CHEMISTRY FOR PROTECTION OF THE ENVIRONMENT
Other volumes in this series
1 Atmospheric Pollution 1978 edited by M.M. Benarie 2 Air Pollution Reference Measurement Methods and Systems edited by T. Schneider, H.W. de Koning and L.J. Brasser 3 Biogeochemical Cycling of Mineral-Forming Elements edited by P.A. Trudinger and D.J. Swaine 4 Potential Industrial Carcinogens and Mutagens by L. Fishbein 5 Industrial Waste Managements by S.E. Jflrgensen 6 Trade and Environment: A Theoretical Enquiry by H. Siebert, J. Eichberger, R. Gronych and R. Pethig
7 Field Worker Exposure during Pesticide Application edited by W.F. Tordoir and E.A.H. van Heemstra-Lequin 8 Atmospheric Pollution 1980 edited by M.M. Benarie 9 Energetics and Technology of Biological Elimination of Wastes edited by G. Milazzo 10 Bioengineering, Thermal Physiology and Comfort edited by K. Cena and J.A. Clark 11 Atmospheric Chemistry. Fundamental Aspects by E. Meszdros 12 Water Supply and Health edited by H. van Lelyveld and B.C.J. Zoeteman 13 Man under Vibration. Suffering and Protection edited by G. Bianchi, K.V. Frolov and A. Oledzki 14 Principles of Environmental Science and Technology by S.E. J@rgensenand I. Johnsen 15 Disposal of Radioactive Wastes by 2. Dlouliy 16 Mankind and Energy edited by A. Blanc-Lapierre 17 Quality of Groundwater edited by W. van Duijvenbooden, P. Glasbergen and H. van Lelyveld 18 Education and Safe Handling in Pesticide Application edited by E.A.H. van HeemstraLequin and W.F. Tordoir 19 PhysicochemicalMethods for Water and Wastewater Treatment edited by L. Pawlowski 20 Atmospheric Pollution 1982 edited by M.M. Benarie 21 Air Pollution by Nitrogen Oxides edited by T. Schneider and L. Grant 22 Environmental Radioanalysis by H.A. Das, A. Faanhof and H.A. van der Sloot
Studies in Environmental Science 23
CHEMISTRY FOR PROTECTION OF THE ENVIRONMENT Proceedings of an International Conference, Toulouse, France, 19-25 September 1983 Organized under sponsorship of The Federation of European Chemical Societies, U n i t e d States Environmental Protection Agency and Ministere FranGais d e I'Environnement b y the l n s t i t u t National des Sciences Appliquees, Universite Paul Sabatier, Toulouse and The Polish Chemical Society
Edited by
1. Pawlowski institute of Chemistry, Maria Curie-Sklodowska University, PI. M. C. Sklodowskiej 3, 20-03 1 Lublin, Poland
A.J. Verdier Ecole Nationale Supgrieure de Chimie, 118 Route de Narbonne, 31077 Toulouse Cedex, Fral
W.J. Lacy US. Environmental Protection Agency, Office of Research and Development, Washington, DC 20460, U.S.A.
ELSEVl E R Amsterdam
1984
- Oxford - New York
- Tokyo
ELSEVIER SCIENCE PUBLISHERS B.V. Molenwerf 1, P.O. Box 21 1, 1000 A E Amsterdam, The Netherlands Distributors for the United States and Canada: ELSEVIER SCIENCE PUBLISHING COMPANY INC. 52, Vanderbilt Avenue New York, N.Y. 10017
Library of Congress Cataloging in Publication Data
Main entry under title: Chemistry for protection of the environment. (Studies in environmental science ; 23) Biblfography: p . Includes indexes. 1. Sewage--Purification-Congresses. 2. Environmental chemistry--Congresses. 3. Pollutants--Enviromental aspects--Congresses. I. Pawlowski, Zucjan. 11. Verdier, A. (Alan) 111. Lacy, W. J. (William J.) IV. Series.
T'D'745.C43 1984 ISBN 0-444-42347-8
628
84-6051
(U.S.)
ISBN 0 4 4 4 4 2 3 4 7 - 8 (V01.23) ISBN 0 4 4 4 4 1 6 9 6 - X (Series)
0 Elsevier Science Publishers B.V., 1984 A l l rights reserved. No part of t h i s publication may be reproduced, stored in a retrieval system or transmitted in any f o r m or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior written permission of the publisher, Elsevier Science Publishers B.V., P.O. Box 330, 1000 A H Amsterdam, The Netherlands Printed in The Netherlands
V
Lucjan PAWLOWSKI (Ph.D. D.Sc.) is Docent - eq. to Associate Professor of Chemistry and Environmental Science at Technical University of Lublin (Politechnika Lubelska). Born in 1946 in Poland. He received his D.Sc. (1980) in EnvironmentalEngineering, Ph.D. (1976) in Environmental Chemistry both from Technical University of Wroctaw and M.Sc. (1969) in Chemistry from Maria Curie-Sklodowska University. He has adited 2 books, was co-author of 4 books, 78 scientific papers and 39 patents. He is coeditor of an international journal Rective Polymers. He is a chairman of the Division of Environmental Chemistry and Engineering of the Polish Chemical Society. His research interest are directed toward application of chemical methods for recycling of wastes.
Alain J. Verdier (Ph.D.) is Prafessor of Chemistry in Institut National des Sciences Appliquees (INSA) - Toulouse. Born in 1939 in France. He received his Ph.D. (1966) and M.Sc. (1960) in Chemistry both from Paul Sabatier University. He has edited 1 book, was author or co-author of 28 papers and 3 patents. He is a director of the Chemistry Department of INSA. His research interest are directed toward application of chemical methods for wastewater treatment.
William J. LACY (Ph.D.) is a director, Water and Waste Management Monitoring Research, ORD, US EPA. Born in 1928. He received BS (1950) in Chemistry and Ph.D. (1959) in Chemical Engineering from University of Connectitut. He is author or co-author of 149 publications, 2 patents and 4 books, editor 3 text books and contributes t o 5 other books. He serves on the Editorial Advisory Boards of Industrial Wastewater Engineering, Environmental International Journal and Pollution Engineering. He is a president of PTA; Board Director of Mansion House Swim and Tennis Club, Officer in Mari'sion House Yacht Club, Vice President Mount Vernon Citizens Association, Foster Parent (to date 29 childrzn have been under his home care).
This Page Intentionally Left Blank
VII
ADVISORY SCIENTIFIC COMMITTEE A. ABADIE, I.N.S.A., Toulouse, France E. ANGELIER, UniversitePaul Sabatiw, Toulouse, France G. ALEARTS, Katholieke UniversiteitLeuven, Belgium J . BARCICKI, Maria Curie-Sklodowska University,Poland D. BARNES, The University of New South Wales, Sidney, Australia R. BENAIM,I.F.T.S. - Agen, France B. A. BOLTO, CSIRO - Melbourne, Australia W. J. COOPER, Horida International University, Miami, USA U. COSKUNER, Transturk Holdings A S . , Istanbul, Turkey P . DOLEJS, CzechoslovakAcademy of Science, Czechoslovakia J, FONTAN, UniversitePaul Sabatier, Toulouse,France F . EL-GOHARY, National Research Centre, Cairo, Egypt P. GRAMMONT, Duolite International, Chauny, France R. GRINNEL, Boston State University,Boston, USA M. GROMIEC, IMG W - Wmsaw, Poland J. M. HEFTY, Rohrn and Hass, USA L. JACKSON, Dpt of Energy, USA A. L. KOWAL, Technical University of Wroclaw, Poland R. G. LANZA, The University of Texas ar Dallas, USA L. LIBERTI, Instituto di Ricerca Sulle Acoue, Italy K. MELLANBY, Monks Wood Experimental Station, U.K. G . MOUVIER , UniversiteParis WI,France N . L. NEMEROW, University of Miami, USA A. PORANEK, Rivers Stare University,Port Harcourt, Nigeria N. RAMANATHAN, Government of India, Dept. of Environment, New Delhi, India A. RODRIGUEZ, University of Porto, Porto, Portugal L. D. ROLAND, Foster Wheeler Limited, U.K. G. SHELEF, Israel Institute of Technology, Haifa, Israel K . SNIDVONGS, Office of the National Environment Board, Bangkok, Thailand V. SOLDATOV, Academy of Science, Minsk, USSR G. TIRAVANTI, Instituto di Ricerca Sulle Acoue, Italy J. K. WALTERS, University of Nottingham, U.K. T. WINNICKI, Technical University of Wroclaw, Poland M. H . WONG, The Chinese University of Hong Kong, Hong Kong G. YOUSIF SIR EL KAHTIM, University of Khartoum, Khartoum, Sudan
Toulouse, France
Toulouse, France
Toulouse, France
This Page Intentionally Left Blank
IX
CONTENTS
FOREWORD
................................................
1
Chapter I POLLUTANTS IN THE ENVIRONMENT: GENERAL ASPECTS B. A. Bolto, J. Barcicki, Z. Kozak, L. Pawlowski TECHNOLOGY AND THE ENVIRONMENT: ALLIES OR ANTAGONISTS?
D. Blasco THE STATE OF THE ENVIRONMENT - REPORT OF UNEP
......
5
...................
33
G. Alaerts A DIFFERENT APPRECIATION: WATER QUALITY AND TREATMENT IN DEVELOPING COUNTRIES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
41
Chae-Shik, RHO CHEMICAL POLLUTION IN KOREA AN EXPERIENCE O F THE DEVELOPING COUNTRIES
65
C h a p t e r I1 POLLUTANTS IN THE ENVIRONMENT: IMPACT AND CONTROL H. Zimny ECOLOGICAL EFFECTS O F INDUSTRIAL POLLUTANTS AND THEIR EFFECT ON CULTIVATEDPLANTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
19
R. F. Holmes, W. J. Lacy MONITORING AND QUALITY ASSURANCE FOR HAZARDOUS WASTE SITE ASSESSMENT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
93
J. M. Bowron, M. L. Richardson CATCHMENTQUALITYCONTROL . . . . . . . . . . . . . . . . . . . . . . . . . .
:.......
F. Y. Saleh, K. L. Dickson, J. H. Rodgers, jr. TRANSPORT PROCESSES O F NAPHTHALENE IN THE AQUATIC ENVIROQMENT .
...
109 119
M. Lamotte, P. Masclet INTERLABORATORY COMPARISON O F QUANTIFICATION O F PAH IN ATMOSPHERIC AEROSOLS BY GAS AND LIQUID CHROMATOGRAPHIES AND BY SHPOL’SKII FLUORIMETRY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
133
J. C. Synnott, S. J. West, J. W. Ross COMPARISON OF ION-SELECTIVE ELECTRODE AND GAS-SENSING ELECTRODE TECHNIQUES FOR MEASUREMENT O F NITRATE IN ENVIRONMENTAL SAMPLES
143
...
X A. Baleix, B. Caussade, J. George, J. Mathieu, A. Reynes, L. Torres STUDY O F MOLECULAR DIFFUSIVITY INFLUENCE ON MASS TRANSFER RATE AT A WATER - ATMOSPHERE INTERFACE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
155
U. Zoller ABU NDANCE O F NONIONIC SURFACTANTS IN ISRAEL MUNICIPAL SEWAGE
16 1
......
C h a p t e r 111 PHYSICOCHEMICAL TREATMENT O F SUSPENSIONS P. Dolejs INTERACTION O F TEMPERATURE, ALKALINITY AND ALUM DOSE BY COAGULATION O F HUMIC WATER . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
169
D. R. Dixon, L. 0. Kolarik MAGNETIC MICROPARTICLES FOR TREATMENT O F NATURAL WATERS AND WASTEWATERS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
179
Sung-bin Wang, K. Y. Chen BITTERNS AS COAGULANTS FOR TREATMENT O F COLOR EFFLUENTS
...
.........
A. M. Dziubek, A. L. Kowal WATER TREATMENT BY COAGULATION-ADSORPTIONWITH DOLOMITE.
........
D. A. Wilms, A. A. Van Haute PRIMARY FLOCCULATION O F WASTEWATER WITH Al,(SO,), AND NaAlO, SALTS RECUPERATED FROM SPENT ALUMINIUM ANODISING BATHS . . . . . . . . . . . . .
..
193
205
21 3
R. J. Francois, A. A. Van Haute FLOC STRENGTH MEASUREMENTS GIVING EXPERIMENTAL SUPPORT FOR A FOUR LEVELHYDROXYDEFLOCSTRUCTURE . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
221
G. Tiravanti, F. Lore, N. Palmisano MEASUREMENT O F THE CHARGE DENSITY O F POLYELECTROLYTES BY A DIFFERENTIAL CONDUCTOMETRIC METHOD . . . . . . . . . . . . . . . . . . . . . . . . . .
235
D. Miskovic, E. Karlovic, B. Dalmacija THE INVESTIGATION O F APPLICATION O F DISSOLVED AIR PRECIPIT+TE FLOTATION IN THE ABSENCE O F COLLECTOR AND FROTHER FOR THE PURIWATION O F WASTEWATER CONTAINING METAL IONS . . . . . . . . . . . . . . . . . . . . . . . . . . .
245
J. Hupka, A. G. Oblad, J. D. Miller HOT WATER PROCESSING O F U.S. TAR SANDS-WATER RECYCLE AND TAILINGS DISPOSAL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
253
J. Hupka THE ROLE O F DILUENT IN OILY WATER TREATMENT IN BED COALESCERS
269
......
B. Gutkowski, St. Mydlarczyk, M. Kowalska, J. Hupka SATURATION PROFILES IN COALESCENCE BED . . . . . . . . . . . . . . . . . . . . . . . . .
285
XI C h a p t e r IV PHYSICO-CHEMICAL TREATMENT: ADSORPTION A. Derylo, M. Jaroniec THEORETICAL FOUNDATIONS O F SOLUTE ADSORPTION FROM DILUTE SOLUTIONS ONSOLIDS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
297
P. Lafrance, M. Mazet, D. Villessot SPECIFIC ADSORF'TION O F ORGANIC MICROPOLLUTANTS ONTO ACTIVATED CARBON: A STUDY O F ELECTROKINETIC PHENOMENA DUE TO MULTICOMPONENT SYSTEMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
313
R. J. Martin, W. J. NG FURTHER STUDIES ON THE USE O F CHEMICALS TO REGENERATE EXCHAUSED ACTIVATEDCARBON . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
329
A. L. Kowal ON UNIT PROCESSES DURING INFILTRATION
343
..........................
Chapter V PHYSICO-CHEMICAL TREATMENT: ION EXCHANGE V. S. Soldatov NEW FIBROUS ION EXCHANGERS FOR PURIFICATION O F LIQUIDS AND GASES
...
35 3
H. Tanaka, M. Nakayama, M. Chikuma, T. Tanaka, K. Itoh, H. Sakurai SELECTIVE COLLECTION OF SELENIUM (IV) FROM ENVIRONMENTAL WATER BY FUNCTIONALIZED ION-EXCHANGE RESIN . . . . . . . . . . . . . . . . . . . . . . . . . . . .
365
J. Hlavay, K. Foldi-Polyfik, J . InczBdy REMOVAL O F ARSENIC FROM NATURAL WATERS
373
.......................
C. Sarzanini, E. Morengo, M. C. Gennard, C. Baioochi, E. Mentasti PRECONCENTRATION AND SEPARATION O F Cr(II1) AND Cr(V1) FROM AQUEOUS SOLUTIONS BY COMPLEX FORMATION-ION EXCHANGE . . . . . . . . . . . . . . . .
...
K. Majewska-Nowak, T. Winnicki THE USABILITY O F POLYSULFONE MEMBRANES FOR REMOVAL O F ORGANIC DYES FROMAQUEOUSSOLUTIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
38 1
387
C h a p t e r VI PHYSICO-CHEMICI L TREA TMENT: OXLDATION A. Poranek, A. MikaGibafa INTENSIFICATION O F THE OXYGENATION PROCESS WITH THE USE O F POLYMERIC DIFFUSERS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
399
T. D. Waite, K. A. Gray OXIDATION AND COAGULATION O F WASTEWATER EFFLUENT UTILIZING FERRATE(V1)ION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
407
XI E. M. Grochulska-Segal, M. M. Sozahski THERMODYNAMIC CRITERIA FOR THE OPTIMIZATION OF IRON AND MANGANESE REMOVALS IN EXPERIMENTAL PROCEDURES
421
N. M. Surida, J. Sugijanto IDENTIFICATION OF TIN FROM TIN-SMELTING REFRACTORY-WASTE AFTER ALKALINESOLVATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
421
Z. Gorzka, M. Kairnierczak, E. Filipiak THE TREATMENT DETERGENTS IN INDUSTRIAL WASTEWATERS ON A PILOT-PLANT SCALE BY CATALYTIC OXIDATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
431
I. Pollo, J. Jaroszyhka-Wolkska ENHANCEMENT OF NITROGEN OXIDES ABSORPTION FROM WASTE GASES USING OZONE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
445
..........................
A. Socha, Z. Gorzka ELECTROCHEMICAL OXIDATION O F ROKAPHENOL N-6 A FLOW ELECTROLYZER
..
45 1
M. Mansour, H. Parlar, F. Korte REMOVAL OF POLLUTANTS FROM THE AQUATIC ENVIRONMENT BY PHOTOOXIDATION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
45I
P. Le Cloirec, G. Martin MODELING OF IMMERGED FILTERS IN THE CASE OF AERATED FILTERS
46 3
........
Chapter VII RECYCLLNG OF WASTE MATERLALS AND POLLUTION FREE TECHNOLOGIES J. Z. Nie6ko WASTE-FREE METHOD OF CADMIUM CARBONATE PRODUCTION L. Pawlowski, M. Kotowski, B. A. Bolto, R. McNeill RECLAMATION OF CHROMIUM FROM WASTES
.............
........................
485 491
L. Liberti, A. Lopez, R. Passino APPLICATIONS OF SELECTIVE ION EXCHANGE TO RECOVER MgNH,PO, FROM SEWAGE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
513
M. R. Stevens, M. Parnazari, F. Saavedra REMOVAL OF SILVER FROM PHOTOGRAPHIC SOLUTIONS .
5 25
.................
Chapter VIII PHYSICO-CHEMICALASPECTS OF BIOLOGICAL TREATMENT
W. M. Drew, G. H. A. Holder MICROBIAL SULPHATE REDUCTION
................................
J. Y. Plat, D. Sayag, L. Andre EFFECT OF SOME PHYSICAL PARAMETERS ON COMPOSITING RATE AND YIELD
531
...
553
XI11 D. Barnes, P. J. Bliss, R. B. Grauer, C. H. Kuo, K. Robins TREATMENT O F HIGH STRENGTH WASTEWATERS BY AN ANAEROBIC FLUIDISED BEDPROCESS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
559
P. Le Calve, N. Therien A RESPIROMETRIC STUDY O F THE INFLUENCE O F ALIPHATIC ALCOHOLS ON ACTIVATEDSLUDGES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
569
C h a p t e r IX AIR POLLUTION A. Trier AIRBORNE POLLUTION PROBLEMS IN SANTIAGO CHILE E. M. Bulewicz, C. Juryb, S. Kandefer FLUE GAS DESULPHURISATION USING LIME WASTE
..................
.....................
D. Barreteau, C. Laguerie DESULPHURIZATION O F GAS BY SORPTION O F SO, ON CUPRIC OXIDE DEPOSITED ON ALUMINA PARTICLES IN A FLUIDIZED BED REACTOR . . . . . . . . . . . . . . . . .
585
589
.
B. J. Mayland, L. D. Roland NITROGEN OXIDES EMMISION CONTROL CDL/VITOK ENHANCED ABSORPTION PROCESS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
597
605
Chapter X PANEL DISCUSSION POSSIBLE DIRECTION O F RESEARCH AND DEVELOPMENT RELATING TO CHEMISTRY FOR THE PROTECTION O F THE ENVIRONMENT . . . . . . . .
........
6 15
Author Index
................................................
621
Subject Index
................................................
623
This Page Intentionally Left Blank
1
FOREWORD
The first conference of this series ws organized in I976 at the Maria Curie-Sktodowska University in Lublin, Poland, and was on a national level. During this conference discussions were held on the findings and results of the many Polish-US. funded environmental research projects. However, the great interest aroused in the conclusions and recommendations discussed at that conference stimulated the participants to oRanize the next conference in I9 79, which evolved into one involving international participation. The proceedings of this second conference which contains 32 peer selected papers were published by Pergamon Press in I980. The third multi-national conference in this series w a s organized and held at the MCK University in Lublin, Poland, in 1981. The attendees were a significant group of leading scientists from all of Europe and the USA. The proceedings of this conference, which contains 36 selected peer-reviewed papers, were published by Elsevier in I982 in their well known-series “Studiesin Environmental Science’: During the organizational phase o f the fourth international conference, the Executive Committee decided to change its title for a broader one i.e. ‘CHEMISTRY FOR PROTECTION OF THE ENVIRONMENT’: The basic reason underlying that decision was the comments received from various scientists who indicated that a need existed to provide an international forum for all chemists and chemical engineers involved in environmental protection activities. The title used for the first three conferences, i.e. ‘FHYSICOCHEMICAL METHODS FOR WATER AND WASTEWATER TREAWENT”, was somewhat limiting because it tended to exclude those who utilize chemical processes outside of the conventional water environment. It is realized that any pollution control action t o h y must take into consideration not only the water problems but also proper environmental handling o f the sludge and related air pollution problems, In all these cases chemical processes normally play a highly significant role. It was therefore concluded by the Executive Committee that there is indeed a need to provide a forum for chemical scientists and engineers who are dedicated to the worthy mission of making a cleaner, healthier world for everyone. This fourth conference thus focused on the application of chemical methods for environmental protection and its main purpose was to provide a setting for the exchange of scientific and technical knowledge between those who know chemistry and those who know about environmental problems. In other words, the fourth Conferencewas to help link knowledge of chemistry with the environmental problems to be solved by chemical methods. The Executive Committee hoped that this conference would attract the interest of those talented scientists and engineers currently not involved in such problems, but who might have useful ideas to contribute on how to clean our environment. This fourth conference attracted participants from 34 countries representing all continents and included scientists and engineers from the many developing nations of Africa, Asia and South Afnca. Nevertheless, the biggest contribution to the fourth international
2 conference came mainly from France, Poland and the USA. The participants obtained a good overview on how the chemical processes are used in different countries of the world, and made this scientific conference of special interest and an extremely valuable source of informationand technology transfer, The United States Environmental Protection Agency, through its Office of Research and Development, provided this conference with special funds allowing us to bring wide and diversified scientific representation .from different nations. Therefore, on beharf of all the sponsored particimnts, we would like to thank the US.EnvironmentalProtection Agency for its support. The Executive Committees agree that it is a good idea to organize and continue this endeavor with a similar conference biennial&. The next one, the fifth international conference, will thus be held 9-13 September 1985, at the Gztholic Universi@ of Leuven, Belgium We hope thst this series of scientific conferences will continue to grow, receive more scientific contributions and attract an even wider representation of chemists and chemical engineers from all over the world. LUCJAN PAWLOWSKI ALAIN VERDIER WILLIAM J. LACY Editors
CHAPTER I
POLLUTANTS IN THE ENVIRONMENT: GENERAL ASPECTS
This Page Intentionally Left Blank
TECHNOLOGY AND THE ENVIRONMENT:ALLIES OR ANTAGONISTS? B. A. BOLT0
CSIRO Division of Chemical and Wood Technology, Private Bag 10, Clayton, Victoria 3168, Australia J. BARCICKI, Z. KOZAK
Department of Chemical Technology, Institite of Chemistiy, Maria Cbie-Sklodowska University, 20-031 Lublin, Poland L. PAWLOWSKI Visiting Scientist at CSIRO from Maria &rie-Sklodowska University, 20-031 Lublin. Poland
ABSTRACT The phenomenally rapid development of technology, especially in the 20th century, has enormously increased Man’s ability to produce goods which have enhanced his standard of living. However, this development has also generated a secondary phenomenon, the pollution of Man’s environment. This has had the contrary effect of leading to a deterioration in the quality of life. For much of history, an enhancement in the quality of life arising from new technology has overshadowed its negative effects upon the environment. Recently there has been some doubt as to whether the further development of technology will necessarily guarantee an improvement in the quality of life. The authors discuss the relationship between technology and the main components of the environment, water and air, and its impact on Man’s living conditions. Technology, appropriately used, can be a powerful force for the improvement of the environment. Relevant actions for upgrading the quality of water and air are put forward. The authors also discuss problems arising from the exploitation of energy resources, with particular reference to the surrounding water and air.
1. INTRODUCTION
Actions being taken to protect the environment are similar to those pursued in the development of new technology (which itself contributes to a deterioration in the environment). They ensure a better quality of life for Man by providing him with healthier conditions, both physically and psychologically. Therefore, discussions about which is the more important - new technology or a healthy environment - are senseless as both are essential for the well-being of Man.
6 The rapid development of technology, especially in the 20th century, has increased enormously Man’s ability to produce goods to enhance his standard of living. This ability to produce goods more efficiently has made by it necessary to emphasize marketing. As Man’s basic needs are limited, there has been a great effort to create psychologically a demand for unnecessary goods which are supposed to make for a decent life. This artificial demand, in turn, increases productivity. However, an increase in productivity accelerates not only the exhaustion of raw materials, but also the deterioration of the environment through the discharge of wastes. Some doubt must arise as to whether further increases in consumption really lead to an improvement in the quality of life, especially when there is a natural limitation of the resources necessary to establish consumption at the level reached by developed countries. Therefore, we believe that the protection of the environment, or improvements to the quality of life, require the development of an adequate model of life. This should be done by influencing human needs and priorities rather than by strict regulations. Chemistry plays a particular role amongst the scientific and technological disciplines. It development has had a great impact on the environment, as understood in the broadest terms; surely, the providing of medicines and disinfectants has contributed enormously to an improvement in the control and even the virtual elimination of many diseases, and the development of synthetic fertilizers has increased the efficiency of food production. However, the development of chemistry has also created new non-biodegradable chemicals, previously unknown in nature. Some of these chemicals have seriously interfered with many forms of life on our planet, causing new diseases. Chemistry itself has thus become an important part of the environment, and its influence should be carefully investigated. Of course, chemistry serves also t o preserve the environment from pollution. It provides us with the means of eliminating some of the negative physical influences of our civilization. It is to show this very beneficial role of chemistry in the protection of the environment that we meet here. A lot of detailed chemical processes dealing not only with environmental protection, but also with ways of slowing down the exhaustion of raw materials by recovering them from wastes, wdl be presented during this meeting. Our intention in this paper is to put forward some quite general, strategic aims and how chemical methods may be utilized in achieving these aims.
2. RESOURCES FOR HUMAN NEEDS
Natural, non-renewable resources are continuously being dispersed throughout the world by Man. The total quantity of resources available does note decrease because of this activity, but those resources are transformed from a concentrated into a diluted form so that they become virtually irretrievable. Theoretically it is possible to imagine a process which would allow one to extract the material for further use once diluted. However, apart from economic difficulties there are some natural limitations, such as the amount of energy which would be required. One might think that scientific and technical knowledge would allow Mankind to find substitutes for exhausted materials and this may well be so in many cases. However, it would be irresponsible to totally base our continued existence on such a belief. It is more
7
likely that there will be a natural limit to expansion. Therefore, we need to classify problems and to define some of the more important aims for the development of mankind, not by alarming and frightening people, but by indicating alternative methods of development, each with their positive and negative aspects. Technology’s strongest point is that it may help to solve some of the problems that Mankind is faced with; yet it is also its weakest point in that it may just as easily lead to disaster - it depends on how Man uses technology. Let us look back to the ancient Greek conception of the four ‘elements’: Water, Air, Earth and Fire. This approach emphasizes our present problems. It is likely that the Greek philosophers, living in rather arid areas at a time when there was very limited technology available, were strongly influenced by the major components of their environment. The singling out of Water and Air from the other chemical and mineral resources, which can be considered as combined in the ‘element’ Earth, seems to highlght that these two are of special importance. For many years civilization, with its gradually improving technology, did not realise the importance of Water and Air. These two ‘elements’ had been too freely available. Nowadays they again attract attention, as it has been realized that they are basic items which influence all aspects of our life. Highly developed nations have rediscovered what had been previously found by the Greeks - the vital importance of these two ‘elements’. Since they both deteriorate or become polluted by the indiscriminate redistribution of our general resources, their clean-up should be combined with the recovery of these resources. Decreasing the consumption of resources is equivalent to the preservation of the Greeks’ third ‘element’, Earth. Therefore, it seems that one of the most important requirements of Man’s future activities should be recycling. The implementation of recycling would achieve two goals: (a) the preservation of a pure environment, and (b) a decrease in the use of resources. It is inevitable that our civilization will interfere with both the water and air environment by discharging pollutants to them. Therefore, there is a need for suitable purification techniques. Figure 1 depicts the general recycling concept. From an environmental protection standpoint, the critical part is the separation process which enables either the resource, water or air, to be recycled. However, such processes are energy consuming, and the energy requirement is often the limiting feature of recycling techniques; it is exemplified by the Greeks’ fourth ‘element’, Fire. From the above it can be seen that the influence of technology on the quality of human life, through its different interactions on the environment, may be related to the ancient Greeks’ ‘elements’: Water, Air, Earth (mineral sources) and Fire (energy).
3. THE WATER ENVIRONMENT
Water is one of the most important components of our environment, without which our present life forms could not exist. Its importance has been recognized from the very beginning of human existence, and is strongly depicted in seminal philosophical theories. It was assuredly not chance that led Thales, the Ionian philosopher, to teach that water
8
or a i r
Resource c
b
1) .-
L
P a r t ic ipatiop in civilization s activity
0 L
0 L
W c
W
2
3 0
0
-
u)
L
0)
Polluted water or air + resource
U
L
W
w
0
0
2
W
>
0
0 0
0 W
c
W
I
I I
Suitable separation technology
1
E
I
Fig. 1. Flowsheet for the complete recycling of water or air and a recovered resource.
or moisture is the mineral from which all things evolved. Aristotle, sometime later, included water among the ‘elements’. What further evidence is needed that civilised Man has long recognized the vital role played by water in the affairs of this planet? Perhaps nowhere is this quite SO obvious as in those parts of the world where water is in short supply. Australians, probably above all others, will understand most clearly Landor’s message: we are what suns and winds and waters make us. Although the importance of water was realized very early in Mankind’s history, the impact of water quality on Man’s well being was recognized only recently. However, the knowledge accumulating from everyday experience made Man appreciate the need for water treatment, even in prehistoric times. Quality, as an important attribute of water supply, began to be identified to some extent in ancient Egypt, India, Palestine, Persia, and China, about 2000 BC. The most quoted evidence of this awareness comes from the Sanskrit medical philosophers, whose writings can be paraphrased: Impure water should be boiled, heated by the sun, and filtered through sand and gravel and charcoal. The Chinese resorted to boiling and adding dried leaves of shrubs to improve the taste. Those shrubs are known today as tea plants.
9 Other ancient treatments included sedimentation and storage in metal vessels. Both Aristotle and Hippocrates saw the need to improve water quality for drinking purposes as did others, but no one of that period between 400 and 300 BC knew the real connection between polluted water and health. The problems of water sources, their procurement and distribution remained of paramount interest until the 17th century, although there was some sporadic use of single, double, and even triple filtration about 50 AD, and the Bible refers to infiltration galleries and the use of salts to purify water. More recently, in 1627, Sir Francis Bacon, describing the scientific developments of the previous ten centuries, mentioned several treatment processes for improving water quality: percolation, boiling, distillation, aeration and clarification. Almost 50 years later, Leeuwenhoek invented the microscope and discovered ‘little animals’ in water, but he had no idea of what the discovery meant, and these ‘little animals’ were unknown to Porzio, when he invented his filter in 1685. It seems that the biggest contribution to enhancing the health of the human population was made, not medical doctors, but by sanitary engineers who significantly improved sanitary conditions by the development of healthy water supplies on the one hand, and sewage discharge systems on the other. In this regard the biggest gain was made in the 19th and the first half of the 20th centuries. Even today over a billion people in remote rural areas and in the urban slums of the Third World lack safe drinking water and elementary sanitation facilities. Figures published by the United Nations and the World Health Organization indicate that three quarters of the world’s inhabitants drink something less than potable-quality water. These same agencies report that each approximately 15 million children die from waterborne diseases before reaching the age of five. The problem is not a technical one, as we already know how to purify water to an acceptable level. It is more a problem of economics and politics. The shortage of a natural resource of healthy water, or one which would need only minimal treatment, appears common in poor, over-populated countries which cannot afford to install and operate expensive water treatment facilities. Therefore, there is still a real need for new, more economic methods of water purification. Recently, a new problem has arisen in well-developed countries. Because of the concentration of population in big cities and intensive industrial activity, natural water reservoirs have become more and more contaminated by man-made chemicals. The presence of these chemicals in waterways leads to the disturbance of natural self-purification processes; water from polluted reservoirs can become unsuitable for drinking purposes. The most serious problen is caused by non-biodegradable chemicals which tend to accumulate in the bodies of living creatures. The most dangerous of these chemicals are heavy metals and chlorinated organic compounds. The need for their removal could result in a continuous increase in the cost of providing safe drinking water. Elimination of the negative impact on human health of these new pollutants is clne of the important tasks of today. The implementation of water reuse by communities or industries is expanding; the problem should be recognized so that we caii work out a suitable strategy. One solution is to decrease the influence of industrial activity on the water environment by implementing closed-loop water systems. Such an approach can reduce and even eliminate the quantity of man-made pollutants in natural water systems.
10 3.1. Water in Industrial Systems
The direct consumption of water by industry, in the sense of its conversion to another material, is not hgh. There are only a few processes in which water is an integral component of the product, as for example, in the production of lime from calcium oxide: CaO + H20 S= Ca(OW2 In most systems water is required as the reaction medium, for transportation of materials or for energy production (mainly for cooling), rinsing, etc. In these operations the water is passing through ’unit operations’ and is not actually consumed. It is discharged together with some Chemicals taken up from the ‘unit operation’. The use of water converts it into wastewater, the introduction of which into natural waterways reduces their quality. To eliminate, or at least to minimize, the negative impact of industrial activity on the water environment, all effluents should be purified and preferably recycled.
Production unit
n L 0)
c
0
Wastewater
3
n Q
1 Separation unit
0 ul
3 0)
a
Fig. 2. Flowsheet for the recycling of all the component parts of a wastewater.
Figure 2 shows the most satisfactory way of dealing with wastewater. The use of suitable separation methods allows one to purify wastewater to a stage where it can be recycled as reusable water. The chemicals recovered should be separated in a form which allows them to be recycled to the production system. A crucial point in recycling is the proper,use of separation techniques which will allow the separation of waste streams into usable products. Among the different kinds of industrial wastewaters it is possible to distinguish three groups: - spent processing liquors - rinsing waters - condensates Spent processing liquors, being hghly concentrated wastes arising for example from plating baths, are strong solutions which become contaminated during manufacturing.
11 When the contamination reaches a specified level, the bath has to be replaced by a fresh one. On discharge to the sewer the untreated exhausted baths cause a harmful peak increase in pollution, As the treatment of wastewater is based on the separation of pollutants from water, so plating baths should be considered as solutions to be purified, by separation of the contaminants from the useful bath components. The treated bath, after adjustment of its composition, is then reused in the process. Renovation permits the recycling of the concentrated solution, and avoids the difficulties of disposal of large amounts of pollutants. Dilute industrial wastewater - another typical source of pollution - is formed for example during different rinsing operations by the uptake of chemicals during the rinsing process (whether the chemicals are the raw materials, products or by-products); or during the use of water for cooling or energy production which results in the accumulation of anti-fouling and anti-corrosion agents in the water. A typical example of such a wastewater is the rinsing water coming from metal finishing plants. The concentration of solutions by evaporation produces condensates which may contain several to a few thousand milligrams of solutes per litre. One of the best known examples is the recycling of condensate constituents (water, ammonia and nitrates) from the fertilizer industry, Two kinds of wastewater are encountered: - the condensate from the processing of ammonium nitrate or sulfate which has a pH of 8-9 and contains 200-5000 mg NH3/L and 100-4000 mg NOJL, - the condensate from the processing of ammonia which has a pH of 7-9 and contains 100-3000 mg NH3/L, mostly as the carbonate. Often, up to several hundred milligrams per litre of SO,--, C1-and Si02 are also present. As our aim is not to review all possible cases of recycling, we will merely show how one particular separation technique, namely ion exchange, can be used to ensure resources conservation.
3.1.1. Renovation of Spent Processing Liquors
An example is the renovation of baths used in the metal finishing industry. The contaminating ions originate from the dissolution of metals in acidic media. Typical examples are plating and anodizing baths which may contain up to 400 g Cr03/L. The baths are contaminated with trivalent chromium, formed as a result of the reduction of hexavalent chromium, and other metal ions such as magnesium, aluminium, zinc, copper, nickel and iron which arise from the dissolution of the metal parts being treated. A strongly acidic cation exchanger is able to take up cations from this bath. The flowsheet for a typical bath renovation process is shown in Figure 3. The cations of metals responsible for the deterioration of the bath are taken up by the resin. Once freed of these cations the bath is ready for reuse. After exhaustion the cation exchanger is regenerated with mineral acid. One can treat all analogous exhausted acidic metal finishing baths in a similar manner to chromium baths. For example, a mixture of acetic acid and sodium nitrate is used for the pickling of magnesium sheets. An increase in the magnesium concentration and a decrease in the acetic acid concentration give progressively slower pickling rates and a variation in the nature of the surface produced. An ion-exchange method can be used
12
10%yse -for-2nd-- --part of rogon. 1
I I
Exhausted chromic bath
Ir Treated both
1st fraction o f regen. effluent ( t o sewer 1
Fjg. 3. Flowsheet for the renovation of chromic baths.
for the renovation of the solution (see Figure 4). In this process the magnesium ions are taken up by a strongly acidic cation exchanger, so that the magnesium acetate is converted to acetic acid, ready for reuse. Another example of the purification of a processing solution for reuse is the renovation of cooling water from locomotives. This water contains corrosion inhibitors such as NazCr0,. Normally, locomotives take on water ‘en route’, and this water, which is usually hard, is drained from the engine after return to the workshops; a harmful wastewater is obtained from which chromium and other toxic chemicals must be removed before it is discharged to the sewer. However, instead of regarding the contaminated cooling water as a wastewater, we should consider it as an exhausted processing solution - although much more dilute than that encountered in the metal finishing industries - that can be purified. For this purpose a strongly acidic gel-type cation exchanger in the Na form-can be used. During the passage of such a solution through the ion-exchange bed, all multivalent cations are exchanged for sodium ions. After adjustment of the composition of the treated solution, the latter can be reused as cooling water. Regeneration of the cation-exchange bed is performed with a 10%solution of NaCl, similar to conventional water softening, This method has been in operation at one station in Poland since 1978. The analysis of the resin used in the process (Amberlite IR 120) has not revealed any deterioration greater than that which occurs in conventional water softening. A common feature of the methods presented in this section is the selective removal of
13 Partly exhourted bath
conc.:
cn,coong 20% tcn,coo)z Mg > o NONO,
N
8%
Exhaurtrd bath
conc.:
cn,coon( 10% (CH,CW&Mg 18% No NO, r y 8%
>
Treated both
conc.: CH,COOH
-
28%
NO,6% H NO, > o n NO
Fig. 4. Flowsheet for the renovation of magnesium sheet pickle solution.
ionic species, the presence of which makes further exploitation of the processing solutions impossible. The processes deal mostly with highly concentrated spent liquors. However, as has been shown for the renovation of cooling water from locomotives, this approach can be extended to dilite solutions, where the removal of some impurities is essential fo water reuse.
3.1.2. Recycling of Rinse Water Constituents
To recycle water one has to purify wastewater to a level which permits its reuse industry. From an economic point of view, the recycling of water becomes more attractive when one deals with large amounts, because the cost of the special treatments needed to meet the reuse standard is counterbalanced by the value of the recycled water and the decreased cost for the discharge of wastewater. With small amounts of recovered
14
water is it not easy to obtain this balance. In many cases the design of wastewater purification plants that the use ion exchange can be based on demineralization plants. Frequently, recovery of chemicals is also feasible. The economics of chemical recovery are affected by the value of the chemical itself, the possibility of its reuse within the plant and its concentration. As a general rule a plant can recycle only certain salts. Therefore, the selection of the regenerating agent is determined by the form of salt which can be reused in the plant. For example, the recycling of ammonia in the nitrogen industry requires that sulfuric or nitric acid be used for the regeneration of the cation exchanger, because only ammonium nitrate o r sulfate can be reused in production. Process economics greatly depend on the concentration of the solution recovered in the ion-exchange operation (the regeneration effluent). Usually there is a need for an evaporation step to increase the concentration of the solution. This significantly complicates the process and increases both investment and running costs. In the rinse operation, water flows through a rinsing tank, which washes out chemicals from the manufactured products, and leaves the chemicals in the wastewater. The scheme for recycling wastewater constituents is illustrated in Figure 5. A significant decrease in the amount of wastewater discharged can be obtained by in-plant modification (e.g., counter-current rinsing). However, at least some wastewater will leave the system which has to be treated. The treatments for various rinse waters are discussed below. A typical example of the recirculation of a rinse water is in chrome plating, where in once-through rinse operations, tap water is used. However, this significantly complicates the recovery of chemicals because the water contains ions which will contaminate the recovered solu-
r------
Manufacturing
I I I
I T Rinse water I
-Regen. - - -agent - - -1
+
I
I
+
Ion exchange separation unit
t
- - - - - -concentrated chemicals Recovered
~
1 I Recovered
I
I
Fig. 5. A general flowsheet for the recycling of the components of a rinse water.
15 tions. Therefore, for best operation of a closed-loop rinse system one strictly has to use demineralized water for rinsing. The treatment steps consist of: - the separation of suspended matter - the removal of organic matter - the decationization of the rinse water, followed by - the deanionization of the rinse water. For the separation of suspended and organic matter, one of the established conventional methods can be used without any difficulty. The decationization step is similar to draa- out
Filter (tor suspended and organic matter)
Caustic soda
-torrcgen - - - - -I
Recovered water (demineralized 1
r---
Strongly acidic Cation
Regen I-----
effluent (
waste 1
Fig. 6. Flowsheet for the recycling of the components in the reuse water from chrome plating operations.
16 all operations performed in the decationization of water. A strongly acidic cation exChanger has to be used because of the low pH (< 2) of the rinse water, The flowsheet for a typical process is shown in Figure 6. The regeneration of the exhausted cation-exchange bed is performed in a conventional way, preferably with the cheaper sulfuric acid, but hydrochloric acid can also be used. . The anion-exchange unit is the more important for the recovery of chemicals. There are two possibilities for performing deanionization: with or without the recovery of chromic acid. If one has small amounts of rinse water, it may be simpler not to recycle. In this case regeneration of the anion exchanger is performed with a four per cent solution of NaOH. Chromates present in the regeneration effluent are then reduced to trivalent chromic ions, and after neutralization with lime, the clarified wastewater is discharged into the sewer.Forthe recovery of chromic acid, the regeneration effluent from the anion exchanger, a mixture of NazCr04 and NaOH, has to be decationized with a strongly acidic cation exchanger for conversion of the chromate to chromic acid. This example drag
- out
I I
I I
I I
I I
I I
I I
--------
I for 1st port of reg
Regen e f f l u e n t
------
Fig. 7. Flowsheet for the recycling of the components from nickel plating rinse water.
17
shows that, not only is recycling possible, but there are also opportunities to convert a less usable by-product (NazCrO,) into a more desirable one (chromic acid) by additional treatment of the regeneration effluent with the strongly acidic resin. Another approach to the recirculation of rinse water constituents is the application of ion exchange for the recovery of water and nickel salts from the rinse water coming from nickel plating operations. The flowsheet of a typical process for recycling nickel and water is shown in Figure 7. The nickel is taken up on a strongly acidic cation exchanger in the decationizer part of the ion-exchange plant. Because the regeneration effluent contains too much sulfuric acid, a twostep regeneration is recommended. The first part of the regeneration effluent, rich in nickel sulfate, is neutralized with NiC03 and reused for make-up of the bath. The second part is collected and used for the first part of the regeneration. The decationized effluent is deanionized on an anion-exchange unit and the demineralized water is recycled. The regeneration effluent from the anion exchange unit is discharged as waste.
3.1.3. Recycling of Condensate Constituents
A typical example of recycling water and chemicals from a condensate is the recovery of water, ammonia and nitrates from fertilizer industry effluents. The flowsheets, which show the general rules of application of ion-exchange methods for condensate purification, are presented in Figures 8-10. Full demineralization of the condensates by ion exchange leads to the production of water and concentrated salt. The water is used for make-up of cooling water, or after polishing with a mixed bed is used for boiler feedwater. Decationized condensate from the processing of ammonium nitrate can be used instead of demineralized water for absorbing nitrogen oxides in the processing of nitric acid. The ion-exchange processes used for the recycling of condensate constituents in the fertilizer industry can be arbitrarily and conveniently divided into two parts: the recovery unit and the water polishing unit. This division is artificial. However, it can greatly help in arranging the published data. The main aims of the recovery unit are to achieve a rough separation of water and solutes, and to increase the concentration of the latter to the highest possible level; i.e., to divide the condensate into roughly demineralized water and concentrated salts. The partially demineralized water can be used for some purposes without further treatment. However, for many potential uses such as boiler feedwater, it is too contaminated and a polishing unit is often required. A mixed bed can be used for the polishing step in the manufacture of ammonium nitrate or sulfate. However, polishing of the analogous demineralized condensate from the manufacture of ammonia depends on the chemical composition of the treated condensate, which varies from plant to plant. in some cases the condensate contains only ammonium carbonate, but often significant amounts of chloride and sulfate are present also. Efficient polihng methods are based on processes used for the demineralization of low-salinity waters. In many cases the placement of a weakly basic anion-exchange bed before the mixed bed leads to savings in caustic soda and to prolonged working cycles for the mixed bed. The regeneration of the polishing units should be carried out in the same manner as in conventional demineralization plants. It is also possible to mix the partially demineralized water with water to be deminera-
18 I
r-
I
1
Processing of ammonium nitratr
t
L
1
Condenro te NH,+ + NO;
-
I
I
I I I I I
I I
I I
c-I I
I
o f f luent for regen.
I I
m Weakly bosic
I I I I
’//
exchanger
t e f f I uent
-4 ’
Roughly deminerolized woter
u I
Fig. 8. Flowsheet for the recovery of water, ammonia and nitrates from condensates obtained in the manufacture of ammonium nitrate.
lized in a conventional ion-exchange plant, as all nitrogen plants possess such units for the processing of boiler feedwater as well as for some process requirements. Because the roughly demineralized water has a low salinity it is possible to increase the flow rate through the conventional ion-exchange bed without any loss in quality of the demineralized water produced. Also, the low salinity means that it does not contribute much to the loading of the ion-exchange bed. The use of aconventionalion-exchange plant in this way significantly simplifies the system used for the recycling of condensate constituents.
3.2. Sewage
A very important environmental problem exists because of the rising levels of nu-
19
Condensate NH,* + C0;-
I ---
I I
Regen. effluent to processing nitrate or sulfote
for C02 removal
Roughly demineralized water
Fig. 9. Flowsheet for the recovery of water and ammonia from condensates obtained in the manufacture of ammonium nitrate (variant 1).
trients such as nitrate and phosphate in surface waters. Their presence has caused a serious deterioration in the water quality of many rivers, lakes and reservoirs. Biological growth has been greatly accelerated in many natural waterways, to the extent that eutrophication is commonplace. The reasons for this problem are the increased use of inorganic fertilizers and the expense associated with adequate treatment of domestic and industrial wastes from densely populated cities. Biological methods for the treatment of sewage are well known, but they lead to the mineralization of organics so that the sewage effluent generally contains nitrate and phosphate. More sophisticated and costly biological methods now exist for the removal of both these species. Thkir discharge to the sewer can be reduced by recycling wastewater constituents from the fertilizer industry, as already outlined. By far the larger share of nutrients, however, arises from sewage and fertilizers applied to the land. The latter non-point sources of pollution cannot readily be controlled. Attention has therefore been given to the removal of these materials from sewage effluents by ion exchange. An Italian proposal for the adsorption of ammonium ions on clinophilolite, and of phosphate ions on an anion exchanger, is of interest. As shown in Figure 11, the system
20 Condensate NH3: 140- 270, CO,-- :160-340. CI, No; < 1.6 mg/l
r - - - - - - -1 I
1
I I I I I I
upto 30%
I
I I
ammonium nitrate or
I
rulfote
I I I
I
I I
tL - Roughly
demineralized water
Fig. 10. Flowsheet for the recovery of water and ammonia from the condensates obtained in the manufacture of ammonium nitrate (variant 2).
involves the recovery of these species as MgNH4P04, which is precipitated from the regeneration effluents. Another proposal, from South Africa, uses ammonia and nitric acid to regenerate anion and cation exchangers employed for desalting sewage effluent. The combined regeneration effluents, being rich in ammonium and nitrate ions, and containing some phosphate, are recommended for use as a fertilizer. The reclaiming of water from sewage effluents has received considerable attention. The incremental load of salt which accumulates in domestic wastewater, usually about 400 mg/L, can lead to the production of effluents which are too saline for reuse by industry, especially when the original water supply is already highly mineralized. The salinity is then at corrosive levels. Furthermore, the presence of nutrients may lead to eutrophication problems. The most common application in industry for sewage effluents is as cooling water. The upgrading of these effluents for higher uses has received some study, but the use of ion exchange in this area is a most difficult task, and attempts to use such pro-
21
L Sewage
.....
- -niltoline ~arre- NOCl - --
r
I
Clinopttlolite
I I
I
L
I I
- - - -Regen - - - - - -J effluent
- -No - -CI- -
I
for r e g e n .
'I-----
Ammonia
1 Adjustment of composition p r e c i p i t a t i o n of M g NH4 PO4
A
Effluent ( free of nutrients)
Fert I I i zer M g NHs PO4
Fig. 11. Flowsheet for the recovery of ammonium and phosphate ions from sewage effluent.
cesses have not yet reached full scale implementation. However, pilot studies have been made in several parts of the world. Resin fouling, especially of anion exchangers, can cause problems, but these are not crucial. For example, it has been found that organic compounds present in the effluent obtained from physicochemical treatment of municipal sewage are adsorbed onto thermally regenerable resins in the cold, and desorbed in the hot regeneration stage. Operation with a trap resin is desirable, as very small amounts (0.2 mg/L) of organic anions akin to humate ions are not thermally desorbed. Magnetic ion exchangers offer a new per-
22
Grafted polymer conta;ning ion exchange groups
-
Cross1inked
Polyvinyl Alcohol
7- Fe, O3
Fig. 12. Diagrammatic representation of a magnetic shell resin.
spective as they can easily withstand high levels of suspended matter when used in truly continuous fluidized-beds. Also they should be less readily fouled because the ion-exchange polymer can be present in an uncrosslinked form which is grafted to the exterior of a magnetic polymeric core, to form an active shell as shown in Figure 12. There is an urgent need for processes which will remove specific pollutants such as heavy metals as part of the sewage treatment system, whether by contact with raw sewage, sewage effluent, or sludge. The high content of calcium and magnesium ions in these wastewaters, often accompanied by ferric ions, makes the equilibrium unfavourable when conventional ion-exchange resins are employed. The innocuous multivalent cations tend to occupy a significant proportion of the ion-exchange sites, to the exclusion of the offending species. Chelating resins have possibilities in this area, and magnetic forms are again an advantage as they may be more easily separated from slurries and sludges. To date, no large scale demonstration of the use of ion exchange in renovating sewage effluents has been conducted for prolonged periods to determine long-term fouling characteristics. Opportunities exist here for the application of ion exchange, as well as membrane processes, which have also been explored in this area.
3.3. General Remarks
We have not reviewed all the known methods for the conservation of water by recycling. The main aim of this presentation is to show that the same technology can be used as is currently employed for the preservation of the water environment. The crucial item is to find suitable efficient methods to separate and concentrate pollutants in ;L form which makes their reuse possible. Many chemical methods of separation have been developed. Ion exchange is one which allows ionic impurities to be separated from liquids and obtained in a more concentrated form. It is similar in efficiency as reverse osmosis, and in some cases, evaporation and precipitation. For the separation of non-ionic impurities a sorption process based o n activated carbon o r a polymeric adsorbent may be feasible. Liquid-liquid extraction may be used for both types of impurities. In general, one may say that technology can provide efficient tools for decreasing or even halting deterioration of the water environment.
23 4. THE AIR ENVIRONMENT
The Earth’s atmosphere, along with water, is the main component of the environment, as evidenced by its selection as one of the ancient Greeks’ four ‘elements’. It forms one common environment for the life of our Planet, is an irreplaceable reservoir of free oxygen and acts as a filter protecting life on Earth against lethal shortwave radiation from space. The Ionian philosopher Anaximenes, 100 years before Empedokles (one of the originators of the four ‘elements’ theory) said that everything that exists has come into being from the air and will turn into air. There is not much overstatement in this pronouncement. In relation to life on Earth, in some respects it is still quite true. According to the belief of a significant part of the scientific community, the Earth came into being from a gaseous dust cloud and the first reaction leading to the formation of an aminoacid took place in the primordial atmosphere. Every living organism has come into being from the air in the sense that changes in the composition of the atmosphere strongly influenced the climate and thus determined the way for biological evolution. In addition, as the 8 S , it is possible, at least Earth’s biomass has the empirical formula C1480H296001480N16P1 in theory, to convert this biomass into a gaseous product. That is, in accordance with Anaximenes it will turn into air. The present composition of the atmosphere allows for the existence of a variety of different biological life forms. However, the influence of the Earth’s atmosphere on living organisms in the past shows that any changes in the atmosphere caused by Man’s activity can affect the present living forms just as strongly. For centuries changes in the composition of the Earth’s atmosphere were relatively slow, and temperature changes were reversible. Most organisms therefore had enough time for adaptation. Although some organisms have disappeared, these changes have not impoverished biological forms on our Planet, as new organisms, better adapted to the changed environment, came into being simultaneously. The extent of atmospheric changes which have occurred in the last few decades are the same as those which have occurred during the past few hundred thousand years. The present changes are not only rapid and profound, they seem to be permanent. If the changes do not yet exceed the biological ability of the organisms to adapt, it seems that this may happen very shortly, as recently the rate of disappearance of some biological varieties and the decrease in population of others has accelerated. As far as the present changes in the composition of the Earth’s atmosphere are concerned, they influence Man mostly by a deterioration in living conditions, which causes among other things problems with new diseases, the so-called ‘civilization diseases’. The consequence of these diseases for Man is not yet known. It is likely that the biological ability of organisms to adapt to fast changes in the Earth’s atmosphere is reaching a limit, so that it is time to change Man’s attitude towards the environment. It seem that it is necessary to set back changes and restore more advantageous conditions for living biological forms. If such changes are not implemented today, tomorrow may be too late. 4.1. Causes of Air Pollution and the Possibility of Prevention
One essential cause of pollution of the air is the tendency to decrease the cost of ma-
24 nufacturing goods by the use of contaminated raw materials without purifying or enriching them before use, For example, a preliminary desulfurization of coal is still very rare. The problem exists particularly in processes where air is used as a source of oxygen. In such cases the nitrogen of the air is a diluent which, after the oxygen has been consumed, is discharged t o the-atmosphere together with some impurities. On approaching close to thermodynamic equilibrium the reaction rate converges to zero so that the intensity of production drops. Therefore, very often attempts to obtain a yield higher than 90 per cent is, from an economic point of view, not justified. As economics are the govering factor, the residual unreacted gases are discharged to the atmosphere. The amount of polluting chemicals is proportional to the amount of gases discharged. Hence the replacement in such processes of air by pure oxygen can lead to a substantial reduction in the amount of gases employed (about four times) and therefore to a decrease in the amount of pollutants discharged to the atmosphere. In some cases, it is even possible to eliminate pollution as recirculation of unreacted gases free of nitrogen can be economic. Often the use of oxygen instead of air can avoid the formation of large amounts of hazardous compounds, as for example in the formation of nitrogen oxides in combustion processes, thus virtually e h i n a t i n g their discharge to the atmosphere. Small quantities can be formed from impurities in the coal. There are two examples which show that the replacement of air by pure oxygen is a solution which may significantly decrease the deterioration of the atmosphere. There is no technological difficulty in implementing such processes as the use of oxygen on an industrial scale is quite common for the production of synthesis gas, for the manufacture of steel by the converter method or for the supply of oxygen in submarines. These examples show that the processes of oxygen production on both large and small scales are well documented. Therefore if oxygen is not employed in processes where its use would decrease pollution of the atmosphere, it is not because there is a technological barrier to the generation of the oxygen, but because its leads to an increase in the cost of production. The choice is only a matter of economics. Therefore implementation of appropriate regulations leading to an increase in the cost of air use, as well as discharge to the atmosphere of gases used in production, would enforce the increased use of pure oxygen for production. The manufacture of pure oxygen itself is free of pollution. It is worthy to note that the world uses about 10 billion tonnes of oxygen for industrial purposes. The generation of this amount of oxygen would require the use of 5 X 10” kWhr of energy, which is equivalent to 610 million tonnes of ‘fuel units’ or to about 20% of the amount of coal used at present. From this, it is apparent that the main limitation to the use of oxygen is an energy barrier, and it does not seem that at present it is possible to assign such an amount of energy for the production of oxygen. Nevertheless, protection of the Earth’s atmosphere requires the wider use of pure oxygen in industry, especially in those processes which are most harmful to the environment. From this point of view there is a need to develop both cheaper and more efficient methods for the generation of oxygen and cleaner, cheaper sources of energy, which would improve the economics of replacing air with pure oxygen. We have seen that energy is a significant component of our environment, which justifies its place as one of the ancient Greeks’ ‘elements’, Fire. Another approach to decreasing the extent of air pollution is to use interstage separation of reaction products. Examples of such technological solutions, where emission of pollutants to the atmosphere is almost e h n a t e d are the multistage Claus’ process for
25
the removal of sulfur compounds from petroleum and the production of sulfuric acid by the contact method with double (interstage) absorption of SO3. From the above, it can be seen that there are two main actions necessary to minimize pollution of the atmosphere: - use of pure oxygen instead of air - interstage separation of reaction products. The processes used for purification of gases discharged to the atmosphere do not very often result in the utilization of the separated impurities, with the result that the pollutants are merely transferred from the air to water or to soil. This does not solve the problem of environment protection. Dusts and fogs are another group of air impurities. Modern techniques provide efficient means for the removal of dust. Therefore, if the present state of removal of a dust from discharged gases is not satisfactory, it is not because of any technological limitation but because of insufficient legislation or lack of enforcement. The ways of protecting the Earth’s atmosphere discussed above require the use of significant amounts of energy, and do not lead to the elimination of COz emissions. An increase in the amount of heat being dispersed in the Earth’s atmosphere and the continuous increase in COz content can lead to a rise in the Earth’s temperature, followed by a change of climate (the so-called ‘greenhouse effect’). However, no-one can say whether such an effect will really occur in the near future and what the consequences will be. Hence the present ideas on counteracting the ‘greenhouse effect’ are closer to science fiction than to real technical solutions. There is no real conception of how to deal with the problem of a constinuous increase in the COz content of the atmosphere, although it cannot be denied that the use of renewable energy resources could ameliorate the effect. 5. THE PROBLEM OF ENERGY
The demand for energy will increase parallel to the growth in population and economic development. One may expect that the consumption of energy in developed countries will grow in proportion to their production capacity and standard of living. This growth seems to be accompanied by an effort to decrease the consumption of energy resources, especially the consumption of oil. In underdeveloped countries, where the majority of the human population is living, to ensure an adequate supply of food as well an improvement in the standard of living, a substantial increase in energy consumption will be necessary. In general, many countries will need an increase in the consumption of energy and even extensive exploitation of new, less convenient sources. Some decrease in energy consumption during the last few years arose from a sharp increase in the price of oil during the early 1970s. It was followed by a rise in almost all prices and initiated a recession in many countries. In underdeveloped regions the increase in oil price halted a lot of economic development programs. 5.1. Structure of and Use of Energy Resources
The most important energy resource is oil. It is not only highly concentrated and easy
26
to transport, but during its processing the products obtained are very convenient materials for chemical synthesis. Thus the production of plastics and synthesis fibres on a large scale was possible only because of the availability of the starting materials from oil. By the end of the 1970s the world wide consumption of energy resources was distributed on a percentage basis as follows: Coal Oil 1979 (actual) 24.5 41.8 2000 (predicted) 23.7 25.2
Natural Gas
Hydroelectricity
Nuclear
Others
17.3
5.1
2.1
9.2
19.2
5 .o
21.0
5.9
According to forecasts, the total energy consumption by the year 2000 will have doubled, and the use of the particular resources will be as above. According to this forecast, the consumption of all kinds of energy resources will increase, but the expansion of oil consumption is going to be substantially slower. The largest use of oil will be in motor fuel and material for chemical synthesis. It is expected that the consumption of coal will have doubled like the consumption of all energy sources and therefore its percentage of the total will be about the same. It is also expected that there will be a sharp increase in the utilization of nuclear power for the generation of electricity and for the supply of heat to large cities. Its total production will have increased about six-fold. It seems that energy supply is a limiting factor in the economic growth of many societies. Mining and converting energy resources to useful forms like electricity or heat require a large financial investment, of which the proportion necessary for the protection of the environment is substantial. Governments, being under considerable economic pressure, do not tend to take serious note of the threat to the environment coming from such processes.
5 . 2 . Environmental Threats from Energy Resources
5.2.1. Oil
During the drilling and tapping of oil wells located in the sea, an accidental out-flow of oil, which is difficult to contro1,can cause damage to biological life in the sea and- result in a local ecological catastrophe. Transportation of oil by ships is another significant threat. Accidents to such vessels may cause enormous damage, especially when they take place close to an urban area, as with the disaster involving an oil tanker close tg the French coast. Apart from the above mentioned threats caused by accidents, there is also substantial pollution of the sea which arises from normal shipping operations. Emulsified hydrocarbons cover the surface of the water close to busy seaways and ports. Oil is spilt during pumping from and into ships, during the cleaning of tankers, etc. These hydrocarbons cause significant damage when deposited on beaches. This is not a consequence of their use, but happens because of thoughtlessness or a lack of protection facilities. The processing of oil in a refinery also threatens the environment because it involves the possibility of hydrocarbons spillage and their discharge with the purified wastewater
27 when the effluent treatment plant is not adequate. The atmosphere becomes polluted with hydrocarbons via evaporation, as well as with combustion gases coming from the burning of excess volatiles. Pollution by carbon monoxide, unbumt hydrocarbons, nitrogen oxides and lead emitted by motors using liquid fuels is a separate issue. The use of petroleum-based fuels, especially those for automobiles, is a substantial consumer of oil. The air pollution generated by these motors becomes a serious problem in big cities. Recently, enormous progress has been achieved in reducing this pollution by improvements in engine design which give greater combustion efficiency, and by the introduction of lead-free gasoline. However, these innovations apply only to well developed countries, so there is still a need to promote them in the less developed countries. Diesel engines, which have a higher thermodynamic efficiency and generate less pollution, are less harmful, but need to be kept in proper tune. They should be more widely used, to run small cars as well as heavy vehicles. A significant amount of fuel oil is used, especially in well developed countries, for the heating of houses, for the production of heat for industry and for electrical power generation. The main pollution coming from the combustion of heating oil is the emission of sulfur dioxide and, if the combustion is not optimized, some amount of unburnt hydrocarbons. The problem of sulfur dioxide emission is serious, as it is mainly responsible for ‘acid rain’, one of the most harmful forms of pollution of the atmosphere. Sulfur dioxide is formed during combustion by oxidation of sulfur compounds such as hydrogen sulfide, sulfides, bisulfides and mercaptans present in the oil. Desulfurization of oil is technologically possible, but the process of desulfurization is expensive and therefore rarely used.
5.2.2. Natural Gas
Natural gas is widely used for municipal purposes, in the steel industry and to a lesser extent in the generation of heat for industry. Another economical way of using it is for the production of synthesis gas, and it is widely used for this purpose. Natural gas very often contains hydrogen sulfide and low molecular weight mercaptans, which are oxidized to sulfur dioxide during combustion. These are harmful to the environment as they are responsible for ‘acid rain’. Therefore desulfurization of natural gas is the only way to protect our environment. It is worth noting that a substantial amount of the sulfur used around the world is obtained from the desulfurization of gas and oil. Some natural gas, especially that having a low content of sulfur compounds, contains mercury, which has to be removed before use as it is very toxic and has the ability to accumulate in living organisms.
5.2.3. Coal
World resources of coal are considered to be several times larger than oil and natural gas resources. The use of coal should be encouraged for the generation of electricity and to produce heat for cities. It is also expected that in future the processes of liquification and gasification of coal will be used on a much larger scale. The economics of the present
28 processes are still unfavourable for the replacement of oil and natural gas by liquid and gaseous products from coal. It seems that coal is and will be the main source of energy to generate electricity and to provide heat for the centralized heating systems of cities. The major usage of coal is for these purposes. The combustion of coal generates two types of air pollution: - ash and dust which are separated from combustion gases by the use of cyclones or electrofilters - sulfur dioxide coming from sulfur compounds present in the coal. Both the amount of dust and sulfur dioxide are determined by the quality of the coal. The amount of dust emitted depends on the mineral content of the coal, which varies from 5 to 30 per cent in hard coal and up to 45 per cent in brown coal. The sulfur is present in the form of pyrites (FeS?), and the total sulfur content varies from 0.8 to 5 per cent. The mass use of coal seems to be limited by the present technological system i.e., problems associated with the mining and transportation of huge amounts of coal and the emission of sulfur dioxide as well as the disposal of ash and dust. Storage of huge amounts of ash has become a serious ecological problem. Some attempts have been made to use the ash for the production of building materials. However, this can create a new problem, as some mdioactive elements, isotopes of U, Th and K, tend to concentrate in the ash. Building materials made with such ash can be radioactive. If we assume that a country is burning 50 million tonnes of coal per year, and the coal contains an average of 1.5 % sulfur, its combustion will generate 750,000 tonnes of sulfur dioxide. This amount after oxidation with oxygen from the air will form 1,100,000 tonnes of sulfuric acid, which will come down with the rain. Desulfurization of combustion gases with alkaline solutions or suspensions, as well as with dry processes utilizing the oxides of calcium or magnesium, require large and expensive facilities because of the huge volume of gases to be treated. One way to bum coal containing sulfur is to use a mixture of powdered coal and the oxides of calcium and magnesium. These oxides combine with sulfur dioxide and form sulphates. Even if this process could be made to work satisfactorily, it does not eliminate a pollution problem for it generates a substantial amount of solid waste. It is also possible to remove sulfur directly from coal. However, it is necessary to use two different methods: one for the removal of inorganic sulfur compounds and the other for organic sulfur compounds. Inorganic compounds of sulfur (pyrites) are removed from powdered coal by the use of sedimentation and magnetic separation techniques It is also possible to oxidize pyrites by a solution of ferric sulfate, when ferrous sulfate and sulfur are formed as the reaction products. Regeneration of the ferrous sulfate is possible by aeration. The removal of organic sulfur compounds is possible by pressurized extraction with suitable fractions of oil in the presence of hydrogen. Under these conditions almost all the sulfur is reduced to hydrogen sulfide, and in addition a substantial part of the coal is transformed into a liquid which can be used as a desulfurized coal, or can be further hydrogenated to produce a lighter fraction of hydrocarbons. Thus the extensive use of coal seems to have significant limitations: costly mining and transportation on one side and pollution of the environment on the other - especially by sulfur dioxide emissions. Resources of coal having a low content of sulfur are limited.
29 5.2.4. Nuclear Energy
Generation of electricity and production of heat by the use of nuclear energy has met with protest from a significant part of society, especially in the well developed countries where the issue is highly controversial. A well made and properly operated nuclear power station is less harmful to the environment than a power station using conventional fossil fuel. Apart from a minimal emission of small amounts of gaseous radionuclides at almost undetectable levels, and a substantial amount of hot water from the cooling system, a nuclear power station does not cause pollution. Nuclear power stations can be a threat to the environment in the case of an accident, as it is possible to get a leakage of radioactive elements to the environment which could not happen with any other kind of power station. Also, it is possible to pollute the environment during processing of the exhausted nuclear fuel if particular care is not taken, and during storage of radioactive waste. Radioactive isotopes of long life are produced by the nuclear reactions which take place. Radioactive waste can hence be of potential harm to the environment for many generations. Its disposal is the most serious problem confronting the industry. At the beginning of the development of nuclear energy, the radioactive waste was immobilized in concrete containers and buried in the ocean bed. However, this method is not safe as the cbncrete containers corrode and leakage of the radioactive materials is possible. A more recent and safer method consists of placing the waste in containers made of a special glass which are stored underground in old salt mines. It is worth mentioning that a futuristic proposal has been made to send radioactive waste into space, towards the sun. Nevertheless it seems that rapid development of the use of nuclear energy faces a major problem in the disposal of nuclear waste.
6. THE IMPACT OF CHEMISTRY ON THE ENVIRONMENT
We have been talking mostly about the application of chemical methods for the protection of the environment. Frequently, pollutants are generated by plants using chemical processes for the manufacture of goods. We have emphasised the positive side of chemistry in providing methods for the neutralization of various wastes. Now we would like to glance at the role of chemical products in Man’s life. The number and variety of chemical products used in every day life is growing rapidly. For their manufacture, new chemicals are sometimes used which have unknown or detrimental health effects. Regulation of the use of chemical products in everyday life is even more difficult than regulation of industrial activity, because there are too many products and a great variety of items. In addition, people having a lack of awareness of this danger often have a strong need to use the chemical products. Because a lot of chemical products are very appealing they attract attention. To a certain degree it is like primitive societies who used to exchange their gold for trinkets with the early European explorers. Attempts to introduce effective ways of prohibiting the use of products known to be hazardous, such as tobacco and drugs, are examples of unsuccessful control by legislation. A typical example of a threat to the environment coming from chemicals can be seen from the history of the use of pesticides. The introduction of a new pesticide onto the
30 market is determined by the ratio of its cost of manufacturing and use to the expected profit. If the profit is hgher and the research of 2-3 years does not show a detrimental effect on the health of Man, the use of a particular pesticide is considered justified. Very often the grounds on which a decision to use a chemical is made are subjective and controversial. After such a short research period it is hard to determine unequivocally whether or not a chemical has mutagenic and/or carcinogenic properties. To find such effects often requires a time equivalent to a few generations of the species, and the results obtained from experiments on animals surely do not indicate that the chemical will have the same effect on Man. The permissible content of chemicals in our environment is based, in essence, on Paracelsus’ belief that everything and nothing is poison, as a poisonous effect depends on the dose. However, taking into account present day knowledge of the accumulation of poisons in organisms and of the existence of highly active substances like pheromones, one molecule of which can initiate a physiological reaction even in quite complicated organisms, it is hard to say what the long term influence will be of the new chemicals being brought into contact with Man. Therefore a belief that the present permissible dose of a substance is ‘safe’ may not be fully justified. In the same way, some doubt may arise when we look through the so-called ‘safe conditions of use’ of a chemical. For example, it was ‘scientifically’ established that when the land-based spraying of chemicals was carried out, the diffusion of chemicals was contained within a zone of 300 m radius. It was later found, however, that the chemicals could be smelt several hundred metres distant from the safe zone, and were eventually detectable by chemical means at distances some thousands of kilometres from their point of use. Therefore, there is a need to establish strict procedures for the acceptance of new chemicals for common use. It seems that from the above standpoint the following items are of special importance: - to better understand the toxicology of the environment - t o introduce into school curricula, mandatory lessons on problems of the environment (both of the threat to the human population and how to prevent deterioration of the environment) - to establish one uniform regulation on an international level which describes the procedure for accepting new chemicals for common use - to make up a balance sheet of gains and losses in the use of a particular chemical, taking into account as well the long term effects of possible mutagenic, teratogenic or carcinogenic properties in order to prevent ‘suicide by instalment’.
I. A FEW LAST WORDS
As one can see it is not easy to give a simple answer to the question are technology and the environment allies or antagonists? It seems that the development of technology, for all its consequences, has improved the quality of life by making Man’s existence richer and more meaningful. In this sense technology has not been an antagonist to the environment. However, some doubt arises as to whether this will remain so. It seems that we are at a stage where it is not clear whether the further development of technology will really improve our quality of life or cause it to deteriorate because of
31 the environment becoming less healthy. Pollution in some parts of the world has reached a level close to ecological catastrophe. In these places the new, very modern goods which the people have do not compensate for the deterioration in their way of life caused by pollution. Therefore, it seems necessary to influence the existing model of human life, so that we balance the commonly called ‘standard of life’, depicted by possession of goods (both necessary and unnecessary), against the quality of the environment. It would be difficult, if not impossible, to have both a continuously growing consumption of goods and a healthy environment. The most obvious limitation is the finite amount of mineral resources available. But more serious is the energy limitation, as the transformation of energy from available fossil fuel resources into usable forms contributes significantly to the deterioration of our environment. The pollution load will be even higher when poorer quality energy resources of t h s type are used. We have not discussed renewable energy sources here, including those derived from biomass and solar systems. These are certainly less polluting sources, but are capable of meeting only a minor fraction of our energy needs. Anti-pollution measures are inseparably bound up with an increased consumption of energy. We have already quoted the example of pollution of the atmosphere being decreased significantly by replacing air as a source of oxygen with pure oxygen. Such a move requires a great amount of energy to be consumed, the production of which is not pollution free. In general, to decrease the pollution generated during the production of usable forms of energy from fossil fuels an additional consumption of energy is necessary. Therefore, to ensure that processes are less polluting, only a part of our energy resources would be used for direct human needs and the rest for antipollution measures. Also, they would not eliminate the ‘greenhouse effect’. Therefore, in answer to our question as to whether technology and the environment are allies or antagonists we reply with some hesitation that they are still allies provided that Man will use his present knowledge, developed and strengthened by technology, more for the production of the essentials of life than for the generation of more and more, less and less necessary goods, so that a balance can be kept between manufacturing requirements and the quality of the environment.
This Page Intentionally Left Blank
33
THE STATE OF THE ENVIRONMENT REPORT OF UNEP
Prepared by ENVIRONMENT LIAISON CENTRE*
ABSTRACT Hazardous wastes that endanger human life and health, acid rain which is damaging soil and water over large areas of industrialised countries, and the potential environmental impact of hardnessing the energy of plants through the so-called energy farms, are the issues analysed in this year’s state of the environment report prepared by the Nairobi based United Nations Environment Programme (UNEP).
1. HAZARDOUS WASTE
It has been estimated that over five million chemical substances have been identified; about 70,000 of these are marketed, maybe only half of them in quanity. Several thousand new ones are found every year, and about a tenth of the new discoveries reach the market. For example, the total production of synthetic organic chemicals rose more than 50% in the past decade. Those chemicals have brought immense benefit to society, but they have also brought new dangers, largely through the wastes generated in their manufacture. Tens of millions of tons of toxic or otherwise hazardous substances enter the environment every year. One of the most worrying features of the problem is that very little is known about the long term consequences of exposure to the chemicals. We know now that over longer periods some can cause cancer, delayed nervous damage, malformations in unborn children, and mutagenic changes. Many other chemicals are likely to have similar effects, but because these take time to show and their causes are hard to pinpoint, we do not yet know which substances are the dangerous ones. The situation is made even more difficult because, once they are in the environment, chemicals spread in a very complex way and may be converted into other substances which have different effects. Until recently, many hazardous wastes were disposed of without proper evaluation of the environmental consequences such as fires, explosions, air, water and land pollution, contamination of food and drinking water, damage to people and harm to plants and animals. In practice, most of the things that could go wrong have indeed occurred, and in
* The editors like to express appreciation to Mr Delmar Blasco, Executive Officer, Environment Liaison Centre, Nairobi, who made this report available for readers of this book.
34
fact, the incidents that have hit the headlines are probably only a few of those that have actually taken place. Perhaps one of the most notorious incidents was that of the “Minimata Disease” in Japan, where discharge of methylmercury to the sea caused the contamination of the fish, which in turn caused neurological disorders to nearly two thousand people; about 400 of them have died. In the USA an area of at least 30 square miles was contaminated with wastes from manufacture of defoliants, pesticides and chemical warfare agents, causing irrigated crops to die and livestock drinking from the wells to perish. Also in the USA, at Love Canal, people have to be evacuated from homes built on a former dump containing pesticides, chemicals used in the making of plastics and the sludge from the bottom of stills. In the Netherlands, drums of paint solvent were included in rubble used to reclaim land on which houses were built; here also, people had to be evacuated and in both cases decontamination measures cost tens of millions of dollars. These incidents may seem anecdotal and devoid of any statistical significance. This is a faor reflection of the truth. There is a lack of information on the magnitude and frequency of such incidents. Many authorities speculate that the full extent of the problem has yet to be revealed. They point to the damaging incidents that have been reported and the fact that they have only been discovered when things have gone badly wrong. Sometimes the discoveries have been by chance. Incidents may well go unreported unless someone has diligently investigated them. The major actions taken to deal with the problem are the national laws for controlling the disposal of hazardous waste now in force in the developed countries and international agreements on limiting marine pollution either from disposal at sea (the London Convention) or from discharges from the land (the Paris Convention). These conventions on pollution of the sea have been supplemented by regional agreements where countries bordering particular seas like the Baltic the North Sea and the Mediterranean make possible more effective action to solve their local problems. On land, there have also been major clean-up operations on hazardous waste sites in a few developed countries. The laws have in general laid down that the disposal of hazardous waste should be approved by regulatory authorities, either on a case-by-case basis or through general regulations. They have brought about a marked improvement, by setting standards where few existed before. Companies have had to pay greater disposal costs, and this has encouraged them to save money by introducing better ways of dealing with wastes. They have increasingly included good waste management in the overall design of new processes, and chosen ones that generate as little waste as possible. They have paid more attention to using waste as a resource, reclaiming materials and fuels from them for rereuse. They have even begun waste exchange schemes, where companies advertise their wastes for scale to other firms that can use them as raw materials. And they have increasingly separated out and segregated different types of waste so that they can be reused or disposed of more economically. The immediate need is to make sure that the laws are enforced in a cost-effective and environmentally sound way. Some developed countries have still to create an effective enforcement system staffed with adequately trained people. Developing countries following the same legislative path may have greater problems in recruiting the right staff. International organisations could consider publishing manuals and providing training facilities
35 though these should be directed at dealing with the actual wastes generated in developing countries rather than at establishing comprehensive theoretical principles. The lack of trained staff is only part of the problem. There are so many companies carrying out so many operations with hazardous waste that even well-staffed authorities cannot guarantee full inspection. In the United States of America, for example, there are about 57,000 firms licensed to generate the waste, 14,000 licensed to transport it, and another 14,000 facilities licensed for disposal. Much, therefore, depends on the integrity and competence of firms - reinforced by the fear that they will lose their licences if they are caught misbehaving. As e.:t controls have tightened in many countries, chemical industries have had to pay more for getting rid of their wastes. Some have been tempted to avoid these extra costs by moving their operations or exporting their wastes to countries where the laws are less strict, or less strictly enforced. These countries could well become international dustbins, and end up with the same sort of problems that brought the strict legislation in the first place. There have even been a few cases where companies have shipped waste to another country, ostensibly for storage, and then abandoned it. Waste from the Netherlands ended up in the United Kingdom in this way, and wastes from the USA have been stored in a warehouse in Mexico. Developing countries would be particularly vulnerable to such pollution exports. Companies setting up in developing countries often stipulate that their processes must remain a secret. If they insist that the composition of their wastes should also be cloaked in secrecy the countries may never know exactly what hazardous substances, in what quantities, have been put into their disposal sites - and will find it almost impossible to control the situation. In fact, so much secrecy can rarely be necessary and, if companies do insist on it, countries should require them to give assurances about the hazards posed by their wastes, and to accept financial liability for the problems caused. Disposal practices and the degree to which they are controlled differ widely among developed countries. This is no doubt a reflection of differing public attitudes to the risks of environmental contamination by low levels of chemicals with largely unknown longterm effects. The differences are important enough economically to affect industry’s costs. So there is a threat that some chemical processes and some hazardous wastes may be moved to less demanding locations, including developing countries which may not be fully equipped to control these polluting activities. Hazardous waste can easily be slipped across frontiers. National definitions of it differ. It can be shipped, without difficulty, under labels which fail to inform customs officers what it consists of (there is no international agreement on how wastes-should be characterised), and it is very difficult for them to carry out checks. So the export of hazardous wastes is a distinct possibility - and incidents caused by careless disposal may follow.
2. ACID RAIN
Without the 110,000 cubic kilometres of rain that fall each year, the continents would be barred. Yet now the rain in parts of the earth has taken on a new and threatening complexity. It mixes in the air with pollution from burning fossil fuels - particularly in power sta-
36 tions, factories and motor vehicles - and brings down dilute sulphuric and nitric acid. This is killing fish and other water life, and corroding buildings, including some of the world’s most important ancient monuments. It may also damage forests and cmplands, and possibly pose a substantial threat to health. Acid rain is not a new phenomenon; what is new is the realisation that it is an international problem. The air of towns like Machester has been largely cleaned, partly by building tall chimneys at power stations and factories, which push pollution hgh into the air. These chimneys have made things better locally, by dispersing the pollutants, but aggravated the international difficulties. For the sulphur and nitrogen compounds emitted by burning fossil fuels can be blown thousands of kilometres by the winds, to cause acid rain in countries far from their points of origin. Acidification is an environmental problem, or becoming one, in parts of Europe and North America. Around five to ten million square kilometres of these continents are affected. Similarly, polluted areas are likely to exist elsewhere in the world, especially around large urban and industrial conglomerations. We do not yet know where they are, because so far no evidence on them is available. Industrial regions of the world suffer much more acidic fall-out than they did before the industrial revolution. This is because power plants, some industrial processes, vehicles and homes emit sulphur and nitrogen compounds, mainly from the burning of fossil fuels, and have greatly increased the amount of them in the environment. Natural processes also put sulphur and nitrogen compounds into the air besides manmade sources. Nobody knows precisely how much they contribute around the globe. Estimates vary between 78 and 284 million tons of sulphur a year in the form of sulphur oxides, and between 20 and 90 million tons of nitrogen a year in nitrogen oxides. In comparison, man emits between 75 and 100 million tonnes of sulphur a year. So, despite the differences in estimates natural sources, it can be concluded that man-made and natural emissions of sulphur are, globally, of the same order of magnitude. Burning coal provides about 60 per cent of the man-made emissions, burning petroleum products gives rise to another 30 per cent, and various industrial processes acount for the remaining 10 per cent. Approximate estimates indicate that burning fuel in electric power stations and industry provides almost three quarters of sulphur emissions in the European Economic Commission countries. Like sulphur oxide pollution, pollution from nitrogen oxides is also of the same order compared to natural sources. Fossil fuel combustion yields about 20 million tons of nitrogen a year which have already caused environmental problems on a regional and local basis in industrialised countries. Not all the pollution is acid rain, is sulphuric and nitric acid dissolved in precipitation. Some of it happnes when the sulphur and nitrogen oxides themselves fall out on the land, in what is known as “dry deposition”. In general this tends to be the main form of tHe pollution near its source, and the longer the gases stay in the air, the more likely they are to go through the complex changes that will turn them into acid rain (or wet deposition), to fall perhaps thousands of kilometres from where they began their journey. Wet deposition rates are fairly well known, but dry deposition is harder to calculate and rates remain more uncertain. Both types of deposition can be intercepted by vegetation canopies. The canopies of evergreen forests, in particular, can be subjected to high deposition rates. Each country gets part of its acid fall out from its own pollutants, but receives the rest
37
on the winds from neighbouring countries. EMEP has worked out estimates of how much sulphur is emitted, and how much deposited, in individual European countries. This reveals which of them are “net importers” and which “net exporters” of air pollutants. Lakes and rivers were the first victims of acid rain to become evident. Hundreds of lakes in parts of Scandinavia, the north-east USA, south-east Canada and south-west Scotland have turned acid. Parts of these areas are particularly vulnerable because their soil and bedrock offer little protection against acidic rain.They are made up of minerals like granite, gneiss and quartzrich rocks which contain little lime and do not weather easily, and therefore can do little to neutralise the acid when it falls. In Sweden, damage to fisheries attributed to acidification has been observed in 2,500 lakes, and is assumed to have taken place in another 6,500 where signs of the process have been found. Meanwhile, out of 5,000 lakes scattered over 28,000 square kilometres of southern Norway. 1,750 have lost all their fish and 900 others are seriously affected. In Canada, nearly 20 per cent of all the lakes that so far have been examined in Ontario have either been turned acid, or are extremely sensitive to the process. Between 30 and 60 per cent of the lakes in various areas of south-west Quebec are considered to be sensitive or hghly sensitive. And in Canada’s Atlantic provinces many lakes have been turned 10 to 30 times more acid during the past two decades. Similar situations have been observed in the north-east of the USA. As the water becomes more acid, the amount of aluminium in it starts to increase rapidly. Concentrations as low as 0.2 milligrams per litre of the metal in acid water kill fish. Large-scale fish kills have been recorded in some Swedish lakes, and these have been attributed to aluminium poisoning rather than to h g h acidity alone. At the same time, phosphates, which nourish phytoplankton and other aquatic plants, attach themselves to the aluminium and become less available as a nutrient. So increasing aluminium levels may reduce primary production on which all other water life depends. As the water gets more acid still, other metals, like cadmium, zinc, lead and mercury also become increasingly soluble. Several of them are hghly toxic, and some may be taken up by water life through food chains, though little evidence of this is available so far. Soils are normally much better able to resist acidification than lakes, rivers and streams, and so can take much more acid without noticeable ecological drawbacks. The acidification of soils is not merely due to acid deposition; it arises from a natural process as well as the result of biological processes within the soils. Normally, the acids thus produced are neutralised during the weathering of mineral soil particles but, depending on the composition of the soils, their capacity to neutralise more than a definite amount of acidity is limited. Acidification may cause nutrients like potassium, magnesium, calcium and other micronutrients to leach more rapidly out of the soil, decreasing soil fertility. Aluminium concentrations would rise, just as they do in water, thus damaging plants and reducing the availability of phosphorus to them. As in water too, metals like cadmium, zinc, lead, mercury, iron and manganese would spread through the environment more readily with acidification. Acid fall-out does seem to have a distinct effect on soil microbiology, chemistry and fauna - but the effects on the growth of plants, including trees, are far less clear. Indeed, depositions of nitrogen may’ even have a fertilising effect and increase productivity significantly, at least in the short term.
38 In the Federal Republic of Germany, on the other hand, 7.7 per cent of the forest area has been reported in 1982 to be damaged (7.5 per cent of the damage being light, 19 per cent medium and 6 per cent heavy) by a wasting disease due to the consequences of deposition and accumulation of air pollutants (20). In addition, trees have suffered more storm damage and experienced regeneration difficulties. These forests receive much more fall-out than Scandinavian ones because they are close to cities and big industrial areas, such as the Ruhr, with many polluting sources. As well as the health of important ecosystems, human health may also be put at risk by pollution. HI& concentrations of sulphur dioxide, nitrogen oxides and dust have long been known to be harmful. This issue is only marginally related to the problem of acid rain, since such concentrations are usually only found close to the sources of pollution, and sulphur oxide levels in many European and North American cities have been decreasing recently. Other, indirect, health hazards are suspected. These would be caused by the metals like lead, copper, zince, cadmium and mercury released from soils and sediments by increased acidification. They can get into groundwater, rivers, lakes and streams used for drinking water, and be taken up in food chains leading ultimately to man. The release of cadmium in particular may give rise to a growing problem as acidity increases, as normal levels in human food are already close to the acceptable daily intake. Acid water may also cause galvanised steel and copper water pipes to release metals, and it seems that the risk arises as soon as the acidity of the water rises above normal. Most drinking water in industrialised countries, however, is supplied by public water works which eliminate this problem with proper treatment techniques but much remains to be done in developing countries. Meanwhile, acid accelerates corrosion in most materials used in the construction of buildings, bridges, dams, industrial equipment, water supply networks, underground storage tanks, hydroelectric turbines and power and telecommunications cables. It can also severely damage ancient monuments, historic buildings, sculptures, ornaments and other important cultural objects. Some of the world’s greatest cultural treasures, including the Parthenon in Athens and Trajan’s Column in Rome, are being eaten away by acid fall-out. The damage to water can be alleviated by adding lime to lakes, rivers and streams and/ or their catchment areas. Many chemicals such as caustic soda, sodium carbonate, slaked lime, limestone, or dolomite can be used to counteract the acidity. Slaked lime and limestone are the most popular. Sweden began a liming programme in the autumn of 1976, and by the summer of 1982 about 1500 Swedish lakes had been limited at a total cost of about $ 15 million. Liming alleviates some of the symptoms of acidification, but it is no real cure, is not practicable for many lakes and running waters, and does not attack the causes of the problem. It should, however, be considered as an interim measure which offers some defence until the emissions of pollutants can be reduced to a satisfactory level, Liming can also be used to counterbalance the increasing acidification of cropland. Lime and calcium-poor soils for centuries. The cost of the extra liming needed to offset acid fall-out in Europe ranges from less than $ 1 to about $ 10 per hectare a year. The only lasting solution is to reduce the emissions of the pollutants in the first place. Apart from the effect that strict controls would have in protecting waters and forests,
39 they could save millions of dollars by avoiding corrosion. The Organisation for Economic Co-operation and Development (OECD) made a first attempt in 1981 to find a way of quantifying corrosion costs. This came up with estimate that strict emission control measures in 13 European countries could save about $ 1.2 billion in corrosion costs every year. But the report acknowledges that this is a very approximate figure and more work is being carried out to improve the estimates. The easiest way to control the pollution is to use fuels that are low in sulphur; but this will not be feasible for long because the world supply of these fuels is believe to be limited. A more permanent solution is to use other sources of energy instead of fossil fuels, and to improve energy conservation. Many users could reduce their energy consumption, and technical improvements could ensure that various processes burned fuel more efficiently. These measures will help cut down emissions of sulphur and nitrogen oxides, but, obviously, acidity of rainfall will not be reduced to agreed acceptable levels unless work is done to remove sulphur from fossil fuels, emission gases, or both and such work is therefore likely to acquire increasing importance in the future. It may also be necessary to remove nitrogen oxides from emission gases. Removing sulphur from fuel and gases creates waste products - solids and slurries which have to be disposed of properly to avoid water, groundwater, or soil pollution. Naturally, this problem grows as emission controls are increased. According to preliminary calculation by OECD, it costs a total of some $800, on average, to stop a ton of sulphur from getting into the air. More recent estimates suggest that the costs may be somewhat higher. If north-western and southern European countries were to cut their annual sulphur emissions by about half (around 5.9 million tons) within the next 10 to 25 years by controlling the emissions of conventional power stations, it would cost them about 10 per cent of the total cost of producing their electricity. OECD has also made estimates of the benefits resulting from emission controls, as part of a methodological study. These suggest that the benefits would outwegh the costs, but the uncertainties surrounding the estimates are so wide that they cannot be used to provide a quantitative and reliable evaluation of the balance between costs and benefits. There are, moreover, other factors that complicate such analyses. One - common to many other instances of damage to shared natural resources - is that the countries which would benefit from the reduced pollution would often be different from those that would have to bear the cost of cutting it back. Another is that all the estimates of benefits assume that the damage caused by acid rain can readily be reversed if enough pollution control is implemented - and in reality this may not be so. It might be a long time before the ecological damage, in particular, began to be reversed. Unfortunately, scientific information on the recovery process is extremely scanty. Furthermore, there must be more effort to research and developed ways of improving energy conservation, environmentally appropriate technologies for producing heat and power, and techniques for removing sulphur from fossil fuels and gaseous emissions. So far the acidification of the environment has been seen a regional problem, restricted to parts of Europe and North America. Other industrialised areas are almost certainly exposed to the same problem, but there is too little information to assess it. Besides, the problem may well spread to new areas as a result of rapid industrialisation and the growth of cities in other parts of the world, particularly developing countries. So it is important
40
that areas affected by acid deposition and susceptible to damage from acidification are identified as soon as possible. If they are, the damage could be mitigated, or even avoided, at a minimum cost to society by initiating research, and applying what we already know about pollution control and environmentally appropriate energy production technologies at an early stage. In the near future the crucial issue is whether countries are ready to take measures needed to cut back emissions to an acceptable level. The consensus reached by the Ministerial Conference on Acidification of the Environment, held in Stockholm in June 1982, was extremely encouraging about this. Representative o f 21 countries agreed that urgent action should be taken under the Convention of Long-Range Transboundary Air Pollution, including: a) establishing and implementing concerted programmes to reduce sulphur and, as soon as possible, nitrogen oxide emissions; b) using the best technology available that is economically feasible to reduce these emissions, taking account of the need to minimise the production of wastes and pollution in other ways; c) supporting research and development of advanced control techniques; d) developing and implementing energy conservation measures further; e) developing the North American monitoring programmes and EMEP further, through better geographical coverage, improved data on emissions, standardising sampling and measurement techniques, and improved modelling, among other measures.
3. ENERGY FARMS
There is an enormous amount of biomass on the globe; every year natural productivity adds enough energy to meet at least ten times all the world’s commercial energy demands, in theory. In practice, of course, it is very unevenly distributed around the world, and in some regions the reserves are being rapidly dissipated because they are being burned faster than they can be replaced by growth. In practice, too, biomass must be economic to harvest and turn into fuel. Almost any crop produced by farmers anywhere in the world, whether its main use is for food, animal feed, fibre or other products, can technically be turned into some form of energy. But in most cases it would not be economic, practical, or even sensible, to do this. Many countries are seeking to solve these problems by paying special attention to energy. Fuel crops can be fast growing trees, conventional crops, or water plants - any plSnts, or mixture of plants, which are more valuable as fuel than as anything else. Many countries have recently been paying a good deal of attention to wood plantations. A recent assessment by F A 0 and UNEP shows that plantations of wood lor industry, fuelwood, charcoal, other products and soil protection, cover about 2.7 million hectares in Indonesia, 2.6 million in India, 400,000 in Bangladesh, 300,000 in the Philippines and over 200,000 in Thailand. Plantations were also under way in Sri Lanka, Pakistan, Malaysia, Nepal, Burma and Brunei. Some suitable tree species for energy farms have already been identified. One importing group of them is the tropical legumes, which automatically provide their own fertiliser by “fixing” nitrogen from the air. As they grow they therefore actually improve the
41
soil. Leucaena leucocephala - the giant “ipil ipil” is one variety - is the best known member of the family. It is particularly bountiful, for it also produces high protein seeds and foliage which makes, an excellent animal feed supplement and green fertiliser. Other tropical legume trees, such as various acacia, calliandra and sesbania species, also grow fast. So do some trees from other families; eucalyptus and casuarina trees are among the most promising. Fast growing trees like these are enormously productive if they are well matched to local condotions. Many species will produce more than 20 cubic metres of wood per hectare every year when grown on reasonably good soil. In exceptional cases leucaena, eucalyptus and several other species are reported to have produced as much as 50 cubic metres per hectare a year. Once the wood is harvested it can be used to generate heat, steam or electricity. “Ipil ipil” plantations have already been established in the Philippines to produce electric power. They are intended to fuel several stations with a total capacity of 200 megawatts by 1987 - and by the year 2000, 700,000 hectares of wood plantations and 2,000 megawatts of electricity are planned. In Tamil Nader, India, 11,500 hectares of casuarina trees would provide fuel for a power plant generating 100 megawatts. Alternatively, wood can be turned into charcoal, gas or liquid fuel. None of these uses is new. Chrcoal has been produced since the dawn of civilisation; some say that making it was the first chemical process ever discovered by man. Relatively large eucalyptus energy farms (8,000 hectares), delicated to charcoal production for steel mills, have been in operation since the early 1950s in Argentina and Brazil. Wood alcohol was used as a liquid fuel for most of the eighteenth and part of the nineteenth centuries. Kerosene and other fuels took its place; but now there is new interest in it, mixing with gasoline for motor fuel. Making gas from wood and chrcoal is another old technology that is being revived. Several European countries and Japan had projects for fuelling engines with the gas before the Second World War. By late 1941, 70,000 cars in Sweden alone used it, as did 55 per cent of the trucks and 70 per cent of the buses. Farm equipment also made good use of it. After the war most vehicles switched to gasoline because it was a better and more convenient fuel. Now the gas is coming into its own again in several countries, particularly in the countryside. In the Philippines, for example, the process is used to run fishing boats, water pumps and some public service vehicles, and to produce electricity. Wood may be the most obvious and widely used form of “green energy”, but it is far from being the only one. Some plants, for example, produce materials like hydrocarbons, some of which are the main constituents of oil and natural gas. Once again, this has been known for centuries - pre-Colombian civilisations in Latin America systematically cultivated trees which produced liquid that could be made into rubber. Even today the natural rubber tree, hevea braziliensis, is much the best known of these plants. Its latex is no use for energy farming, since it is made up of hydrocarbons of a very heavy molecular weight; hydrocarbons of lower molecular weight are needed for fuel. Some studies have shown, however, that many species of plants may produce just what is required. Bushes of the euphorbia group seem to be particularly promising. Experiments with two species (euphorbia lathyris and euphorbia tirucalli) have shown that they can yield between 17 and 36 barrels per hectare a year. Better still, many of the 8,000 or so known species in the family will grow on semi-arid land, which means that they can be cultivated where many
42 other plants will not flourish, and can be particularly useful in developing countries that are prone to droughts. Other plants, including soya beans, sunflowers and groundnuts, also produce oils. Most of them can be used to fuel diesel engines without further processing, either by the themselves or blended with diesel fuel. But such “peanut power” or “beanzal” is unlikely to do much to replace diesel, because the oils cost much more and are, of course, wanted for food. Some crops contain sugars and starches that can be turned into fuel by fermentation. Sugar cane and sweet sorghum are the main sugar crops, and both grow fast in good conditions when farmed by modern agricultural methods. Fifty tons of sugar cane will grow each year per hectare in Brazil, and yields may go up to as much as 120 tons per hectare per year as they do in Hawaii. Sweet sorghum will produce an annual crop of about 45 tons per hectare. Both their sugars can be directly fermented to produce ethanol, a form of alcohol. Sugar cane will produce about 3,600 litres of alcohol per hectare and sweet sorghum about 3,500. Cassava (mandioca) - a subsistence crop in many developing countries - is the primary starch crop of interest. It has many advantages. It tolerates poor soil and adverse weather conditions much better than the sugar crops mentioned above, and unlike them, it does not need h g h levels of fertiliser or pesticides to give good yields. About ten to twelve tons of cassava are produced per hectare each year - and this can be turned into about 2,160 litres of alcohol. Sometimes even soft, green herbaceous plants may be a viable source of “green energy”. Some tropical and savannah grasses, ideally adapted to their conditions, grow extremely fast. Elephant grass (Napier Grass) is one notable example. Not only land crops but water plants as well, can be farmed for energy. At present, plants like water hyacinth (eichhornia crassipes) and duckweed (species of the genus lemna) are a major environmental problem in many countries because they spread and clog up lakes and waterways. So harvesting them for animal feed or energy would be a happy solution. Seaweeds can be found too. Large scale seaweed farms are already moored near the Japanese and Chinese coats, producing the food that has been a delicacy in the Orient for centuries. The open oceans are much the biggest under-used part of the world’s surface, and so plans for farming them for energy, as well as food, are receiving considerable attention. Experiments on growing kelp (macrocystic) for fuel have reported yields of as much as 90 tons per hectare a year. Ocean energy farms would be strange looking structures, enormous offshore platforms with long spokes and ropes sticking out of them. The kelp would grow on these protuberances, fed by nutrient-rich water specially pumped up from great depths. It would then be harvested, and like the fresh-water energy crops, turned into methane. The Ocean Farm Project, California, foresees that there will be a vast demonstration sea farm, covering some 40,000 hectares, by the end of the century. All energy production presents practical and environmental problems, and “green power”, for all its attractiveness, is no exception. Conditions have to be right for growing energy crops. When they are, the energy farms could absorb resources needed for food production. Energy crops could improverish the soil and destroy important wildlife habi-
‘
43 tats. Some could use up more energy to grow and harvest than they would ever produce, and others could cause pollution and possibly affect the climate. Trees, of course, are much the most important energy crop, and they present the most immediate problems. They are a renewable resource, provided they are managed and conserved properly - but, as the steady advance of the deserts and the retreat of the tropical forests bear witness, this rarely happens. Feasible ways of using wood for energy will vary from to country, and there are many different approaches. Countries that have large existing forests merely need to harvest and manage them on a sustained yield basis for a new use - fuel. Others will need to develop new plantations, and this takes capital, energy and land. Environmental protection is an important part of any successful plan for energy farms because, despite appearances, growing trees is not always benign. The whole process of producing wood for fuel can cause environmental problems, from preparing the sites and planting the trees to managing the forests with fertilisers and pest controls, from cutting the wood and coverting it to energy to burning the fuel and disposing of its wastes. Intensive tree farming might improverish the soil in the long term by removing nutrients and organic matter from it. This process which, it is feared, could be caused by the shorter growing times and more thorough harvesting of modern methods, would, of course, cause productivity to decline as the soil grew poorer. Intensive harvesting will remove many times more nutrients than conventional methods. Soil erosion and leaching can also follow felling, with far-reaching consequences. The result again, is that the soil becomes poorer as it loses nutrients, and productivity eventually declines. Once good, nourishing soil is lost, natural processes can take several thousands of years to replace it. Wildlife habitats can also be changed. Usually there is merely a temporary shift after felling when species that like open areas replace their neighbours from the forest until new planted trees grow up. The effects are much more severe when the habitats of endangered species are totally eliminated, so that they can never return. Whatever the crop, whether wood or anything else, it is important t o work out how much net energy it produces. This is because the crops themselves consume energy as they are produced, transported, turned into usable fuel, stored and distributed. Several studies have shown that sugar cane shows the best energy balance sheet in ethanol production, followed by sweet sorghum and cassava. All of them produce more energy than they consume. (It should be noted that the net energy balance for producing ethanol from the same crop will vary from place to place depending on what techniques are used to grow and convert it and how much energy they use). When ethanol is produced from corn, on the other hand, there is a net energy loss, generally because a great deal of energy is used to grow the crop. Besides all the energy needed to turn sugar cane into ethanol can be provided by burning its own by-products, while cobs and stalks provide at most a third of the energy needed to d o the same for corn. Many countries do not view energy production in strict economic terms. They see dependence on other countries for energy supplies and the chronic balance of payment deficits incurred by buying expensive fuel from abroad as major threats to their social and economic development and indeed, to their political stability. So some countries promote domestic energy production even though it costs more than buyling fuel on international markets, as long as the bulk of the money invested is in their own currencies and the pro-
44 jects are expected to produce, a net balance of payments surplus. This climate favours the development of biomass energy programmes, as the Brazilian alcohol production programme shows. In strict microeconomic terms no ethanol should be used as fuel in Brazil today, because it is much more expensive than gasoline. Recently there has been concern that energy crop plantations may absorb agricultural resources that would otherwise be devoted to food production. In Brazil, sugar cane agriculture has expanded to some degree at the expense of food cropland because incentives were given to ethanol production when sugar prices were low (before 1980).The United States faces a rather different situation. Producing ethanol from crops like corn there poses no direct competition to its food supplies, because about 60 per cent of its corn is fed to animals and there is surplus grain; but some argue that it will mop up grain that could otherwise be provided as aid to impoverished third world countries, or that it will affect international trade. Ethanol production can be very polluting. Every cubic metre of the fuel produced from sugar cane generates 12 to 13 cubic metres of effluent into inland or coastal waters, they can cause as much pollution. in terms of biological oxygen demand, as the sewage produced by 6,000 to 6,500 people in one day. The potential impact of this pollution is so great that strict measures to control the discharge of this effluent, called stillage, must be used from the start. Discharging stillage is, anyway, causing waste in more senses than one, for it can be a valuable raw material. It is normally not contaminated by diseasecarrying organisms or toxic compounds, so recovery of mineral and organic substances from it is a potentially attractive undertaking. It is technically feasible to turn it into methane gas to provide energy for the ethanol-producing process or other means, or to convert it into marketable products like fertilisers or feed additives. Water energy farms present a different set of problems. Growing water hyacinth or duckweed on public lakes or streams would be unpopular, because they are thought to be such objectionable weeds. Artificial ponds would be expensive to make, particularly since vast areas of them would be needed, though they would provide new sanctuaries for a great variety of water life and birds. Using salt marshes and coastal lagoons, in direct contrast, would be cheap but could well be environmentally destructive. They are hghly important ecological systems, providing both valuable wildlife refuges and vital stages in food chains on which much ocean life depends. Ocean farms may have other undersirable effects. Pumping huge amounts of cold water, rich in nutrients and supersaturated in dissolved carbon dioxide, to bathe and fertilise the seaweed could change the temperature patterns and chemical balance of the water, the network of life, and the direction of currents both in the immediate area of the farm and in its surroundings. An immediate consequence of this is cooling of the air and increasing cloudiness naar the farm. These changes in turn could cause regional climate changes and even affect the climate of the whole globe. The process will also release large quantities of carbon dioxide which could also influence the climate. Many countries have recently given proper recognition to the importance of developing renewable sources of energy, and both developed and developing nations are giving special attention to harnessing it from biomass. France, for example, gives “green energy” a high priority and is concentrating research on energy crops like coppices, Dona reed, sweet sorghum and water hyacinths. It expects to get fuel equal to 4 million tons of oil
45 a year from biomass by 1990. Energy plantations are being intensively researched in Sweden, and special plantations of willow (salix) and popular (populus) trees are expected to produce 20 tons of dry material per hectare a year. Ireland claims to be a world leader in rapid-harvest forestry, and some 400,000 hectares will be planted there by the year 2000. Alcohol is already being produced in the United States, mainly from corn, and extensive research and development work is being done there on energy farms, especially wood plantations and ocean farm systems. A great deal of activity is also under way in developing countries. Many countries, including Chma, Kenya, the Philippines, the Republic of Korea, the Sudan and Thailand, have reafforestation programmes aimed at producing new firewood as fast as it is used. The Philippines is also one of the countries giving considerable attention to energy farms using fast-growing species. Co-operation between countries over fuelwood plantations is remarkably strong. Other developing countries have recently embarked on alcohol production programmes similar to the Brazilian sugar cane one. Colombia, Cuba, Kenya and the Philippines are among those that have made a start, though the Brazilian programme remains much the largest.
This Page Intentionally Left Blank
47
A DIFFERENT APPRECIATION : WATER QUALITY AND TREATMENT IN DEVELOPING COUNTRIES
G. ALAERTS
Institute of Industrial Cbemistry, Katholieke Universiteit Leuven, de Croylaan 2, Hsverlee, Belgium
ABSTRACT The term “developing” indicates here the process in which a traditional, village-scale technology and its cultural environment evolve into large-scale, sophisticated structures. Three basic differences distinguish the perception of water quality in developing countries from that in developed countries: the notion of waste which has nothing objectionable in a traditional culture because it is fully reused, the mystical connotation of pure water, and the more static conception of life. Developing countries feature three distinct zones on their territory, each with its own level of education, technical facilities and needs. Depending on the zone, water treatment methods must be advanced and efficient, or simple and reliable. In each case, care has to be taken to integrate the facility in the existing social-cultural pattern.
1. INTRODUCTION
Development is a controversial term with different definitions depending on the usage one wants to make of it, and carrying sometimes emotional connotations. It is certainly not suited to describe a complete society however; here it will be used to indicate the process in which a traditional, village-scale technology and its corresponding cultural environment evolve into large-scale and sophisticated structures at a nations level. The notion development is closely linked to modern technical capabilities, but still has to be strongly relativized. The modern brand and level of technology represent only one stage in a longer process. To secure the survival and expansion of previous and contemporary societies, certain other levels of technology and know-how had to be generated before. Although matters like agriculture, personal hygiene, etc. look obvious to us now, they are in fact fairly complicated and marked at the momenr of their development many centuries ago an impressive advance for society. Secondly, even apart of the so-called industrialized countries, many fundamental differences arise between developing countries or even between regions of one such country, as far as the attitude towards technology and environmental quality is concerned. The Third World is far more heterogeneous than the two other Worlds; there are no two countries or provinces with the same historical, cultural or economical background. Hence each country features its own. degree
48
of development, requires its own approach and poses its own characteristic problems and possibilities. Just because of this heterogeneity, solidarity between and mutual interest of developing countries remain, generally speaking, minimal. On the contrary, similar particularities between now industrialized countries (Europe, North America and to a lesser extent Japan) have been smoothed out two to three centuries ago by some largescale political projects, the creation of the civil state and the gradual absorption of the international technological culture. Technology is a culture indeed: it not only specifies the model of the car and TV-set, but it irrevocably promotes a.0. communication and health, and organizes our way of living and, to some extent, of thinking. Presenting a complete discussion of matters related to environment and water quality in the developing countries is by consequence impossible. However, in this contribution it is tried to discern some basic similarities that exist between most developing countries and at the same time deviate from customs in the industrialized countries. For practical reasons only, many examples are taken from one country, Indonesia, but this should not derogate from their general relevance. The perception of water quality will be treated here as a particular case of a broader issue: the functioning of modem technology in a traditional society. 2. CULTURAL DIFFERENCES
A traditional society can show strong positive or negative feelings towards a newly introduced piece of technology. Technology is a foreign object, and expected or suspected to affect the habits of the village and its hierarchical equilibrium; it may render some customs suddenly superfluous (e.g. hauling clean water over a long distance) and create new tasks and opportunities (e.g. a change in farming techniques after the introduction of irrigation). It is often not acknowledged that such novelties are fully interpreted in terms of the local people’s own habits and experience, and not in terms of the scientific knowledge and ideals of those who introduced them. Even in a locality where water is scarce, a new public water supply has certainly not always the same positive meaning as to the people in the capital who planned and financed it. In addition, in many cultures one doesn’t give or receive valuable things from a foreigner just for free. Before, everything in the village was made by and for the villagers themselves. The sudden erection of a water supply, although forming part of the central government’s sanitation programme, disturbs the usual course of things; it will remain a foreign body used with restriction only (in any case: not in the proper way), unless the programme was thoroughly prepared with the villagers beforehand, their real needs accurately assessed and their eventual lack of motivation countered. During the implementation of a regional water supply prpgramme in South Sumatra (Indonesia) government officials of the capital visited a rather isolated village to inquire into their needs. On the question whether clean water was available, a negative answer was given. After a few months a long expensive pipe supplied the village with good quality water from a distant open well. However it turned out that the village actually didn’t need this kind of water (the village owned two deep wells near its border) but irrigation water. The village headman knew that irrigation water could be adduced through a canal system but didn’t realize the same can be done with drinking water; from the beginning on he interpreted the Officials’ proposal the wrong way.
49
Whether a project for the improvement of the environment or hygiene will be successful depends on many variables. A large part of them have only local implications. At the general level one important factor may be added to the ones mentioned before. Traditional communities in fact live fairly isolated from their neighbours as well as from the administrative centre(s) of the country. This not only holds for what are called primitive cultures (tribes) but also for well developed towns offering full housing, complicated farming techniques and local, small-scale industry, and that are accessible by car and via television. Yet each town has its peculiar local hierarchy, organization and habits developed over many centuries (in Indonesia called adat). The resulting rules (laws) and priorities for that specific town may deviate considerably from the prevailing official standards in the country. An ever present government, representative for the country and with continuous authority over the complete territory, is only a modern concept, closely linked to a technocratic state organization. The same particularism is also met in cities: they consist of a ”modern”, cosmopolitan backbone but at least three quarters of its population live in the mosaic of small districts organized like villages (in Indonesia the kampung, consisting of 50 to 200 families). Although so many differing cultures are involved, it is possible to discern three basic attitudes that affect the appreciation of water, hygiene and the environment in general. They are related to the notion of waste, the definition of water, and the static conception of life in traditional societies. I
2.1. The Notion ”Waste”
”Waste” can be defined either the economical or hygienic way. As long as the hygienic connotation doesn’t play a role, like in traditional communities, waste as such is not recognized. It doesn’t exist in terms of something objectionable, but as a valuable product with its specific place in the natural (mainly agricultural) cycles. The consumer of waste enjoys the same social esteem as its producer. Almost every kind of waste is reused in some way or another. This is obvious in the small-scale, traditional rural communities. Organic garbage from the kitchen is used elsewhere, e.g. as cattle-fodder; human and animal excrements are utilized as fertilizer, and the latter in dried form also as firing and as construction material. Excrements can also be discharged directly isfhkponds asis the case e.g. in Central Java (Indonesia). Especially in China, Korea and Japan, night-soil collection is a highly developed technique and socially respected. Real waste, i.e. which cannot be used again somewhere, doesn’t exist. Waste production and its consumption are part of a balanced closed circuit, in which waste is solely considered an economical asset, not some end product. Of course, this is only possible because previously personal hygiene rules have been adapted unconsciously so that potentially hazardous waste like faecal matter, is kept just far enough to eliminate risks. 2.1.1. Unhygienic Waste in the City
The habit of not recognizing waste is taken along by the new inhabitants of the rapidly growing cities, together with other rural customs. This enforces the village-like character
50
of the city’s districts. However, as the traditional consumers of waste are absent, garbage tends to pile up. It forms breeding places for disease carrying animals (mosquitoes, flies, rats, etc.), obstructs roads and clogs drainage canals expected to conceal and convey the garbage. In addition, a city with its higher living standard produces disproportionally larger amounts and new forms of garbage. The waste doesn’t fit anymore into an economical closed loop, although it still keeps its neutral connotation; after some time however it becomes an annoyance and gets recognized as unhygienic. In short, it becomes real ”waste”, something to be carefully discarded. The same reasoning holds for human excrements. In the rural context (low production, high demand) its economical value conceded its hygienically unfavourable characteristics, but in a densily populated city’ this situation is different. Nevertheless, it appears to be very difficult to alter the deep-rooted attitude which considers faecal matter harmless: provided personal hygiene is maintained and the excrements are out of sight (e.g. under the water surface); the invisible relationship bacteria - faeces - water pollution - disease is arduous to demonstrate. This fact may explain why millions of people in developing countries don’t object to live under ostentatively unfavourable conditions. Even if they are aware of the causal relation with high child death rate and disease incidence (for Surabaya: infant (0-1 year old) death rate is app. 50°/00 as compared to 6 to 10 for industrialized countries [2]) the traditional habits and knowledge would hardly incite to search for an explanation and a remedy (see also 2.3). Beside the economical and hygienic definition of waste, a third, superhygienic approach appeared recently in developed countries. Its rationale seems less clear-cut from the scientific point of view and not devoid of emotional fear for any kind of contamination. It is reflected (especially in the U.S.) a.0. in relatively high disinfectant dosage in drinking water supply [3], the rule to chlorinate each wastewater treatment effluent and the extensive measures in private and public toilets to eliminate the last living germs. Apart from the consideration that near-sterilization of the environment might weaken man’s natural immunity system, this approach also hurts on economical and feasability restrictions. Similarly, it has been shown tentatively [4]for the case of some West European countries that waste has been increasingly defined over the last three centuries in terms of growing individual privacy.
2.1.2. Waste Recycling
Whatever its hygienic implications, solid or liquid waste retains its Value as an economical asset. In industrialized countries recycling aims at reducing the cost of the waste treatment, safeguarding the environment as well as responding to the economicalethical call to refrain from squandering the earth’s resources. These motives are more or less integrated in one major government policy. In developing countries on the other hand, the In the City of Surbaya (Indonesia) population density ranges from 47,000 to 180 personslkm’ in various districts, with an overall average of 5,500 [ 11.
51
initiative for the collection and dumping of garbage is often mainly left to the city districts. The system functions at a variable degree of success. The contribution of the municipality is restricted to providing some official dump sites and maybe the transportation of locally collected garbage. But even if the natural balanced cycle between waste production and consumption is disturbed, large fractions of the waste remain economically attractive. Without formal organization, individuals start recuperating and selling them to primarily small-scale industries and workshops. This process is further enhanced by (1) the contrast between the high and low incomes’, the first producing vast quantities of waste containing parts that however little their value, still may provide a subsistence income for the scavenger, and (2) low prices of the salvaged matter when compared to unused raw materials, especially when simple, cheap commodities are to be produced (cans, lamps, belts, tools, etc.). The scavenging, transportation and selling is organized through an elaborate though informal network in which thousands of individuals cooperate, compete, in short: do business. There’s also room for specialization; at one place can buy salvaged tyres and other rubber products, and at other places bottles or cans of specified size. It has been shown [6] for a.0. Cairo, Istambul, New Delhi and Seoul that the waste of the 10 to 20% richest inhabitants provides a living for 1 to 2% of their fellow townsmen. A World Bank mission sent to Addis Abeba to investigate the solid waste problem had to report that no problem existed as everything got recycled [7]. Of course, beside the aristan-like salvaging some materials are collected in an industrial way (scrap metal, glass, etc.). In Surbaya ( 1 9 7 9 , 1,029 ton refuse is produced daily by 2.5 million inhabitants and its industry; 34%is dumped in an organized way, 35% littered or burned on the spot, 6% turned into compost and 25% salvaged [8].
2.1.3. The Objectionable Waste
The salvaging of solid refuse at least partially relieves the local environment, and in fact continues to strip waste of its second connotation as an unhygienic, thus objectionable matter. Yet solid waste sooner or later piles up and cannot be ignored anymore. Liquid waste on the contrary is more easily flushed away and forgotten. Again, the link between waste and unhygienic conditions is not obvious. Rural areas in tropical countries are characterized by low population density, high temperatures and, often, many water streams. The cities themselves constituted far smaller agglomerations as well two to three generations ago. The three factors mentioned guarantee rapid oxidation of the organic matter and dying-off of pathogenicaganisms. Together with the described recycling methods, this contributes to conceal the relation between domestic wastewater and diseases. Obnoxious long-distante effects caused by faecal matter are not recognized (i.e. outside a perimeter of a few metres diameter around the individual). As an example, in Indonesia few Nould care about tens of thousands of people defecating directly into the drainage canals in a densely populated
* In 1980 the official average income of the residents of Surbaya was 37,000 Rp/month (app. 400 Rp/US $) with 11% of the population earning less than 12,000 Rp [ S ] . However, because of lack of control of both very high and very low incomes (street dwellers) it would probably be correct to estimate that 10% of the richest earn more than 100 times the income of the 10%poorest.
52
Fig. 1. It is estimated that 61% of the population of the City of Surabaya makes use of its open drainage canals for the often combined purpose of bathing and defecating. People are convinced that “out of sight is harmless”, yet their personal hygiene standards are strict (note the clean underwear).
