil solutions. Tartaric acid (HOOC-CH(OH)-CH(OH)-COOH) is a low-molecularweight dicarbonic hydroxy acid found in root exudates of many plant species (Riviere, 1960; Smith, 1969; Ivanov, 1973; Hale and Moore, 1978), and in leachates from decomposed leaf litter (Nykvist, 1963). The anion derived from dissociation of tartaric acid forms relatively stable complexes with many metals. Therefore, the presence of tartrates may affect the mobility of heavy metals in the root zone and their uptake by plants. Fulvic acids are probably the most common high molecular weight compounds present in soil solutions. Being polyelectrolytes with a variety of functional groups, they are able to form chelate complexes with HM. Baker (1973) has proposed that the metal transport mechanism in soils involves mainly complexes with fulvic and humic acids. For our case study, FA was extracted from the soil samples by 0.1 M KOH with the subsequent sorption on Amberlite XAD-8, as it was recommended by IHSS (Kuwatsuka et aI., 1992), and was purified by dialysis with dialysis membrane "Film 100"
1 16
Fate and Transport of Heavy Metals in the Vadose Zone 150 - - - , - - - - - - - - - - - - - - - - - - - - ; ; ,
CU,eleased =4.29+0,921 *Cu'otoIned
";'
(r=0.989)
120
Cl
~
'0
E E 'ti' Q) tJ)
90
~
~
... Q)
c.. c..
8
60
30~----.----_.----_,---~
30
60
90
120
150
Cu sorbed, mmol kg-1
Figure 5.4. Efficiency of the retained Cu(lI) extraction from the soil by 3 treatments with 0.02 male L-1 Na 2 EDTA.
CPolymersintez," Vladimir, Russia). The solution was freed of salts by passing through cation-exchange resin KU-2 in H-form. Then, the sample was freeze dried. The stock solution used for the experiments contained 2 g L -1 dry FA. Tartrate added forms a number of complexes with Cu(II) (Ponizovsky et aI., 1997) decreasing its activity in the solution. That can be illustrated by the data obtained by adding tartaric acid to 0.001 M CU(N03)2 solution, measuring Cu 2+ activity by ion selective electrode, and calculating the activity coefficient as a ratio of activity to concentration (Figure 5.5). To evaluate the influence of TA and FA on Cu(II) retention, tartaric acid, or FA solution was added to the soil suspension in 3 mM Ca(N03h Then some amounts of 0.1 M CU(N03)2 solution were added, pH value was adjusted adding HN0 3 or KOH, and the suspensions were shaken at 25± 1°C for 1 day. In some intervals the pH value was corrected by titration with HN0 3 or KOH. The suspensions were centrifuged and supernatant solutions were analyzed. Retained copper was calculated as above. In soil suspensions at copper concentration less than 0.5 mM, tartrate diminished Cu(II) retention by soil at all the pH studied (Figure 5.6 a, b, c). At higher copper concentration, the difference between the retention in containing tartrate and tartratefree suspensions is less, and at some Cu(II) level we observed even increase in retention, probably due to the precipitation of copper tartrate. Fulvic acid decreased copper retention at low Cu(II) concentration (Figure 5.6 a, b, c). That can be also attributed to complex formation. At higher Cu(II) concentration an increase in retention was observed, probably due to precipitation of Cu-fulvate.
Copper Retention as Affected by Complex Formation
117
0.9 . . . . , . . - - - - - - - - - - - - - - - - - - - ,
0.8
~ 0.7
·u
==CD 8
~
0.6
.~
U m
.t.
8
0.5
0.4
0.3 -;-----.----,----,-----,-----,--------1 0.0
0.2
0.4
0.8
1.0
1.2
Figure 5.5. Cu 2+ ion activity coefficient as a function of tartrate to Cu(lI) ratio in solution at mcu=l mM (calculated from the measured Cu 2 + activity).
100,--------------------,
80
~
(a)
60
"0
E E
en"a
40
-e-
without ligands ___ with tartrate added ----y- with FA added
20
O~-_.--,_-~-_.--,_-_r-~-~
0.0
0.2
0.4
0.6
Figure 5.6a. Copper retention isotherms at pH 4.
0.8
1.0
1.2
1.4
1.6
118
Fate and Transport of Heavy Metals in the Vadose Zone 160.-------------------------------------, 140 120 100
~
Cl
..>0::
(5
E E :. u
CJ)
80 60 40 20 O~----_r----_,----_.------r_----,_-----
0.0
0.4
0.2
0.6
0.8
1.0
1.2
Figure 5.6b. Copper retention isotherms at pH 5. 600.-------------------------------------,
500
400 ~
Cl
..>0::
(5
E E
300
:.u
CJ)
200
100
o ~---.----._--_r--_.----,_--_.--_.--~ 0.00
0.02
0.04
0.06
Figure 5.6(. Copper retention isotherms at pH 6.
0.08
0.10
0.12
0.14
0.16
Copper Retention as Affected by Complex Formation
119
As it can be seen from the cation balance (Table 5.1), in the presence of tartrate and FA, copper also replaced both exchangeable Ca2 + and H+, and within the experimental error the exchange seems to be equivalent.
Modeling of Cu(lI) Retention (Exchange) by Soil Copper(II) retention isotherms were described by Freundlich and Vanselow equations. The Freundlich equation was used for heavy metal retention by Van Riemsdijk and Van der Zee (1989), Selim (1992), etc., though this equation was derived not for ion exchange but for sorption on energetically nonhomogeneous surface, and does not take into account the pH dependency of the process. The authors use it as
(1) or
(2) Here Scu' mcu, and acu are amount of copper sorbed (mmol kg-I), Cu(II) concentration (mmol L- I), and activity, respectively; K I, K 2, b, and d are coefficients. The Vanselow approach was applied using the following equations: acu Scu _ k Sca - Cu-ca aCa
SCu
k
acu
-S2 = Cu-H -2H
aH
2<:fS cu + 2<:fSea + SH = Q Here Sea and SH are amounts of Ca2+and H+ adsorbed, and Q is some operational value of cation exchange capacity (mmol e kg-I). kCu-ea and kCu_H are Cu-Ca and Cu-H ion exchange selectivity coefficients. Q differs from CEC in the value of SH. Copper and hydrogen activities were measured by ISE, and aea calculated based on the solution composition using the ion association model described by Mironenko et al. (1996). This model takes into account association of inorganic cations both with inorganic anions and organic anions derived from the dissociation of tartaric or fulvic acids, and precipitation of slightly soluble salts. Freundlich plots of Cu(II) retention isotherms on logarithmic scale are really very close to linear at all the pH values studied (Figure 5.7). Coefficients values are presented in Table 5.2. It can be seen that both Equations 1 and 2 are approximately equally suitable to describe the isotherms. The slopes of the lines for pH 4 and 5 are similar, and the intercepts differ from one another. A pH-dependent Freundlich isotherm was used by several authors to describe adsorption of bivalent heavy metal in acid soils (Lexmond, 1980; Van Riemsdijk and Van der Zee, 1989)
120
'8 VJ ""-'
Cl
Fate and Transport of Heavy Metals in the Vadose Zone -3.0
-3.0 - , - - - - - - - - - - - - - - - ,
-3.5
-3.5
-4.0
-4.0
..Q
••
-4.5
A
pH=4 pH=5 pH=6
-4.5
-5.0 -+----.---..,----.----.---'
-6
-3
-5
-5.0 -+---..,----,----,----,----'
-2
-6
-3
-5
-2
Figure 5.7. Freundlich equation plot for copper retention isotherms: (a) Equation 1, and (b) Equation 2.
Table 5.2. Freundlich Equation Parameters for Cu(1I) Retention Isotherms; r - Correlation Coefficients pH
Ligand
logK
Equation 1 b
r
4
Nl
2.94 1.92 1.88 3.09 2.18 2.04 7.02
0.339 0.417 0.345 0.339 0.511 0.338 1.18
0.991 0.994 0.987 0.993 0.978 0.962 0.912
TA FA
5
Nl TA FA
6
Nl
log K
Equation 2 b
r
3.02 2.02 1.94 2.11 2.31 2.11 3.27
0.350 0.344 0.330 0.340 0.389 0.292 0.630
0.997 0.995 0.992 0.995 0.985 0.991 0.993
(3) where Q(M) is amount of heavy metal adsorbed (mol kg-I), (H) and (M) are the proton and heavy metal activity, respectively, and K3, m, and n are constants. Taking K2 as a function of pH, one can obtain the equation similar to Equation 3 to calculate Cu retention in some pH range, but the validity of this approximation will be limited, and related only to some level of Ca2 +and other cations concentrations in the solution. Such approximation could be useful for some applications, but it needs data for some more pH values. The Vanselow equation can be applied to describe Cu(II) retention if assuming that (i) retention is a reversible ion exchange process; (ii) Cu(II) displaces from soil both exchangeable H+ and Ca2 +; (iii) the exchange is equivalent, i.e., 1 mole Cu(II) displaces 1 mole of (H+ + Ca 2+). Vanselow equation parameters were evaluated based on the obtained isotherms (Table 5.3). The precision of the data does not enable rejecting the hypothesis that parameters values are independent on the pH. To describe copper
Copper Retention as Affected by Complex Formation
121
retention, this equation is probably the most universal, but it needs a lot of experimental data to estimate all the parameters, and some model to calculate the ion activities in the solutions. Isotherms of Cu(II) retention in the presence of tartrate and FA also can be fitted by the Freundlich equation (Table 5.2) but the obtained parameters probably are related only to the studied levels of Cu(II), tartrate or FA, Ca2 +, and other ion concentrations. Application of the Vanselow equation for description of Cu 2+retention in the presence of complex forming substances could be more efficient, but in solutions containing tartrate available models does not allow calculating precise values of Ca2+ and Cu2+ activities, as far as a number of forming complexes is rather large and their properties are not well known (Ponizovsky et al., 1997). So we could only evaluate the parameters of this equation (Table 5.3). That does not enable rejecting the hypothesis of their identity with the ones obtained for the exchange without ligands. The model designed by Mironenko et al. (1996) allows evaluating Cu 2+activity in the solutions containing FA, but the precision of Ca 2+ activity calculation is rather low. So the attempts to use the Vanselow equation to calculate Cu(II) retention by soil can be only semiquantitative (Table 5.3). Available data do not allow estimating whether the parameters of this equation are different for ion exchange with and without ligands.
SUMMARY Copper(II) retention in soil in all cases can be suggested as accompanied by simultaneous displacement of both Ca2+ and H+ ions. Estimated ratios of amounts Cu(II) retained and (Ca2 ++ H+) displaced at pH 4,5, and 6 both in the absence and in the presence of T A and FA were from 0.77 to 1.05 molc per molc. That doesn't enable rejecting the hypothesis of equivalent exchange. Thus "nonequivalent" retention of copper, e.g., on montmorillonite, mentioned by Harter (1992) and Sposito et al. (1981), could be caused by displacement by Cu 2+ of not only exchangeable Na+, but H+. Traditional methods used to estimate the ion exchange capacity and exchangeable cations contents in clays and soils do not enable evaluating the content of exchangeable H+. However, Cu 2+ ions can easily displace them, so the amount of copper retained should be higher than amounts of Ca2+ or Na+ released. This effect did not appear in the experiments of Bingham et al. (1964), who found that copper retention by H-montmorillonite in the concentration range from 10--5 to 10-3 M at pH<3 was equivalent to H+ displaced, probably because they observed Cu 2+-H+ and not Cu 2+-Na+ exchange. Decrease in Cu(II) activity due to complexation with tartraric and fulvic acids leads to diminishing of Cu(II) retention in soil at concentrations below 0.5 mM. Nevertheless, at higher copper concentration the retention in tartrate and FA containing suspensions becomes even higher than in the suspensions without these substances, probably due to the precipitation of copper tartrate and fulvate. The Freundlich equation can be applied for approximation of retention isotherms, but its parameters values can be valid only in a very narrow range of pH, cations concentrations in the solution, and available complex forming substances. The Vanselow equation is much more universal to describe Cu(II) retention in soil even in the presence of some complexes, though its abilities are restricted by the lack of the models to calculate the ion activities in solutions. Increased retention of Cu 2+ at el-
122
Fate and Transport of Heavy Metals in the Vadose Zone Table 5.3. Vanselow Equation Parameters for Cu 2+ - Ca 2+ - H+ Ion Exchange Isotherms
pH
4
5
Ligand NL TA FA NL TA
k 43.9 53.6 28.3
72.4 82.3
CU-H
4.83*10-6 6.23*10-6 4.60*10-5 5.03*10-8 1.94*10-7
k
Cu-Ca
0.996*10 4 1.15*104 9.44*104 1.27*104 5.24*10 4
evated pH can be caused by the shift of Cu 2+-H+ exchange equilibrium. At pH 6 probably some additional mechanisms of retention are involved. Single extraction even by N~EDTA solution doesn't enable releasing all the Cu(II) amount retained, probably as far as it is bound in a strong surface complex. However, 3 treatments with the solution containing N~EDTA and Ca(N0 3)2 enable displacing all the copper retained. This extraction can be suggested for direct estimation of exchangeable Cu(II) content in the experiments on copper exchange and in the contaminated soils.
REFERENCES Allard, B., V. Moulin, L. Basso, M.T. Tran, and D. Stammore. Americium sorption on alumina in presence of humic materials. Geoderma 44, pp. 181-187, 1989. BaHI, F., M.C. Ianni, M. Ravera, and E. Magi. Study of the influence of free dissolved amino acids on copper(II) adsorptionlremobilization from inorganic fractions of marine sediments using a reversed phase liquid chromatographic procedure. AnaL. Chim. Acta. 294, pp. 127-134, 1994. Baker, W.E. The role of humic acids from Tasmanian podzolic soils in mineral degradation and metal mobilization. Geochim. COdmochim. Acta. 37, pp. 269-281, 1973. Bender, M.E., W.R Matson, and RA. Jordan. On the significance of metal complexing agents in secondary sewage effluents. Env. Sci. TechnoL. 4(6), pp. 520-521, 1970. Bingham, F.T., A.L. Page, and J.R Sims. Retention ofCu and Zn by H-montmorillonite. Soil Sci. Soc. Amer. Proc. 28, pp. 351-354, 1964. Bizri, y', M. Cromer, and J.P. Scharff. Constantes de stabilite de complexes organo-mineraux. Interactions des ions plombeux avec les composes organiques hydrosolubles des eaux gravitaires de podsol. Geochim. COdmochim. Acta. 48, pp. 227-234, 1984. Bloomfield, C., W.1. Kelso, and G. Pruden. Reactions between metals and humified organic matter. J. SoiL Sci. 27, pp. 16--31, 1976. Bower, C.A. and E. Truog. Base exchange capacity determination as influenced by nature of cation employed and formation of basic exchange salts. SoiL Sci. Soc. Am. J. 20, pp. 86-89, 1941. Chakrabarti, C.L., Lu Yanjia, D.C. Gregoire, M.H. Back, and W.H. Schroeder. Kinetic studies of metal speciation using Chelex cation exchange resin: application to cadmium, copper and lead speciation in river water and snow. Environ. Sci. Techno!. 28(11), pp. 1957-1967, 1984. Davies, J.A. Adsorption of Natural Organic Matter from Fresh-Water Environments by Aluminum Oxide, in Contaminantd and Sedimentd, 2, R.A. Baker, Ed., Ann Arbor Science, Ann Arbor, MI, 1984. Davies, J.A. and J.O. Leckie. Effect of adsorbed complexing ligands on trace metal uptake by hydrous oxides. Environ. Sci. TechnoL., 12, pp. 1309-1315, 1978. DeMumbrum, L.E. and M.L. Jackson. Copper and zinc exchange from dilute neutral solutions by soil colloidal electrolytes. SoiL Sci. 81, pp. 353-357, 1956.
Copper Retention as Affected by Complex Formation
123
Elliot, H.A. and C.P. Huang. The adsorption characteristics of Cu (II) in the presence of chelating agents. J. ColloiJ Inter/. Sci. 70, pp. 29-45, 1979. Elliot, H.A. and C.P. Huang. Adsorption of some copper(II) - amino acid complexes at the solidsolution interface. Effect ofligand and surface hydrophobicity. Environ. Sci. TechnoL. 14, pp. 8793, 1980. Elliot, H.A. and C.P. Huang. Adsorption characteristics of some Cu(II) complexes on aluminosilicates. Water lW. 15, pp. 849-855, 1981. Ephraim, J.H. and J.A. Marinsky. Ultrafiltration as a technique for studying metal-humate interactions: studies with iron and copper. Anal. Chim. Acta. 232, pp. 171-180, 1990. Florence, T.M. and G.E. Batley. Determination of the chemical forms of trace elements in natural waters, with special reference to copper, lead, cadmium and zinc. Tabnta 24, pp. 151-158, 1977. Gamble, D.S., C.H. Langford, and A.W. Wonderdown. The Interrelationship of Aggregation and Cation Binding of Fulvic Acids, in Complexation 0/ Trace Metau in Natural Waterd, C.J.M. Kramer and J.C. Duinker, Eds., Martinus Nijhoff/Dr. W. Junk, The Hague, 1984. Grimme, H. Die Adsorption von Mn, Co, Cu, und Zn durch Goethit aus verdunnten Losungen. Z PJbntzenernahr. und Bodenlc. 121, pp. 58-65, 1968. Guy, R.D. and D.L. Chakrabarti. Studies of metal-organic interactions in model systems pertaining to natural waters. Can. J. Chem. 54, pp. 2600-2611. 1976. Hale, M.G. and L.D. Moore. Factors affecting root exudation. Adv. Agron. ll(31), pp. 93-124, 1970-1978. Harter, R.D. Competitive sorption of cobalt, copper, and nickel ions by a calcium-saturated soil. Soil Sci. Soc. Am. J. 56, pp. 444-449, 1992. Hirose, K., Y. Dokiya, and Y. Sugimura. Determination of conditional stability constants of organic copper and zinc complexes dissolved in sea water using ligand exchange method with EDTA. Marine Chem. ll, pp. 343-354, 1982. Ivanov, V.P. Pbnt Exudated and Their Role in the Life 0/ PhytocenOded. Nauka, Moscow, 1973 (in Russian). Janvion, P., S. Motellier, and H. Pitsch. Ion-exchange mechanisms of some transition metals on a mixed-bed resin with a complexing eluent. J. Chromatography A., 715, pp. 105-ll5, 1995. Khitrov, N.B. Evaluation of the Pfeffer method modified by Molodtsov and Ignatova to determine exchangeable cations content in a single soil sample. Pochvovyedeniye. 6, pp. 105-111, 1982 (in Russian). Klimova, G.M. and Yu.1. Tarasevitch. Absorption of heavy metals ions from water by sorbents on the basis of layered silicates modified by polyphosphates. Khimiya i Telchnologiya Vody. 14(12), pp. 929-936, 1992 (in Russian). Kuwatsuka, S., A. Watanabe, K. Itoh, and S. Arai. Comparison of two methods of preparation of humic and fulvic acids, IHSS method and NAGGY method. Soil Sci. Pbnt Nutr. 38, pp. 2330, 1992. Laxen, P.H. Trace elements adsorption co-precipitation on hydrous ferric oxide under realistic conditions. Water lW. 19, pp. 1229-1236, 1985. Lexmond, Th.M. The effect of soil pH on copper toxicity to forage maize grown under field conditions. Neth. J. Agric. Sci. 28, pp. 164-183, 1980. McLaren, R.G. and D.V. Crawford. Studies on soil copper. I. The fractionation of copper in soils. J. Soil Sci. 24, pp. 172-181. 1973a. McLaren, R.G. and D.V. Crawford. Studies on soil copper. II. The specific adsorption of copper by soils. J. Soil Sci. 24, pp. 443-452, 1973b. Mironenko, E.V., A.A. Ponizovsky, and T.A. Studenikina. Modeling of Chemical Equilibria in Copper Contaminated Soil Containing Low Molecular Weight Organic Acids and Fulvic Acids, in Heavy Metau in the Environment: Proceedingd 0/ International Sympodium, PUdhchino, RUddia, 1996. ONTI, Pushchino, 1996, pp. 153-154.
124
Fate and Transport of Heavy Metals in the Vadose Zone
Murphy, E.M., J.M. Zachara, and S.c. Smith. Influence of mineral-bound humic substances on the sorption of hydrophobic organic compounds. Environ. Sci. Techno!. 24, pp. 1507-1516, 1990. Murray, J.W. The interaction of metal ions at the manganese dioxide-solution interface. Geochim. COdm. Acta. 39, pp. 505-519, 1975. Nykvist, N. Leaching and decomposition of water-soluble organic substances from different types of leaf and needle. Litter StuJia Fourtolia Succia. Stockholm, 3, pp. 27-31, 1963. Pampura, T.V. The Influence of Low Molecular Weight Organic Acids on Cu and Zn Sorption by Chernozem, in PhYdical ChemiJtry ofMadd-&cchange ProcNdN in Soiu: Proceedingd ofInternational Conference, PU.Jhchino, 1992. aNTI, Pushchino, 1993, pp. 85-90. Parfitt, R.L. and J.D. Russel. Adsorption on hydrous oxides. III. Adsorption of various ions on goethite. J. Soil Sci. 28, pp. 297-305, 1977. Parfitt, RL., A.R. Fraser, and V.C. Farmer. Adsorption on hydrous oxides. III. Fulvic acid and humic acid on goethite, gibbsite, and imogolite. J. Soil Sci. 28, pp. 289-296, 1977. Perdue, E.M. and C.R Lytle. Distribution model for binding of protons and metal ions by humic substances. Environ. Sci. Techno!. 17, pp. 654-660, 1983. Ponizovsky, A.A., E.V. Mironenko, and T.A. Studenikina. Complex formation in calcium and copper tartrate solutions. Zh. Neorgan. Khim. 42, pp. 632-637, 1997 (in Russian). Riviere, J. Etude de la rhizosphere de ble. Ann. Agron. 11, pp. 397-440, 1960. Rodda, D.P., J.D. Wells, and B.B. Johnson. Anomalous adsorption of copper(II) on goethite. J. Collow Inteff, Sci. 184, pp. 564-569, 1996. Selim, H.M. Transport and Retention of Solutes in Soils: Multireaction and Multicomponent Models, in Engineering Adpectd ofMetal- Wadte Management, I.K Iskandar and H.M. Selim, Eds., Lewis Publishers, Boca Raton, FL, 1992. Shindler, P.W., B. Furst, R Dick, and P.U.J. Wolf. Ligand properties of surface silanol groups. I. Surface complex formation with Fe3 ., Cu 2+, Cd 2., and Pb 2+. J. ColLow Inteif. Sci. 55, pp. 469475, 1976. Sillen, L.G. and A.E. Martell. Stability COndtantd of Metal-Ion Comp!exN. Special Publ. No. 25, Suppl. No.1 to Special Publ. No. 17, The Chemical Society, London, 1970. Smith, W.H. Release of organic materials from the roots of tree seedlings. For. Sci. 15, pp. 138143, 1969. Soares, H.M.V.M. and M.T.S.D. Vasconcelos. Study of the lability of copper(II)-fulvic acid complexes by ion selective electrodes and potentiometric stripping analysis. Anal. Chim. Acta. 293, pp. 261-270, 1994. Sposito, G., KN. Holtzclaw, C.T. Johnston, and C.S. Le Vesque-Madore. Thermodynamics of sodium-copper exchange on Wyoming bentonite at 298 K Soil Sci. Soc. Am. J. 45, pp. 10791084, 1981. Stumm W., H. HohL and F. Dalang. Interaction of metal ions with hydrous oxide surfaces. Croatica Chemica Acta. 48, pp. 491-504, 1976. Theng, B.KG. and H.W. Scharpensel. The Adsorption of 14C-Labelled Humic Acid by Montmorillonite, in Proceedingd of the International Clay Conference, Mexico City. Applied Science Publishers, Barking, UK, 1975, pp. 643-653. Tipping, E., J.R Griffith, and J. Hilton. The effect of adsorbed humic substances on the uptake of copper(II) by goethite. Croatica Chem. Acta. 56, pp. 613-621, 1983. Van Riemsdijk, W.H. and S.E.A.T.M. Van der Zee. Multicomponent transport modelling of enhanced metal leaching using synthetic ligands. Geoderma. 44, pp. 143-158, 1989. Vuceta, J. and J.J. Morgan. Chemical modelling of trace metals in fresh waters: role of complexation and adsorption. Environ. SCi: Technol. 12, pp. 1302-1309, 1978. Wershaw, RL., D.M. McKnight, and D.J. Pinkney. The Speciation of Copper in Natural Water Systems. I. Evidence of Presence of Copper(II)-Fulvic Acid Charge Transfer Complex, in
Copper Retention as Affected by Complex Formation
125
ProceedingJ of the Second InternationaL SympoJium on Peat in AgricuLture and HorticuLture, K.M.
Schallinger, Ed., Hebrew University, Jerusalem, 1983. Xu, H., J. Ephraim, A. Ledin, and B. Allard. Effects of fulvic acid on the adsorption of Cd(II) on alumina. Sci. Tot. Environ. 81/82, pp. 653-660, 1989. Yatsimirsky, K.B. and V.P. Vasilyev. Stability COnJtantJ of Complex CompOUndJ. Academic Science Publishers, Moscow, 1959 (in Russian). Young, S.D., B.W. Bache, and D.J. Linehan. The potentiometric measurement of stability constants of soil polycarboxylate-Cu 2 + chelates. J. SoiL Sci. 33, pp. 467-479, 1982. Zachara, J.M., C.L. Resch, and S.C. Smith. Influence of humic substances on Co 2+ sorption by subsurface mineral separate and its mineralogic components. Geochim. COJmochim. Acta. 58, pp. 553-566, 1994. Zhang, Z.Z. and D.L. Sparks. Sodium-copper exchange on Wyoming montmorillonite in chloride, perchlorite, nitrate, and sulfate solutions. SoiL Sci. Soc. Am. J. 60, pp. 1750-1757, 1996.
CH{\PTER 6
Copper Mobility and Bioavailability in Relation with Chemical Speciation in Sandy Soil E.J.M. Temminghoff, M.P.J.C. Marinussen, and S.E.A.T.M. Van der Zee
INTRODUCTION In many urban and rural areas, heavy metals have accumulated in soil. Whereas some heavy metals such as Cu and Zn may be micronutrient elements, others are not essential for growth (e.g., Cd, Pb). In case the heavy metals become abundant in soil, adverse effects may occur. Due to heavy metal toxicity, adverse effects may concern both soil organisms and plants in situ, which threatens soil ecology and agricultural production. However, also agricultural product quality may be adversely affected without observable effects on yield, whereas leaching may cause a deterioration of the quality of groundwater; e.g., used for drinking or industrial water. Consequently, much research has been devoted to quantifying the effects of heavy metal contamination with regard to soil as a resource and with regard to the transfer of these compounds into the food chain. It has been well established that heavy metals may be retained by soil due to chemical interactions with the solid phase. This retention is a major factor that controls heavy metal biological availability as well as mobility. The key variable appears to be the heavy metal concentration in the soil solution which is directly related with the displacement in the flowing soil solution. Also for plants, the supply of heavy metals to the root system is strongly affected by the transport in solution toward roots and root hairs. Chemical interactions between the solid and the solution phases of soil control the level at which metal concentrations are buffered in the solution. This has led to the use of unbuffered salt extractants; e.g., Ca(N0 3h CaCI 2, NaN0 3, NH 4N0 3 solutions, to provide a measure of the bioavailable fraction for, e.g., crops (Sanders et aI., 1987; Novozamsky et aI., 1993; Aten and Gupta, 1996). Unfortunately, it is hard to predict the extractable quantities for a range of soil types and contaminant levels because these selective extractants have not been interpreted quantitatively with regard to the involved soil chemical processes. Therefore, understanding of bioavailability and mobility still necessitate the understanding of chemical behavior of the soiIJextractant solution-system. 127
128
Fate and Transport of Heavy Metals in the Vadose Zone
The occurrence of heavy metals in soil solution and solid phase in the form of different chemical species is known as chemical speciation. Speciation depends on the heavy metal of interest, the composition of the soil solution and of the solid phase. Thus, pH, ionic strength, and the presence of cations, anions, organic and inorganic ligands are important parameters with regard to the soil solution and types and quantities of primary, clay, and metal(hydr)oxide minerals. Chemical precipitates as well as organic matter are important with regard to the solid phase. In general, these parameters are spatially variable in natural soils. For this reason, a practically adequate chemical speciation model should enable the prediction of behavior, taking such spatial variability into account. This implies that a model tested in laboratory experiments should be appropriate for describing observations made in the field. It is our scope to consider in more detail the chemical speciation of copper in sandy soils (for which organic matter is the main sorbent in the acid pH-range) and to use this understanding for the assessment ofbioavailability and mobility. The aim of this contribution is to present two heavy metal sorption models that adequately capture the main phenomena for the binding by organic matter. Mter parameterization of these models for copper sorption by purified humic acid, we show how sorption by other types of organic matter [i.e., solid (soil) and dissolved organic matter] can be described with a plausible adaptation of only two parameters. The applicability of the developed models to describe copper retention and mobility is then assessed. In particular, we show that the models that have been parameterized with laboratory experiments is adequate in predicting the retention in a soil profile in the field. Since the agreement between the extracted quantity of metals for a neutral unbuffered salt extractant (e.g., 0.01 M CaCI 2) has been shown in the literature for plants, the bioavailability of copper is discussed for earthworms on the basis of some recent publications. In the soilecotoxicological literature, it is generally agreed that observations made in laboratory experiments cannot be readily extrapolated to field situations, among others, in view of soil heterogeneity. We show that in principle such an extrapolation may be feasible, as no essential differences were observed in copper accumulation at a field site and in laboratory studies. However, spatial variability may necessitate a large experimental effort.
SORPTION MODELS Both empirical and semimechanistic models have been developed to describe heavy metals binding by soil, by dissolved organic matter, and by organisms. Recently, a semimechanistic Non-Ideal Competitive Adsorption model (NICA) has been published which accounts for adsorption by heterogeneous surfaces such as (soil) organic matter (Van Riemsdijk, 1994; Koopal et al., 1994; Benedetti et aI., 1995). This model can be characterized by the following isotherm equation
(1)
Copper Mobility and Bioavailability in Sandy Soil
129
CR.1,2 c·1 f;,2
+Q i(max,2)
L (K -2 ),
n· C ' ) ),2 )
where Qi and Qi max are the adsorbed quantity and the maximum adsorption of component i (mol/kg), j is the median of the affinity constants (K) of component;', where;' includes component i. The parameter n reflects the nonideal behavior of component;, and p reflects the intrinsic heterogeneity of the heterogeneous surfaces. By Cj we denote the "free" concentration of component;' (mollL). Subscript 1 is considered to be related to the first type groups (e.g., "carboxylic") and subscript 2 to the second type groups (e.g., "phenolic"). Equation 1 has the advantage that it is valid over a wide pH range (210) and free metal ion concentration range (10-2_10- 14 mollL) and that it can easily be extended to the multimetal binding case. At median pH ranges and low heavy metal concentrations the NICA equation can be simplified to a Two Species Freundlich equation (TSF) if we take only proton competition with one heavy metal into account (Van Riemsdijk, 1994; Temminghoff et al., 1997)
K
(2) where we defined
W1 = Q.l,max (it1 )n; (KH )nH(p-l)
(3)
and
(4) In the case of heavy metal binding by sandy soils, where organic matter has chemically the most reactive surface, Qi can be written as
(5) where Q.soil is the adsorbed quantity of component i by soil (mol/kg soil), and fOM the fraction organic matter of the soil (kg/kg).
PARAMETER ASSESSMENT SORPTION MODELS As a model substance for the binding studies onto organic matter we used purified humic acid that was extracted from forest floor material taken from the Tongbersven forest (near Oisterwijk, NL). It has been used in previous studies and is described by
130
Fate and Transport of Heavy Metals in the Vadose Zone
Van Dobben et al. (1992) and Mulder et al. (1994). The upper 2 cm of the podzol B horizon from the forest floor material was collected in plastic bags. Upon return to the lab, the sample was homogenized, sieved « 2 mm), and stored at 4°C (field moist). The humic acid extraction was carried out according to the recommendation of the International Humic Substance Society (IHSS). The obtained purified humic acid was freeze dried and stored until use. Before use, each sample was dissolved in demi water. Data for proton and copper binding by purified humic acid (dissolved organic matter) were measured potentiometrically using a pH electrode and/or a copper ion selective electrode (Cu-ISE) and a double junction calomel reference electrode, all connected to a programmable Wallingford titrator (Kinniburgh et al., 1995). All experiments were performed at 25°C in 0.003 M NaN0 3 background electrolyte. The Cu-ISE was linear for pCu between 3 and 13 (r2 = 0.9996), using ethylenediamine for low Cu activities (Avdeef et al., 1983; Benedetti et al., 1995). Copper binding was measured at constant pH values of 4, 6, and 8. To keep the pH at a certain level we used the pH stat mode of the titrator. The total Cu concentration in solution was calculated by the buret additions. The Cu binding by humic acid can be calculated by subtracting the Cu 2+ from the total Cu concentration in solution. Above pH 6, Cu hydroxide species should also be taken into account (Benedetti et al., 1995). A detailed description has been given by Temminghoff et al. (1997). Copper binding by purified humic acid is shown in Figure 6.1a for pH 4, 6, and 8 on logarithmic scales. Copper binding depends nonlinearly on pH and Cu2+ concentration and is well described by both the (bimodal) NICA model (Eq. 1) and the TSF model (Eq. 2). The fitted parameters are given in Table 6.1 for both models. The NICA model described the Cu adsorption slightly better because it accounts for an adsorption maximum. Proton binding by the purified humic acid is shown in Figure 6.1 b where the relative charge is given as a function of the pH. The fit of the proton binding data by NICA (solid line) was excellent (r2 = 0.99998). The fitted NICA parameters for protons are similar to those determined by Benedetti et al. (1995) despite the fact that their humic acid is from a different location and that they used a different ionic strength. COPPER SPECIATION IN A COPPER CONTAMINATED SOIL
With the parameters derived for copper binding by purified humic acid, we tried to describe copper desorption by a copper contaminated sandy soil, where the most important sorption component is solid organic matter (in terms of reactivity and abundancy). Soil samples were taken from an arable field near Wageningen in The Netherlands (Wildekamp site) which has been used previously to study copper sorption and toxicity effects on earthworms, nematodes, and collembola (Temminghoff et al., 1994, 1997; Marinussen et al., 1997a; Korthals et al., 1996; Bruus Pedersen et al., 1997). The soil was a Spodosol in a slightly loamy moderately fine sand. The field contains a randomized block design of four copper concentrations and four pH adjustments in 6 X 11 m plots. Four pH levels coded A-D for pH KC1 between 4.0 and 6.1 were established in 1982 using calcium carbonate or sulfur, followed by establishment of four total Cu levels coded 1-4 between 0 and 750 kglha, added as CUS04' at each pH-level (Lexmond, 1980). We assumed that an equilibrium between the copper pollution and the soil has been more or less established since the "pollution" occurred more than 14 years ago. Soil samples were collected from the plots with the highest (code D) and lowest (code A) pH, for each pHI
Copper Mobility and Bioavailability in Sandy Soil
1 31
0.50
NICA
TSF 0.00
-0.50
~
CY
b.O
.....o
-1.00 o + +
-1.50
-12
-10
-4
-6
-8
log[Cu 2+] (moIlL) 4
0
,-...
g
data HNICA
3
~
'-" «)
~ .c:::
2
()
~ ..... «)
'0
o
L -_ _ _ _ _ _ _ _. -_ _ _ _ _ _ _ _- ._ _ _ _ _ _ _ _ _ _. -_ _ _ _ _ _ _ _- ._ _ _ _ _ _ _ _
2
4
6
8
10
~
12
pH =igure 6.1. (a) Copper binding by DOC (purified humic acid) as a function of log(Cu 2+) at pH 4, 6, md 8 for data, the NICA and the TSF model at 1=0.003 (0.003 M NaN0 3); (b) Proton binding by )OC (purified humic acid), calculated as delta charge, as function of pH [-log(W)] for data and the \lICA model at 1=0.003 (0.003 M NaN0 3).
:::'u combination, at a depth of 0-0.20 m. All soil samples were air-dried and sieved « 2 nm). Typical soil properties are given in Table 6.3. Neutral unbuffered salt solution
132
Fate and Transport of Heavy Metals in the Vadose Zone Table 6.1. Parameters for the NICA and the TSF Model to Describe the Copper Binding by Purified Humic Acid at I = 0.003 NICA log KH log K Cu nH ncu
P Qcu,max (mol/kg)
TSF
Site 1
Site 2
5.12 5.40 0.81 0.46 0.57 2.18
9.78 10.82 0.83 0.46 0.58 1.54
ncu
mH log Ki
0.36 -0.40 -0.348
NICA: H; coefficient of determination (r2) = 0.99998. Cu and H; coefficient of determination (r2) = 0.98. TSF: coefficient of determination (r2) = 0.97.
extractions are carried out with 0.001 M Ca(N03)2 and 0.01 M CaCl 2 (soil: solution ratio 1:10; wN). Three grams of air-dried soil was equilibrated with 30 mL 0.001 M Ca(N03)2 and 0.01 M CaCl2 for 20 h in an end-over-end shaker. Soil pH was determined in this suspension with a glass-calomel electrode. Mter centrifugation at 2,000 g for 15 min, Cu concentration in the supernatant (Cu ex) was measured on a flame atomic absorption spectrometer (Instrumental Laboratory ANAE spectrophotometer S 11). Total extractable soil Cu content (CUT) was determined by equilibration of 3 g of air-dried soil with 30 mL 0.43 M HN0 3 for 20 h in an end-over-end shaker. Since DOC coagulate in 0.01 M CaCI2, we used 0.001 M Ca(N0 3)2 for the Cu mobility experiments. The concentration of copper extracted from the soil [Cuex ; 0.001 M Ca(N03hJ extraction varied from 0.18 ± 0.04 to 10.3 ± 0.2 J..lmollL and corresponds to "free" copper (Cu 2 +), copper bound by dissolved organic matter (CuDOC) and copper bound by inorganic species. The Cu binding by the inorganic species (e.g., N0 3) is negligible. To assess the Cu 2 + and CuDOC fractions, speciation techniques are required (Temminghoff et aI., 1994). Assuming that Cu binding by dissolved organic matter behaves as Cu binding by humic acid, the "free" Cu 2 + can be calculated at every pH, CUex and DOC concentration by either the NICA or the TSF model. In agreement with Kinniburgh et al. (1996), we assumed that Ca competition is negligible for the copper binding by organic matter. Dissolved organic carbon concentrations for all the 0-0.20 m top layer soil samples were equal (8.9 ± 1.4 mg/L). The calculated pCu 2 + concentrations in the top layer varied from 5.1 (soil4A) up to 9.9 (soillD). For sandy soils, free copper (Cu 2+) has been shown to be in equilibrium with Cu bound by solid organic matter of the soil and by dissolved organic matter of the soil solution (Temminghoff et aI., 1994). The Cu2+ binding by the (soil) solid organic matter in the top layer 0-0.20 m was also described by the NICA and TSF models where all parameters, except Qcu,max and the intrinsic heterogeneity (p), were kept equal to the case of Cu 2 + binding by the humic acids (Table 6.2). By fitting only these two parameters for NICA and the two corresponding parameters for TSF (Table 6.2) we described the Cu 2+ binding by soil for the layer 0-0.20 m (Figure 6.2). The maximum Cu sorption capacity for (soil) solid organic matter had to be reduced to 30% of the Qcu,max of the purified humic acid for both the TSF and the NICA model. This is plausible because the number of reactive sites (sites
Copper Mobility and Bioavailability in Sandy Soil
133
Table 6.2. Adjustable Parameters for the NICA and the TSF Model to Describe the Copper Binding by Soil Organic Matter at I = 0.003 NICA
p Qcu,max
(mol/kg)
TSF
Site 1
Site 2
0.50 0.65
0.50 0.46
-0.58 -1.77
NICA: CU and H; coefficient of determination (r2) = 0.97. TSF: coefficient of determination (r2) = 0.93.
density) is probably smaller for solid organic matter than for dissolved organic matter since solid organic matter is less humified (Gooddy et aI., 1995). For the NICA model the factor p had to be adjusted from 0.57 and 0.58 to 0.50, which indicates larger intrinsic heterogeneity of (soil) solid organic matter than purified humic acid. A change in Qcu,max and parameter p gives a change in K/ and mH (see Eqs. 3 and 4) for the TSF model. The NICA and the TSF model described the "free" copper desorption data excellently. Only for mean pH 5.6 and Cu 2+ concentrations of approximately 10- 10 moliL, the TSF model shows slight discrepancies.
MOBILITY DOC Mobility Enhanced Copper Mobility To show the effect of DOC concentration on Cu mobility, column experiments were carried out with top layer soils of plot 4A (1.610 mmollkg Cu total) and 3D (1.887 mmoli kg Cu). We used perspex columns 4.5 cm in diameter and about 20 cm long which were filled with 300 g of both soils. The top of the column was connected via a tube to the influent solution (0.001 M Ca(N03)2; pH 5.7). A fraction collector (ISCO fraction collector) was used to collect the effluent samples of about 15 mL. Besides total Cu, we measured DOC and the pH in the effluent. More detailed information is given by Temminghoff et al. (1997). In Figure 6.3 the total copper and DOC concentration and pH in the effluent of the columns are given as a function ofleached pore volumes 0INo) for soil4Al, with mean pH 3.9 (Figure 6.3a), and soil 3D 1, with mean pH 6.6 (Figure 6.3b). The effluent pH for both columns is given in Figure 6.3c. The effluent pH differs from the soil pH determined by batch experiment, especially at large pH. At large pH the difference is about 1 pH, while at small pH the difference is negligible. The Cu breakthrough curve (BTC) at pH 6.6 shows one Cu peak that occurs simultaneous with the DOC peak (Figure 6.3b). At pH 3.9 two Cu peaks are visible, the first one at the same moment as the DOC peak, followed by a second Cu peak that corresponds with the decline of pH during the first pore volumes (Figure 6.3c). The Cu concentration at pH 3.9 reached a maximum of approximately 25 IlmoliL Cu after 7 pore volumes and decreased slowly to 8 IlmoliL after 45 pore volumes. At pH 6.6 the Cu concentration was 11 IlmoliL (after 1 pore volume) and decreased fast below 5llmol/L after 5 pore volumes. Maximum DOC concentration in the effluent was found after 1 pore volume both for pH 3.9 (83 mg/L DOC) and for pH 6.6 (120 mg/L DOC) and decreased fast to concentrations below 10 mg/L.
134
Fate and Transport of Heavy Metals in the Vadose Zone -2.00
.• 0
-2.50
data TSF NIeA
,-., bIl
~~
pH 5.6
-
1'*
."
~,~
~
-3.00
0
S
pH 3.9
'-'
~
0
.~
8
-3.50
bIl
-
•
0
• • c•
c•
•
-4.00
-4.50 -10
-12
-4
-6
-8
log[Cu 2+] (moIlL)
Figure 6.2. Log (copper binding by soil) as a function of log ("free" copper soil solution) at two average soil pH levels; data, NICA and TSF model at 1=0.003 [0.001 M Ca(N0 3H 40
~------------------------------------------------~
a) soil column 4A 1
:::J ;:::,
0
Cu data
" " :;' ::I
100
30 80
~
......, C
120
:::J
'bb
60
20
§ U
0
E
Cl 40
S:!. o
.........
... ---------...:.
......
~- --o
-
..,... ............... . .............. .
~----------~-----------.------------~----------~ 40 o 10 30
o
Figure 6.3a. Copper and DOC concentration and pH in the effluent as a function of pore volumes for (a) soil column 4A 1, (b) soil column 301, and (c) effluent pH for 4A 1 and 301 for 0.001 M Ca(N0 3h as influent (1=0.003). Line and dashed line is prediction of copper by NICA and TSF model, respectively.
The solid and the dashed lines are the Cu concentration predictions by the NICA and the TSF model in the effluent that take both pH and DOC effects into account. For the NICA and the TSF calculations, Cu2 + desorption by soil organic matter and Cu 2+ bind-
Copper Mobility and Bioavailability in Sandy Soil ------------ - ~
~------------------------------------------------,
b) soil column 301
:3 :::.
•
30
" !€'"
120
Cu data
0
! d
135
-----
100
DOC data CuNICA
80
CuTSF
::J
E
60
20
:;-
8
~
~
10 20
w
~
0
0 0
~
30
20
10
VNo
8
c) pH effluent soil columns
. .. .... .. ..... . .. ...... .... .... .
7
......... . ".... . ... . ... ... ......... . . . .
... ... .., .........
soil column 3Dl
6
-~... __ a;, .............. ,,",~ ~v~ • ., ..,v
4
.V
99 Vq.
;VV04P4'ncP~
V9V
• • VV • •
v v'" v
soil column 4Al 3
2
o
10
30
40
Figure 6.3b and c. Copper and DOC concentration and pH in the effluent as a function of pore volumes for (a) soil column 4A 1, (b) soil column 3Dl, and (c) effluent pH for 4A 1 and 3Dl for 0.001 M Ca(N0 3b as influent (1=0.003). Line and dashed line is prediction of copper by NICA and TSF model, respectively.
ing by DOC was calculated at each effluent pH and DOC concentration with the use of the parameters of Tables 6.1 and 6.2. The Cu effluent concentration is the sum of Cu 2 + and CuDOC. The Cu effluent concentration is predicted rather well for both models. The first peak for both pH values is slightly overestimated, possibly due to nonequilibrium during the first pore volume percolation in the columns. At pH 3.9 after about 10 pore volumes, approximately 70% of the total Cu in the effluent solution is Cu 2 + and about 30% is present as CuDOC, as calculated with the NICA and/or the TSF model. During the first ten pore volumes, Cu 2+ varied from 20 to 70% due to changes in DOC
136
Fate and Transport of Heavy Metals in the Vadose Zone
and pH. At pH 6.6 only a small part of total Cu is present as Cu 2+ (0.01 %) and most is of the form CuDOC (> 99%), which shows that the effect of DOC on Cu mobility is of greater importance at higher pH than at lower pH. Quite different conditions than in batch/titration studies are found in column experiments. Still, the model and parameterization for a model substance (purified humic acid) yield an excellent description of Cu in leachate. Provided DOC is mobile itself, we observe that facilitated Cu-transport occurs that leads to faster Cu-Ieaching.
Field Site Accumulation in Soil If the previous speciation modeling captures the main phenomena, not only for conditioned laboratory experiments but also for field situations, we should be able to predict the extractable copper concentration in the soil solution at different depth. From two plots (4A and 3D), soil was sampled from the layers 0-0.20,0.20-0040,0040-0.60,0.600.80, and 0.80-0.90 m. All soil samples were air-dried and sieved « 2 mm). Dissolved organic matter (DOC) varied as a function of depth in the field between 16 and 1 mg/L, solid organic matter (SOC) content between 37.8 and 3.8 g/kg, pH between 3.83 and 5.91, and total Cu content between 1.89 and 0.013 mmollkg. In Figure 604 total Cu content and pH (Figure 6Aa) and solid and dissolved organic matter (Figure 6Ab) are given as a function of depth. Using parameters of Tables 6.1 and 6.2 for describing Cu binding by DOC and Cu binding by (soil) solid organic matter (parameters only determined for the top layers 0-0.20 m), we predicted the extractable Cu concentration in the soil solution at each depth between 0 and 0.90 m with the NICA and the TSF models. In Figure 6.5 the predicted CU ex concentration is given as a function of the measured CUex concentration. The agreement is good for both models. For the NICA model the prediction was slightly better than the TSF model since the correlation coefficients were 0.97 and 0.93, respectively. At the original moment of contamination (1982), soils of plot A and D with the same number had the same (total) Cu content in the plow layer, but currently differences are observed. For soil4A (small pH) the total Cu content in 1994 was already smaller than in soil 3D (large pH) in layer 0-0040 m although the added Cu in 1982 was higher for soil 4A (750 kg/ha) than for soil 3D (500 kg/ha). However, below a depth of 0040 m the copper content in soil 4A is much larger than in soil 3D in 1994. The extractable Cu concentration (CueJ, determined via 0.001 M Ca(N03)2 extraction and the models, is for soil 4A (layer 0-0.20 m), about five times larger than for soil for 3D (large pH), which illustrates the increase in Cu mobility at small pH.
BIOAVAILABILlTV Bioavailability for Soil Organisms Another objective of our research is to determine whether the mobile fraction corresponds with the fraction that is available for organisms. In soil fertility as well as environmental studies involving plant uptake it has been shown that the 0.01 M CaCl2 extractable contents in soil is a better indicator of the fraction that is available for uptake than total extractable contents (Novozamsky et al., 1993). Thus, it may be useful as an indication of availability for soil dwelling organisms too. Marinussen et al. (l997a) ex-
Copper Mobility and Bioavailability in Sandy Soil
137
2.0 , . - - - - - - - - - - - - - - - - - - , 7
,-..
eo
~0 S S '-'
5
1.0
.. CJ
6
1.5
:I:: p..
Cu3D Cu4A pH 3D
::I
u
----
4
0.5
0.0
0-20
20-40
40-60
60-80
pH4A
3
80-90
depth (em)
20
- -EJ- -
SOM 3D
-0-
SOM 4A
g --.--
DOC 3D
-15 ;J'
-10
"& ....... o
- e - DOC4A
o
0-20
20-40
40-60
, 60-80
depth (em)
Figure 6.4. Total copper content (mmollkg) and pH as a function of soil depth (Figure 6.4a) and total organic matter and DOC as a function of soil depth for field plot 4A and 3D (Figure 6.4b).
posed earthworms (LumbriclM rubeLLw) to soil plots 2A, 2D, 3A, and 4 C that were sampled from the Wildekamp site as described above (Table 6.3). The earthworms were exposed under laboratory conditions for 1, 7, 14,28, or 56 days to study tissue Cu accumulation. After sampling, earthworms were rinsed with distilled water and kept for three days in a petri dish on moist filter papers to empty their gut. Earthworms were killed by immersion in liquid nitrogen and dried in an oven at 105°C. Earthworms were individually digested in 5 mL 65% HN03 and 4 mL 20% H2 0 2 . These solutions were analyzed for
eu on a furnace AAS.
138
Fate and Transport of Heavy Metals in the Vadose Zone
-4
]
r-------------------------------------------~
o
NICA
+
TSF
-5
u
t;:s ~
g
-6
§: bO
.£
-7
-8
~--------.----------.---------,--------~
~
~
~
~
4
log[Cu] (moVL) measured Figure 6.5. Extractable copper concentration calculated with the TSF and the NICA model in the Spodosol soil profile as a function of the measured extractable copper concentration for five layer soil 4A and 3D up to 0.9 m depth at 1=0.003 [0.001 M Ca(N03)~. Solid line is the 1:1 line.
Table 6.3. Several Characteristics of the Soil Used in the Experiments [pH CaCI 2, Clay (%), Organic Matter (% C), and 'Total" Cu with Standard Deviation (mmol/kglI Clay
C
pH-CaCI 2
(%)
(%)
1A 2A 3A 4A
3.80 3.91 3.84 3.77
2.9 2.2 1.7 3.4
1.98 2.20 2.16 2.15
0.42 0.98 1.69 1.66
18 28 38 48
4.33 4.24 4.18 3.85
2.4 2.9 1.9 2.4
1.89 2.02 2.14 2.02
0.27 1.13 1.49 1.93
1C 2C 3C 4C
4.73 4.24 4.18 3.85
5.0 3.8 3.3 5.0
2.26 2.34 1.87 2.07
0.38 1.18 1.77 2.26
10 20 3D 40
5.24 5.28 4.95 5.12
3.9 3.8 4.5 3.3
2.21 2.24 2.18 2.10
0.16 1.23 1.77 2.64
3.3 1.0
2.11 0.13
Plot Code
Mean Std
CUT (mmol/kg)
± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ±
0.02 0.02 0.05 0.09 0.01 0.01 0.02 0.02 0.01 0.03 0.01 0.04 0.005 0.04 0.01 0.06
Copper Mobility and Bioavailability in Sandy Soil
139
Tissue Cu concentrations increased as a function of time and proportionally with total extractable soil Cu content (r2=0.9). Neither the correlation between accumulation and soil pH, nor between accumulation and CU ex (0.01 M CaC1 2), however, was significant. Marinussen et al. (l997a) observed a large mortality of L. rubeLLIM in soils that were high in CU ex (0.01 M CaC1 2). Their data enable us to demonstrate the relationship between mortality and extractable soil Cu content. Speciation of Cu in the soil solution plays a possible role in mortality but this was not investigated. Therefore, we show in Figure 6.6 the average mortality rate in the five contamination levels. It is obvious that the average mortality rate correlates better to CU ex than to CUT. From these observations we conclude that the exposure route for uptake differs from the exposure route that causes mortality. The latter seems to correlate with Cu concentration in soil solution which depends on total soil Cu content and soil pH, as illustrated by Equation 2.
Field Site Accumulation by Earthworms Generally, soil contamination is spatially variable. Hence, in contaminated field sites, exposure of soil-dwelling organisms to soil contamination varies as a function of time. The larger the spatial variability is, the more variation in exposure should be anticipated. Marinussen and Van der Zee (1996) showed that effects of such variation on accumulation of contaminants in organisms depends on the degree of spatial variability and the size of the home-range of the organism. Earthworms are organisms with a limited home range. The mobility of earthworms is influenced by the earthworms' ecology and by environmental conditions; e.g., soil humidity, soil temperature, and food availability (Sims and Gerard, 1985). Mobility and spatial variability are complicating factors for predicting accumulation of heavy metal in earthworms exposed to spatially variable soil contamination. However, this may be a matter of effort rather than of principle. The involved effort may be the reason why field studies on effects of spatial variability of soil contamination on exposure of organisms are rare. Marinussen et al. (l997a,b,c) studied heavy metal accumulation in earthworms under both laboratory and field conditions. The main objective of their studies was to determine whether data on heavy metal accumulation obtained by laboratory studies can be used for predicting heavy metal accumulation under field conditions. Marinussen et al. (l997a) introduced in each of the four differently contaminated plots (Table 6.3), 500 specimens of the earthworm L. rubeLIm. To determine tissue Cu accumulation under field conditions, earthworms were sampled at three times (14, 28, and 70 days after introduction). In Figure 6.7, we show that the decline in tissue Cu concentration between the first and the second sample time coincided with relatively low soil temperatures. These low soil temperatures may have caused a downward migration of L. ruheLLIM. As shown in Figure 6.4a, soil in the upper 40 cm layer is considerably more contaminated than soil at greater depth. Hence, earthworms that move downward are exposed to less contaminated soil and therefore accumulate less copper, which is in agreementwith the data (Figure 6.7). Marinussen et al. (1997a) found that tissue Cu accumulation was significantly correlated with Cup whereas it was correlated neither with soil pH nor with CUex (0.01 M CaCI 2). They also observed large mortality in plots where soil was high in CU ex • These results are in agreement with the laboratory experiments, described above.
Figure 6.6. Average mortality rate of earthworms exposed to soil samples obtained from the Wildekamp site (see text for details) as a function of CUT (6A) or CU ex (0.01 M CaCI~ (68).
Copper Mobility and Bioavailability in Sandy Soil
141
"--------
,---------------------------.10 1.5 temperature plot 4C
Oil
8
.§.
6
........
~0 1.2 8
0.g 0.9
150 0
plot 2D
d 0
4 0.7
0 ;:l
2
U -----
plot 3A
i
8
d 0
plot2A
IT o
0 ;:l
'"'"
'.0
~----~~T---------------~O
-'0
'"
0.1
o
14
28
42
56
exposure time [days] Figure 6.7. Tissue Cu accumulation under field conditions in earthworms Lumbricus rubel/us exposed to contaminated soil at the Wildekamp site. See text for details.
Marinussen et al. (1997b,c) exposed earthworms Dendrohaena veneta to heavy metal (Cu, Pb, Zn) contaminated soils under both laboratory and field conditions. The soil samples used in the laboratory experiments were obtained from a contaminated field site in Doetinchem, NL [sandy loam soil, 7% clay, 3% organic matter (loss on ignition)]. After homogenization of the soil, subsamples were taken and analyzed for CUT (12.8 ± 1.8 mmollkg), CUex (0.01 M CaCI 2; below detection limit = 0.6 IlmollL), and pH-CaCI 2 (7.0 ± 0.06). Dendrohaena veneta were exposed to the soil for 1, 2, 3, 7, 14,28,56, or 112 days to study tissue heavy metal accumulation under laboratory conditions. To study heavy metal excretion, D. veneta were transferred to uncontaminated soil after exposure to the contaminated soil for 112 days. Both accumulation and excretion of Cu appeared to be fast processes (Marinussen et aI., 1997c). An equilibrium in tissue Cu accumulation was achieved 14 days after introduction in the contaminated soil. Three days after being transferred to uncontaminated soil, earthworms lost about 70% of the accumulated Cu. Lead was accumulated to a very small extent, and Pb excretion stagnated at 40% 56 days after transferring to uncontaminated soil. Zinc was not accumulated. From these data, we conclude that tissue Cu concentration in D. veneta adapt rapidly to changes in exposure which are common in spatially variable soil contamination. In another study of Marinussen et al. (l997b), D. veneta were exposed for 14 days to 10 differently contaminated soil samples (field site in Doetinchem, NL) to determine the relationship between soil Cu content and tissue Cu concentrations. They found that in soil containing 0.16 to 1.57 mmollkg Cu, the earthworm tissue Cu concentration increased proportionally to the total extractable soil Cu content (CUT)' Earthworms seemed to achieve a maximum tissue Cu concentration (Figure 6.8).
ILl2
Fate and Transport of Heavy Metals in the Vadose Zone
1.4
r-------------------------.
•
1.2
j
1
c o
0.8
t
0.6
'.0
•
• •
u
::s
U
g
0.4
'"'" '.0 0.2
o ~---~----~----~----~----~----~----~----~ o 2 4 16 14 6 8 12 10 0.43 M HN03 extractable eu [mmol/kg]
Figure 6.S. Tissue Cu concentrations in earthworms Dendrobaena veneta exposed to Cu contaminated soil under laboratory conditions.
Additionally, Marinussen et al. (1997b) introduced about 100 specimens of D. veneta at each of 20 homogeneously distributed locations in the field site in Doetinchem. The spatial variability of soil Cu contamination in this field site was considerable (Figure 6.9). At three different times, earthworms were sampled and analyzed for tissue Cu concentration (procedure is described above). For each earthworm, the Cu concentration factor was calculated (CFcu is the ratio tissue Cu concentration to the total extractable soil Cu content). An accurate estimation of the total extractable soil Cu content at each location was obtained by geostatistical interpolation (in CadU disjunctive kriging). They also calculated CFCU for earthworms exposed under laboratory conditions (derived from Figure 6.S). It appeared that CF CU values under field conditions were in good agreement with CFcu values under laboratory conditions (Figure 6.10). This successful extrapolation from laboratory to field scale was a result of a high soil sampling density. They took Sl soil samples in the top layer (0-20 cm) in an Sl m 2 experimental plot. The large variation in CFCU under field conditions may be explained by a considerable decrease of soil Cu contamination as a function of depth at this site (Figure 6.11). Since D. veneta moves up and down through the upper layer of soil, the latest exposure level is uncertain as a result of this kind of spatial variability. The field studies by Marinussen et aI. (1997 a,b) show that earthworm heavy metal accumulation under field conditions can be predicted using relationships between soil heavy metal contamination and tissue heavy metal concentrations determined under laboratory conditions. However, a high soil sampling density is required to obtain accurate estimations of exposure levels of individual specimens.
Copper Mobility and Bioavailability in Sandy Soil
143
Figure 6.9. Copper contamination at the field site in Doetinchem, NL. The vertical axis is the total extractable copper in soil (mmollkg). Soil samples were taken from the top layer (0 to 0.20 m). The soil sampling scheme was a squared 8 by 8 grid with 1.0 m node distance, resulting in 81 soil samples at 64 m2.
SUMMARY We presented two models, the NICA and the TSF models that have been developed to describe, among others, heavy metal sorption by soil. For both models, data on pHdependent copper binding by purified humic acid were fitted. The description was good and the determined parameters are in agreement with literature for other humic acids and conditions such as ionic strength. For natural solid soil organic matter it is plausible that the reactivit;y is smaller, whereas the sorption site heterogeneit;y is larger than for dissolved purified humic acid. Adapting
144
Fate and Transport of Heavy Metals in the Vadose Zone
2.5
r--------------------------,
2 1- •
•
....
*
0
u
~
1.5
e::
l-
0
•
'.;::l
o:s
b
e::
(l.l
u
e:: 0
u
1
I-
0.5
t-
;::l
u
0.43 M HN03 extractable eu in soil [mmol/kg]
Figure 6.10. Copper concentration factors in earthworms Dendrobaena veneta exposed to contaminated soil under laboratory conditions (circles) or field conditions (asterisks).
100
eil ~ 0
§
~
;::1
u
10
(l.l
::0 CIl
.... g
~
(1)
0
,.,
~
:E M
~ 0
0.1
V
IV
VI
III
VII
II
VIII
IX
Figure 6.11. Total extractable soil Cu content (CUT) in four consecutive layers of 5 cm thickness at 9 spatially distributed locations in the Doetinchem field site; open = 0-5 cm, hatched = 5-10 cm, cross-hatched = 10-15 cm; fine-hatched = 15-20 cm.