Fig. 2. Swimming boys some 50 m downstream of the spot where the previous picture was taken (at the right: a local uncontrolled garbage dump).
city, yet the left hand is considered impure because used to cleanse (with water) a persons posterior; it is very impolite to shake hands or eat with (Fig. 1-2). Personal cleanliness is an earmark of a.0. all South-East Asian countries [ 9 ] . Even today in large cities the self-purification capacity in the open waterways is such that hygienic effects of the domestic wastewater can be ignored if one cares, like in
53 the rural tradition, for his own quarter only. As an example, Surbaya (1976) counts app. 2.5 million inhabitants and 1,500 km of open waterways, of which 80%are shallow drainage canals. The estimated BOD-load is 500 g Oz/s (domestic: 380 g/s); roughly 100 g/s is removed by auto-purification [lo]. 0.2% of the city inhabitants discharge into a sewerage, 37% evacuate into a septic tank (most of them leaching cesspools however), 2% has access to public bathdtoilets and 60% makes daily use of the drainage canals. It is expected that these figures will only slowly improve. Under these conditions water quality is still much better than would be expected. BOD in the major parts of the city rarely exceeds 50 mg 02/1, to attain 300 to 400 mg/l only in the lower reaches of the canals (usually coinciding with poor quarters). Total coli count per 100 ml is 5 . lo5 and lo7 resp. Obviously some contagious diseases are endemic (cholera, typhus, etc.) but confmed to the poor districts; larger outbreaks which may have a shock effect rarely occur. Partly because of the discontinuous social tissue of the city and partly because of lack of education this unfavourable health condition is not fully acknowledged. Another well-studied example of high selfepuration is the Ganges [ 11 1.
2.2. The Mystic Meaning of Clean Water
The fascination for water is a constant in every culture. In traditional societies it is often regarded as a direct gift from God, an interpretation supported by the fact that rain falls down from the sky. In almost every mythology water is one of the elements, together with earth and fire. Water makes crops grow, which is especially essential to e.g. South Asia’s wet agriculture, and nourishes people. Pure water is the sole means to cleanse the body completely. The metaphysical step from physical cleansing to spiritual purgeing is easy to make. Rituals in which water is used as a symbol for the purgeing of the soul are found in every religion (bathing in the Holy Ganges for Indian Hindus, ritual washing for Moslems, baptism for Christians). The discovery of the beneficient qualities of clean water together with the establishment of personal hygiene rules at a certain stage of mads development, must have had a revolutionary meaning. This may explain why those civil prescripts of common sense so often were corroborated by religious laws. In Ancient Java (up to the 16th century) and stiU in Bali, Hinduism is considered the agama tirta, or water religion. Temples were frequently located near wells, and the numerous public bathing places with a semi-religious signification played an important role in everydays life (Fig. 3). The prodigious surfacing of a spring, the perfect clarity of pure water, its organoleptic qualities, this all contributed to the mystification. The low and well-spread organic pollution load in combination with the self-purification capacity of the surface waters, could hardly affect this propitious image until a few decades ago. This attitude still determines the appreciation of water ”which cannot be treacherous’: It is a logical part of a consistent philosophical system in which not a relation between a cause and its consequence is sought, but a link (though not a fully intelligible one) between Nature, Man and their Creator is accepted. Traditionally speaking water may be regarded as the life bringing element, precious and capricious at the same time. On the other hand, a technically educated person doesn’t see Water, but H,O. He analyzes its composition and proposes standards to which the water should comply. Springs and
54
Fig. 3. A still respected semi-religious public bathing place near Bedulu on the island of Bali, Indonesia (Buddhist-Hindu, 13th-15th century).
rain suddenly have their own physical cause and origin. Whereas previously water was to be taken as it came, now it becomes a quantifiable system to be put to use. It is not regarded anymore as pure and purifying per se. but as an object to be understood and managed.
2.3. Other Cultural Factors
On the basis of the preceding consideration one might conclude that an education programme for the improvement of the local environment should mainly consist of courses at an appropriate level to make clear causes and consequences of pollution and to provide remedies. As a matter of fact this approach has proven ineffective, as was experienced by the first one and a half decade of development cooperation, because it tried to change only a small portion of a large complex machinery. Indeed, many other cultural customs are affected too by this kind of development because they are all interrelated and support each other. It should be understood that development intrinsically means changing of the complete existing oulture. Hence each single education project covering a particular subject hurts on the more fundamental concepts in that culture, like those defining mans place with respect to Nature. In the industridzed countries a major philosophical axiom is based on change and a linear concept of history, in which old turns into new (though it is admjtted that
55 Tab. 1. A brief comparison of some important attitudes between traditional (“developing”) and industrialized (“developed”) countries. The descriptions are only very rough indications, ~~~~
~~~~~
Attitude in traditional society
Attitude in industrialized society
Nature is incomprehensible and animate
Nature consists of interrelated phenomena and deterministic processes
Coincidence has always a supernatural reason
Coincidence is a matter of stochastic processes
Thorough study and accuracy are less important
Thorough study and accuracy are very important
Form is very important
Content is very important
Decisions largely based on subjective criteria like status, hierarchical position and social relation, at the expense of efficiency etc.
Decisions usually based on objective criteria like efficiency, capability, scientific approach, at the expense of the human dimension
Social pressure for taking responsibility for the own family or group
Social pressure for taking responsibility for the large society or the nation
”history repeats itself ’). Traditional societies on the other hand would rather consider history as a cyclic happening, creating nothing really new; their principal activity, agriculture, follows the seasonal cycle (see also the reincarnation theory). By consequence their basic attitude towards Nature is definitely static. Man has to adapt himself to the condition of Nature and should not interfere in its course. At its best he can ask favours through veneration and gifts. Even if nominally Moslem, many Indonesians are still guided by animistic or Hindu principles especially when it concerns the direct relationship with the surrounding nature. North American Indians, just like many other cultures, attribute a soul to objects that are considered inanimate in e.g. Christian and Moslem tradition; not only plants but also earth are thought to possess a spirit and hence cannot be owned or bought by man. The existing culture patterns in developing countries rather call for meditation and acceptance; those in the industrialized countries appeal to creative action and change. Obviously, the reference pattern in developing countries is closely related to the rural experience. This experience also determines which kind of social hierarchical organization wiU be appropriate; in most industrialized countries several types of structures are recognized, some very sophisticated (family, state, factory, etc.), but in developing countries the model of the enlarged family and small-scale community is applied in almost every case. Some of the typical differences generated this way are indicated in Table 1. Another reference to the rural experience of developing countries is the very limited time perspective, i.e. the capability or wish to look into the future and prepare for tomorrow. For the case of Java (Indonesia) for instance, it has been“ascertained [12] that the attention of the farmers wadis exclusively oriented towards the present in order to provide enough food, and with a fatalistic attitude regarding future. The priyayi’s (aristocrat) interest on the other hand is primarily confined to the past aiming at the safeguarding of the traditions. This contrasts sharply with the habits of industrialized countries where planning for the future is considered extremely important; present, past and future are explored in a scientific, objective manner. The foregoing factors also determine convictions on health and death, and affect the existing demeanour towards hygiene and environmental quality. In traditional societies
56 the diagnosis of an ailment is based on external appearences; symptoms of the disease are not reduced to natural causes. Illness is the result of the intervention of spirits or magical forces and healing must be attempted by the medicine man through contact with that other world. Death is not a natural event but the consequence of the action of higher powers. One accepts death and is aware of fitting into a certain all-embracing structural pattern [13]. How to explain hygiene to people guided by a maybe rather extreme interpretation of Hinduism (the honestly living poor will reincarnate in a more prosperous person) or Islam (fatalism)? People organizing their life in accordance with their traditional customs are equally convinced of the value and absolute validity of these customs, as are ”developed” people convinced of the validity of their scientific model. Culture acts as a reference grid through which natural phenomena and social behaviour are interpreted. It also governs one’s perception of water quality, his wish for action and, together with economical and technical constraints, the kind of technology appropriate to solve the problems.
3. APPLICABLE TECHNOLOGIES IN WATER TREATMENT
Most developed countries already posses a history as a rather centrally organized nation. Such a nation is not defined by consanguinity or tribal links,but by the objective concept of the sovereign territory; in principle its inhabitants are all equal for the Law. Because this state structure has since long become integrated in its inhabitants daily life, the country looks homogeneous from a technical point of view: education follows everywhere the same standards, the same law is operative over the whole territory, economical activity is well spread, etc. The expectations regarding water quality are essentially the same in each part of the country, and so are the means to build and operate water treatment facilities. A similar technocraticldemocratic experience lacks in developing countries, conferring to them a more heterogeneous appereance. Three areas can be discerned: (1) priority zones, featuring technical approaches of almost the same standard as in developed countries, (2) the much larger zones where an intermediate technology has to be practiced, and ( 3 ) the largest zone where physical constraints and the lack of education of the population prohibit a classical technical approach. 3.1. The Priority Zones
These zones cover part of the capital and one or two other major cities, as well as the larger industrial complexes. Here only technical and economical criteria determine which type of water or wastewater treatment will be required. High-standard technology is usually available in the form of imported technology and goods; nevertheless most developing countries restrict as much as possible the consumption of their foreign currency reserves for this purpose, and simultaneously stimulate local industries and know-how. Therefore, when treatment plants are designed and constructed by foreign companies,theyusually must join a local contractor and utilize locally made machinery, electrical appliances, etc. This often requires special adaptation and not always improves the plants reliability.
57
Highly qualified staff personnel and trained technicians are readily available nowadays; nonetheless, efforts are still necessary to ensure the plants reliability and efficiency without resorting too much to complicated control methods or treatment principles. This is partly due to the fact that, though well trained, staff personnel often lacks an appropriate specialization, or is not allowed to become thoroughly acquainted with only one part of the total plant because of understaffing or because of social habits (see 2.3). Expensive and complex hard-ware and soft-ware control systems are to be avoided if the expected loss of efficiency is only small. They are preferably substituted by simple and unquestionable rules of thumb. This is especially true if their malfunctioning or faulty usage would reduce efficiency severely. The supply of spare parts or a reparation by a hired specialist may proceed slower and less well organized than expected; these priority zones are indeed only isolated islands of technology, still closely linked to the rest of the country. The consequence for water and wastewater treatment is a need for techniques that combine high efficiency and relatively low investment and operation cost, in which the cheap labour plays a favourable role. The level of sophistication will depend on local conditions, but very often slightly outdated but cheaper installations will be preferred (cheaper because of lower royalty rights, or because constructed "second-hand" by a less expensive contractor and consultant). This approach may e.g. result in the purchase of a rapid multi-layer filter operated at only 90% of its optimal capacity because the backwash programme is not fully computerized (due to expected repair problems and a shortage of trained electronicians). In another case an ionexchange unit has to be installed which will be fded with either the new resin X or the classical resin Y , far cheaper but less efficient in its regeneration creating problems of disposal of the regeneration brine. Such a situation would strongly favour resin X in developed countries, but in developing countries resin Y would be preferred because discharge regulations are less tight (and too often not enforced because of pragmatic economic priorities) and the buffer capacity of the environment not yet exhausted. It should be noted however, that industry usually is concentrated on a small fraction of the territory; this adds to the sharp contrast between the rapidly decreasing environmental quality near those industrial and urban zones, and the untouched environment of the remaining 99%of the territory. Because capital is still very scarce, old plants are kept in function much longer than would be the case elsewhere. This policy often concurs with a, though not particularly creative, yet very careful maintenance by the selected operators who are trained to be responsible for only a limited and well defined task.
A remarkable example of stage-wise technical development is provided by the drinking water treatment facility of Surbaya which is composed of three plants. Plant I is of simple though effective classical design; its main body was erected in 1922, supplying from 1935 on 0.35 m3/s [14]. Because personnel was carefully selected and trained it is still in perfect working order. In 1976 its capacity was increased to 1 m3/s by introducing a sinuous open flocculation canal (length 80 m) after the alum addition and rapid-mix, installing tilted PVC-tubes or corrugated asbestoscement plates in the 4 rectangular settlers @till manually cleaned), and increasing the loading and backwash frequency of the 96 rapid sand filters. A tripling of the flow could thus be reached with relatively simple measures and locally available material. Capital investment was low and the additional operational
58 cost marginal (some more energy, and much more but still cheap manpower for cleaning the settlers). Surabaya’s Plant I1 is of typical Degremont design and erected in 1963-1969; it supplies 1 m3/s. A higher level of technology is introduced, maybe slightly too high for being convenient at that moment. It is made up of (apart from large presedimentation tanks like in Plant I and 111) 4 treatment lines each consisting of a circular pre-clarifier (in fact used as a preflocculator because of the high suspended solids load in the form of very fine clay particles), an Accelator or up-flow clarifier after in-line alum addition, and rapid sand filters. As is the case with Plant I, filters are designed to be backwashed at a fned time interval; this was previously set at 12 h but could recently be raised to 24 h. Head loss nor effluent turbidity are monitored, keeping installation and operation confined to the essentials at the expense of fairly high in-plant water losses (10 to 13%). Finally, Plant I11 is a compact direct-filtration unit constructed in 1981 and designed to deliver 1 m3/s as well. The filters have multi-layer beds and are surface- and backwashed; they require the addition of alum and organic polyelectrolyte. Influent turbidity was reduced by 90 to 99% to app. 20 Ntu through the prevision of an alum flocculation in the already existing large presedimentation basins. This plant features typical high-rate technology, which nevertheless was mastered by its operators in a short period. However, here again backwashing takes place at a set frequency (once per 12 h). Between the previous and the new plant a shift from civil to mechanical construction, and from many low quahfied personnel to few hghly qualified, is patent. Also, more parts had or still have to be imported (e.g. the polyelectrolyte). Whether this type of plant is the most appropriate under the given conditions can hardly be predicted. At present quality of the combined effluent is good with turbidity ranging between 0.1 and 0.5 Ntu; at the end of the dry season however, quality deteriorates drastically.
3.2. The Intermediate Zone
This second zone covers smaller cities, some industrial estates and in fact also large parts of the cited major city(ies) (Fig. 4). In contrast to the priority zones, both technology users and the consumers of its products prone to follow from time to time traditional attitudes which are not always compatible with sound management. Nevertheless most kinds of water treatment and, to a lesser extent, wastewater treatment facilities can be supplied, installed and operated, provided their technology is not sophisticated. Electronic control devices, high-rate techniques, high investment and high operation cost (mainly for power, qualified maintenance, imported chemicals, etc.) are less desirable than in the priority zones. The design of the plant should aim at reliability rather than at high efficiency. Although they are often well operated in practice, the number of critical spots in the equipment and the number of operational problems that require a good knowledge of the fundamentals, should be restricted. Standardized and fully detailed instructions are necessary. Partly because the area in and around such cities is less densely populated and because industrialization has not yet reached a hgh concentration, people fail to recognize industrial or domestic wastewater and its consequences. This attitude enhances uncareful water use and indiscriminate wastewater evacuation without any re-use or treatment.
59
Fig. 4. Contrasts in the city. Approximately 30% 0.f Surbaya’s city dwellers receive (mainly) water from their private tap, 40% buy it from vendors (picture) or public hydrants and 30% depend on shallow wells 1141. The vendor system is a socially efficient; system, creating an income for thousands of people. 37% of the wells are positive as for presence of coli bacteria, with a maximum found in a random sample of 3,000/100 ml[14].
Whereas concern for environment quality is steadily growing in the priority zones, here drinking water supply is still the top priority. With the financial and technical restrictions mentioned above, treatment plants consist of the elementary purification steps, designed in a conservative way and sacrificing some versatility and efficiency to operational reliability. In addition design and construction of such plants can often be fully handled by local contractors and consultants. The scheme flocculation with alum/settling/fdtration is most common (Fig. 5). Rapid-mix is carried out by a propeller or turbine-type mixer, or by a waterfall or hydraulic jump (often combined with a flow measuring weir). Flocculation is brought about by the turbulence in a sinuous canal or labyrinth. Although requiring more surface because less deep than a paddled flocculator in order to facilitate cleaning, its efficiency is almost as high and its maintenance far easier (no moving parts); obviously flow variations should be restricted. Settling tanks are usually rectangular and cleaned manually. Tilted plates may improve their performance when required. Filters are single-layer rapid sand filters, usually operated at a d e c h ing rate; by situating the filter bed app. 2 m below the water level @ the settler, simple filters can be operated at a constant rate (2 to 4 m/h) and increasing head. Those plants are most often constructed in reinforced concrete and lined brick work; they are open and cover a large surface. Recently, relatively cheap and uncomplicated package plants have been introduced, delivering some 0.02 m3/s, and mainly consisting of inland manufactured components, possibly under licence. Future diversification and optimalization of these plants will definitely contribute to the increase of technological know-how of the concerned factories as well as to the service of the public. At several reserach institu-
60
Fig. 5 . A not complicated yet very efficient water treatment facility at Bangkalan, Indonesia. The rapid-mix is followed by a flocculation labyrinth, settlers and rapid sand filters (output app. 0.2 m3/h).
tions efforts are made to upgrade en reshape existing techniques in function of a higher reliability combined with low-cost operation [e.g. 15-1 71. The increasing amount of wastewater will further deteriorate the environments quality during the next decade. As most of the industries located in these areas are agro-industries discharges are mainly of organic nature and almost exclusively suited for biological treatment. The high ambient temperature favours lagooning and anaerobic digestion with biogas . recovery. Up to now, one of the major obstacles on the path of development has not been mentioned yet, i.e. the uncontrolled population growth, especially in the urban settlements. This puts a heavy burden on the financial and technical capabilities of the country. In Indonesia e.g. (population growth rate in 1982: 2.1%),even the recent heavy funding of the drinking water supply programme cannot avoid that proportionally the number of people enjoying good water supply in the urban areas recedes temporarily (Fig. 6 ) .
3.3. The Third Zone Isolated Rural Areas
Areas that are not in direct contact with the two previously described zones are essentially rural and very much characterized by the traditional customs and attitudes described above. Capital is not avadable for traditional economy is based on the barter-trade, and has to be injected by the Central or Regional Government. As already shown waste problems are absent or not recognized; an exception is constituted by e.g. timber felling, wood processing or mining. The effort of the Government to raise the drinking water supply in inspired by public health motives only as the local conditions don’t allow for a return o n investment; neither does the facility pay itself off indirectly by improving the local economical activity.
61
~
-
.
80
I
t o t a l u r b a n i r u r a l = 100 %
Lo
m
-E
60 TJ
> ar L
m
3
0 -
40
m
-
U
-
- - 20
m
c 0
C
0 c
m -
20
A u r a l
10
2
a 0
a 0 1950
I
60
70
80
90
Fig. 6. Evolution of water supply capacity in Indonesia [18]. Urban population growth off-sets the important supply increase in absolute terms.
Although the Government may be convinced of the value of a better water supply, the local community very often remains suspicious and must be thoroughly convinced of its benefits. An important culture gap exists between the rural, traditional people and the ”educated” Government officials and engineers; the latter often act in a condescending manner. Misunder standings and by consequence partial failure of the projects are very common. The local community not only has to be trained on how to maintain the system but also on how to use it properly. Faucets at public hydrants are frequently broken down to make the water run freely (often considered a sign of wealth) or to use its components elsewhere. Broken pipes are most often not repaired by the community, which abandons the public taps and resorts to using the water now freely flowing in the gutter . In the Javanese village N. [ 191, situated in a dry region, five large rain water collection tanks were constructed to provide the village with good quality water nearby instead of the water of a distant, muddy river. All technical aspects concerning the maintenance were extensively discussed and rehearsed with a selected group of the inhabitants. After a year the system was still in perfect condition, but the water appeared to be excessively turbid. This was explained by the villagers by claiming that they were so used to the mud flavour in their drinking water that they regularly filed the tanks partially with ordinary soil. In a Central African country a small dam had to be built for river erosion control as well as irrigation and potable water supply. Its finishing was delayed several times due to sabotage acts, probably by people from the surrounding villages. A staff member of the project went to those villages and after careful investigation found out that a sacred bush inhabited by the ancestral spirits was situated in the midst of the area to be flooded by the reservoir. Patient negotiations eventually led the villages sorcerer to ask the spirits permission to move them, in which they agreed. Shortly afterwards they were officially escorted to their new bush, outside the threatened area. The dam’s construction could be finished without further difficulties. Both examples demonstrate that such technological initiatives are considered foreign,
62
not part of the local identity. They are looked upon as cumbersome, magical or as symbols of wealth, rarely as tools to be used and managed. They can only succeed in their aims if they are carefully integrated beforehand into the local cultural tissue with its peculiar convictions, habits and hierarchy. The new facility should be fully incorporated in the already existing daily life. To insure a good functioning each project has to be followed up and the population guided; drop-andgo formulas are ineffective. Obviously such guidance may take many generations before results are visible. The installations are abviously of the simplest but sturdy design aiming at hgh reliability and self-support. Transportation costs in this zone are often prohibitively high, so preferably locally found or produced building material is to be used. Sometimes a separate power supply has to be provided, either in the form of a generator, wind mills, water rams, etc. Very often no treatment at all is provided and the water is pumped from the well, lake or river directly into the distribution system via a central reservoir. In other regions rain water collection tanks form the best solution. If the water is really of too low a quality simple and cheap adaptations can be proposed. A slow sand filter frequently proves to be an effcaceous solution both in terms of the water quality improvement and the social integration of its operational requirements.
4. CONCLUSION
The upgrading of public hygiene and environmental quality in developed countries has become a major concern at present, and runs parallel with increased efforts for a generalized better education. However, awareness of environmental quality requires certain attitudes; similarly, the motivation for improvement presupposes several traditional customs and priorities that allow such change. In most developing countries these attitudes are found to a limited extent only. By consequence, a successful development project should not only include the factual training or the mere explanation on how to run and maintain the facility, but has to aim as well at a gradual shift in some of the community’s attitudes through a prolonged and adapted education. Inevitably its culture and other customs will also be affected; it is mandatory that simultaneously measures are taken to prevent some kind of a loss of identity. The notion development, as it has been used here, is a very relative one. Between the developing countries many individual differences appear; this also holds often to the various ethnic groups in one country. The rate at which development is going on is obviously almost impossible to assess, though impressive progress was and is being made in several countries. Yet it is very difficult to compare for instance Central African and South-East Asian nations. It has to be reminded also that the development of the industrialized countries is an historic process as well which in fact was initiated not too long ago, and has not yet halted its evolution either. Wastewater treatment, contrary to the drinking water supply, is certainly no priority in developing countries. Public sensitivity to uncontrolled discharging is poor and the cost for treatment often prohibitively elevated. In connection with its agreindustries, occurrence of wastewaters with a high organic load is frequent; lagooning and anaerobic digestion seem to be very appropriate. Industrial water treatment and recycling use imported techniques which should also be adapted to the local economical, technical
63
and cultural constraints. However, water treatment is usually confined to drinking water supply which should also be designed and introduced according to existing cultural habits and knowlegde. Each aspect of its operation, maintenance and usage should be compatible with the social-cultural tissue of the locality. Non-technical criteria may become more decisive in design than its mathematical aspects. There is no doubt that progress is made in most parts of the world. However, within each country the evolution doesn’t proceed at the same rate for all areas. It may be feared that the already existing contrast between the urban priority zones and the rural areas will further exacerbate and once overshadow the problems related to the presently existing gap between the North and South hemisphere.
REFERENCES
1 City Monographic Survey, Surabaya Development Planning Board and Airlangga University, 1974 (in Indonesian). 2 Health Services Plan through Public Health Centers, East Java Provincial Health Dept., 1974 (in Indonesian). 3 Committee Report, J. Am. Water Wks.Ass., 1 (1983), 51-56. 4 A. Corbin, Le miasme et la jonquille, Ed. Aubier, Paris, 1982. 5 J. M. Soeroto and I. D. P. Sukarda, Problems of Solid Waste and Environmental Sanitation, Kotamadya Daerah Tingkat I1 Surabaya, Surabaya, 1980, p. 6 (in Indonesian). 6 D. Miller, Forum du Dkvelopment, 6 (1982), 35. 7 Consultants NV De Koninckx, Antwerp. Personal communication (1981). 8 Surabaya Water, Wastewater, Drainage and Solid Wastes, Vol. IV, Program for Handling Solid Wastes, Camp, Dresser & McKee Inc., Surabaya, 1976. 9 E. F. Lowry, J. Am. Water Wks. Ass., 12 (1980), 672-677. 10 McDonald, Sir M. and Partners, Brantas Pollution Study, Directorate General of Water Resources Development, Jakarta, 1976. 11 R. Chopra (Ed.), The State of India’s Environment 1982, Centre for Science and Environment, New Delhi, 1982, p. 23-26. 12 R. M. Hadjiwibowo, The Degree of Fit Between Business and a Given Culture, Regional Seminar on Business and Culture in South and South-East Asia, lakarta, 1976, p. 14. 13 W. Brand, Differential Mortality in the Town of Bandung, in W. F. Wertheim et al. (Ed.), The Indonesian Town, W. Van Hoeve Ltd., The Hague, 1958, p. 264-273. 14 Surabaya Water, Wastewater, Drainage and Solid Waste, Vol. 11, Water Supply Master Plan, Camp. Dresser & McKee Inc., Surabaya, 1976, Ch. 2-4. 15 Small Community Water Supplies, International Reference Centre for Community Water Supply and Sanitation, The Hague, 1981. 16 H. T. Mann and D. Williamson, Water Treatment and Sanitation-Simple Methods for Rural Areas, 2nd rev. ed., Intermediate Technology Publications, London, 1979. 17 J. N. Kardile, Aqua, 1 (1981), 226-229. 18 Perpamsi, Aqua, 6 (1981), 26-29. 19 A. Weka, JurnalTeknik Penyehatan (Journal of Sanitary Engineering), 1 (19+32), 20-23. 20 Buletin IATP. Jawa Timur (Bulletin of the Indonesian Sanitary Engineering Ass., East Java Chapter), 1 (1982), 42.
This Page Intentionally Left Blank
65
CHEMICAL POLLUTION IN KOREA AN EXPERIENCE OF THE DEVELOPING COUNTRIES
CHAESHIK, RHO
Site & Environment Department, Nuclear Safety Center Korea Advanced Energy Research Institute, Soeul,Republic of Korea
ABSTRACT This report summarizes the results of environmental surveys cmried out in connection with chemical pollution in Korea since the latter half of the 1960s. There is substantial air pollution in the larger industrial cities. However, less readily observed but potentially more serious is the deteriorating quality of surface water and soil, and potential hazards to inshore waters. Although pollution in Korea’s main rivers has not yet caused major problems, the definite trend of worsening water quality could lead to this and have a great social and political impact. To ensure lasting improvements in environmental quality as well as maintaining economic growth and well-being, Korean environmental scientists have played a vital role in recommending and imple menting a national environmental programms. It is hoped that the Korean experience of preserving our environment will be widely acknowledged and used as a reference for studying environmental problems, particularly in the developing countries.
1. INTRODUCTION
Korea, as a nation, has been committed to massive industrialization since 1962, and has mobilized its full capacity of both human and national resources for rapid economic growth. Economic development, which achieved a miraculous expansion and modernization through four Five-year Economic Development Plans, has already contributed significantly to the nation-wide environmental deterioration. As a matter of fact, environmental pollution is a function of population and economic growth. Korea, with its high population density is no exception. After the independence in 1945 there were about 25 million people inhabiting the Korean peninsula. The population in the southern half of the country was estimated as 20 million in 1949, and increased to 38 million by 1981 with a mean annual growth rate of 1.6 percent. On the other hand, the gross national product (GNP) rose from 1,220 billion in 1962 to 44,770 billion (equivalent to about $ 62 billion), and the per capita national income
66 increased from about $A0 in 1960 to more than $ 1,636 in 1981, 37 and twenty fold increases over the early 1960s values, respectively. This amounts to an almost 10 percent annual growth rate of the GNF' since the beginning of the Economic Development Plans. This extraordinary economic. growth, despite insufficient development capital and technology as well as an almost complete dependence on imported industries, shows that Korea is a dynamic nation with the potential to become a self-sufficient industrialized country. However, as economic growth accelerates and industries concentrate in large cities, many poor rural dwellers move to the larger cities out of economic necessity. This rapid influx of population naturally increases the demand for water, energy, transportation, and the volume of solid waste. This generates various environmental problems; air pollution related to energy and motor vehicles, and water pollution to sewage and industrial waste. For example, Seoul, the nations capital, the center of political and economic decisionmaking and education, has attracted people from almost every part of Korea. Currently, Seoul contains approximately 20 percent of the Korean population with an average population density of slightly over 13,000 per km2 in its 627 square kilometer area. Seouls population, rose from slightly over 3 million in 1962 to 8.9 million in 1982. In fact, the city has been almost doubling in size every ten years since 1945, and this rapid population growth (with an average annual growth rate of 2.74 percent) has not slowed significantly. Likewise, the number of motor vehicles in Seoul increased from 38,000 in 1969 to 300,000 in 1982, about eight fold increase over the 1969 value. Approximately 28.5 percent of the Gross National Product (45 percent in 1969) and 55.8 percent (20.1 percent in 1976) of total export business was generated in metropolitan Seoul and its surrounding areas in 1982. Nevertheless, Seoul with its rapid growth rate is facing enormous problems, such as public health, transportation, housing, and above all many environmental problems. In addition to environmental deterioration due to urban growth, the impact of industrial development on the Korean living environment has also accelerated at an alarming rate. According to survey made in 1982, Korean industry discharges 2,293,000 m3 of industrial waste water per day and 3,255,000 m 3 of total fuel combustion pollutants per day. Consequently, air pollution (NO,, SO,, and heavy metals) and water pollution are very high in comparison with rural and residential areas away from the industrial bases and urban areas. Frequently, reports are heard which allege that massive fish kills taking place in coastal waters and estuaries are caused by industrial effluents which contain acids, alkali, and other toxic organic and inorganic constituents.
In Korea, major rivers and streams have been rapidly polluted for the last fifteen years, and an adequate, high-quality water supply has been a focus of the national developme nt plan. The increasing trend of heavy metal concentrations in rivers and coastal waters and in the air over industrial areas is also alarming. In this paper, the author will place an emphasis on chemical pollution with special reference to heavy metal pollution in Korea.
67 Tab. 1. Monthly Average Concentrations of Chemical Constituents in the Nakdong River Year Constituent Mercury Arsenic Copper zinc Cadmium Lead ABS
c1-
PO, -P NH, -N NO, -N NO3 -N BOD COD
1978
1979
-
ND
ND < 0.03 0.007-0.011 0.018 -0.029 ND < 0.001 ND < 0.010 0.1 -0.7 13.4-18.2 0.05 2 -0.092 0.46-1.87 0.013-0.017 0.3-1.0 2.3-3.7 7.0-12.1
< 0.1
-
0.005 -0.01 5 0.015-0.028 ND < 0.001 ND < 0.016 0.3-0.9 7.5 -18.1 0.095-0.164 0.77-3.54 0.012-0.027 0.4-0.7 1.9-5.4 6.4-1 1.8
1980 ND < 0.1 0.006-0.009 0.013-0.030 ND < 0.001 ND < 0.001 0.1-0.4 8.5 - 14.6 0.08-0.014 0.84-3.31 0.010-0.036 0.4-0.7 1.1-4.4 5.7- 14.5
Unit: mg per liter, except for p g per liter for mercury
2. CHEMICAL POLLUTION IN KOREA
In Korea, the magnitude of environmental pollution has warranted serious national attention since the latter half of the 1960s. Perhaps, a report entitled ”A Study on Public Nuisance” was the first comprehensive report which dealt’ with not only pollution inventories by sources, but also with levels of water pollution in the Han River as well as with the exhaust gases and noise problems from various types of automobiles [l]. A report entitled ”A Study on the Water Pollution of the Han River in Seoul City”, was also published by Kwon et al. [2] in which the results of an annalysis of the physical parameters (turbidity, temperature, etc.), chemical pollutants (pH, alkalinity, chloride), Biochemical Oxygen Demand and other measures of sewage wastewater were reported based on Standard Methods for the Examination of Water, Sewage and Industrial Waste, APHA, AWWA, AFIWA of 1955. Since Lee [3-41 published his two papers pertaining to water quality of the Han River in 1969 and 1970, respectively, quite a few papers of similar nature have been published. ”A Study on the Pollution of Seoul Han River” [5] and ”Studies on Stream Pollution by Industrial Waste Water” [6], are examples. From 1978 to 1980, a comprehensive Water Quality Survey at the Nakdong River Basin was carried out by the authors laboratory of the Korea Atomic Energy Research Institute. It was undertaken to provide basic information for the preservation of water quality in the Nakdong River from which more than one third of the national water demand is taken. The resultshave been published annually as a series of research reports [7-91 of the Korea Atomic Energy Research Institute and was also published as an integrated summary report [lo]. According to the survey, the upstream water quality is quite good and clean, while an eutrophication was evident in the water downstream due to the stagnant flow of the river itself (Table 1).
68
10
8 6
5 4 3 2
J
s'
iodang
-
loc
2
6
Ok
Z a
.-6
5
+ 0, 2
4 3
a,
u
r
0
2
0
cn
:I0a 6
5 4
3 2
I o-#
Gueui
Lr '3 75
73 75 Year
Fig. 1. Increasing Trend of ABS Concentration in the Han River and its Tributary Streams
Since 1973, the level of Akyl Benzene Sulfonate (ABS) in the Han River has been monitored by several institutes [ 111. The deterioration of mainstream water quality due to ABS appears to become heavier as the river passes through the city, as is clearly shown by the summary of annual mean values in Fig. 1. This figure also revealed that the main drainage network is incomplete, which results in many of the tributary streams within the City of Seoul being heavily polluted. There was an enormous increase in the annual production rate of ABS from 402 tons per annum in 1966 to 31,892 tons per annum 1975, approximately an eighty fold increase. These figures have effectively been utilized in a series of campaigns against its use, and in August 1980 the Government put a ban on the production and use of ABS. Lee et al. have undertaken a series of studies on the biological conversion and removal of environmental pollutants. They surveyed the residue levels of organochlorine in-
69 Tab. 2. Heavy Metal Contents in Food Stuffs in Korea (ppm) Heavy Metal Sample
Year
AS^
1968
10.20
Rice
1970
0.20-0.34
1971
1973
t 0.09 (0.17) 0.1 3-0.20 (0.25) 0.15 -0.40
1968
t 0.02
1972
Chinese Cabbage
1970 1972 1973 1974
0.12-0.40 (0.05) t 0.12 (0.01) t 0.06 (0.01) t 0.03
Hgb (0.23) 0.14-0.36 (0.14) 0.05 -0.27 (0.05) 0.02-0.08 (0.02) t 0.05 (0.04) 0.02-0.07 (0.01) 0.01-0.02 (0.04) t 0.03 (0.01) t 0.06 (0.03) t 0.07
cu
Pb
Cd
-
-
ND
(1.25) 1.0-1.6 (1.72) 0.89-3.09 (1.45) 0.12- 3.5 0 (0.83) 0.65-1.05
(0.29) 0.18-0.40
ND
-
ND
-
(0.52) 0.3 1-0.60 (0.57) 0.33-0.86 (0.35) t 0.76 (0.05) 0.02-0.14
(0.22) 0.10-0.54 (0.21) 0.09-0.45 (0.41) 0.06-0.91 (0.02) t 0.03 (0.07) t 0.29 (0.12) t 0.20
ND ND -
ND ND ND (0 t 0.01
a the highest natural concentration: 0.24 ppm t <0.005 ppm b Permissible Concentration: 0.05 ppm (New Zealand), 0.1 ppm (Australia)
secticides (i.e., DDT, heptachlor and its epoxide. BHC, endrin, dieldrin and aldrin) in 43 samples of vegetable oils [I21 as the first years project, and obtained the following results: Chemical constituents
Average concentration (ppm)
DDT Heptachlor and its epoxide BHC Endrin Dieldrin Aldrin
0.0279 0.0120 0.0101 0.0084 0.0057 0.0002
In 1978, they studied the fate of BHC residues in rice during various processing steps and the suppressive effect of alginate on the cadmium absorption. The residue rate of BHC originally present in the brown rice was 8%and 20% after polishing, 2.7% and 6.2% after washing, 2.3% and 4.3% after bofing for 100% and 70% polished rice samples, respectively [13]. In 1979 and 1980, Lee et al. further determined the average residue levels of organo: chlorine insecticides for 66 cow milk samples [14] and 80 samples of various meats consumed in Korea [15]. By analizing fat, the results of the former, were 0.196 ppm of a-BHC, 0.159 ppm of 0-BHC, 0.066 ppm of heptachlor epoxide and 0.042 ppm of DDE, respectively, and those of the latter for four organochlorine insecticides were 0.466 ppm
70
... .. b
Y
c
E n
n Y
u
.-0
L
.-c 0 In 3
U .e
(10
2 3
i
Year
Fig. 2. Consumption of mercurial fungiddes and mercury residues in Korean rice. Full circle (0) is the data from whole area and open circle (o), from restricted area
in domestic beef, 0.145 ppm in imported beef, 0.264 ppm in pork, and 0.106 ppm in chicken, respectively. The behavior of BHC was also studied in 1980 by applying the labelled insecticide (i.e., y(U-'%)-BHC) to paddy rice plants in a pot culture, and it was confirmed that 2.8% of 14C was transfered to the straw and grain and of which 9.4% was found in the brown rice. The National Institute of Health, Seoul, has conducted a series of studies on toxic pesticides residues in various food stuffs since 1968 [16-241. In Table 2, heavy metal contents in rice and Chinese cabbage are summarized from those studies. In 1976, mercury and cadmium concentrations were analysed for 112 brown rice samples collected from various production areas in Korea [25] and the results obtained were as follows: 1) Mercury concentrations measured ranged from a low of nondetect-
71
Fig. 3. Contents of Heavy Metals in Fresh-Water Fish and Waters along the Han River
able to a high of 0.310 ppm with a mean of 0.053 ppm for the entire sample, while two groups of samples from the Kimpo area (downstream Han River) showed 5 to 6 times higher than the average values; 2) Cadmium concentrations measured ranged from a low of trace to a high of 0.029 ppm with a mean of 0.021 pprn for the entire sample. In comparison withlevels of mercury residue in rice in the latter half of the 1960s (e.g., 0.131 ppm,range: 0.026-0.261 ppm) [26], it is clear that the level of mercury pollution after a decade became lower by almost one tenth (Fig. 2). This decreasing trend seemed to be solely due to the positive drive of government policy which put a ban on any organomercury fungicides for spraying rice plants in March 1971 [27]. Of the freshwater fish tested, the snakehead contained the highest mercury content of 1.9 f 1.6 ppm (range 0.71-5.5). Carp collected at the Second Han River Bridge station showed a much higher level of arsenic content by a factor of-more than 50 compared with the sample from the Soyang River, Kangwon Province. The levels of cadmium and mercury tend to increase drastically downstream as an industrial complex is located along the river between Seoul and the Han tributary. The mercury content of water at the Soyang River, a tributary of the North Han River was 0.12 f 0.07 p p b , while that of the Second Han River Bridge was 2.7 0.2 ppb [28]. A similar pattern was also shown for cadmium and mercury in fish samples as is clearly illustrated in Fig. 3. The total mercury content of fresh-water fishes collected at 13 sampling points along
*
72 the Han River from upstream (Yang-Gu) to downstream (Kimpo) was recently determined by the quartz tube combustion-gold amalgamation method [29]. The average total mercury content of 140 fresh-water fish was 0.167 0.0054 ppm ranging from a high of 0.3 11 ppm to a low of 0.046 ppm, and none of the samples have had more than 0.5 ppm for mercury, the criteria (body burden) for freshwater quality (aquatic life) proposed by USEPA and USMAS. In 1975, with a survey carried out from January 1972 to December 1974, the Korea Atomic Energy Research Institute (KAERI) showed that marine samples from coastal waters of some littoral industrial cities, contained relatively high concentrations of mercury and cadmium [28]. This report showed that marine fish samples contained 7.2 7.1 ppm (range 2.0-44) of arsenic and 0.31 0.28 ppm (range 0.070-1.5) of mercury; and shellfish had 6.6 3.6 ppm (range 3.3-14) of arsenic and 0.18 0.21 ppm (range 0.016-0.087) of mercury. Similarly, freshwater fish samples contained 0.35 f 0.21 ppm (0.010-0.63) of arsenic, 0.62 f 0.87 ppm (range 0.17-2.9) of mercury, and 0.40 2 0.60 ppm (range 0.23-1.2) of cadmium. Recently, the Korea Ocean Research and Development Institute (KORDI) [30] carried out a study on the heavy metal pollution in the Korean coastal waters using indicator organisms such as mussels and oysters during 1981-1982. The range of heavy metal levels in the soft tissues of mussels was: Cd, 1.8-2.9 ppm; Cr, 2.3-4.4 ppm; Cu, 3.7-6.6 ppm; Fe, 170-790 ppm; Mn, 5.5-21.2 ppm; Pb, 0.65-2.75 ppni and Zn, 50-108 ppni, respectively. The range of biomagnification factors in the soft tissues of mussels and oysters against the dissolved heavy metal concentrations in seawater was; Cd, 2,900-28,600 for mussels, while Cd, 9,300-50,000 and Zn, 3,700-56,900 for oysters, respectively. More recently a study on the heavy metal concentrations in sea sediments of the Ulsan, Busan and Mogpo Bay was made during the Summer (June-Aug.) of 1982 [31]. The highest and lowest mean concentrations of heavy metals were found in those samples taken in Ulsan Bay and Mogpo Bay, respectively, except for the lowest concentration of chromium in Busan Bay. The results showed that sea sediment samples taken in Ulsan Bay contained 0.409 f 0.145 ppm (range 0.197-0.582) of cadmium, 63.692 9.974 ppm (range 56.670-85.542) of chromium, 58.473 f 23.296 ppm (range 33.217-97.583) of copper, 19.532 6,691 ppm (range 12.160-32.973) of lead, 0.244 5 0.190 ppm (range 0.075-0.581) of mercury, and 137.522 6.691 ppm (range 68.555-233.385) of zinc, respectively. Various soil samples from the Seoul area were collected from April 1, 1977 through Dec. 31, 1978, and were analysed for concentrations of heavy metals [32]. The results obtained by the Seoul Metropolitan Government Institute of Public Health are shown in Table 3. It can be said that the pollution levels of various heavy metals in Seoul area sgils are quite high and serious in comparison with normal soils, with a greater copper concentration in the neighborhood of a chemical plant, while higher zinc, cadmium and lead concentrations were found at sites near a metal industrial plant, a rubber industrial plant, and a medical center, respectively. The 1981 data of paddy rice plant soils near several zinc mining plants are also shown in Table 3 for comparison [33]. From October 1977 to January 1978, a sampling was made at 13 sites to determine the level of heavy metals in Seouls air [34]. The average concentrations of each chemical constituent in the air by zone are shown in Table 4. +_
*
*
*
*
*
+_
*
73 Tab. 3. Heavy Metal Concentration in Seoul Area Soil (in ppm) (April 1, 1977 - Dec. 31, 1978) Area Industrial Residential Nos. Sampling Site Heavy Metals 18 Sulfides Cadmium Copper Manganese Chrome Zinc Lead Mercury
2.0-29.97 Trace 12.49 Trace 30,437.9 32.8-394.22 Trace 99.95 4.98-7 870.6 Trace 159.6 0.088-0.967
Commercial
Open space Near Zinc Mining Plants (1981 data)
13
12
3
1.98-10.68 0.5 - 1.60 14.9-59.9 35.9-163.7 Trace 50.84 4.97-63.79 39.84-198.83 0.009-0.254
2.00-11.40 Trace 2.88 25.0-54.8 24.9-219.6 5.0-56.74 4.97-370.50 39.9-279.06 0.04-0.29
2.39-3.97 0.50-0.90 5.0-25.0 19.9- 148.4 2.99-13.0 7.48-29.89 7 9.74- 109.7 2 0.216-0.260
0.013-0.336 4.03-47.02 -
12.20-69.82 5.06-29.36 0.004-0.422
Tab. 4. Summary Table of Heavy Metal Concentration (microgram per cubic meter of air) in the Air, Seoul ( 1977- 1978) Zone Residential Commercial Industrial (11) Industrial (I) Overall Remarks (range)
Total Dust bg/m3)
297.86 253.08 258.37 541.65 337.74 2 times higher than that of in Tokyo in 1969
Lead Cadmium
Chromium Nickel Iron
1.24 1.81 1.50 3.18 1.93 0.05
0.006 0.012 0.009 0.009 0.009 0.001
J
8.16
0.006 0.008 0.007 0.020 0.01 1 0.026
I
I
0.027
0.039
0.029 0.056 0.042 0.104 0.058 0.007
I
0.240
Copper Zinc Manganese
2.18 0.38 2.77 0.43 2.10 0.55 3.83 0.91 2.72 0.57 0.31 0.035
I
5.96
I
2.34
0.227 0.44 0.38 0.66 0.424 0.012
0.023 0.054 0.088 0.102 0.065 0.010
I
S
0.93 0.19
Tab. 5. Mcrcury Residues of the Korean Birds (ppm, wet weight)
M c r c u r y R e s i d u e s in p p m
Locality/Species
Nos’ Breast Muscle of Sample Mean Range
Nakdong River Delta LarusCrassirostris 10 EroliaAlpina 36 E. Alpina 2 Icheon &Kangwha Kyonggido Anas Poecilorhyncha 16 A.Platyrhyncha 2 Phasianus Colehicus 13
1.81 0.24-5.11 3.20 0.48-10.47 -
-
Liver
S t o m a c h ( D i e t ) Wing F e a t h e r
Mean Range
Mean Range
0.71 -
Mean Range
0.36-1.32 3.14 0.32 -
0.19-6.09, 0.88
-
-
-
-
-
-
5.86
4.99-6.74
-
0.34 0.57
0.16-0.44 0.59-0.68
0.92
0.44-1.58 0.83
-
-
-
-
-
-
-
-
-
0.62
-
-
-
-
74 In the meantime, a report o n the birds of the nakdong River Delta was published in 1974 by the Institute of Ornithology, Kyung Hee University, Seoul. This report implied that the birds in the Nakdong River Delta are at least several times more contaminated by mercury residues than those birds in Kyong-Gi Province as for as the breast muscle of the birds is concerned Table 5 [35]. Recently, a studv has been made of changes in the pH values of acid precipitation thatfell in northeastern Seoul for the last five years [36]. The annual average pH of precipitation ranges from 5.5 0.2 in 1977 to 5.0 0.7 in 1981 with a decreasing rate of about 0.1 unit per year, which is greater than the mean annual acidification rates of soft-water lakes in Scandinavia and in the Adirondack Mountains of North America. It was noteworthy that the distribution of the pH in the precipitation is congruent with the distribution of SOz concentrations in the air near ground level.
*
3. EPILOQUE
It is evident and widely recognized that the data and information discussed above has been the basis for the development of appropriate environmental policies to control environmental deterioration in Korea. First, the Korea Development Institute (an umbrella institute of the Economic Planning Board which is responsible for compilation of the national budget) began to take the environmental component into the government budget shortly after the National Conference on Human Environment which was held in 1975. Secondly, late President Park gave an instruction on May 17, 1979 to the Prime Minister that he was to establish a new independent department within the Central Government, which would operate as a National Environmental Management Authority responsible for all aspects of environmental affairs. This instruction was given just one day after he read a headline article in the Dong A Ilbo (a daily newspaper published in Seoul which has the largest circulation in Korea), which discussed the heavy pollution of the Nakdong River, the second largest source for water in Korea. The story was based on the contents of an Annual Report produced by the AuthorS laboratory entitled "Water Quality Survey of the Nakdong River Basins". In view of the magnitude and complexity of the environmental problem in Korea, many scientists have often recommended that responsibility for all aspects of environmental management should be coordinated under a single ministry responsible to the prime minister. This ministry was not formed because it was felt that it might impede the primary goal of the Government which has been to promote economic growth as measured by gain in the Gross National Product (GNP). But, ultimately, the Korean Government decided to provide some subsidies for small and medium industries (needed for pollution control since 1978). Following the endless pressure of academic circles, government assistance was promised because it was believed to be unrealistic to expect small and medium industries t o cope alone with stringent environmental regulations. The introduction of Environmental Rights into the Korean Constitution was also initiated and realized by the Authors KEAF (Korea Environmental Affairs Forum) in collaboration with one of the leaders of the House of Parliament in 1980. In this connection, it is the Authors belief that in the developing countries in particular
75 the academic sector should play a primary role in recommending, drafting, establishing and implementing a national environmental programme as the leaders of non-governmental organizations. It is therefore hoped that the Korean experiences in preserving our environment be widely acknowledged and used as a reference for studying environmental problems, particularly in the developing countries. Finally, the Author would like to emphasize the importance of a technology transfer from the developed to developing countries. In Korea, practical expertise in conducting chemical studies of toxic pollutants was quite limited even in the early 197Os, but it has been reinforced by overseas training. However, undue emphasis upon these aspects, especially to the extent of duplication of research in progress elsewhere, only represented a waste of scarce manpower resources and should have been discouraged. It is, therefore, imperative that selected staff should be seconded for periods of training in practical research and management in environmentally advanced countries. The technologies which have already been established in the developed countries should be transfered through appropriate technical cooperation to the developing countries because the environmental problem is basically an international one and ultimately a global one as well. If the developed countries gain unilateral benefits not only by exporting industrial products and technology but also by exporting technology relevant to the control of environmental deterioration which has occurred in association with the imported industrial products and technology, the developing countries can only exist as less developed countries forever and ever. Furthermore, if the enhancement of our environment is one-sidedly pursued and realized only in the developed countries, the globe, the so called "One and Only Earth", can only be a malformed planet of constantly bright spots and ever dark ones. To increase knowledge of the environment and to protect and improve its quality, research, monitoring and an information exchange must take place and assistance be given to those countries requesting it. Therefore, regional, international and even global cooperation is essential to ensure lasting improvements in our environmental quality and sustainable economic growth and well-being. Promotion of cooperation among the developing countries themselves may be an essential component of this work.
REFERENCES 1 Y. T. Choi et al., A Study on Public Nuisance, prepared by the Korean Industrial Health Association for Ministry of Science & Technology, MOST Report E68-66, 1968, pp. 50. 2 S. P. Kwon et al., Choishin Uihak 11, (1968), 61-70. 3 Y. K . Lee, Yonsei NonChong, 6 (1969), 337-348. 4 Y. K. Lee and K. J. Whang, J. Korean Chemical SOC.,16 (1972), 219-228. 5 D. M. Kim et al., A Study on the Pollution of Seoul Han River, Sponsored by Ministry of Health and Social Affairs, 1972, pp. 36. 6 M. Zong, Studies on Stream Pollution by Industrial Waste Water, Sponsored by National Institute of Health, 1972, pp. 14. 7 S. R. Lee et al., Water Quality Survey of the Nakdong River Basin, KAERI/RR-110/79, 1979, pp. 104. 8 S. R. Lee et al., ibid, KAERI/389/RR-122/80, 1980, pp. 86. 9 S. R. Lee et al., ibid, KAERI/RR-193/80, 1981, pp. 92.
76 10 E. H. Choi and S. R. Lee, Korean J. Environmental Agriculture, 1 (1982), 31 -38. 11 Y. S. Hwang et al., ABS Contamination in the Han River, Report of NIH, Korea, 10 (1973), 285-289. 12 S. R. Lee et al., Studies on the Biological Conversion and Removal of Environmental Pollutants (I), KAERI/114/RR-39/78, 1977, pp. 40. 1 3 S. R. Lee et al., ibid (11), KAERI/242/RR-111/79, 1978, pp. 42. 14 S. R. Lee et al., ibid (110, KAERI/390/RR-l23/80, 1979, pp. 47. 15 S. R. Lee et al., ibid (IV), KAERI/RR-l94/80, 1980, pp. 53. 16 C. B. Ro et al., Study on Toxic Pesticide Residues in Foods (I), MOST Report E68-63, 1968, pp. 19. 17 C. B. Ro et al., ibid (11), Report of NIH, 6 (1969), 237-250. 18 C . B. Ro et al., ibid (110, Report of NIH, 7 (1970), 237-248. 19 R. B. Ro et al., ibid (IV), Report of NIH, 8 (1971), 261-268. 20 I. S. KO et al., ibid, 9 (1972), 389-406. 21 I. S. KO et al., ibid, 10 (1973), 437-453. 22 C. B. Ro et al., ibid, 10 (1973), 257-268. 23 C. B . Ro et al., ibid, 11 (1974), 171-180. 24 C. Song et al., ibid, 11 ( 1 9 7 9 , 141-152. 25 J. S.Young,S. R. Lee and C. S. Rho, Korean J . Food Science and Technology, 11 (1979), 176-181. 26 M. K. Kim, K . D. Woo, and S. S. Han, A Study on Mercury Residues in Polished Rice Grain, Research Reports of the Office of Rural Development 12, 1969, 55-61. 27 C. S. Rho, Current Status of Environmental Pollution in Korea, Proceedings of the 6th Integrated Academic Meeting of Korean Scientists and Engineers, Seoul Science Park, July 26 -Aug. 6, 1976, p. 27-51. 28 C. Lee et al., A Survey of Some Trace Elements in Biological Samples by Neutron Activation Analysis, KAERI Technical Report 2, 1975, pp. 79. 29 D. H. Sohn and H. Y. Jung, J . Kor. Environ. Preservation Assoc., 2 (1981,, 86-96. 30 K. W. Lee et al., A Preliminary Study on the Korean Musselwatch, KORDI Report, 1982, 1-17. 31 H. J. Yun, Study on the Heavy Metal in the Sea Sediments of the Ulsan, Busan and Mogpo Bay, 1982, pp. 1-26. 32 The City of Seoul Government, Survey on the Soil Pollution of Seoul Area, 1978, pp. 28. 33 Office of Environment, Environment Preservation, White Paper, 1982, p. 102. 34 C, W. Cha et al., A Study on the Status of Air Pollution (Heavy Metals) in Seoul, 1978, pp. 46. 35 Institute of Ornithology, Report on the Birds of the Nakdong River Delta, the Southeastern Tip of the Korean Peninsula, Institute of Ornithology, Kyung Hee Univ., Seoul Korea, 1974, p. 133. 36 C. S. Rho, J . Kor. Environ.Preservation Assoc., 3, 1982, 1-12.
CHAPTER I1
POLLUTANTS IN THE ENVIRONMENT: IMPACT AND CONTROL
This Page Intentionally Left Blank
79
ECOLOGICAL EFFECTS OF INDUSTRIAL POLLUTANTS AND THEIR EFFECT ON CULTIVATED PLANTS
H. ZIMNY
Department of Environment Protection, Warsaw Agriculture University 02-766 Warsaw, Nowoursynowska 166, Poland
ABSTRACT The ecological effects of air pollution o n environmcnt may be simplified to the degradation of soil, and as well as to reduction of the basic processes of biomas production, contamination of the crop, and cventually total destruction of the ecosystem.
1. INTRODUCTION
Industry gives immense emission of various solid, liquid and gaseous pollutants into the environment. These emissions contaminate the atmosphere, hydrosphere, lithosphere and biosphere. The variety of emitted compounds and their toxicity constitute a particularly great threat on the biosphere. Negative effects of environmental pollution on basic processes in the biosphere and ecosystems are well known phenomenon, confirmed by numerous authors [ 4 , 7 , 9 , 15, 19, 2 4 , 2 8 , 3 0 , 3 1 , 3 3 ] . The main sources of air pollution in Poland are emissions of the power, metallurgical and chemical industries. Annually, industry in Poland emits to the atmosphere about 2.5 million tons of dust and about 4.8 million tons of gaseous pollutants. Dust pollutants are mainly ashes, primarily calcium compounds and metal oxides, gases include mainly sulphur and carbon oxides. Chemical contents of emissions depend on the type of industry and the size of plant. The effects of industrial air pollution on environment may be simplified to the degradation of soil, and as well as reduction of the basic processes of biomass production, contaminations of the crop, and eventually total destruction of ecosystem. The process of industrial pollutants action is a complex function, involving a synergistic effect of a mixture of compounds of various concentrations over a longer period of time. The action of these pollutants, however, disturbs the basic life processes auch as photosynthesis, respiration and transpiration [ 11, 12, 201.
80 2. CONTAMINATION OF THE SOIL
The soil is contaminated also by pollutants contained in precipitation [29, 34, 351. As shown by studies in the Plock petrochemical region, precipitation waters are contaminated there, with the following ions arranged in a declining succession [35].
S(S0,)
> Ca > Na > K > Mg > Zn > Cu > C1
The contamination of rainfall in the vicinity of Warsaw thorough-fares contained the following succession io ions [34]. Ca > S ( S 0 , )
> Na > K > C1> Mg > Mn > Zn >Po
The pH of precipitation waters examined by the above authors varied from 3.5 to 5.0. Tab. 1. Accumulation of certain trace pollutants in top soils oround smelters in Poland (according to Kabata-Pendias, 181) Smelter
Element
cu
Cd cu Pb Zn Cd Pb Zn F
Zn A1
Pollutant in soil lo-:. g-'
0.201.0 200.0- 1200.0 50.0- 1500.0 130.0- 170.0 10.075.0 100.0- 4700.0 200.0-10000.0 1500.0-1 3000.0
Enrichment factor against backround value
01- 05 07- 40 03-100 03- 05 50-335 07-313 05-250 15-130
Dust and gases contained in the atmosphere diminish the pH of precipitation waters to 2-6 and intensify the processes of the acidification or alkalization of soils (in effect the accumulation of a large quantity of dusts). Metal oxides accumulated in soils in the vicinity of smelters (Table 1) become chemically and biologically activated under the influence of an increasing acidity of soils, which in consequence changes soil processes, especially the biological actival of the soil [36]. Studies have shown that contaminated soils contain much lower levels of the dehydrogenase-group enzymes (Fig. 1) and urease (Fig. 2). Toxins contained in the soil penetrate plants and contaminate the tissues, changing the quantity and quality of the crops. The soil is a complex system - its degradation by metal oxides contamination can produce irreversible consequences. Particulary dangerous here are lead, mercury, cadmium, vanadium, strontium, copper, and as concerns gaseous pollutants - fluorine compounds. Detoxification of soils is often totally unfeasilable, especially on a larger scale. Losses in agricultural ecosystems, which are due to this phenomenon, are estimated at billions of dollars.
81
100
I
Fig. 1. Changes in time (April - November 1976) of dehydrogenases activity (in lo-' dm3 H,. 10 g dry wt-') in soils of parks (A) and street lawns (B) at depth 5-10 cm (showed as an average and range of value)
1-
3 x
L
3,O IT
2 I
z
A
M
J
J
A
S
0
N
Months
Fig. 2. Changes in time (April-November 1976) of urease activity (in g N-NH,. 10 g dry wt-') in soils of parks (A) and street lawns (B) at depth 5-10 cm (showed as an average and range of value)
3. EFFECT OF INDUSTRIAL POLLUTANTS ON CULTIVATED PLANTS
The pollution of the atmosphere with sulphur oxides, carbon oxides, fluorine compounds, hydrocarbons and metal oxides distrurb basic life processes in plant organisms. Details of these disturbances are more or less known and include adverse effect on photosynthesis, respiration, transpiration, etc. These problems have been discussed in detail by many researchers [ l l , 12, 18, 191. The disturbances of the basic life processes of plants diminish their yields. The accumulation of toxins in biomass affects the nutritional value of the crop (Table 2). General changes caused by industrial pollutants or combustion gases of the atmosphere can be reduced to qualitative and quantitative changes in the crop as well as the degradation of ecosystems.
82 Tab. 2. Major air pollutant effects on vegetation (according to KabataPendias, [8]) Pollutants capable of:
SO, SO,
0, NOx
Wide dispersion and longdistance transport Dissolving in living tissues Reacting in living B ioaccumulation Reduction photosynthesis Resistance to metabolic detoxification
x x
x
x x
x
X
x X
Pb
Hg Cd
As
x xa
x xa
x
x x
x
x
x
x
x
x
x x
x x
HF
x
x
Xa
X
X
x
a - in organic form
3.1. Effect of Plant Yields
For growth and development cultivated plants require adequate soil and climatic conditions. Disturbances in the chemistry and biology of the soil, as well as in the composition of the atmosphere constitute a factor impeding growth and development of plants. Negative effects of industrial emissions on cultivated plant yields was found by, amongst others [lo, 1 3 , 2 1 , 2 7 , 31,331. Experiments carried out by Tingey et al. and Krause [ 131 have revealed that synergistic action of a mixture of various gases and liquids is more toxic to plants than the action of each compound separately. In the environment there are no conditions for the action of only one compound, one compound can dominate, but even small concentrations of other toxins fill the cup and deform the yield. Examining the effect of arsenic, vanadium and copper [ 131 found that these elements did not cause a decline in crops if applied separately, though their coupled application resulted in a significant drop in yields: tomatoes by 28-39 per cent, radishes (root) by 22-81 per cent, lettuce by 17-24 per cent. Negative effects of sulphur oxides on winter wheat and rye yields were found by Przybylski [21] in the vicinity of a sulphuric acid plant. In turns out that the SOz concentration of 0.86 mg/m3 during the eighthour exposure affects assimilation and reduces the yield. Plant yields in industrialized regions depend on the concentration and type of emissions as well as the cultivated plant species. Contamination of the atmosphere with sulphur oxides caused a drop in grain yields by 30 to 90 per cent, in sugar beets by 20 to 40 per cent, in lettuce and spinach - some 20 per cent [ 101. Studies carried out by the Environment Protection Institute of the Warsaw Agriculture University in the vicinity of the Oil Refinery and Petrochemical Works in Plock have shown that gaseous emissions caused a decline in yields within the 1.5 kilometre radius from the source of pollution (Fig. 3). Wheat yield losses amounted to 40-90 per cent, sugar beets 20-80 per cent, and spring rape 20-60 per cent. Contamination of the atmospheric air with SO2 varied from two to three-fold concentration at the concentration at the control point [33]. Studies carried out by Warteresiewicz [31] showed that in the vicinity of Szopienice, where the concentration of SOz was 2.7 mg/dm2/24 hours, losses in potato yields amounted to some 60 per cent. Yield losses in were also observed in the vicinity of an iron
83
so- 1
J€tL 727374
@ -t
, Years Fig. 3. Relative yields of plants in the vicinity of petrochemical works as depending upon SO, concentration in the atmosphere (Zimny et al. 1980). I. Total yields of wheat; 11. Total yields of spring rape; 111. Total yields of sugar beets; IV.SO, concentration in the atmosphere in g/dm2 4
50 6o
727374
I t
0,6
3P
i 17,O
E krn
Fig. 4. Crops yields versus distance from the source of pollution in the region of Nowa Huta Steel Plant (according to Warteresiewicz 1978). 1. Potato tubers; 2. Sugar beets; 3. Barley; 4. Beans
smelter (Nowa Huta). Average losses during three-year-long studies are presented in Fig. 4. The lowest yields were obtained at a distance of 0.6 kilometer from the'smelter, whereas at a distance of 3 kilometres yield losses were no longer observed. Fluorine compounds as highly toxic, also negatively affect plant yields. Studies carried out in the region of zinc smelter [27] revealed considerable yields losses: potatoes - 60 per cent, beans - 50 per cent, sugar beet - 45 per cent, clover - 40 per cent, barley 36 per cent. The concentration of fluorine compounds in the area under study varied from 1,000 g F/dm2, while in the control area - 6 to 3,000 . g.
84
A negative effect of dusts on plant yields was also observed in the vicinity of a cement plant, particularly as its concerned acidophilous plants. Generally, it mght be stated that both gaseous and dust pollutants demarcate the growth and development of cultivated plants. Cultivated plants, being annual ones with non-lignification, live tissues are total exposed to toxins and are much more frequently toxiced. Gaseous pollutants, especially sulphur compounds and trace elements, can also have a positive effect on plant yields providing they are emitted in small quantities, and there is a shortage of these compounds in the soil. Both cultivated and wild plants show a certain differentiation of reaction to toxic agents. In Poland and in some other European countries sulphur compounds are the main industrial pollutants. Jefree [5] compared the results of numerous authors as concerns of the sensitivity of cultivated plants to SOz. He divided the plants into four groups: very sensitive, sensitive, little sensitive and resistant, while 27 were sensitive and very sensitive to sulphur oxides. The resistant ones included: maize, celery and rape, while very sensitive: chicory, hops, lettuce, ryegrass, radishes, rye, rhubarb, parsnip, turnip and paprika. The problem of plant resistance to toxic agrents, however, remains open. There are no evidence species which can resist large concentrations of monomial compounds or their mixtures. Particulary the synergistic action is a great; apart from this, plants with limited sensitivity to a toxic agent accumulate considerable quantities of this compound and affect the quality of the crop.
3.2. Industrial Pollutant Effect on Crop Quality
The accumulation of toxic compounds in the biomass of cultivated plants threatens life processes of plants and the health of a consumer, lowers and quite often discredits the nutritional value of the crop. Althought it is mixture of various dusts and gaseous compounds which act in the environment.
3.3. Contamination of Plants with Sulphur Compounds
Industrial and municipal sources give emission of mainly sulphur oxides to the atmosphere. Sulphur oxides are the most frequent gaseous pollutants in Poland. They constitute 60-80 per cent of all gaseous pollutants emitted to the atmosphere. Sulphur penetrate plant tissues and through roots from the soil. The sulphur content in the biomass varies and depends on the concentration of sulphur in the air. the duration of exposure, as well as plant species. A considerable differentation exists depending on plant organs [31,33]. It is quite important for the chemical composition of plants whether the accumulated sulphur is in the mineral or organic form. The reserches have not made such a differentiation, and the information given concern usually sulphur in general. As revealed by studies carried out in the region of Plock [33] assimilation organs leaves contained a considerable quantity of sulphur while fruits and roots - rather small quantities (Table 3). The biggest concentration of sulphur was observed in the plants of
85 Tab. 3. Content of sulphur in the biomass of plants cultivated in the region of petrochemical works (in % of dry matter) Plant species Celery Savoy Spring rape Apple tree (Jonathan) Lettuce Spring beans Red beans Winter wheat
leaves roots leaves straw seed leaves fruit leaves Pod leaves roots straw grain
On non-imminent areas
On contaminated areas
0.40 0.17 0.54 0.26 0.64 0.19 0.04 0.27 0.20 0.19 0.04 0.06 0.06
2.30 0.26 1.09 0.89 1.02 0.41 0.06 0.39 0.37 0.30 0.08 0.21 0.17
the Cruciferae family, as well as in the leaves of root celery; medium quantities in the leaves and fruits of apple-trees, lettuce and beans, while small quantities in wheat (both strow and grain). Wheat belongs to those plant species which revealed quite a considerable differentation in the accumulation of sulphur. Studies carried out in the vicinity of zinc and iron smelters and a thermal electric power plants [31] revealed that sulphur content varied from 0.34 to 1.01 per cent in roots and from 0.09 per cent to 0.18 per cent in stalks, whereas in other plants sulphur content, was as follows: alfalfa 0.32 per cent, sunflower 0.80 to 1.57 per cent, buckwheat from 0.28 to 0.83 per cent. Plant cultivation in contaminated atmosphere yields crops of quality not only due to the accumulation of sulphur, but also as a result of a smaller content of fat, fibere, carbohydrates, betha carotene, chlorophyll and vitamins [ 161.
3.4. Plant Contamination with Nitrogen Compounds (NO,)
The source of contamination with nitrogen oxides are industrial processes based on the energy of liquid, solid and gaseous fuels. Nitrogen oxides are emitted in considerably smaller quantities than sulphur oxides. They constitute some 3 per cent of all gases emitted to the atmosphere in Poland. These compounds (NO,) reveal manifold greater toxicity than sulphur oxides, and they are also precursors of nitrozoamines carcinogenic compounds [61. Nitrozoamines contained in the grass of mountain pasture in Austria, Switzerland, FRC, Czechoslovakia, Romania and Poland caused fatal poisoning of cattle [25]. Industrial and transport contamination of the atmospheric air with nitrogen oxides is a cause of the contamination of cultivated plant crops. As revealed by the examination g to 11 g-' /kg-' of of vegetable on sale in Lodz, nitrate content varied from 8 * vegetables. According to this research beets, radishes, lettuce, fennel, spinach, carrots and kohlrabi accumulate considerable quantities of nitrates. Studies carried out in the vicinity of Plock and Warsaw revealed that the content of nitrate nitrogen in the roots of radishes, red beets and cucumber fruits varied depending
86 on the distance from the emission source and plant species. The biggest quantity of nitrate nitrogen was found in radish roots from Plock (0.29% in dry matter), and only residual quantities in the same species from Wilanow (Warsaw). In red beet roots the situation was reversal. Beets from Wilanow contained 0.15 per cent of nitrates, while those from Plock only 0.04 per cent in dry matter. In the case of cucumber fruits the quantities of N-N03 varied between 0.02-0.04 per cent in dry matter [33]. Nitrogen oxides inhibit also the processes of photosynthesis. Nitrogen oxide (NO) acts more violently, while NOz less violently though more effectively. Studies cobdducted by Beunett [3] showed that 0.6 mg kg-' is the threshold value for each oxide, while the synergistic action of NOz considerably intensifies the inhibition of photosynthesis processes. Barley and clover were also found to be especially sensitive to nitrogen oxides. These phenomena should be further investigated.
3.5. Plant Contamination with Fluorine Compounds
The source of fluorine compounds in the atmospheric air are mainly industrial plants, particularly zinc smelters, enamelling plants, glass mills, mineral fertilizer plants as well as steel works, coal processing plants and others. In Poland (1982) some 4 thousand tons of fluorine compounds were emitted to the atmosphere, which constitutes barely 0.1 per cent of the total quantity of emitted gases. In the US fluorine compounds emission amounts to some 200 thousand tons i.e. some 1 per cent of emitted gases. In terms of quantity this is a relatively low level of emission. Fluorine compounds, however, are exceptionally toxic and phytotoxic. As revealed by studies [26, 271 fluorine compounds show a differentiation in the distribution of concentrations vertically. Much hlgher concentrations of this compound were found at the heigt of above 12 metres drom the soil surface than just above ground. This phenomenon creates a particularly unfavourable situation for trees rather cultivated plants. Fluorine compounds easily penetrate plant tissue and accumulate in the biomass. Fluorine content in plants is directly proportional to the content of this compound in the atmosphere (Fig. 5). The biggest amount of fluorine is accumulated by plants in a zone situated at the smallest distance from the emission source [17, 271. The quantitative distribution of fluorine in plant organs is similar to that of other toxins. The biggest quantities are found in the assimilation organs, smaller in stalks, roots and fruits. As revealed by studies carried out in the vicinity of zinc smelters fluorine compound content in the leaves of cultivated vegetables, fruit trees and farm plants varied from 20 mg to 4544 mg * kg-' d.m. The species accumulating the biggest quantities of fluorine compounds include plum-trees, apple-trees, mangolds, alfalfa; medium: red beets, peartrees, maize, hay, cabbage, relatively small: clover [17,27]. Plant contamination with fluorine compounds is a cause of lowering or discrediting the crop, since the biomass poses a threat to the health and life of consumers. It must be stated that consumer plants should not be cultivated in the zone up to three kilometres from the emission source because of the considerable contamination and the accumulation of big quantities of fluorine in the biomass.
87
0,s
1,o
3,O
km
Fig. 5. Fluorine content in maize depending upon the distance and concentration of emissions (according to Szalonek 1978)
Although the standards of certain counties, e.g. EEC, allow the feed containing 50mg * kg-' of fluorine to be used, these concentrations have adverse effects on farm animals and cause losses in animal breeding [27]. Fluorine compounds are particularly dangerous for animals and humans.
3.6. Plant Contamination with Hydrocarbons
The source of the contamination of the atmospheric air with hydrocarbon compounds are petrochemical plants and combustion transport. A particularly dangerous hydrocarbon is 3.4 benzo(a)pirene (BaP), which accumulates in the biomass and poses a numer of threats in the food chain including carcinogenic effect. The BaP content in the biomass, as in the case of other toxins depends, on the distance from the source of emission and plant species. Studies carried out in the vicinity of petrochemical works in Plock revealed that the BaF' content in low emission zones varied from g kg-' of dry matter, while in hydrocarbon-contaminated zones from 50 to 244 * 180 to 730 . g kg-' of dry matter (Table 4), [32]. Much greater quantities of 3.4 benzo(a)pirene were found in green-house vegetables in the area of an oil refinery. g kg-' of dry matter, radish roots The biomass of cauliflower contained 1100 * 830 * g kg-' of dry matter [32].
-
-
Tab. 4. 3.4 benzo(a)pirene content in cultivated plants in the region of petrochemical works in g . kg-' of dry matter On non-imminent Plant species
areas
On areas contaminated with hydrocarbons
50.0 244.0 145.0 158.0 200.0
320.0 735.0 178.0 280.0 312.0
~
Maize Cucumbers Red beets - roots Radishes - roots Olive trees - leaves
88 Tab. 5. Content of heavy metals in the biomass of cultivated plants in the region of a zinc foundry in mg . kg-' d.m. Pl ant s p e c i e s Lettuce Metal Fe B Pb Mn Mo V
sr
cu Zn Ba Cr
non-imminent area 20 20 05 20 01 01 10 10 150 10 01
Carrot roots
Rye
imminent area
non-imminent area
imminent area
non imminent area
imminent area
1760 140 134 1250 04.2 20
10 10 02 20
600 60 65
04 06 01 -
318 90
01
120
02 50 01
300
-
-
06 11 02 16 12 19 -
840 450 72
traces 01
90 02.5 09 10 70
-
10 03 05 -
Red beet crops revealed large quantities of 3.4 benzo(a)pirene from 901 to 1256 g kg-' of dry matter both in the vicinity of Plock and Warsaw [32]. Research projects on the effect of hydrocarbons are not numerous and there is a need for more investigation, particulary concerning all the phenomena in agroecosystems in exposure zones.
3.7. Metal Oxides and the Quality of Crops
Agroecosystems are contaminated with metal oxides from industrial sources (metallurgy and power industry) and from transport. The degree of contamination and its range depends on the emission source. Industrial sources have a wider range of effect while transport at smaller distances. Plants contamination with metals can be large in crops situated at a distance of 2 to 4 kilometres from a source, or just 5 to 30 metres from a road. It also depends on plant species. Studies carried out [ l ] in the vicinity of a zinc smelter confirm that biggest quantities of metals were contained in lettuce, relatively big in carrot roots and small in rye (Table 5). Roszyk's [23] studies in the vicinity of a copper smelter revealed a considerable diversification, particularly of copper and lead, depending on a distance from emission sources (Fig. 6). It was found that the biggest concentration of metals occurreil in the over-ground parts of plants and grain within the 2-kilometres radius from the emission source. Lead is one of the most dangerous metals, especially for consumers. Lead content in plant biomass varies depending on the concentration of emission and plant species, or even plant organ. Leaves are the most exposed organs. Studies carried out in the vicinity of Cracow revealed that within the radius of 60 kilometers lead concentration in plant leaves varied from 1.40 to 109.0 mg . kg-' of dry matter [2]. The smallest content was
89
cu
\
\
\
\
I
2,o
4P
6.0
40
10,O
km
Fig. 6. Content of heavy metals in above ground plant parts in the region of a copper foundry (arccording to Roszyk 1978)
in grass leaves as well as inside leaves of cabbage, whereas the biggest quantities - in lettuce and potato leaves. Transport is also an important source of contamination with lead compounds. Studies carried out in the vicinity of the Lausanne-Geneva hghway revealed that plant contamination with lead concentrated mainly at a distance of 10-30 metres from the road. Lead content in the biomass considerably varied as regards mono- and dicotyledons. Many times bigger quantities of lead were found in clover and alfalfa than in grass [ 2 2 ] . Plant contamination with metals deteriorates their utility value. Animal feeding with contaminated crops causes various disturbances in their development, as well as in the development of humans. Particularly dangerous are compounds of arsenium, lead, copper, cadmium, vanadium, mercury, strontium, nickel, chromium and others. In many countries studies have been recently conducted on determining the admissible metal contents in farm produce, the problem has not been fully solved, however. The problem of admissible concentration of toxic carcinogenic compounds in food for animals and humans is pending an urgent solution. Agriculture does not function in the vacuum, and it also depends on the action of toxic agents emitted by industry and transport. We must reduce the toxication because of the unaccountable cosequences in ecosystems and the human population.
4. CONCLUSIONS
Industrial emissions comprehensively affect the structure and functioning of ecosystems. Ecological effects of pollutants are to be calculated on longterm evaluating the accumulation of toxins and their circulation between the environment and the biotic components. In general, the effect of atmospheric pollution boils down to the reduction of the efficiency of basic bio-processes such as synthesis, metabolism and decomposition.
90 On the basis of the studies carried out by numerous authors it can be concluded that the contamination of the atmosphere causes: 1. The reduction of primary production - crop. Yield reduction varies from a dozen or so to several score per cent. Yield losses depend on the concentration of pollutants, duration of exposure and plant species. 2. Chemical contamination of the crop with various toxins. 3. The introduction of toxins and metabolites in the food chain, which in consequence threatens the heterothropic levels of the agroecosystem, as well as man. 4. Contamination of the soil and disturbances of the basic biochemical processes in soils. Particularly dangerous is the accumulation of contaminations in the soil, since the processes of soil degradation initiated by this agents can turn out to be durable and irreversible.
REFERENCES
1 S. Baran, Symposium on the protection of the urban environment, Agricultural University of Warsaw, 1976, p. 290. 2 J. Curzydlo, Acta Agraria et Silvestria, Vol. XVIII, p. 2 (1979). 3 A. C. Hill,J. A. Bennett, Environ., 1970,4, p. 341. 4 S. Huttunen, Acta Univ. A. Oulu, 1975, vol. 35, Biol., 2. 5 C. E. Jeffree, U.N., E.C.E., Papers presented to the Symp. on the effects of air-borne pollution on vegetation, 1980, Warsaw p. 328. 6 T. Juszkiewicz, Prob. of Hyg. Warsaw, 1978, p. 169. 7 A. Kabata-Pendias, H. Pendias, Inst. Geol. Warsaw, p. 380. 8 A. Kabata-Pendias, U.N., E.C.E., Paper presented to the Symp. on the effects of air-borne pollution on vegetation, Warsaw, p. 134 (1980). 9 W. Knabe, Ambio, 5, p. 213 (1976). 10 M. Kochniarz, M. Kosicka, Bull. of Research Inst. of Environ. Development, Warsaw, 1978, No. 6. 11 N. Kondo, Y. Akiyama, M. Fujiwara and K. Sugahara, Studies on the effects of air pollutants on plants and mechanisms of phytotoxicity, Res. Rep. Natl. Inst. Environ. Stud. No. 11, p. 137 (1980). 12 N. Kondo, J. Marnta, K. Sugahara, Res. Rep. Natl. Inst. Environ. Stud. 1980, No. 1 1 , p. 129. 13 G. H. M. Krause, U.N., E.C.E., Papers presented to the symposium on the effects of air-borne pollution on vegetation, Warsaw, 1980, p. 219. 14 V. Maly, Sci. Agric. Bohemoslovaca, No. 3. 15 T. A. Mansfield, Cambridge University Press, Cambridge, 1976. 16 J. L. Man, Sci. (American Association), Vol. 187, No. 4178 (1975). 17 B. Micihski, Bull. Inst. of Plants Protection, Poznan - Poland, 1971, No. 50. 18 J. B. Mudd, U.N., E.C.E., Papers presented to the symposium on the effects of air-borne pollution on vegetation, Warsaw, 1980, p. 80. 19 J. B. Mudd, T. T. Kozlowski, Responses of plants to air pollution, Academic Press, N. York, 1975, p. 383. 20 T. Oku, K. Shimazaki, K. Sugahara, Res. Rep. Natl. Environ. Stud., 1980, p. 151. 21 Z. Przybylski, Kosmos, ser. A, an. 26, No. 3 (1977). 22 I. P. Quinche, J. Curzydlo, Revue Suisse d’Agiculture, 1972, No. 6. 23 E. Roszyk, Effect of Industrial Emissions on Useful Plants and Environment, PWN, Warsaw, 1978, No. 206, p. 76. 24 G. Scholl, Stanb. 1975, Vol. 35, No. 5. 25 B. Smyk, Aura, No. 9, Cracow (Poland) 1978. 26 I. Szalonek, XIX International Congress of Horticultural Plants, Warsaw, 1974.
91 27 I. Szalonek, Effect of Industrial Emissions on Useful Plants and Environment, Pol. Acad. of Sci., PWN, Warsaw, 1978, No. 206, p. 55. 28 0. C. Taylor, U.N., E.C.E., Papers presented to the symposium on the effects of air-borne pollution on vegetation, Warsaw, 1980, p. 51. 29 Vermeulen, Sci. and Tech., Netherland, 1978, Vol. 12, p. 1017. 30 J. Wolak, U.N., E.C.E., Papers presented to the symposium on the effects of air-borne pollution on vegetation, Warsaw, 1980, p. 224. 31 M. Warteresiewicz, Effect of Industrial Emissions on Useful Plants and Environment, Pol. Acad. of Sci., PWN, Warsaw, 1978, NO. 206, p. 42. 32 H. Zimny, W. Nowakowski, Prob. of Hyg., PWN, Warsaw, 1977, p. 96. 33 H. Zimny, B. Sikora, U.N., E.C.E., Papers presented to the symposium on the effects of air-borne pollution on vegetation, Warsaw, 1980, p. 160. 34 H. Zimny, D. Zukowska-Wieszczek, Proceedings of the IIIrd Inter. Conference Bioindicat. Deteriorisationis Regionis 12-16th September 1977, Libice near Prague, Czechoslovakia, Praha 1980, p. 151. 35 D. Zukowska-Wiesznek, H. Zimny, Plant and soil,The Hague 1980, Vol. 55, p. 3 2 36 D. Zukowska-Wieszczek,Ecol. Pol., Warsaw, 1980, Vol. 28, p. 267.
This Page Intentionally Left Blank
93
MONITORING AND QUALITY ASSURANCE FOR HAZARDOUS WASTE SITE ASSESSMENT R. F. HOLMES and W. J. LACY
US.Environmental Protection Agency 401 M Street, S. W., Washington, D.C. 20460
ABSTRACT
In response to the chemical and legal requirements of the Comprehensive Environmental Response, Compensation and Liability Act of 1980 (Superfund), the Office of Research and Development, U.S. Environmental Protection Agency, has established a monitoring and quality assurance program for the sampling and analysis of chemicals and complex mixtures obtained from hazardous waste sites. This includes the production of documents that will provide information on the special equipment and unique protocols necessary to sample and analyze the complex mixtures obtained by site monitoring. It also provides for laboratory evaluations, contractor capability evaluations, reference samples, intercomparison studies and data audit. The objective being, the production of data of a known quality. In parallel programs, a high hazard materials laboratory has been established to provide quick response analyses to complex samples and to act as a referee laboratory in the intercomparison of State, Federal and industry protocols. Procedures from private industry, State and other Federal Agencies were adapted to design a comprehensive remote sensing program to provide data on abandoned waste sites prior to entry. The use of geophysical monitoring, and remote sensing are being investigated for application to ensure that the population, surface and groundwater are not contaminated by a closed hazardous releases site. The long term program includes establishing quality assurance procedures to be used in enforcement settlement agreements between EPA and private industry, and the continued monitoring of sites covered by enforcement actions.
1. INTRODUCTION
Studies performed by the U S . Enyironmental Protection Agency estimate that more than 100 million metric tons of hazardous waste are generated annually by U.S. chemical industries. Until 1981 these wastes were disposed of with very little control. Open dumping in landfills, quarries, along streams and in fields was common. Not all industries indiscriminately disposed or their wastes, however, what was considered to be “safe” practices in the 1940’s have proven to be less then benificial to the environment. Years later, the surface and ground water surrounding many disposal areas sites have been found to be severely if not irreparably polluted. In 1978 past practices started to haunt the american chemical industry. Love Canal,
94
Fig. 1. 1980 low oblique photograph of Lowe Canal. Niagara River in the foreground. Buffalo airport is at top of the photograph.
the Valley of Drums, Elizabeth, New Jersey became household names. Environmentalists, lawyers and doctors became very concerned over the welfare of the public living in close proximity to these contaminated areas. The first hazardous site to gain national attention was Love Canal. It was used by a chemical company in the 1950’s and 1960’s as a disposal area for liquid and solid chemical residue from industrial processes. The site shown in Figure 1 was closed and capped. Subsequently, houses and a school were built over the site. By 1978, there were numerous reports of leaking chemicals and chemical vapor concentrations in the houses built around and over the disposal area. In 1980, a study was made to determine the amounts and sources of chemical concentrations at the site. A major multi-laboratory program was undertaken by the EPA. The Office of Research and Development laboratory in Research Triangle Park, North Carolina was responsible for determining chemical concentrations of the air inside and outside the residence. The EPA laboratory in Ada, Oklahoma was responsible for determining the effects on groundwater. The EPA laboratory in Las Vegas, Nevada concentrated on soils and biological monitoring. A fourth Laboratory in Cincinnati, Ohio, provided support in the analysis of environmental samples.
95
1. 2.
3.
Figure 2. Love Canal Sampling Sites.
4. Tap water 6. B8sement sump 6. Soil 7 . Groundwater
More than 3000 environmental samples were collected at the homes within the canal area. The multimedia monitoring protocol at the residences is shown in Figure 2 . It soon became apparent that in order to obtain data of known quality that an exhaustive quality assurance program must be instituted. The QA program should cover all phases of the monitoring Brogram: Sampling, analytical analysis, field surveys, remote sensing and site assessment. In 1981 the United States Congress passed a bill which intended to clean up the estimated 20,000 abandoned and uncontrolled sites around the United States. The law, “Comprehensive Environmental Response, Compensation and Liability Act of 1980” called Superfund, was designed to bring to bear the resources of the Federal Government and the States on what has been called “perhaps the most serious environmental problem facing the Nation”. The Environmental Protection Agency (EPA) was designated as the lead Agency to initiate actions to implement the Superfund Program. EPA together with the 14 other Federal Agencies revised the National Contingency Plan which detailed the steps necessary to meet an environmental emergency. EPA also established a program within the Office of Research and Development, Office of Monitoring Systems, to provide an independent quality assurance support function. This office is responsible for evaluating contractor performance and analytical data in order to obtain meaningful analytical data of “known quality”. The objectives include providing less costly, more timely and less perishable data. In addition, the program is designed to improve the Agency understanding of sitespecific problems and improve the capability to interpret and apply monitoring information. The program is comprised of three major areas:
96 1. Provide quality assurance support to analytical methods 2. Provide remote sensing support to the site assessment 3. Provide technical assistance in monitoring air, water and soils plus high hazard chemical analysis laboratory for referee analysis of questionable and complex samples.
2. LEGAL MANDATES
The actions of the EPA in the control and clean-up of hazardous waste problems have been mandated by a number of laws. Section 311 of the Clean Water Act relates to oil spills and releases of hazardous substances into navigable waters of the United States. Beginning in 1972 the EPA responded to an increasing number of spills in all sections of the country. Since 1978 the clean up of hazardous waste sites such as “Valley of the Drums”. Louisville Sewers, Oswego, New York have been conducted under sections of the Clean Water Act. In 1976, the Resource Conservation Recovery Act was passed to insure proper handling and disposal of wastes in the future. The Act gives EPA responsibility to regulate waste from generation to ultimate disposal i.e., cradle to grave. However, the act does not cover old or abandoned dump sites. To fill this gap and to direct the efforts of the Federal Government and the States to abandonded waste sites the Superfund was enacted in Dec. 1981. It provides for a $ 1.6 billion dollar fund that will allow Federal Agencies to clean up abandoned hazardous waste sites. The act provides a revolving fund for clean up first and then collection from responsible parties. If the responsible parties are not known, the clean up resources are still to be provided. It also provides for costs to identify, investigate and take enforcement and abatement actions when the public health is in jeopardy.
3. SUPERFUND OPERATION
In a effort to clarify the Superfund Program the EPA has published numerous definitions and reviews. An official summary of the program follows: “The Superfund is built on the recognition that responses and cleanups must be tailored to the specific needs of each site or each release of hazardous substances. EPA’s strong enforcement effort seeks to ensure that private responsible parties finance cleanup actions when possible. Direct government action, when called for, can take the following forms: 1. Immediate removals, when a prompt response is needed to prevent harm to public health or welfare or the environment. For example, immediate removals maybe ordered to avert fires or explosions, to prevent exposure to acutely toxic substances, or to protect a drinking water supply from contamination. Actions may include the installation of security fencing, the construction of physical barriers to control a discharge, or the removal of hazardous substances off the site. Ordinarily, immediate removals are limited by law to six months and a total cost of $ 1 million. 2. Planned removals, when an expedited, but not necessarily immediate, response is needed. These actions are intended to minimize increases in danger or exposure that
97 would otherwise occur if response were delayed. Planned removals are subject to the same time and cost limits as immediate removals. 3. Remedial actions, which are longer-term and usually more expensive, aimed at permanent remedies. They may be taken only at sites identified as national priorities. EPA published an interim list of 115 national priority sites in October 1981 ;the list will eventually be expanded to include some 400 sites. Specific actions may include the removal of drums containing wastes from the site, the installation of a clay “cap” over the site, the construction of ditches and dikes to control surface water or drains, liners, and grout “curtains” to control groundwater, the provision of an alternate water supply, or the temporary or permanent relocation of residents”.
4. MONITORING PROGRAM
The Office of Research and Development plants for Quality Assurance in Superfund consists of five projects: - Laboratory Evaluation - Standards Repository - Sampling Methods - Remote Sensing - QA Support Laboratory
4.1. Laboratory Evaluation
An independent quality assurance evaluation of the contractors performing the analysis of samples taken at hazardous waste sites was identified as the primary need for Superfund. Through the use of “third party review” it was considered possible to deliver data of “known quality” suitable for enforcement actions as well as for monitoring. The project includes a comprehensive pre-award review of all laboratory capabilities including personnel, methods and equipment. This review is then repeated on a quarterly basis. These reviews include audits of sample analysis reports, performance of blind sample analyses and on-site evaluation of the laboratory operation. Blind performance evaluation samples are used throughout the evaluation process. Field teams submit blind samples with their normal field samples. The results of these analyses are carefully reviewed prior to each on-site review. A major task of the evaluation process is the audit of the analytical data. Presently, appsoximately 15% of the organic analyses and 35% of the inorganic analyses are reviewed. This review includes completeness of each report, spectra matching iuality, review of surrogate matching spikes and tuning criteria used in each analysis. It is planned that the data obtained from the audits will be computerized so that patterns of common errors can be identified and intra and interlaboratory performance can be compared. Table 1, provides an overview of the scale of this program and its growth. The figures illustrate that a 2-4 fold growth is anticipated by 1984. From 1985 on the program is expected to level off. Because of the number of sites and the detail required for clean-up, a reduction in sample load is not expected for many years.
98 Tab. 1. Quality Assurance Audit Program FY'83 Contractors Pre-award Quarterly on-site evaluations Data audit (samples) Preaward performance evaluation samples Quarterly Performance samples Analytical Standards organics
3 4 -8 14-16. 10-1 5,000 40-60 50-60 120
FY'84 12 8-12 24-40 25 -5 0,000 60-100 100-160 170-220
4.2. Standards and Reference Material Repository
The standards and reference material project was established as a means to provide a single, traceable source of known purity standards for the analytical program. Through the use of reference materials it is possible to: 1) provide standards for instrument calibration and analytical quality control 2) provide reference solutions for performance evaluation 3) provide reference materials (soils, sludges etc.) of known composition for intercomparison studies, methods evaluation and analytical quality control. During 1982 approximately 10,000 ampules of analytical reference standards were distributed. 500 ampules were provided to enforcement cases. The program in 1983 purchased 120 compounds on the Superfund hazardous substances list. The compounds are to be analyzed to verify identification and purity. Once verified, the compounds will be distributed to contractors for instrument calibration. The standards will consist of solutions of single components to be used for instrument calibration or solutions containing several chemicals for use as spiking samples to obtain recovery accuracies. The use of the standards and reference materials has become an integral part of the quality assurance program. It is anticipated that the program will expand as the States become more involved in the analysis of sites.
4.3. Sampling Methods
A continuing program of the Office of Research and Development is the transfer of technology and guidance to the operational functions of the Agency and to State and local authorities. A Superfund contribution to this ORD program is a Compendium of Sampling Methods and a manual on the Characterization of Hazardous Sites. Both are being prepared as a means of setting standard approaches for field investigations. The characterization manual will discuss in general terms the many aspects required to determine the environmental problems present at a hazardous waste site. Preliminary assessment, initial inspection, administration, off-site procedures, document control and chain-ofcustody procedures are included. The document on sampling methods discusses in detail, methods of selection and types of samples, establishment of a sampling plan and chain of custody. Methods for sampling
99 of solids, liquids, gases, and radiation are covered. The section on solids includes soils, sludges and sediments. Surface water and ground water are discussed in the liquid section. Sampling of air at hazardous waste sites as a means of determining potential safety problems and for screening the presence of airborne contaminants is included. The sampling document includes ambient monitoring, container head space and soil gases. The final section of the sampling compendium is a basic summary of sampling instruments for radiation. As stated in the document, “radiation monitorhg” should be one of the first tasks performed because of the high risk to human health.
4.4. Remote Sensing
The operating organizations involved in the Superfund program set a very high priority on the need for information on the present condition of sites and the historical development of sites. Only aerial photography can provide the field team with detailed information of a site without risking exposure. In a similar vein, only aerial photography obtained from archives can provide an accurate history of past management practices. Figure 3 illustrates the product provided to field investigators. The Environmental Protection Agency has been involved in the development of remote sensing systems and techniques since 1971. Through the Office of Research and Development, the Agency first investigated the use aerial photography for surveillance and prevention of oil spills. In 1973 the ORD took another step in investigating aerial surveillance during sub-optimum weather. Following the completion of these two programs, the ORD proceeded to propose and organize an operational program to meet the needs of the Spill Prevention Control and Countermeasure Program. In 1975 the first study was made to utilize aerial photography under diverse situations, utilizing various fdm/fdter combinations to study leachates from landfills and waste sites. Concurrently, the ORD prepared a survey of all known waste dump sites in Delaware and New Jersey. In 1976 a major step was taken to provide all EPA field offices with a remote sensing capability. The ORD designed and had fabricated the first Enviro-pod. With the advent of the system each EPA Region could acquire its own imagery and thus have a close working arrangement between the photo-interpreter and the environmental scientist. In 1979 the ORD prepared a document “Aerial Reconnaissance of Hazardous Substances Spills and Spill-Threat Conditions” that demonstrated that aerial reconnaissance procedures currently utilized in monitoring oil production and storage facilities could also be applied to monitoring chemical production and storage facilities. At the same time, the ORD undertook an extensive program to demonstrate overhead remote sensing applications in compliance monitoring of hazardous waste sites. As Love Canal, and the Valley of the Drums, became common terms of EPA lawyers and environmentalists so did photographic interpretation terms such as trend analysis, historical search, vegetation stress and thermal scanning. The f i s t site-specific landfill study was a joint effort between the Environmental Photographic Interpretation Center, (EPIC), a field station of the Environmental Monitoring Systems Laboratory, Las Vegas, Nevada, and the School of Civil and Environmental Engineering, Cornell University. The EPIC provided photo acquisition services and consultation on the application of photography to be analysis of environmental features.
100
Fig. 3. Typical lo-oblique photograph of a hazardous waste processing site. PCB’s leaked out of the holding ponds killed trees on right of photo and then drained into streams flowing to the Hackensak River (Top of Photo)
Cornell undertook the detailed analysis, ground verification and site analysis of thirteen landfills in Central New York. The program was conducted over four seasons so that the effect of surface and subsurface water flow as well as heat, cold and rainfall on the presence of leachate could be determined. Acquisition of the photographic and thermal infrared data took place during daytime hours as well as early morning periods to test the full utility of the sensors. The primary conclusion of the program was that “photography and thermal scanners can be effective tools for detecting leachates and determining the monitoring required to ascertain the environmental problems involved in landfills”. A major advantage of
101 using remote monitoring techniques was found to be the reduction in costs and time required to survey the problem sites. With the use of aerial photography it was possible to determine the presence or absence of possible problem areas, to establish the magnitude of the problem and prepare a tenative monitoring plan prior to entry into the site. One point should be emphasized, the use of remote sensing did not exclude the use of ground sampling and laboratory analysis to determine the true extent of the problems involved. The findings of the report were: a) The best sensor is the aerial camera producing photography at a scale of 1:SOO or larger; b) Cartographic cameras and panoramic cameras are not essential but they provide better resolution characteristics than hand held 35 mm or 70 mm cameras; c) Color, color infrared film and thermal infrared scanners can all the used to determine leachate problems; d) If only one camera system is available color infrared film will provide the most information; e) Multi-spectral scanners can be utilized but the cost involved in data acquisition and computer processing does not make the systems cost-effective. The present remote sensing program of hazardous wastes has evolved into four tasks: 1) emergency response projects, 2) single data analyses, 3) intensive analysis and 4) wide area inventories. In 1983, alone, the ORD completed 98 studies covering 751 sites. Emergency requests, are prepared in response to a spill or to an abandonded site that is recognized as an immanent hazard to the surrounding population. Aerial photos are acquired and the extent and location of spillage, vegetation damage and threats to drainage are annoted and delivered to the request or within 24 hours. Weather, type of coverage and amount of analysis sometimes cause delays of 1 to 5 days. Single date products, are prepared with newly acquired photography or with recently acquired photos acquired from other Federal or State Agencies. Each analysis addresses drainage, topography, management practices, evidence of leachate, vegetation damage and adequacy of containment. As shown in Figure 4, the products include photographs with interpretation and topographic maps with overlays. Intensive studies, are performed when the need for past practices, condition of original drainage before site opening and the location of buried materials are needed. The discussion under trend Analysis provides more details or these products. Wide-area inventories, are provided when problems involving groundwater or surface features such as lakes have been identified. The sources of the pollution are not known, therefore, it is necessary to identify all possible sources and then perform intensive field investigations of selected small areas rather than the entire area. Trend Analysis: during the period 1975 through 1978, the Office of Research and Development investigated the use of “trend analysis” photographic interpretation to study lake shore erosion on the Great Lakes, urban development surrounding cities and the growth and development of dumps and iandfills. Through the analysis of historical photographs it is possible to determine what and when significant actions were taken at a suspected hazardous waste site. Further, it is possible to determine when housing has been constructed over waste sites or, as in the case of Love Canal, what portion of the disposal area has been covered by housing and schools.
102
Fig. 4. Typical end product of a site investigation. Annotations are generally made on an overlay to the photograph.
103
Fig. 5 . 1948 photo, of a hazardous waste site. Creek has been channelized and roads are built.
104
Fig. 6 . 1980 photographic of area shown in Figure 5. Chemical processing facilities have been constructed and disposal areas opened.
105 The location and acquisition of historical aerial photographs is a relatively straight forward proposition since only a few Federal organizations maintain libraries of imagery flown of the United States. Analysis of the archives held by the Department of Agriculture, the U.S. Geological Survey and the National Archives will provide eighty percent of available imagery. If there are gaps in the coverage, investigation of State archives and private contractor holdings will often fill out the historical requirements. As illustrated in Figure 5 and Figure 6 the process then amounts to determining when a site was opened, finding photography prior to that time and delineating the soils, drainage and cultural patterns on overlays and maps. The process is then repeated on imagery acquired at later dates. In a very short period of time, a detailed history of the site development and management can be prepared. By comparing a 1938 analysis, with 1960 or more recent photography an analyst can tell what drainage has been disrupted, what valleys have been filled and where houses or shopping centers are constructed over former sites. The study of waste sites depends on the extensive use ancillary data to prepare detailed studies. Fire insurance maps, soils and geology maps, topographic maps and past land use studies all contribute greatly to final product.
4.5. Quality Assurance Laboratory
Quality assurance of Suspended programs relies heavily on the independent analysis of data. In order to avoid conflict of interest in the evaluation of contractors, ORD established a cooperative agreement with the University of Nevada, Las Vegas to establish an independent analytical laboratory. The functions of the Laboratory include: 1) Prepare samples for pre-award, blinds and intercomparison of contractor labs; 2) Assist with on-site evaluations; 3) Test new analytical protocols general application; 4) Analyze split samples; 5) Act as a referee laboratory; 6 ) Provide quick response methods development; 7) Analyze complex samples in support of compliance and enforcement investigations.
5 . FUTURE ACTIVITIES
The monitoring program to support the Superfund has not been completely developed. Support of enforcement actions, geophysical monitoring and a quality assurance overview of the sampling program for hazardous waste sites is in planning. A summary of each new program follows:
5.1. Enforcement Activities
Not all enforcement cases end up in court. A mechanism that is currently being used at EPA is the settlement agreement. The agreement requires the firm concerned to clean
106 up a hazardous waste site rather than having the Federal Government perform the cleanup and then going to court for reimbursement of the costs. In all cases, Agency overview of the industry clean-up is required. The Office of Research and Development task in this program will be to provide quality assurance and quality control review of clean-up plans and data. The largest and most complicated settlement for EPA was initiated on April 30, 1982. The Agreement covers the Hyde Park hazardous waste site in the town of Niagara New York. The Agreement between EPA, the State of New York and the Hooker Chemical Co., requires Hooker Chemical Company to: - contain, collect, and treat essentially all contaminated ground water migrating from the Hyde Park landfd; - contain, collect, and treat the relatively insoluble chemicals that have migrated and may continue to migrate from the site; - cap all surface soil which contains any detectable concentration of dioxin attributable to the landfill; - either excavate or cap all soil in the Bloody Run drainage basin containing materials which have migrated from the landfill; - implement a stringent health and safety plan to protect the residents and workers during remedial activities; - maintain the remedial technology and monitor the area around the landfill for as long as it is necessary to protect against any endangerment to the public health or the environment; and - take corrective steps-including possible adoption of additional or different technology-of effectiveness tests show that any remedial technology is not accomplishing the result required by the agreement.
These activities are of great interest to the Federal Courts, the State of New York, the EPA and several local communities. The ORD also reviews the clean-up to ensure scientific principles are adheared to and that valid protocols are followed. The task includes: - review and evaluation of sampling experimental designs, methodology, and data interpretation, - review and evaluation of QA/QC plans and analytical techniques used by the fi to monitor remedial activities, - review and evaluation of the f i s internal QA/QC and QA/QC of any subcontractor involved, - conduct of periodic inspections and on-site evaluations of laboratories conducting analyses supporting the action, - analyses of split samples to verify acceptability of analytical data, - review of raw data from analytical laboratories to evaluate the level of QA/QC achieved by the laboratory, - provision of expert withnesses in support of continuing court actions.
It is anticipated that the Hooker Agreement will be the first of many during the next few years.
107 5.2. Geophysical Monitoring
For years the petroleum and mineral exploration industries have been using sub-surface geophysical measurement methods as a tool for exploration. Recently has the EPA and several small firms applyed the methods to hazardous waste site surveys. The methods to be used include simple and complex resistivity, conductivity and ground penetrating radar to identify sub-surface anomalies, map leachate plumes in the vadose zone and locate buried wastes such as barrels. Sub-surface Geophysical Surveys, when used in conjunction with aerial photography will make it possible and it will provide valuable information for planning monitoring wells and other remedial actions which may be necessary.
5.3. Sampling Quality Assurance
A quality assurance program directed at sampling techniques wdl be implemented to compliment the quality assurance for analytical programs. Presently the Superfund Program lacks consistant and comprehensive sampling guidelines and standard operating procedures. The emergency nature of the program requires the establishment of unique standard operating procedures. Therefore, the program will provide guidelines for monitoring design and strategy, sampling methods, sample handling and shippment and documentation. A program will be initiated to training field teams on methods of on-site audits, procedures and protocols. This will be accompained by a review of sampling plans, QA plans and data validation procedures. The sampling QA program will eventually encompass all media. Guidelines and standard operating procedures involving soils will be prepared first. This document will be followed by similar documents on surface and ground water.
6. SUMMARY
This Monitoring and Quality Assurance Program program has two objectives: 1) to provide information to the field teams and the headquarters decision maker so intellgent decisions may be made on the clean-up of hghly dangerous abandonded waste sites and 2) to establish, for enforcement purposes as well as monitoring, that the data obtained through field sampling and laboratory analysis in of a “known quality”. Through the use of standard operating proceedures protocols, intensive on-site reviews and a high percentage data audit, it has been possible to reduce analytic$ costs, expand the number of competent laboratories and improve immersurably the quality and quantity of data required for Superfund. REFERENCES
Bills, H. M . and Holmes, R. F., 1981, Development of Hazardous Waste Sites Monitoring Methods and Characterization, International Conference on Clean Air and Pure Water - Tomorrows Luxuries, Jonkoping, Sweden.
108 Colby, B. N., 1982, Quality Assurance Guidance for the Determination of Organic Parameters in Soil, Sediment and Water Samples, EPA document 600/x-82-030. Ford et al, 1983, Available Sampling Methods, EPA contract #68-03-3050. Gurka D. F., et al, 1982, Analytical and Quality Control Procedures - The Hazardous Waste Sites Contract Program, 3rd National Conf. on Management of Uncontrolled Hazardous Waste Sites. Holmes, R. F., 1981, Hazardous Waste Site Discovery and Characterization, International Conference on Solid Waste, Sludges and Residual Materials: Monitoring Technology and Management, Rome, Italy. 96th Congress, 1980, Comprehensive Environmental Response Compensation and Liability Act of 1980,94 stat 2767,42 USC 9601. U.S. Environmental Protection Agency, 1980, Assessing Environmental Hazards at Love Canal, Office of Research and Development Fact Sheet. U.S. Environmental Protection Agency, 1981, Using Aerial Photography for Locating and Investigating Hazardous Waste Sites, Office of Research Fact Sheet. U.S. Environmental Protection Agency, 1981, Aerial Photography to Support Chemical Exposure Assessments, Office of Research and Development Fact Sheet. U.S. Environmental Protection Agency, 1982, Hooker - Hyde Park Settement Agreement, Environmental News Release. U.S. Environmental Protection Agency, 1982, Environmental Monitoring at Love Canal, Vol I, EPA Document 600/4/8 2ii030a. U.S. Environmental Protection Agency, 1982-83, Operating Plan, Office of Research and Development.
109
CATCHMENT QUALITY CONTROL
J. M. BOWRON and M. L. RICHARDSON
Thames WaterAuthority, Directorate of Scientific Senices, New River Head Laboratories, Rosebery Avenue, London ECIR 4TP, England
ABSTRACT Catchment Quality Control (CQC) is a lowland river management concept for predicting the presence of organic micro-pollutants. Initially, CQC dealt with industrial or point discharges and full details of the procedure are given. More recently, attention has been given to non-point or diffuse discharges such as those of household and allied chemicals. The technique supplements those of analytical chemistry by utilising biodegradation and mutagenicity studies and these are particularly of value where no analytical chemical techniques are available. Epidemiological studies are also briefly outlined. The importance of pharmacokinetics, metabolites and chemical interactions where applicable are also described. Finally, an outline of riskbenefit assessments are given.
1. INTRODUCTION
There is justifiable concern [ 1-31 when water supplies for large industrial and domestic areas are drawn from lowland rivers which may contain substantial quantities of sewage works effluents. During the past decade there have been a number of attempts to identify, by analysis, e.g. [4], the nature and quantity of many organic micropollutants which may enter such rivers from a variety of sources. One of the difficulties associated with this type of study, even when using the most sophisticated analytical techniques such as gas chromatography-mass spectrometry, is that these techniques are only capable of determinig some - 25% of the compounds present. They rely on the compounds being volatile, or being derivatised to volatile compound. Latterly high performance liquid chromatography has enabled a number of other compounds to be determined.Very recently, electrochemical techniques, such as solid state polarography, appear to be entering a revival phase. In view of the difficulties associated with analytical techniques, a novel mangement technique was devised by Thames Water Authority - this is called Catchment Quality Control (CQC) [S-81.
110 CQC is a technique for predicting the nature and quantity of micro-pollutants which may enter the water cycle. Further details of this technique are given below. In order to give some idea of the magnitude of the problem, it is worthwhile considering the number of organic compounds which are likely to enter the aquatic environment. In addition to the very numerous, ca lo6,compounds occurring naturally as products of plants or micro-organisms, many manufactured compounds such as medicines, pesticides, food additives and industrial chemicals now find their way into the river systems. Our present concern is the occurrence and fate of these compounds, xenobiotics, in the living organism. Xenobiotics, previously called foreign compounds, anutrients and environmental chemicals are chemicals not formally functioning as nutrients. These xenobiotics, when polar and non-lipophilic, can be eliminated from animal organisms, but lipophilic compounds tend to remain in the living system unless they are metabolised to more polar compounds and then excreted. There are at least 10,000 chemicals in regular industrial use, in quantities of one tonne or more per year and this figure is enhanced to 100,000 when lesser used chemicals are considered. It has been estimated that between 200 and 1000 new chemical compositions are marketed each year [9]. As a result an increasing range of xenobiotics are reaching the aquatic environment via sewage treatment works, or directly from industrial and agricultural activities, from chimney effluvia and highway run-off [7,8]. Many of these xenobiotics are not removed by water purification processes, in fact, some organic substances which may represent a public health hazard, such as the trihalomethanes, are produced during the process of chlorination. Very little is known about the harmful effects of these substances at concentrations occurring in drinking water. It can be argued that the great emphasis on chlorinated pesticides is partly due to the fact that they are relatively easy to identify and quantify at low levels. However, many other compounds, possibly present at higher concentrations, are more difficult or impossible to analyse and so have not become the subject of suchintense investigation in relation to their possible adverse effects on human health and on flora and fauna. It is in this area that CQC has significant advantages in being able to predict the presence of xenobiotics. It is worth remembering that when these chemicals are present in water at low levels (0.1 pg/l) and are ingested for 70 years at 2 litres/day, adverse health effects may be caused. Some pharmaceuticals would not normally be prescribed to pregnant or nursing mothers because they may have adverse effects on the foetus, neonate, young infants or to those in advanced years. Furthermore, there are a number of minority groups of the population who are deficient in certain enzyme systems, e.g. glucose-6-phosphate dehydrogenase or may suffer allergies to pharmaceuticals or other xenobiotics. Effects on dialysis patients should not be ignored. A number of xenobiotics, especially halogenated compounds, e.g. hexachlorobenzene, hexabromodiphenyl etc., are metabolised with great difficulty by mammals and only very slowly by certain micro-organisms. These compounds are characterised by environmental persistence, very long biological half-lives and progresive accumulation in the fatty tissues of mammals exposed to these compounds. This can lead to toxicosis when such fatty tissue is utilised. Guidelines [9, 101 have been drawn up for drinking water standards and during the last few years it has been recognised that the increasing concern relating to micro-organic pollutants must
111 necessitate a major review of the existing WHO [I 1J and other drinking water standards.
2. CATCHMENT QUALITY CONTROL
The concepts of CQC introduced by this Authority are the first of their kind in this country, and probably in the world. In the TWA area there are particular problems as both the River Lee and River Thames contain substantial quantities of sewage works effluent. The River Lee, under dry weather flow conditions, can contain in excess of 50% of sewage works effluents and the River Thames in abstracted at a number of locations. The basic concept of CQC include a concentration criterion of 0.1 pg/l for each xenobiotic. This was chosen for the following reasons: a) Perusal of worldwide quality criteria did not indicate a limit more stringent at that time. b) The concentration criteria of 0.1 pg/l ignores any possible biodegeneration and is, therefore, additionally stringent. Briefly, in the light of current knowledge, it was concluded that no chemical, unless radioactive, was likely to cause harm in drinking water at a concentration of < 0.1 pg/l. More recent findings [101 suggest that certain polycyclic aromatic hydrocarbons, dioxins (TCDD) etc. may require lower criteria. The general principles of CQC have been reported in the literature [5-8, 12 -151. A flow diagram ilustrating the scheme of CQC operations is shown in the following figure.
CATCHMENT QUALITY CONTROL SCHEME OF PROCEDURES
Are unkmwn
popletarympcund6
on list I
YES
-+NO
1
112
t
113 Tab. 1. Dischargers Flow Data Average daily flow (m3)
Desjgnated (F)
2 2- 10 10- 30 30-100 100-250 >250
Tab. 2. Points given for various industries (examples only) (organic chemical load) (PI 1
2
3
4
Boiler blowdown Dairy washings Dry cleaning Laundering Textile Manufacture Vehicle washings
Anodising Brewing Food waste Leather works Paint Manufacture Photographic processing
Adhesives Bone and fat rendering Farm wastes Fertiliser Manufacture Pjgment Manufacture Plastics Manufacture
Chemical Manufacture Cosmetic Manufacture Dyeing Paper Manufacture Pesticide Manufacture Pharmaceutical Manufacture
The allocation of points in Table I1 is based on experience of the nature of the above and the materials and chemicals used.
Tab. 3. Final classification - T. This is obtained by combining the flow marks (F) Table I with points for organic chemical load (P) Table I1
Innocouous Traces only Moderate concentration Significant concentration High concentrations
Combination marks (F + P)
Classification (TI
0-1
0 1 2 3
2-3 4-5
6-7 8+
4
114 2.1. River Lee Catchment
In the River Lee Catchment initial studies gave 500 proprietary and other compounds which were resolved into 600 organic chemicals which after initial consideration were subsequently reduced to 30 in number. After obtaining further information a short list of 14 compounds was produced and these were studied in greater detail [12-151.
2.2. River Thames Catchment
Consideration of the appreciably larger River Thames is in hand. Over 3000 proprietary compounds have been investigated, to date.
2.3. Diffuse (non-point) Discharges
Currently, attempts are being made to asertain the effects of compounds originating from other sources. Many cheqjcals are used in the household - detergents, disinfectants, cosmetics and toiletries - not forgetting the most active group of substances, pharmaceutical chemicals. Recently consideration has been given to the last mentioned, and a list of all prescribed drugs from National Health Service prescriptions for the year 1976 was used to calculate the quantities of active substances so prescribed and hence deduced to enter the sewer either after ingestion or directly (i.e. excess drugs discharged), A list of 150 ethical pharmaceutical chemicals, used at 1 tonne or more, per annum, was then derived. The 25 most used by weght, together with 14 chemicals similarly used in proprietary medicines, are being considered in greater detail [8, 13, 161. Certain drugs, such as aspirin are known to degrade [17, 181; tetracyclines are likely to be persistent and lead to resistance factor mediation [ 191. Other antibiotics, e.g. penicillins, may induce allergic response, even after partial degradation [20], and hypoglycaemic agents not normally prescribed in pregnancy have been shown to be teratogenic [21]. Our invertigations into the potential long-term level ingestion of these is well in hand.
2.4. Techniques Available to Supplement Analytical Chemistry
Early in this paper, it was stated that analytical chemistry is only possible for some 25% of organic chemicals when present in the pg/l range in water samples. In view of this difficulty consideration has been given to other techniques (q.v. Fig.). Probably the most important of these alternative techniques are biodegradation studies [22, 231 as they are very relevant in predicting the probable fate of compounds in a simulated aquatic environment. One of the criticisms of these tests is that, owing to the current state of the art they can only be undertaken in the mg/l range, whereas most xenobiotics are likely to be present in the aquatic environment in the pg/l range. A co-operative research project with the Water Research Centre is in hand to investigate the
115 technique in the latter range using “ C labelled compounds, which, if successful may be used to investigate the fate of compounds where analytical chemical procedures are likely to be particularly difficult. Another valuable technique is genotoxicological testing. Perhaps the best known is the Ames test [24],using various species of Salmonella typhimurium to indicate base-pair or frame-shift mutations. Sensitivity can be enhanced by a modification called the fluctuation modification [25],and another variation is to use Escherichia coli WP2. Other tests, such as host mediated assay, cell transformation etc. may prove to be of greater value in the future. There are many of these tests and a battery of them will be much cheaper and less emotive than animal tests and in this difficult area provide considerable useful information, When animal data are available they are not necessarily very useful. For example an LD,, value is not really relevant. It is very rare that costly, long-term studies, to ascertain the effects over three generations in more than one species of animal, have been undertaken. If long-term studies have been performed it may be possible to calculate a tentative ‘no effect’ level but this will only be for the species concerned. Epidemiological studies [26] are also important in considering the effects of environmental chemicals. They are obviously more appropriate in the assessment of the environmental causes of cancer then any indirect system, but they are often beset by many confusing variables. Other techniques, such as pharmacokinetic studies can be of value when appropriate data are available. By means of such an approach in has been possible to predict that the known, weak carcinogen, 1,4-dioxaneYwill have no effect to man if present in water to a level of below 100 pg/l [27,281.
2.5. Miscellaneous
In studies of this nature it is also important to consider metabolites [8, 13, 14,15, 16, 271 and chemical interactions [13,15, 291. Additionally, antagonism, synergism [8] and potentation [8, 13, 141 should not be forgotten.
3. RISK/BENEFIT ASSESSMENT
CQC studies are continuously involved with risk assessment. If a significant risk is established with a particular chemical, analytical results should be obtained to confirm its presence. Its presence must then be balanced against the benefits to society. Risk-benefit decisions [30,31]have wide-ranging implications both for industry and civilisation in general and must be carefully considered. Most chemicals studied by CQC are likely to fall between possible extremes, viz. they have some risks and some benefits. If a situation arose where a toxic compound predicted to be present was subsequently confirmed to be present by analysis in river and/or potable water, at a level which could adversely affect human health, action would be necessary. If the substance originated from a point source, the discharge could be kept to an acceptable level by normal consent procedures. However, if the xenobiotic were entering
116 the environment from diffuse sources it would probably be necessary to submit the problem to central Government for appropriate action to be taken.
4. CONCLUDING REMARKS
CQC can assist in assessing some the questions concerning the occurrence, fate and potential risk of xenobiotics in the aquatic environment. It is not the ultimate technique but helps in combination with information derived from other sources, e.g. analysis [4], mutagenicity studies [25], and epidemioligical studies [26] in ‘risk/benefit’ considerations. CQC has particular advantages assessing the wholesomeness of water drawn from lowland rivers containing significant proportions of sewage effluent. Whether the CQC kind of approach should, and could, be adopted in other countries, or in other catchments, depends on the amount of industry, water supply and organisational position applicable in these circumstances. It is likely that in any catchment where rivers draining development areas are abstracted for public water supply, events will dictate that CQC, or some similar process would have to be estabilished sooner or later, if the public are to be given reasonable assurance regarding the acceptability, or otherwise, of the water they drink.
5. ACKNOWLEDGEMENTS
The Authors wish to express considerable appreciation of the continuing assistance of scientists in industry, trade and research associations for readily providing information for the studies.
REFERENCES 1 ‘Cleaning our Environment: a Chemical Perspective’, Amer. Chem. SOC.2nd Ed. Ch 5, p. 88,1978. 2 R. B. Pojasek, ‘Drinking Water Quality Enhancement through Source Protection’. Anr. Arbor Science, Michigan USA 1977. 3 N. J. Nicolson, ‘Water Re-use and Public Supply in the Thames Basin’. 3rd Conference Water Quality and Technology, The Hungarian Hydrological Society, Budapest 9-13 October 1979. 4 M. Fielding, T. M. Gibson, H. A. James, K. McLaughlin and C. P. Steel, WRC TR159 Organic Micropollutants in Drinking Water’, February, 1981. 5 H. Fish and S. Torrance, J. Nat. Wat. Coun. 1977, (15) 15. 6 H. Fish and S. Torrance, Int. Wat. Supply Assn. Kyoto, 1978. 7 L. B. Wood, M. L. Richardson, Chem in Brit. 1978, 14 (10) 491. 8 L. B. Wood, M. L. Richardosn, Prog. Wat. Tech. 1980,12, 1. 9 J. I. Waddington ‘Organic Micropollutants in Water - A WHO Viewpoint’, Institute of Biology, 13 March 1981. 10 International Standards for Drinking Water, 2nd Edition, p. 33, Geneva, WHO 1970; ibid 3rd ed. p. 31, 1971; D. Stofen, Toxicology, 1973, 1, 187; Gosidarstvenny standart Soyuza SSR, izdaniye ofitsialnoye, GOST 2874-73; Drinking Water Standards Washington US Public Health Service, 1962; Drinking Water Standards, Canada 1968, Official Journal of the European Communities 18 Sept. 1975; Drinking Water Standards, Washington: US-EPA 1975.
117 11 W. M. Lewis and J. I. Waddington WHO International Drinking Water Standards Revised - International Symposum ‘Water Supply and Health.’ Noordwijkerhout, Netherlands 27-29. 8. 1980. 12 M. L. Richardson ‘Xenobiochemistry of Some of the Microorganic Contaminants in the River Lee Catchment in Chemical Pathways in the Environment’ FECS Paris 22-26 September, 1980 p. 29. 13 M. L. Richardson, ‘Catchment Quality Control in Trace Organics in Water’ IWES London 21st October, 1980. 14 M. L. Richardson ‘The Probable Fate of Some Xenobiotics in the R. Lee Catchment’ in Proceedings of the Nonvich Symposium Association of Colsulting Scientists ‘The Food Industry and its Effluents’ November, 1980 - consulting. 15 N. J . Nicolson, P. Casapieri and M. L. Richardson ‘Some Organic Micropollutants in the River Lee Cathment’, Water Industry ’81, Brighton June 1981. 16 M. L. Richardson, ‘Catchment Quality Control - Pharmaceutical Chemical Consideractions in Chemical Pathways in the Environment’. FECS Paris 22-26 Sept. 1980, p. 80. 17 I. S. Wilson, J. Proc. Inst. Sew. Purif 1954, 86. 18 H. Gubser, Gas. Wass. Abwass. 1969,49,175. 19 W. 0. K. Grabow, 0. W. Prozesky and L. S. Smith, Water. Res. 1973, 7,1589. 20 J . W. Bridges and others private communications. 21 S. Schoff, J . V. Aranda, L. Stern, J. Peadriatrics, 1970, 77, (3) 457. 22 E. F. King, Biodegradability tests, Notes on Water Research No. 28, WRC Medmenham, August 1981. 23 Assessment of Biodegradabw NWC/D of E. Standing Committee of Analysts HMSO - to be published. 24 B. N. Ames, J. McCann and E. Yamasaki, Mutation Research, 1975, 31, 347. 25 R. Forster and I. Wilson, J. IWES 1981, 35, (3) 259. 26 S. A. A. Beresford ‘Water Re-use and Health in the London Area’ Technical Report TR138, WRC. 1981. 27 P. G. Watanabe, J. C. Ramsey and P. J. Gehring, ‘Pharmacokinetics and Metabolism of Industrial Chemicals’ in ‘Progress in Drug Methabolism’, 1980, Vol. 5, Ed. J. W. Bridges and L. F. Chausseaud, J. Wiley and Son Ltd. 28 W. T. Scott and P. G. Watanabe, Dow Chemicals Inc. USA Private communication. 29 M. L. Richardson, T. Gough and K. Webb, Ecotox. and Environmental Safety, 1980,4, 207. 30 D. J. Brunswick, Chemosphere 1978, (5) 403. 31 P. J. Gehring, P. G. Watanabe and G. E. Blau, Ann. N. Y. Acad. Sci. 1979, 329, 137.
This Page Intentionally Left Blank
119
TRANSPORT PROCESSES OF NAPHTHALENE IN THE AQUATIC ENVIRONMENT F. Y. SALEH, K. L. DICKSON, J. H. RODGERS, jr. Institute of Applied Sciences North Texas State University Denton, T e r n 76203, USA
ABSTRACT A study was conducted on the environmental transport processes of naphthalene in three surface water-sediment systems and in MilliQ water-clay system. Solubility, Henry's law constant (H), volatilization rate constant, photolysis rate constant, and sorptiondesorption coefficients, were determined. Aqueous solubilities ranged from 28.0 to 29.4 mg/L and did not show a significant difference in the different waters at a = 0.5. Experimentally determined H were lower in the natural waters than in MilliQ water possibility due to decrease in surface tension. Estimation of volatilization rates by measuring K:/K? was subject to photolysis side effects. Photolysis rate followed first order kinetics and was enhanced in natural waters by factors from 2 to 7. Data on naphthalene sorption-desorption coefficients were not conclusive. Translation of the laboratory data to field or microcosm conditions may reflect different situations than the apparent differences noted in individual process determined in the laboratory. Under well characterized microcosm conditions, volatilization.is the major transport process of naphthalene. Photolysis and biotransformation, each account for less than 15% of the kinetic process. The natural water matrix retarded volatilization rates and enhanced photolysis rates. The microcosm sediment partition coefficient was lower than values derived from laboratory data.
1. INTRODUCTION
Mathematical fate models are becoming essential tools for prediction of the hazards associated with the introduction of chemicals into the aquatic environment. Environmental models such as Exposure Analysis Modeling Systems (EXAMS) [l], and Simplified Lake and Stream Analysis (SLSA) [2] are frequently used for hazard assessment of new chemicals and for site specific waste load allocation. These models are based on the law of conservation of mass and the kinetic theory. The models usually include transformation and transport mechanisms between air, water, sediment, and biological compartments. Reactions in the water phase include volatilization, oxidation, photolysis, hydrolysis, biotransformation, bioconcentration, and sorption-desorption [3]. The latter two reactions are considered to be rapid and thus are expressed as equilibrium coefficients between sediment and water or biota and water compartments. Input data to all the models include physical, chemical, and biological characteristics of the environment
120
and the chemical to be studied. The quality of the models predictions can be no more accurate than the quality of the input data. This paper presents the results of a study of transport processes of naphthalene, a widely spread contaminant from oil refineries. Table 1 shows the physical and chemical properties of naphthalene, The approach used in this study was to measure the transport processes in three different types of aquatic systems and in a milli-Q water/pure clay system. Data on rate constants and coefficients as well as data on physical characteristics of a two compartment microcosm were used as input parameters in the SLSA model. The magnitude of differences due to the water matrices are evaluated. Tab. 1. Physical and Chemical(a) Properties of Naphthalene Structure Molecular Weight Melting Point Vapor Pressure at 25°C Solubility in Water at 25OC Log Octanol/Water Partion Coefficient Bioconcentration Factor
128.19 85.55 1.5 X at m(b) 31.7 to 34.4 mg/L 3.37 2 12(c)
(a) From Ref [4] (b) From Ref [5] (c) From Ref (61
2. EXPERIMENTAL 2.1. Materials and Methods
Samples. The three water-sediment systems include an eutrophic pond in Texas (Roselawn Cemetery, RC), a dystrophic reservoir in Louisiana (Cross Lake, XL), and an oligotrophic rock quarry in Indiana (Indiana Quarry, IQ). Surface water samples were collected in 5 gallon carboys from each site. All waters were filtered through glass fiber filters (Schleilcher and Achuell = 30) and stored at 4'C. Sediment samples were wet seived to < 60 pm particle size and stored frozen until needed. Materials. High purity (> 99%) naphthalene Baker Analyzed Reagent was used. Pesticide quality 2,2,4-trimethylpentane, hexane, and pentane were used for preparing stock standard solutions and for extraction of naphthalene. High purity water from a Millipore-Milli-Q system was used. Wyoming montmorillonite (bentonite) of particle size < 60 pm was obtained from Atlantic Richfield Company, Dallas, Texas. Equipment. A Hewlett-Packard model 57 10A gas chromatograph with flame ionization detector was used for naphthalene analysis in water and sediment samples. The column used was glass of 180 cmL X 2 mm ID and was packed with 10% SP 2100 on Supelcoport 100/120. Aqueous naphthalene solution was analyzed isothermally at 140'C. Pentane extracted naphthalene was analyzed by temperature programming set at 90'C for 2 minutes men increased to 16OoCat rate of 8'C/minute. The lower detection limits for naphthalene in water and sediment were 5 pg/L and 0.30 mg/kg, respectively. Beck-
121 man model 915 Total Organic Carbon Analyzer was used for measuring organic carbon in water samples. Leco Carbon Analyzer model IR-12 was used for measuring the organic carbon in the sediments. Procedures. Water and sediment characteristics were run according to published Standard Methods [7, 81. Tables 2 and 3 show the general chemical characteristics of the experimental water and sediments. Protocols for measuring solubility, Henry’s law constant, volatilization rate constants, photolysis, and sorption-desorption were essentially those outlined in the Federal Register [9], Smith et al. [lo], and Mackay et al. [ l 11. Details of the procedures used are described elsewhere [12]. Tab. 2. Characteristics of Experimental Waters(a)
pH at 25°C Turbidity (NTU) Specific Conductance p mho at 25°C Acidity as C a m 3 mg/L Alkalinity as CaCO, mg/L Hardness as CaCO, mg/L Orthophosphate as P mg/L Total Phosphate as P mg/L Soluble Organic Carbon mg/L Nitrate as N mg/L Sulfate as SO:- mg/L Chlorophyll a mg/L Plate Counts (cfu/mL)
Roselawn Cemetery
Cross Lake
Indiana Quarry
9.3 5 884
I. 3 67 164 0 37 156 0.01 0.01 8.2 0.37 47 0.3 3.5 x 103
1.8 4 311 3 90 440 0 0 1.9 0.21 42 0.4x 10-3 1.4 x 104
0 370 122 0.3 0.3 7.1 1.33 44 0.5 x 10-3 3.1 x 105
(a) All unfiltered except dissolved organic carbon Tab. 3. Characteristic of Experimental Sediments(a) ~~~
~
~
Roselawn Cemetery pH at 25°C Total Organic Carbon % 1.34 Cation Exchange Capacity (m eq/100 g) Partice Size % Clay % Silt % Sand
~
Cross Lake
8.5
7.6 0.04’)
1.33
~~
Indiana Quarry 6.4 0.03’)
0.55
O.OO@)
43.3
48.0
22.8
60 34 6
I0 26 4
61 38 1
(a) All sediments wet seived to less than 60 pm particle size. (b) Standard deviation of triplicate analyses.
For each water solubility was determined in triplicate plus a procedure blank. Saturated aqueous solutions were prepared. At 48 hour intervals, aliquots of the solution were transferred to a sealed tube and centrifuged at high speed for 30 minutes. The supernatent was carefully withdrawn and extracted in another sealed tube and then analyzed for dissolved naphthalene. Henry’s law constant was determined by two methods. In the first, H was calculated from the experimental aqueous solubility and published data on naphthalene vapor pressure. In the second method, H was directly measured using Mac-
122
kay et al. [ 111 method. This method involves measurement of the compound concentration in the water phase as being stripped isothermally from the solution at known gas flow rates. Also the method of oxygen reaeration [ 131 was used to estimate naphthalene volatilization rate. The method involves measurement- of the ratio of naphthalene volatilization rate constant K$ to oxygen reaeration rate constant K,O under controlled laboratory conditions. The laboratory measured Kt/K: is multiplied by field or microcosm oxygen reaeration coefficient in order to estimate the compound’s rate of volatilization. Photolysis experiments were run at a naphthalene concentration of about 5.0 mg/L. Methyl parathian was used as an actinometer [14]. The test was run by exposing the samples to direct sunlight acccording to the procedure described in the Federal Register [7]. Duplicate samples were prepared for both the light exposed and dark controls. Procedure blanks containing no chemical were also prepared. Temperature and weather conditions were monitored throughout the experiment. The pH was measured at the start and end of the experiment. Sorption-desorption experiments were all run at a sediment concentration of 1 g/L. The percent dry weight of each sediment was determined and the calculated weight of sediment was added to the appropriate volume of naphthalene solution in a sealed container. Extreme precautions were required in order to minimize loss due to volatilization. These included the utilization of completely sealed 500 mL flasks and bottles during sorption, desorption, and centrifugation. Duplicates were run for each concentration level. Equilibration time was 24 hours for both the sorption and desorption. Blanks and controls for the sorption-desorption test included: test water - no chemical, test water and sediment - no chemical, and an aqueous solution of naphthalene at the same concentration level.
3. RESULTS AND DISCUSSION 3.1. Aqueous Solubility
Data on naphthalene aqueous solubility in Milli-Q and experimental waters are shown in Table 4. The solubility ranges from 28 mg/L in XL to 29.6 mg/L in RC water. Statistical analysis of the data showed no significant difference between the results at a = 0.05. Published data [15] on naphthalene aqueous solubility range from 30.3 to 34.4 mg/L. Tab. 4. Aqueous Solubility of Naphthalene at 25 f l ? C mg/L in Experimental Waters Naphthalene Experimental Waters
X
s
MiUiQ Water Roselawn Cemetery Pond Cross Lake Indiana Quarry
28.7 29.4 28.0 28.4
1.93 2.12 2.48 0.80
X s
Mean of 12 measurements Standard Deviation
123 3.2. Henry's Law Constant
Data on H values in different types of water are shown in Tab 5. In milli-Q water there is an excellent agreement between the calculated H (0.513 X lod3 from solubility and vapor pressure) and the experimentally derived value of 0.532 X These values are also in close agreement with the 0.483 X reported by Mackay et al. [ 113. The experimentally determined H values in the natural waters are all less than the H value in Milli-Q water by a factor of about 2. Two factors are expected to influence the experimental H values in natural waters. One is the presence of electrolytes which reduces solubility and thus may result in increase of H. The second is the presence of natural organics which reduces the surface tension and/or interact with 'naphthalene and thus result in decrease of the experimental value of H. Since the solubility values are not significantly different, it is reasonable to conclude that the electrolytes effect insignificant in this case. The reduction of H values in the experimental waters may be attributed to the decrease of surface tension due to natural organic matrix [16]. Some workers [ 171 have reported that the presence of surface active agents reduces the volatilizatiqn rates of certain chemicals. Smith et al. [ 181 summarized different approaches used to explain the effects of surfactants on volatilization rates and indicated the complexities of these effects. The experimental H values were used as input parameters into the SLSA lake model, to estimate naphthalene volatilization rates from different types of water, The SLSA volatilization model is described by the two film mass transfer model. Table 6 and Figure 1 show naphthalene volatilization rate constants and half lives from the SLSA [q] model lake of 5.22 m depth. Tab. 5. Henry's Law Constants for Naphthalene Henry's Law Constant ' H m3atm gmo1-I Experimental Waters
Molar Solubility Molar/m3
Vapor Pressure(a) atm
Calculated(b)
Experimentally Derived
MilliQ Water Roselawn Cemetery Cross Lake Indiana Quarry
0.224 0.227 0.224 0.222
1.15 x 1.15 x 1.15 x 1.15 x
0.513 X 0.507 X 0.513 X 0.518 X
0.532 X 0.292 X 0.350 X 0.330 X
10-4 10-4 10-4 10-4
(a) Value from ref. [5] @) Calculated from molar solubility and vapor pressure Tab. 6. SLSA(a) Lake Model Estimation of Naphthalene Volatilization Rate Constants and Half Lives in the Experimental Waters Experimental Water
Rate Constants hr-'
MilliQ Water Roselawn Cemetery Cross Lake Indiana Quarry
0.425 X 0.233 X 0.280 X 0.265 X
lo-' lo-' lo-' lo-'
HalfLife t 1/2 hr
163 297 248 262
(a) The model assumes liquid phase mass transfer coefficient (kp Ilm/day) and gas phase mass transfer coefficient of Kg I1000 m/day. The water column depth is 5.22 m.
124
291
300
200
Derived from Two-film Theory
262
-
150 -
163 12: -
100 50
-
0, Cross lake
Roselawn Cemetery
Milli-Q Watr
Indiana Quarry
Fig. 1. Estimated Naphthalene Volatilization Half Lives in a Model Lake Using Experimental Waters.
3.3. Volatilization Rates by Measuring K:/K:
Table 7 and Figure 1 show the K&/K; and the estimated volatilization rates and lives of naphthalene from a lake of average oxygen reaeration coefficient of 0.01 hr-' . It is noted that naphthalene volatilization rate is faster in the experimental waters and follow the order CLi% RC > IQ > MQ. These results appear to be in contrast with the volatilizaTab. 7. Naphthakne Kt/K: and Estimated Volatilization Rate Constants and Half Lives in the Experimental Waters Experimental Waters MilliQ Water Roselawn Cemetery Cross Lake Indiana Quarry
K: hr-'
DOC(a) mg/L
Ratio K:/K:
X
S
Lake@)
t 112 hr Lake
0.0
0.602 1.320 1.510 0.633
0.79 0.061 0.082 0.160
0.56 X 1.32X 1.51 X 0.63 X
122 53 46 110
7.1 8.2 1.9
(a) DOC - Dissolved organic carbon (b) Oxygen reaeration coefficient of 0.01 hr-I
lo-' lo-'
125 tion rates derived from the experimental H values. However, if the photolysis side effect is taken into account, the results can be explained. As will be seen under the photolysis data, the natural water matrix has enhanced the photolysis rates of naphthalene in the same order as found in the K,C/K: experiments. The duration of the oxygen reaeration experiment is 5 to 6 hours during which the dissolved oxygen increases from 2 mg/L to saturation, 8.4 mg/L at 25'C. Under these conditions the photolysis side effect may be pronounced. For light sensitive volatile compounds, the photolysis side effect may be a significant factor affecting estimation of volatilization rates by measuring the K:/K: ratio.
3.4. Photolysis Rate Constants
Naphthalene is a bicyclic aromatic hydrocarbon and is expected to strongly absorb solar radiation at wavelengths above the solar cutoff of 290 nm and therefore may undergo direct photolysis or photooxidation. Table 8 and Figure 2 show the results of determinations of photolysis rate constants and half lives. It is noted that photolysis rates followed first order kinetics and that K values increase in the order MQ > IQ > XL. The results indicate photolysis rate enhancement in the natural waters by a factor of 2.3 in the case of IQ and by a factor of 6.9 in the case of XL. The extent of rate enhancement followed the same order as the organic carbon content of the water. The data suggest that the organic carbon matrix acts as a sensitizer which enhances naphthalene photolysis.
700r
673
600k
M i l t i - Q Water
NAPHTHALENE ADJUSTED M I DW I NTER HALF L I V E S
Cross Lake
Fig. 2. Naphthalene Photolysis Adjusted Midwinter Half Lives.
Indiana Quarry
126 Tab. 8. Kinetic Data for Naphthalene Photolysis(a) Milli-Q Water Naphthalene initial concentration mg/L pH at initial time at 25°C pH at final time 246 hr at 25°C Extent of reaction (percent) Photolysis rate constant Kp hr-' Half life (t 1/2) hr (1st order) Adjusted mid winter t 1/2 hr@) Field photolysis rate constant(@Kp hr-I Field photolysis half life hr
Roselawn Cemetery Cross Lake
5.10 6.98 6.45 92 2.07 x 10-3 335 673 0.623 X lo-" 11.12 x 103
5.10 7.30 6.41 90(d) 1.46 X 48.1 97 4.32 x 10-4 1.60 x 103
5.20 7.80 7.12 54(d) 4.7 x 10-3 147 295 1.41 X lo-" 4.92 x 103
(a) Mean sample temperature during the experiment 36.6 k24"C Using method described in ref. [14] and latitude 33"N (d) At 174 hrs (c) Photolysis conversion factor 0.03
(b)
Natural organics (humic and fulvic acids) through absorption of lght, may enhance photolysis rates due to photosensitized or photoinitiated free radical processes [ 191. These results are consistant with the findings of Smith et al. [8], who reported photolysis rate enhancement of benzo (a) anthracene and benzo (a) pyrene in natural waters when compared to rates in pure water. The experimental photolysis data were used as input parameters into SLSA lake model, to estimate naphthalene photolysis in different types of water. The field photolysis rate (Table 8), were calculated using the following expression: Kphoto lab (~-e-~e~l)f Kphotofield = Ke HI
where Ke - field extinction coefficient for the UV radiation wavelengths responsible for photolysis. In natural waters of medium depth of 5.22 m, K e range from 2.2 to
5.5 m-' photolysis period, 0.5 day H1 - water column depth (5.22 m) f
-
3.5. Sorption-Desorption
Naphthalene is a relatively non-polar compound and is expected to strongly sorb on sediments and partition into organic matter. Based on naphthalene octanol-water partition coefficient, estimated naphthalene sediment partition coefficients are 19.4, 19.2, and 8 for RC, XL, and IQ, respectively [20]. Sorption-desorption experiments were successfully run on Milli-Q water-montmorillonite clay system and only sorption experiments were run on RC water-sediment systems. Table 9 shows the experimental results.
127 Tab.191. Data for Naphthalene Sorption-DesorptionExperiments ExperimentalSystem Milli-QWaterMontmorillonite Clay
Sorption
Naphthalenein Water PglL Naphthalene in Sediment PgFg Desorption Naphthalene in Water PglL Naphthalene in Sediment Pglkg
331
426
550
1780
120x 103 i 7 o x 103 i 8 2 x 103 269x 103 31
50
20
22
40X 10’
28X lo3 83X lo3 l l O X lo3
K(4
ExperimentalSystem MiUi-Q WaterMontmorillonite Clay
Sorption
Naphthalene in Water Pg/L 3310 2950 6310 266 Naphthalene in Sediment Pglkg 513X lo3 813X lo3 1148X lo3
ExperimentalSystem Roselawn Cemetery Sorption WaterSediment
Naphthalene in Water Pg/L 121 Naphthalene in Sediment Pglkg 27x 103
125
543
549
39x 103 i o 7 x 103 121x 103
K(4
Experimental System Roselawncemetery Sorption WaterSediment
Napthalene in Water PglL 820 1600 1900 215 Naphthalene in Sediment Pglkg i 8 o x 103 200x 103 4oox 103
(a) Partition Coefficient Based on I/n = 1
In soils and sediments, the Freundlich’s isotherm is the most commonly used expression. X/M = K where amount of chemical adsorbed in g/kg of dry sediment
X/M
-
Ce
- chemical aqueous equilibrium concentration in g/L - adsorption coefficient
K l/n
-
constant parameter
A plot of log X/M vs. log Ce should give a straght line with a slope of l/n and an intercept of K, if the Freundlich’s expression is valid. The data derived from Table 9 are plotted in Figure 3. Using linear regression, the isotherm has a slope (l/n) of 0.53 and an intercept log K of 3.78. The latter value corresponds to naphthalene sorption-desorption coefficient of K = 6.03 X lo3. The Freundlich‘s expression is an empirical relation and is not derived from thermodynamic or kinetic concepts [21]. No methods are available for estimating l/n, if a mea-
128
7 6
5 E
\
-
4 A
Sorption Desorption Sorption, R.C.
3 r=0.88 Slope=0.53 Intercept=3.78
2
1
0 0
1
2
3
4
5
6
Log Ce Fig. 3. Freundlich Isotherm for Naphthalene Sorption-Desorption Experiments.
sured value is not available. A wide range (0.3 to 1.7) is reported in the literature [22] for l/n for a number of hydrophobic chemicals [20]. Some investigators, prefer tcassingn a value of 1 to the l/n in the Freundlich's equation [23]. Table 9 shows naphthalene sorption and desorption coefficients based on l/n = 1 . Naphthalene sorption coefficient in Mill-Q water-montmorillonite clay system and in RC water-sediment system are 266 and 215, respectively. The desorption coefficient in the first system, 2038, is significantly different from the sorption coefficients. Apparently variable K values can be estimated or derived from the sorption-desorption experimental data. Precaution should be taken in interpretting sorption data. There are several inherent problems and uncertainities associated with the sorption expression when it is applied to natural water-sediment systems. Assumption of reversibility of adsorption or linearity of the isotherm or the value of the slope are major sources of errors [23,24].
129
Also some investigators [23] have reported an inverse relationship between partition coefficient and solids concentration. We have observed the same phenomenon in the case of sorption experiments on mirex [25]. Based on the forementioned discussion it is difficult to assess the role of sorption as a pathway affecting the fate of naphthalene in the aquatic environment. However, a microcosm study [2q has shown that sediment is a minor compartment for naphthalene.
3.6. Use of the Kinetics and Equilibrium Data in SLSA Model
Input parameters to the SLSA model include physical transport characteristics of the aquatic environment and rate constants and coefficients of the chemical. Both parameters are interrelated and should be carefully determined for evaluation of the models predictive ability. In a subsequent phase of this research activity, a set of continous flow through microcosm consisting of water and sediment compartments, were used for SLSA model validation [25]. The microcosm physical transport characteristics in terms of mixing, air-water rate of transfer, and water-sediment rate of diffusion were determined. These data along with data on naphthalene rate constants and coefficients were used as input parameters to the SLSA model. Results are summarized in Table 10. It is apparent that volatilization is the major transport process for naphthalene under the microcosms operating conditions. Volatilization accounts for 78 to 98% of the total kinetic process. The large contribution of volatilization in the microcosm is essentially due to the high oxygen reaeration coefficient (0.1 18 hr -'). Under more quiescent conditions, for example in a lake situation with an average reaeration coefficient of 0.01, the contribution of volatilization would be close to other kinetic processes. Photolysis and biotransformation play minor roles and their contribution to the total kinetic processes are less than 25% in the microcosm experiments. However, the exact contribution may differ in different water matrices. The role of naphthalene sorption was difficult to predict, since variable values were derived from the sorption data. The experimental microcosm results indicate that sediment is only a minor compartment for naphthalene. Tab, 10. Kinetic Data for Naphthalene in the Microcosm and Percent Contribution of Each Process Milli-Q Water Volatilization Rate Constada) Kvol hr -' Photolysis Rate Constant@) Kphot hr-' Biotransformation Rate Constanddl Kbio hr-' Total Kinetic Processes hr-' Half Life of Total Kinetic Processes Percent Contribution of Kvol hr-' Percent Contribution of Kphoto hr-' Percent Contribution of K B hr-' ~ ~ ~
~~
~
0.10580 0.00256 0 0.10836 6.40 97.64 2.36 0
Roselawn Cemetery
Cross Lake
0.09326 0.09730 0.01 783(c) 0.0 1783 0.00853 0.00574 0.11956 0.12086 5.80 5.73 77.95 80.51 14.91 14.75 7.13 4.74
~
(a) Using microcosm 0, reaeration coefficient of 0.118 hr-' @) Microcosm photolysis conversion factor 0.05 16 (c) Kp was not experimentally determined, value used is estimated (d) From ref [ 121
Indiana Quarry 0.09610 0.00582 0.00506 0.10698 6.48 89.83 5.44 4.73
130 It is apparent that translation of the laboratory data to field or experimental microcosm, provides an overall picture of the processes controlling the fate of chemicals. Individual rate constants and coefficients may reflect variation due to natural water matrix. Magnitude of differences due to the overall processes under specific environmental conditions, may be different from the apparent variation in a single process determined in the lab.
4. CONCLUSIONS
Transport processes for naphthalene were determined in three surface water-sediment systems and in a Mu-Q clay system. Naphthalene aqueous solubilities in natural waiers were not significantly different from values determined in Milli-Q water and 25 *1 C. The experimentally determined H values in the natural waters were less than the value in Milli-Q water by a factor of about 2. In Milli-Q water, there is an excellent agreement between the calculated and experimental H values and both values agree well with published H values. Major kinetic processes for naphthalene are volatilization, photolysis, and biotransformation. The natural water matrix retarded the volatilization rates as calculated from H values. Estimation of naphthalene volatilization rates by measuring K$/K? may be subject to photolysis side effects. Photolysis rate followed first order kinetics and was enhanced by factors ranging from 2-7 in the natural water systems. Data on naphthalene sorption-desorption were not conclusive. Microcosm experiments indicated that sediment is only a minor compartment for naphthalene. The relative importance of each transport process is dependent on the system’s physical transport characteristics. Translation of the laboratory data to field or microcosm conditions may reflect a different situation than the apparent differences noted in individual processes measured in the laboratory.
ACKNOWLEDGMENT
We wish to acknowledge the excellent laboratory technical support of Ms. S. Rafferty, K. Pace, and R. Stolley. Funds supporting this research were provided by the Chemical Manufactures Association and North Texas State University Faculty Research Committee.
REFERENCES 1 L. A., Cline, D. M., and Lassiter, R. R. ‘Exposure Analysis and Modeling Systems (EXAMS)’. User Manual and System Documentation. USEPA Athens, GA (1981). 2 D. M., Di Toro, O’Connor, D. J., Thomann, R. V. and St. John, J. P. ‘Analysis o f the Fate of Chemicals in Receiving Waters Phase 1’. Report to the Chemical Manufactures Association, CMA Project ENV-7-W (1981). 3 G. L. Baughman, and Burns, L. A. ‘Transport and Transformation o f Chemicals: A Perspective’ In: The Handbook of Environmental Chemistry, Reactions and Processes. 0. Hutzinger (ed.) (1980). 4 U.S. Environmental Protection Agency, Water-related Environmental Fate of 129 Priority Pollutantsvolume 1, (1979).
131 5 Handbook of Chemistry and Physics CRC (1980) D-173-175. 6 G. D., Veith, Defoe, D. L., and Bergstedt, B. V. ‘Measuring and Estimating the Bioconcentration Factor of Chemicals in Fish’, J. Fish. Res. Board Can. (1979) 36, 1040. 7 Standard Methods for Examination of Water and Wastewater, 15th Edition (1980). 8 C. A., Black, Evans, D. D., White, L., Ensminger, L. E., Clark, F. E., and Dinauer, R. C. Methods of Soil Analysis Parts 1 and 2 (1965). 9 Federal Register, ‘Toxic Substances Control: Discussion of Premanufacture Testing Policy and Technical Issues’(1979) 44 (53): 16250-16277. 10 J. H., Smith, Mabey, W. R., Bohonos, N., Holt, B. R., Lee, S. S., Chou, T. w., Bomberger, D. C., and Mill, T. ‘Environmental Pathways of Selected Chemicals Fresh Water Systems, Part 11: Laboratory Studies’ (1978) EPA 60017-78-74. 11 D. Mackay, Shiu, W. Y., and Sutherland, R. P. ‘Determination of Air-water Henry’s Law Constants for Hydrophobic Pollutants’. Environ. Sci. and TechnoL, (1979) 13: 333-337. 12 K. L., Dickson, Rodgers, J. H., and Saleh, F. Y.‘Measuring Rate Constants for Chemicals in Simple Model Aquatic Laboratory System’. Report to the Chemical Manufactures Association, CMA Roject ENV-8-W, March (1981). 13 J., Hill, KO&, D. F., Paris, D. F., Wolfe, N. L. and Zepp, R. G. ‘Dynamic Behavior of Vinyl Chloride in Aquatic Ecosystems’ U.S. (1976) EPA-600-3-76001. 14 T. Mill, and Mabey, W. R. ‘Laboratory Protocols for Evaluating the Fate of Chemicals in Air and Water’ SRI International EPA Contract 68-03-2227 (1 980). 15 W. E., May ‘The Solubility Behavior of Polycyclic Aromatic Hydrocarbons in Aqueous Systems’ In: Petroleum in the Marine Environment. ACS Series 185 (1980). 16 R. L., Wershaw, Burcar, P. J. and Goldberg, M. C. ‘Interaction of Pesticides with Natural Organic Material’ Environ. Sci. & Technol. (1969) 3: 273. 17 D. D., Lee, Masters Thesis, Iowa State University, Ames, L4 (1973). 18 J. H., Smith, Bomberger, D. C., Jr. and Haynes, D. L. ‘Prediction of the Volatilization Rates of High Volatility Chemicals from Natural Water Bodies’, Environ. Sci. & Technol. (1980) 14, 1332. 19 R. G., Zepp, and Cline, D. R. ‘Rates of Direct Photolysis in Aquatic Environments’. Environ. Sci. & Technol. (1977) 11, 359. 20 S. W., Karickhoff, Brown, D. S., and Scott, T. A. ‘Sorption of Hydrophobic Pollutants in Natural Sediments’ Water Res. (1979) 13, 241. 21 W. J., Weber, Jr., ‘Principles and Applications of Adsorption’ In: Manual of Treatment Processes VoL 1: Eckenfelder Jr. Ed. Environ. Sci. Serv. Corp., Stamford, Conn. (1968) Ch. 15. 22 J. W. Hamaker, and J. M. Thompson ‘Adsorption’ In: Organic Chemicals in the Soil Environment Vol. 1. C. A. I. Coring and J. W. Hamaker (eds) MarcelDekker Inc, New York (1972). 23 D. M., Di Toro, and Horzempa, L. M. ‘Reversible and Resistant Components of PCB AdsorptionDesorption Isotherms’ Environ. Sci. and Technol. (1982) 16,594. 24 P. S. C., Rao, and Davidson, J. M. ‘Estimation of Pesticide Retention and Transformation Parameters Required in Nonpoint Source Pollution Models’ In: hironmental Impact of Nonpoint Source Pollution, M. R. Overcash and J. M. Davidson (eds) Ann Arbm Science Publishers Inc Ann Arbor, MI (1980). 25 F. Y., Saleh, Staples, C. A., Dickson, K. L., and Rodgers, J. H. ‘Determining the Compartmentalization and Fate of Chemicals in Aquatic Microcosms’ Final Report to the Chemical Manufao turers Association, CMA Project No. ENV-8-W (May 1982). 26 C. A., Staples, Dickson, K. L.,Saleh, F. Y.and Rodgers, J. H. ‘A Microcosm Study of Lindane ami Naphthalene for Model Validation’ Proc on Sjxth Annual Symp. on Aquatic Toxico. Oct. 1981. ASTM Public (in press).
This Page Intentionally Left Blank
133
INTERLABORATORY COMPARISON OF QUANTIFICATION OF PAH IN ATMOSPHERIC AEROSOLS BY GAS AND LIQUID CHROMATOGRAPHIES AND BY SHPOL’SKII FLUOROMETRY
M. LAMOTTE
Lab. de Chimie Physique A - Universitk de Bordeaux I - LA 348 351, cours de la libkration - 33405 Talence ckdex, France P. MASCLET
Lab. de Physico-chimie Instrumentale - Universitk Paris VII 2, place Jussieu - 75251 Paris ckdex 05, France
ABSTRACT An intercomparison exercise on polynuclear aromatic hydrocarbons (PAH) analysis in atmospheric aerosols is reported, The sample was taken from an underground parking lot in Jussieu (Paris University) using glass filter and HIVOL Sampler. After extraction (cyclohexane/dichloromethane, 60/40 in volume), the dried sample was dissolved in methanoL Three techniques have been used for the analysis: - HPLC on reversed phase and fluorimetric detection (4 laboratories) - Shpol’skii fluorimetry at 15 K (2 laboratories) - Capillary gas chromatohraphy (3 laboratories). 16 aromatic including the ones recommended by the EPA has been quantified. There respective concentrations were found to be about 2 to 6 ng/m’ of air. With respect to the mean values, the deviations ranged from 20 to loo%, they are however more reduced when only results obtained by HPLC are compared. The results obtained from Shpol’skii fluonmetry at low temperature are in good agreement with the HPLC values. However, one laboratory gave some values which are slighly lower. Values obtained by capillary gas chromatography appears, for most PAH, substantially overestimated. As no previous separation was performed, these results must be caused by th,e complexity of the chromatograms. Whereas the determination of relative concentrations and/or their variations have a widely accepted significant character, the present intercomparison points to the difficulties in determining absolute concentrations. The discrepancies in the reported results are for the main part due to intrinsic defects of each technique but must be also attributed to the difficulties in preparing homogeneous reference materials and samples: evaporation loss occuring when changing the solvent, photochemical degxadation, etc., are problems which are currently encountered in environmental monitoring programs.
134 1. INTRODUCI'ION
Eight french laboratories took part in an intercomparison exercise on the analysis of PAH in an extract of atmospheric aerosols. PAH are compounds, most often highly carcinogenic, whose origin is the combustion of fossil fuels. Three criteria have been considered for this study: - the reliability of the method - the maximum sensitivity for the assay of traces in the atmosphere - the rapidly of the measurement, in order to approach a routine technique. The sample was taken in the underground car-park of Jussieu on a Whatman GF/A glass filter by a HIVOL sampler. The amount collectec corresponds to about 200 m3 of heavity polluted air where the PAH come from motor vehicles. The sample was prepared by only one laboratory, by extraction with a cyclohexane/dichloromethane mixture (60/40), then diluted in methanol after evaporation to dryness. It is known that the sample preparation method affect the results markedly. This amount concerns only the discussion of the advantages and disadvantages of the analytical methods: - reverse phase High Performance Liquid Chromatography (HPLC) with fluorometric or W detection (4 laboratories) - low temperature Shpol'skii fluorimetry (2 laboratories) - capillary gas phase chromatography (GPC) with flame ionisation detection or coupled with mass spectroscopy (2 laboratories). The analysis concerned 16 PAH, including those listed by the Environmental Protection Agency (EPA). The results are given in Tab. 1. The names of the laboratories do not appear; only the technique employed is indicated.
2. HPLC ANALYSIS
The chromatographic conditions and those of identification and assay differ for each laboratory. In the laboratory I the analysis was performed in two stages under isocratic conditions with an acetonitrile - water mixture. The column used was a Merck Lichrosorb (5 m). In laboratory I1 the analysis was carried out in one step using a particulary economical methanol/water gradient. In laboratory I11 the separation required two chromatograms on a Whatman 10 ODS column of 9 mm internal diameter and 50 cm long. These are rather susprising conditions fot the analysis of traces and more appropriate to preparative chromatography. Nevertheless the resolution is very good. The laboratory IV used a Vydac column reputed to be very suitable for the separation of PAH. In all the cases it appears that the chromatographic conditions present no particular problem, at least for the tetracyclic compounds and bigger. All peaks are resolved and their retention times always differ by at least 30 s. This separation allows the peaks to be integrated under good conditions.
135 Tab. 1. Concentrations of PAH (ng in the sample
-
20 m 3 air)
Lab o ra t o irie s Compound Fluorthe PhBnanthrhe Anthracene Fluo ranthene Pyrhe BaAnthrache Chryskne B bFluoranth6ne BkFluoranthkne BePyrhe BaPyrkne DibzahAnthracSne Peryline BghiF'6rylkne Ind6noPyrhe Coronene
I
25 25 25 135 110 270 330 430 185 260 295 15 65 460 245 180
I1
111
220 6 390 40 60 10 230 150 170 110 150 200 270 270 270 450 130 180 170 270 230 240 35 100 50 50 220 400 320 290 700 80
IV
v
-
-
VI
VII
VIII
Meanvalue
600 1970
-
-
-
-
2 7 40 20 standard 170 265 130 80 125 90 40
-
-
-
950 580 180 300 600 210 430 40 590 550 530 330
-
-
440 270
-
-
140 110 70
390 180 -
230
380 5 10 -
390 -
1200 -
240 900
-
-
1000 1000 600 170 400 200
230 240 200 320 380 160 250 380 240 80 370 250
-
-
-
850 1350 -
It must be noted, however, that the background intensity depends on the method chosen. The detection conditions are as follows: all the laboratories used variable wavelength fluorometric detection Laboratory IV used fluorometricand W detection. Laboratory I used 7 pairs of wavelengths making it possible be measure the PAH under the most sensitives conditions; only 3 or 4 PAH can be measured from each chromatogram, with the consequent disadvantage that analysis is too long. Laboratory I1 used 4 pairs of wavelengths chosen so that they could be changed during the chromatogram without stopping it; the analysis is shorter but the highest sensitivity is not always achieved. An example of a chromatogram is given in Fig. 1. Laboratory I11 used a different pair of wavelengths for each PAH which was identified by stopping the chromatogram just before the top of the peak and recording the fluorescence spectrum. The assay was perfomed on the spectrum itself, usually by adjusting the wavelengths and the slits. The obvious advantage is in the unambignous identification of each compound eluted, but its complexity and the excessive duration of the analysis make this method unsuitable for routine measurements. Laboratory IV used fluorometewith a fEed wavelength of 3 7 0 4 1 0 nm; only a few PAH can be measured under these conditions. The others were analysed by W; in this case the aspect of the chromatogram changes and since the sample has not been retreated, many compounds can be coeluted, hindering peak assignment and, in particular, introducing errors into the peak area calculations. An example of a chromatogram is given in Fig. 2. This method can still be applied to a very polluted sample but is no longer applicable when the samples come from little polluted sites. The calibration methods are various, either by measurements of the fluorescence spectrum with a calibration curve, or by means of external or internal standards. The
136
0
P
Mttt
1 t t t t t -?l
ii
10mn
20mn
t 30mn
co
'II
Fig. 1. HPLC of PAH in airborne particulates with variable fluorimetric detection
first is the most reliable but obviously the longest. The second is easy and give good results but does not allow for the correction of possible variations in the luminous intensity in the fluorometei. This problem arises frequently and the method is somewhat risky. The internal calibration method gives good results, but requires a supplementary chromatogram if the standard chosen is present in the sample (as it happens, fluoranthene for laboratory 11). If the site is only slightly polluted a concentration ratio of 100 between the standard and the sample avoids the use of corrections. The results are of good quality for laboratories I, I1 and 111, the values being close to the average and always in the range of experimental error. Some low values are however obtained by laboratory 11. For coronene, nevertheless, the values are not consistent. The results of laboratory IV are generally bad; the errors rising to as much as a factor of 10. This laboratory carried out the analysis 3 months after the samples were taken, in contrast to the others laboratories. This can explain in part the very low value for BaP (40 ng for an average of 380 ng) since it is known that BaP is rapidly degraded in solution. For fluoranthene (950 for an average of 230) only an error of calihation or identification can explain this difference. We have already seen the causes of error for compoundsassayedby W. 2. ANALYSIS BY LOW TEMPERATURE SHPOL'SKII SPECTROSCOPY
Laboratories V and VI used this method at 10°K or 16°K in an n-octane matrix. One
137
UV
,254nm
FLU0 Xex = 370nm
Fig. 2. HPLC of PAH in airborne particulates with UV detection and fixed wavelength fluorimetric detection
interest of the method is obvious; it requires no chromatographic separation since each PAH appears in the fluorescence spectrum as one or more peaks (multiplets due to different insertion of the PAH in the n-octane matrix). The lines are very precise and characteristic. The analysis itself is short but is requires calibration curves set up by the technique of adding known amounts (Fig. 3). This clearly requires time and many experiments. The appearance of the spectrum given in Fig. 4 shows that the background noise is considerable. This does not hinder peak assignment but makes the method less sensitive (except for some PAH such as BaP) than HPLC coupled with fluorescence. The two laboratories using Shpol’skii spectroscopy did not employ the same calibration, and for this reason there may be differences in the results. Laboratory V obtains excellent results (always very close to the average) but few measurements were performed. Laboratory VI often obtains values somewhat less than the average; note however that the ratios betyeen the concentrations of the PAH stay the same. This is no doubt due to a small systematic error which difficult to explain (perhaps pyrene was not a good choice for the internal standard). 3. ANALYSIS BY CAPILLARY GAS CHROMATOGRAPHY
Laboratories VII and VIII employed this method. Laboratory VII coupled the chro-
Extract SO
1
398
392
BkF
. . . . . . . . . . I
392
390 nm
nC8 15K
Fig. 3. Calibration curves in Sholp’skii fluorometryat 15 K (B k Fluoranthene in n-octane)
1
1
I
400
1
I
I
I
425
Fig. 4. Sholp’skii spectrometry of an extract of PAH in airborne particulates
1
450
1
hhm)
139
P ht EM :ester m&hylique Pht: phtalate A119 6 A130: n alcanes
P
:M 36 EM
I
2A:
Pht
Fig. 5. Ion chromatogram of airborne particulates (GC/MS)
matograph to a mass spectrometer. The chromatogram obtained (Fig. 5 ) shows that the PAH are not the most important compounds but that esters, phtalates and aliphatic compounds (C19 to C, alkanes) predominate. Some PAH were identified by their molecular weight in particular small peaks difficult to assay. This laboratory concludes that it is impossible to use the sample in the state it was supplied. In several cases, when GPC is used, the authors carriedout a pre-separation, either by liquid-liquid extraction with DMSO (for example) or by chromatography on a column or on a silica gel plate. The two methods can be used together and lead to a marked decrease in the non-aromatic compounds but also to (small) losses of PAH. The chromatogram is simplified and can be exploited. Laboratory VIII used GPC with a capillary column (CP SIL 5) taking care to correct the base line as the temperature varied; conventional flame ionisation detection was used and the PAH are assayed by the addition ot two internal standards. Despite these precautions, since the sample had not been preseparated, the results are frankly poor. Only B g h ~perylene, indenopyrene and fluoranthene are correctly evaluated. All the others are seriously overestimated by factors 2 to 10 (BeP : 1000 ng for an average of 250 ng). The cause is obviously the complexity of the chromatogram (no selective detection) and the systematic overestimation of concentration due to the superposition of coeluted components.
140 4. THE PARTICULAR CASE OF TRICYCLIC COMPOUNDS
The results which are on the whole consistent for tetracyclic or bigger PAH are not for the tricyclic compounds: phenanthrene anthracene and fluorene. Several causes can be suggested: the most likely is due to the preparation of the sample by evaporation of the solvent, leading to considerable losses (up to 60% for phenanthrene). Next, the time between the preparation of the sample and the analysis seems always to be too long. The only laboratory (11) which performed the analysis immediately after the preparation (the same day) finds high values. The low concentrations observed in some cases can also be explained by evaporation during a change of solvent, in Shpol’skii spectroscopy. An other, equally plausible, reason lies in the possibility of interference of the peaks of the tricyclic compounds. In reverse HPLC, these compounds are rapidly eluted in a region where polar aromatic compounds appear, identification becomes difficult and peaks are likely to be superposed.
5. CONCLUSIONS
h Conolusion, three methods have been used. Generally speaking, it appears that the results are satisfactory for 12 of the 16 PAH studies (20 to 100% deviation on average), but that each method is shown to have particular difficulties. HPLC with fluirimetric detection is a classical, relatively long method, and one it obliged to limit the number of wavelengths used for the assay. The internal calibration method seems preferable, the analysis time can now be reduced by employing ”high speed analysis” column; in this case it constitues a routine method, but the sensitivity is not very good. In our study the sensitivity is excellent (threshold detection for BaP is 10 pg/m3 air). Shpol’skii spectroscopy also gives very good results and promise to see wider use in the future because it is well adapted to PAH measurements. The sensitivity of the Shpol’skii is comparable to that obtained in HLPC with fluorimetric detection for many PAH; however, since the apparatus is more specialised it cannot be used for the analysis of others types of compounds. In conclusion, on the three criteria considered two are satisfied: 1 - the methods are reliable (HLPC and Shpol’skii; perhaps GPC would be also after pretreatment of the sample), 2 - the methods are sensitive; the detection threshold allows short sampling times in rural area or even in marine aerosols, 3 - on the other hand it appears that, at present, no method is really adapted to routine analysis; in the most favorable case the sampling, the preparation and the analysis of the sample require a day’s work. It is in this direction that the studies must be furthered. It appeared that the problem of assaying the light PAH was not resolved at the level of the analysis itself. Since moreover, the concentrations of these volatile compounds depends markedly on the sampling conditions (temperature, etc.) three are serious doubts about the validity of measurements on these compounds.
141 LABORATORIES TAKING PART IN THE INTERCOMPARISON EXERCISE: UNIVERSITE BORDEAUX I(33); laboratoire de Chimie Physique A; Lamotte M. et Joussot-Dubien J . UNIVERSITE PARIS 7 (75); laboratoire de Physico-Chimie Instrumentale; Masclet P. et Mouvier G. UNIVERSITE DE SAVOIE - CHAMBERRY (73), service de Chimie; Martin Bouyer M., Jarosz J., Paturel L. et Wittenberg M. INSTITUT CURIE (91), ORSAY; section de biologie; Muel B. UNIVERSITE D’AIX-MARSEILLE (13); centre de Spectroscopie Moleculaire; Mille G. LABORATOIRE D’HYGIENE DE LA VILLE DE PARIS (75); Person A. et Festy B. UNIVERSITE DE METZ (57), service d’analyse commun LAMMA; MulIer J. F. INSTITUT FRANCAIS DU PETROLE DE RUEIL (92); analyses; Petroff N.
This Page Intentionally Left Blank
143
COWARTSON OF IONSELECTIVE ELECTRODE AND GASSENSING ELECTRODE TECHNIQUES FOR MEASUREMENT OF NITRATE IN ENVIRONMENTAL SAMPLES
J. C. SYNNOTT, S. J. WEST, J. W. ROSS
Orion Research, Incorporated, Cambridge, M A , USA
ABSTRACT Work [l] published recently in Germany describes a method for nitrate analysis in which nitrate is reduced, using a solution of titanous chloride, to ammonia. An ammonia gas-sensing electrode is used to make the measurement, so interferences usually associated with nitrate determination are minimized. A decrease in the time required for analysis is also reported. We have evaluated the method in an attempt to simplify it, and we have compared it with techniques which utilize a liquid membrane, nitrate-selective electrode, both with and without chemical suppressionof interferences. The methods were developed with particular regard for the chloride interference to the nitrate electrode and for temperature effects in the ammonia measurement. Samples tested using these methods included drinking water, river and lake waters, and seawater. This paper tabulates the comparative test results and reports on reproducibility at low concentrations, spike recoveries, and calibration stability. The advantages and limitations of the methods are discussed.
1. INTROEUCTION
Concern for the adverse health effects of excess nitrate in the environment has provided the impetus for development of suitable analytical measurement techniques. ‘Suitable’ has come to mean fast and simple, as well as precise and accurate, since larger numbers of samples must be tested. Traditional colorimetric and spectrophotometric methods [2] tend to be tedious, and suffer from interference from sample color and turbidity. Analyses can be made quite simply using a liquid membrane ion-selective electrode [3], but suppression of interferences, particularly from chloride and carbonate, is necessary, adding to the cost of measurement. The method published by Braunstein et al. [l] seemed to respond to the problems associated with other techniques. It uses a solution of titanous chloride to reduce nitrate to ammonia which is measured using a gas-sensing electrode. Although somewhat complicated to run, it showed promise, especially in regard to elimination of interferences. We have attempted to simplify the method for use by less sophisticated technicians, and to evaluate the quality of data obtained, using the modified method, on a variety of environmental samples. We believe the method to be both practical and predictable.
144 2. EXPERIMENTAL 2.1. Apparatus
Digital mV-pH meter, Orion model 901. Printer, Orion model 95 1. Nitrate ion-selective electrode, Orion model 93-07. Ammonia gas-sensing electrode, Orion model 95-10. Chloride ion-selective electrode, Orion model 94-17B. RossTM pH electrode, Orion model 81-02. Reference electrode, double-junction, sleeve type, Orion model 90-02. Standard laboratory glassware and equipment.
2.2. Reagents
Titanous chloride solution, 20% practical grade. Nitrate interference suppressor solution, Orion 93-07-10. Sodium hydroxide solution, 1ON, reagent grade. Calibration solutions for each of the electrodes were prepared from reagent grade chemicals. Electrode filling solutions were those recommended by the manufacturer.
3. METHOD DEVELOPMENT
The intent, from the start of this work, has been the development of a practical, simple test method. Braunstein’s work used an ammonia gas-sensing electrode for measurement at pH near 13. This pH is higher than that needed for ammonia measurement, but was suggested as necessary to achieve complete reduction of nitrate to ammonia. Except for the addition of titanous chloride, the approach appeared to be direct measurement as described in the electrode manual [4]. We hoped to simplify the technique by incorporating nitrate standards and direct digital readout (in mg/L NOT-N). The initial experimental work was carried out using commercially available titanous chloride solution (20%), but two problems became apparent immediately. In using concentrations of hydroxide and Tic13 that were appropriate to the measurement levels, we were unable to attain pH of 13, and a substantial thermal excursion (- 10°C) was observed. Both problems were traceable to the titanous chloride solution which was found to be 10 M in acid. We presume that the solution is prepared by dissolving titanium metal in concentrated hydrochloric acid, though this has not yet been verified. Even if the hydroxide level in the measurement were sufficient to offset the acid, the heat of neutralization would adversely affect the performance of the electrode. We addressed the problem by ‘preneutralization’ of the titanous chloride to pH 0.25, using 10 N sodium hydroxide. The RossTM pH electrode was used to monitor the addition of hydroxide since it is not affected by heat of neutralization. Little titanium is lost to precipitation at this pH level, and the conversion to pH = 13 can be accomplished
-
-
145
without extraordinary temperature problems. Preneutralized titanous chloride (P. N. TiCIB) was used throughout this series of analyses. It remained to choose a P. N. TiC13 level that would be adequate for a reasonable nitrate range, and a hydroxide level sufficient to set pH at 13 for both conversion and measurement. In ordinary ammonia-measurements, a fmed .volume of hydroxide solution is added to a fured volume of standard or sample, and the ammonia concentration is determined from a calibration curve (Fig. 1). Reproducibility is expected to be within one millivolt over the calibration range. We chose, arbitrarily, a sample volume of 100 ml, and 10 ml of P. N. TiCIB as a starting point, and determined that the solution would have to be made at least 0.5 M in hydroxide to achieve pH > 13. For simplicity, we decided to use 10 ml additions of 10 M hydroxide. By fixing all the volumes used in measurement, it is possible to provide for direct readout on a digital meter or a calibration curve. -1(
I
I
I
I
E
REPRODUCIBILITY: 5 RUNS OF 10-3 M SAMPLE -15.8 f 0.3 m\’
1(
SLOPE = 57 mV
Fig. 1. Standard ammonia electrode calibration in 1.0 M KOH.
Calibration curves (Fig, 2) were generated, within the above constraints, using both ammonia and nitrate standards. The reproducibility values indicated were obtained using nitrate standards, and are about the same for ammonia standards. They include contributions from temperature, volume error, ammonia loss, etc., so are representative of what a user could expect to achieve. The range for nitrate is adequate for wastewater analysis but would need improvement for low level measurement as in the case of most drinking water samples. Reducing the level of P. N. TiC13 (Fig. 3) does not, of itself, solve the problem. More work is being done to tailor a low level method and results will be reported at a later date. The ‘method’, at this point, seemed simple, reasonabIy fast, and adaptable to direct readout, so we began testing on environmental samples.
146
1
f
1
I
-50 E
(mV) 2
1.8mV
-
0-
0.83M KOH - 1.67% TiCI,
+ 50 .o1
I .1
w-r
NH, STANDARDS
0-0
NO,
I 10
I
1
STANDARDS
I
100
-
lo00
(N]. mglL
Fig. 2. Comparison of NO; and NH, as calibration standards.
1 /" 1
1.0M KOH 1.8% TiCI,
/"
7"
*/-*
1.0M KOH
0.2% TiCI,
I .01
I
.l
I 1 (NO,--N]. mglL
I
10
J 1
Fig. 3. Effect of TiCl, concentration on nitrate determination.
4. SAMPLE CHARACTERIZATION
Samples were collected from a variety of natural and processed waters as described briefly in Table 1. These were intended, primarily, to provide matrices with which to compare methods of analysis, both directly and by spike recovery (recovery of known
147 Tab. 1. Description of Samples ~
~~
~~
Approximate Concentration WlL Sample
Source
C. R.
Tidal River Sampled Upstream Fresh Water Pond Qcean Sampled At Shore Municipal Water SUPPlY Waste Treatment Clarifier Effluent Waste Treatment Clarifier Effluent
J. P. W. B. T. W.
F. E. D. M.
NO;
NH,
Cl-
<1
< 0.1 < 0.1
67
I0 (see text)
0.3
<1
< 0.1
51
15
0.3
87
13
0.1
45
<1
68
2457
additions of nitrate). No special precautions were taken in sample handling or preservation. The nitrate test methods included use of a liquid-membrane, nitrate-selective electrode, both with and without chemical suppression of interferences, so all samples were analyzed for chloride, a major interference for that electrode. Ammonia content was determined, since it would affect the gas-sensing, Tic13 method. In addition, a rough estimate of nitrate was made to facilitate the comparative testing. The ammonia level in each case was low enough to avoid a major contribution to the spike recovery tests, so further testing for it was not carried out.
5. COMPARATIVE TESTING OF METHODS
The first experimental series evaluated the effect of interference suppression on the liquid membrane electrode method. The commercial suppressor solution contains aluminum sulfate (to complex organic anions), silver sulfate (to minimize halogen, sulfide, and cyanide interferences), sulfamic acid (to reduce nitrite levels), boric acid (to act as a preservative), and a pH adjustor (to limit CO, interference). The efficacy of this solution depends on the total of the interfering ions in the sample and on the concentrations of the suppressor constituents. As a practical matter, chloride, the most significant interference, will cause 10% measurement error when its concentration exceeds nitrate concentrations by a factor of approximately 20. The presence of other interferences, however, can effect the capacity of the suppressor solution. When run without suppression, the method uses ammonium sulfate as an ionic strength adjustor. Calibration curves for each technique, with typical reproducibility values, are shown in Figures 4 and 5. Samples form four sources were tested (Tables 2 and 3). In each case, we prepared a spike to assess recovery, and measurements were replicated five times. The reproducibility of potential values was remarkably good at all levels, and littie drift was observed. This is especially noteworthy in view of the recovery results in tests without interference suppression. These clearly show that one can obtain very reproducible, stable, wrong answers if sample condition is not considered. Recoveries where suppression was used show that
-
148
/I
100-
200E (mV)
/”
5 Ax /’
300.bl
A. 0 SLOPE = 58 mV I
REPRODUCIBILITY: RUNS OF 14 mglL SAMPLE 153.94 f 0.6 mV
I
.1
1
I
10
I
100
Fig. 4. Liquid membrane nitrate electrode calibration without chemical suppression for interferants
I
I
I
I
1
/‘
100-
200E (mV)
$4”
A ,’
% - X I x
,’.
’SLOPE = 58mV 300.01
I
.1
I
1
REPRODUCIBILITY: 5 RUNS OF 14 mglL SAMPLE 155.98 2 0.3 mV I
10
I 100
1
Fjg. 5 . Liquid membrane nitrate electrode calibration with chemical suppression for interferants.
the method is adequate as a check on the gas-sensing technique to be tested. An exception to this is sample W.B., a seawater sample, which will be discussed shortly. The gas-sensing electrode/TiC13 procedure is fully described in Appecdix 1, and can be easily adapted for direct readout of concentration. It was used in this measurement series on six samples after calibration as shown in Fig. 6. Spike recovery tests and replication of measurements were as in the fnst series. The nitrate levels were verified using the
149 Tab. 2. Nitrate Sample Analysis Liquid Membrane Electrode Without Interference Suppression As Collected Sample
EActual
C. R.
240.8 mV 239.8 239.0 238.9 238.7 250.3 249.9 249.1 248.3 247.4 169.0 169.3 169.3 169.5 169.7 247.1 246.9 246.9 246.0 245.9
J. P.
W. B.
T. W.
As Spiked EAverage 239.44 * 0.87 mV
249.00
* 1.18 mV 169.36
* 0.26 mV 246.56 k0.56 mV
EActual 196.8 mV 196.7 1.47 mg/L 197.0 196.6 196.4 201.1 200.8 0.74 mg/L 201.0 200.9 201.1 151.6 152.7 67.79 mg/L 152.8 (see text) 153.1 153.6 199.9 199.7 1.34 mg/L 200.2 199.5 199.7
%
EAverage
"0, -N1
196.70 k0.26 mV
19.26 mg/L 120.7
200.98 k0.13 mV
15.87 mg/L 107.7
k
152.76 0.74 mV
141.7 mg/L 105.4
199.80 k0.26 mV
18.18 mg/L 118.6
EAverage
[NOi-N]
199.72 *0.60 mV
14.64 mg/L
202.32 *0.44 mV
14.48 mg/L 101.2
k
144.10 0.25 mV
140.2 mg/L
99.9
204.12 kO.11 mV
14.17 mg/L
99.0
Tab. 3. Nitrate Sample Analysis Liquid Membrane Electrode With Interference Suppression As Collected Sample
EActual
C.R.
245.9mV 246.7 246.4 246.0 246.1 251.4 251.3 250.9 251.2 250.7 160.4 159.6 159.4 159.0 158.8 245.4 246.4 246.4 246.9 246.3
J.P.
W.B.
T.W.
As Spiked EAverage 246.22 0.33 mV
*
251.10
* 0.29 mV 159.44
* 0.62 mV 246.28 0.54 mV
*
"0;-Nl
EActual
199.1 mV 199.4 0.65 mg/L 199.4 200.5 200.2 201.6 202.2 0.31 mg/L 202.5 202.7 202.6 143.8 144.1 73.95 mg/L 144.4 (see text) 144.3 143.9 204.2 204.2 0.31 mg/L 204.2 204.0 204.0
% Recovery
99.9
150
Fig. 6. Calibration of ammonia electrode for TiCl, /gassensing method
25 20
/ /
-
/
//i/ /
/
/
THEORETICAL -
/ / /
/
/
//
/
/
0.01
0.1 SAMPLElSUPPRESSOR RATIO
1
Fig. 7. Measure of nitrate in seawater by nitrate-selective electrode with interference suppression
151
suppressed, liquid-membrane electrode technique. The results, summarized in Table 4, show that the method provides for reproducible, accurate measurement of nitrate. The results for seawater, however, do not agree with the liquid membrane electrode technique. We checked this technique by analyzing a series of sample dilutions as shown in Figure 7. The deviation observed suggests that the method is providing false information, probably due to uncompensated chloride interference, and, therefore, that the gas-sensing data is likely to be correct.
Tab. 4. Nitrate Sample Analysis TiC1, /Gas-Sensing Electrode Technique
C. R.
J. P.
W.B.
T.W.
F. E.
D.M.
43.4 mV 44 .O 44.1 43.6 44 .O 48.2 49.7 49.0 49.5 49.8 55.1 55.6 56.1 55.6 57.1 47.1 47.5 48.5 48.8 49.2 -19.8 -19.1 -18.5 -18.6 -19.5 -15.2 -15.5 -16.0 -16.1 -15.9
43.82 i0.30 mV
0.60 mg/L
49.24 *0.66 mV
0.34 mg/L
55.90 *0.76 mV
0.08 mg/L (see text)
48.22 k0.89 mV
0.37 mg/L
-19.10 i0.56 mV
15.24 mg/L
-15.74 i0.38 mV
13.08 mg/L
-2.2mV -3.3 -2.1 -2.9 -3.1 -1 8.6 -18.3 -17.7 -18.5 -16.5 -1 .o + 0.5 -0.9 -0.5 -1.2 -3.2 -2.5 -2.1 -2.0 -2.4
-2.72 i0.54 mV
1.56 mg/L
-17.92 i0.86 mV
14.40 mg/L 100.4
-0.6 2 i0.68 mV
6.90 mg/L
98.5
6.10 mg/L
99.3
99.5
-2.44 f 0.47
mV
152 6. CONCLUSIONS
The experimental work reported here results in a simple, practical method for the analysis of nitrate in a wide variety of environmental water samples. It is generally free from the interferences encountered in most nitrate measurement techniques, and is quite precise and accurate. Ammonia, a potential interference, can be determined separately and substracted where significant. The method requires no sophisticated instrumentation or personnel, can be adapted for direct concentration readout, and is usable in the field as well as in the laboratory. The method has been developed for use in an on-line monitor, and is being refined for low level measurement.
REFERENCES 1 L. Braunstein, K. Hochmiiller, and K. Spengler, Special Printing from ‘Vom Wasser’, 54, 1980, Verlag Chemie GmbH, Publishers. 2 Methods 418, Standard Methods for the Examination of Water and Wastewater, 15th edn., 1980, Published jointly by the American Public Health Association, the American Water Works Association, and the Water Pollution Control Federation, Washington, D.C. 3 ‘Nitrate in Potable and Ground Water’, Orion Technical Service Notes, Orion Research, Inc., Cambridge, MA, 1982. 4 Model 95-10 Instruction Manual, Orion Research, Inc., Cambridge, MA.
APPENDIX 1 Method o f Test Using a Gas-SensingElectrode 1. Application 1.1. This method is applicable to the determination of nitrate ion in water, within the interference constraints specified (See. 3). 1.2. Samples containing 0.5 to 100 mgllitre of nitrate-nitrogen can be analyzed by this method 2. Summary of Method 2.1. Nitrate ion is determined potentiometrically using an ammonia gas-sensing electrode after reduction of the nitrate to ammonia. The reduction is accomplished by use of titanous c h b ride reagent. 2.2. The electrode is calibrated in known nitrate solutions, and the concentrations of unknows are determined in solutions with the same background (containing potassium hydpxide and titanous chloride).
3. Interferences 3.1. Ammonia, if present in the sample, is measured with nitrate. If its concentration is significant, relative to the nitrate concentration, it must be nensured separately and substracted from the nitrate results.
153 4. pppamtus 4.1. pH meter, with expanded millivolt scale, or a specific ion meter. 4.2. Ammonia gas-sensing electrode. 4.3. Mixer, magnetic, with a TFE fktorocarbon-cmted stirring bur. 4.4. Loboratory glassware. 5. Reagents 5.1. Titanous chloride solution, 20%,practical grade. 5.2. Titanous chloride reagent. Reneutralize commercial TiCl, solution, using 10 M KOH,to a pH value of 0.25. Note: Preneuiralization generates hest and H , gas and should be cam'ed out carefully and in
-
a safety hood. 5.3. Potassium hydroxide solution, 10 N. Dissolve 561.1 g reagent grade KOH pellets and dilute to 1000 ml with distilled water. 5.4. Stock nitrate solution (I0 gllitre NO;-N). Dissolve 72.186 grams o f anhydrous potassium nitrate, KNO, ,and dilute to I000 ml with distilled water.
6. Galibration 6.1. Assemble and check the ammonia electmde according to the manufacturer's instructions. 6.2. Repare, by serial dilution o f the nitrate stock solution,nitrate standards at . I , I , 10 and I00 mg/L. ( I 00 ml volumetric j h k s are adequate for this purpose). 6.3. Transfer 80 ml o f 0.1 mg/L standmd into a 150 ml beaker, add a stirring bar and stir moderately with a magnetic stirrer. 6.4. Pipette I0 ml10 M KOH and 10 ml preneuiralized TiCl, reagent into the beaker. 6.5. Immerse the electrode in the stirring solution. Allow the potential to stabilize and record the value. This willbe plotted as 0.1 mg/L NO;-N. 6.6. Repeat steps 6.3-6.5 for each of the remaining stanabrds. Each potential value will be plotted as the nomino1 standard concentration. 6.7. Prepare a calibmtion curve by plotting, on semilogarithmicgmph paper, the potential observed vs. the concentration of the standard used. Note that volume corrections are incorporated into calibration, so that samples analyzed according to the procedure in section 7 can be read directly. 7. Procedure 7.1. Transfer 80 ml sample into a 150 ml beaker, add a stirring bar and stir moderately using
a magnetic stirrer. 7.2. Pipette 10 ml I0 M KOH and I0 ml preneutralized TiCl, reagent into the beaker. 7.3. Immerse the electrode in the stim'ng solution. Allow the potential to stabilize and record the
value. Sample concentration is read directly from the calibmtion curve generated in 6. Z Rinse the electrode with distilled water between samples.
This Page Intentionally Left Blank
155
STUDY OF MOLECULAR DIFFUSMTY INFLUENCE ON MASS TRANSFER RATE AT A WATER -ATMOSPHERE INTERFACE A. BALEIX, B. CAUSSADE and J. GEORGE I.M.F. Laboratoire Associk au C.N.R.S., 2 rue C. Chmichel, Toulouse, France J. MATHIEU, A. REYNES and L. TORRES
E.N.S.C.T. 118, route de Narbonne, Toulouse, France
ABSTRACT
In environmental Engineering, it is necessary to construct and improve mathematical models for predicting and controlling pollution, and therefore to study certain basic phenomena such as mass transfer at an interface. For gases of relatively low solubility in water, resistance to penetration occurs at the interface in the viscous sublayer of the liquid phase. To emphasize the importance of molecular diffusivity for mass transfer rate we have worked with wide range of molecular diffusivities using gaseslike oxygen and carbondioxideand, in contrast, helium. However, while it is relatively easy to measure oxygen and carbon dioxide concentrations in water, the determination of helium concentrations is very difficult. A solid-gas chromatography method has been used, and problems like pumping, storing ans injecting samples have been overcome. Absorption rates are shown to be strongly related t o Schmidt number and wind velocity.
1. INTRODUrnION
The phenomenon of absorption plays an important role in the maintenance of the natural balance in the environment: ocean and rivers being responsible for most of the air purification by absorption of excess substances such as carbon dioxide [ 1,2], sulfur dioxide, ammonia, radon, oils, pesticides ... Studies that have been carried out in laboratories as well as ‘in situ’ have shown that mass transfer is controlled by resistance on the gas side (volatile liquids) or on the liquid side (absorbed gases) of the interface depending on the value of the Schmidt number (Sc). Sc is defined as Sc = v/D where v is the kinematic viscosity of water and D the gas molecular diffusivity.
156
In the case of gas absorption (liquid side resistance, Sc % 1) the shape of the interface and thecharacteristics of the absorbed gas are particularly important. The shape of the interface is, affected by surface tension, presence of waves, wind created interfacial shear stress ( T ~ )and turbulence in the liquid phase. The effect of wind appears to be most important and there is a good correlation between the interfacial friction velocity USL (USL = ~ T S / P where L p~ is the liquid density) and the mass transfer coefficient KL. This is shown in Fig. 1 [3 -81. The nature of the absorbed gas is also of
r)
LlSS (1 973) (1978)
xo BROECKER loe3
JAHNE (1979) 0
n
L l S S (1979) MERLIVAT(1979) i.M.E(i 980)
$'
t
A
A
10-4
* A
0
Fig. 1. Double logarithmic plot of the transfer velocity KL against friction velocity in water USL.
prime importance and it is represented by the Sc value. Experiments as well as theoretical models (diffusive film, surface renewal ...) have shown that when the interface is stressed by a wind,KL appears to be roughly proportional to Scn where n is a number from - 0,67 to - 0,5 ; different values being quoted by various authors [S]. In order to elucidate the role of the Schmidt number in absorption we have used helium for which Sc is half of that for carbon dioxide (COz) and oxygen (0,) we had formerly been using. The present study was mainly carried out at the Institut de MBcanique des Fluides de Toulouse. In the case of experiments with helium the samples were analysed at the laboratoire de Chimie Analytique de 1'Ecole Nationale SupCrieure de Chimie de Toulouse.
157 2. EXPERIMENTAL FACILITIES
Absorption experiments were conducted using two gas liquid flumes of different sizes. The channel in which experiments with helium were conducted is shown in Fig 2 . It was
n
PUMP
I
HSRIFICE
Fig. 2. Schematic representation of the flume.
an' enclosed rectangular plexiglass transparent channel so that the conditions of the interface might be observed, its slope was m/m. The dimensions of the channel were: 0,l m high, 0,2 m wide and 12 m long. Liquid flowed along the bottom of the channel and gas flowed cocurrently. The gas sent through the test section was either air enriched with oxygen (up to 40% of 0,) or a mixture of air and carbon dioxide (up to 25%) or heluim (up to 80%). Gas was supplied from a recirculation loop by means of a blower and the flow rate of it was conrolled using a venturi meter. Water was also pumped from a recirculation loop and the flow rate of it controlled through a calibrated orifice plate. Because of high turbulence level in the liquid phase concentration profiles were homogeneous over each cross-section, and therefore the absorbed gases were analysed for by sampling at the bottom of the channel for the liquid phase,at the top of the channel for the gaseous phase. The mean and root mean square (R.M.S.) velocity profiles and the turbulent shear stress were measured directly using hot wire anemometers for the gaseous phase and laser doppler velocimeter for the liquid phase. The experimental parameters varied over the ranges: wind velocities, 4 to 16 m/ s; liquid velocities, 0,.2 to 0.5 m/s; and R.M.S. velocities of waves, 0 to 2.5 cm. There were no major difficulties in measuring C 0 2 or O2 concentrations. Selective BECKMAN probes were used to measure oxygen concentrations in water originally highly deficient in 0 2 .The injected C 0 2 was analysed for using an infrared analyser for the gaseous phase and, for the liquid phase a conductimeter (initial conductivity of deionized water: x < 3.10" 6 . 52-' cm-' ) [9].
158
The chromatographic method, believed to be original, used to determine helium concentrations, is presented below.
3. CHROMATOGRAPHICMETHOD
To perform the analysis, we utilized the assembling presented in Fig. 3. This one
Fig. 3. Experimental setup for operation on fluids and injection into the chromatograph.
included two parts which permit to titrate respectively helium dissolved in the water and helium in the gaseous phase.
3.1. Analysis of Liquid Samples
The special ampoule (1) containing the liquid sample was linked to the incoming circuit of the chromatograph by means of two needles (2) through a membrane. Volumes (20 to 50 11) of water were injected by means of the pump (3) using a valve (4) and a loop the volume of which was known precisely. Dissolved gases were first exhausted through a porous polymer (Porapak N) (5) and could be detected, Fig. 4. Under our experimental conditions the time of retention of water ranged from 15 to 20 minutes, which was long enough not to disturb the signal relating to helium and all other dissolved gases. Once helium was detected, the gas flow was detoured to the exit of the chromatograph through the valve (6); water was then prevented from running past thermal conductivity detector (7). As soon as the water was out normal flow patterns were restored. The signal for helium was a measure of the quantity of it dissolved in the water provided the volume was precisely known.
159
He
L 20
t (min)
Fig. 4. Chromatogramobtained by injection of 20 pl of water containing desolved air and helium.
3.2. Analysis of Gaseous Samples. Standardization.
A standardization was made using a gas loop presented on the left hand side of Fig 3. The loop was first exhausted using a pump (8). The vacuum was measured by means of the electronic gauge (9). The pump being separated, a quantity of helium (lo), for which the volume, pressure and temperature (therefore the number of molecules) were accurately determined, could be introduced into the loop through the valve (1 1). This process was repeated for different volumes of helium and the signals obtained were used as standards.
The concentration of helium in the gaseous phase was measured the same way, injecting a gaseous sample (10).
4. CONCLUSION
We would like to emphasize the fact that using our technique, which could still be improved it is possible to measure concentrations of helium down a few tens of parts per billon (p.p.b.) with a good reproducibility when in-situ sampling injecting problems are overcome. It should also be noted that an analysis of our results confirms a direct dependence of
KL on the waves and on the interfacial shear stress. It also shows (see Fig. 5) that the nature of the gas is very important and that KL increases with gas diffusivity, decreasing when the Schmidt number increases. We cannot, however, draw any conclusion from our preliminary results concerning the power n of the Schmidt number.
160
0
0 0 0
a
0 0
t
* 1
1
0.00 5
0.010
0.0 15
USL ( m / s ) Fig. 5. Mass transfer velocity KL against friction velocity in water USL. (* : CO, ,o : He).
REFERENCES 1 L. Merlivat et L. Memery, Gas exchange across an air-water interface: Experimental results and modeling of bubble contribution to transfer. Journal of Geophysical Research, 1983, Vol. 88, n"C1, pp. 707-724. 2 L. Merlivat, Le probleme de l'environnement i 1'6chelle planhire: une composante oceanique du cycle du CO, . Suppl6ment au no 52 du Courrier du CNRS. Images de l'Environnement, 1983, pp. 21-25. 3 P. S . Liss, Processes of gas exchange across an air-water interface. Deapsea research 20, 1973, pp. 221-238. 4 H. C. Broecker, J. Petermann and W. Siems, The influence of wind on CO, exchange in a wind-wave tunnel, including the effects of monolayers. J. Marine Res. 36, 1978, pp. 595-610. 5 B. Jiihne, K. 0. Munich and U. Siegenthaler, Measurements of gas exchange and momentum transfer in a circular wind-water tunnel. Tellus 31,1979, pp. 321-329. 6 L. Merlivat, Study of gas exchange in a wing tunnel. Preliminary results. Symposium on capillary wavesand gas exchange (Hambouig), 1979. 7 L. Hasse, P. S. Liss, Gas exchange across the air-sea interface. Symposium on capillary waves and gas exchange (Hambourg), 1979. 8 L. Aisa, B. Caussade, J. George et L. Masbernat, Echanges de gas dissous en koulements stratifies de gaz e t d e liquide, Int. J. of Heat and Mass Transfer, 1981, Vol. 24, n06, pp. 1005-1018. 9 L. Aisa, These de Docteur-Inghieur, Institut National Polytechnique de Toulouse, 1978.
161
ABUNDANCE OF NONIONIC SURFACTANTS IN ISRAEL MUNICIPAL SEWAGE
U. ZOLLER
Division of Chemical Studies University of Haifa-Omnim, P.O.Kiryat Tivon 36 910, Israel
ABSTRACT ‘Hard’ (nonbiodegradable) nonionic surfactants are currently the most commonly used nonionics in Israel, which probably has no parallel in other western industrial countries. These synthetic detergents constitute a significant factor of the municipal sewage profile and, hence, determine both the environmental consequences and the possibility of the sewage water’s reuse after appropriate purification processes. A study of mapping Israel municipal sewage in respect to the distribution and content of nonionic surfactants in them has been undertaken, and determinations of the nonionic surfactants in different municipal sewage systems of all parts of the country have been conducted. Modified chemical methods based on the SDACTAS procedures and calibration curves specifically prepared - in the presence of various representative concentrations of anionic detergents - have been used for these determinations. Typical concentrations of nonionic surfactants in Israel municipal sewage have been found to be within the range of 1.5-3.5 mg/liter. The ratio between bio- and nonbiodegradable nonionic surfactants in the checked sewages has also been monitored.
1. INTRODUCTION
Alcohol ethoxylates (AE) and alkylphenol ethoxylates (APE) are the two largest classes of nonionic surfactants in current use [ 1-21. Furthermore, nonionic surfactants, led by synthetic 12-18-carbon alcohol ethoxylates are expected to make the biggest gains in the market in the next 3-5 years [ 11. Primary alcohol ethoxylates (F‘AE), the fastet growing major surfactant, contain alkyl groups which are essentially linear - and, therefore, are biodegradable [3], while most commercial alkylphenol ethoxylates contain hghly branched alkyl groups and, consequently, are considered to be nonbiodegradable [4]. It turns out, that ‘hard’ (nonbiodegradable) nonionic surfactants of the APE type are currently the most commonly used nonionics in Israel [5], which probably has no parallel in other western industrial countries. In the latter, PAE are the largest volume nonionic in household detergent-based formulations. In the industrial sector, APE are the major
162 surfactants. In Israel, however, APE is dominant in both sectors. These nonbiodegradable nonionic surfactants, ultimately reach the municipal sewage systems causing serious foaming problems and also interfere with sewage treatment processes [6]. They are also responsible for the various aspects of short- and long-term environmental problems involved in their use. The increase of environmental concerns worldwide, as well as the urgent need to use more and more reclaimed water from various sources, has made the monitoring of wastewaters with respect to their detergent content an imperative. This refers not only to the primary biodegradation of detergents (i.e., the loss of a measurable physical or chemical property of the intact surfactant when exposed to microbial attack [7], but also to the ultimate biodegradation (defined as microbial attack of an organic substrate to produce carbon dioxide and water) - on which recent studies have focused [8]. It was clearly shown either in prior laboratory tests or in field studies that APE resists both primary and ultimate biodegradation much more than PAE under a variety of conditions [4, 81. Whatever the case may be, no extensive monitoring of nonionic detergents in sewage effluents under ‘in vivo’ conditions has been reported, particularly because the determination of nonionic surfactants in the presence of a complex mixture of other detergents present in sewage has been unavailable until rather recently [4, 91. Our study of the current abundance distribution, and content of nonionic surfactants in Israel municipal sewage has been initiated in view of the following: a. synthetic detergents - including nonionic surfactants - constitute a significant factor of the municipal wastewater profie, b. in Israel, which is a model of efficient water utilization [lo], the reuse of reclaimed wastewater is probably the most effective response to the scarcity of water resources and the overexploitation of existing resources in the country [ 1I], c. the content of nonionic surfactants - nonbiodegradable in particular - in municipal sewage determines both the possibility of the wastewater reuse after appropriate processes and the long-term environmental consequences, d. the chemical monitoring of nonionic detergents under real conditions is an extremely complicated task [4], and e. excluding one earlier pilot study in the subject [12], the on-going abundance and distribution of nonionic surfactants in Israel municipals sewage have been as yet unknown.
2. OBJECTIVES
The main objectives of our study were twofold: a. To map Israel municipal sewage (influents) in respect to their content of nonionic surfactants. b. To establish - to the first approximation - the ratio between bio- and nonbiodegradable nonionic surfactants in the checked sewages (basically, the initial ratio between straight- and branched-chain nonionic ethoxylates used in Israel).
163 3. EXPERIMENTAL PROCEDURES AND METHODOLOGY
Some typical ‘representative’ municipal sewage ‘systems’ or ‘stations’ distributed over all parts of Israel (except the far south), have been selected. Wastewater of the influents
of the selected sewage works were sampled and brought to the laboratory for nonionic analysis. Formaldehyde was used as a preservative in those cases when the determination was not carried out immediately following the sampling. A modified version of the SDA-CTAS procedure of nonionic surfactants determination [ 131 was used [9]. Thus, weighted calibration curves (Absorbance versus concentration) specifically prepared - based on a mixture of the nonionic surfactants most commonly used in Israel - have been used. These calibration curves were prepared in the presence of various concentrations of anionic detergent (primarily linear alkylbenzene sulfonate LABS) typically present in Israeli wastewater systems. Once the concentration of the anionic surfactants in the sample checked has been determined (using the conventional MBAS method [ 14]), the appropriate calibration curve for the determination of the nonionic surfactant in the sample could be selected. The same procedure was used for the determination of the ratio betweeb the bio- and nonbiodegradable nonionic surfactants in the wastewater samples collected. The content of the nonionic surfactants in these nonpreserved samples was determined immediately after sampling, and then - 24 hrs., 48 hrs., and one week later, respectively. The ratio between the bio- and the nonbiodegradable nonionic surfactants could be thus established.
4. SELECTED RESULTS AND DISCUSSION
The three calibration curves prepared and used in this study are shown in Figure 1 below. The procedure used for their preparation was described elsewhere [9]. The three curves shown correspond to anionic surfactant concentrations of 5(I), 1 q I I ) and lS(II1) mg/liter. These concentrations correspond to anionics-nonionics ratio within the range of about 2: 1-6: 1 respectively. Based on previous experience [9], each determination obtained based on the appropriate weighted curve should be corrected by a factor of 1.093 (i.e. the concentration reading in the calibration curve should be multiplied by the above factor). This is so, since the sublation method used to extract the nonionic surfactants from the sewage sample [13], gives at the average, only 91.5% of the nonionics actually present. Representing results of nonionic surfactants’ concentrations in the influents of some selected municipal sewage systems in Israel are given in Table 1 , below. One can see that contemporary concentrations of nonionic surfactants in Israel municipal wastewaters are within the range of 2-3.5 mg/liter. These concentrations are far below those found in wastewater in Israel in a pilot study agout five years ago (10-15 mg/liter) [ 121. Indeed, the current low concentrations found are rather surprising in view of the continuing world trend of using more and more nonionic surfactants at the expense of anionic detergents [15]. However, one should keep in mind, that simultaneously with the above trend, the straight-chain nonionic surfactants (i.e. PAE and closely-related synthetic detergents) take the largest share in this growth [ l , 2, 81, at the expense of the nonbiodegradable branched chain APE. Apparently, this is what we find in Israel too.
164
I
Nonionics [ m g / L l Fig. 1. Calibration curves for the determination of Nonionic Surfactants (Absorbance versus Concentration). The nonionic mixture consisted of : Nonylphenol ethoxylate containing an average number of 14, 12, 10 and 8 units of ethylene oxide respectively: (NPI4) - 15%; NP,, - 25%; NP,, - 4076, and NP, - 20%. (i) a 1: 1 ratio between the anionic detergent (LABS) and the nonionic surfactants. (ii) a 2 : 1 ratio. (iii) a 3: 1 ratio. Tab. 1. Total Concentration of Nonionic Surfactants in Israel Municipal Sewages Location of Sewage Influent
Nonionic Surfactants mg/liter
Tiberias (upper) Tiberias (lower) Nahariya Acre Haifa Nazereth Oranim Netanya Hadera Tel Aviv (north) Tel Aviv (south) b
3.2 2.2 1.8 1.7 1.8 2.7 0.85 a 3.2 1.6 2.6 3.5
a A very s m d college community. Including Jafa, Holon, Bat-Yam.
165
Nonetheless, the results given in Table 1 were obtained during winter time in Israel. We are now in the midst of a second round of sample collections and determinations. The results of this second round being conducted in spring time (no rainfall which probably is responsible for some ‘dilution effect’), should be delineated with the concentrations obtained in the first round. Finally, preliminary results concenrning the ratio between bio- and nonbiodegradable nonionicmfbctantsinthe c h e c k e d w s showedit to be within the range of 2-3 :8-7. l h latter result should be considered only as a ‘first approximation’, since in this stage it is still based on a small number of determinations. Any attempt to generalize, to extrapolate, or to draw any meaningful and reliable conclusions concerning this ratio, should wait for the accumulation of more data currently being collected and assessed. At any rate, the ratio between bio- and nonbiodegradable nonionic surfactants provides valuable data from which the real scope of the problem of nonionic detergents in Israel municipal wastewater can be estimated including all the implications involved.
5. SUMMARY
The use of chemistry for wastewater monitoring proved to be valid and reliable even under ‘in vivo’, uncontrolled continuously changing field conditions that exist in sewage systems. A successful application requires several modifications according to the specific set of the local constraints and the particular procedure should be worked out locally. It appears that the concentrations of nonionic surfactants in municipal sewage in Israel are decreasing. This accords with the current world trend in detergent consumption and use but somewhat unexpected in view of the unique situation in Israel in this respect. In view of the expected increase in the use of reclaimed sewage water particularly for unlimited irrigation purposes, the presence of nonionic surfactants in municipal wastewater has its long-term environmental trade offs - as a result of cumulative processes of past and present man-made operations. The monitoring of wastewaters with respect to their detergent content as a basis for decision-making for action is clearly an imperative.
ACKNOWLEDGEMENT
This project was supported by the EPA of Israel, the Office of the Interior.
REFERENCES
1 2 3 4 5 6
P. L. Layman, C & Eng. News, Jan. 11 (1982) 13-16. J . Amer. OilChem SOC.,59 (1982) 554A-557A. R. D. Swisher, Surfactant Biodegradation, Marcel Dekker, 1970. L. Kravetz, J . Amer. Oil Chem SOC.,58 (1981) 58A-65A. U. Zoller and R. Romano, J. Amer. OilChem. SOL,60 (1983) (to be published). N. Narkis and B. Ben David, Research No. 013-445, The Technion R & D Foundation Ltd., and Water Commissioner’sOffice, Israel, 1980.
166 7 0.E. C. D., Determinations of the Biodegradability of Surfactants used in Synthetic Detergents, 1976. Paris, France. 8 L. Kravetz, H. Chung, K. F. Guin,W. T. Shebs, and L. S. Smith, A paper presented at the 1982 Annual Meeting of the Soap & Detergent Assoc., Boca Raton, Florida, January, 1982. 9 U. Zoller and R. Romano,Environ. Inter., 9 (1983) 55-61. 10 S. Arlosoroff, A paper presented in the United Nations Water Conference, Mar del Plata, Argentina, 1977. 11 E. Idelovitch, Munic. Wastewater Reuse News, 49 (1981) 12-21. 12 N. Narkis and S. Henefeld-Furie, Water & Sewage Works, March (1977), 69-71. 13 S. L. Boyer, K. R. Guin, R. M. Kelley, M. L. Mausner, H. F. Robinson, T. M. Schmitt, C. R. Stahl, and E. A. Stezkorn, Environ. Sci.& Tech., 11 (1977) 1167-1171. 14 B. M. Milwidsky, Practical Detergent Analysis, McNairDorland, New York, 1970. 15 J. Amer. Oil Chem SOC.,58 (1981) 837A-838A.
CHAPTER I11
PHYSICO-CHEMICALTREATMENT OF SUSPENSIONS
This Page Intentionally Left Blank
169
INTERACTION OF TEMPERATURE, ALKALINITY AND ALUM DOSE BY COAGULATION OF HUMIC WATER
P. DOLEJS
Institute of Ecology, Czechoslovak Academy of Sciences, Na sbdkcich 7,370 05 Ceski Budejovice, Czechoslovakia
ABSTRACT The results of pilot-plant studies about coagulation of naturally coloured humic water are presented.The aim of the work is to investigate mutual interaction between water temperature, alkalinity and alum dose. Their influence upon the residual concentration of A13+, residual colour, residual COD and particle size distribution was experimentally studied. The dose of coagulant can be optimised in relations to the criteria investigated or to the methods of separation. The results indicate that there is not one 'universal' pH for treating the humic water with the same composition of organics, nor is there any simple equivalence between changes of alkalinity and dose of coagulant. A certain optimal combination of alum dose and alkalinity was found to be different for every criterion studied. It also changes with the temperature of water and the effect of temperature is not the same for each criterion studied.
1. INTRODUCTION
Humic substances are in many cases the main polluting material in raw water, which has to be treated to provide drinking water. Theoretical studies of water purification in the last 10-20 years (often confirmed by model experiments), show great potential for controlling the process of water purification. They were mostly obtained for turbid waters, where model pollutants e.g. kaolin, bentonite or silica were used. The mathematical model of separation of these substances can be complicated in view of commonly considered fact of today, that the characteristics of humic substances differ according to the place, origin and time [ 11. For studying the influence of temperature on coagulation-aggregation of humic water an experimental approach was chosen. To explain the influence of temperature and to gain knowledge about the mutual interaction between singular parameters, a complete experimental study of the whole problem is necessary. This is a similar approach to that of T. Ishibashi [2] who said that
170
'the theory of coagulation has been developed hypothetically rather than experimentally'. The author presents here the results of an experimental investigation of interacttion between water temperature, initial alkalinity and alum dose.
2. EXPERIMENTAL PROCEDURES
2.1. Pilot-plant
The experiments were performed in a pilot-plant, the esential parts of which were rapid and slow mixing tanks. The diameter of the rapid mixing tank was 90 mm and with 1 l/min water flow rate; the theoretical residence time being 56 s. The slow mixing tank diameter was 140 mm and the residence time for these experiments was 905 s. The mean turbulent velocity gradient (G) was kept constant in both rapid and slow mixing tanks through the whole study. To highlight the effects of the criteria investigated, the G values were kept relatively low. G of rapid and slow mixing was 95 s-' and 18.5 s-' respectively. For the G calibration the strain gauge toryuemeter was used. A homogeniser was placed in the pipe just in front of the inflow into the rapid mixing and just behind the dosage of coagulant. The homogeniser consists of three holed plates and its residence time is about one second. The detailed scheme of this pilot-plant is published in reference 3.
2.2. Model Humic Water
A stock of humic concentrate was prepared by extraction of humic substances from peat by demineralised water. Alkaline extraction was not used because a better simulation of this process occured in nature. The extraction with e.g. NaOH gives much better quantitative results but the resulting composition of organic matter is more or less different from that in natural waters. Tap water was diluted four times with demineralised water and the humic stock solution was added to get the colour equivalent to 50OPt. The total concentration of Ca2' and Mg2' ions was 0.5 mmol/l and the COD (Mn) of the model water was 7.5 mg/l. The desired alkalinity of model water was adjusted by NaHC03 or HCl. After the addition of reagent the water was kept overnght to reach equilibrium.
2.3. Methods and Investjgated Criteria
The experiments were planned according to a three factor composite design. From the experimental results, coefficients of the second order polynomial (Eq. 1) were evaluated, describing the mutual relations of three variables (temperature, alkalinity, alum dose):
171 The influence of the temperature was studied in the range of 1.7°C-16.30C, the alum dose between 16-52 mg/l and alkalinity between 0.29-0.91 mmol/l. The residual COD(Mn), residual colour, the particle size distribution after mixing and ‘the coefficient of the efficiency of aggregation (a)’were estimated to evaluate the influence of the three mriables mentioned above. The last criterion (a)was developed by Hereit, Mult and Vigner [4, 51, It is defined as:
where: Co - the total concentration of A1* (mg/l) doses Cf - the concentration of A1* in the supernatant after centrifugation at defined conditions (time, g value) (rngll). This criterion is based on the assumption that separable aggregates of water impurities and alum have a certain sedimentation velocity. Separation of such particles in waterworks is achieved by filtration, which is used either as the only separating operation or the final separating operation for the smallest particles. The destabilised and aggregated particles with diameter greater than about 0.5 pm can be separated by centrifugation. Not destabilised and therefore non-aggregated and not separable particles, organic molecules and residual A1* remain in the supernatant. The coefficient of the efficiency of aggregation a may be determined at every step and every phase of the treatment process after coagulant injection. It can provide valuable information about the level of separability of coagulant and about the possible residual concentration of A1* (or Few) after separation. In laboratory and pilotplant experiments the samples were centrifugated for 5 minutes at 2,000 g. The sampling point was located at a depth of 4 cm from the level of the liquid in the slow mixing tank. Particle size distribution was determined from sedimentation characteristics of flocs suspension. This was derived from the decrease of total A1* concentration during sedimentation of flocs. Particles which disappear from the 4 cm depth after 5 minutes of sedimentation are called macroparticles. Their sedimentation velocity is greater then 0.13 mm/s. Such particles can be efficiently separated by sedimentation in waterworks. Microparticles can be separated in a clarifier and their sedimentation velocity is greater than 0.01 1 mm/s and their diameter is of the order of 0.1 mm. They disappear from the sampling point depth by sedimentation in the interval of 5-60 min. Primary particles are the smallest ones and in waterworks are separated by sand filtration, Their sedimentation velocity is about mm/s and their diameter is of lo-’ mm order. They are separated from the sample taken in 4 cm depth after 60 minutes sedimentation by centrifugation. In the supernatant of this sample the non-aggregated A1* fraction, residual COD(Mn) and residual colour were determined. The concentration of A13+ in samples was determined spectrophotometrically with ‘aluminon’. COD(Mn) was determined by oxidation of 100 ml sample with 20 ml of 0 . 0 2 M KMn04 and 5 ml of 30% H2S04boiling the mixture in water bath for 20 mi-
172
nutes. This ‘old’ procedure was considered to be good enough for low concentrations of humic substances in very similar samples to be determined. The alkalinity is expressed as a concentration of HCO; in mmol/l. It was estimated by titration with 0.1 M HCl to pH 4.3.
3. RESULTS AND DISCUSSION
Fig. 1-3 show the influence of initial alkalinity and alum dose for temperatures of 3, 9, 1SoC on the three most important criteria investigated in jar tests or pilot-plant studies, e.g. - portion of macroparticles formed after rapid and slow mixing - residual A13+(expressed as non-aggregated fraction) - residual COD(Mn).
Fig. 1. Influence of initial alkalinity and alum dose on the macroparticles fraction, non-aggregated fraction and COD. T = 3°C
Fig. 2. Influence of initial alkalinity and alum dose on the macroparticles fraction, non-aggregated fraction and COD. T = 9°C
173
0,30 0,25
0,20 0,15 0,lO
0,m Of0
26 c
----COO(Mn) lmg/l 1 ALUM DOSE Ima/t 1
32
9
% !
Fig. 3. Influence of initial alkalinity and alum dose on the macroparticles fraction, non-aggregated fraction and COD.T = 15°C
Presentation of more than these three variables into the figure would be confusing. In these figures, the dotted isolines represent values of the non-aggregated fraction, the dashed isolines represent the residual COD(Mn) and full lines represent the macroparticles fraction value. Dotted, dashed and full small rings indicate the centre of corresponding elipsoids. Comparing the macroparticle fractions in these figures we see that for temperatures 9OC and 15OC a maximum of macroparticles is formed at high alum doses and at high alkalinities, out of the investigation range. 'Cold' water (3OC) has its optimal combination of alkalinity and alum dose for macroparticles within the investigation range. In all three figures, we see changes of optimal pH (see Fig. 5 for pH values) and of optimal alum dose (influencing formation of macroparticles) for various initial alkalinities. Optimal pH is at low alkalinities lower than at hgher alkalinities. E.g. for a temperature of 3OC is at initial alkalinity 0.35 optimal dose for max. of macroparticles 32 final pH 5.5 ,, 0.60 41 ,, 5.8 ,, 0.85 >, 48 ,, 6.0 9,
For residual alkalinity see Fig. 4. The residual A13+ concentration is also influenced by temperature. By drop of temperature optimal dose of coagulant decreases, optimal initial alkalinity increases and final pH increases too. The most important reason for this result may be the dependence of the rate of hydrolysis of coagulant on temperature and its possible compensation by increase of concentration of HCO,.The residual COD(Mn) has the most efficient combination of initial alkalinity and coagulant dose at temp. 9, 9,
3OC is optimal alkalinity 0.37 optimal dose 29 final pH 5.6 residual COD 1.35 1.25 35 5.9 9OC 7) 0.57 ,, 6.1 ,, 1.05 15°C ,, 0.80 ,, 41 ,, 3)
9,
1 74
I 32
26
I 41
37 43 40 ALUM DOSE [mq/il
54
Fig. 4. Effect of alum dose and initial alkalinity on residual alkalinity. T = 3°C
non-aggregated fraction -x - _
0 -COD
v - colour
c
o-colour (,,~=12)
5
.
- a after rapid mixing
M - macroparticles
c L
E ,-
1.3-
>.I
t
.-C
25
35
45
55
6
A l u m dose [mq/~]
Fig. 5. Positions of optima points of some dependent variables for temperatures 1.7; 3;6;9 ;12;
15; 16.3’C
The relations of optima of some most important dependent variables for various temperatures in co-ordinates alkalinity - alum dose in these experiments are presented in Fig. 5. With rising temperature there is an apparent shift of optima of factors describing organics removal (COD and colour) to hgher alum doses, to higher initial alkalities and to higher pH values. Similar is the tendency of the shift of optima of macroparticles formation. Quite another is the trend relating residuals of dosed coagulant e.g. non-aggregated fraction and the coefficient of the efficiency of aggregation a. With rising temperature the optimal dose is increasing but the initial alkalinity and resulting final pH is decreasing. Comparison of the influence of water temperature, alkalinity and alum dose on three investigated criteria (macroparticles, non-aggregated fraction and residual COD(Mn)) shows, that there is not a ‘universal’ pH or optimum dose of coagulant or initial alkalinity for these criteria in the coagulation-flocculation process.
175
Fig. 6. Influence of initial alkalinity and alum dose on the coefficient of the efficiency of aggregation (ol). T = 3OC
Fig. 6 illustrates the influence of initial alkalinity and alum dose on efficiency coefficient a!. The dot and dash line indicates the change of position of maximal local values of a! with change of initial alkalinity and alum dose. The point at this line indicates the absolute maximum of a for the given temperature (compare with Fig. 5). For a description of the influences of alkalinity and alum dose on the particle-size distribution, there are three three-dimensional schemes, Fig. 7-9. They are drawn for temperatures 3,9, and 15OC. A strong influence of temperature on formation of macroparticles especially in the range of higher alkalinities is evident. The residual A13+concentration is little influenced by temperature in the vicinity of 'optimal' combinations of alkalinity and alum dose. The author's opinion as a result of these experiments (accompanied by series of experiments with other independent variables published previously [3] or still unpublished [9]) is that the coagulation-flocculation of humic substances is more complex than that T = 3OC
MACROPARTICLES MICROPARTICLES
Fig. 7. Influence of initialalkalinity and alum dose on the particle size distribution. T = 3°C
176
Fig. 8. Influence of initial alkalinity and alum dose on the particle size distribution T = 9°C
T = 1SoC
PRIMARY FARTICLES NON-AGGREGATE0 FRACTION
Fig. 9. Influence of initial alkalinity and alum dose on the particle size distribution. T = 15°C
of turbid waters. In coagulation of humic waters there occur probably more reaction steps giving rise to more possibilities of affecting than in the processes of turbidity removal. Several hypotheses about the influence of some variables upon the humic water purification processes can be formulated: - Hydrolysis of alum produces various species with various efficiencies on various fractions of humic substances - Temperature and pH influence the shape and size and reaction properties of humic substances changing also their coagulation behavior - Collision efficiency factor [ 6 ] changes during the coagulation process (similar ,to [6, 71) because some parts of molecules of humic substances are more effective for destabilization (or bridging) and other parts of the same molecule are not so effective (for chemical and sterical reasons). The drop of concentration of places with high effectivity is faster than that of places with low effectivity. It can result in the decrease of mean efficiency of collisions during the coagulation pr2cess. The optimum of COD(Mn) removal is at 3 C at a relatively low dose of coagulant
177 which can be explained with decrease of ionization of humic molecules. Better separation is also enabled by low pH value and by low initial alkalinity. The alkalinity at 3OC plays an essential role in the kinetics of hydrolysis of alum and for formation of separable particles. This explains the fact, that for low temperatures minimum of residual A13+ needs higher initial alkalinity rather that would be necessary for optimal COD(Mn) removal. The shift of optima of non-aggregated fraction and a with temperature (Fig. 5) can be explained by the same way with kinetics of reaction of alum with HCO, ions. The macroparticle fraction is most influenced by temperature. Not so much influenced by temperature is the residual A13+ concentration. It is possible that the temperature is not the parameter influencing only perikinetic phase [6], but by the treatment of humic waters it can also affect the collision efficiency factor. At higher temperatures (Fig. 3) in the range of h g h doses and h g h alkalinities (e.g. 50 mg/l and 0.9 mmol/l) such primary particles are formed, which can aggregate with better efficiency than in other parts of the range investigated. These primary particles form after flocculation hgher percentage of large flocs, but they do not cause higher removal of COD(Mn) or residual A13+. The most probable explanation of this finding, according to Stenquist and Kaufman [8], is an inadequacy of initial mixing with hgher alkalinities which detrimentally affects the process performance. Experiments with independent variables such as temperature, the C value of rapid mixing and alkalinity, confirmed the adequacy of homogenisation and of initial mixing in the pilot-plant used. A strong effect of temperature only was demonstrated even at alkalinities above 1 mmol/l. The resulting particle size distribution and residual organics were influenced by the C value of rapid mixer in the whole range 1.7-16.3OC [9].
4. CONCLUDING REMARKS
By treatment of humic water with certain alkalinity and composition of organics and by using only alum as coagulant, it was found that there are different optimal alum doses for each of the criteria investigated. This can be explained by taking into consideration the fact that any criterion investigated has its own mechanism and kinetics. The process of coagulation of humic water can be further optimised if there is a possibility of controlling the alkalinity of raw water by acid or alkali addition. It is apparent from this series of experiments that each of the studied criteria has different optimal combination of alkalinity and alum dose, resulting in various final pH values. The macroparticles formation is strongly influenced by temperature. Temperature decrease causes a decrease of the optimal dose of coagulant (for the fsllowing criteria COD(Mn), residual A13+ and the fraction of macroparticles). This formerly published dependence [3] is in force over the whole range of alkalinities as shown here. The results presented in t h i s study were verified by comparison with experiments using natural water from the Vltava River in Prague, The trends of the influences of the variables investigated were found to be in good agreement with that obtained using model humic water.
178 ACKNOWLEDGEMENT
This work has been experimentally carried out at Prague Institute of Chemical Technology, Dept. of Water Technology and Environmental Engineering at the pilot-plant built by author. I want express my thanks to Ing. N. Kalouskovi for drawing the figures in the final form.
REFERENCES 1 V. Pennanen, Hydrobiologia, 86 (1982) 73-80 2 T. Ishibashi, Jour.AWA,72 (1980) 514-518 3 P. Dolejs, in J. A. Oleskiewicz, J. Przewlocki and T. Winnicki (Eds.), Environmental Protection Engineering, Tech. Univ. Wroclaw, in print 4 F. Hereit, S. Mutl and V. Vkner, Vodnihospodbstvf(Water management), 27 (1977) Ser. B 80-86 5 F. Hereit, S. Mutl and V. V&ner, Aqua, No. 5 (1980) 95-99 6 H. H. Hahn and W. Stumm, Advances in Chemistry Series, No. 79, ‘Adsorption from Aqueous Solution’, American Chemical Society, Cambridge, Mass., 1968, 91-1 11 7 N. Tambo and Y.Watanabe, Water Res., 13 (1979) 429-439 8 R. J. Stenquist and W. Kaufman, Initial Mixing in Coagulation Processes, U.S. EPA, Washington, D.C., 1972, 91 9 P. Dolejs, Interaction between Temperature and Operational Variables by Treatment of Humic Water. Ph.D. Thesis, Prague Institute of Chemical Technology, 1980, 1-103
179
MAGNETIC MICROPARTICLES FOR TREATMENT OF NATURAL WATERS AND WASTEWATERS
D. R. DIXON and L. 0. KOLARIK
Division of Chemical and Wood Technology, CSIRO,Bayview Avenue, Clayton, Vic. 31 68, Australia
ABSTRACT This article describes a new approach to the treatment of natural waters and wastewaters. Details of the steps involved and of the progress of the method from laboratory experiments to full-scale commercial plants are combined with the results of fundamental colloid science studies. Also other possible practical applications for this novel process are suggested.
1. INSOLUBLE MICROPARTICLES IN WATER TREATMENT
Clarification of low-turbidity, coloured waters may be difficult. Due to the small number of particles, the rate of floc formation in these systems is slow and the settling characteristics are poor. To improve the process efficiency other solids, are sometimes added [ 13. The role of the added solid-phase is described as twofold: (a) the added solids may influence the rate of floc formation, which is called the "nucleation effect", and (b) they may improve the settling characteristics of the formed aggregates which is called the "weighting effect". In practice this can be achieved by recirculating the settled sludge from the previously treated water or by adding a fresh solid-phase such as insoluble microparticles. Only the latter will be discussed here. The solids frequently used include frnely divided sand, activated silica, bentonite clays and activated carbon [ l , 21. Demeter et al. [3] have proposed the use of sand particles ranging in size from 10-200 pm in conjunction with polyelectrolyte or inorganic coagulants, or both. Generally these materials alone do not promote the destabilization of colloidal dispersions to any appreciable degree. Their handling, recovery and reuse are difficult. Hydrocyclones have been used to separate sand form attached impurities prior to recycling and reuse.
180 2. MAGNETIC MICROPARTICLES
With added magnetic particles it is possible to exploit the rapid kinetics and the weighting action, as well as the magnetic properties of the microparticles themselves, by applying a magnetic field to effect separation of the particles from the liquid phase. A search of the available patent literature over the past forty years shows that in 1941 Urbain and Stemen [4] proposed the use of magnetite in a water clarification process, to aid sedimentation. Magnets placed at the base of a settler were intended to increase the settling rate of the agglomerates. Since then a number of patents have been granted covering the use of magnetite and other magnetic particles for the removal of various waterborne impurities [5-111. The magnetic particles were generally used in conjunction with inorganic coagulants or organic flocculants. A range of magnetic devices was used for the separation of the particles with attached impurities from the water. No activation of the magnetic material was suggested, illustrating that the earlier use of magnetite involved exploitation of its magnetic properties only. Magnetic particles have also found use a few special applications. Magnetic ionexchange resins have been successfully employed in desalination [ 121 and dealkalization [ 131. Magnetic carbon was prepared and used for the removal of unwanted organic substances from a food processing stream [14-161. Processes for controlling surface pollutants, e.g., oil in water, were proposed by Weiss and Battaerd [7]; MitcheU and Chet [18] developed a magnetic separation method for recovering protein from single cell organisms. The latter suggested the use of alkali (2M NaOH) for the recovery of protein, but apparently not for the activation of magnetite. Magnetic particles have also been used as substrates for the immobilization of various inorganic and organic compounds. Nonporous magnetic materials were used as enzyme supports [ 191, e.g., chymotrypsin was successfully attached to precipitated magnetic particles including magnetite. At present there are two promising water treatment processes utilizing magnetic microparticles: high gradient magnetic separation (HGMS) and the technique featured in this paper. Oneexample of a HGMS process is high gradient magnetic filtration (HGMF) which is based on the Kolm type separator [20]. The principles and applications of magnetic fitration are further discussed by Oberteuffer [21]. In HGMF, magnetically susceptible particles are collected on steel wool or on an expanded metal matrix. The background magnetic field magnetises the matrix and produces strong magnetic gradients that converge on the matrix fibrous strands. The method has been used in the mineral benefication of semitaconite [22] and for the removal of sulphur from coal [23]. Bitton et al. [24] showed that phosphate concentration in water could be significantly reduced, using magnetite in conjunction with aluminium sulphate and montmorillonite clay. In 1976, de Latour [25] described the use of the HGMS technique in the water and wastewater field. Magnetite particles were used in conjunction with alum to remove a number of impurities including colour, turbidity, bacteria and phosphate. The magnetite acted as a ”seeding material” in such interactions. But the particles were not reused. The advantages of the HGMS method are its simplicity and very high-fdtration rates.
181 2.1. Reusable Magnetic Particles
The concept of reusable magnetic particles was first applied in ionexchange to overcome the difficulties associated with the handling of very fine particles [ 121. This approach also led to the development of a dealkalisation process for non-clarified wastewater [13] and to the use of non-functional magnetic polymers as filter aids [26]. In 1975-1976 Bolto et al. summarized the development and use of magnetic polymer-coated particles in various water treatment processes [27, 281. The potential of magnetic ionexchange particles for the removal of colour and turbidity from water was first studied using a magnetic cation exchange resin [29]. A further development was the use of positively charged magnetic ”whisker resins” [30]. Deposition of an amphoteric, amorphous Fe(OH)3 gel coating onto magnetite particles was also investigated [3 1]. Although such coated magnetite particles showed excellent coagulationadsorption properties, economic or technical difficulties (i.e., mechanical stability, and fouling) have prevented their commercial applications.
3. THE”SIROFL0C” PROCESS*
The use of alkali-treated magnetite in water purification processes was described for the first time in 1977 [32]. It was shown that treatment of fine particles (1-10 pm) of magnetite with a dilute (0.1 M) sodium hydroxide solution imparted excellent coagulation-adsorption characteristics to the particles which can be described as a solid reusable coagulant - adsorbent. The properties of these materials derive from the amphoteric behaviour of the hydroxyl groups present at the oxide surface [33,34]. This process has been patented [35,36]. Alkali-treated magnetite interacts strongly with turbidity particles and the colour substances present in natural waters. Depending on the initial colour and turbidity values, such particles can be used alone as a solid coagulant-adsorbent or in conjunction with another primary coagulant. In the latter case significant savings of the coagulant can be achieved. Both inorganic coagulants (alum or ferric chloride) and organic synthetic polyelectrolytes may be used successfully. Water is treated in a fraction of the time normally needed. The magnetite particles used in the coagulation-adsorption steps are demagnetised. To separate magnetite from the water the particles are magnetized. The material is reused after reactivation with alkali which releases the attached impurities. A relatively small volume of a fmal alkaline effluent is produced. The block diagram of the process is shown in Fig. 1. There are two main stages in the process; (i) removal of impurities from the feedwater, which consists of three steps: PREMIX, COAGULANT/FLOCCULANT ADDITION AND AFTERMIX and SEPARATION, and (ii) regeneration prior to reuse of the particles, which includes ALKALI REACTIVATION and WASHING.
* SIROFLOC is an AUSTEP Pty Ltd registered trademark for the clarification and decolourization of water with regenerable magnetic particles.
182
Fig. 1. Steps in the Sirofloc process.
3.1. Premix
In the first step of the process, activated magnetite and feedwater are contacted at pH values 4-6 for ten minutes. This time is sufficient for a significant reduction in colour and turbidity, and for waters containing low colour and turbidity this step is often effective enough to produce water of acceptable quality. It was also shown that colour and turbidity removal is strongly pH dependent and usually increases with decreasing pH. 3.2. Coagulant/Flocculant Addition -Aftermix
If the amount of impurities exceeds certain levels, more magnetite may be added to complete the treatment. However, the addition of a coagulant or flocculant is a preferable alternative. After the addition of the coagulant, further mixing (4 to 5 minutes) completes the effective attachment of the remaining colloidal particles to the magnetite. 3.3. Separation
The magnetite particles with attached impurities and coaqulant are magnetiz'ed. The magnetically-flocculated agglomerates then settle out rapidly, leaving clarified water. 3.4. Alkali Regeneration
The magnetite slurry is then treated with a dilute solution of sodium hydroxide.
183
Fig. 2. Full size plant a t Mirrabooka, Western Australia.
Under laboratory conditions, contacting the magnetite slurry with 0.1 M NaOH solution for 10 minutes results in separation of attached impurities.
3.5. Washing
The particles must be washed prior to their reuse to remove excess alkali. Vigorous mixing and magnetization - demagnetization during regeneration and washing, exhances the separation of the impurities from magnetite.
4. DEVELOPMENT OF THE PROCESS
Since 1977, the Sirofloc process has undergone considerable engineering development. The advances in design and operation of a pilot plant have been discussed [38, 391, as well as important commercial considerations [40]. Further engineering developments of the process and commissioning of the first demonstration full-size plant have recently been reported [41]. The demonstration plant at Mirrabooka in Western Australia (as shown in Fig. 2 ) is designed to treat 35 ML/day of underground water. This plant was officially operated in July 1981. More recently (February 1983) another large plant has been
184 Tab. 1. Species commonly found in or added t o natural waters
Component
1. Soluble Species - hardness cation Caz+,Mg2+
Mechanism of Interaction with magnetite
Application
Treatment of hard feedwaters Metal ion recovery from effluents - anions PO:-, SO:Adsorption Removal of PO:- from agricutural effluents - organic acids, humic, fulvic acids Adsorption Removal of colour from feedwaters and on occasions pesticides, surfactants and other man-made pollutants. 2. Suspended Solids - clays, silica Heterocoagulation Removal of turbidity - biocolloids, algae, viruses, bacteria Heterocoagulation Algal harvesting, removal of pathogens 3. Additives - AP+, Fe3+ salts Adsorption As secondary coagulants or surface coatings - polyelectrolytes Adsorption Improved clarification and regeneration - oxidants Cl, , H,O,, 0, Redox reactions Treatment of anaerobic waters removal of trihalomethanes -
heavy metal ions PbZ+,Co2+,Zn2+,Mn2+
Adsorption Adsorption
established at Bell Bay in Tasmania which treats 20 ML/day of highly coloured, lowturbidity surface water. 5. COLLOID AND SURFACE CHEMISTRY
While the practical aspects of the process have been well established, the study of the fundamental aspects is in its early stages. It has long been recognized that the phenomena of coagulation and flocculation occur at solid-liquid interfaces [42]. They involve particles of colloidal dimensions whose existence in natural systems depend on the presence of an electrical charge at the surface. The magnitude of this surface charge depends not only on the nature of the colloid but also on the composition of the aqueous phase. Adsorption or binding of solutes to the colloid surface may increase, decrease or even reverse the effective charge on the solid [43]. Both coagulation and flocculation rely upon neutralization of the surface charge to achieve destabilization. In natural waters the suspended solids are usually negatively charged and thus alum and/or cationic polyelectrolytes are added to bring about coagulation and/or flocculation. To obtain a detailed understanding of these phenomena one must examine the colloid and surface chemistry of the system. This is no less true for the Sirofloc process of water treatment in which magnetite is used to replace some or all of the alum or polyelectrolytes commonly used. Much of the research carried out in t h i s study has been directed towards an understanding of the process occurring on the magnetite surface in an attempt to improve the efficiency of the process in various plant situations.
185
I0
W
N
-20-
-30-
-GO-507
4
5
6
?,,
8
b
(0
11
Fig. 3. Surface charge development and isoelectric point of magnetite.
At this point it may be instructive to list the compounds and species likely to be found in a natural water. These are included in Tab. 1.
6. MAGNETITE
The starting point for any study of the Sirofloc process is magnetite and attention has been focussed on the magnetite-water interface [44,45]. The surface characteristics of magnerite in contact with water are determined by the dissociation of the surface hydroxyl groups as shown in the following equations. Development of a positive charge is by reaction with protons Fe-OH+H**Fe-OHl and of a negative charge by reaction with hydroxyl ions Fe - OH +OH-+ Fe - 0 - +H,O
(2)
The surface charge of magnetite is therefore pH dependent. Experimentally the surface charge of a mineral can be measured directly by potentiometric titration methods. More frequently microelectrophoresis is used to determine the zeta potential of the particles which can be related to surface charge. At a certain pH, the
186 zeta potential is zero and this point is called the isoelectric point - i.e.p. as shown in Fig. 3. Standard colloid chemistry techniques such as microelectrophoresis and streaming potential measurement have been used to determine the characteristics of the magnetite surface [45]. These studies have shown that the raw mineral obtained from Savage River, Tasmania, contains anionic impurities which can be removed by treatment with acid, alkali ot both. There is an accompanying increase in the i.e.p. of the oxide, which is reflected in clarification perfomance (i.e., jar tests). The correlation between the degree of pretreatment, the i.e.p. and the efficiency in jar tests is now well established. It has also been demonstrated that inefficient regeneration, i.e., the failure to release adsorbed colloids, lowers the i.e.p. and eventually leads to poor clarification. The importance of surface chemistry to the process is illustrated by the routine use of electrophoretic data to compare magnetite samples from different sources, from different stages of the pilot plant or after different procedures have been used in the pilot plant. In keeping with the declared objective of bringing the laboratory system as near as possible to the real situation, the effect of introducing other soluble components on the surface properties of magnetite has been examined in a number of projects. The results from these studies will be briefly reviewed.
6.1. Cation Adsorption
The ability of oxides to adsorb heavy metal ions is well known [46]. Magnetite is also capable of removing ions such as Cu2+,Pb2+, Zn2', Mn" from solution. The rate and extent of the adsorption and the corresponding effects on the magnetite depend upon pH, metal ion concentration and the solid-liquid ratio [47]. Desorption of these ions is best achieved at low pH and thus in a practical situation, a two-stage regeneration scheme would be needed - acid treatment to recover the metal ions and alkali treatment to reactivate the magnetite surface. Likely applications for the recovery of metal ions include treatment of metal refinery effluents, electroplating effluents, sewage effluents and sludges. A preliminary study has shown that addition of Alp and Fe3+ions has a similar effect on the surface properties of magnetite, i.e., increasing the positive charge to an extent dependent on parameters such as pH and the cation concentration. Thus these metal ions may be used as secondary coagulants to increase the capacity of magnetite to treat any turbid or highly coloured feedwaters. Alternatively by incorporation of these ions into the magnetite surface it may be possible to modify the surface properties and increase clarification efficiency. Such surface modifications are only possible when the product is of a higher market value than that currently given to drinking water, More recently the effect of Caz+and Mg2+ions (hardness), which are common to many feedwaters, on the magnetite surface properties has been studied [48]. It was found that these ions also adsorb strongly, markedly affecting the surface properties of magnetite. In many waters, there is sufficient hardness present to reverse the surface charge of magnetite, thereby aiding clarification but hindering regeneration. This has now been verified both in laboratory experiments (e.g. microelectrophoresis) and in jar tests on
187
a number of feedwaters. For a pilot plant treating such hard feedwaters, this effect has necessitated changes in regeneration and in some cases, the introduction of an acid desorption stage prior to regeneration.