Copper Mobility and Bioavailability in Sandy Soil
145
only the two parameters that are related with reactivity (sorption maximum) and heterogeneity, the two models describe pH-dependent copper binding by natural organic matter also. To further ascertain the applicability of the models, the agreement between model prediction and measured data was considered for two rather different situations. The first of these concerned the leaching of copper for two sandy topsoil columns that have different pH and DOC levels. The other situation was the retention that is apparent in a field soil for depths up to 0.90 m, with significant variations in dissolved and soil organic matter, total copper, and pH. In both cases the agreement between model predictions and observations was good. This indicates that for sandy soil, the two models capture the main phenomena. Hence, with regard to both mobility and the chemical interpretation of a neutral unbuffered salt extraction, an interpretation using the NICA and the TSF models may improve our understanding. The chemical speciation modeling may be necessary to be able to predict, e.g., copper uptake by plants. However, the available information with regard to accumulation of copper earthworms indicates that bioavailability for earthworms depends on the total copper content rather than the fraction that can be extracted with a mild extractant. The latter fraction does appear to control the short-term toxicity of copper for earthworms, possibly due to oral uptake. Copper accumulation by earthworms may be controlled mainly by dermal uptake and is not strongly related with short-term toxicity effects. As the total copper levels in field soils are often spatially variable, copper accumulation in field situations may be difficult to predict. For a field site, we showed that it is in principle feasible to predict copper accumulation using observations from laboratory experiments and a reliable map of total copper contents of the involved site. Unfortunately, to obtain a reliable map may require a high density of soil sampling. In summary, we conclude that tools have been developed to translate laboratory data such that they have relevance for soil in situ. To apply these tools for practical predictions of bioavailability and mobility of heavy metals is currently under investigation.
REFERENCES Aten, C.F. and S.K. Gupta. On heavy metals in soil; rationalization of extractions by dilute salt solutions, comparison of the extracted concentrations with uptake by rye grass and lettuce, and the possible influence of pyrophosphate on plant uptake. Sci. Tot. Environ. 178, pp. 45-53, 1996. Avdeef, A., J. Zabronsky, and H.H. Stuting. Calibration of copper ion selective electrode response to pCu 19. AnaL. Chem. 55, pp. 298-304, 1983. Benedetti, M.F., C.J. Milne, D.G. Kinniburgh, W.H. Van Riemsdijk, and L.K. Koopal. Metal ion binding to humic substances: Application of the non-ideal competitive adsorption model. Environ. Sci. TechnoL. 29, pp. 446-457, 1995. Bruus Pedersen, M., E.J.M. Temminghoff, M.p.J.e. Marinussen, N. Elmegaard, and C.A.M. van Gestel. Copper uptake and fitness of Fouomia candUJa willem in a copper contaminated sandy soil as affected by pH and soil moisture AppL. SoiL &oL. 2(6), pp. 135-146, 1997. Gooddy, D.C., P. Shand, D.G. Kinniburgh, and W.H. Van Riemsdijk. Field-based partition coefficients for trace elements in soil solutions. European J. SoiL Sci., 46, pp. 265-285, 1995. Kinniburgh, D.G., C.J. Milne, and P. Venema. Design and construction of a personal-computerbased automatic titrator. SoiL Sci. Soc. Am J. 59, pp. 417-422, 1995.
146
Fate and Transport of Heavy Metals in the Vadose Zone
Kinniburgh. D.G .• C.J. Milne. M.F. Benedetti. J.P. Pinheiro. J. Filius. L.K. KoopaL and W.H. Van Riemsdijk. Metal ion binding by humic acid: Application of the NICA-Donnan model. Environ. Sci. Techno!. 30. pp. 1687-1698. 1996. Koopal, L.K.. W.H. Van Riemsdyk. J.C.M. De Wit. and M.F. Beneditti. Analytical isotherm equations for multicomponent adsorption to heterogeneous surfaces. J. Coll. Interface Sci. 66. pp. 51-60. 1994. Korthals. G.W.• A.D. Alexiev. T.M. Lexmond. J.E. Kammenga. and T. Bongers. Long-term effects of copper and pH on the nematode community in an agroecosystem. Environ. ToxicoL. Chem. 15. pp. 979-985. 1996. Lexmond. Th.M. The effect of soil pH on copper toxicity for forage maize as grown under field conditions Neth. J. Agric. Sci. 28. pp. 164-183. 1980. Marinussen. M.P.J.C. and S.E.A.T.M. Van der Zee. Conceptual approach to estimating the effect of home-range size on the exposure of organisms to spatial variable soil contamination. &ol. MOdelling. 87. pp. 83-89. 1996. Marinussen. M.P.J.C.• S.E.A.T.M. Van der Zee. and F.A.M. De Haan. Cu accumulation in LumbriclM rubelllM under laboratory conditions compared with accumulation under field conditions. &otox. Environ. Safety 36. pp. 17-26. 1997a. Marinussen. M.P.J.C.• S.E.A.T.M. Van der Zee. and F.A.M. De Haan. Cu accumulation in the earthworm Dendrobaena veneta in a heavy metal (Cu. PB. Zn) contaminated site compared to Cu accumulation in laboratory experiments. Environ. PoLL. 96(2). pp. 227-233, 1997b. Marinussen. M.P.J.C.. S.E.A.T.M. Van der Zee. F.A.M. De Haan. L.M. Bouwman. and M.M. Hefting. Heavy metal (Copper. Lead and Zinc) accumulation and excretion by the earthworm. Dendrobaena veneta. J. Environ. Qual. 26. pp. 278-284. 1997c. Mulder. J .• D. Van den Burg. and E.J.M. Temminghoff. Depodzolization Due to Acid Rain: Does Aluminium Decomplexation Affect to Solubility of Humic Substances? In Humic SubdtancN in the GWbal Environment and Implicationd on Humic Health. N. Senesi and T.M. Miano, Eds .• Elsevier Science. 1994. pp. 1163-1168. Novozamsky. 1.. Th.M. Lexmond. and v.J.G. Houba. A single extraction procedure of soil for evaluation of uptake of some heavy metals by plants Int. J. Environ. Anal. Chem. 55. pp. 47-58, 1993. Sanders. J.R .• S.P. McGrath. and Mc.M. Adams. Zinc. copper and nickel concentrations in soil extracts and crops grown on four soil treated metal loaded sewage sludges. Environ. PoLL. 44, pp. 193-2lO. 1987. Sims. R.W. and B.M. Gerard. Earthwormd, Linnean Society Synopses of the British Fauna (Ne'w Series) No. 31. London and Leiden. E.J. Brill/Dr. W. Backhuys. 1985. Stevenson. F.J. HumUd ChemiJtryj GenNiJ, CompOdition, Reactiond. John Wiley & Sons. Canada. 1982. Ch. 14. Temminghoff. E.J.M.• S.E.A.T.M. Van der Zee. and M.G. Keizer. The influence of pH on the desorption and speciation of copper in a sandy soil. Soil Sci. 158. pp. 398-408. 1994. Temminghoff. E.J.M.• S.E.A.T.M. Van der Zee. and F.A.M. de Haan. Copper mobility in a copper contaminated sandy soil as affected by pH. solid and dissolved organic matter. Environ. Sci. Techno!. 31(4). pp. 1109-1115. 1997. Tipping. E .• A. Fitch. and F.J. Stevenson. Proton and copper binding by humic acid: application of a discrete-site/electrostatic ion-binding model. Eur. J. Soil Sci. 46. p. 95. 1995. Van Dobben. H.F .• J. Mulder. H. Van Dam. and H. Houweling. In Impact of AcwAtmodphm; Depodition on the BiogeochemiJtry of Moorland Poou and Surrounding Terredtrial Environment. Pudoe Scientific Publishers. Wageningen. 1992. Chapter 2. Van Riemsdijk. W.H. Keynote Lecture. 15th World CongrNJ of Soil Science. Acapulco. Mexico; The International Society of Soil Science. Madison. WI. Vol. 1, 1994. p. 46.
CHAPTER 7
Selenium Speciation in Soil Water: Experimental and Model Predictions Katta J. Reddy
INTRODUCTION Selenium (Se) occurs naturally in soils. The main geological source of Se in soils is cretaceous shales. The common range ofSe in soils is between 0.01 and 2 mg/kg-1 (Lakin, 1972). However, in seleniferous soils Se concentrations can be as high as 1200 mg/kg- 1 (Adriano, 1986). Selenium is a required micronutrient for humans and animals. Its requirement, however, for plants is not clearly understood. Human activities introduce Se into soils in many ways. These include burning fossil fuels (coal), disposal of coal combustion by-products, mineral extraction activities, and application of fertilizers (Nriagu, 1989). Selenium as a naturally occurring element is gaining national and international attention because of its potential deficiency and toxicity problems to humans and animals. For example, in China two types of Se human diseases, cardiomyopathy (Se deficiency) and selenosis (Se toxicity) were reported (Yang et al., 1983). In another case, disposal of agricultural drainage water into wetlands of Kesterson National Wildlife Refuge in California, caused bioaccumulation of Se by plants, fish, waterfowl, and animals at levels that were harmful (Ohlendorf, 1989). In soil water Se may exist in different oxidation states. These include Se (+6), Se (+4), Se (0), and Se (-2). Among these, the Se (+6) and Se (+4) oxidation states are thermodynamically stable under the pH and redox conditions that are found in most soils (Elrashidi et al., 1987). However, in low redox environments Se (0) and Se (-2) species may be expected. The Se (+6) and Se (+4) oxidation states in soil water may be comprised of free ions and complexes. These include SeO/-, HSe04-' H 2SeO/, CaSe04o, MgSe040 and SeOl-, HSe03-' H2Se03o, CaSe030, and MgSe03o. Additionally, soil water contains dissolved organic carbon (DOC) due to the plant, animal, and biological activity; therefore, DOC-Se complexes are expected. Very little information exists on the speciation of dissolved Se in soil water because it is difficult to separate Se (+6) and Se (+4) oxidation states without destroying their 147
148
Fate and Transport of Heavy Metals in the Vadose Zone
natural distribution. However, research in surface and groundwater suggest that dissolved Se consists of not only Se042- and Se032- but also metal-Se complexes and DOCSe complexes (Siu and Berman, 1989; Cooke and Bruland, 1987; Tanzer and Heumann, 1991; Wang et al., 1994; Reddy et al., 1995a). Similarly, we can expect different dissolved Se species in soil water because soil water contains higher ionic strength than surface water or groundwater due to an increased concentration of dissolved salts. Thus, isolation, extraction, and measurement of dissolved Se species in soil water are important. Such information may help in predicting the fate (availability, toxicity, adsorption, and precipitation) and transport (mobility) of dissolved Se species in soil vadose zones (Reddy, 1998). To date, there is little documentation on the quantification and model verification of dissolved Se speciation in soil water. Therefore, in this chapter we review procedures proposed for the speciation of dissolved Se, compare experimentally measured dissolved Se speciation with a model prediction, and identiry further research needs in the speciation of Se in soil water. The emphasis will be placed on SeOl- because it is predominant and more toxic than SeOl- in natural environments.
SPECIATION OF DISSOLVED Se The dissolved Se speciation in an aqueous solution can be performed with analytical methods and/or geochemical models. Several methods including hydride generation atomic absorption spectrometry (HG-AAS), fluorometry, (FM), high pressure liquid chromatography (HPLC), and ion chromatography (IC) are available for the speciation of Se. Among these methods, HG-AAS is the most commonly used method for the speciation of dissolved Se in aqueous solutions because it can detect as low as 1 pg L- 1 of Se. The HG-AAS method measures dissolved Se in an aqueous solution as Se (+4), from which Se (+6) and DOC-Se can be measured by altering the pretreatment steps as described below (Cutter, 1978; Workman and Soltanpour, 1980). The concentration of Se (+4) in a sample is measured by generating H 2Se with a NaBH4 solution and 7 M HCl (undigested). Another aliquot of sample is heated for 20 minutes at 85°C with 7M HCl to reduce Se (+6) to Se (+4). The concentration of Se in this solution is considered as the sum of Se (+6) and Se (+4) (digested). Difference between the concentration of Se in digested and undigested samples is considered as the concentration of Se (+6). The undigested or digested samples do not include DOC-Se. The total Se concentration in an aqueous solution is measured by oxidizing organic matter with H 20 2 for 20 minutes at 85°C and then digesting with 7M HCl for another 20 minutes at 85°C. This total Se is the sum of Se (+6), Se (+4), and DOC-Se. The difference between total Se and digested Se is considered as DOC-Se species. However, the concentrations of Se (+6) and Se (+4) in soil water, determined with HG-AAS, may consist of SeOl- and SeOl- species and their solution complexes. If SeOl- and Se032- can be isolated and extracted directly from soil water, then they can be measured using HG-AAS. A geochemical model (e.g., MINTEQA2, GEOCHEM, WATEQFC) could be used to calculate the speciation of dissolved Se in soil waters. Chemical data such as dissolved concentrations of cations and anions, pH, and redox potential of the soil water are required by these models to calculate the chemical speciation (i.e., activity or concentra-
Selenium Speciation in Soil Water: Experimental and Model Predictions
149
tion of free ions and complexes) by solving a series of mass-balance equations through an iterative procedure. However, models are based on the assumption of equilibrium; therefore, soil water should be close to a steady-state condition. The mass-balance equations for each dissolved species should contain all possible solution species to ensure accurate calculation of the speciation; omission of any significant solution species from the mass-balance equation will cause overestimation of the activity of dissolved free ions. Reported values for the equilibrium constants for solution species might vary and the constants for some species that may be present are not known, thus the species cannot be included in the model. All these factors could lead to the misinterpretation of the speciation of dissolved chemicals in soil water. Excellent discussions on this topic are presented by Amacher (1984), Baham (1984), and Sposito (1994). Few studies have examined the speciation of dissolved Se in soil water (See et aI., 1995; Fio and Fujii, 1980; Reddy et aI., 1995a), despite its importance in understanding Se solubility, availability, toxicity, and mobility. Reddy et al. (l99Sa) reported that increasing the pH of CuCl2 solution containing SeOl-, SeOl-, and sulfate (SOl-) by adding NaOH precipitates cupric oxide (CuO). The zero point of charge (ZPC) for CuO occurs at pH 9.S. As illustrated in Figure 7.1, CuO particles adsorb SeOl- and SeOl- at pH 6 and desorb them at pH 13. Reddy et al. (l99Sa) successfully used this phenomenon and isolated SeO42- and Se032- from groundwater samples containing and DOC concentrations greater than 10,000 and SO mglL, respectively.
sol-
EXPERIMENTAL AND MODEL PREDICTIONS Dissolved Se Speciation with CuO The procedure for the comparison of experimentally measured Se speciation in soil water with model predictions is outlined in Figure 7.2. Soil samples were extracted with distilled-deionized H 20 after reacting for 24 hours on a mechanical shaker at 200 rpm under the laboratory temperature. Each soil H 20 suspension filtered and each filtered sample was divided into two subsamples. One subsample was analyzed for Se (+6), Se (+4), and DOC-Se with the HG-AA in addition to major cations and anions, as well as pH (Figure 7.2). The other subsample was used for extracting SeOl- and SeOl- with the CuO particles. The experimental procedure to isolate and extract SeOl- and SeOl- species from soil water involved adding CuO particles to solutions and lowering the pH to 6.0 and reacting for 4 hours, followed by separating the CuO particles from solution and increasing the pH to 12.S and analyzing solutions for SeOl- and SeOl- by HG-AAS. The concentrations of metal-SeO/- and metal-SeOl- complexes in soil water were calculated by subtracting the concentrations of Se042- and SeOl- from Se (+6) and Se (+4) concentrations, respectively. From SeOl-, SeOl-, metal-SeOl-, metal-Se032-, and DOC-Se, the speciation of dissolved Se in soil water samples was calculated.
Dissolved Se Speciation with GEOCHEM As discussed earlier, several geochemical models including GEOCHEM and MINTEQA2 are available to model the speciation of dissolved trace elements in soil
150
Fate and Transport of Heavy Metals in the Vadose Zone
pH=6.0 pH- 13.0 Aqueous Solution Figure 7.1. Illustration of seO/- and SeO/- adsorption and desorption mechanism by the (uO particles in aqueous solution.
water. For this research the program GEOCHEM was used to calculate the speciation of dissolved Se (free ions and metal-SeOl- and metal-SeO/- complexes) in soil water because it computes the highest number of solution species. The pH and the concentrations of cations and anions were used as input to the model to calculate the speciation. For Se input, Se (+6) and Se (+4) concentrations (determined with HG-AAS) were used without redox potential because SeOl- and SeOl- reduction and oxidation reactions are very slow (Reddy et al., 1995b). From the input of pH and concentration of Se (+6), Se (+4), cations, and other anions, GEOCHEM calculates the concentration of free ionic species (e.g., SeOl- and Se032-) and metal-SeO/- and metal-SeO/- complexes using the thermodynamic data of solution species. The thermodynamic data used to calculate metal-Se042- and metal-SeO/complexes in soil water are shown in Table 7.1. The concentration of free SeOl- and SeO/- ionic species and metal-SeO/- and metal-SeO/- complexes predicted by the GEOCHEM were compared with the CuO extraction method (Figure 7.2).
Comparison Selected chemical data of soil water, which are used for the discussion, are presented in Table 7.2. The pH of soil water ranged between 5.8 and 8.4. Total dissolved Se concentrations were between 11 and 162 pg L- 1• These concentrations are well below the limit of quantification of Se analysis with IC (Blaylock and James, 1993). Soil water 1 contained high concentrations of dissolved Mg, Na, and DOC when compared with other soil water samples. Soil water 1, 2, 5, and 6 contained higher concentrations of dissolved Ca than soil water 3 and 4. Dissolved SO/- concentrations in soil water ranged between 15 and 1666 mg L- 1• Dissolved Se analyses with HG-AAS are shown in Figure 7.3. It should be noted that concentrations of Se (+6) and Se (+4) also include SeO/- and SeO/- species plus metal complexes. The soil waters examined in this study were dominated by Se (+6) concentrations. The Se (+6) concentrations ranged between 5 and 136 pg L- 1, whereas Se (+4) concentrations ranged between < 1 and 7 pg L -I. The DOC-Se concentrations were found between 1 and 19 pg L- 1• Similar distribution for dissolved Se species was observed by Fio and Fujii (1990) in soil water from California. Results from the isolation and extraction of SeOl- and SeO/- with CuO are presented in Figures 7.4 and 7.5. These results suggest that the removal of SeO/- ranged between 70 and 83%, except soil water 4, when compared with Se (+6) concentrations
Selenium Speciation in Soil Water: Experimental and Model Predictions
Analyze pH, CatioDl, AniODS, Se (6+), Se (+4), and DOC-Se
151
Isolate and Extract Selenate and Selenite with CuO
Se Speciation with GEOCHEM Input pH, Cations, ADions Se (+6), and Se (+4)
Se Speciation with COO Determine Metal Selenate Complexes Selenite Complexes
Se Speciation Selenate and Selenite Ions Metal Selenate and Selenite Complexes
Se Speciation Selenate and Selenite Ions Metal Selenate and Selenite Complexes
Figure 7.2. Procedures for the dissolved Se speciation comparison between the CuO/HG-MS and geochemical modeling.
Table 7.1. Metal-SeO/- and Metal-SeO/- Complexation Reactions used in the Thermodynamic Database of GEOCHEMa No.
1 2 3
4
Reaction Ca 2+ + SeD 42- Mg2+ + SeD 4 2- Ca 2+ + SeD 3 2- Mg2+ + SeD 3 2- -
J(
CaSeD 40 MgSeD4 0 CaSeD 30 MgSeD 30
102 .8 102 .4 104.2 105.0
a Sposito and Mattigod, 1980.
(Figure 7.4). The Se analysis of the Cu 0 supernatant solutions suggested that 17 to 30% of Se042-was left in the solutions. However, for Se032-the removal rate is 100%, except soil water 2, when compared with Se (+4) concentrations. For soil water 5 and 6, Se032concentrations were below the detection limit of 1 pg L- 1 (Figure 7.5). These results also suggest that other anions such as and DOC did not interfere in the SeOi- and Se032- removal process by CuO particles. If S042- and DOC compete with SeOi- and Se032- for adsorption sites, one would expect no adsorption of these species by the Cu 0 particles, because the ratio of S04 2- and DOC to Se is very high. The results also suggest that metal-SeOl- complexes are not significant. The observed 70 to 83% removal of SeOi- by the CuO particles from the soil water could be due to the presence of other Se (+6) species (e.g., MgSe040, CaSe040), which may not be adsorbed by the CuO particles (Reddy and Gloss, 1993). For example, Giordano et al. (1983) showed that formation of neutral complexes (e.g., CdCI2°) lowers
sOi-
152
Fate and Transport of Heavy Metals in the Vadose Zone ""---,,--,-------
-"'-""""'---------."---
"
~-------
----------
"
--------
Table 7.2. Selected Chemical Data of Soil Water (SW)a Parameter pH Calcium Magnesium Sodium Potassium Sulfate DOC Selenium ~g L- 1
SWl
SW2
SW3
SW4
SW5
SW6
7.5 188 922 745 47 1666 144 162
7.9 214 48 31 14 166 60 21
8.4 58 26 36 16 16 33 13
8.0 39 26 29 7.2 15 24 11
6.1 243 63 18 14 960 19 69
5.8 275 76 31 9.4 1130 7.6 120
a Units are mg L-1. Data for SW2, SW5, and SW6 adapted from Reddy (1998). Reprinted
with permission of John Wiley 8- Sons, Inc.
Legend
-",*--
Se(+6) Se(+4)
.............
DOC-Se
.i
1 ~
..............
....
o - - - . ::::.:::.~-.=.'"".:::-.::-:-:...==I................-...-...-...-...-.. ..... SW1
SW3
SW4
SW5
SW6
Soil Water Samples Figure 7.3. Dissolved Se concentration in soil water as measured by HG..AAS.
the concentration of Cd2+, which decreases the adsorption and increases the mobility of Cd in sewage sludge amended soils (see also Mattigod et aI., 1979; Bowman and O'Connor, 1982; Elrashidi and O'Connor, 1982). There may be a number of reasons why CuO particles adsorb both SeOi- and SeOlin the presence of other ions; however, the most possible reasons include: • On a time scale, the metal-SeO/- and metal-Se032- complexation reactions in aqueous solutions are much faster than adsorption reactions of these species by the CuO particles. Also, SeOl- and Se052- adsorption reactions by the CuO particles are much faster than reduction and oxidation of these species in aque-
Selenlul1l Speciation in Soil Water: Experimental and ModeJ Predictions
D
Initial
•
CuOMethod
SoN3
SlN4
Soil Water Samples Figure 7.4. Extraction of Se (+6) from soil water with CuO method.
SW1
D
Initial
•
CuOMethod
SW2.
Soil Water Samples Figure 7.5. Extraction of Se (+4) from soil water with CuO method.
153
Fate and Transport of Heavy Metals in the Vadose Zone
154 -
-----,-,,--
""--"""---~--.'"-"---"
-".,----""'----
---~----"'-""-,~--
""~--,--'"
-
""
---"'''''"-----,."---
Table 7.3. Speciation of Se in Soil Water with CuO/HG-AASa Species pH SeO/Se0 32 Metal-SeO/DOC-Se Total a
b
SW1
SW2
SW3
SW4
SW5
SW6
7.5 59 4 25 12 100
7.9 43 19 19 19 100
8.4 38 38 8 16 100
8.0 46 46 NSb 9 100
6.1 58 NS 15 27 100
5.8 72 NS 22 6 100
Units are %. Data for SW2, SW5, and SW6 adapted from Reddy (1998). Reprinted with permission of John Wiley a Sons, Inc. NS=not significant <1 %.
ous solutions. For example, less than 30 minutes are required for the adsorption of Se species by the CuO particles in aqueous solutions, whereas Se oxidation and reduction reactions take much longer times (Reddy et al., 1995b; Reddy, 1998). • The CuO particles are stable in the pH range of 6 and 13 and show low affinity for SO/-. The speciation of dissolved Se concentrations in soil water samples by CuO/HG-AAS is presented in Table 7.3. These results show that total dissolved Se concentrations were dominated by SeO/- (38 to 72%). The concentration of SeOl- was between d to 46%. The DOC-Se complexes ranged between 6 and 27% and the concentrations of metalSeO/- complexes were between d to 25%. The DOC-Se concentrations can be further separated into hydrophobic base, acidic, and neutral fractions by following the procedure as outlined by Wang et al. (1994). Comparisons between the CuO method and GEOCHEM calculations for SeO/- and metal-SeO/- complexes are shown in Figures 7.6 and 7.7. The agreement between these two methods for SeO/- and metal-SeO/- complexes is excellent. These results strongly suggest that the CuO particles extracted only free SeO/- ions from the soil water. The accurate determination of SeO/- and metal-SeO/- complexes in soil water is more important than SeOl-, because SeO/- is more toxic than SeOl- (Page and Bingham, 1986). Additionally, studies have shown that SeO/- is more mobile than Se032- in soils (Ahlrich and Hossner, 1987; Balistrieri and Chao, 1987) because it is not adsorbed by the soil mineral phases (Neal and Sposito, 1989). The agreement for Se032- and metal-SeOl- complexes between the two methods was poor. The CuO procedure predicted metal-SeOl- complexes as insignificant, because almost 100% of Se (+4) was extracted from soil water (see Figure 7.5). GEOCHEM predicted free ion Se032- concentration as less than 1% of the total Se (+4) concentration and 99% as metal-SeOl- complexes. One possible explanation for this may be due to the weak strength of metal-SeOl- complexes, which might have dissociated during the Cu 0 extraction process. Nevertheless, both experimental and modeling studies on dissolved Se strongly suggest that soil water may consist of not only SeO/- and SeOl- ionic species but also metal-SeO/- and SeOl- complexes. The results in this study also suggest that SeO,/-
Selenium Speciation in Soil Water: Experimental and Model Predictions
155
0.16 r----7"'"---------"'II::-----,
Legend
o
0.14
SW1
Model
SW2
ExperimeDtal
•
SW3
SW4
SW5
SW6
Soil Water Samples Figure 7.6. Comparison between CuO method and GEOCHEM for seol- concentrations in soil water.
60
Legend Model
50
•
ExperimeIIIal
"0'
..
"-'
j
40
! ~
~ CIl
1 SWI
SW2
SW3
SW4
SW5
SW6
Soil Water Samples
Figure 7.7. Comparison between CuO method and GEOCHEM for metal-SeO/- complexes in soil water.
may form strong complexes with divalent cations. In contrast, Se032- may form weak metal complexes in soil water. These results have important implications for our understanding of Se biogeochemistry in soils. One such implication is that the formation of strong metal-SeOl- complexes in soil water may increase their potential mobility in the soil vadose zone. This may be the one of the reasons why dissolved SeOl- concentra-
156
Fate and Transport of Heavy Metals in the Vadose Zone
tions in soil water, surface water, and groundwater often dominate the dissolved Se concentrations.
FUTURE RESEARCH Currently, speciation of dissolved Se in soil water is not well understood. This is partially due to the difficulty involved in separation of Se (+6) and Se (+4) species without disturbing their natural distribution. However, quantitative description of dissolved Se speciation is required for predicting the fate and transport of Se in the soil vadose zone. Selective Se adsorption mechanisms by the CuO particles provided an approach for the speciation of dissolved Se in soil water. Our results suggest that a significant portion of dissolved Se in soil water is comprised of SeO/- and metal-SeO/- complexes, which may not be adsorbed by the mineral surfaces, and thus move along with water flux in the soil vadose zone. Further research to separate the dissolved Se associated with hydrophilic base, acidic, and neutral fractions is needed. Finally, research to determine the stability of metal-dissolved Se and DOC-Se complexes in soil water will be invaluable.
REFERENCES Adriano, D. C. Trace Element" in the TerredtriaL Environment. Springer-Verlag, New York, pp. 329361, 1986. Ahlrich, J.S. and L.R. Hossner. Selenate and selenite mobility in overburden by saturated flow. J. Environ. QuaL. 16, pp. 95-98, 1987. Amacher, M.C. Determination of ion activities in soil solutions and suspensions: Principal limitations. SoiL Sci. Soc. Am. J. 48, pp. 519-524, 1984. Baham, J. Prediction of ion activities in soil solutions: Computer equilibrium modeling. Soil Sci. Soc. Am. J. 48, pp. 525-531, 1984. Balistrieri, L. and T. Chao. Selenium adsorption by goethite. Soil Sci. Soc. Am. J. 51, pp. 11451151, 1987. Blaylock, M.J. and B.R. James. Selenite and selenate quantification by hydride generationatomic absorption spectrometry, ion chromatography, and colorimetry. J. Environ. Qua!. 22, pp. 851-857, 1993. Bowman, R.S. and G.A. O'Connor. Control of nickel and strontium sorption by free metal ion activity. Soi! Sci. Soc. Am. J. 46, pp. 933-936, 1982. Cooke, T.D. and K.W. Bruland. Aquatic chemistry of selenium: Evidence of biomethylation. Environ. Sci. Techno!. 21, pp. 1214-1219, 1987. Cutter, G.A. Species determination of selenium in natural waters. Ana!ytica Chemica Acta. 98, pp. 59-66, 1978. Elrashidi, M.A., D.C. Adriano, S.M. Workman, and W.L. Lindsay. Chemical equilibria of selenium in soils: A theoretical development. Soil Sci. 144, pp. 141-152, 1987. Elrashidi, M.A. and G.A. O'Connor. Influence of solution composition on sorption of zinc by soils. Soi! Sci. Soc. Am. J. 46, pp. 1153-1158, 1982. Fio, J.L. and R. Fujii. Selenium speciation methods and application to soil saturation extracts from San Joaquin Valley, California. SoiL Sci. Soc. Am. J. 54, pp. 363-369, 1990. Giordano, P.M., A.D. Behel, J.E. Lawrence, J.M. Soileau, and B.N. Bradford. Mobility in soil and plant availability of metals derived from incinerated municipal refuse. Environ. Sci. Techno!. 17, pp. 193-198, 1983.
Selenium Speciation in Soil Water: Experimental and Model Predictions
157
Lakin, H.W. Selenium Accumulation in Soils and Its Absorption by Plants and Animals, in GeochemicaL Environment in Re&ztion to HeaLth and DifeaJe, H.L. Cannon and H.C. Hopps., Eds., Geological Society of America, Boulder, CO, 1972. Mattigod, S.V., A.S. Gibali, and A.L. Page. Effects of ionic strength and ion-pair formation on the adsorption of nickel by kaolinite. C&zyd C&zy Mineral!. 27, pp. 411-416, 1979. NeaL RH. and G. Sposito. Selenate adsorption on alluvial soils. Soil Sci. Soc. Am. J. 53, pp. 7074, 1989. Nriagu, J.O. Global Cycling of Selenium, in Occurrence and Diftribution of Selenium, M. Ihnat, Ed., CRC Press, Boca Raton, FL, 1989. Ohlendorf, H.M. Bioaccumulation and Effects of Selenium in Wildlife, SeLeniwn in AgricuLture and the Environment, in L.W. Jacobs, Ed., Soil Science Society of America, Madison, WI, 1989. Page, A.L. and F.T. Bingham. Availability and Phytotoxicity of Selenium to Crops in Relation to Chemical Form and Concentration, in 1985-86 TechnicaL Progre.J.J Report, KK Tanji, Ed., Univ. of California Salinity Drainage Task Force, University of California Davis, California, 1986. Reddy, KJ. and S.P. Gloss. Geochemical speciation as related to the mobility of F, Mo and Se in soilleachates. AppL. Geochem. SuppL. I.J.Jue. 2, pp. 159-163, 1993. Reddy, KJ., Z. Zhang, M.J. Blaylock, and G.F. Vance. Method for detecting selenium speciation in ground water. Environ. Sci. TechnoL. 29, pp. 1754-1759, 1995a. Reddy, KJ., M.J. Blaylock, G.F. Vance, and RB. See. Effects of Redox Potential on the Speciation of Selenium in Groundwater and Coal Mine Backfill Materials, in Decade.J Later: A Time for ReaJde.Jdment: Selenium: Mining, Rec&zmation and EnvironmentaL Impactd, G.E. Schuman and G.F. Vance, Eds., ASSMR, Princeton, WV, 1995b. Reddy, KJ. Selenium Speciation in Natural Waters, in Encyclopedia of EnvironmentaL AnaLYdif and Remediation, RA. Meyers, Ed., John Wiley & Sons, Inc., New York, 1998, pp. 4291-4300. See, RB., KJ. Reddy, G.F. Vance, A.A. Fadlelmawla, and M.J. Blaylock. Geochemical Processes and the Role of Natural Organic Solutes on the Solubility of Selenium in Coal Mine Backfill Aquifers, Powder River Basin. FinaL Report, Ahandoned CoaL Mine &zndd Re.Jearch Program, Office of Surface Mining, Denver, CO, 1995. Siu, K W.M. and S.S. Berman. The Marine Environment, in Occurrence and Diftribution of Selenium, M. Ihnat, Ed., CRC Press, Boca Raton, FL, 1989. Sposito, G. and S.V. Mattigod. GEOCHEM: A Computer Program for the Calculation of Chemical Equilibria in Soil Solutions and other Natural Water Systems. The Kearney Foundation of Soil Science, Univ. of California, Riverside, 1980. Sposito, G. ChemicaL Equilibria and Kineticd in Soil!: Chapter 2. Oxford University Press, New York, 1994. Tanzer, D. and KG. Heumann. Determination of dissolved selenium species in environmental water samples using isotope dilution mass spectrometry. AnaL. Chem. 63, pp. 1984-1989, 1991. Wang, D., G. Alfthan, and A. Aro. Determination of total selenium and dissolved selenium species in natural waters by fluorometry. Environ. Sci. TechnoL. 28, pp. 383-387, 1994. Workman, S.M. and P.N. Soltanpour. Importance of prereducing selenium (VI) to selenium (IV) and decomposing organic matter in soil extracts prior to determination of selenium using hydride generation. SoiL Sci. Soc. Am. J. 44, pp. 1331-1333, 1980. Yang, G., S. Wang, R Zhou, and S. Sun. Endemic selenium intoxication of humans in China. Am. J. CLin. Nutr. 37, pp. 872-881, 1983.
CHAPTER 8
Influence of Reducing Conditions on the Mobility of Divalent Trace Metals in Soils Philippe Cam bier and Rayna Charlatchka
INTRODUCTION The behavior of trace elements in soils has been the subject of many studies at first related to plant nutrition, and now often justified by the ecotoxicologic question raised by their possible accumulation in soils and in the food chain. The importance of their bioavailability remains, whatever trace elements are considered as nutrients or pollutants. It is also important to consider their mobility, which can be defined as the ability to change their speciation, and more concern is about mobility toward soluble or gaseous species. In fact, many trace elements encountered as soil pollutants, particularly heavy metals, are generally little mobile; then concern is more about the long-term behavior of elements likely to accumulate in soils, or already concentrated in polluted sites. The main factor that influences the behavior of trace elements and which can change with time and soil use is the pH. Other important factors such as soil texture and constitution appear more constant; however, physical properties, organic content, and redox conditions, can also quickly change, or be changed. With respect to this latter influence, trace elements must be separated between at least two groups. Some elements can change their oxidation number in common soil conditions; they are mainly those occurring as stable oxoanions (As, Mo, Se, V, etc.), Cr and Hg. The influence of redox conditions on their speciation appears direct, and they will not be considered here. Other environmentally important trace elements are divalent in common conditions. For them, reducing conditions induce indirect effects on their speciation and mobility, through changing pH and soil constituents. For example, Fe and Mn oxides, and organic compartments are affected by reducing conditions and affect the behavior of all trace metals. Published studies on this topic often present overall effects and hypothetical mechanisms, so that little consistency is obtained: the mobility or the bioavailability of Cd, Zn, etc., may increase, or decrease, with the redox potential. The present contribution is to review and illustrate the redox processes which influence the mobility of 1::0
160
Fate and Transport of Heavy Metals in the Vadose Zone
divalent trace metals in soils, and to propose a few qualitative rules for understanding and prediction.
CONTROVERSIAL STUDIES ON SOIL-PLANT SYSTEMS Forno et al. (1975) showed that Zn uptake by rice was generally lowered during the first weeks of flooding, without relationship to total soluble Zn which can increase at the same time. Reddy and Patrick's (1977) experiments on soil suspensions showed that soluble Cd and total uptake of Cd by rice decreased when redox potential Eh decreased, while the opposite was observed for Pb. Iu et al. (1981) found Zn and Cu less soluble and exchangeable when Eh decreased in a waterlogged brown soil; however, the variation of their availability to plants depended on the species (Iu et al., 1982). These authors concluded that Co, like Mn and Fe, but not Cu and Zn, become more mobile and available under reducing conditions. Bjerre and Schierup (1985) also reported effects of waterlogging on the chemical extractability and the plant availability of Zn, Cd, Pb, Cu, in 3 different soils: exchangeable Zn increased while the exchangeable fraction of the others remained very low; however, the total plant uptake of these trace metals generally decreased, although roots and leaves showed different patterns. Again from chemical extractions, applied to flooded rice soils after adding organic matter, Ghanem and Mikkelsen (1987) found that exchangeable Zn decreased with the redox potential while Zn bound to oxides increased; curiously, Zn estimated bound to organic substances also decreased. Using the same approach on an acid red soil loaded with Cd sulfate, Xiong and Lu (1993) found Cd less exchangeable after flooding. Ahumada and Schalscha (1993) applied chemical extractions to near neutral soils spiked or not with Cd and Cu, after incubating soil suspensions in different conditions; they found that Cd extractability varied with Eh, whereas no marked influence was observed for Cu. Brown et al. (1989), performed flooding experiments on Luvisol and Cambisol with or without sludge amendment; they concluded that reducing conditions generally increase or have little effect on the phytoavailability and solubility of Cd and Ni, depending on plant species and metal source. Iu et al. (1981) interpreted that changing Fe and Mn oxides by reduction can increase their adsorptive capacity for Cu and Zn. Ghanem and Mikkelsen also interpreted that reduction can generate more amorphous solids that were able to sorb more Zn. Mter Brown et al. (1989), the dissolution of Fe and Mn compounds can induce the release of divalent trace metals in the aqueous phase but also the competition with them for plant assimilation. Bjerre and Schierup (1985) added to these mechanisms that root functioning is affected when Eh and oxygen decrease. MandaI et al. (1987, 1988) attributed their observations to the reductive dissolution of specific sorbents of Zn. It is often estimated that reductive dissolutions of Fe and Mn oxides can release trace metals (Me Bride, 1989; Reddy and Patrick, 1977, for Pb). Alloway (1995) reported after Bingham et al. (1976), that Cd uptake by rice from contaminated paddy soils is strongly decreased when conditions are anoxic, due to precipitation with sulfide. He also estimated that gleying processes could decrease adsorptive capacities of Cd due to lowering contents of Fe and Mn oxides. In the same book, McGrath (1995) recalled that a good relationship exists between exchangeable Ni, i.e., extracted by unbuffered salt solutions, and phytoavailable Ni, and that both generally increase in poorly drained soils. Kiekens (1995)
Influence of Reducing Conditions on Mobility of Divalent Trace Metals in Soils
161
hardly considered the effect of reducing conditions on Zn behavior in soil-plant systems, except to recall that much available Fe is antagonistic to Zn plant uptake. The above review suggests that reducing conditions indirectly affect the mobility and availability of divalent trace metals through several mechanisms: formation of insoluble sulfides, release from sorption sites of reduced-dissolved Fe and Mn oxides, or increased fixation by them after alteration, displacement of divalent trace metal from any exchange site by increasing soluble Mn2+ and Fe2+, and antagonisms between the later and the other cations for plant uptake. The last mechanism concerns only bioavailability of metals; the two later ones make the concentration of trace cation in the soil solution and its plant availability vary in opposite directions: high concentrations of dissolved Fe 2+ or Mn2+ should increase concentrations of, e.g., Cd2+, Zn 2+ in the soil solution but should compete against their uptake by plants. The first conclusion is that it is difficult to clear up at the same time the effects of reducing conditions on the solubility of divalent metals, and on their bioavailability. So from this point, we will focus more on the chemical mobility and processes than on bioavailability. The attempts to follow the changing solid speciation of trace metals by chemical extractions encounter the usual limitations of this method (e.g., Tessier and Campbell, 1991) and a few others, as the possible occurrence of two oxidizable compartments, i.e., organic substances and sulfides. Another general chemical approach consists in studying the possible solubility and redox equilibria. This approach has been more often applied and with more details to Fe and Mn. However, some important works concerned trace divalent metals.
FORMATION OF INSOLUBLE SULFIDES AND OTHER SOLUBILITY EQUILIBRIA When S2- activity rises, formation of insoluble sulfide is expected with "heavy metals," "soft," and "borderline" cations according to the HSAB theory (Sposito, 1981). This solubility mechanism, which can take place in strongly reducing conditions, was often studied with sediments, and also in particular cultivated systems, i.e., paddy soils. Sulfide formation can occur at redox potential between -0.1 and -0.2 V at near neutral pH (Ponnamperuma, 1972; Patrick and Jugsujinda, 1992), i.e., pE around -3 (Lindsay, 1979; Sposito, 1989). pH also determines S2- activity since H 2S dissociates at pH 7. Bingham et al. (1976) grew rice on soil amended with CdS04 enriched sludge, flooded or not, and found Cd less available and much less soluble at the end of flooded culture; at the same time, sulfate became undetectable in solution of flooded samples so that the results were explained by the formation of sulfide. Huaiman (1984) determined the ion activity products of Cd and S2-, and other solutes, at equilibrium with anaerobic soil suspension and concluded that CdS rather than CdC03 controlled the solubility of this metal. Recently, Brennan and Lindsay (1996) with soils polluted by metallic ores, and Morse and Arakaki (1993) with synthetic samples, have shown the importance of CdS, CuS, NiS, PbS, ZnS, with respect to the solubility of heavy metals in highly reducing conditions. They also underline the importance of ferrous sulfides either as sorbents for these heavy metals or because they control S2- activity. Under intermediate reducing conditions, obtained by the former authors by adding only organic matter to the soil, solubilities of Cd, Pb, Zn, appeared controlled by carbonates or by reactions with iron oxides.
162
Fate and Transport of Heavy Metals in the Vadose Zone
The enhanced mobility of trace metals from oxidizing sulfides is clear, because this oxidation and the precipitation of Fe (III) induce acidification, and sulfate compounds are much more soluble than sulfides. This was often observed for sediments (e.g., Forstner, 1987; Tack et al., 1996; Altmann and Bourg, 1997). After Tack et al. (1996), the solubilities of Cd, Cu, Pb, Zn, remain very low when sediments are kept flooded and reduced, due to the presence of sulfides, but Co, Ni, and Mn seem always associated to carbonates. Mter oxidation and induced acidity, the increasing solubilities of Cd, Cu, Pb, Zn, become controlled by iron oxides and possibly organic substances. Other works, without particular attention to reducing conditions, put in evidence the role of phosphate in limiting the solubility of at least Pb (e.g., Joponyand Young, 1994; Lindsay, 1979), and the role of Zn silicates at basic pH (BrUmmer et al., 1983). These ligands are not produced by reduction like S2-; however, indirect effects of reduction could be through the activity of these ligands which are known to be released by reductive dissolution of iron oxide (Balzer, 1982; Ponnamperuma, 1972). So it is difficult to separate the role of trace metal sulfides and the reactivity of Fe and Mn compounds-sulfides, and also oxides, carbonates (Forstner, 1987). Since Fe and Mn compounds are involved in trace metal mobility, it is important to recall briefly how they are ruled by precipitation-dissolution and redox phenomena. The solubility of Fe can be determined by sulfides at very low Eh. In moderately reducing conditions, the activity of Fe (II) and Fe (III) are controlled by pH and pE through iron oxide solubility and the Nerst equation (Balzer, 1982). However, mixed Fe(II)-Fe(III) oxides and carbonates can also impose the final "stable" concentrations (Ponnamperuma, 1972). Schwab and Lindsay (1983) from soil suspension studies found that Fe(I!) activity can be controlled by FeC03 (siderite) or Fe3(OH)s. Recently, Trolard et al. (1996) showed that another mixed solid, known when synthetic as green rust, can control Fe(II) in temporary hydromorphic soil horizons. The case of Mn appears still more complex with an important role played by intermediate species involving the less stable valency 3 of Mn (Bartlett and James, 1993); however, the more common situations deal either with poorly soluble, but very reactive, Mn(IV) oxides, in oxidizing conditions, or with Mn(II), whose the solubility is controlled by carbonate (Ponnamperuma, 1972; Balzer, 1982; Sadana and Takkar, 1988; Sposito, 1989; Bartlett and James, 1993). The very low activities, as for Fe 3+ at common pH, and S2- in presence of much Fe2+, often are not measured during the experiments, they are deduced from some chemical equilibria. However precipitation-dissolution of definite compounds seem well established mechanisms regarding the effects of Eh on trace metal mobility in soils. Kinetic studies can also put in evidence some mechanisms. Charlatchka and Cambier (1996; 1998) recorded the chemical evolutions of suspensions of polluted soil; when pH 6.2 is maintained, using only HCI without adding any oxidizing or reducing reagent, soluble Cd, Pb, Zn, seemed more related to Mn and Fe evolutions than to any other solutes; particularly, remained practically constant, so that it was unlikely that sulfide appeared, despite the low Eh (Figure 8.1).
sOi-
ROLE OF Fe AND Mn OXIDES AS TRACE METAL SORBENTS The solubilities of Fe and Mn are ruled by redox and precipitation-dissolution processes. The same consensus cannot be brought out for divalent trace metals except in the
Influence of Reducing Conditions on Mobility of Divalent Trace Metals in Soils
163
0,20 r--------------,-r:Z=-n-----, 1,6
Pb
0,15
1,2
~ ~
Cl
E
.ri D-
Cl
E
0,8
0,10
Cd
"0
c: as
c:
N
"0 (,)
0,4
0,05
0,00 4
0
12
8
0,0 16 400
10
b) ~~
300
8
Eh
Cl
.
E
200
0 en
>
"0
~
E
----
c: as c:
::::?:
4
..o----oMn
oj
100 w 0
u.
2
-100
0 0
4
8
12
-200 16
Time (days)
Figure 8.1. Evolution of solutes and redox potential in a soil:suspension (1:10) under nitrogen atmosphere, at pH 6.2 controlled by HCI addition (cultivated layer polluted by atmospheric industrial fallout).
case of S2- formation. The role of Fe(I!) in this case can be through controlling S2activity, or iron sulfides could act as specific sorbents. In less reducing conditions, it is often assumed that Fe and Mn oxides intervene as specific sorbents of trace elements. From several studies on soils or on soil-plant systems, it seems that divalent trace metals can be fixed back by amorphous or altered iron oxides under reducing conditions (Ghanem and Mikkelsen, 1987; Iu et al., 1981; Reddy and Patrick, 1977, for Cd). Stud-
164
Fate and Transport of Heavy Metals in the Vadose lone
ies carried out on more or less heterogeneous systems, like pot experiments, hardly sustained that reoxidation in part or whole of analyzed samples was impossible. So reoxidation-precipitation processes could explain partly decreased solubility, and increased "oxide fraction," under, or rather after reducing conditions. Many authors rather expect that reductive dissolution of Fe and Mn compounds decreases the soil sorption capacity of trace metals, but few presented detailed chemical data for Fe, Mn, and other metals. Working with synthetic systems, Francis and Dodge (1990) have shown that anaerobic processes can release heavy metals (Cd, Ni, Zn, and to a lesser extent Cr, Pb) trapped by goethites. Such a process should increase the solubility of divalent trace metals in soils as surely as it is ruled by adsorption on these oxides rather than by other reactions. Recently, Chuan et al. (1996), using a device similar to Patrick and Jugsujinda's and a sandy loam paddy soil highly contaminated by industrial waters, put in evidence that Eh has less influence than pH on the solubility of Cd, Cu, Pb, Zn, although lower Ehs do increase the solubility of these elements. The ranges of variation were [3.3, 8] and [0.33, -0.1 V] for pH and Eh, respectively. Chuan et al. (1996) reported the increasing solubility of Fe at one pH value (pH 3.3). Chuan et al. concluded that the main mechanisms involved were the reductive dissolution of ferric and manganic oxides. Charlatchka and Cambier (1998) incubated soil suspensions during similar periods and imposed only pH. Different Eh values were reached at equilibrium, depending on reagents added (HC!, HN0 3, or NaOH, and glucose between 0 and 3% by weight of soil). Dissolved Fe, Mn, Pb, Cd, and less systematically Zn, varied like -Eh (Figure 8.2), whereas Ca was practically constant.
REDUCING PROCESSES CHANGE pH The influence of pH on trace element solubility is well known. Either through solubility equilibria, or due to complexation by soluble and surface ligands, increasing pH, within an ordinary range, decreases solubility and bioavailability of divalent trace metals. Most environmental redox reactions involve protons. More precisely, considering half-redox reactions toward reduction, they consume H+ (e.g., Lindsay, 1979; Sposito 1981; Bartlett and James, 1993). However, only complete redox reactions actually occur, so that a right idea on the effect of reducing conditions on pH can be obtained only if main reduced and oxidized substances are identified and the main redox reaction is written. In order to write a possible redox process, 2 half-redox reactions can be added with the following conditions: • electron exchange must be balanced • the reaction must be favored thermodynamically • some physicochemical and biological conditions must be fulfilled. The first condition is easily fulfilled by adding 2 half-redox reactions where always exactly 1 electron is involved, like in many published lists, one toward reduction, the second toward oxidation. For example, nitrate reduction can be theoretically obtained by oxidation of organic carbon (written CH 20). This oxidation can be more or less complete, leading to different end-products, and to consumption or production of protons:
Influence of Reducing Conditions on Mobility of Divalent Trace Metals in Soils
-,------------
."
.
- - - - - --
- - -- - - - -
-----
""--""------~-
--,--
9
-
--"'" ,-
165
-,----" -----------._-" - - - - - - - -
-550 -300
6
>
~
-50
bO
e
-
3
e
Gi
200 0 U
U
~
~
<'!
~ 0
'+
U
U:::;-
~
~~
~
u
450
:s
~
+
6
-550 -300
--
4
~
-50
bO
e
2
>e
Gi
200 0 U
U
~
~
<'!
~
~
'+
210
U
U:::;-
~
~n~
~
u
.~
+~
Ca
--"-
-550
\
140
70
o
1\
450
:s
-300
Ik
r/
V~
-50
f\
:e Gi
200
1\
450
Figure 8.2. Concentrations of dissolved metals and Eh values at pH 6.2, in soil suspensions after 14 days following different treatments: +%C indicates glucose addition; 4 first treatments under nitrogen, 5th treatment under decarbonated air; pH controlled by NaOH or HN0 3 ; except 4th one (+HCI). Eh scale is reversed.
l66
Fate and Transport of Heavy Metals in the Vadose Zone -
--- ---,,- -----
0,3
-550 -300
......
0,2
>
~
-50
CI)
e
0,1
e
~
200 0
u ~
<'!
'+
u ~
"'~"
u
~
u:::~u
oX
+'-'
450 u
:E
~
0,2
-550
0,15
-300
0,1
-50
0,05
200
......
~
CI)
e
°
u
~
u:::-
~~
>e
~
450
:E
.~
3
-550 -300
......
2
>
~
-50
e1
e
~
200
°
u ~
<;
u:::~u
<;e
450 u
~..
Figure 8.2 (continued). Concentrations of dissolved metals and Eh values at pH 6.2, in soil suspensions after 14 days following different treatments: +%C indicates glucose addition; 4 first treatments under nitrogen, 5th treatment under decarbonated air; pH controlled by NaOH or HN0 3 ; except 4th one (+HCI). Eh scale is reversed.
Influence of Reducing Conditions on Mobility of Divalent Trace Metals in Soils
167
or
(reactions written for a moderately acid environment, where CO2 does not dissociate) Keeping the rules for writing redox reactions in mind, it appears that reductive dissolutions consume more protons than the other half-redox reactions. In other words, processes involving reduction-dissolution of Mn and Fe should increase pH, whatever the electron donor. On the contrary, when solid sulfides oxidize and oxidation overpasses the level of iron transformation, acidification is observed which is mainly due to oxidation of FeS2 (pyrite) and precipitation of ferric oxides. The last important redox reaction involving phase changes is the formation of CO 2 which partly sinks into Mn(II), Fe(II), or Ca carbonate solids; this also produces protons. Formation of CO2 does not occur only during aerobic processes since organic carbon can be the electron donor also under reducing conditions. The oxidation of carbon can be limited to the formation of some organic acids, which generally also tends to acidification. The effect of changing Eh on pH depends on the main complete redox reaction involved, on the reagents and the final products, and finally also on the average pH because of the acid-base reactions. About this last point, acid-base reactions between carbonate ions, or sulfide, ammonia, or organic species, Ponnamperuma (1972) estimates that the former play the major role in relation to the importance of respiration. This also appears from global microbial "respiration" equations written by Krumbein and Swart (1983). A slight acidification is often observed as a first step after flooding, which can be attributed to aerobic processes: biological activity, enhanced by moisture and using trapped O 2, produces CO 2. The next step uses nitrate as oxidant and soil organic matter. As long as N0 3- is present, no volatile organic acids and few ammonia form (Ponnamperuma), and no methane (Christensen and Kjelsen, 1989). Whatever the factors that determine the ratio N 21N20 (Firestone et aI., 1979; Smith et aI., 1983), the consumption of H+ for denitrification is the same (reaction 1 or 3 below; van Breemen et aI., 1983).
(3) The more reducing reactions involving N -organic N and not nitrate or nitrite-can lead to ammonia accumulation, the end-products being CO2, fatty acids, and possibly CH4 (Ponnamperuma, 1972). The exact balance of protons can vary. After denitrification, when reducing conditions go on, reductive dissolutions of Mn, then Fe, occur and, as already noticed, consume H+ and tend to increase the pH (Konsten and van Breemen, 1994). However, particularly Mn(II) can precipitate as MnC0 3 and this should in turn release H+ and limit the first effect on pH (Sadana and Takkar, 1988). What can be summarized after Ponnamperuma (1972), Bartlett and James (1993), is that hydromorphic conditions tend to shift soil pH toward neutrality. pH of acid soils, particularly sulfate soils, increases when reduced (Konsten and van Breemen, 1994). Basic soils, calcic or sodic, tend to acidify, because resultant evolution of pH depends much on reactions with carbonate (Ponnamperuma, 1972).
168
Fate and Transport of Heavy Metals in the Vadose Zone ~
-------~-
~~---~~~-
~---~~~--
--
~~--~~-------~-~----
..
~---
The interpretation of the effects of reducing conditions on trace metal solubility through changing pH was present in several studies reviewed above (Bjerre and Schierup, 1985; Francis and Dodge, 1990). This influence is obviously taken into account within the solubility approach. Other examples are given below. U sing soil columns water saturated or not, Welch and Lund (1987) found that Ni was less mobile in saturated conditions, and that this was related to higher pH. Dutta et al. (1989) reported that flooding samples from 26 alluvial rice fields increased pH and decreased DTP A extractable Cu and Zn; decrease in extractable Zn was correlated to pH increase, among other variables (total carbonate, sulfide, DTPA-extractable Fe and Mn). Dudley et al. (1986) followed nitrogen transformations, changes in pH, soluble carbon and Cu, Ni, Zn, during the aerobic incubation of soils amended with anaerobic sludge. At first, organic C, NH3 and NH4 +releases were accompanied by increasing pH and soluble Cu organic complexes, then nilrification occurred, which led to decreasing pH and coherently increasing soluble Cu and Zn. Chanmugathas and Bollag (1987) subjected columns of 4 different soils spiked with Cd(N03)2 to aerobic or anaerobic incubations. The evolution of pH during anaerobic incubations varied; however, Cd retention was generally consistent with it; i.e., Cd retention was lower at low pH and increased when pH increased. Flooding an undisturbed block of silt loam soil from a wheat field polluted by industrial atmospheric fallout, Charlatchka et al. (1995) noticed that soluble Cd, Pb, and Zn, varied more or less according to pH variations (Figure 8.3). Evolution of Mn and soluble carbonate were also consistent to common observations. Then, the pH was controlled on suspensions of the same soil, but different reagents favored different reactions involving H+: anaerobic transformation of glucose can produce fatty acids (Figure 8.4), as in Glinski et al. (1996), so that pH spontaneously decreased by more than one unit and NaOH was used to stabilize the pH. Without glucose, different mineral acids were used, but more nitric acid was needed as compared to HCl to maintain the same pH because denitrification strongly occurred with HN03 and consumed protons. It is noteworthy that anaerobic transformations of organic substances containing much available carbon (about 1% by weight of soil) can lead to acidification.
ROLE OF SOLUBLE ORGANIC LIGANDS The presence of soluble organic ligands (whatever their origin and the redox conditions) should move the chemical equilibria for trace metals toward the aqueous phase and it has been found that they can actually increase their mobility (e.g., Chanmugathas and Bollag, 1988; Dudley et al., 1986; Dunnivant et al., 1992; Gerritse, 1996; Lamy et al., 1993; Reddy et al., 1995). However, high concentrations of fulvic acids can also increase retention of metals by building mineral-organic complexes (e.g., Chubin and Street, 1981). The final distribution and effects depend much on pH (Japenga et al., 1992; Lamy et al., 1993). Under reducing conditions, several low mass organic acids can increase the solubility of trace metals, the most abundant and persistent being acetic acid (Ponnamperuma, 1972). Dudley et al. (1986) and Francis and Dodge (1990), evoked the role of soluble organic ligands, metabolic products, to explain increased mobility under anaerobic conditions. One may also recall that complexed Zn or other trace element can be in solution but not available to plants (Forno et al., 1975). Mter these authors, and Ponnamperuma, anaerobic conditions in fields induce a concentration peak of acetic acid during the first
Influence of Reducing Conditions on Mobility of Divalent Trace Metals in Soils -------~-~--~--------,,-'""-
..
-,-~-
"-"
,
~
~---
--
~----------
600
-
169
-
9 a) 8
~ 400 Cl E C 0 .0 «l
...
7 I
a.
6
200
0
5 4
0 100
0
200
400
300
640
150 b)
Ca
~
480
Cl
100
E
c
:::2: "0 c
~---
50
«l Q)
u.
...----------
--
_Fe....
~
320
cO 0
160
Mn
I
0
0 100
0
200
300
400
80
800 c)
\Pb ~
I I I I I
60
Cl :::L
.0
a.. 40
600 ~
400
"0
0
Cl :::L
C
C
«l "0
Cl
E
N
20
200
0
0 0
100
200
300
400
Time (days) Figure 8.3. Evolution of waters draining from a flooded undisturbed block of soil (from the same cultivated layer as in Figure 8.1).
weeks. Enhancement of anaerobic incubations of soil by adding 1% glucose produced similar phenomena, or, with 3% glucose, stabilized high levels of butyric and acetic acids (Figure 804) (Charlatchka et aI., 1995). With respect to the formation of soluble complexes with trace metals, acetate appeared important since in the experiment of Figure 804, about 40% of soluble Cd should have been complexed by acetate, and 70% of Pb (Figure 8.5). Despite the competition with major cations, particularly Ca, at the end of
170
Fate and Transport of Heavy Metals in the Vadose Zone
1400 ~
~----------------------------~
1200
~ 1000 c:
0 .0 ....
C'tl () 0
'cC'tl
Toe
800
Sum e (Ac+But)
600
e (Butyric Acid
Cl .... 400 0
e (Acetic Acid
200 e (Glucose
o~~~~~~~~~~~~~
o
4
8
12
16
Time (days) Figure 8.4. Evolution of organic solutes during anaerobic incubation of a soil:suspension with 3% glucose (by weight of soil).
flooding experiments of Figure 8.3, about 20% of Cd was complexed by acetate after chemical analyses and calculated speciation.
TRANSFORMATION OF INSOLUBLE ORGANICS We presented with enough accuracy the influence of major soil mineral transformations under reducing conditions on trace metal behavior. Also, organic substances are deeply affected when anaerobic processes take place instead of aerobic ones. Unfortunately, there is much less accuracy about these transformations and still less on how they affect trace metals. The previous part presented only some processes concerning the aqueous phase. The retention properties of natural insoluble organic substances for trace elements, particularly Cu and Pb among divalent trace metals, have been extensively studied. However, there is no universal modeling of involved reactions, like for solubility equilibria. Besides, the organic substances are expected to change in nature and reactivity. The chemical extraction approach provided some information: after adding organic matter and flooding rice soils, Ghanem and Mikkelsen (1987) found that the organic fraction of Zn progressively decreased. Iu et al. (1981) also estimated that solid organic fractions of Zn and Cu decreased under hydromorphic conditions, whereas soluble organic Cu increased. Following a more global approach, and focusing on biological processes, Francis and Dodge (1990) noticed that Pb was not remobilized by anoxic microbial activity like other trace metals from a synthetic iron oxide trap. They interpreted that Pb was fIxed back by biosorption. Chanmugathas and Bollag (1987) attempted to clear up the role of microbial activities, aerobic or anaerobic, on the retention of Cd (added as nitrate) by soil materials. Different trends were observed at the beginning of their experiments or later, and depending on soil properties; anyway, by comparison with sterile materials, microbes generally increased the retention of Cd, through lower pHs and by more direct intervening.