6.2. Anion Adsorption
The interactions between iron oxides such as goethite and hematite and anions such as phosphate, sulphate and chloride have been examined and are reported in the literature [49]. However, it has been assumed that for the Sirofloc process, the adsorption of these simple ions would be secondary; the larger organic anions which exist in higher concentrations were thought likely to be adsorbed preferentially. While this assumption remains untested, recent results [47] have indicated that the presence of phosphate ions lower the i.e.p. of magnetite due to strong adsorption. This occurs even in the presence of competing solutes such as humic acid. Thus magnetite could be used to remove unwanted phosphate ions that are present in waters or wastewaters. Another possible application is the use of magnetite for removal of chromate ions from electroplating effluents [50].
6.3. Organics
The colour bodies present in most waters are generally defined as organic acids, humic and fulvic acids. The higher molecular weight, humic compounds which are of colloidal dimensions may be regarded as organic colloids and treated in a way similar to that for the inorganic clays and silicates present. The smaller fulvic acids can be viewed as complex anions and their interaction with magnetite treated not as heterocoagulation, but rather as anion adsorption, similar to phosphate adsorption. The qualitative evidence that exists at present favours the latter mechanism, emphasizing the specific nature of the interaction between Fe ions on the magnetite surface and these complex anions which suggests the forniation of a metal complex. The removal of these colour bodies has assumed additional importance since the discovery that some are precursors of the trihalomethanes produced upon disinfection of the product water with chlorine.
6.4. Polymer Adsorption
The adsorption of polyelectrolytes by inorganic substrates has been neglected as an area of research and it is only in the last decade that efforts have been made to unravel the mechanism by which polymers affect the stability of colloidal materials. The concept of polymer bridging between particles and the more recent idea of charge neutralization as the dominant factor are the two main theories to have evolved from thistesearch [51, 521. One of the main difficulties encountered with research in this area is the lack of suitable experimental techniques to provide information about polymer configurations and the particle size of the floc during adsorption. There is hope that modern instruments may overcome these problems.
188 Preliminary data indicate that cationic polyelectrolytes behave similarly to inorganic cations, raising the positive zeta potential of magnetite and increasing its i.e.p. Part of the Division’s work has been to correlate such data with polyelectrolyte structure and ultimately with jar-test and pilot-plant results. A number of commercial polyelectrolytes were used. It must be emphasized that in the plant situation, the polymers perform more than one task; not only do they assist in turbidity removal during clarification and influence turbidity release during regeneration, but they also affect turbidity shearing and magnetite carryover during separation of the loaded magnetite from the product water.
6.5. Two-component Systems -Heterocoagulation
The object of some of the Division’s work has been to investigate the interaction of magnetic particles with other colloids likely to be present in both natural waters and domestic and industrial effluents. Examples of such colloids include silica, clay, organic colloids (e.g. humic acids), bacteria, virus and algae. Some of these systems have been investigated and a brief summary of the results will be presented.
6.5.1. Inorganic colloids
The clay used in this study was bentonite, which was found to be negatively charged over the pH range 2-10. Experimentally, the phenomenon ofheterocoagulation was examined by the use of a light scattering technique to d e r e d n e the extent of coagulation of the non-magnetic colloid, immediately following the removal of magnetite by the application of a magnetic field. Preliminary results indicate that heterocoagulation does occur when the two colloids are oppositely charged, and is dependent upon pH, particle size and sofids ratio. It remains to be seen what effect the addition of other components has in this system, and how the results correlate with data on the removal of turbidity obtained from jar test experiments.
6.5.2. Biocolloids
Although the Sirofloc process employing treated magnetite, was designed for the removal of undesirable colour and turbidity from water, the conditions of this process are such that a satisfactory proportion of some viruses likely to be present b contaminated water. will also be removed [53]. Furthermore there is evidence that some viruses are disrupted at the pH values employed in magnetite regeneration, so that infectious particles are not released from magnetite during regeneration. The efficiency of virus adsorption by magnetite is affected by ions, suspended matter such as clay and the components of sewage effluent. These variables represent those likely to be encountered. in the purification treatment of some natural waters or in sewage effluent ”renewal”. It has been shown that the addition of a low concentration of polyelectrolyte will largely counter the interfering effect of the water components listed.
189
Hence, it appears that the Sirofloc process can be used to treat otherwise unusable water containing a small number of viruses. Further, "renewal" of even heavily COIItaminated water may be achieved in a multistage process involving the use of polyelectrolyte. In the latter case it is likely that process conditions will need to be achieved to suit the particular water under treatment. In order to provide a further safety margin, terminal disinfection by some conventional means, such as chlorination, would be desirable when Sirofloc is used to adsorb viruses from clarified sewage effluents. It has also been established that magnetite is efficient in adsorbing bacteria from aqueous suspensions [54]. The process of adsorption is chiefly electrostatic. The bacteria and algae tested have a negatively charged surface above their i.e.p. values (approximately pH 3.0) and adsorb readily to the positively charged magnetite at neutral or mildly acidic pH values. The capacity of the magnetite to adsorb bacterial cells is large. Suspended bacterial cells of concentrations less than 200 pg/ml dry weight of cell material can be removed in one step with 10 g/l magnetite. Bacterial suspensions of higher concentration must be subjected to several adsorption steps to produce a clear supernatant. Certain green algae can also be removed with magnetite. This may enable algal harvesting to be carried out more efficiently. The capacity of magnetite to remove suspended microbes may have applications other than for the treatment of effluent waters. It may be of value in microbial processes where the separation of the cells from the liquid phase by centrifugation or filtration is either difficult or expensive. Recent results have indicated that microbes adsorbed by magnetite can be gainfully employed in processes such as denitrification and dehalogenation. It has been demonstrated that numerous species of microbes adsorbed to magnetitecan accumulate the insecticides lindane and DDT from an aqueous solution. The same occurs to a lesser extent with the herbicide 2,4D. The nature of the process is not known but it appears to a partitioning of the halo-organic compounds into lipid portions of the cells. It may be practical to remove the lindanecontaining cells by desorption at pH 10. Evidence obtained suggests that adsorbed bacterial cells could be used several times before desorption.
7. CONCLUSION
The potential of the Sirofloc processes for the treatment of natural waters has been demonstrated for a range of feedwaters on laboratory and pilot plant scales. The commercial viabihty of the process has now been proven by two large-scale plants operating on vastly different raw waters. In this article we have attempted to extrapolate from fundamential colloid studies of the mechanism by which the process Sperates, to suggestions for other practical applications. Some of the more promising of these include recovery of metal ions from effluents and wastes, removal of phosphate from domestic and agricultural effluents, algal harvesting, removal of pathogens from drinking water supplies and the use of microbes attached to magnetite. It will be of interest to follow the progress of these applications in future years.
190 ACKNOWLEDGEMENTS
The authors readily acknowledge the support and assistance of all members of the water group within the CSIRO Division of Chemical and Wood Technology, Clayton, Victoria. In particular the technical expertise of Mr T. C. Ha and Ms P. A. Freeman are gratefully recorded.
REFERENCES
1 American Waters Works Association, Inc. Water Quality and Treatment, 3rd Ed., 1971,p. 94. 2 K. J. Ives, Effluent and Water Treatment Jour., November 1974,636. 3 L. Demeter, and D. Galgaczi, U.S. Patent 3350302, 1967, Hungarian Patent, Ser. No. 396.895, 1964. 4 0.M. Urbain, and W. R. Stemen, U.S.Patent No. 2,232,294,Feb. 18, 1941. 5 M. Mamula, et aL, Czechoslovakian Patent No. 132624,June 15,1969. 6 J. A. Bartnik, and A. F. San Miguel, U.S. Patent No. 3,536,198,Oct. 27, 1970. 7 D. F. Peck, et al., U.S. Patent No. 3,549,527,Dec. 22, 1970. 8 D. S. Blaisdell, and R. E. B. Klaas, U.S. Patent No. 3,697,420,Oct. 10, 1972. 9 Ch. de Latour, U.S. Patent No. 3,983,033,Sept. 28, 1976. 10 Mitugi Miura et al., U.S. Patent No. 4,039,447,Aug. 2, 1977. 11 A. B. Boliden, Belgium Patent No. 871,458,Apr. 23, 1979. 12 N. V. Blesing, B. A. Bolto, D. L. Ford, R. McNeill, A. S. Macpherson, J. D. Melbourne, F. Mort, R. Siudak, E. A. Swinton, D. E. Weiss and D. Willis, Ion Exchange in the Process Industries, SOC. Chem Ind., London 371,1970. 13 B. A. Bolto, D. R. Dixon, A. J..Priestley, and E. A. Swlnton, Prog. Water Techn., 9,833,1977, Pergamon Press, London.
14 G. Sutherland, U.S. Patent No. 3,803,033,Apr. 9, 1974. 15 D. E. Weiss, S. M. West, and D. R. Dixon, Australian Patent Appl. 20760/76. 16 D. R. Dixon, J. Lydiate, and J. Lubbock, Australian Pat. AppL 39951/78. 17 D. E. Weiss, and H. A. J. Battaerd,U.S. Patent No. 3,890,224,June 17, 1975. 18 R. Mitchell, and I. Chet, U.S. Patent No. 4,001,197,Jan. 4, 1977. 19 P. A. Munro, P. Dunnill, and M. D. Lilly, Biotech. Bioengineering XIX, 101,1977. 20 H.H. Kolm, Magnetic Monopoles, Science Jour. 4,9,60,1968,and H. H.Kolm, E. Maxwell, J. A. Oberteuffer, D. R. Kelland, C. de Latour, and E. P. Marston, 17th Ann. Conf. on Magnetism and Magnetic Materials, API, IEEE, AIME, ONR and ASTM, Chicago, Nov. 1971. 21 J. A. Oberteuffer, IEEE Transactions on Magnetics, Vol. Mag. 9,No. 3, 303, 1973. 22 D. R. Kelland, IEEE Transactions on Magnetics, Vol. Mag. 9,No. 3, 307, 1973. 23 S. C. Trindade, and H. H. Kolm, IEEE Transactions on Magnetics, Vol. Mag. 9,No. 3, 310, 1073. 24 G.Bitton, R. Mitchell, C. de Latour, and E. Maxwell, Water Research, 8, 107, 1974. 25 C. de Latour, J.A.W.W.A., 68,325,443,498,1976. 26 B. A. Bolto, K. W. V. Cross, R. J. Eldridge, E. A. Swinton, and D. E. Weiss, Chem. Eng. Prog., 71,
47,1975. 27 B. A. Bolto, D. R. Dixon, R. J. Eldridge, E. A. Swinton, D. E. Weiss, D. Willis, H. A. J. Battaerd, and P. H. Young, J. Polymer Sci., Symp. No. 49, 211, 1975. 28 B. A. Bolto, D. R. Dixon, R. J. Eldridge, L. 0. Kolarik, A. J. Priestley, W. G. C. Raper, J. E. Rowney, E. A. Swinton, and D. E. Weiss, The Theory and Practice of Ion Exchange, SOC.Chem. Ind., London, 271, 1976. 29 N. J. Anderson, R. J. Eldridge, L. 0. Kolarik, E. A. Swinton, and D. E. Weiss, Water Research 14,
959,1980. 30 N. J. Anderson, B. A. Bolto, R. J. Eldridge, L. 0. Kolarik, and E. A. Swinton, Water Research 14, 967,1980. 31 N. J. Anderson, L. 0. Kolarik, E. A. Swinton, and D. E. Weiss, Water Research 16, 1327, 1982.
191 32 L. 0. Kolarik, A. J. Priestley, and D. E. Weiss, Proc. 7th Federal Convention of Australian Water and Wastewater Association, Canberra 21-24 Sept., 1977. 33 G. A. Parks, Equilibrium Concepts in Natural Water Systems, Ch. 6, Advances in Chem. Series, No. 67, 1967. 34 G. A. Parks, Chem Rev., 65,177,1965. 35 D. E. Weiss, L. 0. Kolarik, and A. J. Priestley, Aust. Patent No. 512,553, U.K. Patent No. 1,583,881. 36 L. 0. Kolarik, N. J. Anderson, D. E. Weiss, and A. J. Priestley, Aust. Patent No. 518,159, U.S. Patent No. 4,279,756 and 4,363,749. 37 L. 0. Kolarik, Water Research, 17,141,1983. 38 N. J. Anderson, N. V. Blesing, B. A. Bolto, L. 0. Kolarik, Proc. of 8th Federal Conv. of Australian Water and Wastewater Association 12-16 Nov. 1979. 39 A. J. Priestley, and P. R. Nadebaum, Water, Jour. of Australian Water and Wastewater Association, 7, No. 3, 19, 1980. 40 P. R. Nadebaum, and H. V. Nguyen, Proc. of 8th Federal Conv. of Australian Water and Waste water Association, Gold Coast, 12-16 Nov. 1979. 4 1 P. R. Nadebaum, N. V. Blesing, A. Easton, and R. Vaughan, Proc. of 9th Federal Conv. of Australian Water and Wastewater Association, Perth, April 1981. 42 J. Th. Overbeek, Colloid Science, ed H. R. Kruyt, Elsevier, Amsterdam, Vol. 1, 1952. 43 D.C.Grahame, Chem Rev. 41,441,1947. 44 L. 0. Kolarik, D. R. Dixon, P. A. Freeman, D. N. Furlong and T. W. Healy, Presented at the AIME Conference on Fine Particles Processing, Las Vegas, Febr. 1980, P; Somasundaran, ed. Ch. 34, 652. 45 L. 0. Kolarik, Ph. D. Thesis, University of Melbourne, 1983. 46 E. Matijevii, Principles and Applications of Water Chemistry, ed Faust, S. D. and Hunter J. V., John Wiley, New York 1962. 47 D. R. Dixon, Water Research, accepted for publication 1983. 48 D. R. Dixon, Colloids and surfaces, submitted for publication 1982. 49 H.P. Boehm, Disc. Faraday SOC.,82,264,1971. 50 N. J. Anderson, L. Pawlowski, and B. A. Bolto, Nuclear and Chemical Waste Management, submitted for publication 1983. 5 1 V. K. La Mer, and T.W. Healy, J. Phys. Chem, 67,2417,1963. 52 D. R. Kasper, Ph. D. Thesis, Californian Inst. of Techn. 1971. 53 J. G. Atherton, and S. S. Bell, Water Research, 17, 943, 949, 1983. 54 I. C. MacRae, and S. Evans, Water Research, 17, 271, 1983.
This Page Intentionally Left Blank
193
BITTERNS AS COAGULANTS FOR TREATMENT OF COLOR,EFFLUENTS
Sux-bin WANG
Planning and Management Division Taiwan Water Pollution ControlAgency, Taiwan, Republic of China K. Y . CHEN
Environmental Engineering Program University of Southern Gdifomia, Gdifomia, USA
ABSTRACT Bitterns, by-products of solar salt production, are found to be enriched in magnesium and of b h ionic strength. The concentration of magnesium in the liquid form of bittern is 1.32 to 2.85 mol/kg N salinityo8 M. The in Baume degree of 29.1 to 34.9. The ionic strength of bittern is 4.05 X solid form of bittern contains about 10.1% magnesium. Bitterns have been found to be effective and economic coagulants for color purification in the pulp and textile industries. About 85% of color reduction is achieved in alkaline and acid bleach pulp industrial effluents with 2% addition of liquid form bittern using lime to maintain an appropriate pH value. Color removals of approximately 80% are obtained at a dosage of 2% of liquid form bittern introduction to textile waste waters in which acid and direct azo dyestuffs are used in coloration. Coagulation for color removal is quite chemically specific. The development of chemical interactions between magnesium and colorcausing compounds leading to the formation of insoluble aggregates are vital reactions in achieving color clarification.
1. INTRODUCTION
Dyes and pigments are used extensively in many industries for coloring. Color effluents also arise from the processing of raw material to products. Many dyes, pigments, and color-causing compounds are inert and non-toxic when discharged into receiving waters at the concentrations normally encountered in the industrial processes. However, the color can impart undesirable quality in water and reduce its usefulness. The clarification of colored water is an important problem in water pollution control. In aqueous solution, the color causing components are negatively charged colloids. Since similarly chraged particles repell each other, the colloidal particles remain suspended in water. One can, however, obtain clean water by coagulation. In this process, color is removed by the agglomeration of colloids, which can then be settled by gravity. The process of coagulation in color removal is generally very expensive, and the resulting sludge is difficult to dispose of.
194 In this paper, we propose an alternative solution for removing color from water, which is very effective and is less expensive than coagulation using the usual reagents. This new process 'for color removal from solution involves the utilization of bitterns. Bitterns are the residues after sea water has been partially evaporated by the sun. Many salts, especially magnesium compounds, are still dissolved in this solution. The objectives of this study were the following: a) to evaluate bitterns as practical and effective coagulants for color removal in treating industrial wastewaters. b) to examine the possibility of chemical interaction between color-causing compounds and bitterns which are added to bring about color purification.
2. EXPERIMENTAL 2.1. Materials
Liquid form bitterns were collected from Pu-tai solar salt production field in the southwestern part of Taiwan. Solid form bitterns were sampled from Ann-pin solar salt production and processing fields near Tainan, Taiwan. All of these samples were stored in a constant temperature and humidity chamber. Two industrial pulp wastewaters were obtained directly from the pulp mill effluents in Tao-yen and Chung-hwa, Taiwan. Four textile industrial wastewaters were sampled directly from the wash-boxes on the dye range at textile finishing mills in Taipei and Tao-yen, Taiwan. . The color-causing compounds (Concorde S and M-B) evaluated in this study were supplied by the Concorde Manufacture Company in Taiwan. Alkaline lignin was supplied by the Taiwan Chemical Industry.
2.2. Methods
Analyses of chemical compositions of the bitterns were performed according to procedures described in Standard Methods, 14th edition [l] and Sea Water Analysis Manual prepared by Institute of Oceanography, National Taiwan University. Baume degree of bittern was directly determined using Baume degree meter (Tsurumi Co.). The density of bittern was calculated by the equation: Density (g/mL) = 144.3/ /( 144.3-Baume degree). The conductivity of bittern was measured using a (Tacussel Co. Type CD 6N) conductivity meter. Double-distilled and demineralized water served as reference. The salinity of bittern was determined using a digital salinometer (Tsurumi-Seiki Co., Model 1506). The characteristic wavelength for each sample was determined by running a scan of the wastewaters on a Perkin Elmer, Model 124 spectrophotometer and determining the wavelength of maximum absorbance. The percentage of color remaining was calculated by comparing the absorbance value for the wastewater after treatment to the absorbance value for the original wastewater. In the coagulation experiments, one-litre beakers containing 800 mL of wastewaters
195 were placed on a six position mechanical stirrer. The reagent grade of calcium oxide was added to raise the pH values. This followed by the addition of a desired dosage of bittern, in the form of liquid or solid. Two minites were allowed for rapid mixing of the reactants at a stirring speed of 100 rev/min. The solutions were then allowed to settle for 30 minutes. After settling, the magnesium content of the supernatant and precipitates were determined by atomic absorption spectrophotometer (Perkin Elmer, Model 305 B). The precipitates that were used for infrared spectrum measurement were thoroughly mixed with 300 times as much powdered KBr. The mixture was then placed in a pellet die, vacuumed to remove entrapped air and pressed at 23,000 psi while under vacuum. A transparent pellet about 13 mm in diameter and 1 mm thick was formed to run an infrared spectrum (Perkin Elmer, Model 710B). For the analysis of organic complexed magnesium, one hundred and fifty mL of organic solvent (methyl alcohol, pyridine, 1,2-dichloroethane ethanol, 1,2-dichloroethane ether and tetrahydrofuran) was added to a glass-stopped flask containing 0.5 gram of finely ground precipitates. After shaking for twenty hours, the sample was filtered. The filtrate was then transferred to a breaker and evaporated under a blower hood. The dried residue was digested with sulfuric and nitric acid and flame atomic absorption used for the analysis of magnesium.
3. RESULTS AND DISCUSSION
The chemical composition of bittern liquids that were sampled from the Pu-tai solar salt fields are shown in Table 1. The total magnesium concentration of the four different bittern liquids are 1.32, 1.76, 2.36, and 2.85 mol/kg. There would have been more magnesium if more concentrated bitterns were obtained. Bittern liquids are easy to obtain for industrial use with approximately 30.6 Baume degree, because this kind of bittern liquid is a by-product of solar salt production. The magnesium concentration of this bittern is about 40 times greater than that of sea water. The major constituents of bittern solids collected from the Ann-pin solar salt production and processing fields are ahown in Table 2. Two ways of producing bittern solids are also listed in the same table. The bittern solid that was used in this study was the Tab. 1. Chemical Composition of Liquid Bitterns Type
A
Density (g/mL) 1.25 BaumeDegree 29.1 Major ions
(mol/kg)
c1-
4.19 0.44 2.18 0.24 1.32 0.002
so:-
Na
+
K+ Mga+
Ca'+
* Used in this study.
B
C
D
*1.27 30.6
1.30 33.8
1.31 34.9
4.11 0.59 1.43 0.34 1.76 0.002
4.49 0.72 0.74 OS7 2.36 -
5.29 0.53 0.18 0.47 2.85 -
196 Tab. 2. Chemical Composition of Bittern Solids ChemicalComponents
C1-
SO:- Na+ K +
Mgz+ H,O
%
33.5
2.3
10.1 5 3
Processes: 1. Bittern liquids % Bittern Solids 2. Bittern liquids + CaC1, CaSO, Liquid complex % Bittern Solids
0.5
0.6
+ liquid complex
Tab. 3. Chemical Composition of Synthetic Liquid Bittern MajorCompounds
NaCl
KCI
MgCl,
Weight (g/Kg Water)
83.6
25.2 112
Specific Gravity: 1.27 (g/mL) Baume Degree: 30.6 Ionic Strength (1 X lo-' mL/mL) = 2.18 X ( p mho) = 0.68 M
MgSO,
CaSO.
71.0
0.34
X Conductivity
pmduct of process 1. The major cations of bittern solids are magnesium, potassium and sodium. Chloride and sulfate are the major anions. About half the weight of bittern solid is water. Magnesium constitutes about 10.1%of the total solid weight. The effect of ionic strength on the destabilization of colloids can be explained using the Gouy-Chapman theory of the electric double layer [2-31. The thickness of the double layer decreases with the increasing ionic strength in a way consistent with the increasing efficiency of colloid removal with increasing ionic strength. Snoeyink and Jenkins [4] poinzed out that for water treatment purposes an approximation of the ionic strength can be derived from a correlation with conductivity. Table 3 presents the ionic strength of synthetic bitterns based on the calculation using the equation I = 1/2 X C i Z f . The ionic strength (1 X 10 -+ mL/mL) is calculated as follows:
I = 2.18 X lo-' X Conductivity (p mho)
(1)
Salinity is another general bittern characteristic. Salinity and conductivity are hghly related in the synthetic and natural bitterns in this experiment. The conductivity of bittern, then expressed by salinity is as follows: Conductivity (p mho) = Salinity'
m6
X lo327
(2)
Substituting equation (2) into equation (l), the ionic strength of natural bittern (1 X lo-' mL/mL) is derived as follows:
I = 4.05 X
X Salinityom6
The ionic strength of bittern is approximately ten times hgher than sea water.
197 In Taiwan, about 40% of the total industrial wastewater discharges are from pulp industries. Textile industries contribute about 25% of the total industrial wastewaters. The major pollutants in both pulp and textile mill effluents are suspended solids, soluble organics, color, and high pH values. As shown in Figure 1, with 10 mL bittern added to 1000 mL alkaline bleach effluent, about 85% of color reduction is obtained with 1000 mg/L lime added. However, with a 20 mL bittern addition, a color reduction of 82% is achieved with a 600 mg/L lime addition. From Figure 2, it can be seen that the color of the acid bleach effluent is strongly dependent on the amount of lime addition, When the effluent is neutralized with a lime addition of about 600 mg/L, its color becomes more intensive. With further lime additions, the color of the supernatant is effectively reduced to less than 20%. A color removal of approximately 80% is achieved at a lime dosage of 400 mg/L and of 20 mL of bittern to 1000 mL of the acid dye effluent (Figure 3). Textile mill direct dye effluent, like acid dye waste, is easy to treat with bittern for color removal(Figure4). As shown in Figure 5 and Figure 6, the sulfur dye effluent and dispersed dye effluent do not respond to treatment in manner similar to the direct and acid dye effluents. Both sulfur and dispersed wastewaters fail to yield expected color removal with bittern. In order to establish the importance of various ions in color purification with bittern, we have used both synthetic and natural bitterns to treat sythetic colored water. The color removal of lignin solution using the sythetic bittern is almost identical to that achieved with the natural bittern. These results demonstrate that the natural and synthetic bittern have the same ability to clarify color because of their similar chemical composition and ionic strength. As shown in Figure 7, the lignin removal is strongly related to magnesium removal from the alkaline solution which is consistent with the specific involvement of magnesium in the coagulation process. The infrared spectrum of an organic compound gives pertinent information of its structure by signifying what functional groups are present or absent from the molecule [4]. The infrared spectrum of the Concorde S is shown in the upper part of Figure 8. The lower part of this figure is the infrared spectrum of the Concorde S precipitates in which magnesium chloride is used as the coagulant. The presence of the band at 3200 to 3600 cm-' indicates that the hydroxyl group of this color-causing compound is labilized at alkaline condition, to provide the coordination site for reaction with magnesium, and settle as precipitates in the color clarification. Alkalimetric titration can be used to demonstrate specific chemical effect in coagulation. Alkalimetric titration of metal ions, if carried out in the presence of different complex forming ions, can give reliable information on the extent of complex formation and on the influence of these constituents on the pH for optimum precipitation. The extent of the displacement of the alkalimetric titration curve to lower pH in the presence of the coordinating anions is a measure of the tendency to form complexes or precipitates with the added metal ions [5]. Figure 9 shows that the presence of the coordinating anions of Concorde M-B has an effect on the shift of the alkalimetric titration curve. The color-causing compound of Concorde M-B contains two aromatic hydroxyl groups [6]. The effect of magnesium upon the shape of this compound titration curve may be explained by assuming that a magnesium-color complex is formed: 2 MgZ+f OH-R-OH
++MgO-R-OMg+
t
2 Ht
198
1
1
1
1
1
1
1
1
-
----
o N O Bittern o 10ml Bittern 2Oml Bittern
0' 0
I
'
I
I
I
'
-
I
I
'
400 800 1200 1600 mg CaO/Litre of Effluent
Fig. 1. Effect of Addition of Bitterns on the Color Reduction in Alkaline Bleach Effluent
140
c
120
-
$100 m c .-
.g
80
5
a
0
60
U 0
40
20
~. 0' 0
I
I
. I
'
I
I
'
800 1200 1600 mg CaOlLitre o f Effluent
600
Fig. 2. Effect of Addition of Bitterns on the Color Reduction in Acid Bleach Effluent
199
1
I
1
I
I
-
-
No Bittern o 10ml Bittern 0 2Oml Bittern 0
-
0
0
-
200 400 600 800 1000 mg CaOILitre of Effluent
Fig. 3. Effect of Addition of Bitterns on the Color Reduction in Textile Mill Acid Dye Effluent
I
c .-
-
I
I
I
1
201111 Bittern Bittern composition
0
E"
50
-
0
0
200
400
600
800
1000
rng CaOlLitre o f Effluent
Fig. 4. Effect of Addition of Bitterns on the Color Reduction in Textile Mill Direct Dye Effluent
200
30
o No Bittern o 10ml Bittern 0 2Oml Bittern Bittern composition See Type B, Table 1
10
"
0
200
400
600
800
1000
mg CaOILitre of Effluent Fig. 5 . Effect of Addition of Bitterns on the Color Reduction in Textile Mill Sulfur Dye Effluent
1
I
1
I
I
1 1
*......... o......
90
...
30
o No Bittern o lOml Bittern 0 20ml Bittern Bittern composition See Type B. Table 1
0
200
400 600 800 1000 mg CaO/Litre of Effluent
Fig. 6 . Effect of Addition of Bitterns on the Color Reduction in Textile Mill Dispersed Dye Effluent
201
--P
-
50
40
c CII .-
.c 30
-
-
1 -20 Lignin S O O m g l l .-s -Bittern 1 x 10”
a
:
g
:
v/v, 1Oml
lo ’Bittern composition
2
-See Type B, Table 1
0 140 120 CII 100 .2 80 2 60 b 40 2 20
-
.g
-
4 6 8 NaOH, O.OSH ( m l )
10
Fig. 7. Reduction of Color and Magnesium in Solution after Bittern Treatment at Different Alkaline Levels
Fig. 8. Infrared Spectra of Concorde S (Upper Chart) and of Concorde S Coagulated with Magnesium Chloride at pH = 12
202
11.5
11.0
I
10.5
I
n 0.1 x ~ O - ~Concorde M M-8 o 5 x 1 0 - 3 M Mg" 0 Mixture
10.0
9.5
1
10
I
1
I
20 30 40 NaOH, 0.01 M ( m l )
1
50
Fig. 9. Alkalimetric Titration of Concorde M-B, Magnesium Chloride and their Mixtures
and that the neutralization is
The magnesium-color complex ion is soluble and precipitation does not take place until the reaction 'MgO-R-OMg'
-I-2 OH- + HOMgO-R-OMgOH
has taken place to a significant extent. The additional OH- ions are consumed for charge neutralization of the magnesium-color complex ion. Since alkalimetfic titration of Concorde M-B has pronounced effects on the displacement of metal titration curves indicating the extent of complex formation, it was desirable to determine if magnesium coordinates with the color-causing molecules to form precipitates. The direct approach was to conduct magnesium analyses of the precipitates in addition to the infrared spectrum and deprotonation titration analyses discussed in the proceding sections. The recovery of magnesium from precipitates has been investigated by means of the dissolution of precipitates into organic solvents. In general, inorganic magnesium salts do not dissolve in organic solvent, thus, the amount of magnesium
203 obtained by dissolving the coagulated precipitates in the organic solvent is assumed to be the portion of magnesium that hascoordinated color-causing molecules to form magnesium complex. The results of the treatment of Concorde S and M-B precipitates with methyl alcohol to recover magnesium are illustrated in Table 4. The amounts of magnesium recovery from the precipitates for Concorde S and M-B are 0.81-0.82 mM and 0.88-0.90 mM. The molar ratio between magnesium and Concorde S is 1.86, for magnesium and Concorde M-B is 1.91. In reference to the results of infrared spectrum analysis (Figure 8) and alkalimetric titration (Figure 9), it is likely that the reaction pathways of complex formation between magnesium and Concorde S and M-B are through coordination with hydroxyl, which are functional groups containing oxygen as the electron donor. Tab. 4. Molar Ratio of Complexed Magnesium to Color-causing Compound in the Precipitates Concentration of Magnesium (mM)
15
Concorde S Concorde M-B
1.86 1.85 1.86 1.91 1.92 1.91
20
25
The conclusions from these studies seems to be that complex formation between magnesium and hydroxyl functional groups of color-causing compounds forming precipitates are the major removal mechanisms for Concorde S and M-B and alkaline lignin solution. From a chemical standpoint, the original ratio of water-miscible to water-immiscible portions of both Concorde S and Concorde M-B molecules are very small. Complex formations between hydroxyl groups and magnesium create a more limited aqueous solubility resulting in precipitation. Lignin are largely insoluble in their original form but are degraded to smaller fragments by the action of the cooking liquor to expose water-miscible portions which then dissolve in aqueous solution. With the introduction of magnesium, the hydroxyl groups of fragments coordinate with magnesium and hydroxide to form insoluble complexes precipitating out of the solution.
REFERENCES Standard Method for the Examination of Water and Wastewater, 14th edn., APHA, Washington D.C., 1976,Part 400, pp. 273-508. D. W. Sundstorm and H. E. Klei, Wastewater treatment, Prentice Hall, Englewood Cliffs, N.J., 1980,pp. 335-355. A. P. Black, Jour, AWWA, 52 (1960), 492-504. R. T. Morrison and R. N. Boyd, Organic Chemistry, 3rd edn., Allyn and Bacon, Boston, 1978, pp. 410-412. W. Stumm and J. J. Morgan, Aquatic Chemistry, 2nd edn., Wiley Interscience, New York, 1980, pp. 633-636. Concorde Manufacture Company, private communication, Tainan, Taiwan, 1981.
This Page Intentionally Left Blank
205
WATER TREATMENT BY COAGULATION-ADSORPTION WITH DOLOMITE
A. M. DZIUBEK and A. L. KOWAL
Institute of Environment Protection Engineering, Technical University of Wrociaw, 50-370 Wrociaw, Poland
ABSTRACT Experiments were run on natural surface water samples for removal of turbidity, coloured matter and organics by coagulation-adsorption with soft-burned and wet-slaked dolomite as coagulant. The optimum coagulant dose was found to depend on initial alkalinity provided that turbidity and coloured matter concentration are moderately high. Empirical relationships are developed between optimum dolomite dose and alkalinity, and between pH and dolomite dose. The optimum pH for dolomite coagulation ranges from 10.2 to 10.8 and is lower than that for lime coagulation with precipitation of magnesium. The relationship between removals of TOC and CODp and dolomite dose can be described by generalized BET adsorption isotherms.
1. INTRODUCTION
During the past decades chemical methods have become a vital part of practically all technological systems for advanced water and wastewater treatment. By far the most common reagent to be encountered in water and wastewater technologies is lime, which has found considerable application in softening processes, and can also be successfully used for removal of algal nutrients. Since lime is amongst the most common reagents, it is also amongst the ones which receive serious attention of many investigators who described the relationship between the presence of magnesium hydroxide precipitated during lime coagulation and the removal efficiency in wastewater treatment. The results reported earlier [ 11 indicate that the degree of clarification increases when the lime dose employed is sufficient to precipitate magnesium in the form of Mg(OH)*. This phenomenon has later been explained [ 2 ] as being due to the adsorption of precipitated CaC03 particles on the Mg(OH)2 flocs, a process which improves the degree of clarification. Taking into account the coagulating and adsorbing capacities of magnesium hydroxide, Black and Thompson [3, 41 developed a water treatment method involving magnesium carbonate as coagulant, which is precipitated to magnesium hydroxide by lime addition. The method was quite successful in producing good treatment effects both in soft and hard waters. It is also claimed to have the advantage of recirculating the coagulant which has been recovered from sludges. A recent study [ 5 ] substantiates the great ease of mag-
206 nesium hydroxide in coagulating and adsorbing organic substances present in water and wastewaters. It is generally known that the precipitation of Mg(OwZ from aqueous solutions is significant at a pH of about 10.5; a practically complete precipitation takes place at a pH range between 11.2 and.. 11.5, which can be achieved in a &h-lime treatment. On the other hand, it is obvious that, in this process, the pH level in the water or wastewater to be treated should be decreased to the pH of stabilization. Comparative investigations [6, 71 on the use of pure COz and atmospheric air in the recarbonation process have revealed that atmospheric air employed for decreasing pH in alkaline wastewaters yields CaC03 particles which show a greater settleability than those obtained with pure COz. Furthermore, when air is used for pH adjustment, the precipitated CaCO3 does not solubilize. Based on the results of both theoretical and experimental studies, the following generalization can be made: the total effect of water treatment involving lime coagulation is made up of two effects, the effect of CaC03 and the effect of Mg(OH)z on the pollutants occurring in the water to be treated. The objective of the present study was to determine the efficiency of turbidity, colour and organic matter removal in a coagulation-adsorption process with dolomite as coagulant. The results reported here form a part of a large research programme (sponsored by the Polish Government) which is concerned with the applications of domestic dolomites in water and wastewater treatment systems, as well as with the usability of dolomites in water renovation. 2. EXPERIMENTAL
The experiments were run with soft-burned and wet-slaked dolomite as coagulant. The parameters of coagulant preparation from crude dolomite have been reported in an earlier study [8] dealing with thermal dissociation and hydration. The results show that a highly reactive product will be achieved when crude, grinded dolomite is burned for three hours at 1073 K, and then slaked at a water to dolomite ratio of 2: 1. With these parameters, the total degree of hydration (for CaO and MgO) equaled 98.6 percent. The experiments reported in this paper involved 2-percent dolomite suspensions containing about 60 percent of Ca(Ow2 and 30 percent of Mg(OW2. Dolomite doses varied from 40 to 600 gm-3. The experimental water was surface water; its composition is given in Table 1. The water samples had a volume of 1.5 dm3. The experiments were conducted using the jar-test method. The process parameters were the following: rapid mix with the use of a 3 x 8 cm paddle mixer at a speed of 80 rev min-' (G = 158 s-' ) during two minutes; slow mix (flocculation-adsorption) with the same paddle, at a speed of 20 rev min-' (G = 20 s-'), during 20 minutes, and sedimentation, during 30 minutes. After completion of the sedimentation process, the samples were analyzed for turbidity, coloured matter, pH, alkalinity, presence of Ca2+and Mg" ions, and concentration of organics; the latter in terms of total organic carbon (TOC) and chemical oxygen demand-permanganate (COD,). The determinations were carried out according to Standard Methods (TOC being determined with the use of a Beckman Analyzer, TOC 915 A). The criterion for optimizing the dolomite dose was the decrease of turbidity and coloured matter to a level of 10 gm-' and 20 gm-3 (Pt-Co) respectively after sedimentation.
207 Tab, 1. Chemical composition of the raw water (except pH, other values are in gm-') Parameters
Average'
Range
Turbidity Colour (PtCo) PH Alkalinity as CaCO, Total hardness as CaCO, Chlorides as C1Sulphates as SO:TOC CODp as 0, TDS Calcium as Ca Magnesium as Mg Ammonia as N Nitrates as N Nitrites as N
30 35 1.2 120 220 150 130 12.8 10.2
15-20 20-45 6.9 -1.4 60-150 200-260 125-185 125-165 1.2-24.5 5.6-21.5 520-920 64.3-83.6 9.4-13.3 0.6- 1.4 0.1-3.9 0.015-0.200
I00 61.8 12.0 0.9 2.0 0.050
3. RESULTS AND DISCUSSION
The removal of turbidity and coloured matter showed a typical behaviour, which is usually observed during water coagulation. The dolomite doses employed in the experiments enabled a practically complete decolorization and clarification to be achieved. The optimum dolomite dose depends primarily on the initial alkalinity level. No correlation was found between initial turbidity or coloured matter content and the optimum quantity of dolomite to be used. Thus, the optimum dolomite dose is a function of alkalinity provided that the levels of turbidity and coloured matter in raw water are moderately high. Based on a statistical analysis of the experimental results (for alkalinity range of 50 to 150 gm-3 as CaC03), the relationship of interest can be described by the following empiricial equation: D o p t = 0.014 A*
+ 0.8 A
where D o p t = optimum dolomite dose, gm-3, = alkalinity of raw water, gm-3 as C ~ C O ~ . A
The calculated correlation coefficient, r, equals 0.975. Thus, the optimum dolomite dose ranges from 100 to 400 gm-3, depending on the alkalinity level in raw water. Analysis of the pH behaviour during dolomite coagulation indicates that the optimum pH equivalent to the optimum dolomite dose is between 10.2 and 10.8. At this pH range, precipitation of magnesium contained in the water was insignificant. Fig. 1 shows the relationship between pH and dolomite dose for water of various alkalinity levels. Hence,
208
a
11
P" 10 initial alkallmty. g ~ nas - ~CaC03
9
0-0
60-90
0-0
110-120
&-A 7
I
0
m
1
I
150 I
I
zm 300 mo dolomite dose D.
50.3
I
600
7 0 ~
gG3
Fjg. 1. pH versus dolomite dose.
for alkalinities ranging from 60 t o 90 gm-3 as CaC03, the relation pH = f(D) becomes l.l)-' (when D E ) 50 gm-3), and for alkalinity levels varying from pH = D(0.083 D 110 to 150 gm-3 as CaC03, this relation takes the form pH=D(0.078 D t 5.4)-' (when D 2 200 gm-' ), where pH indicates the value obtained after coagulation, and D denotes dolomite dose, gm-3 . The calculated correlation coefficients are 0.999 for both equations. This means that the correlation between pH and dolomite dose is very good from the pH level of about 9.5 on (which conforms with water softening). For comparative purpose, another experimental series with two coagulants, dolomite and lime, was run. While there was almost no difference in the range of the optimum dose between the two coagulants, this difference became more pronounced in the ranges of optimum pH. The oprimum pH for lime coagulation is always above 11.O (it usually falls between 11.1 to 11S ) , whereas that for dolomite coagulation never exceeds 11.0 (it varies, at most, from 10.2 t o 10.8). This difference in optimum pH between lime and dolomite (even though the water contained in the samples and the dosage of both coagulants were identical) can be attributed to the fact that in dolomite Ca(OH)? accounts for some 60 percent, whereas the remainder consists predominantly of Mg(OH)?. As the optimum pH for lime coagulation (particularly of water with low or moderate alkalinity levels) exceeds 11.O, there is a need to precipitate all of the Mg" ions in the form of Mg(OH),, which is an important factor affecting the removal efficiency, as well as the degree of water clarification. When the coagulation process involves dolomite, Mg(OH)2 is contained in the coagulant. Conducting coagulation at a pH of about 10.5 is advantageous in that it not only brings about a softening of the water and an increase in the degree of flocculation of CaC03 particles, but it is also sufficiently high to prevent Mg(OH), from dissolving (together with dolomite) in the water. The decrease of opt% mum pH level in dolomite coagulation as compared to lime coagulation is a problem of great significance for the following two reasons: (1) When pH is higher than 10.5, the concentration of hydroxide ions OH- increases rapidly. Thus, increasing the pH level from 10.5 to 11.5 gives an almost tenfold increase in OH- ions concentration in the water. In other words, to convert OH- to COi- at pH 11.5 the consumption of COz is ten times higher. Hence, water subjected to high-lime
+
209 8 0
u
e
dolomite dose D, gni3 Fig. 2. TOC and COD, versus dolomite dose.
coagulation will need much more COz for recarbonation than when subjected to a coagulation process using dolomite. (2) Water treatment carried out at a pH less than 11.O prevents magnesium ions present in the water from a complete precipitation, and this is also important to the human organisms. In addition to turbidity and coloured matter removal achieved predominantly by coagulation, water systems involving chemical methods may yield removals of organics, due to adsorption. In Fig. 2, the removals of TOC and COD, are plotted against dolomite dose. For TOC, the relation Ce = f D takes the form CFoc = D (0.19 D - lo)-', and for COD, this relation becomes CZ0!' = D (0.27 D - 22)-' in which CToc and CZoDP are residual concentrations of TOC and COD,, respectively, which persist in the water after the process, and D is dolomite dose greater than 200 gm-3. The calculated correlation coefficients are 0.997 and 0.977 for TOC and COD,, respectively. This indicates that TOC and COD, and dolomite dose are highly correlated, beginning from > 200 gmF3 doses, above which TOC and COD, removal efficiencies are no longer dependent of their initial concentration in the water. As can be seen from the plots of Fig. 2, TOC and COD, removal tends asymplotically to certain values which are referred to as non-removable concentrations (Cn). The values of Cn, calculated from the equations of the relation Ce = f(D) for an assumed coagulant dose of 1000 gm-3, are C;foc = 5.6 gm-3 and CEOD, = 4.0 gm-3 as 02, this accounts for some 40 percent of the initial TOC and COD, concentrations. Analyzing TOC and COD, concentrations removed per unit mass of dolomite as a function of equilibrium concentrations (C,) permitted respective curves to be plotted. The shape of the curves is similar to that of the BET multilayer adsorption isotherms. In these considerations the effect of Mg(OH)z precipitated from water at pH > 11 on TOC
210
and COD, removal efficiencies is insignificant as compared to the amounts af Mg(OH)* entering the water together with the dolomite dose. Based on the assumption that the solution contains certain amounts of C,, and that the process of TOC and COD, removals satisfies the model of multilayer adsorption, a
generalized equation of the BET isotherm was derived. The final formula with the introduction of Cn has fhe form:
where:
X
of grams of solute adsorbed per gram of dolomite at adsorbate concentration C,) Xm - number of grams of solute adsorbed in forming a complete monolayer on the adsorbent surface K - constant expressing the energy of interaction with the surface C o - initial adsorbate concentration Ce - equilibrium adsorbate concentration Cn - non-removable adsorbate concentration. - adsorption capacity (number
Figure 3 gives the isotherms of adsorption for TOC and COD, along with respective equations. The calculated correlation coefficients are 0.997 and 0.916 for TOC and COD,, respectively. Having these in mind, it may be concluded that the removal of organics by coagulation-adsorption with dolomite obeys the model of multilayer adsorption and can be described by adsorption isotherms.
n nn
equilibriun concentration
lo .g$
Fig. 3. Adsorption isotherms for TOC and COD,. Co values as in Fig. 2.
11
a
21 1 4. CONCLUSIONS
The experimental results show that dolomite prepared in an appropriate manner is an effective coagulant which may be successfully employed in the treatment of surface waters. In this method, the process of water softening has been combined with the coagulation-adsorption process. The optimum dose of coagulant depends on the alkalinity level in raw water, but only if turbidity and coloured matter content are moderately hgh. Optimum dolomite doses range from 100 to 400 gm-3, depending on initial alkalinity. With these doses, hgh degrees of decolorization and clarification are achieved; TOC and COD, removal efficiencies may be almost complete. In dolomite coagulation optimum pH varied from 10.2 to 10.8 and was lower than that recommended for lime coagulation with precipitation of magnesium. Dolomite coagulation conducted at optimum pH has the inherent advantage that the quantities of COz required for recarbonation are considerably lower than those needed in hgh-lime coagulation, and magnesium present in the water will not be precipitated. The main mechanism governing TOC and COD, removals is adsorption both on CaC03, which has been precipitated in the course of the process, and on Mg(OH)2, which enters the water together with the dolomite dose. The non-removable fractions persisting in the water after completion of the dolomite coagulation-adsorption process were up to 40 percent of the initial TOC and COD, values. TOC and COD, removal efficiencies can be plotted as isotherms of adsorption, and may be described by a generalized equation of BET isotherm. The overall efficiency of adsorption on CaC03 and Mg(OH), was found to be considerably lower than that on typical activated carbons. Coagulation-adsorption involving dolomite as coagulant can be successfully applied to the treatment of surface waters irrespective of their hardness and initial magnesium concentration. Current studies of dolomite coagulation-adsorption deal with the application of sludge blanket clarifiers, the modification of the Mg(OH)2 structure, and the management of precipitation sludges.
REFERENCES
1 2 3 4 5 6
M.E. Flentje, J. Am. Wat.Wks. Ass., 17,1927,253-260. A. P. Black and C. G. Thompson, Grant Project 12120 ESW, EPA, 1971. C. G. Thompson, J. E. Singley and A. P. Black, J. Am. Wat. Wks. Ass., 64, 1972, 11 -20. C. G. Thompson, J. E. Singley and A. P. Black, J. Am. Wat. Wks. Ass., 64, 1972, 93-100. J. Leentvaar and M. Rebhun, Water Res., 16,1982,655-662. B. Dziegielewski, A. M. Dziubek and A. L. Kowal, in L. Pawlowski (Ed.), Physicochemical Methods for Water and Wastewater Treatment, Pergamon Press, Oxford and New York, 1980, pp. 283-289. 7 A. M. Dziubek and A. L. Kowal, in P. S. Hansen (Ed.), Proc. Int. C o d . Coal Fired Power Plants and the Aquatic Environment, Copenhagen, 1982, pp. 330-338. 8 A. M. Dziubek and A. L. Kowal, Government Project PR 7 - 03.04.02.02.120., Inst. of Env. Prot. Engng., Wroclaw Technical University, 1980 (in Polish).
This Page Intentionally Left Blank
213
PRIMARY FLOCCULATION OF WASTEWATER WITH Alz(S04)S AND NaAlOz SALTS RECUPERATED FROM SPENT ALUMINIUM ANODISING BATHS
D. A. WILMS and A. A. VAN HAUTE Institute of Industrial Chemistly, Katholieke Universiteit Leuven de O o y h 2, B-3030 Heverlee, Belgium
ABSTRACT The coagulation effect of aluminium (10g/m3 A13+as 50 mol% Al,(SO,), and 40 mol% NaA10,) for treatment of sewage has been examined during two months at a wastewater treatment plant for 350,000 p.e. in Antwerp. Half of the wastewater was treated in the classical way (presettling and activated sludge), whereas to the other half of the incoming stream primary flocculation was applied. After primary treatment, the average removal efficiencies for BOD, COD, suspended solids and phosphates were 13, 5, 26 and O%, compared to 40,36,60and 72% when A13+was added. The biological effluent also was significantly better than in case no flocculation was applied. A total production of 140 g of sludge per ms of wastewater handled was found in the conventional process, and of 170 g when 10 g A13+/m3 was added. Provided sewage treatment plants have spare capacity for handling the excess sludge, application of aluminium salts from the anodising industry is especially interesting for these plants that are temporally or constantly overloaded.
1. INTRODUCTION
In 1978 S. Wajc developed a process for treating the concentrated spent baths of the aluminium anodising industry in order to recuperate technical pure aluminium sulphate and sodium aluminate, both as concentrated solutions. The scope of our study is to investigate the benefits of applying these aluminium salts as flocculants in the physicochemical pretreatment of wastewaters. The feasability and the optimal conditions (dosis, pH of flocculation, composition of the mixture of aluminium salts) for primary flocculation of domestic wastewater have been studied as a first stage in a number of jar tests and in a series of continuous runs on a pilot plant scale [l]. From these experimental results and from economical considerations [2] it was concluded that the optimal A13' dosis was about 10 g/m3, and that if an appropriate mixture of acid aluminium sulphate and alkaline sodium aluminate was used, there was no need to correct pH, which is advantageous in practice. To test these optimum conditions on a full scale, a sewage treatment plant has been searched that can easily be adapted in such a way that the whole treatment is splitted up in two equal parallel parts: one half of the installations would work in the classical
214 way, whereas in the other half the normal pre-sedimentation stage would incorporate a flocculation step. In cooperation with the engineering office De Koninckx, N.V., of Antwerp, which also designed the necessary adaptations, the sewage treatment station "Schijnpoort" in Antwerp was chosen. This treatment plant has a design capacity of 350,000 p.e., the actual BOD-load is 365,000 p.e., whereas the hydraulic load has increased in time to 600,000 p.e. or 90,000 m3/day. The conventional system consists of a presetthg, biological treatment with activated sludge and aeration by mammouth rotors. The waste activated sludge is added to the incoming sewage so that it is withdrawn together with the primary solids. The mixed waste sludge is thickened, digested anaerobically, press-filtered and incinerated. During the runs, that lasted almost two months, the concentrated solutions of recuperated aluminium sulphate and sodium aluminate were pumped from two 25 m3 containers by two metering pumps to the inlet channel before the primary settling tanks of one half of the plant. As the pumping rate was made proportional to the incoming flow, a constant dosis of 10 g A13'/m3 wastewater has been applied during the whole experimental period, and a constant composition of 60 mol% A12(S04), and 40 mol% NaA10, has been installed. The impact of dosing A13' salts on the treatment efficiency could be evaluated by comparing the quality of the primary and secondary effluent in both parts of the plant. Therefore 24 h mixed samples have been taken daily on 5 points: one sample of the incoming sewage (common for the two parts of the plant) and separately for both parts a sample of the effluent after primary and one after biological treatment.
2. RESULTS
2.1. Water Treatment
Table 1 shows the averages of the experimental results of the jar tests, the pilot plant runs as well as the runs on fill scale. One can see that there is a great similarity between the results from the three kinds of experiments, which is very important since from it one may conclude that the optimalisation of the parameters can be done perfectly well on laboratory scale, which is a much more convenient way. In a large number of jar tests the specific influence of the composition of the flocculant (pure aluminium sulphate, pure sodium aluminate or mixtures of both salts), of the flocculant dosis and of the pH of flocculation, have been investigated. As can be seen from Fig. 1, the effluent quality of a wastewater flocculated with an equimolar mixture of aluminium sulphate and sodium aluminate, improves on increasing the flocculant dosis. Generally speaking, the greatest effect is obtained by applying the smallest dosis. For COD-removal there is no much gain in increasing the dosis above 20 g/m3 A13+, whereas the removal of suspended solids and of phosphates decrease further by increasing the Aldosis. The amount of sludge produced increases steeply with the Aldosis too. The influence of Aldose and flocculation-pH on COD remoyal has been studied on a number of wastewater samples. Both AI,(SO,), and NaAlO, salts have been studied
215 Tab. 1 . Effect of A13+(10 g/m3 as 60 mol% Al,(SO,), and 40 mol%NaAlO,) on removal of BOD, COD, TSS and Phosphorus from sewage effluent: comparison of data BOD
COD
TSS
Po:-
In Out Removal In Out Removal In Out Removal In Out Removal (mg/l) % (mg P/1) (mg/l) % (mg/l) % ~~
JarTests
without withA13+ Pilot Plant without withA13+ Schijnpoort without withAP+
243 243 249 249 299 299
226 137 329 148 260 180
7 44
80 80
0
66
41 13 40
125 68 46 149 141 5 149 95 36
66 115 115
76 27 107 27 85 46
5 66 0
59 26 60
41 37 37 8.3 8.3
12 29 16 11 2.3
72 23 58 0
72
and the results are shown in Eig. 2. Here CODf stands for the COD-value of the filtered wastewater: a COD/CODf ratio < 1 means that the effluent after flocculation has a lower COD than can be obtained by filtration. From this one can see that the influence of pH is most pronounced at low dosis of flocculant, and that for a given flocculation-pH and a given dosis the COD-removal with aluminium sulphate is generally 10 to 20% better than with sodium aluminate. The influence of Aldose and flocculation-pH on removal of suspended solids is shown in Fig. 3. Here also the lowest TSS-values after flocculation are obtained with the highest Al-dosis and at the lower pH-values. NaAlO, seems to perform better than Al,(SO,),, except at pH 9. Although pure aluminium sulphate gives better results than sodium aluminate, the application of a mixture of both aluminium salts is considered since both aluminium compounds are wasted in the aluminium anodising industry, and since the application of aluminium sulphate alone may necessitate a pH-correction. Based on the pilot plant results a mixture of 60 mol% aluminium sulphate and 40 mol% sodium aluminate and a dosis of 10 g A13+/m3have chosen for plant operation. As shown in Table 1: The COD removal efficiency by sedimentation alone was on average 13% whereas by primary flocculation the COD removal efficiency varied between 23 and 61% with an average of 40%. A similar result was observed for the BOD removal: the efficiency varies between 23 and 69% with an average of 36%, to compare with 5% removal by sedimentation alone. The average removal of suspended solids by sedimentation was 26%, whereas by flocculation this varies between 39 and 89%, with 60%as an average. By sedimentation there was no removal of orthophosphates observed; with the addition of 10 g/m3 A13' 57 to 97% of the phosphates are precipitated, with an average or 72%. 2.2. Sludge Production
As the flocculation of raw wastewater with A13'-salts results in an important increase in the removal of COD, suspended solids and orthophosphates, it is evident that the amount of sludge produced will increase too.
216
01
I
I
I
5
10
1
15 20
I
I
30
40
1
80 g/m3 AI
'+
Fig. la. Reduction of COD as a function of the A13+dosis
I
I
5
10
l
l
15 2 0
I
30
1 40
Fig. lb. Reduction of TSS as a function of the A13+dosis
i
0
9,m3 A I ~ *
217
30-
25
-
20-
15
-
5-
Fig. lc. Reduction of ortho-phosphates as a function of the A13+dosis Sludgo .mount
, g/m3
ao-
300-
200-
OL
I
I
5
10
I
1
15 20
1
I
30
40
Fig. Id. Sludge production as a function of the A13+dosis
1 bo g ~ m AP 3
218
Fig. 2a. Ratio of COD after flocculation to COD after filtration as a function of flocculation pH and dosis NaAlO,
-
Fig. 2b. Ratio of COD after flocculation to COD after filtration as a function of flocculation pH and dosis Al,(SO,),
-
219
Fig. 3a. Suspended solids after flocculation as a function of pH and dosis Al,(SO,),
Fig. 3b. Suspended solids after flocculation as a function of pH and dosis NaAlO,
First, there is the production of an extra amount of "chemicd sludge" in the form of aluminium phosphate and aluminium hydroxide. Then an extra amount of primary sludge is formed, that is the difference between the suspended solids load of the incoming sewage and of the effluent from the primary settling tanks. On the other hand, due to the increased removal of organic material by flocculation, there is a reduced production of excess activated sludge. Plant scale experiments confirmed that there is an excess production of sludge; by
220 addition of 10 g/m3 A13’ the amount of sludge produced is 20% higher than produced by the conventional treatment. The amount of sludge per m3 of the wastewater is: conventional primary W m 3) excess activated sludge (g/m3) chemical sludge (dm’) Total
Wm’)
30 110 -
flocculation with 10 g/m3 Al’+
70 60 40
-
-
140
170
3. CONCLUSIONS
Primary flocculation of wastewater with 10 g/m3 A13+,as a mixture of 60 mol% aluminium sulphate and 40 mol% sodium aluminate results in: - 30%more removal of BOD in the first stage - at least 35%more removal of COD in the primary settling - a removal of 70% or more of the orthophosphates. In the conventional primary treatment the phosphate removal is practically zero. The production of a better primary effluent makes it possible to decrease the volume of the aeration tank by at least 30%. The energy consumption also decreases of approx. 30%. This means for a sewage treatment plant of 100,000p.e. an energy saving of 500,000 kwh per year, which is, by all means, not negligible. Primary flocculation may also be advantageously applied in cases where a treatment plant is systematically overloaded: the provision of a primary flocculation, which does not mean an important investment, can give the plant a relief so that an effluent can be produced that again meets the required standards. In both cases one must be able to handle the extra amount of sludge. For the dose of 10 g/m3 A13+ this represents approximately 20% more sludge, compared with the conventional treatment.
REFERENCES
G. Alaerts, D. Wilms, A. Van Haute, Alkaline and acid A13++it3 in the flocculation of water and wastewater, In: Physico-chemical methods for water and wastewater treatment, L. Pawlowski Ed., Elsevier Scient. Publ. Co., 1982, p. 43-53. G. Alaerts, E. Missine, D. Wilms,A. Van Haute, Economical comparison of BOD-removal from wastewater by physico-chemical flocculation, biological and combined treatment, In: Physicochemical methods for water and wastewater treatment, L. Pawlowski Ed., Elsevier Scient. Publ. CO., 1982, p. 31-41.
22 1
FLOC STRENGTH MEASUREMENTS GIVING EXPERIMENTAL SUPPORT FOR A FOUR LEVEL HYDROXYDE FLOC STRUCTURE
R. J. FRANCOIS and A. A. VAN HAUTE
Katholieke Universiteit Leuven, Institute of Industrial Chemistry de Croylaun 2, B-3030 Heverlee, Belgium
ABSTRACT A literature survey about floc structure research is given. In this investigation, hydrolysing metal salts were used during the coagulation-flocculation of very diluted kaolinite suspensions. The different experiments are briefly described and explained. The influence of the kinetic process parameters o n the floc strength is used to prove the validity of a four level organisation of hydroxyde floc aggregates. The different levels of organisation are: primary particles, flocculi, flocs and floc aggregates. The bonds between the particles are elastic. The influence of the kinetic process parameters on the floc structure is deduced as well.
1. INTRODUCTION
Only a limited number of investigations treat the problematic nature of floc structure. A plurality of those investigators are active in the domaine of rheology. A short review of the different investigations will be given below and summarized in Table 1. By computer simulation Vold [ 11 obtained a rigid floc structure on three levels. Flocs were formed by successive random addition of individual spherical particles. No internal rearrangements occurred. The particles not included in the core formed projecting tentacles which gave the total floc a rough surface and a mean extent about five particle diameters larger than the core diameter. Those tentacles could entangle to form weak aggregates. Vold controlled the obtained shape with the shape of flocculated unstable silica colloids in organic solvents. Sutherland [2] demonstrated that in Volds model the random addition of the primary particles was not truly random. After a correction of the errors, Volds model was found unable to fit the experimental findings. The predicted floc density didn’t increase for flocs larger than 500 units. Using Smoluchowski’s rate equation for perikinetic flocculation, Sutherland [3] treated flocculation as a series of random collisions between primary particles and particle clusters. This mechanism leads to an open network since the bonds are considered as rigid. The previous “Single Smoluchowski Model” was modified by Sutherland [4]. He modified the random rotation of the clusters and chose the col-
222 Tab. 1 . Survey of different investigations on floc structure. (a: volume fraction) Authors
Type of suspension
Type of used data
M. J. Vold [ I ]
Silica sols in organic solvents 0.00475 Carbon Black suspension Kaolinite suspension @kaolinite = 0.005 @floes = 0.0395-0.0515 @aggregates = 0.325 -0.175 Iron hydroxide flocs with kaolinite
Computer simulation
Aluminium hydroxyde flocs with kaolinite
Floc density measurements
Different suspensions in water and organics
Energy dissipation
@ =
D. N. Sutherland [3,4] A. S. Michaels, J. C. Bolger [6, 71
A. L. Lagvankar, R. S. Gemmel[8] N. Tambo, Y. Watanabe [9, 101 B. A. Firth, R. J. Hunter [ l l ] T. G . van de Ven R. J. Hunter [I21 R. J. Francois, A. A. Van Haute (this work)
Computer simulation Sedimentation experiments Plastic flow behaviour
Floc density measurements
0 = 0.022-0.15
Aluminium hydroxyde flocs with kaolinite @kaolinite = 0.0000283 @aggregates < 0.005
Floc strength measurements
lision sequence in a slightly different manner. Instead of using the same collision rate for all the particles, he altered it for each pair of aggregates. The different cluster models all were in qualitative agreement with experiments on carbon black suspensions [5]. Michaels and Bolger [ 6 , 71 used a multilayer floc model for fitting experimental data of sedimentation and plastic flow of concentrated kaolinite suspensions. The observed phenomena were explained satisfactorily. More direct experimental indications for a multi-level floc structure are found in the floc density - floc diameter relationships as measured by Lagvankar and Gemmel[8] for iron hydroxyde flocs and by Tambo and Watanabe [9, 101 for aluminium hydroxyde flocs. The first group found a discontinuity for a floc diameter of about 1.1 to 1.3 mm. For a floc smaller than that diameter, the Vold model gave a good prediction. The other group found a discontinuity with transition diameters between 4 and 100 pm, depending on the ratio of aluminium added to the quantity of solids in suspension. Similarly to Michaels and Bolger [7], Firth and Hunter [ 111 used a Bingham model to describe the flow of an electrically charged colloidal sol. The flow can be represented by three parameters (Figure 1): the point at which the flow curve becomes linear ( G o ) , the Bingham yield value (TB) and the slope of the linear curve, the Bingham viscosity ( 7 7 ~ ) . They compared the single particle model, the hard floc model and the elastic floc model with experimental results. Only an elastic floc could give a satisfactory flow pattern. So, the unit of flow can’t be a single sphere or a hard non-deformable floc. A detailed calculation of the energy dissipation in a flowing sol exhibiting plastic behaviour shdwed again the validity of the previous statement. Van de Ven and Hunter [12] perfected the previous calculations. They also defined an elaborated four level floc structure. A floc aggregate is composed of flocs. These closed packed flocs consist of a number of flocculi which are formed at the highest rate of shear to which the system was ever subjected. The bonds between the different parts are elastic.
223
=0
G
Fig. 1. Schematic plot of shear stress T versus shear rate G for coagulated sols.
2. PRESENT WORK 2.1. Experimental Set-up and Methods
A coagulation-flocculationprocess consists of three steps: a coagulation of the suspended solids, a flocculation of the destabilized particles and an elimination of the formed floc aggregates. Besides the raw water properties, the process is strongly influenced by the kinetic process parameters: duration of the rapid mixing and slow mixing phase and the energy input during the different phases. In the experiments discussed below, the slow mixing phase is always long enough to form full-grown flocs. Two groups of experiments are developed. The first group consists of sedimentation experiments. For a given set of flocculation conditions one obtains a floc mixture with an average floc diameter. In relation to the dosage there is a variation of the floc dimensions as shown in Figure 2. A change of one of the variables has an impact not only on dosage 1 and diameters 1 and 2, but also on floc build-up time and sludge production. An extended review of the obtained results and a detailed description of the experiment can be found in reference 13. For the second group of experiments one has used the same flocculator as in the previous experiments. So, the experimental conditions are exactly the same in both investigations. For these experiments dosage 2 is used (dosage 2 = 0.80 * dosage 1). The floc diameters and their distribution were measured each 15 seconds with a Malvern Particle Sizer type 2200. In this group, three types of experiments were executed, For a given set of process parameters, the floc growth was measured. An example is presented in Figure 13. From such an experiment one obtains information about floc build-up time, coagulation kinetics and floc growth [ 141. The coagulation and orthokinetic flocculation have first order kinetics [ 15, 161:
dN = -(Yo dt
4 * (du/dz)
n
-
N
A graphical example is shown in Figure 11.
224
A diuter 1
3-----------w
! i 1 diut.r2
-----
I
I
I i
I I
I
&*age
1
2
dosage 1
D
dosage ( A l l
Fig. 2. Schematic graph of the influence of the dosage on floc diameter.
Knowing the build-up time for a given set of process parameters, one can start investigating the floc strength. Full-grown flocs are broken by a known shear stress, expressed by the velocity gradient (G). After rupture the fragments are allowed to regrow. A series of such experiments provides information about the floc strength. Examples are shown in Figure 3 and 15. The last type of experiments will be referred to as the stripping experiments. The fullgrown flocs are subjected to a discontinuous increase of the velocity gradient. The fractured parts are measured. All the dimensions of the flocs, used in this text, are volume based medium values. There is no need to search for the maximum diameter since in a flocculated suspension one has a floc population. The properties of the suspension are rather due to the statistical properties of the whole population than to that of the largest floc [17]. Nevertheless, whether one uses the maximum floc diameter or one uses the volume medium average floc diameter, one obtains the same trends for the floc strength relation between floc dimension and velocity gradient in the suspension (d = a G-b). This was photographically observed by Leentvaar [ 181. It's beyond the scope of this paper to produce a survey of floc strength results obtained. Only those relevant to support the model for the floc structure will be given.
-
Tab. 2. Standard conditions of the different process parameters during the experiments Standard conditions Raw water conditions: - a 75 mg/l suspension of kaolinite in distilled water -temperature: 25°C - pH: 7.0 Kinetic conditions : - time of rapid mixing: 60 seconds - velocity gradient (Grapid) during rapid mixing: 389 s-' - velocity gradient (Gslow) during slow mixing: 34 s-'
225
The experiments which will be discussed in the following section were all executed under the “standard conditions”, as given in Table 2.
3. RESULTS AND DISCUSSIONS
The discontinuity in the slope of the diameters after rupture and diameters after regrow vs G value during the rupture (Fig. 3) is strong evidence for a non-homogeneous floc structure. So, apart from the primary particles, a floc must be organized in at least two other levels. In stripping experiments it is observed that for G values smaller than that of the bending point (k220 s-l) only large floc fragments are formed (> 80 pm). At larger G values also small fragments are observed. When one measures in detail the small fragments during a stripping experiment, then one observes that primary particles and some very small fragments are formed only at G r u p w values hgher than the Grapg value of the floc formation. This is a good indica-
+ dimebars
after
regrow
100..
T +diameters a f t e r
rupture
30 10 60 80 100
200
400 600 800 1000 2000
Fig. 3. Diameters after rupture and regrow vs Grupture
Tab. 3. Fragments measured after a Grupture of 221 s-’ . The Grapid value during the floc growth was 1018 s-’ Upper and lower diameter for the different detection rings (pm)
Fragments measured after a Grupture = 221 s-’
5.8 7.2 9.1 11.4 14.5 18.5 23.7 30.3 39.0 50.2
constant quantity of non-coagulated primary particles empty empty empty empty 0.183 vol% fraction I empty 2.138 vol% fraction I1 1.231 vol% fraction 111 3.147 vol% fraction IV
1.2 5.8 7.2 9.1 11.4 14.5 18.5 23.7 30.3 39.0
226
d ~ M t i 0 I l
I
0
10
20
30
40
._
50
60
urn
Fig. 4. Schematic survey of the methodology.
tion for the existence of flocculi as proposed by van de Ven and Hunter [ 121. From such measurements one can determine the dimension of the flocculi. An example is given in Table 3 and Figure 4. From fraction (I) it is clear that a single flocculus has a diameter between 18.5 p m and 14.5 pm. If so, a doublet must have a diameter between 37 pm and 29 pm and a triplet between 55.5 p m and 43.5 pm. The measurements show only possible triplets (fraction IV) in the range 39 p m to 50.2 pm. So, it is clear that a triplet has a diameter between 43.5 pm and 50.2 pm. Thus, a single flocculus is limited in diameter to between 14.5 pm and 16.73 pm, with a high probability of being smaller than 15.15 p m because vol % fraction (11) > vol % fraction (111). Some more results are shown in Table 4 and Figure 5. This demonstrates also the influence of the velocity gradient during the formation of the flocculi. A similar effect on the dimensions of the aggregates has been observed (Figure 6 ) . The dosage decreases for a decrease of the Grapavalue. Nevertheless increases the aggregate dimension. Both observed curves (Figures 5 and 6) can be explained by the concept of the elastic floc. Vadas et al. [19] demonstrated that an increase of the velocity gradient resulted in more spherical aggregates. He used for his experiments 2 pm diameter polyvinyl toluene latex spheres in water and biconcave 8 pm diameter human red cells in plasma. Mason [20] obtained analogous results with elastomer threads. Tomi and Bagster [21] used a structural frame analysis computer program to calculate the forces in elastic links between nodes. Four nodes composed a regular tetrahedron. The regular tetrahedra together formed a chain or a spherical aggregate. The maximum tension as a function of the number of nodes in the different configurations is given in Figure 7. The larger the aggregates, the higher the stresses in the bonds. The stresses are larger in chains than in compact aggregates. Although a chain will have less nodes than a comTab. 4. Flocculus diameter, as influenced by the rapid mixing energy Grapid ( P I )
Flocculus diameter
Floc diameter (pm)
280 3 89 542 696 843 1018
18.5 pm < flocculus < 19.5 pm 14.5 pm < flocculus < 15.15 pm 10.1 pm < flocculus < 11.35 pm 11.8 pm < flocculus < 14.5 pm 14.5 pm < flocculus < 16.73 pm 14.5pm < flocculus < 16.73 pm
175.7 192.12 166.9 167.84 187.33 204.15
227
0
250
750
lo00
1250
Fig. 5. Flocculus diameter, as influenced by the rapid mixing energy.
pact aggregate in a fixed shear field, nevertheless both aggregates will have about the same dimensions. This is demonstrated in Figure 8. This graph is calculated from data of Figure 7, assuming that the nodes are flocculi with a diameter of 15 pm and that the flocculi touch each other. The chain aggregates, formed at lower Grapa, are large because of their linear shape. The compact aggregates, formed at high Grapd, are large because of the large number of particles per aggregate. With intermediate Grapid values one can expect more tortuous aggregates with intermediate dimensions. Tortuous flocculi can be expected to interwine better than linear or spherical ones. Thus, flocculi formed at intermediate Grapd values can group themselves better. This explains Figure 9. The number of flocculi per floc is calculated by dividing the volume of a floc (Table 4) by the volume of the corresponding flocculus. For this, one can use Tambo's [ 101 equation relating floc porosity to floc diameter to calculate the volume of solids in the large floc. The differences in shape do not interfere further with the combination of flocs to form floc aggregates. This is proved by Figure 10. From Figure 3 it is already known that floc aggregates are completely disrupted in flocs at G r u p m values of more than 221 s - l . The graphs of Figure 10 show that a constant ratio of the floc diameter to the floc aggregate diameter exists except for the lowest Grapd value of 280 s-l. This exception is probably due to insufficient mixing, as demonstrated by the different reaction constant (k) for that small Grapavalue (Fig. 11). Sufficient mixing depends not only on the G value but also on the duration of the mixing phase. No rapid mixing at all (t = 0 s ) yields very different results: the flocs are smaller (ref. 13, Figures 12 and 14), the floc build-up time is longer (ref. 13) and the reaction constant is smaller (ref. 14). This is self-evident because of the flocculi, the building stones of the structure, are formed at the w e s t G value to which the system is subjected. For mixing times of 30 seconds or longer, one observes a constant average floc dia-
228
0
2 50
500
750
1000
Grapid
Fk.6. Influence of GIapid on dosage,
(0).
diameter, (m) and diameter,
1250
(2) (A).
meter for very high G r u p m values. This means that the build-up of flocculi does not take more than 30 seconds. For G r u p b e values between 221 s-l and 696 s-l only mixing times of more than 150 seconds display a constant floc diameter. Mixing times of 60 and 30 seconds seem to result in more resistant flocs. An explanation for this is that mixing times of more than 120 seconds disturb the growth of flocs to floc aggregates [ 2 2 ] . Figure 13 demonstrates the temporary arrest of floc growth for a rapid mixing time of 30 seconds, the growth of the flocculi is the same, only the growth from floc to floc aggregate can be disturbed. So, it is clear why there isn't a constant rupture factor for the rupture of floc aggregates to flocs and why the flocculus diameter is the same for rapid mixing times of more than 30 seconds. Although the flocculi who are the building units of the structure have the same dimensions and shqe, nevertheless the recovery capability of the flocs formed by them is not the same (Figure 14). A final kinetic variable is the Gsbw value. Three series of experiments were carried out. Gslow values of 21 s-' , 34 s-' and 54 s-l were tested. After rupture, the floc frag-
229 200 100
60 40
20 10
6 4
2 1 1
2
4
10
6
20
40 60
100 200 400 100 1000
nurnbc,r of n o d e s
Fig. 7. Maximum tension in the elastic links between the nodes vs number of nodes. Results from reference 21, for spherical aggregates ( 0 ) and chains (m).
ITmax k
.c 1
I
a00
100
60 4u 20
10
6 4
2 1
20
40
60
80
100
dinwnsions
200 (
Fig. 8. Diameters ( 0 ) of a spherical aggregate and length of a chain (m) vs the maximum tension in the elastic links between the nodes.
230
rl
7
0 0 rl 0 44
W
0
Fig. 9. Influence of Grapid on the number of flocculi per floc.
0
2 50
500
750
1000
Grapid k3-l) Fig. 10. Velocity gradients during rapid mixing vs rupture factors.
diameter after rupture diameter after undisturbed grow legend of the symbols : + Gmpture 100 S-' 0 Grupture 221 s-' 0 Grupture 389 C' 0 Grupture 696 s-' A Grupture 1018 S-' Grupture 1398 s-' rupture factor =
1250
23 1
1E-2 6E-3 4E-3 2E-3 c,
$ 1E-3 u1
8
6E-4 4E-4 2E-4 1E-4t 0
2 50
I
1
1
500
750
1000
I 1250
-1
Grapid
Fig. 11. Grapid vs constants, reaction constant k
-
E 1
(’
.,
destabilisationfactor a0 A.
350 300
cn ru
250
:200 bd
2 4J
3” 150 bd
ru 0
100
bd
2 8
50
; 0 “ 0
0
50
100
150
200
250
300
350
400
time of rapid mixinq ( s )
Fig. 12. Time of rapid mixing vs diameter of ruptured flocs. The key is the same as in Fig. 10.
232 300
di ln-
Fig. 13. Floc growth vs coagulation-flocculation time for a rapid mixing time of 300 seconds.
Fig. 14. Time of rapid mixing vs diameter of the regrown flocs. The key is the same as in Figure 10.
233
diameters after regrow
,diameters a f t e r rupture
30 40
60 80 100
200
400 600
1000
Gdes trucc ion ( s
-1
2000 1
Fig. 15. Diameters after rupture and diameter after regrow vs Gmpture for flocs formed under different slow mixing conditions. Legend of the symbols: 0 Gslow 21 s-' 0 Gslow 34 S-' A Gslow 54 s"
ments were allowed to regrow under a G value of 34 s-'. The dimensions of both the ruptured and the regrown flocs are the same (Figure 15). This agrees completely with the four level elastic floc model. Only the aggregation of flocs to aggregates is influenced by the slow mixing energy input. Differences in floc aggregates are only noticable at Gruptm values smaller than the Gsbw value. There seems to be no difference once an aggregate is disrupted with aGmpture larger than G slow.
4. CONCLUSIONS
Aggregates formed as a result of a coagulation-flocculation process have a four level structure: primary particles, flocculi, flocs and floc aggregates. The bonds in the complete structure are elastic. This model is now proved to be also valid for flocs formed during coagulation-flocculation processes in extremely dilute suspensions, and with the use of hydrolysing metal salts. A constant number of flocculi form a floc. If the nature of the flocculus is changed, then the number of flocculi in a floc changes. There is also a constant number of flocs in a floc aggregate. This number increases with a decrease of the mixing intensity during the formation of the aggregates and vice versa. A disturbance of the growth also influences the ratio number of flocs per floc aggregate.
234 NOMENCLATURE empirical constants in the floc strength relation diameter (pm) velocity gradient (s-') reaction constant (s-' ) force proportionality constant number of primary particles number of primary particles at time 0 seconds number of primary particles at time i seconds time (s) largest force in a structure, either t e n d or compressive of all orientations orthokinetic collision efficiency, destabilisation factor (-) Bingham viscosity link length shear stress Bingham yield value volume fraction
REFERENCES
1 M. J . Vold, J . Coll. Sci., 18 (1963), pp. 684-695. 2 D. N. Sutherland, J. Coll. Interf. Sci., 22 (1966), pp. 300-302. 3 D. N. Sutherland, J. Coll. Interf. Sci., 25 (1967), pp. 373-380. 4 D. N. Sutherland, I. Goodarz-Nia, Chem. Eng. Sci., 26 (1971), pp. 2071-2085. 5 A. I. Medalia, Carbon, 7 (1969), pp. 567 e.v. 6 A. S . Michaels, J. C. Bolger, Ind. Eng. Chem. Fund., 1 (1962), pp. 24-33. 7 A. S . Michaels, J. C. Bolger, Ind. Eng. Chem. Fund., 1 (1962), pp. 153-162. 8 A. L. Lagvankar, R. S . Gemmel, J. Am. Wat. Wks. Ass., 9 (1968), pp. 1040-1046. 9 N. Tambo, Y. Watanabe, J. Japan Wat. Wks. Ass., 397 (1967), pp. 2-10; 410 (1968), pp. 14-17; 445 (1971), pp. 2-9. 10 N. Tambo, Y. Watanabe, Wat. Res., 13 (1979), pp. 409-419. 11 B. A. Firth, R. J. Hunter, J. Coll. Interf. Sci., 57 (1976), pp. 248-275. 12 T. G. van de Ven, R. J. Hunter, Rheol. Acta, 16 (1977), pp. 534-543. 13 R. J . Franpois, A. A. Van Haute, G . C. Winderickx, Proc. Water Filtration, Antwerp, April 21-23, 1982, K. VIV Antwerp 1982, pp. 1.55-1.64. 14 R. J. Francois, Proc. Dag der Jongeren, Leuven, April 20, 1983, V.C.V.-Tijdingen special edition, 8 pages (in dutch). 15 T. R. Camp, P. C. Stein, J. Bost. SOC.Civ. Engrs., 30 (1943), pp. 219-237. 16 M. von Smoluchowski, Z. Physik. Chem., XCII (1917), pp. 129-168. 17 J. D. Pandya, L. A. Spielman, Proc. IUTAM-IUPAC symp. on Interaction of Particles in Colloidal Canberra (Australia), March 16-21, 1981, 38 pages. 18 J. Leentvaar, M. Rebhun, Proc. Water Filtration, Antwerp, April 21-23, 1982, K. VIV Antwerp 1982, pp. 1.45-1.54;Wat. Res., 17 (1983), pp. 895-902. 19 E. B. Vadas, H. L. Goldsmith, S. G. Mason, J. Coll. Interf. Sci., 43 (1973), pp. 630-648. 20 S. G. Mason, J. Coll. Interf. Sci., 58 (1977), pp. 275-285. 21 D. Tomi, D. F. Bagster, Chem. Eng. Sci., 30 (1975), pp. 269-278. 22 R. J. Franpois, A. A. Van Haute, Proc. 29th International Congress of Pure and Applied Chemistry, Cdrsptq kzng 5 4 Q ,19%3(&sj.iwJ>, .
235
MEASUREMENT OF THE CHARGE DENSITY OF POLYELECTROLYTES BY A DIFFERENTIAL CONDUCTOMETRIC METHOD
G. TIRAVANTI, F. LORE, N. PALMISANO
Water Research Institute, CNR, 5 Via De Blasio, Bari, Italy
ABSTRACT The reaction between two polyelectrolytes of opposite charge proceeds via the formation of a polyelectrolyte complex (PEC) which precipitates near the isoelectric point. During this reaction equivalent amount of counter-ions are released in the solution, making the indication of the end point by conductometric titration possible. An electronic device has been developed in order to carry out a differential titration. The method has been successfully applied to determine, at different pH values, the charge density and the degree of ionization of many cationic polyelectrolytes usually applied in wastewater treatment. The comparison at neutral pH with the colorimetric method gives results which agree between *3%. The proposed method requires a careful control of the temperature (* 0.01"C); it is more precise and accurate than the colorimetric one, and can be applied to dilute solutions in the whole pH range.
1. INTRODUCTION
The water soluble ionic polyelectrolytes with high molecular weights (1-10 Millions) are useful for many processes, as water quality control, sludge conditioning, corrosion control, oil/water separation, etc. In these processes the polyelectrolytes are used to neutralize suspended impurities and to agglomerate them into larger masses for rapid solid-liquid separation by sedimentation, flotation, centrifugation, etc. In spite of this great interest, many information are lacking on the part of the manufacturers, including the type of polymer or copolymer, the concentration of active sites, the degree of ionization, the molecular weight and the intrinsic viscosity under specified conditions. The knowledge of these data is particularly important in order to develop rational procedures for the selection of the material and of the optimum dosage for a particular task. Most often polymer selection and dosage are made by means of empirical data obtained from ,jar-test*, experiments, whose results are not always reliable and transferable to a plant design. In particular, polyelectrolytes are used in water treatment either as primary coagulants or in addition to metal salts. In these cases the knowledge of the value of the polyelectrolyte charge density allows the prediction of its behaviour, on the basis of the mechanism of surface charge neutralization and subsequent adsorption of colloidal particles. Generally the electrophoretic mobility is considered to be a useful
236 tool in determining the charge characteristics of colloidal particles, for coagulant dosage control. However, the system under study is often constituted by several kinds of different colloids, including the nature, shape, and size of the particles, having different electrophoretic mobilities. This implies a measurement of an ,,average,;migration velocity, which, on the other hand, is not easily determinable as it requires experience, time, sophisticated equipment and specialized personnel. A different approach was followed by Kawamura and Hanna [ 11 and by Wang and coworkers [2]. They study a colloidal titration technique, for determining the charge density of colloids, in which a positive colloid reacts stoichiometrically with a negative standard colloid forming a precipitate polyelectrolyte complex (PEC) at the isoelectric point, using the toluidine blue (TB) as indicator. The end-point, determined by the metachromatic color change of TB from light blue to bluish purple, is well defined during the direct titration of positive colloid suspensions with PVSK (potassium polyvinylsulphate). The charge density of negative colloids can be determined by an indirect method, where a known excess of a positive colloid must be added first and then back-titrated with PVSK, otherwise the color change at the end-point is indistinct. The application of this technique has some limitations as it cannot be applied to hghly colored wastes, to dilute solutions, and to acidic or alkaline media, conditions quite frequent in wastewater treatment. In this paper a differential conductometric method has been developed to indicate the end-point of the collid titration, following a modified version of the method reported in [3]. The principle of the method is based on the release of the counterions of the polyelectrolytes during the PEC formation, according to the reaction: R-OSO;K+
+ R'-NH;Cl--+
R-OS03-NH3-R'
+ K' + C1-
These ions increase the electrical conductivity of the solution, until the equivalence point is reached; further addition of the reagent, having lower conductivity, gives rise to a decrease in conductivity. The equivalence point can be determined, as usual, by plotting the differential conductivity values as a function of the volume of the reagent; the points thus obtained can be interpolated by two straight lines intersecting at the equivalence point. The value of the charge density of the polyelectrolyte under study is easily calculated if the charge density of the standard colloid is known. Colloid charges are most conveniently expressed in terms of meq/g of positive or negative charge. 2. EXPERIMENTAL 2.1. Apparatus
The electrical conductivity was measured at 1000 Hz with a conductometer CDM3 by Radiometer, Copenhagen. A water thermostated (+O.Ol"C) glass cell with 2 platinum electrodes of about 0.5 cm2 each at a distance of about 0.4 cm from each other served as a measuring cell. Differential conductometric measurements were carried out making use of an operational amplifier, connected to the conductometer as reported in Fig. 1. This arrangement
237 allowed the reset of the conductivity of the initial sample, so that only the variations related to the addition of the reagent were measured. An automatic titration apparatus, with a microburette (accuracy 0.001 ml) was used.
2.2. Reagents
1. Stock (Potassium Polyvinylsulfate) PVSK solution, 0.01 M. Dissolve 1.622 g of salt (Eastman Kodak, Cat. No. 8587, or Serva Feinbiochemica, Cat. No. 33426) on a 100% active bases in 1 1 of distilled water. This solution must be standardized with Polybren solution. 2. Toluidine Blue4 solution (TB), 1000 mg/l. Weight 1.000 g of Toluidine Blue 0 (Eastman Kodak, Cat. No. C1756) on a 100% active basis. Dissolve in distilled water and dilute to 1 1.1 .OO ml = 1.OO mg Tb. 3. Stock (1,5dimethyl-l,5diazaundecarnethylenepolymethobromide) Polybren solution, 0.01 M.Weight 1.871 g of salt (Ega-Chemie, cat. No. 10,768-9) and dissolve in 1 1 of distilled water.
2.3. Polyelectrolytes
The cationic polymers used in this study are listed in Table 1. The number-average molecular weights and active ingredients are also reported. The aqueous solutions of these polymers were prepared just before use at concentration of 1 g/l.
2.4. Procedure
Place 400 ml of distilled water into the thermostated glass cell, gently stir (-300 Hz) by mechanical means, and add 2000 ml of a solution of cationic polyelectrolyte at 1 g/l. The pH of the solution is adjusted to the desired value with small additions of HCl or NaOH 0.1 M. Wait until the temperature of the solution reaches a constant value, by measuring it with an accuracy of +O.Ol"C. The initial value of the conductivity is reset making use of the potentiometer a (see Fig. 1) which controls the offset voltage so that Vo = Vr. Start with the automatic titration by adding the standard polyanion PVSK 0.01 M, previously adjusted to the same pH value of the solution under study. Record the differential titration curve. The same procedure is applied to anionic polyelectrolytes, the only difference beeing the titrant solution which, in this case, can be either the standard Polybren solution 0.01 M or every previously standardized polycatkn.
3. RESULTS AND DISCUSSION
PVSK made by two different firms (Eastman and Serva) have been standardized by colorimetric and conductometiic titrations. A solution of Polybren 0.00582 M (20.00 ml
Fig. 1. Electric lay-out of the differential conductometric method for measuring the charge density of polyelectrolytes. VR conductivity meter output VO offset voltage VU (VR-VO) differential voltage output a) variable potentiometer b) high stability power supply NS conductivity meter A.O. operational amplifier mV voltmeter REC recorder
in 400 ml of water) was used as standard polycation. Table 2 shows experimental data obtained from the comparison between both reagents. From these data the degree of esterification of PVSK Kodak and Serva has been estimated to be 92.2% and 76.6%, respectively. The concentration of potassium determined on both theoretical 0.01 M PVSK solutions gave results which, in the case of Kodak only, were in good agreement with those obtained with the conductometric titration. PVSK Kodak was then chosen as standard polyanion, due to its greater degree of esterification and the lower conductivity of its aqueous solutions. Fig. 2. shows an example of differential conductometric curve obtained for the polyelectrolyte Primafloc C7, titrated with the PVSK 0.0092 M solution at pH 7. The first branch of the graph reflects the appearance of ions having greater conductivity, but after the end point has been reached the graph becomes almost horizontal, since the excess of aqueous PVSK is not appreciably ionized. The accuracy of the method is better than 2%. The precision is demonstrated by the following example. Ten determinations of the charge density of the Primafloc C7 at pH 7 gave a mean value of 3.32 k0.08 meq/g, with a standard deviation of 2.5%. The kinetics of the reaction between polyelectrolytes of opposite charge is rather fast: it has been verified by rapidly adding 11.0 ml of PVSK 0.00654 M solution to a Primafloc C7 aqueous solution (20.0 ml at 1 g/l in 400 ml water). The reaction, followed
Tab. 1. Characteristic and charge densities of cationic polyelectrolytes ~~
~~
~
Polymer
Pro deflo c Cl O-I05
Praestol 444K
Praestol 523K
Praestol 4233
Ecoclar AR-8017
Primafloc c7
Nymco 540
Reference Supplyed by
14 I Prod e co
15 1
181 Nymco
> 106 solid Polyacryl amide
> lo6 liquid Polyamine
4.106 solid Polyacryl amide
PH (a) Specific conductivity (b) Cationic charge fraction Charge density (c)
3.75 32.0 100% 3.74 kO.09
4.4 35.0 100% 3.10 f 0.08
5.5 16.6 50% 2.25 2 0.05
I51 Chem. Fab. Stockhausen <3.106 solid Organic co-polymer of acrylamide 4.6 11.6 50% 1.98 f 0.05
171 Rohm & Haas
5 . lo6 solid Polyacryl amide
Chem. Fab. Stockhausen <10.106 solid Polyamine
16 1 Montedison
Molecular weight Physical state Matrix
15 1 Cheni. Fab. Stockhausen 1-3 . solid Polyamine
4.3 200 100% 3.60 f 0.09
2.5 55.5 100% 3.32 f0.08
4.1 6.4 50-100% 2.25 f 0.05
(a) 1% aqueous solution (b) T = 25"C, 1% solution, p K ' cm-' (c) determined at pH equilibrium, meq/g
240 Tab. 2. Results obtained from the standardization of PVSK made by two different firms (Kodak and Serva)
Theoretical concentration (mol/l) pH of the 0.01 M solution Electrical conductivity of the 0.01 M solution at 25°C ( p a - ' cm-') Experimental concentration (mol/l) Theoretical charge density (meq/g) Experimental charge density (meq/g) Potassium concentration (mol/l) Degree of esterification (%)
Kodak
Serva
0.01 4.9 610 0.00921 6.16 5.68 0.0092 92.2
0.01 2.85 1,100 0.00765 6.16 4.72 0.0089 76.6
conductometrically, was completed between 1 0 seconds. This allowed to utilize the automatic titrator. A comparison between the colorimetric titration, carried out according to the procedure described in [2] and the conductometric one gave results in agreement between *3% at neutral pH values; however, working with solutions at pH lower thgn 5 and higher than 8, the scattering of the data given by the indicator method wasin the order of *2Wo. The last line on Table 1 reports the values of the charge density of the studied polyelectrolytes, determined at the pH of equilibrium of their aqueous solutions (1 g/l). In order to test the applicability of the proposed method as a function of the ionic strenght of the solution, measurements of charge density of polyelectrolytes were performed by dissolving them in aqueous solutions containing different amounts of NaCl, in the range between 1 and 50 g/l. In these experiments the pH and conductivity of the standard PVSK solution were corrected by adding acid or base and NaCl until reaching the same values of the solution to be titrated. The results were satisfactory up to an initial conductivity of about 10,000 pS; for higher values, the constancy of the temperature during the run became the critical factor. In these conditions it can be calculated that a 0.0l"C variation in the temperature of the solution gives rise to a 4% error in the measurements. Fig. 3. shows the trend of the charge density of the selected polycations as a function of pH. The curves indicate that, for pH values lower than 5, the amino groups are converted into the ammonium form. With increasing pH, some primary, secondary or tertiary amino groups rearrange into the indissociated form with consequent decrease of the charge density. This phenomenon is less evident for strong polyelectrolytes (Ecoclar 8017, Praestol 444K, Nymco 540) while it is considerable for the weak ones: the charge density variations, between pH 5 and 9, are about 20%, 37% and 7576, for the Prodefloc C10-105, the Primafloc C7 and the Praestol 423K, respectively. This leads to the conclusion that the cationic polyelectrolytes, when utilized for water treatment, perform best in acidic pH media, where they have the maximum of the positive charges. Moreover, at a given pH value, several polyelectrolytes classined in the literature as 100% active cationic form, give charge density values which differ from each other in the order of 20%. Greater differences can be found for the polymers with the 50% of cationic charges.
24 1
20
15
v, %
a
10
5
0
5
10
15
m l PVSAK 0.009N Fjg. 2. Differential conductometric titration curve for determining the charge of the cationic Polyelectrolyte Primafloc C-7. a) equivalence point.
The proposed differential technique can be applied for the determination of the charge density of polyelectrolytes as low as 0.5 X18" meq/g, by using a titrant solution at 2.5 X M concentration.
4. CONCLUSIONS
The results of this study indicate that the differential conductometric end-point method performs better than the colorimetric one, in the titration of colloids of opposite charge, as it allows the determination of the charge density of polyelectrolytes and the degree of ionization in the whole pH range with an accuracy better than 2% and a precisione of +2.5%. The proposed method requires a tailored temperature control (+0.01 "C), it cannot be applied to solutions having ionic strenght higher then 0.05 M NaCl.
242
P
I
\
\
\
\
f\
\t
\
1-
\
NYMCO 540 h- PRODECO CIO-105 PRIMAFLOC C7 *--PRAESTOL K444 *-..- ECOLAR AR8017 +PRAESTOL 423K PRAESTO L 523 K
\\
-e
b
e
\ \ \
-
1
\
‘t
.---O! 3
\
I
I
I
I
I
4
5
6
7
8
I
9 PH
Fig. 3. Charge density of some polycations as a function of pH.
The kinetics of the charge neutralization reaction between colloids is so fast that an automatic titrator can be utihzed to record the titration curve. This technique has been successfully applied in termining the characteristics of the total charge and of the isoelectric point of colloidal particles in process waters, industrial effluents and sludge conditioning [ 9 ] ,even if they are colored and at pH far from neutrality. Research is still in progress in order to apply this method to the automatic control of the optimum dosage needed to attain the isoelectric point for a given wastewater, with response times in the order of 10 minutes. No t e: Fig. 3
-
the symbol o (open) refers to NYMCO; 0 (full) refers to PRODECO
243 REFERENCES
1 S. Kawamura, G. P. Hanna, Coafulant dosage control by colloid titration technique, Procees. 21st Ind. Waste Conference, Purdue Univ. (1966) 381. 2 L. K. Wang, M. H. Wang, J . F. Kao, Water, Air, Soil Pollution, 9 (1978) 337. 3 K. Toei and T. Kohara, Anal. C h i n Acta, 83 (1976) 59. 4 Anon., Technical Bulletin of Prodeco, S. Donato Milanese (1982). 5 Anon., Technical Bulletin of Chemiche Fabrik Stockhausen & C. Krefeld, (1974). 6 Anon., Technical Bulletin of Montedison, Milan (1973). 7 J. Gregory, Trans. Faraday SOC.,65 (1969) 2260. 8 Anon., Technical Bulletin of Nymco, Cormano Milan (1980). 9 F. Lore’, N. Palmisano, G . Sonnante and G . Tiravanti, Influence of charge density of cationic polyelectrolytes on sludge conditioning, Proceedings of the Meeting on sludge treatment, Bari, Nov. 24, 1982 (in Italian), (in press).
This Page Intentionally Left Blank
245
THE INVESTIGATION OF APPLICATION OF DISSOLVED AIR PRECIPITATE FLOTATION IN THE ABSENCE OF COLLECTOR AND FROTHER FOR THE PURIFICATION OF WASTEWATER CONTAINING METAL IONS
D. MISKOVIC, E. KARLOVIC, B. DALMACIJA Institute of Chemistry, 21000 Novi Sad, Yugoslavia
ABSTRACT A study of batch dissolved air flotation of mixtures of following hydroxides (model solution of a wastewater): iron, copper, nickel, zinc, cadmium, and chromium is reported. Collector and frother were not used in this series of experiments. The influence of pH on the flotation of Cu(OH), in the absence of collector and frother was also investigated. The best recoveries, in the range of 47.7-73.0% (depending on the metal hydroxide), and 89.4% for Cu(OH), , were achieved at pH 8. Concentration degrees of the meta! hydroxides (defined as the ratio of the metal ion concentration in the foam and flotated solution) were in the range from 45.3 for Ni(OH), to 119.2 for Fe(OH), , and for Cu(OH), from 67.1 (pH 7) to 250.7 (pH 8). Percentage of solution in the foam for a mixture of hydroxides was 3.4, and in the case of Cu(OH), in the range from 2.0 (pH 6) to 16.1 (pH 8). Kinetics of flotation of Cu(OH), are linear at pH 6, and power law at pH 7, 8 and 9. These results were compared with those of flotation in presence of dodecylbenzene sulphonate and ethanol, used as collector and frother, respectively. It was shown that the flotation process is only to a small extent more efficient in the presence of dodecylbenzene sulphonate and ethanol. In spite of that, the general goal of the present investigation, i.e. the possibility of precipitate flotation in the absence of collector and frother, was achieved with good recovery both of metal hydroxides in mixture and Cu(0H) 2 , treated separately.
1. INTRODUCTION
Precipitate flotation encompases all the processes in which an ionic species is concentrated from an aqueous solution by formation of a precipitate, which is then separated by flotation [ l , 21. It is primarily employed as a separation method, recently in wastewater treatment [3, 41. It has been shown earlier that with the use of surfactants as collectors, metal hydroxides can be separated from water suspensions by precipitation flotation, i.e. colloid flotation [l-131. The objective of this work is the separation of a mixture of metal hydroxides by their flotation with no collector and frother present in the system. Some basic parameters
246 describing the efficiency of the process, such as recoveiy as a function of time, concentration degree, and percentage of the solution passing into foam, have been determined. The influence of pH on the kinetics and flotation efficiency parameters has been investigated on Cu(OH), , treated separately. For purposes of comparison, flotation of the metal hydroxides in the mixture and Cu(OH), have been also carried out in the presence of dodecylbenzene sulphonate (DBS) and ethanol as the collector and frother, respectively. In this way we have investigated the application of precipitate flotation without collector and frother for the purification of wastewater containing metal ions. Some aspects of the mechanism of the process are discussed.
2. EXPERIMENTAL 2.1. Chemicals and Solutions
The suspensions of hydroxides of copper, zinc, chromium, cadmium, nickel, and iron, each at a concentration of 60 mg/dm3, and Cu(OH), when treated separately at a concentration of 400 mg/dm3, were prepared from the corresponding salts by precipitation with 5 M NaOH. The stock solution of DBS was prepared from the purified preparation (by Sohxlet extraction of DBS with ethanol [ l l , 12]), by dissolving it in distilled water, or in a water-ethanol (1 : 1) mixture.
2.2. Apparatus
The glass apparatus for dissolved air flotation is schematically depicted in Figure 1 . It consisted of the cylinder 1 and flotation cell 2. The cylinder had an inlet for flotation
i'T 7
L
Fig. 1. Apparatus for dissolved air flotation
I
I OLUMN 2. FLOTATION C E L L 3. C O M P R E S S E D A I R INL ET L. F L O T A T I O N PULP I N L E T 5,6, 7.8. T A P S
247 pulp 4, tap 7 and tube 3 for the compressed air supply, tap 5 for air outlet, and tap 8 for outlet of the flotation pulp. The flotation cell had also tap 6 for sampling. Concentrations of DBS and copper, when Cu(OH), was treated separately, were determined by spectrophotometric methods (described below), on a Spekol 20, Carl Zeiss. Concentrations of nickel, zinc, cadmium, chromium, iron, as well as copper from the metal hydroxides mixture, were determined by atomic absorption spectrometry (AAS), using an SP-19 1, Pye Unicam.
2.3. Procedure
The starting volume of the solutions was 0.800 dm3. Conditioning of the flotation pulp at the pH of investigation was carried out for 10 minutes. The flotation pulp in saturation cylinder 1 (Figure 1) was satured by compressed air at 200 kPa for 5 minutes. After that the flotation pulp was transferred into cell 2 by closing tap 7 and opening tap 8. During the 120 second period of flotation the samples for analysis were taken through tap 6. The concentrations of metals in the mixture were determined by AAS in the usual way. Spectrophotometric determination of copper was carried out by measuring the absorbance of its tetraammino complex [ 141, while DBS concentration was determined by the methylene blue method [15]. All experiments were carried out at 293 k0.5 K. The statistical data treatment was performed on a Hewlett-Packard, HP 25 programmable pocket calculator.
3. RESULTS AND DISCUSSION
The diagrams given in Figure 2 show the recoveries of the metal hydroxides from their mixtures as a function of time when flotation was carried out in the absence of DBS
.
T u
60
> K
W
40
>
0 V
W
20
a: 20
LO
60
80
TIME
[scc]
100
120
Fig. 2. Kinetics of recovery of metal hydroxides in the absence of DBS and ethanol
248 and ethanol. It can be seen that the maximal recovery was'within the range of 47.7 to 73.0%. The highest recovery was achieved for iron and copper (73.0 and 72.5%); recovery was somewhat lower for chromium and zinc (64.0 and 62.4%); and was lowest for cadmium and nickel (53.6 and 47.7%). The lower recoveries for cadmium, zinc and nickel hydroxides can be ascribed to the higher values of their solubility products. On the other hand, nickel and cadmium do not behave in the same way in flotation, although they have same solubilities. This can be explained by the different competing effects in the complex flotation system of metal hydroxide mixtures. In Figure 3 are shown the results of the separation of the metal hydroxides from their mixtures in the presence of DBS (4 mg/dm3) and ethanol (0.6%, vol.). We find [ I 1, 121, that this relatively small concentration of DBS in presence of frother ethanol, is sufficient for high recovery of the metal hydroxides, despite of large specific surface areas of these precipitates. The maximal recoveries in this case were within the range from 59.2 (Cd) to 91.6% (Cu and Cr). From a comparison of the curves in Figure 2 and Figure 3 it is evident that the metal hydroxides are separated in the same sequence and their recoveries are increasing with time. Also, it is evident that the recoveries are higher in the flotation using DBS and ethanol. The maximal separation of DBS after 120 seconds of flotation is 66.0%. In Figure 4 is depicted the dependence of the recovery of Cu(OH), on flotation time; the process was carried out at pH 8, with and without presence of DBS and ethanol. As can be seen, the maximal recovery without DBS was achieved after 120 sec. of flotation. The maximum recovery in the presence of collector and frother occurs in about half this time (60 sec). These results, as well as those shown in Figure 3, can be explained on the basis of the following fact. Flotation of both Cu(OH), and the mixture of hydroxides is a physicochemical process which is favored by presence of surfactants. Thus, DBS and ethanol, acting as the surfactants, facilitate flotation processes which take place at the interphases liquid-gas, gas-solid, and liquid-solid.
80
.
T:
U
60
> U
W
> 0 0 W
U
- Cd
20
& 8 -
-
I
I
20
I
I
10
I
I
I
60 TIME
Ni DBS I
I
80
I
100
I
I
120
Cscc3
Fig. 3. Kinetics of recovery o f metal hydroxides in the presence of DBS and ethanol
249
80 n
s
U
> a
60
W
>
LO
0 0 W
0-
20
WITHOUT DBS AND ETHANOL
ETHANOL
30
60
T IM E
90
120
Csccl
Fig. 4. Kinetics of recovery of Cu(OH), in absence and in presence of DBS and ethanol
Although the flotation rates of Cu(OH), and hydroxides mixtures id the absence of collector and frother are lower than in their presence, our results differ in a qualitative sense, from those on investigation of precipitate flotation carried out to the present. We showed that the application of the dissolved air technique can be fairly successful, though no collector and frother are present in the system. Further, the use of dispersed air flotation gave no satisfactory results [ 11, 121. The above findings can be explained by the following facts. When the dissolved air flotation technique is applied, the size of the air bubbles formed is smaller than are obtained with the dispersed air technique. This was confirmed by measuring the bubble diameters after photographing then by means of an MBR-1 microscope (Optlkomekancheskoe obedinenie Leningrad). The average air bubble diameter was obtained by a statistical treatment of the measurement results. Under the experimental conditions of the m, while the average air bubble diadissolved air technique it had a value of 4.38 X meter in dispersed air flotation was 7.47 X 10 m [ l l]. This difference in size of the air bubbles can be explained by the difference in flotation conditions. In the dispersed air flotation, by bubbling air through the porous plate, an overpressure is formed. Due to it the larger air bubbles, moving with a higher velocity, are formed. Contrarily, in the dissolved air technique, during decompression of the pulp, the smaller air bubbles are formed spontaneously and their upward velocity in the flotation cell is considerably smaller than in the previous case. Also, the processes in precipitate flotation of the hydroxide mixture and Cu(OH), by dissolved air technique proceed by a mechanism differing from that in dispersed air flotation. Formation of the particle-bubble complexes in dissolved air flotation can be explained in two ways [ 16, 171: by attaching the resing air bubbles to the particles in the solution; by separation, i.e. ‘precipitation’ of the bubbles directly on particle surface from the air saturated solution. ~
-
250 In contrast to this, precipitate flotation by dispersed air proceeds according to a collision mechanism. Formation of the particle-bubble complex by collision is generally less probable compared to the above mentioned ways. Also, the probability of re-dispersion of the already floated aggregates back into the pulp is higher when the process is going according to the collision mechanism, than in dissolved air flotation. The concentration degrees of the individual metals from their hydroxide mixture for flotation carried out in the absence and in the presence of DBS and ethanol are given in Table 1 and Table 2 , respectively. The percentages of the solutions passing into the foam are also given. The concentration degrees were calculated with respect to the maximum recovery as the ratio of metal (or DBS) concentrations in the foam and in the floated solution. The basis for calculation were the results shown in Figures 2 and 3. It is evident from Table 1 that the concentration degrees of copper and iron are hghest, and they, at same time, show the best recovery. Further, chromium and zinc have somewhat lower concentration degrees, while cadmium and nickel exhibit the lowest values, which is in agreement with their recoveries. Concentration degrees of the metals are higher in the presence of DBS and ethanol (Table 2 ) , than in the absence o f collector and frother (Table 1). At the same time, the percentage of the floated solution passing into the foam is in this case half as large, compared to the runs in the absence of surfactant. The results presented in Table 3 show the concentration degree of Cu(OH)? and percentage of the solution passing into the foam in flotations with and without DBS and ethanol. They were calculated on the basis of the results given in Figure 4. It is evident (Table 3) that the percentage of the solution carried in the foam is higher when flotation Tab. 1. Concentration degrees of metal hydroxides and percentage of solution in foam; flotation without DBS and ethanol Hydroxidesof
Cu
Zn
Cr
Conc.degrees
116.5
73.3
78.6 51.1
% of foam
Cd
Ni
Fe
45.3
119.3
3.4
Tab. 2. Concentration degrees of metal hydroxides and DBS, and percentage of solution in foam; flotation in the presence of DBS and ethanol ~~
Hydroxidesof(andDBS)
Cu
Zn
Cr
Cd
Ni
Fe
DBS
Conc. Degrees
242.2
147.3
244.2
32.4
44.8
234.9
43.2
% of foam
1.9
Tab. 3. Concentration degrees of Cu(OH), and percentage of foam in absence and in presence of DBS and ethanol
Without DBS and ethanol DBS + ethanol
Conc. degree
% of foam
250.7 359.3
16.1 11.3
25 1
of Cu(OH), was carried out in the absence of collector and frother, which is in agreement with the results in Table 1 and Table 2 . In the absence of collector and frother the Cu(OH), particles, as well as the particles of metal hydroxides in their mixtures, are more hydrophillic, causing thus a larger amount of water molecules to be brought up to the top of the flotation system. This then leads to a higher percentage of solution in the foam. This also means a decrease of the concentration degree for Cu(OH), in the foam, if flotation is carried out in the absence of collector and frother. In Figure 5 is illustrated the influence of pH on Cu(OH), flotation without DBS and ethanol. The optimal separation was achieved at pH 8, while at pH 6 the process had the lowest efficiency. This can be explained by the fact that at a pH near 8 copper is already precipitated quantitatively, while at pH 6 the precipitation process is just starting. The results obtained are in agreement (in a qualitative sense, only) with our former results on the pH dependence of Cu(OH), flotation [ l l , 121. But the recoveries achieved here, especially those at lower pH values, are higher than those obtained before. This is due to the above mentioned differences between mechanisms of dissolved and dispersed air flotation. The separation process of Cu(OH), in the absence of DBS and ethanol at pH 6 showed a linear dependence of the recovery on the flotation time, This corresponds to an empirical equation of the type y = ax + b. The same process at pH 7, 8 and 9 followed power law dependence of the recovery on time, expressed by an empirical equation of the type y = axb'. Table 4 contains the a and b' values for the kinetics of Cu(OH), separation at different pH values. It is evident that the highest separation rate for Cu(OH), was at pH 8, which corresponds to the optimal flotation conditions. The dependence of concentration degree for Cu(OH), and the percentage of solution passing into the foam on pH values are given in Table 5. The maximum concentration degree at pH 8 corresponds to the highest recovery of Cu(OH), at the same pH value. In view of the hydrophilic properties of Cu(OH), at the maximum recovery, the maximal amount of the foamate, corresponds also to the same pH value.
.
80
6
U
60
LO 20
30
60 T IME
90
120
[scc]
Fig. 5. Dependence of flotation of Cu(OH), on pH in absence of DBS and ethanol
252 Tab. 4. Dependence of flotation rate constants from pH values in absence of DBS and ethanol PH
6
7
8
9
Constants (%/set.)
0.12
2.53
2.62 2.55
Tab. 5 . Dependence of concentration degrees of Cu(OH), and percentage of foam on pH values
PH
6
Conc. degree % of foam
70.6 67.1 250.7 16.1 2.0 10.6
7
8
9 78.0 13.6
As a conclusion, it could be said that the results of these investigations indicated the possibility of their practical application for purification of the wastewaters containing metal ions. Further investigation will be carried out on actual wastewaters.
REFERENCES
1 R. S. Baarson, C. L. Ray, Precipitate flotation, a new metal extraction and concentration technique, American Institute of Mining, Metalurgical and Petroleum Engineers Symposium, Dallas, 1963. 2 F. Sebba, Ion flotation, Elsevier, Amsterdam, 1962. 3 A. I. Macnev, Ochistka stochnikh vod flotaciej, Budiveljnik, Kiev, 1976. 4 D. J. Wilson, E. L. Thackston, Foam flotation treatment of industrial wastewaters: Laboratory and pilot scale, EPA-600/2-80-138, June 1980. 5 R. Lemlich, Adsorptive bubble separation techniques, Academic Press, New York, 1972. 6 T. A. Pinfold, Separation Science, 5 (1970) 379-384. 7 D. Bhattacharya, J. A. Carlton, R. B. Grieves,A. I. Ch. E. Journal, 17 (1971) 419-424. 8 R. B. Grieves, S. M. Schwarz, J. Appl. Chem., 16 (1966) 14-17. 9 K. S . Kalman,G. A. Ratcliff, Canad. J. Chem. Eng., 49 (1971) 626-632. 10 L. Dobrescu, V. Dobrescu, Epurarea apelor uzate prin flotatie, Ministerul agriculturii, industriei alimentare, silviculturi si apelar, Bucuresti, 197 1. 1 1 D. Miskovid, The recovery of metal precipitates and surfactants by flotation, Ph. D. Thesis, Faculty of Sciences of Novi Sad, Yugoslavia, Novi Sad 1977. 12 E. Karlovid, The recovery of heavy metal hydroxides from metal finishing industries wastewaters by flotation with surfactants, M. Sc. Thesis, Faculty of Sciences of Novi Sad, Yugoslavia, Novi Sad 1977. 13 A. N. Clarke and D. J. Wilson, Foam flotation: Theory and applications, Marcel Dekker, New York 1983. 14 IUPAC, Tables of spectrophotometric absorption data of compounds used for the colorimetric determination of elements, Butterworth, London, 1963, p. 148. 15 AmericanF’ublic Health Assocciation, Standard methods for the exanlination of water and wastewater, XI1 Edition, New York 1965, p. 296. 16 V. I. Klassen, An Introduction to the theory of flotation, Butterworth, London, 1963, pp. 89-138. 17 I. L. Schmidt, J. A. Hajnman, A. V. Proskujakov, Zh. Prikl. Khim, 4 3 (1970) 2553-2558.
253
HOT WATER PROCESSING OF US.TAR SANDS WATER RECYCLE AND TAILINGS DISPOSAL
J. HUPKA, A. G. OBLAD Department of Fuels Engineering J. D. MILLER
Department of Metallurgy and Metallurgical Engineering University of Utah, < Lake City, Utah 84112-1I83
ABSTRACT Tailings disposal and water recycle are vital factors for the industrial development of a hot water process for the recovery of bitumen from US. tar sands. Based on batch laboratory experiments, with 4 to 12 kg tar sand charge, tailings sedimentation behaviours are presented and discussed for tar sand samples originating from six U.S. deposits. Recycle of 90% of the water is possible after 10-20 minutes clarification assuming 10% solids content in the flotation cell. Sand and silt, which make up about 95 -99% of the total quartz minerals present in five domestic tar sands (with the exception of McKittrick), settle completely (40% porosity with water saturation). Very fine particles (< 25 pm) which amount to up to 10% of the tailings stream still retain about 30 wt% water after 2 weeks of thickening.
1. INTRODUCTION
Future energy demand may be the development of technology dedicated to the production of synthetic crude oil from alternate fossil fuels such as coal, oil shale, and tar sands (oil sands). Heavy oil reserves in tar sand deposits in the world are estimated at one and a half times those of light oil [l]. In the USA, about 85 per cent of all known tar sand resources are located in Utah [ 2 , 31. Of the above-ground recovery technologies of bitumen from oil sands, the hot water separation process ( W P ) appears to be the most promising for separation of bitumen from sand and has been used industrially for the past fifteen years in Canada [4, 5 , 6 , 71. One of the weak points in the water processing of tar sands is the application of large amounts of water, part of which may be lost with the spent sand, creating the additional problem of tailings disposal. This concern is compounded by the limited water resources available for industrial development in arid regions such as Utah. Athabasca tar sands
254 Lumps of mined tarsand
Tailings
10 mesh
C LAR IFICATI0 N 5 30 min 506OOC
I
LT’1 50-60 O C
Slime ’L
D ILUENT RECOVERY
d ih e nt
further proccessing
I
w
WATER 8 SAND SE PARAT I0N
THICKENING (ponds)
’ -
SKIMMING Water for recycling Sand for land reclamation or. further utilization
Fig. 1. Process flow sheet for bitumen recovery from tar sands.
contain a significant amount of fine clay minerals which after easy processing remain suspended indefinitely. This problem attracted special attention to the ”tailings problem” [8, 9, 10, 11, 121 and even fostered the opinion that hot water separation should be abandoned due to difficulties associated with tailings treatment and disposal. However, not all tar sands contain such clay minerals which would diminish their value as feed for hot water processing. US. tar sands are substantially less than those of Canada. The U.S. tar sands seem to contain a smaller amount of clay minerals. In this regard, the potential of US. tar sands as an energy resource is under intensive investigation for bitumen
255
SAND GRAIN
FINE SAND BITUMEN AQUEOUS PHASE
Fig. 2. Tar sand pulp after digestion.
recovery by many methods [ 131. In this paper the question of tailings disposal and water recycle in the H W P of U.S. tar sands is considered in terms of the impact of these processing steps on the environment.
2. STRATEGY AND ENVIRONMENTAL CONSIDERATIONS IN HOT WATER PROCESSING OF TAR SANDS
Tar sands are deposits of consolidated or unconsolidated clastic sediments (e.g., sandstone, limestone, diatomite) that have pore spaces partially or completely saturated with a heavy, tar-like hydrocarbon mixture known as bitumen. Similar to heavy oil, tar s a n d i t u m e n i s a member of the petroleum family of organic substances that cannot be economically recovered by the relatively simple techniques used for recovery of lightcrude oil [ l , 141. As discussed elsewhere [15 J tar sands, which contain bitumen of viscosity less than lo3 Pas at 90°C with the mineral matter composed mostly of quartz sand, can be processed with high recovery using a hot water separation technique. A moderate temperature process (SO-60°C) developed at the University of Utah is especially promising because of significant energy savings. A flow sheet of this process, applicable for many tar sands and independent of the bitumen grade and bitumen viscosity, is shown in Fig. 1 . A drawing of the tar sand pulp after digestion is presented in Fig. 2. The bitumen is completely disengaged and separated from sand grains. The bitumen particles have a hydrophobic character which accounts for easy recovery by flotation. Large bitumen droplets float within the first 20 seconds. The small bitumen droplets require a longer flotation time. The sand particles are clean and largely liberated from the bitumen phase. The tailings stream contains sand and varying amounts of suspended fines (depending on the tar sand deposit). The behaviour of this suspension is the subject of our investigation, and discussion of experimental results are presented in the next sections. Tailings disposal behaviour, based on 4 to 12 kg batch experiments, is discussed for the sand deposits described in Tab. 1 . Residual water present in the tailings must be entirely recovered or eventually removed from the sludge by evaporation in order to avoid penetration to ground water. Environmental problems connected with tar sands processing can be related to three
256 Tab. 1. Characteristicsof tar sand samples used in separation tests
Origin of tar sand
Bitumen content (wt%)
McKittrick (CA) 13.2* Asphalt Ridge (UT) 11.5 9.5 Sunnyside (UT) 11.0 Wyoming I 3.7 M. West VII 5.8 M. West VIII
Bitumen viscosityfPas)
Fine sand < 100 pm content
50°C
90°C
(wt%)
Sand alkalinity** (pH)
1 .o 48 1500 43 25 24
0.1 1.2 18 1.2 1.0 0.8
35 8 25 43 12 11
8.8 8.3 7.7 7.8 7.7 7.8
* toulene extract, but due to diatomaceous earth content a few additional percents of organic material are adsorbed and not included ** pH of aqueous extract of sand after bitumen extraction with toulene basic areas [8, 16, 17, 18, 191: air, water and land. In the not water processing of tar sands air pollutants include sulfur in various compounds, particulates, hydrocarbons, carbon monoxide and products of combustion [20]. Land-surface contamination is due to the amount of area needed for mining or processing operations [17]. Surface-water or ground-water contamination, however, as a result of accidental leakage or spillage, as well as deliberate discharge of tailings slurry, is the greatest threat for the environment. In a study that involved the discharge of tailings sludge into a small part of a river in northern Alberta, Canada, a 60% reduction in the standing stock of benthic invertebrates occurred throughout a four-week period in the immediate area of spillage. The tailings sludge was characterized as at thikslurry of inorganic particles in the clay, mixed with globules of tar-like hydrocarbon material suspended in an aqueous solution of variety of organic and inorganic compounds. The material affected the benthic invertebrates and other aquatic life in two basic ways: (1) through the toxic effects due to the organic and inorganic compounds, and (2) by covering breathing, feeding or living surfaces with fine particulates. It was recommended that appropriate measures be established to prevent the addition of oil sands tailings sludge, be either accident or design, to lakes or rivers because the tailings sludge constituted a principal hazard to the aquaric communities.