Influence of Reducing Conditions on Mobility of Divalent Trace Metals in Soils
171
100% - , - - - -
80%
60%
40%
20%
0%
+---Cd • free cation
Pb • inorganic complexes
0 acetate complexes
Zn III other organic complexes
Figure 8.5. Calculated speciation of dissolved trace metals in suspension incubated after glucose amendment (see Figure 8.4).
Charlatchka (1996) used a particular extraction technique to put in evidence the role of biomass on trace metal fixation, based on fumigation with chloroform and aqueous extraction. This technique is recommended to quantifY C from the biomass (Vance et aI., 1987). Applied at the same time on Zn and Cd, it showed that Zn had been partly fixed back during some anaerobic incubation by favored biomass.
SUMMARY When reducing conditions occur in soils, several processes start which modifY soil components and the mobility of trace elements, even trace metals which practically keep the same valency. The intensity of the various processes depend on starting conditions; their sequence, if reducing conditions last, is rather constant, except for organic substance transformations. Obviously, soil organic matter, and organic amendments, are heterogeneous and intervene in many redox reactions, at different times. The influence of some redox processes on trace metal behavior can be qualitatively predicted, following the sequence below: A first step after limiting oxygen, or rather after water saturation, is often observed: pH decreases and soluble metals increase. Then, nitrate and Mn oxides begin to be reduced, and consume protons. When nitrate is abundant, denitrification producing different gas generally consumes protons, and metal solubility should decrease. However, pH increase can be limited by other H+ transfer. If much available C is present, like in case of recent organic amendment, formation of weak acids, including CO 2, dominates redox processes. Dissociated acids decrease the pH, and despite being also
172
Fate and Transport of Heavy Metals in the Vadose Zone
rather weak ligands for trace metals, they increase their solubility by complexation and induced acidity. When both nitrate and organic carbon are available, competing effects occur. pH variation should be determined by their ratio and by final pH, as it appears in reactions written by Krumbein and Swart (1983). Another process becomes detectable, the retention by developing anaerobic organisms of a few trace elements. Unless soils contain much nitrate, Mn reduction quickly intervenes, and often seems to come with more soluble trace metals. This is certainly explained by the high retention capacity of Mn oxides for trace cations, and by the limited increase of pH, if any, which occurs. H+ consumption is limited because dissolution concerns a small amount of reactive Mn oxides, which can partly precipitate as carbonate of low reactivity. Again other H+ transfers can intervene, like mineralization of organic matter to reduce Mn, and Mn 2+/H+ exchange in soil-plant systems. Reduction of iron takes more time and more electron donor, and certainly can release more trace metals adsorbed or trapped by ferric oxides. However, the expected increase of pH should favor the precipitation of these metals or their fixation back on other soil constituents. Under strong reducing conditions, sulfide formation immobilizes most divalent trace metals. This process could appear as the more irreversible in some cases of permanent flooding because it involves stable pure minerals, and the sequential picture given above can appear as concerning short-term influence of reducing conditions. It is true that redox processes in soils are generally cyclic, so that the mechanisms described are also temporary. However, anaerobic-aerobic cycles certainly result in some translocation of trace elements. The resultant of one cycle is not always null because soils are open systems, and there are examples of dramatic evolution regarding soil acidification and chemistry (van Breemen et aI., 1983). Other contributions seem necessary for improving our knowledge of long-term effects of redox cycles on trace metal behavior. REFERENCES
Ahumada, I. T. and E.B. Schalscha. Fractionation of cadmium and copper in soils. Effect of redox potential. Agrochimica .37, pp. 281-289, 199.3. Alloway, B.J. Cadmium, in HearyMetau in SOlM, 2nd ed., B.J. Alloway, Ed., Blackie Academic & Professional, London, 1995, pp. 122-151. Altmann, RS. and A.C.M. Bourg. Cadmium mobilisation under conditions simulating anaerobic to aerobic transition in a landfill leachate-polluted aquifer. Water Air Soil Potfut. 94, pp . .385.392, 1997. Balzer, W. On the distribution of iron and manganese at the sediment/water interface: Thermodynamic versus kinetic control. Geochim. Co.Jmochim. Acta 46, pp. 115.3-1161, 1982. Bartlett, RJ. and B.R James. Redox chemistry of soils. Ad!'. Agron. 50, pp. 151-208, 199.3. Bingham, F.T., A.L. Page, RJ. Mahler, and T.J. Ganje. Cadmium availability to rice in sludgeamended soil under "flood" and "nonflood" culture. Soil Sci. Soc. Amer. J. 40, pp. 715-719, 1976. Bjerre, G.K. and H.H. Schierup. Influence of waterlogging on availability and uptake of heavy metals by oat grown in different soils. Plant Soil 88, pp. 45-59, 1985. Brennan, E.W. and W.L. Lindsay. The role of pyrite in controlling metal ion activities in highly reduced soils. Geochim. Co.Jmochim. Acta 60, pp . .3609-.3618, 1996. Brown, P.H., L. Dunemann, R Schulz, and H. Marschner. Influence of redox potential and plant species on the uptake of nickel and cadmium from soils. Z Pjlanzenerniihr. Bodenl.:. 152, pp. 8591, 1989.
Influence of Reducing Conditions on Mobility of Divalent Trace Metals in Soils
173
Brummer, G., K.G. Tiller, U. Herms, and P.M. Clayton. Adsorption-desorption and/or precipitation-dissolution processes of zinc in soils. Geooerma 31, pp. 337-354, 1983. Chanmugathas, P. and J.M. Bollag. Microbial mobilization of cadmium in soil under aerobic and anaerobic conditions. J. Environ. QuaL. 16, pp. 161-167, 1987. Chanmugathas, P. and J .M. Bollag. A column study of the biological mobilization and speciation of cadmium in soil. Arch. Environ. Contam. ToxicoL. 17, pp. 229-237, 1988. Charlatchka, R Mobilite des metaux dans un sol contamine. Influence des conditions reductrices. These de l'Universite de Paris 12, 1996. Charlatchka, R, P. Cambier, and S. Bourgeois. Mobilization of Trace Metals in Contaminated Soils Under Anaerobic Conditions, in Contaminateo Soi&, Third Intern. ConE. Biogeochem. Trace Elements, Paris, May 15-19, 1995, R Prost, Ed., INRA Editions, Paris, 1997, pp. 159-174. Charlatchka, Rand P. Cambier. Influence of Reducing Conditions on Solubility of Trace Metals in Contaminateo Soi&. Accepted for Water Air Soil Pollut. 1998. Christensen, T.H. and P. Kjelsen. Basic Biochemical Processes in Landfill, in Sanitary Lano/iLLing: Proce.Jd, Technology ano EnvironmentaL Impact, 1989, pp. 29-49. Chuan, M.C., G.y' Shu, and J.C. Liu. Solubility of heavy metals in a contaminated soil: Effects of redox potential and pH. Water Air Soil PoLLat. 90, pp. 543-556, 1996. Chubin, RG. and J.J. Street. Adsorption of cadmium on soil constituents in presence of complexing ligands. J. Environ. Qua!. 10, pp. 225-228, 198!. Dudley, L.M., B.L. McNeaL and J.E. Baham. Time-dependent changes in soluble organics, copper, nickeL and zinc from sludge amended soils. J. Environ. Qua!. 15, pp. 188-192, 1986. Dunnivant F.M., P.M. Jardine, D.L. Taylor, and J.F. McCarthy. Cotransport of cadmium and hexachlorobiphenyl by dissolved organic carbon through columns containing aquifer material. Environ. Sci. Techno!. 26, pp. 360-368, 1992. Dutta, D., B. MandaL and L.N. MandaI. Decrease in availability of zinc and copper in acidic to near neutral soils on submergence. Soil Sci. 147, pp. 187-195, 1989. Firestone, M.K., M.S. Smith, RB. Firestone, and J.M. Tiedje. The influence of nitrate, nitrite, and oxygen on the composition of the gaseous products of denitrification in soil. Soil Sci. Soc. Am. J. 43, pp. 1140-1144, 1979. Forno, D.A., S. Yoshida, and C.J. Asher. Zinc deficiency in rice. I. Soil factors associated with the deficiency. Plant Soil 42, pp. 537-550, 1975. Forstner, U. Changes in Metal Solubilities in Aquatic and Terrestrial Cycles, in MetaL Speciation, Separation, ano Recovery, J.W. Patterson and R Passino, Eds., Lewis Publishers, Boca Raton, FL, 1987, pp. 3-26. Francis, J.A. and C.J. Dodge. Anaerobic microbial remobilisation of toxic metals coprecipitated with iron oxide. Environ. Sci. TechnoL. 24, pp. 373-378, 1990. Gerritse, RG. Dispersion of cadmium in columns of saturated sandy soils. J. Environ. QuaL. 25, pp.
1344-1349, 1996. Ghanem, S.A. and D.S. Mikkelsen. Effect of organic matter on changes in soil zinc fractions found in wetland soils. Comman. Soil Sci. Plant Anal. 18, pp. 1217-1234, 1987. Glinski J., K. Stahr, Z. Stpeniewska, and M. Brzezinska. Changes of redox and pH conditions in a flooded soil amended with glucose and manganese oxide or iron oxide under laboratory conditions. Z. P/Lanzenerniihr. Booenlc. 159, pp. 297-304, 1996. Huaiman, C. Studies on kinetics and equilibria of CdC03 and CdS in soil solution. Acta PeooLogica Sinica 21(3), pp. 258-267, 1984. Iu, K.L., I.D. Pulford, and H.J. Duncan. Influence of waterlogging and lime or organic matter additions on the distribution of trace metals in an acid soil. II. Zinc and copper. Plant Soil 59, pp. 327-333, 198!. Iu, K.L., I.D. Pulford, and H.J. Duncan. Influence of soil waterlogging on subsequent plant growth and trace metal content. Plant Soil 66, pp. 423-427, 1982.
174
Fate and Transport of Heavy Metals in the Vadose Zone
Japenga, J., J.W. Dalenberg, D. Wiersma, S.D. Scheltens, D. Hesterberg, and W. Salomons. Effect of liquid animal manure application on the solubilization of heavy metals from soil. Intern. J. Environ. AnaL. Chem. 46, pp. 25-39, 1992. Jopony, M. and S.D. Young. The solid-solution equilibria oElead and cadmium in polluted soils previously treated with sewage-sludges. Plant Soil 132, pp. 179-196, 1994. Kiekens, L. Zinc, in Heavy Metau in Soiu, 2nd ed., Blackie Academic & ProfessionaL London, 1995, pp. 284-305. Konsten, C.J.M., N. van Breemen, S. Supping, LB. Aribawa, and J.E. Groenenberg. Effects of flooding on pH of rice-producing, acid sulfate soils in Indonesia. Soil Sci. Soc. Am. J. 58, pp. 871--883, 1994. Krumbein, W.E. and P.K. Swart. The microbial carbon cycle, in Microbial Geochemidtry, W.E. Krumbein, Ed., Blackwell Scientific, London, 1983, pp. 5-62. Lamy, I., S. Bourgeois, and A. Bermond. Soil cadmium mobility as a consequence of sewage sludge disposal. J. Environ. Qual. 22, pp. 731-737, 1993. Lamy, I., P. Cambier, and S. Bourgeois. Pb and Cd complexation with soluble organic carbon and speciation in alkaline 'soilleachates. Environ. Geochem. Health 16, pp. 1-16, 1984. Lindsay, W.L. Chemical Equilibria in Soiu. John Wiley & Sons, New York, 1979. MandaL L.N. and B. MandaI. Transformation of zinc fractions on rice soils. Soil Sci.. 143, pp. 205212, 1987. MandaL B., G.C. Hazra, and A.K. Pal. Transformation of zinc in soils under submerged condition and its relation with zinc nutrition of rice. Plant Soil 106, pp. 121-126, 1988. McBride, M.B. Reactions controlling heavy metal solubility in soils. Adv. Soil Sci.. 10, pp. 1-35, 1989. McGrath, S.P. Chromium and Nickel, in Heavy Metau in Soiu, 2nd ed., Blackie Academic & Professional, London, 1995, pp. 152-177. Morse, J.W. and T. Arakaki. Adsorption and coprecipitation of divalent metals with mackinawite (FeS). Geochim. COdmochim. Acta 57, pp. 3635-3640, 1993. Patrick, W.H., Jr. and A. Jugsujinda. Sequential reduction and oxidation of inorganic nitrogen, manganese, and iron in flooded soil. Soil Sci. Soc. Am. J. 56, pp. 1071-1073, 1992. Ponnamperuma, F.N. The chemistry of submerged soils. Adv. Agron. 24, pp. 29-96, 1972. Reddy, C.N. and W.H. Patrick, Jr. Effect of redox potential and pH on the uptake of cadmium by rice plants. J. Environ. QuaL. 6, pp. 259-262, 1977. Reddy, K.J., L. Wang, and S.P. Gloss. Solubility and mobility of copper, zinc, and lead in acidic environments. Plant Soil 171, pp. 53-58, 1995. Sadana, U.S. and P.N. Takkar. Effect of sodicity and zinc on soil solution chemistry of manganese under submerged conditions. J. Agric. Sci. Camb. Ill, pp. 51-55, 1988. Sajwan, K.S. and W.L. Lindsay. Effects of redox on zinc deficiency in paddy rice. Soil Sci. Soc. Am. J. 50, pp. 1264-1269, 1986. Schwab, A.P. and W.L. Lindsay. Effect of redox on the solubility and availability of iron. Soil Sci. Soc. Am. J. 47, pp. 201-205, 1983. Smith, C.J., M.F. Wright, and W.H. Patrick, Jr. The effect of soil redox potential and pH on the reduction, and production of nitrous oxide. J. Environ. Qual. 12, pp. 186-188, 1983. Sposito, G. The Thermodynamic of Soil Solutwnd. Oxford, 1981. Sposito, G. The Chemidtry of Soiu. Oxford, 1989. Tack, F.M., O.W.J.J. Callewaert, and M.G. Verloo. Metal solubility as a function of pH in a contaminated, dredged sediment affected by oxidation. Environ. PolLut. 91, pp. 199-208, 1996. Tessier, A. and P.G.C. Campbell. Comments on "Pitfalls of Sequential Extractions" by P.M.V. Nirel and F.M.M. Morel, Water &1.24, pp. 1055-1056, Water Red. 25, pp. 115-117, 1991. Trolard, F., J.M.R. Genin, M. Abdelmoula, G. Bourrie, B. Humbert, and A. Herbillon. Identification of a green rust mineral in a reductomorphic soil by Mosssbauer and Raman spectroscopies. Geochim. COdmochim. Acta 61, pp. 1107-1111, 1996.
Influence of Reducing Conditions on Mobility of Divalent Trace Metals in Soils
175
van Breemen, N., J. Mulder, and C. T. Driscoll. Acidification and alkalinisation of soils. Plant SoiL 75, pp. 283-308, 1983. Vance E.D., P.C. Brookes, and D.S. Jenkinson. An extraction method for measuring soil microbial biomass C. SoiL BioL. Biochem. 19, pp. 703-707, 1987. Welch, J.E. and L.J. Lund. Soil properties, irrigation water quality, and soil moisture level influences on the movement of nickel in sewage sludge-treated soils. J. Environ. QuaL. 16, pp. 403-410, 1987. Xiang & Banin. Long term transformations of toxic heary metals in arid-zone soils incubated. 1. Under saturated conditions. Water Air SoiL PoLLut. 96, p. 367, 1997. Xiong, L.M. and R.K. Lu. Effect of liming on plant accumulation of cadmium under upland or flooded conditions. Environ. PolLut. 79, pp. 199-203, 1993.
CHAPTER 9
Lead Mobilization in Calcareous Agricultural Soils Carmen Perez-Sirvent, Josefa Martinez-Sanchez, and Carolina Garcia-Rizo
INTRODUCTION Soil provides the physical support and the nutrients necessary for plant growth although it can also provide undesirable concentrations of certain elements which, as well as causing phytotoxicity problems, may pass into the food chain and affect human health (Kabata-Pendias and Pendias, 1992; Thunus and Lejeune, 1994). There have been numerous studies published concerning the mobility of heavy metals and of their adverse environmental impact (Allen et al., 1995; Salomons and Forstner, 1995; Ure and Davidson, 1995). However, most studies refer to noncalcareous soils and when similar techniques to those described in the bibliography are used for highly calcareous soils, the results may not accurately indicate metal availability because of the peculiar physicochemical and mineralogical characteristics of such soils. Indeed, the results may be ambiguous and even contradictory, not fully describing the behavior of the metal in the soil, so that their environmental impact will not be fully understood. While the present situation in these soils might be reflected, neither potential future adverse effects can be predicted nor possible measures for reducing such hazards be proposed. There are large areas of calcareous soils in SE Spain and many are found in places where Pb and Zn are mined. These soils often present very high levels of these metals with a strong spatial dispersion of values. At the same time, these areas often have strong agricultural traditions and may be dedicated to the growth of horticultural crops, since the local climate is very favorable to such activities. The area around Mazarron (Murcia, SE Spain) presents ideal characteristics for a study such as that undertaken in this chapter. It has geomorphological, edaphological and climatic characteristics similar to other mining areas in the Mediterranean area with a scant, degraded vegetation. There are many old mining installations; the soil is calcareous and often has a very high Pb content, and much is known about the original material of these soils. Agriculture is intensive, both in the open air and in greenhouse, and is 177
178
Fate and Transport of Heavy Metals in the Vadose Zone
mainly dedicated to horticultural crops. In this chapter we describe the origin, geochemical process, natural mobility of Pb, and the mobilization provoked by reagents in different acid and Eh conditions.
SOIL FORMATION FACTORS Environmental Conditions A good example of calcareous agricultural soils (Figure 9.1), is found in Las Moreras Rambla (wadi complex), which may occasionally be flooded after heavy rain. The soils of the area feeding the rambla are calcic and petrocalcic xerosols, and calcareous regosols (Alias et al., 1986). Nearby there is a drainage channel which carries substantial quantities of materials (urbic antrosols) from the mining area of Cabezo de San Cristobal and Las Pedreras. Piping type erosion is present in slags and reveals the existence of surface particulate flows and infiltrations of subterranean waters. The flood area is wide (500-1000 m) and three levels of terraces have been identified (Figure 9.2). The frequent braids of the present bed indicate continuous modification of the channel. It is here where there are high drainage flow densities and where the most intensive agricultural activity takes place. The soils are sandy-calcaric fluvisols (coarse texture) and saltcalcaric (high salt content) (Alias et al., 1986). Erosion and sedimentation is substantial in the last stretches of the wadi. Evidence of this is the water erosion observed over a short period of time (1956-78) by RodriguezEstrella (1993) in Sierra de las Moreras, where there was a very clear modification of the detritic fans at the pediments and in the terraces. The spatial distribution of Pb minerals is related to the mining history of SE Spain which stretches over more than 2500 years, starting with the exploitation of these deposits by the Phoenicians and Carthaginians in the fifth and fourth centuries B.C. The Romans opened up shafts of up to 500 m depth. After a relatively quiet period, mining started again in the twentieth century but ceased in 1962. Annual rainfall is 185-312 mm and the average temperature is 16.5 to 18.8°C. Mean potential evapotranspiration is 869-935 mm and there is a very pronounced hydric deficit. Rainfall is scarce and may be torrential, resulting in flash floods with occasionally disastrous consequences. For example, on 7 September 1989 rainfall in the wadi area exceeded 100 mmlh. In the climatic conditions of the area, the high temperatures cause the water contained near the surface of the soil to evaporate rapidly. Weathering occurs extremely slowly in the dry air. The fundamental alteration and mobilization processes occur at greater depths, in the unsaturated zone or aeration zone, where the water is in the form of vadose water. The water remains here for a longer time and there is greater root development.
Nature of the Materials The materials which participate in the formation of these soils are sediments from soils developed on carbonated rocks (limestones, dolomites, sandy marls and conglomerates), volcanic rocks and slags which are found in the areas around the gully (Figure 9.1) and which have undergone a process of erosion, transport, and sedimentation.
lead Mobilization in Calcareous Agricultural Soils
t
N
Moreras
Mediterranean Sea
Legend 1
0
2
3
4
SKm
I • I So.;, Pb2+, W
Figure 9.1. Studied zone.
Figure 9.2. Schematic diagram of the wadi complex.
179
180
Fate and Transport of Heavy Metals in the Vadose Zone
The scarce organic material accompanying these soils (particularly in horizon A) undergoes subsequent oxidation and mineralization, both of which are very important biological degradation processes in the soils of semiarid areas (Tudela and Martinez-Sanchez, 1997). In this way, N0 3-, PO/-, K+, Na+, Mg2+, Ca2+, and Fe 2+, are made more available to the plants. The materials which contribute Pb to the study area come from mine wastes, which are a mixture of materials including unaltered parent material (dacites and riodacites), altered parent materials, primary mineralization products (metallic sulfIdes), secondary mineralization products (hydrothermal alteration and products of the supergenic alteration of sulfIdes), and remains of the mechanical and metallurgical treatment of exploited ores. These materials therefore are of a very varied nature with a heterogeneous granulometry. Tables 9.1 and 9.2 show the chemical and mineralogical composition of the fresh and altered rocks and the predominant mineralogy of the mining waste studied in the zone (Arana et al., 1993). The eroded xerosols contribute calcite and dolomite as principal components, accompanied by quartz and feldspars, small quantities of clay minerals such as illite and kaolinite together with chlorite and smectite, iron oxides and organic materials. The calcareous regosols contribute abundant quantities of clay minerals (illite and smectite), calcite, dolomite, small quantities of quartz and feldspar and organic materials. Figure 9.3 shows the sequence of the formation processes which take place. The calcareous soils of the reference area (Table 9.3) had a low organic matter content and a slightly basic pH, reflecting their high calcium carbonate content. The texture was sandy loam to clay loam and the cationic exchange capacity values were low, as was to be expected in soils with a low proportion of organic matter and containing illite type clay. In some cases, the concentration of soluble salts was very high for cultivated soils. A mineralogical study of the fIne earth fraction showed that all the samples were of a similar composition: calcite (30-40%), quartz (20-35%), illite (10-25%), feldspars (1020%), and dolomite (0-15%). Representing less than 5% in all samples and therefore not easily observed by XRD were the oxides and oxyhydroxides of Fe and Mn, which are of low crystallinity or amorphous. Some samples contained gypsum. Figure 9.4a shows the X-ray diffractograms of a representative sample. It must also be noted that total Pb levels were high (3000 ppm) for an agricultural soil (Table 9.6). Examination of the samples using a scanning electron microscope equipped with EDS analyzer revealed quartz, calcite, clays, plagioclase, and feldspars of different particle sizes. Mapping of the chemical element distribution (Figure 9.5) showed that the lead was distributed widely but with no set morphology.
TRANSPORT The above materials are carried by flowing water to the study area in the following way: • as dissolved load from sediment-forming minerals in solution • as suspended load from solid sedimentary materials carried along in suspension • as bed load dragged along the bottom of the stream channel. The degree and rate of movement of suspended sedimentary material in streams depend on the velocity of the water flow and the settling velocity of the particles in suspen-
Table 9.1. Chemical Composition of Volcanic Rocks Si0 2
%
AI 2 0 3 %
Unaltered Volcanic Rocks Max 61.46 17.58 Min 52.86 12.91 Mean 56.1 3 1 5.87 Altered Volcanic Rocks Max 67.81 Min 21.78 Mean 47.26
Fe 2 0 3 %
CaO %
MgO %
Ti0 2 %
K2 0 %
Na2 0 %
PbO ppm
DlOOO°C %
' r
'/I)
III
0.
5.98 3.08 4.70
4.39 2.64 3.37
2.67 1.23 1.63
0.96 0.44 0.78
5.32 3.32 3.50
5.57 2.82 3.50
391 70 180
9.43 4.38 6.20
31.80 4.35 10.62
3.16 0.30 0.84
2.05 0.20 0.60
1.47 0.29 0.98
7.80 2.45 6.80
1.44 0.58 0.89
5365 285 2358
28.11 5.26 12.34
3: !O 1
101=
... c)" N
III
::l
17.10 7.30 12.80
5' (")
III
i=i QJ
....
/I)
0
c
'"
~ ....
;:;.
t: ;:; t:
.... a!.
VI
9. Vi
.... 0:1 ....
182
Fate and Transport of Heavy Metals in the Vadose Zone
Table 9.2. Summary of the Mineralogical Composition in Mine Waste Mineral
Volcanic Rock
Altered Rock
Material
Silicates
Quartz Cristobalite Plagioclase Sanidine Phyllosilicates Amphiboles
Quartz Sanidine Phyllosilicates
Quartz Cristobalite Plagioclase Sanidine Phyllosilicates
Sulfides
Oxides and oxyhydroxides Carbonates
Insoluble sulfates
Sphalerite Galene Pyrite Marcasite
Calcite
Hematites Goethite Calcite
Jarosite Group Alunite Group Gypsum
Soluble sulfates
Elements
Sulfur
Cerussite Smithsonite Azurite Malachite Jarosite Group Alunite Group Gypsum Anglesite Copiapite Group Halotrichite Group Botryogen Group Romerite Alunogen Coquimbite ButIerite Bianchite Melanterite Goslarite Epsomite Ferrohexahydrite Group Szomolnokite Kieserite Goslarite Sulfur
sion. Under normal conditions, finely divided silt, clay, or sand make up most of the suspended loam, although larger particles are transported when the water flows rapidly. The ability of a stream to carry sediment increases with both the overall rate of flow of the water (mass per unit time) and the velocity of the water. Both of these are higher under flood conditions, and so floods are particularly important in the transport of sediment. Thus the sporadic torrential rains of the area transport large quantities of carbonated materials and materials from the slags. Excessive rates of water flow prevent infiltration, lead to flash floods, and cause soil erosion.
Dissolved load Slowly flowing water results in dams being formed on the slopes of the mining areas, leading to new alteration phenomena. This water flows in the vadose zone for several
lead Mobilization in Calcareous Agricultural Soils
1
1 ... Li.m.es.t.on.e_...
c:::J
l
Slags
H20 C02
Particulate Material Clay, Calcite, Dolomite ...
J
Weathering Products:
t Dissolution Cr, HCOi, SO: Co.~ Ca~Mg: Na: K+
Volcanic rock Sulfides Sulfates Oxides ..
Oxldates Hydrates Hydrolyzates Sulfates
I Dissolution: ~ A1~So:,
183
_~
H;
pn Partie. Material Clay, Jarosite, ...
Precipitation Neutralization Hydrolysis Hydration Adsorption Oxidation
Sedimentation
Figure 9.3. Main processes involved in the soil evolution.
Table 9.3. Analytical Data of Soil Samples
Min Max Mean
%O.M
C/N
1.0 1.9 1.3
3.3 6.7 4.4
meql %CaC0 3 100g Equivalent T 27.9 66.7 39.2
7.6 16.0 11.0
ds/m E.C.
H2 O
KCI
Particle %Sand
2.7 9.4 3.4
7.7 8.0 7.9
7.6 7.8 7.7
23.1 55.0 42.5
pH
Size Distribution %SiIt %Clay 11.0 49.4 24.5
12.1 21.5 16.4
days after rain has fallen. Table 9.4 shows the mean composition of the soluble products contained in the waters that leach from the slags. The behavior of the elements was studied by correlation techniques. There was a positive correlation between Pb and Fe(II), and As and Sb, but a negative correlation between Pb and pH, reflecting the decrease in Pb concentration at lower acidity values (Lopez-Aguayo et al., 1992).
184
Fate and Transport of Heavy Metals in the Vadose Zone
a)
b)
c)
d)
o
10
20
30
40
28
Figure 9.4. X-ray diffractograms obtained from a soil sample (a) and from the remainders after the first, second, and third extraction steps (b, c, and d, respectively) recommended by BCR.
SE
Fe
K
Mg
Na
o
Pb
S
Ca
Si
Figure 9.5. Mapping of the chemical element distribution obtained by SEM-EDS (raw sample).
When gramineae covered the soils developed from limestones and dolomite, which occurred mainly in spring, the carbon dioxide concentration increased in the soil solution, leading to the solubilization of calcium carbonate:
Table 9.4. Chemical Composition of Acidic Waters Major Compounds mmol/L
Max Min Mean
Si
Fe(lII)
Fe(1I)
AI
Mg
Zn
Mn
Cu
S04
pH
5 0.9 3
947 67 349
258 6 45
1001 44 256
477 16 200
627 26 189
55 4 27
6 1 3
1200 826 915
2.2 1.5 1.8
Na
K
Minor Compounds mmol/L
Max Min Mean
Pb
As
Sb
Ca
Ge
Sn
Cr
Cd
0.030 0.005 0.001
7.34 0.44 2.60
0.213 0.008 0.120
1.996 0.200 0.500
0.964 0.027 0.324
0.826 0.042 0.337
0.025 0.002 0.010
1.156 0.027 0.780
Mean values for 30 samples collected during a year.
0.2610 0.026 0.100
0.084 0.002 0.061
r
<'1l III
0.
s:o
g; N' III ....
o·
:::l :::l
(')
III
i=i
III
.:; o
= VI
~
C1Q
;:;. '""
= ;:;: = '"" !l:!.
III
9.
v;
-'
00 U1
186
Fate and Transport of Heavy Metals in the Vadose Zone
It is well known that water with a high dissolved carbon dioxide content in contact with calcium carbonate contains Ca2+ and HC0 3- ions. Flowing water containing calcium may become more basic through the loss of CO 2 to the atmosphere or through contact with dissolved bases, resulting in the deposition of insoluble CaC03 :
This is very important throughout SE Spain (Martinez-Sanchez, 1982). In sloping limestone areas there was a substantial mobilization of carbonates due to the transport of water-soluble bicarbonates 15-30 cm below the surface, which precipitate in the lower zones to form calcic, sometimes strongly cemented, horizons. An example of this are the calcic and petrocalcic xerosols at the foot of the Sierra de las Moreras (Figure 9.1). Soluble salts are dissolved rapidly in marls. Given the semiarid climate, scant vegetation, and low permeability of the marls, the dissolution of alkaline-earth carbonates takes place very slowly and may even be prevented. There was no sign of carbonate movement through the profIle of developed soils (calcaric regosols).
Particulate Forms: Suspended and Bed Loads When the water flow is rapid, soluble and particulated forms are dragged unchanged from their place of origin as far as the gully. As mentioned above, the mineralized materials contribute and generate insoluble sulfates (anglesite, jarosite, alunite), oxides and oxyhydroxides, and clay minerals such as kaolinite and chlorite, which are transported along with the soluble materials (Table 9.2). In this way, unaltered materials are incorporated into this phase as fragments of unaltered rocks, resistant minerals of encasing rocks and sulfIdes. The Pb is bound to: • sulfates and sulfIdes (as a main component) • oxides and hydroxides of iron and manganese (adsorbed and guest in the frame) • silicates (included in the crystalline structure and adsorbed). The most noteworthy contribution of carbonated particulate materials to the soil is calcite, with dolomite and phyllosilicates being the most common.
GEOCHEMICAL PROCESSES When the kinetic energy of the transported system diminishes, materials are sedimented. Both suspended particles and solutes which are being carried along by the water are sedimented regardless of their origin, with coagulation and flocculation phenomena and heterogeneous chemical reactions occurring. The important chemical actors in the medium studied are: • • • •
acid waters waters containing bicarbonates very saline waters carbonated particles
lead Mobilization in Calcareous Agricultural Soils
187
• silicated particles • mineralized particles (Pb) Chemical weathering, as a chemical phenomenon, can be regarded as the result of the tendency of a rock/water/mineral system to attain equilibrium. This occurs through the usual chemical mechanisms of dissolution/precipitation, acid-base reactions, complexation, hydrolysis, and oxidation-reduction, the kinetics of which is conditioned by pH, Eh, and temperature. Water increases the rate of weathering for several reasons. Water, itself, is a chemically active substance in the weathering process. Rainwater is free of mineral solutes but is slightly acidic due to the dissolved carbon dioxide it contains. This makes it chemically aggressive, particularly when compounded by the presence of the sulfuric acid which results from the supergenic alteration of metallic sulfides.
Mobilization-Physical Weathering-Hydration Relations As regards the way in which physical weathering affects materials, it is known that rocks tend to weather more rapidly when there are pronounced differences in the physical conditions, such as alternate freezing and thawing, alternating wet and dry periods, and roots growing through cracks. The most important mechanical aspect is the swelling and shrinking of minerals which accompany hydration and dehydration. Such processes are clearly in evidence in the studied zone since the marls contain alkaline and alkaline-earth sulfates and chlorides, which give rise to polyhydrated salts (gypsum, hexahydrite, epsomite, thenardite, etc.):
Similarly, a large number of sulfates of different degrees of hydration are found in the products which originate in the mineralized zone and which are formed in the final stages of the supergenic alteration process. In some cases, these do not remain unaltered but change with the environmental conditions of humidity and temperature (Lopez-Aguayo and Arana, 1992). All the above favors contact between the solid particles and the alteration agents, and results in faster chemical reactions:
Fibroferrite [FeS04(OH) 5H 2 0] :=:} Hohmannite + 1.5 H 2 0 :=:} Amarantite + 0.5 H 2 0
Soluble Pb-Adsorbent Precursor Ratio The mineralized materials dissolve to produce acid waters, which contain SO/, H+, Fe+ 3, Fe+ 2, Al+ 3, and Pb+ 2, whose evaporation products in a confined atmosphere are shown in Table 9.5.
188
Fate and Transport of Heavy Metals in the Vadose Zone
Table 9.5. Mineralogy of the Precipitates in Acid Waters Formula
Copiapite Halotrichite Alunogen Paracoquimblte Bianchite
(Zn,Fe,Mg)(AI,Fe)i50,J6(OHh' 20 H20 FeAI 2 (50 4)4' 22 H2 0 AI 2(50 4h' 17 H2 0 Fe2 (50 4h' 17 H20 (Zn,Fe, Mg) 50 4 • H20
Max
Mean
50 35 15 14 40
14 18 8 10
15
The mineralogy of these compounds is closely related with the overall chemical composition of the samples, as Zodrow (1980) found in Nova Scotia and Yushkin (1984) in Pai-Khoi. The following groups can be differentiated (a) sulfates of AI (alunogene) or Fe(III) (coquimbite and paracoquimbite), (b) sulfates corresponding to the isomorphous series of Zn, Fe(II) and Mg (bianchite), and (c) hydrated double sulfates of trivalent metals (Fe, AI) and divalent metals (Mg, Fe, Zn), with a molar ratio, M(lII)/M(II), of 4 in the copiapite group and of 2 in the halotrichite group. These compounds also occur in the alteration products of the slags (Table 9.2). The same phases were found in in vitro experiments (Perez-Sirvent et al., 1993). S04
=
+
R3+
+
R2+
+
H 20
+
H+ =}R2(S04h n H 20 + RS04. n H 20 R(S04Hh n H 20 + R (S04) . n H 20 R+ 2R+ 32(S04)4' n H 20 R+ 2R+ 34(S04)6 (OHh. n H 20
where R+3 = AI+ 3, Fe+ 3 and R+2 = Fe+ 2, Mg+ 2, Zn+ 2, Mn, Cu are the principal elements, incorporate smaller quantities of Pb, Ge, Sn, Cd, Co, Ni, Cr. A simplified model reflecting this process should at least take into account the following two factors: the solubility of the crystallized phases and the stoichiometric ratios of the sulfates. The first sulfates to crystallize are those of the divalent elements since their solubility product is lower, resulting in the formation of Fe (S04)'7 H 20 (melanterite), which is only stable at very acidic pH. The importance of this mineral species is that the Fe(II) sulfates are precursors of the genesis of Fe (OHh, which is adsorbent and which transports Pb.
Bicarbonated-Acidic Water Interaction The water which runs off and leaches from calcareous materials has a high HC0 3content. When this water reacts with the above-mentioned acidic waters there is a coprecipitation of basic iron carbonates which retain different proportions of Pb. Depending on the pH, the precipitation of Fe(OHh, which acts as adsorbent of Pb, can be gIven as:
'--
lead Mobilization in Calcareolls Agricultural Soils
189
The soluble Pb can react with HC03-:
a reaction which is favored as the acidity of the medium is reduced by bicarbonate water. If, in addition to HC03-, ligands such as CI-, S04 =, C03=, OH-, humic and fulvic acids, and CN- (used in sulfide flotation processes) are also present, Pb complexation reactions may occur.
Acid Water-Mineralized Particulate Material-0 2 -C0 2 Interaction The instability of sulfides in an oxidizing and acidic medium results in their oxidation. The most important oxidation reactions for the mobilization of lead are those involving pyrite and galena:
The carbon dioxide acts in the oxidation process by carbonating the metal and supplying sulfate ions, which catalyze subsequent oxidation reactions, and by increasing the medium's acidity:
Of special interest is the redox chemistry of the Fe and Mn present in these outcrops in relation to the mobility of other metals. Their hydroxides and oxyhydroxides play an important role in Pb adsorption phenomena (Manahan, 1994). The Fe(OH)3 is a link in the process of jarosite formation:
This H-jarosite includes Na, K, and Al in the general jarosite formula (Na,K, Xh(Fe,Alh (S04)2 (OH)6 where X is Ca, Pb, Ag ... The Pb enters the sulfate frame either because it is retained in the Fe (OHh by adsorption or because it is brought into the reaction as soluble Pb (plumbojarosite). As regards the genesis of alunite, the most probable formation mechanism is the transformation of potassium micas via kaolinite (Hueso et al., 1981). Biogeochemical reactions involving thwhaciLLIM !errooxwafM and thwhaciLLIM thwoxwafM type bacteria may take place in these processes (Lundgren and Silver, 1980; Allan, 1995).
Acid Water-Carbonated Particulate Material-0 2 -CO z Interaction The calcite present acts as pH buffer in the solutions, counteracts the acidification of the medium brought about by the H+ ions and, by reacting with the sulfate ions present, enables cerussite to exist in these media.
190
Fate and Transport of Heavy Metals in the Vadose Zone
The acidic waters react with the carbonates to bring about precipitation of the Fe and AI hydroxides, which can react with sulfate ions to form jarosite and gypsum. For this reason, the materials that make up the bed of the wadi in this zone have a pH in water of 3, an EC of 24 mS/cm, and are made up of quartz, feldspar, illite, gypsum, and jarosite (Perez-Sirvent et al., 1995). In areas of the gully with a pH<3 and slightly oxidizing conditions, the Pb undergoes a process of Fe and Mn oxyhydroxide desorption, as mentioned above. If, on the other hand, the pH is slightly alkaline, the Fe and Mn can be precipitated as carbonates, which are less adsorbent than the corresponding oxyhydroxides. In unusual situations of anaerobiosis caused by flooding or the rising of the water table and in acid conditions, Pb will be mobilized more readily because the conditions will be ideal for dissolving the Pb compound mentioned above.
Pb Sorption-Desorption In the above interactions, Pb sorption/desorption reactions occur, with those involving the oxideslhydroxides of AI, Fe, and Si being especially important. The sorption of Pb on solid surfaces typically shows an "adsorption edge" onto iron hydrous oxide (Liang and McCarthy, 1995). The mechanism regulating these processes can be explained by complexation. The surface of oxides and hydroxides such as Mn02 can adsorb Pb+ 2, giving rise to surface complexation,
or chelation M-OH
I
+ Pb 2+ ~
M-OH
The pH plays an important role in the sorption/desorption reactions. At high pH (pH> 6), Pb is preferentially sorbed onto iron oxide, but at pH < 3.0, almost all the Pb remains in solution. Between pH 3 and 6, sharp increases in Pb adsorption accompany very small increases in pH. This is why, in our case, the process of Pb sorption is important, since the increased pH caused by the mixing of acidic waters with carbonated suspensions leads to the desorption/adsorption oflead (Lumsdon and Evans, 1987). The phyllosilicates may have similar action mechanisms to those reported above, and act as possible adsorbents of: (1) the soluble Pb in the leaching waters, (2) the Pb liberated in situ by the oxidation of particulate Pb sulfide, and (3) the Pb liberated in the desorption of the particulated Fe and Mn oxidelhydroxides produced by a decrease in pH following contact with acid waters. No effect of the ionic strength on Pb adsorption on a goethite surface has been observed (Hayes and Leckie, 1995). From this, we can deduce that the high salt concentration of the waters from the zone containing marls has no effect on the Pb adsorbed on the oxideslhydroxides.
lead Mobilization in Calcareous Agricultural Soils
191
Although humic substances act as adsorbents of Pb in soil, this was of little importance in our case, because of the very small quantities concerned.
MOBILITY
Provoked Pb MObility: Speciation Study The method traditionally used for heavy metal mobilization studies is based on speciation studies (sequential and selective extraction procedures) (Tessier et al, 1979; Forstner, 1985). Chemical extraction sequences were applied to differentiate between the exchangeable, carbonated, reducible (hydrous Fe/Mn oxides), oxidizable (sulfides and organic phases), and residual fractions. The undisputed advantage of this approach to estimate the long-term effects on metal mobilities lies in the fact that the rearrangement of specific solid "phases" can be evaluated prior to the actual remobilization of certain proportions of an element into the dissolved phase (Forstner, 1985). Figure 9.6 shows the fractions (extracted Pb"'lOO/total Pb) extracted in the three sequential extraction steps applied to eight representative samples. The first step corresponds to the acid soluble fraction which represents the metal bound in the exchange positions, water soluble metal and carbonate-bound metal; the second step corresponds to the reducible fraction (metal bound to iron and manganese oxides), and the third step to the oxidizable fraction of organic matter and sulfide-bound metal, although other unidentified phases may also be extracted (Tessier et al., 1979; Forstner, 1985; Kersten and Forstner, 1986; Ure et al., 1993). There may also be other reactions involving desorption, coprecipitation, or neutralization during the three extraction steps (Khebonian and Bauer, 1987; Martinez-Sanchez et al., 1996). Experiments with calcareous soils applying sequential extractions as recommended by the BCR procedure (Ure et al., 1993) show that the fractions extracted from each sample during the different steps vary greatly and there seems to be a relation between the total carbonate content and the mineralogical species in question (calcite, dolomite). In the first step, acetic acid attack neutralizes the calcite when it represents up to 20% of the total mineral content. If the calcite and dolomite contents are above this level they remain in the residue, as can be seen from the diffractogram of Figure 9Ab. All the above is in agreement with electron microscopy observations. Figure 9.7 shows the distribution mappings of lead, which may be partly superimposed on other elements such as S, 0, and Fe, possibly as jarosite crystal, or be widely dispersed. In the second step, nitric acid neutralizes the residual calcite fraction with a large particle size and part of the dolomite (diffractogram Figure 9Ac). SEM reveals that the lead may be associated with dolomite crystals, sulfur compounds, and silicates. Acetic acid is also involved in the third step where, judging from the mineral composition of the residue (diffractogram, Figure 9Ad), it neutralizes the rest of the calcite and dolomite present. Distribution mapping of the lead shows it to be associated with sulfate and silicate crystals, which belong to the least mobile residual lead phase. The quartz/calcite and quartz/dolomite ratios, as deduced from the XRD study of the untreated samples and the residues of the sequential extractions (Figure 9.8), lend weight to the above affirmation. The first ratio increases slightly during the first steps to reach a maximum during the third step, while the quartz/dolomite ratio remains constant at
192
Fate and Transport of Heavy Metals in the Vadose Zone
% Pb extr/Pb total 100
!m3rd EXTR 1IlIID2nd EXTR §1st EXTR
80 60
40
20 OL-----------------------------~
1
2
3
5 sample 4
6
7
8
Figure 9.6. Pb recovered from eight samples using BCR methodology.
SE
Na
Fe
o
K
Mg
Ca
Ph
S
Si
Figure 9.7. Mapping of the chemical element distribution obtained by SEM-EDS (after the first step extraction, acetic acid).
first, rises during the second step, and shows the same behavior as the quartz/calcite during the third step. The BCR sequential extraction method applied to highly calcareous soils does not relate lead to the different speciation forms described in the bibliography, since carbonates are present in the different extraction steps. To summarize, according to the scheme described in Table 9.7, the first step corresponds to the most soluble lead, the exchangeable form, and that bound to part of the carbonates in the form of calcite. However, it seems that significant and substantial errors exist as regards lead speciation when using BCR methodology in soils with a high carbonate content. The carbonate-bound lead is not totally extracted during the first
Lead Mobilization in Calcareolls Agricultural Soils ratio
sample
193
ratio
step 1
step 2
step 3
a
sample
step 2
step 1
step 3
b
Figure 9.8. Evolution of the quartz/calcite (a) and quartz/dolomite ratios (b) in soil samples after each one of the three extraction steps.
step, and remains unquantified in the following extraction stages. Consequently, the lead bound to reducible forms (which is given theoretically by the result found for the second extraction stage) also includes a part of the metal still bound to calcite and dolomite. Lastly, the lead bound to organic matter and sulfides (third step) is also higher than the true value, because it also includes the metal remaining in the carbonates which have not been neutralized in the previous steps. It is clear, then, that the results for lead mobilization given by BCR in such soils are erroneous since, in fact, lead is extracted from carbonate during all three steps. Lastly, the application of these methods to calcareous soils does not predict the mobilization provoked by different environmental conditions, such as changes in pH, redox potential, etc. The reagents used in all the steps were acids, which react with the carbonates present in these soils and so mask the effect which is to be evaluated.
Mobility in the Vadose Zone Due to the above-mentioned difficulties in obtaining a reliable speciation scheme by means of sequential extractions, different reagents which mimic environmental conditions were used to study mobility in calcareous soils. Table 9.6 depicts the main values obtained in selective extractions. The mobile fraction is extremely slight (0.01 to 0.1%) in aqueous medium, which agrees with the described mineralogy and pH of the soils studied. The exchangeable fraction as determined by using ammonium acetate (pH 7) (Simard, 1993) is low but higher than the soluble phase. It varies with the mineralogy and texture of the clays; that is, with the exchange capacity of these soils, which is a function of these two factors, since the organic matter content is less and the amorphous components scarce. The amount of lead bound to the amorphous forms of iron oxides and extracted with oxalic-oxalate at pH 3 (Chapman, 1965) is very small in spite of the fact that both soluble and exchangeable forms, and part of the lead arising from the mobilization of carbonates are included. These results agree with those obtained in the exchange phase. Given the
194
Fate and Transport of Heavy Metals in the Vadose Zone
Table 9.6. Values Obtained in the Selective Extractions (Soluble x 100ITotal) Total Min Max Mean
mg/kg
H2O
NH4 Ac
HAc 0.5 N
EOTA
254 6513 375
<0.01 0.1 0.02
1,3 3,2 2,1
4,4 37,6 19,9
12,7 70,1 30,2
OTPA 3,1 8,4 4,3
Bicar
Oxal
Cit-Oit
Pyro
0.8 3.1 1.8
1,4 8,2 3,5
5,3 64,7 27,4
4,7 37,3 14,7
Table 9.7. sequential Extraction Schemes Kersten and FOrstner, 1986
EC/BCR; Ure et al., 1993
EC/BCR; Martinez et al., 1996
Exchangeable ions
1 M NH4 OAc
0.1 M HOAc
Soluble + Exchangeable + Carbonate-Associated
Carbonateassociated metals
1 M NaOAc, pH 5 w/ HOAc
Easily reducible (Mn-oxides) Moderately reducible (amorph. Fe-oxides)
0.01 M NH2 OH HCI w/ 0.Q1 M HN03 0.1 M oxalate buffer pH 3
0.1 M NH2 OH HCI HN0 3 pH 2
Carbonate - Associated + Reducible
Sulfidic/organic
30% H20 2 pH 2/ 0.02 M HN0 3
8.8 M H20 2 (2x); NH 4OAc/HOAc
Carbonate - Associated + Sulfidic /Organic
Residual
Hot HN03 cone.
adsorption properties of the amorphous substances and their great contribution to the soil's exchange capacity, it might be expected that high exchangeable lead values would result from high levels of amorphous substances, although, in fact, the opposite occurs. This extractant also acts on active calcium carbonate since, due to its very small particle size, the compound is partially dissolved, leading to a partial precipitation of the Ca and Pb as oxalate and to the mobilization of the metals in the crystalline structure of this carbonate. This is revealed in the XRD diagrams of the treated samples, in which reflections, which can be assigned to phases of calcium oxalate in different states of hydration appear. The quartz/calcite and quartz/dolomite ratios also increase in the treated samples. Citrate-dithionite is a reducing reagent with a pH near to neutrality, which extracts easily extractable amorphous forms of low crystallinity (Mehra and Jackson, 1960), such as Fe and Mn oxides and oxyhydroxides and the metals associated with them. The complexes that citrate forms with the metals studied are very stable. Since the metals are released from amorphous forms or those oflow crystallinity by reduction, they are easily extracted, and are, in some cases, the most abundant fraction. Several carbonates are also solubilized in this extraction, which is seen when calculating the quartz/calcite ratio although there are no precipitation reactions as in the case with oxalates, and the dolomite is hardly affected. The water soluble fractions are included in this fraction (PerezSirvent et al., 1997).
lead Mobilization in Calcareous Agricultural Soils
195
Lastly, sodium pyrophosphate is a reagent which extracts forms bound to humic substances. As indicated above, it dissolves part of the calcium carbonate and dolomite, precipitating the calcium phosphate and extracting other materials contained in the carbonate structure by complexation. In our case, because of the low humic matter content of the soils studied, it acts as a carbonate complexer and extractor.
Pb Assimilation by Plants It is generally accepted that the lead levels in soil extracts obtained when using acetic acid M 0.5 (MAFF, 1981), EDTA-ammonium acetate-acetic acid (pH 4.65) (Lakamen and Ervio, 1971), or with DTPA (Linsay and Norvell, 1978) provide a good assessment of metal assimilation by plants. To check that these particular extraction stages really indicate the quantity of lead assimilated by plants growing in calcareous soils, both the vegetal material and the extracts obtained by using these recommended extractants were analyzed. Table 9.6 shows the results obtained for the soil extracts. It is clear that the results considerably differ from one extractant to another, which means that the above assumption is not valid when dealing with the particular type of soils discussed here. EDT A-ammonium acetate-acetic acid (pH 4.65) acts by complexing metals which seem to be associated to inorganic forms of Fe and Mn with a low crystallinity and fine granulometry; that is to say, plant-available metals. In our experiments this reagent also acted by neutralizing carbonates, as can be seen from Figure 9.9, which represents the diffractogram of the residue after the reagent has acted on a soil sample. The diminution in the calcite reflection and background noise, which can be put down to the presence of amorphous substances, is clear. For this reason, the values for plant-assimilable lead which are obtained by this procedure (12-70% extractable) are too high to represent the real in situ absorption by the plant. Both the soluble metal and the metal associated to the exchange complex and to finely divided forms are affected by complexation with DTP A. For this reason, the low values observed in this fraction are the result of the above-mentioned behavior of Pb. The ray diagrams obtained after treatment show decreased background noise, (as do the EDT A residues), increased illite crystallinity, and a constant calcite content. Acetic acid is the most problematic reagent when evaluating plant assimilable Pb in carbonated soils, since it dissolves calcite totally and dolomite only when it is finely divided. This extraction cannot be considered as suitable for establishing plant availability. The high solubility of lead in its presence is evident when it is bound to calcite, reflecting the effects of complexation (high solubility of lead acetate) and of neutralization. As indicated above, in order to ascertain the reagent that best reflects the plant availability of Pb, the metal content of plants cultivated in the studied soils was analyzed. The low assimilation of metal by plants in the cultivation conditions used is reflected in the levels found in the root, leaves, and fruit. These are illustrated in Table 9.8, in which the mean values of all the plant samples are depicted. The most suitable reagent in such soils tends to be DTP A, and the least suitable 0.5 M acetic acid and EDTA. Pb mobility was assayed in a bicarbonate medium of pH 8.5 (Olsen and Sommers, 1982) since this reagent mobilizes the plant-available phosphorus fraction. The lead in the above assay also can be easily mobilized in a carbon dioxide rich atmosphere and the
x-
J 96
Fate and Transport of Heavy Metals in the Vadose Zone Q
Qxal
Cltr Pyro
DTPA EDTA HAc RS
o
20
10
30
40
28 Figure 9.9. X-ray diffractograms obtained from a soil sample and from the remainders after selective extractions.
Table 9.8. Mean Lead Concentrations in Tomato Plants (JIg g-l)a Cultivated in the Soils Studied Sample Root Leaf Fruit a
Min
Max
Mean
1.8 1.2 1.8
4.5 2.8
3.0 2.6
3.2
2.4
Results for dried material (OW).
bicarbonate medium would be similar to the bicarbonate environment of the soil solution in the rhizosphere. The experiment would simulate the mobilization of lead fractions provoked by the lateral leaching waters from the most superficial zones containing bicarbonate ions in solution. Even when the extracted fractions are on the order of 1-3%, these are mobilization conditions near those existing in carbonated soils. Indeed, the subterranean water extracted from nearby wells for irrigation purposes is rich in bicarbonate (40 meq/L) and the lead that we identifY in this experiment could pass to the soil solution.
CONCLUSION Because of the high amount of carbonated materials which are supplied by the wadi complex, an immobilization of the lead lixiviated from the slags takes place.
Lead Mobilization in Calcareous Agricultural Soils
197
The lead speciation methods recommended for other soils and based on the use of conventional sequential extractions are unreliable. A different approach which takes into account the particular physicochemical characteristics and mineralogy of these highly calcareous soils is necessary.
REFERENCES Alias, L.J., R Ortiz, J. Martinez, and P. Linares. Mapa de Suelod, Mazarrtfn, 976. ICONA, Universidad de Murcia, 1986. Allan, RJ. Impact of Mining Activities on the Terrestrial and Aquatic Environment with Emphasis on Mitigation and Remedial Measures, in Heavy Metau, W. Salomons, U. Forstner, and P. Mader, Eds., Springer-Verlag. Berlin, 1995, p. 412. Allen, E.H., C.P. Huang, G.W. Bailey, and A.R. Bowers. MetaL Speciation and Contamination of Soil, CRC Press, Inc., Boca Raton, FL, 1995, p. 358. Arana, R., C. Perez-Sirvent, and R. Ortiz-Gonzalez. Explotaciones mineras e impacto ambiental en el sector de Mazarron (Murcia), in Pr06lemdtica Geoam6ientaL y DedarroLio. Part 2. R OrtizSilla, Ed., Actas V Reunion de Geologia Ambientaly Ordenaci6n del Territorio. Murcia. 1993, pp. 811-835. Chapman, H.D. Methodd of SoiLAnaLYdi..l, Part 2. C.A. Black, Ed., American Society of Agronomy, Inc., Madison, WI, 1965, pp. 891-900. Forstner, U. Chemical Forms and Reactivities of Metal in Sediments, in ChemicaL Methodd for AJdedding Bio-Availa6le Metau in SLUJged and Soiu, R Leschber, RD. Davis, and P. L'Hermite, Eds., Elsevier Applied Science, London, 1985, pp. 1-30. Hayes, K.F. and J.O. Leckie. Modeling ionic strength effects on cation adsorption at hydrous oxide/solution interfaces. J. CoLloUJ Inter:/. Sci. 115, pp. 564-572, 1987. Hueso, R, J. Rodriguez-Gordillo, and F. Lopez-Aguayo. Las jarositas de Sierra de Almagrera (Almeria). Mineralogia y genesis. BoL. Soc. Elp. Mineralogta, 4, pp. 29-36, 1981. Kabata-Pendias, A. and H. Pendias. Trace Elementd in Soiu and Plantd. 2nd ed., CRC Press, Inc., Boca Raton, FL, 1992, p. 365. Kersten, M. and U. Forstner. Chemical fractionation of heavy metals in anoxic estuarine and coastal sediments. Water Sci. Techno!. 18, pp. 121-130, 1986. Khebonian, C. and C.F. Bauer. Accuracy of selective extraction procedures for metal speciation in model aquatic sediments. AnaL. Chem., 59, pp. 1417-1423, 1987. Lakanem, E. and R. Ervio. A comparison of eight extractants for the determination of plant available micronutrients in soils. Acta Agric. Ferm. 123, pp. 223-232, 1971. Liang, L. and J.F. McCarthy. Colloidal Transport of Metal Contaminants in Groundwater, in MetaL Speciation and Contamination of Soil, E.H. Allen, C.P. Huang, G.W. Bailey, and A.R Bowers, Eds., CRC Press, Inc., Boca Raton, FL, 1995, p. 358. Linsay, W.L. and W.A. Norvell. Development of a DTPA test for zinc, iron, manganese and copper. SoiL Sci. Am. J., 42, pp. 421-428, 1978. Lopez-Aguayo, F. and R Arana. Las etapas finales en la alteraci6n supergenica de sulfuros, in RecurdOd Mineraw de Elpaiia. J. Garcia-Guinea and J. Martinez-Frias, Eds. Textos U niversitarios, C.S.I.C., Madrid, 1992. Lopez-Aguayo, F., C. Perez-Sirvent, R. Ortiz Gonzalez, and R Arana. Composici6n quimica de las aguas de lixiviaci6n minera en el Cabezo de San Cristobal (Mazarron, Murcia). Rev. Soc. GeoL. Elpana 5(3-4), pp. 73-79, 1992. Lumsdon, D.G. and L.J. Evans. Predicting Chemical Speciation and Computer Simulation, in ChemicaL Speciation in the Environment, A.M. Ure and C.M. Davidson, Blackie Academic & Professional, Chapman & Hall, New York, 1987, p. 408.
! 98
Fate and Transport of Heavy Metals in the Vadose Zone
Lundgren, D.G. and M. Silver. Ore leaching by bacteria. Ann. Rev. MicrobioL., 34, pp. 263-283, 1980. Manahan, S.E. Environmental Chemutry, 6th ed. CRC Press, Inc., Boca Raton, FL, 1994, p. 811. Martinez-Sanchez, J. FAudio Eddjico de la.J SierrM oe Orce y Maria, Tesis Doctoral. 1982. Martinez-Sanchez, J., C. Perez-Sirvent, and C. Garcia Rizo. Errores de Evaluaci6n de riesgos en la movilizaci6n de metales pesados en suelos carbonatados. Acta.! III Congruo del Meoio Ambiente, Madrid, 1996, pp. 1053-1101. Martinez-Sanchez, J., C. Perez-Sirvent, and C. Garcia Rizo. La problematica del Zn y Pb en el estudio de la posible contaminaci6n por metales pesados en suelos agricolas de zonas aridas, in Recur,}o,} Natura0 y Medio Ambiente en el Surute Penin.Jular, L. Garcia-Rossell and A. Navarro, Eds., Instituto de Estudios Almerienses, 1997, pp. 445-454. Mehra, O.P. and M.L. Jackson. Clay,} and Clay Minerau, Monograph No.5, Earth Science. Series: 1960, pp. 317-327. Ministry of Agriculture, Fisheries and Food. The Analy,}u ofAgricuLturaL Materiau. Her Majesty's Stationery Office, London, 1981. Olsen, S.R. and L.E. Sommers. Phosphorous, in Method,} of Soil Analy,}u, Part 2, 2nd ed., A.L. Page, Ed., American Society of Agronomy, .Madison, WI, 1982. Ortiz Gonzalez, R. Quimumo de !O,} producto'} de aLteracwn ,}uperginica en el dutrito minero de Mazarron (Murcia). Tesis Doctoral. Universidad de Murcia. 1991, p. 466. Perez-Sirvent, C., M.M. Garrido, R. Arana, and F. Lopez-Aguayo. Cristalizaci6n de Sulfatos de Fe (III), Zn y Mg: grupo de la Copiapita. Bo!. Soc. Elp. MineraL. 16, pp. 143-153, 1993. Perez-Sirvent, C., J. Martinez-Sanchez, and C. Garcia-Rizo. Assessment of the Risk of Heavy Metal Mobilization in Calcareous Agricultural Soils, in Proceeding'} of the Third InternationaL Conference on the Biogeochemutry of Trace Element,}, 110-PDF. Contaminated Soils. INRA, 1997. Ritchie, G.S.P. and G. Sposito. Speciation in Soils, in ChemicaL Speciation in the Environment, A.M. Ure and C.M. Davidson, Blackie Academic & Professional, Chapman & Hall, New York, 1995, p. 408. Rodriguez-Estrella, T. EI caracter torrencial de la Rambla de las Moreras (Murcia) y su incidencia en la ordenaci6n del territorio, in Problemdtica Geoambiental y Duarrollo. Part 2. R. Ortiz-Silla, Ed., Actas V Reunion de Geologia Ambiental y Ordenaci6n del Territorio. Murcia. 1993, pp. 835-853. Salomons, W. and U. Forstner. Metau in the Hydrocycle, Springer, Berlin, 1981. Salomons, W., U. Forstner, and P. Mader. Heavy Metau. Springer-Verlag, Berlin, 1995, p. 412. Simard, H. Ammonium Acetate-Extractable Element, in Soil SampLing and MethOd,} of AnaLy,}u, M.R. Carter, Ed., for Canadian Society of Soil Science, Lewis Publishers, Boca Raton, FL, 1993, p. 823. Sposito, G. and S. Mattigod. Geochem, University of California, 1980. Sposito, G. The Chemutry of Soiu, Oxford University Press, New York, 1989, p. 277. Tessier, A., P.G.C. Campbell, and M. Bisson. Sequential extraction procedure for the speciation of particulate trace metals. AnaL. Chem. 51, pp. 844-851, 1979. Thunus, L. and R. Lejeune. In Handbook on Metau in CLinicaL and AnaLyticaL Chemutry, H.G. Seiler, A. Sigel, and H. Sigel, Eds., Marcel Dekker, Inc., New York, 1994. Tudela, M.L. and J. Martinez-Sanchez. Desertizaci6n progresiva en el sureste peninsular y su relaci6n con la puesta en cultivo, in Recur,}o,} Naturale,} y Medio Ambiente en eL Surute Penin.Jular, L. Garcia-Rossell and A. Navarro, Eds. Instituto de Estudios Almerienses, 1997, pp. 433-444. Ure, A.M., Ph. Quevauviller, H. Muntau, and B. Griepink. Speciation of heavy metals in soils and sediments. An account of improvement and harmonization of extraction techniques undertaken under the auspices of the BCR of the Commission of the European Communities. Int. J. Environ. Anal. Chem., 51, pp. 135-151, 1993.
lead Mobilization in Calcareous Agricultural Soils
199
Ure, A.M. and C.M. Davidson. Chemica! Speciation in the Environment, Blackie Academic & Professional, Chapman & HalL New York, 1995, p. 408. Yushkin, N.P. Mineralogy of Pai-Khoi copiapite, Akad. Nauk. SSSR Komi Fi!, InA. Ceo!., 45, pp. 79--86, 1984. Zodrow, E. Hydrated sulfates from Sydney Coalfield, Cape-Breton Island, Nova-Scotia, Canada. The copiapite group, Am. Mineralogidt, 65, pp. 961-967, 1980.
CHAPTER 10
Metal Retention and Mobility as Influenced by Some Organic Residues Added to Soils: A Case Study Luis Madrid
INTRODUCTION This chapter provides a review of the present knowledge of how retention and mobility of metals in soils and other environmental systems are influenced by organic substances, especially those present in residues added to soils for recycling purposes, and is further illustrated with examples of recent research carried out with an agricultural residue which represents a major concern in Mediterranean countries.
Soil as a Sink for Trace Metals As Domergue and Vedy (1992) say, "mobility is a concept frequently used in soil science to estimate the risk of contamination of other environmental compartments by toxins, and especially by heavy metals." However, the definition of mobility depends to a large extent on the scientific field where it is used. In the following text, we shall use this term in the context of the dissolved metal forms in opposition to those bound to solid phases, as the former are likely to be those responsible for immediate environmental risks. Nevertheless, it must be kept in mind that long distance transport of heavy metals occurs largely as part of fine solid particles of earth materials, causing a slow buildup of soil pollution (Tiller, 1989). Soils receive trace metals from various sources, mostly from lithogenic origin, i.e., from the parent material, or from anthropogenic origin; e.g., present in fertilizers, as part of wastes (industrial, agricultural, or otherwise), irrigation waters, etc. (Singh and Steinnes, 1994). When metals enter the soils, they undergo a number of reactions, namely dissolution/precipitation, adsorption/desorption, complexation, inclusion in minerals through formation of solid solutions (Alloway, 1990), which determine their "final" distribution in various chemical forms in the soil with very different solubilities. This final 201
202
Fate and Transport of Heavy Metals in the Vadose Zone
distribution is frequently described in terms of several fractions, operationally defined according to a sequential extraction technique. Although different fractionation techniques have been proposed, most of them distribute the soil metals in fractions called "soluble," "exchangeable," "bound to carbonates," "bound to iron/manganese oxides," "bound to organic matter," or similar names, and that part of the metal content which cannot be extracted but by strong concentrated acids is called "residual." Although these names refer to a given chemical nature, in fact the fractions cannot usually be identified with a definite group of chemical compounds. In the past, the term "speciation" was used for the process of distinguishing different species of an element, and such species could be defined either operationally (by the procedure or extractant used) or functionally (bioavailable, exchangeable, etc.). Nowadays, the term "speciation" is preferred to be applied to the distinction of specific chemical forms of an element (Ure et al., 1993). Figure 10.1 is an example of the fractionation of some metals in Polish soils, where the solubility or availability of each fraction increases from bottom to top (data from KabataPendias, 1995). Regardless of how the fractions are called, those fractions that are more easily released usually account for a minor proportion of the metal content. Therefore, as a consequence of the set of reactions mentioned above, soil acts as a geochemical sink for metals and controls their transport to other media (Kabata-Pendias, 1995), so that their persistence in soils is high: while the residence time of easily leachable elements like Ca, Mg, etc., is usually below 3pO years, for many heavy metals residence times of several millennia have been reporte~J:3owen, 1979). Consequently, the budgets for some trace metals in soils are positive in many cases. Table 10.1 gives some examples for European soils (from Kabata-Pendias, 1995). Only in the case of some acid forest soils, losses of Mn or Zn are found. Summarizing, metals tend to accumulate in soils, either by long-distance transport of fine solid particles or by immobilization of the soluble forms, so that a sort of "chemical time bomb" is building up. The environmental risk of such a chemical time bomb depends o~ whether various chemical "fuses" may cause a more or less sudden increase in solubility, passing to ground or surface waters, or becoming available to plants.
Modeling Approaches for Retention of Metals by Soils The use of mathematical models for a quantitative description of the various reactions involved in a given process is a common practice in the study of environmental problems (Basset and Melchior, 1990). Various models have been specifically developed for understanding surface reactions involving metals or other ionic species. Some of them, which can be illustrated by those used by Davis and Leckie (1978) and by Barrow et aI. (1981), assume some "mechanistic" characteristics of the retention process, e.g., pairwise association between adsorbing ions and oppositely charged surface groups, or location of adsorbing ions in planes of a given value of the electric potential. In both models, a number of parameters that usually cannot be estimated by independent experiments have to be adjusted by some iterative procedure. Another group of models (Amacher et aI., 1986) do not assume any particular mechanism for the retention reaction. Some of them use simple equations to describe equilibrium situations, e.g., one- or two-surface Langmuir equation, Freundlich equation, etc. Others try to describe the time dependence of the solution concentration of the sorbing
Metal Retention and Mobility as Influenced by Some Organic Residues
203
C:=:J Residual
rzza
Organic matter
~
Bound to oxides ~ Exchangeable IIIJJ Soluble
100
-!= -.s...8
80
co
....
60
...= 5Col
"" ~
~
40
20 '~---
Cd
Zn
Cu
Pb
Figure 10.1. Example of fractionation of some metals in soils (from Kabata-Pendias, 1995, with permission).
Table 10.1. Metal Budgets of Surface Soils in Europe (g ha-1 yr- 1)a Ecosystem (Method)
country
Cd
Cu
Mn
Pb
Zn
Pine forest (seepage water) Spruce forest (Iysimeter) Farmland (drainage water) Farmland (seepage water)
W. Germany Sweden Denmark Poland
3 -1 3
10 5
-360 -600
2
14
90
104 75 260 160
134 -130 130 360
a
From Kabata-Pendias, A., in Heavy Metals, Problems and Solutions, W. Salomons, U. Forstner, and P. Mader, Eds., Springer-Verlag, Berlin, 1995, pp. 3-18, with permission.
species, assuming reversibility or irreversibility and a given reaction order. In general, these models are too simple and are not likely to represent the experimental data of retention of metals by soils satisfactorily, probably because such retention is due to many
204
Fate and Transport of Heavy Metals in the Vadose Zone
different reactions, as mentioned above. That is why models considering the contribution of various simultaneous, concurrent reactions are probably more realistic. In an exhaustive review, Selim (1992) described a number of multireaction kinetic models, empirical in nature, based on the assumption that a fraction of the total sites are highly kinetic, and the remaining sites interact slowly, irreversibly, or instantaneously with the adsorbing species. Some models include concurrent and concurrent-consecutive processes. Many metal retention processes are well described by this kind of model (Selim et aI., 1990).
METAL CONCENTRATIONS IN THE SOIL SOLUTION The various metal retention processes that may occur in soils cause the solution concentration of many metals to be generally below the values corresponding to the solubility of pure components. A detailed description of the reactions controlling heary metal solubility can be found elsewhere (McBride, 1989). The occurrence of coprecipitation of metals with the major metal oxides from soil solutions is accepted, but the nature of the precipitate is not well defined. It can be easily shown (Driessens, 1986) that the solubility of a metal can be lowered in a mixed ionic compound relative to the solubility of the pure compound, and the activity of the metal compound is a function of its mole fraction. That means that when a metal is incorporated in an iron or aluminum oxide in trace proportions, the solution concentration of the former may be several orders of magnitude below that expected from the solubility of the trace metal compound alone (McBride, 1989). Various observations suggest that metals retained by oxides are concentrated at solid surfaces (McBride, 1978), and little or no difference is found whether they are "coprecipitated" or adsorbed on preexisting oxide surfaces. Formation of surface solid solutions has been shown to be the cause for Cu adsorption by calcium carbonate (Papadopoulos and Rowell, 1989), and by soils rich in lime (Madrid and Diaz-Barrientos, 1992), and in both cases the resulting equilibrium concentrations are several orders of magnitude below that corresponding to the solubility of Cu carbonate. Figure 10.2 shows that, while Ca solutions in the presence of CaC03 produced equilibria located close to the solubility line of CaC03 in solubility diagrams, Cu solutions showed a definite tendency to be in equilibrium with Cu(OHh, considerably less soluble than the metal carbonate. Madrid and Diaz-Barrientos (1992) treated calcareous soils with acetate buffer at pH 5 to eliminate carbonates and equilibrated the original and treated soils with Cu solutions and subsequently with Cu-free solutions, and found that all the equilibria with the original samples corresponded with solubilities considerably below those of the oxide, hydroxide or basic carbonate, while the equilibria with treated soils were close to the line for CuO (Figure 10.3). Lindsay (1979) estimated that the solubility of Cu in soils could be described by an equilibrium:
which causes a value of the ratio Cu 2+/(H+)2 of 10 2.8, giving Cu 2+ concentrations several orders of magnitude lower than those corresponding to the solubility of most Soil-Cu /mmerals. Lindsay (1979) considered it likely that "soil-Cu" could correspond to cupric fer0te, CuFe204' and suggested that the corresponding Zn compound, Zn-ferrite, could
Metal Retention and Mobility as Influenced by Some Organic Residues
• ..
205
Equilibrated with CU2+ Equilibrated with Ca2+
• CuO
10
CU2(OH)2COJ CU(OH)2
8
~
~
c. CuCOJ
6
CaC03
4
2L---....lL.._--+----L_ _- - l
6
7
8
9
pH Figure 10.2. Solubility diagrams for Cu 2+ or Ca 2 + solutions equilibrated with A.R. CaC0 3 (from Papadopoulos and Rowell, 1989, with permission).
also account for the solubility of Zn in soils. However, other authors (Baker, 1990; McBride, 1989) believe that formation of mixed oxides similar to cupric ferrite is unlikely in the conditions found in soils. There is general consensus that adsorption on the various surfaces present in soils is the main process responsible for the low metal solution concentrations often found in many soils. Only in some cases with high metal loadings do some solid phases determine metal solubility; e.g., malachite in soils with very high eu contents (McBride and Bouldin, 1984; Lund and Fobian, 1991).
FACTORS CAUSING A REVERSAL OF IMMOBILIZATION
Altho~;):\m a large scale, soils tend to immobilize trace metals, some conditions of the soil envir~nment can undergo changes driving to a localized reversal of the process. Metals with different valences are frequently more soluble in their lower oxidation states, so that reductive conditions may cause a local increase in solubility, and a decrease in pH will also cause an enhanced solubilization of heavy metals. A decrease in pH also causes a decrease in negative electric charge of variable-charge surfaces, so that a decrease in
206
Fate and Transport of Heavy Metals in tile Vadose Zone
Adsorption Desorption o c Untreated soils ..
'II'
Treated soils
0
-1
.CU(OH)l -2
C
= -=
-3
U
'-' ~
-4
:t:
0
%
-5
0
0
0 0
c
0
c
0
c
0
CC c Cc C C
-6
Cc
c
CC
-7 4
5
pH
6
7
Figure 10.3. Solubility diagrams for adsorption and desorption equilibria of Cu with soils rich in lime, before and after being treated with acetate buffer at pH 5 (from Madrid and Diaz-Barrientos, 1992, with permission).
adsorption of cations also contributes to an increase in the presence of metals in solution (McBride, 1989). Such effects are usually localized and easily reversed when the cause disappears. An example of such reversible solubilizations can be found in the case of some rivers close to mining areas, which may show low pH values and high concentrations of several metals immediately downstream from the mining site, but a short distance further downstream the pH increases several units, metal concentrations decrease sharply and the bank sediments show high contents in metal compounds of low solubility (Arambarri et aI., 1984). The presence of organic compounds with functional groups with ability to form stable complexes with transition elements is another factor which may increase the solution concentration of heavy metals, and is a known fact since the end of the past century. In contrast with acidification or with reducing conditions, formation of soluble complexes may enhance the transport of metals in solution to long distances, due to the high stability frequently shown by such complexes. Frequently a high selectivity is found in the interaction of complexants with metals, as shown by Xue et aI. (1995), who found that the concentration of free Cu 2 + in a eutrophic lake was about 106 times lower than total Cu concentration, whereas a substantial part of Zn was present as free Zn ions and ~ak complexes. I
Metal Retention and Mobility as Influenced by Some Organic Residues
207
The presence of a complexing agent in waters in contact with soils or sediments may actively desorb metals from the solids, or hinder the adsorption processes responsible for metal retention. This was shown to happen in the case of synthetic complexants added to waters: in a study of nitriloacetic acid (NT A) as a possible alternative to polyphosphate in detergents, it was found that the presence of NT A in waters resulted in a significant increase in solubility of metals in river sediments and a decrease in their metal-sorbing ability (Salomons and Forstner, 1984). Also, Davis and Singh (1995) found that when soil columns contaminated with Zn were leached with EDT A solutions, the metal retained was almost quantitatively removed. Recently, we have seen that eu added to soil columns is strongly removed by glyphosate, a herbicide with zwitterion structure.
INTERACTION WITH NATURAL ORGANIC MATTER The behavior of natural organic polymers as humic or fulvic acids is not as straightforward as that of simple complexants like NT A. Their action on heavy metals is twofold: on the one hand, soluble components of the humic substances can form complexes which will be responsible for an enhanced solubility of metals, and, on the other hand, fine solid particles of humic substances or organic coatings of minerals often contribute to the retention (and immobilization) of metals. In either case, the interaction with the metal ions occurs through various kinds of functional groups present in the organic polymers (Senesi, 1992), and whether the interaction results in metal mobilization or immobilization is strongly related with the size and solubility of the organics rather than with the functional groups involved. Fulvic acids play an important role in the transport of heavy metals in water, due to their lower molecular mass and much greater solubility as compared to humic acids (Forstner and Wittman, 1983, and references therein). The stability sequence of the complexes for some selected cations is considered to be (Klamberg et aI., 1989)
Therefore, organic matterlheavy metal interactions determine three broad, dynamically interrelated groups of species with different influence on the availability of the metals for the living organisms (Figure lOA): (a) Solid organic surfaces can retain metals, which are not immediately available to plants. From this point of view, metals retained in this way behave similarly to those retained by inorganic solid surfaces. However, some metals, e.g., cadmium, may show a preferential association with aqueous solution components as compared to solid surfaces (Neal and Sposito, 1986). (b) Some components of the natural organic matter can eventually be dissolved and form metal complexes in solution, which in turn will be affected by adsorption/desorption equilibria with the solid surfaces. Monomeric organic molecules (e.g., released from plant roots) or large polymers can contribute to these equilibria. If such complexes are strong enough, usually the toxic effects of t~ metal are ameliorated, but it has been observed that some lipid-soluble comjexes can rapidly penetrate a biomembrane, so that substances forming
208
Fate and Transport of Heavy Metals in the Vadose Zone
Increasing bioavailability
:> Low mol. weight complexes
Solid surfaces
...E 1 - - - - - - - -
Free hydrated metal
High mol. weight complexes Figure 10.4. Schematic distribution of metal species from the point of view of their availability to biological systems.
this kind of complexes can cause an enhanced toxicity of the metal (CarlsonEkvall and Morrison, 1995). (c) The free or weakly complexed metal ions will usually be the most bioavailable species. Their concentration will be influenced by the various equilibria with complexing species in solution and by adsorption/desorption reactions on organic as well as inorganic solid surfaces. Another important feature that distinguishes natural organic matter is the heterogeneity of its complexing sites. Figure 10.5 shows schematically the dependence of the cumulative distribution of complexing sites of fulvic acid upon their stability constant~ (based upon data of Buffle, 1988). Therefore, a single value for the stability constan: cannot be defined, although different graphical approaches for estimating apparent o~ "average" constants have been proposed (Fitch and Stevenson, 1984). Another useft:. parameter in this context is the complexation capacity, defined as the maximum quantity 0: a given metal that can be bound per gram of the substance (Perdue, 1988). The amoun r and distribution of complexing sites of natural complexants have been estimated by vari0us techniques, especially those able to distinguish free from complexed metals. The use of ion-selective electrodes (e.g., Buffle et al., 1977, 1980; Gamble et al., 1980; Stevensor. and Chen, 1991; Stevenson et al., 1993) appeared to give reliable results in terms at applicability and sensitivity (Sterrit and Lester, 1984). Polarographic techniques han proved to give information of the lability of the complexes, and have been successfull.,. applied to the characterization of metal complexes with humic and fulvic substance, (Greter et al., 1979; Wilson et al., 1980; Filella et al., 1990; Pinheiro et al., 1994). Equilibration with ion exchangers has also been used for studying metal complexation with organic matter. The method proposed by Schubert (1948) could be used only for mononuclear complexes with respect to the metal ion, but Zunino et al. (1972a,b) proposed some changes in the equations involved that allowed them to be applied to complexes at the type Ma(ligandh, where a and 6 are integers;::: 1.
Metal Retention and Mobility as Influenced by Some Organic Residues --
. --- -
-
---
---
--~
------------
---
209
--
80
:::t.
r Il
70
....... ...= 60 tIj ~
rIl
Cil
-a=-
50
u
40
Q
""'a. Q
~
~
a:=
30
= 20 ~
~
-a ~
«S
:= :=
U
10 0
4
5
6
7
10gKc Figure 10.5. Example of cumulative distribution of metal complexing sites in a solution of fulvic acid as a function of their stability constants.
EFFECT OF ORGANIC RESIDUES ON METAL SOLUBILITY Land application is a widely used practice for disposal of many organic wastes. They frequently contain substances which improve soil fertility, although other components can have undesirable consequences for crops, so that research on many different aspects of their addition to soils has increased in the last few years. One of the points that must be carefully considered when such wastes are added to soils is their possible effect on metal mobility or solubility: such organic residues are likely to contain polymers with not very different properties from those of natural humic or fulvic substances, and therefore it can be expected that metal complexes similar to those known to form with the latter will also be formed with such residues. In soils receiving dairy cattle-manure slurry, Del Castilho et al. (1993) found significant increases in tKe concentrations of Cd, Zn, and Cu in soil solutions. In the case of Cu, its concentrat4>ns were correlated with the dissolved organic carbon, while Zn and Cd were also infl~ed by low pH and high levels of electric conductivity. They concluded that Cu complex~~ showed high stability, and a considerable part were of high molecular weight and nonlabile. Japenga et al. (1992) also studied the effect of the liquid fraction of animal manure on heavy metal solubilization in soil, and found a significant relationship between dissolved organic carbon and Cu concentrations in aqueous extracts (Figure 10.6). They concluded that, together with pH, complexation involving dissolved, high molecular weight organic matter is the most important driving force for heavy metal solubilization. Metal complexation was also considered to be one of the causes of metal leaching from a soil in a reed bed
210
Fate and Transport of Heavy Metals in the Vadose Zone
1.2
.------,I----~ I - - - - rI -- - - . -I -- - - - ,
1.0 -
-
!l
0
-
00
00
0.8
0
00%
f-
0
0
bi)
e
--g
0
-
0 0
0.6
f-
0.4
f-
0.2
-
-
0
Q (J
=
U
o
o ~o 0
0.0 0
-
o
-
f8
I
I
I
I
200
400
600
800
1000
DOC (mgI;l) Figure 10.6. Relationship between dissolved organic carbon and Cu in aqueous extracts of a soil mixed with liquid animal manure (selected data from Japenga et aI., 1992).
system continuously flooded with sewage (Wenzel et al., 1992). Barrado et al. (1995) also concluded that extracts from eucalyptus and oak leaf litter showed complexation ability for metals, and could estimate the complexing constants for various metals.
THE CASE OF SEWAGE SLUDGE Addition of sewage sludge to soils was found to decrease the sorption of Cd at low concentrations of this metal (Neal and Sposito, 1986). In soils treated with sewage sludge and artificially contaminated with high doses of Cu in the form of Cu carbonate, Cheshire et al. (1994), using electron magnetic resonance, found evidence of Cu solubilization through complexation. The results for organically bound Cu in the soil solution indicate oxygen ligand coordination in equatorial arrangement. Keefer et al. (1994) also found significant metal-organic association in soils amended with sewage sludge, and McGrath and Cegarra (1992) observed large increases in the most soluble fractions of metals in a soil with long-term applications of sewage sludge. They found that the fractionation of metals in the original sewage sludge differed from that observed in the soil treated with the residue. Frequently, sewage sludges have relatively high metal contents, so that their effect on metal mobility in soils has been often attributed to the metals present in the residue itself: Sposito et al. (1982) concluded that the accumulatienohnetals in soils amended with sewage sludge was governed by the metal content in the sludge, and Cavallaro et al.
Metal Retention and Mobility as Influenced by Some Organic Residues
211
Table 10.2. Examples of Maximum Permissible Concentrations of Some Metals in Soil after Application of Sewage Sludge (mg kg- 1 ) Soil pH
5.0-5.5 5.5-6.0 6.0-7.0 > 7.0 > 5.0 < 7.0 > 7.0 a b
Country
Cd
a
UK UKa UKa UKa UKa Spain b Spain b
3 1 3
Cu
Ni
80 100 135 200
50 60 75 11P
50 210
30 112
Pb
Zn
200 250 300 450 300 50 300
150 450
Data selected from Department of the Environment, Code of Practice for Agricultural Use of Sewage Sludge, HMSO Publications, London, 1992, p. 6. Data selected from Boletin Ofloal del Estado, No. 262, Madrid, 1990, p. 3234.
(1993) found that increases in Mehlich-3 extractable eu and Zn in soils treated with sewage sludge were similar to the amounts of these metals added in the residue. Most countries have established regulations concerning the use of sewage sludge on the basis of the maximum permissible contents of potentially toxic elements in soil after application of sewage sludge (Table 10.2) and annual rate of addition of such toxic elements, so that no legal limit exists if the sewage sludge added to a soil shows a low content in toxic metals. It is thus forgotten that solubilization of the metals already present in the soil can be enhanced by complexation, as shown by some of the authors quoted in the previous paragraph. This lack of attention paid to the effect of soil management practices, especially the use of sewage sludge, on the solubility of the metals present in amounts below the legal limits in the soil has been claimed by several authors (McBride, 1994; Evans et aI., 1995), and has been favored by the conclusions of some authors, who even found a decrease in metal mobility in some cases (Emmerich et aI., 1982; Saviozzi et aI., 1983; Hooda and Alloway, 1994).
A MEDITERRANEAN CONCERN: OLIVE Mill WASTEWATER
Setting Up the Problem In areas with extensive production of olive oil, disposing of the residues from manufacturing plants for this agricultural product represent a major concern. The traditional procedure implies generating large amounts of wastewater (called aLpecbin, from now on OMW) with extremely high BOD and other undesirable properties which have caused the existence of regulations prohibiting its disposal in rivers since 1981. Everyyear, about 10 million tons of this waste have to be disposed of in the Mediterranean countries, mainly by storing them in evaporation ponds, composting the resulting sludge with other plant refuse or, in countries where the production of this residue is not especially high, discharging them into watercourses. In recent years, olive oil production plants are being adapted for new techniques using much smaller volumes of water, so that production of OMW is decreasing sharply, but its disposal must still be considered until total substitution of the old manufacturing plants, and the existence of small factories which cannot afford the changes cannot be forgotten, at least during several years in the near future.
212
Fate and Transport of Heavy Metals in the Vadose Zone
While the effect of sewage sludge on heavy metal availability has been extensively studied, as summarized in the previous section, literature on the relationships between heavy metals and OMW is scarce, although in the last few years some authors have found evidence of significant metal solubilizing effects of this residue, both when added to soils and when present in freshwaters. OMW is a slightly acid (pH 4-5), dark-colored aqueous phase with highly variable composition, containing 10-15% organic matter and 1-2% of mineral salts. Its contents in heavy metals' is usually negligible, except Fe (1020 mg L- 1), Mn « 5 mg L-1), and Zn « 2 mg L- 1). Several authors have given detailed descriptions of the composition of this waste (e.g., Gonzalez-Vila et aI., 1992; MartinezNieto and Garrido-Hoyos, 1994).
Effect of OMW on Metal Retention Properties of Soils As with other organic wastes, one of the first ideas that emerge when recycling OMW is considered is its application to soils as fertilizer, and it has been frequently used to irrigate olive trees. Considering previous knowledge of the nature of this residue, SaizJimenez et aI. (1987) deemed it of interest to carry out a detailed study of its chemical composition in order to evaluate its potential value as soil fertilizer. They concluded that the composition of the humic fraction was different from soil humic acids, but still suggested that the residue had good properties as fertilizer. On a relatively short-term basis, applications of a composted olive mill sludge to soils have been found to cause no harmful effects on plants, the improvement of soil physical properties is apparent, and significant increases are found in soil organic N. Also, increases in available Cu, Zn, Mn, and Fe determined by DTPA extraction have been observed (Martin-Olmedo et al., 1995). In a study specifically oriented toward the effect of OMW on metal availability, Perez and Gallardo-Lara (1993) found that although OMW initially caused a slight decrease in Zn availability and hardly any effect on Cu availability, a significant residual increase in Cu availability was observed after growing barley and ryegrass. A fundamental aspect that must be considered is whether the presence of OMW affects the action of soils as a sink of heavy metals which are added in soluble forms. Madrid and Diaz-Barrientos (1994) chose three soils (called A, B, and C) with widely differing contents in organic matter, carbonate, and clay fraction and CECs for that purpose. Soil A had been manured in the field with 150,000 kg ha- I of a compost obtained from OMW and other plant refuse, and soil B had received a similar dose of raw OMW. Soil C was untreated. Moreover, samples of the three soils were aged in vitro with freeze-dried OMW in a proportion corresponding to twice the dose received by soils A and B in the field. The reaction of several metals with the original, manured, and aged samples was studied. Figure 10.7 shows the results for the adsorption isotherms obtained for Cu and Zn. The adsorption of these two metals was strongly decreased by mixing and aging the soils with OMW, while manuring with OMW or compost obtained from it only caused a significant decrease in the case of Zn. Manuring even caused a slight increase in Cu adsorption by soil A. The pH values of the adsorption experiments did not show differences large enough to explain the differences in adsorption. The authors suggested that the decrease in adsorption when OMW was added in large doses could be the result of coating the sorbing surfaces with organic matter. However, in the samples containing
o
::~~ -
'7bt) ~
80 60
1
40 20
!.
0
II
~
150
.f!
\I
)7
l
Manured samples
90
f
Cu/A
jl ~
~o
30
40
5~
\l
Aged samples
/ J::
~~
~
~ ~ ~ ;0 I'll I'll :::l
.... ....
150
o·
-I 100
r:f
60 f-'Jf
if
10
-
~ ::[~
~t
] 200,0, a.. ~
!
;
rl~
I
Original samples
:::l III
d
30 fg 0 150
£" •• fD
t7" t
~}
4
6
8
10
12
-I
50
j
0
1~ 150
:::I C.
~
o
g
0
5
10
15
20
'Ill VI
Ii
Zn/A
100 f-
/
~
aE"
_ -I 100
I'll
:::l
~ 100
50
;:;: '<
I"'l
~?
_____________
o ~f: I o 10 20
~
j 30 40 50 60 70 80
50
0
I'll
iI/
~ ZnlB
~ 0
10
20
30
40
C. c;
50
'< VI
~ 50
o
3
I'll
0 '---'-----"- - L - L - L - - - - ' - - - - - - - - - - ' - - - L 0
10 20
30 40
50 60
70
Final metal concentration (J.&mol 1;1)
o....
ttQ III :::l
n°
?i VI
Figure 10.7. Adsorption isotherms of Cu and Zn by three soils (A, B, and C), before and after different treatments: Soils A and B were manured with composted and raw OMW, respectively, and the three soils were aged in vitro with a high dose of freeze-dried OMW (from Madrid and DiazBarrientos, 1994, with permission).
C.' C
I'll VI
....
N
w
214
Fate and Transport of Heavy Metals in the Vadose Zone
OMW some oversaturation was found in eu experiments with two of the soils, both in the adsorption and the desorption steps, with respect to the eu compounds likely to form. This result suggests that formation of soluble complexes could also contribute to the decrease in adsorption of this metal. Selim et al. (1990) considered the adsorption reactions caused by several kinds of surfaces, with different kinetic characteristics. Madrid and Diaz-Barrientos (1996) examined whether the effect of OMW on the metal-sorbing properties of soils manifested itself when the multireaction model (MRM) of Selim et al. (1990) was applied to kinetic data of eu adsorption by two soils which had received OMW compost. The MRM model distinguished several fractions in the amount sorbed: Se' that reacts "instantaneously" with the sorbate in solution, described by a Freundlich equation with a distribution coefficient ~ and exponent" (Equation 1); s, and S2' which react kinetically and reversibly with the solution at different rate, described by the forward and backward rate coefficients k, and k2 for s, and k3 and k4 for S2' and reaction orders nand m respectively (Equations 2 and 3); and finally Sirr' which reacts kinetically and irreversibly with the solution by a first order reaction, described by a rate coefficient ks (Equation 4). The original model considered that some of the fractions could undergo other consecutive reactions, generating further fractions, but Madrid and Diaz-Barrientos (1996) found out that this latter assumption did not increase the percentage of variance explained when the model was applied to their data. The mathematical structure of the model as used by Madrid and Diaz-Barrientos (1996) is summarized below, where p and represent the bulk density and water cont~nt of the soil in the experiments.
e
(1)
(2)
(3) (4) Table 10.3 shows the parameters estimated by Madrid and Diaz-Barrientos (1996) for their original samples and those treated with OMW compost. It can be observed that addition of OMW to the soils causes significant changes on some of the parameters: the instantaneous distribution coefficient ~ is strongly increased by a factor of 50 in both soils; the irreversible fraction Sirr becomes irrelevant, and the rate coefficients of the kinetic fraction s, are significantly decreased. S2 does not show any significant effect. Madrid and Diaz-Barrientos (1996) concluded that the eu-immobilizing action of the soils was altered by the presence of composted OMW, probably by the presence of new solid, organic surfaces, which react instantaneously with the metal in solution. The dependence of the instantaneous fraction upon the solution concentration of eu means that the metal adsorbed by the soil with OMW must easily come into solution if its concentration decreases, in contrast with the behavior of the untreated soils, which hardly release eu by dilution. The disappearance of fraction Sirr also means a decrease in the immobilization of the metal.
Metal Retention and Mobility as Influenced by Some Organic Residues
215
Table 10.3. Average Values of the Model Parameters. k, to k4 and kSl h-'; b, nand m, dimensionless; Ku mg kg-' a
Soil
Parameters of Each Fraction of the Model S, se b k, n Kd k2 k3
A orig. A compo C orig. C compo
9 440 18 925
a
0.4 0.4 0.4 0.9
24.5 1.82 28.4 9.7
2.79 0.79 1.15 0.45
0.3 0.4 0.4 0.4
0.33 0.50 0.61 0.77
S2 k4
m
sirr ks
R2
0.13 0.04 0.08 0.05
0.7 0.5 0.7 0.4
0.04 0 0.15 0
0.998 0.997 0.995 0.994
From Madrid, L. and E. Diaz-Barrientos, Taxicol. Environ. Chern. 54, pp. 93-98, 1996, with permission.
OMW in the Aqueous Phase as a Mobilizing Agent of Insoluble Metal Forms Although discharges ofOMW in watercourses is prohibited, its accumulation in ponds may cause a slow migration of its soluble components to groundwaters. Moreover, accidental releases of significant amounts into rivers do occur. Thus, another important point that must be considered is the solubilization of metals in "immobile" forms when in contact with water containing OMW. Bejarano and Madrid (1992, 1996a) studied the solubilization of heavy metals from river sediments, with high metal contents due to their location close to mining sites, when treated in vitro with dilute solutions of OMW. These solutions were prepared from freeze-dried OMW and adjllsted at pH values between 3 and 5, considering the slightly acid pH of the residue. They found that some metals, e.g., Mo or Zn, were not solubilized by the residue, and in the case of Mo the sediment even retained part of the metal originally present in the OMW. On the contrary, Cu, Fe, and Pb from the sediments were solubilized when in contact with OMW solutions. The presence of OMW favored Pb solubilization at any pH, while Cu and Fe were dissolved to a greater extent than in the absence of OMW only at the higher pH tested. Considering that the solubility of these metals usually increases at lower pH values, this result suggests that the solubilization of these two metals can be related with the formation of soluble complexes with OMW components. Figure 10.8 shows a summary of the results obtained by the authors for Pb and Cu. In the second paper mentioned (Bejarano and Madrid, 1996a) the authors showed that the amounts solubilized by OMW were comparable with the metals originally present in the sediments in forms bound to carbonates and to oxides, according to a conventional fractionation technique. The hypothesis of complexation by OMW components had been previously checked by Cabrera et al. (1986). Using the cation-exchange resin method of Zunino et al. (1972b), they found that the freeze-dried residue showed a complexing ability of 0.66 mmol of Cu per gram of OMW. Bejarano and Madrid (1996b) studied the time-dependence of the release of several metals by solutions with three different OMW concentrations from a river sediment, and the resulting solutions were filtered through C-18 reverse-phase cartridges. The metals complexed by polymers present in the OMW, especially those forming less labile complexes, were supposed to be retained by the cartridges, together with the uncomplexed organic polymers. Previously they checked that no free metal was retained in the cartridges. Figure 10.9 summarizes the results for Cu, Mo, and Zn. The
216
Fate and Transport of Heavy Metals in the Vadose Zone - - - - - - . -----
-- - - - - - - - - - - - - -
---
0.7
-
....,!..;l
10
0.6
-a..
....,!..;l
0.5
OJ)
!.
-.... 1
Q "-1 "-1
"C
,Q
8
OJ)
0.4
6
"C
0.2
...-
0.1
U
Q,j
0.3
Q
~
4
"C
~
='
2
0.0
o
5
10
15
20
25
30
35
o
5
10
15
20
25
30
35
OMW concentration (g C 1) Figure 10.S. Pb and Cu dissolved from a river sediment by solutions of various OMW concentrations and pH values (from Bejarano and Madrid, 1992, with permission).
authors concluded that most of the Cu and Zn released into solution by the OMW was in complexed forms, as nearly 100% of both metals were retained by the cartridges. In the case of Mn, only about 50% of the metal dissolved by OMW was retained by the cartridges, suggesting that complexation was less complete or that the complexes formed were more labile. Considering that some of the results commented on in this and the previous section suggest that some components of OMW do form complexes with several metals, Bejarano and Madrid (1996c) appfied the techniques of Zunino et al. (1972a,b) to determining the complexation parameters of this residue for several metals. They found that the maximum complexing ability (MCA) was inversely related with the ionic radii of the metal ions, and a direct dependence between the logarithm of the conditional stability constants and the metal electronegativity (Figure 10.10). This latter result agrees with the fact that the stability of complexes formed by a given ligand with a series of metals is expected to increase with the electronegativity of the metals (Irving and Williams, 1948), thus showing indirectly that complexation by OMW does occur. The authors concluded that the "average" complex formed is mono-nuclear, with a bidentate bond for Cu 2+. For other larger M2+ ions, a progressive steric hindrance seems to exist. This simple model of a mono-nuclear, bidentate complex between the "average" component of OMW and a metal ion was further developed elsewhere (Bejarano and Madrid, 1996d). The reaction was assumed to be
• o
...
•
'"
v
o
x
1,2,3 g L- I OMW and blank, total extracted 1,2,3 g L- I OMW and blank after filtration
3:
rc ....
to;
;0
...
rc 1.6
...o
~
4
::::l
1.2
to;
-
f
o
~
;:;: '<
-=
~
B 2
B 0.8
•
3:
g
'...:l
'...:l
•
3
....
....
r~l f a 0.4
·J ~
1.2
:::l C.
to; VI
N
0.4 f- 11\
r
-!
S' ..... i: rc
1
::::l ,-.,
rc
0.0
o
0.0
o
5
10
t(days)
15
20
C. 0-
o
5
10
t (days)
15
20
'<
III
o
o
5
10
15
3t'P
20
t (days)
o
I
cia
III
Figure 10.9. Time dependence of metals found in solutions of OMW in contact with a river sediment, and effect of filtering the solutions through (-18 reverse-phase cartridges (from Bejarano and Madrid, 1996, with permission),
::I
n' ;0 1'0 VI
is: c 1'0 \II
N ....
'1
218
Fate and Transport of Heavy Metals in the Vadose Zone
0.5 r--O---r,-----r-,----,r-"'!
--~
5 r--'I"--'I--'I--"--"---'
Cu 0.4
Cu Pb
-
I-
~
I
~
0.3
Zn t-
4
o
o -
I-
o OMn Zn
O
o
~
:::: 0.2 Q
3 -
.! ~
Pb
0.1 -
~
Mn
-
o
o 1 _1 I 0.0 '-----'------'-----~ 0.6 0.8 1.0 1.2
M2+ ionic radius (A)
__
2L-_~~
1.4
1.5
~1_ _L-1_~1L-_~1__~
1~
1~
1~
13
2.0
Metal electronegativity
Figure 10.10. Relationship between maximum complexing ability of OMW and the radius of each metal ion, and dependence of the stability constant on the metal electronegativity (from Bejarano and Madrid, 1996, with permission).
A double acid dissociation was assumed for the organic ligand, defined by two acid dissociation constants. From Cu 2+ and Zn 2+ data of experiments similar to those mentioned above with reverse-phase cartridges, estimates of free metal concentration [M2+] and, by difference, of complexed metal [ML] were obtained. Total ligand concentrations were considered to be those of the residue (in g L -I). By an iterative, computer simplex method the authors obtained values for the conditional constant KML and the two acid dissociation constants. The graphs in Figure 10.11 show the experimental data and the corresponding calculated solution compositions. As can be seen, the model was a good approximation to the behavior of OMW as complexant. The values of the stability constants obtained by this model (Bejarano and Madrid, 1996d) and by a cation exchange resin (Bejarano and Madrid, 1996c) were reasonably congruent despite the different techniques and conditions, and agreed with that previously obtained for Cu2+ by Bejarano et al. (1994) using voltammetric techniques.
SUMMARY During the last decades, the view of land as a sink for any waste has been ruled out as erroneous, and concern for the long-term environmental hazards of accumulation of wastes has gained increasing importance. In the preceding pages we have tried to show that, even though soils can "fix" large quantities of potentially toxic metals, acting as a barrier against metal pollution of ground and surface waters, organic matter, either natural or, especially, added to soils, is a very important factor able to change the status of the metals in the system. Whether such change is in the direction of increasing the fixation of metals or of mobilizing them depends on several circumstances, but many studies suggest that the presence of soluble organic matter generally increases metal mobility through formation of soluble complexes. Therefore, those processes for maturing organic wastes
Metal Retention and Mobility as Influenced by Some Organic Residues
•
o
Ig L-1 OMW;
n v
2g L-1 OMW; •
-4.00 r----,,--"--,..,--'--,--,'-'
I
-4.0 r
-
-4.25 -
-4.5 -4.50 _
-
~ -4.75
ill ~ •
.
o 3gC 1 0MW
I-
,
I
,
I-
•o~
-
~
•
-
'Y
~.
'\}
o
pH
-
'\}
-6.0 -5.25 L-_...L_I_ _-,---l_...ll_ _-'_...J'LJ '5.0 5.5 6.0 6.5 7.0 7.5
-
'Y.
-
-5.00 r
, -
~
-
oi ~
219
o I
I
I
I
I
5.25
6.00
6.75
7.50
8.25
pH
Figure 10.11. Comparison of the metal concentrations found in solutions of OMW equilibrated with a river sediment and subsequently filtered through C-18 reverse-phase cartridges (hollow symbols) and values calculated by a simple model of mono-nuclear bidentate complexation (filled symbols) (From Bejarano and Madrid, 1996, with permission).
previous to their use as soil amendments must be aimed, among other purposes, at minimizing the proportion of soluble components. Thus, the mobilizing action will also be minimized and the resulting highly polymerized, sparingly soluble organic wastes will contribute to immobilize metals and consequently will help to keep a low bioavailability of such potentially toxic elements.
REFERENCES Alloway, B.J. Soil Processes and the Behaviour of Metals, in Heavy Metau in Soiu, B.J. Alloway, Ed., Blackie and Son, Glasgow, 1990, pp. 261-279. Amacher, M.C., J. Kotuby-Amacher, H.M. Selim, and 1.K. Iskandar. Retention and release of metals by soils-Evaluation of several models. Geooerma, 38, pp. 131-154, 1986. Arambarri, P., F. Cabrera, and C.G. Toea, La ContaminaciOn oet Rio Guaoiamar Y .Ill Zona oe InfLuencia, MarumaJ oeL GuaoaLqllirir y Coto 1)onana, por ReJiOuoJ Oe InollJtrlaJ MineraJ y AgrfcolM. CS I C, Madrid, 1984. Baker, D.E. Copper, in Heavy Metau in Soiu, B.J. Alloway, Ed., Blackie and Son, Glasgow, 1990, pp. 151-176. Barrado, E., M.H. Vela, R. Pardo, and F. Rey. Determination of acidity and metal ion complexing constants for eucalyptus and oak extracts. Commlln. Sotl Sci. Plant AnaL., 26, pp. 2067-2078,
1995. Barrow, N.J., J.W. Bowden, A.M. Posner, and J.P. Quirk. Describing the adsorption of copper, zinc and lead on a variable charge mineral surface. AwtraL. J. SoiL Red., 19, pp. 309-321, 1981.
220
Fate and Transport of Heavy Metals in the Vadose Zone
Basset, R.L. and D.C. Melchior, Chemical Modeling of Aqueous Systems: An Overview, in ChemicaL MOdeling ofAqu£oUJ SY.Jtenu II, D.C. Melchior and R.L. Basset, Eds., ACS Symposium Series No. 416, American Chemical Society, Washington, DC, 1990, pp. 1-14. Bejarano, M. and L. Madrid. Solubilization of heavy metals from a river sediment by a residue from olive oil industry. Environ. TechnoL., 13, pp. 979-985, 1992. Bejarano, M. and L. Madrid. Solubilization of heavy metals from a river sediment by an olive mill effluent at different pH values. Environ. Techno!., 17, pp. 427-432, 1996a. Bejarano, M. and L. Madrid. Release of heavy metals from a river sediment by a synthetic polymer and an agricultural residue: Variation with time of contact. ToxicoL. Environ. Chem., 55, pp. 95-102, 1996b. Bejarano, M. and L. Madrid. Complexation parameters of heavy metals by olive mill wastewater determined by a cation exchange resin. J. Environ. SCt: HeaLth, B31. pp. 1085-1101, 1996c. Bejarano, M. and L. Madrid. Solubilization of Zn, Cu, Mn and Fe from a river sediment by olive mill wastewater: Influence of Cu or Zn. Toa:ico!. Environ. Chem., 55, pp. 83-93, 1996d. Bejarano, M., A.M. Mota, M.L.S. Gonr;alves, and L. Madrid. Complexation ofPb(II) and Cu(II) with a residue from the olive-oil industry and a synthetic polymer by DPASV. Sci. TotaL Environ., 158, pp. 9-19, 1994. Boldin OficiaL deL &tado, Madrid, No. 262, p. 3234, 1990. Bowen, H.J.M. EnvironmentaL Chemutry of the ELement.J. Academic Press, London, 1979, p. 333. Buffle, J. Complexation Reaction,} in Aquatic SY.Jtenu:AnAnaLyticaLApproach. Ellis Horwood, Chichester, 1988, p. 33. Buffle, J., F.-L. Greter, and W. Haerdi. Measurement of complexation properties of humic and fulvic acids in natural waters with lead and copper ion-selective electrodes. AnaL. Chem., 49, pp. 216-222, 1977. Buffle, J., P. Deladoey, F.L. Greter, and W. Haerdi. Study of the complex formation of copper(II) by humic and fulvic substances. AnaLytica Chimica Acta, 116, pp. 255-274, 1980. Cabrera, F., M. Soldevilla, F. Osta, and P. Arambarri. Interacci6n de Cobre y Alpechines. Limnitica, 2, pp. 311-316, 1986. Carlson-EkvalL C.E.A. and G.M. Morrison. Toxicity of copper in the presence of organic substances in sewage sludge. Environ. Techno!., 16, pp. 243-251, 1995. Cavallaro, N., N. Padilla, and J. Villarrubia. Sewage sludge effects on chemical properties of acid soils. SoiL Sci., 156, pp. 63-70, 1993. Cheshire, M.V., D.B. McPhaiL and M.L. Berrow. Organic matter-copper complexes in soils treated with sewage sludge. Sci. TotaL Enriron., 152, pp. 63-72, 1994. Davis, A.P. and I. Singh. Washing of zinc(II) from contaminated soil column. J. Environ. Eng., 121, pp. 174-185, 1995. Davis, J.A. and J.O. Leckie. Surface ionization and complexation at the oxide/water interface. 2. Surface properties of amorphous iron oxyhydroxide and adsorption of metal ions. J. CoLloid Inteiface Sci., 67, pp. 90-107, 1978. Del Castilho, P., W.J. Chardon, and W. Salomons. Influence of cattle-manure slurry application on the solubility of cadmium, copper, and zinc in a manured acidic, loamy-sand soil. J. Environ. QuaL., 22, pp. 689-697, 1993. Department of the Environment, COde of Practice for AgricuLturaL U.Je of Sewage SLudge. HMSO Publications, London, 1992, p. 6. Domergue, F.-L. and J.-C. Vedy. Mobility of heavy metals in soil profiles. Int. J. Environ. AnaL. Chem., 46, pp. 13-23, 1992. Driessens, F.C.M. Ionic Solid Solutions in Contact with Aqueous Solutions, in Geochemical Proce.J.Je.J at MineraL Suiface.J, J.A. Davis and K.F. Hayes, Eds., ACS Symposium Series No. 323, American Chemical Society, Washington, DC, 1986, pp. 524-560.
Metal Retention and Mobility as Influenced hy Some Organic Residues
221
Emmerich, W.E., L.J. Lund, A.L. Page, and A.C. Chang. Solid phase forms of heavy metals in sewage sludge-treated soils. J. Environ. Qual., 11, pp. 178-181, 1982. Evans, L.J., G.A. Spiers, and G. Zhao. Chemical aspects of heavy metal solubility with reference to sewage sludge amended soils. Int. J. Environ. Anal. Chem., 59, pp. 291-302, 1995. Filella, M., J. Buffle, and H.P. van Leeuwen. Effect of physico-chemical heterogeneity of natural complexants. Part 1. Voltammetry oflabile metal-fulvic complexes. Analytica ChimicaActa, 232, pp. 209-223, 1990. Fitch, A. and F.J. Stevenson. Comparison of models for determining stability constants of metal complexes with humic substances. Soil Sci. Soc. Am. J., 48, pp. 1044-1050, 1984. Forstner, U. and G.T.W. Wittman. Metal Pollution in the Aquatic Environment, 2nd ed., SpringerVerlag, Berlin, 1983, p. 223. Gamble, D.S., A.W. Underdown, and C.H. Langford. Copper(II) titration of fulvic acid ligand sites with theoretical, potentiometric, and spectrophotometric analysis. Anal. Chem., 52, pp. 1901-1908, 1980. Gonzalez-Vila, F.J., T. Verdejo, and F. Martin. Characterization of wastes from olive and sugarbeet processing industries and effects of their application upon the organic fraction of agricultural soils. Int. J. Environ. Anal. Chem., 46, pp. 213-222, 1992. Greter, F.-L., J. Buffle, and W. Haerdi. Voltammetric study of humic and fulvic substances. Part I. Study of the factors influencing the measurement of their complexing properties with lead. J. Electroanal. Chem., 101, pp. 211-229, 1979. Hooda, P.S. and B.J. Alloway. Sorption of Cd and Pb by selected temperate and semi-arid soils: Effects of sludge application and aging of sludged soils. Water Air Soil Pollut., 74, pp. 235-250, 1994. Irving, H. and R.J.P. Williams. Order of stability of metal complexes. Nature, 162, pp. 746-747, 1948. Japenga, J., J.W. Dalenberg, D. Wiersma, S.D. Scheltens, D. Hesterberg, and W. Salomons. Effect of liquid animal manure application on the solubilization of heavy metals from soil. Int. J. Environ. Anal. Chem., 46, pp. 25-39, 1992. Kabata-Pendias, A. Agricultural Problems Related to Excessive Trace Metal Contents of Soils, in Heavy Metau, Problem! and Solutiow, W. Salomons, U. Forstner, and P. Mader, Eds., Springer-Verlag, Berlin, 1995, pp. 3-18. Keefer, R.F., S.M. Mushiri, and R.N. Singh. Metal-organic associations in two extracts from nine soils amended with three sewage sludges. Agric., &o.JY.Jt. Environ., 50, pp. 151-163, 1994. Klamberg, H., G. Matthess, and A. Pekdeger. Organo-Metal Complexes as Mobility-Determining Factors of Inorganic Toxic Elements in Porous Media, in Inorganic Contaminant.J in the Vado.Je Zone, B. Bar-Yosef, N.J. Barrow, and J. Goldshmid, Eds., Ecological Studies Series, Vol. 74, Springer-Verlag, Berlin, 1989, pp. 3-17. Lindsay, W.L. Chemical Equilibria in Soiu. John Wiley & Sons, New York. 1979, p. 222. Lund, U. and A. Fobian. Pollution of two soils by arsenic, chromium and copper. Geoderma, 49, pp. 83-103, 1991. Madrid, L. and E. Diaz-Barrientos. Influence of carbonate on the reaction of heavy metals in soils. J. Soil Sci., 43, pp. 709-721, 1992. Madrid, L. and E. Diaz-Barrientos. Retention of heavy metals by soils in the presence of a residue from the olive-oil industry. Eur. J. Soil Sci., 45, pp. 71-77, 1994. Madrid, L. and . Diaz-Barrientos. Nature of the action of a compost from olive mill wastewater on Cu sorpti n by soils. Toxicol. Environ. Chern., 54, pp. 93-98, 1996. Martin-Olme ,P., F. Cabrera, R. LOpez, and J.M. Murillo. Successive applications of a compo ed olive oil mill sludge: Effect on some selected soil characteristics. FreJeniw Environ. Bull., 4, pp. 221-226, 1995.
222
Fate and Transport of Heavy Metals in the Vadose Zone
Martinez-Nieto, L. and S.E. Garrido-Hoyos. El Alpechin, un Problema Medioambiental en Vias de Solucion (I). Quimica e I,wlMtria, 17-28, November 1994. McBride, M.B. Retention of Cu 2+, Ca2+, Mg2+, and Mn 2+by amorphous alumina. Soil Sci. Society Am. J., 42, pp. 27-31, 1978. McBride, M.B. Reactions controlling heavy metal solubility in soils. Aov. Soil Sci., 10, pp. 1-56, 1989. McBride, M.B. Toxic metal accumulation from agricultural use of sludge: Are USEPA regulations protective? J. Environ. QuaL., 24, pp. 5-18, 1994. McBride, M.B. and D.R Bouldin. Long-term reactions of copper(II) in a contaminated calcareous soil. SoiL Sci. Soc. Am. J., 48, pp. 56-59, 1984. McGrath, S.P. and J. Cegarra. Chemical extractability of heavy metals during and after longterm applications of sewage sludge to soil. J. Soil Sci., 43, pp. 313-321, 1992. Neal. RH. and G. Sposito. Effects of soluble organic matter and sewage sludge amendments on cadmium sorption by soils at low cadmium concentrations. SoiL Sci., 142, pp. 164-172, 1986. Papadopoulos, P. and D.L. Rowell. The reactions of copper and zinc with calcium carbonate surfaces. J. SoiL Sci., 40, pp. 39-48, 1989. Perdue, E.M. Measurements of Binding Site Concentrations in Humic Substances, in Meta! Speciation: Theory, AnaLY.JiI anO AppLication, J.R Kramer and H.E. Allen, Eds., Lewis Publishers, Boca Raton, FL, 1988, pp. 135-154. Perez, J.D. and F. Gallardo-Lara. Direct, delayed and residual effects of applied wastewater from olive processing on zinc and copper availability in the soil-plant system. J. Environ. Sci. HeaLth, B28, pp. 305-324, 1993. Pinheiro, J.P., A.M. Mota, and M.L. Simoes Gon~alves. Complexation study of humic acids with cadmium(II) and lead(II). AnaLytica Chimica Acta, 284, pp. 525-537, 1994. Saiz-Jimenez, C., J. W. De Leeuw, and G. Gomez-Alarcon. Sludge from the waste water of the olive processing industry: A potential soil fertilizer? Set: TotaL Environ., 62, pp. 445-452, 1987. Salomons, W. and U. Forstner. Metal.! in the Hyorocycle. Springer-Verlag, Berlin, 1984, pp. 176179. Saviozzi, A., R. Levi-Minzi, and R Riffaldi. How organic matter sources affect cadmium movement in soil. Bwcycle, pp. 29-31, May/June 1983. Schubert, J. The use of ion exchangers for the determination of physical chemical properties of substances particularly radio tracers in solution. J. Phy.J. Colloid Chem., 52, pp. 340-356, 1948. Selim, H.M. Modeling the transport and retention of inorganics in soils. Aov. Agron., 47, pp. 331384, 1992. Selim, H.M., M.C. Amacher, and 1.K. Iskandar. MooeLing the Tran.Jport 0/Heavy Metal.! in Soil.!. U.S. Army Corps of Engineers, CRREL Monograph 90-2, U.S. Government Printing Office, Washington~, 1990. Senesi, N. Metal-H~ic Substance Complexes in the Environment. Molecular and Mechanistic Aspects by MUltiPl~spectroscoPiC Approach, in Bwgeochemiltry 0/ Trace Metal.!, D. C. Adriano, Ed., Springer-Verla, New York, 1992, pp. 429-496. Singh, B.R and E. St .nnes. Soil and Water Contamination by Heavy Metals, in Soil Proce.J.Je.J ano Water QuaLity, R Lal and B.A. Stewart, Eds., Advances in Soil Science Series, Lewis Publishers, Boca Raton, FL, 1994, pp. 233-271. Sposito, G., L.J. Lund, and A.C. Chang. Trace metal chemistry in arid-zone field soils amended with sewage sludge: 1. Fractionation of Ni, Cu, Zn, Cd, and Pb in solid phases. SoiL Sci. Soc. Am. J., 46, pp. 260-264, 1982. Sterrit, R.M. and J.N. Lester. Comparison of methods for the determination of conditional stability constants of heavy metal-fulvic acid complexes. Water R£.J., 18, pp. 1149-1153, 1984. Stevenson, F.J. and Y. Chen. Stability constants of copper(II)-humate complexes determined by modified potentiometric titration. SoiL Sci. Soc. Am. J., 55, pp. 1586-1591, 1991.
Metal Retention and Mobility as Influenced by Some Organic Residues
223
Stevenson, F,J., A. Fitch, and M.S. Brar. Stability constants of Cu(II)-humate complexes: Comparison of select models. SoiL Sci., 155, pp. 77-91, 1993. Tiller, K.G. Heavy metals in soils and their environmental significance. Adv. SoiL Sci., 9, pp. 113142, 1989. Ure, A.M., Ph. Quevauviller, H. Muntau, and B. Griepink. Speciation of heavy metals in soils and sediments. An account of the improvement and harmonization of extraction techniques undertaken under the auspices of the BCR of the Commission of the European Communities. Int. J. Environ. AnaL. Chern., 51 (Special Issue), pp. 135-151, 1993. WenzeL W.W., M.A. Pollak, and W.E.H. Blum. Dynamics of heavy metals in soils of a reed bed system. Int. J. Environ. AnaL. Chern., 46, pp. 41-52, 1992. Wilson, S.A., T.C. Huth, R.E. Arndt, and R.K. Skogerboe. Voltammetric methods for determination of metal binding by fulvic acid. AnaL. Chern., 52, pp. 1515-1518, 1980. Xue, H.B., D. Kistler, and L. Sigg. Competition of copper and zinc for strong ligands in a eutrophic lake. LirnnoL. Oceanogr., 40, pp. 1142-1152, 1995. Zunino, H., G. Galindo, P. Peirano, and M. Aguilera. Use of the resin exchange method for the determination of stability constants of metal-soil organic matter complexes. SoiL Sci., 114, pp. 229-233, 1972a. Zunino, H., P. Peirano, M. Aguilera, and 1. Escobar. Determination of maximum complexing ability of water-soluble complexants. Soil Sci., 114, pp. 414-416, 1972b.
CHAP"I'ER 11
The Rhizosphere and Trace Element Acquisition in Soils George R. Gobran, Stephen Clegg, and Francois Courchesne
INTRODUCTION Soil chemical analyses indicate the potential nutrient availability under favorable nutrient conditions for root growth and activity (Marschner, 1986). Conventional soil tests give no information about root induced changes in the rhizosphere due to exudation. Such processes may ameliorate toxic ions such as Al (Gobran and Clegg, 1996; Inskeep and Comfort, 1986). Therefore, the use of conventional soil and soil solution tests in both forestry and agriculture can be unsatisfactory for the prediction of plant responses to fertilization, acidification, or to other external stresses (Mahendrappa et al., 1986; Marschner, 1986; Hogberg and Jensen, 1994). Beier and Cummins (1993) indicated "difficult-to-study" areas in ecosystem manipulation experiments that require more attention in order to understand relationships between biological and chemical parameters. These areas were roots, soil, in-plot variability, soil microbiology, and ecophysiology. Studies of the soil, rhizosphere, and roots in manipulation experiments can directly address these areas, but have hitherto been poorly investigated in forests (Vogt et al., 1993). The scope of this chapter is to demonstrate the importance of the rhizosphere in the soil-plant system and to emphasize its role on the cycling of both macro and trace elements. T iterature will first be reviewed to characterize the rhizosphere and to contrast th~ prope . es of the bulk and rhizosphere soil environments. A conceptual model for nutrient availa. ility in the mineral soil-root system will subsequently be described. The model focuse~ on dynamic feedback processes between plant roots and the surrounding soil materials to illustrate the soil response to environmental stresses. The results from a series of case studies encompassing a range of field observations from hydrologically and chemically manipulated forest soils also will be presented. Finally, the implications of the available data on the rhizosphere dynamics and of the concep~25
226
Fate and Transport of Heavy Metals in the Vadose Zone
tual model on our understanding of trace element cycling in the soil-plant system will be assessed.
History One of the most notable milestones in rhizosphere research was the isolation and identification of the nitrogen fixing bacteria (genus Rhizobium) in leguminous plants (Beijerink, 1888; Hellriegel and Wilfarth, 1888). These studies attracted the attention of L. Hiltner, who later coined the term, rhizosphere (Hiltner, 1904) due to his opinion that this zone was unique for soil organisms. Other substantial contributions to rhizosphere research stem from the work of Bowen (1961), Roriva and Bowen (1966), and Roriva (1969), who increased understanding of the role and nature of root exudates on the rhizosphere effect. We now see that the rhizosphere has become an area of intense interest to soil scientists, ecologists, and agronomists.
izosphere-Definitions The term stems from two Greek words rhizo for root and dphere (sphaira), the environment in which one acts or exists. Although the term rhizosphere seems self-explanatory, many conceptual and operational definitions exist. Curl and Truelove (1986) described the rhizosphere as "that narrow zone of soil subject to the influence of living roots, as manifested by the leakage or exudation of substances that affect microbial activity." Alternatively, Lynch (1990) stated that "the total rhizosphere environment is determined by an interacting trinity of the soil, the plant and the organisms associated with the roots." According to the Glossary of Soil Science Terms (SSSA, 1997), the rhizosphere is "the zone of soil immediately adjacent to plant roots in which the kinds, numbers, or activities of microorganisms differ from that of the bulk soil." The term rhizosphere has also been macroscopically referred to as that portion of the soil profile where most roots are located (Hedley et al., 1982; Mengel et aI., 1990). Ulrich (1987) stated that "Morphologically, roots and soil, or micro-organisms and soil, usually could be clearly separated," but "From a functional point of view this clear boundary does not exist." However, this has not stopped the formulation of a number of conceptual and operational subdivisions of the rhizosphere based on distance from the root or on the inclusion of root materials. This is primarily due to the recognition that the root surface is a critical site for soil-plant-microbe interactions. This surface has been called the rhizoplane (Clark, 1949; Richards, 1987; Paul and Clark, 1989), or the soil-root interface (Gobran and Clegg, 1996), while the term rhizosphere is more commonly used to describe a region dominated by soil material. The root-free material surrounding the rhizosphere soil is termed the bulk soil although the extension of the rhizosphere and its boundary with the bulk soil are difficult to define precisely. Also, the spatial distribution of the rhizosphere soil changes with time as roots die or colonize new areas of the soil profile.