3. TAILINGS TREATMENT
3.1.Water Loss and Mass Balance
Hot water separation of bitumen from tar sands involves a large amount of process water, a part of which is discarded with the damp sand discharge, provided that fine mineral particles settle well. In solvent extraction processes, water loss with sand discharged is a problem of similar importance when solvent is recovered from spent sand by steam stripping or hot water displacement. Assuming that water completely flls all free spaces between sand particles and the porosity is 40%, the theoretical loss of water depends on tar sand grade. For 10% tar sand grade water loss with sand is 2.2 bbl H,O/ bbl bit.
257
Fig. 3. Water mass balance for Sunnyside tar sand processing. Intrinsic water present in the ore has been ignored.
By filtration (e.g., belt filter), about 75% of this water can be removed which makes the loss significantly smaller (0.6 bbl H,O/bbl bit for 10% bitumen in tar sand). An inherent assumption in the case of filtration is made that residual bitumen and fine sand will not permanently clog the filter cloth. Spills and/or evaporation may also significantly contribute to overall water loss when appropriate precautions are not considered. Any improvements in the process resulting in higher recycle ratio will importantly influence processing costs. Water balance for laboratory experiments carried out for 4 kg charge of tar sand is presented in Fig. 3. Flotation is the unit operation which requires the largest amount of water. In the present laboratory set-up the flotation separation is accomplished at 10% weight solids. Experiments with a 15 dm3 capacity reactor and 12 kg charge of tar sand has shown that the solids present during flotation can be increased to 30% with no significant decrease in bitumen recovery.
258 3.2. Sand Sedimentation
The first step in tailings treatment - screening and recycle of middlings (see Fig. 1) is aimed at the recovery of undigested feed material. Undigested aggregates are characteristic of consolidated sandstone [21] but also may appear under certain conditions in the unconsolidated tar sand. The next important steps in tailings treatment are : - fast separation of sand from water in settler of residence time not longer than 0.5 to 2.0 hours (clarification), and - water removal from sand in tailings ponds (thickening). These two unit operations, though very similar in basic principles, differ significantly in the final objective which is intended to be achieved. Preliminary seetling aims at as much water recycle as possible, thus avoiding a dilute sludge thickening in settling ponds. Tailings discharge to ponds for final sedimentation will cool down to ambient temperature, which means significant energy loss. Tailings ponds periodically filled with sand slurry should serve for sand dewatering to final water content below 15-20 weight %. This value corresponds to nearly complete water saturation of the sand. The last unit operation in tailings treatment is oil removal from the surface of tailings ponds. Residual diluted bitumen (oil) always accompanies the tadings stream directed for thickening. Part of the oil is released and floats to the surface. This slick creates a threat for aquatic birds, slows down water evaporation (though this may be beneficial in some circumstances), and may cause secondary contamination of sand when water is recovered through the drainage system. Unlike Athabasca tar sands [22], tailings from Utah tar sands separation contain only 0.6% and less bitumen in laboratory scale [23] and below 0.4% in pilot plant operation [24] due to low clay content. Investigations of sand sedimentation in tailings water have accompanied most hot water separation experiments. Two-liter 50 cm cylinders and Imhoff Cones served as laboratory settlers. Content of solids in water was within the 10-30% level. Two types of sedimentation behaviour were observed in these experiments which are characterized by the drawings in Fig. 4. The type of settling designated as "A" pertains to fines which easily coagulate and form large flocs. Middlewest U.S. (M. West) tar sands are an example of such settling behaviour. At the beginning of sedimentation, Zone 11 is the most voluminous, but after about 30 minutes Zone I (supernatant zone), containing nearly clear water (< 100 mg/l), prevails. In group B in Fig. 4, another behaviour of sand during sedimentation is presented. This type of settling was observed for most Utah tar sands. Fine solid particles flocculate slowly and settle as tiny flocs. Zone I1 does not exist at a l l in the very beginning. However, after several minutes it is noticeable and increases slowly during the first 60 minutes, despite occurring compression. Residual concentration of sand in Zone I gradually decreases, reaching, after two hours, values close to those obtained for Type A. Settling curves describing fines content in Zone I (Sunnyside, Asphalt Ridge and Wyoming tar sands) or the velocity of compression of Zone I1 (M. West tar sands) are presented in Figs. 5-7. Fast clarification of the aqueous phase was found for all tar sands tested at given processing conditions. Procesing temperature, kerosene application in the pretreatment step, and even 30% solids in the flotation cell (instead of 10%) have only a minor effect on the settling time. Higher concentration of solids in the flotation cell
259
A.
ZONE 1 I (Vz) ZONE Ill (V,) 1min
5 min
20 min
v, =v,+v,+v,
>
120 min
( V -Votume)
B.
ZONE I (V,l
ZONE I I (V,)
Z O N E I I I (Vs) 20 min
5min
1min
> 120 min
Fig. 4. Settling behaviour of sand in tailings from hot water processing of tar sands.
I
I
I
I
1
I
A 7OoC 00
0
ASPHALT RIDGE
6OoC 30"hSOLIDS 55OC 107LSOLlDS
00
0
0
20
40
60
80
100
120
SETTLING TIME, rnin Fig. 5 . Change in fine sand content in Zone I during Sunnyside and Asphalt Ridge tar sand tailings sedimentation.
260
800
d
600
a
PROCESSING TEMPERATURE. o 37OC 25’/0 KEROSENE A 86OC NO KEROSENE
I
I
I
I
I
20
40
60
80
100
120
SETTLING TIME, min Fig. 6. Change in fine sand content in Zone I during Wyoming I tar sand tailings sedimentation.
PROCESSING TEMPERATURE 1
I
I
0
20
I
40
0
620C]MWESTVII 55OC
0
55OC M.WEST Vlll
I
1
60 80 TIME, min.
Fig. 7. Seetling curves for tailings from M. West tar sands.
1
100
1
1
1 I 120
261
+30
+ 20 0 Sand,Sunnyside + Dispersed kerosene
+10
> E
0 -1 0 Bitumen
-20
\
-30 0
a
-4 0
Q I-
-50
N
-6 0
w
-70
+
-+-++--+F-
Sand II
I
l
l
I
I
I
I
I
l
l
Fig. 8. Zeta potential curves for fine particles present in tailings from hot water processing in tar sands.
automatically results in a smaller flow to be treated. Therefore, the size of settlers is the same as for more diluted tailings. Only the residence time varies. Tar sands from various deposits contain different amounts of fine sand and different mineralogy. This fact, besides changing solids concentration is tailing water, influences the settling behaviour of tailings. Nevertheless, tailings from tar sand which contains less fine sand should not necessarily settle faster than tailings from tar sand of higher fines content as can be seen in Fig. 5 for Sunnyside and Asphalt Ridge ore. Tailings from McKittrick tar sand showed a combined settling mechanisms. Most fine sand is aggregated in large flocs and settles in 15 minutes forming Zone 11. Zone I1 after 15 minutes undergoes only small volume changes. The residual content of fines in Zone I is much higher than for other tar sands and remains as high as 500 mg/l after 2 hours sedimentation. These fine particles of colloidal dispersion require 6-12 hours to reach a concentration of less than 100 mg/l. The ability of suspended fine particles to aggregate can be directly related to the surface charge. Electromobility of mineral particles from domestic tar sand tailings is much lower than for Athabasca tar sands. For well-settling Sunnyside and Wyoming tailings the zeta potential as below -25 mV, while for Athabasca tailings it is -45 mV and -65 mV (pH = 7-8) as presented in Fig. 8. Lower surface charge accounts for easier flocculation and faster sedimentation of fines in the tailings from the H W P of U.S. tar sands. Kerosene addition in the process may hinder the flocculation process to some degree, as can be seen in Fig. 6 . The surface charge on kerosene droplets suspended in centri-
262 fuged aqueous phase from the flotation cell is higher than for Ties originating from
HWP of U.S. tar sands. Eventual kerosene adsorption on tailings sand can be responsible for slower sedimentation; therefore, excessive addition of diluent to the ore in the pretreatment step should be avoided. Clarification rates in Zone I indicate how much water can be recycled to the process in a short period of time. Sand settling in Zones I1 and I11 provides preliminary data with respect to tailings thickening in tailings ponds. Effective dewatering, especially of very fine and colloidal sand from Zone I1 will determine the size of tailings ponds and the ultimate land reclamation system. For Athabasca tar sands, tailings occupy up to 30% larger volume than the original ore due to water adsorption by clay components [9]. When the slurry is used for refilling excavation pits the pit volume is insufficient and some extra space for tailings is required. Study of thickening of domestic tar sand tailings reveals that water content in Zones I1 and I11 should not cause difficulties similar in scale to those experienced with Athabasca tar sands. Experiments were performed in the following way; after 2 weeks of thickening in laboratory conditions, water from Zone I was allowed to drain through Zone I1 and Zone I11 by means of a drip valve at the bottom of the cylinders. When Zone I had disappeared, water content was determined in the slime from Zone I1 and coarser sand in Zone 111. The results for Sunnyside tar sand indicate that water content in Zone I1 was 32 wt% after 1 hour percolation time and 29 wt% after 24 hours percolation time. Residual saturation of sand with water was 168% after 1 hour and 153% after 24 hours percolation time. Coarse sand in Zone 111 was saturated in 104% which corresponds to 20 wt% water content. Saturation was independent on percolation time. The flow rate measured for our experimental cylinders containing approximately a 2 cm thick slime layer and 10 cm thick sand layer was 6 m3/mZh. The data have shown that coarse sand does not tend to retain excessive amounts of water, unlike the very fine sand forming Zone 11. Microscopic analysis revealed that mineral particles found in Zone I1 are smaller than 25 pm, most of them being in the 5-10 pm region. Clay particles, making up to 1-2% of the total sand, are below 2 pm and remain attached to the surface of larger grains which consist of 95% quartz and oligoclase. Sand in Zone I1 retains about 30% of water after 2 weeks of thickening and several hours of dripping under the influence of gravity. The weight ratio of sand in Zone I1 to sand in Zone I11 is approximately 1:6 for Sunnyside tar sand and smaller for other tested tar sands (relatively less sand is contained in Zone 11). Therefore, it can be concluded that the total volume of dewatered tailing slurry should not occupy a larger space than the original ore. An exception may be the McKittrick deposit. Due to the diatomaceous earth content, Zone I1 is very voluminous for this tar sand and retains much more water. Further study on settling behaviour for McKittrick tar sand, eventually with flocculants application, is necessary. Tailings ponds with drainage systems are recommended for other domestic tar sands in which the mineral part is mostly composed of quartz sand. Thickening in tailings ponds probably will take several weeks, as data in Fig. 9 indicates. A drainage system should reduce this period several times and a study on this problem is in progress. Due to relatively fast water percolation such ponds can be used periodically for: slurry filling in and dewatered sand removal cycles.
263
Preliminary investigation of properties of tailings produced in the Enercor* pilot plant (50 bbl per day of bitumen), while operating at low alkalinity, confimied our expectations about the amount of water which can be recycled and sand utilization for land reclamation. Alkalinity of the aqueous phase during processing was consistently within the pH range of 7.5-8.6. Such alkalinity is mainly a result of the presence in the ore of carbonate minerals and only 0.5-2.0 g of Na,CO, per kg of tar sand were added in some experiments. Nevertheless, a highly alkaline process mode can be used for bitumen recovery from selected tar sands [26]. High alkalinity induces the release of a substantial amount of surface active compounds which in turn facilitates bitumen disengagement from the sand. For less viscous bitumen with extraction proceeding at elevated temperature, (95"C), n o diluent has to be used prior to or during digestion and flotation. However, the tailings treatment may face serious difficulties due to slow sand settling, lack of colloidal fines flocculation and strong water retention in Zone I1 (see Tab. 2 for data on Asphalt Ridge tar sand).
3.3. Dissolved Organic Compounds and Water Recycling
The presence of organic compounds in the aqueous phase, dissolved from bitumen, was observed indirectly by the change in color of the aqueous solution originating from the digester and from the flotation cell. The color intensity of the aqueous solution is a function of the temperature of digestion, penetration time and alkalinity. With increasing temperature and alkalinity the color of the aqueous phase changes from light yellow (temperature 50-60°C and pH = 8.4) to dark brown (temperature 90-95°C and pH > 10). The chemical composition of bitumen for specific tar sands determines the color shade and intensity. In Tab. 3, the processing conditions are correlated with total organic carbon content (TOCC) for Sunnyside and Wyoming I tar sands. Surface tension of the aqueous phase changes only slightly with increasing content of dissolved organic compounds, which means that these compounds do not exhibit strong surface activity in higher concentrations. Samples of air, digester water and bulk bitumen were collected during some experiments which were carried out at a processing temperature of 50°C. These samples were analyzed for benzene solubles and polynuclear aromatic hydrocarbon (PNA) content [27]. Standardized analytical methods were used to deterfnine the benzene solubles and PNA. Benzene solubles and PNA identification in the samples of air and digester water were neghgible. Bulk bitumen samples contained a relatively large quantity of benzene soluble compounds and nine different PNA's were identified [27]. Since potentially carcinogenic PNA's were identified in the bulk bitumen, further research is necessary to determine the hazards both apparent and potential in the production of synthetic fuel from tar sands. From an industrial hygiene standpoint, very little research has been carried out by those interested in tar sand commercialization as judged by available literature published [ 17,281.
* Enercor - a corporation engaged in tar sand development located in Salt Lake City, Utah. The pilot plant was based on University of Utah hot water extraction technology.
264
100 0
z - 80 t-
Z w + t
A
<5
0
Wyoming Asphalt Ridge Sunnyside
- 60 -
00
nz
3 40
s
20
W 0 I- N
r
01 0
I
I
100
200
I
300
400
THICKENtNG TIME (hours) Fig. 9. Compression of Zone I1 with time for Wyoming I, Asphalt Ridge and Sunnyside tar sands.
Tab. 2. Comparison of settling ability of sand for high and low alkalinity hot water process for Asphalt Ridge tar sand Alkalinity of aqueous phase in flotation cell (pH) Settling ability of sand ~~~
7.8
> 10
~
Settling time of coarse sand 0.5-1.0 Zone 111 (min) Final formation of stable 1-2h Zone I1 Water content in Zone 11 (wt%) 2 hours 90 1 day 15 14 days 34 Residual fines content in Zone I (mg/l) 2 hours 220 24 hours < 10
0.5-1.0
- 14days Zone 1 &I1 cannot be distinguished clearly 99
91
> 500
- 100 stable colloid
A high water recycling ratio may cause accumulation of fine sand and dissolved organic compounds in the aqueous phase. Data presented in Tab. 4 reveal, however, that separation efficiency expressed by the coefficient of separation is comparable tb that with fresh water. Increase of dissolved organic compounds, leached from the bitumen, as well as a moderate increase in the amount of fines in the aqueous phase, does not affect the process. All water, with the exception of that discarded with dump sand, was recycled in the Enercor pilot plant through the entire operation time. No deleterious side effects resulting in inferior separation efficiency were observed.
265 Tab. 3.Properties of the aqueous phase from hot water processing of Sunnyside and Wyoming I tar sands Digestion time: 15 min Flotation time: 10 min Percent solids in reactor: 75 Percent solids in flotation cell: 10 Diluent used: kerosene Feed size: < 5 mm Na,CO, addition: 0.75 g/kg t.s. Processing temperature CC)
Penetration time (hours)
71 72 72** 77 64* 38
0 3 6 16 100 16
Alkalinity (pH)
Surface tension (mN/m)
Reactor
Flot. cell
8.3 8.4 8.1 8.3 8.4 7.9
8.4 8.4 8.4 8.4 8.3 8.3
7.5 8.4 8.5
8.6 8.0 8.4
Reactor
Total organic carbon
Flot. cell
Reactor
Flot. cell
62 64 61 55 53 62
160 190 320 390 600 145
19 18 27 38 61 17
55
306 95 120
28
Sunnyside
86** 58** 37**
60 70 58 54 51 58
Wyoming I
0 0.5 22
43 56 56
-
67
-
15
* feed size: 7-3
** no Na,CO,
mm added
Tab. 4.Water recycling tests in hot water separation of Asphalt Ridge tar sand Processing temperature: 50°C Diluent addition: 20% (with respect to bitumen) Tailing settling time: 1 hour (90% water recycled) Number of times recycled Fresh water
I I1 I11
X
Akalinity (pH)
Surface tension (mN/m)
Digestion
Flotation
Digestion
Flotation
Coefficient of separation*
8.4 8.3 8.4 8.4 8.5
8.1 8.0 8.1 8.2 8.2
49 51 46 49 50
62 64 65 64 63
0.91 0.89 0.90 0.92 0.91
* measure of process efficiency, percent of the feed which undergoes perfect
separation
4. FINAL COMMENTS AND CONCLUSIONS
The low level of clay content in domestic tar sands allows optimistic predictions with facilities. Our laboratory respect to tailings disposal and water recycle experiments support such promising expectations. Tailings from domestic tar sands contain 0.2-0.6% bitumen whereas tailings from Athabasca tar sands have 0.9-1.1%
266 bitumen [29]. One has tb bear in mind, however, that results from Athabasca tar sands processing were obtained on a commercial scale, whereas domestic tar sands were processed only a batch laboratory set-up or in a pilot plant. Water recycle and tailings disposal seemed simple and not worthy of great attention before the opening of the first commercial-scale plant in Canada [30].Today, after 16 years of operation, the tailings problem is the greatest of the entire hot water technology and is still waiting for a successful solution, though numerous patents and papers suggest vital improvements. The promising results with water recycling and tailing thickening obtained in the present work should therefore be viewed cautiously. On the basis of the laboratory experiments the following conclusions can be drawn: - Natural flocdation and settling of fine particles in a low alkalinity environment is fast and after 10-20 minutes 90% of the water can be recycled. Residual concentration of sand in recycled water is lower than 500 mg/l. - Two mechanisms of fine sand sedimentation were observed according to its ability to form large flocs. - Thickening of coarse sand in settling Zone 111 resulted in complete sedimentation, while fine sand in Zone I1 retained about 30% of water. Nevertheless, the total volume of dewatered tailing sand should not be larger than the original tar sand ore. -Water recycle had no deleterious effect on the separation efficiency despite the slight increase of dissolved and suspended matter in the recycled aqueous phase.
ACKNOWLEDGEMENT
We acknowledge gratefully the funding for this work by the US. Dept. of Energy (Laramie Energy Technology Center) contract # RE-HS20-82LC10942 and the University of Utah.
REFERENCES
1 2 3 4 5 6 7
W. Bonse, Erdoel-Erdgas-Zeitschrift,7 (1) (1981) 19. C. Q. Cupps, L. C. Marchant, World Oil, Sept. 1979, 73. L. C. Marchant, D. Jackson, Engineering and Mining J., June 1981, 130. N. Berkowitz, J. G. Speight, Fuel, 5 4 (1975) 138. L. Gabinet, Canadian Petroleum, May 1981,28. J. Muir, Canadian Petroleum, May 1981, 40. G. L. Baughman, “Oil Sands” in Synthetic Fuels Data Handbook, Cameron Engineers, Inc., Denver 1978, USA, p. 259. 8 C. W. Bowman, G. W. Govier, Tenth World Energy Conference, 19-24 Sept. 1977, Istanbul. 9 F. W. Camp, Can. J. Chem. Eng., 55 (1977) 581. 10 M. B. Hocking,G. W. Lee, Fuel, 56 (1977) 325. 11 A. Thomas, Canadian Petroleum, April 1981, 51. 12 E. S . Hall, E. L. Tollefson, Can. J. Chem. Eng., 6 0 (1982) 812. 13 P. M. Kohn, G. Parkinson, W. P. Stadig, Chemical Eng., April 1983, 18. 14 R. J. Byramjee, Oil Gas J., 81 (27) (1983) 78. 15 J. Hupka, J. D. Miller, A. Cortes, Transactions AIME, accepted for publication. 16 S. S. Penner, S. W. Benson, F. W. Camp, J. Clardy, J. Deutch, A. E. Kelley, F. X. Mayer, A. G. Oblad, R. P. Sieg, W. C. Skinner, D. D. Whitehurst, Energy, 7 (7) (1982) 567.
267 17 J. I. Daniels, L. R. Anspaugh, Y.E. Ricker, G. J. Rotariu, Health Impacts of Different Sources of Energy, International Atomic Energy Agency, Vienna, 1982, p. 595. 18 E. Lbsz16. Ambio, 10 (5) (1981) 254. 19 S. Aronoff, G. A. Ross, W. A. Ross, Photogrammetria, 38 (3) (1982) 77. 20 A. Kumar, Environ. Sci. Technol., 13 (6) (1979) 651. 21 J. Hupka, A. G. Oblad, J. D. Miller, to be published. 22 C. E. Denis, M. A. St., Kessick, Can. J. Chem Eng., 60 (1982) 675. 23 J. Hupka, J. D. Miller, SME-AIME Annual Meeting, 14-18 Feb. 1982, Dallas, Texas, USA, Preprint No. 8283. 24 K. Hatfield, J. D. Miller, A. G. Oblad, Seventh Symposium of the Rocky Maountain Fuel Society, Feb. 19-21, 1982, Salt Lake City, UT, USA. 25 C. W. Bowman, Seventh World Petroleum Congress, Mexico City, 1967, Proceedings Vol. ;, p. 583. 26 J. E. Sepulveda, J. D. Miller, Mining Eng., 30 (9) (1978) 1311. 27 M. C. Geraci, M. S. thesis, University of Utah, 1983. 28 W. H. Calkins, J. F. Deye, R. W. Hartgrove, C. F. King, D. F. Krahn, Fuel, 62 (1983) 857. 29 E. S. Hall, E. L. Tollefson, Energy Processing, 72 (1980) 39. 30 E. D. Innes, I. V. D. Fear, Seventh World Petroleum Congress, Mexico City, 1967, Proceedings, VoL 3, p. 633.
This Page Intentionally Left Blank
269
THE ROLE OF DILUENT IN OILY WASTE WATER TREATMENT IN BED COALESCERS
J . HUPKA
Department of Fuels Engineering University of Utah Salt Lake City, UT.84112, USA
ABSTRACT Periodic or continuous kerosene injection into an oily waste water stream has been shown to cause liberation of a coalescer bed from accumulated oil and suspended matter. Thus the lifetime of the bed was extended several times with respect to an unregenerated one, without a decrease of coalescence efficiency.
1. INTRODUCTION
Mathematical approaches to evaluate the filtration coefficient proposed by various authors [ 1 - 41 describing a steady state coalescence process do not show direct dependence of the coalescence efficiency on the viscosity of the dispersed phase. A general assumption, however, lies at the basis of these models, which is that viscosity of the dispersed phase is close to that of the continuous phase. The continuous phase is usually water, liquid fuels or hydrocarbon solvents, which means that liquids of low viscosity are expected to be involved in separation in a coalescer bed. There are only a few reports in the literature on the dependence of filtration coefficient on oil viscosity [4, 5, 61 indicating generally hardly any influence of dispersed phase viscosity on the coalescence effectiveness. An assumption can therefore be made on the basis of existing models describing performance of bed coalescers, that diluent dispersion should ungergo coalescence in the same degree as droplets of heavy oil. The capacity factor of the coalescer depends strongly, however, on the oil viscosity. The relationship between capacity factor and oil viscosity presented in Fig. 1 is based on many laboratory experiments and field tests. Although operating conditions and breakthrough time varied for different experiments, the decrease of the capacity factor with increasing oil viscosity can be seen clearly. For coalescer beds formed from very fine and densely packed fibres the capacity factor may even not exceed 10 for light oil of viscosity close to that of water. The present work focuses on the coalescence in bed efficiency for O/W dispersions when korosene as diluent is applied as the process aid. The coalescer bed is formed from granular and fibrous material preferentially wetted by the dispersed phase.
270 2. DILUENT APPLICATION IN THE PROCESS
Hydrocarbon solvent used for heavy oil dilution can be blended with the O/W emulsion (dispersion) stream by means of an agitated tank preceding the coalescer separator [7] or injected into the emulsion a moment prior to its entry into the coalescer bed [8]. Both techniques are presented schematically in Fig. 2. Diluent application during oily water treatment in coalescers can involve two different strategies of bed regeneration: 1. Continuous diluent addition in amounts depending on oil content and viscosity. 2. Periodic diluent addition according to maximal and minimal pressure drop across the bed. In the first method diluent prevents excessive deposition of heavy oil in the bed because coalescemce and dilution occur simultaneously. Diluent is introduced into the waste water as long as it contains heavy oil. In the second method oil is allowed to build up in the bed until maximal pressure drop across the bed is obtained. From this moment diluent is added into the emulsion until the head loss in the bed decreases to its minimum value. Pressure limits are necessary to avoid irreversible clogging of the bed with oil (maximum limit) and to sustain a residual saturation of the bed with oil (minimum limit) necessary for continuation of effective coalescence. Numerical values for maximal and minimal pressure drop vary due to coalescer bed properties and processing parameters. Another regeneration technique may be considered involving coalescer bed washing with an organic solvent [9, lo]. A significant volume of solvent is used during bed regeneration and coalescer is not fed with emulsion at that time. Such a process is not presented in this work. Backwashing of the bed with clear water and steam application were briefly discussed earlier [8]. 3. FLOW OF O/W DISPERSION AND DILUTION CONDITIONS IN COALESCER BED
A coalescer bed formed from granules or fibres has a different structure (see Tab. 1) which determines coalescence efficiency and should influence the extent of oil dilution when hydrocarbon solvent is injected into the O/W dispersion. A single-phase flow through porous media, assuming that there are no fluid-bed interactions, can be described by the following equation : 1c dp
p=---
dx
...@
where :
Q
v=-
A
apparent fluid velocity
k
bed permeability
P
viscosity o f the flowing liquid
dP dx Q A
pressure gradient in the direction of flow volumetric flow rate cross-sectional area of the bed
(1)
27 I
0.01 0.1 OIL VISCOSITY [Pa
0.001
1.o 53
Fig. 1. Capacity factor vs. oil viscosity.
5- -7 COALESCE Oil R
@
AGITATED TANK
-
ggj
Water
1
>
-
<
Diluent
PUMP
PUMP
A
Fig. 2. Two methods of introducing dilutent into the system: A - using agitated tank, B - using nozzle.
Piluent
B
272 Tab. 1. Comparison of granular and fibrous beds in coalescer Diameter of grain or fibre Kind of bed (mm) Granular
Fibrous
1.25 0.75 0.50 0.018 0.010 0.002
Packing density* g/cm3
Porosity
Specific surface** m2/g
polyethylene
glass
POlYethylene
POlYglass ethylene
0.58
1.58
0.38
0.38
0.24
0.64
0.75
0.75
0.5X lo-* 0.9X lo-' 1.3X lo-' 0.24 0.43 2.13
glass
Approx. size of void spaces (mm)
0.2X 10" 0.3X lo-' 0.5X lo-' 0.09 0.16 0.78
0.2-0.5 0.1-0.3 0.08-0.2 0.01 -0.04 0.005 -0.03 0.001-0.004
* Calculated for polyethylene density of 0.94 g/cm3 and glass density 2.55 g/cm3
** Calculated for smooth surface
from which the Darcy equation for laminar flow is derived :
An examination of the Darcy equation shows that for a given pressure drop (AP= P, - P,), lack of alteration in the structure of a porous medium (expressed by k) and length (L), the flow is inversely proportional to fluid viscosity. In the case of two-phase flow, when the dispersed phase is much more viscous than the continuous phase, the pressure drop necessary to cause a flow of the continuous phase is insufficient to force the coalesced phase through the porous medium at such a rate that equilibrium saturation of the bed can be achieved. Moreover, compressible beds (like fibrous beds) undergo structural changes when the head loss exceeds a certain value. Decreasing porosity additionally reduces bed permeability. For two different moments in O/W dispersion flow through the bed the relationship betwenn pressure drop and porosity can be determined from the Kozeny-Carman equation [ 1 1] resulting in: - A p t et" 1 -eo I -Et
APo
E:
where :
APo, APt E,, et
pressure drop initially and after time, t, respectively bed porosity initially and after time, t, respectively
The definition of saturation is:
(3)
273 Tab. 2. Characteristics of artificial and industrial waste water and properties of diluent used in oil coalescence tests
O/W dispersion in waste water Description
Units
artificial
Oil density at 20°C Oil viscosity at 20°C Suspended solids (sediment) in oil Oil concentration in emulsion (do < 60 wm) Mean oil droplet diameter O/W interfacial tension at 20°C Zeta potential of fine oil droplets at pH = 7.5 pH of aqueous phase Suspended solids in aqueous phase Temperature Diluent used
kg/m3 Pa s mg/kg mg/l wm mN/m mV mg/l "C -
895 0.12* 4900 300 5-8 28 -68 6.9
Diluent density at 20°C Diluent viscosity at 20°C Zeta potential of diluent droplets at pH = 7.5
kg/m3 Pas mV
industrial
890-930 0.07-1.5 4800-25,000 80-600 5-20 20-31 (- 40) -( - 7 0) 6.8-7.8 5 60-380 20 18-28 kerosene (70% paraffins, 25% naphthenes, 5% aromatics) 821 1.2 x 10-3 - 62
* unless stated differently and after substituting e t in (3), the final dependence of actual pressure drop on bed saturation with oil is:
-_
1 - e o (1 - S ) APo ( l - ~ o ) ( l - S ) ~ apt -
When diluent is injected into waste water for bed regeneration the change in pressure drop can be used as a measure of bed liberation from heavy oil. The calculated value of saturation will be closer to the real one the less the diluent viscosity differs from that of water and the larger are the void spaces within the bed (capillary forces may not be taken into account). The coalescer bed must not be compressible as well. Two zones may be distinguished in a coalescer bed in which an ideal separation takes place, with no residual concentration of fine oil in the effluent, as shown in Fig. 3 [4, 12, 131. In the first zone discrete oil globules exist that are captured by the two-phase matrix of solid and held-up oil. When moving deeper into the hed, unbroken threads (channels) of oil gradually formed which are characteristic of the second zone. Both zones were observed for granular bed not preferentially wetted by oil [4] and for fibrous beds in which filaments of a highly oleophilic surface were employed [13]. Oil globules in zone I partially remain spherical and partially spread on the packing elements [ 141. Because of the highly developed oil surface dilution may be very effective. The size of pores within the bed may have an influence on the dilution kinetics. Diluent drops, much larger in diameter than oil droplets, can completely fill some narrower capillaries. A partially capillary-conducted flow of diluent may result in a poorer heavy oil dilution due to the reduced contact surface. In the entry part of the bed diluent is
274
I
WATER
ZONE I
_1 t
D I R E C T I O N OF F L O W Fig. 3. O w dispersion flow through coalescer bed.
in large excess with respect to heavy oil. Diluent migrating deeper into the bed contains gradually more dissolved heavy oil, which results in reduced dissolution ability and extended dissolution time.
4. SELECTION OF DILUENT
regeneration efficiency point of view, the substance used as a From the diluent for oil should have a viscosity close to that of the continuous phase, be insoluble in water, and be able to rapidly dissolve the dispersed phase. Another requirement is that the diluent should have no deleterious effect on the properties of the recovered oil with respect to its utilization. Aliphatic and naphthenic hydrocarbons possessing 10-1 2 carbon atom molecules are therefore considered for such a role. Kerosene is the commercially available product containing the above-mentioned hydrocarbons. Solubility parameters calculated for kerosene and for mineral oils, which are often present in waste water, are close enough to secure complete mutual solubility [16, 171. Diluent application in water deoiling must not create further processing problems, which can be much more serious than that of the heavy oil itself. Solvent may be lost
275 due to dissolution and emulsification in water causing simultaneously such a hgh organic carbon content in the effluent that subsequent additional treatment steps are necessary. Light hydrocarbons are flammable and easily explode in mixtures with air at elevated temperature. Diluent application does not, however, require the waste water to be heated before processing and hydrocarbons as found in kerosene are safe. Chlorinated hydrocarbons as inflammable solvents may be considered as diluents, but such a solution introduces additional heteroatoms into the oil. Some components of the regenerative medium may be harmful to health, mainly because of carcinogenic properties, or may cause changes in the bed (swelling of the packing material or loss of desirable surface properties of the bed due to dissolution or adsorption). Proper examination of the potential diluent with respect to such undesired properties will allow one to avoid later technological difficulties. To enhance the gravitational separation in the settling zone in the coalescer, the lowest possible density of the diluent is desirable. An ideal solvent which fulfills all of the above requirements does not exist, and kerosene seems t o be an acceptable compromise, especially when its low cost and recovery possibility are taken into account.
5. EXPERIMENTAL PROCEDURE
In the experiments reported here, oil droplets suspended in waste water were coalesced in a laboratory scale filter (Fig. 4) and in a pilot separator equipped with exchangable cylindrical cartridges (Fig. 5). Artificial O/W dispersion was prepared in a laboratory homogenizer using oil recovered from industrial waste water. The pilot scale coalescer was tested on oily waste water from a liquid fuel recovery plant. The characteristics of artificial and industrial waste water are given in Tab. 2. Polyethylene granules with a diameter of 0.75 and 1.25 mni and 18 pni diameter (average) fibres prepared from cord of spent tires were used as packing media [8]. The bed porosity obtained with granulated polyethylene was 0.36 for 0.75 mni grains and 0.40 for 1.25 mni ones. The packing density of fibres equal to 0.31 g/cm3 resulting in 0.75 porosity was used in most tests, unless stated differently. The same materials were used in the laboratory and field experiments. The depth of the bed in the laboratory coalescer was 10 cm and in the cylindrical cartridges 7 cm. In most laboratory and field experiments kerosence was injected periodically into the waste water when the pressure drop across the bed exceeded a fixed value. In some experiments diluent was blended with the influent continuously. Diluent injection was carried out by means of a single nozzle in the laboratory assembly and a single of set of nozzles in the 3 m3/h flow rate pilot separator. For the jetting velocities within the range of 1.0-10.0 m/s, the diluent minimal-maximal droplet diameter varied from 20 pin to 2 mni, with the average value between 80 pin and 300 pni. Analytical methods involving determinations necessary for oil waste water characterization presented in Tab. 2 are described elsewhere [ 131. The degree of bed saturation with oil was meassured by extraction of the entire filling or (in the laboratory) 1 cm thick layers with CCl,. Saturation was correlated with the rate of pressure drop changes accross the bed.
276
0i\
COAL E S C E R HOF O G E N
Fig. 4. Laboratory experimental setup.
Oil
8 FIRST STAGE
SECOND STAGE
vg I
h l
MANOME TER
Wate r >
M E T E R I N G PU M P
Fig. 5. Coalescer of 3 m3/h capacity equipped with regenerable cartrigdes.
277 6 . RESULTS AND DISCUSSION
In Fig. 6 more detailed evidence is given (than in Fig. 1) as to the influence of oil viscosity on the capacity factor*, related to different packing density in a fibrous bed. The capacity factor is several times larger for less viscous oil than for heavy oil and increases with decreasing packing density of fibres. The capacity factor increases for higher porosity in the bed, but simultaneously, oil removal efficiency from waste water decreases as indicated by the increasing C / C F ratio. In another experiment, oil of viscosity 0.30 Pas was steadily diluted. during coalescence process by continuous kerosene injection. When assuming that uniform and perfect dilution took place in the bed the resulting viscosity of the solution should be 0.07 Pas. The capacity factor, however, did not reach the value as for oil of original viscosity equal to 0.07 Pas, but was slightly lower. This implies that dilution was not complete; nevertheless, significant improvement in the lifetime of the bed has been achieved. When the apparent velocity of O/W dispersion in the bed was lowered from 10 m / h to 1 m/h, dilution was more effective, and the capacity factor was equal to that of oil of viscosity 0.07 Pas (data not plotted in Fig. 6). Analysis of the lower part of the graph in Fig. 6 reveals that for the same packing density, quality of oil removal practically does not depend on oil viscosity, which is also the case when diluent was applied (though the total organic phase content was higher because of diluent addition). Untreated fibres originating from the cord of spent tires cannot secure the rigid structure of the bed below a packing density of 0.2 g/cm (porosity 0.084). The porous medium undergoes compression soon after oil begins to saturate it. Fibrous and granular beds respond differently to diluent injection into the O/W dispersion. Fig. 7 presents a change in bed saturation with oil during periodic regeneration attempted in a laboratory coalescer. Similar to field experiments [8], it is also true in a laboratory coalescer that a fibrous bed already clogged with oil does not release oil when diluent is injected. A coalescer packed with polyethylene granules is sensitive to the presence of solvent in fed O/W dispersion, but the degree of oil liberation depends on the grain size and oil viscosity. Ten minutes of regeneration was insufficient to remove oil of viscosity 1.O Pas. Decreasing regeneration efficiency when considering coarse granular bed, fine granular bed, and fibrous bed, may be related to the decreasing size of void spaces within the porous material (see Tab. 1). Saturation with oil, as shown in Fig. 7, represents a mean value for the entire bed. In reality bed saturation with oil is a function of depth as presented in Fig. 8. The entrance part of the bed is saturated to about 50-70%, whereas in the middle a less than average part of the porosity is occupied
by oil (also compare [4,17, IS]). Fine pores and high saturation of zone I in the bed may account for the failure in the regeneration attempt in a fibrous bed due to the very limited porosity which is available for the flow of the continuous phase with suspended diluent droplets. Pores in a granular bed are larger, closer to each other in size, and more uniformly distributed in the bed; therefore, oil saturating bed has a higher contact surface with the diluent. Larger capillaries in bed facilitate flow of partially diluted oil, which
* Capacity factor-amount of dispersed phase introduced into bed related to amount of dispersed phase hould-up in the bed at the break-through point. ** C/Co - concentration ratio, measure of separation efficiency for coalescer, C - oil content-in the effluent, Co - oil content in the feed.
278 103
I
Oil viscosity
\
\ \
0 0.07
Pa s
I 10 -
\
\\
0.8 -
-
0.6
0.4 0.2 0
0 (1)
I
I\'\
0.1 (0,921
0,2 (0.84)
0.3
0,4
(0.76)
(0,691
PACKING DENSITY (glcm3) (BED P O R O S I T Y 1 Fig. 6 . Influence of oil viscosity on the performance of laboratory coalescer with fibrous bed.
explains better regeneration efficiency and higher capacity factor for coarse filling. Oil existing in zone I1 in continuum form is much less susceptible to dilution. Diluent droplets also undergo coalescence; therefore, a significant part of the diluent is expected to not be present in zone I1 in discrete form. Special attention was paid to the behaviour of the entire coalescence system during diluent injection. Introduction of an additional amount of organic phase means a serious disturbance in the quasiequilibrium conditions already achieved in the bed fed with heavy oil. Fig. 9 reveals that diluent injection into the influent caused a rapid increase of oil content in the effluent. This oil exists as a fine suspension and is not separated in the settling zone of the coalescer. The appearance of a multitude of fine droplets in the effluent may be explained by liberation of uncoalesced droplets [8] or by phase inversion occurring locally in the bed and resulting in subsequent production of fine droplets at the bed outlet, as reported by Jeater et al. [18]. Phase inversion during coalescence of pure engine oil (viscosity 0.18 Pas at 20°C) dispersed in distilled water was also observed for a fibrous bed when oil content was changed rapidly from 0.5 t o 20 g/l [ 191. Created
279
50
Oil viscosity 1.0 Pa s 0 7 5 mm granular bed
40
30 20 10
0
-a
I 0
1.25 mrn granular bed o 0 7 5 mm aranular bed A 1 8 p m fib;ous bed ( 6 =0,76 1
I 18 20
I
2
6
4
8 10 12 TIME (rnin)
16
14
Fig. 7. Change in bed saturation with oil during kerosene injection into O/W dispersion for fibrous and granular packing in laboratory coalescer.
90
80 70 h
e 60
- 50 a 40 30 m 20 10
0
0
1
2
3
4
BED
5
6
7
a
g
DEPTH ( m m )
Fig. 8. Saturation profile for total (discrete and continuous) oil in the bed.
-
1
0
280
-2.5
5
0
10
20
15
TIME ( m i d
Fig. 9. Saturated granular and fibrous bed response to kerosene injection in laboratory coalescer.
I
0,005 0,Ol
I
I
I I I I I I
0,051 0,l
I
I
I
I 1 I I I I
0,5
1
DILUENT MEAN DROPLET DIAMETER d,(mm) Fig. 10. Difference in pressure drop across 0.75 m m granular bed vs. kerosene mean droplet size.
28 1
Fig. 11. Influence of continous kerosene injection on the performance of fibrous bed in laboratory coalescer.
(E
= 0.72)
oil membrane was extremely stable and water droplets were larger than 1 mm. Final rupture of the membrane did not produce a noticeable amount of microdroplets. The presence of water dispersion in coalesced oil was also noticed at the bed exit during kerosene injection into granular polyethylene bed, but agah, fine oil droplets were not created as a result of membrane rupture [20]. The influence of kerosene droplets size on regeneration efficiency is presented in Fig. 10. The decrease of the pressure drop across the bed after 10 min. of diluent injection was used as a simple measure of regeneration efficiency. Data plotted in Fig. 10 were obtained in a laboratory coalescer for 0.75 mm granular packing. Injection of dilluent was started when the pressure drop exceeded 0.3 X lo5 Pa, which corresponds to 35% bed saturation with oil. The average diluent droplet size dd has practically no influence on the rate of pressure drop change A(AP) across the bed, as long as the dd remains above 20 pm. Below 20 pm dilution ability drops sharply due to impeded coalescence between oil and diluent droplets. The impact of the degree of diluent dispersion in waste water can be seen even more clearly in Fig. 1 1. In this case diluent was injected into the feed from the very beginning of the coalescence process in the proportion 1 : 1 with respect to the amount of heavy oil. Such a strategy is the only effective one in preventing rapid heavy oil build-up in the fibrous bed. The capacity factor curve in Fig. 11 possesses maximum, because both fine and large diluent droplets indicate lower dilution ability. Diluent droplets of average diameter below 10 pm are too small to undergo effective coalescence. Droplets above 500 pm cannot reach the majority of free spaces in the bed and a significant amount of oil remains undissolved, resulting in a lower capacity factor. Due to constantly changing properties of industrial waste water no systematic series of experiments could be planned with application of the 3 m3/h flow rate pilot coalescer. The liquid fuel recovery plant, which was the site of field tests, accepts ballast and bilge water from ships, washings from fuel tanks, oil-water mixtures from oil spill recovery operations, etc. Despite averaging the feedstock in tanks, the viscosity of oil in waste water varies in between wide limits, which gave an unique opportunity to study its in-
282 Tab. 3. Influence of oil viscosity on regeneration possibility of cartridge filled with coarse granular polyethylene ~~~
Oil viscosity Pa s
Critical Pressure drop change A(AP) pressure drop Critical Pax saturation Pa x lo-’
0.07 0.12 0.18 0.30 1.oo 1.50
0.52 0.48 0.45 0.42 0.25 0.20
0.50 0.49 0.46 0.42 0.31 0.30
0.37 0.33 0.30 0.27 0.10 0.05
~~~~~~
Method of diluent injection* periodic
periodic or continuous continuous
* The amount of diluent is adjusted according to oil viscosity, resulting in final viscosity of the mixture < 0.05 Pas at the temperature of processing. fluence on the coalescer performance. Suspended matter (mostly fine sand and products of corrosion) was also present in aqueous and organic phase (see Tab. 2 ) . All experiments reported here were carried out with coarse granular polyethylene cartridge working as the first stage and fine granular polyethylene cartridge or fibrous cartridge in the second stage. Coarse polyethylene bed served as a preliminary coalescer and protector of the second cartridge against suspensions of solids. Oil content reduction in the first stage usually remained at 50% level. Field tests have shown that for oil exceeding viscosity of 1.5 Pas, the coarse granular bed was not liberated from heavy oil after 1 0 min. of diluent injection. Application of a triple amount o f diluent had practically no influence on oil fluidity, but continuous injection of diluent, when a new cartridge was installed, prevented oil build-up in the bed. The data in Tab. 3 reveals that the higher is the oil viscosity the lower critical pressure drop is required for starting diluent injection. The critical values of pressure drop and the corresponding saturation make up the limit of still achievable bed liberation from oil. With increasing oil viscosity A(AP) decreases resulting in more frequent regeneration. Finally, continuous diluent injection has to be applied. One must emphasize that values given in Tab. 3 were strongly influenced by fluctuations in oil and suspended matter content and represent only the described experimental system. In the second stage, oil content in the effluent was reduced to approx. 45 mg/l. Cartridge was fed with oil and diluent still contained in effluent from the first stage. Regeneration was less frequent but with inferior results. Coalescer of the 3 m3/h capacity required more frequent regeneration than in the laboratory one, mostly due to presence of suspended matter in waste water. Build-up of fine solid particles in bed was due to deep bed filtration, enhanced with presence of oil saturating the bed. Suspended matter was not released during bed operation, even during back-wash. In such a case diluent injection resulted in over 90% liberation of the bed from suspended matter. The fibrous cartridge did not release solids clogging the bed during continuous diluent application. From small scale tests it seemed that regeneration of the bed saturated with heavy oil can be repeated many times. In the pilot installation, however, granular beds were quicker or slower losing coalescence ability even when the concentration of suspended matter was low. A coarse granular bed had a longer lifetime than a fine one, but for both of them continuous or periodic injection of kerosene did not prevent gradual growth of head
283 loss across the bed. Changing properties of the treated waste water were found to be responsible for fluctuations in the cartridge lifetime. Close examination of spent bed in a granular cartridge revealed a bituminous-like substance coating polyethylene grains. Dissolution attempts on the coating in n-pentane showed only 20% by weight solubility. 85% of the substance was soluble in toulene. In 15% of the insoluble part 90% was of inorganic origin (suspended matter entrapped in oil). This coating, insoluble in diluent, caused gradual reduction in porosity resulting, at some moment, in permanent loss of coalescence efficiency. Besides bituminous materials, solid particles of a size comparable to the diameter of capillaries can irreversibly clog the bed with no chance for their release during diluent injection. These particles were sometimes carried in with the feed as oil-solid agglomerates. In such cases even the coarse granular bed had to be replaced promptly.
7 . REMARKS ON COALESCER DESIGN
Oil-water dispersions which can be met in industrial waste water, unlike other O/W dispersions (present in cooling water, condensed steam, and solvent extraction systems), have to be treated separately when coalescers are being applied in their processing. To improve coalescer efficiency materials of special properties, such as resins of procured high oleophilicity [2 13, positively charged aluminium filaments [22], dissimilar materials packing [23] or combined metal-metal and carbon-metal fillings [24, 251 have been employed in the bed. These materials, advantageous in many applications, are often of limited use in a coalescer working on an actual pollutant stream. When industrial waste water contains heavy fuel oil or crude oil the surface of packing in coalescer bed may be covered by a thin layer of well-adsorbed oil or its components, like asphaltenes, resins or waxes, completely eliminating the beneficial surface properties of the original material. The solution to this problem can be found in application of inexpensive coalescing materials, easily available from utilization of waste materials. Regeneration of the bed is another "must" for an industrial coalescer. When light petroleum products are present periodic back-wash of the bed allows for suspended matter removal. Wasre water containing heavy oil requires diluent injection during processing. Coalescer used for waste water treatment and equipped with periodically regenerated bed should always be considered as two- even three-stage device. The first and second stages employ granular regenerable filling, and the polishing disposable fibrous bed makes the third stage. Granular material in spent regenerable cartridges can still be recycled after cartridge replacement in separator and suitable treatment of the packing.
8. CONCLUSIONS
1. Fibrous bed already saturated with viscous oil appeared not to be susceptable to regeneration performed by diluent addition t o the influent. Diluent injection into O/W dispersion allowing both heavy oil and diluent droplets to simultaneously saturate the bed resulted in prolonged lifetime of the bed.
284
2. Granular bed could be liberated from viscous oil by diluent injection into the feed stream when the saturation remained below 40-45%. 3. Periodic injection of diluent requires recycling of the effluent during regeneration period due to rapid increase of oil concentration. 4. The degree of diluent dispersion in the feed does not influence the regeneration efficiency assuming that droplets are not too fine or too large. Tiny droplets do not coalesce fast enough, and large droplets cannot easily migrate into the entire bed. 5. Kerosene was successfully used as a diluent, but another regenerative medium can be used, bearing in mind all the limitations connected with such a process aid.
REFERENCES
1 D. F. Sherony and R. C. Kintner,Can. J C h e m Eng., 49, 314, 1971. 2 J. I. Rosenfeld and D. T. Wasan, Can. J. Chem Eng., 52, 3, 1974. 3 L. A. Spielman and S . L. Goren, Ind. Eng. Chem Fundam, 11 (l), 66, 1972. 4 L. A. Spielman and Y. P. Su, Ind. Eng. Chem Fundam., 16 (2), 272,1977. 5 C. G . Vinson and S. W. Churchill, Chem Eng. J., 1, 110, 1970. 6 J. N . Chien, E. F. Gloyna and R. A. Schechter, J . Environ. Eng. Div., April 1977, 163. 7 A. Tribellini, Circ. Inf. Tech. Cent. Doc. Sid., 35 (3), 537, 1978. 8 J. Hupka, World Filtration Congress 111, Sept. 13-17, 1982, Downington, PA, USA, Proceedings, Vol. 11, p. 669. 9 W. M. Langdon, T. Sumpatchalit, V. Sampath, and D. T. Wasan, Proceedings of the International Solvent Extraction Conference, Toronto, Sept. 5-16, 1977, CIM Special Volume 21, 413. 10 B. Shah, W. Langdon and D. Wasan, Environ. Sci. Technol., 11 (6), 167, 1977. 11 W. M. Langdon and D. T. Wasan, Recent Developments in Separation Science, Vol. V, Chapter 10, CRC Press, Inc., West Palm Beach, Florida, 1979. 12 R. N. Hazlett, and H. W. Carhart, Filtr. Sep. 9, 456, 1972. 13 J. Hupka. P h D. The+, Technical University of Gdahsk, 1978, 14 k. Aurelle and H. Ruques, 2nd World Congress of Chemical Engineering, Oct. 4-9, 1981, Montreal, Canada, Proceedings Vol. IV, 238. 15 H . J. Neumann, Erdol und Kohle - Erdgas - Petrochemie, 34 (8), 336, 1981. 16 D. L. Mitchell and J. G . Speight, Fuel, 52 (4), 149, 1973. 17 J . F. Bitten and E. G . J. Fochtnian, J. Colloid Interface Sci., 37, 312, 1971. 18 P. Jeater, E. Rushton and G . A. Davies, World Filtration Congress 11, London, 1979, Proceedings, p. 605. 19 J . Hupka, unpublished results, Technical University of Gdahsk, 1977. 20 M. Kubiszowa, M. S . thesis, Technical University of Gdahsk, 1980. 21 Y. Aurelle, A. Abadie and H. Roques, Tribune du CEBEDEAU, 352, 108, 1973. 22 D. B. Chambers, Chem. Ind., 21,834,4 Nov. 1978. 23 D. P. Bayley and G . A . Davies, Environ. Pollution Management, July/August 1978, 91. 24 F. M. Fowkes, F. W. Anderson and J . E. Berger, Environ. Sci. Technol., 4 (6), 510, 1970. 25 M. M. Ghosh and W. P. Brown, J. Water Pollution Control Fed., 47 (8), 2101, 1975.
285
SATURATION PROFILES IN COALESCENCE BED
B. GUTKOWSKI, St. MYDLARCZYK, M. KOWALSKA, J . HUPKA Institute of Inorganic Chemistry and Technology Technical University of Gdansk, 80-952 Gdarisk, Poland
ABSTRACT An influence of oil viscosity on the degree and profile of saturation of the coalescence bed was studied. Three types of oil of viscosities 4.6, 7 and 340 mPas were dispersed in fresh water. Granular polyethylene with particle diameters between of 0.6 and 0.75 mm was used as a filling in a model coalescer. The degree of saturation was determined by extracting the oil with an organic solvent from each centimetre of the 10 cm deep bed. It was found that the degree of saturation is the greatest in the entrance part of the bed and then decreases along the bed. Unlike the degree of saturation, the saturation profile depends o n oil viscosity. Two kinds of oil flow in the bed were observed: “moving layer” flow for very viscous oil and “channel” flow for oil of low viscosity.
1. INTRODUCTION
Coalescence process involves changes in oil dispersion in the emulsion introduced into the bed as well as oil retaining in the packing. Progressive saturation of the bed with oil has a great influence on the effective time of operation [l]. Therefore the nature of the hold-up oil and the oil flow in the bed are of basic significance, if only from the economical point of view [2,3]. The effect of saturation on the efficiency of the coalescence process has been considered by several authors. Sherony and Kintner [4] point to the dependence of the filter coefficient on the degree of saturation, noting little or no coalescence for low saturated beds. Rosenfeld and Wasan [S], in theoretical equation for the filter coefficent, also indicate lack of effect of saturation on the coalescence process. Experiqents carried out by Spielman and Goren [6], however, suggest an existence of a minimum saturation Smin = 0.3, which is indispensable for conducting the oil held-up in a fibreous bed. Later studies by Spielman and Su [7] carried out with granular beds has shown that even Smin = 0.1 is sufficient for capillary conduction of oil retained in the bed. Although Spielman and Goren [8] suggest a uniform degree of saturation for the entire coalescence bed on the basis of a theoretical model, this has not been confirmed by other authors [7, 91. Studies by Bitten and Fochtman [9] concerned with the distribution of water
286
f Fig. 1 . A schematic diagram of experimental equipment.
Tab. 1. Physicochemical Characteristics of Oils Used in Tests Typ eo f o i l
Density kg/m3
Viscosity mPas
Interfacial tension Wetting angle mN/m deg
Mean droplet diameter wn
Lux 10 Recovered oil Diesel oil
897 895 832
340 7 4.6
27.2 10.7 28.9
3.0 3.2 2.9
30.3 27.0 20.4
held-up in a fibrous bed show a high degree of saturation at the inlet face of the bed and 3 to 5 times lower saturation in the middle and at the end. Similar saturation profiles were obtained by Spielman and Su [7], remaining also in agreement with their theoretical considerations. Our present work was aimed at determining saturation profile changes in a granular bed during the coalescence process bearing in mind the flow pattern of the dispersed phase through the packing.
2. EXPERIMENTAL METHODS AND APPARATUS
Laboratory tests were carried out in a glass column, 80 cm high and 7 cm dia shown with auxiliary equipment in Fig. 1. Lower part of the column was filled with 10 cm of packing. Emulsion was fed from a tank ( 2 ) by means of a pump (3). Model emulsion was prepared in a unit composed of a tank of oil (4), a pump (5) for metering of oil to a homogenizer (6) and a tap water supply equipped with a rotameter (7).
287 The homogenizer revolutions were set in the range 5000 to 10,000 rpm depending on the kind of oil being used in order to maintain a similar droplet size distribution. The concentration of oil vaned between 350 and 450 ppm and the emulsion flow rate was 2.5 X m/s. Three kinds of oil were used during tests: "Lux 10" lubricating oil, diesel oil and oil recovered from industrial waste water. The physicochemical characteristics of oils used are presented in Tab. 1. Short chain granular polyethylene with particle size from 0.6 to 0.74 mm was used as a packing, resulting in porosity E , = 0.44. After each run the polyethylene filling was washed with an organic solvent and recycled. However, fresh lots of granular polyethylene were applied for each particular kind of oil. Analytical control of the process included fluorescence determination of oil at the column inlet and outlet as well as observations of the pressure drop changes across the bed. The amount of oil held-up in the bed was determined by direct extraction. The coalescence was stopped when the desired pressure drop was attained. The bed subsequently was removed from the column and divided into 10 one-centimetre thick layers. Then oil was extracted from each layer with the aid of carbon tetrachloride. Saturation was determined on the basis of the following formula [lo]:
s. =
moi p p
mbi Po €0 where : Si fraction saturation of total held-up oil in the i-th centimetre of the bed, dimensionless moi weight of held-up oil in the i-th centimetre of the bed, g mbi weight of the i-th centimetre of the bed after oil extraction, g po oil density, g/cm3 weight of the bed pp packing density defined as , g/cm3 volume of the bed e0 bed porosity, dimensionless.
3. RESULTS AND DISCUSSION
In order to study the degree of saturation of the polyethylene bed each experiment was continued till the moment at which the desired pressure drop across the bed equal to AP = 0.095 X lo5 Pa, 0.16 X lo5 Pa, 0.26 X 10s Pa was obtained. Then, the process was stopped and the saturation of the bed with oil was determined. The initial pressure drop was equal to AP, = 0.07 X lo5 Pa in each run. The results are shown in Figs.'2, 3 and 4 as a function S = f(h), for various values of the pressure drop increase defined as A(AP) = AP - AP,, (h-is the distance through the bed). Plots in Figs. 2, 3 and 4 reveal that the saturation decreases with distance through the bed. For A(AP) = 0.025 X lo5Pa (see Fig. 2) the shape of the curves for all kinds of oil is similar. The highest overall saturation was obtained for the "Lux 10" oil. The largest amount of oil was retained in the entrance part of the bed (the first 3 cm).
-s
70-
z
-t0
a
6C-
A
(
APP) = 0.025.105 Pa
a
3 t-
3: 50-
L
2 4 6 8 10 DISTANCE T H R O U G H T H E B E D , t c m l
Fig. 2. Results showing fraction saturation of held-up oil for A(AP) = 0.025 l o 5 Pa. 0/0Lux 10 oil, recovered oil, o/= diesel oil. Open and dark symbols respectively correspond to duplicated experiments under similar conditions.
A/A
In the case of A(AP) = 0.09 X lo5 Pa (Fig. 3) a further increase of the bed saturation takes place. For such A(AP) differences in the saturation profiles begin to be more distinctive depending on the kind of oil used. The ”Lux 10” oil continued to be most intensively retained one in the bed. However, its linear saturation profile clearly differs from the shape of the curves for the less viscous oils. In the case of the ”Lux 10” oil the bed saturation changes in relatively narrow limits from S = 0.35 to S = 0.55. In the case of diesel oil and recovered oil the saturation differences between the initial and final parts of the bed are in the range of S = 0.05 to 0.55. Continuation of the process resulted in the lack of essential changes in the saturation of the bed with oil. As can be seen in Fig. 4 the saturation profile obtained at A(AP) = 0.16 X lo5 Pa for ”Lux lo” oil is linear. The differences in the degree of saturation between the inlet and outlet part of the bed are negligible. For diesel oil the increase of
289
Fig. 3. Results showing fraction saturation of held-up oil for A(AP) = 0.09 10' Pa. 010 Lux 10 oil, recovered oil, o/m diesel oil. Open and dark symbols respectively correspond to duplicated experiments under similar conditions.
A/A
retained oil is uniform along the bed with respect to the lower A(AP), therefore the shape of the curve S = f(h) did not charge. As far as recovered oil is concerned, the degree of saturation at the entrance of the bed increased considerably to a value of S = 0.7 while the saturation in its final part remained almost completely unchanged ( S = 0.05). Thus, an analysis of the diagrams points to the existence of distinct differences in the bed saturation profile with the "Lux 10" oil and with the other oils. Attempts were made to elucidate this phenomenon by carring out additional observations of the bed during each run. In the subsequent tests the flow of the O/W emulsion was continued till the pressure drop reached A(AP) = 0.9 X lo5 Pa. At that moment the process was interrupted for 30 min. Then the coalescer was fed with water only, but the flow rate was maintained
290
DISTANCE THROUGH T H E B E D ,
[cml
Fig. 4. Results showing fraction saturation o f held-up oil for A(AP) = 0.16 10' Pa. 0 / 0 Lux 10 oil, recovered oil, fi/= diescl oil. Open and dark symbols respectively corresponds to duplicated expcriments under similar conditions.
n/r
twice or three times higher than that at the beginning. A cloud of oil droplets of diameter less than 100 pm appeared immediately at the outlet. Fine oil droplets gradually disappeared in the effluent after about 5 niin., leaving transparent water. Microscopic examination of the elements of the filling reyealed thin film around each grain especially those coming from the entrance part of the bed. Thus, our observations point t o the existence of a discrete and a continuous forin of oil in the bed. Presence of a continuous form of oil, rendered doubtful by Calteau et al. [ 113, is of a basic significance for explaining tl; i' differences which appear in the oil saturation profile. Experimental data in Fig. 2, 3 and 4 show high and low saturated zones in the bed for low-viscosity oils. The first 4 to 6 crn of the bed, saturated at least in 50%, contains oil in discrete and
29 1
Fig. 5. Photograph of cross-sectional area of the 5th, the 6th and the 7th centimetre of the bed. Recovered oil. A(AP) = 0.09 lo5 Pa.
continuous form. The oil film covers uniformly practically all packing elements in the plane perpendicular to the flow direction. Thus, there is a saturation gradient exclusively in the direction of flow. In the low saturated zone oil does not cover filling elements with a continuous film but flows through 3 to 6 mm dia channels as a result of capillary conduction. The photographs in Fig. 5 show cross-section of the bed at certain depth at which the flow channels can be seen. The diminishing number of channels with increasing height of the bed indicaties that a steady-state of the process has not yet been attained. At the entrance part of the bed oil forms a film due to a sieving process and coalescence. Under the influence of friction between the waterioil film interface a flow of the continuous oil occurs deeper into the bed. It is also this part of the bed, to which there is a flow of oil droplets, grown larger by coalescence with their neighbouring droplets, and not spread on the packing elements nor absorbed by the oil film. The final part of the bed is used merely for transport of the held-up oil. High fluidity of less viscous oils facilitates outflow from the first zone, in which the major portion of oil is held-up. The degree of saturation in the second part of the bed is much lower than that in the first one because due to presence of oil channels the amount of oil retained in its cross-section is not uniform. The flow of low viscosity oil may be therefore described by ”channel” flow. Unlike the oils of viscosity 7 and 4.6 mPas, the bed saturation with the ”Lux 10” oil shows lack of decrease in the middle part of the packing. The dependence S = f(h) is linear and the curve becomes more parallel to abscissa with increasing pressure drop across the bed. This is a different behaviour of the system with respect to that found for the others oils. It is difficult to distinguish zones in the bed similar to those previously described. The bed in the crosssection located at the outlet is saturated in a uniform manner and no channels can be noticed. The flow of the herd-up oil take place through most of the Void spaces in the entire packing. The high viscosity of oil resulting in hindered flow through the bed may account for the lack of channels. The oil movement in the bed can be described by a ”moving layer” flow, well explaining the liner character of the saturation profile. Fig. 6 shows the dependence of the de-oiling efficiency on the increase in the pressure drop across the bed. It follows from the diagrams that the efficiency of the process
292
1.0
1
I
I
I
Z J
a a-
2:2
0
- L U X 10
- REGENERATED OIL A - D I E S E L OIL o
1 I
I
I
I
0.0 0.05 0.10 0.1 5 0.20 INCREASE IN PRESSURE DROP ACROSS THE B E D A (A ~ i . i o * r r n ]
Fig. 6. Coalescence efficiency vs. increase in pressure drop across the bed.
increases with the increasing pressure drop caused by the growing saturation. Our observations confirm the theories of the coalescence process so far proposed, indicating the dependence of the filter coeffcent on the degree of saturation. The relatively small increase of the efficiency of the process found in the case of the "Lux 10" oil may be due to the moderately stable saturation in the inlet part of the bed w h c h plays a basic role in the hold-up of the dispersed phase.
4 . CONCLUSIONS AND COMMENTS
On the basis of performed experiments with the 10 cm deep coalescer bed the following conclusions and comments can be drown.
293 - Flow of an O/W dispersion through a coalescer bed results in a different saturation degree according to the position in bed. The amount of oil retained in bed is the highest at the entrance part and diminishes with the bed depth. - The saturation profile depends on the oil viscosity. In the case of low viscosity oils (4.6 and 7 mPas) the saturation diminishes rapidly and shows a sharp fall in the centre of the bed. The amount of high viscosity oil retained in the bed is a linear function o f the distance through the bed. The degree of saturation in the initial and fiial parts of the packing approaches the same value for extended processing time. - Different patterns of oil flow through the bed were observed depending on the oil viscosity. Oil of low viscosity migrates into the all void spaces in the first half of the bed, and then flows through the other half in several 3 to 6 mm dia channels. Oil of high viscosity was uniformely distributed in the bed and no channels were created. Flow of that oil takes place the whole transverse area of the bed. - Eventual remarks about the selection of a coalescer bed for application in systems containing oil of varying viscosity can be presented after accomplishement of our study on other parameters influencing the coalescence technique. Surface properties of the dispersed phase and that of the filling are among the most important ones.
REFERENCES
1 J. Hupka, Ph. D. Thesis, Technical University of Gdahsk, Gdahsk, Poland 1978. 2 B.Gutkowski,Przem. Chem, 2 (1983) 112-115. 3 B . Gutkowski, St. Mydlarczyk, ibid., 4 (1983) 236-238. 4 D. F. Sherony, R. C. Kintner, Can. Jour. Chem. Eng., 49 (1971) 321-325. 5 J. J . Rosenfeld, D. T. Wasan, ibid., l ( 1 9 7 4 ) 8-10. 6 L. A. Spielman, S. L. Goren, Ind. Eng. Chem. Fundam., l ( 1 9 7 2 ) 73-83. 7 L. A. Spielman, Y. P. Su, ibid., 2 (1977) 272-282. 8 L. A . Spielman, S. L. Goren, ibid., 1 (1972) 66-72. 9 J. F. Bitten,E. G. Fochtman, Jour. Coll. Int. Sci., 2 (1971) 312-317. 10 M. Kowalska, M. S. Thesis, Technical University of Gdahsk. Gdahsk, Poland, 1983. 11 J. P. Calteau, Y . Aurelle, H. L. Roques, AIChE Journal, 4 (1978) 741-745.
This Page Intentionally Left Blank
C H A P T E R IV
PHYSICO-CHEMICAL TREA T M E " : ADSORPTION
This Page Intentionally Left Blank
297
THEORETICAL FOUNDATIONS OF SOLUTE ADSORPTION FROM DILUTE SOLUTIONS ON SOLIDS
A. DERYLO and M. JARONIEC Institute of Chemistry, Maria Curie-Sklodowska University, 20-03 1 Lublin (Poland)
ABSTRACT The simple theoretical description of adsorption from dilute solutions on solids is presented. Some important factors determining the adsorption process, such as energetic heterogeneity of the adsorbent, multilayer formation, differences in molecular sizes of solutes and pH of the solution are discussed. A method of selection of the isotherm equation giving the best representation of a given experimental system is proposed. Some of the relationships presented are examined by using single- and bi-solute experimental data.
1. INTRODUCTION
The process of adsorption from solutions is becoming more and more popular from a practical point of view. In particular, adsorption systems consisting of dilute solutions in contact with different kinds of activated carbon are the most frequently applied in many branches of science and technology. A typical example of their utility is environmental protection, in which adsorption is an important stage of drinking water or wastewater treatment. Moreover, this process is also used for special purposes in medicine and industry. Control of the adsorption process utilized in these practical applications requires its mechanism and parameters to be known. Therefore, apart from experimental measurements the theoretical studies of adsorption are necessary to its full description. In the majority of papers concerning the adsorption of nonelectrolytes from dilute solutions, isotherm equations obtained analogously to gas adsorption isotherms were used. Lately, some theoretically advanced papers have been published in this field. Radke and Prausnitz [ 11 used the methods of statistical mechanics to describe adsorption from dilute solutions. A very important group of papers dealts with the rhermodynamic description of adsorption from dilute solutions. In 1972 Radke and Prausnitz [2] applying the theory of ideal adsorbed solution [3] proposed a method of predicting adsorption from multicomponent solutions by using only the experimental single-solute adsorption data. The Polanyi potential theory was adapted to adsorption from dilute solutions by Manes et al. “41. Other interesting publications present an attempt to describe solvent effects in adsorption process. This problem was considered by Belfort [5] and indepen-
298 dently by Melander and Horvath [6]. All these theoretical approaches are consistent from the thermodynamic point of view. They do not introduce any special model of the adsorption process. On the other hand, they do not formulate any form of the isotherm equation and are not very suitable to study the influence on the adsorption process of factors such as surface heterogeneity, solute-solvent interactions or differences in molecular sizes of solutes. In the light of these remarks, simplified theoretical approaches such as for example Langniuir theory and its modifications may be very useful for predicting multi-solute adsorption. The main advantages of these approaches are that they make it possible to estimate the effect of different factors on the adsorption equilibria and that they give simple analytical forms of adsorption isotherms. This paper presents a simplified theoretical description of adsorption from dilute solutions. I t tries to include some important factors, such as energetic heterogeneity of the adsorbent, differences in areas occupied by solute molecules in adsorption space, pH of the solution and multilayer formation. The proposed relationships make it possible to estimate the influence of each of these factors on the course of the adsorption isotherm. Moreover. a method of selection of the single-solute isotherm equation, which may be treated as a characteristic of a given adsorption system, is presented.
2. ENERGETIC HETEROGENEITY OF THE ADSORBENT SURFACE
Heterogeneity of the solid surface is one of the most important factors determining the equilibrium of the adsorption process. This problem was discussed with particular reference to the adsorption of single gases and their mixtures, for which heterogeneity effects play an important role [7]. I n the case of adsorption from solutions the heterogeneity effects are not so noticeable because of competitive character of adsorption process in liquid/solid systems. However, it ought to be remarked that for the majority of systems investigated, this factor should be taken into consideration. In the case of liquid adsorption on heterogeneous surfaces many theoretical papers have been published on adsorption from liquid mixtures over the whole concentration range [7]. Adsorption systems containing dilute solutions were also discussed in some cases [8,9]. The discussion of adsorption from dilute solutions on heterogeneous solids presented here is a generalization of the results of our earlier papers [ 10-1 51. It is based on research methods developed for adsorption from gases and solutions over the whole concentration range [7, 16-20]. In the first stage, monolayer adsorption of n non-dissociated solutes of equal sizes from a dilute solution will be considered. The proposed isotherm equations describing the behaviour of such systems may be regarded as the special cases of two main relationships. The first of them, so called the generalized Langmuir (GL) equation [?I], extended for multicomponent adsorption, has the following form:
where 0 is the total relative adsorption of n components from a dilute solution: n
O= ZOi i=l
299 The variable x is defined as follows:
where
Ki is a Langniuir-type constant, proportional to the equilibrium constant of the adsorption process describing transfer o f a niolecule o f the i-th solute from the bulk phase to the adsorption space with tlie simultaneous transfer of a solvent molecule from the adsorption space t o the bulk solution. In tlie above relationships c, is the concentration of the i-th solute in the solution. K O , is the preexponential factor and E, is the adsorption cnergy of the i-th solute expressed with respect t o the adsorption energy of the solvent. However, El is a constant connected with the characteristic energy E l o f the energy distribution function; El defines the position of this distribution on the energy axis. The heterogeneity parameters m and ni‘ describe the shape of the energy distribution function. The parameter m characterizes the widening of the distribution function in the direction of lower adsorption energies, while m’- in the direction of higher energies; both parameters can vary from zero to unity. For special values of the parameters ni and ni’ eqn. (1) becomes one of the following relationships: (1) for ni = m’ = 1 it becomes a Langmuir-type isotherm (L), which describes the adsorption on homogenous solids ( 3 ) for it1 = ni’ E ( 0 , 1) it becomes the Langniuir-Freundlicli equation (LF) corresponding to a symmetrical quasigaussian energy distribution (3) for m = 1 and ni’ E ( 0 , 1) it becomes the generalized Freundlich equation (GF) relating to the exponential energy distribution (4) for in’ = 1 and m E ( 0 , 1) it becomes the Toth isotherm (T) corresponding t o an asymmetrical quasigaussian energy distribution. The second group of adsorption isotherms may be obtained from the so-called exponential equation [ 131:
x .
J
0 = exp [ Z Bj (RT j=1
111
_)’I X1
where B, are the heterogeneity parameters and X, is a constant connected with the niinimum value of the adsorption energy E l . For special values of the parameters Bj, the exponential equation (4) gives the following types o f adsorption isotherms: (1) when B1 > 0 and B, = 0 for j = 2, 3, ..., J it beconies the classical Freundlich isotherm (F) corresponding to an exponential energy distribution (2) when B, < 0 and B, = 0 for j = 1, 3, ..., J it becomes a Dubinin-Radushkevich-type equation (DR) relating t o an asymmetrical quasigaussian energy distribution (3) when B, < 0 and Bj = 0 for j = 1, 2, ..., J and j # n it becomes the Dubinin-Astakhov equation (DA). All o f these special forms of eqns. (1) and (4) were obtained by assuming that adsorption
3 00 of all solutes is characterized by identical energy distribution functions, which are shifted on the energy axis only. In the theory of liquid adsorption on heterogeneous surfaces, another method [17, 201 was also proposed for deriving the isotherm equations. It was first used by Jaroniec [17, 201 who obtained the following form of Langmuir-Freundlich (LF) equation:
Equation (5) corresponds to a symmetrical quasigaussian energy distribution. For the parameter m equal to unity, the Langmuir-Freundlich isotherm becomes a Langmuir-type equation, which gives the following relationship:
where ni ( c l , c 2 , ..., c n ) is the adsorbed amount of the i-th solute from an n-component mixture of solutes, and nF(cl + c2 t .... + c,) is the adsorbed amount of the i-th solute from the single-solute dilute solution. Equation ( 6 ) , similar to a thermodynamic relationship used in mixed-gas adsorption [22], may be very useful in studies of adsorption from dilute solutions asa heterogeneity test. In the case of experimental systems fulfilling eqn. (6) a Langmuir-type isotherm may be used. However, a system for which experimental points deviate from the theoretical relationship (6) ought to be characterized by an isotherm equation involving energetic heterogeneity of the adsorbent surface. The relationships presented in this paper describe the behaviour of multi-solute dilute solutions in contact with a solid. In addition, they have a special mathematical form they contain parameters characterizing adsorption from single-solute dilute solutions. This fact makes possible the prediction of adsorption from multicomponent solutions by using data for suitable single-solute adsorption systems. Such a procedure eliminates the time-consuming measurement of multi-solute adsorption. The usefulness of the proposed isotherm equations in such studies was discussed in earlier papers [ 10-13, 15, 161. The main difficulty in the use of the above procedure is a correct selection of the single-solute isotherm equation. The following method of such adjustment may be proposed: (1) In the first stage of this procedure the experimental system is investigated by means of the heterogeneity test relationship (6), (2) when the experimental points do not fulfil the relationship (6), the energy distribution function is calculated by using a general method. Knowledge of the distribution function shape makes possible selection of a simple isotherm equation giving best representation of the experimental data. The usefulness of this procedure is demonstrated for two dilute aqueous solutions of phenol and p-nitrophenol adsorbed on activated carbon B 10 at 293 K [23]. Firstly, the
301 above experimental systems were investigated by using the heterogeneity test (6). In the case of adsorption from a dilute solution containing two solutes, the relationship (6) becomes:
The results obtained by means of the relationship (7) are shown in Fig. 1. Evident deviations of the experimental points from the straight line predicted by eqn. (7) are observed. This means that the system must be described by the isotherm equations involving energetic heterogeneity of the adsorbent. Next, the energy distribution functions for both single-solute systems were calculated by using the known procedure of Jaroniec [ 24, 251. The shapes of these functions are presented in Fig. 2. In the case of both systems, the distribution functions are quasigaussian with a widening in the direction of higher energies. This indicates that a DR-type equation may be a most favourable for describing the above systems, because this relationship corresponds to such a distribution function. The above conclusions have been confirmed by additional numerical calculations made for four isotherm equations: Langmuir (L), Langmuir-Freundlich (LF), Dubinin-Radushkevich (DR) and classical Freundlich (F). Table 1 contains the values of the ratios of the standard deviations SD for different relationships to the minimal standard deviation SD,h. It follows from Table 1 that DR equation gives the best representation for the above data, whereas, in the case of Langmuir isotherm great deviations of the experimental points from the theoretical isotherm are observed. The LF equation also shows a good fit to the experiment, but worse in comparison to the DR relationship. As was expected, the Freundlich isotherm gives a poor agreement with the experimental data.
Fig. 1 . T h c linear relationship (7) for adsorption of phenol (1) + ynitroplienol(2) from dilute aqueous solution o n activated carbon B 10 at 293 K.
302
f (a
02
0.2 01
0
2
4
-€
6
Fig. 2 . Energy distribution functions f(e) characterizing the energetic heterogeneity of the activated carbon B 1 0 in relation to dilute aqueous solutions of phenol (-) and p-nitrophenol (---).
Tab. 1. The ratios of the standard deviations SD/SDmin for different isotherm equations calculated for single-solute adsorption from dilute aqueous solutions o n activated carbon B 1 0 at 293 K
phenol p-nitrophenol
0.0206 0.0240
21.80 13.76
1.11 1.19
1.00 6.53 1.00 3.30
SD - thc standard deviation calculated for a given adsorption system and a givcn isotherm SDnlin - minimal value of SD chosen from the standard deviations for a givcn adsorption system
W
-
number of cxperimcntal points
One may conclude that the proposed procedure for selecting the best isotherm equation may be useful in practice. Figs. 3 and 4 present the linear forms of the DR isotherm for the systems investigated: for single-solute systems: -
(- In @i)1’2 = B i1’2 RT In Xi
for bi-solute system:
-
Bi1’2
RT In ci
303 [-
(0,
+ 02)]1/2 = B1'*
RT In F1 - B'12 RT In (cl
+ KZ1 c2)
(9)
where B i and B are the heterogeneity parameters for single- and bi-solute systems, respectively. In both types of systems a good agreement between the experimental points and the theoretical straight lines is observed.
8
4
0
ln c;
4
Fig. 3. Thc linear relationship (8) for adsorption of phenol (a) and p-nitrophenol(0) from dilute aqueous solutions o n activated carbon B 10 at 293 K.
Fig. 4. The lincar relationship (9) for adsorption of phenol (1) aqueous solution o n activated carbon B 10 at 293 K. o - lo^ conccntration range 0 - high concentration range
+ p-nitrophenol(2) from dilute
3. MULTILAYER EFFECTS
Many experimental data indicate that the process of niultilayer formation on an adsorbent surface occurs frequently in adsorption from dilute solutions [7-6-18]. However.
304 this problem is rarely studied. A first approach to the description of such adsorption systems has been made by Hansen et al. [26, 271 and Schwuger [28], based on the BET-type equation. A relationship analogous to the Harkins-Jura isotherm was also used to describe adsorption from dilute solutions [ 291. Nevertheless, all these attempts are only an automatic adaptation of the gas adsorption isotherms to liquid adsorption and they neglect the specificity of the liquid/solid systems. Quite recently, multilayer single-solute adsorption was discussed by Jaroniec and Dqbrowski [30, 311. They postulated the use of the generalized Toth isotherm for describing adsorption systems containing dilute solutions [30]. In the next paper [31] they derived new isotherms for single-solute adsorption on heterogeneous solids using the BET-type relationship as the local isotherm in the integral equation. In this paper another approach to multilayer adsorption from dilute solutions will be presented. These considerations are based on the previously published theory of multilayer adsorption from multicomponent solutions over the whole concentration range [32]. Applying this model to adsorption from dilute solutions, the following relationship is obtained for the adsorption isotherm on a homogeneous surface:
In the above equation r is the number of adsorption layers and L(:) is proportional to the constant characterizing the transfer of the i-th solute molecule from the bulk phase to the k-th adsorption layer with the simultaneous transfer of a solvent molecule from the surface space to the solution. These constants are equal to:
L(r) = Kf-1) 1
(12)
1
where Kim)is proportional to the constant describing the transfer of the i-th solute molecule from the (k + 1)-th to the k-th adsorbed layer with the simultaneous transfer of a solvent molecule from the k-th to the (k + 1)-th layer. Taking into account studies of multilayer gas adsorption on heterogeneous solids [33] we assume that adsorbent heterogeneity exerts considerable influence on the formation of the first adsorbed layer only and may be neglected in the process of formation of the higher layers. Applying the method described in the papers [ 17, 201 the following form of the isotherm for multilayer multi-solute adsorption on heterogeneous adsorbents may be obtained:
j = 1'
J
J'
j=1
J
This relationship corresponds to a symmetrical yuasigaussian distribution of adsorption energies.
305
In
-10
-8
-6
-4
-2
Ln (c/col
0
Fig. 5. Theoretical isotherm eqn. (13) calculated for two values of the parameter m: 1.0 (-) 0.5 (......) and r = 1 , 3 , 5, 7.
and
Fig. 5 present the results of model calculations performed using eqns. (12) and (13) for the case of single-solute adsorption from dilute solution on homogeneous (m = 1) and heterogeneous (m = 0.5) surfaces. The theoretical curves, plotted in logarithmic scale, correspond to the monolayer isotherm equation (r = 1) and multilayer ones (r = 3, 5 , 7). The initial parts of the curves, corresponding to the region of monolayer formation, show greater adsorption on a heterogeneous surface than on a homogeneous surface. However, in the region of multilayer formation the heterogeneous effects are very small. These results are in a good agreement with the assumption of the proposed model limitation of the influence of adsorbent heterogeneity to the first adsorption layer only. Recapitulating, it may be stated that the model of multilayer adsorption reproduces the experimental multilayer isotherms and may be useful in their interpretation.
4. DIFFERENCES IN THE MOLECULAR SIZES O F SOLUTES
Although many experimental studies show that differences in cross-sectional areas of the solutes exert a significant influence on the adsorption equilibrium, theoretical investigations of liquid adsorption are usually carried out by assuming equality of their cross-sectional areas. Recently Jaroniec et al. [34, 35 J proposed a simple model for adsorption from dilute solutions involving energetic heterogeneity of the solid as well as the differences in molecular sizes of the solutes. They obtained a relationship defining the dependence between the adsorbed amounts and concentrations of two solutes. This relationship may be presented in two linear forms:
3 06 and In ( n i / c y ) = rn In Ki:
+ s In ( n j / c J ” )
(16)
where s is the ratio of the cross-sectional areas of the i-th and j-th solutes and KTi is the constant proportional to the equilibrium constant K i j (see eqn. (3)). where a proportionality factor contains quantities resulting from recalculation the mole fractions of solutes in both phases t o the numbers of moles ni, nj and concentrations ci, cj. The usefulness o f the proposed model for the description of the adsorption systems is examined by using the following experimental data: A p-chlorophenol(1) + plienylacetic acid (2) B p-nitrophenol (1) + o-phenyl phenol (2). These substances were adsorbed on activated carbon B 10 at 293 K [ 1 3 ] . Tab. 2. Parametcrs 111, s and I n K , characterizing adsoiption from bi-solute aqucous solutions o n activated carbon B 10 at 293 K ~
System
s
A
0.97 0.73 0.70 0.92
B
-4
in
In K,,
SD
1.60 -1.01
0.16 0.20
-2
0
2 0
a
I
Fig. 6 . The lincar relationship (15) for adsorption of p-chlorophenol(1) + phenylacetic acid (2) (a) and p-nitrophenol (1) + o-phenyl phenol (2) ( b ) from dilute aqueous solutions o n activated carbon B 10 at 293 K.
Table 2 contains the values of the parameters s, ni and In K:2. The linear relationships (15) and (16) are presented in Figs. 6 and 7. Now, we shall discuss the values of the para-
307
0
2
6
Fig. 7 . The linear relationship (16) for adsorption of p-chlorophenol(1) + phenylacetic acid (2) (a) and p-nitrophenol ( 1 ) + o-phenyl phenol (2) (b) from dilute aqueous solutions on activated carbon B 10 at 293 K.
meters m and s for both investigated systems. In the case of the system A the parameter s is close to unity and m is equal to 0.73. This means that heterogeneity effects play an important role and can not be neglected in the description of this adsorption system, whereas the role of differences in the solute sizes is very small. However, in the case of the system B the parameter s should be taken into consideration, but the heterogeneity effects play a smaller role (m = 0.92).
5 . DISSOCIATION EFFECTS
The influence of pH on adsorption from solutions was noticed some time ago [36,37]. A first attempt at the quantitative description of the dependence of adsorption on the pH of the solution was proposed by Getzen and Ward [38, 391, who considered this problem on the basis of the Langmuir equation. They treated a solution of a weak electrolyte as a bi-component system. However, Baldauf et al. [40], applying the same concept, used the modified Freundlich isotherm to include the effect of pH on adsorption. Rosene and Manes [41] discussed this problem on the basis of the Polanyi potential theory. The most advanced attempt at the description of adsorption from dilute solutions of weak organic electrolytes was presented by Muller et al. [42]. These authors presented a model to predict the influence of pH on adsorption, including energetic heterogeneity of the solid as well as surface charge. They considered an exponential energy distribution to characterize the adsorbent heterogeneity. Our discussion of the adsorption process from dilute solutions of the weak electrolytes takes into consideration the solid heterogeneity only, ignoring surface charge. Let us consider a simple model of adsorption from a dilute solution of an organic acid for example. The reactions representing the fundamental processes are the following: - the dissociation process in the bulk solution
K,.
RH - R
-
+ H+
308 where K a is the dissociation constant
and CH+ is the concentration of H’. - the adsorption of R-
(R-)* + (w)’
KR
* (R?’ + (w)’
(19)
In the above the superscripts ‘1’ and ‘s’ refer to the bulk phase and the adsorption space, ‘W denotes a molecule of solvent and K R is the equilibrium constant defined as follows:
- the adsorption of RH
KRH
( R H ) ~+ ( w ) ~ * ( R H ) ~+ (w)’ where K R H is the equilibrium constant defined by:
The total adsorption in the system investigated is the sum of the surface coverages of the ionized and neutral forms:
The total adsorption of both forms of the weak electrolyte may be expressed in terms of the total concentration c of the solute:
The concentrations of the two forms of the solute depend on pH as follows: C R = [a /
(1
-(-
a)] c
c ~ ~ = c / ( a) l +
(‘5)
where a = K, / C H
Putting the dependences (25) in the relationship ( 2 3 ) , the following form of the isotherm equation for adsorption from the dilute solution of a weak electrolyte on a homogeneous solid may be obtained:
The parameters KR and KRH in the above equation are the equilibrium constants for adsorption of the ionized and neutral solute forms, respectively. Their values may be obtained experimentally in the range of very low and high pH values, where only the ionized or neutral molecules exist in the solution. In this way, the eqn. (26) gives rise to the possibility of calculating the adsorption from solutions of different pH. Equation (26) may easily be generalized to the case of adsorption on heterogeneous solids by applying the method described earlier [ 17, 201:
(-1” 9=
1
1 +a
(KF a m +KgH
cm
The above relationship gives rise to computational possibilities analogous to eqn. (26).
6 . NOMENCLATURE ratio of dissociation constant and concentration of hydrogen ions heterogeneity parameter of eqn. (4) total concentration of ionized and neutral forms of concentration of hydrogen ions
eak electrolyte
concentration of solute i in dilute solution
sa tura tin:: concent rat ion of solute concentration of ionized form of \\.eak electrolyte concentration of neutral form of weak elcctrolytc concentration o f solvent adsorption energy of solute i exprcssed u.ith respect to adsorption encrgy of solvent characteristic energy of distribution function of adsorption energies dissociation constant Langmuir-typc constant, proportional to equilibrium constant of adsorption proccss of solute i equilibrium collstilllt reprcscnting compctitivc cliaractcr ot‘ adsorption preexponential factor in cqn. ( 3 ) constant conncctcd with charactcristic cncrgy of di,iriburion function equilibrium coiibtant describing transition of molcculc o f solute i fi-om ( k laycr of adsorption ,pace cquilibrium constrlnt defined by cqn. (20) cquilibrium constant defined by cqn. ( 2 2 )
+
I ) laycr to k-th
310 - constant connected with characteristic energy o f distribution function - constant connected with characteristic energy of distribution function - constant characterizing exchange of molecule of solute i from bulk phase with solvcnt molecule from k-th adsorbed layer m , ni’ - heterogeneity parameters of eqn. (1) - number of solutes in dilute solution n adsorbed amount of solute from single-solute solution n* - adsorbed amount o f i - t h solute from n-component mixture (eqn. (6)) ni - adsorbed amount of i-th solute from single-solute system (eqn. (6)) nT ~
I
- number of adsorption layers
R
-
-
S
T X
-
x,
universal gas constant ratio of molecular sizes of solutes i and j temperature variable expressed by eqn. ( 2 ) constant of eqn. (4) connected with minimum value of adsorption energy
adsorption energy total relative adsorption of n solutes from dilute solution relative adsorption of solute i from dilute solution 0i -- relative adsorption of ionized form of weak electrolyte OR @RH - relative adsorption of neutral form of weak electrolyte E
0
i k
-
1 R
RH S
vv
7.
solute adsorbed layer - bulk phase - ionized form of \\ eak electrolyte - neutral form of weak clcctrolytc -
- adsorption space - solvent
IIEFERLNCES
1. C. J. Kadkc and J . M. Prausnitz, J . Cheni. Phys., 57 (1972) 714-722. 2. C. J. Radkc and J. M. Prausnitz, AIChF J ., 18 ( 1 972) 76 1-768. 3. A. L. Meycrs and J . M. Prausnitz, AIChE J., I 1 (1965) 121-132. 4. M . Manes, in I . H. Suffet and M. J. MLGuire (Eds.), Activated Carbon Adsorption of Organics from the Aqueous Phase, vol. 1, Ann Arbor Science: Ann Arbor, Michigan, 1980, pp. 43-64. 5 . G . Belfort, Environ. Sci. Technol., 1 3 (1979) 939-946. 6. W. Melander and Cs. Horvith, in I. H. Suffet and M. J. McCuire (Eds.), Activated Carbon Adsorption of Organics from the Aqueous Phase, vol. 1, Ann Arbor Science: Ann Arbor, Michigan, 1980 131’. 65 -89. 7. M. Jaroniec, A. Patrykiejco and M. Bor6u-ko, in Progress in Surface and Membrane Science, vol. 14, Academic Press, Nem York, 1981, pp. 1-67. 8. Ch. Shcindorf, M. Rebhun and M. Sheintuch, J . Colloid Interface Sci., 79 (1981) 136-142. 9. M. Okazaki, H. Kape and R. Toei, 1. Chcni. Eng. Jap., 1 3 (1980) 286-291. 10. M . Jaroniecand M. Dcryto, Chcm. Ilngng. Sci., 36 (1981) 1017-1019. 11. A. Derylo and M. Jaroniec, Przem. Chcm., 6 0 (1981) 43-45. 12. M . Jaroniec and A. Derylo, J . Colloid IntcrFace Sci.. 84 (1981) 191-195.
31 1 13. A. Deryko and M. Jaroniec, C h c m Scripta, 19 (1982) 108--115. 14. M. Jaroniec, J. Colloid Interface Sci., 86 (1982) 588-589. 15. M. Jaroniec and A. Deryko, in 1.Pawkowski (Ed.), Physicochemical Methods for Water and Wastewater Treatment, Elsevier, Amsterdam, 1982, pp. 361-368. 16. M. Jaroniec, A. Deryko and A. W. Marczewski, Mh. Chem., 114 (1983) 393-397). 17. M. Jaroniec, J . Rcs. Inst. CatalysisHokkaido Univ., 26 (1978) 155-162. 18. M. Borbwko, M. Jaroniec, J. O k i k and R. Kusak, J. Colloid Interface Sci., 69 (1979) 311-317. 19. M. Borbwko, M. Jaroniec, and W. Rudzihski, 2. Phys. Chenue, Leipzig, 260 (1979) 1027-1032. 20. M. Jaroniec, J. O k i k and A. Derylo, Acta Chiin. Acad. Sci. Hung., 106 (1981) 257-265. 21. A. W . Marczcwski and M. Jaroniec, Mh. Cheni., in press. 22. D. M. Young and A. D. Crowell, Physical Adsorption of Gases, Butteru.orths, London, 1962, p. 402. 23. F. A. DiGiano, G. Baldauf, B. Frick and H. Sontheimer, Cheni. Engng. Sci., 33 (1978) 16671673. 24. M. Jaroniec, Surface Sci., 50 (1975) 553-564. 25. M. Jaroniec, A. Deryko and J. Czarniecki, Przeni. Chem., 61 (1982) 463-465. 26. Y. Fu, R. S. Hanscn and F. E. Bartell, J. Phys. Chem., 52 (1948) 374-387. 27. R. S. Hansen, Y . Fu and F. E. Bartell, J. Phys. Chem., 53 (1949) 769-785. 28. M. J. Von Schwuger, Koll. Z . , 234 (1969) 1048-1052. 29. K. P. Iyer and N. S. Wariyar, J. Inst. C h e m , Calcutta, 44 (1972) 111-116. 30. A. Dqbrowski, M. Jaronicc and J. T6th, J . Colloid Interface Sci., in prcs,. 31. M . Jaroniec, A. Dqbrowski and J. Toth, Cheni. Engng. Sci., in press. 32. M. Jaroniec, J. OScik and J. Deryko, Mh. C h e m , 112 (1981) 175-185. 33. M . Jaronie and W. Rudzihski, Acta Chim. Acad. Sci. Hung., 88 (1976) 351-362. 34. M. Jaroniec, Thin Solid Films, 81 (1981) L97-199. 35. M. Jaroniec, A. Deryko and A. W. Marczwski, Chem. 1:ngng. Sci., 38 (1983) 307-311. 36. J . J . Kipling, J . Chem. Soc., part 2, (1948) 1493-1499. 37. W. J . Weber and J. C. Morris, J . Sanit. Eng. Div., ASCE, 90 (1964) 79-86. 38. I:. W. Cetzcn and T. M. Ward, J. Colloid lnterface Sci., 31 (1969) 441-452. 39. T. M. Ward and F. W. Getzen, Environ. Sci. Tcchnol., 4 (1970) 64-67. 40. G. Baldauf, B. Frick and H. Sontheimer, Vom Wasscr, 49 (1977) 315-326. 41. M. R. Rosene and M. Manes, J . Phys. Cheni., 81 (1077) 1651-1657. 42. G. Miiller, C. J. Radke and J . M. Prausnitz, J. Phys. Chem., 84 (1980) 369-376.
This Page Intentionally Left Blank
313
SPECIFIC ADSORPTION OF ORGANIC MICROPOLLUTANTS ONTO ACTIVATED CARBON : A STUDY OF ELECTROKINETIC PHENOMENA DUE TO MULTICOMPONENT SYSTEMS
P. LAFRANCE, M. MAZET and D. VILLESSOT
Laboratoire de Gdnie Chirnique Appliqud aux Daitements des Eaux, Universite' de Lirnoges 123, rue Albert Thomas, 8 7060 Lirnoges Cedex, France
ABSTRACT Activated carbon adsorption is an important unit operation for the removal of micropollutants found in water supplies. Most of the adsorption studies that have been reported concern thc effectiveness of adsorption for removing a single solute from pure solution or a selected pollutant from a well-defined heterogeneous system. However, little has been done to investigate the physico-chemical factors responsible for adsorption affinities and capacities from multicomponent equilibria on activated carbon. The purpose of this study is to examine the adsorption specifity from two binary systems of pollutants in the micromilar concentration range. Experimental data showed an enhancement of adsorption capacity an anionic surfactant (sodium dodecyl sulphate) by cations (Na' and Ca2'), and an inhibition of adsorption capacity of 2-naphtol by sodium dodecyl sulphate. The influence of co-adsorption on the apparent adsorptive capacities for the single solutes is discussed in terms of surface potential changes of carbon particles during simultaneous adsorption of ionic compounds in solution. It is shown that a specific adsorption of one solute over another can be enhanced by electrostatic interactions due to the ionizable functional groups on the carbon surface and ion adsorption from solution. Complexation of the anionic surfactant to a bivalent cation surface species provides a cooperative adsorption on the hydrophilic sites of the carbon. Competitive adsorption and non-competitive adsorption of 2-naphtol by the anionic surfactant could illustrate the dependance of the adsorption equilibria on the interactions between species adsorbcd on the carbon surface.
1. INTRODUCTION
Adsorption onto activated carbon provides an effective technique for the removal of organic matter from domestic and industrial wastewaters, and also for final purification of drinking water. Developed as an important physico-chemical process for micropollutant recovery, the use of granular activated carbon (GAC) in large-scale adsorption plants can be improved by a fundamental exploration of chemical interactions at the solid-solution interface. The performance of activated carbon as an adsorbent is not only related to the specific surface area adsorption, the pore volume and pore-size distribution of GAC particles, but
314 also to the adsorption affinities for the single dissolved solutes. Furthermore, mutual effects of the adsorbates on the carbon surface properties, and interactions between micropollutants in bulk solution or adsorbed on the surface sites, can effect the adsorption selectivity of organic compounds found in water supplies. The mechanisms of co-adsorption from mixed aqueous solutions have had little study in the water treatment literature. Therefore, it is essential in the design of an adsorption technique that appropriate considerations should be given to interactions between the components of a multisolute system and a contact bed of GAC. The purpose of this study is to investigate the adsorption specificity from two binary systems of pollutants in the micromolar concentration range. Laboratory experiments showed an enhancement of adsorption capacity of an anionic surfactant (sodium dodecyl sulphate) by metal ions (Na’ and Ca”), and a reduction o f adsorption capacity of 2-naphtol by sodium dodecyl sulphate. This paper describes the experiments which were performed in order to assess the effects of co-adsorption on the apparent adsorption capacities of the single solutes, and the electrokinetic properties of carbon particles. Adsorption is described in terms of ultimate quantity of substsnce adsorbed as deduced by Langmuir isotherms, and surface potential changes of powdered activated carbon by zeta potential measurements. The influence of specific sites of adsorption and electrostatic interactions on the co-adsorption of solutes at the carbon interface is discussed.
2. EXPERIMENTAL METHODS 2.1. Material and Analysis
The adsorbent used throughout this study was Chemviron Filtrasorb-400 active carbon, which is marketed for use in water purification. The carbon granules are manufactured in 0.4-1.7 mm in diameter (12 X 40 U.S. Standard mesh size), with a BET-N2 surface area of approximatively 1200 mz /g. To prepare the powdered carbon for the experiments, the granules were mechanically ground and sieved to < 50 pm. Experiments showed that drying of this fraction at 150°C for 3 hours did not affect the adsorption capacity of the powdered carbon. The adsorbates that were used in the first binary system included reagent-grade NaCl and CaClz. A technical grade sample of sodium dodecyl sulphate (SDS) was purified twice by crystallization in anhydrous ethanol and found to be 2 99.5% pure in terms of elementary analysis. Colorimetric determination of the surfactant concentration was carried out by benzene extraction with crystal violet and was read at 605 nm. The second binary system of pollutants included the purified salt of dodecyl sulphate and reagent-grade 2-naphtol. Individual concentrations of 2-naphtol were determined by ultraviolet spectrophotometry at 224 nm (1 cm cell length), using a BECKMAN model 25 apparatus. The chloroform extraction with Caminoantipyrine and colorimetric determination at 460 nm did not improve the detection level of 2-naphtol. Stock solutions and working solutions of the experimental systems were prepared with distilled water at pH = 4.5-5.0.
315 2.2. Adsorption Isotherms
The adsorption isotherm tests consisted of mixing a solution containing a known concentration of one or more solutes with several weighed amounts of powdered activated carbon in 1 liter of distilled water. Batch studies were performed at room temperature with continuous stirring of the samples. A contact time of 3 hours was allowed to ensure an apparent adsorption equilibrium, related to the rapid surface transfer of the adsorbates to the carbon macropores. These apparent equilibrium isotherms. called "pseudo-isotherms", account only for that fraction of the adsorption capacity utilized in the initial rapid adsorption period. After equilibration the powdered carbon was removed by filtration through fiberglass filter paper, and the concentration of the adsorbate was measured. In the first bisolute batch. a constant initial concentration of SDS ( 1 .O mg . 1 ) was mixed with different concentrations of the second solute, Na' or CaZ+.After simple filtration of the binary system, no change of pH and no loss of SDS by precipitation, as ionic strength changed, were observed. The second bisolute batch contained a constant initial concentration of 3-naphtol (1.0 mg . 1-') with different concentrations of SDS as co-adsorbate. After equilibration, the extent of adsorption for SDS (first binary system) and 3-naphto1 (second binary system) was determined and the ultimate adsorption capacity for the carbon calculated using the Langmuir model for individual components.
-'
2.3. Electrophoretic Measurements
The zeta potential ({) value of activated carbon particles was obtained directly using a Pen-Kem LAZER ZEE METER model 500 apparatus. Measurements were carried out on a batch of 40 mg . 1-' suspension of activated carbon in distilled water, containing different concentrations of one or more solutes. Some variation of the zeta potential value can occur with time for the carbon suspension blank, and determinations on agitated samples were made at initial contact time and after 3 hours contact time, at room temperature. The zeta potential data were corrected for all temperature changes and expressed at a reference temperature of 20"C, according to the Pen-Kem instructions.
3. RESULTS AND DISCUSSION
The Langmuir equation was used to describe single-solute adsorption from dilute aqueous solutions [ 1-31. The ultimate adsorption capacity is then obtained at the intercept, r,', of the straight line by plotting the reciprocal adsorption density,F'versus the reciprocal equilibrium concentration of the solute. 3.1. Adsorption Isotherms for Individual Solutes
The adsorption capacity for the anionic surfactant is determined in distilled water, in tap water, and in two solutions of varying pH (Figure 1).
316 2.0
1.8
I
," 1.6 2
-E ?--
I
1.L
0 distilled
water
tap
1.2
1 .o
0
A
p H = 9.8
v
p H = 4.0
1
1
I
1.0
2.0
3 .O
(SDSI-'
watet
1
4.0 (pM)-'
L 5 .o
Fig. 1. Lanpnuir isotherms for SDS adsorption o n activated carbon. Effect of pH and water mineralization.
Langmuir pseudo-isotherms for SDS in distilled water (conductivity = 2 pS * cni-') and tap water (conductivity = 75-80 pS . a n - ' ) show sinular values of ro,approxiniatively 0.70 niM/g. However, isotherm plots showed that micromolar quantities of SDS ( S 1.O mg . 1 - I ) will be preferentially eliminated in distilled water. For dilute aqueous solutions of anionic surfactant, the effect of pH on the carbon efficiency is limited in constant ionic strength media of 80 pS . cm-' . Thus, only a minor decrease of the ultimate adsorption capacity value, from 0.78 mM/g to 0.76 niM/g, is obtained for a pH change from 4.0 to 9.8. The difference observed between isotherm plots for pH values of 4.0 and 9.8 is in accord with the adsorption behavior of typical organic pollutants, resulting from repulsive electrostatic interactions between their polar or negatively charged groups and the anionic surface functions on the carbon [4-51. The Langniuir pseudo-isotherm for 2-naphtol (K, = 1.17 . lo-'' at 20" C) in distilled water is presented in Figure 2. The ultimate adsorption capacity for 3-naphtol is 1.70 mM/g (UV method) with a carbon contact time of 3 hours, and 2.95 mM/g after an adsorption equilibrium time of 3 days. Such results corroborate the influence of slow adsorption kinetics (internal diffusion of the adsorbate to macropores of the carbon) on
317
0.8
0.6
0 .L 7
I
L
0.2
'l 3 days
-
1 -
0 C.=1.0
mgil
o Co=l.O
mg/l
G-aminoant.
-
C0=2.0 m g / l 1 U . V .
A
method
method
Co=l.Omg/l I
0 0
~~~~
0.2
~
1
I
0.L
0.6
( 2 - n a p h t o l 1-1
I
0.8
1 .o
IpM 1-l
Fig. 2. Langmuir isotherms for 2-naphtol adsorption o n activated carbon.
the extent of adsorption isotherm parameters [6]. This kinetic-determining step of the adsorption process greatly affects the predictive evaluation of GAC contactors performance [7-81. For identical carbon quantities, a much slower approach t o a true adsorption equilibrium will be observed for higher initial concentrations of the solute [6]. Isotherm data obtained for a fixed time of 3 days show a decrease of the apparent adsorption capacity with an elevation of the initial concentration of 2-naphtol. The slow uptake capacity for micropollutants illustrates the difficulty in defining the attainment of true equilibrium in activated carbon isotherm studies. Batch co-adsorption of organic compounds present in heterogeneous systems can induce modifications of the isotherm equilibrium parameters found for the single solutes. Since the linear form of experimental pseudo-isotherms established for single-solute systems was satisfactorily conserved in the presence of a co-adsorbate, we will consider the variation of Po parameter for one micropollutant simultaneously adsorbed with a promotor compound, and an inhbitor compound of adsorption.
3.2. Adsorption from Bisolutc System
The first binary system concerns adsorption efficiency of SDS onto activated carbon. This was considerably improved by Na' ions and Ca2' ions as co-adsorbates. Many re-
318
>
10
1.0
8
0.9
7 0
E
-
6
.
a
d
4
0.7
2
0.6
0
0.5 0
0.02
0.04
0.06 CNi3'
0.08
0.1 0
(Mi
Fig. 3. Effect of sodium ion concentration on maximum adsorption capacity of SDS (a), and <-potential of activated carbon particles at initial contact time ( 0 ) and 3 hour contact time ( 0 ) .
-
1.o
m
2
0.9
E
0.8 0
L
0.7
0.6 0.5 PCa Fig. 4. Effect of calcium ion concentration o n maximum adsorption capacity of SDS on activated carbon.
ports have observed an elevation of the anionic surfactant adsorption on kaolinite [9] and SDS adsorption on activated carbon [lo-121 in the presence of simple electrolytes. This enhancement of the SDS uptake by activated carbon can be seen in Figure 3 for Na' co-adsorption and Figure 4 for CaZ+co-adsorption.
319 In the case of Na' ions, the promotion of the roparameter requires a high concentration of the univalent ion. For an ionic strength value of 0.10 M, the attainment of the largest value of ro= 1.02 mM/g represents an elevation in ultimate capacity for SDS of 44%. The linear increase of adsorption capacity for SDS in the presence of a high Na' concentration, strongly suggests that this favoured co-adsorption is probably due to a modification of the electrostatic environment of the carbon surface, rather than a particular SDS-Na' complex formation on adsorption sites. t o an hyperbolic In the case of Ca" ions, the promotion of adsorption for SDS form of ro data versus Ca" concentration. Thus, micromolar concentrations of divalent ions are sufficient to improve the carbon adsorption efficiency for the anionic surfactant. The largest value of ro = 1.05 mM/g,'representing an increase in ultimate capacity for SDS of 47%, can be reached at a molar ratio Ca2'/SDS = 10 (13.3 mg . 1-' in Ca"). In the binary system divalent cation-anionic surfactant, previous work [ 13-14] showed that the solubility product of the (DS)zCa complex is much higher than the solutes concentrations used in our study. Furthermore, the hydrolytic stability of SDS is not affected in dilute aqueous solutions and moderately acidic conditions [ 151. One might expect that the low calcium ion concentration required to obtain a promotion of adsorption for SDS. allows an inter-adsorbate association mechanism affecting: (1) the relative solubility of the surfactant by formation of associate species (DS),.Ca,; ( 2 ) the ionic environment of macropores by lowering the surface potential of carbon particles, and; (3) the electrostatic binding between functional sites of the carbon and the anionic group of the surfactant (ionic bridges by direct adjunction of Ca"). The second binary system concerns the adsorption efficiency of 2-naphtol onto activated carbon. This was reduced by an anionic surfactant as co-adsorbate (Figure 5 ) .
--.
1.0 0
6.0
5.8
5.6
5.4
p SDS Fig. 5 . Effect of SDS concentration o n maximum adsorption capacity of 2-naphtol on activated carbon.
5.2
5.
320 The inhibitive co-adsorption for some aromatic compounds in the presence o f an anionic surfactant has been studied b y many workers. Weber and Morris [2] showed that for equimolar concentrations of micropollutants, the alkyl benzene sulfonates (ABS) have a much more pronounced effect o n the rate o f adsorption of aromatic compounds than the latter has o n t h e rate o f uptake of t h e surfactant. The authors attribute this difference of mutual inhibition to the retarding o f intraparticle transport o f small molecules by steric hindrance due t o ABS large molecules. However, this adsorption selectivity caused b y the accessibility of the porous structure cannot account entirely for t h e “competitive” and “noncompetitive” aspects of coadsorption. The examination of t h e binary system neutral p-nitrophenol/anionic salt of benzene sulfonate b y Jain and Snoeyink [3] has demonstrated the noncompetitive aspect of inhibitive co-adsorption for organic compounds of similar molecular size. Noncompetitive co-adsorption (for adsorption sites accessible for all adsorbates but specific for only one of the solutes) can be explained by the electrostatic influence o f functional surface groups of the carbon on the adsorption affinity for neutral o r ionic compounds in solution. The promotion and inhibition of adsorption from binary systems of micropollutants are then dependent o n the selective adsorption properties presented by functional sites on the surface o f the carbon. Also, interactions between adsorbates in bulk solution o r adsorbed o n surface sites, can affect the adsorption specificity of activated carbon for heterogeneous systems of micropollutants. The modification of the apparent surface potential of carbon particles during adsorption from single or niultisolute systems suggests the probable mechanism o f equilibria relations in those adsorption systems.
3.3. Influence of Individual Adsorbates on Zeta Potential of the Carbon
The experimental measurement o f the zeta potential of carbon particles is approximately -20 mV in distilled water, indicating an “acid-type” carbon, according to [j]. Among the ionic species used in this study, we shall distinguish: ( 1 ) indifferent electrolytes which are not expected t o enter into a special interaction with the electrical double layer of an activated carbon particle, such as Na’ ions: (2) potential-determining ions (p.d.i.) which exert a fundamental control o n the surface charge and the potential at the surface of the dispersed phase, such as H’ and Ca2’ ions, and: (3) organic species intermediate between these extreme ions which appear t o interact in some chemical \\.a)’ with the surface, and are referred to as specifically adsorbed ions, such as DS- ions. Neutral species 2-naphto1, in the range of concentration used for isotherm studies. sho\ved no influence o n the zeta potential value o f carbon particles. In Figure 3, it can be seen that a linear reduction of zeta potential of carbon particles is obtained as a function o f ionic strength o f the indifferent electrolyte, h’a’. The important charge density developed by high concentrations o f univalent cation is then acsonipanied by a corresponding compression o f the coulombic influence zone of the ionic surface sites o f the carbon. I n the case of potential-determining ions. the zeta potential change of carbon particles is indicated in Figure 6 for H‘ ions and in Figure 8 for Ca” ions. The H’ concentration required to make the zeta potential zero led to an apparent isoelectric point (i.e.p.1 at
32 1
0
- 10
>
-20
-E u
-
30
- 40
- 50 2
3
4
5
6
7
8
9
10
PH Fig. 6 . Effect o f pH on f-potential o f activated carbon particles.
pH = 2.6. For CaZ+ions, the compensating effect of small quantities of this divalent cation on the surface charge provide a marked potential drop of carbon particles after initial contact time. After 3 hours contact time, examination of Figure 8 shows that there is a fall in the zeta potential change, which tends to be close to zero with much longer contact times. On the other hand, in contact with small concentrations of SDS, the zeta potential value of carbon particles is sharply lowered in the negative region (Figure 10). A negative variation of the apparent charge of different kinds of surfaces, due to anionic surfactant adsorption, has been indicated by many other investigators [ 17-20]. As in the case of calcium potential-determining ion, it should be noted that the zeta potential change of carbon particles decreases markedly after a 3 hour contact time (Figures 7 and lo), to become relatively insensitive to the SDS concentration at a longer equilibrium time. The active surface for adsorption developed by carbon particles may be described, in a simplified model, by different proportions of hydrophobic and hydrophilic surfaces, respectively related to the aromatic rings of the graphite carbon [5] and to the polar o r ionic functional groups at the carbon surface [21]. The modification of zeta potentcal
322
- 25
-30 > E
-35
-40
0
1 .o
0.5 time
1.5
3.0
(hours)
Fig. 7. Dependence of f-potential of carbon particles on varying contact time with adsorbates.
CSDS= 1 . 0 m g . 1 . ' .
of carbon particles in binary systems of micropollutants can provide information about the influence of a co-adsorbate on the affinity and the specificity of adsorption sites for another organic compound in solution.
3.4. Specific Adsorption and Zeta Potential due to Multicomponent Systems
The zeta potential modification of carbon particles in heterogeneous adsorption systems is indicated in Figure 9 for the Ca2'/SDS system, and in Figure 10 for the SDS/ 2-naphtol system. As stated in the isotherm studies section, the improvement of adsorption for SDS by simple electrolytes needs a high ionic strength of Na' ions, whereas micromolar quantities of Ca" ions induce an hyperbolic promotion of the SDS adsorbed layer. For a molar ratio Ca2'/SDS < 10, examination of Figure 9 shows that the divalent cation, present in the binary system, causes a depression of the carbon surface potential similar in magnitude to that obtained by Ca2+ions alone (Figure.8). Accordingly, the "neutralization" of the negative surface groups on the carbon by Ca" ions can provide a decrease of the repulsive forces towards the anionic group of the surfactant. This will favour electrostatic interactions between (DS)2nCa, species and func-
323
14 12
-E
10
>
Fig. 8. Effect of calcium ion concentration onf-potential of activated carbon particles. o initial contact time; 0 3 hours contact time (broken line: with derivation of the carbon suspension blank).
tional sites of the carbon. The difference observed between ultimate adsorption capacities for SDS alone and SDS in a binary system with cations, could then represent the restriction to the SDS adsorption caused by ionic surface groups of the carbon. Since calcium ions are the potential-determining ionic species in the binary system, it is quite possible that CaZ+will be first preferentially adsorbed, to allow a calcium-surfactant complexe formation on the surface of the carbon. The formation of (DS)Ca’ and (DS)z Ca species, with increasing ionic strength of the solution, could then lead to a sequential adsorption (probably competitive between all transitory (DS)2nCa, associates) of different forms of surfactant, free or complexed with cations [ 121. Thus, micromolar concentrations in CaZ+ions will provide a cooperative co-adsorption of monomers and subsequently dimers of anionic surfactant onto activated carbon. In the binary system SDS/2-naphtol, the modification of the zeta potential of carbon particles by SDS co-adsorption, is presented in Figure 10. In such a system with a molar ratio SDS/2-naphtol< 1.0, it can be seen that the zeta potential change in the presence of a constant initial concentration of 2-naphtol is more pronounced than that obtained with SDS alone. According to [5],the fixation of phenolic compounds onto activated carbon is expected to proceed by donor-acceptor interactions through the 7-r-electron of the aromatic rings of the system. This “charge-transfer’’ complex of the nonspecific n-bonding can be
3 24
12
Q
6
2
0
Fis. 9. Effect of calcium ion conccntration 011 <-potential of activated carbon particles (binary s)’strm with CSDS = 1.O mg . 1 ). o initial contact time: 3 hours contact time (brokcn linc: \\.itli derivation of thc carbon suspension blank).
affected by t h e vicinity o f ionized surface groups of the carbon, which can change or deactivate the overall electron density in the n-system of the aromatic rings o f graphite. Such mechanisms could influence the stability of the donor-acceptor complex and consequently reduce the apparent adsorption affinity of the carbon for the phenolic compounds. Previous work [ 121 showed that simple electrolytes Na’ and CaZ+did not affect the adsorption rate of micromolar quantities of 7-naphtol onto activated carbon. Furthermore, no modification o f t h e surface potential o f activated carbon particles occurs.with 7-naphtol adsorption from solution. The above experimental observations suggest that co-adsorption of 3-naphtol with DS- ionic species led to a competitive and a noncompetitive inhibition of the 7-naphtol uptake by activated carbon. Competitive inhibition of the 7-naphtol adsorption could be related t o that fraction of SDS adsorbed o n the carbon through hydrophobic interactions between t h e long hydrocarbon chain and the underlying surface also accessible for t h e phenolic compound. Noncompetitive inhibition of 9,-naphtol adsorption can occur for that fraction o f SDS adsorbed o n the specificauy hydrophilic sites of the carbon, through chemical interactions even with an unfavourable surface charge [ 2 2 ] . This noncompetitive aspect of co-adsorption could then illustrate the possible “longrange” electrostatic influences of surface functional groups o n the apparent adsorption
325 p SDS
6.0
5.8
5.6
5.4
5.2
5.0
0
- 2 - 4
< - 6
-
-d
8
- 10
-1 2
-14
Fig. 10. Effect of SDS concentration o n r-potential of activated carbon particles. a no added 2-naplitol; with 2-naphtol= 1.0 nig . 1 -' . Open points from initial contact time, filled points from 3 hour co tit act time.
affinity of carbon for aromatic compounds in solution. Finally, other cheniical factors, such as possible interactions between adsorption sites being occupied and specific for only one of the solutes (the drop of the interfacial tension and the increase in external negative charge due to the SDS adsorption, being unfavourable to the 3-naphtol adsorption), can improve tlie selectivity o f adsorption presented by activated carbon for an heterogeneous system o f micropollutants. The origin of chemical interactions which are important in determining the degree of adsorption from a complex natural water system is not yet entirely explained by the measurement of the total diffuse doublc-layer potential of carbon particles. A viable interpretation of equilibria mechanisms occuring during the rapid phase of co-adsorption of complex solutes needs a more detailed study about the chemical significance of isotherm parameters. There is no doubt that tlie knowledge of the adsorption process will be significantly improved by a more consistent n~olecular interpretation o f adsorption isotherm parameters used t o described the activated carbon performance.
4. CONCLUSION
A description of phenomena occuring at the solid-liquid interface during bisolute adsorption from dilute aqueous solutions, can be approached by electrokinetic measure-
326 nients of potential changes of carbon particles. Preferential adsorption o f one species over another can result from electrostatic interactions due to ionizable functional groups o f the carbon, and ion adsorption from solution. The observations made on two bisolute systems have provided an experimental corroboration of the dependence of adsorption capacity onto activated carbon on the zeta potential value of the carbon surface. It appears from these results that: (1) an enhancement of adsorption capacity o f an anionic surfactant can be realized by the compression o f the electrical double layer due t o the increase in ionic strength, thus making it easier for the surfactant to approach the surface. The complexation o f SDS to a CaZ' surface species provides a dramatic promotion of the adsorption capacity o f the surfactant on hydrophilic sites o f the carbon, ( 2 ) an inhibition of adsorption capacity of 2-naphtol by SDS can originate from competitive and noncompetitive fixation o f the solutes on the carbon surface. Competition accounts for that fraction of surface available for both adsorbates, and noncompetition for different kinds of sites specific for each solute, but being affected by electrostatic changes in their neighbourhood during co-adsorption. Further investigations are required t o make the adsorption specificity from multisolute equilibria, encountered in natural water systems, comprehensive. The adsorptive properties of activated carbon are the result of a variety of mechanisms concerning adsorption kinetics and adsorption equilibria in heterogeneous systems o f micropollutants. Optimum design and evaluation o f the suitability o f carbon contactors as a unit process for water treatment, can be obtained only if factors that control capacities for adsorption are well defined and understood.