Methods of Rhizospheric Study A wide range of methods was used to study the rhizosphere in the laboratory and in the field. Plant growth experiments in the greenhouse have been extensively employed (Spyridakis et al., 1967; Berthelin and Leyval, 1982; Kirlew and Bouldin, 1987; Mengel
The Rhizosphere and Trace Element Acquisition in
et aI., 1990; Chung and Zasoski, 1994). Some researchers used homogenized bulk soil as the growth medium, while others added a reference mineral like biotite, muscovite, or phlogopite to the soil materials to study the rhizosphere effect on mineral weathering (Boyle and Vogt, 1973; Mojallali and Weed, 1978). The more elaborate methods demand reconstruction of the soil profile with the subsequent changes to biological and chemical conditions. Examples include porous plastic envelopes where roots have no physical contact with the soil but where nutrients and water pass through the pores (Brown and UI-Haq, 1984). Soil columns were used where homogenized soil materials were packed in a container and divided by stainless steel screens to create a series of soilroot zones (Helal and Sauerbeck, 1983; Dormaar, 1988; Mengel et aI., 1990). High root densities were achieved in the central zone (under the growing plant), while adjacent zones were root-free but subjected to the influence of the rhizosphere. Hinsinger et al. (1992) adapted the Kuchenbuch and Jungk (1982) method to produce a two-dimensional root mat on a polyamide net to simulate a macroscopic root surface. The mat was then laid down on an agar-mica substrate. This method has enabled detailed examination of the rhizospheric substrate following sectioning of the agar-mica with a microtome (Hinsinger and Jaillard, 1993; Hinsinger et al., 1993). Comparison of the mineralogy of bulk and rhizosphere soil materials was also achieved using field samples collected from agricultural (Kodama et aI., 1994) and forested sites (April and Keller, 1990; Gobran and Clegg, 1996; Courchesne and Gobran, 1997). The separation of the rhizosphere and bulk fractions involved drying, gentle shaking in a plastic container (Hendriks and Junk, 1981; Haussling and Marschner, 1989; Kirlew and Bouldin, 1987), and brushing of roots to free adhering rhizosphere soil (Haussling and Marschner, 1989; Clemensson-Lindell and Persson, 1992). The soil particles directly contacting the root surfaces were consider as rhizoplane soil (April and Keller, 1990). Another useful method that has often been overlooked is the rhizocylinder method (Riley and Barber, 1969, 1971; Hoffmann and Barber, 1971) in which plant root plus adjacent soil particles were examined as a separate fraction (the rhizocylinder) and compared with rhizosphere and bulk soil samples (Gobran and Clegg 1996). Kosola (1996) used pressurized-wall minirhizotron tubes equipped with a borescope and a laparoscopic sampler for collecting roots of known age in the field. This new technique could help refine the measurement of the effect of specific living roots on the surrounding soil. The detailed identification and quantification of chemical, physical, and mineralogical changes occurring in the rhizosphere not only relied on macroscopic approaches (chemical extraction, titration, ion exchange, X-ray diffraction) but also involved the observation of thin sections using microscopic techniques such as transmission electron microscopy (TEM), scanning electron microscopy (SEM) , SEM with energy dispersive spectrometry (EDS), and SEM in the backscattered electron mode (BESI) (April and Keller, 1990; Kodama et aI., 1994; Bruand et aI., 1996). Conkling and Blanchar (1989) and Conkling et al. (1991) constructed glass microelectrodes with sensing tips 0.02 mm in diameter and 0.05 mm long for in situ pH measurements. The technique was successfully used in conjunction with minirhizotrons to measure the rhizosphere pH 'of alfalfa, corn, and soybean. Future developments in the area of microelectrode technology will facilitate the direct and continuous measurement of chemical changes in the rhizospheric environment.
228
Fate and Transport of Heavy Metals in tile Vadose Zone
RHIZODEPOSITION Root Distribution and Longevity Despite a long history of study, our understanding of root distribution and processes is poor (Jackson et al., 1996; Vogt et al., 1993). Yet, together with litterfall, root production provides the greatest input of organic carbon to many soils, which can store carbon at twice the rate of the above-ground biomass (Waring and Schlesinger, 1985). Fine roots with their high production of exudates and variable life span contribute the major portion of the total carbon input (Grayston et al., 1996). Root distribution patterns vary greatly with depth across the major biomes of the world. For example, Jackson et al. (1996) found that tundra, boreal forest, and temperate grassland have the shallowest rooting profiles with 80-90% of roots in the top 30 cm of the soil, whereas deserts and temperate coniferous forests had deep profiles with -50% of roots in the top 30 cm of the soil. Moreover, many plant species may have a wide root distribution which can cause difficulties in identifying which zones roots acquire most of their nutrients. For example, it was shown that spring wheat during the later periods of the growing season took 3040% of total P from the subsoil despite a higher P content in the topsoil and a reasonably uniform rainfall during the growing season (Fleige et al., 1981). Fox and Lipps (1961) found that 3% of the total root mass in alfalfa took up 60% of total nutrients from the subsoil during periods of drought. On a global scale, the average root distribution has been estimated as being 30%, 50%, and 74% in the top 10 cm, 20 cm, and 40 cm, respectively (Jackson et al., 1996). Yet in many forest ecosystem studies there has been a tendency to emphasize the role of the upper organic horizons as a major source of nutrients. Despite this, 40 to 50% of the fine root biomass is found in the first 30 cm of mineral soil at many forest sites (Clemensson-Lindell and Persson, 1992; Haussling and Marschner, 1989; Persson et al., 1995; Wood et al., 1984). This root distribution is possibly related to the large pool ofN, P, and S retained in the mineral horizons, especially in spodosols (Stevenson, 1991). Clearly, more investigation is needed on the role of roots in the mineral soil and the processes by which they acquire elements and alter their availability. Minirhizotron studies of Norway spruce have shown that roots can live for nine months or more (Majdi, 1994) and frequently reoccupy old root channels (H. Persson, pers. comm.). Fahey and Hughes (1994) estimated the median fine root (d-mm) longevity (50% survivorship) in the forest floor under maple and beech at Hubbard Brook, New Hampshire, to be about 6 months. In a maple-beech forest of south-central New York, 59% of the spring root cohort were still alive after 5 months. Moreover, Hendrick and Pregitzer (1993) working in northern hardwood forests indicated that roots born in spring live longer than the average fine root. Persson (1983) suggested that roots growing deeper in the soil profile live longer than roots in surface horizons. The seasonal decline in total root biomass is less pronounced in mineral than in organic horizons although rapid root disappearance can occur in both (Hendrick and Pregitzer, 1992). This indicates that the rhizosphere may not always be an ephemeral environment in the soil. The apparent longevity of roots in forests would allow more time for the establishment of rhizospheric processes than is normally considered in experimental and agricultural systems. The reestablishment of rhizospheric conditions in unoccupied root channels may also be enhanced due to the priming effects of dead roots.
The Rhizosphere and Trace Element Acquisition in Soils
229
Such effects may include physicochemical properties such as high porosity and readily decomposed organic matter as well as biological priming with the propagules of rhizospheric organisms.
Belowground Carbon Flux Although the rhizosphere constitutes only 1-3% of the soil volume (Coleman et al., 1978; Gobran and Clegg, 1996), a large proportion of plant-assimilated carbon is released in the belowground system as dead roots, exudates, and as substrates for mycorrhizae. It is also interesting and important to compare this belowground flux to that aboveground. Studies indicate that loss of fine roots and mycorrhizae returns two to five times more organic matter to the soil than the aerial biomass (Fogel and Hunt, 1983; Waring and Schlesinger, 1985). The magnitude of C fluxes to and from soils vary with latitude. For example, C cycling in litterfall increases by two- to threefold from 65 to 45°N (Van Cleve and Powers, 1995). The annual C input derived from the activity of root and mycorrhizae and from root decay (estimated as the difference between CO 2 fluxes from the soil surface -litterfall) also increases with decreasing latitude but more rapidly than litterfall fluxes. At Hubbard Brook the aboveground and belowground C pools were 9900 and 2300 kmol C ha-I y -I, respectively (Johnson et al., 1995). As for the C fluxes, 124 kmol C ha- I y-I was deposited to the soil as litterfall, while 60 kmol C ha- I y-I came from root decomposition. Trettin et al. (1995) compiled C flux values from a range of northern forested wetlands. The aboveground biomass ranged from 15 to 55 ton C ha- I y-I, while belowground C pools totaled 50 to 1300 ton C ha- I y-I, of which 7 to 21 ton C ha- I y-I came from root biomass. Above- and belowground C fluxes were estimated at 0.4 to 1.6 and 0.1 to 0.4 ton C ha- I y-I. Rhizodeposition varies due to other factors such as climate, the presence of symbiotic organisms and plant age and type. For instance, annual plants release less carbon than perennials (Grayston et al., 1996).
Exudates in the Rhizosphere Apart from the cellulose, lignin, and other compounds released to the soil by dead roots, there are a wide variety of compounds released by live roots which are collectively called exudates (Tables 11.1 and 11.2). Exudates and their diversity have been the subject of many reviews (e.g., Marschner, 1986; Uren and Reisenauer, 1988; Bowen and Roriva, 1991; Grayston et al., 1996). Organic acids are believed to be quantitatively the most important component of plant exudates (Table 11.3). They can serve as carbon nutrient sources for the microbial population and playa major role in weathering and complexation of micro- and macronutrients. Organic acids of low molecular weight are ubiquitous in soils, yet the type and quantity vary not only in the bulk and rhizosphere soil (Fox and Comerford, 1990; Grierson, 1992; Szmigielska et al., 1996), but also within the biosphere compartments such as plant canopy, forest litter, surface horizons, soil solutions, rhizosphere, rock surfaces, etc. (Drever and Vance, 1994).
Acid-Base Changes in the Rhizosphere This has major direct and indirect consequences for the availability of nutrients and toxic elements and their uptake by roots. Five important factors affecting acid-base condi-
230
Fate and Transport of Heavy Metals in tile Vadose ZOlle
Table 11.1. Root Products Released in the Rhizosphere a Product
Release
Compound
Excretions
Leakage from or between epidermal cells Active excretion
Secretions
Active secretion
Sugars, inorganic acids, amino acids, water inorganic ions, oxygen, etc. Carbon dioxide, bicarbonate, protons, electrons, ethylene, etc. Mucilage, protons, electrons, enzymes, siderophores, alleopathic compounds, etc. Root-cap cells, cell contents, etc.
Root exudates Diffusates
Root debris
a
Cell or root death, sloughing
After Uren and Reisenauer, 1988.
Table 11.2. Examples of Root Products Exuded to the Rhizosphere a Some Organic Compounds Exuded by Roots Carbohydrates Amino acids Aliphatic acids Aromatic acids Fatty acids Sterols Enzymes Miscellaneous a
Arabinose, fructose, glucose, maltose, ribose, sucrose All 20 amino acids Acetic, citric, fumaric, malic, oxalic, tartariC, valeric p-Hydroxybenzoic, p-coumaric, gallic, salycyclic Lineolic, palmitic, stearic Campesterol, cholesterol, sitosterol Amylase, deoxyribnuclease, peroxidase, phosphatase Plant and microbial growth regulators, stimulators and inhibitors
After Uren and Reisenauer, 1988.
Table 11.3. Examples of Organic Acids and Other Complexing Compounds Found in the Rhizospherea Compound
Occurrence
Citric, tartaric, lactic, and malic acids Oxalic acid
Produced by roots and bacteria in the rhizosphere. Present in litter extracts and canopy throughfall. Produced by fungi, including mycorrhizae. Abundant in acid soils. Produced by rhizosphere and ecomycorrhyzal fungi. Produced under conditions of Fe stress. Formed through the decay of Iignins. Involved in the mobilization and transport of Fe in acid soils. Synthesized by bacteria on rock surfaces and in the rhizosphere. Abundant in habitats rich in decaying organic matter.
Siderophores Phenolic acids 2-Ketogluconic acid a
After Stevenson, 1991.
tions in the rhizosphere will be discussed here: (1) the production of CO 2 from respiration, (2) the excretion of organic acids, (3) microbial production of acids following assimilation of released root carbon, (4) ion uptake, and (5) plant genotype.
The Rhizosphere and Trace Element Acquisition in
231
The reviews ofNye (1986) and Marschner and Romheld (1996) do not emphasize the role of the first three factors. For instance, the differences in air pressure between the root surface and the bulk soil in normally aerated soils will rapidly diffuse away CO 2 with a negligible effect on pH (Nye, 1986). It has been suggested that low molecular weight acids released to the rhizosphere by roots and microorganisms may have an acidifYing effect (Mench and Martin, 1991; Petersen and Bottger, 1991). However, the quantities required to have a significant effect in the rhizosphere have not been found, either because they are produced in too small amounts by roots or because they are rapidly metabolized by microorganisms (Nye, 1986). Also, the quantity of acidity produced by microbes which use carbon deposited in the rhizosphere is probably small, due to the vast quantity of carbon that would be required. Calculations suggest that one-third of the total carbon released by roots would have to be converted to acids by microbes to produce a significant change in pH (Nye, 1986). Finally, Hedley et al. (1982) found that neither the total number of microbial colonies nor the number of acid-producing colonies were related to rhizosphere pH. The consensus is that the dominant mechanism responsible for pH changes in the rhizosphere is the net release of H+, HC0 3-, or OH- in response to the imbalance between cation and anion uptake by roots (Tinker, 1990). The most evident factors affecting acid-base conditions of the rhizosphere are ion uptake and plant genotype. For example, when NH4 + rather than N0 3- ions were supplied to soils (Riley and Barber, 1971; Soon and Miller, 1977; Rollwagen and Zasoski, 1988), a drop in pH of two units was reported in the soil close to root surface. In acid soils, such as spodosols, where the rate of nitrification is very low, the form of nitrogen taken up will mostly be NH/. In this situation cation uptake, will exceed anion uptake resulting in net H+ excretion and in a pH decrease in the rhizosphere relative to the bulk soil. Additionally, plants associated with N-fixing organisms (e.g., legumes, ALnlM, and CaJuarina) also acidify the rhizosphere since the uncharged dinitrogen molecule crosses the soil-nodule or soil-root, resulting in a higher uptake of cations than anions. In soils where N0 3- is the primary N-form, the amounts of anions taken up by plants tends to exceed cations, thus plants are required to release HC0 3- or OH- to maintain electrical neutrality across the soil-root interface. This causes increases in rhizosphere pH compared to the bulk soil. This phenomenon is so well-established that a method to manipulate rhizosphere pH has been elaborated by using different N sources (Riley and Barber, 1971; Sarkar and Wyn Jones, 1982). Given the same soil and form of nitrogen supply, large differences can arise between differing species or cultivars to acidify their rhizosphere (Marschner and Romheld, 1996). For example, it has been shown that lime-induced chlorosis (iron deficiency) in different cultivars of soybean, maize, and peanut supplied with N0 3- were related to rhizosphere pH between genotypes. Two cultivars of soybean supplied with N0 3- in a soil with a pH of 6.0 showed basal and apical rhizosphere pH values of 6.8 and 5.8 in the iron inefficient cultivar and 5.6 and 5.3 in the iron efficient cultivar. However, no difference was observed between three corn hybrids suggesting that they do not differ with respect to the mechanism controlling rhizosphere pH (Kirlew and Bouldin, 1987). Rollwagen and Zasoski (1988) also observed significant differences in rhizosphere pH values between Douglas fir, Sitka spru<;e, and western hemlock seedlings planted in the same soil. The ecological relevance of these differences can be assumed to apply to other plants, genotypes, and nutrients such as Zn and Mn (Macar and Graham, 1987) but it has been little investigated.
232
Fate and Transport of Heavy Metals in the Vadose Zone
Changing the pH of the rhizosphere has direct implications for forest nutrition and crop management. For instance, a decreasing pH will favor the dissolution of metals and their uptake by plants. Sarkar and Wyn Jones (1982) showed that the Fe, Zn, and Mn content of the shoot and roots of dwarf French beans was inversely proportional to rhizosphere pH. The dissolution of Al solids and KCI-extractable Allevels increased as excess H + ions were released by the roots of corn seedlings (Kirlew and Bouldin, 1987). Rollwagen and Zasoski (1988) reported a dramatic increase in foliar Mn content of conifers as their rhizosphere pH was decreased by ammonium addition to the soil. Cation exchange equilibria will also be affected by the pH gradient existing between the root surface and the bulk soil. Chung and Zasoski (1994) demonstrated that the selectivity for NH4 over Ca increased as pH decreased but no effect was observed for NH4-K exchange.
RHIZOSPHERIC FEEDBACK lOOPS
Regulating Processes In spite of the small volume that the rhizosphere occupies in the mineral soil (Gobran and Clegg, 1996), it plays a central role in the maintenance of the soil-plant system (Clegg and Gobran, 1997; Clegg et aI., 1997) and influencing the biogeochemistry of forest ecosystems (Gobran et aI., 1998). These root effects on soils suggest to some investigators that soil can be considered, in part, as a product of plants and soil biota (Van Breemen, 1993). It has been suggested that interactions between roots, microbial communities, and the soil under forest conditions are characterized by feedback loops driven by photosynthate released by roots (Perry et aI., 1989; DeAngelis et aI., 1986; Hobbie, 1992). Under such circumstances, the rhizospheric community "continually pull themselves up by their own bootstraps" (Perry et al., 1989) so that nutrient cycling and availability in the rhizosphere is higher than in the bulk soil, thus buffering the ecosystem against disturbances. An indication of this mechanism is reflected by the ratio of organisms in the rhizosphere soil to counts in bulk soil (R/S), which were found to range from 10 to 50 under varying plant species at different stages of development and in different soils and climates (Paul and Clark, 1989). Moreover, the R/S is typically 10 to 50 for bacteria and 5 to 10 for fungi (Richards, 1987). Certain tree species actively control nutrient availability by complex feedback loops (Perry et aI., 1989) between trees, microbial communities, and the soil to maintain a competitive advantage. This could be achieved, for example, by decreasing the availability of essential nutrients such as P in the bulk soil by acidification (Cole, 1995; Van Breemen, 1993) or by decreasing mineralization rates. This may be achieved by production of secondary metabolites which are leached from foliage and litter (Van Breemen, 1995) while simultaneously favoring their own root systems by the establishment of nutrient-rich rhizospheric conditions and mycorrhizal association. It is apparent that accumulation of nutrients in the rhizosphere is a natural result of these many feedback processes occurring under forest conditions. Due to these regulating processes, the rhizospheric nutrient supply and demand are more carefully balanced in nutrient-poor forests than under agricultural and short-term experimental conditions. It is generally believed that in infertile soils (e.g., spodosols), the roots and microorganisms render the soilless favorable for plant growth (Chapin, 1980; 1993). Although this observation holds when the bulk soil is considered, it is not
The Rhizosphere and Trace Element Acquisition in Soils
233
supported by examination of the rhizospheric chemical properties (Gob ran and Clegg, 1996). At this scale of investigation, it seems that roots and microorganisms do increase the capacity of soils to support plant in nutrient-poor environments. Since most mineral nutrients must pass through the rhizosphere before assimilation by plants, we consider that this small zone has great potential as a regulator of plant nutrient availability and flux. Yet, in order to have any significant effect on nutrient cycling in forest ecosystems the rhizosphere must exhibit two features: persistence in time and resilience against perturbation, which are known properties of ecosystem stability (Richards, 1987). We believe the rhizosphere has such features and that they are reflected in the consistent difference observed between rhizosphere and bulk chemistry and mineralogy which remain after a wide range of field treatments (e.g., drought, irrigation, and ammonium sulfate) (Clegg and Gobran, 1997; Clegg et al., 1997). In addition to long-term rhizospheric effects, trees adapted to nutrient-poor sites have low nutrient absorption rates as well as efficient internal cycling (Van Breemen, 1995). This strategy of trees under such conditions contrasts with rapidly growing agricultural and ruderal species growing on intensively managed and fertile soils where mass flow may equal that of diffusion in supplying certain mobile nutrients (Binkley, 1986). Moreover, a relatively large and long-lived root biomass may benefit from the accumulation of organic matter observed in the rhizosphere (Gobran and Clegg, 1996; Chung and Zasoski, 1994) which could act as both a source and sink of available nutrients and potentially toxic ions.
Element Supply and Mobility in the Rhizosphere Three major mechanisms that may be connected with the supply and mobility of elements from immobile phases in the rhizosphere are (1) disturbance of equilibria between the solid phases of the soil and the soil solution; for example, under acid conditions there is a relative abundance of Fe, Mn, Zn, Bo, Mo, and Cu, which could be present in sufficiently high concentrations in solution to be toxic to common plants (Brady, 1990); (2) dissolution of sparingly soluble minerals containing Fe and P by the release of hydrogen ions (Kochian, 1991); and (3) the proliferation of organic acids in the rhizosphere could play two roles by either raising or lowering ion mobility. They can, for example, be synthesized and release micronutrients from the rhizosphere soil or act as chelating agents, thus reducing the mobility and activity of ions such as Al, Fe, Mn, Cu, and Zn. The type of biochemical chelator varies with their mobility and life span in the soil. Approximate concentrations of organic compounds and acids in soil solutions are found in Table liA. The literature shows that organic acids released to the rhizosphere by roots and mycorrhizae can play an important part in trace element nutrition to plants (trees/forests). Thermodynamic calculations from experiments with hypothetical rhizosphere solutions from PinlM radiata and Hordeum vulgare containing amino and low molecular weight acids (e.g., oxalic and acetic acids) showed that exudates make a significant contribution to the total soluble Fe (III), Cu (II), and Zn (II) (Inskeep and Comfort, 1986). Field studies also give a similar picture. In a study with and without the presence of ectomycorrhizal mats in Douglas fir (PJeudotJuga menziuil), it was shown that P, S, H, Al, Fe, Cu, Mn, and Zn were in significantly higher concentrations in the soil solution in the presence than in the absence of ectomycorrhizal mats. It was concluded that dissolved organic carbon and particularly oxalate released to the rhizosphere could provide a local weathering environment which raises the availability of P, S, and trace elements.
234
Fate and Transport of Heavy Metals in the Vadose Zone Table 11.4. Approximate Concentrations of Some Organic Acids and Compounds in Soil Solutionsa Compound
Concentration
Simple organic acids Amino acids Phenolic acids Siderophores
1 X 10-3 8 X 10-5 5 X 10-5 1 X 10-8
a
to to to to
4 X 10-3 6 X 10-4 3 X 10-4 1 X 10-7
M M M M
After Stevenson, 1991.
Since the oxidized states of Fe, Mn, and Cu are generally less soluble than the reduced states (Brady, 1990), microorganisms in the rhizosphere can alter the availability or toxicity to plants of Fe, Mn, and Cu by changing their oxidation reduction states and by mineralization of soil organic matter. Recent work by Bartlett (1996) stated that the rhizosphere is a poised redox system that can be either a reducing or oxidizing milieu for Mn. This is important since Mn oxides affect the oxidation mobility and toxicity of trace elements such as Cr (II), Pu (III), and Co (II). Moreover, dissolved Mn (II) in the presence of Fe influences Fe crystallization, hence affecting genesis of Fe oxides. Manganese oxides also promote decomposition of organics, thus affecting C and N transformations (Huang, 1991). Therefore, these processes in the rhizosphere deserve further attention for the study of soils and environmental quality. There is a concern about high weathering and mineralization releasing toxic elements to the rhizosphere/root. However, it appears that organic acids produced in the rhizosphere protect plants from the toxic effects of AI and other metals. This is due to effective complexation of the free metal species by organic matter, even though total metal concentrations in the rhizosphere may be higher than in the bulk soil (Gobran and Clegg, 1996). Additionally, high mineralization rates in the rhizosphere may have a positive effect on degrading toxic organic (Boyle and Shann, 1995; Pidgeon et aI., 1996; Reilley et aI., 1996). Indeed, the use of rhizosphere systems in site bioremediation is attractive and a great attention should be focused on the development of plant root systems for better understanding site remediation technology (Skladany and Metting, 1993).
Microbial Activity and Element Accumulation in the Rhizosphere Mass flow and diffusion are assumed to be the most important ion transport processes in supplying roots with nutrients (Nye and Tinker, 1977; Barber, 1984), although the magnitude of these pathways can vary greatly from one system to another as indicated in Binkley (1986). Moreover, the magnitudes of these pathways are difficult to quantifY and there may be considerable nutrient and water transport by biological agents such as mycorrhizae (Richards, 1987; Finlay and S6derst6m, 1989). The roots of most soil-grown plants are usually mycorrhizal (Grayston et al., 1996) and the role of mycorrhizal symbiosis in nutrient uptake has been well documented. There are several groups of mycorrhizae; those most commonly associated with trees and shrubs include vesicular-arbuscular mycorrhizae (VAM), ectomycorrhizae (ECM), and ericoid mycorrhizae (EM). For example, Marschner and Dell (1994) implicated the three mycorrhizal groups with the ability to take up and deliver nutrients to plant roots
The Rhizosphere and Trace Element Acquisition in
235
(e.g., P, NH4, N0 3, K, Ca, S04' Cu, Zn, and Fe). Depending on tree species, a variable proportion of the root supply of mineral nutrients pass through the fungal hyphae. For example, in ECM of plants such as Norway spruce, more than 90% of the root apical zones are enclosed by a fungal sheath. The interactions between soil acidification, plant growth, and nutrient uptake in ectomycorrhizal associations of forest trees has recently been reviewed by Finlay (1995). In this review it was proposed that the ECM mediated the effects of acidification on forest trees through modification of the soil chemical environment, altering patterns of plant nutrient uptake and increasing tolerance to, or detoxifying, the increased levels of heavy metals or aluminum often associated with soil acidification. Moreover, ECM greatly extend the life span of absorptive roots beyond the few days that root hairs are active; ECM persist for 6-9 months and even up to 13 years (Fogel, 1983). Mycorrhizae can also chemically modify the rhizosphere environment by direct and indirect means to facilitate P uptake. Ectomycorrhizal fungi release large quantities of oxalic acid which can mobilize P from sparingly soluble Ca phosphates in calcareous soils and AI and Fe phosphates in acid soils (Marschner and Dell, 1994). It is also possible that mycorrhizae in the rhizosphere have a large quantity of exchange sites with a higher affinity for phosphate and other ions (Richards, 1987; Bolan, 1991; Jakobsen et aI., 1992). These exchange sites may act both as source and sink for nutrients, thus easing seasonal variations in nutrient supply (Grayston et aI., 1996), and buffer aboveground parts from the effects of toxic metals, such as AI, Cu, and Zn (Wilkins, 1991; Marschner and Dell, 1994). Mycorrhizae also have several biochemical mechanisms to access and supply nutrients through the production of enzymes. For example, the excretion of phosphatases may play an important part in mineralizing the large organic P pool in the rhizosphere (Clegg and Gobran, 1997; Marschner and Dell, 1994; Haussling and Marschner, 1989).
CASE STUDIES The case studies in this chapter are based on the most important results presented in the following papers; Gobran and Clegg (1996), Clegg and Gobran (1997), Clegg et aI. (1997), and Courchesne and Gobran (1997). Our studies were based on the view that data from routine soil chemical analyses do not always correlate well to tree health, nutrient uptake, and leaf chemical composition (Binkley, 1986; Mahendrappa et aI., 1986). This may be due to routine soil sampling missing the rhizosphere fraction. Since the rhizosphere soil is altered by many interdependent processes between roots and microorganisms, rhizospheric processes have a large potential to change soil fertility and nutrient cycling, particularly in undisturbed forest soils. Additionally, much of the current knowledge about the rhizosphere is based upon research using agricultural species (often seedlings) grown under controlled conditions on homogeneous nutrient-rich substrates (Chapin, 1980; Parmelee et al., 1993). This has enabled description and modeling of ion transport to roots to be based on solid theory and experiment (Nye and Tinker, 1977; Barber, 1984). However, linking model predictions to nutrient uptake under field conditions may not be straightforward (Wild, 1989; Hogberg and Jensen, 1994).
236
Fate and Transport of Heavy Metals in the Vadose Zone
The Conceptual Model Due to the close linkage between plants, microbes, and soil, the study of rhizospheric processes will be a key to resolving problems related to understanding and predicting soil/plant interactions in a changing environment. Therefore, we proposed a conceptual model for nutrient availability in the mineral soil-root system (Cobran and Clegg, 1996). Our hypothesis was that fine roots and their associated organisms maintain a higher level of nutrient availability in the rhizosphere (Rhizo) than in the bulk soil (Bulk) (Figure 11.1). This was accomplished by the release, transport, and accumulation of reactive soil organic matter and inorganic compounds in the soil-root interface (SRI) and rhizosphere. The interaction between soil, microorganisms, and roots creates a mutually supportive system that can raise nutrient availability by increasing moisture content, mineralization, and enriching the pool of cations and anions through increased exchange sites. We envisioned the soil fractions as a multiple-phase system comprised of a gas phase, a solution phase, and a surface phase. The surface phase of the bulk soil with the largest volume of the soil body is represented by a large box, yet has a lower charge per unit mass, thus a lower cation exchange capacity (CEC) than the rhizosphere and SRI. In contrast, the rhizosphere and SRI represent a smaller fraction of the soil body but have a larger CEC due to higher organic matter (OM), clay mineral, and amorphous oxide content (CM). The organic matter component probably differs from the bulk soil since it may contain of a higher proportion of easily mineralized and reactive root material, exudates, mycorrhizae, and other associated microorganisms. We hypothesized that organic matter in the rhizosphere and SRI is the most dynamic part of the system since it acts as both a source and sink for elements, is involved in weathering, and fuels biological reactions. Accordingly, CEC follows the same trend as organic matter content, increasing from the bulk soil to the SRI (Figure 11.1). The arrows represent the three major transport mechanism between the soil fractions and phases; mass flow, diffusion, and biological transport by mycorrhizae. Our studies focused on mineralogical and chemical changes in the solution and surface phases which were considered as products of both chemical and biological interactions.
Field Site and Treatments In a field investigation, soil samples were taken from a 30-yr Norway spruce [Puea abie.J (L.) Karst.] stand situated at Skogaby in southwest Sweden (Cobran and Clegg, 1996). The soil was classified as a Haplic podzol (FAO-UNESCO, 1988) with a silty loam texture throughout the profile. Manipulated field treatments control (C), ammonium sulfate (NS), and irrigation (1) are described in Table 11.5. These treatments caused rapid increases in growth and are discussed in more detail below.
Soil Fractionation Separation of the soil fractions was conducted by carefully removing all roots by hand from the field moist mineral soil, which was then passed through a 2-mm mesh to give the bulk fraction (Bulk). The remaining fine roots « 2-mm) and soil were gently shaken to separate the soil aggregates (0.5 to 5 mm) from the roots to give the rhizosphere fraction. This method is similar to that described by Hendriks and Junk (1981). We
The Rhizosphere and Trace Element Acquisition in Soils
O CO 2
II(
t+
II(
Rhizosphere (Rhlzo)
•
2
ca, Mi' K, AI, H ~
Ca, Mg, K, Na, AI, H
~t Roots, OM and CM
CEC
Bulk
O Cft 2
II(
SOLP
Ca, Mi' K
II(
Na,AI,H
~t
ca Mg K NaAI Na Ca Mg NaAI Na
I
GP
caMgKNaAINa
I~I
.~
237
~t
•
GP
• SOLP
Gas phase (GP) Solution ~hase SOLP)
caMgKNaAIH
OM and CM ~ Organi~ Matter (OM) ~ Cia Minerals (eM
CEC
Surface phase
CEC
Figure 11.1. A conceptual model for nutrient availability in the mineral soil-root system.
Table 11.5. Manipulation Treatments from the Skogaby Sitea Treatment
Description
Control (e) Irrigation (I)
No treatment A sprinkler system prevented water storage deficits greater than 20 mm during May to September. 100 kg N ha-1 and 114 kg S ha- 1 were added as solid ammonium sulfate annually.
Ammonium sulfate (NS)
a
After Nilsson and Wiklund, 1992.
defined the soil-root interface fraction (SRI) as apparent free space within fine roots and adhering rhizoplane soil « 0.5 mm thick), which is similar to the rhizocylinder fraction described by Riley and Barber (1969) and Hoffmann and Barber (1971). Before chemical extraction, the SRI was cut into 5-mm pieces to homogenize the samples and to facilitate weighing.
Chemical Properties of the Soil Fractions The three soil fractions generally differed greatly in chemical composition (Figure 11.2) throughout the upper three mineral horizons. The pH (KCI) was lower in the rhizosphere and SRI than in the bulk soil and tended to increase with depth; however, the differences were not significant in many cases. Titratable acidity (TA = AI + H, cmole kg-I)
L,.:
mE
_Bs
oBh
J C!i
5.0
45,
-l
•
4.5
=a
:::; 0..
i
4.0
~ 15
3.5
o
3.0 Bulk
Rhlzo
:::l V1
It
30
1
ci
b
'"Cl
0
:l
....
0
:r. It)
w <:: '< '
I
~
~
SRI
Bulk
::.;
SRI
Rhlzo
V'.
~
45
2.01
•
1.5 ; 30 ID
rn
c(
ID
(.)
15
1.0 0.5 0.0
0 Bulk
Rhlzo
SRI
j
b
b
I
b
- -
•• •• Bulk
Rhlzo
rp
.< ru
Q.
0
V1 It)
~
• SRI
Figure 11.2. The pH, titratable acidity (TA), base saturation (BS), and calcium-aluminum balance (CAB) in three soil fractions and horizons from the control plots at Skogaby. At a given horizon, soil fractions with differing letters indicate significant differences at the 5% level. The absence of letters indicates no significant difference between soil fractions.
N C
:::;
ro
The Rhizosphere and Trace Element Acquisition in Soils
239
increased significantly in the order Bulk
Weathering in Bulk and Rhizosphere Soil We compared the mineralogy of the bulk and rhizosphere soil fractions of two profiles collected from the untreated plots (Courchesne and Gobran, 1997). The working hypothesis was that the mineral assemblage of the two fractions would differ, reflecting the enhancement of mineral weathering in the immediate vicinity of roots in forest soils. In each profile, samples were collected from the E (0-5 cm), the Bh (5-15 cm), and the Bs (15-30 cm). The mineralogy of the clay-sized particles of both fractions was determined by X-ray diffraction (XRD) of oriented specimens after removal of coatings, saturation with Mg, Mg-ethylene glycol, or K and heating of the K-saturated specimens to 300 and 550°C (Whittig and Allardice, 1986). The integrated intensity of each mineral (I) was normalized relative to the intensity of the (100) peak (J = 0.426 nm) of quartz (I Qz) to calculate a mineral intensity ratio (1/I Qz). Iron and Al were extracted with acid-ammonium oxalate (Ala, Fea) and analyzed by atomic absorption spectrophotometry (AAS). Oxalate dissolves amorphous organic and inorganic solid phases, most of which are of pedogenic origin: they either accumulated as in situ weathering products or precipitated from solution in podzolic soils. Mineral abundance (l/I Qz) in the rhizosphere differed consistently from that in the bulk soil (Figure 11.3). The magnitude of changes was controlled by the relative stability of primary minerals in a weathering environment stimulated by root activity, and followed the order: amphiboles > plagioclases > K-feldspars. The rhizosphere contained significantly lower amounts of amphiboles (I < 0.10) relative to the bulk soil. The abundance of plagioclase was also found to decrease in the rhizosphere for five of the six horizons, but the decrease was not significant. However, XRD showed no rhizosphere effect for K-feldspars. Our results also indicated that expandable phyllosilicates were
240
fate and Transport of Heavy Metals in the Vadose Zone
OBulk
2.5
•
2
~
II Rhizo
a
1.5
S
1
0.5 a
b
0 Amphibole
Int. Vermiculite
Plagioclase
K-Feldspar
Figure 11.3. Average mineral composition of bulk and rhizosphere soil for the three horizons. Soil fractions with differing letters indicate significant differences at the 10% level.
less abundant in the rhizosphere (I < 0.10) and thus less stable than the plagioclase and K-feldspars. This observation is in agreement with Sarkar et al. (1979) but contrary to Kodama et al. (1994). Oxalate extractable Al and Fe were systematically higher in the rhizosphere than in the bulk soil (Figure 11.4), a fact also noted by Sarkar et al. (1979) and Chung and Zasoski (1994). These results support the XRD data by pointing toward an accelerated degradation of mineral structures in the rhizosphere zone. The depletion of weatherable minerals and the concomitant preferential accumulation of weathering products (Al o and Feo) close to root surfaces indicate that the weathering regime was stimulated by root activity. Apart from the work of April and Keller (1990) who observed the preferential dissolution of biotite compared to muscovite close to root surfaces, we are aware of no other field study on the impact of roots on mineral weathering in forest soils. All the knowledge on the weathering of minerals in rhizospheric environments is based on experiments conducted in greenhouses, and mostly using agricultural plants and K-bearing minerals. For example, the accelerated weathering of biotite, phlogopite, or illite and the subsequent release of nutrients (K, Mg, Ca, Fe) in the rhizosphere of wheat (Mortland et al., 1956), pine seedlings (Boyle and Vogt, 1973), soybean grown with mycorrhizal association (Mojallali and Weed, 1978), corn grown with symbiotic and nonsymbiotic microflora (Berthelin and Leyval, 1982), and clover (Tributh et al., 1987) were demonstrated. The most resistant K-bearing minerals (muscovite, K-feldspar) were not significantly affected by the rhizospheric environment. The micas were generally altered to vermiculite, although Spyridakis et al. (1967) reported the transformation of biotite to kaolinite in the rhizosphere of coniferous (cedar, pine, hemlock, spruce) and deciduous (oak, maple) seedlings. Hinsinger and coworkers (Hinsinger and Jaillard, 1993; Hinsinger et al., 1991, 1992, 1993) further showed that the vermiculitization of phlogopite in the rhizosphere of ryegrass and rape was a rapid reaction, on the order of days.
The Rhizosphere and Trace Element Acquisition in Soils
DBulk
241
IIIII Rhlzosphere
80
_ 60 ~~
en
-
If
40
CI/:I
< 20 o AI
AI
E
Bh
AI Bs
Fe
Fe
Fe
E
Bh
Bs
Figure 11.4. Acid ammonium oxalate extractable aluminium and iron in bulk and rhizosphere soil for the three soil fractions and horizons from the control plots at Skogaby. Soil fractions with differing letters indicate significant differences at the 10% level.
These mineralogical changes between rhizosphere and bulk soils were accompanied by physical transformations. Sarkar et al. (1979) and Chung and Zasoski (1994) recorded a reduction in particle size in the vicinity of plant roots. Whether the increased clay content was due to root-induced weathering remains unclear. Bruand et al. (1996) reported the compression of soil particles close to the soil-root interface and measured an increase in bulk density of up to 20%. April and Keller (1990) observed the tangential alignment and bending of phyllosilicate minerals, and the fracturing of grains in contact with roots. Kodama et al. (1994) demonstrated the existence of specific mineral-root associations. Clay particles were shown to be attached on root surfaces with binding preferentially occurring where mucilages accumulated. In some cases, clay-size particles were found to penetrate into the mucilage layer and even into the root surface. Our field data from forest soils and the results of greenhouse experiments indicate that the rhizosphere is more corrosive for weatherable minerals than the bulk soil. It appears that the preferential accumulation and subsequent transformation of organic tissues (organic matter of plant or animal origin, microflora, microbes) in the rhizosphere plays a central role in mineral alteration. The production of acidic and/or complexing substances during the decomposition of organic matter, as a consequence of nutrient uptake or root exudation and following microbial activity, is well established and was discussed earlier. These processes generate a series of organic acids that can efficiently attack mineral structures and complex the dissolution products (e.g., metals). In some instances, inorganic acids such as nitric or sulfuric acids produced by bacteria during nitrification or the oxidation of sulfides, can also contribute to create an even more corrosive environment. However, the dominant acids and complexing agents in the rhizosphere are organic in nature. The ability of low molecular weight (e.g., citric, oxalic) and complex (e.g., fulvic) organic acids to weather soil minerals has long been recognized (Stumm et aI., 1985;
242
Fate and Transport of Heavy Metals in the Vadose Zone
Colman and Dethier, 1986; Tan, 1986; Berthelin, 1988). However, the rhizosphere is of special interest because the total organic acid concentration is higher than in the bulk soil. Moreover, the nature and distribution of these acids differ from that observed in the bulk soil (Mench and Martin, 1991; Petersen and Bottger, 1991; Szmigielska et aI., 1996). It follows that mineral weathering in the rhizosphere can be enhanced, thus favoring the release of soluble weathering products like nutrient cations or trace metals and their uptake by the biota.
Tree Growth and Rhizosphere Chemistry Soil nutrient availability is one of the most important factors influencing tree nutrition, distribution, and health. However, traditional methods of measuring nutrient availability in soils are too often poorly correlated with plant nutrient uptake and status. The following example demonstrates that rhizosphere sampling provides more useful information about ecosystem changes following perturbation than bulk soil sampling alone. For example, we present a case in which we link increased nutrient demands due to stimulated tree growth and rhizospheric chemistry. Rapid growth could deplete nutrients from the rhizosphere when rhizospheric processes cannot keep pace with high demand, particularly if additional nutrient supply is mediated by weathering. The effect of rapid growth was studied in three field treatments (Table 11.5). The NS and I caused significant growth increases of 31 % and 20% within the first three years of application when compared to the control (Nilsson and Wiklund, 1992). Since soil mineralogy, as indicated in the previous section, showed that the weatherable K-bearing minerals were scarce, leaving only feldspars, potassium in the soil is presented here as an example. Exchangeable K and K saturation (K% = KlCEC) had similar trends within soil fractions and horizons. Figure 11.5 presents only K% data and shows that in the control, K% increased in the order Bulk
IMPLICATIONS AND FUTURE RESEARCH A number of observations point to significant changes in the rhizosphere of soils compared to the bulk fraction. Root-induced changes include higher organic matter and dissolved organic carbon levels, enhancement of microbial diversity and population counts, alteration of soil pH and redox potential, accelerated mineral weathering and neoformation, increased bulk density, and major and trace elements enrichment in both the dissolved and exchangeable phases. The rhizosphere thus appears as a distinct soil environment, enriched in various organic substances that favor an intense microbial activity which, in turn, together with exudates and decomposition products, mobilize nutrients and potentially toxic elements through acidification, dissolution, reduction, and complexation. From this point of view, the major reason for conferring a central role to the rhizosphere is the abundance of organic matter and microorganisms; both having a fundamental influence on the bio-
The Rhizosphere and Trace Element Acquisition in Soils
o Bulk
lID Rhizo
243
EI SRI E Horizon
6.0
a
4.0 2.0 0.0 +-'--
c
NS
a
6.0
Bh Horizon
4.0 ~
~
2.0 0.0 C
15.0
NS
8
B8 Horizon
10.0 0~
~
5.0 b
0.0 C
NS
Figure 11.5. The potassium saturation (K%) in three horizons and soil fractions from the control (e), ammonium sulfate (NS), and irrigated (I) plots at Skogaby. Soil fractions with differing letters indicate significant differences at the 5% level.
geochemical cycle of most major and trace elements. Although this constitutes a reasonable assessment of the causes of the rhizosphere effect, it is too static and, thus, has a limited value. The rhizosphere is also an extremely dynamic environment characterized by a range of feedback processes linking phenomena in the biosphere with processes in the pedosphere. For example, reduced nutrient availability in the soil can trigger an
244
Fate and Transport of Heavy Metals in the Vadose Zone
increase in the release of exudates by roots. Under Fe-deficient conditions, roots were shown to increase the exudation of phenolic compounds (e.g., caffeic acids) or of organic chelators (phytosiderophore) that are highly effective at solubilizing Fe from the soil solid phase (Marschner and Romheld, 1996). Moreover, these substances not only form complexes with Fe but they can also mobilize micronutrients like Zn, Cu, and Mn, thus possibly preventing other types of deficiencies. Similarly, tree growth, the uptake of nutrient cations by roots, H+ release, and mineral weathering closely interact to assure a steady flow of matter in the rhizosphere. These interrelated processes contribute to the nutritional status and productivity of the forest and allow the soil-plant system to withstand a range of disturbances and environmental stresses. Therefore, rhizosphere studies have great implications for the understanding and modeling of element acquisition in forest ecosystems. Unfortunately, most forest ecosystem models do not incorporate the functional aspects of the rhizospheric environment of the soil/plant system they describe. For example, Hogberg and Jensen (1994) questioned the prediction of future forest ecosystem conditions based on models using bulk soil chemistry and culture solution input data because they may not represent the buffered nutritional condition of the soil surrounding the roots. In this context, due to the critical interactions that bind plants, microorganisms, and soils, we believe that the study of rhizospheric processes will provide a key in addressing emerging issues like the prediction of soil quality, trace element mobility, bioremediation of contaminated soils, and soil-plant relationships in a changing environment. Future research needs are varied but we believe that our understanding of the biogeochemistry of trace elements in terrestrial ecosystems would strongly benefit from focusing on:
1. The identification, quantification, and characterization of the organic compounds excreted by roots and the dissolved organic substances present in the rhizosphere 2. The quantification and speciation of the dissolved trace elements found in the rhizosphere and their response to chemical changes (pH, organic acids) 3. The characterization of the reactive solid phases (organic and inorganic) produced in the rhizosphere and the mechanisms responsible for their formation 4. The measurement of changes in redox conditions in the rhizosphere with special emphasis on their impact on the availability of macro- and micro nutrients 5. The evaluation of the temporal and spatial dynamics of microbial populations in the rhizosphere 6. The development of small-scale sampling and analytical techniques adapted to the rhizospheric environment that allow the in situ monitoring of processes 7. The measurement of the rhizosphere effect under a range of field conditions (disturbed, manipulated, natural) and for a variety of vegetation types (coniferous, deciduous, agricultural) with emphasis on forested ecosystems. These experiments are required to support and validate the observations obtained from controlled experiments (laboratory or greenhouse), to reduce the lack of information on forest systems, and to help constrain the scaling of knowledge from the laboratory to the field.
The Rhizosphere and Trace Element Acquisition in Soils
245
REFERENCES Alva, A.A., D.A. Eelwards, C.B. Asher, and P.F.C. Blarney. Effects of phosphorus/aluminum molar ratio and calcium concentration on plant response to aluminum toxicity. SoiL Sci. Soc. Am. J. 50, pp. 133-137, 1986. April, Rand D. Keller. Mineralogy of the rhizosphere in forest soils of the eastern United States. Biogeochemiltry. 9, pp. 1-18, 1990 Barber, S.A. SoiL Nutrient BioavaiLahiLty. A Mechaniltic Approach. John Wiley & Sons, New York, 1984. Bartlett, R Manganese Redox in the Rhizosphere. Agronomy Ahdtractd, AnnuaLMeeting, IndianapoLil, Indiana, USA, Nov. 3-8, 1996, p. 204. Beier, C. and T. Cummins. The Future, and Current Limitations, of Manipulations and Monitoring of Terrestrial Ecosystems. EcOdYdtemd Re.Jearch Report No.4. ExperimentaLManipuLationd ofBiota and BiogeochemicaL CycLing in EcodYdtemd. Commission of the European Communities, 1993, pp. 338-341. Beijerink, M.W. Die Bakterien der Papilionaccenknollchen. Bot. Ztg. 46, pp. 725-804, 1888. Berthelin, J. Microbial Weathering Processes in Natural Environments, in PhYdicaL and ChemicaL Weathering in GeochemicaL Cyc!e.J, A. Lerman and M. Meybeck, Eels., NATO ASI Series, Kluwer Academic Publishers, Dordrecht, 1988. Berthelin, J. and C. Leyval. Ability of symbiotic and non-symbiotic rhizospheric microflora of maize (Zea mayd) to weather micas and to promote plant growth and plant nutrition. Plant Soil 68, pp. 369-377, 1982. Binkley, D. Foredt Nutrition Management. John Wiley & Sons, New York, 1986, p. 290. Bolan, N.S. A critical review on the role of mycorrhizal fungi in the uptake of phosphorus by plants. Plant SoiL 134, pp. 189-207, 1991. Bowen, G.D. The toxicity of legume seed diffusates toward rhizobia and other bacteria. Plant SoiL 15, pp. 155-165, 1961. Bowen, G.D. and A.D. Roriva. The Rhizosphere. The Hidden Half of the Hidden Half, in PLant Rootd: The Hi2den Half, Y. Waisel, A. Eshel, and T. Kafkafi, Eels., Marcel Dekker, Inc., 1991, pp. 641-669. Boyle, J.J. and J.R. Shann. Biodegradation of phenol, 2,4-DCP, 2,4-D, and 2,4,5-T in fieldcollected rhizosphere and nonrhizosphere soils. J. Environ. QuaL. 24, pp. 782-785, 1995. Boyle, J.R. and G.K Vogt. Biological weathering of silicate minerals: Implications for tree nutrition and soil genesis. Plant SoiL. 38, pp. 191-201, 1973. Brady, N.C. The Nature and Propertied of Soiu, 10th ed., Macmillan Publishing Co., New York, 1990. Brown, D.A. and A. Ul-Haq. A porous membrane-root-root culture technique for growing plants under controlled soil conditions. Soil Sci. Soc. Am. J. 48, pp. 692-695, 1984. Bruand, A., I. Cousin, B. Nicoullaud, o. Duval, and J.-C. Begon. Backscattcred electron scanning images of soil porosity for analyzing soil compaction around roots. SoiL Sci. Soc. Am. J. 60, pp. 895-901, 1996. Chapin, F.S. The mineral nutrition of wild plants. Ann. Rev. Eco!. SYdt. 11, pp. 233-260, 1980. Chapin, F.S. The evolutionary basis of biogeochemical soil development. SoiL Sci. 57, pp. 223-227, 1993. Chung, J.B. and RJ. Zasoski. Ammonium-potassium and ammonium-calcium exchange equilibria in bulk and rhizosphere soil. Soil Sci. Soc. Am. J. 58, pp. 1368-1375, 1994. Clarholm, M. The Microbial Loop in Soil, in Beyond The BiomaM, K Ritz, I. Dighton, and KE. Giller, Eels., Wiley-Savce Publications, 1984. Clark, F.E. Soil microorganisms and plant roots. Adv. Agron. 1, pp. 241-288, 1949.
246
Fate and Transp0l1 of Heavy Metals in the Vadose Zone
Clegg, S. and G.R. Gobran. Rhizospheric P and K in forest soil manipulated with ammonium sulfate and water. Can. J. Soil Sci., 77, pp. 525-533, 1997. Clegg, S., G.R. Gobran, and X. Guan. Rhizosphere chemistry in an ammonium sulfate and water stressed Norway spruce [PiCea abie.:J (L.) Karst.] forest. Can. J. SoiL Sci., 77, pp. 515-523, 1997. Clemensson-Lindell, A. and H. Persson. Effects of freezing on the rhizosphere and root nutrient content using two sampling methods. Plant Soil, 139, pp. 39-45, 1992. Cole, D.W. Soil nutrient supply in natural and managed forests. Plant SoiL, 168-169, pp. 43-53,
1995. Coleman, D.C., c.v. Cole, H.W. Hunt, and D.A. Klein. Trophic interactions as they affect energy and nutrient dynamics. 1. Introduction. MiCrob. &01. 4, pp. 345-349, 1978. Colman, S.M. and D.P. Dethier. Rate.J 0/ ChemiCaL Weathering 0/Rock.! ano Minerau. Academic Press, New York, 1986. Conkling, B.L. and R.W. Blanchar. Glass microelectrode techniques for in.!itu pH measurements. SoiL Sci. Soc. Am. J. 53, pp. 58-62, 1989. Conkling, B.L., R.W. Blanchar, and T.L. Niblack. Effects of foliar and soil acidity on the rhizosphere pH of alfalfa, corn, and soybean. J. Environ. QuaL. 20, pp. 381-386, 1991. Courchesne, F. and G.R. Gobran. Mineralogical variations of bulk and rhizosphere soils from a Norway spruce stand, southwestern Sweden. SoiL Sci. Soc. Am. J. 61, pp. 1245-1249, 1997. Cronan, C.S. and D.F. Grigal. Use of calcium/aluminum ratios as indicators of stress in forest ecosystems. J. Environ. QuaL. 24, pp. 209-226, 1995. Curl, E.A. and B. Truelove. The RhizNphere. Springer-Verlag, Berlin, 1986. DeAngelis, D.L., W.M. Post, and C.C. Travis. POditive Feeoback in NaturaL Sy.!temd. SpringerVerlag, Berlin, 1986. Dormaar, J.F. Effect of plant roots on chemical and biochemical properties of surrounding discrete soil zones. Can. J. Soil Sci. 68, pp. 233-242, 1988. Drever, J.1. and G.F. Vance. Role of Soil Organic Acids in Mineral Weathering Processes, in OrganiC ACid.! in GeologiCaL Procedde.J. E.D. Pittmsan and M.D. Lewan, Eds., Springer-Verlag, 1994, pp. 138-161. Fahey, T.J. and J.W. Hughes. Fine root dynamics in a northern hardwood forest ecosystem, Hubbard Brook Experimental Forest, NH. J. &oL. 82, pp. 533-548, 1994. FAa-UNESCO. SoiL Map 0/ the WorLo. Revised legend. FAa, Rome, 1988. Finlay, R.D. Interactions between soil acidification, plant growth and nutrient uptake in ectomycorrhizal associations of forest trees. &oL. BuLL. 44, pp. 197-214, 1995. Finlay, R.D. Uptake and Mycelial Translocation of Nutrients by Ectomycorrhizal Fungi, in MycorrhizaJ in &o.!ydlemd. D.J. Read, D.H. Lewis, A.H. Fitter, and I.J. Alexander, Eds., C.A.B. InternationaL Wallingford, 1992. Finlay, R.D. and B. Soderstom. Mycorrhizal Mycelia and Their Role in Soil and Plant Communities, in &ofo.qy 0/ Arable Lano. M. Clarholm and L. Bergstrom, Eds., Kluwer Academic Publishers, 1989. Fleige, H., O. StrebeL M. Regner, and H. Grimme. Die potentielle P-Anlieferung durch Diffusion als Funktion von Tiefe, Zeit unt Durchwurzelung bei einer Parabraunerde aus Loss. Mitt. Dt.!ch. Booenko. Ge.J. 32, pp. 381-386, 1981. FogeL R. Root turnover and productivity of forests. Plant SoiL 71, pp. 75-85, 1983. FogeL R. and G. Hunt. Contribution of mycorrhizae and soil fungi to nutrient cycling in a Douglas-fir ecosystem. Can. J. For. 1W. 12, pp. 219-232, 1983. Fox, R.L. and R.C. Lipps. Distribution and activities of roots in relation to soil properties. Trnd.
Int. Congr. SoiL Sci. 7th, 1960, 1961. Fox, T.R. and N.B. Comerford. Low molecular weight organic acids in selected forest soils of the southeastern USA. SoiL Sci. Soc. Am. J. 54, pp. 1139-1144, 1990.
The Rhizosphere and Trace Element Acquisition in Soils
247
Gobran. G.R and E. Bosatta. Cation depletion rate as a measure of soil sensitivity to acidic deposition: Theory. Ecol. MooelL. 40. pp. 25-36. 1988. Gobran. G.R and S. Clegg. A conceptual model for nutrient availability in the mineral soil-root system. Can. J. SOlt Set: 76. pp. 125-131, 1996. Gobran. G.R. S. Clegg. and F. Courchesne. Rhizospheric processes influencing the biogeochemistry of forest ecosystems. BiogeochemiJtry, in press. 1998. Gobran G.R. L.B. Fenn. and I. Al-Windi. Nutrition response of Norway spruce and willow to varying levels of calcium and aluminum. Fert. Ru. 34. pp. 181-193. 1993. Grayston S.J .• D. Vaughn. and D. Jones. Rhizosphere carbon flow in trees. in comparison with annual plants: The importance of root exudation and its impact on microbial activity and nutrient availability. App. SoiL. &ol. 5. pp. 29-56. 1996. Grierson. P. F. Organic acids in the rhizosphere of Bank.1ia intergrifolia LJ. Plant Soil, 144. pp. 259265. 1992. Haby. P.A. and D.E. Crowley. Biodegradation of 3-chlorobenzoate as affected by rhizodeposition and selected carbon substrates. J. Environ. QuaL.. 25. pp. 304-310. 1996. Haussling. M. and H. Marschner. Organic and inorganic soil phosphates and acid phosphatase activity in the rhizosphere of 80-year-old Norway spruce [PtCea ab0 (L.) Karst.] trees. Biol. FertiL. Sow 8. pp. 128-133. 1989. Hedley. MJ .. P.H. Nye. and RE. White. Plant-induced changes in the rhizosphere of rape (BrtMJtCa napLM var. emerald) seedlings II. Origin of the pH change. New PhytoL. 91, pp. 31-44. 1982. HeiaL H.M. and D.R Sauerbeck. Method to study turnover processes in soil layers of different proximity to roots. Soil Biol. Biochem. 15. pp. 223-225. 1983. HellriegeL H. and H. Wilfarth. Beilageheft zu oer Z Ver RubelkJzucker-Ino DtJch RetChJ. 1888. p. 234. Hendrick. R L. and K S. Pregitzer. The demography of fine roots in a northern hardwood forest. &owgy 73. pp. 1094-1104. 1992. Hendrick. RL. and KS. Pregitzer. Patterns of fine root mortality in two sugar maple forests. Nature 361. pp. 59-61, 1993. Hendriks. L. and A. Junk. Erfassung den Mineralstoffverteilung in Wurzeln durch getrennte Analyse von Rhizo- und Restboden. Z Pflanzenern Booenlco. 144. pp. 195-202. 1981. Hiltner. L. Uber neuere Erfahrungen und Probleme auf dem Gibiet der Bodenbakeriologie und unter besonderer Berucksichtigung der Grundungung und Brache. Arb DtJch. Lanowirt GeJ. 98. pp. 59-78. 1904. Hinsinger. Ph. and B. Jaillard. Root-induced release of interlayer potassium and vermiculitization of phlogopite as related to potassium depletion in the rhizosphere of ryegrass. J. Soil Sci. 44. pp. 525-534. 1993. Hinsinger. Ph .• J.E. Dufey. and B. Jaillard. Biological Weathering of Micas in the Rhizosphere as Related to Potassium Absorption by Plant Roots. in Plant RootJ ano Their Environment. B.L. McMichael and H. Persson. Eds .. Elsevier. Amsterdam. 1991. Hinsinger. Ph .• B. Jaillard. and J.E. Duffey. Rapid weathering of trioctahedral mica by the roots of ryegrass. Soil Sci. Soc. Am J. 56. pp. 977-982. 1992. Hinsinger. Ph .. F. Elsass. B. Jaillard. and M. Robert. Root-induced irreversible transformation of a trioctahedral mica in the rhizosphere of rape. J. Soil Sci. 44. pp. 535-545. 1993. Hobbie. S.E. Effects of plant species on nutrient cycling. TrenoJ &0L. Evol. 6. pp. 336-339. 1992. Hoffmann. W.F. and S.A. Barber. Phosphorus uptake by wheat (TrittCumaeJitivllm) as influenced by ion accumulation in the rhizocylinder. Soil Sci. 112. pp. 256-262. 1971. Hogberg. P. and P. Jensen. Aluminium and uptake of base cations by tree roots: A critique of the model proposed by Sverdrup et al. Water Air Soil PoLLut. 75. pp. 121-125. 1994. Huang. P.M. Kinetics of Redox Reactions on Manganese Oxides and Its Impact on Environmental Quality. in RateJ of Soil ChemtCal ProceJJeJ, D.L. Sparks and D.L. Suarez. Eds .. SSSA Special Publication No. 27. 1991.
248
Fate and Transport of Heavy Metals in the Vadose Zone
Inskeep, W.P. and S.D. Comfort. Thermodynamic predictions for the effects of root exudates on metal speciation in the rhizosphere. J. Pi£1nt Nutr. 9, pp. 567-586, 1986. Jackson, R.B., J. Canadell, J.R. Ehleringer, H.A. Mooney, O.E. Sala, and E.D. Schulze. A global analysis of root distributions for terrestrial biomes. Oecologica 108, pp. 389-411, 1996. Jakobsen, I., L.K Abbot, and A.D. Robbson. External hyphae of vesicular-arbuscular mycorrhizal fungi associated with Tri/oLium,fU6terraneum L. 1. Spread of hyphae in phosphorus inflow to roots. New PbytoL. 120, pp. 371-380, 1992. Johnson, C.E., C.T. Driscoll, T.J Fahey, T.G. Siccama, and J.W. Hughes. Carbon Dynamics Following Clear-Cutting of a Northern Hardwood Forest, in Car60n ForfTkl and Functiofhl in Forut Sow, W.W. McFee and J.M. Kelly, Eds., Soil Science Society of America, Madison, WI, 1995. Kirlew, P.W. and D.R Bouldin. Chemical properties of the rhizosphere in acid sub-soil. SoiL Sci. Soc. Am. J. 51, pp. 128-132, 1987. Kochian, L.v. Mechanisms of Micronutrient Uptake and Translocation in Plants, inMicronutrientd in AgricuLture, 2nd ed., J.J. Mortvedt, Ed., Soil Science Society of America, Madison, WI, 1991. Kodama, H., S. Nelson, A.F. Yang, and N. Kohyama. Mineralogy of rhizospheric and nonrhizospheric soils in corn fields. Ci£1yd Ci£1y Minerau 42, pp. 755-763. 1994. Kosola, KR Sampling Roots of Known Age in the Field. Agronomy A6dtract, American Society of Agronomy 88tb AnnuaL Meeting 1996. Indianapolis, IN. 1996. Kuchenbuch. Rand A. Jungk. A method for determining concentration profiles at the soil-root interface by thin slicing rhizospheric soil. Pi£1nt SoiL 68, pp. 391-394, 1982. Lyncb, J.M. T6e RiJizodP6ere. Jonn Wiley & Sons. 1990. Macar. N.E. and RD. Graham. Genotypic variation for manganese efficiency in wheat. J. Pi£1nt Nutr. 10, pp. 2049-2055. 1987. Mahendrappa. M.K. N.W. Foster. G.F. Weetman. and H.H. Krause. Nutrient cycling and availability in forest soils. Can. J. Soil Sci. 66. pp. 547-572, 1986. Majdi, H. E/fectd of Nutrient AppLicationd on Fine-Root Dynamic.! and RootlRbiw.Jpbere Cbemi.Jtry in a Norway Spruce Stand. Swedish University of Agricultural Sciences. Report 71, 1994. Marschner, H. MineraL Nutrition of Higber Pi£1ntd, Academic Press. London. 1986. Marschner, H. and B. Dell. Nutrient uptake in mycorrhizal symbiosis. Pi£1nt Soil 159. pp. 89-102. 1994. Marschner. H. and V. Romheld. Root Induced Changes in the Availability of Micronutrients in the Rhizosphere. inpi£1nt Rootd: Tbe HWdenHa!j. 2nd ed .• Y. Waisel. E. Amram. and U. Kafkafi. Eds .. Marcel Dekker. Inc .• 1996. Mench, M. and E. Martin. Mobilization of cadmium and other metals from two soils by root exudates by Zea maYd L.. Nicotina ta6acum L.. and Nicotina rUdtica L. Pi£1nt Soil. 132. pp. 187-196. 1991. Mengel. K, D. Horn, and H. Tributh. Availability of interlayer ammonium as related to root vicinity and mineral type. SoiL Sci. 149. pp. 131-137. 1990. Mojallali. H. and S.B. Weed. Weathering of micas by mycorrhizal soybean plants. SoiL Sci. Soc. Am. J. 42. pp. 367-372. 1978. Mortland, M.M .• K Lawton, and G. Uehara. AI teration of biotite to vermiculite by plant growth. Soil Sci. 82, pp. 477-481, 1956. Nilsson, L.a. and K Wiklund. Influence of nutrient and water stress on Norway spruce production in south Sweden- The role of air pollutants. Pi£1nt SoiL 147. pp. 251-265. 1992. Noble, A.D and M.E. Sumner. Calcium and aluminum interactions and soybean growth in nutrient solutions. Commun. SoiL Sci. Pi£1nt AnaL. 19. pp. 1119-1131. 1988. Nye. P.H. Acid-base changes in the rhizosphere. Adv. Pi£1nt Nutr. 2. pp. 129-153. 1986. Nye, P.H. and P.B. Tinker. SoLute Movement in tbe SoiL Root SYdtem. Blackwell Scientific Publications. Oxford, London. Edinburgh, Melbourne. 1977.
The Rhizosphere and Trace Element Acquisition in Soils
249
Parmelee, RW., J.G. Eherenfeld, and RL. Tate. Effects of pine roots microorganisms and nitrogen availability in two horizons of a coniferous spodosol. BioL FertiL Soiu. 15, pp. 113-119, 1993. PauL E.A. and F.E. Clark. SoiL Microbiology and Biochemidtry. Academic Press, Inc., 1989, p. 273. Perry, D.A., M.P. Amaranthus, J.G. Borchers, S.L. Borchers, and RE. Brainerd. Bootstrapping in ecosystems. Biodcience 39, pp. 230-237, 1989. Persson, H. The distribution and productivity of fine roots in boreal forests. Plant SoiL 71, pp. 87101, 1983. Persson, H., Y. Von Fricks, and H. Majdi. Root distribution in a Norway spruce [Picea died (L.) Karst.] stand subjected to drought and ammonium sulfate application. PLant Soil 168-169, pp. 161-165, 1995. Petersen, W. and M. Bottger. Contribution of organic acids to the acidification of the rhizosphere of maize seedlings. Plant SoiL 132, pp. 159-163, 1991. Pidgeon, C.S., C.M. Reynolds, l.K Iskandar, and F.R Magdoff. Rhizosphere Enhanced Bioremediation of an Alaska SoiL in Agronomy Abdtractd, AnnuaL Meeting, IndianapoLid, Indiana, USA, Nov. 3-8, 1996, 1996, p. 229. Reilley, KA., M.K Banks, and A.P. Schwab. Dissipation of polycyclic aromatic hydrocarbons in the rhizosphere. J. Environ. QuaL 25, pp. 212-219, 1996. Richards, B.N. The Microbiology of TerredtriaL &odYJtemd. John Wiley & Sons, New York, 1987. Riley, D. and S.A. Barber. Bicarbonate accumulation and pH changes at the soybean [GLycine nuz.x (L.) Merr.] soil-root interface. SoiL Sci. Soc. Amer. Proc. 33, pp. 905-908, 1969. Riley, D. and S.A. Barber. Effect of ammonium and nitrate fertilization on phosphorus uptake as related to root-induced pH changes at the root-soil interface. SoiL Sci. Soc. Am. Proc. 35, pp. 301-306, 1971. Rollwagen, B.A. and R.J. Zasoski. Nitrogen source effects on rhizosphere pH and nutrient accumulation by Pacific Northwest conifers. Plant SoiL 105, pp. 79--86, 1988. Roriva, A.D. Plant root exudates. Bot. Rev. 35, pp. 35-57, 1969. Roriva, A.D. and G.D. Bowen. The effects of microorganisms upon plant growth. II. Detoxification of sterilized soils by fungi and bacteria. Plant Soil 14, pp. 199-214, 1966. Sarkar, A.N., D.A. Jenkins, and R.G. Wyn Jones. Modifications to mechanical and mineralogical composition of soil within the rhizosphere, in The SoiL-Root Interface, J.L. Harvey and R Scott-Russell, Eds., Academic Press, London, 1979. Sarkar A.N. and RG. Wyn Jones. Effect of rhizosphere pH on the availability and uptake of Fe, Mn, and Zn. Plant SoiL 66, pp. 361-372, 1982. Skladany G.J. and F.B. Meeting, Jr. Bioremediation of Contaminated SoiL in Soil MicrobiaL &oLogy. F.B. Meeting, Jr., Ed., Marcel Dekker, Inc. 1993, pp. 483-513. Soil Science Society of America. GLoddary of SoiL Science Termd. Soil Science Society of America, Madison, WI, 1997, p. 138. Soon, Y.K and M.H. Miller. Changes in the rhizosphere due to NH4 + and N0 3- fertilization and phosphorus uptake by corn seedlings (Zea maYd L.). SoiL Sci. Soc. Am. J. 41, pp. 77--80, 1977. Spyridakis, D.E., G. Chesters, and S.A. Wilde. Kaolinitization of biotite as a result of coniferous and deciduous seedling growth. SoiL Sci. Soc. Am. Proc. 31, pp. 203-210, 1967. Stevenson, F.J. Organic Matter-Micronutrient Reactions in SoiL in Micronutrientd in AgricuLture. SSSA Book Series, No.4. 1991, pp. 145-186. Stumm, W., G. Furrer, E. Wieland, and B. Zinder. In The Chemutry of Weathering. NATO ASI Seried, J.A. Drever, Ed., D. Reidel Publishing Company, Dordrecht, 1985, pp. 55-74. Szmigielska, A.M., KC.J. Van Rees, G. Cieslinski, and P.M. Huang. Low molecular weight dicarboxylic acids in rhizosphere soil of durum wheat. J. Agric. FOOd Chem. 44, pp. 1036-1040, 1996. Tan, KK Degradation of Soil Minerals by Organic Acids, in Interactiond of SoiL Minerau with NaturaL Organic.J and Microbu, P.M. Huang and M. Schnitzer, Eds., Soil Science Society of America, Madison, WI, 1986.
250
Fate and Transport of Heavy Metals in the Vadose Zone
Tinker. P.B. Nutrient Uptake by Plant Roots in Natural Systems. in Nutrient CycLing in Terr&ftriaL Eco.!y.!tenu. A.F. Harrison. P. Ineson. and O.W. Heal. Eds .• Elsevier Applied Science. London. 1990. Trettin. C.C .• M.F. Jurgensen. M.R Gale. and J.W. McLaughlin. Soil Carbon in Northern Forested Wetlands: Impacts of Silvicultural Practices. in Carbon Fornu and Function.! in For&ft Soiu. W. W. McFee and J .M. Kelly. Eds .. Soil Science Society of America. Madison. WI. 1995. Tributh. H .• E. von Bogulawski. A. von Lieres. D. Steffens. and K Mengel. Effect of potassium removal by crops on transformation of illitic clay minerals. SoiL Sci. 143. pp. 404-409. 1987. Ulrich. B. Stability. Elasticity. and Resilience of Terrestrial Ecosystems with Respect to Matter Balance. in EcologicaL StUdied 61, E.-D. Shultze and H. Zolfer. Eds .• Springer-Verlag. Berlin. Heidelberg. 1987. Uren. N. C. and H.M. Reisenauer. The role of nutrient exudates in nutrient acquisition. Adv. Plant Nutr. 3. pp. 79-114. 1988. Van Breemen. N. Soils as biotic constructs favoring net primary production. Geoderma 57. pp. 183-211. 1993. Van Breemen. N. Nutrient cycling strategies. Plant SoiL 168-169. pp. 321-326. 1995. Van Cleve. K and RF. Powers. Soil Carbon. Soil Formation and Ecosystem Development. in Carbon Fornu and Function.! in For&ft Sow. W.W. McFee and J.M. Kelly. Eds .• Soil Science Society of America. Madison. WI. 1995. Vogt. KA.• D.A. Publicover. J. Bloomfield. J.M. Prerez. D.J. Vogt. and W.L. Silver. Belowground responses as indicators of environmental change. Environ. Exp. Botany 33, pp. 189-205. 1993. Waring, RH. and W.H. Schlesinger. For&ft Eco.!y.!tenu, Academic Press, 1985. Whittig, L.D. and W.R Allardice. X-ray diffraction techniques. in MethOd.! of SoiLAnaLy.!iJ. Part 1. Phy.!icaL and MineralogicaL MethOd.!, 2nd ed., A. Klute, Ed., Agron. Monogr. 9. ASA and SSSA, Madison, WI. 1986. Wild, A. Soil scientists as members of the scientific community. J. SoiL Sci. 40, pp. 209-221, 1989. Wilkins, D.A. The influence of sheathing (ecto) mycorrhizas of trees on the uptake and toxicity of metals. Agric. EcO.l. Environ. 35, pp. 145-160, 1991. Wood, T., F.H. Bormann, and G.J. Vogt. Phosphorus cycling in a northern hardwood forest: Biological and chemical control. Science. 223, pp. 391-393. 1984.
CHAPTER 12
Distribution of Ecologically Significant Fractions of Selected Heavy Metals in the Soil Profile T. Nemeth, K. Bujtas, J. Csillag, G. Partay, A. Lukacs, and M.Th. van Genuchten
INTRODUCTION The amount of wastes, wastewaters, and sewage sludges produced by agricultural, industrial, and municipal activities is rapidly increasing worldwide. In the developing regions of the world this may be simply the result of the improving supply of clean tap water and canalization. Because of increasing environmental awareness, dumping of sewage into surface waters is subject to more strict regulations, thus the amount of wastewaters subjected to treatments is also increasing in nonindustrialized countries. In consequence, a large growth in sewage sludge production may be expected, especially when taking into account the higher requirements and standards for wastewater treatment. In Hungary approximately 1000 million cubic meters (mS) of wastewater were produced per year in the middle of the '80s, of which only 187 million m 3 were sufficiently treated, the majority only partially treated, and 173 million m 3 not treated at all. At that time, there was an increasing gap between the development of municipal water supply and of sewage systems, with the latter lagging behind the substantial improvements in the water supply. To develop the collection and proper treatment of liquid wastes is still a problem for many smaller municipalities. According to recent data, the amount of sewage sludges in Hungary was above 1 million m 3 per year. About 40% of these sludges were deposited on agricultural fields and on forest plantations. One reasonable and economic way to dispose of wastewaters and sludges is to apply them to agricultural fields, thereby exploiting their water and nutrient content. Currently, this practice is becoming increasingly important in many countries. In the early nineties about 30-50% of the sewage sludges were disposed by land application in the majority of the industrialized European countries, which compares to 33% of the annual sludge production in the United States (McGrath et aI., 1994). 251
252
Fate and Transport of Heavy Metals in the Vadose Zone
Sludge Application Excessive application of wastewater and sewage sludge to agricultural land may cause soil and groundwater pollution problems, since heavy metals and other, potentially toxic elements may be regarded as characteristic contaminants of sewage sludges and wastewaters. These elements often occur in large amounts in these wastes, limiting their applicability in agriculture (Chang et aI., 1984; Juste and Mench, 1992). Sludge application in Hungary is officially regulated in order to prevent or limit the pollution of soils, surface waters and groundwater (Hung. Techn. Dir., 1990; Molnar et aI., 1995). The regulations use a 3-step system in which water and nitrogen content of the sludge, and the concentration of any toxic element in the sludge and in the sludgetreated soil are the limiting factors. Also, limiting conditions have been defined for each soil factor related to the application of wastewaters and sewage sludge on a specific agricultural field. Taking into consideration the hazard of contamination of surface and subsurface waters, a permanent monitoring is required on fields where these materials are used, in order to control the load effect. The upper limits for the toxic elements used in the case study discussed in later sections are 15 mg Cd, 1000 mg Cr, 200 mg Ni, 1000 mg Pb, and 3000 mg Zn per kg sludge dry matter (Hung. Techn. Dir., 1990). Assuming average sludge application practices (i.e., incorporating 500 tlha sewage sludge containing 5% dry matter into a 20-cm surface soil layer), these limits are equal to loadings of 0.375 kg Cd, 25 kg Cr, 5 kg Ni, 25 kg Pb, and 75 kg Zn per hectare. The cumulative amounts of the toxic elements after prolonged application of sludge or of any other contaminant sources must not exceed their maximum acceptable concentrations in the soil, which for the metals selected for the present study are the following: 6.4 kg Cd, 320 kg Cr, 64 kg Ni, 320 kg Pb, and 800 kg Zn per ha for soils with an adsorption capacity of 15-25 cmol/kg soil as calculated from maximum allowed limits of soil concentrations. Although the limit values relating to sludge application practice are generally lower in Hungary than in the European Community or in the United States (Hung. Techn. Dir., 1990; USEP A, 1989; Chang et aI., 1992; McGrath et al., 1994), improper or illegal deposition of sludges and industrial wastes may lead to serious contamination of the Hungarian environment with toxic substances.
Adsorption and Mobility The potential risks for heavy metal contamination in agricultural fields depend on the relative amounts of metals being adsorbed versus amounts in the soil solution. Potentially toxic elements may become especially hazardous to the environment, when they enter the liquid phase of the soil. Soil solution concentrations are regarded to be indicators of the mobile pool of metals in soils (Kabata-Pendias and Adriano, 1995). Watersoluble forms of an element may move through the vadose zone (Kabata-Pendias and Pendias, 1992). Although the removal of metals from sludge-treated soils by leaching to the subsurface is normally very small (McGrath et aI., 1994; Kabata- Pendias and Adriano, 1995), some literature data indicate that sludge-born metals may get further from the site of contamination under specific circumstances (McBride, 1995). There are contradictory views in the literature regarding the extent to which sludgeborn metals are vulnerable to leaching. Differences among experimental conditions may be one reason for the differences observed in the movement of various metals in the soil.
Distribution of Ecologically Significant Fractions of Selected Heavy Metals
253
For instance, Legret et al. (1988) found significant migration of Ni and especially of Cd, but little or no movement of Pb and Cr during a field experiment on a sludge-treated, coarse-textured soil. Dudka and Chlopecka (1990) observed leaching of Cd, Cr, Ni, Pb, and Zn in a lysimeter experiment on a sandy loam contaminated with sewage sludge containing large amounts of these elements. In contrast, Dowdy and Yolk (1983), Chang et al. (1984), and Alloway (1990) reported retention of these elements within the zone of sludge incorporation during both column and field experiments. Metals originating from anthropogenic sources usually occur in the soil in forms different from the original, native metal content of the soil, thus their mobility and availability is also different. This may change when the freshly added metals enter the dynamic equilibria among the various forms of the elements in the soil chemical processes. When evaluating the biological and ecological impacts of soil contamination, it is necessary to estimate the total amounts of the toxic elements which may become available even after longer periods. Treatment of soil samples with 2 mollL HN0 3 at 100°C as proposed by Andersson (1976) was shown to be a suitable method to determine the totaL potentiaLLy available /ractwn of the metals in the soil.
Extractions and Bioavailability The plant-available concentratwtU of heavy metals occurring in the soil are typically estimated by different chemical extraction methods. Several extractants and extraction procedures have been proposed to determine the availability of essential elements, as well as of some elements potentially toxic to plants. Much research has been carried out to develop a universal extractant characterizing plant availability, and applicable to several elements in soils having widely different properties. One of the most generally useful extractants of this type is acid ammonium acetate +EDT A (AAAc- EDT A). This extractant has been found suitable for the simultaneous extraction of both macro- and micronutrients (Lakanen and Ervio, 1971), and is now included in the procedures adopted by the Soil Advisory Service in Hungary. Extensive F AO studies involving 30 countries proved that this extraction methodology is appropriate for the micronutrients Zn, Fe, Cu, and Mn (Sillanpaa, 1982), and also for some toxic elements such as Cd, Pb, Co, and Se (Sillanpaa and Jansson, 1992). Theoretically, the best measure of the availability of elements in a specific soil-plant system should be their actual amounts taken up by the plants. These amounts are often poorly correlated with extractable amounts of the elements in the soil (Marschner, 1991; Chang et aI., 1992; McBride, 1995), since they depend not only on soil factors but also on several plant properties and are governed by physiological and biochemical processes. In contrast, soil extraction methods depend on the laws of soil chemistry. Thus, the fractions of the element content utilized by the plants and assessed by plant analyses are not the same as those measured by the soil chemist. However, plant analyses have many practical limitations, and plant uptake processes are susceptible also to environmental factors (Sillanpaa and Jansson, 1992). Hence, chemical extraction methods remain useful tools to assess phytoavailability of the elements in the soil. Recent evidence suggests that elemental concentrations in the soil solution itself may serve as a useful diagnostic tool for plant uptake of the various elements. It is reasonable to expect that plant uptake will be a function of the soil solution concentrations since
254
Fate and Transport of Heavy Metals in the Vadose Zone
water-soluble forms of an element are generally most easily and immediately available for plant uptake (Petruzzelli, 1989). However, there is only a limited amount of data in the literature to support this idea for the toxic elements (Alloway, 1990). Several laboratory methods have been developed for obtaining the liquid phase from undisturbed bulk soils or from air-dried, ground and rewetted soil samples. Many investigators attempted to characterize the energy status of the liquid phase to be separated from soils. Some of these studies focused mainly on macro-elements (Zabowski and Ugolini, 1990; Jones and Edwards, 1993; Csillag et al., 1995). Others have determined the heavy metal content of the soil solution (e.g., Mullins and Sommers, 1986; cit. in Campbell and Beckett, 1988; cit. in Kabata-Pendias and Pendias, 1992).
CASE STUDY The initial hypothesis of this case study presented here is as follows. Following the application of heavy metals to the soils, quasi-equilibrium concentration in the soil solution fraction which is held in the soil by forces corresponding to less than -1500 kPa (pF 4.2 = the conventional wilting point of plants), i.e., which are directly accessible to the roots, may be regarded as the most important variable characterizing plant availability of the elements. To test this hypothesis, sewage sludge spiked with Cd, Cr, Ni, Pb, and Zn nitrates was applied to the top layer of large undisturbed soil monoliths. The distributions of the total potentially available and the plant-available fractions (characterized by the AAAc-EDTA extractable amounts and also by the directly plant-accessible soil solution concentrations) of the applied metals in the soil profile of the monoliths are discussed here. Other aspects of the study were presented elsewhere (Nemeth et al., 1994; Bujtas et al., 1995). A brown forest soil (Ochrept, from Godollo) and a slightly acidic sandy soil (Psamment, from Somogysard) were included in the experiments, each one with four monoliths. The major physical and chemical properties of these soil types are shown in Table 12.1. The undisturbed, 40-cm diameter, lOO-cm long soil monoliths were prepared following the methods proposed by Homeyer et al. (1973) and modified by Nemeth et al. (1991). The monoliths were excavated at the selected field sites and their cylindrical surfaces were coated with fiberglass cloth impregnated with a synthetic resin. After the coatings solidified, the monoliths were lifted, and the bottoms were similarly coated. The monoliths were subsequently transported by truck to the laboratory. The coatings made extremely close contact with the soil surface by imbibing the outer micropores, thus eliminating possible "wall effects" during the experiments. The excellent insulating properties of the coating also eliminated leakage from the monoliths. Weights of the columns at the beginning and at the end of the experimental procedure were recorded on a movable scale. Changes in temperature, soil moisture content, and gas composition along the soil profile were followed by sensors inserted into the monoliths at various depths through holes drilled in the coatings. Temperature was recorded daily. Ambient temperature during the experiments was about 25°C, with 4.1 °C difference between the two extreme values. Variation in temperatures of the monoliths was less than 2.0°C. Soil water content was followed by time-domain reflectometry (TDR), and was regulated along the soil profile by saturating the columns from the bottom through a special built-in valve connected to a hanging water-column or by sprinkler irrigation at the
Distribution of Ecologically Significant Fractions of Selected Heavy Metals
255
Table 12.1. Some Chemical and Physical Properties of the Soils Used in This Study Depth (em)
pH Horizon
H2 O
Organic Matter (g/kg) KCI
Brown forest soil, Oehrept (Godollo) 0-8 5.90 5.02 10 Ap 8-16 6.26 5.29 12 A1 16-43 B 5.85 5.01 11 a 43-66 BC 6.51 5.24 a >66 C 7.15 5.75 Slightly acidic sandy soil, Psamment (Somogysard) 4.48 5-15 5.63 13 Ap 40-50 B 6.07 4.84 3 a 80-90 C 6.38 5.56
<0.02 mm
CEC (cmol/kg soil)
SP
(%)
9.0 8.5 8.5 10.1 58.0
25 28 28 28 74
19.5 18.9 20.5 23.3 54.4
6.7 4.7 2.3
27 30 24
20.2 12.2 5.0
a No measurements.
soil surface. Irrigation water was added on the basis ofTDR measurements and/or weighting. Soil gas phase was sampled by special capillary microsensors, and concentrations of water vapor, N 2 , O 2 , and CO 2 were measured by quadrupole mass spectrometry (QMS) (Partay et aI., 1994). Supplemental light was provided in 12 hours day/night cycles. The monoliths were air-dried to constant weight, then via a bottom-valve gradually filled up to the surface with deionized water. This step was followed immediately by gravitational drainage through the bottom-valve. The aim of these procedures was to bring the soils to a fairly uniform physical status with a moisture content of maximum water capacity (pF = 0). Also, the procedure compensated the temporary shrinkage of monoliths caused by the long initial air-drying period of the field moist soils. When gas composition and soil moisture data remained constant for several days after completion of drainage, communal sewage sludge spiked with Cd-, Cr-, Ni-, Pb-, and Zn-nitrates was mixed into the upper 10 cm of the soil. Dry matter content of the compressed sludge was 20.6%, and the inorganic matter content 48.2%. Original concentrations of the selected metals in the sludge were: 12.3 mg Cd, 217 mg Cr, 109 mg Ni, 210 mg Pb, and 3026 mg Zn per kg d.m. These values are comparable to or less than the limits specified in the Hung. Techn. Dir. (1990). The metal nitrates were added to the sludge so that the final metal loading rates in the soil were equivalent to 10, 30, and 100 times the permitted loading limits (L-values) specified in the Hungarian Technical Directive assuming average sludge application practices in Hungary (i.e., 500 tlha sewage sludge containing 5% dry matter incorporated into a 20-cm surface soil layer). Loadings corresponding to 1L are 0.125 mg Cd, 8.33 mg Cr, 1.67 mg Ni, 8.33 mg Pb, and 25 mg Zn per kg soil. Original, unspiked sludge was used as control treatment at rates of 400 g sludge dry matter per column, resulting in loadings of 0.25 mg Cd, 4.48 mg Cr, 2.25 mg Ni, 4.34 mg Pb, and 62.5 mg Zn per kg soil. Identical amounts of the same sludge were used for each treatment in order to obtain as uniform conditions as possible in terms of such additional factors in the sludge as organic matter content, nutrient levels, and concentrations of other elements. One week after sludge application, nine corn (Zea may.:! L.) seeds were sown per monolith. When the plants reached a suitable developmental stage, microsensors of the QMS
256
Fate and TranspOIi of Heavy Metals in the Vadose Zone
system were implanted into the stems (one per monolith) to study the gas metabolism in vivo inside the plants. By the end of the experiment the plants had already finished their vegetation period. Mter harvesting the mature corn plants, the upper parts of the monoliths were cut into four consecutive soil layers at the 0-10, 10-15, 15-20, and 20-30 cm depth intervals. Water potentials of the soil samples were determined, as were relative root distributions. Soil pH and total potentially available metal concentrations in the soil (after 2 mol/L HN0 3 extraction as described by Andersson, 1976) were measured in air-dried soil samples, and plant-availability of the metals was estimated by: (a) metal concentrations in acid ammonium acetate-EDTA soil extracts (Lakanen and Ervi6, 1971) of air-dried soil samples, and by (b) directly plant-available concentrations in the soil solution obtained by centrifugation of the moist soil samples immediately after the cutting procedure, from each layer of the monoliths in triplicates (Csillag et al., 1995). Elemental concentrations in the 2 mol/L HN0 3 and AAAc-EDTA soil extracts, and in the soil solution, were measured by inductively coupled plasma atomic emission spectrometry (ICP-AES). Concentration values in the soil solution were related to the mass of the dry soil instead of the volume of the liquid phase, in order to eliminate the differences caused by the slightly different moisture content of the soil samples (water potential == pF 0). Recovery of the elements in the potentially available forms was related to added metal amounts. Metal budgets, based on the total potentially available amounts were calculated for each depth increments and for the whole columns. The so-called "soil available factor," SA, expressing the "percentage of the total content of an element in the soil which is available for uptake to plants" (Coughtrey et aI., 1985), is often used when describing the distribution and transport of radionuclides in terrestrial ecosystems, but it can be applied similarly to the stable forms of the elements. A similar approach was adopted to evaluate the availability of the applied metals in our experiments. Metal concentrations measured either in the plant-available soil solution fractions (cs) or in the AAAc-EDTA extracts (CAE) were related to the total potentially available metal concentrations measured in the 2 mol/L HN0 3 soil extract (c m ):
The calculated percentages (SAs and SAAE) characterize the proportion of the directly plant-available and of the supposedly plant-available metal forms, respectively, in the total potentially available metal pool of the soil. Also, the total amounts of the metals found in the liquid phase of the soil layers were calculated from the metal concentrations of the soil solution samples, for each depth increments and for the whole columns. These values were compared to the total potentially available metal contents of the columns.
NITRIC ACID EXTRACTION
The 2 moLIL nitric acW-extractahLe concentrations of all the five metals increased proportionally to the initial loading rates in the upper 10 cm layer, i.e., in the initially contaminated zone containing the metal-spiked sludge. Figure 12.1 shows linear relationships between the extracted total metal concentrations and the metal loadings (L) for the top layer in the acidic sandy soil. Similar linear relationships were obtained also for the
Distribution of Ecologically Significant Fractions of Selected Heavy Metals
257
2000
.... ....
1500
Zn
=5
.' .'
1000
II)
~ Cl
E
........
~
ID
~
~
. ..
'
........
...•..
•....
500
.... .........
.
..........
....•
Pb
O+--.......~F..;.;.;:;;.;.;....;.;.;,--"T"""-'r"'---,r--~---r-.., 0
10
20
30
40
50
60
70
80
90
100
o
10
20
30
40
50
60
70
80
90
100
I-
~
LOADING RATE (1 OL, 30L, 1OOL) Figure 12.1. Correlations between 2 maUL HN0 3-extractable soil concentrations and loading rates of the metals in the 0-10 cm layer of the slightly acidic sandy soil.
brown forest soil. The measured total metal concentrations in the top 10 cm soil layer accurately reflected at each loading level the differences among the application rates of the five metals, being the highest for Zn and the lowest for Cd, and very similar for Cr and Pb which were applied in identical amounts (Figures 12.1, 12.2a.). The differences are shown by the values of the slopes of the linear regression equations, calculated for the relationship shown in Figure 12.1, where Y is the total extracted metal concentration and X the applied metal loading (L): Cd: Cr: Ni: Pb: Zn:
Y Y Y Y Y
.= 0.076 X + 1.807 =
= = =
5.17 1.03 5.63 16.3
X X X X
+ + +
+
13.87 11.75 8.75 102.5
R2 R2 R2 R2 R2
= = = = =
0.9945 0.9984 0.9977 0.9984 0.9983
The slopes correspond closely to the ratios among the loading rates of the metals, i.e., Cd : Cr: Ni : Pb : Zn = 0.075 : 5 : 1 : 5 : 15. The brown forest soil showed no substantial increases in metal concentrations in the deeper layers, even at the higher metal application rates, with the exception ofNi and Zn which had slightly elevated concentrations in
258
Fate and Transport of Heavy Metals in the Vadose Zone
2000
a. 0-10 em depth
1500 1000 500
Cr
Ni
Pb
Zn
Pb
Zn
200
b. 10-15 em depth
150
W
100
--l
en
;:: ~ l-
50
Cr
X
W
Ni
c. 15-20 em depth 50 ]
0- CTJ:O
Cr
Ni
,rTUJ Pb
Zn
Cd
.1 = 1
a. 0-10cm
b. 10-15 cm
c. 15-20 cm
~~~~I brown forest soil • slightly acidic sandy soil Each set of bars represents the applied loadings from left to right: control(L), 10L, 30L and 100L
Figure 12.2. Metals extracted with 2 maUL HN0 3 .
the 10-15 cm and 15-20 cm depth interval at the highest contamination level. For the more acidic sandy soil we observed a somewhat more pronounced downward movement of the elements below 10 cm at the 1OOL loading rate (Figure 12.2b). Distribution of the total metal contents among the sampled layers of the upper 20 cm also shows this feature (Table 12.2). Although most of the metals (generally more than 95% of the total potentially available contents of the upper 20 cm) were found between oand 10 cm, small but significant amounts of the applied metals from the metal nitrate enrichments of the sludge were recovered in the layers between 10 and 15 cm, especially
Distribution of Ecologically Significant Fractions of Selected
Metals
--
Table 12.2. Recovery of the 2 mol/L-Extractable Metal Forms in the Top 20 cm from the Metal Nitrates Added as Enrichment of the Sewage Sludge, in the Slightly Acidic Sandy Soil
0-10 cm Loadings
Cd Cr 10 L
Ni
Pb Zn
Cd Cr 30 L
Ni
Pb Zn
Cd Cr 100 L
Ni
Pb Zn
18.8 1190 252 1240 4080 38.2 2770 548 2980 9030 145 9850 1960 10600 31300
Recovered Amount of the Metals 10-15 cm 15-20 cm Total mg/column mg/column
0.1 25.7 7.2 35.0 36.6 -1.0 31.6 5.9 48.6 76.6 3.5 266 105 291 1160
-0.1 0 2.4 2.8 4.5 -1.6 -4.5 -4.7 0.2 -14.9 -0.4 4.2 6.2 3.8 65.2
18.8 1220 261 1280 4120 35.6 2800 549 3020 9090 148 10100 2070 10900 32500
%
Added Metals mg/column
77 75 80 79 126 49 57 56 62 93 60 62 64 67 100
24.4 1630 326 1630 3260 73.3 4890 978 4890 9780 244 16300 3260 16300 32600
at the highest loading rate. Metal budgets for the brown forest soil showed similar features, with somewhat higher total recovery but less downward movement of the metals. Metals originating from the sludge might not have reached the deeper layers as it is indicated by the distribution of the total potentially available metal contents in the control treatments of both soils. Although in these treatments the total amounts of the metals, reflecting both the native metal content of the soils and the metal amounts added in the original sewage sludge, were different in the two soils; metal contents in percentage of the total content of the upper 20 cm were nearly identical, with about 60-70% of the metals found in the sludge-containing zone, and about 15-20% in each of the two deeper 5-cm-Iayers (Table 12.3). The similarity of the metal contents in the two deeper layers indicates that no metal may have moved downward from the sludge. Andersson (1976) showed that the 2 mol/L HN0 3 extraction procedure gave a good estimate of the total pollution potential of the soil from heavy metals: the method released 57% (Cr)-86% (Cd) of the total content of various heavy metals from normal, unpolluted soils, and between 65 and 92% of the total content from sewage sludge-treated soils. The differences between the extracted amounts accounted for 82-96% of the total amounts of Cd, Cr, Cu, Ni, Pb, and Zn accumulated during several years of sewage sludge application. Similar results were obtained in our laboratory (A. Lukacs, personal communication), where after 2 mol/L HN0 3 extraction the recoveries of Cr, Cu, Ni, Pb, and Zn were between 69 and 99% of the amounts measured in aqua regia extracts in a sewage sludge amended soil (BCR No. 143, 1983), but only between 21 and 77% when compared to the total metal content in an unpolluted reference soil sample (CCRMP, 1979) that contained the metals mostly in various minerals. The results suggest that nearly the total elemental content, with the exception of the most strongly fixed, residual forms are extracted from the soil by this method, which thus may be regarded as a good estimation of the total potentially available amount of the elements in the soil.
260
Fate anci Transport of Heavy Metals in the Vadose Zone
Table 12.3. Distribution of the 2 mol/l Nitric Acid-Extractable Metal Contents in the Upper 20 em of the Control Columns 0-10 em mgleolumn % Cd Cr Ni Pb Zn
25 256 292 300 1880
63.4 62.2 57.6 58.1 74.3
Cd Cr Ni Pb Zn
34 200 203 148 1440
54.2 58.6 54.9 63.6 72.2
10-15 em mg/eolumn % Brown forest soil 8 19.5 80 19.5 22.3 113 111 21.6 327 12.9
15-20 em % mgleolumn
Total mg/eolumn
7 75 102 105 326
17.1 18.3 20.1 20.3 12.8
39 412 507 516 2540
Slightly acidic sandy soil 15 23.7 14 73 21.4 68 86 23.3 80 42 18.2 42 316 15.8 240
22.1 20.0 21.8 18.2 12.0
63 341 369 233 2000
In our experiments where the soil was contaminated by a low-metal sewage sludge spiked with metal nitrates and the extraction procedure took place about 3 months after the metal contamination, relatively less of the metals was extracted by this method at the extremely high contamination levels, but recovery of the metals was remarkably uniform for Cd, Cr, Ni, and Pb. When calculating the metal budgets, it was found that at 10L loading rate in the slightly acidic sandy soil, about 80% of the applied metal-nitrate enrichment of the sludge was recovered from these four metals by the 2 mol/L nitric acid-extraction in the upper 20 cm of the monoliths (Table 12.2). Total recovery was somewhat smaller, about 50-60% at the two higher loading rates. The added Zn was practically totally recovered in the upper 20 cm soil layer at all application rates.
AAAc-EDT A EXTRACTION This extraction provided similar results to the nitric acid extraction. In the slightly acidic sandy soil the concentrations of the metals increased in the layer between 0-10 cm when the initial loading rates were higher (Figure 12.3). Very small increases (detectable only at the highest load) were found also in the layer below the application zone (Figure 12.4), that is between 10-15 cm for the more mobile Cd, Ni, and Zn, and for Pb as well. Similar results were obtained for the brown forest soil (Figure 12.3), but the AAAc-EDTA extractable amounts did not increase in the deeper layer even at the highest application rate as it is shown on the example of Zn (Figure 12.4). The ratios among the metal concentrations in the AAAc-EDTA extracts were somewhat different from the application ratios. The AAAc- EDT A extractable concentrations of Cd, Ni, and Zn reflected fairly well the original application rates. In contrast, relatively more Pb was extracted than Cr (Figure 12.3) although the initial application rates of these elements were identical. The concentration of Pb in the extracts followed well the loading rates, while the concentration of Cr was relatively smaller and decreased with increasing loading rates as compared to the applied amounts. The difference between these two elements at the lower application rates was about twofold, while at the
Distribution of Ecologically Significant Fractions of Selected Heavy Metals --------
~
-------
261
.,.
1400 :::::- 1200 ·0 III Cl
-~
1000
Cl
E
800
UJ
...J
co
600
~
400
X
200
()
r-
UJ
0
Cd
Cr
Ni
Pb
Zn
brown forest soil slightly acidic sandy soil Each set of bars represents the applied loadings from left to right: control(L), lOL, 30L and 100L
Figure 12.3. Metals extracted with AAAc-EDTA at 0-10 em depth.
highest application rate the extractability of Cr was only 1/6 that of Pb in both soils. In contrast, 2 mol/L HN0 3-extracted amounts of Cr and Pb were similar, in agreement with the application rates as was shown on Figures 12.1 and 12.2a. The similar behavior of Cd, Ni, Pb, and Zn with respect to extractability by the acid ammonium acetate + EDT A is in agreement with the stability constants of the reactions of these metals with EDTA (Me + L +--+ MeL), given in Lindsay (1979) as 16.36 for Cd, 18.52 for Ni, 17.88 for Pb, and 16.44 for Zn. Chromium forms more stable complexes with EDTA; value of the stability constant for Cr(Ill) EDTA is 24.0. However, complexes of Cr(Ill) are described as "inert" since they attain the equilibrium very slowly (Dwyer and Mellor, 1964). Such behavior might have resulted in a smaller extractability of Cr during the limited period of the extraction procedure, despite its stronger affinity to EDTA. Among the components of the AAAc-EDTA extractant, the acid and neutral ammonium acetate extract the readily soluble and exchangeable fractions of the trace elements. The addition of EDTA as a suitable chelating agent makes possible the extraction of trace elements bound by soil organic matter (Lakanen and Ervi6, 1971). However, only about 5% of the total Cr in soil was shown to be available to common extracting solutions such as acetic acid or EDT A, and the majority of Cr entering the soil is rapidly immobilized (Coughtrey and Thorne, 1983). Neutral and acidic ammonium acetate were shown to dissolve only very small fractions of Cr (Andersson, 1976). In sludge-treated soils, a great proportion of Cr was bound to Fe-Mn-oxides (Dudka and Chlopecka,
262
Fate and Transport of Heavy Metals in the Vadose Zone
1500
a. 0-10 cm depth
1200 ..-..
·0
brown forest soil
900
!/)
Cl
~
600
Cl
E
'-"'
300
- --
W
--l
a:l
~
-13 •
. -. -slightly acidic sandy soil
0 0
0
100
50
« 0:: l-
X
W
150 b. 10-15 cm depth
..
100 slightly acidic sandy soil
50
••
o
.. brown forest soil
50
100
LOADING RATE (L) Figure 12.4. AAAc-EDTA extractable amounts of Zn in the contaminated zone and in the layer below it at different loading rates.
1990) and/or was found in residual forms (Legret et al., 1988). Thus, the AAAc-EDTA mixture may not be expected to extract most of the Cr.
CONCENTRATIONS IN SOIL SOLUTION Metal concentrations in the soil's liquid phase were determined using a centrifugation sampling technique to obtain soil solution fractions available for plant uptake. The centrifugal speed was calculated by applying an equation used by Cassel and Nielsen (1986), to correspond to -1500 kPa water potential, which is the conventional value of the wilting point of plants. Plants generally cannot take up those fractions of soil water which are held in the soil more strongly than -1500 kPa; thus the soil solution fractions separated with -1500 kPa are considered to represent the liquid phase available for the plants at natural soil water contents because they are retained in the soil with suctions less than the suction corresponding to the wilting point. From water-saturated soil, for example, solutions at soil water potentials between -1500 and -0.1 kPa are separated with this technique, while from samples at field capacity, solutions between -1500 and20 kPa can be obtained. By analyzing these solution fractions, the quantity and chemical
Distribution of Ecologically Significant Fractions of Selected Heavy Metals
263
composition of the soil solution utilizable for the plant (in the sense of energy conditions) is modeled. Other methods applied widely for the separation of the soil solution do not represent soil moisture available to plants. In the case of suction methods, for example, the maximum suction that can be exerted is -100 kPa, whereas plants can exert much higher values. Displacement methods also do not give information about the quantity and energy status of the solution remaining in the soil after the extraction. It is also probable that the extreme high pressures applied to the moist soil samples in a hydraulic pressure apparatus during the extraction of the soil solution, disturb the prevailing equilibrium between the soil phases. Metal concentrations in the soil solution were generally several orders of magnitude lower than the 2 mollL HN0 3 extractable total concentrations, with the exceptions of Cd, Ni, and Zn in the sandy soil and of Ni and Zn in the brown forest soil at the highest loading rate (Table 12.4). Notice that in Table 12.4 we used the more convenient unit of Jlg/kg for the soil solution concentration, rather than mg/kg for the total concentrations in Figures 12.1 and 12.2. Chromium entered the liquid phase of the soils in negligible amounts, even at the highest metal loading. Lead showed similar low soil solution concentrations as did Cr in the brown forest soil, but in the sandy soil Pb concentrations were significantly higher than Cr concentrations, for all loading rates. Compared to loading rates, release of Cd, Ni, and Zn into the soil solution was much higher than of Cr and Pb, and increased substantially at the higher metal application rates in the top 10 cm. Similar increases also occurred in the originally uncontaminated soil layers directly below the application zone. Concentrations of Cd, Ni and Zn in the soil solution were several orders of magnitude higher at the 30 Land 100 L loading rates than in the control treatment, in the top 10 cm. Of these three metals, Cd was found in much lower concentrations in the soil solution. The observed low Cd concentrations are consistent with the relatively low application rate of this metal as compared to those for the other elements. Release of the metals into the soil solution at all metal application rates was significantly higher in the slightly acidic sandy soil than in the brown forest soil (Table 12.4). It is difficult to compare our data, which hold for soil solution concentrations in heavily contaminated soils, with the highly variable literature data obtained under different experimental conditions and using a wide array of methods. Still, our results show similar tendencies to those cited by Kabata-Pendias and Pendias (1992) or by Kabata-Pendias and Adriano (1995); i.e., the relatively mobile metals, Cd, Ni, and Zn occur in a relatively larger proportion in the solution phase than the less mobile Cr and Pb. These findings cited were obtained for natural soil solutions separated by centrifugation from different soils. Soil solution concentrations expressed as percentage of the total potentially available concentrations (SAs) indicated very low availability of Pb and especially of Cr in those fractions of the soil's liquid phase which are directly accessible for plant uptake, since only negligible amounts of these elements were released into the soil solution (Table 12.5). In contrast, Cd, Ni, and Zn were more readily available for uptake in both soils, and the proportion of the directly plant-available amounts of these metals increased sharply at the highest application rate (Figure 12.6); they were more than one order of magnitude higher than in the control treatment. Such increases were observed not only for the
264
Fate and Transport of Heavy Metals in the Vadose Zone Table 12.4a. Concentrations of the Metals in the Soil Solution in Brown Forest Soil a (~g/kg dry soil)
Element
Cd Cr Ni Pb Zn a b
Depth (cm)
0-10 10-15 15-20 0-10 10-15 15-20 0-10 10-15 15-20 0-10 10-15 15-20 0-10 10-15 15-20
Loading Control
10 L
b
0.4
30L
1.2
b
b
b
b
b
b
1.9 0.6 b
4.3 2.1 5.6 b b b
50.8 81.0 130
1.4 3.0 2.4 12.9 4.9 3.5 5.9 3.3 6.5 120 82.6 140
0.7 b
3.7 39.2 6.8 11.3 b
6.1 2.8 219 127 125
100 L
90.2 11.6 b
7.0 2.3 b
2600 225 13.7 7.8 11.5 b
28600 761 232
Detection limits in mg/L: Cd - 0.005, Cr - 0.005, Pb - 0.05. Concentrations below detection limit.
Table 12.4b. Concentrations of the Metals in the Soil Solution in Slightly Acidic Sandy Soil (~g/kg dry soil)
Element
Cd Cr Ni Pb Zn
Loading
Depth (cm)
Control
0-10 10-15 15-20 0-10 10-15 15-20 0-10 10-15 15-20 0-10 10-15 15-20 0-10 10-15 15-20
2.1 1.7 0.7 0.4 3.0 3.2 24.5 27.0 17.0 18.2 20.1 34.3 283 395 184
10 L
2.0 1.4 1.0 2.9 17.4 2.9 49.3 23.6 12.8 10.4 30.1 14.0 500 273 78
30L
100L
16.5 2.2 1.1 7.2 4.8 1.7 375 39.0 22.5 14.1 16.9 9.7 3492 342 168
357 13.2 1.2 10.7 10.8 6.0 9197 623 29.4 154 33.6 28.1 106200 6249 266
application zone of the metal-enriched sludge, but also for the originally uncontaminated 10-15 cmdepth interval (Figure 12.6). The breakthrough-like increases in the relative amounts of the directly phytoavailable metal forms indicate a decrease in the metal-buffering capacity of the soil at this extreme metal application rate. The calcula-
Distribution of Ecologically Significant Fractions of Selected Heavy Metals - -- -
""-,-""-'-'"--'---.
-
-" ---
-~---
-------_ ...
265
"-""----"
Table 12.5. Recovery of the Metals in the Liquid Phase of the Slightly Acidic Sandy Soil per Thousand of the Total Potentially Available Amounts
Amount of the Metals Recovered in the Liquid Phase 0-lOcm 10-15 cm 15-20 cm Total mg/column mg/column %0
Loadings
Total Potentially Available Amounts mg/column
Brown forest soil
Control
100 L
Control
100 L
a
Cd Cr Ni Pb Zn Cd Cr Ni Pb Zn Cd Cr Ni Pb Zn Cd Cr Ni Pb Zn
a
a
a
0.04 0.08
0.Q1 0.02
a
0.05
0.05 0.15
a
a
a
a
a
0.96 1.70 0.13 49.0 0.15 539 0.04 0.01 0.46 0.34 5.33 6.74 a
173 2.90 2000
0.76 1.23 2.95 0.11 1.81 0.02 0.15 2.12 0.13 51.3 0.11 0.26 7.17 2.19 548 Slightly acidic sandy soil 0.Q1 0.07 0.02 0.03 0.03 0.07 0.25 0.87 0.16 0.19 0.32 0.85 3.73 1.73 10.8 0.12 0.01 6.87 0.10 0.06 0.16 5.88 0.28 179 0.32 0.26 3.48 58.9 2.51 2060
0 0.1 0.3 0 1.2 8.1 0.01 16.3 0.02 13.6
40 411 507 516 2530 224 13100 3140 14100 40200
1.1 0.2 2.4 3.7 5.4 33 0.02 73 0.3 60
63 341 369 232 2000 211 10400 2440 11200 34500
Values below detection limit (see Table 12.4a).
tions revealed a higher availability and mobility of the elements in the liquid phase of the slightly acidic sandy soil as compared to the brown forest soil (Table 12.4, Figure 12.6). Such differences are shown not only by the metal concentrations but also by the distribution of the total amounts of the metals in the soil's liquid phase. Data for the control treatment and for the highest contamination level are shown in Table 12.5. While distribution of the metals among the liquid phases of the three sampled soil layers was fairly uniform in the control treatments, at the highest contamination level most of the metals were found in the top 10 cm, with slight increases of the Ni and Zn amounts between 10 and 15 cm. The soil pH in our experiments decreased somewhat as the metal loading rate increased (Figure 12.5). This decrease is likely caused by the acidity of the nitrate salts being used in our study. The lower pH may have contributed to the increased mobility and availability of the metals at the higher loadings. Also, the somewhat more acidic character of the sandy soil may have been responsible for the higher availability of the metals in this soil. This explanation is in agreement with literature data about the effect of acidification on the availability and mobility of heavy metals (LJzsbersli et aI., 1991; Marschner, 1991). The influence of soil pH on the mobility of trace metals depends also
266
Fate and Transport of Heavy Metals in the Vadose Zone 7.5 7 brown forest soil
~ I
6.5
I
c.
slightly acidic sandy soil
............ .....- ... - ........................ . .
6
·"·-"'--" .. e
5.5 0
20
40
60
80
100
LOADING RATE (L) Figure 12.5. Soil pH in the 0-1 0 em layer at the end of the experiment.
upon the geochemical properties of the metal: for example, mobility of Cd has been classified as medium up to pH 6, and that of Ni and Zn up to pH 5, while Cr and Pb have only weak mobility above pH 4.5 and 4, respectively (Kabata-Pendias and Pendias, 1992). Thus, the differences among mobilities of various metals may have been influenced also by the pH of the soil liquid phase. Amounts of the elements entering the soil solution seemed to be rather low in comparison to literature data. For instance, the highest value of Zn in our experiments was only 5.8% of the total potentially available amounts, while values up to 50% have been reported in the literature (Coughtrey et aI., 1985). One likely reason for this disparity is that literature data are generally based on extraction procedures that use chemical desorption and wide soil:extractant ratios. Such methods should give much higher elemental concentrations than those which occur in the soil liquid phase at natural field soil water contents. The proportion of the phytoavailable metal fractions was by about one order of magnitude higher when the calculations were based on the AAAc- EDTA extractable metal concentrations (SAAE) as compared to the phytoavailabilities calculated from the soil solution concentrations (SAs). In an extensive study comparing the availabilities of the elements in the Hungarian soils (1013 samples), using different extractants, the AAAc+ EDT A-available amounts of the metals included in our study were between 39 and 91 % of their 0.5 M HN0 3-extractable amounts (Marth, 1990).
MOVEMENT While SAs of Cd, Ni, and Zn showed a breakthrough-like phenomenon at the highest loading rate (Figure 12.6), especially in the slightly acidic sandy soil, values of SAAE increased linearly or to a smaller extent when the metal loading rates increased. A further difference between the results of the two methods used to estimate the phytoavailability of the metals was that while the directly available amounts of Cr and Pb in the soil solution were similar to each other (in agreement with the similar application rates), the AAAc-EDTA-extractable amounts of these elements were much differ-
Distribution of Ecologically Significant Fractions of Selected
%
HP;nfU
Metals
267
a. 0-10 em depth
8 6 4 2
0
Cd
%
~1
Zn
Ni
b. 10-15 em depth
0
o
1,-1]·1
Cd
Ni
Zn
Cd
Ni
Zn
%
0:] brown forest soil slightly acidic sandy soil Each set of bars represents the applied loadings from left to right: control(L), lOL, 30L and 100L Y axis: % of total potentially available concentration (SA s = 100 Cs I Cm , where Cs is the metal concentration measured in the plantavailable soil solution fraction, Cm is the total potentially available metal concentration measured in the 2 mollL HND3 soil extract)
Figure 12.6. Soil availability factors (SAs) of the metals.
ent. The availability of added Pb, as estimated by the acidic ammonium acetate-EDTA extraction, was comparable to the availability of Zn and Cd, and somewhat exceeded that ofNi in both soils (Figure 12.7). In the brown forest soil it increased only slightly at the higher loading rates, while in the slightly acidic sandy soil it followed more closely the increasing metal pollution levels. Again, in this respect its behavior was similar to those of Cd, Ni, and Zn. Metal budgets calculated for the liquid phase of the upper 20 cm also show an increased proportion of the directly plant-available forms of Cd, Ni, and Zn in per thousand of their total potentially available amounts in these layers, at highly elevated metal application rates (Table 12.5). Literature data on the possible leaching of the elements show only small downward movement of the heavy metals, and mostly in coarse-textured, sandy, or gravelly soils. However, in many cases the applied metal concentrations
268
Fate and Transport of Heavy Metals in the Vadose Zone
% 80 70 60 50 40 30 20 10 0
Cd
Cr
Ni
Pb
Zn
brown forest soil slightly acidic sandy soil Each set of bars represents the applied loadings from left to right: control(L), IOL, 30L and IOOL y axis: % of total potentially available concentration (SAAE = 100 cAE/ Cm, where CAE is the metal concentration measured in in the AAAc-EDTA extracts, Cm is the total potentially available metal concentration measured in the 2 moVL HN03 soil extract)
Figure 12.7. Soil availability factors for the AAAe-EDTA extractions (SAAEl at 0-10 em depth.
were much smaller than the provocative overloading of 100 L in our experiments. Careful evaluation of some data interpreted as being proof of no metal movement reveals such a slight movement of the metals, which is in the range expectable from the relatively low metal application rates, and which are sometimes interpreted as resulting from inadvertent mixing of the soil layers (e.g., in the papers summarized in Dowdy and Yolk, 1983). Downward movement of the metals involves the leaching of the water-soluble forms. The movement of toxic metals was studied most often in the percolate water of lysimeter experiments. Our experiments, conducted under natural soil moisture conditions, with no addition of extra water that might have caused leaching (as shown by the TDR measurements along the soil profile), showed that only a very small percentage of the polluting metals may enter the soil solution at field soil moisture contents, even at provocative overloadings. Thus, the expectable concentrations of the metals in the liquid phase of the deeper layers are small. Such small values are regarded often as proofs of no substantial movement. Although in many experiments no metal movement was found in the deeper layers below 60 cm, in some instances significant increases were reported. In a sludge-amended clay loam soil under a forest vegetation and with a surface pH of 5, about 3% of the applied Zn and 4-7% of the applied Cd was leached beyond 120 cm (cit. in Dowdy and Yolk, 1983). In a lysimeter experiment on a sandy loam contaminated with sewage sludge containing large amounts of Cd, Cr, Ni, Pb, and Zn, the mobility of these elements, measured as their concentration in the percolate water, increased two times in sludgeamended soil (Dudka and Chlopecka, 1990). Since in this chapter various metal inputs
Distribution of Ecologically Significant Fractions of Selected Heavy Metals
269
and outputs (g/ha) are also presented, the amount of the leached, water-soluble forms in percentage of the total input can be calculated. These values are 5.4% for Cd, 0.003% for Cr, and 0.17% for Ni and Zn. The Cd and Cr data are comparable to the relative availabilities of these metals in the soil solution in our experiments (Figure 12.6).
SUMMARY The biological and ecological effects of heavy metal pollution of soils depend not only on the total amount of the contaminating metals, but to a great extent on their biologically and ecologically active, easily soluble and mobile fractions. The approach discussed in this chapter focused on determining those fractions of the soil's metal content which are directly and easily available for plant uptake (using two different methods to characterize the phytoavailability), or which are potentially becoming available for the plants during longer periods. Ratios of the directly available and potentially available metal concentrations were used to estimate the bioavailability of the elements in soil columns contaminated with sewage sludge spiked with (Cd + Cr + Ni + Pb + Zn)-nitrates. The total potentially available amounts of the metals reflected well the application rates, and did not increase or only slightly increased in the soil layers lying directly below the application zone. Proportion of the directly plant-available amounts of Cd, Ni, and Zn (measured in the soil solution considered as directly available for the plants) increased sharply at the highest application rate, and such increases were observed not only in the contaminated layer but also in the originally uncontaminated 10-15 cm depth interval. We want to stress the ecological significance of the movement of these watersoluble forms. Since the solubility of the heavy metals in soils has great significance in their bioavailability and their migration (Kabata-Pendias and Pendias, 1992), relatively small increases in solution metal concentrations may have an impact on the environment. The breakthrough-like increases in the relative amounts of the directly phytoavailable metal forms might also indicate a decrease in the metal-buffering capacity of the soil at this extreme metal application rate. In contrast, Cr and Pb entered the liquid phase (i.e., were directly available for the plants) in negligible amounts, even at the provocative overloadings. This is in agreement with the majority of the literature data about the transport of these elements into the plants. The AAAc + EDTA extraction which is a method used to estimate the plant-availabilities of both macro- and micronutrients simultaneously, after a single extraction procedure, and which was shown to be appropriate also for several toxic elements, indicated much greater availability of the metals than was estimated from the metal concentrations measured in the soil's liquid phase. Estimated availability (i.e., extractability) of Cr was less and that of Pb was higher than expected on the basis of the application rates. In the AAAc + EDTA extraction procedure Pb behaved similarly to Cd, Ni, and Zn, which are regarded as mobile elements in the soil. Thus, the two methods used to assess the plant availability of the selected metals gave different estimations of the phytoavailable proportions. Comparison of the estimated phytoavailabilities with the actual metal uptake by plants should give support in favor of one of these methods. However, further studies have to consider that plant concentrations measured at a specific point during the vegetation period reflect the cumulative uptake of the elements until that time, while soil solution concentrations pertain only to
270
Fate and Transport of Heavy Metals in the Vadose Zone
a specific situation (soil water content, temperature, etc.) at a specific moment and soil extraction techniques cannot take into account the plant physiological processes.
ACKNOWLEDGMENTS Supported by HNSRF, grant numbers T23221 and T23360.
REFERENCES Alloway, B.J. Heavy Metal! in Soil!. Blackie and J. Wiley and Sons, Glasgow, 1990. Andersson, A. On the determination of ecologically significant fractions of some heavy metals in soils. SweoiJh J. Agric. &1. 6, pp. 19-25, 1976. BCR No. 143. The certification of the contents of cadmium, copper, mercury, nickel, lead and zinc in a sewage sludge amended soil. Report. Community Bureau of Reference - BCR. Brussels, 1983. Bujtas, K, J. Csillag, G. Partay, and A. Lukacs. Distribution of Selected Metals in a Soil-Plant Experimental System Mter Application of Metal-Spiked Sewage Sludge, in Proc. XXVth AnnuaL Meeting of ESNA, M.H. Gerzabek, Ed., Castelnuovo Fogliani (PiacenzaiItaly), Seibersdorf: Osterreichisches Forschungszentrum Ges. m.b.H., Austria, 1995, pp. 99-105. Campbell, D.J. and P.H.T. Beckett. The soil solution in a soil treated with digested sewage sludge. J. SoiL Sci. 39, pp. 283-298, 1988. Cassel, D.K and D.R. Nielsen. Field Capacity and Available Water Capacity, in Methood of Soil AnaLYdiJ, Part I, 2nd ed., A. Klute, Ed., American Society of Agronomy, Madison, WI, 1986, pp. 913-915. CCRMP (Canadian Certified Reference Materials Project). Certificate of Analysis. Reference Soil Sample SO-4. CANMET Report 79-3. Ottawa, Canada, 1979. Chang, A.C., J.E. Warneke, A.L. Page, and L.J. Lund. Accumulation of heavy metals in sewage sludge-treated soils. J. Environ. QuaL. 13, pp. 87-91, 1984. Chang, A.C., T.C. Granato, and A.L. Page. A methodology for establishing phytotoxicity criteria for chromium, copper, nickeL and zinc in agricultural land application of municipal sewage sludge. J. Environ. QuaL. 21, pp. 521-536, 1992. Coughtrey, P.J. and M.C. Thorne. RaownucLwe DiJtrilllltwn ano Trandport in TerrutriaL ano Aquatic &odYdtefnd. A CriticaL Review of Data. Vol. 2., A.A. Balkema, Rotterdam, 1983. Coughtrey, P.J., D. Jackson, and M.C. Thorne. RaownucLweDiJtrwutwn anO Trandport in TerredtriaL ano Aquatic &odYdtefnd. A Compenoium 0/ Data. Vol. 6. A.A. Balkema, Rotterdam, Boston, 1985. Csillag, J., T. T6th, and M. Redly. Relationships between soil solution composition and soil water content of Hungarian salt-affected soils. Arw SoiL &1. RehabiLitatwn 9, pp. 245-260, 1995. Dowdy, R. H. and V. V. Volk. Movement of Heavy Metals, in Proc. Symp. on ChemicaL MObility anO Reactivity in SoiL SYdtefnd, SSSA Spec. PuM. No. n., Atlanta, 1981, D.W. Nelson et al., Eds., pp. 229-240, 1983. Dudka,S. and A. Chlopecka. Effect of solid-phase speciation on metal mobility and phytoavailability in sludge-amended soil. Water Air SoiL PoLLut. 51, pp. 153-160, 1990. Dwyer, F. P. and D.P. Mellor, Eds. Chelating Agentd anO MetaL Chelated. Academic Press, N ew York and London, 1964. Homeyer, B., KO. Labenski, B. Meyer, and A. Thormann. Herstellung von Lysimetern mit Boden in naturlicher Lagerung (Monolith-Lysimeter) als Durchlauf-, Unterdruck-oder Grundwasserlysimeter. Z Pjlanzenernahr. Booenko. 136, pp. 242-245, 1973. Hungarian Technical Directive. Land and Forest Applications of Waste Waters and Sewage Sludges. MI-08-1735-1990 (in Hungarian).