REFERENCES
1 W. J . Weber Jr., and J. C. Morris, J. San. Eng. Div., Am. Soc. Civ. Eng., 90 SA 3 (1964), 79-107. 2 W. J . Webcr Jr., and J . C. Morris, J. Am. Water Works Ass., 4 (1964), 447-456. 3 J . S. Jain and V . L. Snoeyink, J. Water Pollut. Control I'cd., 12 (1973), 2463-2479. 4 W. J. Weber Jr., Pliysico-chemical Processes for Water Quality Control, Wiley-Intcrscicnce, N.Y., 1972, Cli. 5. 5 J. S. Marrson and H. B. Mark Jr., Activated Carbon. Surface Chemistry and Adsorption from Solution, M. Dekker (Ed.), N.Y., 1971. 6 R. G. Peel and A. Bcnedck, Environ. Sci. Technol., 1 (1980), 66-71. 7 J . Malleviale, T h e Interaction of Slow Adsorption Kinetics and Bioactivity in Full-scale Activated Carbon Filters: The Development of a Nebv Prcdicive Model, paper submitted to the Cliemviron A\vard Comittee, February 1980. 8 F. Fiessinger and Y. Richard, Proc. Annual Conference of the American Water Works Association, Atlanta, Georgia, June 15-20, 1980. 9 P. E. Figdore, J. Colloid Interface Sci., 2 (1982), 500-517. 10 M. J. Schwuger and H. B. Sniolka, Colloid Polymer Sci., 6 (1977). 589-594. 11 I. Abe, K. Hayashi and M. Kitagawa, Nippon Kagaku Kaishi, 12 (1977), 1905-1910. 12 P. LaFrance, M. Mazet and D. Villcssot, submitted for publication 13 R. S. Lee and I. D. Robb, Faraday Trans. 1, Cliem. SOC.,9 (1979), 2116-2125. 14 I. Krznarik, J . Bozi6 and N. Kallay, Croat. Chem. Acta, 3 (1979), 183-189. 15 J. Sanchez, J. J. Garcia Donihguez and J . L. P a m , Invest. I d . Text. Tensioactivos, Z (1981), 141-148. 16 J. J . McCreary and V. L. Snoeyink, J. Am. Water Works Ass., 8 (1977), 437-444. 17 H. M. Rcndall, A. L. Smith and L. A. Williams, Faraday Trans. 1, CIwni. SOC., 3 (1979), 669-678. 18 B. D o b i k , Colloid Polymer Sci., 7 (1977), 682-690.
327 19 R. K. Mishra, S. Chander and D. W. Fuerstenau, Colloids and Surfaces, 1 (1980), 105-1 19. 20 T. Murata and Y. Matsuda, Denki Kagaku Oyobi Kogyo Butsuri Kagaku, 2 (1981), 127-128. 21 J. T. Cookson Jr., C. Ishizakiand C. R. Jones, AIChe Symp. Series, 68 (124), Water-1971, 157-168. 22 R. J . Hunter, Zeta Potential in Colloid Sciencc. Principles and Applications, Acadcmic Press, 1981, Ch. 8.
This Page Intentionally Left Blank
329
FURTHER STUDIES ON THE USE OF CHEMICALS TO REGENERATE EXHAUSTED ACTIVATED CARBON
R. J. MARTIN*
Department of Civil Engineering, University of Birmingham, P.0. Box 363, Birmingham, England W. J . NG Department of Civil Engineering, National University of Singapore, Singapore 051 1
ABSTRACT This paper represents the second of a series of reports o n the chemical regeneration of exhausted activated carbon follouing experimental research studies carried out at the University of Birmingham, U K . A \tide range of regenerants, inorganic and organic, \\as evaluated in the treatment of carbon samples exhausted \\ith 2-naplithol,2methoxyphenol,k h l o r o p h e n o l , o-cresol and 2-nitrophenol, in order to compare the effects of the introduction of a second benzene ring (i.e. in effect thc addition of C, H, ), OCH ) , Cl, CH, and NOIg o u p s to phenol o n susceptibility to desorption. Organic chcmical regenerants \L ith solubilising pou c'rs \yere found to be generally much more effective than inorganic chemical rcgenerants \\ ith oxidisitif: powers. In gcneral, the efficacy of the organic regenerants \! ithin a group decreased as their molecular \wights increased: the smaller the regenerant, tlie further it could penetrate into the micropores of the carbonand displace the adsorbate. For the adsorbates studied, there \\as a n m k e d correlation between decreasing molecular \vc'ight of adsorbate and decreasing value of regeneration efficiency; the smaller the adsorbate. the furthcr i t could penetrate into the micropores of thc carbon thereby resisting displaccnicnt by the regenerant. Boiling \later \ \ a s observed to be of regenerating value. The efficiency of regencration b!' boiling \varcr \\as influenced by the initial adsorption capacity of the carbon for the adsorbate: lo\\ adsorption capacites i n turn resulted in lo\\ desorption capacities.
1. INTRODUCTION
Activated carbon adsorption has developed as an important unit process for the removal of a wide range of organic pollutants from waters and wastewaters in both municipal and industrial sectors of areas where tlie concentration of population in urban centres has resulted in pollution and water resource problems.
* To
horn corrcspondencc should be addressed.
330
Although the internal pore structure of an activated carbon provides a very high surface area, this surface area is finite and must eventually become covered with adsorbate; adsorption, which is a surface phenomenon, ceases at this stage. When granular activated carbon (GAC) becomes exhausted or when the effluect from a GAC bed reaches the maximum allowable discharge level imposed, the spent carbon must be processed to remove the adsorbate thereby regenerating the carbon for subsequent use as an adsorbent again. The most common technique practised in regeneration is thermal volatilization in which adsorbed organics are desorbed by volatilization and oxidation at high temperature; this thermal regeneration technique is characterized by the loss of carbon (perhaps 5-10%) due to oxidation and attrition, and by the cost of energy in heating the carbon to around 800-850°C [I]. An alternative technique is that of chemical regeneration in which chemical reagents are applied to the exhausted carbon. Two main categories of reagents may be used: inorganic chemical regenerants with oxidising powers and organic chemical regenerants with solubilising powers. A literature survey on chemical regeneration as an alternative to thermal regeneration indicated that relatively little work has been done. There remain areas where either little is known or conclusions drawn by various workers are in conflict. The essential objective of the research, therefore, was to explore, in a detailed and logical manner, the chemical regeneration of exhausted activated carbon for a wide range of both adsorbates and regenerants and in particular to examine the possible importance of the relationship of the molecular size, structure and properties of the adsorbate to those of the regenerant. T h s paper presented by the authors represents the second of a series of reports on chemical regeneration following research studies carried out at the University of Birmingham. The first report by Martin and Ng [2] considered the effect water may have on the powers of organic regenerants in regenerating carbon samples exhausted with nitrobenzene; ihe report also contained studies on the regeneration of carbon samples exhausted with a range of mono-substituted benzene compounds. A literature survey on chemical regeneration and a critical appraisal of the literature were features of the initial report of the authors [2]; it is not intended that this appraisal of the literature be reproduced here. The work described here considers the third phase of the research; studies were performed using a series of adsorbates that were all substituted phenols-2-naphthol, 2-methoxyphenol, 2-chlorophenol, 2-nitrophenol and o-cresol. Samples of carbon were exhausted with these adsorbates and regenerated with various regenerants to evaluate the significance of the molecular properties of the compounds involved.
2. EXPERIMENTAL PROCEDURE
Filtrasorb 400 GAC supplied by Chemviron Ltd. was used throughout the experimental research. It has been used in previous adsorption research studies and has been found to be effective for the adsorption of small molecular weight organic pollutants (as used in the regeneration research reported here) and for the adsorption of much higher molecular weight organics (as used in the regeneration research but to be reported in the future).
33 1
The exhaustion of the carbon was carried out using established techniques [3, 41. After exhaustion, the carbon was subjected to the regeneration phase [ 2 ] . Following exhaustion and regeneration, the carbon was then subjected to re-exhaustion. The equilibrium concentrations of the adsorbates reported here were determined by gas-liquid chromatography with a flame ionization detector; direct injection of the aqueous samples was employed. This has been found to be an excellent technique for monitoring the adsorption process for mono- and multi-solute solutions [3, 5 - 71. The fact that gas-liquid chromatography is an organic-specific method of analysis means that in the application of organic solubilising regenerants, differentiation between residual solute and regenerant can be properly made. 3. CALCULATION OF REGENERATION EFFICIENCY
The efficiency of any one regenerant is judged on the extent that it effects a recovery of the adsorptive powers of the carbon. The following method of calculation was employed to quantify this recovery. The original capacity (A,,) of the carbon for a particular adsorbate was deemed to be that quantity of solute adsorbed from solution per unit weight of carbon (experimental conditions chosen to ensure equilibration and exhaustion). The capacity of the regenerated carbon (A,) was deemed to be that quantity of the same solute adsorbed from solution per unit weight of regenerated carbon (again ensuring both equilibration and exhaustion). Regeneration efficiency (RE %) = (A,/A,) X 100. 4. RESULTS A N D DISCUSSION
4.1. Boiling with Water. Following exhaustion with their respecthe adsorbates, the carbon samples were rinsed and then placed in 200 id of boiling distilled water and boiled for various times. The 200 ml volume of boiling water was maintained throughout. The results are presented in Figure 1. It was observed that maximum REs were achieved after approximately 4.5 hours for three of the five adsorbates studied; this observation was also recorded in experiments with the mono-substituted benzene compounds [2]. The curve for 2-nitrophenol showed that little would be gained, in terms of RE, by boiling beyond approximately 4.5 hours. The exception to this general trend was o-cresol; as is shown in Figure 1, RE values for carbon samples exhausted with o-cresol increased as the duration of boiling was extended. As a check, the tests were repeated; the same rising curve and absence of a plateau were noted. Excluding o-cresol, on the basis of maximum REs achieved, the ease with which adsorbates could be removed decreased in the following order.
NO,
> C1> C4H4> OCH,
It is interesting to compare this progression with that for decreasing adsorption capacities which is presented as follows.
NO,
> C,H,
and C1> OCH,
> CH,
332
-t- 2-Nitrophenol
-.......=-......- 2 - Chlorophenol
.-.-
--*--..-.. 22-..-“Q
*
‘0
A
1
o-cresot Naphthol Melhoxyphenol
---_--2
3 -Boiling
4
5
6
7
Time (hours 1
Fig. 1. Effect of Boiling o n Adsorbates.
The two progressions may be seen to be fairly similar. This would suggest that the efficiency of regeneration by boiling water was influenced to a certain degree by the initial adsorption capacity of the carbon for the adsorbate; low adsorption capacities in turn resulted in low desorption capacities. This trend was observed in the studies on the mono-substituted benzene compounds [2]. N. B. Previous work by Al-Bahrani and Martin [5] compared the adsorption of the five substituted phenols using the activated carbon CAL supplied by Chemviron Ltd; adsorption capacities decreased in the order of C4H,
> OCH, > C1> CH, and NO,
when monolayer capacities were compared on a mass basis (mg/mg carbon), and in the order of C4H4 and OCH,
> CH, > C1> NG,
when monolayer capacities were compared on a molar basis (mole/g carbon). The two carbons, CAL and Filtrasorb 400, had been thought to be virtually identical [8, 91 and, therefore, the difference in the adsorption capacity progressions for the five substituted phenols by the two carbons was Surprising. Studies on the selection of a particular commercial activated carbon for a particular treatment problem have stressed the need for access to pore distribution and other data on adsorptive powers [4, 91. The solubilities of the various compounds in water appeared to have no influence over the REs achieved, as the order of decreasing solubilities in water shows. C1> CH,
> OCH, > NO, > C4H4
333 Some correlation between REs and molecular weights does appear to be shown however. Molecular weights decrease in the following order.
C,H,
> NO, > C1>
OCH,
> CH,
With the exception of '-naphthol. a polycyclic compound, this is a sinular progression to that of the RE values; the 'regenerating efficiency of boiling water decreased with decrease in molecular weight of the adsorbate being desorbed. Results for the monosubstituted benzene coni2ounds, excepting the relatively highly water-soluble phenol. showed a similar correlation [ 2 ] . It is possible that tlie smaller molecules, in penetrating further into the micropores of the carbon, are more difficult to remove. Comparison of the adsorption of the five substituted phenols and justification of the differences were very difficult [ 5 ] ; solubility, polarity and hydrogen bonding were all observed to be of significance but in an unpredictable way. It must be concluded that the desorption of these compounds is similarly difficult to compare and tiiat the differences are siniilarly difficult to justify. What can be stated with certainty is that the introduction of a second benzene ring (i.e. in effect the addition of C,H,), OCH,, C1. CI1, or a NO, group to phenol improves its adsorption and hinders its desorption. 4.3. Oxzdzsitzg Agcnrs. The exhausted carbon samples were regenerated with SO nd of the selected regenerants. The oxidising agents, potassium pernianganate, potassium dichromate and sodium hypochlorite, were used in concentrations sinular to those employed by Johnson et al. [ 101 and Beccari et al. [ 1 I ] . The treated carbon was then rinsed with three SO nd rinses of boiling distilled water. The results obtained are presented in Table 1. Tab. 1. Osidisinp Apcnts and thcir R E s RE ( 7 r ) for carbon exhausted \\it11 Rcgcnerant
2-naphthol
2-metlioxyplienol
2-chlorophrnol
o-creeol
2-nitrophenol
10% KMnO, 10% K , C r l O . 0.35% NaOCl
26.2 18.6 35.0
6.6 2.4 8.4
1.2 9.9 6.2
11.3 -7.5 -26.3
20.6 17.6 27.5
The REs achieved were low and in some cases negative; a negative RE indicated that the carbon not only failed to adsorb the organic from solution in the re-exhaustion phase, but furtliermore, desorption from the carbon took place giving a lugher equilibrium solute concentration at the end of tlie phase than at the beginning. Similar results (i.e. low REs and, in some cases, negative REs) were recorded for the mono-substituted benzene coinpounds [2]. In examining these results, it is necessary not only to consider the effects of the oxidising agents on tlie absorbates, but also their effects on the adsorbent itself. Treatment with oxidising solutions will increase the concentration of surface oxides on the surface of the carbon. These surface oxides inhibit adsorption. The formation of these surface oxides and their role in tlie inhibition of adsorption have been discussed in earlier studies by Martin and Ng [ 2 ] . In attempting to consider the chemistry of the re-
334 actions betweeen the reagents and the organic adsorbates, it is virtually impossible to allow for the contribution of the chemical role of the carbon surface with any degree of certainty. One of the most important chemical properties of phenols in general is their susceptibility to electrophilic substitution; this susceptibility to attack by electrophilic reagents such as oxidising agents is significantly influenced by the nature of the phenol [ 121. The presence of electron-releasing groups such as OCH,, CH, or another b e m n e ring should facilitate ease of attack whilst the presence of electron-viithdrawing groups such as NO, and C1 should have a deactivating influence. From knowledge of ionization constants (Appendix 1 lists the adsorbates used in this study, together with their relevant properties including pK, data; Appendix 2 lists the regenerants used, again with relevant properties) and from general organic chemistry principles of relative reactivities, it is likely that the order of susceptibility to attack by the oxidising agents may be represented by the following progression. OCH, and C4H, > CH,
> C1>
NO,
Table 1 clearly shows that the efficacy of desorption of the five organic compounds does not follow this progression. Other factors must therefore have an influence. It is possible that the greater penetration of the smaller molecules into the micropores of the carbon prevented ready access of, and subsequent reaction with, the oxidising agents. In general, the trend in Table 1 for decreasing RE values supports this possibility, and is as follows. C4H4> NO,
> OCH, > C1> CH,
This order is very similar to that for decreasing molecular weights. Thus, incomplete oxidation of the smaller adsorbates would have resulted; some adsorbate molecules would have remained within the carbon pores. The negative RE values for o-cresol would indicate the release of the retained adsorbate during the re-exhaustion phase. A second factor is the possibiiitv of sterjc hindrance arising 2s a consequence of the oxidation of the adsorbate by the oxidising agent; capillary condensation of oxidation products within the pores would mean a loss in available surface area for the organic compound of interest in the subsequent re-exhaustion phase if not effectively removed by rinsing with water at the end of the regeneration phase. These oxidation products could be organic or inorganic and could hinder the adsorption of the organic compound of interest by competitive adsorption [3, 71 or by physical sterichindrance [2]. It is of course possible that the products of reactions between regenerant and adsorbate could be more desorbable than the parent adsorbate; such a phenomenon has been observed in previous studies by the authors [2], although the poor RE values in Table 1 would indicate that this did not take place here. Of the three reagents, it is thought that oxidation by potassium permanganate would have been the most effective in attacking the substituted phenols and causing structural changes; the extent of these changes and the nature of the organic oxidation products are not easily predicted [ 131.
335 Appendix 1
- Characteristicsaf
adsorbates
Compound
M. W t .
Solubility in water at 20" C approx. (g/lOOg H,O)
2-Naphthol 2-Methoxyphenol 2Chlorophenol oCresol 2-Nitrophenol Phenol
144.19 124.15 128.56 108.15 139.11 94.11
0.068 1.85 2.80 2.50 0.20 9.30
Dipole moment 1.30-1.53 1.30-1.43 1.41 -1.54 3.10 1.55 1.7 3 ~
Ionization constant pKa 9.5 1 9.98 8.10 10.20 7.17 9.89
Footnote (a) Dipole moment data taken from Transactions of the Faraday Society, 30, (1934), Appendix. (b) Ionization constant data taken mainly from 'Handbook o f Chemistry and Physics', 53rd. edition (1972).
Appendix 2
- Characteristics
of organic regenerants
Compound
M.Wt.
B. Pt. ("C)
Formic acid Acetic acid Propionic acid n-Butyric acid iso-Butyric acid n-Valeric acid iso-Valeric acid ti-Hexanoic acid Dichloromethane Chloroform Carbon tetrachloride Met hano 1 Ethanol 1Propanol 2-Pro panol 1-But ano I 2-B u t an01 Acetone Benzene n-Prop ylaniinc Ethanolamine n-But ylamine Triethylaniine
46.03 60.05 74.08 88.11 88.11 102.13 102.13 116.16 84.93 119.38 153.82 32.04 46.07 60.10 60.10 74.12 74.12 58.08 78.11 59.11 61.08 73.14 101.19
100.7 1 17.9 141.0 163.5 153.2 186.1 176.7 205.0 40.0 61.7 76.5 65 .O 78.5 97.4 82.4 117.3 99.5 56.2 80.1 47.8 170.0 77.8 89.3
Solubility in water a t 20" C approx. (p/IOOgH.O) m
m m m
20.0 3.7 4.2 1.1 2.0 0.82 0.08 m m
m m
7.9 12.5 m
0.07 m m m
14.0
Footnote Solubility data and boiling point data taken from ti\ o main sources (a) 'Lange's Handbook of Chemistry', 12th edition (1979). (b) 'Handboakof Chemistry and Physics', 5 3 rd edition (1972).
336 The use of potassium permanganate invariably resulted in the formation of a brown precipitate of insoluble manganese dioxide; this inorganic reaction product could have blocked carbon pores, and in particular the micropores in which the adsorption predominantly takes place. Without exception, the five substituted phenols proved less amenable to desorption from carbon by the action of the three oxidising agents than did phenol [2]. It should be stated, however, that the far greater water-solubility of phenol makes any comparison somewhat artificial.
4.3. Car6oxyhc A C I ~ SThe . exhausted carbon samples were regenerated with 25 ml of the selected regenerants. The results obtained are presented in Table 2. The trend for decreasing RE values may be represented by the following progression. C4H, > NO,
> C1>
OCH,
> CH,
Tllis order corresponds to that for decreasing molecular weights of the adsorbates. Tab. 2. Carboxylic Acids and their REs ~~~~
-
~~
RE(%) for carbon exhausted with Regenerant
2-naphthol
2-methoxyphenoi
2-chlorophenol
67.7 92.7 84.1 79.9 73.2 70.1 75.6 56.1
32.9 30.5 27.5 17.4 13.2 16.2 12.6 -1.8
41.9 51.6 51.1 52.6 50.9 41.2 41.8 26.6
o-cresol
2-nitrophenol
~~
Formic acid Acetic acid Propionic acid n-Butyric acid iso-Butyric acid n-Valcric acid iso-Valeric acid n-Hexanoic acid
29.1 17.2 31.3 23.9 20.9 5.2 -3.1 -11.9
75.6 82.5 80.6 81.1 71.9 73.7 69.6 41.0
The studies on the mono-substituted benzene compounds [2] showed that the RE of an organic regenerant was, in general, dependent on its molecular weight; the higher the molecular weight was, tlie less effective it was as a regenerant. Furthermore, the relationship between the molecular weight of the adsorbate and that of the organic regenerant was observed to be of significance; it was found desirable to use a regenerant of molecular weight smaller than that of tile adsorbate so that physical displacement could be facilitat ed. The results for tlie straight-chain acids are presented in Figure 2. Beyond the first three acids in the homologous series, the REs of the acids generally decline with increase in molecular weight of acid. The molecular weights of the substituted phenols are, with the exception of o-cresol, larger than the molecular weight of the highest acid in the range of acids employed, and therefore the adsorbate/regenerant size relationship is not as sensitive a factor as it was in the case of the mono-substituted benzene compounds of molecular weights comparable with those of the middle and higher acids.
-
Chenucal reactions between the substituted phenols and the carboxylic acids were unlikely to have occurred.
337
.
- -
\
--- 2-Naph tho1 2-Nitrophenol
\
...........-.........-2-Chlorophenol ..- ..-.. -..- 2-Methoxyphenol Cresol
.-.-.o-
............u.....*%._
-..
-
>.
.-
a
e
a
-201 0
20
-
\ ' I
60
40
80
100
Molecular Weight of
120
Acid
Fig. 2. Carboxylic Acid M.Wt./RE Relationships.
4.4. Amines. The exhausted carbon samples were regenerated with 25 nil of the selected regenerants. The results obtained are presented in Table 3. The trend for decreasing RE values may be represented by the following progression.
C4H4> C1> NO, and OCH,
> CH,
This order approximates fairly closely to that for decreasing molecular weights of the adsorbates . Tab. 3. Amines and their REs RE(%) for carbon exhausted with Regenerant
2-naphthol
2-niethoxyphenol
2-chlorophenol
o-cresol
2-nitrophenol
nPropylaniine E thanolamine n B u tylamine Tr ie t hylaminc
83.1 63.4 12.7 54.1
44.4 33.9 36.3 28.1
54.3 55.6 53.1 45.1
43.6 5.3 51.9 -3.0
42.6 44.6 34.8 40.2
In general, the lower the molecular weight of the amine regenerant, the more effective it was as a regenerant. Ethanolamine and n-propylanine have very similar molecular weights (61 and 59 respectively), but the results show that etlianolamine, for three of the five adsorbates, was significantly less successful than n-propylanune. A similar observation was made in the studies on the mono-substituted benzene compounds [ 2 ] . It is thought that the presence of the OH group in etlianolanine exerted an adverse effect on the efficiency of ethanolanine as a regenerant.
338 The molecular weights of the regenerants were all smaller than those of the adsorbates and thus the adsorbate/regenerant size relationship is not likely to be of significance here. Chemical reactions between the substituted phenols and the amines were unlikely to have occurred. The success achieved by an organic regenerant is partly governed by its ease of removal at the end of the regeneration phase; carry-over of regenerant into the re-exhaustion phase reduces the adsorption capacity per unit weight of carbon for a susequent application of adsorbate. Rinsing with hot water was employed at the end of tile regeneration phase; for ease of removal, the regenerant should be readily water-soluble. As the watersolubility data in Appendix 2 show, the solubilities of the amines (including that of triethylamine) are sufficiently high to ensure ease of removal with aqueous rinses, and it is therefore thought unlikely that amine carry-over was of any significance. In the application of the carboxylic acids. it is possible that the higher acids (valeric and hexanoic acids) were carried over into the re-exhaustion phase because of their relatively low water-solubilities; acid carry-over could have reduced RE values for the adsorbates under study. 4.5. Hydrochloric Acid and Sodium Hvdroxide. The exhausted carbon samples were regenerated with 50 ml of the selected regenerants. The concentrations of the reagents are similar to those employed by Beccari et al. [ 111. The results obtained are presented in Table 4. Chemical reactions between the substituted phenols and HC1 were not likely. Tab. 4. HCI and NaOH and their REs RE(%) for carbon exhausted with Regenerant
2-naphthol
2-rnethoxyphenol
2-chlorophenol
o-cresol
2-nitrophenol
10%HCI 6M NaOH
35.0 66.1
22.2 26.3
28.4 36.4
-27.6 5.2
1.4 72.4
Experimental studies by Martin and Iwugo [ 141 have confirmed that for those organic compounds which possess acidic or basic properties in solution, adsorption is at its strongest in the pH region which yields the highest proportion of undissociated molecules. Thus, the acid conditions afforded by the application of hydrochloric acid would have encouraged the adsorption of the phenols, thereby discouraging subsequent desorption; the low RE values were therefore expected. Chemical reactions between the substituted phenols and NaOH undoubtedly took place; phenols are converted into their salts by aqueous hydroxide solutions because of the acidic nature of the phenols. The formation of the water-soluble sodium salts meant that desorption was facilitated. Sodium hydroxide has been shown to be commercially effective in the regeneration of GAC exhausted with phenol [15]. Later tests with other adsorbates showed that weaker solutions of sodium hydroxide ( 1 M and 3M solutions were evaluated) were more effective, sometimes considerably so, compared to the strong 6 M solution reported here. It is likely that adsorption of hydroxyl ions (arising from dissociation of the sodium hydroxide in aqueous solution) interfered with the adsorption
339
of the organic compounds [ 2 , 61; the use of weaker hydroxide solutions avoids excessive interference by hydroxyl ions thereby permitting more effective regeneration. Just as the superiority of NaOH as compared with HCl was expected for the particular adsorbates under study, so also was the order of decreasing RE values achieved with NaOH. This order is represented as follows. NO,
> C4H, > C1> OCH, > CH,
The stronger the substituted phenol acidity is, the more complete will be its conversion to the sodium salt of the parent phenol. The order of ascending pKa values is represented as follows (the smaller the numerical value of pKa, the stronger is the acid to which it refers). NO,
< C1< C4H4< OCH, < CH,
The similarity of the two progressions is readily apparent. The position of 2-naphthol in the progression of RE values was somewhat more elevated than had been expected. It is probable that the molecular weight and structure of 2-naphthol limited its penetration into the micropores, thereby making the adsorbate more accessible to reaction with the sodium hydroxide regenerant.
4.6. Chloromethanes. The exhausted carbon samples were regenerated with 25 ml of the selected regenerants. The results obtained are presented in Table 5. Tab. 5.Chloromethanes and their REs RE(%) for carbon exhausted with Regenerant
2-naphthol
2-methoxyphenol
2-chlorophenol
o-cresol
2-nitrophenol
Dichloromethane Chloroform Carbon tetrachloride
81.4 12.1
45.5 35.1 9.1
55.3 48.4 31.7
43.3 31.3 -4.5
19.9 19.4 34.3
39.3
In common with the carboxylic acids and amines the efficacy of the chloromethanes as regenerants appeared to be influenced by their molecular weights; increasing molecular weight of cliloromethane corresponded to a decrease in RE values. This influence of molecular weight of regenerant is clearly shown in Figure 3. Further examination of the results in Figure 3 reveals that the adsorbate/regenerant size relationship is a sensitive factor. The molecular weights of the adsorbates cover the range 108-144 which is entirely within 85 and 154, the molecular weights of dichloromethane and carbon tetrachloride respectively. For four of the five adsorbates, a sharp drop in RE occurs on reachmg a chloromethane regenerant with a molecular weight larger than that of the adsorbate. Because the molecular weight of o-cresol falls between those of dichloromethane and chloroform, a more significant difference was expected in the use of CH,Cl, and CHC1, to regenerate carbon exhausted with o-cresol. The trend for decreasing RE values my be represented by the following progression.
100
---- 2-Naphthol 2- Nitrophenol .......................... 2 - Chlorophenol .._.._.._.._.2 - Methoxyphem
00 R E (%)
t
-Cresol
.-.-.-o
60-
40. 20 -
N
3 N
I
V
0-
1
,
1
-20
100
80
120
140
160
Fig. 3. Chloromethane M.Wt./RE Relationships.
C,H, and NO,
> C1>
OCH,
> CH,
Again, the parallel between decreasing molecular weights and decreasing RE values may be noted. Chemical reactions between the adsorbates and the chloromethanes were unlikely to have occurred.
4.7. Acetone and Benzene. The exhausted carbon samples were regenerated with 25 1111 of the selected regenerants. The results obtained are presented in Table 6. Tab. 6. Acetone and Benzene and their REs ~~_______ ~
~~~
~
RE(%) for carbon exhausted uith Rcgenerant
2-naphthol
2-methoxyphenol
2-chlorophenol
o-cresol
2-nitrophenol
Acetone Benzene
85.4 13.2
44.2 22.7
55.7 41.8
42.9 1.5
81.1 55.3
The superiority of acetone as a regenerant compared with benzene was also observed in previous studies by the authors [ 2 ] .It could be argued that this superiority arises from the smaller molecular size and weight of acetone, thereby giving superior penetrating and adsorbate-displacing powers. As the molecular weight of benzene is well below those of the adsorbates, it is thought unlikely that this factor is significant. The difference in water-solubilities of the two regenerants is more likely to be the reason for the difference in results achieved with acetone and benzene. Insufficient removal of benzene when the
341
carbon was rinsed with hot water at tlie end of the regeneration phase is thought to have been responsible for tlie lower results, the water-solubility of acetone is such that the acetone would have been very efficiently removed at the end of the regeneration phase. The residual benzene would have adversely affected the adsorption of the organic compound under study in the re-exhaustion phase; the presence of benzene was detected by gas-liquid chromatography when the re-exhaustion solution was analysed for the equilibrium concentration of the residual solute. Whilst there are numerical differences between the RE values obtained with acetone and benzene for each adsorbate, it is significant that the order of decreasing RE values is identical for each solvent; this order is as follows. C4H, > NO,
> C1>
OCH,
> CH,
Once again, the parallel between decreasing molecular weights and decreasing RE values is exhibited. Chemical reactions between the adsorbates and the two regenerants were unlikely to have occurred.
4.8. Alcohols. The exhausted carbon samples were regenerated with 25 nd of the selected regenerants. The results are presented in Table 7. The trend for decreasing RE values may be represented by the following progression. C4H4> NO,
> C1> OCH, > CH,
The RE values achieved by the alcohols showed no obvious correlation with their molecular weights. Tab. 7. Alcohols and their REs
RE(%) for carbon exhausted nit11 Regenerant
2-naphthol
2-nicthoxyphenol
2-chlorophenol
o-cresol
2-nitrophenol
Methanol Ethanol 1-Propanol 2-Propanol 1-Butanol 2-Butallol
79.1 74.4 11.5 67.5 64.5 63.4
31.4 36.3 42.1 30.4 31.4 33.9
50.6 52.8
25.4 27.6 34.6 10.4 20.1 11.9
61.1 66.8 65.4 63.6 68.2 63.1
-
53.1 56.2 66.0
The results show that for the isomers of the higher alcohols, the straight-chain isomer was generally more effective than the branched-chain isomer. The results for the isomers of the higher carboxylic acids also show the superiority, in general, of the straight-chain isomer. It is possible that the more linear isomer would more easily gain access into the micropores of the carbon and dislodge the adsorbate molecules held therein. Previous regeneration studies by Martin and Ng [2] showed that the same range of alcohols achieved very high RE values with carbon samples exhausted with phenol: RE values between 90 and 100% were recorded.
342 Chemical reactions between the adsorbates and the alcohols were unlikely to have
occuned.
5. SUMMARY
The results of the studies may be summarized as follows. 1. Organic chemical regenerants with solubilising powers were found to be generally much more effective than inorganic chemical regenerants with oxidising powers. 2. The success of inorganic regenerants used to induce high or low pH conditions was found to be dependent on the type and properties of the adsorbate; the pH effect on the dissociation of the adsorbate together with the possibility of chemical reaction between adsorbate and regenerant were of importance. 3. In general, the efficacy of the organic regenerants within a group decreased as their molecular weights increased; the smaller the regenerant, the further it could penetrate into the micropores of the carbon and displace the adsorbate. 4. For the adsorbates studied, there was a marked correlation between decreasing molecular weight of adsorbate and decreasing RE value; the smaller the adsorbate, the further it could penetrate into the micropores of the carbon thereby resisting displacement by the regenerant.
REFERENCES 1 F. J. Guymont, in M. J . McCuire and I. H. Suffet (Eds.), Activated Carbon Adsorption oforganics frDm the Aqueous Phase, Vol. 11, Ann. Arbor Science, Michigan, 1980, Chapter 23, 531-538. 2 R. J. Martin and W. J. Ng, in Proc. 1st. Int. Conf. Industrial Pollution and Control, Vol. I, Singapore, December 15-17, 1982, pp. 39. 3 R. J. Martin and K. S. Al-Bahrani, Wat. Res., 11, 1977, 991-999. 4 R. J. Martin and K . 0. Iwugo, Publ. Hlth. Engr., 7, 1979, 176-182. 5 K . S. Al-Bahrani and R. J. Martin, Wat. Res., 10, 1976, 731-736. 6 R. J . Martin and K . S. Al-Bahrani, Wat. Res., 12, 1978, 879-888. 7 R. J. Martin and K . S. Al-Bahrani, Wat. Res., 13, 1979, 1301 -1 304. 8 G. Wallcr, personal communication, 1979. 9 R. J. Martin, Ind. Eng. Chem. Prod. Res. Dev., 19, 1980, 435-441. 10 R. L. Johnson, F. J. Lowes, R. M. Smith and T. J. Powers, United States Public Health Service Publication No. 999-WP-13, 1964. 11 M.Beccari, A. E. Paoliniand G. Variali, Effl. Wat. Treat. J., 17, 1977, 287-294. 12 R. T . Morrison and R. N. Boyd, Organic Chemistry, Allyn and Bacon, Boston, 1959. 13 J. Burdon, pcrsonal communication, 1983. 14 R. J . Martin and K. 0. hvugo, Wat. Res., 16, 1982, 73-82. 15 K . J. Himmelstein, R. D. Fox and T. H. Winter, Chem. Engng. Prog., 69, 1973, No. 11, 65-69.
343
ON UNIT PROCESSES DURING INFILTRATION
A. L. KOWAL
Institute of Environment Protection Engineering, Technical University of Wroclaw 50-370 Wroclaw. Poland
ABSTRACT Ground water infiltration is cmployed in municipal water supply and also in industrial when high quality water or constant temperature are needed. Phenomena taking place during infiltration fall into the following groups of unit processes: biochemical, chemical, physicochemical and physical. All of these unit processes are discussed in detail, and their contribution to water quality changes during infiltration is shown.
1. INTRODUCTION
Infiltration is usually defined as a water treatment process which has some additional advantages over other treatment methods. They are the following: alimentation of ground water, equalization of chemical composition, possibility o f water storage in the ground, make-up o f ground water resources and equalization of water temperature. Of these factors, the equalization of water temperature is unique, since it cannot be achieved in any other technology. The possibility of water storage in the ground is likewise advantageous, because this water may be used for many purposes, especially when high pollution levels d o not permit intake of surfaces waters. In addition to the advantages mentioned, of equal importance from the technological point of view is that each o f the unit processes involved in infiltration contributes t o the improvement o f water quality. Changes in quality take place both in the infiltration basin during percolation from the basin bottom t o the ground water level, and in the ground during a horizontal flow t o the infiltration water intake. The phenomena taking place during infiltration fall into the following groups of unit processes: biochemical, chemical, physicochemical and physical (Fig. 1).
2. PROCESSES OCCURING IN INFILTRATION BASINS
Biochemical processes involved in infiltration are the most intensive in the infiltration
344 UNIT
PROCESSES
4&----infiltration boSm bottom
sludge
L ground
water
7
Fig. 1. Unit processes involved in water infiltration.
basins, in the bottom sludge layer and in the aerobic layer situated in the ground immediately beneath the basin. This intensity tends t o decrease in the anaerobic layers of the ground, and the decrease becomes more pronounced during flow through the ground, which is due t o the loss o f substrate, and the decreasing number o f microorganisms. The intensity of the process varies with water temperature, following seasonal patterns, this is particularly evident in infiltration basins as well as in the predominantly aerobic layer situated immediately beneath the basin and extending t o a depth of at most 3.0 m. The intensity of bioche~nicalprocesses decreases with depth to approach zero in the lowest parts of' the ground. Biochemical phenomena occurring in infiltration basins are quite the same as those observed in shallow ponds. The presence o f plankton is o f utmost importance there. A shallow basin having a depth less than 3.0 m is exposed to solar radiation from top to bottom. As the water usually contains sufficient amounts o f nutrients, plankton blooming may often be very intensive. Algae blooming frequently leads to oxygen supersaturation of the water (up t o 225 percent), which has a favourable effect on the depth of the aerobic layer in the ground. Excess oxygen and elevated water temperature bring about an acceleration of biological processes in the pond, in the bottom sludge and in the ground layer beneath the bottom of the pond. Algae blooming is responsible for the depletion of carbon dioxide followed by decomposition o f bicarbonates and precipitation of calciuin carbonate. In infiltration basins pH may rise even to a level pf 10.0. Calcium carbonate content in bottom sludge is markedly higher in summer than in winter. Under these conditions, also heavy metals are precipitated to the bottom sludge in the form of
I
345 sparingly soluble hydroxides. In the winter months, both calcium carbonates and heavy metal hydroxides may dissolve and penetrate into the deeper layers of the ground [ 1 , 21. In addition to natural suspended solids and calcium carbonates, detritus is an essential component of the bottom sediments. The degradation of organics present there, leads to an evident increase of permanganate COD in the water infiltering into the subbottom layer. Although in this situation the harvesting of algal populations seems to be best suited, propagation of fish in order to maintain the biological balance in the basin has the advantage of greater ease in producing the effect required. This is because plankton consumption by fish approaches 30 kilograms for each 1.0 kilogram increment in weight. Moreover, plankton organizms both live and dead have a high ability of adsorbing radionuclides, heavy metals and biocides [3, 4, 51. Fish feeding with plankton will therefore be a successive step in the process of making toxic substances more concentrated. Another process of significant importance in the standing water environment is that of photodecomposition, and one of the principal factors affecting the length of its course is light penetration. Thus, in summer, photodecomposition of nitrosoamines may be completed in a period of several hours or, at most, of several days. In winter, the presence of ice cover will limit light penetration, extending the decomposition process even to several months [ 6 ] .Except for fluorantene, most of the PAHC compounds are subject to photodecomposition. The standing water basins receive inflow of surface waters carrying suspended matter. On the other hand, the basin produces its own suspensions of dead plankton, which settle in the bottom part. Hence, maintaining biological equilibrium in a standing water body becomes a problem of great significance. Suspended matter present in the pond is subject to complete sedimentation. The incoming suspended solids are retained by sedimentation or filtration. To prevent silting-up of the bottom, it is necessary to remove the inflowing suspended matter by pre-sedimentation or filtration, silting-up by dead plankton may be avoided by fish propagation in the basin. In some instances, this pre-treatment procedure does not differ from routine practices employed in the treatment of surface waters. But the choice of the method depends not so much on the silting-up problem, as on the desired quality of infiltration water. Thus, for infiltration purposes it is sufficient to use waters of second-class purity, whereas a conventional coagulation-filtration system requires first-class purity water. Pretreatment is especially recommended when the infiltration water has t o be transported through a pipe system over a long distance from the point of intake to the infiltration area. But in that particular case, of importance is the problem of pipeline protection. The time of water retention varies with time of operation. After completion of the cleaning procedure (removal of bottom sediments) the incoming water percolates through the bottom of the pond and forms a pool at the inflow, until the silting-up the bottom reappears. Then, as the silting-up increases, so does the water table of the pond to reach its maximum value at a given period of time. The depth of the water varies from 1.5 to 3.0 m. With this depth and with an infiltration rate ranging from 0.1 to 0.5 m/day. the residence time may vary from several minutes at the onset of operation up to several tens of days. I t follows that the effect of the processes occuring in the pond on water quality will depend on the time of residence. This points to the necessity of successive cleaning and operating at least three basins. Such a system enables an optimum utilization of the processes that occur in the basin.
346 Infiltration basins can be considered almost ideal for equalizing the chemical composition of the water as it percolates over the whole surface area of the bottom. The equalization rate increases with the increasing time of water flow in the basin. Assuming that the water composition does not change during infiltration, the quality of infiltrated water may be evaluated on the basis of equalization in the ground and mixing with ground water. The process of equalization can be described by mathematical models. An extended residence time may be disadvantageous, because in the summer season water in a shallow pond warms more rapidly than in deep basins. 3. PROCESSES OCCURRING IN THE GROUND AND IN THE BOTTOM SEDIMENTS
Percolating the pond bottom, water leaves the solid matter conrained in it to form a layer of bottom sediments before entering the ground. Although the formation of bottom sediments is considered disadvantageous from many aspects [7], their sorbing capacity may counteract some of the limitations. As the bottom sediment layer and the top soil provide aerobic conditions, the organic substances present, stimulate growth of proteolytic and nitrifying bacteria [8]. In the ground layers where oxygen absorbed by the water is sufficient to facilitate this process, nitrification is initiated. Thus, while ammonia nitrogen and oxygen are depleted, the concentrations of nitrates and carbon dioxide continue to increase. COD behaves in the following manner: increases in the ground layer adjacent to the bottom sediments and then decreases rapidly. Biochemical phenomena occuring in ground layers, where aerobic processes are predominant, resemble those taking place in slow-sand filters. With ground stratification there is a gradual decrease of dissolved oxygen, which create favourable conditions for denitrification and brings about a considerable drop in the concentration of total nitrogen. Reports on the investigation of the Wroclaw water supply system indicate [9, 101 that total nitrogen measured after infiltration varies from 0.4 to 0.7 mg N/dm3 which is the average from many years’ measurements, and there exists no relationship between these values and those measured prior to infiltration when total nitrogen ranged from 0.7 to 6.2 mg N/dm3. As shown by these data, total nitrogen concentration has decreased substantially. Thls decrease should be attributed to the interaction of a number of unit processes, like denitrification, sorption, ion exchange, and dilution with ground waters which exhibit low concentrations of nitrogen compounds. This was not so in the case of slow-sand filtration, yielding a good balance of nitrogen compounds (without any loss whatsoever), which persist in the effluent of the filter bed in the form of nitrates. The thickness of the serobic ground layer depends on the content of dissolved oxygen, on the level of water pollution (permanganate COD, dichromate COD, BODS, TOC), as well as on the infiltration rate, and may vary from several tens of centimeters to several meters. The intensity of biochemical processes in the ground is temperature-dependent, yet the differences are in the main negligible, so the factor of flow duration is claimed to be of greater importance. The survival rate for bacteria and viruses in the ground is also a function of temperature, and increases when water temperature decreases. Penetration of microorganisms to the intakes of infiltration water depends primarily on the filtrability of soil and ground. In general, this filtering capacity is sufficient to provide bacteria free water meeting sanitary regulations. Pathogenic organisms are retained and they die out.
347 The distance over which microorganisms are transported in the soil is influenced by the variation in the table of ground waters. These water table variations are also responsible for the change of redox potential in the soil layer in which they occur. Thus, favourable conditions are created for iron and manganese removals. Investigations are reported on the intensive growth of fungi which has been observed in summer under the floor of an infiltration basin in Poznan [ 111. This was due to the accumulation of organics during winter when biochemical processes were inhibited by decreased temperature. In unsaturated subbottom layers biochemical processes develop more intensively than they do in a saturated layer. However, the infiltration rate is higher in a saturated layer. Adsorption on bottom sediments and in the ground layer has an advantageous influence on water quality. Soils consisting of dusty sand and clays are characterized by good sorption and ionexchange capacity but, at the same time, they are hardly permeable, thus decreasing the infiltration rate. Bottom sediments in an infiltration basin have all of the properties typical of a good sorbent. Determinations for bottom sediment samples show enrichment with heavy metals, organics and refractants [7]. The layer itself is in the most part only several centimeters thick, and its organic zone creates advantageous conditions for biodegradation, which initiates occurrence of desorption phenomena. It is therefore advisable to remove the bottom sedinients as frequently as possible in order to prevent wash-down of the adsorbate. A physical factor indicating the efficiency of the infiltration process is the ability to retain micropopullutants. The adsorption of hydrophobic compounds in soil is stronger that that of hydrophylic substances, which remain in the solution. Very strong hydrophobes are, e.g., DDT and PCB. Since chlororganic compounds are difficult to remove by infiltration [12], it does not seem advisable to employ chlorination as a prior step. Chlororganic solvents are also amongst chemical substances characterized by a high stability [ 131, they are resistant to biodegradation and sorption in the ground. On the contrary, other manmade organics entering the soil along with the infiltration water are easily sorbed and biodegraded. In order to determine which of the two process is predominant it is necessary either to perform experiments with a given sorption-ion exchange complex or to assess the resistance to sorption and biodegradation. Heavy metals can be removed by precipitation (in the form of sparingly soluble hydroxides), by ion-exchange with the minerals occurring in the ground, and by sorption [14, 151. The principal factors contributing to the predominance of one the processes over the remaining two are the following: redox potential of the complex, composition of the minerals present in the soil [16, 17, 181, and composition of the infiltration water. It may also be expected that ammonium ions present in the soil will be subject to sorption and that they can be exchanged with calcium ions contained in the minerals. Hydrolysis of minerals brings about an insignificant yet noticeable increase in the hardness of the infiltration water. However, under certain conditions, when ferric sulphides occurring in the soil are oxidized to ferric sulphates with evolvement of sulphuric acid, thus contributing to a decrease of pH, water hardness may increase rapidly together with the concentrations of sulphates and iron compounds. For these reasons, I recommend chemical analyses of soil and bore-hole samples (especially determinations of the forms in which iron
348
-16
-12
-8
-4
0
8
4
temperature,
12
16
20
i
OC
Fig. 2. Plots of air and water temperatures measured prior to, and after, infiltration in the area of Legnica Water Works in 1975; (1) air temperature, (2) water temperature in the basin, (3) water temperature in the well.
0
C
water
8
temp
12
20
16
in the well,
OC
Fig. 3. Relationship between water temperature in the well (after infiltration) and in the basin (before infiltration) for 1975.
349 and manganese may occur), and estimates of the resistance of the two compounds t o oxidation and dissolution. The equalization o f water composition, which takes place in the soil. is interfered with by the mixing of infiltering and ground waters. When the difference of chloride concentrations between the two types of water is significant, they will be determined analytically at the intake. Equalization of water temperature is a unique phenomenon. which cannot be achieved in any other treatment technology. In the area of infiltration water intakes for the city of Legnica. air temperature, water temperature in the infiltration pond and water temperature in the principal well were measured after infiltration during a period o f two years. Measurements were performed every day at 7.00 a.m. The results obtained in 197.5 are plotted in Fig. 2 (frequency distribution o f water temperature). As shown by this figure. the median temperature o f infiltration water in the principal well amounts t o 7.75"C and is slightly higher than that in the basin. Figure 3 gives the relationship between water temperature in the pond and in the well. The lowest temperatures measured in the basin and in the well are 1°C and S°C, respectively, whereas the highest temperatures measured there are 2 1OC and 15" C, respectively. As shown by these data, the degree of equalization is significant. Similar values of minimum and maximum temperatures o f infiltration water were measured in the area of the Poznan Water Works in 1969 [8]. Great differences between actual and median temperatures indicate that the percentage of surface water in the ground water nlixture is very high. 111the area o f the Legnica infiltration intakes, even though the percentage o f surface water nught have been high, no unfavourable changes were found in the physicochemical and bacteriological composition o f the water at the point o f intake. 4. CONCLUSION
1. Since infiltration has a nuniber of inherent advantages, it is one of the treatment methods which should receive serious attention in engineering practice, the more so as the process may employ second-class purity water. 2 . A rapid short-term decrease in the quality of the infiltering water, which can soiiietimes be observed, would be more than outweighed by the advantage of equalizing the composition o f the water, and by the influence of the unit processes involved in infiltration. 3. Periodical increase o f water pollution has no dramatical influence on the infiltration intake, because the inflow to the infiltration basins may be stopped. Water stored in the basins and in the ground sufficiently supplies the wells in the emergency period. 4. Infiltration is a unique method of equalizing the temperature of the delivered water. which has not been achieved so far in any other treatment technology. 5. lnfiltration consists of a number o f unit processes. Their effect should be considered for each of them separately, and re-examined with each change of conditions. 6. When the infiltration process involves highly polluted water, the soil acts as an overloaded sorption bed with sorption front moving toward the wells. In the summer months fungi can grow in the sub-bottom ground layer and decompose the excess o f adsorbed organic matter.
350 REFERENCES
1 A. L. Kowal, Composition of bottom sludge in infiltration basins, Technical University of Wroclaw (in preparation). 2 M. Blaiejewski, Environment Protection Engineering 6 (1980) pp. 362-382. 3 C. E. Cushing, Limnology and Oceanography 12, 1967. 4 U. Bauer, Uber das Verhalten von Bioziden by der Wasseraufbereitung unter Desondere Beriicksichtigung der Langsamsandfiltration. Veroff. des Inst. fur Wasserforschung GmbH der Dortmunder Stadtwerke A.G. Nr 19, 1972. 5 L. J. Goldwater, Scientific American, 224, 1971. 6 G. Dure, L. Weil, K.B. Quentin, Vom Wasser, 1978. 7 Micropollutants in river sediments. WHO Copenhagen 1982. 8 S. T. Kolaczkowski, S. Spandowska, Osterreichische Wasserwirtschaft 24 (1972) pp. 238-246. 9 A. L. Kowal, A. Serwach, Effectiveness of infiltration as compared to slow sand filtration. Technical University of Wroclaw (in preparation). 10 A. L. Kowal, Proc. of Int. Symposium Ground Water Recharge, May 14-18, 1979, Dortmund, pp. 71 -72. 11 M. BlGejewski, Sztuczna infiltracja w uzdatnianiu w6d powierzchniowych. Research Institute for Environmental Development (RIED), Warszawa 1982. 12 G. J . Piet, Zoeteman B.C.J. JAWWA July 1980. 13 W. Kohn, JAWWA 70 (1980), 326-331. 14 T. Maruyama, S. Hannah, J. M. Cohen, J.W.P.C.F. 47 (1975), 962-975. 15 U. Forstner, G . Muller, Schwermetalle in Fliissen und Seen. Berlin - Heidelberg - New York 1974. 16 M. Schmitzer, S. J . M. Shinner, Soil Science, vol. 102/6, p. 361. 17 M. Schmitzer, S. J. M. Shinner, Soil Science, vol. 103/4, p. 247. 18 E. Jargenson, Water Research 1975/2, p. 163.
CHAPTER V
PH YSICO-CHEMICALTREA TMENT:ION EXCHANGE
This Page Intentionally Left Blank
353
NEW FIBROUS ION EXCHANGERS FOR PURIFICATION OF LIQUIDS AND GASES
V. S. SOLDATOV
Institute of Physico-Organic Chemistry of the BSSR Academy of Sciences, 220603 Minsk,USSR
ABSTRACT The application of ion exchange fibres opens new possibilities in technology of purification of air and water. Their main advantages are: a high rate of an ion exchange process (caused by small diameters of monofilaments ranging from 3 t o 25 mcm), low and easily controlled resistance to the flow of liquids and gases, the possibility o f using in a variety of physical forms, such as column filters, conveyer belts, thin layers etc. The present communication is a review of studies performed in the author's laboratory on the properties and possible application of some new fibrous ion exchangers for the environment protection. They are strong acid, strong and medium base exchangers based on different industrial fibrous matcrials. The characteristics of their ion exchange properties, osmotic, thermal and chemical stability, mechanical strength and structure have been studied. It has been shoh n that in niany processes concerned with the environment protection the fibrous exchangers have great advantages compared to traditional ion exchange resins.
1. INTRODUCTION
The fibrous ion exchangers and sorbents can be efficiently used for purification o f liquids and gases and are of great importance for environment protection. During recent years they draw attention of researchers and practicians. A number of studies in this field continuously increases. The recent monographs [ 1 --31 give a clear representation of the modern state in this branch of science. The fibrous ion exchangers can be produced in tlie form of filaments, cloths o r nonwoven materials, which opens many possibilities for tlie ion exchange technology. Alondside with the traditional column processes, it is possible t o use ion-exchange fibers in the form of continuous conveqer belts, devices combining tlie functions of mechanical and ion-exchange filtefs, thin flat layers etc. A small diameter of the monofilaments (3-25 mcm for tlie commercial fibers) provides a high rate of the ion exchange and high efficiency of tlie ion exchange apparatuses. The commonly used ion exchangers with granules of tlie same size have a diqensity corresyondin;: to 700--300 mesh. The application of such materials could provide very good parameters for the ion exchange processes. how ever. their usage is liardly possible in large-scale processes due to high resistance of the ion exchange filters t o the flows of
354 liquids and gases. This difficulty is eliminated when the ion exchange fibers are used since the resistance of a layer is easily controlled by density of the fiber packing in a filter and may be predetermined according to the engineering requirements. We consider the fibrous ion exchangers as having valuable properties specific to microspherical ion exchangers but deprived of their shortcomings. The fibrous ion exchangers are very promising in the environment protection where it is necessary to treat large volumes of water or air containing impurities in low concentrations. In spite of the valuable properties mentioned above, up till now the fibrous ion exchangers have not found a wide application in the ion exchange engineering and are not produced commercially in quantities. There are some reasons for that: - the assortment of fibrous ion exchangers is very scarce; in particular, there are no good methods for production of the most widespread ion exchangers, strong-base and strong acid ones; - exchange capacity of fibrous ion exchangers is substantially smaller compared to the granulous ones; mechanical strength. osmotic and chemical stability of ion-exchange fibers are often not high enough to realize their potential advantages; - technology of synthesis of ion-exchange fibers and their application are not developed enough and their physico-chemical, selective and other properties are insufficiently studied. During the recent five years in the Laboratory of Ion Exchange and Sorption, the intensive studies have been carried out aimed at the development of methods for preparation of fibrous ion exchangers of various types and the search for the fields of their efficient application. As a result of these studies, rather simple methods have been developed for preparation of the fibrous sulphonic-type cation exchangers (the analogues of Dowex-50). strong base anion exchangers (the analogues of Dowex-1), weak base and medium base anion exchangers, and weak acid cation exchangers with complexing properties. In the present paper the following ion exchange materials will be discussed: the strong base (conventionally named SBF), the medium base (MBF), and the strong acid (SAF) fibers. The SAF and the SBF exchangers are based on graft copolymers of polystyrene and polypropylene. The MBF exchangers have been synthesized by chemical modification of acrylic fibers. The main part of the present work has been carried out on the ionexchange fibers with a monofilament thickness from 20 to 40 mcm. All these materials have been used in various textile forms depending on a concrete problem to be solved. Filaments, cloths, non-woven materials or staple have been used in a variety of experiments. ~
2. PROPERTIES OF THE FIBROUS ION EXCHANGERS
The most important properties of the developed ion-exchange fibers are given in Table l*. Below these properties will be discussed in more detail.
* The SBF exchangers have been synthesized and investigated by A. Shunkevich and 0. Popova: the SAT: - by A. Shunkevich and A. Pokrovskaya: the MBT: - by G . Serseev and A. Shunkevich. The data in this paper are published \I ith the pcrniission of the authors.
355 Tab. 1 . Properties of fibrous ion exchangers Ion Functional exchanger group
SA F SB F MBF(1) MBF(I1)
-so ; -N(CH,): -NH; =NH; ENH
Exchange optimum mg-cqv/g
3.5 3.2 9
5 -7
Capacity Swelling maximum g/g opt.
4.6 4.4 11 11
0.8 0.8 1 .o 1.5
Tensile strength kg/mm’ opt. 9
1s 18 15
Elastisity modulus kg/mm’
Elongation at rupture %
200 2 00 200 600
18 45 40 28
+
Some examples of the potentiometric titration curves are given in Fig. 1. A strong base and a strong acidic nature of the SBF and the SAF as well as polyfunctionality of the MBF are clearly seen. The exchange capacity and the swelling of the SAF and the SBF can be varied in a wide range by changing the amount of graft polystyrene which can reach 600% relative to the mass of the polypropylene fiber. These parameters can be also controlled by other conditions of the synthesis. A degree of sulphonation is usually about 1 sulphonic group per phenyl ring. At the same time more than one (up to 1.3) trimethyl ammonium group per phenyl ring can be incorporated in the fibrous graft copolymers. Therefore, the exchange capacity of the strong base fibers are often higher than that of the commercial granulous resins of the same chemical type. No special means have been used to prepare the ion exchange fibers of an equal thickness. Only commercially available fibers have been used as the initial materials. After chemical modification the scatter of the monofilament diameters was observed to increase compared to that of the initial material. Nevertheless, the scatter is much smaller than in the case of fractionated granulous resins generally used in industry. (See Table 2 ) .
Fig. 1. Potentiometric titration curves for the fibrous ion exchangers in the 1 N solutions of KC!. pH is a function of a quantity of milligram equivalents of HCl(K0H) per gram of the H+(OH-) exchanger: 1 - SAF; 2 - SBF; 3 - MBF(I1).
Tab. 2. I f f e c t of a quantity of the grafted polystyrene on the propertics of ion exchatigc fibcrs
Ion exchatigcr E xch a npc
Graft cd co po lymcr
Ion exchanger SAT:
SBT:
_______~______
AP*
D iamct er mcm
Tensile st rcng t h kg/nirn'
Elongat ion %'
capacity mg-eqv/g
g/g
Diarnctcr mcni
0 104 171 263 104 350
22 f 1 34 + 1 34 f 3 42 + 2 34 + 1 44k3
50.8 23.1 20.1 18.5 23.1 15.0
36 34 60 37 34 31
3.25 3.38 4.10 3.13 3.58
1.2 3.2 0.9 0.4 0.6
49+3 50 + 4 54 i s 41 + I 50 * ?
Swelling
Tensile strcngth kg/mrn'
9.6 5.7 7.2 7.6 10.4
Breaking strcss kg
18.1 11.2 10.5 I 0.0 20.4
Elonpa t io t i %
18 10 20 ~
22
357
Fig. 2. Microphotograph of the SBFC1- sample with 173% of grafted polystyrene A magnification of 900 power.
0
b
0.2
0.4
-
0.6
0.0
-
1.0
-
" +
Fig, 3. Selectivity coefficients for K f - H' exchange on thc SAF exchangers (1-3) and the KRC-10 granulous resin (4). Water sorption, g H , O per g H*: 1 - 2.3; 2 - 1.2; 3 - 0.7;4 - 1.0.
The tensile strength of the ion exchange fibers is high enough to allow their textile reprocessing. Neither destruction of the fibers nor a loss in the exchange capacity has been observed after repeated bending or drying and changing in their swelling. Table 2 also illustrates mechanical properties of the air-dry fibers on the base of polypropylene in the H' and C1- forms. It is seen that the specific tensile strength and elastisity of the fibers decrease with increasing exchange capacity. It is to be noted that a decrease in the specific tensile strength is mainly due to an increasing thickness of filaments after the grafting of polystyrene and its further chemical modification. The breaking stress of the fibers does not substantially vary with increasing capacity. It shows that mechanical
358 8
I
.
0
0
. 42
.
-
Q4
.
.
a6
.
_
_
0.8
-X
. t.0
No;
Fig. 4. Selectivity coefficients for NO; - C1-exchange on the SBF exchangers (1-4), Dowex-1x4 (5). Water sorption, g H,O per g Cl-: 1 - 1.09; 2 - 0.80; 3 - 0.36; 4 - 0.26; 5 - 1.08.
4ol 'G
I
/
Fig. 5. Sorption of methylene blue as a function of time by the SAF (1) and KPC-4 resin (2). Water sorption: SAF - 2.3 g H,O/g H+;KPC4 - 1.99 g H,O/g H+.
strength of the ion exchange fibers of this kind is provided by the polypropylene skeleton. The structure of the ion exchange fibers based on polypropylene strongly depends on a preparation procedure of the initial graft copolymers. It can be more or less dense providing different mechanical and osmotic stability of the fibrous exchangers. All the exchangers described in the present paper are macroscopically homogenous materials. The grafted polystyrene is regularly distributed in the volume' of the polypropylene fibers. Nevertheless, the observation in an optical microscope indicated a stripped structure with a period of about 5-6 mcm. A typical example of such a structure can be seen in Fig. 2. The nature of this phenomenon is under investigation. The X-ray diffraction analysis showed that the graft copolymers have a domain structure, the linear dimention
359
Fig. 6. Chromatohraphic separation on the fibrous exchangers against granulous resins. The experimental conditions: bed height, 16 cm; diameter, 0.8 cm; flow rate, 2.15 ml/min. a) K' - Cs separation. Eluent - 0.4 N HCl. 1. SAF. Capacity, 3.25 mg-qv/g H'; water sorption, 1.2 g H,O/g; filament diameter, 49 3 mcm, a quantity of the exchanger in the column, 3 g. 2. SAF' Capacity, 2.02 mgeqv/g H '; water sorption, 1.2 g H,O/g; filament diameter, 3 1 mcm; a quantity of the exchanger in the column, 2.3 g. 3. Dowex-50x2, 100-200 mesh (a partice diameter is 75-150 mcm), a quantity of theresin in the column, 1.5 g; b) C1- - Br- separation. Eluent - 0.4 N KNO, for Dowex-1x2 and 0.6 N KNO, for the SBF. 1. SBF. Capacity, 3.13 mg-eqv/g C1-; water sorption, 0.42 g H,)/g; filament diameter, 41 1 mcm; a quantity of the exchanger in the column, 3.1 g. 2. Dowex-1x2, 100-200 mesh (apartlole diameter is 75-150 mcm)
*
*
*
.
360
.-
r, hours
Fig. 7. Breaktgrough curves for SO, adsorbed from the air by ion exchangers. The experimental conditions: flow rate, 7 m/min; Co = 200 mg/m3 ; the bed height, 8-9 mm; relative humidity, 80%. Thc dashed line shows MPC (10 m&/m3)for SO, in tllc air. 1 - VION KH-I, the commercial fibrous ion , commercial n c a k base anion exchanger; 3 - SBF; 4 - AV-17x8, exchanger; 2 - A H - 5 1 1 ~ 8 / 1 0 0 the the granulous strong base anion exchanger; 5 - MBF(I1); 6 - MBF(1).
,
hours
Fig. 8. Breakthrough curves for H, S absorbed from the air. T h c experimental conditions: flow rate, 3.6 m/min; H , S concentration, (C,) - 1000 mg/m3; relative air humidity, 85%; the bed height, 10 cm. The dashed line indicates MPC of H I S in the air. 1 - MBF(I1); 2 - AV-17: 3 - SBF.
of the domains being in the range of 50-3-00 8. The propylene matrix preserves its crystalline structure. The strong acid and strong base ion exchange fibers have such an inert base as propylene, therefore their chemical stability under common temperature conditions towards strong acids, bases and oxidants is similar to that of the ordinary styrene-divinylbenzene resins. Some difference becomes evident at high temperatures when the polystyrene
36 1 Tab. 3. Capacity of ion exchangers to SO2 and H: S. The experimental conditions are given in the captions to Fig. 7 and 8 Moisture capacity %
Capacity to 0, , mg/g
Ion exchanger
Total exchanger capacity mg-eqv/g
a;
a;*
AB-17x80 H B OHXH-I MBF MBF (amine) MBF (amine)
4.13 6.00 4.31 9.6 9.02
60.2 65 .O 60.0 51.7 51.6
125 62 135 115 157
218 144 206 294 278
Capacity to H , , mg/g a;
a r
50
81
-
-
103
141 28
-
4
matrix starts melting and destructing (> 170'C). Shunkevich and Prokopchuk have performed a careful investigation o f therniostability and thermomechanical properties o f fibers of this type and came t o a conclusion that they can be used in the same temperature range and under the same conditions as the granulous styrene divinylbenzene resins. The ion exchange selectivity o f the fibrous exchangers towards inorganic ions appeared t o be rather close t o that of the styrene divinylbenzene resins with the same swelling; however, complete agreement was not observed (Figs. 3, 4). Fig. 5 gives some information on the kinetic properties and permeability of the fibrous exchangers towards large ions. Martsinkevich and Pokrovskaya showed that the fibrous ion exchangers based on polystyrene fibers can be successfully used for chromatographic separation of inorganic ions (see Fig. 6 ) .
3. SORPTION OF ACID GASES FROM AIR
This series of studies have been carried out by Elinson and Tsigankov. The strong and medium base fibers in the form o f staple have been tested for absorption o f acid gases: S O 2 . H,S etc. Figs. 6. 7, 8 reproduce breakthrough curves for SOz and H2S absorbed from air by the fibrous ion exchangers against those for the granulous ion exchange resins, the columns being of the same size. The data on breakthrough capacity, some characteristics of ion exchangers and the experimental conditions are summarized in Table 3. It is evident from the above data that the fibrous ion exchangers can be efficiently used for air purification from S O 2 . The SBF ion exchanger does not practically yield in capacity t o its granulous analogue. The MBF(I1) ion exchanger substantially exceeds other ion exchangers in capacity. At the same time this ion exchanger is inefficient for air purification from H 2 S . In this case the SBF ion exchanger offers the best properties. It is a valuable feature o f the SAF exchangers that their breakthrough capacity towards the gases studied is practically independent o f the SO2 and H z S concentration in the air, their contents may be much lower than MPC (maximum permissible concentration).
362 4. SORPTION OF IONS FROM WATER SOLUTIONS
The fibrous strong base and strong acid ion exchangers as well as their granulous analogues can be successfully used for the environment protection. Again, it is their advantage that they display a higher rate of absorption and desorption which makes it possible to increase the flow rate through the columns. The possibility of using strong acid and strong base granulous resins in this field has been widely discussed it is unnecessary to give additional comments. At the same time, the application of new medium base exchangers, MBF, opens some new possibilities that are illustrated by the following examples. 4.1. Removal of Cr(V1) from Waste Waters of the Electroplating Baths
Pestrak and Bulatskaya studied the application of the MBF exchangers for treatment of the waste water discharged after washing some chrome-plated industrial goods. The water composition was typical for that kind of industries. It contained the following ions: Cr3+,Cr", Ca*+?C1-, K'. The experiments were carried out on both model solutions and a real waste water. The MBF sample of the total exchange capacity, 7.2 mg-eyv/g, was used in the form of staple in the columns, 10.6 cm high and 1.9 cm in diameter. The flow rate was 25.5 ml/min or 5.4 m/h. The parameters chosen were specific to a commercial installation used for the same purpose. The industrial methods applied in the Soviet Union for treatment of the Cr(V1)-containing waste waters (4) involve the removal of Cr3+ by the sulphonic-type cation exchanger, (KU-2), and the extraction of anions of CrO:-(Cr, O;-) by the strong base anion exchanger, (AV-17). After the anion exchanger saturated, chromium anions should be washed out by OH- or CO:-. The strong base exchangers have high selectivity to the chromate and bichromate ions, and a large excess of a strong alkali solution is required for the regeneration of the anion exchange column. (See Table 4). The weak base resins cannot be used due to their intensive oxidation and destruction in the course of regeneration. Fig. 9 illustrates the breakthrough curves of Cr(V1) absorbed by the MBF resin from the waste water against the AB-17 commercial resin. Table 4 gives some additional information on the characteristics of these processes. It is seen that the NBF sorption capacity is 4 times as high as that of AV-17 (0.250 g Cr/g resin and 0.060 g Cr/g resin, respectively). The use of the fibrous ion exchangers makes it possible to intensify the process by increasing a flow rate; no change in the dynamic activity of the MBF has been noticed when the flow rate increased by a factor of two. Regeneration of the columns with the MBF exchanger is much easier than that with AV-17. Several times smaller Table 4 Ion exchanger type
Breakthrough capacity g Crh+ / g ion cxchange
mgeqv NaOH/g ion exchanger for regeneration
M BF AV-17 EDE-lOP*
0.250 0.063 0.067 0.150
17 5 00 65 decompose
AH-31*
* the commercial weak base anion exchangers
363
Fig. 9. Breakthrough curves for the Cr(V1) sorption from waste water. Flow rate, 25.5 ml/min (5.4 m/h). Column dimensions: d = 1.9 cm; 11 = 8 . 3 cm. The MBF(1) mass, 6.2750 g; the AV-17 mass, 3833 g. 1 - AV-17; capacity, 3.45 mg-eqv/g; 2 - MBF(I), capacity, 6.50 mg-eqv/g.
amounts of alkali a t lower concentrations are required in this case. The MBF exchangers were not destructed in the chromium (VI) solutions. The origin of the MBF stability is not clearly understood at present. We have observed a change in colour (from yellow t o grey-green) after a continuous contact of the exchanger with the Cr(V1)-solutions. At the same time, n o decrease in the exchange capacity towards chromium anions have been detected. Even a small increase in capacity has been observed during 10 complete cycles of the sorption-desorption processes. The nature of this phenomenon is under investigation.
4.2. Selective Removal of Complexing Cations from Water Solutions
It has been established that the MBF exchanger analogous t o many weak base exchangers can absorb the complexing cations such as Cu2+,Ni”, Coz+ etc. The sorption runs fast and can be efficiently controlled by pH of the solution. Fig. 10 presents the sorption of Cu2+,Ni” and Coz+ plotted against pH of the solutions. All the cations are absorbed from the 0.01 N solution in 1 N NaCl. A possibility o f the selective removal of these cations from aqueous solutions is seen from the figure. These problems are of actual importance for many industries dealing with electrodeposition, metal etching etc.
5. CONCLUSIONS
The fibrous ion exchangers described in the present paper can serve as an efficient means for air and water purification. They feature a uniform structure, a small monofilament diameter, a high capacity, chemical and thermal stability, satisfactory strength.
364
45
0
Fig. 10. Equilibrium sorption of Cuz*, Ni’’ and Co” ion exchangers from the 0.01 N solutions of their chlorides in 1 N NaCl as a function of pH. 1 - Cu2’, 2 - Ni2’, 3 - Co”.
The ion exchange filters on the base of the fibrous ion exchangers have a high rate of sorption combined with weak resistance to the flows of liquids and gases.
ACKNOWLEDGEMENTS
The author would like to thank A. A. Shunkevich, G. I. Sergeev, I. S. Elinson, V. I. Tsigankov, A. I . Pokrovskaya, 0. P. Popova, R. V. Martsinkevich, A. F . Pestrak, N. G. Bulatskaya, V. A. Litvinenko for presenting the experimental data and technical assistance in peparing the paper.
REFERENCES
1 M. P. Zverev, Chemosorption Fibers (Russ.), M., Khimiya, 1981, p. 191. 2 L. A. Volf (Ed.), Fibers with Special Properties (Russ.), M., Khimiya, 1980. 3 I. N. Ermolenko, E. D. Buglov, I. P. Lyubliner, S . P. Dovgilev, New Fibrous Sorbents for Medical Application (Russ.), Minsk, Nauka i Tekhnika, 1978, p. 215. 4 T. G. Suslina, S. I. Chagina, V. B. Vojtovich, R. F. Kambarova, Ion Exchange Purification of AcidBase Sewage of the Eletsk Tractor Hydraulic Unit Plant (Russ.), Teoriya i Praktika Sorbtsionnykh Protsessov, Voronezh, 1976, Vol. 11, p. 96.
365
SELECTIVE COLLECTION OF SELENIUM (IV)FROM ENVIRONMENTAL WATER BY FUNCTIONALIZED ION-EXCHANGE RESIN
H . TANAKA
Faculty of Pharmaceutical Sciences, Kyoto University, Sakyo-ku, K y o t o 606, Japan M. NAKAYAMA
Department of Industrial Chemistry, Kumarnoto University, Kurokami. Kumamoto 860, Japan
M . CHIKUMA Chest Disease Research Institute, Kyoto University, Sakyo-ku, K y o t o 606, Japan
T. TANAKA Megi College of Pharmacy, Nozawa, Setagaya-ku, Tokyo 154, Japan
K. ITOH Environmental Pollution Research Institite, c i t y of Nagoya, Minami-ku, Nagoya 45 7, Japan
H. SAKURAI Faculty of Pharmaceutical Sciences, Tokushima University, Shomachi, Tokushima 770, Japan
ABSTRACT A versatile method for the preparation of useful functional resins for the collection of various kinds of environmental pollutants has been developed. Common ion-exchange resins could be converted into the functional resins by the simple treatment with somc reagents, which possess terfunctional property; namely selective reactivity with a metal ion, anion or organic compound which we want to collect o n the functional resin, capability of counter-ion-exchange with ion-exchange resins and strong physical adsorption t o ion-exchange resins. As an example of the new functional resin, bismuthiol-I1 resin, which is effective for the selective collection of selenium (I\?, is presented. Collection of selenium (IV) was found t o be based on the formation of selcnotrisulfide from mercapto group of the reagent. Elution of selenium (IV) was achieved by concentrated nitric acid, cysteine and penicillamine. Application of bismuthiol-I1 resin to the collection of selenium (IV) from some environmental water samples gave reasonable results.
366 1. INTRODUCTION
We have developed a simple and versatile method for the preparation o f functional resins for various purposes by the conversion of commonly used ion-exchange resins. Ion-exchange resins can be converted into various functional resins by treatment with some reagents wliiclt possess terfunctional property, namely capabilities of a highly selective reaction with the substance, which we want to collect on the functional resin, ion-exchange reaction with the ion-exchange resin and further strong physical adsorption to the ion-exchange resin matrix. Our idea of the preparation of the functional resin is illustrated in Fig. 1. We have developed some useful resins for the collection of mercury [ l ] , silver and fluoride [?] by the use of the terfunctional reagent and an anion-exchange resin. This paper deals with the collection of selenium(1V) by the use of bisniuthiol-11 and azothiopyrinsulfonic acid (ATPS) (Fig. 2 ) as terfunctional reagents and conimonly used anion-exchange resin, as an example of the development and application of the terfunctional reagent and the functional resin. Selenium is an element which has been widely
A N I O K - EXCHANCE RESIP;
TERFI:NCTIONAL REA2ENl' O ( = F L Y C T I O N A L GRDI'P)
EXC!iANCE
I'L'NCT InNAL
RESIN
RESIS
Fig. 1. Preparation and regeneration of anion-exchange resin functionalized - ~-physical intcraction.
N----- N
Fig. 2. Terfunctional r e g e n t .
\I
ith terfunctional rcagcnt.
,LD
367 used in industry and has attracted keen interest in both biological essentiality and high toxicity. Some sorbents, such as XAD-resin [3] and activated carbon [4] have been used for the collection of selenium of very low concentration from water samples. but these sorbents lack in selectivity. In an attempt t o develop a new functional resin for the selective collection o f selenium(1V) by the use o f a terfunctional reagent and an ion-exchange resin, we have examined the applicability of the reagents mentioned above. Further, application of the functional resin to the collection o f selenium from environmental water samples has been investigated.
2 . I X P E RIM ENTAL 2.1. Preparation of Functional Resin
The anion-exchange resin (Amberliie IRA-400. 100-200 mesh) was added to the solution of the terfunctional reagent and the mixture was shaken 3t 30' for about 1 hour. The resin was separated by filtration, washed with water and methanol. air-dried. and stored in a refrigerator. The resin. on which 0.2 mmole of bismuthiol-11 per one gram of the anion-exchange resin was loaded. was used unless otherwise stated. 2 . 2 . Detcrmination of Selenium
Selenium(1V) in solution \ + a s clcterniined by thc fluorometric method b! thc use of 2.3-dianinonaplitlialene with a Shimadzu spectrofluoropliotometer RF-500.
Bismutliiol-II resin (50 mg; bismutliiol-I1 0.20. 0.30. 0.60 and 1 .O ininole per one gram of resin) was shaken with excess o f selenium(1V). fl).drochloric acid concentration o f this sample solution was adjusted by the addition of 5 M hydrochloric acid prior to the shaking with the resin. The amount o f seleniuin(1V) left in solution was deterininzd. 2.4. Adsorption Isotherm
Bismuthiol-II resin (100 mg) was shaken with 50 ml o f 0.3 M hydrochloric acid solution containing known amounts of selenium(1V) for 5 hr at 30'. After equilibrium was reached, aliyuots of the solution were taken and the concentration of selenium(1V) left in the solution was determined by fluorometry. 2.5. Elution of Sclcnium(1V) Adsorbed o n Bisniutliiol-I1 Resin
Method 1: Bismuthiol-I1 resin (200 mg) was shaken with 50 ml of 0.3 M hydrochloric acid solution containing 100 pg of selenium(1V) for 5 hr at 30°. Bismuthiol-I1 resin
368 which adsorbed completely selenium(1V) was packed in a glass column (0.7 cm diameter). The column was eluted with 20 ml of 6-13 M nitric acid. The aliquots of the eluate were taken and the concentration of selenium(1V) was determined by fluorometry after adjustment of pH value of the solution to 1 with ammonium hydroxide solution. Method 11: Bismuthiol-I1 resh (100 or 200 mg) was shaken with 50 ml of 0.3 M hydrochloric acid solution containing 100 or 500 pg of selenium(1V) for 5 hr at 30'. Bismuthiol-I1 resin which adsorbed completely selenium(1V) was packed in a glass column (0.7 cm diameter). Aqueous solution of cysteine or penicillamine (20 ml) was passed through the column at a flow-rate of 1.0 ml/min, and the column was washed with 50 ml of distilled and demineralized water. Solution of 10 M nitric acid (20 ml) was continuously passed through the column, and the amount of selenium in eluate, which corresponds with the amount of selenium remaining on the resin, was determined.
2.6. Collection of Selenium(1V) by Column Operation
A column (1.0 cm diameter) was packed with bismuthiol-I1 resin to a height of 5.0 cni. After the column had been washed with 100 ml of distilled and demineralized water, 10 mg/l selenium(1V) solution in 0.3 M hydrochloric acid was passed through the column at a flow-rate of 1.O ml/min. The concentration of selenium(1V) solution passed through the column was determined by fluorometry. The resin was washed with 50 ml of distilled and demineralized water. Selenium(1V) adsorbed on the resin was eluted with 0.1 M penicillamine at pH 5 (flow-rate: 1.0 ml/min). After washing the column with 50 ml of distilled and demineralized water, 10 mg/l selenium(1V) in 0.3 M hydrochloric acid was again passed through the column for the examination of the capacity of the regenerated resin in repeated use.
3. RESULTS AND DISCUSSION
The exchange-capacity for bismuthiol-I1 was found to be 3.1 mmole/g-resin. When bismuthiol-I1 reacted with the anion-exchange resin in the chloride form, the ratio of chloride ion released to bismuthiol-I1 adsorbed on the resin was about 1 : 1 . This result indicates that bismuthiol-I1 was bound to the anion-exchange resin with its thiolate ion by the ion-exchange reaction. Bismuthiol-I1 was found to be retained on the resin, even when the resin was exposed to 0.1, 0.5 and 1.0 M sodium chloride solution. This strong fixation of bismuthiol-I1 on the resin is attributed to some physical interaction between bismuthiol-I1 and the ion-exchange resin. The binding capacity for selenium(1V) in 0.5 M hydrochloric acid is shown in Fig. 3. The time required for 50% uptake of selenium(1V) was less than 20 min. Upon adsorption of selenium(1V). the color of the resin changed from light brown to pale yellow, The binding-capacity for selenium(1V) increased linearly with the increase of bismuthiol-I1 loaded, and the binding-ratio of selenium(1V) to bismuthiol-11 on the resin was about 1 : 4. The reaction between bismuthiol-I1 and selenium(1V) can be expressed as fo 110ws.
369
0.251
Amount of Bismuthiol-U exchanged , mmolelg-resin Fig. 3. Binding capacity of bismuthiol-I1 resin for selenium (IV).
4 RSFI ‘t H, SeO 3 -* RS-Se-SR + RS-SR + 3 H, 0 (RSH: free form of bismuthiol-11) Complete adsorption of selenium(1V) below pH 2 can be attributed probably to the formation of selenotrisulfide on bismuthiol-11 resin. The binding capacity of selenium(1V) in the presence of metal ions was determined from the amount of selenium(1V) left in the solution. Except for copper(l1) and chromium(VI), the effect of the coexistence of the metal ion was found to be slight as shown in Fig. 4.
x
L
c
>
0 U
a
t 0
1
1 1
1
50 Concentration of metal ion, mq/( 10
Fig. 4. Effect of metal ion on recovery of scleniuni (IV). c Cu(ll), Cr(VI), 0 Fe(III), 0 Zn(II), x Mn(I1).
370 U
a4 D
5
10-
v)
U
.$
a
h
a
4
2;.
v
alm
m -
0
1:
Y3E + C
Y
0
E
a
0.1
1
1
I
1
Equilibrium concentration of Se(N), mg I I Fig. 5. Adsorption isotherms for selenium (IV) in 0.3 M hydrochloric acid. Concentration of sodium chloride: o 0 M, 0 0.5 M.
The adsorption isotherm for selenium(1V) in 0.3 M hydrochloric acid is shown in Fig. 5. Two adsorption isotherms obtained in the solution of different ionic strength gave agreement. Bismuthiol-11 resin is expected to be applicable to the collection of trace amounts of selenium(1V) from sea water, based o n the results shown in Fig. 5. Two possibilities can be taken into account in the elution of selenium(1V) adsorbed and the regeneration of the resin as follows. Method I: (resin-bisniuthiol-11-selenium) (resin) bismuthiol-11 + selenium Method 11: (resin-bismuthiol-11-selenium) (resin-bismuthiol-11) selenium where the solid state is designated in parentheses. Seleniuni(1V) was eluted completely with 8 M nitric acid by method 1. When cysteine or penicillamine was used as eluting agent, satisfactory results were obtained by method I1 as shown in Fig. 6. In this case, the formation of selenotrisulfide with these thiols was supported by the result of 77Se n.1n.r. study on the reaction product of this reaction [5]. In the case of method 1, the anion-exchange resin is regenerated and the eluate can be applied to the fluorometric determination of selenium(1V) directly after the proper pH adjustment. In the case of method 11, bismuthiol-11 resin is regenerated and the eluate can be applied to the fluorometric determination of selenium(1V) after the digestion. The collection of selenium(1V) by the column operation with bismuthiol-11 resin, and with that regenerated gave good results as shown in Fig. 7. The method presented here gave reasonable results in the collection of selenium(1V) from some environmental water samples taken from several points in the Japanese sea coast and river, as shown in Table 1. When ATPS loaded resin was used, the collection of selenium(1V) was not satisfactorily high. However, complete collection of selenium(1V) was achieved when ATPS was added to the sample solution which contains selenium(1V) and the reaction product was adsorbed to the ion-exchange resin [6].
-
-+
+
+
37 1
\
,-" >5
5 500
V al
m
0 - 1
I
II
1
I
I
- 1
Fig. 6 . Elution of selenium o n bismuthiol-I1 resin with thiols. o 0.1 M penicillamine, 0 0.05 M cysteine.
Effluent volume ,
ml
Fig. 7. Break-through curves for selenium (IV). Bismuthiol-11 resin: (I) bismuthiol-I1 0.2 mmole/g-resin, (11) regenerated r a i n \\ith 0.1 M penicillaminc at pH 5. Column: 1 0 x 5 0 mm. Selenium (IV) solution: 10 mg/l Se(1V) in 0.3 M HCI. Flou rate: 4 0 ml/hr.
I n conclusion, selective and effective collection of seleniuin(1V) was achieved w i t h bismuthiol-I1 resin. The method presented here may be valuable for the practical technology dealing with environmental water based on the following several points. Uismuthiol-I1 is synthesized easily from phenylhydrazine and carbondisulfide and commercially available. Bismuthiol-11 resin can be prepared and recovered easily and used repeatedly. Bismuthiol-11 resin was found to be stable for several months on the resin and selenium
372 Tab. 1. Determination of selenium (IV) in environmental water
Shirakawa River, Kokai, Kumamoto October, 1982 Shirakawa River, Kokai, Kumamoto March, 1983 Amakusa Basin, Amakusa, Kumamoto November, 1982 Ise Bay, Nagoya, Aichi December, 1982
0.075 0.099 0.016 0.06 3
(IV) adsorbed on the resin was confirmed to be stable for about three weeks. The reaction is rapid enough for the column operation, although detailed kinetic study has not yet been carried out. The molecular design and the choice of the terfunctional reagent which is highly reactive with the substance to be collected on the resin may lead to the development of the functional resins useful for the collection of various hazardous substances in the chemical protection of environment.
REFERENCES
1 M. Nakayma, M. Chikuma, H . Tanaka and T. Tanaka, Talanta, 29 (1982) 503-506. 2 H . Tanaka, M. Chikuma and M. Nakayama, in J. Albaiges (Ed.), Proc. 2nd Int. Congress, Analytical Techniques in Environmental Chemistry, Barcelona, November 23 -25, 1981, Pergamon Press, Oxford, 1982, pp. 381 -388. 3 Y . Sugimura and Y. Suzuki, J . Oceanogr. Soc. Jap., 33 (1977) 23-29. 4 H. J . Robberecht and R. E. Van Grieken, Anal. Chem., 52 (1980) 449-453. 5 H. Sakurai, in preparation. 6 M. Nakayama, M. Chikuma, H. Tanaka and T. Tanaka, Talanta, in press.
373
REhlOVAL OF ARSENIC FROM NATURAL WATERS
J . HLAVAY, K . FOLDI-POLYAK, J . INCZEDY
Institute for Analytical Chemistry, University of Chemical Engineering, P.O. Box 158, VeszprPm 8201, Hungary
ABSTRACT Use of some new adsorbents for arsenic removal from synthetic and dcep wcll waters was studied. Adsorbcnts were prepared from titanium dioxide, aluminium oxide and their mixtures, respcctively, as porous support materials and were treated chemically by frcshly precipitating iron (111) hydroxide onto their surfaces. The breakthrough capacities at 0.05 mg As/dm’ concentration ranged betwcen 3.15 and 8.68 mg As/g adsorbent. In laboratory experiments up t o 3100 BV synthetic solution spiked with 2.5-2.9 nig As(III)/dm3 can be loaded o n t o the column without reaching thc maximum permissible limit. For regeneration of the spent adsorbents, 30-40 BV of 1 M NaOH solution was used. In field experiments deep well water containing high amounts of humic acids, iron ions and dissolved gases was purified and the safe operation limit could be kept during the 20-22 days continuous cxhaustion period. Presumably, arsenic is removed by adsorption but chemisorption and occlusion can also occur.
1. INTRODUCTION
In certain geographical areas drinking water is contaminated with arsenic in concentrations high enough t o pose long term health hazards. Such contamination is believed to arise from natural leaching of arsenical rocks by the percolating water. In their survey McCabe et al. [ 11 found that 0.5%of the treated water samples exceeded the 0.01 mgfdm3 recommended limiting concentration and 0.2% of the samples exceeded the 0.05 nig/bm3 maximum permissible limit for drinking waters. More than 18,000 community water supplies in the USA were investigated in this survey. Similar problems were encountered in other parts of the world i.e. in Cordoba, Argentina, in Nova Scottia, Canada, and in Taiwan [2]. I n natural waters inorganic a r s e n i c is p r e s e n t as a r s e n i t e a n d a r s e n a t e . I n o x y g e n a t e d waters, arsenic acid species ( H 3 A s 0 4 , H2 AsO;) are stable while under mildly reducing conditions, arsenious acid species (H3 A s 0 3 , H2 AsO; and f-1AsOi-) become stable. The chemistry of arsenic in the aqueous environment has been reviewed in great detail [3]. Arsenic (111) is more toxic than arsenic (V), therefore, knowledge of the dominant arsenic species in drinking water is a very important analytical task.
374 Technologies used to remove arsenic from drinking water supplies have been summarized b y Patterson and Minear 141. Mostly chemical processes were successfull, in particular those including coagulation and precipitation with ferric salts and lime. Ion exchange and adsorption methods have also been used, but their application for large scale water purification is limited [5- lo]. The aim of our work was to develop new adsorbents and to investigate their characteristics and capacities for removal of arsenic ions from drinking water.
2. I~:XI’ERIMLNTAL
2.1. Prcparation ot‘ the Adsorbents
Adsorbents were prepared from porous supports after chemically treatment of their surface. First, granules of 0.5 ~ - .O1 mm particle size were formed from titanium dioxide, alutiiiaiutn oxide and their mixtures in different ratio, respectively. then iron (111) hydroxicie was freshly precipitated onto the surface of the particles. The resulting iron (111) hydroxide impregnated porous adsorbents were dried at room temperature. packed into xi ion-exchange coluiiin and washed with water t o remove the excess reagent. In one case, titanium dioxide and freshly precipitated iron (If I) hydroxide were mixed and granulated. Thus, not only the surface of the support material, rather inside o f the granules contained as well a great amount o f iron (111) hydroxide [ 171. Support inaterials were prepared from titanium and aluminium oxides as follows: 10052 titanium dioxide, 100% y-alumina and their 1 t o 1 mixture, respectively. The preparation metliod and the symbol of the adsorbents are summarized in Table 1. Tab. 1. Prcpxation nictliod and synibol of different adsorbents Raib material
Preparation
Sy ntbo1
TiO: TiO? 50% TiO, -50’2 Al,O, AI20 ,
TiO: -iron (Ill) hydroxide mixed TiO: -iron (111) hydroxide on surface Support-iron (Ill) hydroxide o n surface Al:O,-iron (111) hydroxidc 011surface
TIM TIS 50-50 TAlS AIS
Laboratory studies were carried out with 3.0 cni i.d. X 10 cni and 3.1 ciii i.d. X 15 ciii columns. The adsorbent bed volumes (BV) were 50 and 80 m i 3 , respectively, and the columns were operated in a downllow mode. The exhaustion flow rates ranged from 5 to 10 BV 1 1 - l . Synthetic solutions were prepared from tap water spiked with As(1ll) in the range of 0.1 3.9 m g / d m 3 . The stock solution was kept in a flask flushed with N, and resealed after each use t o prevent oxidation. Fresh stock solution was prepared daily and kept stoppered. I n field experiments drinking water from a deep well containing high amounts of dissolved gas, Iiuiiiic acids, ammonia, iron and manganese ions was purified. The water contained mostly As(V) ions in an average concentration of 0.2 mg/dni3. For these field experiments 9.5 cm X 92 cni columns were used and the adsorbent bed volume was
375 4 dni3. Tlie columns were operated in downflow mode during exhaustion and u p f l o ~ mode during regeneration. Regeneration of the spent adsorbents was carried out by 1 M NaOH solution. The regenerant flow rate was similar t o that used in the exhaustion process. After treatment with NaOH solution, t h e adsorbents were backwaslied with t a p water until neutral pH was reached.
2.2. Aiialy t ical Proced tire
Samples for determination of the adsorbent capacity were taken daily. A spectrophotometric method using the silver diethyldithiocarbamate (SDDC) complex of arsine was applied for the determination of arsenic in the 1-100 pg range. Arsenic is reduced to arsine by granular zinc in hydrochloric acid and arsine reacts with SDDC in pyridine. Tlie adsorbance of the red coloured coniplexisreadat 5 3 3 nni [ 2 ] .
3 . RESULTS AND DISCUSSION
It is known that colloidal iron (111) hydroxide is a good scavenger of arsenic ions. and can be used for the purification of waters [ 11-13]. According to the precipitation method. iron (111) chloride is added t o the water, iron (111) hydroxide tlocs are formed and a high amounts o f sludge have to be disposed. The arsenic content of this sludge can be as high as 0.5% w/w [ 11 1, so the sludge is a potential toxic pollutant. Disposal of toxic pollutants is a great problem all over the world. In our adsorption processes the granules are chemically treated o n their surface and colloidal iron (111) hydroxide is formed there. Freshly precipitated iron (111) hydroxide effectively adsorbs or chcmisorbs arsenic. The great advantage o f our method is that no sludge is produced and tlie adsorbents can partially be regenerated. Precipitation of arsenic froiii the spent regenerant will be discussed later.
3.1. Laboratory Experiments
First, the titanium-dioxide-iron (111) hydroxide mixed (TIM) and titanium-dioxideiron (111) hydroxide o n surface (TIS) adsorbents were compared. The results of tlie exhaustion experiments are shown in Fig. 1. Breakthrougli concentration (0.05 nig As/dm3) was reached after loading 15 10 BV and 3080 BV drinking water onto the TIM and the TIS columns, respectively. The breakthrough capacity of TIM adsorbent was about 35% of that of the TIS adsorbent. This means that colloidal iron (111) hydroxide can effectively remove arsenic from drinking water if it is freshly precipitated onto the surface of the support. The adsorption capaand therefore the operation efficiency of the adsorbent is much less when city iron (111) hydroxide is granulated with titanium dioxide. In further experiments, only the surface treated adsorbents were used. -
-
376
Fig. 1. Exhaustion experiment on TIM and TIS adsorbents. Influent As(II1) conc: 2.50 mg As/dni3 for TIM and 2.94 mg As/dm3 for TIS; flow rate: 5 B V h'l , pH =7.21; breakthrough capacity: 3.15 mg As/g TIM and 8.68 rng As/g TIS.
A 5050TAIS AIS
IS00
lB00 2000 loxh-g[Bv]
I:&. 2. Lxhaustion cxpcrimcnt o n 50-50 TAIS and AIS adsorbents. Influent As(II1) conc: 3.97 nig As/dni3 for 50-50 TAIS and 3.48 mg As/dm3 for AIS; flow rate: 10 BV 11.' : breakthrough capacity: 4.22 mg As/g 50-50 TAIS and 3.48 mg As/g AIS, pH = 7.23.
377 The results o f exhaustion experiments carried o u t with 50-50 TAIS and AIS adsorbents are summarized in Fig. 2. In this case, the influent As(II1) concentration was considerably increased (up t o 4 mg As/dm3) and the flow rate was doubled. The effect of higher arsenic concentration and flow rate can be seen o n t h e breakthrough volumes, i.e. 990 BV for tlie AIS and 1360 BV for tlie 50-50 TAIS adsorbents. These are much lower than the 3030 BV values for the TIS adsorbent but, considering tlie experimental conditions, the 4.22 mg As/g 50-50 TAIS and 3.48 nig As/g AIS breakthrough capacities are high enough t o justify their use in field experiments.
3.2. Regeneration Study
The efficiency and the simplicity of the regeneration process is a very important factor in water purification techniques. Spent adsorbents were regenerated by 1 M NaOH of a flow rate of 5 BV h-' . The results of the regeneration are shown in Table 2 . Tab. 2. Regeneration of spent adsorbents
Adsorbent
Adsorbed
Regcnerant
Back\vash water
As deaorbed (nig)
As(lI1) mg/dm3
BV
dm3
BV din3
1 M NaOH
Water
30 40 40 30
1.5 2.0 3.2 1.5
90 50 40 40
11.79 157.82 314.38 143.96
6.57 4.26 2.46 0.65
TIM 213.8 473.21 TIS 50-50 TAIS 545.44 194.37 AIS
4.5 2.5 3.2 2.0
Rcgeiicrat ion efficiency Total %
18.36 8.59 162.08 34.25 316.84 58.09 144.61 74.40
It can be seen that adsorbed arsenic could be removed by 30-40 BV regenerant but tlie recovery efficiency was poor. Even excess regenerant could not rcmove more As from the adsorbent. In their experiments, Shigetonii et al. 171 used 0.5 M NaOH solution as rcgenerant to remove the adsorbed arsenate from the polyacrylamide-bound hydrous iron (111) oxide. Regeneration efficiency is not quoted. except that the original adsorption capacity could not be achieved in successive treatments. The renioval tneclianism was not discussed either. Considering tlie efficiency of the laboratory scale regeneration process it can be concluded that not only adsorption has taken placc. I t is known that at high pl4 the surface of iron (111) hydroxide is negatively charged and anions can be desorbed 1141. Arsenic is removed froni drinking water by adsorption and simultaneous other processes. Fergusan et al. 1151 gave no answer whether arsenic adsorption can be e'xplained solely in terms of molecule-surface interactions o r whether other niechanisni. e.g. precipitation or occlusion, are also operative. They used amorphous ferric and aluminium hydroxide precipitates in adsorption experiments where the precipitates and arseniccontaining samples were stirred in a flask at 25 +0.SoC. Their conclusion, concerning the removal meclianisni, was drawn only from tile adsorption experiments because n o regeneration or desorption experiments were done. Our regeneration results (in which regenera-
378 tion efficiency ranged between 8.5%-74.4%) suggest that part of the arsenite is bound by molecule-surface interactions, the rest b y other mechanisms. Since similar adsorbents, i.e. titanium-dioxide o r aluminium-oxide chemically treated by iron (111) hydroxide o n its surface, have not been used yet, comparison o f o u r results with literature data is difficult. Further experiments are needed to confirm these results and conclusions.
3.3. Field Experiments
The average ionic composition of the deep-well water used in the field experiments is listed in Table 3. It should be pointed o u t that there is an unusally high concentration of humic acid, iron ions and dissolved gases in this water. No pretreatment steps were carried o u t o n the water t o be purified before its loading onto the columns. Tab. 3. Average ionic composition of a dcep well water Coiiiponent
Conccnt ration (mg/dni3 ) 45 -70 40 -70 2,2-2,6 11,5-22 10-18 0.3-4.2 0.05-1.0 3s -37 0,2-0,6 Y,2-13,2 45-55 7,6-8,05
Two exhaustion and a regeneration experiment were carried out in a 72 days continuous investigation. Siniilarly to the laboratory experiments, 1 M NaOH was used as a regenerant solution in 35 BV in both cases. Only two adsorbents, the TIS and the AIS, were used in the field experiments. The results of this study are shown in Table 4. The first exhaustion cycle lasted 37 days while the second, after regeneration, 35 days. Comparing the operation of the two adsorbents, the TIS shows a higher removal efficiency and can be used safely for more than 20 days (1250-1300 BV of treated water) without reaching the 0.05 mg As/dni3 breakthrough limit. The removal efficiencies were not as high as measured with synthetic solutions but since adsorbents were exli’austed more after they had reached the breakthrough levels. the average As concentrations in the effluent seem t o be low enough to warrant trying the adsorbents in pilot plant scale tests. I n the adsorption process discussed earlier. not only arsenate and arsenite ions but great amounts of humic acids of large molecular weight were removed from the water. Humic acids form many complexes with metal ions present in natural water, and these
379 Tab. 4. Results of the field experiments AIS
TIS Adsorbents ~~~
~
1. set
2. set*
1. set
2. set*
37 9.04 2.5 1
35 8.46 2.50
37 9.16 2.48
35 10.4 3.08
~
Experimental time (d) Loaded volume (m’) Flow rate (BV/h) Influent As conc. average (mg/dm3) Effluent As conc. avcrage (mddm’) Removal efficicncy (%) BV for 0.05 mg/dm’ breakthrough conc. Exhaustion time for breakthrough (d) Regeneration efficiency %
0.165
0.281
0.074 56.7
0.094 65.9
1250-1300 20-22 41.9
0.189
0.285
0.077
5 8.0
0.145 48.0
750-800
1000
400
12
17 30.2
8
* after regeneration are also adsorbed by iron (111) hydroxide freshly precipitated onto the surface of a support. In the field experiments the amount of humic acid of natural water decreased in a range of 25-35% due t o adsorption. Therefore, the adsorption sites can be blocked by humic acid molecules decreasing the possibilities of arsenic ion adsorption. In the second set of the experiments even less favorable results were obtained. Since the regeneration was far from complete (41.9% for TIS and 30.2% for AIS) the safe operation period was only 12 and 8 days, respectively. This means that this regeneration process is not suitable for a long operation period. Development of new regeneration method is in progress and the results will be discussed in the near future. It is known that the removal of arsenic ions from waters by adsorption is a pH dependent process [14]. Since the pH of drinking water does not change widely, the pH of the treated synthetic solutions and deep well water was not adjusted. The initial pH of water was kept and the final pH of the purified solution was recorded. No significant pl-l change was noted during the exhaustion processes. The economy of our adsorption process has not been established yet. Experiments were carried out only to develop this new idea, i.e. the use of chemically treated adsorbents for the removal of arsenic ions from water. Compared t o the coagulation and precipitation processes one advantage is immediately apparent: no excess sludge is produced. The high amount of arsenic in the small volume of the regenerant solution might be removed by lime precipitation [16]. Backwash water is also needed t o reach the original pH of the adsorbent but much less of it is needed than for the backwash of the filter bed in the coagulation technique. There is a patent pending for the process discussed here [ 171.
REFERENCES 1 L. J . McCabe, J . M . Synions, R. D. LCCand G . G. Robeck, J . Anier. Water Works Assoc.. 62, 1970. 670.
2 Environmental Health Criteria 18, Arsenic, WHO, Geneva, 1981, 44. 3 J. F. Ferguson and J. Gavis, Water Research, 6, 1972, 1259. 4 J . W. Patterson and R. A. Minear, Wastewater Treatment Technology, 2nd Ed. Illinois Institute for Environmental Quality Document 73-1, February, 1973. 5 Sato, H., Shigeta, S. and Uchida, H., Jpn, Kokai Tokkyo Koho 80 08 843, 22 Jan. 1980. 6 E. Bellack, J . Amer. Water Works Assoc. 63, 1971, 454. 7 Y . Shigetonii, Y. Hori and T. Kojima, Bull. Chem. SOC.Jpn. 53, 1980, 1475. 8 C. M. Elson, D. H. Davies, and E. R . Hayes, Water Research 14, 1980, 1307. 9 J , H. Gulledge and J . T. O’Connor, J. Amer. Water Works Assoc., 65, 1973,548. 10 S. K. Gupta and K. Y. Chcn, J . of Water Poll. Control Federation, 50, 1978, 493. 11 M. Csanady and B. Kelemen, J . Hungarian Hydrological Society 26, 1982, 378. 12 K. Kermer, I. Roske, Acta Hydrochim. Hydrobiol., 5, 1977, 55 1. 13 I. Roske, K. Kermer, Acta Hydrochim. Hydrobiol., 7, 1979, 115. 14 J. I. Morgan and W. Stumrn. in 0. Jaag (Ed.) Advances in Water Pollution Research 1st. edn. Vol. 1. Perganion Press Inc., 1965, p. 103. 15 J. F. Fergusan and M. A. Anderson, in A. J. Rubin (Ed.), Chemistry of Water Supply, Treatment and Distribution. Ann Arbor Science Publisher Inc. 1974, p. 137. 16 L. M. Magnusen, T. C. Waugli, 0. K. Galle and J. Bredfeldt, Science 168, 1970, 389. 17 J. Hlavay et al. Patent pending (1982).
38 1
PRECONCENTRATION AND SEPARATION OF Cr(III) AND Cr(VI) FROM AQUEOUS SOLUTIONS BY COMPLEX FORMATION-ION EXCHANGE
C. SARZANINI, E. MARENGO, M. C. GENNARO, C. BAIOCCHI and E. MBNTASTl
Istituto di Chimica Analitica, Universita di Torino, Via P. Giuria 5, 101 25 Torino, Italy
ABSTRACT A preconcentration and separation procedure for Cr(II1) and Cr(V1) has been developed. 1,2-dihydroxybenzene-3,5disulphonic acid (Tiron) is an effective chelating agent for Cr(II1) and, in addition, allows the adsorption of the resulting complex on anion-exchange resin through the interaction of the -SO; groups which are not involved in the coordination act. The elution of aqueous samples containing mixturcs of Cr(II1) and Cr(V1) through an anionexchange resin (pH 5) will result in the adsorption of Cr(V1). The recovered solution is then treated with Tiron and brought to pH 9. Elution will produce the adsorption of Cr(II1) in the form of its Tiron complex. After fixation, the adsorbed species are eluted with the appropriate reagent (HC10, solution for Cr(II1) and NH,CI solution, brought to pH 8, for Cr(V1)). According t o the described procedures, solutions of t h e two analytes at ppb levels, wcre concentrated by a factor of lo', as tested by AAS, with yields of 104 + 9 and 94 +8%, respectively for Cr(II1) and Cr(VI).
1 . INTRODUCTION
The separation and determination of metals ions at trace levels including the same species in different oxidation states, is important both environmentally and geochemically. A method has been described for the preconcentration of Fe, Cu, Pb, Cd, Ca and Mg [ 13 based on the formation of metal-l,2-dihydroxy-3,5-benzenedisulphonic acid complexes and their subsequent adsorption onto an anion-exchange macroporous resin. We report here a detailed investigation on the applicability of this method to the preconcentration and determination of Cr(II1) and Cr(V1). Chromium exists in solution mainly in two oxidation states Cr(II1) and Cr(V1). The first is the more stable and in the presence of non-complexing anions (e.g. C104) exists as Cr(Hk0)l' and its hydrolysis products [2], chromium (VI) undergoes a series of eyuilibria [3-51.
* Work supported by
the National Research Council of Italy.
382 Already considerable attention has been devoted t o the separation and determination of chroniium at trace levels, extraction procedures are preferred [6] and preconcentration steps have also been described for neutron activation analysis [7]. A determination of Cr(I1l) and Cr(V1) by anion exchange was performed by Cresser and Hargitt [8] at 0.5 ppni levels. 2 . EXPI-RIMENTAL
2.1. Appararus. A Perkin Elmer 2380 atomic absorption spectrophotometer was used with air-acetylene burner, and a chromiuni hollow cathode lamp operated at 25 mA. 2.2.Reagetzts. The resin, macroscopus Bio-Rad AC MP1 100-200 mesh, was used in the nitrate form. Chromium (111) and (VI) solutions were prepared from analytical reagent-grade chromium (111) chloride and potassium dichromate (1,000 pg Criml) and these stock solutions. after appropriate dilution, showed identical absorbances in the air-acetylene flame. Demineralized and doubly distilled water from quartz (DDW) was used throughout. The disodium salt of 1.2-dihydroxy-3,5-benzenedisulphonic acid (Tiron) was purchased from C. Erba as reagent grade chemical (RP). 2.3. Procedure. The glass column (8 mni i d . borosilicate glass tube 30 cni height) was slurry loaded with the proper amount (1 g) of resin suspended in DDW. The desired volume of sample was added, when required, with 5 ml o f 0.050 M Tiron; the pH was adjusted t o the appropriate value and the solution was passed through the column at a constant flow 10 ml/min ensured by a rotary vacuum pump and by-pass flowmeter. After fixation the metals were released by elution with 10 in1 o f reagent (see below). The concentration of the investigated metal ion was measured in the eluate in order to evaluate the recovery efficiency. In order t o obtain the best precision and accuracy in the spectroscopic determinations, solutions with or without Tiron, for Cr(II1) and Cr(V1) respectively, have been used for zeroing and dilution. Blanks were periodically run following the same procedure with unspiked solutions. Each experiment was conuucted in triplicate. Between runs the column did not need t o be reslurried and was washed with saturated NaN03 solution in order t o remove the free ligand and repeatedly washed with DDW. Glassware, polythene and polypropylene laboratory materials were cleaned and washed according t o suggested procedures [9]. 3. RESULTS AND DISCUSSION
The half-life of the exchange of co-ordinated water for chromium (111) has been reported [ 101 t o be 40 hr with rate constants ranging from [ 1 1 1 2 lo-' t o 4.8 . 10 set' [lo]. For this reason a preliminary study was made t o define thc optimutn pH and reaction time o f Cr(I1I)-Tiron eluting solution, using 100 nil solutions of 1 ppiii concentration. In Fig. 1 the behaviour o f retention yield is reported for samples cluted after 2 Iirs of reaction. At pH 9 the recovery yield was 96.0 -+2.7%1 and a t thc uiiie pll witti elution performed immediately after the ligand addition, 94.5 &0.5%.Tliis dciiionstrates that the recovery is unaffected by the reaction time. '
383
100
80
60
8 d 3
40
20
0
3
5
7
PH
Fig. 1. Uptake efficiency (u.c. %) for prcconcentratioii and recovery ofCr(ll1) as a t'unction of p H : with ( 0 ) or \\ itllout ( 0 ) rccycliiig tiic first elution.
In order to optimize the anion-exchange the same experiments weic performed by recycling the solution after the first elution. The recovery resulted 98.3 50.5% and 98.7 +0.3% at pH 9 and 8 respectively. thus showing the effectiveness of the recycling procedure. The release of Cr(lI1) immobilized on the resin involves the Cr(II1)-Tiron complex dissociation. To this purpose the reagents reported in Table 1 have been tested and the results are reported. The lower recovery obtained with HCI and H,S04 may well be due to competitive complex formation with the stripping reagent. Thus HCIO: has been used as the reagent and was preferred t o N H 0 3 , which, even if it showed good yields, might have attacked the polymeric structure of the resin. 3.1. Cr( VZ). Since Cr(V1) is stable in the form o f anionic species its adsorption on the anion exchange resin is effective in the absence o f any reagent. In order to determine the
3 84 Tab. 1. % Cr(II1) recovery efficiency, as a function of the stripping reagent concentration
% Recovery efficiency
Reagent H,SO, HCI HNO, HCIO,
1M 2M 1M 2M
60.9 73.3 89.2 96.0
pH conditions for highest recovery efficiency experiments on 100 ml aliquots of chromate solution (1 .O ppm) were performed at pH 2, 5 and 9. Table 2 reports the uptake efficiens obtained by monitoring Cr(V1) recovered from the resin with the appropriate stripping reagent (see below). Tab. 2. Preconcentration efficiency of Cr(V1) at different pH's; stripping reagent NH,CI 2M at pH 8 Adsorption PH
% Uptake recovery
No. measurements
2 5 9
91.6 * 1.8 99.2 k0.2 99.2 c 1.7
3 6 3
An increase in yield is observed between pH 2 and 5, while the yield remained unaffected above pH 5. This behaviour may be clearly correlated with the presence of equilibria such as:
which makes the anion exchange interaction less efficient below pH 5. Table 3 reports the results obtained with the different stripping procedures. As can be seen 2.OM N b C I previously brought to pH 8 was found the most effective reagent. Satisfactory results have been also achieved with 1 M NaCl solutions but these were avoided in order to reduce the Na' interference in the spectroscopic determinations. Measurements were conducted also in solutions containing either Cr(II1) and Cr(V1) Tab. 3. % Cr(V1) recovery efficiency, as a function of the stripping reagent and concentration
% Recovery efficiency
Reagent HCI NaCl NH,Cl NH,C1
-
2M 2M 1M 2M pH 8
89.5 102.4 80.4 99.2
at a concentration of 0.1 ppin in the presence of a high salinity (NaCl 0.5M) which resembles sea water. The results, which appear satisfactory, are: Cr(II1) pH 9 88.5 +0.3% Cr(V1) pH 5 67.6 *0.7% Table 4 reports the results obtained on mixtures of Cr(II1) and Cr(VI), Tab. 5 collects some data obtained on solutions of low concentration (0.005 ppm). For these experiments the following procedure has been adopted: the solution (1 1) was first brought to pH 5 and eluted through the resin in order to recover Cr(V1). The appropriate volume of Tiron solution was then added to the eluate, which was subsequently brought to pH 9. Elution on the anion exchange resin allowed the recovery of Cr(II1). Tab. 4. Preconcentration efficiency of Cr(II1) and Cr(V1) mixtures Added pg/ml
Found* pg/ml
Cr(II1) 0.05 0.10 0.10
Cr(II1) 4.76 t.0.13 8.97 t.0.02 9.38 kO.01
Cr(V1) 0.10 0.10 0.05
Cr (VI) 9.49 50.04 9.49 +0.52 4.86 k0.17
* Mean
and standard deviation for 3 samples (preconcentration 100 fold).
Tab. 5. Preconcentration efficiency of Cr(II1) and Cr(V1) mixtures at ppb levels Added ng/ml
Found* ng/ml
Cr(II1) Cr(V1) 5.00 5.00 Recovery %
Cr(II1) 5.22 k0.47 104.4 t. 9.3
Cr(V1) 4.72 +0.38 94.40 + 7.6
*
Mean and standard deviation for 3 samples (preconcentration 100 fold).
REFERENCES 1 C. Sarzanini, E. Mentasti, M. C. Gennaro and C. Baiocchi, Ann. Chini. (Rome), accepted for pubblication. 2 K. Emerson and W. M. Gravcn, J . Inorg. Nucl. Chem., 1 1 (1959) 309. 3 W. G. Davies and J . E. Prue, Trans. Faraday SOC.,51 (1955) 45. 4 J . D. Neuss and W. Rieman 111, J . Am. Cheni. SOC.,56 (1934) 2238. 5 J . Y . Tong and E. L. King, Trans. Faraday SOC.,51 (1955) 1045. 6 V. M. Rao and M. N. Sastri, Talanta, 27 (1980) 771. 7 J. F. Pankow and G . E. Janauer, Anal. Chini. Acta, 69 (1974) 97. 8 M. S. Cresser and R . Hargitt, Anal. Chin. Acta, 81 (1976) 196. 9 R. V. Moore, Anal. Cheni., 54 (1982) 1890. 10 1. P. Hunt and H . Taube, J. Cheni. Pliys., 1 9 (1951) 602. 11 R. E. Dickerson, H. B. Gray and G. P. Haight, Jr., Chemical Principles, 2nd edn., Bcnjanun, Mcnlo Park, 1974, p. 796.
This Page Intentionally Left Blank
387
THE USABILITY OF POLYSULFONE MEMBRANES FOR REMOVAL OF ORGANIC DYES FROM AQUEOUS SOLUTIONS
K. MAJ EW SKA-NOWAK
Institute of Environment Protection Engineering, Technical University of Wroclaw Wybrzeie Wyspialiskiego 27, 50-370 Wroclaw, Poland T. WINNICKI Rivers State University of Science and Technology, Chem./Petrochem. Department P.M.B. 5080 Port Harcourt, Nigeria
ABSTRACT Polysulfone membranes were synthesized and their separation properties were studied to check the usability of this type of membranes in decolorizing aqueous solutions of organic dyes by ultrafiltration. The membranes were prepared from 15-percent (wt.%) polysulfone solution in dimethylformamide and formed o n a glass support. The casting solution had different initial temperatures. The membrane thickness was between 50 and 115 pm. Experiments were carried o u t under static conditions and involved six organic dyes having molecular weights between 291 and 1048.2. The dye concentration in the aqueous solution was 100 gm-3. The pressure applied Mas varied between 0.5 and 2.5 MPa. Separation propcrties of the membranes were found to be better with thinner membranes and lower temperatures of the casting solution. The membranes are capable of retaining 100 percent of organic dyes (of a molecular weight of above 800) irrespective of the casting parameters and of the pressure applied. The volume flows of dye solutions through the membranes varied from 0.05 t o 0.20 1n3m-’ day-’ (at 2.5 MPa) dcpcnding on tlic casting parametcrs.
1. lNTRODUCTION
Conventional treatment methods w h c h have been used so far to remove hardly degradable organic dyes from industrial wastewaters, actually do not yield satisfactory decolorization effects. Thus, the need for more efficient techniques has directed the attention of many investigators to the application of membrane processes (such as ultra- or hyperfiltration), which involve the selective action of semipermeable membranes to separate organic dyes from aqueous solutions. Membrane processes not only permit an abatement in the pollution load, but they also allow for the reuse of the water purified via this route and for recovery of some other valuable substances. At the present time, polysulfone membranes are accepted as being the most promising among the
388 variety of types employed, because they may be used in a very wide range of pH [l], as well as at elevated temperatures (up to 378 K) [2]. Our preliminary results have shown that polysulfone membranes perform very high removal ability (from 90 to 100 percent) toward some high-molecular-weight dyes (e.g., direct meta black of molecular weight of 781.2) [3]. The investigations reported in this paper, being the continuation of our previous studies, are aimed at final evaluating the utility of polysulfone membranes in ultrafiltration decolorizing and concentrating aqueous solutions of organic dyes.
2. EXPERIMENTAL 2.1. Preparation of Polysulfone Membranes
The casting solution containing 15 wt.% of P 3500 aromatic polysulfone (Union Carbide) dissolved in dimethylformamide (DMF) was prepared according to the procedure described by Koenst and Mitchell [4]. The membranes were cast by means of a typical device consisting of a leveled glass sheet and movable frame with two micrometric screws fixed to it with their upper ends and to a casting knife with the lower ones. The screws were used to control the distance between the knife edge and the glass surface and, consequently, the membrane thickness. The frame was moved slowly along a leveled bench by means of a small electric motor and the knife spread the solution uniformly on the glass surface. The speed of the frame movement was also controlled. Based on the previous results [3], membranes 50-1 15 pm thick were cast at four different initial temperatures of the casting solution (303, 318, 333, and 348 K) and then the membranes were left at room temperature for 60 seconds to evaporate the solvent from their surface.
2.2. Testing Device
The polysulfone membranes prepared according to the method described above were tested under static conditions and without stirring a feed solution at the membrane surface. The testing device (Fig. 1) was a pressure apparatus (2) made of stainless steel. The membrane sample (1 1) of an effective surface area of 15.9 cmz was fixed in the lower part of the apparatus and supported by a porous PVC sinter. The solution t o be tested was fed to the apparatus in its upper part through an inlet pipe (3). To obtain and maintain the pressure value required, nitrogen was supplied from a cylinder (7) and passed through a reducing valve (6) to an inlet (4).
2.3. Testing Method
Each experimental cycle was preceded by conditioning the membrane samples. In this process distilled water was passed through the membrane under a pressure of 2.5
389
Fig. 1. Apparatus for membrane testing under static conditions: (1) outlet of permeate; (2) pressure apparatus; (3) stub pipe for solution supply; (4) stub pipe for gas supply; (5, 6) reducing valves; (7) gas cylinder; (8) needle valve; (9) receiver; (10) rubber gasket; (11) membrane; (12) filter paper; (13) sinter;(l4) grid.
MPa, until a steady volume flow was achieved, which usually occurred after a period of 30 to 80 h. Model solutions of selected organic dyes (Table 1) of concentration of 100 gm-3 were used to examine the separation properties of the membranes. The ultrafiltration tests were carried out under the following pressures: 0.5, 1.0, 1.5, 2.0, and 2.5 MPa (after steady state conditions had been achieved). Dyes concentrations in the model solutions were measured colorimetrically at respective wavelenghts, Amax, (Table 1) using a Carl-Zeiss-Jena Spectrophotometer.
Tab. 1. Characterization of the selected dyes used in the ultrafiltration tests Dye
Molecular weight
Amax, nm Symbol of dye
Direct Red 79 (C.I.* 29065) Direct Green 81 (C.I.* 30315) Reactive Yellow Eriochrome Black T Acide Orange 52 (C.I.* 13025) Methyl Red
1048.2 878.1 593.6 461.0 327.0 291.0
555 390 430 560 485 495
* Colour Index
DR DG
RY EB A0 MR
390
100
-
8060-
4020-
c
= 100
0 C .-
.-E
------
80-
d
a,
604020.
303 3i8 333 348 temperature of casting solution, K Fig. 2 . Elimination coefficient versus temperature of casting solution plots for various dyes: DR (1); DG ( 2 ) ; RY (3); EB (4); A 0 ( 5 ) ; and MR (6). Measurements were performed at 0.5 MPa (A); 1.5 MPa (B) and 2.5 MPa (C) for 65 pm thick membranes.
3. RESULTS AND DISCUSSION 3.1. Effect of the casting solution temperature on the membrane selectivity
The membranes tested had a thickness of 65 pm. It was observed that increasing the temperature of casting solution decreases the elimination coefficient. This tendency becomes more pronounced for the higher pressures applied as well as for the lower molecular weight of the dyes tested (Fig. 2). It is worth noting that the high-molecular-
39 1
40
\s
20 A
c
0 .c
k
20-
B
0 -
.100.-E
A
aJ
80 60
\
-
I
- \
40 20 P
40
80 I00 120 membrane thickness, Dm
60
Fig. 3. Elimination coefficient versus membrane thickness plots for various dyes: DR (1); DG (2); RY (3); EB (4); A 0 (5); and MR (6). Measurements were performed at 0.5 MPa (A); 1.5 MPa (B), and 2.5 MPa (C) for membranes cast from a solution at 318 K.
weight dyes (above 8003 were eliminated in 95-100 percent by all of the membranes, irrespective of the pressure applied, while those having low molecular weight (helow 400) were hardly retained by the membranes, if at all. The obtained results indicate that raising the temperature of the casting solution leads to an increase of the average diameter of the membrane pores. It seems that the mechanism governing the membrane separation in this process can be explained in terms of a porous flow model, in which the membrane is considered as a "molecular sieve". According to this model, particles of high-molecular-weight dyes are assumed to be greater in size than the pore diameter in the membrane. Particle size
392
.a-
u-
a, 0
A
0.5
1.5
0.5
1.5
25 1
Q5
1.5
1.5 p r e s s u r e , ~ p a
25
2.5
Fig. 4. Elimination coefficient versus pressure plots for membranes prepared from casting solutions having various temperatures and for various dyes: 303 K (A); 318 K (B): 333 K (C); and 348 K (D); DR ( 1 ) ; DC (2); R Y (3), EB (4); A 0 (5); and MR ( 6 ) . Membrane thickness: 65 pni.
of low-molecular-weight dyes is considered to be comparable with the pore diameter in the membranes cast from solution having an initial temperature of 303 or 3 18 K, and to be much smaller than the pores occurring in the membranes cast at higher initial temperatures (333 or 348 K) of the polysulfone solution. 3.2. E f f e c t of Membrane Thickness on Separation Properties
This effect was studied by utilising the membranes cast at the same initial temperature of 318 K and six different thicknesses: 50, 60, 70, 8 5 , 100, and 115 pm. It was found that the elimination coefficient tends to decrease with increasing membrane thickness. This effect is enhanced as the pressure acting on the membrane is increased and molecular weight of the dyes decreased (Fig. 3 ) . It should be pointed out that high elinlination coefficients (95-1 00 percent) were obtained for high-molecular-weight dyes with membranes of a thickness equal to, or less than 100 pm, irrespective of the pressure applied. The results also indicate that the membrane thickness affects its pore size which is greater for thicker membranes.
393
A
C
Y I
I
B
I
D
pre ss u r e
, MPa
Fig. 5. Elimination coefficient versus pressure plots for membranes of various thicknesses and for various dyes: 50 pm (A); 60 pm (B); 70 p m (C); 85 p m (D); 100 pin (E): and 115 pm (F); DR (1): DG (2); RY (3); EB (4); A 0 (5); and MR (6). Temperature of casting solution: 318 K.
Assuming that the porous flow model is valid for the process studied we may draw conclusions similar to those presented in the previous paragraph. Particle size of highmolecular-weight dyes is apparently greater than the pore diameters in all of the membranes. Particle size of low-molecular-weight dyes is comparable wirh the membrane pore size when the membrane thickness falls between 50 and 70 pm and is much smaller when the membrane are thicker than 80 pm. 3.3. Relationship Between the Selectivity of the Membranes and the Pressure Acting Upon Them
The dye elimination coefficients obtained for the selected membranes by the pressure
394
Fig. 6. Elimination coefficient versus molecular weight of dye plots for membranes prepared from casting solutions o f various temperatures: 3 0 3 K (1); 3 1 8 K (2), 333 K ( 3 ) ; and 348 K (4). Measurements were performed at a pressure of 2.5 MPa for 65 pni thick membranes.
Fig. 7. Elimination coefficient versus molecular weight of dye plots for selected membranes of various thicknesses: 50 pm (1); 8 5 pni ( 2 ) ; and 115 pm (3). Measurements were performed at a prcssure of 2.5 MPa for membranes prepared from casting solution having a temperature of 3 1 8 K.
tests were plotted as a function of the pressure applied (Fig. 4 and 5). As shown by these plots, the assumed porous flow model is valid for the high-molecular-weight dyes; the elimination coefficients measured for them are essentially constant and pressure-independent. The situation becomes more complexed when the low-molecular-weight dyes (below 400) are tested. The pressure-dependent elimination observed for these dyes makes the applicability of the porous flow model less evident. Some efforts are made to develop another model and to verify it experimentally.
395 3.4. Elimination Characteristics
The elimination characteristics of pblysulfone membranes (i.e., the relationship between the elimination coefficient and the molecular weight of the dye) were plotted on the basis of the experimental results, as well as on the preliminary results previously reported [3]. Some of the characteristics obtained for the selected membranes are given in Figs. 6 and 7. The shape of the curves gives an indication of the pore size distribution in the membrane. Steep plots indicate that difference among pore sizes is less pronounced, i.e., the pore size distribution is comprised within a very narrow range. The elimination characteristics allow to determine the values of the nominal molecular weight limit (NMWL), i.e., the lowest-molecular-weight compound being completely retained by the membrane. The increase in the membrane thickness and in the temperature of the casting solution shifts the elimination characteristics to the higher NMWL values. Depending on tne casting parameters of the membrane tested, the NMWL values obtained are between 700 and 1500.
3.5. General Characteristics of the Polysulfone Membranes
A previous study [3] showed that the transport properties of polysulfone membranes determined for distilled water increase with increasing the casting solution temperature, membrane thickness, and pressure acting upon the membranes. In the present study the membrane permeabilities in relation to the organic dyes solutions (Table 1) were also determined. Thus, the volume flows of the dye solutions (under a pressure of 2.5 MPa) were by some 15 to 25 percent smaller than that of distilled water (obtained under the same pressure) and ranged from 0.05 to 0.20 m3m-'day-' depending on the casting parameters. Considering the results presented in this paycr together with those reported previously [3] it can be stated that the membranes cast from a solution which has an initial temperature of 318 K are the most promising. If their thickness is equal to, or less than, 100 pni, they can yield 95 to 100 percent elimination of high-molecular-weight dyes (above 700). Polysulfone membranes 90 to 100 pin thick exhibit the best transport properties; at a pressure of 2.5 MPa the flow ranges from 0.12 to 0.14 m3m-'day-' [3].
4 . CONCLUSIONS
1 . Polysulfone membranes cast on a glass surface can be used for removing organic dyes from aqueous solutions by ultrafiltration. The membranes are able to retain completely the dyes having a molecular weight higher than 800. 2. The separation properties of polysulfone membranes depend on the casting parameters. When the temperature of the casting solution and the thickness of the membrane increase, the selectivity decreases. The pressure acting on the membrane has no significant effect on the degree of elimination for high-molecular-weight dyes (above 700); in the case of low-molecular-weight dyes, tile increase in the pressure applied decreases the eliniina t ion coefficient .
396 3. Polysulfone membranes cast from a solution which has an initial temperature of 3 18 K are the most promising. If their thickness is equal to, or less than, 100 pm, they can yield 95 to 100 percent elimination of high-molecular-weight dyes (above 700). Polysulfone membranes 90 to 100 pm thick exhibit the best transport properties.
ACKNOWLEDGEMENT
The authors thank the Union Carbide for providing them with samples of polysulfone. The authors also thank the Research Programme ”Environment Protection” for the financial support.
REFERENCES 1 D. Spatz and R. H. Fricdlander, Rating of chemical stability of U.C. RO/UF membrane materials, Water and Sewage Works, 2 (1978), 36-40. 2 I. K. Bansal, Reverse osmosis and ultrafiltration of oily and pulping effluents, Ind. Wastes, 5 (1977), 32-37. 3 K . Majewska, T. Winnicki and J. Wihiewski, Ultrafiltration of dyes by polysulfone membranes, in: L. Pawlowski (Ed.), Studies in Environmental Science 19, Proc. 3rd Int. Conf. Physicochemical Methods for Water and Wastewater Treatment, Lublin, Poland, Sept. 14-19, 1981, Elsevier Sc. Publ. Co., Amsterdam and New York, 1982, pp. 321-331. 4 1. W. Kocnst and E. Mitchell, Method of casting tubular polysulfone ultrafiltration membranes in sand modules, U.S. Patent 4,038,351, 26 July 1977.
CHAPTER VI
PHYSICO-CHEMICAL TREA TMENT: OXIDA TION
This Page Intentionally Left Blank
399
INTENSIFICATION OF THE OXYGENATION PROCESS WITH THE USE OF POLYMERIC DIFFUSERS
ALOJZY PORANEK
River State University of Science and Technology, Port Harcourt, Ngeria ALICJA MIKAGIBALA
Environment Protection Engineering Institute, Technical University o f Wroclaw, Poland ABSTRACT A new method of manufacturing plastic diffusers is discussed. Thrce po\vdered polymers, polymethyl methacrylate (PMMA), polystyrenc PS, and polyethylene (I’E 110 LD), \\ere chosen for the sintering process in order to obtain tabular diffusers. Pressure drop, oxygenation capacity (OC) and oxygen absorption efficiency (OAE) were tested and compared \\ it11 the same properties of two commercially available diffusers, HKP 600 and Brandol 600. Diffusers madc of PS were tested in pilot-plant experiments.
1. INTRODUCTION
Oxygenation plays an important role in the treatment o f municipal sewage, industrial wastewaters, and drinking water (in an ozonization process). This is a unit process in which oxygen is subject to diffusion through the bubble-liquid interfacial surface. Uptake of oxygen by a liquid phase, e.g. water o r wastewater, depends directly on the interfacial surface area and time of contact between air bubbles and the liquid phase. Both the factors are related to the diameter of air bubbles. For example, a diffuser generating 2 mm diameter bubbles makes for the interfacial surface area six times as large as a diffuser generating bubbles of a 1 2 mm diameter. Small diameter air bubbles ascent in the liquid at a slow rate, thus bringing about a longer contact time. Oxygenation involving small-diameter bubbles has become widely used throughout the world. A detailed account of various air diffusers and oxygenation systems has been given by Morgan and Bewtra [ 11 and Pasveer and Sweeris [ 2 ] . In this paper, the application of some porous plastic materials to the design of air diffusers and their basic properties such as oxygenation capacity (OC), oxygen absorption efficiency (OAE), and pressure losses are discussed. The study presented here has been initiated and carried out at the Institute of Environment Protection Engineering, Technical University of Wroclaw, with the aim to developed a technology of air diffuser production on the basis of porous plastic materials [3, 41.
400 2. PREPARATION O F PLASTIC DIFFUSERS
Four fundamental demands are usually made upon diffusers: high porosity, small pore diameter, low resistance to flow and good mechanical properties. Although there are a number of porous materials, polymers seem t o have tlie most hopeful possibilities. Four our eyperiments chosen were three polymeric materials: pol) methyl metliacrylate (PMMA). polystyrene (PS) - both prepared by a domestic form - and polyetliylene (PE 110 LD) made by Plastic Coating Company. The experimental diffusers were tubular in shape and had been sintered in a steel form (Fig. 1). A certain portion of tlie
Fig. 1. Steel form for the sintering of diffusers: 1 tered polymer, 6 - bottom of tlic form.
-
screw, 2 - cover, 3
-
core. 4 - female, 5
-
sin-
pc,lymer was placed in such a form, which was shaken so as to achieve the apparent density rcquired. The form was thenheated in an oven with automatic temperature control. After completion of the sintering process, tlie steel form was cooled for 20 minutes at room temperature and dismantled t o ren:ove the diffuser. The experimental diffusers thus obtained had a length of 0.50 ni, an outside diameter o f 0.07 ni and an internal diameter o f 0.05 in. The sintering parameters depended on the type of the polymer. They are listed in Table 1. Process temperature may vary from glassy to breakdown. Sintering temperature and time are strongly interrelated and they may be changed in the ranges shown in Table 1. A detailed account o f the technological parameters for sintering powdered polymers is given in References [3, 4, 6 and 71.
Tab. 1. Sintcring Parameters ~~
~
~~~
~~~
Type of polymer
Apparent density kg/m3
Average pore size 10.' ni
Sintering temperature K
Sint cr ing time
PMMA PS PE 1 1 0 L D
800 800 500
40 48 28
423-453 398-420 383-398
48 48 4
11
40 1 3. BASIC PROPERTIES O F THE DIFFUSERS AND METHODS OF DETERMINING THEM
A further step was to determine the resistance t o air flow. tlie oxygenation capacity (OC) and the oxygen absorption efficiency (OAE). Pressure losses were measured for wetted and not wetted diffusers. The pressure tneasuring niethod is shown schematically in Fig. 3. Pressure drop for the rate o f air flow through the diffuser between 0 and 20 m 3 / h was read twice: once with no diffuser mounted t o the measuring set. and once after attachement of the diffuser. The difference in pressure drop with and without diffuser for these same air flow rates was assumed t o be tlie pressure loss of tlie diffuser.
Fig. 2. Mcasurement of pressure loss o f diffusers: 1 - manometer.
-
compressor, 2
-
rotamcter, 3
-
diffuser,
4
The Pasveer method [ 3 , 51 was used to measure OC. The experiments were run in a 1 m diameter aeration tank filled up with 1 ni3 of tap water. The experimental tube was installed at a submergence depth of 1 in. For this study the water was deoxygenated by adding sodium sulphite in solution with cobalt sulphate as catalyst in an amount o f g C 0 + ~ / n 1 ~Increase . of oxygen concentration in tlie deoxygenated water was measured for various air volume flow through the diffusers. The CO and OAE values were calculated on the basis of measured results. Hence,
OC = 36.1
~
1 t, - t i
log
cS
- cO
~
C, - C t
’
4-K283 KT
oxygenation capacity, g 0, /m3 11 time, 11 diffusion constants oxygen concentration in water after deoxygenation, g 0, /m3 oxygen concentration in water at saturated condition, g 0 2 / m 3 oxygen concentration in water at measured time, g 0,/ni3
402 The OAE value was calculated as a percentage ratio of oxygen absorbed in water t o total oxygen passed through the diffuser. The experiments were run for unfractionated pulverized PMMA, PS, PE 110 LD, and for HKP 600 (OY Nokia Company, Finland) and Brandol 60 (Schumacher Company, FRG) commercially available tubes, 0.50 m long, with a diameter of 0.067/0.050 m and 0.070/0.040 m for HKP 600 and Brandel 60, respectively.
4. PILOT PLANT TESTING
The diffusers were tested in the biological system of a small wastewater treatment plant involving a volume flow rate of 70 m3/d (Fig. 3). The total volume of the aeration
excess sludge treated wastewaters Fig. 3. Schcme of pilot plant: 1 - crusher, 2 - aeration tank, 3 - secondary clarifiers, 4 contact chamber, 5 - sludge stabilizing tank.
-
chlorine
tank was 11.2 m 3 ( 4 x 0.8 x 3.5 m), whereas the volume of wastewater treated in the tank amounted to 8.5 m 3 . In this tank were installed six tubular diffusers (0.07 x 0.05 x 0,50 m) made of unfractionated PS. The submergence depth was 3 m. Like laboratory investigations, these tests involved tap water to fill the tank and the Pasveer method to measure the OAE for different values of air flow through the diffusers. The OAE values for that kind of diffusers are gathered in Table 2 . Tab. 2. Oxygen Absorption Efficiency of Diffusers Made of Unfractionated PS at a Submcrgence Depth of 3 m (in water) Rate of air flow through the diffuser Nm’/h 4 7
10
Oxygen absorption efficiency 74
13.7 12.5 13.0
The pilot-scale experiments aimed at answering the question whether or not PS diffusers are applicable to the aeration of municipal sewage in a biological treatment plant. Measurements were carried out over a period of three months. During that tiine the work of this type of diffusers was found to be good. The technological parameters of the pilot plant are given in Table 3 .
403 Tab. 3. Technological Parameters of Biological Pilot Plant Avcrage wastewater flow rate m3/d
Air flow rate Nm3/li
Sludge BOD, of loading wastekgBOD, /dry water sludge g o,/n13
Oxygen concentrationin tank
BOD, treated wastes
go,/n13
go,/n13
Average decrease of BOD %
55-60
55 -65
0.4-0.5
2-3.5
18-20
80-90
150-500
5. RESULTS AND DISCUSSION
Experiments were conducted for air diffusers made of unfractionated powdered polymers. For comparison, some experimental series were run with HKP 600 and Brandol 60 tubes. Resistance to air flow was measured for each of the diffusers under study, both ‘dry’ and submerged (Figs. 4 and 5). The values of pressure loss in submerged tubes d o not include the hydrostatic pressure. Of the air diffusers studied, PMMA tubes, both submerged and dry, exhibited the highest pressure losses. This could be expected because, compared to PS and PE 110 LD, powdered PMMA sliows a higher hydrophilicity and, eventually, a better water absorption capacity. The lowest pressure loss was measured in dry Brandol 60 tubes, which is due t o the greater average pore diameter ( I 80 x m according to suppliers’ technical literature). I11 submerged Brandol 60 tubes, the increment in pressure loss with the increasing air flow rate is more rapid tlian in tlie remaining diffusers under test. Again, this can be attributed to the high hydrophilicity o f the material used. Pressure losses in dry PS and PMMA diffusers have more or less the same values. This is also true for PE 110 LD and HKP 600 tubes. No significant difference was observed in tlie pressure loss b e m e n submerged PS and HKP 600 diffusers (Fig. 5). Submerged PE 110 LD tubes show the lowest pressure losses. They are due to the hetero-
hPa
10
5
0 4
8
12
16
2 0 Nm3/h
Fig. 4. Pressure loss of not \vetted diffusers: 1 - unfractionatcd PMMA, linc 2 - unfractionatcd PS, linc 3 - unfractionated PE, line 4 HKP 600, linc 5 - Brandol 60, 70/40. ~
404
h Pa
50 40
30 20
4
8
12
16
2 0 Nmsh
Fig. 5. Pressurc loss of wetted diffusers: line 1 - unfractionated PMMA, line 2 - unfractionated PS, line 3 - unfractionated PE, line 4 - HKP 600, line 5 - Brandol60, 70/40.
geneous structure of the diffusers, which should be attributed to some manufacture problems. OC values are shown in Fig. 6 Figure 7 presents the relationship between OAE and the rate of air flow through a tube mounted in the tank at a submergence depth of 1 ni. The highest OAE and OC values were obtained for HKP 600 and PMMA diffusers. This is to be attributed t o the properties of PMMA and probably t o the small pore diameters in the HKP 600 tubes (in the suppliers’ technical literature no information is given about the average pore diameter). While OC and OAE obtained for PS and Brandol 60 diffusers had approximately the same values, the values o f these properties for PE 110 LD tubes were very low because of the inhomogeneity o f their structure, whicli is due to some manufacture problems. Pilot-scale experiments involved PS diffusers which are characterized by low pressure losses and relatively good oxygenation properties. The OAE values measured for this kind of diffusers are given in Table 3. As shown by this table, OAE values obtained under pilot-scale conditions are proportionally higher than those determined by laboratory experinents, which could be expected, considering the difference in the submergence depth. Another behavioral feature o f the OAE values is that they d o not decrease when the rate o f air flow through the diffuser increases. This is likely to be due to the shape of the aeration tank and a better mixing of water at higher air flow rates. Under pilot plant conditions, unfractionated PS diffusers exhibited good oxygenation properties and enabled oxygen concentration in the aeration tank to be kept on the level desired throughout the experimental run. Although the duration of the pilot-scale experiments was too short t o substantiate the usability o f the air diffusers in the aeration o f municipal sewage, the results obtained so far (Table 2) seem to be promising. Hence,
405
250 L
YE
200
21 5 0 UI
0 0
100
50
2
r,
6
8
1 0 Nm3/h
Fig. 6 . Oxygenation capacity of diffusers at a submergence depth of 1 m (in water): line 1 - unfractionated PMMA, line 2 - unfractionated PS, line 3 - unfractionated PE, line 4 - HKP 600, line 5 Brandol60, 70/40.
2
4
6
8
10 Ndlh
Fig. 7. Oxygen absorption efficiency in terms of iar tloa pcr dii!'i:w:r ;it a subiiicrpencc depth of 1 ni (in vater): line 1 - unfractionatcd PMMA, line 2 - unfraction;ited I'S, linc 3 - unfractionated PE, line 4 HKP 600, line 5 - Brandol 60, 70/40. ~
406
there is hope that further investigations will support these expectations. It is interesting to note that PS and PMMA are found to be highly resistant to the action of oxygen and ozone. Having these in mind, there is no doubt that they may be successfully applied to oxygenation processes involving pure oxygen or ozone.
6. CONCLUSIONS
1. Of the three materials tested (PMMA, P S and PE 110 LD), P S is best suited for the manufacture of air diffusers, because this polymer accounts both for the low pressure loss and for the good oxygenation capacity of the diffuser. 2. Diffusers made of polystyrene exhibit good oxygenation properties and enable the technological parameters of the biological treatment process to be maintained on the levels desired. 3. Diffusers made of polystyrene or polymethyl methacrylate have the inherent advantage that they are highly resistant to the action of oxygen and ozone. It can therefore be expected that they will be successfully applied to the oxygenation of water or wastewater employing either of the two oxygenizing agents.
REFERENCES
1 . P. F. Morgan and J . K. Bewtra,JWPCF, 10 (1960) 1047-1059. 2. A . Pasveer and S: Sweeris, JWPCF, 9 (1965) 1267-1274. 3. A. Mika-Gibala, A . Poranek, T. Winnicki, et al., Report of the Inst. of Envir. Prot. Engng, Technical Univ. of Wroctaw, PWr 1-15/K-333/78 (1978). 4 . A . Poranek, T. Winnicki, et al., Report of the Inst. of Envir. Prot. Engng, Technical Univ. of Wroctaw, PWr I-15/P-32/79 (1979). 5 . J . Kurbiel, Oznaczanie rzeczywistej zdolnoki urzqdzeri napowietrzajqcych do natleniania, in: V1 Conference ,,Postgp Techniczny w Oczyszczaniu Sciekow”, Arkady, Warszawa, 1979. 6 . A . Mika-Gibala, T . Winnicki, et al., Report of the Inst. of Envir. Prot. Engng, Technical Univ. of Wroclaw, PWr 1-15/K-221/76 (1976). 7. J . Malczewski, A. Mika, Tworzywo spiekane z polichlorku winylu. Otrzymywanie w skali laboratoryjnej i wkakiwoki. Pr. nauk. Inst. Technol. Org. PWr, 1 9 (1975) 3-20.
407
OXIDATION AND COAGULATION OF WASTEWATER EFFLUENT UTILIZING FERRATE (VI) ION
T. D. WAITE and K. A. GRAY Department of Civil Engineering University of Miami, Coral Gables, Florida 331 24
ABSTRACT For the past several years our laboratory has been developing the technology for thc usc of fcrratc (VI), [FeO:-], in water and wastewater treatment schemes. TWOproperties of fcrratc (VI), \vliich arc displayed by its oxidation and coagulation reactions, indicate that ferrate (VI)could be an cffective multi-purpose treatment chemical. Studies evaluating the reaction of fcrratc (VI) with various organic compounds have been reviewed. In addition t o being a selective oxidant of organic compounds, ferrate (VI) has been demonstrated to reduce THM potential and oxidize a model USEPA priority pollutant. Data has been collected under a number of different scenarios to illustrate the effectiveness of ferrate (VI) as a coagulant of turbid water systems, including secondary effluent. Current research is aimed at defining solid phase formation with ferrate ion decay in order to elucidate the mechanism of coagulation in colloidal systems. Comparative studies have been conducted which evaluate the efficiency and effectiveness of ferrous, ferric, and ferrate salts in destabilization of a model colloid.
1. INTRODUCTION
For the past several years our laboratory has been developing the technology for the use of ferrate (VI), [FeO:-], in water and wastewater treatment schemes. Ferrate (VI) has many appealing characteristics including its strong oxidizing potential (Ea = 2.3 V) and concomitant formation of reduced iron species which will act as coagulants of suspended materials. In addition, the solid phase formed, when ferrate (VI) is reduced, is a good scavenger of metals and anions such as phosphate. The above properties indicate that ferrate (VI) could be an effective multi-purpose treatment chemical for water and wastewater. This paper will discuss two properties of ferrate (VI), i.e., oxidation and coagulation reactions, as these properties are the most important for water and wastewater treatment. The disinfection capacity of ferrate (VI) has been described in detail in earlier publications.
2. FERRATE CHEMISTRY
Iron in its familiar forms exists in the +2 and +3 oxidation states; However, in a
strong oxidizing environment it is possible t o obtain higher oxidation states of iron. Compounds o f iron (IV), (V) and (VI) have been isolated as the metal salts of ferric acid, however, ir is the hexalent form of iron t!iat is o f interest for water treatment. This seemingly exotic form of iron has been of interest t o analytical chemists since 1841 when Fremy [ 11 first synthesized potassium ferrate. By 1925 a wide variety of metallic iron (VI) salts had been synthesized. It was not until 1948, however, that procedures were developed whereby a stable, crystalline solid of high purity could be synthesized, and analyzed for its iron (VI) content [ 2 , 3, 4, 51. As a result of work by Sclireyer, physical chemists and kineticists have been able t o establish the structure for iron (VI) ferrate and find evidence t o support its existence [6, 71. Although ferrate chemistry is in a state of infancy, several U.S. patents are currently held that relate to the use of ferrate in aqueous solutions. Three of these patents include: removal of color from industrial electrolytic baths [8]. use in making catalysts for the Fischer-Tropsch process [9, l o ] , and purification of hemicellulose [ 111. Ferrate (VI) ion has the molecular formula FeOi- and is a powerful oxidizing agent through the entire pH range. Wood [7] has reported the redox potential of ferrate to vary from -2.2 V to --0.7 V in acid and hase, respectively. The standard electromotive force for the half reaction is: F e 3 ' + 4 H z O + F e O ~ - + 8 H ' + 3e-.E0=-2.2+0.03V
(1)
Latimer [l?] gives a calculated potential estimate for the reaction Fe(OH)3
+ 5 0 H - - + FeO:-+
4 H z 0 -k 3 e - o f E, =-0.77 L0.03 V
Nearly 30 metallic salts of ferric acid (FeOz-) have beeii prepared, but only a few of these compounds are found t o yield a highly pure and staole product. As a matter of interest, ferrate compounds containing Ag, Al, Zn,Cr, Cu, Co, Pb, Mn. Ni. Hg, or T1 have been synthesized by double decomposition of B a F e 0 4 , and the correspondin6 metal nitrate in aqueous solution [ 131. For example: BaFe04
+ 2A1(N03)
+
Ba(N03)z
+ Alz FeO4
(3)
It is difficult to isolate most of these compounds from solution as they are subject t o decomposition at 3OoC, and react rapidly with COz while being dried in air. Of more practical interest are the metal ferrates which form either stable solutions o r stable crystalling solids. These compounds vary widely in their aqueous solubilities. Lithium, sodium, calcium, and magnesium ferrate are reported to be extremely soluble and can be synthesized by double decomposition with alkali metal perchlorate (MCIO,) and potassium ferrate [14]. Products vary in purity from 15% to 69%. Another procedure developed by Schreyer, et al. [4]. employs wet chemical oxidation of Fe (111) by hypochlorite, follwed by chenlical precipitation of F e O i - with KOH, forming K z FeO, . Recrystallization results in a high purity crystalline solid. This method was utilized in generating the potassium ferrate for experiments in this study. Aqueous solutions of ferrate ion have a characteristic violet color much like that of permanganate. Spectroscopic analysis of visible spectra of aqueous ferrate solutions show
409 one maximum peak at 505 nm and two minima at 390 nm and 670 nm. The molar extinction coefficient as determined by Wood [7] is 1070 *30 in lo-.' M NaOH. Potassium ferrate decomposes in aqueous solution generating hydroxide ion and molecular oxygen. The overall decomposition of ferrate (VI) ion in aqueous medium is described by Equation (4): 2Fe0,2-+ 3 H 2 0 + 2FeO(OH)
+ 3/2 O2 + 4 0 H -
(4)
The decomposition rate is strongly dependent on pH, initial ferrate concentration, temperature, and to some extent on the surface character of the hydrous iron oxide formed upon decomposition. Ferrate is most stable in strong base with two regions of maximum stability, one at pH 10-1 1 and the other in solutions greater than 3 M in base [7], although this is highly dependent on the initial Fe(II1) [IS]. Studies on the stability of ferrate in aqueous solution have shown that dilute solutions of ferrate are more stable than concentrated solutions [16]. Wagner, et al. [17], found 1.9 . M ferrate solutions to be only 37.4% decomposed after three hours and 50 minutes at 25OC. Ferrate decomposition rate has also been found to decline markedly in the presence of phosphate, and at low temperatures [7, 16, 171.
-
20
LO [minute]
30 Time
50
60
Fig. 1. Effect of various salts on ferrate (VI) stability (from Ref. 2).
Figure 1 illustrates the effects of several salts, and ferric hydroxide on ferrate (VI) stability. Other solute domains probably exist in the presence of SO:-, F - and dissolved or colloidal organic matter, and these groups can form stable complexes with Fe(II1) which may also alter the decomposition rate of K2 F e 0 4 . Ferrate reacts rapidly with most inorganic reducing agents under both acid and basic conditions. Reactions involving inorganic ammonia have been studied in detail, and oxida-
410
tion o f ammonia appears t o have a n optimum conversion in the pH range 9.5 t o 11.2, although losses t o the gaseous phase might be suspected at the higher pH. Strong [ 181 reported the degree o f conversion of ammonia t o increase as the molar ratio of ferrate to ammonia became greater, and as temperature was increased. Murmann [ 141 reports a pseudo first order rate constant for ferrate oxidation of NH3 t o be 7 .0 . sec-' at pH 10.6 and 2.5 . lo-* sec-' at pH 9.0.
3. OXIDATION OF ORGANICS
Iron (11) and iron (HI)enter into a wide variety of reactions with organic compounds which can include complexation, chelation, precipitation and oxidation-reduction. Although some work has been completed which evaluates these reactions, and provides some insight into mechanisms, the extent of organo-iron (VI) interactions is still largely unknown. The degree of oxidation of amino acids by FeO: - varies with initial ferrate concentration [ 141. Cystine and glycine react completely with excess ferrate forming C 0 2 and N 2 . When the amino acid is in excess, a variety of oxidation products are generated. Most sugars, and glycol are slowly oxidized t o organic acids. Certain organo-ferrate (VI) reactions have been studied, and in one study oxidation of organics by ferrate (VI) was evaluated by monitoring the disappearance of substrate using gas chromatohraphy [19]. Tests were conducted over a pH range of 2 t o 10.5 at 2OoC. A wide variety of substrate t o ferrate molar ratios were examined, utilizing the following substrates: benzene, allybenzene, chlorobenzene and 1-hexene-4-01. The data were evaluated in terms o f pH dependency, effect of substrate-ferrate molar ratio, and synergistic effects in two substrate systems. Ferrate was found t o significantly reduce the concentrations of allybenzene and chlorobenzene, while benzene and 1-1iexene-4-ol were converted by about 50% t o products. The range of maxima for per cent oxidation of substrates for reactions occuring below pH 8 are shown in Table 1. The oxidations are deTab. 1, Oxiddtion of Organic Substrates by Ferrate (VI) at pH values < 8 (from Ref. 19) Coinpound
% oxidation
Benzene Chlorobenzene Allylbenzene 1-hexene-4-oi
18-47 23-76 85-100 32-55
pendent on S : Fe(V1) molar ratios, where an excess of ferrate is shown to be most effective in reducing substrate concentrations. Molar ratios of s:Fe(V1) greater than 1 : 3 did not significantly enhance conversions. This points to the formation of products such as organic acids, rather than complete oxidation t o C 0 2 . It is possible that more complete oxidation of substrates would be obtained with multi-stage additions of ferrate. Studies evaluating the reaction of ferrate (VI) with phenol have also been undertaken. Variable ratios of ferrate (VI) t o phenol were investigated at variable pH; then, per cent
41 1
103
Secondary effluent FeOZ-added in phosphate buffer *--TOC -609
90
(initial 12 ng/L) [initial 12 8 mg/L)
80
70 0
c
60 al
a
$ 50 40
30
20
Ot
I
L -t
8
I
:2 16 20 24 FeO?dose [mg./L a s Fe:
-
Fig. 2. Oxidation of TOC and BOD in sccondary cfflueiit wit11 ferrate (VI).
removal o f phenol, and COD were determined. The following conclusions were made from tlie study: 1) There is a general increase in ferrate reactivity with increasing pH. 9) Initial ferrate (VI) concentration is an important factor at high ferrate t o phenol ratios. 3) The relative ratio of phenol to phenolate species (C, H, OH/C, H5 0-) regulates reactivity as the ionized species reacts more readily with ferrate (VI). 4) Efficient phenol oxidation occurs when tlie ferrate (VI) to phenol molar ratio is > 10.
412
1
-Unfiltered (TOC = 17 porn) &-Filtered ( T O C = l L ppm) Chiorme 30 ppm PH
'
8.5
Contact Time 4Hrs 25
/
6 20-
/*
c
0
3
D Q)
CK
15-
< O
10-
/r
//
Fig. 3. Oxidation of THM precursors with ferrate (VI).
Ferrate (VI) in concentrations of less than 10 mg/l as Fe is also able to oxidize biodegradable organics (BOD) in domestic secondary effluent. Figure 2 shows data from an experiment where ferrate (VI) was added at different concentrations t o a secondary effluent with a total carbon content of 13 mg/l and a BOD o f 12.8 mg/l. It can be seen that all of tlie biodegradable carbon was oxidized by ferrate (VI), and approximately 35% of the TOC was removed. It should be noted here that no filtration o f the effluent after ferrate (VI) addition was attempted; therefore TOC and BOD removals are due t o oxidation only. It is anticipated that even greater removals could be achieved if the coagulation capacity of ferrate (VI) were taken into account. Preliminary studies have also been undertaken t o evaluate the ability of ferrate (VI) t o oxidize organic precursors o f trihaloniethane (THM). Water samples were collected from the Fox River wliich is located in Northern Illinois (U.S.A.). The water samples which were analyzed averaged 48 turbidity units (TU), pH = 8.5, and TOC = 17 nig/l. The effect of ferrate (VI) o n trihalonietlime potential was examined at several Fe0:- doses. Ferrate was applied 30 nlinutes prior t o chlorination. In all tests a dose of 30 ppm chlorine was used which represented the approximate demand of the raw water. Trihaloniethane concentrations were measured four hours following chlorination. Figure 3 shows a summary of the experimental data. Ferrate (VI) was able to reduce THM potential up t o approximately 25% in this system. The data indicate that no optimum ferrate (VI) did exist, so it is somewhat difficult to interpret the results. It should also be pointed out that the reduction in formation of THM shown here is due entirely to oxidation of precursors. There was no filtration of the samples, thus, no removal due
413
to coagulation was measured, In addition, the amount o f chlorine added was held constant at 30 ppm even though ferrate (VI) was added. Because the ferrate (VI) would normally perform most o f the disinfection, very little chlorine would have t o be added after ferrate (VI) treatment. This would further reduce THM formation in treated secondary effluents. The above data indicate that ferrate (VI) is a selective oxidant of organic compounds, and may have use as an oxidant of toxic organics in waste streams. One recent study has evaluated ferrate (VI) oxidation of one of the 129 priority pollutants listed by the USEPA [20]. Naphthalene was selected in this study as a model compound, and was reacted with ferrate (Vl) at different molar ratios. Table 2 below shows a summary of the data, and it can be seen that ferrate (VI) was an efficient oxidizing agent for the model priority pollutant. Tab. 2. Oxidation of Naphthalene by Ferrate (VI) (from Ref. 20) Perccntage Removal Molar Ratio Perratc/Naphthalenc
N a p h t h a l e n e (nig/L) 100 320 1000
10 20 30 40 60
22 43.5 63.2 15.5
82.2
40 66 16 83.5 95
46 90 100 100 100
4. COAGULATION
The removal of turbidity is also a major objective in water and wastewater treatment. In natural waters turbidity is largely the result of discrete, negatively charged particles and macromolecules which are stabilized by charge repulsion. Coagulation is the process where the surface chemistry of colloids is modified to pernlit aggregation and subsequent removal by gravity settling. The mechanism of coagulation can involve a number of reactions, but in general, two distinct phenomena probably occur: (a) the potential energy of repulsion is reduced; and (b) particles become enmeshed in a precipitate as it is formed [21]. A great deal of research has been conducted o n iron (11) and iron (111) coagulation which has promoted a widely accepted theory describing the probable n~echanismof colloid destabilization. It has been clearly established that salts of iron and alunlinium undergo hydrolysis in aqueous solution. In turn, the resultant aquometal complex will undergo polynierization by successive elimination of coordinated water molecules by Iiydroxide groups. The extent of polymerization is pH dependent. Under conditions uhich exceed the solubility limit of the metal hydroxide, these various polynuclear hydrolysis products may be considered soluble kinetic intermediates in the gradual precipitation of the metal hydroxide [??I. It has been shown that polynuclear hydroxo-metal complexes are readily adsorbed at the liquid-solid interface, and are more effective than non-hydro-
414
lyzed ions in destabilizing colloids. Although these intermediate metal species could be regarded as indifferent electrolytes producing coagulation by double layer compression, this is thought to be less significant than the reduction of zeta potential [21, 231. If similar reasoning is applied to the iron (VI) system, it is plausible that coagulation may involve a greater variety of intermediate hydrolytic species, and possibly, species of greater net, positive charge. It is known that the ferrate (VI) ion rapidly decomposes in acid solution, and its stability increases with increasing pH above 7 [16]. Although the kinetics of Fe5+ and Fe4' formation have never been specifically defined, differences in both the mechanism of precipitation of solid iron and the coagulation behavior between ferrate (VI), iron (IlI), and iron (11) systems would provide a basis for inferring the existence of such intermediate species. Preliminary studies have demonstrated that ferrate (VI) will effectively coagulate turbid water systems. Coagulation jar tests have been carried out on lake water systems to which bentonite clay was added to increase initial turbidity. The results of these tests are presented in Figure 4 which illustrates final turbidity values for a range of ferrate (VI) dose. These data demonstrate a general trend that turbidity removal increases with ferrate (VI) dose to an optimum value of 5 mg/l Fe0:- as Fe. Beyond this minimum, turbidity removal decreases with increasing ferrate dose. When ferrate was added in phosphate buffer the turbidity was reduced by 95%, while for the carbonate buffer and distilled water applications, the turbidity was reduced 79% and 84% respectively. It appears that the presence of phosphate has a positive effect on
1
0
-b
FeO:-odded
in carbonate bLffer
4
-.-
Fee:-added
:n distil,ed w a t e r
!,
Lake W a t e r Bentonite CLay 2Fe04 added in phosphate Duffer
06.1
In
2 05aJ
1
z. 0" 4 -\ L
0
\
ul
-n 0 3 - \ 0
L
~
T__
6
.r-
8
.
7
-
12 16 20 2L FeOZ- dose I r n g I L as F e l
Fig. 4. Coagulation bentonite augmented lake water by ferrate (VI).
r
41 5 ferrate (VI) ability to destabilize colloids. I n order to further investigate this phenomenon, coagulation tests were performed on the lake water and bentonite clay system using the optimum ferrate dose (5 mg/l Fe0;- as Fe) with varying amounts of orthophosphate (0-10 mg/l as P). In the presence of increasing amounts of orthophosphate, turbidity removal improved. The enhanced efficacy of ferrate (VI) in the presence of phosphate may be due to a combination of phenomena; e.g. ( 1 ) the stabilization of the ferrate (VI) ion in the presence of phosphate, due to its chelation of Fe3+,which otherwise accelerates the decomposition of Fe6+; and (2) the formation of a mixed hydroxo-phosphatometal precipitate, which enhances colloid enmeshment or coprecipitation. Comparative coagulation jar tests were carried out on the lake water and bentonite clay system using iron (11) and iron (111) salts. At low coagulant dose, i.e., up to 10 mg/1 as Fe, ferrate and ferrous iron remove turbidity more efficiently than ferric iron. At the optimum ferrate dose of 5 mg/l as Fe, greater turbidity removal is achieved with ferrate than with either ferrous or ferric iron. In order to accomplish the same degree of coagulation, a coagulant dosage greater than 15 mg/l as Fe must be used with both ferrous and ferric iron. Furthermore, at ferrate doses greater than 8 mg/l as Fe, the final turbidities of the coagulated system increase with increasing ferrate dose. This behavior contrasts the trend of increasing turbidity removal with increasing dose at low ferrate concentrations and at all tested doses of iron (11) and iron (111). The ferrate system exhibits either colloid restabilization or the formation of a stable hydrous iron oxide colloid with increasing ferrate concentration. Data collected from a small (10 l/h) bench-scale pilot plant also indicated that ferrate can be an effective chemical for suspended solids removal in tertiary treatment of secondary effluent [24]. The pilot facility included a flash nux reactor, where potassium ferrate, of 90% or greater purity, was added, a flocculation unit, and a sedimentation clarifier. The residence time in the flash mix unit was approximately two minutes, and in the three chambered flocculation unit, approximately 45 minutes. The total sedimentation time in the final clarifier wzs four hours. The plant was operated at steady state
--
t
Ferrate concentratlon
Orng/L
2Q/L
o-.- Cmg/L
6mg/~ 8 mg/L c.....1.0 mg/L c..-
'\.
v)
U
L C
-H!%lVG
Fig. 5. Coagulation of secondary effluent in a pilot treatment plant.
416 throughout testing. Figure 5 shows suspended solids removal through the bench-scale pilot plant at various ferrate doses. Despite the variability of secondary effluent quality, the data demonstrate that the pilot system operated optimally at a ferrate dose of 8 mg/l with approximately 86% suspended solids removal. Better than 80% solids removal was observed in all tests utilizing ferrate in concentrations greater than 6 iiigil. However, higher ferrate dosages seemed t o contribute to the turbidity of the effluent stream. This turbidity was not measurable in the solids deternunation, and was obviously due to the formation of colloidal iron from ferrate addition. The behavior of ferrate in water systems with a wide array of organic and inorganic constituents is extremely complex, as a combination of oxidation, coagulation, and precipitation reactions occur. In order t o focus singly on ferrate’s coagulative capabilities. current research is investigating the behavior of ferrate in a defined colloidal system, buffered at various pH and maintained at constant ionic strength. The objective of this new research is t o consider the operative behavior of ferrate (Vl) decay in coagulation in contrast to t h e behavior of iron (11) and iron (111) salts. Since, in practice, the solubility limit o f metal hydroxide is normally exceeded by the dosage of metal salts required t o destabilize colloids, careful consideration has been given to the kinetics of hydrous iron oxide generation for iron (VI), iron (HI), and iron (11) salts. Therefore, an approach has been developed to consider the decay of the ferrate ion, and the precipitation of the resultant metal hydroxide in comparison t o the iron precipitation reactions of iron (11) and iron (111) salts. F o r equivalent dosages of iron, the rate of turbidity generation is believed to reflect the duration of soluble iron species, intermediate t o the formation of iron hydroxide. The formation of hydrous iron oxide with time, as measured by light scattering, has been studied at a variety of iron salt doses and pH. The appropriate dose of iron was added t o one liter of biocarbonate buffer adjusted t o the specified pH and ionic strength. After disper;ing the dose throughout the volume, a small aliquot was transferred to a sealed vial where iron precipitation and sedimentation was followed under quiescent conditions. Data from experiments conducted for potassium ferrate. ferrous sulfate, and ferric nitrate at doses of 15 mg/l as Fe are presented in Figure 6. These results reflect the rate o f generation of an insoluble phase in aqueous solution, buffered at pH 7. If the turbidity profile of iron (111) is regarded as a standard illustration of the hydrolysis, olation, and precipitation of the ferric ion, then comparison between this and profiles similarly developed for iron (11) and iron (111) will demonstrate any differences in the reaction kinetics of iron hydrolysis. Although turbidity measured by light scattering represents the net result of solid iron formation, coagulation, and sedimentation, the formation of solid iron is considered to be the predominant mechanism in iron (11) and iron (Vl) systems within the first few minutes o f turbidity monitoring. Figure 6 shows that a ferric salt generated a maximum m u n t of turbidity within 30 seconds o f the salt addition. Flocs were visible immediately, and throughout the rest of the analysis, turbidity was seen to decrease, due to the flocculation and settling o f the hydrous iron oxide. The turbidity profile of the same dose of ferrous salt in the buffered aqueous system describes behavior different from that o f the iron (111) system. The Fe” must undergo oxidation to Fe3+ prior to hydrolysis, polymerization, and precipitation of an insoluble iron. Under atmospheric conditions at pH = 7 significant oxidation o f iron (11) takes approximately 20 minutes [25]. However, t h e turbidity profile in Figure 6 may reflect t h e competitive for-
41 7
15mg/L a s Fe pH 17
50 -
0-
x-0-.-
Fe(NO3I3 Fe S O 4 . 7H20 K 2Fe O4
-z 3 5 x
c
30-
i
+? 3
/
I-
25. 20
-
0
i
/
5
10
15
rime [ m i n u t e ] Fig. 6. Formation of insoluble phase by different iron salts.
mation of a ferrous carbonate complex due to the use of a sodium bicarbonate buffer. Visible flocs were not observed for the first 60 minutes of monitoring, although the color of the system deepended with time. The iron (11) system generated an iron colloid which was stable for approximately two hours, at which time fine flocs became visible and rapidly settled. The ferrate (VI) system demonstrated turbidity formation distinct from both the iron (11) and iron (111) systems. Decomposition of the ferrate ion occurred within the first 3.5 minutes, and fine flocs were observed after the first minute of monitoring. Initially, turbidity increases at a rate slightly greater than that of the ferrous system, but this rate slows after two minutes and an iron colloid is formed. This colloid was observed to be stable for approximately 90 minutes before significant settling occurred. Figure 7 reports data collected from colloid destabilization experiments. A colloidal silica suspension was developed, buffered at pH 7, and maintained at constant ionic
418
\
\
\
/
-s
100
- 90 a,
c
5 80
E"
f
\
\
70
\
1 ,
c
{
60
f ?
50
40 30 20 10
t
10
20
LO 30 Time [ m l n u t e l
50
60
Fig. 7. Destabilization of a silica colloid by different iron salts.
strength. Colloidal destabilization was determined for each iron salt at 15 mg/l as F e and measured as the amount of turbidity remaining in the systems at various times. After the addition of each salt, t o one liter of colloid suspension stabilized over 24 hours, the system was rapidly mixed for one minute, slowly flocculated for 30 minutes, and allowed t o settle for 30 minutes. Ferrate (VI) achieved the greatest amount of turbidity removal, 89% after 30 minutes of settling, while ferrous and ferric salts produced 76% and 58% removal, respectively. It is thought that ferrate accomplished the largest degree of colloid destabilization because o f the greater coagulating efficiency o f soluble species intermediate to its decay t o Fe3'. The 20% decrease in turbidity within one minute of ferrate addition may be due t o the surface activity of these species. After the initial flash mixing, turbidity increased t o a constant level throughout flocculation. This increase reflected the growth o f metal colloid flocs. The iron (11) system, which generated a great deal of turbidity in the ab-
419 sence of the silica colloid, showed a very small, initial decrease,in turbidity and a large increase in turbidity with flocculation. Throughout flocculation very large flocs were visible. The iron (111) system, which had demonstrated immediate hydrous iron oxide formation and gradual settling, behaved similarly under colloidal destabilizing conditions. There was no increase in turbidity with flocculation, and the least amount of silica turbidity was removed. In fact, only in the system destabilized with iron (111) was the white silica colloid still visible after 30 minutes of settling. In conclusion, both ferrate (VI) and iron (11) were better coagulants than iron (111) for this test system. This may be due to the duration of soluble species and the lesser rate of solid iron generation. Also, ferrate (VI) was observed to remove more turbidity from this model colloid system than iron (11). There may have been some diminishment in the coagulation ability of the ferrous salt due to the interference of the bicarbonate buffer. However, most of the iron (11) is believed to have been oxidized to iron (111) under the conditions of this experiment. It one considers the behavior of iron (11) and iron (VI) in buffered aqueous systems, it may be that the stability of the solid iron generated by iron (11) and iron (VI) is mediated by the variety of soluble iron species intermediate to the transition to ferric hydroxide. It may be those differences in the nature of the soluble species generated by ferrate and ferrous salts that also mediates coagulation efficiency. It is also interesting to note that even though a large amount of turbidity was generated by the iron (11) system it did not effectively coagulate the silica colloid. These observations support the theory that the most efficient mechanism of coagulation proceeds via adsorption of soluble aquometal species at the colloid surface.
REFERENCES 1 E. F. Freniy, Coiiipt. Rend. 12, 23, 1841. 2 J. M. Schreyer, L. T. Ockernian and G. W. Thompson, Anal. Chcm., 22, 1950, 691. 3 Inorganic Synthesis, 4, 1953, 164. 4 J. M. Schreyer, et. al., Anal. Cheni., 22, 1951, 1426. 5 J. M. Schreyer, et. al., Jour. Anier. Cheni. SOC., 73, 1951, 1379. 6 H. J. Hrostowski and A. B. Scott, Mourn. Chcm. Phys., 18, 1950, 105. 7 R. H. Wood, The Heat, Free Energy and Entrophy of the Ferrate (V1) Ion, Journ. Amcr. Chcm. SOC.,80, 1958, 2038. 8 J. M. Schreyer, USP, 2,536,703, 1951. 9 M. A. Mosesnian, USP 2,470,784, 1949. 10 M. A. Mosesnian, USP 2,455,696, 1948. 11 J. B. Harrison, USP 2,728,695, 1965. 12 W. M. Latinier, Oxidation Potentials, Prentice Hall, N.Y., 1952. 13 L. Lozana, Acido ferric0 e ferrati (VI), Gazz. Chim. Ital., 55, 1925, 468. 14 R. K. Murmann, The Preparation and Oxidation Propcrtics of Fcrrate (FeO;.), NTIS Publication PB-2 3 8-057, 1974. 15 R. G . Hairc, A Study of thc Deconiposition of Potassium Ferratc (Vl) in Aqueous Solution, Doctoral Abstracts, 1965. 16 J. M. Schreyer and L. T. Ockernim, Stability of Ferratc (Vl) Ion in Aqueous Solution, Anal. Cliem., 24, 1950, 1498. 17 W. F. Wagner, J . K. Gump and E. N. Hart, Factors Affecting the Stability of Aqueous Potassium Ferrate (V1) Solutions, Anal. Chcm., 24, 1952, 1397.
420 18 A. W. Strong, An Exploratory Work on the Oxidation of Ammonia by Potassium Ferrate (VI), NTIS Publication PB 231873,1973. 19 T. D. Waite and M. Gilbert, Oxidative Destruction of Phenol and Other Organic Residuals by Iron (VI) Ferrate, J. Water Poll. Contrl. Fed., 1978, p. 543. 20 S. Deluca, A. C. Chao and C. Smallwood, Removal of Selected Pollutants with Potassium Ferrate, Proc. of 13th Mid-Atlantic Conf. on Indust. Waste, ed. by C. P. Huang, Ann Arbor Sci. Pub., 1981. 21 C. O’Melia and W. Stumm, Aggregation of Silica Dispersions by Iron (111), J. Colloid and Interface Sci., 23, 1967, 437. 22 C. R. O’Melia, “Coagulation and Flocculation”, in Physicochemical Processes, W. J. Weber, Jr., New York: Wiley-Interscience, 1972. 23 G. R. Wiese and T. W. Healy, Adsorption of AI(II1) at the TiO, -H,O Interface, Jour. Colloid and Interface Sci., 51, 1974, 434. 24 T. D. Waite, Feasibility of Wastewater Treatment with Ferrate, ASCE J. of Environmental Engg. Div., 105, 1979, 1023. 25 J. T. O’Connor, Iron and Manganese, in Water Quality Treatment, AWWA, New York, McGrawHill, 1971.
421
THERMODYNAMIC CRITERIA FOR THE OPTXMIZATION OF IRON AND MANGANESE REMOVALS IN EXPERIMENTAL PROCEDURES
E. M. GROCHULSKA-SEGAL, M. M. SOZANSKI
Institute of Environment Protection Engineering, Technical University of Wroclaw, SO-370 Wroclaw, Poland
ABSTRACT Presented are methods of determining an optimum technology for the processes of iron and manganese removals. The methods do not involve mathematical models of the processes under study. They take into account not only some relationships determined by electrochemical equilibrium curves redox potential-pH for Fe-H,O and Mn-H,O systems, but also the criterion of maximum decrease. Such methods may be of utility in the planning of experiments, as well as in the process design.
1. INTRODUCTION
The technological processes in which iron and manganese are removed have the disadvantage of being complex and troublesome. This is because the water to be treated contains a great number of substances (such as Fez', Mn", H z C 0 3 , HCO;, Cog, SO,, C1-, OH-, Ca2+, Mg2+, Fe(OH), , Mn(OH)z, to name just a few), which react with one another during the treatment process. A separate study of these unit reactions (both chemical and electrochemical) is insufficient to enable either the identification or the optimization of the complex treatment technology. However, recently the application of electrochemical thermodynamic methods has made it possible to describe the mechanism governing the processes of interest. In these methods the direction of the reactions can be predicted in terms of the redox potential Eh and pH of the water. These two parameters indicate the content of the elementary particles that are active in the chemical and electrochemical reactions under study. Despite its great usefulness in describing the mechanisms of iron and manganese removals, the redox potential has been inadequately appreciated so far in the planning of the experiment and in process design. In the studies reported in this paper, the redox potential was used for determining the optimum technology of iron and manganese removals. 2. GENERAL REQUIREMENTS FOR THE OXIDATION OF IRON AND MANGANESE
The equilibrium of the oxidation reactions for iron and manganese may be represented
422
2
0
2
4
6
8
10
12
1416
Fig. 1. Iron-water (A) and manganese-water (B) systems at 25°C: a, b = lines describing the thermodynamic zone of water stability; full lines indicate the equilibrium state for substances on both sides; numben in circles refer to respective equations in the quoted literature: 0, -2, -4 and -6 denote the and mol d m - ) , activity of the metals of interest in a dissolved form, which is lo", respectively.
423 graphically in the Eh = f(pH) system. In this way we can determine the regions of relative predominance for ions or particles occurring in simple (as [Fe”] S [Fe3’]) and complex forms, as well as the regions for the relative thermodynanuc stability of the solid phase e.g. Fe(OH),, M n 0 2 . The plots describing the relationships among Eh, pH, activity of ions (predominantly of Fe”, Mn”, Fe3+,Mn3+,Mn4’) and the solid phase generated by them are given in Fig. 1 [ I ] . From these plots it is seen that: for pH < 2 and Eh > t0.771 V iron occurs in the form of Fe3+, - i n ground waters which usually have a pH lower than 7.0 and a redox potential between -0.440 and +0.771 V, iron takes the form of Fe2’; - the conditions for the conversion of the bivalent form of iron Fe” to the sparingly soluble compound Fe(OH), are given by the equations -
Eht, = 0.271 -0.0591 pHt, Ehtw = 1.057-0.1773 pHt,
(1) -
0.0591 log [Fe2’];
(3
- in natural waters manganese appears in the form of Mn” ions, and the region of Mn” occurrence in the pH, Eh coordinate system is significantly larger than that of Fez+ occurrence ; - the oxidation conditions for manganese are described by the equations
Eht, = 1.014-0.0591 pHtw Eht, = 1.443-0.1773 pHt,
(3) -
0.0591 log [Mn”].
(4)
The equilibrium plots presented here determine the iron-pure water and manganesepure water systems. In the presence of pollutants the stability of individual iron and manganese forms will be subject to variations. In ground waters such changes are primarily due to the presence of carbon dioxide and sulphur compounds. The effects of various concentrations of dissolved C 0 2 and sulphur compounds on the Eh = f(pH) relationship are reported in the literature [2]. As shown in this paper, increasing the concentrations of carbonates brings about an increase in the region of thermodynamic stability for FeCO, at the cost of the region for Fe(OH),. This phenomenon, however, is disadvantageous in that the oxidation of iron to Fe(OH)3 will be rendered difficult, thus making the technology of iron removal a troublesome operation.
3. CRITERION FOR OPTIMIZING THE PROCESSES OF IRON AND MANGANESE REMOVAL
Optimization criteria are given for the following cases: (1) the cost of reagents required for increasing the redox potential, Eh, by a unit value is identical to the cost of reagents required for increasing pH by a unit value, and ( 2 ) the cost of increasing the redox potential by one umit is alpha times higher than the cost of increasing pH by one unit. The optimum solution to the problem of iron removal is to carry out the process perpendicular to the straight line between Fe(OH)2 and Fe(OH)3, or between Fe” and
424 Fe(OH), . Thus, to determine the direction of the process, it is necessary to formulate the equation o f the line which is perpendicular to the straight line intersecting the point of the pH and redox values for raw water (pH,, and Eh,, , respectively). Hence, while for Equ. (1) the expression to describe the perpendicular line takes the form
for Equ. (2) it becomes
In this way we obtain a set of equations which includes the equation of the equilibrium line and the equation of the line which is perpendicular to the latter. The solution t o this set gives the values of the coordinates of point pHt,, EhtIv, which define the pH and redox potential of the water after treatment, respectively. The points of coordinates pH,,, Eli,, and pHt,, Eht, indicate the optimum direction for the process of iron removal from ground waters. For the case described by Equ. ( l ) , pHt, and Elit, may be written as
pH,,
Eht, =-5.89 .
+ 3.46 . lo-,
+ 0.27
Eh,,,-
(8)
respectively. For the case defined by Equ. ( 2 ) , pHt, and Eht, take the form pHt, =0.9696 pH,,
-
0.1716 Eh,,
+ 3.041.
Ehtw =-0.1719 pH,,
-
lg [Fez’]
1.014.
Ehrw
-
5.73.
(9)
lg [Fe”]
(10)
respectively. The calculated values of pHt, and E h t , are the final points for the process of optimum iron removal. But when the costs of reagents for pH and Eh increment differ from one another, we cannot fail to take into account the following equation
Thus, we obtain a set consisting of Equ. (1) and Equ. (1 1). The solution t o this set becomes
a pH,, PHtw =
u -5.91
Eht, =
-
+ 0.271
Eh,,
+ 5.91 . . lo-*
(Y
pHrw (Y
+ 5.91 . + 5.91 .
Ehrw
-
1 . 6 . lo-’
+ 0.271
(13)
425 The solution to the set incorporating Eyu. (2) takes the form
a pHrw - E h , ,
-
PHt, =
Eht, =
5.91 . lo-’ lg [Fez’]
+ 1.057
a + 0.1773
+ 0.1773 Eh,, + 1.048. lo-’
-0.1773 a pH,,
a
- 5.91 .
lg [Fe”]
+ 0.1773
lg [Fe2’]- 0.1874 -
+ 1.057
(15)
The final point pHtw, Eht, obtained via the above route lies along the optimum direction of the process of iron removal, adequate to the difference in the costs of reagents between pH and Eh increments. Like iron removal, the process of removing manganese involves a set of equations to describe the optimum directions. They are the following: pHt, = 0.9965 pH,,. Eht,, =-5.89
-
. lo-’ pH,,
pHt, = 0.9696 pH,, Eht, = 4 . 1 7 1 9 pH,,
-
+ 5.944
5.862 . lo-’ Eh,,,
+ 3.46 .
0.1716 Eh,,
+ 3.042. lo-’
Eh,, -
1.014.
Eh,,
-
lo-’
(16)
+ 1.0105
(17)
*
lg [Mn2’]
+ 0.2476
5.73. lo-’ lg [Mn”]
+ 1.399
(18)
(19)
As shown by the relations presented in this paper, the optimum directions for iron and manganese removals are identical. There is only a shift in the optimum direction at the beginning of the process of manganese removal, as compared to the removal of iron. Thus, ApH = 1.014 . lo-’ (log [Fez’] AEh = 5.73 . lo-’ (log [Fez’]
~
-
log [Mn”])
log [Mn”])
+ 6.62 . l o - * , and
+ 0.3742.
(70)
(21)
Note that Eyus. (20) and (21) were determined for a = 1.
4. SUMMARIZING COMMENTS
The theoretical Eh = f(pH) plots represent ideal thermodynamic conditions, and they are of utility only in approximate estimations of the Eh and pH values defining the requirements for iron and manganese removals. It follows that the derived optimum directions for the processes of iron manganese removals are also approximate. To give a full answer to the question of whether or not the knowledge of the redox potential enables a better understanding of the mechanisms governing the removals of iron and manganese (and, consequently, permits the optimization of the processes), it is necessary to develop
426
accurate and simple methods of measurement. This is likewise true for the redox potential measurements in the aquiferous layer, i.e., before the ground waters enter the surface of the ground. Adequate methods of determining the optimum processes of iron and manganese removals are of particular significance, when high-cost and troublesome treatment technologies have t o be employed. This is so when iron and manganese occur in combinations with organic substances.
REFERENCES
1 M . Pourbaix, Electrochemical Corrosion: Lectures, PWN, Warsaw, 1978 (Polish translation). 2 J. D. Hem, Equilibrium chemistry of iron in ground water, Principles and Applications of Water Chemistry, John Wiley, New York, 1967.
427
IDENTIFICATION OF TIN FROM TIN-SMELTING REFRACTORY-WASTE AFTER ALKALINE SOLVATION
N. M . SURDIA
Chemistry Department, Institute of Tecflnology,Bandung (ITB), Indonesia J. SUGIJANTO
Directorate of Mineral Resources, .Bandung, Indonesia
ABSTRACT Chrome magnesite refractories arc used in tin smelters. Tin is obtained by reduction of cassiterite with coal in a smelter made of magncsite refractories at around 1100-1400°C. At this high tcmpcrature and with increasing time, tin diffuses into the refractory material. After about a year the tincontent in the refractory waste was about 1076, which was dctermined by X-ray Fluorcsccncy and Atomic Absorption Spectroscopy. The solubility of tin \vas small within a pH range of 3.5-9.0, so the solvation had been tried out in a NaOH solution using some oxidizing agents, e.g. KNO, and NaNO,. The variables examined wcre tlie type of oxidizing agents, concentration of NaOH, and temperature. However, the tin recovered was only 35-45% of the total. So to get more information on the type of tin in tlie refractory material, identification had been carried out by X-ray Diffraction Techniques (XRD) and Scanning Electron Microscopy (SEM).
1. INTRODUCTION
In nature, tin deposits occur as cassiterite, SnO,, which has a tetragonal crystal structure. The pure metal can be obtained by reduction of the tin ore with coal in a smelter made of chrome magnesite refractories at around 1 100-14OO0C, or by reduction with hydrogen at 700-800°C. Within 1 0 months, working at that high temperature, causes 1/3 of the refractories to be erroded, and tin diffuses into the refractory material. At the refractory-waste, one can observe clearly the presence of tin at the surface by Scanning Electron Microscopy. After about a year, the tin-content in the refractory-waste was about lo%, which was determined by wet analysis, X-ray Fluorescence Spectrometry, and Atomic Absorption Spectroscopy. In an attempt for tin-recovery, the refractory-waste was solvated in a sodium hydroxide solution using some oxidizing agents, e.g. KNO, and NaNO,. The variables examined were
428
the type of oxidizing agents, concentration of NaOH, and the temperature, to be able to find the optimum condition of solvation. However, the tin recovered was only 35-45% of the total. Several methods had been tried out in tin-recovery, but the results were not as one would expect. So in order to get more informations about the kind of deposit, some characterizations with X-ray Diffraction Spectrometry and Scanning Electron Miscroscopy were carried out.
2. THEORETICAL BACKGROUND 2.1. Solvation of Tin
Tin as an element has amphoteric properties and it has three allotropic modifications, i.e. face-centered a-tin, which changes at 18°C to tetragonal @in, and at 161°C changes into rhombic, brittle y-tin [l]. Its solubility is low within the range of pH 3.5-9.0. Due to its amplioteric behavior, tin can be dissolved either in strong acids or in strong bases. Tin dissolves readily in warm hydrochloric acid to form stannous chloride. Then Sn(I1) can be oxidized into Sn(IV) in the presence of oxidizing agents like peroxide [ 2 ] . The method of tin solvation in hydrochloric acid has been used to determine yuantitatively tin and the other elements in the refractory sample. For recovery purposes, the use of hydrochloric acid is not favourable, because most metallic ions besides tin are soluble in acid, so that these ions might interfere the purity of tin obtained by the electrolytic process in tin recovery. Instead of hydrochloric acid, an alkaline solution can be used. Normally, virgin tin is not very easily soluble in a caustic solution, so the addition of oxidizing agents will promote solvation and will change tin completely into Sn(1V). The equation for the oxidation of tin by NO; is as follows:
The dissolved tin can exist as Sn0; or as Sn (OH),. Factors influencing the solvation of tin and its compounds are, among others, the concentration of alkali, type and concentration of oxidizing agents, and temperature.
2.2. Elcctrolytic Pioccss of Tin-Recovery [ 3 I
After tin has been dissolved, pure tin can be obtained after electrolysis using iron as anode and stainless steel as cathode, at which tin will be deposited. The equation for recovering the tin is as follows:
-
- ekctrolyds
Sn(0H);
Sn + 2 0 H - + 2 H 2 0 +O,
Tin is passive between pH 3.5-9.0 due to the formation of a tin dioxide layer on the surface, which is blue-black colored. At a pH < 3.5 tin is active and it will dissolve, and
429
at the same time iron will dissolve too. At a pH > 13 iron is passive and the deposition of tin can be controlled. For that reason electrolysis is carried out in alkaline solution. To increase the rate of diffusion of ions toward the cathode, electrolysis is carried out at higher temperatures, e.g. about 65-80°C. Factors influencing the electrolytic process are the current density, and the temperature. The greater the current density and the hgher the temperature, the finer the crystals, and so the better quality of deposit obtained.
2.3. Characterization of the Tin Deposit by Scanning Electron Microscopy 141
One of the methods used in characterizing the deposited tin is by Scanning Electron Miscroscopy (SEM). When a focused electron beam impinges on a specimen surface, elastic and inelastic scattering may occur. In inelastic scattering, the moving electrons lose energy and loosely bound electrons are ejected, forming secondary electrons with energy less than 50 eV. These secondary electrons are very useful in getting topographical informations of a surface. However, with this technique tin and other heavy elements can not be distinguished from the lighter ones, like Ca, Mg, and so on. Elastic scattering on the other hand, have higher energies, and may occur as single or multiple scattering. Multiple scattering may result in a large change of direction of the impinging electron beam. This is the process of backscattering. The total number of electrons backscattered depends on the atomic number of the sample. As the atomic number increases, more electrons will be backscattered, with means that heavier elements will give a brighter image. So backscattered electrons can be used to reveal the chemical composition of the surface (COMPO technique). COMPO technique can not give any qualitative or quantitative informations of the deposit. This can be obtained by the Energy Dispersive Spectrometry (EDS technique), which is based upon the fact that when electrons of sufficiently high energy bombards an atom, they will generate X-rays, which are characteristic of the element bombarded. Scanning Electron Microscopy, however, can not give the kind of compounds formed, which can only be characterized by X-ray diffraction technique. So this technique has to be done too.
3.1:XPERIMENTAL PROCEDURE AND RESULTS 3.1. Analysis of the Refractory-Waste
Qualitative analysis has been done by emission spectrography. The instrument used consisted of a Standard Varisource, Jarrel Ash, Division Serial No. 1514-75-3 Cat. No. 42651, a 1.5 meter Wards worth Grating Spectrograph model 78-090, and a 19-300, 311 Series Arc Spark Stand. The most important elements found were Fe, Mg, Ca, Cr, Sn, Ti, Mn, Zr, Sb and Ni. For the quantitative analysis 2 g of a 200 mesh powder of the refractory waste was mixed with 4 g Na,CO, and 8 g Na,O,. The mixture was fused and later on dissolved in
430 Tab. 1. Quantitative Analysis of Refractory Waste Wet Analysis
Atomic Absorption Spectrometry Percentage of element (%)
Type of oxide SiO ALO3 Fez 0, MgO CaO TiO, H2O Cr20,
2.56 5.31 15.97 57.10 1.67 0.17 0.47 5.73
Percentage of Element Element (%) Mn Ni cu
0.22 0.015 0.002
Tab. 2. The influence of NaOH concentration in the solvation of tin Concentration of NaOH (%)
Sn percentage (with KNO,)
Sn percentage (with NaNO,)
5 10 20 30 40 50
0.34 1.19 1.17 1.21 1.23 1.23
0.40 1.06 1.06 1.09 1.13 1.13
Tab. 3. Tin Analysis after Solvation in NaOH solution by AAS method Temperature of solvation
Percentage of Sn (with KNO,)
Percentage of Sn (with NaNO,)
75°C 80°C 85°C 90°C
2.26 2.45 2.20 2.23
0.83 0.73 0.90 0.83
HCI, and afterwards filtered. The precipitate is treated with HF to dissolve silicates and this solution is mixed with the first one. All elements are determined from the final solution by wet analysis. For Mn, Ni, and Cu, Atomic Absorption Spectrometry was used (see Table 1). The total percentage of elements obtained by wet analysis is 88.98%, and by Atomic Absorption Spectrometry is 0.237%. Assuming that the elements present are totally loo%, then the tin content in the original refractory waste is 100%- (88.98 + 0.237)% = 10.78%. 3.2. Solvation in Sodium Hydroxide Solution
Two grams of sample were dissolved in 250 ml solution. The concentration of solution
43 1
kn03 12
1.o
0.8
0.6
C
W.
0.4
x
0.2
0
10
20 30 40 50 C o n c e n t r a t i o n o f NaOH (%)
Fig. 1 . The influence of NaOH concentration in the solvation process
c
5.0
I
50
60
70
80
temperature ' 9 0 ~ )
Fig. 2. The influence of temperature in the solvation process
100
432 Tab. 4. Expcrimental XRD data of Refractory Material (d values in A) Original refractory
Refractory waste
Wastc after solvation ( + KNO,)
Waste after solvation (+ NaNO,)
2.07 A 2.106 2.31 2.49
1.492 A 1.595 1.731 2.017 2.065 2.433 2.501 2.56 4.18
2.106 A 2.32 2.49 4.67
2.106 A 2.32 2.43 2.49 2.55 4.79
Tab. 5. Theoretical XRD data of tin, tin compounds, and some silicates (d values in A) &tin
SnO,
MgSn(OH),
MgO
NaMg,Cr Si,O,,,
donathite
1.484 1.659 2.017 2.063 2.793 2.915
1.498 1.593 1.765
2.07 2.33 2.44 2.14
1.489 2.106 2.431
2.08 2.30 2.50
2.077 2.086 2.502
was varied and the oxidants used were KNO, and NaNO,, 0.25 g respectively (see Table 2 and Fig. 1). The optimum concentration of NaOH found was 10%. Two grams of sample were dissolved in 250 ml 10% NaOH solution. Then the temperature of the solution was varied (see Table 3 and Fig. 2). The best temperature for KNO, as oxidant was 80-85"C, while for NaNO, there is nearly no influence, so the same temperature range can be taken.
3.3. Analysis of Tin after the Solvation Process
After solvation of refractory waste in NaOH solution, tin is supposed to be dissolved in solution, and so the amount of tin dissolved is determined by AAS techniques and the amount of tin left at the refractory material is determined by X-ray Fluorescence methods. The atomic absorption spectrometer used is AAS model A-66, and the X-ray fluoresceme spectrometer is the Phillips type. The results of tin analysis from solution can be seen at Table 3. From these results one can conclude that KNO, as oxidant is better than NaNO,. Quantitative analysis by X-ray Fluorescency (XRF) is done by the internal standard method with a standard of tin ore that contains 74.8% SnO, or 58.92% Sn. By this technique it was found that the amount of tin not soluble in 10% NaOH solution was 5.82% with KNO, as oxidant and 7.31% with NaNO, as oxidant.
433
Fig. 3. SEM image of the surface of the refractory waste
3.4. Characterization of the Tin Deposit at the Refractory Waste
To get more informations about the type of tin deposit at the refractory material, X-ray Diffraction patterns ( X R D ) were taken from the rerfactory material before and after being used in the smelters. XRD patterns were also taken from the refractory material after the solvation process in 10% NaOH solution. The patterns were made using a Cu tube with K, = 1.542 A. These data were compared with the XRD file data [8,9, 101and the results are given at Tables 4 and 5 . Comparing the experimental d values with the XRD file data shows that tin can exist as 0-tin, SnO,, and MgSn(OH), at the refractory material. However overlap with refractory coniponents may occur. Further on, 0-tin is absent in the refractory waste that has been solvated in NaOH solution, which means that 0-tin is soluble in alkali, but not the other tin compounds. To make sure that tin and its compounds are present in the refractory material, investigation with a scanning electron microscope type Jeol CX-100 was carried out (see Fig. 3 and 4). Fig. 3 is SEM image of the surface of the refractory waste, which however does not give any informations about the presence of tin. Using backscattered electrons, a brighter image of the metal is obtained (Fig. 4). The pictures were taken at a potential of 4 kV and a magnification of 500X.
434
Fig. 4.COMPO technique of the refraetory waste
4:DISCUSSION AND CONCLUSIONS
1. In an attempt for tin-recovery from refractory-waste of tin smelters, the tin adhering at the chrome magnesite refractory has been tried to be dissolved in a NaOH solution. 2. The variables of the solvation process, i.e. concentration of NaOH, type of oxidizing agent, and temperature, has been determined. Two oxidizing agents have been used, i.e. KNO, and NaNO,. It was found that the most suitable condition for solvation was a 10% NaOH solution with KNO, as oxidant at a temperature of 80-85°C. 3 . The elements present in the original refractory material have been determined qualitatively and quantitatively by emission spectrography, wet analysis, and atomic absorption spectroscopy. It was found that the tin content was 10.78%. 4. The tin content at the used refractory was determined by X-ray Fluorescency, and the tin content present in the alkaline solution was determined by atomic absorption spectroscopy. The tin recovered was only 34-45% of the total tin present at the refractory waste. 5. The possible types of compounds formed at the refractory material were examined by X-ray Diffraction Spectrometry, and it was found that tin was present as 0-Sn, SnO,, and MgSn(OH),. 6. To identify the tin at the refractory material, several pictures of its surface were made by Scanning Electron Microscopy (SEM). The composition of the surface could only be identified by using backscattered electrons in the SEM (COMPO technique). The Energy Dispersive Spectrometry (EDS) device could identify the tin present.
43 5 REFERENCES 1 E. S. Hedges (Ed), Tin and its Alloys, Edward Arnold Publ. Ltd., London, 1959. 2 C. L. Mantell, Tin, its Mining, Production, Technology, and Applications, 2nd edn., Reinhold Publ. Corp., New York, 1949. 3 N. M. Surdia, Asiah Hussain, T. Surdia, Salim, and Buchari, in Murray Moo-Young and Grahame J . Farquhar (Eds), Proc. 1st Int. Symp. Waste Treatment and Utilization, Waterloo, Canada, July 5-7, 1978, Pergamon Press, Oxford, 1979, pp 141-147. 4 J. I. Goldstein and H. Ykawitz (Eds), Practical Scanning Electron Microscopy. Electron and Ion Microscope Analysis, Plenum Press, 1977. 5 A. I. Vogel, A Textbook of Quantitative Inorganic Analysis, 3rd edn., Longnian, London, 1971, pp 864-870,503-506. 6 F. Sutton, A Systematic Handbook of Volumetric Analysis, 1st edn., J. & A. Churchill Ltd., London, 1935, pp 353-358. 7 W. W. Scott, Standard Methods of Chemical Analysis, 5th edn, Van Nostrand Comp. Inc., New York 1959, pp 954-974. 8 J. H. Fang and F. D. Bloss, X-ray Diffraction Tables, Southern Illinois University Press, Carbondale and Edwardsville, 1966. 9 Powder Diffraction File Search Manual Hanawalt Method Inorganic Compounds, Publication SMH-27, Southern Illinois University Press, Carbondale and Edwardsville, 1977. 10 Selected Poi\.der Diffraction Data for Minerals, 1st cd., Joint Committee on Powder Diffraction Standards, Pensylvania, 1974.
This Page Intentionally Left Blank
437
THE TREATMENT OF DETERGENTS IN INDUSTRIAL WASTEWATERS ON A PILOT PLANT SCALE BY CATALYTIC OXIDATION
Z. GORZKA, M. KAZMIERCZAK, E. FILIPIAK
Institute of General Chemistry, Technical University of Eodi, 90-924 t d d i , Poland
ABSTRACT A method for the treatment of detergents and other toxic organic substances occurring in industrial wastewaters in high concentrations has been worked out. The wastes being treated contained detergents with non-ionic surface active substances (NSAS) addition products of ethylene oxide with alkylophenols in concentrations up t o ca 70,000 mg/dm3, and also oils and grease from washing operations. Large quantities of impurities in the wastes are confirmed by the values of COD u p t o ca 123,000 mg 0, /dm3. In the method described, aqueous solutions of the wastes are evaporated from the surface of a vaporizer packing at a temperature of ca 500°C and then the organic substances are oxidized t o CO, and H,O over a copper-zinc catalyst in t h e presence of excess air at a temperature of 400-500°C. The reduction of the impurities obtained in t h e pilot plant installation, expressed as the change of NSAS and COD values, was 97% o n average. The installation was heated electrically and was able t o treat the wastes continuously at a flow rate of ca 1.2 m 3 / 2 4 h. The heat of the gaseous products of the reaction is used for preheating the air and wastes. O n the basis of the results of investigations of the pilot plant over one and a half years, a material and energy balance o n the system has been worked out. The consumption of electrical energy for the reduction of 1 kg of COD from the wastes was betwecn 10 and 50 kWh.
1. INTRODUCTION
Detergents belong to a group of organic materials which can pose a great threat to the aqueous environment. Their use is continually increasing making necessary more and more effective methods of waste disposal. One of the reliable methods is the thermalcatalytic oxidation of organic substances contained in industrial waste waters [ 1-41. A prototype installation for the thermal-catalytic oxidation of used waste waters containing non-ionic detergents from washing baths was constructed in 1981 in the Factory of Transport Equipment in Kalisz. Baths are used in this plant for washing metal parts. They contain surface-active substances of the ethoxylated alkylophenols type, and also both anticorrosive and antifoaming agents. Used baths contain, in addition, considerable amounts of oils and greases removed from the metallic parts.
43 8 The wastes treated in the installation were characterized by great variations in composition. The concentration of non-ionic surface active substances (NSAS) ranged from 5,000 to 70,000 mg/dm3. The concentration of organic substances measured as chemical oxygen demand (COD) ranged from 8,500 to 123,000 mg O2 /dm3 and the residue after calcination at 550°C was from 1.5 to 15 mg/dm3.
2. DESCRIPTION OF THE INSTALLATION
The installation, shown diagrammatically in Fig. 1 was constructed according to the design worked out by a team consisting of the workers of the Teclmical University of t d i and the Factory of Transport Equipment [ 5 , 61. According to the design specification the maximum values of the parameters were: flow rate of the wastes 50 dm3/h, air flow rate 35 m3/h, temperature of the evaporator 500°C and temperature of the reactor 400°C. Tlie major items of the waste treatment installation are: - evaporator made of acid-proof pipe 450 mm in diameter filled with Rashing rings, heated by an electric furnace of 52 kW power, - reactor made of acid-proof pipe 250 mm in diameter, containing 75 kg of Cu-Zn catalyst, heated electrically by a furnace of 35 kW power. An oxide catalyst containing 49% Cu and 20% Zn in the form of tablets 6 X 8 nim was used, - air heater constructed as a cross - flow heat exchanger with a heat exchange area of 2.7 ni2, - wastes heater built of a ribbed pipe placed inside the duct through which the waste gases flow. Heat exchange area 3.4 m2. The rig is operated in the following way (Fig. 1). The air, heated in the air heater AH, is passed to the evaporator E. Simultaneously the wastes, heated in the wastes heater WH, are delivered to the upper part of the evaporator E. The evaporated wastes, together with the air, flow through the reactor R. The waste gases leaving the reactor heat the air and the wastes in the heaters AH and WH, and then are discharged outside through the chimney. The installation described is a prototype device enabling measurement of temperature and pressure to be made at many points, and also allowing easy observation of the surface of the packing in both the evaporator and the reactor.
3. RESULTS OF THE INVESTIGATIONS
Tlie main problem of the methods described is high energy consumption caused by the neccessity of evaporating the water contained in wastes. The specific energy consumption decreases with increasing flow rate of the wastes, as shown in Fig. 2. It foliows that it is desirable to deliver as much of the wastes per unit of time as can be evaporated in the evaporator. At a flow rate of 50 dm3/h, the energy consumption was 1440 kWh/m? Further increase in the flow rate would result in lowering the temperature of the evaporator and then in flooding the packing. The energy consumption is also affected by the concentration of organic substances in the wastes. Fig. 3 presents the dependence of electrical energy demand on the COD
439
Fig. 1. Schematic diagram of the installation for the thermal-catalytic destruction of detergents.
I m
$8mx C"
.-0
6000-
4-
E
2 C
4000-
0 0
x
iaJ? 2oco
C W
0
0.01
0.02 Flow
0.03 0.04 0.05 r a t e , n$'h
Fig. 2 . Dependence of the specific energy consumption o n the f l o ~rate of the wastes. The COD of the wastes n a s 20,000 mgO, /dm3
440
I
40
1
I
I
100 120 Chemical oxygen demand, kg OZ/m3
60
80
-
Fig. 3. Dependence of the electrical energy demand o n the concentration of organic substances in the wastes. Flow rate of the wastes 50 d m 3 /h, air flow rate 35 m’/h.
1
40
1
1
1
I
100 120 Chemical oxygen demand, k g O2&
60
80
-
Fig. 4. Spccific consumption of clectrical energy as a function of the concentration of organic substances in the u astes. Flo\i rate of the tvastes 50 d m 3 /11, air flow ratc 35 i n 3 /h.
44 1
Fig. 5. Heat balance on the installation for thermal-catalytic oxidation of detergents, kWh/m”
of the delivered wastes. As the COD increases from 20,000 mgQ/dm3 to 120,000mg0,/dm3 the energy consumption decreases by about 25%. With higher concentrations of organic substances greater amounts of energy are produced in their oxidation, and thus there is a lower consumption of electrical energy. Fig. 4 presents the specific energy consumption expressed as kWh/kg COD as a function of the COD of the wastes from the baths. For adequately high concentrations of organic substances the energy consumption decreases to about 10 kWh/kg COD.
442
The costs of energy, that amount to above 40% of operating costs of installation described, decrease therefore to ca 5 zlotylkg COD (zloty - Polish monetary unit, energy price in 1981). Operating costs of biological purification plant for municipal wastes were 10-50 zloty/kg COD [7]. Thus the costs of treatment by means of thermal-catalytic oxidation, are competitive in relation to those calculated for biological methods of treatment. The method described also has a considerably higher reliability. The process of catalytic oxidation is stable even when substantial changes occur in the composition of the wastes to be treated. The degree of reduction of the impurities obtained in the installation, expressed as the change of NSAS and COD values, is 97% on the average. It seems interesting that slightly higher degrees of conversion were obtained for greater initial concentrations of organic substances in the used baths and for higher flow rates of the wastes. Thus it seems advantageous to use high flow rates of the wastes containing organic substances in high concentrations.
4. HEAT BALANCE ON THE INSTALLATION
Fig. 5 shows the heat balance on the installation for thermal-catalytic oxidation of the used washing baths wastes. The data presented in the Sankey diagram were obtained for the following parameters: flow rate of the wastes 50 dm3/h, air flow rate 38 m3/h, temperature of the evaporator 5OO0C,temperature of the reactor 400°C, COD of wastes from the baths 20,000 mgO, /dm3. The installation described is relatively small and that is why heat losses are rather great (above 40%) and heat recovery is inconsiderable (ca 7%). Greater reactors would be characterized by better utilization of energy.
5. CONCLUSIONS
Investigations carried out in the pilot plant installation for treatment of the used wastes from the washing baths confirmed the effectiveness and reliability of the method of thermal-catalytic oxidation. However, if this method of treatment of wastes is to be economically justified several conditions must be satisfied. The wastes must contain toxic organic substances in high concentrations. It is inadmissible to dilute the wastes with waste waters with low concentrations of organic substances. Too low a flow rate of the reactants should also be avoided. Equipment of the type described is particularly useful as a preliminary treatment plant for the treatment of wastes before mixing them with the entire factory wastes. In the case of the installation described it proved possible to treat 50 dm3/h of the used washing baths wastes which had a COD ranging from 40,000 to 120,000 mg O 2/dm3 and an NSAS ranging from 10,000 to 70,000 mg/dm3. Energy consumption ranged from 1200 to 1700 kWh/m3 of the wastes or from 10 to 50 kWh/kg COD. As can be seen the cost of treatment for organic substances by thermal-catalytic oxidation per kg of COD is comparable with the costs of biological treatment. However, in order to judge the profitability of the two methods the waste water economics of the whole plant should be analysed.
443 Application of this reliable method of treatment of detergents allowed washing of the metallic parts with aqueous solutions instead of the previously used preparations containing organic solvents which are more expensive and more troublesome to use.
REFERENCES B. Borkowski, Water Research 115 (1967) 367-385. H. Quillmann, Chem. Ind. XXV/August (1973) 447-8. M. A. Walsh, J. R. Katzer, Ind. Eng. Chem. Proc. Des. Develop. 1 2 , 4 (1973) 477-481. Z. Gorzka, K. Janio, M. Kaimierczak, Materiaty Zjazdu Naukowego PTCh oraz STIPCh, Toruri 1974, p. 319. 5 Z. Gorzka, M. Kaimierczak, Physiochemical Methods for Water and Wastewater Treatment, Pergamon Press, Oxford and N. York, 1980, p. 175. 6 Z. Gorzka, E. Filipiak, R. Nowak, M.Kaimierczak, Sprawozdanie Instytutu Chemii Ogolnej, Politechnika todzka, 1978. 7 W. Motoniewicz, T. Sedzikowski, T . Bonikowski, Mate oczyszczalnie Sciekow, Arkady, Warszawa 1979.
1 2 3 4
This Page Intentionally Left Blank
44 5
ENHANCEMENT OF NITROGEN OXIDES ABSORPTION FROM WASTE GASES USING OZONE
1. POLL0 and J . JAROSZYNSKA-WOLINSKA
Technical University, Lublin, Poland
ABSTRACT The absorption of nitrogen oxide from diluted gases using ozone as compared u i t h absorption in the presence of oxygen has been studied. The nitrogen oxide concentration varried bet\\ een 0.05 inol 1 1 1 ~and ~ 0.5 in01 nC3 and ozone to NO ratio between 0 and 1.0. In the experiments a packed tower laboratory reactor was used. The tilass transfer coefficient for the \\~holcoxydation-absorption process increased \%hen ozone was used in excess (2.6 times). This makes it possible to absorb near 90% NO, even by relatively short contact times.
1. INTRODUCTION
The use of ozone for the removal o f nitrogen oxide from waste gases found application in last ten years. The have appeared several patents that use the processess of nitrogen oxide oxidation with absorption [ 1-41. The strong oxidant ozone, accelerated tlie oxidation, the slowest stage of absorption in nitric acid manufacturing. The oxidation of nitrogen monooxide by oxygen is a very slow third order reaction, and at concentrations normally encountered the absorption time and volume of absorber would be prohibiting. In the presence of ozone the equilibrium state is shifted, the oxidation is much faster and then nitrogen oxides in water as well as in nitric acid of higher concentrations can be treated. When using ozone, there is no problem with the disposal of the waste liquid [4]. Although ozone is expensive, there is interest in tlie details o f a sequential process, including the oxidation by ozone (and oxygen) as well as the absorption in water phase. In our laboratory we are working on this problem. some results have been published [5-81. In this paper we should like to present some new experimental data, that extends the conclusions we have reported t o date. The goal of the research has been to determine effect of ozone at different concentrations on the whole process kinectics. 2. EXPERIMENTAL
The experiments have been performed in a fixed bed column with water as sprinkling liquid. The nitrogen monooxide was formed by action of sulphuric acid on sodium ni-
446 trite in presence of iron -11, into a stream of pure nitrogen. A regulated concentration of NO, containing no NO, was obtained. This gas was mixed with oxygen-ozone inmediately before the inlet of the column. Nitrogen oxides were photometrically determined at the entrance on the exit of gas, as well as on the exit of sprinkling water. In the presented paper we have demonstrated a new form of Shenvood's equation, through experimental results.
3. RESULTS
The device used (Fig. 1) consisted of an ozone generator, the nitric oxide producer and feeder, an ozone-nitric oxide mixer, situated on the inlet of a laboratory packed
1 ozone generot 3 Pump 4 heat exchanger 5 absorber 6 ventilator
-I Fig. 1 . Schema of experimental set up
column absorber, sprinkled by water at a contact flowrate. The following parameters were varried: the nitric oxide concentration on the input, the ozone to nitric oxide ratio, the rate of gas flow (the nitrogen as diluting gas). The temperature during all experiments was constant 2 0 ° C , several experiments were conducted at elevated temperature to verify the value of the Schmidt's number. The results for two column heights of packing and the dependences of the mass transfer coefficient on the O3 to NO molar ratio at the inlet of the column are shown in Figs. 2 and 3. From the experimental data the total rates of absorption and then the rates of absorption per unit of area have been calculated. The mass transfer coefficient for each experiment the absorption process has been deduced. The total process has been analyzed using dimensional analysis and theory o f similarity. The concentration of oxidizing agents was taken into account. It contained two components: oxygen at high concentration that varied only slightly and ozone at low concentration that varied over a large range. The number of units, used for expressing the properties of gas, the parameters of flow, the concentrations of components and geometry of column, has been taken into account
447
~~-
0.3
0.6
0.9
1.2
Fig. 2. The dependence of mass transfer coefficients from the ratio of ozone concentration to nitric oxide concentration for different conditions t , = 12.5 t, = 19 t , = 31.5 contact times (s) concentration of CNO = 0.2 curve 1 4 7 nitric oxide (mol/mJ) CNO = 0.1 curve 2 5 8 CNO = 0.025 curve 3 6 9 heights of packing h = 0.14 m
0.3
0.6
0.9
1.2
Fig. 3. The dependence of mass transfer coefficients from the ratio of ozonc concentration to nitric oxide concentration for different conditions t , = 36 t: = 54 t,* = 90 contact times (s) concentration of nitric oxide (n101/m3) CNO = 0.5 curve 1 4 7 CNO = 0.1 curve 2 5 8 CNO = 0.05 curve 3 6 9 heights of packing h = 0.39 ni
448 and compared with the number of indispensable dimensions. According to the theory the final equation should contain five dimension-less terms. The form of this equation is similar to the other one, used for physical absorption, but containg one more term. This term express the oxidant to nitrogen oxide ratio. Coefficients and indexes were computed by linear correlation to experimental data. It is impossible to introduce a simplex, in the form of concentration ratios with a common index. As expected the proposed form limits, to some extent the validity of the equation over a larger range of parameters values. The theoretical equation has the from: a b c d Sh = A . R e . Sc . S1 . S, where: Sh = (k . de)/6
-
k de
Sherwood’s number
- mass transfer coefficient kg/m2 s
l/a’ per m2/m3 of packing 6 diffusion NO, coefficient dynamic kg/ms Re Reynold’s number - Schmidt’s number sc S1 = h / d e h - heigthofpacking s2 = ( c u t / c N O ) : c u t , c03,CO, CNO - nitric oxide concentration Cut - oxygen and ozone concentration
a’
~
~
The coefficients x, y, s, were optimized and were found using the experimental data for mixtures of oxygen and ozone. The following values for the coefficients for the packed reactors were:
A=0.22, a=0.43, s = 0.35
b =0.33,
c = -0.43,
d =-0.22
and
x = 2.7,
y =0.35,
Hence, the equation describing the process which takes place in packed reactors will be: 043
Sh = 0.12 Re
033 -043
Sc
S,
-022
S,
whcre: the value C u t in the parameter S, is described by a function
449 4. CONCLUSIONS
The process of diluted oxides sorption in presence of ozone in a water sprinkled packed tower has been studied. The mass transfer coefficient was 2.6 times greater, when ozone was used in subsoichiometric quantities to NO, as compared to the case with 02. The resulting equation contains an unconventional term, describing the effect of oxidizing agents.
REFERENCES Okabe T., Jap. Patent 7384,796 1974. Piechelauri E., USSR Patent 394 300 1974. Diemer P., DDR Patent 2123993 1972. I. E. Kuznetsov and T. M. Troickaya, Zascita vozdusnogo basseyna ot zagraznien3a wrednymi wieszczestwami chimiczieskich predprijatij, Chimia, Moskva, 1979. I. Pollo and J. Jaroszyhska-Wolihska, Information-Atmosphere Protection, Dresden DDR, 4, 73, 1978. I. Pollo and J. Jaroszyhska-Wolihska, ibid., 1, 53, 1982. I.' Pollo and J. Jaroszyhska-Wolihska, Proc. Symp. Chemical Pathways in the Environment, 1980, Paris-Palaiseau, p. 7. I. Pollo and J. Jaroszyhska-Wolihska, Proc. Symp. Chemical Technology Chemical Cybernectics, 1982, Wroclaw - Poland, Works of Technical University, 24, p. 113.
This Page Intentionally Left Blank
451
ELECTROCHEMICAL OXIDATION OF ROKAPHENOL N-6IN A FLOW ELECTROLYZER A. SOCHA, Z. GORZKA Institute of General Chemistry, Technical Uniqersity of t o d i , 90-924 t o d z , Zwirki 36 (Poland)
ABSTRACT Rokaphenol N-6 is the name commonly used in Poland for 18-p-nonylphenyl-3,6,9,12,15,18-hexaoxaoctadecan-1-01. This compound belongs to the group of non-ionic surface active substances (NSAS). Detergents containing NSAS of this type are toxic in the aqueous environment and their biodegradation is very slow. The lack of effective methods for treatment of waste waters containing NSAS at higher concentrations has led us to investigate this subject. For electrooxidation platinum and oxide electrodes (RuO, 40% and TiO, 60%) have been used. It was found in batch experiments that electric charge is most effectively used for oxidation of Rokaphenol N-6 in the range of substrate concentrations 160 to 500 mg/dm3. The flow of an electric charge greater than 0.025 Ah through the electrolyzer produces, in fact, a constant degree of oxidation about 90% in the batch experiments. Similar results were obtained by oxidizing Rokaphenols N-8, N-14 and detergent Roksol IT. Parameters for oxidation in the flow electrolyzer were the optimum parameters determined in the batch experiments. The flow rate of electrolyte in the 100 cm3 electrolyzer varied from 0,35 to 5 cm3/min. A current density of 0.025 A/cmz, a temperature of 313 K and a NaCl concentration of 10 g/dm3 were used. These parameters allow obtaining of about 80% degree of conversion in the flow electrolyzer. The products of Rokaphenol N-6 electrolysis show much lower toxicity to aquatic organisms and do not inhibit the growth of algae.
1. INTRODUCTION
Rokaphenol N-6 is the name commonly used in Poland for 18-p-nonylphenyl-3,6,9, 12,15,18-hexaoxaoctadecan-l-ol. This compound is a product of the addition of ethylene oxide (EO) to nonylphenol (NP). It belongs to the group of non-ionic surface active substances (NSAS). Ethoxylated alkylphenols belong to the group of so called hard detergents, highly resistant to biodegeneration. Commonly known and so far applied methods of treatment of industrial waste waters are not useful for concentrated solutions of NSAS (0.1 to 1 g/dm3). It appeared that the method of electrochemical oxidation would considerably lower NSAS concentrations. Use of a platinum electrode at a current density of 0.025 A/cm2 and an electrolyte NaCl concentration not lower than 5 g/dm3 gave 90% conversion of the substrate [ 11. Salinity of the waste waters with sodium chloride in concentration higher than 5 g/dm3 is quite often in factories producing or using detergents.
45 2 When oxide electrodes (RuO, 40% and TiO, 60%) are used as anodes and the above mentioned parameters held constant, the efficiency of the reaction can be increased and the electric energy consumption decreased to about 10 kWh/kg of the oxidized substrate ~ 231. , This paper presents the results of further investigations which aimed at: - determination of the effect of substrate concentration on the degree of conversion - oxidation of Rokaphenol N-8 and N-14, and Roksol IT - oxidation of Rokaphenol N-6 in a flow electrolyzer. 2. METHODS
In order to carry out the above tasks a glass electrolyzer with separated electrode areas was used. The diaphgram was foam glass of G-2 type. The volume of the electrolyte was 100 cm3. Oxide electrodes were used as anodes. The concentration of the substrate was determined by a colorimetric method and the degree of conversion was calculated from the difference in NSAS concentrations before and after the reaction [4]. The optimum parameters found previously were used [2, 31. 3. RESULTS AND DISCUSSION
The dependence of the degree of conversion a on the electric charge Q was determined
Gc 1 100
[ O/O
80.
6 0.
40
20.
1 0.03 0.05 Q [ A h l
Fig. 1. Dependence of the degree of conversion on the charge Q for different initial concentrations.
453 for different initial concentrations of NSAS ranging from 86 to 800 mg/dm3. The results of these measurements are presented in fig. 1 . Using the first Faraday law the apparent electrochemical equivalent k and apparent number of electrons n taking part in the reaction were calculated for two charges: Q = 0.05 Ah equal to 18 C and Q = 0.01 Ah equal to 36 C. The results of these calculations are presented in Table 1. The degree of conversion a, presented in Table 1 was estimated from the difference in concentrations: initial one NSAS C o and concentration of NSAS after the time of reaction t - C t according to the formula: ~
a = (C0
-
C,) lOO/C0
Knowning the value a from the experiments, the mass of oxidized NSAS-mox has been calculated for given Co and Q. From the estimated value of m,, the number of moles of the oxidized NSAS - N and the values k and n have been obtained. Tab. 1. Change in the number of electrons n and electrochemical equivalent k at a defined charge Q depending on the initial concentrdtion Co. m,, - NSAS oxided, N - number of moles oxided CO
Q=0.005
mg/dm’
%
86 160 286 376 648 800
53 43 24 18 13 15
mg
4,558 6,880 6,860 6,710 8,420 12,000
0,95 1,42 1,42 1,40 1,74 2,48
Ah = 36 C
Q = 0.01
Ah=18C
19,80 13,12 13,16 13,34 10,72 7,52
mg/C
%
1%
0,253 0,382 0,381 0,376 0,468 0,668
72 76 44 34
6,19 12,160 12,584 12,784 14,260 24,000
-77 30
mg/C
1,28 2,51 2,60 2,64 2,95 4,96
29,17 14,85 14,35 14,12 12,66 7,52
0,172 0,338 0,350 0,355 0,396 0,667
It can be seen from the calculations and fig. 1 that the degree of conversion of Rokaphenol N-6 in the range of concentrations from 160 to about 500 mg/dm3 increases almost proportionately to the increase in the amount of charge up to 0.01 Ah. If the charge changes from 0.01 Ah to about 0.025 Ah, this proportionality slowly decays. Above a charge of 0.025 Ah the degree of conversion depends little on the charge and initial concentration. When solutions of initial NSAS concentration lower than 160 mg/dm3 are used chlorine and oxygen are evolved. Incomplete consumption of the charge for oxidation of NSAS takes place, resulting in changes in k and n values. In the concentration range from 160 to 500 mg/dm3 NSAS, the charge flowing through the system is used to a maximum degree for oxidation of Rokaphenol N-6. This is in agreement with the view that the presence of organic compounds susceptible to oxidation xonsiderably decreases the evolution of oxygen ahd chlorine [ 5 ] . In the concentration range NSAS from 160 to 500 mg/dm3 practically constant values k and n are observed for the charges investigated. i n considering the dependence of the degree of conversion on the amount of charge it should be noted that, under the conditions of the experiments carried out, the potential of the electrode under study changed during the reaction fig. 2. As can be seen from the fig. 2, after about 20 mins of reaction time i.e. after the charge about 0.01 Ah has flowed, the potential of the electrode changes.
1V l SCE
0.05A
I/0
------1
10
30
50
t
[min]
-
Fig. 2. Changes in the potential of the electrode during electrolysis for different current intensities.
This indicates the occurrence of another reaction, the progress of which is not controlled analytically. Then the change in the degree of conversion of the investigated compound is small. It should thus be concluded that when a charge greater than 0.01 Ah flows through the electrolyzer, it is used also for another electrolytic reaction, uncontrolled analytically, other than the reaction of oxygen or chlorine evolution. This results in a less than proportionate increase in the degree of conversion of the substrate in relation to the increase in the charge above 0.01 Ah. During electrolysis of NSAS solutions in the concentration range from 500-800 mg/dm the value k increases and n decreases. This may indicate a so called ‘shallow’ oxidation of the investigated compound. As for Rokaphenol N-6, the dependence of the degree of conversion of Rokaphenol N-8 and N-14 and the detergent Roksol IT on the amount of charge has been determined. The main components of Roksol IT are Rokaphenols. On the basis of these experiments we have estimated the dependence on the number of etoxyl group EO of the electric charge consumption needed to obtain a given degree of conversion of Rokaphenol. If the concentration of the substrate is expressed in mg/dm3, an increase in the number of ethoxyl groups from 6 to 8 i.e. going from Rokaphenol N-6 to N-8 is accompanied by a considerable increase in the electric charge needed to obtain the same degree of oxidation. For an increase in EO from 8 to 14, the change in the charge needed is small. The high efficiency of electro-oxidation of Rokaphenol N-6 in a batch apparatus led us to begin investigations of oxidation of this compound in a flow electrolyzer. The initial parameters of the reaction carried out in a continuous way were the optimum parameters estimated from the batch experiments. The parameters and conditions were as follows: current density 0.025 A/cm2, temperature 3 13 K , concentration of sodium chloride 10 g/dm3, and the volume of electrolyte 0.1 dm3. Initial concentration of NSAS in all the measurements was 484 mg/d.m3. For investigations of continuous oxidation, the same electrolyzer as for the batch experiments was used. A known volume of electrolyte was supplied and drained gradually by means of a pump. The flow rate of the electrolyte was selected so as to obtain different times of retention - shorter and longer than in the
A
45 5
t
8 0-
60-
40-
1
I
I
I
I
1
2
3
4
-
5 vlcqrninl
Fig. 3. Dependence of the degree of conversion of Rokaphenol N-6 o n the flow rate of the electrolyte.
batch method. The electrolyte flow rate was thus varied over the range from 0.35 to 5 cm3/min, corresponding to retention time of 285 and 20 mins, respectively. Fig. 3 shows the change in the degree of conversion of Rokaphenol N-6 depending on the electrolyte flow rate. The degree of conversion decreases from 80 to about 40% as the flow rate increases from 0.35 to 5 cm3/mi11. A tenfold increase in the flow rate results in about a fifty percent decrease in the degree of conversion. It was found that in order
0.06
0.0 8
Oi l 0
0.i2
J [A]
Fig. 4. Dependence of the degree of conversion of Rokaphenol N-6 on the current intensity at the electrolyte flow rate v = 5 cm3/min.
456 to prevent a lowering of the degree of oxidation, current greater than 0.125 A should flow through an electrode of 2 cm2 surface area fig. 4. Application of the flow rate 5 cm3/min allows three electrolyzer volumes per hour to be treated with a degree of conversion of Rokaphenol N-6 greater than 70% when current intensity higher than 0.125 A is employed. This products of oxidation of Rokaphenol by electrolysis in electrolyte solution were subjected to a test for toxicity. It was found that the toxicity to aquatic organisms of the products of oxidation of Rokaphenol N-6 is about one-seventieth that of the substrate and that the growth of algae is not inhibited [ 6 ] . The authors acknowledge financial support for this work by a grant MRI-11 from University of t o d i .
4.REFERENCES 1 2. Gorzka, K. Jasihska, A. Socha, Physicochemical Methods for Water and Wastewater Treatment, Pergamon Press, Oxford and N. York, 1980, p. 163. 2 Z. Gorzka, A. Socha, Zeszyty Naukowe P.S. z. 91 (1979) 97-106. 3 Z. Gorzka, A. Socha, Polish Patent No. 118222, 1 3 January 1983. 4 W. Hermanowicz, Fizyczno-chemiczna badanie wody i Sciekbw, Warszawa 1976. 5 G . Kortum, Elektrochemia, Warszawa 1970, p. 569. 6 Z. Gorzka, A. Socha, E. Kwiatkowska, Gaz, Woda i Technika Sanitarna, 11 (1980) 321 -323.
45 7
REMOVAL OF POLLUTANTS FROM THE AQUATIC ENVIRONMENT BY PHOTOOXIDATION
M. MANSOUR, H. PARLAR and F. KORTE Gesellschaft fur Strahlen- und Umweltforschung mbH Miinchen, Institut fur Okologische Chemie, 0-8050 Freising-Attaching and Technische Universitat Miinchen, Institut fur Chemie, 0-8050 Freising-Weihenstephan, W. Germany
ABSTRACT Since degradation under the influence of natural light is an important process in the environment, we have investigated the photooxidation of some organic compounds in aqueous hydrogen peroxide (30%) exposed to UV light (A > 290 nm). Photolysis of hydrogen peroxide yields 'OH, which reacts with most classes of chemicals and thus is of special importance in their degradation. The possible reaction mechanism is discussed, and the photodecomposition rate of H:O: in water is estimated.
1. INTRODUCTION
Chemicals introduced into the environment by human activity can cause ecological effects in manifold ways. It is necessary therefore to have methods for quantitatively determining the persistence of chemicals in water t o obtain an ecotoxicological assessment of these chemicals. Persistent substances present difficult problems during effluent treatment, because they are decomposed only very slowly biologically and can not be removed satisfactorily by the physical and chemical means often used. Among possible simple methods for eliminating chemicals from effluent, photo-induced oxidation in aqueous hydrogen peroxide deserves special regard, since it can produce a complete breakdown to harmless or at least biologically degradable compounds. Our experiments showed a complete degradation o f certain organic environmental chemicals in aqueous hydrogen peroxide, thus allowing an assessment of the technical possibilities; we devoted these studies t o this model reaction o f photo-oxidatively eliminating from water organic compounds which are difficult to destroy biologically. The use o f H2 O 2 during photooxidation of indust rial effluent is one possibility, along with other special technical processes, for converting polluted into usable water. It is well established that hydrogen peroxide and ozone, in water irradiated by UVlight, are especially suitable for decomposing environmental chemicals [ 11. It is also known that OH radicals are very reactive towards organic chemicals in water and the atmosphere [?I. They possess a great affinity for electrons and react with a variety of
45 8 Tab. 1. Rate constants for the reaction of OH radicals in water (KoH)(in L/mol, sec)
'OH
Substrate
Ozone
5.1 '10' 3 . 6 . lo9 3.0. lo9 -9.0. lo9 1 . 2 . 10'" 9 . 0 . 10' 1 . 8 . 10'"
Methanol Benzene Toluene Styrene pCresol NH, Phenol
[4] [4] [4] [6] [4] [6] [8]
3.0. lo-? 1 . 7 . lo-' 1 . 0 . lo5 1.0 20.0 .20.0
[5]
[S] [7] [7] [6] [6]
chemical classes by abstraction of a hydrogen atom or by addition to a carbon-carbon double bond to form water and an alkyl radical (Table 1). A comparison of OH with OR and OOR shows OH to be the most effective [3]. Decomposition of H 2 0 2 into radicals is effected by light and is faster at higher pH values. The rate of decomposition of H 2 0 2 into OH radicals thus depends on pH and on the nature and concentration of impurities in the water (Fig. 1). We have investigated the effect of H 2 0 2 on the oxidation of benzene and determined the rate and the product distribution accurately. The decomposition rates of H 2 0 2 were greater than those of benzene. Thus, the observed variation in the rate of decomposition of H 2 0 2 with its concentration shows the presence of a hydrogen peroxide decomposition system that is initiated by OH or other radicals formed by decomposition of the substrate (Fig. 2 ) . The concentration of H 2 0 2 altered only little during the early stages of the reaction, so it may be considered to be constant in this first phase, while the second reaction phase proceeds faster as the initial ratio of substrate to hydrogen peroxide increases. At approximately equal initial substrate and H2 O2 concentrations the reaction slows progressively with time. Higher concentrations of H 2 0 2 also cause the reaction to slow down.
1001
l
0:
V
I'
60
120
I'
180 240
Irradiation time (min)
Fig. 1 . Photodecomposition of 0.532 mol H,O, in l W m l aqueous solution at 25°C.
459
100
+
c L
: 90d
-5
80-
0 I0
m 0 $’
-
70-
0
a x
60-
0
I 01 0.2 0.3 0.4 0.5 0.6 0.7 Hydrogenperoxide(mol)
Fig. 2. Photodegradation rate of 9.2 . lo-’ M Benzene solutions at pH 4.6 and pH 2.3 under h 2 290 nrn in distilled water after 3 hours.
0
1
2
3
Irradiationtime ( h o W
Fig. 3. Photodegradation of 2.45 . lo-’ M Benzene in the presence of 0.53 ml H,O, in 1000 ml distilled water (0)and tap water ( 0 ) under h 2 290 nm.
Investigations at varying pH under selected experimental conditions gave almost twothirds decomposition of the HzOz after 140 minutes. At pH 3.9 the HzOZ decomposition and the oxidation of organic compounds take place quickly (Fig. 3). At pH 2.5 the H z 0 2 reaction rate depends on the organic substrate and its behaviour in aqueous solution. The influence of oxygen and nitrogen on the decomposition was investigated, and it was found that increasing concentrations of oxygen accelerate the H 2 0 z breakdown in comparison with rates after flushing with nitrogen. In order to test whether the OH radicals and other oxygen species can react with model compounds, We
460 held a constant number of parameters: reaction time, molar ratio of H 2 0 2 to substrate and temperature (Tab. 2). The highly chlorinated compounds PCB and cyclodiene insecticides react at slightly different rates. Tab. 2. Photooxidation of organic compounds in aqueous hydrogen peroxide with ultra violet light Substrate dis appearcnce % h
h
Substrate
Concentration of substrate
Concentration of hydrogen peroxidc (mole)
Photoreaction time (inin)
<290nni
>290nn1
Initial pH
Methanol Allylalcohol Benzene Toluene pCresol Phenol Dimethylphthdate
2.15 2.70 1.17 1.80 3.50 2.80 2.87
0.53 0.49 0.57 0.62 0.51 0.42 0.45
180 180 240 300 300 180 300
38.5 41.8 58.6 45.4 60.3 49.7 25.4
10.75 14.85 25.4 18.2 22.3 21.5 7.4
5.3 5.9 6.3 6.5 5.2 5.9 6.7
2. MATERIALS AND METHODS
30% hydrogen peroxide and UV light of wavelengths greater than 290 nm were used in the present work. The only waters used were tap water and distilled water. The experiments were conducted in a thermostated glass cylinder. To improve mixing the reaction solution was flushed with a stream of nitrogen. A Philips HPK 125 lamp giving 17 . mol quanta per hour was employed. The lamp was surrounded by a Pyrex (borosilicate glass) cooling jacket and was immersed to about 95% of its lenght in the reaction solution, which was cooled to 20-2SoC. The solution was stirred by a magnetic stirrer. Hydrogen peroxide concentrations were measured spectrophotometrically and titrimetrically [9]. We used a Pye Unicam SP8-100 spectrophotometer to determine the absorption coefficient. Our experience has shown that the individual compounds could be analysed by direct injection of the water sample into the gas chromatographs without i! sample work-up if the concentration is high enough. In many cases, however, enrichment was performed by extraction with suitable solvents. The gas-chromatographic conditions were selected to match whatever compound was to be analysed. In selecting suitable GC columns, special attention must be devoted to seeing that substances of different polarity give well-resolved peaks. A gas-chromatograph (Carlo Erba Fractovap 2450) equipped with a flame ionisation detector and coupled to a 1 mV recorder (Linseis L 6501) with an electronic integrator (Hewlett Packard 3881 A) was employed. The columns used were 2 X 3 mm ID. Glass columns packed with Tenax GC 60/80 mesh gas-chrom for the determination of methanol, ally1 alcohol, benzene and toluene and 1% SP 1240 deactivated for acidic compounds on SO/ 100 supelcoport for the determination of p-cresol and phenol. 3% SE-30 was used for dimethyl phthalate. The carrier gas was nitrogen (GC grade) at a flow rate of 30 ml/min. Quantitation was based on external standards using the authentic compounds in the analytical solvent.
46 1 The column temperature was programmed from 80 to 120°C at boC/mi11., with the flame ionisation detector at 2SO°C. Nitrogen carrier gas flow was 40 ml/min. All solvents were o f analytical reagent quality and were obtained from E. Merck, D-6100 Darnistadt, F.R.G.; hydrogen peroxide (30%) was obtained from Fluka, D-7910 Neu-Ulm, F.R.G.; Dimethyl phthalate and pcresol, both from Fluka AG, CH-9490 Buchs, Swiss, were of the highest available purity (> 99%). The standards were prepared by dissolving the substance in 15 nil water and mixing well to give concentrations of 0.06 and 0.03 mg/nil. A standard curve was obtained by analyzing the solutions. For the dark reaction a solution of 100 nd destilled water at pH 4.5 of p-cresol(3.1 g) and H 2 0 2 (0.53 ml) was allowed t o stand in the dark, purged with N 2 , for 24 hours. The reaction mixture was extracted twice with 50 ml cyclohexane. the extract was concentrated to 0.5 ml and analysed by gas chromatography. The analysis indicated 86-94% of the p-cresol remained unreacted. Sinlilar reactions in the dark at pH 1.5 and pH 8.9 showed 89% and 84% o f the p-cresol, respectively, t o remain.
REFERENCES
1 W. C. Schuinb, C. N . Satterficld, and R. L. Welitworth (1955) Hydrogen peroxide, ACS Monograph No. 128, Chapter 8, Reinhold Publishing Co., N e n York. 2 H. Taube, ‘Photochemical Reactions of Ozone in Solution’, Trans. Faraday SOC.,53, 1957, p. 656-665. 3 T. Mill, D. G. Hendry, and H. Richardson, Science 207, 1980, p. 886. 4 M. Anbar, and P. Neta, hit. J. Appl. Rad. Isotopes 18, 1967, p. 493-523. 5 T. W. Nakagawa, L. J . A n d r e w , and R. M. Keefcr, J . Am. Cheni. SOC.82, 1960, p. 269-276. 6 J . Hoign6, and H. Bader, Vom Wasser 48, 1977, p. 283-304. 7 D. G. Williamson, and R. J . Cvetanovik, J. Am. Chem. Soc., 90, 1968, p. 3668-3672; 4248-4251. 8 C. 1.: Adams, J . W. Boag, J . Currant, and B. D. Michael, Pulse Radiolysis, Ed. M. Ebert ct al., AcadeniicPress, New York (1965), p. 131-144. 9 A. 1. Vogel, ‘A Text-Book of Quantitative Inorganic Analysis’ 2nd edition (1955) p. 283-284, Longman, Green, and Co., New York.
This Page Intentionally Left Blank
463
MODELING OF IMMERGED FILTERS IN THE CASE OF AERATED FILTERS
P. LE CLOIREC, G. MARTIN Laboratoire Chimie des Nuisances et Gtnie de IEnviroiiment ENSCR Avenue du Gentral Leclerc F 35000 Rennes, France
ABSTRACT The proposed modcling relies on the associated study of the: hydrodynainics of the filters used - kinetics of compound degradation. We first of all determined the effective average residential time in biological aerated filters. It was found to depend both o n the void factor E , that varies with the biomass, and on the space p taken up by the water depending o n the amount of air flo\\.ing into the filters (F). The \\,riters proposed the following formula: -
e = E ( X ) @(I?)
V/Q
The second step was to drew up a balance of the weight factor on the dispersion theory, and on series of pcrfectly mixcd reactors, \ye thus obtain a good correlation between the values comported with thc model and the experimental findings on a pilot unit treating a mixture of cthanol and ammonia nitrogen of concentration 20 n1gl-l and 5 nig1-l .
1. INTRODUCTION
The use of fxed biomass fiters in water treatment tends to develop fast. Media may be plastic or clay fdter, or activated carbon. The fdters concerned may be two phases (water-media) or three phases (water-media-gas) as is the case of denitrification or activated carbon biological filters aerated in the mass [ 11. Modeling of these systems have recently been proposed [ 2 ] . The ELMALEH model [3, 41 introduces degradation of substrate proportional to the biomass in the reactor. The flow of oxygen at the biofilm-liquid interface is proportional to the concentration of oxygen and degradation agents. By plotting and resolving the balance of the different elements, it is possible to dimension the units however certain discrepancies between theory and experimental results may appear, for instance with regard to the dissolved oxygen. The GRASMICK model [ S , 61 is more complex than the previous one, taking into account the resistance to diffusion inside the biofilm. It is a fact that the degradation speed at the biofilm-liquid interface reaches an asymptote when the thickness of the bio-
464 film varies between 60 and 150 p m depending on the substrates and the techniques used [7, 81. However the notion of effective thickness is introduced, limiting the thickness of the biofilm to the possibility of diffusing air inside the biomass. GAID [9] and MARTIN [lo] introduce interaction between the substrate, the media and the bacteria. The different balance equations take account of the building up of activated carbon in the pores followed by the diffusion from the pores to the bacteria. These phenomena are put forwards after analysis with a radioactive substrate [ I I , 121. The model brings up the advantage of activated carbon when the influent to be treated contains a lot of biodegradable and adsorbable matter. The models thus proposed assume plug flow conditions when considering the various parameters. We begin with studying the hydrodynamic parameters of used biological filters (using activated carbon) by measuring residential time distribution. The second step will consist in determining the degradation speed of substrates contained in the filters feed water. We shall then use these two studies to model the phenomena observed.
2. ANALYTICAL METHODS AND EQUIPMENT
The filter flow scheme used during testing is shown in Fig. 1. Characteristics of this assembly are : - column diameter 0.06 m - media height 0.80 m - type of media "Picactif" activated carbon - granulometry 1-1.5 mm porosity mesoporous - specific area 1200 m2g-' - hydraulic load 5-10 mh-' ~
out let
pump
Fig. 1 . Pilot unit schematic diagram.