Distribution of Ecologically Significant Fractions of Selected Heavy Metals
271
Jones, D.L. and A.C. Edwards. Effect of moisture content and preparation technique on the composition of soil solution obtained by centrifugation. Commun. SoiL Sci. Plant AnaL. 24, pp. 171-186, 1993. Juste, C. and M. Mench. Long-Term Application of Sewage Sludge and Its Effects on Metal Uptake by Crops, in BiogeochemutryofTraceMetau, D.C. Adriano, Ed., Lewis Publishers, Boca Raton, FL, 1992, pp. 159-193. Kabata- Pendias, A. and D.C. Adriano. Trace Metals, in SoiLAmendment,} and EnvironmentaL QuaLity, J.E. Rechcigl, Ed., CRC Press/Lewis Publishers, Boca Raton, FL, 1995, pp. 139-167. Kabata-Pendias, A. and H. Pendias. Trace Element,} in SoiU and Plant,}, 2nd ed. CRC Press, Boca Raton, FL, 1992. Lakanen, E. and R. Ervi6. A comparison of eight extractants for the determination of plant available micronutrients in soils. Acta Agr. Fenn. 123, pp. 223-232, 1971. Legret, M., L. Divet, and C. Juste. Migration et speciation des metaux lourd dans un sol soumis a des epandages de boues de station d'epuration a tres forte charge en Cd et Ni. Wat. RiAf. 22, pp. 953-959, 1988. Lindsay, W.L. ChemicaL EquiLibria in Soiu. John Wiley & Sons, New York, 1979. ~bersli, E., E. Gjengedal, and E. Steinnes. Impact of Soil Acidification on the Mobility of Metals in the Soil-Plant System, in Heavy Metau in the Environment. Trace Metau in the Environment. i., J.P. Vernet, Ed., Elsevier, Amsterdam, 1991, pp. 37-53. Marschner, H. Plant-Soil Relationships: Acquisition of Mineral Nutrients by Roots from Soils, in Plant Growth: Interactiond with Nutrition and Environment, J.R. Porter and D.W. Lawlor, Eds., Cambridge University Press, 1991, pp. 125-155. Marth, P. Comparative Study on Soil Extractants. Postgraduate thesis, Agricultural University of G6d6ll6, 1990, (manuscript in Hungarian). McBride, M.B. Toxic metal accumulation from agricultural use of sludge: Are USEPA regulations protective? J. Environ. QuaL. 24, pp. 5-18, 1995. McGrath, S.P., A.C. Chang, A.L. Page, and E. Witter. Land application of sewage sludge: Scientific perspectives of heavy metal loading limits in Europe and the United States. Env. Review,} 2, pp. 1-11, 1994. Molnar, E., T. Nemeth, and o. Palmai. Problems of Heavy Metal Pollution in Hungary- "Stateof-the-Art," in Heavy Metau. Pr06LemJ and SoLutiond, W. Salomons et al., Eds., Springer, Berlin, 1995, pp. 323--344. Mullins, G.L. and L.E. Sommers. Characterization of cadmium and zinc in four soils treated with sewage sludge, J. Environ. QuaL. 15, pp. 382-387, 1986. Nemeth, T., G. Partay, I. Buzas, and H. Gy. Mihalyne. Preparation of undisturbed soil monoliths. Agrolcemia iJ Talajtan 40, pp. 236-242, 1991, (in Hungarian). Nemeth, T., G. Partay, K. Bujtas, and A. Lukacs. Application of Quadrupole Mass Spectrometry to Assess Effects of Sewage Sludge on Gas Composition in Undisturbed Soil Columns, in Biogeochemutry of Trace Element,}, D.C. Adriano, Ch. Zueng-Sang, and Y. Shang-Shyng, Eds., Science and Technology Letters, Environ. Geochem. HeaLth 16, pp. 141-151, 1994. Partay G., A. Lukacs, and T. Nemeth. Soil monolith studies with heavy-metal containing sewage sludge. Agrolcemia iJ Talajtan 43, pp. 211-221, 1994. Petruzzelli, G. Recycling wastes in agriculture: Heavy metal bioavailability. Agric. Eco,}Y,}temJ Environ. 27, pp. 493-503, 1989. Sillanpaa, M. Micronutrients and the Nutrient Status of Soils. FAO Soiu BuLletin 48, 1982. Sillanpaa, M. and H. Jansson. Status of cadmium, lead, cobalt and selenium in soils and plants of thirty countries. FAO SoiU BuLletin 65, 1992. USEPA (U.S. Environmental Protection Agency). Standards for the disposal of sewage sludge; Proposed Rules 40 CFR Parts 257 and 503. Fed. Regut. 54, pp. 5746-5902, 1989. Zabowski, D. and F.C. Ugolini. Lysimeter and centrifuge soil solutions: Seasonal differences between methods. SoiL Sci. Soc. Am. J. 54, pp. 1130-1135, 1990.
CHAPTER 13
Heavy Metal Contamination in Industrial Areas and Old Deserted Sites: Investigation, Monitoring, Evaluation, and Remedial Concepts Irena Twardowska, S. Schulte-Hostede, and Antonius A.F. Kettrup
INTRODUCTION The historical development of industrialization in Germany and Middle European countries based on mining and metallurgy resulted in the creation of thickly populated "hot spots" impacted by high and long-term emission of heavy metals to the environment. One of the major sources of non point contamination of soils and water by heavy metals is dry and wet deposition from industrial stacks. To date, despite the attention of governmental agencies, industry, and environmentalists, as well as of public concern focused on this source, the understanding of the dimension of the human risk potential from soil contamination originating from long-term stack emission is still limited. Due to the long-term effect and frequently occurring unfavorable transformations of anthropogenic metal-bearing waste properties in time, old abandoned industrial sites may display high contamination potential in a postclosure period. Because of the lack of adequate knowledge, environmental protection was not considered when these facilities were sited and operated. The characteristics and utilization of contaminated sites, kind of emission, character of environmental impact and target-oriented protection objectives determine different criteria and methods of site investigation, evaluation, and remedial concepts. The objective of this chapter is to present a different, site-specific approach to the evaluation of heavy metal contamination in areas oflong-term anthropopression, with a special regard to evaluation of mobile/mobilizable forms of heavy metals in soil and the vadose zone matrix. Methods of evaluating human risk potential and selection of appropriate remedial options for the site-specific cases, comprising technical measures and the use of adequate extractants or adsorbents were also discussed.
274
Fate and Transport of Heavy Metals in the Vadose Zone
The presented approach was exemplified in case studies on three sites impacted by emissions of different character and potential: long-term stack emission, and metal emission from a large-area deserted industrial site. Localization of the surveyed sites is presented in Figure 13.l. The scope of applied investigation procedures was highly site-specific, depending on the kind of heavy metal emission, extent of risk, and endangered objectives to be protecyd (e.g., population, agricultural area, groundwater resources). To evaluate actual e¥ent of anthropogenic impact and pathways of investigated heavy metals in the envirdnment, a broad range of preliminary studies and investigations had to be undertaken in each case. The general scope of the preliminary studies consisted of collecting available archival data required for the standard procedure of environmental impact assessment (EIA), and next filling the information gaps and uncertainties with new investigations, with particular regard to the aim of the studies to be conducted in each site. As a part of a comprehensive contamination assessment and risk evaluation in the investigated sites, the following parameters were considered: site localization, area, population, land use, natural conditions (morphology, meteorology, hydrology, hydrogeology), characteristics of soils and the vadose zone matrices as well as the source of heavy metal emission (type, scope, extent, duration).
IMPACT OF LONG-TERM STACK EMISSION The investigation and assessment procedures for the areas impacted by the long-term metal emission from stacks were illustrated in case studies on soil contamination in the areas of Sendzimir steelworks in Poland (Site I) and Irena Glasswork (Site II). In both sites, surface soil and the upper part of the vadose zone profile (subsoil) enrichment with heavy metals from dry and wet deposition of heavy metals from long-term industrial stack emissions was the main objective of the presented studies.
Site Characteristics Site I: Nowa Huta nlCracow, Area Adjacent to the Sendzimir Steelwork Complex, Po/and (Figures 13.1, 13.2A) One of the 27 "hot spot" areas in Poland, classified as ecological emergency areas (Chief Statistical Office, 1995) is the Cracow district, mainly due to the extent of atmospheric pollution. Some industrial plants are subject to extreme controversy due to their unfortunate location. One of these plants, the 40-year-old Sendzimir Steelworks, is located in the highly developed agricultural area near Cracow, a city of great architectural and historical value. The emission of particulate and gases from Sendzimir steelworks have been considered to be one of the major sources of contamination by heavy metals and other pollutants in Cracow. The objective of the study was to assess the steelworks as a source of soil pollution by means of evaluating the concentrations of heavy metals in the soil in its vicinity.
Morphology Vadose Zone Matrix and Soil Characteristics The site of the total area 6219.3 ha (Figure 13.2A) represents an area of long-term anthropogenic impact. Relief is typical for loess areas and consists of a number of table-
Heavy Metal Contamination in Industrial Areas and Old Deserted Sites
.----....
275
-~\
POLAND
\ \
\
I
(""
-., I \
I
\
Site I \ Nowa Huta n/Cracow ) Sendzimir Steelwork;' j
9-/
..r
(
·v....~-,\
Figure 13.1. Location of surveyed sites I, II (Poland) and III (FRG): Site I: Nowa Huta n/Cracow, area adjacent to Sendzimir steelworks, Poland; Site II: area adjacent to Irena Glasswork n/lnowroclaw, Poland; Site III: Marktredwitz urban area with chemical plant site, North Bavaria, FRG.
lands, crosscut by valleys of small streams. The soil and the vadose zone matrix comprise Quaternary loess sediments from Cracow glaciation, 8-12 m thick. Black earth, mostly degraded, in a minor rate diluvial, accounts for 65.7% of soils; 27.7% of the area are brown soils. The residual 6.6% are mainly fen soils (5.1 %). These are almost entirely rich heavy soils (96.2%), composed predominantly of loess, with some clay-like loess formations, of high or medium humus content. The pH value is mainly neutral, except for soils located SE, S, and NW of the site, which are acidic (Table 13.1). The utilization of the land shows its agricultural characteristics and is comprised of: arable land, mainly horticultural, 83.9%; grassland, 10.2%; forests, 0.5%; disused lands, 3.2%; railroad area, 1.9%; compact settlements, 0.3% of the total area.
SOUfce of Anthropogenic Contamination of Soil by Metals Since the late fifties, the area has been affected by the emission from Sendzimir Steelworks complex sited at its SW border, in the direction of dominating winds, blowing from W, NW, and SW (Figure 13.2B-I). The Sendzimir Steelworks is the major source of particulate and gaseous emission in the area of the surveyed site and Nowa Huta satellite district of Cracow. It is considered to be also one of the major sources of air contamination with dust and heavy metals in the historic downtown of Cracow. According to air monitoring data (Chief Statistical Office, 1993; 1995) average mean annual wet and dry particulate deposition in 1994 in Nowa Huta area ranged from 68 to 159 mg km-2 a-I, mean 117 mg km -2 a-I. In Cracow downtown the particulate deposition ranged from 41 to 105 mg km-2 a-I, mean 70 mg km-2 a-I. According to the same source, the average annual concentration of trace metals in air in both districts represented an or-
276
Fate and Transport of Heavy Metals in the Vadose Zone
Figure 13.2A-C. Site I: Nowa Huta n/Cracow, area adjacent to the Sendzimir Steelworks, Poland. (A) General map. Soil sampling paints {op}, soil profiles (* P) and spatial distribution of heavy metals (mg kg-') in surface soil: (B) Fe, (C) Mn. Kriging estimates: linear model, angle 15.
Heavy Metal Contamination in Industrial Areas and Old Deserted Sites
277
Zn P60
•
o
1'61
•
4
P70
•
. . . .
P69
P42
•
6
8
P41
10
P23
P18
12
...........
'"
P10 .~
.
14
P8
•
'"
km
km~E='~--------An~~--~--------~~~------------mrr---------, Pb
10
12
14
km
Figure 13.2D-F. Site I: Nowa Huta n/Cracow, area adjacent to the Sendzimir steelworks, Poland. Soil sampling points (·P), soil profiles (OP) and spatial distribution of heavy metals (mg kg- 1) in surface soil: (D) Zn, (E) Pb, (F) Cr. Kriging estimates: linear model, angle 15.
278
Fate and Transport of Heavy Metals in the Vadose Zone
~~----------~~~--~----------~~--------------~.----------,
6
8
10
12
14
10
12
14
km
o
km~----------~7b----------------~~~------------~r----------'
o
6
8
10
12
14
Figure 13.2G-1. Site I: Nowa Huta n/Cracow, area adjacent to the Sendzimir steelworks, Poland. Soil sampling points (·P), soil profiles (* P) and spatial distribution of heavy metals (mg kg- 1) in surface soil: (G) Ni, (H) Cu, (I) Cd. Kriging estimates: linear model, angle 15.
Heavy Metal Contamination in Industrial Areas and Old Deserted Sites
279
Table 13.1. Physicochemical Properties of Soils in Site I (Nowa Huta n/Cracow, Area Adjacent to Sendzimir Steelworks, Poland) Mean Range
Percentage of the Total Area of Soils (%)
44 33-58
0-10 very light 0
6.3 3.6-7.8
<4.5 very acid 7.6
2.76 1.21-8.44
2.06 0.52-6.98
86.0 34.8-98.6
Content of fraction <0.02 J.lm 11-20 21-35 mean 3.8
light 0
pH value in 1M KCI 4.6-5.5 acid 22.8
5.6-6.5 slightly acidic 15.2
Humus contents (%) 1.01-2.00 2.01-3.50 <1.00 low mean high 0 16.5 77.2 Hydrolytic acidity (Hh, ceq kg-' of soil, dry weight) 1.51-3.00 3.01-4.50 <1.5 low mean high 58.2 17.7 11.4 Base saturation of soils (V) (%) 26-50 51-75 <25 very low low mean 0 1.3 15.2
>35 heavy 96.2 6.6-7.2 neutral 44.3
>7.2 alkaline 10.1
>3.51 very high 6.3 >4.51 very high 12.7 >75 high 83.5
der: Zn > Pb » Cu > Mn » Ni > Cr > Cd (Table 13.2). In downtown, concentrations of Zn, Mn, and Cd appeared to be higher than in Nowa Huta close to the Sendzimir Steelworks, which can be explained by the contribution of other sources of air pollution. The direct measurements of wet and dry deposition and the concentration of Zn, Pb, and Cd in the settled dust in the area taken during 1992 (preceding the survey of the soils for trace metals), recorded dust deposition in the sampling point 1.4 km SW from the Steelworks to be 2 to 4 times higher than in other parts of Cracow (Table 13.3). This proves Steelwork to be a serious source of air contamination in the Cracow area. Concentrations of Zn and Pb in the deposition in Site I were comparable with those recorded in western suburbs and the city center of Cracow, downtown being the highest. These data are in a good agreement with the routine ai~onitoring data that were cited earlier (Chief Statistical Office, 1993, 1995). It should e emphasized that despite similar concentrations, the load of metals in Site I adjacent to endzimir Steelworks is the highest, due to high dust deposition (Table 13.3). Considering that just one defined strong source of non point contamination occurs in the area (high and low emission from the Steelworks), the site was selected as appropriate for the evaluation of the contamination potential of emission from the steelworks. Objectives of a particular interest included: barrier capacity of soils, forms of binding, as well as accumulation and mobility of metals of anthropogenic origin in the vadose zone compared to the natural background. In accordance with the kind of the emission, the metals of concern were Fe, Mn, Zn, Pb, Ni, Cu, Cr, and Cd (Figure 13.2B,C,D,E.F,G,H,I). This order of metals aligned with their descending concentration ranges in soils of Site 1. In Figure 13.2B-I, the
280
Fate and Transport of Heavy Metals in the Vadose Zone
Table 13.2. Mean Annual Particulate Deposition (mg km- 2 a-') and Concentration of Metals (ng m- 3) in Nowa Huta (Sendzimir Steelworks Area) and Cracow Downtown a
Contaminant Wet and dry particulate deposition (mg km-2 a-')
Numberb n N 12
144
Metal concentrations (ng m-3) Zinc, Zn 1 Lead, Pb 1 Cadmium, Cd 1 Copper, Cu 1 Manganese, Mn 1 Nickel, Ni 1 Chromium, Cr 1 a b
12
Sampling Point. Direction and Distance from the Steelworks podg6rze, Nowa Huta, Cracow Downtown, SW10km W 1.0 km Wl0km 1994 1992 1994 1992 1994 1992 106
117
123
540
325 138 2 30 16 14 6
340 160 0.0 50 20 10 0.0
12 12 12 12 12 12
0.0 30 30 10 0.0
70
77
81
542 98 3 25 21 8 4
310 160 0.0 20 20 10 0.0
363 146 2 28 20 4 4
Chief Statistical Office, 1993, 1995. N - number of sampling points; n - number of measurements.
Table 13.3. Wet and Dry Deposition of Dust in Nowa Huta Area (Site I) Compared to Other Districts of Cracow (1992)
Contaminant
Unit
Sampling Points, Directions, and Distance from the Steelworks Cracow KObierzyn Mogila Mistrzejowice Downtown, Bielany SW 1.4 km NW5km W10km W 18km SW 16 km
Annual wet and dry deposition mg km- 2 164.2 Metal concentrations in dust deposition mg kg- 1 Zinc, Zn 2237 Lead, Pb 449 Cadmium, Cd 50.2 Metal deposition kg km-2 Zinc, Zn Lead, Pb Cadmium, Cd
367.3 73.7 7.2
89.1
86.0
46.6
42.9
1960 369 26.5
3104 580 50.2
2254 438 16.5
1967 502 11.3
174.8 32.9 2.4
267.0 49.9 4.3
105.0 20.4 0.77
84.3 21.5 0.48
spatial distribution of ~tals is presented. Figures 13.3 to 13.6 illustrate the vertical distribution and chemical fractionation according to the binding strength of Zn (Figure 13.3), Cd (Figure 13.4), eu (Figure 13.5), and Pb (Figure 13.6). Figure 13.7 displays the structure of these metals enrichment in the mobile/mobilizable fractions that are of particular importance for risk assessment.
Heavy Metal Contamination in Industrial Areas and Old Deserted Sites -
--
281
--.""~---
Zn Profile R 0
100% 80%
50
100
mg/kg
150
0-30
250
60%
mglkg
20-30 40% 30-50 20% 50-70
0% 0
~ N
~
'"
... 0
E
...•"
~
0
>70 em
Profile W 100% 80%
0
50
100
0
50
100
150
0-30
60% 30-70
40% 20%
70-100 0%
~
...0 0
'"
8
~
...
g ~
E
~"
100-130 em
Profile I 100% 90% 80% 70% 60% 50% 40% 30% 20% 10%
mg/kg
150
0-30
30-70
70-100
Figure 13.3. Vertical distribution and chemical fractionation of Zn in soil profiles (0-130 cm) in Site I; Metal binding fractions: FO+ 1(EXC): pore solution and exchangeable; F2 (CARB): carbonate associated; F3+4 (EMRO): easily and moderately reducible Mn- and amorphous Fe-oxides; F5(OM): oxidizable, associated with organic matter and sulfides; F6(R): residual (lithogenic matter).
Site II: Irena Glasswork Inowroclaw, Poland (Figures 13.1, 13.8)
~erous
Lead is one of the most and widespread anthropogenic pollutants. Environmental contamination with lead is associated with processing of zinc and lead ores, combustion of leaded gasoline in car engines, production of accumulators, paints, etc. Of
282
Fate and Transport of Heavy Metals in the Vadose Zone
Cd Profile R o
100% 90% 80% 70% 60%
0,1
0,2
mg/kg
0,3
0-30 I----------~r=___,I
50%
40% 30% 20% 10% 0%
20-30
I=llIII!l~.
30-50
PIl~
••
50-70 ~_.
§ g
•
>70 em
Pllml••
Profile W 100% 90%
80% 70% 60% 50%
40% 30% 20% 10% 0%
30-70
70-100
100-130 em
Profile I o
100%
0,1
0,2
mglkg
0,3
90% 80%
70%
0-30
60% 50%
40% 30% 20% 10% 0%
30-70
70-100
100-130 em
·IFO+11 Figure 13.4. Vertical distribution and chemical fractionation of Cd in soil profiles (0-130 cm) in Site I; Metal binding fractions: FO+ 1(EXC): pore solution and exchangeable; F2 (CARS): carbonate associated; F3+4 (EMRO): easily and moderately reducible Mn- and amorphous Fe-oxides; F5(OM): oxidizable, associated with organic matter and sulfides; F6(R): residual (lithogenic matter).
various industries, lead crystal glassworks are a proven source of lead contamination. The scope of this study comprises evaluating the extent of soil enrichment by lead in the vicinity of Irena Glasswork.
Heavy Metal Contamination in Industrial Areas and Old Deserted Sites
283
Cu Profile R o
100% 90% 80% 70%
~30
60%
~
5
10
15
________________
mg/kg
20
~
2~3O~• • •_
50%
3O-5OleI_ _ __
40% 30%
20% 10%
""""1I
~70~~• • •
0%
>70cm ~• •_ _
Profile W o
100% 90% 80% 70%
5
10
15
mg/kg
20
60%
50% 30-70
40%
30% 20% 10%
7~loo
0% 1~13Ocm
Profile I 100% 90% 80% 70%
20
60%
50% 30-70
40% 30%
20% 10%
7~loo
0% 1~13Ocm
I]FO+1
I
1IIIIIIIIIIIIIF2
. .I,--FS-l
Figure 13.5. Vertical distribution and chemical fractionation of Cu in soil profiles (0-130 cm) in Site I; Metal binding fractions: FO+ 1(EXC): pore solution and exchangeable; F2 (CARB): carbonate associated; F3+4 (EMRO): easily and moderately reducible Mn- and amorphous Fe-oxides; F5(OM): oxidizable, associated with organic matter and sulfides; F6(R): residual (lithogenic matter).
Locotion, Soil C/w(acteristics, and Land Use
Site II is located in Central Poland in the flat area adjacent to Irena Glasswork, in the NW part of Inowroclaw (Figure 13.8). Geomorphologically it belongs to plains of the moraine upland originating from the Baltic glaciation phase. The area lies at the border
284
Fate and Transport of Heavy Metals in the Vadose Zone
Pb Profile R o
100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0%
~30
20-30
5
10
15
20
~________________~~~,I
40 mg/kg
~~~II• • •
3~50 ~~• • •
5~70 ~• • • >70em
•••
bE~
Profile W 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0%
~30
3~70
7~loo
1~13Oem
Profile I o
100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0%
5
10
15
20
~3O
30-70
7~loo
100-130 em
IFO+11
1IIIIIIIIIIIIIF2
fFAjF3+41
IBIL..-FS----,
"L.....
F6 ---,
Figure 13.6. Vertical distribution and chemical fractionation of Pb in soil profiles (0-1 30 cm) in Site I; Metal binding fractions: FO+ 1 (EXC): pore solution and exchangeable; F2 (CARB): carbonate associated; F3+4 (EMRO): easily and moderately reducible Mn- and amorphous Fe-oxides; F5(OM): oxidizable, associated with organic matter and sulfides; F6(R): residual (lithogenic matter).
of a fallow gley podzol on the vadose zone matrix composed of sands and brown soils formed from loess and loessial formations. This results in variable content of clay fraction, ranging from 3.8 to 35.0 wt % and sand fraction occurring within the range from 78.2 to .38.8 wt %. The agricultural land represents mainly wheat complex. Dominating
Heavy Metal Contamination in Industrial Areas and Old Deserted Sites
285
Zn 100%
100%
80%
80%
80%
60%
60%
60%
40%
40%
40%
20%
20%
20%
0% I-
Profile
'--r-
~
0
R~
r
r
g
~
~
0 ....
0
"I
E
0
0" ....
e-
7'
"'
·
100%
r-'
0%
~
0 ....
0
'"O Profile W
8 ;;;....
0
'"
0%
0
....
0
'"
~ Ii
Profile I
0
'"0
8
;;;....
0
'"
g'-,
~ Ii
Cd 100% 80%
100%
.,--,
r
c
~
.... 80%
i
100% 80%
I
•••••••
60%
60%
40%
40%
40%
20%
20%
20%
60%
I
-+-
0% o
Profile R ~
b1
g
4:::!
fi:
E~ 0" ....
•
0%
g
4-
-+~
o
0
ProfileW
4-
8
~
'"
~
0%
0
'"
~ Ii
re
e-
r-
~
; ..
g
Profile I 0
-+-
-+~
g
8
~
4g -
Lj
E
~ "
Pb 100%
r
7'
'"'
100%
100%
;-
:>
80%
80%
80%
60%
60%
60%
40%
40%
40%
20%
20%
20%
0%
g
-+-
4-
ProfileWC!i
-+-
~
~
8
g
'" Pi
g
-~"E
0%
g -+- ~ 4- 8 -+-g
Profile I C!i
g
~
~
--l
Ii
Cu 100%
100%
100%
80%
80%
80%
60%
60%
60%
40%
40%
40%
20%
20%
20%
0%
g
Profile R ~
I
-+-4-4-y ~ ~ E
g
fi:
jFO+1
I
~
•
0%
'--r0
Profile W
~
'--,
~
.... 0
g
8 ;;;....
0
'"- E ~"
0%
g 4- ~ Lr- 8 -+-g
Profile I 0
g
;;;....
--l
~ Ii
1IIIIIIIIIIIIIF2
Figure 13.7. Partition of the mobile fraction of trace metals (Zn, Cd, Pb and Cu) in soil profiles (0130 cm) in Site I; Metal binding mobile fractions: FO+ 1 (EXC): pore solution and exchangeable; F2 (CARB): carbonate associated; F3+4 (EMRO): easily and moderately reducible Mn- and amorphous Fe-oxides; F5(OM): oxidizable, associated with organic matter and sulfides.
286
Fate and Transport of Heavy Metals in the Vadose Zone
E
_o;;;':'s~_ _ _....;c~
0-....
km
c e2l 025
0
Figure 13.8. General plan of Site II: area adjacent to Irena Glasswork n/lnowroclaw, Poland). P - Irena Glasswork; 1 - sampling points along the intersections A, B, C, 0; 2 - compact residential area; 3 - railways; 4 - motorways.
winds blow from W, NW, and SW direction, which means that due to unfavorable location, mainly the compact residential area of Inowroclaw City is affected with the emission from the Glasswork. The surveyed site comprised an area of 5000-m radius in NE, NW, W, and partially SW directions from Irena Glasswork, used for intensive agricultural production (mainly arable land and orchards). In the NE, NW, and W-WS directions the area is crosscut with parallel railroads and motorways.
of Antflropogenic Contarnination of Soil by Lead The major source of soil contamination by Pb in Site II is an emission from the stacks of Irena Glasswork. The glasswork has been in operation already for several decades, but as a source of Pb emission it is considered since 1976, when the production of lead crystal glass started. Pb has been emitted to the atmosphere with particulates, mainly in the form of oxides. Another source of Pb contamination occurring in the area is leaded petrol combustion in motor vehicles. The survey of soil contamination by Pb in Site II was focused on the evaluation of the glasswork impact on the extent and character of undisturbed soil contamination in this area. The results of the survey are illustrated in Figures 13.9 and 13.10, which present spatial and vertical distribution of Pb in the un-
Heavy Metal Contamination in Industrial Areas and Old Deserted Sites
o-2.5cm
2.5 -5.0 em
5.0 -10.0 em
10.0 -15.0 em
15.0 - 20.0 em
35.0 - 40.0 em
287
Figure 13.9. Site II (area adjacent to Irena Glasswork n/lnowroclaw). Spatial and vertical distribution of lead in an undisturbed surface soil layer (0-40 cm).
disturbed surface soil layer 0--40 em, as well as chemical fractionation ofPb accumulated in soil vs. distance from the source of emission (Glasswork stacks).
Soil Enrichment with Heavy Metals in the Areas Impacted by long-Term Stack Emission
it
Screening Survey and Methods An extent of surface soil and the upper part of the vadose zone (subsoil) matrix contamination by heavy metals in Site I (Nowa Huta, area adjacent to Sendzimir Steel-
288
Fate and Transport of Heavy Metals in the Vadose Zone
°
A-1a 0-2,5
100
200
11111111:·· .::.: ......
300
400
500
600
:-:~
mg/kg
2,5-5,0 1 - - - - - - ' 5,0-10,0 1--_-' 10,0-15,0 1--_.... 15,0-20,0
A-1a
Total
Mobile Fractions
35,0-40,0 em
°
8-8a 0-2,5
100
11111:
200
.....:. :..
300
400
500
600
~
mg/k 2,5-5,0 1--_ _--' 5,0-10,0 1--_-' 10,0-15,0 1--_....
Total
8-8a
Mobile Fractions
35,0-40,0 em
C-15a
°
100
°
100
200
300
400
500
600
mg/kg
0-2,5 2,5-5,0 5,0-10,0 10,0-15,0 15,0-20,0
C-15a 100%
Total
Mobile Fractions
,--"r=',.,..,.=----.,.........,.,.",...---,
35,0-40,0 em
0-22a
80%
0-2,5
60%
2,5-5,0
40%
5,0-10,0
20%
10,0-15,0
300
400
500
600
mg/kg
15,0-20,0
0%
0-22a
200
Total
Mobile Fractions
35,0-40,0 em
Figure 13.10. Vertical distribution and chemical fractionation of Pb in an undisturbed surface soil layer (0-40 cm), Site II. Metal binding fractions (McLaren and Crawford, 1973): fO+ 1 (EXC): pore solution and exchangeable; f2(CARB): specifically sorbed, carbonate-associated; f5(OM): oxidizable, associated with organic matter; f3+4(EMRO) free Mn- and Fe-oxides.
Heavy Metal Contamination in Industrial Areas and Old Deserted Sites
289
works in Poland) (Figures 13.1, 13.2A) resulting from the impact of the Steelworks stack emission was evaluated on the basis of metal accumulation and spatial distribution in surface soil layer (Table 13.4, Figure 13.2B-I). Barrier capacity of soil, binding phases, and metal mobility in soiVvadose zone matrix were also taken into consideration (Figures 13.3 to 13.7). The evaluation was based on the random survey of soil for trace metals in node-points of the network of SOxSO m squares carried out according to the EPA guidelines (Barth and Mason, 1984). The survey comprised sampling: (i) surface soil layer in the basic 63 node-points; (ii) surface soil layer in 8 points (one of every 10) taken in duplicate in close proximity to the basic node-points to estimate the variability among sampling units; (iii) soil and subsoil matrix in four layers up to 30 cm thick along the upper part of the vertical profile of the vadose zone up to 130 cm deep, in three points selected at different distances and directions from the Steelworks with respect to the wind rose: Wadow 64-67 (Profile W), Ruszcza 4S-48 (Profile R), and Igolomia 75-78 (Profile I) (Figures 13.2B-I, 13.3 to 13.7). In Site II (Irena Glasswork) (Figures 13.1, 13.8), the undisturbed soil enrichment with Pb against the distance from the source of emission and the depth of soil layer was assessed (Figure 13.9), with a special regard to the vertical migration of Pb into undisturbed soil layers (Figures 13.9, 13.10). The screening survey comprised soil sampling for lead in 2S sampling points along the 4 intersections (A, B, C, D) in the distance from so to SOOO m from the Irena Glasswork, consecutively in 50, 100, 2S0, SOO, 1000, 2S00, and 5000 m from the source ofPb emission (Glasswork stacks). In the investigated area, other sources of Pb than the Glasswork occurred: Pb emission from motor vehicles, pump stations distributing leaded gasoline, and other activities emitting Pb in the compact settlement area of Inowroclaw city, located in the E, SE, and S direction from the Glasswork. The disturbing effect of cultivation on vertical distribution of Pb in the soil layer in the agricultural areas located in SW, W, NW, and N directions from the Glasswork also should have been considered. To exclude effect of these factors that could influence spatial and vertical Pb distribution in the soil layer, it was desisted from the random sampling procedure and from soil sampling within the compacted settlements of Inowroclaw city located in the area most affected by emission from the Irena Glasswork. The samples were thus taken along the intersections laid out in the barren undisturbed land in agricultural area, in the NE, NW, W, and SWW directions from the Glasswork, not closer than SO-100 m from the motorways and buildings. In each point,S consecutive soil layers from 2.5 to S.O cm thick (the uppermost two 2.5 cm thick, the rest 5.0 cm thick), up to the depth of 20 cm, along with the layer 3S-40 cm were sampled and analyzed for Pb. In conformity with the sampling program and a scope of the studies, the material analyzed for trace metals comprised soil and subsoil matrix. Acid-digested (ASTM, D 5198-92, 1992) soil and subsoil matrix samples were analyzed for the total metal content by standard methods using AAS and ICP-AES techniques (AAS Perkin Elmer 1100 B and ICP Perkin Elmer Plasma 40). Binding strength of metals in the selected soiVvadose zone matrix samples from Site I was evaluated using sequential extraction (Tessier et al., 1979, modified by Kersten and Forstner, 1986). Sequential extraction scheme partitions off exchangeable FO+1 (EXC) , carbonate-bound F2(CARB), easily and moderately reducible Mn-oxides and amorphous Fe-oxides F3+4(EMRO), oxidizable sulfidic/organic FS(OM) and residual frac-
N \.0
o
."
III .... fI)
III
::::I C.
~
III
::::I VI "0
o
:4-
a::I:
Table 13.4. Concentrations of Heavy Metals in Soils of Site I (Nowa Huta n/Cracow) Compared to the Geochemical Background in Unpolluted Areas Site Nowa Huta (Site I) Geochemical background"
from-to mean from-to mean
Fe (%)
Mn
1.25-2.50 1.61 0.80-2.78 1.20
254-1160 485 380-700 560
" Kabata Pendias and Pendias, 1992.
Heavy Metals Concentration (mg kg-l) Zn Cu Pb Ni 32-670 120 30-360 65
8.4-22.2 13.8 4.0-53.0 19
10-42 27 19-49 25
11.0-22.8 16.7 10.0-104 25
Cd
Cr
0.1-1.1 0.45 0.08-0.96 0.38
14.4-40.8 24.2 14.0-80.0 38
~re
s:
~ :i" iii
:fI):r
<: c.
III
~
fI)
N
o ::::I
fI)
Heavy Metal Contamination in Industrial Areas and Old Deserted Sites
291
tions F6(R), of subsequently increasing binding strength. For the selected soil samples from Site II, that was investigated several years earlier than other sites, the 5-step sequential procedure by McLaren and Crawford (1973) was used, which partitions off exchangeable (fO+1), specifically sorbed, mainly carbonate-bound (£2), organic (f5), free oxides (f3+4), and residual (f6) fractions. In general, it well corresponds with a chemical fractionation of metals derived from the procedure by Tessier et al. (1979) modified by Kersten and Forstner (1986).
Metal Distribution in Soil
VS.
the Duration and Extent oj Emission
Survey of soils for heavy metals in Site I showed distinct impact of a stack emission from the Steelworks on the trace metal distribution in the undisturbed surface soil layer (0 to 40 cm) (Figure 13.2B-I). The maps of equipotential lines generated by SURFER 5.01 Mapping System for the surveyed metals display clear correlation with the emission from the Steelworks for Zn, Cr, Fe, Cd, and Pb. At the same time, weak impact on Mn and Cu concentrations and practically no effect on Ni content in the area adjacent to the Steelworks in the direction of dominating winds was observed (SW area of Site I in the close proximity to the Steelworks). Zn in the whole impacted area occurred in mean concentrations markedly (about twofold) higher with respect to the mean geochemical background for the similar soils from the unpolluted areas (Table 13.4) (Kabata-Pendias and Pendias, 1992). Besides, Zn was also the only element showing a uniform background concentration range of 50 to 100 mg kg- 1 in the nonimpacted area and the distinct borders of the impact of the Steelworks emission, while no other competitive sources of emission occurred (Figure 13.2D). The Steelworks were apparently responsible for the elevated concentrations of Fe in the soil compared to the geochemical background (Figure 13.2B), though another area in SE of the site (Igolomia) also displays concentration anomaly. Of other 8 surveyed metals, maximum and mean contents of Cd occurred somewhat in excess with respect to the background concentrations in unpolluted soils, both in the impacted area of the Steelworks (SW) and in the Igolomia site (SE) (Figure 13.21). The Steelworks was also a sole source of a substantial (two- to threefold) enrichment of the surface soil layer in the adjacent area by Cr (Figure 13.2F). The background concentrations of Cr in the area were, though, very low and both the maximum and mean concentrations of this element were placed within the concentration range occurring in the unpolluted soils (Table 13.4). The analysis of the equipotential map of heavy metal distribution in the surface soils of Site I gives evidence of occurrence of other sources of metal enrichment in soils, comparable or even of a higher extent than the Steelworks. Such area of significantly elevated concentrations of Fe, Mn, Pb, Ni, Cu, and Cd was identified in Igolomia, SE of Site I (Figure 13.2B,C,E,G,H,I). The no longer existing actual source of this anomaly is probably an old metal processing workshop, which can be assumed from the character and limited area of enrichment. Pb and Ni displayed a vast area of elevated concentrations compared to the local background (Figure 13.2E,G). The distribution of these metals suggests overlapping effect of several enrichment sources, in particular motorways besides the Steelworks and Igolomia abandoned site. In general, from the spatial distribution of surveyed metals and their concentrations in soils of Site I in the vicinity of the Steelworks, it can be assumed that the significance of the Steelworks with respect to soil contamination by metals in Cracow
292
Fate and Transport of Heavy Metals in the Vadose Zone
region is overestimated. Besides Zn and Cd in the limited area, all other surveyed metals emitted from the Steelworks did not exceed the concentrations occurring in unpolluted soils, despite Steelworks being in operation for more than 40 years. Other sources of elevated concentrations of metals with respect to the geochemical background appeared to be of comparable (Igolomia: Fe, Pb, Cd) or higher extent (lgolomia: Mn, Ni, eu; motorways: Pb, Ni; railway: eu). Comparison of the geostatistic analysis of Zn, Pb, and Cd occurrence in the Site I fitted by GEO-EAS computer program (Twardowska, 1995) and by SURFER Version 5.01 Surface Mapping System applied in this study showed a very good consistency and thus proved high compatibility of both software.
Barrier Capacity Of a Surface Soil Layer Strong barrier capacity of a surface soil layer with respect to metals was proved by a distinctly higher content of these species in contaminated surface soil than in deeper parts of the soil profile. A vertical distribution of metals in the surface soil layer clearly reflects an extent of anthropogenic impact and either anthropogenic or geogenic origin of a metal. Good illustration is the vertical distribution of surveyed metals in soil profiles in Site I (Figures 13.3 to 13.6). The elevated concentrations of Zn caused by the emission from the Steelworks occur actually only in the surface soil layer, while Zn content in subsoilloessial matrix is at the uniform background level (Figure 13.3). The excess of Zn concentration in the surface layer appeared to be more than 5 times higher compared to the layer 20 to 30 cm in the profile R (Ruszcza) closest to the Steelworks and located in the disused area. This indicates accumulation of metals from wet and dry deposition in the uppermost layers of undisturbed soil profile. Land cultivation (plowing) causes homogenization of metal concentrations within the cultivated layer. Nevertheless, also in the cultivated area, the uppermost soil layer directly exposed to the deposition will be more enriched with metals. Therefore, any error or inconsistency in sampling procedure can lead to serious errors in evaluation of the environmental impact of the particular source of emission. A pattern of a vertical distribution similar to that of Zn show also other metals of a distinct anthropogenic impact proven by the pattern of spatial distribution, in particular Pb and Cd (Figures 13.4, 13.5). For metals of a low or negligible anthropogenic impact (e.g., eu) the species concentration along the whole profile is uniform (Figure 13.6). For the evaluation of the extent of soil contamination, besides spatial and vertical distribution of metal concentrations in soils, no less important is metal mobility. This quality is characterized by metal fractionation according to the binding strength (Figures 13.3 to 13.7), in particular by mobile/mobilizable fractions (Figure 13.7). Metal mobility and its assessment will be discussed further. The correct site-, use-, and target-specific risk assessment, cost-effective cleanup actions, and correct evaluation of barrier capacity of a surface soil layer with respect to the continuous anthropogenic impact of a particular metal from the stack emission require more detailed information concerning spatial and vertical distribution of species in this layer itself, both in undisturbed and cultivated soil of different types. The pattern of metal accumulation and migration in the undisturbed surface soil layer was exemplified by the study of soil contamination in the vicinity of Irena Glasswork
Heavy Metal Contamination in Industrial Areas and Old Deserted Sites
293
after 7 years' production of lead glass (Site II) (Figures 13.1, 13.8). The survey of the surface layer (0 to 2.5 cm) of undisturbed soils for lead in the vicinity of Irena Glasswork after 7 years' production of lead glass (Site II) showed its extensive enrichment with Pb species (Figure 13.9). The highest concentrations of Pb and their longest range occurred along the intersection A-A, the lowest in the directions C-C, in conformity with wind frequency. In all the investigated area, Pb concentrations in the uppermost soil layer exceeded the natural background contents (20 mg kg-I). The distance range from the source of emission of estimated concentrations above the level considered permissible (100 mg kg-I) (Kabata-Pendias and Pendias, 1992) was from about 3000 m (NE, intersection A-A) to below 500 m (W, intersection C-C). The vertical distribution of Pb shows, though, that high accumulation of this metal occurs predominantly in the uppermost 0 to 2.5 cm layer of surface soil. The Pb concentration decreases sharply (approximately in half) in the next 2.5 to 5 cm layer and gradually in the subsequent layers, generally in accordance with a logarithmic pattern. In the layer 35 to 40 cm, all the Pb concentrations were below 50 mg kg-I. In the area up to about 500 m radius around the Glasswork, the concentrations of Pb in the deepest surveyed layer 35 to 40 cm were still above the background level. Up to over twofold higher Pb contents in this layer were observed in the closest proximity (in the radius of about 250 m) from the source of emission (Figures 13.9, 13.10). The pattern of a vertical distribution of Pb confirms, on the one hand, a very strong barrier capacity of the uppermost, humic layer of the soil. On the other hand, it suggests lack of a leak-tightness of this layer and permanent gradual "unloading" of the uppermost layer due to the vertical migration of a metal. Considering the pattern of Pb vertical distribution, maximum mean concentrations of this metal in the cultivated surface soil layer relevant to the reported values in the undisturbed soil layer are supposed to be considerably lower due to averaging effect of cultivation. The averaged values will range from about 80 mg kg- I in the direction of the lowest wind frequency (W, NW, SW), to about 160 mg kg- I in the directions of the most frequent winds (E, NE, SE). Due to continuous emission of metals, also in cultivated soils the uppermost layer will display higher metal content resulting from accumulation of dry and wet deposition in this layer in the periods between cultivation treatment of soils. The presented data lead to a conclusion that for risk assessment at land use such as playgrounds, sport fields, or recreation areas, where transfer pathways are through the direct uptake and the risk receptors are children, the metal concentrations in the uppermost 0 to 2.5 cm layer of soil should be considered.
Heavy Metal Binding Strength and Mobility in Soils For quality-safe target-specific actual and potential risk assessment, the identification of metal mobility in soils is of fundamental importance. The evaluation of the total species concentration in the matrix does not give enough information, due to different binding strength of various physiochemical associations, in which the metals occur in soils. This feature determines metal ability to mobilization under different exposure conditions with respect to different direct risk receptors, of which humans (adults and children), farm and wildlife, soil organisms and groundwater should be specified. At present, many authors emphasize a necessity for determination of metal fractions of different mobility as a requirement for risk assessment (Gupta et aI., 1996; McGrath, 1996; Ure,
294
Fate and Transport of Heavy Metals in the Vadose Zone
1996). For this purpose, sequential extraction schemes for distinguishing metal-binding fractions appear to be an extremely useful tool. The concept of this technique is that elements occur in the soil matrix in various pools of different binding strength which can be assessed by different reagents (Ure, 1996). Since 1973, more than 10 sequential extraction procedures using different extractants and defining from three to nine extraction steps to identify "forms" of metal binding, have been elaborated. Among them, those developed by McLaren and Crawford (1973), Tessier et al. (1979), Kaszycki and Hall (1996), and Ure (1996) are of general or specific use. The chemical extraction sequences by many authors are still subject to arguments and controversy concerning the selectivity of extractants and the redistribution of metals among phases during fractionation (e.g., Tessier and Campbell, 1991; Xiao-Quan and Bin, 1993; Tack and Verloo, 1996). The indisputable advantage of this method lies in the possibility of evaluating actual and potential availability of metals for site- and use-specific risk receptors, as well as of estimating long-term effects on metal mobility of the changing controlling factors. The attempts of many authors are focused on using a sequential extraction procedure mainly for the identification of chemical associations of pollutants in different mixed organic/inorganic matrices. The greatest virtue of the chemical extraction sequences, though, is a possibility to differentiate between the fractions of different binding strength onto particular matrix and to compare different matrices partitioning with respect to the fractions of adequate binding strength. Mechanisms of metal binding onto these matrices can be different; e.g., metal bonding on the matters predominantly organic like peat and predominantly mineral like fly ash (Twardowska and Kyziol, 1996). In general, for these purposes the optimum sequential fractionation procedure should be simple both analytically and conceptually and display an order of a consecutive increase of binding strength. These properties are found in one of the most widely applied sequential extraction procedures, that of Tessier et al. (1979), modified by Kersten and Forstner (1986) for partitioning sediment samples. This procedure has been also used directly for metal speciation in soils (e.g., Harrison et aI., 1981; McGrath, 1996), also in this study for Site I (Figures 13.3 to 13.7). Comparison of the chemical fractionation of zinc in the anthropogenically impacted surface soil layer, and in consecutive subsoil layers where the Zn concentrations reflect the natural background level, indicates occurrence of considerable qualitative as well as quantitative changes (Figures 13.3, 13.7). With respect to distribution in the surveyed unpolluted subsoil layers (>70 to 130 cm) in mg.kg.- 1, Zn association with fractions of a different binding strength followed the order: F6(R) > F2(CARB) > F3+4(EMRO) » F5(OM) »» FO+l(EXC) The residual fraction accounts for about 50%, while 25 to 30% is associated with mobile carbonate-bound fraction. Variable amounts of Zn are bound with mobilizable F3+4(EMRO) and strongly bound with oxidizable fractions F5(OM): the rate of Zn associated with both these fractions ranged from about 15% to approximately 30%. The association of Zn with easily and moderately reducible Mn- and Fe-oxides comprised 14 to 25%, while in oxidizable fraction (generally associated with organic matter and sulfides, here in particular with organic matter), it occurred in a minor amount (1.2 to 5.0%). The role of the exchangeable fraction in Zn binding was negligible in all layers of the surveyed soil profiles.
Heavy Metal Contamination in Industrial Areas and Old Deserted Sites
295
In overlaying intermediate subsoil layer 20 to 70 cm, both Zn concentrations and chemical fractionation appeared to be similar to that in the underlying layer >70 to 130 cm, though a tendency to increase of Zn rate associated with reducible fraction was observed. In the surface humic soil layer weakly impacted by the emission from the Steelworks (profiles Wand I), the enrichment of most binding fractions with Zn occurred, but to the different extent. It resulted in the rearrangement of the relative and absolute partitioning order according to the sequence (Figures 13.3, 13.7): F3+4(EMRO) > F6(R) » F2(CARB) > F5(OM) » » FO+1(EXC) The association of Zn with mobilizable reducible oxide-bound fractions and oxidizable fraction associated with organic matter in the surface soil increased up to 51 to 56% of the total Zn, the reducible oxide-bound fraction being dominant (41-48%). The amount of Zn associated with carbonate-bound fraction was generally stable along the soil profile and did not show enrichment in the surface soil, which resulted in the substantial decrease of the proportion of mobile fraction in comparison with the subsoil layer. The results of Zn partitioning are generally in line with those reported by other authors for unpolluted and geochemically polluted soils (McGrath, 1996) and sediments (Tack and Verloo, 1996). All these matrices showed domination of Zn associated with reducible oxide fractions, generally minor role of oxidizable fraction, high rate of stable residuum, and relatively low proportion of mobile fractions associated with carbonates. The role of the exchangeable fraction in zinc binding appears to be negligible. To summarize, in the surface soil layer enriched with zinc anthropogenically, from 30 to 35% of this species comprised immobile lithogenic material, while 65 to 70% of the total concentration comprises mobile or mobilizable fractions of different binding strength (Figure 13.7). Partitioning of cadmium (Figures 13.4, 13.7) in the surveyed soil profiles followed the sequence: - in subsoil layer 30 to 130 cm: F6(R) > F3+4(EMRO) > FO+1(EXC) > F2(CARB) » » F5(OM) - in surface soil layer 0 to 30 cm: FO+1(EXC)
z
F3+4(EMRO) > F6(R)
z
F2(CARB) » » F5(OM)
In the subsoil layer, 35 to 45% of Cd was associated with residual fraction. Reducible oxide-bound Cd comprised 27 to 30%. The mobile exchangeable and carbonatic fractions accounted for 19 to 22% and 12 to 17% of Cd, respectively, while the amount of Cd bound to oxidizable organic fraction was negligible. High content of humic organic compounds in the surface soil layer did not enhance binding Cd with oxidizable organic fraction, whereas the amounts of Cd bound in exchangeable, reducible, and carbonate fractions significantly increased as compared to subsoil. Except the oxidizable fraction, the Cd distribution among the fractions of different binding strength was almost uniform, showing great resemblance to the pattern of Cd partitioning in the moderately polluted soil studied by Harrison et al. (1981). Frac-
296
Fate and Transport of Heavy Metals in the Vadose Zone
tionation of Cd adsorbed onto peat; i.e., predominantly organic matter, displayed dominance of binding mainly onto FO+1(EXC) fraction of the weakest binding strength (Twardowska and Kyziol, 1996). It could be therefore admitted that a significant part of the most labile FO+ 1 (EXC) fraction may be associated with organic matter, besides that of clay minerals. In the case of humic-rich matters, the attribution of metal binding to ion exchange mechanism is questionable. This supports an assumption, expressed also by Kersten and Forstner (1988) and Tack and Verloo (1996), that the mechanism of metal binding onto different or transformed matrices also considerably differs, while the most reliable parameter for comparison is binding strength, adequate to the related fractions (Twardowska and Kyziol, 1996). It should be also emphasized that in general, chemical fractionation of Cd in soil and sediments indicates its predominant binding onto mobile and easily reducible phases (Harrison et al., 1981; Kersten and Forstner, 1988; Forstner, 1992; McGrath, 1996; Tack and Verloo, 1996). Therefore, Cd is susceptible to remobilization resulting from the changes of the chemical environment. The partitioning of Cd is thus also subject to strong changes. The reported data on soils and sediments are consistent with respect to role and significance of exchangeable, reducible oxide-associated and carbonate associated mobile and mobilizable fractions in Cd binding (Figures 13.4, 13.7). A bigger difference in the reported data is concerning oxidizable and residual fraction (Forstner, 1992; McGrath, 1996; Tack and Verloo, 1996). Copper occurrence in the soils of the Site 1, as shown by the spatial distribution and concentration range (Figure 13.2H, Table 13.4) displays weak impact of the emission from the Steelworks and in the surveyed soil profiles is of predominantly geogenic origin. It results in uniformity of Cu distribution and partition along the profiles (Figures 13.5, 13.7). From 44 to 54% of Cu is stably bound in the residual fraction. The predominant part of mobilizable species was found in oxidizable fraction (25 to 33%), that seems to be geogenically specific for soils and sediments and is in conformity with other sources (Tack and Verloo, 1996; McGrath, 1996). It should be mentioned that also in some anthropogenic materials, such as municipal solid wastes, domination of specific linkage of Cu to organic matter was observed (Prudent et al., 1996), though not all the matrices show the same binding pattern (Twardowska and KyzioL 1996). The partitioning of Cu in the surveyed profiles, both in surface soil and subsoil layers, follows the general order: F6(R) > F5(OM) > F3+4(EMRO) » F2(CARB) >FO+1(EXC) In the deeper part of Profile I (Igolomia), predominant binding to the reducible fraction occurred (Figures 13.5, 13.7). No visible increase of Cu association with oxidizable organic matter-bound fraction was observed, despite much higher content of organic matter in this layer. Distribution of lead (Figures 13.6, 13.7) reflects its low mobility in soil and subsoil profIles. Opposite to Cd, association of Pb with mobile fractions, both exchangeable and particularly carbonate-bound appeared to be very low. In all layers of the profiles, including surface soil layer, mobile exchangeable and carbonate-bound fractions comprised 2.5 to 5.5% and 0.0 to 3.2% of total Pb, respectively. The highest, though variable, enrichment of subsoil with Pb occurred in the residual (25 to 66%) and reducible oxide-bound fractions (25 to 66%). The highest lead binding in the residual fraction (56 to 66%) and the lowest in reducible oxide-bound one (22 to 29%) was observed in the R
Heavy Metal Contamination in Industrial Areas and Old Deserted Sites
297
(Ruszcza) profile, in the area of the highest impact of the Steelworks. In two other profiles Wand I located in less impacted areas the proportions of the residual and reducible fractions were 25 to 43% and 52 to 66%, respectively. The total Pb contents in the subsoil (> 30 cm) of the R profile ranged from 10 to 12 mg kg-I, while the concentration range in the subsoil of these two other profiles was 10 to 22 mg kg- I and could be thus assumed as falling within the uniform background concentrations. Therefore, it is rather unlikely that enrichment of residual fraction and decrease of reducible one was induced anthropogenically, and probably reflects the geogenic variability of the area in this respect. In the subsoil layer > 70 cm, Pb binding onto organic matter is generally very low or negligible, but increases in the upper transitional subsoil layer (70 to 50 cm). In the surface soil layer, the proportion of Pb associated with organic fraction considerably increases (to 13-23%), which shows good correlation both with the content of organic matter in the soil profile and exposure to the anthropogenic impact. Besides higher rate of organic-bound fraction and general quantitative increase of Pb-enrichment, dependent upon the distance and direction with respect to the emission source, no substantial changes in partitioning of this metal in soil profile was observed. Partitioning of Pb with respect to binding strength and predominant chemical associations in the surveyed soil profiles followed the sequence: - in subsoil layer 30-130 cm: F3+4(EMRO)
><
F6(R) » FO+1(EXC)
><
F6(R)
~
F2(CARB)
~
F5(OM)
- in surface soil layer: F3+4(EMRO)
>
F5(OM) » FO+1(EXC)
~
F2(CARB)
Comparison of chemical fractionation of Pb in the surface soil layer in Site I (Figures 13.6, 13.7) and in Site II adjacent to the Irena Glasswork (Figure 13.8), where the uppermost 0 to 2.5 cm soil layer is highly contaminated by lead (about one order of magnitude compared to the surface soil layer in Site I) (Figure 13.9), displays clear influence of the extent of anthropogenic impact on Pb distribution among the fractions (Figure 13.10). Partitioning of Pb in the least contaminated soil samples (C-15-a and D-22-a) shows high enrichment of the stable bound residual fraction (62 to 63%). The rest of species was almost equally partitioned over mobile and mobilizable fractions of different binding strength: exchangeable fraction comprised 8 to 9%, the fraction associated with organic matter 12 to 15%, while the rest was distributed among oxide-bound and specifically sorbed (mainly carbonate-bound) fractions: F6(R) » F5(OM)
>
F1+0(EXC)
><
F2(CARB)
><
F3+4(EMRO)
In the most contaminated area (samples A-1-a and B-8-a), considerable changes of Pb distribution occur: F6(R) "" F5(OM) » F3+4(EMRO)
><
F2(CARB)
>
FO+1(EXC)
298
Fate and Transport of Heavy Metals in the Vadose Zone
Comparison between the samples of the highest (A-I-a) and the lowest (D-22-a) Pbcontamination showed that the most anthropogenically enriched fractions appeared to be those associated with organic matter (46%) and stable residuum (28%). Much weaker anthropogenic impact displayed, in the descending order, fractions: specifically sorbed F2(CARB) (14%), oxide-bound F3+4(EMRO) (8%), and exchangeable FO+1(EXC) (4%). The anthropogenic enrichment follows, therefore, the sequence: FS(OM) > F6(R) » F2(CARB) > F3+4(EMRO) > FO+1(EXC) The chemical fractions associated with Pb in the soil samples taken from Site II cannot be directly compared with those in Site I due to use of different sequential extraction methods (by McLaren and Crawford, 1973, in Site II and by Tessier et ai., 1979, modified by Kersten and Forstner, 1986, in Site I). The analysis of Pb partition in both sites, though, clearly shows that the highest enrichment due to the anthropogenic impact (stack emission) occurs in the oxidizable organic matter-bound and stable residual fraction. Mobile chemical associations with exchangeable and carbonate fraction, as well as with reducible step associated mainly with manganese oxides and amorphous iron oxyhydroxides, are subject to the anthropogenic enrichment to much lesser extent. To conclude, heavy metal fractionation in surface soil and the vadose zone matrix differs substantially with respect to binding strength. Surface soil has high barrier properties, which cause enrichment of this layer with heavy metals in the areas impacted by anthropogenic emission. In general, anthropogenic enrichment occurs in all binding fractions, though at a different rate. The highest increase has been observed in mobilizable fractions, which results in the elevation of hazard to higher extent than it can be assumed from the quantitative changes. This leads to the conclusion that for quality-safe risk assessment, not only quantitative but also qualitative transformations of metal associations caused by anthropogenic impact should be considered.
Monitoring Program Requirements for Risk Assessment from large-Area Soil Contamination by Trace Metals from Anthropogenic Sources The results of soil survey in two anthropogenically impacted sites show the importance of assessing such parameters as (i) actual and potential land use and risk receptors; (ii) the thickness of an averaged surface soil layer to be exposed to a direct contact with risk receptors and the form of a contact; (iii) the fractions of the total metal content in soil actually available and implying a risk for the risk receptor; (iv) the fractions of the total metal content in the soil potentially available (mobilizable) and probable conditions of the metal(s) mobilization. These parameters are essential for a quality-safe monitoring and evaluation of an extent of soil contamination by trace metals. The monitoring requirements are based on the character of a vertical distribution of metals in anthropogenically impacted soil from the large-area emission, showing high accumulation in the uppermost layers of soil, 1-2 cm thick. In the areas undisturbed by depth-averaging cultivation treatment (e.g., lawns in childrens' playgrounds, meadows used as grazing areas) this layer will be directly exposed to the contact with receptors (children, farm animals, and wildlife). In agricultural land, the direct receptors (e.g.,
Heavy Metal Contamination in Industrial Areas and Old Deserted Sites
299
plants, food crops, fodder crops) will be exposed to a concentration of metal averaged by cultivation treatment. Considering the association of metals in soil matrices with "pools" displaying different binding strength, which reflects direct and potential availability to the different receptors, application of the sequential extraction procedure gives an essential opportunity to avoid false-positive errors in actual risk assessment. Overestimating the risk may be avoided through excluding the rate of metal stably bound in the residual fraction. The correct risk assessment requires an identification of metal-binding fractions directly available to the particular receptors, e.g., mobile fractions FO (pore solution), F1 and F2 displaying the weakest binding strength, and thus susceptible to leaching and groundwater contamination. In general, for the actual risk assessment, evaluation of mobile, mobilizable on/after uptake and immobile (stable) fractions provides adequate required information. For this purpose, the sequential extraction is a proven, reliable tool. The testing in the frame of BCR-interlaboratory studies of two extraction procedures, to be considered as standards by ISO (Quevauviller et al., 1996; Ure, 1996), confirms both their reliability and usefulness for risk assessment needs. As has been shown above, sequential extraction also provides valuable information on quantitative and qualitative changes in partition of a metal in question, resulting from the anthropogenic impact. For the potential risk assessment, long-term prognosis of heavy metal release and selection of the optimal remedial/cleanup actions, not only metal fractionation according to binding strength, but also identification of the geogenic and anthropogenic chemical associations of pollutants, in particular in mobilizable fractions [easily and strongly reducible F3+F4(EMRO) and oxidizable (FS) (OM)] are required. The definition ofparameters, transforming equilibria conditions in matrix, as well as external or internal factors controlling these transformations (e.g., pH, Eh) are also a prerequisite for the correct life cycle risk assessment. The results presented here show the need for a differentiated approach to actual and potential risk assessment from the large-area sources of emission such as stack emission, and point out the pitfalls of data inconsistency in their evaluation. The monitoring program for quality- fe risk assessment and the selection of a both efficient and cost-effective remedial strate minimizing the adverse consequences oElong-term emission should be highly use-specific and target-oriented. Monitoring data on trace metal enrichment in soil and vadose zone matrix caused by wet and dry depositio from industrial sources (mainly stack emission) in the vicinity of operating industrial lants (Sendzimir Steelworks, Site 1, and Irena Glasswork, Site II) showed an essenti role of chemical fractionation of metals in adequate evaluation of soil contaminati n. A substantial part of the total metal load originating from the anthropogenic industrial sources is stably bound in the residual fraction. In some emissions, though, anthropogenic contaminants occur in more labile forms than the species of the lithogenic/geogenic origin, which adequately increases the risk (e.g., anthropogenic enrichment of oxidizable fraction with Pb in Site II). Taking into consideration at risk assessment not only contaminant concentration, but also its chemical fractionation with respect to binding strength could highly improve the classical principle of preliminary evaluation of contaminated sites based on soil threshold values. Application of scientifically proven critical values would also greatly enhance site- and use-specific models of exposure assessment. Up to now, these values are a weak point of the best-constructed
300
Fate and Transport of Heavy Metals in the Vadose Zone
exposure assessment models. Metal fractionation in soil for risk assessment and management has been taken into account in a three-level concept by Gupta et al. (1996).
EVALUATION OF A LARGE-AREA DESERTED INDUSTRIAL SITE Investigation of a large deserted industrial site as a potential human risk was presented in the case study on an abandoned industrial area of Marktredwitz in Germany, impacted by the long-term emission of Hg and Sb from an old chemical plant (Site III) (Figures 13.1, 13.11). The major issue facing old contaminated sites sanitation requirements is the need of a quality-safe evaluation of such areas, taking into account both interests of the environment and nature on one side and economy and industry on the other. Therefore, an optimum model of investigation and assessment of chemical pollution of the site is to be use- and site-specific, in accordance with particular criteria in view of the defined protection objectives, which are determined by further use of the decontaminated area, and corresponding human sensitivities. In Germany, the efforts directed to elaboration of reliable long-term risk assessment methods resulted in developing several models of different applicability. The proposed approach to the assessment of the human risk potential originating from deserted industrial sites has been exemplified in a case study of the large-area soil contamination by mercury and antimony in Marktredwitz city, North Bavaria, FRG (Figures 13.1, 13.11). The study, conducted by the research group of the GSF-Institute of Ecological Chemistry, FRG, has been focused on a site-specific risk assessment and selection of the adequate preventive/remedial action. Unlike the studies in Sites I and II, oriented to one selected measurement endpoint (soil), this study was of a complex character: measurement endpoints included soil, water, air, sediments, dust, plants; while target risk receptors were human: adults and children.
Site Characteristics Site IV is a typical urban area of a city that started to develop in the industrialization period of the end of the eighteenth century as a residential area of one of the oldest plants in Germany, chemical factory Marktredwitz (CFM). Hence the central position of the plant in the town, which is surrounded by a railway (Figure 13.11). The CFM area adjoins the Kossene river course, which belongs to the Elbe River drainage basin. The river was regulated in e 1930s to intercept frequent floods. The reclaimed old riverbed is also adjacent to CFM. ain wind directions are Wand SW. Dominating types ofland development are individu I houses with gardens.
Sources of Heavy
M11 Contamination in the Area
The major sourc~ heavy metal contamination in the area is now the abandoned industrial site of a more than 200-year-old former chemical factory Marktredwitz (CFM) founded in 1788 (Figure 13.11) which used to produce a great variety of inorganic and organic chemicals, among them Hg- and Sb-based compounds and herbicides (Table 13.5) before closure in 1985 for ecological reasons. Maximum concentrations of heavy metals found in soils of the area (Table 13.6) reflect an extent of the environmental damage.
Heavy Metal Contamination in Industrial Areas and Old Deserted Sites
301
Figure 13.11. General map of Site III: Marktredwitz urban area with chemical factory site, North Bavaria, F.R.G. with the location of measurement points for Hg.
Table 13.5. Compendium of the CFM Catalog: Heavy Metal-Bearing Chemicals CFM - Products Inorganic and metalorganic products: Hg numerous products, under it Hg, Hg2C1 2, HgX2 (X=CI, Br, I, CN, SCN), Hg(N0 3h, HgO, RHgCI (R=CH 3, C2Hs, C6 Hs), etc. Sb potassium antimony tartrate, potassium antimony citrate As Hg3 (As0 3h, Hg3 (As0 4h
Zn
Zn 3 P2
Cu
3Cu(OHh·CuCl 2
Among 12 groups of CFM products, Hg-bearing chemicals comprised 10 different groups, of them Fusariol® accounted for 5%. One group represented Sb compounds, of which over 90% consisted of potassiu~ antimonyl tartrate (tartar emetic). Mercurybearing chemicals prevailed in the CFM production: the rate of Hg-products accounted for 94%, and Sb-products, 6 Yo. The Hg- and Sb-products delivery from CFM to customers increased since 1961 to 1982 from 67.0 to 116.1 t and from 0.7 to 9.6 t, respectively.
toi
®
Registered trademark of the Chemical Factory Marktredwitz, Inc., Marktredwitz, Germany.
302
Fate and Transport of Heavy Metals in the Vadose Zone
Table 13.6. Detected Heavy Metal Contaminants from CFM and Their Maximum Concentrations in Soil Metal C (mg kg- 1) Metal C (mg kg- 1 )
As
Cd
Co
Crt
Cu
Hg
684
1.64
18.8
560
673
6,140
Mn
Ni 390
Pb
Sb
Sn
Zn
19,364
36,400
13.2
2,532
770
Monitoring Strategy A complex character of the old contaminated site evaluation determined a broad program of preliminary investigations undertaken in the framework of the Marktredwitz project. Due to historically long-term impact of contaminants under the changing conditions of area development/management and extent of anthropopression, detailed preliminary studies were required to most accurately define the monitoring strategy. These studies comprised measured endpoints and risk receptors, sampling points and parameters assuring quality-safe evaluation of the site. The preliminary studies undertaken in the framework of Marktredwitz project (Site III) were focused on making explicit the factors of concern for site evaluation, in particular the kind and pathway of pollutants, past and future area development, availability, and adequacy of the existing database. The target task was to elaborate an optimum sampling and measurement program adequate for a reliable use-specific risk assessment. The studies comprised historical background investigation and elaboration of a geographical information system (GIS) for an investigated site. The objectives of the historical investigations were to identify precisely vs. time: (i) contaminants inventory: industrial products in the area, delivered amounts; (ii) pathways of contaminants: aquatic (surface- and groundwater), terrestrial (controlled and uncontrolled waste disposal and use), air (wet and dry particulate deposition); (iii) uncontrolled waste disposal and use as common fill or soil amendment; and (iv) causes of environmental damage from uncontrolled sources. To identify possible uncontrolled disposal of contaminated material in the Marktredwitz area, the inventory of industrial and other sites (e.g., quarries) where such material could have been disposed was elaborated. For this purpose also, the available archival aerial views were investigated to detect changes of the cityscape in time, where contaminated material could have been involved as a common fill (relocation of the river bed, urban area development, road construction, changes of land use, position and condition of industrial sites in time, land leveling). For the contaminated SIte eva tion, a GIS visualizing any kind of data with reference to their Gauss- Kruger or loca oordinates in the Windows style appeared to be particularly suitable. It served as a spa· ally allocated data bank of required information: general, pathways and input/output 0 contaminants (emission, imission, utilization), sampling and analysis (e.g., Figures 1 .1 L 13.12).
Survey of Transfer Pathways and Risk Receptors A sampling program was designed with use of the GIS spatially related database obtained from the preliminary historical investigations. It was focused on deriving compre-
,"",,,,,7)'1'\1
•
antimony (new values)
•
mercury
•
railway installations
-
river KOsselne
III
residential area
Metal Contamination in Industrial Areas and Old Deserted Sites
303
• Figure 13.12. Site III: Marktredwitz area with the location of circles D=130 m to define sampling paints for Hg and Sb.
hensive and reliable data for evaluation of the extent and propagation of contamination in the areas suspected of being polluted. In these investigations, a human as a risk receptor was a target assessment endpoint, while groundwater was not considered at this stage due to the lack of elevated concentrations of site-specific pollutants in drinking water. The area to be surveyed for Hg could be roughly estimated on the basis of the available qualitative information, while for Sb no such estimation was possible due to insufficient data and weak correlation between the occurrence both metals. The sampling area was thus planned, starting from the old CFM site and define according to the main transfer pathways; i.e., the Kosseine river flow (E) and predomina t wind direction (S, SW), though E direction of wind transport also occurs. As a joint) effect of the major pathways, in the E direction from the old CFM site, "hot spots" of the highest extent of contamination were expected, while in the W direction a somewhaylesser contamination could not be excluded. With the help of the geostatistical analysts' and available data on Hg, the distance of 130 m in diameter was found to be sufficient for the reliable evaluation of the contaminant expansion. On designing the sampling network, the maximum distances of sampling points accounting for 130 m were therefore generated by means of circles centered in proven contaminated points (Figure 13.12). The maximum distance of 130 m was assumed to be valid also for Sb. Thus, an extensive sampling of the surveyed area could be accomplished with a minimum effort. For the quantitative exposure assessment, all relevant transfer pathways comprising soil, food crops, indoor and outdoor ambient air were sampled. Soil samples for analysis were prepared through averaging of a sufficient number of random samples from the
304
Fate and Transport of Heavy Metals in the Vadose Zone
respective area, therefore the results represented mean values for the area. In total, about 200 areas were sampled. The sampling was performed in accordance with ISO/DIN10381. The soil exposed to air contamination was taken from the top layer 0 to 10 cm. The soil enriched with contaminants from a long-term impact was sampled also from the layer 10 to 30 cm, if risk receptors were children. Additionally, in some points layers of 30 to 60 cm were also taken. Sampled food crops grown in contaminated house gardens comprised mainly vegetables and fruits. Particulate samples were taken from the indoor ambient air in the living rooms and outdoor of the residential houses, and a fraction <2.5 pm (lung penetrating dust) was separated for analysis. In total, 800 soil samples, 200 plant samples, and 146 indoor air and dust samples were collected. In the Marktredwitz site, a human biomonitoring was conducted simultaneously with sampling the transfer pathways, i.e., soil, food crops, and indoor air. The residents of the areas supposed to be highly contaminated were invited to participate voluntarily in the investigations, submitting 24 hours urine and giving blood samples. In total, 264 volunteers participated in the monitoring, among them the most sensitive group: 0 to 4-yearold children accounted for 8 volunteers giving blood, and 21 giving urine. The collected 261 blood samples and 264 urine samples were examined for Hg and Sb. In conformity with the sampling program and a scope of the survey, the material analyzed for Hg and Sb comprised atmospheric particulate, soil, plants, river sediments, indoor and outdoor air, human blood and urine. Samples were analyzed for Hg and Sb content by standard methods using ICP-MS techniques (ICP Perkin Elmer Sciex ELAN 5000 coupled with a high-resolution quadruple mass-spectrometer Finnigen Mat). The results of the screening/monitoring survey were introduced into the GIS Geographical Information System. Currently, GIS is used as a spatially allocated data bank of required input information on the contaminated site. Further development of the system will comprise the integration of a geostatistical estimation of contamination with a quantitative estimation of exposition as equipotential maps.
Human Risk Potential Assessment Approach to Human Risk Potential Assessment For evaluation of Hg contamination in the Marktredwitz area, the method of usespecific and site-specific quantita ·ve exposure assessment (QEA) has been applied. The QEA method adopted by the Insti te of Ecological Chemistry is an individual sitespecific qualitative exposure estimati n model comprising standardized scenarios with defined uptake rates through the tra sfer paths for different kinds of soil utilization developed at the Institute Fresenius GmbH, Taunstein in Germany (Simmleit et al., 1997). The application and deve~pment of the QEA model were a part of the Marktredwitz case study. The essence of the model and the procedure to be followed for assessing the risk potential will be presented and discussed in the example of evaluation of Site III contamination by Hg in the background of a critical discussion of other risk assessment concepts. The Marktredwitz study and applied concept of risk potential assessment exemplifies a new site-specific approach which is aimed at finding a sound compromise between the economy and environment. To facilitate the understanding of differences and specificity
Heavy Metal Contamination in Industrial Areas and Old Deserted Sites
305
Area A - protect: unlimited, local general and multifunctional possibility of use, Area B - tolerate: limited, but site- and protected object-related possibility of use, Area C - remediate: area of toxicity, where a damage of the protected object (plants, animals, humans, ecosystems) might occur; protective means are necessary Figure 13.13. Definition of the Three Areas according to Eikmann, Kloke, and Lilhr (1991).
of this concept, the risk evaluating models most widely applied in Germany (ThreeArea-System and Critical Values by Ewers/LAGA) will be discussed below.
Fundarnental Remarks Soil contamination can be assessed in two essentially different ways: (i) use-nonspecific, protecting soil multifunctionality; (ii) use-specific, which considers further use of a decontaminated area. Most experts consider a use-nonspecific assessment unrealistic. Of site-related and use-specific methods, a Critical Value System by LAGA and ThreeArea-System or Eikmann/Kloke list are of particular significance. Other classical principles of contaminated sites evaluation are based on soil threshold values derived from experience and summarized in assessment lists (e.g., Dutch list, Swiss list). Though widely used by environmental engineers, only a few lists are toxicologically proven. An interesting approach for three-level risk assessment was proposed by Gupta et al. (1996).
Three-Area-System The Three-Area-System, known also as an EikmannlKloke list (Eikmann et al., 1991) has been elaborated to provide a risk assessment based on the land use and sensitivity of objects to be protected against soil contamination. Soil pollution is attributed to three increasing levels (A, B, C) prescribing adequate actions (Figure 13.13). The kind of action following the determination of soil contamination is divided into three categories, depending on the seven soil utilization modes and related risk receptors (e.g., playgrounds for children, house gardens). The latest issue of the Three-Area-System (Eikmann and Kloke, 1993) comprises a list of 19 heavy metals, 5 nonmetal and 3 chloroorganic compounds, while protection measures are focused on human. The system, nevertheless, uses empiric compilation of soil values (e.g., for Hg: Table 13.7) and does not incorporate soil parameters into the EikmannlKloke list. Therefore, It n be considered as a guideline for use-specific, but site-nonspecific assessment without to icological background.
Use-Specific Critical Values by EwerslLAGA (Ewers and ~iereCk-GOethe! 1993) The use-specific critical values (PW) and threshold value~hndicating the necessity of preventive actions (MW) proposed by the Ewers/LAGA e~ert committee, contrary to those of the Three-Area-System, are based on the toxicological data (Hassauer et al., 1993). The respective values for mercury are presented in Table 13.8. However, these values also display some weak points, the major of them being: (i) entirely exposure mode "oral soil intake" by the most sensitive risk receptors (little children) is considered;
306
Fate and Transport of Heavy Metals in the Vadose Zone
Table 13.7. Guide Values for the Use-Dependent Evaluation of Soils Polluted by Hg, According to Eikmann and Kloke (1993) Use
Protected Object
BW
mg Hg kg- 1
Multifunctional
BWI
0.5
Playground for children
human beings, others human beings
2
Private gardens
human beings
3
Sport grounds and football fields
human beings
4
Parks and leisure grounds
human beings
5
Industrial and trade areas
human beings
6
Agricultural land, orchards and market gardening Ecosystems not used agriculturally
others
BW II BW III BW II BW III BW II BW III BW II BW III BW II BW III BW II BW III BW II BW III
0.5 10 2 20 0.5 10 5 15 10 50 10 50 10 50
No.
0
7
others
Table 13.8. Use-Specific Critical Values (PW) and Cleanup Threshold Values (MW) for Assessing Soil Contamination by Hg, According to Ewers/LAGA (Ewers and ViereckGoethe,1993) Playground for Children PW (mg Hg kg- 1) MW (mg Hg kg- 1)
Residential Area
4
8
10
20
Parks and Leisure Grounds
20
(ii) threshold values indicating the necessity of preventive action are based upon the rates of resorbed pollutant dose, while currently there is no standard method for evaluation of resorbable fraction in soil; (iii) the values do not provide an evaluation of "risk situation" in the legal and juridical sense. Hence, the values by Ewers/LAGA are of an entirely informative character and hardly meet the requirements of recommendations for preventive/remedial action.
Applied Model: Quantitative Exposure Assessment (QEA) Bosic Concept /
A concept of quantitative exposure assessme (QEA), applied for evaluation of soil contamination by Hg and Sb in the old site of t e Chemical Factory Marktredwitz is aimed to avoid the simplifications and disadvantag s of the approaches presented above. The method is comparable to the UMS model; that is a German abbreviation of "Toxicological Assessment of the Human Exposure to ollutants from Contaminated Sites" (Simmleit et aI., 1997; Hempfling et aI., 1991). he exposure scenarios accommodated by different federal expert groups were recently published by Stubenrauch et al. (l994a, b).
Heavy Metal Contamination in Industrial Areas and Old Deserted Sites
307
For the evaluation of individual areas in the vicinity of CFM in the Marktredwitz site evaluation, the QEA model modified for the most sensitive land-use scenario, "Living Area with a House Garden," has been used in accordance with the ARGE Foconl Fresenius Institute. It follows the gradual procedure, from exposure estimation and toxicological data (Figure 13.14A) to calculating a risk value describing the degree of hazard (Figure 13.14B). In Table 13.9, the transfer/exposure pathways of Hg to two groups of risk receptors within the scenario "Living Area with a House Garden" and intake rates used for QEA calculations are incorporated. The intake rate of pollutants by human from a particular area has been determined by three factors: (i) frequency and mode of land use; (ii) pollutant concentration in contact media (soil, air, and plants); (iii) ubiquitous pollutant concentration in foodstuffs and contact media, including drinking water. Ubiquitous background contamination of human (UBI) refers to an average intake rate from foodstuff, drinking water, and air. To evaluate the total risk value (RW), UBI values should be added to the site-specific intake rate. There is, though, still lack of statistically confident UBI data; consequently, just estimated value can be used. The reported data on Hg intake and resorbed dose vary considerably. WHO (1990) estimates a total resorbed Hg concentration in the range from 0.08 pg kg- 1 d- 1 up to 0.28 pg kg- 1 d- 1 per 70 kg body weight. In Germany, an overstated UBI value of 0.08 pg kg- 1 d- 1 for Hg was accepted for land use modes such as playground for children, park and sports ground (for protective reasons). At present, taking into account the quality of foodstuff of a proper origin and the indoor residence time, the UBI value for the use scenario "Living Area with a House Garden" has been reduced by the Fresenius Institute to 0.025 pg kg- 1 d- 1• To evaluate tolerable resorbed dose of Hg, a resorption rate of 7% for oral and 80% for inhalative intake of inorganic Hg was assumed (WHO, 1991). As all available data from the Marktredwitz area refer to total Hg concentrations, the provisional inorganic Hg values were estimated as 0.1 % Hgt in soil and 10% Hgt in air, according to analyses by GKSS/LAGA (1993). The observations of methyl-mercury are not actually necessary, unless the concentrations of organic Hg exceed 0.05 pg kg- 1 d- 1• The provisional guide values for Hg used by Hassauer et al. (1993) for the toxicological evaluation of total body dose rates were derived from the lowest observed adverse effect level (LOAEL) during the subchronical and chronical animal testing and the lowest observed impact level of an epidemiological study. These values were defined as "... total body doses of a pollutant, which-at the state of the art-on their own do not exhibit adverse effects on health, or give only little rise to inducing health risks" (Hassauer et al., 1993). The provisional guide values used in this study accounted for 0.08 pg kg- 1 d- 1 (TRD oral) and 0.07 pg kg- 1 d- 1 (TRD inhalative). These values are supposed to be questionable and subject to chan at t lead to different results of evaluation. Assessment of risk caused by the presume 'te contamination is aimed at evaluating actual and potential risk situations with regard tO~he legal and juridical term "risk." This term defines a situation that would provoke a da age, i.e., a violation of the integrity of protected objects if no preventive actions are take (LAGA, 1991). In order to evaluate exposure, uptake pathway specific risk indices are calculated as a ratio of an intake pathway-specific dose (PDI) to tolerable absorbed dose (TRD). The sums of indices calculated this way are then being referred to the ubiquitous background
308
Fate and Transport of Heavy Metals in the Vadose Zone
Quantitative Exposure Assessment for the Scenario "Living Area with a House Garden" (According to HEMPFLING et aI., 1994)
Determination of the relevant exposure pathways
soil- oral;
plant - oral;
indoor air - inhalative;
indoor dust - inhalative
Determination of the pollutant concentration in contact media
/
\
/
plant
soil
indoor air
'"
indoor dust
Calculation of the daily intake rate (DlR) of contact media
----
user groups: -little children - adults
different exposure pathways
Calculation of the potential daily resorbed intake dose (POI) for
"
/
different user groups
different intake pathways
Figure 13.14A. Flow c rt of the q titative exposure assessment (QEA) for Hg in the neighborhood of the former CFM for the expos e scenario "Living Area with a House Garden: Calculation of the potential daily resorbed intake ose POI.
!
I
Heavy Metal Contamination in Industrial Areas and Old Deserted Sites
309
Quantitative Exposure Assessment for the Scenario "Living Area with a House Garden" (According to HEMPFLING et al., 1994)
Risk assessment by calculating the specific risk indices (RI) for each intake pathway using the
~
~
potential daily resorbed dose (POI) for the inhalative and oral intake pathway
tolerable resorbed pollutant dose (TRD) for the inhalative and oral intake pathway
calculation of the total risk index (GRI) by summing up the intake pathway specific risk indices considering the background burden from food, water, air,
Calculation of the risk value (RW) as a base of assessment RW = risk index: risk factor (where the risk factor (GF) depends on the security factors used fOI derivation of TRD-values)
Deduction of site- and use-specific recommendations dependent on
/
\
the background burden
quality of the data base
a risk cannot be excluded, however it is in the range of the ubiquitary background burden
0,25
$
RW" 1
RW
a growing concern must be considered; preventive measures, respectively changes of use, are to be examined
the probability of heath impact is high; remediation action is necessary
Figure 13.148. Flow chart of the quantitative exposure assessment (QEA) for Hg in the neighborhood of the former CFM for the exp~eDario 'Living Area with a House Garden: Calculation of the risk value RW. / .~
\
310
Fate and Transport of Heavy Metals in the Vadose Zone
Table 1 3.9. Survey of the Exposure Pathway-Specific Intake Rates for the Scenario "Living Area with a House Garden" Establishing the Pathway Specific Intake Rate User Group
Exposure Medium Exposure Pathway
BWa
Daily Intake
Exposure Frequency
Intake Rate (average per year)
Little children
soil - oral plant - oral indoor air indoor dust plant - oral indoor air indoor dust
10 10 10 10 70 70 70
1g 3,7 g d-1 ,b 3 mL d-1 3 mL d- 1 20 g d- 1 ,b 20 m d-1 20 ml d-1
200 d a-1 all-year 21 h d-1 21 h d- 1 all-year 21 h d- 1 21 h d-1
55 mg kg- 1 d-1 370 mg kg- 1 d-l,b 0,26 mL kg- 1 d-1 0,26 mL kg- 1 d-1 ,c 290 mg kg- 1 d- 1 0,25 mL kg- 1 d-1 0,25 mL kg- 1 d-l,c
Adults
inhalative - inhalative inhalative - inhalative
kg kg kg kg kg kg kg
BW: body weight. b Homegrown fruits and vegetables (dry weight). C Dust retention: 75%.
a
concentration (UBI). This method, however, is not sufficient for deriving well-founded recommendations of protective measures related to the specific risk receptors. To fill the gap between the toxicological statement on the one hand and the legislative requirements on the other, Konietzka and Dieter (1994) proposed to assess a risk threshold value with regard to the safety factors used for determination of the TRD values. The safety factors are expected to be slightly below the LOAEL of sensitive individuals. According to this proposal, risk indices are transformed into risk values depending on the reliability of data used for estimation of provisional guide values. Estimation of the provisional guide value for the oral uptake of inorganic mercury was based upon oral uptake data for rats, with a safety factor of 200, while the respective values for inhalative uptake of inorganic Hg were derived from LOAEL for humans, with a safety factor of 20. The Fresenius Institute determines risk threshold values by using provisional guide values for oral uptake with a safety factor of 10 and for inhalative uptake with a safety factor of 4.5, referring to Konietzka and Dieter (1994). A risk value (RW) calculated as the total risk index to the risk factor ratio, allows distinguishing between the lack of risk and possible risk for the receptors. To describe the situation for the land-use scenario "Living Area with a House Garden," weighted means are used as risk factors for oral and inhalative uptake. In Figure 15.15, the three RW ranges for Hg of a different probability of risk situations requiring adequate actions are presented. At the lowest stage (RW < 0.25) the risk cannot be entirely neglected, but no specific action is required. At the highest stage of risk, preventive or remedial action has to be taken.
Estimation oj Soil Values jor a Scenario "Living Area with a House Garden. " Marktrec1witz Area (Data up to 1992) The risk assessment using I sensitive scenario "Living Area with a House Garden" applied in Mark edwitz reqUl s database quality for transfer/uptake pathways (soil, plant, indoor concentrations) be evaluated. From Figure 13.16 one can conclude that soil was the only medium exa ined in 61 % of all the areas (quality level
Heavy Metal Contamination in Industrial Areas and Old Deserted Sites
0,25
:s;
< <
RW RW RW
;:0:
0,25 1,0 1,0
311
hazard in the range of the average background burden risk with a growing concern danger (in the sense of police regulations)
Figure 13.15. Scheme of risk values.
PER:ENTAGE OF SINGLE AREA
QUALITY OF DATABASE
70 60 50 40 30
20
10 0 A ANALYTICAL DATA FDR All
B
----+
C
D ANALYTIC AL
DB:RfASING DEGREE OF QUALITY
EXPDSURE PATHWAYS
DATA DNlY
FOR ONE EXPOSURE PATHW AT
Figure 13.16. Database quality for a single area QEA for the scenario "Living Area with a House Garden': decrease of quality from A to 0, expressed in percent of the area surveyed for Hg in soil.
D). The problem of the result compatibility in time for different uptake pathways in individual fields was not considered. The results of an exposure assessment carried out within the project "Site-specific assessment of the contamination in the vicinity of the Marktredwitz Chemical Plant," displayed a virtual significance of the indoor inhalative uptake path for the used scenario. Quantitative exposure assessment for individual areas was accomplished using either actual measurements or statistical values generated from indoor analyses. The indoor concentrations of mercury were found to be strongly influenced by the living habits of residents (e.g., frequency of room cleaning, airing, etc.) and also dependent upon the time of sampling, conducted in 1987 to 1991, due to the altered exposure conditions (demolition and removal of contaminated CFM buildings). Another problem was a small amount of indoor data available. These reasons led to the concept of using indoor data as statistical parameters for evaluating Hg concentrations in air and indoor suspended dust for the QEA of the single areas. Mercury concentrations in food/fodder crops have been also evaluated on a unjformst:atis . al basis and then used to calculate critical values. Consequently, Marktredwitz site-speci IC critical values of Hg concentrations in soil were estimated in a toxicologically base way. Risk values for soil contamination by Hg, derived from the QEA single area asse~~ment for the land use scenario I
I
312
Fate and Transport of Heavy Metals in the Vadose Zone Deduction of site specific assessment values by means of quantitative exposure assessment (scenario: "Living Area with a House Garden")
Formulation of the question: IcBo = f(RW) atgivenRwl
RW
= 0.25
(ascertained empirically)
total risk index
specific risk index for each pathway
GRI
I
1.0
0.75
0.5
Ri (inhalative)
RI (oral)
potential daily intake dose POI pathway
RI pathway
Input: Pollutant concentration in contact media taken from statistical analysiS of site-specific data (90 th percentile, average value)
....
c plant = f (c SOil)
C plant
=constant
C indoor air
=constant ~#
c indoor dust = constant
~exposure scenario (living area with house garden) I·
oral:
soil plant
little children little children, adults
drinking water
not considered in Marktredwitz
indoor air and dust
little children, adults
inhalative: Ir
mercury concentrations in soil depending on the risk values for the receptor group little children:
33
78
127
176
assessment value (Hg)
RPW
MSW(I)
MSW(II)
G
RW (risk value)
0.25
0.5
0.75
1.0
C
IOit
[mg Hg/kg soil]
Figure 13.17. Deduction of site-specific assessment values by means of quantitative exposure and risk assessment. RPW: risk testing value; MSW (I, 11): low- and high frequency examination threshold values; G: value indicating danger. The calculation of these assessment values was carried out in cooperation with the Institute Fresenius Ltd., Taunusstein.
"Living Area with a House Garden," served as a basis for the evaluation of this land-use mode (Figure 13.17). As the estimated soil values (33 to 176 mg kg-I) were so-called "field means" (FDW) and incorporated statistical error of 74%, for exposure calculations, the corrected soil values were used as input dat~lues for the user/risk receptor group "little children" were calculated for two versions of i\ut data for Hg in plants: (i) Hg concen-
Heavy Metal Contamination in Industrial Areas and Old Deserted Sites
3 I3
tration in food/fodder plants is proportional to the relevant concentration in soil; and (ii) constant Hg concentration in plant = 90% of the average for useful plants per field value. In both versions, concentrations of Hg in all species of useful plants, suspended dust and air were considered as input data. The critical values estimated for both versions are presented in Table 13.10. These site- and use-specific critical values, based on the data up to 1992 and considering the variability of on-site concentrations, appeared to be surprisingly high and in most cases significantly exceeded the concentrations of Hg in Marktredwitz soil. These values, though, are consistent with the results of a human biomonitoring. No markedly heightened mercury concentrations were found in numerous samples of body fluids (human blood and urine) taken in 1987 and 1982. The aforementioned critical values were derived from the toxicological models with a high factor of confidence with respect to the body transfer pathways, i.e., resorption of the pollutant (Hg). Consequently, only higher Hg concentrations would cause distinct physical damage. The lack of an adequate amount of data for children should however, be pointed out (Figure 13.18). Only five of them were in the age range of 2 to 4 years, and just one of them was checked for Hg concentrations in soil. For this addressee, a field mean including analytical error of 74% is below 33 mg Hg/kg. It should be added that only a few of the examined group were living in heavily contaminated localities, and the information concerning habits and living conditions was limited. Another human monitoring conducted in 1995 tried to avoid these weak points. The RPW and G values derived for the land-use mode "Living Area with House Garden" have three basic advantages: (i) they have a toxicological fundament, being based upon the exposure analysis and toxicologically related Hg-TRD values by Hassauer et al. (1993); (ii) they enable differentiation between lack of a risk and a possible danger for risk receptors; and (iii) they consider a site-specific situation in Marktredwitz. Four critical values derived site-specifically are linked to the different Hg concentrations in soil. The risk value of 0.25 corresponding to 33 mg Hg kg- 1 in soil already considers a heightened site-nonspecific background concentration. It should be emphasized that the differentiation between RPW and G values has no toxicological reason. Their aim is just to define the time interval of taking measures. The interpretation of soil values corresponding to the different risk level is given in Figure 13.19. On application of these values to the evaluation of the Marktredwitz site based on the land-use mode "Living Area with a House Garden," some insufficiencies had to be considered. The most important of them were: (i) lack of a complete set of data for the area adjacent to CFM, that had to be substituted with average means; (ii) many soil values were acquired using averaged samples for a whole evaluated area; (iii) the toxicological evaluation of Hg contamination was subject to change. Therefore, the evaluation of contamination by Hg should be modified accordingly.
Human Biomoniioring Data Besides models of exposure assessment, a quality-safe risk assessment requires human biomonitoring data to be incorporated into investigations. These data enable an estimation of the internal exposure of a lrGm~n to ntaminants. To evaluate properly the internal exposure data originating from the soil co tamination, other sources con-
314
Fate and Transport of Heavy Metals in the Vadose Zone
Table 13.10. Deduction of Site-Specific Assessment Values for the Scenario "Living Area with a House Garden," Based on Uniform Data; the Assessment Data Underlaid with Gray Were Used to Evaluate House Gardens in Marktredwitz
Risk Values RW
Assessment Values
= 0,0084 • c 5011 (mg Hg kg-1 soil) User Group: Little Children 1 C soil (mg Hg kg- soil) Variant 1
0.25 0.5 0.75 1.0
RPW Hg a MSWHg(l) b MSWHg(1I) GHg c
33 78 127 176
C plant
a b C
Cplant
=
0,016 mg Hg kg-l plant User Group: Little Children 1 C soil (mg Hg kg- soil) Variant 2
34 87 139 191
RPW: risk testing value. MSW (I, II) low- and high-frequency examination threshold values. G: value indicating danger.
NUMBffiOF PROBAMlS
AGE STRUCTURE OF PROBANDS BLOOD SAMPLING 428
450~------------------------------------------,
400
350 300
250 200
150 100
50
6
10
<4
4-7
7-10
10-20
20-50
>50
AGE (YEARS)
Figure 13.18. Age structure of probands of the human biological Hg-monitoring in the surroundings of CFM (blood samples taken between 1987 and 1992).
tributing to the exposure have to be considered and differentiated (Table 13.11). The human biomonitoring data compared to the toxicologically proven reference values serve as a base for recommendations. The evaluation of human biomonitoring data is still in progress and subject to a comprehensive statistical analysis when completed. Here, the partial results are presented and discussed. The concentrations of Sb in the blood and urine were found to be predominantly at the edge of a detectable level, and only in a few samples occurred in the range of a few pg dm -3, although concentrations of Sb in soil were in some areas very high. It can therefore be concluded that high soil contamination by Sb has not been reflected in its occurrence in the body fluids (Figure 13.20). The reasons for it were a short Sb half-period of 70 hours in a human body and a comparativjYJow-fltHn~er of probands in the age when
/
Heavy Metal Contamination in Industrial Areas and Old Deserted
315
Interpretation of the site specific assessment values for the mercury concentration in soil c soli (= average value of an area plus additionally 74 % for potential measurement errors) deduced by means of QEA on a uniform data base for the mercury concentrations in the contact media (analytical data gained from measurements between 1987 and 1992): scenario:"Living Area with a House Garden" I site: Marktredwitz risk testing value:
RPWH9
=
mg Hg/kg soil
33
C soli < RPWHg (RW < 0,25) mercury concentration in soil, up to which no human health hazard above the average· is expected
RPWH9 S C soli < MSWHg(l) (0,25 S RW < 0,5 mercury concentration in soil, posing a stress to human health over the average without short- or medium-term need of actions. low-frequency examination threshold value I
MSWHg(l) =
78
mg Hg/kg soil
MSWHg(l) S c soli < MSWHg (lI) (0,5 S RW < 0,75) mercury concentration in soil from which investigations and health examinations should be started: an increasing stress for the human health occurs. low-frequency examination threshold value II
MSWHg(lI) = 127
mg Hg/kg soil
MSWHg(lI) SCIoli < G H9 (0,75 S RW < 1,0) mercury concentration in soil from which frequent investigations and health examinations should be started:checking if danger is looming. value indicating danger:
= 176
mg Hg/kg soil
csoll ~ G H9 (RW~ 1,0) mercury concentration in soil from which a high probability of danger for health occurs: remediation of contaminants impact should be started immediately
Figure 13.19. Interpretation of site-specific risk assessment values for the mercury concentration in soil.
they ingest soil (crawling infants). Therefore, an important transfer pathway did not contribute to the exposure. The problem of a low participation of the most sensitive group of population in blood sampling is common for almost all investigations, while the participation in urine sampling is definitely higher. The alternative sampling of hair as a sink for Sb was renounced due to the methodological and time problems. The results of the blood and urine analysis for Hg were categorized in accordance with the FRG Health Ministry proposals (Figure 13.21). The measured values appeared to differ significantly from those obtained in the frameworks of the German extensive environmental survey (UWS) conducted in 1990 and 1992 (Table 13.12). The Hg contents in blood of probands from the Marktredwitz area were distinctly elevated in comparison with the UWS, though none of the probands was classified with the category III. The preliminary results show therefore, that soil contamination by Hg is represented in the blood of the inhabitants, while the data on urine did not differ considerably from the UMS results.
316
Fate and Transport of Heavy Metals in the Vadose Zone
Table 13.11. Factors to Be Considered for Human Biomonitoring" 1. Choice of appropriate parameters • relevant contaminants • toxicokinetics, investigation material • consideration of other important possibilities of exposure with respect to the contaminants to be investigated (profession, tobacco smoking, amalgam fillings, etc.) 2. Choice of the probands • criteria of choice • criteria of exclusion • investigation of a group of comparison • criteria of choosing a group of comparison 3. Recording of data relevant for an estimation (questionary) • personal data (age, gender, separately: name and address) • time of living, living conditions • profession(s) • smoking habits • personal customs and behaviors • use of the garden • consumption of the grown vegetables and fruits 4. Criteria of judging the measured data of human biomonitoring • reference values • orientation values • toxicologically derived values of exposure referring to the effect (values of human biomonitoring) a
According to Ewers and Suchenwirth, 1996.
The presented investigation and evaluation of the deserted industrial area (Site III) shows that quality-safe assessment can be accomplished with limited efforts, provided an adequate investigation procedure as well as site- and use-specific exposure assessment is applied. The complex investigation procedure should comprise detailed historical studies aimed at identification of possible sources of contamination and characterization of the site; visualization of data by the geographic information system; and estimation of the results with use of adequate methods. Of the methods currently in use, models for the Quantitative Exposure Assessment (QEA) appears to be the best fit to the purpose of site- and use-specific estimation of human risk potential originating from contaminated soils. Estimation based entirely on the transfer pathways can lead to an overestimation of the actual exposure; for the reliability of the evaluation, a strong linkage between the toxicologically proven threshold values for the contaminant uptake and the results of the exposure assessment is indispensable. As such values are available only for a limited number of contaminants, the best solution to the problem is a human biomonitoring which provides empirical data on the actual internal burden of human risk receptors.
REMEDIAL CONCEPTS The environmental issues exemplified in case studies show the need for a differentiated approach to the selection of appropriate, site- and use-specific remedial actions. On remedial action analysis, based on estimated extent of contamination by heavy metals
Heavy Metal Contamination in Industrial Areas and Old Deserted Sites
317
MJM BER OF PROBANCS
119 120 100 80
60 40
20
3
a It)
o
a
1
a
a
a
a
a
a
N
It)
Mit)...,.
It)
It)
It)
N
a ..;
a
a
a
a
a
1
"4' iii ANTIMONY CON::ENTRA TION IN BlOOD [II9/L]
NUMBEROF PROBANJS 90 80
70 60 50 40 30
20
10
a 0.1
0.2
0.3
0.4
0.5
0.6
a
a
a
0.7
0.8
0.9
a 1.1
1.2
ANTIMONY CONCENTRATION IN URlt£ [1I91L]
Figure 13.20. Antimony concentrations in blood and urine samples.
The categories for evaluating the contents of heavy metals in human samples (Krause et aI., 1987), edited by the Institute of Water, Soil and Air Hygiene, are referred to the measured values in blood and urine. Three categories have been distinguished: Category I
in nonrisky value
Category II
higher value, no evident health risk, examination is recommended
Category III
distinctly elevated value, a health risk cannot be excluded in the long run, cleanup is necessary to remove or at least reduce the sources of a risk
Figure 13.21. Estimation based on the guide values for human
biomonitorin~~ "\
318
Fate and Transport of Heavy Metals in the Vadose Zone
Table 13.1 2. Comparison Between the Human Biomonitoring in Marldredwitz in 1966 and the Environmental Survey (UWS), 1990-1992 Studies
Number of Persons
Measured Value Hg in blood (lJg L-')
CFM-Investigation
UWS (1990-1992)
134 3958 Hg in urine (lJg L-')
CFM-Investigation
UWS (1990-1992)
134 3958
Category I
Category II
Category III
<3
3-10
>10
83,6% (112) 97,9%
16,4% (22) 2,1%
0% (0) 0,03%
<5
5-20
>20
100% (134) 96,8%
0% (0) 3,0%
0% (0) 0,2%
and assessed requirements for such action (further use of the land, groundwater rotection) , a set of efficient and cost-effective options elaborated also by the auth rs was considered. These options are in line with the current trend to developing metho s, which clean up the soil without destroying its properties and fertility, be applied· situ. Such methods are generally very cost-effective. To a great extent, they are based on the vulnerability to mobilization of metals bound onto mobile or easily mobilizable phases. Removal of metals bound in these fractions would render the soil harmless in the most effective way. Basic options that have been analyzed with regard to the cost-effective application for trace metal control in the large-area contaminated sites was polluted soil sanitation by leaching methods. Studies conducted within this task comprised the selection of the novel effective extractants. The reuse of waste material as extractant was also considered. Here, only general concepts are presented. The detailed discussion of methods and data are subject to being published elsewhere, part of them having already been published (Fischer et al., 1994; Leidmann et al., 1995). With regard to the remedial actions for decontamination of soils and prevention of contaminated water infiltration from the waste layer to the groundwater, an application of agricultural and hydrolyzed food engineering residues (silage effluents, residues of a brewing industry, slaughtering offal) have been found promising. These residues can be used either directly (Leidmann et al., 1995) or as raw materials for preparation of extractants for the sanitation of metal polluted soils by leaching method (Fischer et al., 1994). The extraction rates obtained with grass silage juice from two adsorbents (bentonite and peat) were: Cd 74.7%, Zn 55.7%, Cu 53.5%, Ni 38.9%, Cr 12.7%, and Pb 8.9%. The efficiency of metal extraction with hydrolysate applied to contaminated soils was adequate: Cu 50.3% and Ni 38.7%, at initial concentration of metals in soils 279 mg Cu kg- I and 54 mg Ni kg-I. Many other natural or reused materials are now under investigation by numerous authors, in order to apply them either as metal extractants or adsorbents. These options offer the benefit of both the decontamination of soils and leachates and the treatment/ disposal of the waste material in an ecologically tolerable way. Growing interest in mild cleanup technologies results in extending their practical use. Many of these technologies are currently a routine practice, while others need more research and efforts to make them practically applicable.
Heavy Metal Contamination in Industrial Areas and Old Deserted Sites -
--
----_
.. - - - - ...........
- ...- - - - - . - - - - - - -
319
SUMMARY From the presented case studies, investigations, and evaluation concepts, several general conclusions can be derived, which are summarized as follows. Optimization of monitoring programs for heavy metal contaminated site evaluation should be based on a use-specific and site-specific approach. Among the required data on factors and criteria for evaluation of contaminated sites, vertical distribution of a metal in the soil profile, species mobility/availability to the risk receptors and its possible transformations in time are to be considered for estimation of scientifically proven threshold values. The siteand use-specific, exposure-based evaluating model for assessment of the human risk potential originating from old landfills and deserted industrial sites appears to provide a reliable quality-safe estimation of the actual risk resulting from a contaminated site. The remedial methods using appropriate waste material as heavy metal extractant in a controlled, environmentally friendly way are promising cost-effective options for decontamination/preventive actions preserving soil properties and fertility.
REFERENCES ASTM Designation: D 5233-92. Standard Practice/or Nitric Acid Digution 0/ SoLid WlUte, 19 133-135. Barth, D.S. and B.J. Mason. SoiL SampLing QuaLity AJdurance UderJ Guide. Cooperative Agreement CR 81050-01, U.S. EPA Environmental Monitoring Systems Laboratory, Office of R&D, Las Vegas, NY, March 1984, p. 104. Chief Statistical Office. Environment Protection 199J. In/ormation and StatuticaL Data, Warsaw, 1993, p. 450 (in Polish). Chief Statistical Office. Environment Protection 1995. Information and Statistical Data, Warsaw, 1995, p. 490 (in Polish). Eikmann, Th. and A. Kloke. Nutzungs- und schutzgutbezogene Orientierungswerte fur SchadStoffe in Boden-Eikmann-Kloke-Werte, Kennzahl3590. BOdendchutz - erganzharu Handhuch, D. Rosenkranz, G. Bachmann, G. Einsele, and H.-M. HarreB, Eds., Erich-Schmidt-Verlag 14. LEg. Xl93, Berlin, 1993 (in German). Eikmann, Th., A. Kloke, and H.P. Luhr. Grundlagen und Wege zur Ermittlungvon Bodenwerten fur das Drei-Bereiche System, in IWS-Schriftenreihe Band 13: Ahleitung von Sanierungdwerten/iir Icontaminierte BO'den, Erich Schmidt Verlag, Berlin, 1991, pp. 279-360 (in German). Ewers, U. and L. Viereck-Goethe. Ableitung von wissenschaftlich begruendeten nutzungs- und schutzgutbezogenen Pruefwerten fur Bodenverunreiningungen, Bericht im Au/trag dU BayStMLU in Vergindung mit der Arheitgruppe Prue/werte dU LAGA-AuddchUdded "ALtlMten. "Hygiene-Institut des Ruhrbereites, Gelsenkirchen, 1993, p. 79 (in German). Ewers, U. and R. Suchenwirth. Expositionsabschatzung: Human-Biomonitoring Modellrechnungen. UWSF-Zeitdchrift/iir UmweLtchemie und OlcotoxilcoLogie 8(4), pp. 213-220, 1996 (in German). Fischer, K., P. Riemschneider, D. Bieniek, and A. Kettrup. Food engineering residues: Amino acid composition of hydrolysates and application for the decontamination of metal polluted soils. FrueniUd J. AnaL. Chem. 350, pp. 520-527, 1994. Forstner, U. Riverine and estuarine sedimentation of pollutants and leaks from sludges and wastes: Analytical, prognostic, experimental and remedial approaches. Land Degradation RehahiL. 4, pp. 281-296, 1992. GKSS/LAGA. Unterduchung ded Gefahrdungdpotentia& von queclcdiLherlcontaminierten Standorten in Bayern. Olctoher 1992 Forschungsbericht im Auftrag des Bayerische Landesamt fur
320
Fate and Transport of Heavy Metals in the Vadose Zone
Umweltschutz. GKSS-Forschungszentrum Geesthacht und Landesgewerbeanstalt Nurnberg, 1993 (in German). Gupta S.K., M.K. Vollmer, and R Krebs. The importance of mobile, mobilizable and pseudo total heavy metal fractions in soils for three-level risk assessment and risk management, Sci. TotaL Environ. 178, pp. 11-20, 1996. Harrison, RM., D.P.H. Laxen, and S.J. Wilson. Chemical associations oflead, cadmium, copper and zinc in street dust and roadside soils. Environ. Sci. Techno!. 15, pp. 1378-1383, 1981. Hassauer, M., F. Kalberlah, J. Oltmanns, and K. Schneider. Basisdaten Toxikologie fur umweltrelevante Stoffe zur Gefahrenbeurteilung von Altlasten. UBA-Berichte 4/93, E. Schmidt, Berlin, 1993 (in German). Hempfling, R., L. Heming, and S. Stubenrauch. Quantitative Exp0.JitiolLJaIJdchiitzung und-bewertung fur dad Umfet;} der Chemi.lchen Fabrik Marktredwitz auf Badi.l von GSF-au/bereiteten Daten. Februar 1994. Bericht im Auftrag des GSF - Forschungszentrums fur Umwelt und Gesundheit. Institut Fresenius GmbH, Taunusstein, 1994 (in German). Hempfling, R, S. Stubenrauch, U. Mayer, and S. Simmleit. Fallbeispiele fur die Altlastenbewertung mittels UMS, inALtfMtenbewertung-DatenanaLY.Je und Gejahrenbewertung, S. SChUlte-,stede, R Freitag, A. Kettrup, and W. Fresenius, Eds., ECOMED-Verlag, Landsberg, 199 (in German). Kabata-Pendias, A. and H. Pendias. Trace Element.J in Soil! and Plant.J, 2nd ed., CRC ress, Inc., Boca Raton, FL, 1992, p. 365. ntial Extraction Kaszycki, C.A. and G.E.M. Hall. Application of Phase Selective Methodologies in Surficial Geochemistry, in EXTECH I: A MuLtidi.lcipLinary Approach to MadJive SuLphide Re.Jearch in the RUJty Lake-Snow Lake GreenJtone Belt.J, Manitoba. GeowgicaL Survey of Canada, BuLL. 426, G.F. Bonham-Carter, A.G. Galley, and G.E.M. HalL Eds., 1996, pp. 155-168. Kersten, M. and U. Forstner. Chemical fractionation of heavy metals in anoxic estuarine and coastal sediments. Wat. Set: TechnoL. 18, pp. 121-130, 1986. Kersten, M. and U. Forstner. Assessment of Metal Mobility in Dredged Material and Mine Waste by Pore Water Chemistry and Solid Speciation, in Chemi.ltry and Biowgy of SoLid Wadte. Dredged MateriaL and Mine TaiLingJ, W. Salomons and U. Forstner, Eds., Springer-Verlag, Berlin/Heidelberg, 1988. Konietzka, Rand H.H. Dieter. Kriterien fur die Ermittlung gefahrenverknupfter chronischer Schadstoffzufuhren. GeJundheitJweJen 56, pp. 21-28, 1994 (in German). LAGA, Erfassung und Sanierung von Altlasten. Mitteilungen der LA GA. Bd.35, E. Schmidt Verlag, Berlin, 1991 (in German). Leidmann, P., K. Fischer, D. Bieniek, and A. Kettrup. Chemical Characterization of silage effluents and their influence on soil bound heavy metals. Intern. J. Environ. Ana!. Chem. 59, pp. 303-316, 1995. McGrath, D. Application of single and sequential extraction procedures to polluted and unpolluted soils. Sci. TotaL EflIJiron. 178, pp. 37-44, 1996. McLaren, R.G. and D.V. Crawford. Studies on soil copper. I. The fractionation of copper in soils. J. Soil Sci. 24, pp. 172-181, 1973. Prudent P., M. DomeizeL and C. Massani. Chemical sequential extraction as decision-making tool: Application to municipal solid waste and its individual constituents. Sci. TotaL Environ., 178, pp. 55-61, 1996. Quevauviller, Ph., Ed. Leaching/extraction tests for environmental risk assessment, Special issue, Sci. TotaL Environ. 178, p. 145, 1996. Simmleit, N., S. Stubenrauch, U. Mayer, and R Hempfling. UMS-Modell: Einzelflachenbezogene quantitative Expositionsabschatzung und Gefahrenbeurteilung von Altlasten, in ALtfMtenbewertung-DatenanaLY.Je und GejahrenbewertUflg. S. Schulte-Hostede, R Freitag, A. Kettrup, and W. Fresenius, Eds., ECOMED-Verlag, Landsberg, 1997 (in German).
Heavy Metal Contamination in Industrial Areas and Old Deserted Sites ---------"-"-----"--"-"
""----
-
----"---~---
321
-"-----"-
Stubenrauch, S., R. Hempfling, N. Simmleit, T. Mathews, and P. Doetsch. Abschatzung der Schadstoffexposition in Abhangigkeit von Expositionsszenarien und Nutzergruppen. I. Grundlagen und Vorschlage zur Ableitung von Aufnahmeraten am Beispiel von Trinkwasser, UWSF - Z Umweltchem. Okotox. 6, pp. 41-49, 1994a (in German). Stubenrauch, S., R. Hempfling, S. Simmleit, T. Mathews, and P. Doetsch. Abschatzung der Schadstoffexposition in Abhangigkeit von Expositionsszenarien und Nutzergruppen. II. Vorschlage fur orale Aufnahmeraten von Boden, Badewasser und Nahrungsmitteln des Eigenbaus, UWSF - Z Umweltchem. Okotox. 6, pp. 165-174, 1994b (in German). Tack, F.M. and M.G. Verloo. Impact of single reagent extraction using NH 4 0Ac-EDTA on the solid phase distribution of metals in a contaminated dredged sediment. Sci. Total Environ. 178, pp. 29-36, 1996. Tessier, A., P.G.C. Campbell, and M. Bisson. Sequential extraction procedure for the speciation of particulate trace Metals. Anal. Chem. 51, pp. 844--851, 1979. Tessier, A. and P.G.C. Campbell. Comment on Pitfalls of Sequential Extractions by P, M.V. Nirel and F.M.M. MoreL Water Red., 24, pp. 1055-1056, 1990. Water Red., 25. pp. 115-117, 1991. Twardowska, I. Areas of long-lasting anthropopression: Assessment and monitoring of pollution potential to soil and ground water. SPIE, 2504, pp. 253-264, 1995. \ Twardowska, I. and J. Kyziol. Binding and chemical fractionation of heavy metals in typical ptat matter. FredenilM J. AnaL Chem. 354, pp. 580-586, 1996. i Ure, A.M. Single extraction schemes for soil analysis and related applications, in Special Istue: Harmonization of Leaching/Extraction Tests for Environmental Risk Assessment/ Ph. Quevauviller, Ed., Sci. Total Environ. 178, pp. 3-10, 1996. WH 0 (World Health Organization). Environmental Health Criteria 101, Methylmercu . natwnal Programme on Chemical Safety. World Health Organization, Geneva, 1990. WH O. Environmental Health Criteria 118, Inorganic Mercury. IPCS, I nternatwnal Programme on Chemical Safety. World Health Organization, Geneva, 1991. Xiao-Quan, S. and C. Bin. Evaluation of sequential extraction for speciation of trace metals in model soil containing natural minerals and humic acid. Anal. Chem. 65, pp. 802--807, 1993.
INDEX
A Acid ammonium acetate-EDTA (AAAc-EDTA) extraction 253,
256, 266, 260-262, 267-269 Acid-base changes in the rhizosphere
229-232 Acid forest soils 29-55 Aging 5, 6, 11, 19, 20 Aliphatic acids exuded by roots 230 ALnUd 231 Aluminum 20, 23, 24, 41, 45, 185,
225, 232, 233, 240 effect on roots 239 hydrous oxides 29, 61 lomc 207 Aluminum hydroxide 29, 61 Aluminum oxide 13, 14, 24, 110, 181 Amino acids 230, 234 Ammonium ion 231 Ammonium nitrate 127 Amphiboles 239, 240 Antimony 300, 301, 302, 303, 304, 306 in the blood and urine 314, 317 Aromatic acids exuded by roots 230 Arsenate 10, 12 Arsenic 10, 301, 302 deposition 29 B Barrier capacity of a surface soil layer
292-293
Batch versus flow-through systems
71-74 Bentonite 108 Beryllium distribution coefficients for 4 Bicarbonate 186 Binding strength 293-298 Biotite transformation to kaolinite
240 Biphasic sorption reactions 13 Breakthrough curves (BTCs) 76,
77-78, 83, 92, 100, 103, 104, 105, 133 Brunauer-Emmett-Teller (BET) isotherm model 19
C
6, 9, 13, 19, 23, 59-85, 91, 92, 100, 102, 160, 164, 166, 168, 169, 171, 209, 253, 257, 263, 264, 266, 267, 269, 278, 279, 280, 282, 291, 292, 302, 318 desorption kinetics 82-83
Cadmium
isotropic exchangeability 5 retention 96 sorption 63 reversibility 80-83 uptake by soils 4 Cadmium-calcium exchange isotherm
101 Cadmium carbonate 22, 23 Cadmium sulfate 160, 161
324
Fate and Transport of Heavy Metals in the Vadose Zone
Cadmium sulfide 161 Calcareous agricultural soils 177-197 Calcium 23, 68, 100, 152, 164, 165,
167, 169, 194 IOmc 108, 112, 114, 115, 121, 152, 161, 186, 205, 207 Calcium carbonate 186, 204, 205 deposition 29 fractionation 296 in sludge 252, 255, 268 in surface soils 203 mobile fraction 285 partitIOning 295 solubility 162 sorption 210 USEP A drinking water quality standard 59 Calcium chloride 68, 70, 79, 112,
127, 132 Calcium nitrate
desorption 5, 7 distribution coefficients 4 isotropic exchangeability 17 sorption 16-17 Cobalt hydroxide 20, 31 Complexation capacity 208 Convective-dispersion transport equation (CDE) 99 Copper 13, 14, 92-94, 97, 102-104,
105, 160, 164, 168, 170, 185, 206, 207, 209, 210, 212-219, 233, 278, 279, 280, 292, 296, 301, 302, 318 adsorption by calcium carbonate
204 bioavailability 127-145 fractionation 283 in sludge 259 in surface soils 203
ionic
66, 68, 70, 80, 112,
127, 132 Calcium oxide 181 Carbohydrates exuded by roots 230 Carbon 169, 171 flux 229 transformations 234 Carbon dioxide 167, 189, 231 loss 186 Cardiomyopathy 147 CaJuarina 231 Cesium 2,6 desorbed fraction 5 sorption 14, 15 Chelate effect 15 Chlorine ions 64-66, 79 Chromium 19, 92, 100, 164, 234,
253, 257, 263, 264, 266, 269, 277, 279, 280, 291, 318 deposition 29 in sludge 252, 255, 259, 268 sorption 4 Chromium hydroxides 31 Clover 240 Cobalt 3, 6, 8, 19, 20, 22, 23, 65, 81,
160, 162, 234, 302 deposition 29
114, 205-207 mobile fraction 285 mobility 127-145 oxidized states 234 retention 96, 107-122 solubility 162 sorption on noncrystalline aluminum oxide 4 Copper-hydronium exchange equilibrium 122 Copper-magnesium exchange isotherm
103 Copper-sodium exchange 109 Corn (Zea mayd) 231, 240, 255 Cupric chloride 108 Cupric hydroxide 108, 110, 130 Cupric nitrate 112 Cupric oxide 149, 150, 151, 152, 153,
154, 155, 156 Cupric sulfate 103, 130, 161 D Dendrobaena veneta 141, 142, 144 Diffuse double-layer 30 Diffuse ion association 30-32 Diffuse ion complex 11 Diffusion-controlled kinetic reactions
8-24
Index
Dissolved organic carbon (DOC)
131-137, 145, 148-150, 152, 154, 210 Douglas fir 231 Drinking water 307 E Earthworms 139-142 EDTA 115, 122, 194, 195, 207 See auo Acid ammonium acetateEDT A extraction Electronegativity 12, 218 Electron spin resonance (ESR) spectroscopy 13
F Fatty acids exuded by roots 230 Feldspars 239, 240 Ferric hydroxide 110 Ferric oxide 159, 161, 181 Fick's second law 9-10 Flame atomic absorption spectrometer
132 Flow-through systems 73 compared to batch systems 84 Foliar manganese content of conifers
232 Food crops 304 Food engineering residues 318 Forest soil 264 French beans 232 Freundlich equation 93, 102, 119, 120, 121 See auo Two Species Freundlich equation; van Bemmelen-Freundlich equation Fulvic acids (FA) 110-111, 115-119,
121, 168, 241 G Galena 189 Gardens, contaminated 304, 307-315 GEOCHEM 64, 149-150, 154, 155 Geographical information systems (GIS) 302, 304 Gibbs free energy function 10 Gibbsite 20
325
Goethite 70, 108, 164 Grass silage 318 Grid model 79 Gypsum 190
H High resolution transmission electron micrography (HRTEM) 23 Hordeum vulgare 233 Human biomonitoring data 304,
313-316 Human risk potential assessment
304-316 Humic acids (HA) 5, Ill, 195 Hydride generation atomic absorption spectrometry (HG-AAS) 148,
149, 150, 151, 152, 154 Hydrogen 233 effect on roots 239 IOmc 114, 115, 121, 164, 167, 168,
171, 172, 189, 232 release 244 Hydrous ferric oxide (HFO) Hysteresis 5, 6, 8, 96
6
I Indoor air 304, 310-312 Inner-sphere surface complexation
11-13, 15, 30, 32, 61-62 Ion chromatography 41 Ion exchange retention 100-105 Ionic radius 8, 12, 19, 218 Ionic strength 12, 14, 62-63, 66,
67-69, 79, 84, III Iron 8, 19, 41, 45, 160, 164, 165, 167, 169, 188, 189, 195, 212, 215, 233, 240, 276, 291, 292 content of the shoot and roots of dwarf French beans 232 distribution coefficients 4 hydrous oxides 29 Iomc 161, 185, 207 oxidized states 234 preCIpItation 162 solubility 162 Isomorphic substitution 12
328
Fate and Transport of Heavy Meta!s in the vadose Zone
Siderite 162 Siderophores 230, 234 Silicates 182 Silicon 22, 185 backscattering 19 hydrous oxides 29, 61 Silicon dioxide 181 Silicon hydroxide 29, 61 Sludge See Sewage sludge Sodium 65, 109, 150, 152 lOmc 112, 121 Sodium chloride 66 Sodium nitrate 66, 127 Soil depth 137 formation 178-180 fractionation 236-239 Soil-plant systems 160-161 Solid-solution ratio 71-73 Sorption kinetics of trace elements
1-25 Soybean 231, 240 Stack emissions 274-300 Sterols exuded by roots 230 Strontium 9 Sulfate 64, 150, 152, 182, 188 Sulfides 182, 185 formation 172 insoluble 161-162 Sulfur 228, 233 Surface precipitation 18-24
T Tartaric acid (TA) 115, 121 Tartrate 115-119, 121 Tin 302 Titanium 22 Titanium dioxide 22, 23, 181 Trees 232-235 growth and rhizosphere chemistry
242 Two Species Freundlich (TSF) equation 129-136, 138, 143
v Valence charge 12 van Bemmelen-Freundlich equation
34, 39-40 van Bemmelen-Freundlich isotherms
42-46 Vanselow equation 109, 120-122 Vermiculite 240 Volcanic rock 182 chemical composition 182
w Waste sites 273-319 Wastewater 211-218, 252 Weathering in soil 239-242 Wheat 240
x X-ray absorption fine structure (XAFS) spectroscopy 9-10, 11,
13, 19-20, 24 X-ray absorption spectroscopy (XAS)
12-13, 19-20 X-ray photoelectron spectroscopy (XPS) 9
Z Zinc
9, 13, 38, 65, 79, 82, 100, 102,
160, 161, 164, 166, 188, 204-207, 209, 217-219, 233, 253, 269, 277, 279-280, 302, 318
168-171, 185, 212, 213, 215, 257, 263-267, 291-295, 301,
content of the shoot and roots of dwarf French beans 232 distribution coefficients 4 in sludge 252, 255, 259, 268 mobile fraction 285 retention 96 solubility' 162 surface soils 203 uptake by soils 4 Zinc sulfide 161