465 Air is blown into the mass of activated carbon at flow rates between 10 and 100 lh-' . The feedstock is compoused of ethanol at 20 mgl-' and ammonia chloride 5 nigl-' , to which 1 mgl-' of potassium phosphate is added. Due to the presence of air and these biodegradable substrates a growth of biomass appears on the media after two weeks. In order to remove this the filter is regulary washed and the resumption of biological treatment is then quite quick (0.5-2 hours). The used substrates are not adsorbed by the activated carbon [12]. The ethanol is analysed by gas chromatography: - column Tenax - detector flamme ionization - furnace temperature 90" C - injector temperature 300" C - detector temperature 300" C - gas N2 vector pressure 1.5 bar The Nessler method enables the ammonia to be determined.
3. DETERMINATION OF HYDRODYNAMIC PARAMETERS 3.1. Average Residential Time
The average residential time in a reactor is expressed as follows:
V reactor volume Q flow rate.
In aerated biological filters, this formula cannot be applied as it stands. In fact, the real volume of the passing fluid is less than in an empty reactor owing to the presence of the filter media, tho biomass and the air. We propose to define this as a relation that embraces all these parameters.
3.1.1. Influence of Filter and Biomass
The real residential time of the fluid in a non aerated filter is defined as:
E,
void percentage
In biological filters the grains of media are covered by bacteria either dispersed or in the futur of a biological slime that influences the void factor ( E ) . We determine the ratio between E and the biomass quantified in terms of volatile dry solids (VDS).
466
2
4
6
x
m M
Fig. 2. Variation of void factor with biomass.
The activated carbon beds are fed in a continuous process with an ethanol mixture (up flowing air and water). When the biomass on the media has sufficiently developed, the head loss on the reactor is measured. Then the media is washed and the VDS it carries is counted. Measuring the loss of head fines the void factor by application of Leva's equation [ 131.
AP/H = ( 2 f U2/d) [ ( l
-
~ ) ~ - " / ( y e~3-)"]
In the operating velocity range, the Renyold's factor Re = dUp/p is less than 10. The flow is laminar: n = 1 and f = 100/Re. The shape factor y is determined by the following relation y = 4.87 v231 0
V volume of the grain Assuming that the increase biomass on the media does not affect the shape factor we find y = 0.74. Variations in the void factor with biomass are shown in Fig. 2. In filters operating at average or higher load, the concentration of bacteria is between 0 and 2 mg VSS per gram of media. These limits a feasible linear variation between E and X can be assumed according to the following formula: E
= 0.035 X
X E
e0
+ eo
biomass mg VSS per gram of media void factor void factor in clear filter.
467 3.1.2. Influence of Aeration
The reactors are fed by an up flowing stream, hence it can be assumed that all the voids between the grains are wet. When the reactor is aerated, a part of this space is taken up by the air and a part by the water. The welted fraction can be devided into two parts: the part due to internal retention Pi i.e. the fraction of water contained in the pores, and the fraction of water fixed at the points of contact between the granular media. The fraction due to dynamic retention fld consisting of the liquid moving through the filter
is determined by the simultaneous weighting of dry and wet media. We had an average value of pi = 0.30. Pd is measured by collecting liquid at the bottom of the column after stopping the infeed. Different measurements have been effected in the ratio of F (air flow rate/water flow rate). pi
Fig. 3. Variation of residential time according to air delivery.
The result of as a function of F are shown in Fig. 3. During our experiments F is between 0 and 4 and p is then linear, corresponding to the following formula :
p=p0
- 0.027 F
Po 1 in immerged filters and when F = 0. Both tests enable the average residential time in an aerated biological filter to be calculated by the relation :
468
E(X) P(F) V Q
is the void factor depending on the biomass present on the media is the wet fraction of the filter volume depending on the rate of aeration is the useful volume of empty filter is the average inlet flow rate.
3.2. Retention Time Distribution (RTD)
The average residential time, as previously defined is an overall parameter. In effect, each molecule entering the reactor has its own specific residential time. A distribution can be defined having a formula similar to :
Written the following limits J ( 0 ) = Q for 0 = O J ( 0 ) = 1 for 0 - + %
3.2.1. Measuring the RTD
Measurement is performed with the help of a weakly concentrated inert tracer so as not to disturb the current. We use sodium chloride detected at inlet and outlet by a conductometric cell. We proceed by stepped injection. The equation governing the tracer concentration at the inlet (C,) and the concentration (C) has the following form:
J(0) = (C/Co) step The reponse curves are given in Fig. 4 with the two extreme cases that may be encountered : - plug flow characterized by an identical residential time for all particles - “perfectly mixed” flow in which residential times are unspecified and the composition is assumed to be uniform at all points.
3.2.2. Interpretation of Response at One Inlet Step
Several modelizations are proposed for the interpretation of velocity distribution curves: we have applied two feasible models: - dispersion model - perfectly mixed multi-stage model.
Ir/
469
inlet
Fig. 4. Measuring the RTD. __-- Outlet of a perfectly mixed reactor - - _-- - Outlet of a plug flow reactor Outlet of a real reactor - Step input
__
3.2.2. Dispersion Model
The axial and radial dispersion through a filter bed is described by the following differential equation: -Div(-Dgrad
C
+ UC)=(6C/6t)
(111-1)
Like several other authors, we assume the influence of radial dispersion to be negligible [ 181. In these conditions: We obtain the following equation: (111-2) With the following conditions:
C = O atZ>O C = Co at Z < 0
t=O t =0
The resultion of the differential equation 111-2 gives the fcXowing relation :
erf: error function erf(Y) = 2/dr J-:
e-"
dx
Peclet factor Pe = UL/DL
0.75
Fig. 5. Determination of the number of stages.
c/", 0.75
k5
0.2 5
Fig. 6. Determination of Peclet's factor.
47 1 3.2.2.2. Model for Perfectly Mixed Multi-Stage Reactor
In this model, it is assumed that the actual reactor is composed of N series-mounted stages, each with its own response, of velocity distribution in a perfectly mixed reactor. The J reactor balance is expressed as: Cj-1 Q
-
(I I 1-4)
CjQ= Uj(dCj/de)
dCj/dK + N(Cj/F) = (NCj - 1 ) / F
(111-5)
hence :
(111-6) This equation is integrated for each stage, giving:
C/Co = 1 - e
[ l + N O / & + 1/2(NO/&)’
+ . . . + 1/(N
-
I)!
(Ne/F)N-’]
(111-7)
3.2.3.Deterniination of Peclet’s Factor and the Number of Stages
Equations 111-3 and 111-7 enable us t o obtain a set of curves J(O) = f(O/e) for the different Peclet factor or stage numbers. We used a microprocessor Apple I1 for this purpose. The average residential time is determined, taking into account both aeration and the biomass present on the media. The comparison of these ideal curves with those obtained experimentatly makes it possible to approach the number of percently mixed reactors to be mounted in series in order to stimulate the hydrodynamics of the reactor under study. This reactor is covered by biomass and aerated in order to obtain a complete simulation of the column in which the degradation of the ethanol and the assimilation of the ammonia nitrogen is effected. Fig. 5 and 6 show the real life curves that allow us to propose the following parameters: N 8 stages 1/Pe 0.075. It will be noticed that the dispersion model graphs are more difficult t o match with real curves than the perfectly mixed stage curves. 4. KINETICS 01: SUBSTRATE ELIMINATION
In order to modelize treatment on the filters we need the degradation kinetics of the added substrates. We have studied more particularly the removal o f ethanol and ammonia. 4.1. Equipment and Methods Used
The experimental apparatus consisted of small 20 cm reactors containing 10 to 50 g
472
buffer
1
Fig. 7. Experimental apparatus for the kinetics determination.
of activated carbon (see arrangement on Fig. 7). A pump was provided t o recirculate feedstock from a buffer tank. The volume of the latter and of the pipings was assumed to be negligible. It is possible to inject air into the system. The media is civered with a certain layer of biological slime, the thickness of which is measured with a microscope fitted with a micrometric eye glass. The ambient temperature is 180°C.
4.2. Ammonia Elimination Kinetics
Experimental results enable us to put forward a 1/2 order kinetic relation for elimination of ammonia. This corroborates results found by various writters [ 19, 201. - [d(NHi)/dt] = K(NH;)O5
(IV-1)
The results are presented in Fig. 8. We obtain the following equations for different values of F (air flow rate/water flow rate ratio) C : NH,' concentration
I
1
Fig. 8. Ammonia ion elimination kinetics.
473
I
k
Fig. 9. Variation of ethanol removal speed depending on the thickness of the biological film.
F=O
-
(dC/dt) = 0.25 C o 5
(IV-2)
F=l
-
(dC/dt) = 0.525 C o 5
(IV-3)
2
-
(dC/dt) = 0.620 C o 5
(IV-4)
The kinetic constant is dependent on the biomass present on the media. We have plotted the thickness of the biological film in terms of the kinetic constant. An example is given in Fig. 9. The removal of ammonia is proportional to the thickness of the layer of biomass to the extent of a limited value that corresponds to the substrate penetration depth. This agrees with former research by Kornegay and al. [ 2 , 71 on the biofilm theory, beyond a certain thickness, varying between 60 and 150 pni according to the substrate. There is
Fig. 10. Ethanol removal kinetics.
474
a stabilization of the degradation kinetic reaction [22, 231. As far as our tests are concerned we found a limit to the thickness of active bacteria of about 70 p m bearing in mind that, in our experiments, the biomass on the media was in the range of 10 mg VDS, the relations (IV-2, -3, -4) then become : F=O
-
(dC/dt) = 0.025 X Co5
(IV-5)
F=l
-
(dC/dt) = 0.520 X Co5
(IV-6)
2GFG4
- (dC/dt) = 0.062 X C o 5
(IV-7)
c X
NH; concentration in mgl-l biomass on media (VDS) mg
4.3. Ethanol Removal Kinetics
As in the previous case we worked out the kinetics of ethanol degradation. Results are consolidated in Fig. 10. We obtain an order 1 kinetic reaction :
F=l
- [d(EtOH)/dt]=0.25 (EtOH)
(IV-8)
F=O
- [d(EtOH)/dt] =0.18 (EtOH)
(IV-9)
We know the amount of active biomass on the media at the time of measuring. We can therefore propose the following equations intergrating the value X of the micro-organisms in the kinetic constants.
C X
ethanol concentration biomass in test column in mg VDS
F=l
-
(dC/dt) = 0.3 15 X C
(IV-10)
F=O
-
(dC/dt) = 0.225 X C
(IV-11)
The biofilm was also measured, enabling us to demonstrate that we obtain the same type of curve K = f(e) and that the active biofilm thickness is also about 70 pm. These tests enable us to define the various degradation kinetics of ethanol and ammonia nitrogen in biological treatment on media. The velocity constants depend both on the biomass and filter aeration. The biomass only affects this value when irs thickness is G 70 pm. If it is any thicker, KX is a constant value.
5. MODELING OF TREATMENT ON BIOLOGICAL AERATED FILTER
Knowledge of the kinetic orders and constants of the removal of ethanol and ammonia
47 5 ions, as well as a rough estimation of the hydrodynamics of the filter used make it possible to model water treatment in the aerated inhabited by a large growth of biomass. 5.1. Balance Equations
Dispersion Model
The weight balance can be expressed as:
i.e. by dividing by z and taking the extreme case Az + 0 DL(d2C/dz2) - U(dC/dz) = r The resolution of this equation gives the following relations: I Order Kinetics
When: W = (1 + 4k1 BDL/UL) x treatment efficiences 112 Order Kinetics
We are using Burchard’s results [24] and approximations to solve equation (V-2) when the kinetic reaction is of the order 0.5.
x = 1 - [(2 - R)’/4] *
[((2 - R)/2
+ 1)’
1 - [Pe(2 - R)/2 ln(2
-
R)/2]
- R/2 - 13
Where:R=K,, t!fC;, Model of Series Mounted Reactors
The weight balance for a reactor j is expressed as
Where: ei = e / N
N: number of reactors
+ Pe’ R/(2 - R)’
.
476 Order 1 kinetics
x = l - 1 / ( 1 +kl
e/N>N
07-71
112 Order kinetics
In this case we calculate the concentration at the outlet of each reactor for N reactors in series, equivalent to the real filter.
5.2. Comparison between Technical and Calculated Values 5.2.1. Assumed Dates
No SS in feedstock. Oxygen concentration in test column is constant and depends on the degree of aeration in mass of media. The substrate inlet concentration is constant. Modification of the porosity of the bed is due to: - aeration - increased biomass. The amount of biomass washed out by the flow of treated water is negligible. The thickness of the biofilm on the media is constant. The degradation speed of substrate is proportional to the quantity of active bacteria on the fiiter. The last of these hypotheses led us to try to determine the quantity of active bacteria on the media. We proceeded it by measuring the thickness of the biofim in turn related to the corresponding quantity of VDS 9Fig. 11'. By the given weight of the activated carbon this gave us the amount of "active bacteria" on the media and enabled us to define the velocity constants.
Fig. 11. Correlation of biological film with biomass.
47 7 At the time of our experiments, the ethanol and ammonia measurements where performed at the filter outlet after 1 or 2 days operating, i.e. when the biofilm had reached a thickness of 70 pm (maximum limit for active bacteria) which corresponds to 200 mg of biomass in one of the test columns. This figure will be adopted for the determination of the different degradation velocity constants for ethanol and ammonia. Measurements effected just after sampling showed that in our operating conditions VDS in the filters amounted to 200 and 300 mg in the column as a whole.
5.3.2. Comparison between the Model and the Experimental Values 5.3.2.1. Removal of Ammonia
The kinetic constants used for the calculations bearing in mind the biomass clinging to the media are given in the Table 1. Results for the utilization of the dispersion model are reported on Fig. 12. We obtain poor correlation between the model and the experimental values. Far better results are obtained when we use the model for the perfectly mixed series mounted reactors (see results on Figs. 13, 14, 15). Tab. 1. Kinetic constants used for the modeling F
0
1
2-4
KX
10
20
25
Fig. 12. Ammonia ion removal - dispersion model.
478
Fig. 1 3 . Removal of ammonia ion - series reactors model (F = 0).
Fz1
'5
io
n cBI. exp.
25
J 0.y 5
0.1
B
h
Fig. 14. Ammonia ion removal - series reactors model (F = 1).
47 9
Fig. 15. Ammonia ion removal
-
series reactors model ( 2
< F < 4).
5.3.2.2. Removal of Ethanol
With regard to ethanol, one series of tests only was performed in aeration conditions of F = 1 given the weight of biomass, 200-300 g approx. at the time of sampling, we can define a velocity constant of 61.5 h-'. Whatever the model used, dispersion or multistage, we obtain excellent correlation between the experimental and computed values. These results are illustrated in Fig. 16 and 17.
6.CONCLUSION
A study of the hydrodynamics of aerated filters, associated with the determination of the degradation kinetics of the various products seems to be a good approach to modeling: good correlation between actual and computed values is obtained with series-mounted reactors or the dispersion model, in the case of ethanol and ammonia nitrogen removal. Modelling is comparatively easy to apply in connection with pilot filters (the case in point) but can also be extended to industrial scale filters. There sre, in fact, few constants to be determined.
480
8
I 00 5
OJ
6
h
Fig. 16. Removal of ethanol -dispersion model.
a95
0,l
Fig. 17. Removal of ethanol -series reactors model.
h
48 1
A further developnient of the model could be it application to other biodegradable substances or to mixed compounds, such waste water. ACKNOWLEDGMENTS
This study reported in this article was sponsored by the "Compagnie Generale des Eaux, 52 rue d'Anjou 75384 Paris cedex 08 France".
ABBREVIATIONS volume of filter plus biomass volume of filter only concentration C concentration t o time t = 0 CO average diameter of particles d dispersion factor DL axial dispersion factor D, dispersion factor in biomass DR F air flow rate t o water flow rate ratio f Leva equation constant height of media in reactor H J ( 6) residential time distribution function kinetics constant K length of reactor L N number of reactors in series n Leva equation constant pressure P Pe Pelcct's factor Pe = UL/DL flow rate Q chemical reaction speed r Renyold's factor Re = dup/y Re RTD residential time distribution ss suspended solids fluid velocity in empty reactor U V volume VDS volatile dry solids X biomass expressed in dry volatile matter percentage of eliminated substrates X biomass t o time t = 0 XO Z height of reactor 01 kinetics order residential factor P dynamic residential factor Pd internal residential factor Pi percentage of void in media E percentage of void in clear media €0 particulc shape factor Y viscosity of fluid P volume of fluid P residential time e average residential time specific area of particles W A A0
e
RE1:ERENCIS 1 ODA Brcvct no 78-30, 282 (1978). 2 C . Martin, Le point sur I’epuration et le traitment des effluents. Tech. et Documentation, Lavoisier Paris (1982). 3 S. Elmaleh, H. Labaquere, R. Ben Aim, Wat. Res. (1978), 12,41. 4 S. Elmaleh, L’Bpuration biologique en regime transitoire, These de Docteur d’Etat, Montpellier ( 1976). 5 A. Grasmick, Thbse de Docteur-Ingenieur, Montpellier (1978). 6 A.Grasmick, S. Elmaleh, R. Ben Aim, Wat. Res. (1979), 13, 1137-1147. 7 B. H. Kornegay, J. F. Andrews, J. Sanit. Div. Amer. SOC.Civil. Eng. (1979), 95, p. 95. 8 M. Onuma, T. Omura, Wat. Sci. Tech. (1982), 14, 553-568. 9 K. Gaid, These de Docteur d’Etat, Rennes (1981). 10 G . Martin, A. Y. Le Roux, P. Schulhof, Wat. Sci. Tech. (1982), 14, 599-618. 11 K. Gaid, P. Le Cloirec, G. Martin, T. Bernard, Env. Tech. Letters (1982), 3, 8, 329-336. 12 P. Le Cloirec, K. Gaid, G. Martin, J. Sibony, Env. Tech. Letters (1982), 3, 6, 257-262. 13 J.H. Perry, Chemical Engineer’s Handbook, 4th edition, McGraw Hill, New York. 14 J. M. Smith, Chemical Engineering Kinctics, 2nd edition, McGraw Hill, New York (1970). 15 0. Levenspiel, Chemical Reaction Engineering, 2nd edition, J. Willey, USA (1972). 16 J. Willermaux, GBnies de la reaction chimique-conception et fonctionnement des reacteurs. Tech. et Documentation, Lavoisier, Paris (1982). 17 R. Aris, N. R. Amundson, in The Mathematical Understanding Engineering Systems, 1st edition, Pegamon Press (1980), 72-79. 18 M. Crine, A. Degee, J. P. Pirard,Trib. Cebedcau (1976) 391, 246-264. 19 M. Riemer, P. Harremoes, Wat. Res. (1978) 149-165. 20 P. Harremoes, Watten (1977) 2, 22, p. 143. 2 1 B. H. Kornegay, J. F. Andrews, J.W.P.C.F. (1968) 40, 11,460-471. 22 K. Williamson, P. L. McCarty, J.W.P.C.F. (1976) 48, 1,9-24. 23 N. Onuma, T. Omura, Wat. Sci. Tech. (1982) 14, 553-568. 24 A. Burghardt, T. Zaleski, Chemical Engineering Science (1968) 23, 575-591.
CHAPTER VII
RECYCLING OF WASTEMA TERIALS AND POLL W I O N FREE TECHNOLOGIES
This Page Intentionally Left Blank
48 5
WASTE-FREE METHOD OF CADMIUM CARBONATE PRODUCTION
J. 2. NIECKO
University of Nigeria, Chemistry Department, Nsukka, Nigeria
ABSTRACT Cadmium carbonate is used in industry as one of the main components for production of highquality inorganic pigments. CdCO, is manufactured industrially by a precipitation method which requires large quantities of HNO,, Na,CO, or NaHCO, and watcr to remove Na' and NO; ions from a filter-cake of the cadmium carbonate. The precipitation method generates large quantities of wastewaters which are difficult to utilize. A waste-free method of manufacturing the cadmium carbonate presented in this paper is based on a reaction of the CdO/water suspension with gaseous carbon dioxide under elevated pressures. Cadmium oxide, water and carbon dioxide are the only raw materials in this process and no wastewater is generated. Carbonation of the CdO/water suspension has been carried out in an autoclave under pressures of 1.0 X l o 3 , 2.0 X lo', 5.0 X lo3 and 1.0 X 10" hPa. On basis of the studies presented in this paper, a CdCO, production scheme by the waste-free method has been proposed.
1. INTRODUCTION
Cadmium carbonate, CdCO,, is one of the main components used for industrial production of the high quality inorganic pigments. Some amount of the cadmium carbonate is also consumed in chemical laboratories. The cadmium carbonate is manufactured on the industrial scale by a precipitation method shown in Fig. 1 . Cadmium oxide required in this process is obtained by burning cadmium vapour in air. Disadvantages of the method are: (1) Several unit operations and processes are employed. (2) Huge quantities of washing water are required. (3) Great consumption of nitric acid and sodium carbonate or sodium hydrogen carbonate. (4) Generation of waste-water containing NaNO, and, as a rule, some quantities of both Cd2+ and CdC03 in the form of very fine particles. All cadmium species are highly cancerigenic. (5) Time-consuming unit operations of filtration and drying in a chamber drier. (6) Labour-consuming unit operations of grinding and sieving. (7) A phenomenon of occlusion of the Na+ and NO; ions which then are present in the
d
H20
Dissolving
I
I Filtration
1 I
73
aJ
-a-
0
0 W L-
C 3
0
U 0
v I Drying
I I
1
1
Grinding Grinding
I
Sieving
t C~CO, I
I I
U
aJ
L N .-c
F e
: mm
0
Fig. 1 . Flowsheet of CdCO, production by precipitation method
final product. These impurities greatly affect the quality of the inorganic cadmium pigments which may possess undesirable ydirtyycolours. All the above has necessitated investigations into a waste-free method of cadmium carbonate production which could avoid the said disadvantages.
2. THEORETICAL BACKGROUND
A CdO-C02-CdC03 system was studied by Bretsznajder i2]. He found out that the rate of reaction of CdO with gaseous C02 was very low. For instance, the extent of the reaction after 181 hours was found to be only 75 mole percent. A study on possibilities of obtaining hydrozincite, Zn,(CO,),(OH),, in the reaction of ZnO with C 0 2 in aqueous phase shown that the rate of that reaction was reasonably high [I].
487
Ca vaDour
Air
I
Burning Augmentation for water losses CdO trapping in water
scrubber
L
a,
4-
;
U a,
h
s L
a, -0
Drying in spray d r i e r
C 0
u 1
1
1
Cd CO3 Fig. 2. Flowsheet of CdCO, production by waste-free method
Taking the above into account and the fact that CdO and ZnO possess a lot of chemical similarities, a hypothesis and flow scheme of CdCO, production have been suggested by Niecko [4]. The main idea and goal of which were to modify the technology of the CdCO, production in order to: 1. Eliminate a consumption of some chemicals. 2. Reduce drastically a consumption of water both for preparation of Cd(NO,), and Na,CO, solutions, and for washing the cadmium carbonate filter-cake. 3. Eliminate some of long-lasting and labour-consuming unit operations and processes. A flowsheet of the modified process is pictured in Fig. 2. It has also been suggested that following the Henry y s Law, an increase in pressure would increase the concentration of carbonic acid species in aqueous phase which, in turn, would significantly a rate of the CdCO, formation and shorten the time of carbonation required for a high CdO-to-CdCO, conversion.
3. EXPERIMENTAL
The carbonation of CdO/water suspensions under atmospheric pressure was carried out in a vessel equipped with a stirrer, water thermostating jacket, bubbler system and point of sampling in motion at the bottom of the vessel. The bubbler system was fed with gaseous carbon dioxide taken from a CO, cylinder, through a gas regulator and then flowmeter. A suspension head in the vessel was fixed to be 34 cm. The carbonation under elevated pressures was carried out in an autoclave equipped with the same apparatus as above. A suspension head in the autoclave was fixed to be 28 cm.
488
In order to find a relationship between the CdO-to-CdCO, conversion (X) and the carbonation time (t) several series of experiments at chosen and fixed conditions were carried out. Thus, the concentration of CdO in the suspension ranged from 100 to 300 grams per dm3. Temperatures of carbonation were fixed at 293 and 343°K. Pressures ranged from atmospheric up to 1.O X lo4 hPa. The volume of samples of the carbonated suspension up taken in the motion was 50 cm3. Samples were filtered and dried at the temperature of 378°K to a constant mass. The portion of each dried sample was then heated at the temperature of 673°K for 4 11. Known masses of samples before and after heating were employed to find the weight percentage of the fixed residue consisting of CdO. Dried samples at 378°K were analysed for the CO, weight percentage by gasometric method [5]. A weight percentage of water in dried samples was calculated by substraction of the sum of CdO and CO, weight percentages from 100%. Then the CdO-to-CdCO, conversion (X) was calculated from tlie equation: X = n/n + N , where n and N are mole ratios of CdO and CdCO, in the sample. Measurements of volumes of CO, introduced into the bubbler system and then into the suspension were employed for calculations of the total degree of CO, utilization per one cycle of the carbonation process (TDU).
3. RESULTS
A. Four series of carbonation of CdO/water suspension were carried out. The conditions of experiments of each series and results obtained are sunlniarized in tables 1-5. Obtained data of experiments of series 1-4 have proved the suggested hypothesis to come true. Nevertheless, the time of carbonation is long enough and the CdO-to-CdCO, conversion has not reached the value of unity yet, though the difference is less than 0.04 as in the case of runs 1.8, 1.9, 2.6-2.9 and 4.8. The total utilization of CO, in the carbonation process is high at the beginning of the process and then decreases rapidly with increasing time of carbonation. TDU depends greatly on the rate of the CO, feedstock and the lower the rate, the higher the TDU. At tlie end of the carbonation process, TDU is small and equals to several percent only. However, TDU need not be a substantial factor if the carbonation process is modified to recycle CO,, as shown in Fig. 2. B. The CdO concentration in suspension subjected to the carbonation process has been fixed at 300 g/dm3. The rate of the COz carbonation has been fixed at 50.0 dm3/hour X kg CdO (measured at STP), and the temperature at 293°K. Tab. 1. Conditions of carbonation for series 1-4
Series 1 2 3 4
Concentration of CdO in suspension; g CdO/dm'
Average rate of C 0 2 feedstock; dm3/hr X kg CdO
Temperature of carbonation; 'K (" C)
100 100 150 300
17.1 25.1 63.9 71.9
343 (70) 293 (20) 343 (70) 343 (70)
489 Tab. 2. Results of experiments of series 1 Run
t
X
1.o 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9
0.00 1.25 4.00 9.00 20.00 25.00 32.00 43.00 5 1.oo 68.25
0.00506 0.0826 0.256 0.602 0.862 0.919 0.947 0.954 0.962 0.968
Tab. 3. Results of experiments of series 2 TDU
Run
2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9
-
74.7 75.0 79.1 52.9 42.3 33.9 25.9 21.8 15.7
t 1.00 3.78 7.00 21.00 28.00 45.00 58.00 86.50 105.75
X
TDU
0.0432 0.180 0.384 0.899 0.941 0.961 0.971 0.979 0.981
47.1 55.5
60.5 33.6 26.2 16.1 12.3 8.1 7.0
Tab. 5. Results of experiments of series 4
Tab. 4. Results of experiments of series 3 Run
t
X
TDU
Run
t
X
TDU
3.1 3.2 3.3 3.4 3.5
1.00 3.50 5.50 10.00 18.50
0.137 0.447 0.668 0.830 0.944
33.9 38.5 31.3 23.9 14.4
4.1 4.2 4.3 4.4 4.5 4.6 4.7 4.8
0.50 1.oo 2.00 3.50 5.00 7.00 18.75 21.75
0.0584 0.112 0.227 0.410 0.489 0.619 0.951 0.968
29.4 31.6 34.2 34.0 28.9 25.3 14.0 12.3
Tab. 6. Results of experiments conducted in the autoclave X at pressures; hPa X lo3
t
1
0.5 0.049 1 0.083 2 0.177 3 0.284 5 0.432 7 0.555 9 0.664 12 0.775 15 0.96 1 18 0.921
2
5
10
0.068 0.124 0.273 0.389 0.588 0.730 0.855 0.916 0.541 0.972
0.081 0.180 0.354 0.543 0.786 0.896 0.957 0.596 0.988 0.981
0.120 0.233 0.436 0.659 0.944 0.982 0.994 0.996 0.996 0.997
Three series of experiments were conducted at elevated pressures and one series at atmospheric pressure as a mark of reference. The experimental results are summarized in table 6. The obtained set of data proves that a rise in pressure in the CdO-HzO-C0,-CdCO, system increases the reaction rate of CdCO, formation and, as a result, the CdO-to-CdCO, conversion increases with increased
490 pressures for the same time of carbonation. For example, X after 9 hour-carbonation at the pressure of 1.0 X lo4 hPa is 0.994. In order to achieve X of similar value at the atmospheric pressure, the required time of carbonation would be several times greater.
4. CONCLUSIONS
On the basis of studies presented in this paper and studies carried out in the past on the same problem [3,4], it can be stated that the waste-free method of CdC0,production pictured in figure 2, is unexpectedly simple and effective, particularly when the carbonation of the CdO/water suspension is carried out at elevated pressures. Main profits of the wastexree method of CdCO, production are pointed out below: 1. The process is very simple from technological and apparatus points of view. 2. The process is waste-free. It eliminates contamination of the environment with cadmium and nitrate species. 3. The process eliminates a few tiresome and longlasting unit operations and processes, especially precipitation, filtration and washing. 4. The product, cadmium carbonate does not contain other impurities than those present in cadmium metal. 5. The process eliminates the consumption of NHO, and Na.&03 or NaACO,. 6. The process enables recycling of water. 7. The suspension after carbonation possesses the CdCO, concentration up to 8 times higher than that analogous obtained by the precipitation method. 8. The unit process of carbonation can be carried out in a wide temperature range but the best at the room temperature. 9. The hermetic carbonator with recycling of unreacted CO, gives 100% utilization of co,. 10. The carbonation process can be carried out in a wide range of CO, feedstock but the best one is that very close to the critical value. It can be achieved by means of an adequately designed CO, circulating pump. High values of the CO, feedstock guarantee achieving high value of the CdO-to-CdCO, conversion in a relatively short period of time.
REFERENCES 1 GmbH, Cadmium, System-Nummer 33, Weinheim, 1959, p. 647. S. Bretsznajder, Roczniki Chemii, Vol. 12, 1932, p. 799.
2 3 4 5
J. Niecko, et al., Polish Patent 229806, 1981. J . Niecko, Przem. Chem, 61/7,1982, pp. 193-195. Z . Minczewski, Z. Marczenko, Chemia Analityczna. Analiza Ilosciowa, Vol. 2, PWN, Warszawa, 1973, p. 325.
49 1
RECLAMATION OF CHROMIUM FROM WASTES L. PAWLOWSKI
Visiting Scientist at CSIRO, from Maria Curie-Sklodowska University, Lublin, Poland M. KOTOWSKI Maria Curie-Sklodowska University, Institute of Chemistry, 20-031 Lublin, Poland B. A. BOLT0 and R. McNEILL
CSIRO, Division of Chemical Technology, P . 0 . Box 31 0, South Melbourne, Victoria 3205, Australia
ABSTRACT With industrial development comes increasing amounts of hazardous wastes and a simultaneous depletion of raw materials. The storage of hazardous solid and liquid wastes at dump sites presents a serious health and safety risk to the community, and other disposal methods are required. It seems prudent to consider wastes as a resource, which can be detoxified and converted to usable products. One such waste from industrial effluents is chromium. The application of ion exchange for chromium reclamation from effluents is reviewed. However, the main type of chromium waste is sludges from plating plants containing up to 20% Cr20,. There are thousands of tons of this sludge in storage throughout the world. The methods of detoxification of the sludge are reviewed. The most effective method is probably the oxidation of Cr(II1) followed by washing-out of chromic acid or sodium chromate. The solution obtained in the washing operation can be directly re-used in a plating plant. The most important step of the process is oxidation of Cr(1II) to Cr(V1). The are few wet reactions which are practical for the conversion of Cr(II1) to Cr(VI). The most promising method is oxidation by air at 850-9OO0C.
1. INTRODUCTION
Among solutes present in surface and ground waters, heavy metals in their ionic forms belong to the category of pollutants that can be acutely toxic to humans, animals, and plants. Their accumulation in the tissues of the organism can result in chronic illness. Being non-biodegradable, they can be concentrated along the food chain, producing their toxic effects at points often far removed from the source of pollution. One such metal is chromium, widely used in the metal finishing and tanning industries. As these industries are usually widely dispersed, supervision and control of chromium waste disposal is often difficult. One method for dealing with chromium-bearing wastes is to precipitate the metal in an insoluble form for storage in special areas. For example, frequently chromium is precipitated as the hydroxide, and stored in ‘sealed’ land-fill sites. However, the stability of the hydroxide is being questioned and leakage of chromium from these sites would enable the toxic material to re-enter the biosphere. In some
492 countries where rigorous regulations for chromium waste disposal have been implemented handling and disposal costs have increased substantially. Economically workable chromium-ore bodies are not widely distributed in the world, and some countries lack supplies of this essential metal. Thus, the recovery of the metal from wastes is environmetally desiderable and may also be economically feasible. Three main types of chromium reclamation are discussed: - chromium reclamation from wastewaters by direct recycling; - chromium reclamation from sludges which have been stored over the years by transforming the insoluble forms into reusable compounds; and - chromium reclamation from tannery wastes.
2. CHROMIUM RECLAMATION FROM LIQUID WASTES
Two types of wastes are discharged by the metal finishing industry: exhausted plating baths and rinsing waters. The baths, which contain up to 400 g Cr03/L, are contaminated with trivalent chromium, formed when hexavalent chromium is reduced. Salts of other metals such as magnesium, aluminium, zinc, copper, nickel and iron, are formed when the metal parts being treated are dissolved to a slight extent. A strongly acidic cation exchanger can adsorb these contaminating cations from the bath. Because of the strongly oxidizing properties of Cr03, some researchers [ 1 -61 recommend dilution of the solution to prevent deterioration of the cation-exchange resin. They recommend that the concentration of Cr03 be kept below 200 g/L. A typical flowsheet for such a process is shown in Figure 1. This method, however, requires an evaporation unit, which increases the investment cost and complicates the plant’s operation. Recently, it has been found that some macroreticular resins (for example, Amberlite 200) can withstand the higher concentration chromium baths (up to 500 g Cr03/L)[7]. An evaporation stage is not required in this case. However, the bath solution is diluted upon contact with the cation exchanger bed. Firstly, the frontal and tail part of the liquor flowing through the bed is dispersed through diffusion and fingering effects and, secondly, the first part of the rinsing water contributes to the dilution. On average, the concentration of the recovered Cr03 in the treated bath (even in the best case) drops to 60-80% of its original level. The Cr03 solution can be recycled to the bath in this dilute form. An evaporation stage is not required, as the bath is often running at temperatures of 4O-8O0C, and some evaporation occurs anyway. As well as utilization of the chromium, some of the bath is “dragged out” on the treated metal parts to the rinsing tank, so replenishment of the bath with Cr03 is necessary. The concentration of the “make-up” depends on the composition of the actual bath in the process. A solution of higher concentration can be achieved most simply by the addition of some solid Cr03. This-permits simplification of the process. However, because,of the high acidity of the treated bath solution, the efficiency of cation removal is low and, therefore, the bath has to be treated often. The tails of the regeneration effluent are reused in a prelinunary regeneration. Another source of chromium is the rinsing operation, where the metal parts treated in the chromium baths are rinsed once-through with tap water. However, tlus significantly complicates the recovery of chemicals because the water contains other ions that con-
493 Chromic both
_for_2nd _lox-port -yso4 -of- regen. -
I I
Exhausted chromic both
T r e a t e d both
Is1 f r a c t i o n of regen. e f f l u e n t
( t o sewer 1
Fig. 1. Flowshcct for the renovation of chromic baths.
taminate the recovered solutions. Therefore, to have a closed-loop rinse system one strictly has to use demineralized water. The treatment steps consist of (Figure 2): - separation of suspended matter - removal of organic matter - decationization of rinse water - deanionization of rinse water For the separation of suspended and organic matter from the effluent, one of the established conventional methods can be used without any difficulty. For example, the decationization step is similar to all operations performed in the decationization of water. A strongly acidic cation exchanger has to be used because of the low pH (< 2) of the rinse water. The regeneration of the exhausted cation-exchange bed is performed in a coventional way, preferably with the cheaper sulfuric acid, but hydrochloric acid is ako recommended [S, 71. According to Shuldt and Haber [8], regeneration with hydrochloric acid eliminates the difficulty of precipitates forming when the system is refilled with fresh but hard water to compensate for losses in the closed-loop system. Because of the strong bonding of polyvalent heavy metal cations to the resin, a high dose of acid is required for regeneration. Amounts ranging from 320 g at 10% concentration [2] to 490 g at 7% concentration [3] of sulfuric acid per litre of the cation exchanger have been recommended.
494
L
Chromic both
c
Rinse tank
Repen
Courtic sodo
Id r a t e r (demineralized )
Stronply ocidie cation / sachanger
Chromic a c i d
Fig. 2. Flowsheet for the recycling of chromic rinse water constituents.
Chromates can be recovered by the use of anion-exchange resins. Deanionization can be performed with or without conversion to chromic acid. If one has small amounts of rinse water, it may be simpler not to recycle it. In this case regeneration of the anion exchanger is performed with a 4% solution of NaOH. Chromates present in the regeneration effluent are then reduced to trivalent chromic ions, and after neutralization with lime the clarified wastewater is discharged to the sewer. For conversion to chromic acid, the regeneration effluent from the anion exchanger (a mixture of Na2Cr04 and NaOH) is decationized with a strongly acidic cation exchanger. The adsorption of chromic acid onto an anion exchange bed is possible by the use of strongly basic [ 1-61 and weakly basic [6, 8-10] anion-exchange resins. Strongly basic resins achieve low leakage of Cr(V1) (< 1 mg Cr/L), but they require greater amounts of caustic soda for regeneration, as the efficiency of regeneration is lower, whereas a weakly basic resin has a higher level of leakage but consumes less caustic soda. The use of a
49 5
weakly basic resin has some advantages because the closed-loop system does not require a very low content of Cr(VI), in contrast to the requirement for discharging as wastewater. In rise tank systems parts of the metals are immersed in the rinse water, where the concentration of Cr(V1) is kept at a level of 50-200 mg/L. The overflow from the rinse tank can be treated to allow recycle of the rinse water. Significant leakage from the ion-exchange column is tolerated because the resultant increase in the recycle rate keeps chromium levels in the rinse tank at an acceptable level. From the above, one can conclude that in a closed-loop rinse system the level of leakage of Cr(V1) is not the most important item. The decrease in caustic soda consumption is the critical factor in the economics of rinse water pretreatment. Another very important parameter is the concentration of the regeneration effluent. For regeneration of the strongly basic anion exchanger in the conventional manner, a dose of 80-140 g NaOH/L of resin at 4% concentration is recommended [ 11. Furthermore, it is possible to get regeneration effluents containing up to 6 0 g Cr(VI)/L by the use of weakly basic resins regenerated with highly concentrated caustic soda [ 111. Recycling of the regeneration effluent is then possible without concentration by evaporation in cases where a hot plating bath is used. An alternative method for the treatment of chromium rinse water is neutralization with NaOH, followed by removal of the precipitated heavy metal hydroxides and adsorption of Cr, onto a strongly basic anion exchanger. The oxidation of the anion exchanger by Cr03 which can occur under acidic conditions is avoided and there is a decrease in the consumption of sulfuric acid needed to regenerate the cation exchanger. Another source of the wastewaters containing chromium is cooling blow-down water, Here the ion-exchange process may also be used for the recycling of chromates [ 121. The flowsheet of such a process is depicted in Figure 3. Another major source of chromium pollution is the wastewater of the tannery industry. However, due to the complex composition of this wastewater and its high content of organic matter, an ion-exchange method cannot be used directly for recovery of the chromium. Therefore, other more complex methods are used (Section 4 of this paper). Hence ion-exchange methods can be used for the concentration and removal of chromium from wastewater; they therefore enable one to recycle chromium compounds from wastewater and should be more widely employed for this purpose. The steps involved in such treatment of metal-finishing wastes vary considerably for the different installations. The general treatment consists of (1) separation o f grease and oil; (2) reduction of chromates; (3) destruction of cyanides by oxidation; (4) neutralization of the waste stream; and ( 5 ) separation and disposal of the sludges. The metal ions initially present in the wastewaters are precipitated as hydroxides and appear in the sludge, One of the most troublesome and costly aspects in the treatment of metal-finishing wastes is the final disposal of the metal-plating sludges that appear in step ( 5 ) . Increasing congestion of urban areas in which many of the larger plating facilities are located, together with the accompanying increases in the cost of land, add to the problems involved in ultimate disposal. A similar method of treatment is used for the decontamination of cooling blow-down water, and results in the formation of a sludge containing from 5% to 20% chromium, an amount greater than for most sludges, and its total value is significant. Thus, in the USA
496 Fresh wotei
Cooling tower Blowdown
pH o d l u s t r n e n t 1030-45
Regen effluent (forrnoke upwater)
Fig. 3. Flowsheet for the recycling of chromium from cooling blowdown water by ion cxchange.
the value of chromium discarded in chromium-bearing sludges has been estimated at about 25-39 million dollars for the year 1975 [13].
3. WET METHODS OF RECLAMATION OF CHROMIUM FROM SLUDGES
There are two ways of recovering chromium from sludges, The more specific method (Figure 4) consists of an oxidative step which converts Cr(II1) to Cr(VI), followed by a leaching step. The less specific method consists of direct leaching of chromium with a mineral acid, preferably sulfuric acid (Figure 5). Direct leaching with sulfuric acid often results in the production of chromium sulfate and significant amounts of sulfates of other metals. Nevertheless, this method is used in the tannery industry. The oxidation method is more specific as it results in the conversion of Cr(II1) into
497 Chromium bearing sludge
* Preireaimeni ( i f necessary)
Oxidation
e
Z
3
Leaching
Insoluble residual sludge
S o l u t i o n of recovered chromates
Fig. 4. General flowsheet for the recovery of chromium from sludges by oxidation followed by leaching.
C h r o m i u m bearing sludge
Pretreatment ( i fn e c e s s a r y )
Solution of r e c o v e r e d chromium sulfate
L-f-7
Insoluble residual sludge
Fig. 5. General flowsheet for the recovery of chromium from sludges by leaching.
498
chromates, which being soluble in water, are easier to separate in a purer form from the residual sludge. This method tends to be used for recovery of chromium from plating sludges. Oxidation can be carried out in an aquesous environment (wet oxidization) with the use of peroxide, persulphate or permanganate salts, perchloric acis, chlorine, ozone or fluorine. However, the most practical method is air oxidation during calcination of the chromium bearing sludge in the presence of a mixture of soda ash and lime.
3.1. Ox i d at i o n
Wet oxidation is feasible when a solution of the recovered chromates 0 s reused directly in plating. Such a method [ 141 consists of removal and cencentration of chromium by
Rinsing water ~ l O O m g / l C r " ~p,H - 2
I
and precipitation o f C r ( O H ) , by
NaOH
Ox i d a t ion by H,O,+ NaOH
Filtration
Insoluble residuaI sludge
Fig. 6 . Flowsheet for the recovery of chromium from plating rinse water by precipitation followed by oxidation.
499
reduction, precipitation and filtration followed by oxidation of Cr(II1) to chromate with HzOz (Figure 6). Chromates are separated from the residual sludge by filtration. Both water and chromates are recycled. The method is simple and easy to operate. However, the cost of chemicals used is about three times higher than the value of chromates recovered. Senoo and Shimizu [ 151 describe a method for the recovery of chromium from sludge, in which the sludge is first oxidized with Hz Oz . After filtration, the filtrate is decationized on a strongly acidic cation-exchange bed (Figure 7). The recovered solution contains about 10 g Cr03 /L. Similar techniques involving chlorine were used by Tanaka [ 161 and others [ 171, who used NaOCl acidified with HZSO4 instead of chlorine to oxidise Cr(II1) to Cr(V1). Another method for the recovery of chromium from cooling blowdown water was developed by Duffey [ 181. Here (Figure 8), the gaseous chlorine is used for the oxidation Cr(II1) to Cr(V1). Attempts to use gaseous chlorine as a selective oxidizing agent were made by many researchers. Muller and Witzke [ 191 demonstrated a method which consists of the chlorination of a suspension of chromium-bearing sludge at a temperature of
'7: Chromium b e a r i n g sludge
NaOH
HO ,,
Oxidation [Cr (OH),: N a O H . H,O,
= 10 7:63
=
I
I F i I t r o lion
",SO*
r----
'
a c i d i c ,// c a t i o n ,/ exchanger /;/Dowex 5 0 ~ x 8 )
I'
/=
Solution
of r e c o v e r e d c h r o m i c acid
Fig. 7. Flowsheet for the recovery of chromium from sludges by oxidation followed by ion exchange.
500 Make up water I1
Oxidotion Criil-DCrVI
F i It r o t i o n
residual sludge
Fig. 8. Plowsheet for the recycling of chromium from cooling blowdown water by precipitation followed by reoxidation.
20" to 80°C, followed by the separation of the residual insoluble sludge from chromates (Figure 9). These authors have also reconunended the recovery of other metals from the residual sludge. Holland [20, 211 and Deacon [12] have investigated the oxidation of Cr(lI1) to Cr(V1) by other oxidizing agents (sodium oxide, potassium permanganate, sodium persulfate, perchloric acid, sodium peroxide, ozone, fluorine) both in acidic and alkaline media. Drkos and Bahensky [ 2 3 ] developed a method of recovering Cr(V1) as BaCr2O7 by oxidation of Cr(II1) to Cr(IV) with ozone at pH > 10, followed by precipitation with a soluble barium salt to form barium chromate. Some of these methods worked quite well, but suffered from high chemical costs, handling problems, or the addition of potentially more polluting species to the sludge. Thus, wet oxidation has probably very limited practical application.
501 C h r o m i u m beormg
Gaseous CI, for oxid. O x i d o t ion 20'- 80 "c 4
IF
H P 4
A d j u s t m e n t o f pH t o I 0 - 3.0
t
II
Filtration
I
Insoluble sludge
Chromate bearing filtrate 4-8%
NaOH
Fig. 9. Flowsheet for the recycling of chromium from sludge by oxidation with chlorine followed by concentration on an anion exchanger bed.
3.2. Leaching
The leaching of sludges without oxidation is not a selective process for chromium recovery since other heavy metals are also extracted. Thus, the process may be economically feasible if the recovery of all heavy metals is desirable, In this section the recovery of chromium is considered as part of a more complex process of heavy metal recovery. Selective leaching with ammonia solution or ammonium salts causes some metals to redissolve (e.g. Ni, Cu, Zu) and form ammine-complexes. Hallowed at el. [25] have developed a method which involves leaching sludge with a solution of (NH4)zC03 at 50" to redissolve copper and nickel compounds. The levels of these compounds in the sludge are lowered from 10% to less than 1%.The insoluble residual sludge was cilcined with
502
soda ash to convert Cr(II1) to Cr(VI), followed by leaching with water to dissolve the chromates. Komoda et al. [24] have found that leaching sludge with a mixture of ammonia solution and ammonium chloride at pH 9-10 and a temperature of 150" results in the removal of Ni, Cu and Zn compounds. Insoluble residual sludge contains C r 2 0 3 . n H 2 0 . Leaching sludge with ammonia solution or ammonium salts was also discussed by other researchers [ 131. However, there is no full-scale process using this method. Matsumoto and Yukiki [25] have tried to use ammonium sulfate to precipitate chrome ammonium alum from a solution of sludge leached with sulfuric acid. Direct leaching of metals from sludges by mineral acid (usually sulfuric) has also been investigated. At a controlled pH level of "3, it is possible to dissolve most of the metals in the sludge. The insoluble residual sludge contains (in this case) mostly F e 2 0 3 .n H 2 0 and some Cr, 0 3 n. H, 0, and the filtrate contains the salts of other heavy metals.
F i l t r a t e sulfates of Z n o n d N i NaOH
t
+
H P * I
Znextraction with mixture of cyclohexanone oxime ond oleic ocid a t pH 5 7
Recovery of ZnSO,
F i l t r a t e : Sulfate of N i
i Ni extraction with mixture o f cyclohexanone o x i m e a n d o l e i c a c i d a t pH 6.7
__c
Recovery of NiSO,
Fig. 10. Flowsheet for the recovery of chromium from sludge by liquid extraction.
503 S l u d g e of h y d r o x i d e s o f
H CI
Recoverd H C I
for rediss
Fig. 11. Flowsheet for the recovery of chromium from sludge by a combination of ion exchange and liquid extraction.
The recycling of metals from the filtrate is possible only if each metal has been separated into reasonably pure salts or other usable compounds. Liquid extraction processes can be used ( e g Figure 10) to separate-out each metal salt. The advantage of the process is that only one extracting agent (a mixture o f cyclohexanone oxime and oleic acid) is used for the selective extraction of metal ions by adjustment of the pH [26]. The application of other extracting agents to separate-out metals from a solution obtained by leaching of sludges has been described [27-291. Another separation method employs ion exchange. Van Veen [26] describes an ionexchange process combined with liquid extraction (Figure I I ) , which can separate a mixture of metal ions into pure salts (FeC13, CuSO4, P b S 0 4 , Z n S 0 4 , CdS04) or metal (Cr, Ni). On an industrial scale, the process is too energy intensive because of its complexity and because evaporation of the very dilute regeneration effluents from the anion
504 exchanger would be required. Nevertheless, ion exchange may be considered as a separation method in some specific applications [30]. These process (Figures 10 and 1 I), although probably technically feasible, are unlikely to be economically viable. High capital costs and technical complexity would preclude their incorporation in small metal finishing plants. However, liquid-liquid extraction processes could be of potential use in large, central waste-treatment plants. 4. CALCINATION
A well-known commercial process for manufacturing sodium chromate or dichromate is~. roasting the chromite ore (containing aprroximately 44% of C r 2 0 3 )with soda ash (Na (NazC03j and quick lime (Cab) at 1 1OO0-125O0C. Atmospheric oxygen oxidizes Cr(lI1) to Cr(V1). The soda ash alkalizes the mixture, while the quick lime increases the roasting rate appreciably and converts the alumina and silica in the ore to insoluble calcium aluminates nad silicates. The leach solutions then require no additional processing. Since chromium-bearing sludges are precipitated with lime, no quick lime addition is needed before sludge calcination, although soda ash is required. The expected calcination reactions are given in reactions (i-iii). In these equations “M” represents metals and can stand for Ca, Cu, Fe, Mg, or Zn. Primary R e a c t i o n s
2Cr,O3+4NazCO3+3OZ
M(OH),
Heat --+
MO
Heat
4NazCr04+4COz
+ y H20
(9 (ii)
C o rnp e t i n g R e a c t i o n
+
2CrZO3 4 C a 0
+ 3O2
Heat --+
4CaCr04
(iii)
The reaction (i) results in the conversion and oxidation of chromic oxide to sodium chromate. The other metal hydroxides or hydrated oxides in the sludge are simply dehydrated by the calcination to form their corresponding oxides, as in (ii). The calcium chromate formed in the competing reaction (iii) has w lower solubility in water than sodium chromate, so leaching with water is ineffective for recovery of the chromium from the calcined sludge. The use of acid leaching for recovering chromium from the calcined sludge results in a large amount of nonchromium compounds also being dissolved. Despite this difficulty, calcination is probably the most promising method of chromium recovery from sludges [ 131. There are two problems to be solved: optimization of calcination and’, improved methods of chromate leaching. 4.1. D r y O x i d a t i o n
The efficiency of chromium recovery from sludges by calcination depends greatly on the efficiency of oxidation. There are three basic parameters which influence the effi-
505 ciency of the oxidation of Cr(II1) to Cr(V1) - the composition of the sludge as adjusted with admixture of CaO, Na2 C 0 3 and NaOH, the temperature of calcination and the time of calcination. Most experiments have been carried out at temperatures between 850" to 900°C [13, 311. Deacon et al. [31, 321 have shown by a thermal analysis study that sodium chromate formation occurs at about 860°C. However, Bahensky and Kubankova [33] have found that the optimum temperature is around 800°C. Machida and Nakagawa [34], while studying the calcination of exhausted catalyst, found that heating a mixture of catalyst containing 32.5% Cr(II1) with Na2C03 at 760°C for 4 h caused almost full oxidation of Cr(II1) to Cr(V1). Yasuda et al. [35] have found that calcination of plating sludge with a mixture of Na2C03 and CaC03 at 550-730°C for 2 h also gave good results. To elucidate the appropriate amount of CaO, Na2C03 and NaOH to be added is not an easy task because most researchers have employed widely varying amounts of Na2C03. From reaction (i) the proportion of C r 2 0 3 to Na2C03 is in the molar ratio of 1 : 1. It may be lowered by the competing reaction (iii). Machida et al. [34] give the ratio 1 :1.1. Our results [36] show that the amount of Na2C03 required depends on the type of sludge, the ratio of Cr203 to other heavy metals, and the lime content of the sludge. Generally, the amount of Na2C03 required decreases as C r 2 0 3 and CaO levels increase. However, the effect is small and does not exceed 20% of the stochiometric amount (Table 1). TABLE 1 OPTIMAL MOLAR RATIO OF Cr,O,: Na,CO, FOR PLATING SLUDGES CONTAINING DIFFERENT AMOUNTS O F Cr,O, AND CaO ~
~
~~~
Content in dry sludge, % ' Cr20, CaO Optimal molar ratio Cr,O,: Na,CO,
8
2
8 8
10:16 1 0 3
12 2
12
15
15
20
8
2
8
2
10:15
10:14
10:14 10:13
20 8
10:14 10:14
The temperature of calcination is probably the most important parameter. In all the cited cases the maximum efficiency of sodium chromate formation is at a temperature around 860°C. However, the form of the agglomerate (sinter) is greatly influenced by temperature. At temperatures above 870°C a glassy agglomerate tends to form, and the resultant leaching of chromium is a serious problem. Sometimes, even at lower temperatures such gfassy aggfomerates are formed. In all cases the oxidation efficiency is in the range 85-95%.
4.2. L c a c h i n g
The efficiency of leaching greatly depends on the particle size of the agglomerate. Therefore, before leaching, the agglomerate should be milled. Leaching with water gives good results when calcined sludges do not contain calcium, as in exhausted catalysts. If calcium is present in the pretreated sludge, the slightly soluble CaCr04 is formed during
506 calcination. Dissolution of the CaCr04 requires acidification which also solubilizes some other metals. We have devised a special two-step process [36]which produces a regeneration effluent of mainly chromate: firstly, the sludge is leaching with neutral water followed by acid leaching: and secondly, the solution is circulated through an ion-exchange system (Figure 12). The method gives a solution of NaCr0, at a concentration of about 10% containing about 0.5% Na2S04, and a residual sludge substantially free of chromium. The product solution is further processed by evaporation and crystallization, and the residual sludge is discarded. Use of the ion-exchange method greatly improves the leaching process. Although calcination followed by leaching is one of the most promising method of chromium recovery from sludges, its direct use in plants which generate sludges, e.g. small metal finishing plants, is unlikely to be economic. The construction and operation of a large central Ground c a l c i n e d sludge Water
F i I t r a tion
solution
-
of N a,Cr 0, 10% a n d
-
I
0.5%
Filtration
Regen effluent water
(product)
Fig. 12. Flowsheet for the recovery of chromium from sludge by a calcination followed by a combi a combination of gradual and ion exchange.
507
plant for processing sludges is more appropriate as it gives economy of scale. However, we suggest that a small plant specializing in the production of chemicals for small users could use plating sludges as a raw material for processing chromium compounds, mostly for plating plants and the tanning industry. It could be subsidized by sludge producers with a sum of money equivalent to the cost of sludge disposal and storage. A full economic study of the concept has yet to be completed. However, our preliminary evaluation indicates that such a plant should be more profitable than plants using other sources of chromium of the production of chromium compounds.
5. TANNERY WASTES
Chrome compounds are widely used in the tanning industry, which like the metal finishing industry consists of many small plants, which makes economic waste treatment
-
Waste shavings Moisture 17%, Cr203- 3.6 %.
-
C O D - 9 4 0 g / kg, pH 3.1 NaOH
I
( 8 0 % o n dry shavings)
Wet a i r oxidation temp.- 300 OC
1-
Liquid: 96.5% c o n v e r s i o n of C r % CrV' 2.6 g/l COD-259/1 pH-9.5
-
Acidification to pH - 3
NaOH
Fig. 13. Flowsheet for the recovery of chromium from waste shavings.
508
particularly difficult. The amount of wastes could be reduced by direct reuse of some processing liquids [ 371. Such direct reuse systems involve skimming-off the tan clumps, microscreening the waste liquid to remove heavy solids, grease skimming, and using the liquid as the pickle make-up base. Langerwerf et al. [38, 391 have used MgO for precipitation of Cr(I1I) from tannery wastewater. Precipitation with MgO reduces the Cr(II1) content below 5 mg/L and gives high sedimentation rates (0.25 m/h). The precipitated Cr(OH), obtained may be reused as Crz(S04)3 by dissolution with sulfuric acid or, after oxidation, as a chromate. Donati [40] claims the prevention of water pollution by chromium when precipitation takes place at pH 10-14 at ambient temperature. Comino et al. [41] have developed a method of chromium recovery from tannery wastewater: adjusting pH to 7-8 by addition of alkali, warming to a temperature above 35'C, and filtering with a rotating vacuum filter. The filter cake is treated with H2S04 to pH 2 . The resultant solution of CrZ(SO4), can be reused. These methods, consisting of the precipitation, sedimentation and redissolution of the precipitate of Cr(OH), in sulfuric acid, are simple and easy to implement in tannery plant plants. The incorporation of an oxidation step complicates the process. Nevertheless, much work has been carried out on the use of oxidation as one of the steps for the recovery of chromium from tannery wastewater. Another group of processes consists of the oxidation of Cr(II1) to Cr(V1) and the
NaHC03
Fig. 14. Flowsheet for the recycling of chromium from tannery wasers.
509
reuse of recovered chromium as a chromate. A typical process is depicted in Figure 13. The waste shavings from hide treatment, after mixing with NaOH, are oxidized, and the resultant chromate separated by ion exchange [42]. A disadvantage of the process is the clogging of the ion exchange bed by organic material in the chromate liquor. This problem could be partially overcome by the use of magnetic ion exchange resins, which have been used successfully for the treatment of a highly turbid sewage effluent [43]. Another approach to the recovery of chromium from tannery wastewater was developed by Cot and Crane11 [44]. The wastewater, after oxidation with H 2 0 z in alkaline medium, is treated for chromate removal by the use of the anion exchanger Dowex 1 (see Fig. 14). During regeneration, when H 2 0 z is used in an acidic enviroment to reduce Cr(V1) to Cr(III), as well as clogging problems with organic matter, a precipitate may from which blocks the bed when conventional resins are used. The use of a magnetic resin system would improve the performance of the process [43]. A further technique used in the recovery of chromium from the tannery wastewater is incineration, during which Cr(II1) is oxidized to Cr(V1) by O 2 [45, 461. It is necessary to also add NaOH and Na2C 0 3 or NaI03 and H, SO4 [47]. These methods will probably be used on an industrial scale because the equipment is simple and not-specialized. Their use should prevent chromium in tannery wastewaters from entering the environment.
6. OTHER SOURCES OF CHROMIUM
Other important sources of chromium are slags and furmes in the steel industry. They can contain up to 10%chromium. Although disposal of chromium from these sources can be decreased by improvements in process operation and design [48], some chromium waste would still remain. One of the well established techniques could be used to recover chromium from slags and fumes. Special recovery methods include reduction with coke in an ore furnace [49], and the use of CO or H2 as a reducing agent [SO]. In these cases, a mixture of iron and chromium with other metals is produced. This “ferro-chrome” can be used in the production of chrome steel alloys.
I . SUMMARY
The main purpose for treating wastes containing chromium is to prevent the spread of chromium throughout the environment. Recovery of chromium from liquid and solid wastes seems to be a complementary task to environment protection because the reusable material could improve the economics of waste treatment. Furthermore, such recovery methods help prevent exhaustion of non-renewable raw materials, and conserve existing resources for future generations. The easiest method for the recovery of chromium is by direct treatment of waste waters by ion exchange which allows the separation of chromium as a concentrated solution which may be recycled. Removal of chromium by precipitation, although it minimizes water pollution, creates new problems because the hazardous compounds must be stored and other opportunities for contamination of the environment then arise.
510
Several methods for the recovery of chromium from waste waters are described. The methods w h c h are closet to industrial application are calcination and liquid extraction. However, it is difficult to apply these at the site of sludge production, and it seems more feasible to build a central plant for processing such sludges. The economics of plant which produce chemicals from wastes could be justified if industries with waste sludge were to pay the equivalent of the cost of storage of sludge to such processing plants. The protection of the environment depends partly on technical solutions which provide cheaper methods of processing wastes, and partly on an integrated approach which combines legislative action in the form of anti-pollution regulations with the offer of tax incentives or subsidies for waste recovery plants. The community would thus benefit from an improved environment and from the conservation of a scarce resource.
REFERENCES 1. Kunin, R. Ion exchange for the metal products finisher. Prod. Finish. (Cincinnati), (April-June) 1969, 18-28. 2. Stomquist, D. M. and Reents, A. C. Removal of cations from chromic acid solutions. Proc. 6th Ind. Waste Conf., Series 76, Purdue University, Lafayette, Ind., 1952, pp. 289-293. 3. Costa, R. 0. Regeneration of chromic acid solutions by ion exchange. Ind. Eng. Chem., 42, 1950, 308-3 11. 4. Helser, J. C. Recovery and treatment of metal finishing wastes by ion-exchange processes. Proc. 21st Ann. Water Conference, Eng. SOC.West Pa., Pittsburgh, 1960, pp. 89-103. 5. Kunin, R. Ion-exchange process in the plating and allied industries. Electroplating, (Jan.-April), 1953, pp. 176-184. 6. Gold, H. Metal finishing wastes. Ion Exchange for Pollution Control, Vol. I, ed. by C. Calvin and H. Gold. CRS Press, Boca Raton, 1979, p. 173. 7. Pawlowski, L. and Barcicki, J. Cation exchangers as adsorption media in recycling of concentrated chromium plating baths. Conservation and Recycling, 1, 1977, 289-291. 8. Schuldt, A. A . and Haber, R. R. Chromate and water recovery from tin plating operation using ion-exchange technologies. Proc. 37th Int. Water Conf., Eng. SOC.West Pa., 37, 1976, p. 29. 9. Pawlowski, L. Ion Exchange in the recycling of plating effluents. Effl. and Water Treat. J., 20, 1980,581-585. 10. Pawlowski, L. Klepacka, B . and Zalewski, R. A new io-exchange method for recovering highly concentrated solutions of chromates from plating effluents. Nuclear and Chemical Waste Management, 2, 1981, 43-51. 11. Pawlowski, L. , Klepacka, B. and Zalewski, R. A new method of regeneration of anion exchangers used for recovering chromates from wastewater. Water Res., 5, 1981, 1153-1156. 12. Seward, R. B. Chromate removal by ion exchange. Proc. Int. Water Conf. Eng. SOC.West Pa., 1978, 38, 201-210. 13. Hallowell, J. B., Bartlett, E. S. and Cherry, R. J. Ammoniumcarbonate leaching of metal values from water treatment sludges. EPA Report-600/2.77-105, June 1977. 14. Pawlowski, L. Recovery of chromium from wastewater. Polish Pat. 206, 816 (1978). 15. Senoo, Y. and Shimizu, Y. Treatment od sludge containing chromium. Japan Pat. 75, 105, 599 (1975). 16. Tanaka, T., Tanaka, M. and Tanaka, T. Recovery of useful metals from waste sludge in metal coating. Japan Pat. 79, 138, 801 (1979). 17. Recovery of valuable metals from metals plating sludges. Japan Pat. 80, 107, 742 (1980). 18. Duffey, J. C. Chromate recovery process. US Pat. 4, 318, 788 (1982). 19. Muller, W. and Witzke, L. Processing nonferrous metal hydroxide sludge wastes. US Pat. 4, 151, 257 (1949).
51 1 20. Holland, M. E. Removal and recovery of hexavalent chromium from industrial effluents - A literature survey. SAT-786, Goodyear Atomic Corporation, Piketon, Ohio, February 21, 1975. 21. Holland, M. E. Possible methods for recovery of chromium for RCW-produced sludge, GAT-T-22.6, Goodyear Atomic Corporation, Piketon, Ohio, July 5, 1974. 22. Deacon, L. E. Chromium recovery from RCW-produced sludges by open-air calcination, GAT-T-2326, Goodyear Atomic Corporation, Piketon, Ohio, March 31, 1975. 23. Drkos, F. and Bahensky, V. Recovery of chromium compounds from waste sludge after disposal of spent electroplating baths. Czech. Patent 130, 794 (1969). 24. Komoda, S., Yamazaki, K., Yasuda, K. and Natsukawa, K. Extraction of heavy metals from industrial waste. Japan Pat. 77, 152, 803 (1977). 25. Matsumoto, T. and Yukiki, K. Preparation of ammonium chromium alum from Chromium plating sludge. Chiba-Ken Kikai Kinzoku Shikenjo Kenkyu Hokoku, 1977, 24-30. 26. van Veen, F. Recycling of complex heavy metal wastes by solvent extraction and ion exchange as a contribution to the solution of environmental problems. Conservation and Rec., 3, 1980, 46 1-467. 27. Mueller, W. and Witzke, L. Treatment of nonferrous metal hydroxides sludge residues, German Pat. 2, 841, 271 (1980). 28. Cornwell, D. A., Westerhoff, G. P. and Chine, G. P. Batch feasibility testing of heavy metals removal from wastewater sludge with liquid-ion exchange. Mid.4. Ind. Waste Conf., (Proc.), 12th, 1980, pp. 111-119. 29. Kawakami, N., Temma, S . and Kato, T. Recovery of iron and chromium from chromium plating sludge. Japan Pat. 7, 611, 093 (1974). 30. Bolto, B. and Pawlowski, L. Reclamation of wastewater constituents by ion exchange. Parts I to VII. Effluent Water Treatment Journal (in press). 31. Deacon, L. E., Holland, M. E. and Kaplan, R. I. Chromium recovery from sludges produced from chromate-inhibited cooling water. Internal Report, Goodyear Atomic Corporation, Piketon, Ohio, 1975. 32. Deacon, L. E., Holland, M. E. and Kaplan, R. I. Chromium recovery from sludges produced from chromate-inhibited cooling water. Proc. 3rd Environ. Prot. Conf., 2, 1975, 511 -526. 33. Bahensky, V. and Kubankova, E. Removal of sludge from purification plants. Kovoze Ochr. Meter, 17, 1973, 81-83. 34. Machida, T. and Nakagawa, M. Recovering of copper oxide and alkali metal chromate from copper chromate catalysts. Brit. Pat. 1, 223, 734 (1968). 35. Yasuda, K., Komoda, S . , Natsukawa, K. and Yamazaki, K. Recovery of chromium in electroplating sludges. Kagaku Kojo, 1977, 21 (2), 71-74. 36. Bolto, B. A., Kotowski, M. and Pawlowski, L. Recovery of chromium from plating wastes. Intcrn. Symposium on Industrial and Hazardous Wastes, Philadelphia 5-1 1, March, 1983. 37. Maire, M. S . Chrome and sulfide conservation. Leather Ind., 10, 1981, 22-27. 38. Langerwerf, J. S . A. and De Wijs, J . C. Precipitation and reuse of trivalent chromium. Leder, 28 ( l ) , 1977, 1-9. 39. Langerwerf, J. S . A., De Wijs, J . C., Pelckmans, H. H. A. and Koopman, R. D. Precipitation and recycling of trivalent chromium. Rev. Tech. Ind. Cuir, 69 (5), 1977, 154-156, 158-163. 40. Donati, M. Recycling processes of tannery wastes. Cuoio, Pelli, Mater. Concianti, 54 (3), 1978, 4 19-429. 41. Comino, P., Cortinouis, M. and Tacchini, C. Continuous recovery of chrmium from tannery waste water. AES, 2 (6), 1980, 83-85. 42. Okamura, H. and Shirai, K. Recovery of chromium from shavings by wet air oxidation. J. Amer. Leather Chemists Assoc., 71 (4), 1976, 173-179. 43. Bolto, B. A,, Dixon, D. R., Priestley, A. J. and Swinton, E. A. Ion exchange in a moving bed of magnetized resin. Prog. Water Tech., 9, 1977, 833-844. 44. Cot, J. and Granell, J. E. Improvement of tannery wastewater, Part I. Recovery, concentration and reuse of chromium from manufacturing process. AQEIC Bol. Tec., 32 (lo), 1981, 225-247. 45. Jones, B. H. Chromium recovery through incineration of liquid and solid tannery wastes - the ultimate solution. J. Amer. Leather Chem. Assoc. 74 ( l l ) , 1979, 395-403.
512 46. Holloway, D. F. Recovery and separation of nutritious protein hydrolysate and chromium from chrome leather scrap. US Pat. 4, 100, 154 (1978). 47. Yazykov, V. K., Sirenov, V. 1. and Maksimova, L. G . Detanning collagen-containing wastes from production of chrome leathers. Mater. Nauchn Konf., Vost.-Sib. Tekhnol. Inst., Sckts. Khim.Tekhn., l l t h , 1972. 48. Masafumi, M., Sano, N. and Mastushita, Y. Chromium recovery from chromium-containing slags. Conservation and Recycling, 4, 1981, 137-144. 49. Higley, L. W., Neumeie, L. A., Fine, M. M. and Hartnian, J . C. Devcloprnent of a pyronietallurgica1 technique to recycle stainless stcel wastes. Conservation and Recycling, 3, 1979, 53-62. 50. Rizescu, C., Ursu, V. and Grecu, L. Recovery of alloying elements from the scale of high speed steel. Cercet. Met., Inst. Cercet. Met., Bucharest, 14, 1973, 153-162.
513
APPLICATIONS OF SELECTIVE ION EXCHANGE TO RECOVER MgNH4PO4 FROM SEWAGE
L. LIBERTI, A. LOPEZ, R. PASSINO
Istituto Ricerca Sulle Acque4JV.R. 5, Via F. De Blasio, 70123 Ban, Italy
ABSTRACT Two case histories for the application of the RIM-NUT process to remove and recover a magnesium ammonium phosphate fertilizer from wastewater of domestic and zootechnical origin are described. Ammonium and phosphate ion concentration in municipal secondary effluent is reduced to acceptability limits through tertiary treatment by selective ion exchangene.The exhausted resins are regerated with neutral 0.6 M NaCl, from which MgNH,PO, m recovered by precipitation. Anaerobically treated pig farm factory effluent is precipitated to yeld MgNH,PO, and the filtrated effluent polished by aerobic treatment, followed (optionally) by selective ion exchange.
1. INTRODUCTION
A major challenge to those concerned with cleaning and preservation of the environment arises from the increasing need to recycle nonrenewable resources, at the same time learning more about the ultimate fate of pollutants removed during wastewater treatment. This is particularly true for methods of removing P and N compounds which, together with C compounds, are the major contaminants in all municipal and many industrial wastewaters. P compounds (essentially orthophosphates) are usually removed by chemical precipitation (with lime, alum, Fe, etc.), with production of worthless by-products (hydroxylapatite and similar highly insoluble phosphates) [l]. The latter, along with the accompanying considerable amount of other chemical sludges (carbonates, metal hydroxides), continue to exhaust the readiness of the environment to sludge disposal (landfilling), so that the final fate of P through either liquid or solid dischargeto the environment has still to be assessed [2, 31. Similarly, N compounds (essentially ammonia) are quite often converted biologically and discharged as nitrates [4].Very few treatment plants are effectively affording the less Cost-effective and energy consuming biological denitrification, which would eventually allow them to release free nitrogen to the atmosphere [5]. This again means wasting of yaluable material (ammonia), converted (at least partially)
514 into harmful compounds (nitrates, nitrites, chloro-nitroso-amino compounds potentially carcinogenic) during its travel through the environment [6]. The RIM-NUT is a new physicochemical process for removal of ammonia and phosphates from wastewater, with recovery of MgNH$O,, a slow release, premium quality fertilizer [7,8]. It provides a cost-effective alternative to the above mentioned methods, ensuring both an affordable way for ultimate entrance of N and P into the environment and the recycling of precious nutrient material from unlimitedly renewable raw source. This paper describes two typical applications, among the several possible, of the FUMNUT process, namely to municipal secondary effluent and to industrial wastewater of zootechnical origin.
2. PRIOR ART
Worldwide experience in nutrient control needs and technology, on the basis of data collected from full-scale plants, has been reviewed during an USAIUSSR Symphosium held in 1978 [9]. Apart from biological methods, such as the well known nitrification-denitrification for ammonia, and the so-called "luxury uptake" of P [ l o ] obtained occasionally during activated sludge treatment of wastewater at San Antonio [l 11, Forth Worth [12], Baltimore [13], Los Angeles [14], Seneca Falls [15] and Johannesburg [16], several physicochemical methods are available to remove N and P compounds from wastewater. Air stripping and break-point chlorination are effective methods for ammonia removal, particularly under severe climatic conditions, although the potential agronomic value of this species is irremediably lost [ 171. Selective ion exchange, on the contrary, permits the ammonium ion to be removed and concentrated quantitatively, so that its recovery may become economically attractive. After extensive laboratory and pilot-plant attempts [18-221, two large-scale installations, at Upper Occoquan, Virginia (85,000 m3/d) [23] and Tahoe-Truckee, California (22,500 m3/d) [24], both built in 1978, are yielding high removals of ammonia (- 90%) with the use of a natural zeolite (Clinoptilolite) in the tertiary treatment of municipal wastewater. The ion exchange resin is regerated by a merry-go-round procedure with NaCl solution, from which ammonia is recovered by stripping-absorption and sold as a 40% (NH,),S04 liquid fertilizer. The low value of this latter, however, contributes little in cutting down the capital and operational costs for ammonia removal in these plants, substantially higher than those of the biological treatment. Other cationic resins (Phdipsite, Mordenite and related zeolite-type, synthetic and natural) have been studied recently with the aim of reducing media attrition during backwash and fixed-bed clogging by suspended solids [25-281. Among physicochemical methods, selective ion exchange for P removal has also received increasing attention in recent years. The method in question involves either organic or inorganic anion exchangers [29-361. However, very little experience, if any, has been gained nowadays in full-scale application of selective anion exchange for phosphate removal from sewage. As far as we know, a combined removal of both ammonium and phosphate ion from
515 wastewater by selective ion exchange, with the recovery of an ammonium phosphate fertilizer, has never been evaluated in detail, as it is in the RIM-NUT process. On the other hand, due to the high agronomic value of magnesium ammonium phosphate, substantiated in specialised literature [37, 381, the use of it as a slow release fertilizer has been suggested as early as in 1857 [39] and its precipitation as a method to remove scale formers from sea water in evaporation plants has been proposed by W. R. Grace Co. in the early '60s [40, 411 and extensively studied and patented since then [42-491. The cost of chemicals involved, however (ammonia and phosphoric acid), made this descaling treatment economically unattractive. Later, precipitation of magnesium ammonium phosphate was investigated as a means to eliminate phosphates in two troublesome sidestreams from conventional wastewater treatment, i.e., waste activated sludge and digester supernatant [50]. Nowadays, increase of cost of fertilizers following the 1973 oil crisis, inflation wave, growing world demand for fertilizers and limitation of raw material supplies have enhanced the importance of the fertilizer as a component of cost of planting. This, coupled with the need fo fmd more affordable and cost-effective alternatives to remove ammonia and phosphates during wastewater treatment, makes the RIM-NUT process for removal and recovery of magnesium ammonium phosphate from wastewater technically and economically feasible.
3. DESCRIPTION OF THE RIM-NUT PROCESS
After more than 7 years of research, with approx. 50 anionic and cationic resins investigated, including natural zeolites, the RIM-NUT process has been developed at IRSA/CNR of Italy to remove and recover ammonium and/or potassium and/or phosphate nutrient ions from wastewater [51, 521. In its basic configuration, the RIM-NUT process has the following items (Fig. 1): - two ion exchange resins, cationic and anionic in series, treat the wastewater (e.g., municipal secondary effluent) to remove residual suspended matter by fdtration, bioresistant organics and viable organisms by adsorption, and ammonium and/or potassium and/or phosphate ions by selective ion exchange, - neutral NaCl solution at sea water concentration (0.6 M) is used to regenerate both resins, - magnesium ammonium phosphate and/or magnesium potassium phosphate salts are precipitated from resin regeneration eluates, which can then be reused. When the wastewaters, or even side-streams produced during the treatment process, contain concentrated amounts of one of the above mentioned nutrient ions, the corresponding pre-concentration operation through selective ion exchange may be'omitted and the nutrient ion(s) precipitated directly from wastewater (Fig. 2). Depending on local limits for discharge of nutrients, and considering the water solubility of these phosphate salts (approx. 160 mg/l as MgNH$O, * 6H20 and 220 mg/l as MgKPO, . 6H,O at 25"C, influenced by pH and particle size), the final effluent precipitated may receive further polishing treatment by ion exchange (dottedlinesin Fig. 2). More often, an intermediate situation may occur, where the wastewater under con-
516
SELECTIVE REN0VAT E D EFFLUEN (N, K, r
oil.) ..
L
Ly
I4
I
A
w 0
w
PR ECI PITAT ION
+
MgNH4P04 Mg KP04 Fig. 1. Basic scheme of the RIM-NUT process
Fig. 2. Modified scheme of the RIM-NUT process
sideration, or even its side-streams, contain different concentrations of various nutrients. This holds, for instance, for the wastewater from pig factories, where concentrations of up to 1000 ppm N-NH, and 400 ppm K, (that of P-PO, being 10 ppm) occur; or, similarly, in the complete biological treatment (activated sludge with nitrificationdenitrification) of municipal sewage, where the final effluent still contains approx.
-
517 10 ppm P-PO, and virtually no N, but much NH, and/or phosphates may still be found in the supernatant solution from sludge digestion. In such cases the pre-concentration involving ion exchange resins is applied only to the diluted nutrient species, while the concentrated one(s) may be precipitated from the wastewater. One may indicate respectively as < 5 or 2 15 mol/m3 the concentration limit for each nutrient species at which pre-concentration with the use of ion exchange is still (or is no longer) required; in the intermediate range any decision should be reexamined. Various types of ion exchange resins may be used for selective removal of said nutrient ions depending on the nature of the wastewater. N&+ and K' cations may be exchanged selectively by some zeolites either synthetic or natural (Clinoptilolite, Phillipsite, etc.). Porous, strongly basic anion exchange resins, or even inorganic materials such as activated alumina, prove useful for phosphate removal from wastewater.
4. EXAMPLES OF APPLICATION OF THE RIM-NUT PROCESS
There are two case histories to illustrate our general comments. (A) Municipal wastewater represents a more general case of wastewater with appreciable concentrations of both ammonium and phosphate nutrient ions. After primary settling, aerobic bio-oxidation by activated sludge and secondary settling, a typical southern Italian municipal secondary effluent has the composition shown in Tab. 1. As the Italian admissible concentration levels are 10 or 0.5 ppm P and 32 or 10 ppm N for discharge into sea or lakes, respectively, a tertiary treatment is further required. A 200 m3/d RIM/NUT plant was built for demonstration purpose at West Bari Sanitation Station in 1982 to treat that effluent (general plant view in Fig. 3). Two cationic columns (C1 and C2, each containing approx. 0.45 m3 of a natural zeolite in Na form) and two anionic columns (A1 and A2, with 0.42 m3 each of a suitable anionic resin in C1 form) ensure continuous tertiary treatment of the municipal effluent (Cl-A1 are in service when C2-A2 are being regenerated, and vice versa). The final effluent which is obtained steadily is virtually free from nutrients (average removal for both ammonium and phosphate ions ranges from 85 to 95%), with a remarkably lower content of suspended matter, soluble organics and viable organisms, so that its final acceptability to discharge is enormously improved. Adsorption of suspended matter and of COD depends on the influent composition, although mean experimental values vary between 40 and 60% for both parameters. Furthermore, intermittent analyses show that it is possible to achieve even 99% removals of colibacteria and streptococci. After exhaustion, the ion exchangers are backwashed countercurrently and regenerated by neutral 0.6 M NaCl solution, which restores their ion exchange capacity, desorbs organics and kills a great part of the viable organisms. Resin regeneration occurs in parallel: three consecutive fractions (eight BV (bed volumes) each) are used for the cationic resin, coming from the reservoirs SC1, SC2, and SC3, and collected into SC4, SC1 and SC2, respectively; two consecutive fractions (two BV each) regenerate the anionic resin, pumped from SAl and SA2 and collected into SC5
518 Tab. 1. Average composition of West Bari secondary effluent (mg/l) Chlorides (as C1) Alkalinity (as CaCO,) Sulphates (as SO,) Phosphates (as P) Nitrates (as NO,) (as NO,) Ammonium (as NH,) Potassium (as K) Sodium (as Na) Calcium (as Ca) Magnesium (as Mg) BOD, (asO,) COD (as 0,) PH Fecal coliforms (MPN/100 cc) I
161 415 38 11 0 0 60
30 130 30 6 32 107 7.4 170,000
Fig. 3. View of Bari RIM-NUT plant
and SA1 tanks, respectively. Only the head fraction. of the regeneration streams (i.e., the first eight BV of the cationic eluate collected in SC4 and the first two BV of the anionic eluate collected in SC5), which contain 70 to 90% of the nutrients exchanged, are subject to precipitation treatment in order to recover magnesium ammonium phosphate. For this purpose the pH of the cationic eluate headfraction in SC4 is increased to approximately 9, using Na2C0,, MgCO, or a similar base, so that precipitation of Ca
519 Tab. 2. Summary of costs for production o f 5000 t/y of MgNH,PO, by the RIM-NUT ion exchange process from a 40,000 m3/d municipal treatment plant* Expenditures - Investment - amortization** -O&M Total running costs Revenue - from sale of fertilizer*** - from eutrophication control" Total revenues Net profit
(m $/Y) $ 2.18 M
0.42 2.36 2.78
($ 0.23/m3)
2.50 1.20 3.70 0.92
($ 0.31/m3) ($ 0.08/m3)
* 300 working days per year
** 16% per year for civil (25 ys) and electromechanical (10 ys) works 500/t MgNH,PO, 100% $ 0.1/m3 of effluent treated
*** $
"
and metals eventually occurs. The supernatant solution is then transferred to SC5, mixed with the anionic eluate headfraction and by further addition of Mg or/and PO, compound(s) to obtain a stoichiometric ratio Mg : NH, : PO, = 1 : 1 : 1 , at a pH between 8.5 and 9, quantitative precipitation of MgNH,PO, . 6H,O is achieved. The latter compound may be converted into MgNH,PO, . H,O by 1 h digestion at 90°C and then separated from the liquid phase, which is returned into SC3 for further use. Large crystals of magnesium ammonium phosphate, highly seetable and easily filtrable, with a chemical purity of -98%, are obtained in each precipitation cycle, whereas residual concentrations of 10 ppm P-PO, and 16 ppm N-NH, are found in the filtrate. From the preliminary results obtained the favourable economics resumed in Tab. 2 are expected for this application.
-
-
(B) After primary settling, anaerobic digestion and secondary settling, a typical effluent from a pig factory (approximately 30 m3/d discharged during daytime) had the composition shown at point 2 in Fig. 4.This effluent was then submitted to extended aerated lagooning (Ll-L2-L3 in Fig. 4; the overall capacity of the lagoons being 4500 m3/d for a residence time of 5 months), after which NH; concentration decreased from 714 to 126 pprn N-NH, (82% removal). Final sedimentation (point 3 after SED-F) brought about a further decrease in nutrient concentration to 149 ppm N-tot and 62 ppm P (86% and 88% overall removal respectively, as compared to raw sewage composition, point l ) , still unacceptable to public discharge. In the RIM-NUT process the extended aerated lagooning was replaced by a precipitation-sedimentation treatment (reactor PR in Fig. 5 , with 30 m3 net capacity, operated discontinuously during night-time), which, after filtration, ensures a degree of removal for both nutrients comparable with that achievable with the lagoons, including an average 78% abatement of COD. The filtered effluent is then oxidized aerobically (OX in Fig. S ) , so that further removal of COD, P and N occurs for bacterial growth. The final effluent may then be improved, or not, in a further treatment by selective ion exchange as described in example (A).
520 -
8.6 UH ST 42120 sv 34730 coo 481 14 1064 N tot N-NH4 693 ass. NO2 ass. NO3 P 530 HC03 8 5 4 0
--
-
8.2 7620 3270 5346 1008 714 ass. ass. 460 i940
1
-
-
-
8.4
8.3
1741 400 784
1720 176 126
7.8 1370 !loo 895 149 112 31 906 62 525
T
1
7
I
OIG - AN SLUDGE Fig. 4. Sheet and chemical analyses after anaerobic digestion and extended aerated lagoon treatment of a pig factory effluent (GR - screening grid; SED-I - primary sett1er;DIG-AN - anaerobic digester; L1,2,3 - lagoons; SED-F - final settler)
LiJ ’$ COO
-
14940
Na2C03
Fig. 5. Flow-sheet and chemical analyses after anaerobic digestion and RIM-NUT treatment of a pig factory effluent. (PR - chemical precipitation reactor; F - filter; OX - aerobic biological treatment)
521 Tab. 3. Mass and economic balance of chemicals during the treatment o f the anaerobic effluent from a pig factory by the RIM-NUT process (plant size 30 m3/d) A) Chemicals consumed daily unit cost ($/Kg) 0.50 H,PO, 75% MgCO, 89% 0.15 Na,CO, 0.14
B) Fertilizer recovered daily MgNH,PO; H,O 1.10
kg/d 286 216 100
$/d 142 33 14
-
-
602
189
376
414
N.B. It. Lit. 1400/$
As shown by the data reported in Tab. 3, the value of the magnesium ammonium phosphate fertilizer recovered in this plant covers largely the daily cost of chemicals consumption and may contribute significantly to the amortization of the plant.
5. CONCLUSION
Among physico-chemical methods for wastewater treatment, selective ion exchange provides an effective technique for removal of nutrients (i.e., ammonium and/or potassium and/or phosphate ions) virtually to any desired limit. Furthermore, after resin regeneration nutrient ions are found in concentrated solutions from which their recovery may become easily feasible and economically attractive. This is so, in particular, with the RIM-NUT process, in which a combined use of selective ion exchange and/or chemical precipitation makes it possible to remove and recover nutrients from sewage as magnesium ammonium phosphate, a highly valuable, slow release fertilizer. Practical experience gained at two installations (namely a 200 m3/d tertiary treatment of a municipal effluent and a 30 m3/d treatment of anaerobic effluent from a pig factory) shows that the application of the RIM-NUT process may eventually turn wastewater treatment into a profitable operation. Finally, by a proper design of the operating conditions, the process may be applied to remove and recover nutrients from other wastewaters of industrial origin.
ACKNOWLEDGEMENTS
The assistance of Dr. G . Boari during design as well as that of Mr. N. Limoni and Mr. C. Longobardi during construction, automation and operation of the West Bari plant is sincerely appreciated.
522 REFERENCES
1 U.S. Env. Prot. Agency, Process design manual for phosphorus removal, Oct. 1971. 2 D. M. Cavagnaro, Phosphorus removal in sewage treatment, Vol. 1, 1964-1976, and Vol. 2, 19771979. (A bibliography with abstracts), Gov. Rep. Announce Index (US) 7 9 (26) (1979), 233. 3 E. F. Barth, H. D. Stensel, International nutrient control technology for municipal effluents, JWPCF 5 3 , 1 2 (1981) 1691-1701. 4 D. M. Cavagnaro, Nitrogen removal in sewage treatment systems, Vol. 1, 1964-1977, and Vol. 2, 1978-1979. (A bibliography with abstracts), Gov. Rep. Announce Index (US) 79 (24) (1979) 192. 5 R. W. Wilson, K. L. Murphy, P. M. Sutton, S . L. Lackey, Design and cost comparison of biological nitrogen removal processes, JWPCF 53, 8 (1981) 1294-1302. 6 D. J. De Renzo, Nitrogen control and phosphorus removal in sewage treatment, ISBN 0-8155-0711-9, Park Ridge, N. J.: Noyes Data Co., 1978, 704 pp. 7 L. Liberti, Ion exchange advanced treatment to remove nutrients from sewage, in Physicochemical Methods for water and wastewater treatment, L. Pawlowski ed., Elsevier Sci. Publ. Co., Amsterdam, 1982,225-37. 8 L. Liberti, G . Boari, R. Passino, AWT by ion exchange, Effl. Wat. Treat. J., July 1982,253-7. 9 U.S. EPA, Washington, D.C., Seventh Symposium o n Advanced Treatment of Biologically Treated Effluents Including Nutrients Removal, USA/USSR, Moscow, USSR, Nov. 12-1 3, 1978. 10 J. B. Carberry, M. W. Tonney, Luxury uptake of phosphate by activated sludge, JWPCF 45 (1973) 2444 -62. 11 D. Vacker, C. H. Connell, W. N. Wells, Phosphate removal through municipal wastewater treatment at San Antonio, Texas, JWPCF 3 9 , 5 (1967) 750-771. 12 J. L. Witherow, Phosphate removal by activated sludge, Proc. 24th Ind. Waste Conf. Purdue Univ. Ext. Ser. No. 135 (1969) 1169-84. 13 W. F. Milburg et al., Operation of conventional activated sludge for maximum phosphorus removal, JWPCF, 4 3 (1971) 1890. 14 R. D. Bargman et al., Nitrogen-phosphate relationship and removals obtained by treatment process at the Hyperion treatment plant, in Adv. in Water Pollution Research, Proc. 5th Intern. Conf. on Wat. Pollut. Res., I, 14 (1971) 1-14. 15 G. V. Levin, G . J. Topal, A. G . Tarnay, Operation of full-scale biological phosphorus removal plant, JWPCF 47 (1975) 557-90. 16 H. A. Nicholls, D. W. Osborn, Bacterial stress: prerequisite for biological removal of phosphorus, JWPCF 51 (1979) 557-69. 17 U.S. Environmental Protect Agency, Process design manual for nitrogen control, Oct. 1975. 18 Battelle-Northwest/South Tahoe Public Utility District, WW Ammonia removal by IE, USEPA, Water Pollut. Contr. Res. Serv. Rep. No. 17010 EC2 02/71, 1971. 19 P. W. Johnson, J. Mc N. Sieburth, Ammonia removal by selective ion exchange, a back-up system for microbiological filters in closed-system aquaculture, Aquaculture 4 (1974) 61-8. 2 0 J. H. Koon, W. J. Kaufman, Ammonia removal from municipal wastewater by ion exchange, JWPCF 47 (1975) 448-65. 21 S. E. Jorgensen, 0. Libor, K. L. Graber, K. Barkacs, Ammonia removal by use of Clinoptilolite, Water Res., 1 0 (1976) 213-24. 22 R. C. Polta, R. W. De Fore, W. K . Johnson, Evaluation of physical chemical treatment at Rosemount, USEPA Rep. No. EPAd00/2-78-201, 1978. 23 G. A. Gunn, AWT plant makes wastewater potable, Wat. Wastes Eng. 16, 11 (1979) 36-44. 24 0. R. Butterfield, J. Borgerding, Tahoe-Truckee Sanitation Agency: the first three years, TTSA internal report, 1981. 25 J. R. Klieve, M. J. Semmens, An evaluation of pretreated natural zeolites for ammonium removal, Water Res. 14, 2 (1980) 161-8. 26 D. Sherman, R. J . Ross (Union Carbide Co), Phillipsite-type zeolites for ammonia adsorption, US. Patent 4, 344,851 (Aug. 17, 1982). 27 Z. CsikosHartyani, V. Olaszi, Model experiments for wastewater treatment on the system zeolite/ heavy metal ions, Stud. Environ. Sci (1982) 285-90. 28 P. Ciambelli, P. Corbo, C. Porcelli, NH: removal by Phillipsite tuff in relation to acquacultural systems, Intern. Conf. “Zeo-Agriculture ’82” Rochester, USA 1982.
523 29 I. S. Singh, K. K. Jain, G. Prasad, V. N. Singh, Ion exchange for phosphate rich effluents, Proc. Ion Exch. Symp., Feb. 23-25,1978, Barnagar, India, 60-63. 30 Z. Czamy, J . Oszczudlowski, K. Wrbbel, W. Jagielska, Use of ion exchanger to remove phosphate ions from wastewater, Przem Chem 57 (10) (1978) 530-2. 31 S. E. Jorgensen, Renoval of phosphorus by ion exchange, Vatten 3 (1978) 179-182. 32 T. Takeuchi, M. Tsukakawa, R. Kimoto, Phosphate removal from wastewater, Jpn Kokai-Tokkyo Koho 79 149,261 (22 Nov. 1979). 33 H. Sato, S. Shigeta, N. Mochisuki, Treatment of P-containing wastewater, Jpn. Kokai Tokkyo Koho 80 13,170 &13,171 (30 Jan. 1980). 34 Y. Yamada, 0. Mitsui, Phosphate ion removal by adsorption, Jpn Kokai Tokkyo Koho, 80 31,409 (5 March 1980). 35 B. Carbonel, L. Dalle, R. Chatelin, Extraction of polyphosphate ions in water circuits before passage on ion exchangers, Eau Ind., 46 (1980) 49-50. 36 J. C. Brown, Recovery and recycle of phosphoric acid bright dip rinse-water, Ann. Techn. Conf., Proc.-Am. Electroplat. SOC.,1982,69th ( l ) , Paper E-4. 37 Y. Araten, S. Lavie and G. L. Bridger, Magnesium fertilizers, in New Fertilizer Materials, C.I.E.C., Y. Araten ed., Noyes Development Co., Park Ridge, NJ, 1968. 38 K. C. Knudsen, The production of NPK fertilizers by ion exchange, J. appl. Chem. Biotechnol. 24(1974) 701-8. 39 Murray, Notices and Abstracts, Brit. Assoc. Adv. Sci. Report, 27th Meeting, 1857, 54-55 (publ. in 1858). 40 G. L. Bridger, M. L. Salutsky and R. W. Starostka, Metal ammonium phosphates as fertilizers, J. Agr. Food Chem. 10 (1962) 181-8. 41 M. L. Salutsky, M. G. Dunseth, Recovery of minerals from sea water by phosphate precipitation, Adv. Chem. Sci. 38 (1963) 27-39. 42 M. G. Dunseth, M. L. Salutsky, Removal of scale forming elements from sea water, Ind. Eng. Chem., 56 (1964) 56-61. 4 3 G. L. Bridger, N. K. Alfrey (W. R. GraceLCo.) U.S. Patent 3,125, 411 (March 17, 1964). 44 G. L. Bridger, J. F. McCullough, U.S. Patent 3, 174,844 (March 23, 1965). 45 M. L. Salutsky, F. S . Lee, G. L. Bridger, U.S. Patent 3, 126,254 (March 24, 1964). 46 M. L. Salutsky, P. Messina, Removal of scale formers with by-product recovery, Proc First Intern. Symp. Water Desalination, Washington, D.C. 1965, Vol. 2, 695-707. 47 C. Legal, B. L. Mobley (W. R. Grace &Co.) U.S. Patent 3,320,048 (May 6, 1967). 4 8 C. C. Legal, L. P. Schindler (W. R. Grace &Co.) South African Patent 6, 901,327 (Sept. 11, 1969). 49 P. H. Peng, W. R. Emst, G. L. Bridger, E. M. Hartley, Slow release fertilizer materials based on MgNH,PO,. Pilot-plant granulation studies, Ind. Eng. Chem. Proc. Des. Dev. 18 (1979) 453-8. 5 0 K. M. Ries, R. G. Dunseth, M. L. Salutsky, J. J. Shapiro, Ultimate disposal of phosphates from wastewater by recovery as fertilizer, Phase I. Final Report, FWPCA Report, Contract No. 14-12-171 (July 15, 1969);Chem. Eng. Prog. Symp. Ser. No. 107, Vol. 67 (1970) 54-62. 51 L. Liberti, G. Boari, R. Passino, It. Pat. No, 47912-A/81 (27 Feb. 1981). 5 2 L. Liberti, G. Boari, R. Passino, It. Pat. No. 47563-A/83 (14 Jan. 1983).
This Page Intentionally Left Blank
525
REMOVAL OF SILVER FROM PHOTOGRAPHIC SOLUTIONS
M. R. STEVENS Environmental Engineering Program, University of Southern California, Los Angeles, California 90089 M. PIRNAZARI* Civil Engineering Department, University of Southern California, Los Angeles, California 90089 F. SAAVEDRA
Environmental Engineering Program, University o f Southern California, Currently Assistant Professor, University o f Carabobo, Venezuela
ABSTRACT Silver removal efficiencies from photographic spent fix solutions were determined. Three processes including: metallic displacement, electrolysis, and a combined electrolytic-metallic displacement were studied. Analytical methods employed are detailed. The residual silver concentrations obtained by combined electrolytic-metallic displacement system were significantly below Phase I levels of the Sanitation Districts and are within the maximum concentration limits as set forth by the USEPA.
1. INTRODUCTION
Silver recovery, as distinguished from silver removal, from spent photographic solutions has long been practiced by the photographic industry. It represents a considerable economic savings for concerns engaged in large scale photographic processing. Technically, all three aspects of film processing, i.e., developing, fixing, and washing, will yield spent and wash solutions containing silver. For practical purposes it is the spent fix solution which contains the most silver. This is so since it is the fix solution which solubilizes the silver in the unexposed silver halide left on the film. The actual quantity of silver left on the film will depend upon the degree of exposure of the film, which in turn determines the concentration of the silver in the spent fix solution. Published data [ 11 regarding the amount of silver potentially removable from solutions indicate that up to 0.63 kg (1.39 Ib) of silver may be recovered when 0.305 km (1000 ft)
* To whom correspondences must be addressed.
526 of motion picture film are processes. At today’s market value, this represents a sizeable amount of money for large processors. I t also indicates the magnitude of the silver present in the fix solution. Photographic processors are generally located in large urban areas and discharge their wastewater streams directly into the municipal sewer system. Silver recovery systems report silver concentrations in spent fix waste solutions of 10 to 20 ppm [a,41. This quantity of silver present in the wastewater may cause problems for the biological secondary treat me nt operations. Should the processing plant be situated in a rural area and the effluent disposed of directly into nearby streams, serious pollution problems will be encountered. I t *is known that t h e presence o f 0.4 to 1 ppni soluble silver has a pathological effect upon living organisms [5]. Because of this, the United States Environmental Protection Agency (USEPA) published regulations establishing standards for the discharge of soluble silver in wastewater [ 6 ] .Silver recovery is currently secondary t o silver removal. The USEPA pretreatment standard for silver (Ag) is a discharge of 1.? ppni maximum for any one day with the average o f daily values for 30 consecutivenionitorings t o be equal to o r less than 0.5 ppm. While certain operations are exempt from this rule, none are exempt from the Phase I limits of the Sanitation Districts which is 5 ppm discharges of 10,000 gallons per day. Current practices [ 3 , 4, 71 generally focus o n the recovery of silver. A few [?. 8. 91 aim for the removal of silver, however these processes involve either complex systems or the production of compounds with undesirable physical and/or chemical characteristics. The purpose of the study presented here is t o demonstrate a practical way of achieving both silver recovery from photographic wastewater and simultaneously reduce the silver concentration in the spent liquor t o within t h e Sanitation Districts and possibly the USEPA limits. This goal is t o be acheived without resorting t o complex systems which necessitates the use of flocculation. generation o f sulfides or the use of hypochlorite [2, 91. Additionally, the use of cyanide as a coniplexing agent is avoided in achieving the objective of the study.
2. SUMMARY 01: METHODS 01’ RLCOVERY
Most large processors use electrolytic techniques to recover silver. Such systems will be affected by pH and fixer formulation. F o r example, for bleach-fix spent solutions the optimum pH is 7.2, while for non-bleach-fix solutions, the optimum pH is on the acid side (4.5 t o 6.9). Further, precise current density control is required. If the plating current is too high, thiosulfate will be broken down leading t o the formation of silver sulfide (Ag, S) as well as hdyrogen sulfide (H2S) gas. Nonetheless, the residual silver, remaining after electrolysis, is always of the order o f 20 t o 5 0 0 mg/l depending upon whether or not an electrolytic tailing unit is employed [ I ] , i.e., two electrolysis units in tandum, with the first removing the major portion o f the soluble silver from solution. The second popular method utilizes the metallic replacement cartridge, espoused by Kodak and Drew Refining [ 3 , 101. This mcthod, however, will displace only 95% of the silver present and is less effective than electrolysis in removing silver from solution. A third technique, sulfide precipitation, has alsways been regarded as effective method
‘
527 for the recovery of silver from solution. With a silver sulfide solubility product of l O T S 0 , the precipitation is complete. It is the physical properties which, in conjuction with its potential hazardous chemical property, make this method unattractive. Also sodium sulfide, the usual reagent employed, is unpleasant to work with and is also hazardous to use. The resulting Ag2S is difficult to filter and requires use of a fliter aid. All spent fix solutions must be rendered alkaline prior to the addition of the sulfiding reagent. Another technique is the use of sodium borohydride to precipitate the silver [ 2 3. Silver recovery is 99.9% but this means that silver levels in the effluent still equals 10 ppm silver. This is better than the above cited recovery methods, except sulfide precipitation, but does not meet the phase I limit of the Sanitation Districts, and falls far short of the overall USEPA goals. 3. SCREENING OF ANALYTICAL TECHNIQUES
Two methods were investigated for sample preparation. The first was the dithizone extraction of the silver from the spent fix solution. The second was the dichelation rect analysis of the spent fix solution for silver by the wet ash techniques. Two wet-ash methods were investigated for quantitative determination of the silver in the sample prepared by the methods cited. 1) Dithizone Extraction Method: This method involves the chelation of the metal by dithizone, extraction into an organic solvent followed by isolation of the Ag in aqueous thiocyanate (CNS-) solution [5 1. This solution is again treated with dithizone to complex the silver and bring it back into an organic phase. The solution is then analyzed spectrophotometrically. The procedure is long and thus not practical when a large number of samples are involved. To shorten it, the CNS- solution was subject to wet-ashing and subsequent atomic absorption analysis. The method failed to give consistent results. This may be due to interferences caused by chelation phenomena [ 5 ] . 2) Wet-Ash Method: Wet ashing is the application of oxidizing agents in order to destroy all organic matter and oxidize metals to their highest oxidized state. Several oxidizing agents may be employed, individually or in combination, for example, sulfuric acid (H2SO4), hydrogen peroxide ( H 2 0 2 ) , nitric acid ( H N 0 3 ) , etc. One of the several variations of the wet-ash methods tried, i.e., H 2 S 0 4 - H 2 0 2 oxidizing agent, produced a brown-black residue in addition to carbon. Upon oxidation of the carbon, this residue, silver oxide (Ag,O), was difficult to redissolve. The most successful wet-ash method for this work, discussed later, was the one which employed H 2 S 0 4 - HNO, as the oxidizing agents. The reaction product was a clear liquid with sulfur floating on the top. After filtration, the solution was subjected to atomic absorption spectroscopic analysis. This technique provided reproducible data and was employed throughout the experimental program. ~
4. MATERIALS AND METHODS 4.1. Reagents
Bakers Analyzed Reagent Grade ammonium thiocyanate, nitric acid, sulfuric acid (Ba-
528 ker Chemical Company, Phillipsburg, NJ) and Spectrum 30% hydrogen peroxide (Spectrum Chemical Manufacturing Company, Redondo Beach, CA) were employed in the experiments described. All experimental solutions were prepared using deionized distilled water. Spent fuc solutions were obtained from a large volume industrial photographic processor. Reagent Grade silver nitrate (Mallinckrodt Inc., Paris, KY) was used to prepare the calibration solutions for atomic absorption spectroscopic analysis.
4.2. Sample Preparation
Spent fix solutions were withdrawn from waste reservoirs. Each reservoir contained, on the average, 26 gallons of spent fix. One gallon of each reservoir was taken for experilight mentation. The pH of each sample ranged from 5 to 6. In general, all were yellow in color, transparent, and contained suspended matter. Each sample was filtered through Whatman No. 40 as well as through Millipore 0.45 p m filter paper prior to use. All samples were refrigerated prior to analysis. Wet-Ash P r o c e d u r e : This procedure, a modification of the EPA method [ l l ] , was used for both the ammonium thiocyanate (NH4CNS) extract of the Dithizone Method and for the direct analysis of the fix solutions. A 2-3 sample volume was placed in a high-wall Coors crucible. 1 ml concentrated Hz SO4 was added along the sides of the crucible. The frothing was allowed to stop, then heat was added until copious sulfur trioxide (SO3) fumes were apparent. The crucible was cooled and 1 ml concentrated HN03 added. Heat was applied until nitrogen dioxide (NOz) fumes were absent. The last step was repeated until no black residue remained. A yellow precipitate, usually floating on the top of the liquid, indicates the presence of sulfur. The solution was subsequently filtered using Whatman No. 40 paper. The crucible was washed five times with 2 ml of 3% HN03 and the washing passed through filter paper. All filtrates and washings were collected in a 100 ml volumetric flask which was then diluted to volume with 3% HN03. Solution was analyzed for Ag by atomic absorption. A t o m i c A b s o r p t i o n Analysis: This work was carried out using a Perkin-Elmer Model 406 atomic absorption spectrophotometer equipped with a Perkin-Elmer Intensitron Hollow Cathode Lamp. Instrument parameters were:
absorption wavelength - 328nm 0.7 nm monochromator slit opening flame air/acetylene ratio 2 lamp current 20mA 30 gain
Both the lamp current and gain value were set at the above noted levels throughout this work in order to obtain comparable data over the duration of the experimental work. The calibration of thc atomic absorption spectrophotometer was made using silver nitrate crystals, as noted before. Standard solutions were prcparcd using a 1000 ppni stock solution 01' Ag. This solution was subsequently diluted to prepare standards o f 10, 5 , I , 0.25, and 0. I ppm Ag.
529 5. METHODOLOGIES FOR RECOVERY OF SILVER FROM SPENT FIX SOLUTION
5.1. Metallic Displacement
Silver recovery by metallic displacement was accomplished by the addition of iron to 100 ml volumes of spent fix solution. The iron was added in the form of known amounts of steel wool. The amount added was normally 5 times the quantity of Ag present in the fix solution. To determine the maximum displacement the mixture was shaken for 1 hour. After settling, 2 ml of the clear supernatant liquid was taken, wet-ashed and analyzed for Ag by atomic absorption. The sample mixture was again shaken for an hour. After settling another 2 ml sample was treated as above. If the Ag concentration in the original liquor equalled that previously obtained, an additional quantity (1 or 2 grams) of steel wool was added. The mixture was shaken again for one hour and 2 ml of the clear supernatant liquor analyzed. If the silver concentration had decreased, another one to two grams of steel wool was added and the cycle repeated until a minimum constant amount of silver was obtained. The reaction mixture was then filtered. The sludge resulting from the metallic displacement reaction was treated with sodium cyanide (NaCN) which reacts with the silver preferentially. Five times the stoichiometric amount of NaCN was used to insure bringing the silver into solution. The solution was filtered to remove the residual iron precipitate. The silver concentration in the solution was determined by atomic absorption. The quantity of zinc powder used for the displacement reaction was five times the stoichiometric amount. The powder was added to the filtrate which consisted of silver cyanide complex ion [Ag(CN) ] and unreacted sodium cyanide (NaCN). Zinc precipitated the silver and replaced it in the solution. The solution was then filtered and the silver recovered. Completeness of precipitation of the silver was determined by atomic absorption analysis of the solution.
5.2. Silver Recovery by Electrolysis
The recovery efficiency of the electrolytic process was determined by electrolyzing 200 ml of the spent fix solution using an Eberbach Electroanalyzer (Eberbach Corp., Ann Arbor, MI). The silver content of the fix solution was initially determined. Electrolysis was carried out using a current of 0.1 amp for a duration based upon the initial silver concentration of the fix solution. The efficiency of the process was measured in terms of the amount of silver left in solution after electrolysis. 2 ml of the electrolyzed solution were wet-ashed and the resulting solution analyzed by atomic absorption spectroscopy as discussed previously.
6. RESULTS AND DISCUSSION
The experimental results are grouped into three categories: (1) Recovery efficiency by metallic displacement ( 2 ) Recovery efficiency by electrolysis (3) Removal efficiency by combined electrolysis and metallic displacement
530 6.1. Silver Recovery by Metallic Displacement
The displacement of silver from spent fix solutions follows the principle that a more reactive metal, in the electromotive series, will reduce a less active metal in solution. This is the basis for the silver recovery cartridges. Type P and Type 3, marketed by the Kodak Chemical Company [3, 41 and the Drew Refining Company [ 101. Both use iron in the form of steel wool closely packed within a rigid plastic container. The stated efficiency is low [3]. The overall reaction is: AgzS*03 + F e + FeS2O3 + 2 Ag This oversimplified view of the reaction does not take into account the possible effects of the other components of the fix solution. Table 1 presents a listing of possible components and their concentration in fix solutions within the 4-8 pH range [ 121. No one fix solution necessarily contains all of the chemicals shown in any one concentration category. Tab. l.(a) General Composition of Four Concentration Ranges of Fixer Solutions (pH 4 to 8) Less than 1 gram/Q
1 to 10 grams/Q
10 to 100 grams/Q
More than 100 grams/Q
Acetate Borate Carbonate Iodide Phosphate Sequestering agent Solubilizing agent -
Acetate Aldehyde Aluminium Thiosulfate Ammonium Bicarbonate Borate Bromide Phosphate Polyglycol Sequestering agent Solubilizing agent Sulfate Sulfite Tartrate Thio cy anat e
Acetate Aluminium Ammonium Borate Bromide Chloride Phosphate Sequestering agent Sulfate Sulfite Thiocy anat e Thiosulfate
Ammonium Thiocyanatc T h o sulfatc
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
(a) Ref. 12
For these experiments, five 100 ml portions of spent fix solutions were treated with increasing amounts of iron in the form of steel wool portions and the reaction allowed to proceed for two hours. Table 2 summarizes the data obtained. Examination of the data indicates that a ratio of approximately 3: 1 of silver: iron is needed to displace 99.8% of the silver from the spent fix within a two hour reaction time. Doubling or tripling the amount of iron did not increase the efficiency of the reaction. To determine the efficiency as a function of reaction time and iron concentration a second series of determinations were made. In this case 100 ml of spent solution containing 0.3585 grams of silver was reacted with 0.208 grams of iron for 1 hour. At that time
53 1 Tab. 2. Recovery of Silver from Spent Fix Solutions by Iron Displacement" Solution
Ag cone.@) (mg/Q)
Grams Fe added
Ag conc. (mg/Q) after displacement
% displaced
A B C D E
4,530 4,530 4,530 4.530 4,530
0.52 0.79 1.33 2.60 3.25
81.0 45.3 8.8 13.9 37.6
98.2 99.0 99.8 99.6 99.2
(a) Total run time 2 hours @) Conc. = concentration
the solution was sampled and an equivalent amount o f iron added. The reaction was allowed t o proceed for a n additional hour. After sampling the solution, a third 0.208 grams amount of iron was added and the reaction allowed t o proceed for a third hour. The experiment was run in triplicate. Table 3 summarizes the data. As shown in the table, approximately 0.6 grams of iron was required t o remove 99.5 % of the silver with the reaction time of 3 hours. Comparison o f these data with those of Table 2 indicates by 1/3, will achieve approximately the same degree of silver removal as obtained when reacting twice this amount at 2/3 the reaction time.
+
Tab. 3. Recovery of Silver from Spent Fix Solution by Iron Displacement Stage I reaction time: 1 hour Fe added: 0.208 gm Initial Rcmaining ~ g c o n c . ( @ ~g conc. Solution (mg/Q) (mg/Q) A B C
3,585 3,585 3,585
61 141 105
% recovered
98.30 96.07 97.07
Stage I1 reaction time: 2 hours Fe added: 0.208 gm
Stage 111 reaction time: 3 hours Fe added: 0.208 gm
Remaining Agconc. (mg/Q)
Remaining Agconc. (mg/Q)
% recovered
7 6.6 10.2
99.8 99.8 99.7
%
recovered
@)
-
14.2 69.5
99.6 98.1
(a) Concentration @) Error in atomic absorption analysis
It was noted that the action of the fix solution upon the iron resulted in a rapid 'coating' o r discoloration o f the iron surface. This caused the disintegration of the steel wool, forming a rust color sludge which contained black particles, possibly iron sulfide. The extent of the displacement reaction was a function of the availability of fresh iron surface. It was noticed that once the surface was coated the reaction stopped. Addition of fresh iron was needed t o achieve 99 + % removal of silver form the spent fix solution. Separation of the silver from the resulting sludge was accomplished by reacting the sludge with NaCN. Finally the silver was recovered by precipitation of the silver from the cyanide solution using Z n dust. Table 4 presents a tabulation o f the overall recovery for the solution given in Table 3. Based upon these experimental works, the overall metallic displacement recovery process will yield 99.5% recoverable silver from photographic spent fix solutions. It was also determined that the major loss was from the initial iron displacement reaction.
532 Tab. 4. Recovery of Silver from Spent Fix Solutions: Metal Displacement Followed by CNFormation and Zn Precipitation (Complete Recovery System) Initial Ag cone.@ Solution (mg/Q)
A
B C
3585 3585 3585
Total F e (mg) added
Millgrams Ag in sludge
CN-(gm) added
Zn(gm) added
Final Ag conc. (mg/Q)
Ag recovered
0.624 0.624 0.624
3571 3571 3571
0.540 0.310 0.310
0.245 0.244 0.241
3 3 6
99.5 99.5 99.4
(a) Concentration
The filtrate obtained after displacing silver by zinc dust presents a problem since the solution contains cyanide. Removal of the cyanide will require additional steps, for example, oxidation. The additional cost may make the overall process economically undesirable in spite of the current silver prices.
6.2. Silver Recovery by Electrolysis
Typical fix solutions of known silver content were used to deterfine the efficiency of this process. All were electrolyzed by the Eberbach Electroanalyzer under continously mixed batch conditions. It was established that 0.1 amp provided the optimum current to use. At that current the formation of silver oxide (Ag20) was minimum and a reasonable plating time achieved. In each case 200 ml of fix solution was used requiring about a 10 minute run time. It was noted that during electrolysis oxygen was produced at the anode which reacted with the silver ion (A;) rapidly producing Ag20, in the absence of cyanide ion (CN-.I. It was found that a low current would minimize oxygen (0,) formation and hence A g 2 0 product ion. The increase in weight of the cathode should be directly proportional to the amount of silver removed from the solution. It was found that these values were higher than the amount of silver present in the fix solution used in every experiment conducted. This phenomenon has been observed by other investigators [ 131. Consequently the efficiency of the process was measured in terms of the amount of silver left in solution after electrolysis. Table 5 presents the results of electrolysis using fix solutions. The silver content of four of the electrolyzed solutions were well within the Phase I limit of Tab. 5. Silver Recovery by Electrolysis from Spent Fix Solution Initial conc." Solution of Ag (mg/P) 1 2 3 4 5
1,070 ,270 3,500 1,000 1,750
Final conc. of Ag after electrolysis (mg/Q) 1.9 4.4 2.6 1.9 14.5
% Ag removed 99.82 98.37 99.93 99.81 99.17
533 the Sanitation District, ranging from 1.9 to 4.4 ppm. One showed a residual of 14.5 ppm, exceeding the Phase I limit. None met USEPA limits. The electrolytic recovery of silver is far superior to the metalic displacement procedure. Over 99% of the silver is recovered without addition of cyanide. However, this technique cannot be regarded as a silver removal alternative.
6.3. Silver Removal by Combined Electrolysis and Metallic Displacement
For this experimental series, 200 ml of spent fix solution was electrolyzed in a similar manner described ealier. The resulting electrolized solution was then treated twice with 5 grams of iron, except for Run 1 solution, and shaken for one hour. Run 1 solution required only one 5 gram portion of iron to achieve 99.97% silver removal as depicted in Table 6. Tab. 6. Silver Recovery from Spent Fix Solutions by Electrolysis Followed by Metal Precipitation of Residual Silver
Solution
Initial Ag cone.@ (mg/Q)
1 2 3 4 5
1,070 ,270 3,500 1,000 1,750
Conc. of Ag after electrolysis (mg/Q) 1.9 4.4 2.6 1.9 14.5
% removed
Conc. of Ag after metal displacement (mg/Q)
Total % removed
99.82 98.37 99.93 99.81 99.17
0.31 0.63 0.66 1.5 1.19
99.97 99.77 99.98 99.85 99.93
~~
_____
(a) Concentration
The addition of the iron to the electrolyzed spent fix solution effectively removed the residual silver, leaving only a trace quantity in the solution. Furthemore there was no need for additional treatment with cyanide. In retrospect, the metallic displacement system acts soley as ‘clean up’ for the minute amount of silver remaining after electrolysis reaction, yielding a silver concentration far below the Sanitation Districts Phase I limit.
7. SUMMARY AND CONCLUSION - Metallic displacement of silver alone, is the least efficient method. Additionally, separation of the silver from the resulting sludge necessitates the use of cyanide. - Electrolysis is substantially more efficient than metallic displacement. Further, it does not require addition of cyanide. Nonetheless, the residual concentration exceeds the Sanitation District Limits. - Clearly, the combined electrolysis-metallic displacement system is the most desirable technique for removal/recovery of silver from spent photographic fix solution.
534 REFERENCES
1 ‘Potential Silver from Kodak Photographic Products’, Publication J-IOA. Estman Kodak Co., 1981. 2 M. Cook and J. Lander, ‘Use of Sodium Borohydride t o Control Heavy Metal Discharge in the Photographic Industry’, J. Applied Photographic Engineering, 5 (3), 1979. 3 ‘Silver Recovery with Kodak Chemical Recovery Type Cartridge, Type F”, Publication J-9, Eastman Kodak Co., 1980. 4 ‘Silver Recovery a i t h Kodak Chemical Recovery Cartridge, Type 3’, Publication J-9A, Eastman Kodak Co., 1980. 5 APHA, AWWA, WPCF, Standard Methods for the Examination of Water and Waste Water, 15th edition, 1980. 6 ‘EPA Final Pretreatment Regulations for the Elcctroplating Industries’, Federal Register, 1979. 7 ‘Recovering Silver from Photographic Materials’, Publication J-10, Eastman Kodak Co., March, 1980. 8 G. Lorenzo, ‘A Review of Electrolytic Silver Recovery for the Regeneration of Bleach Fix Solution’, J. Applied Photographic Eng. 5 (3), 1979, pp. 141-143. 9 ‘Method Recovers Silver from Photo Processing’, C & E News, Sept. 13, 1982, p. 7. 10 ‘Introducing the New Industry’s Standard’, Technical Brouchc, Drew Refining Company, Glendale, CA, 1982. 11 ‘Iterim Method for Analysis of Priority Pollutants in Sludge’, EMSL - CIN, 1978. 12 ‘Chemical Composition of Photographic Processing Solutions’, Publication 5-47, Eastman Kodak, 1981. 1 3 S. Armstrong, Service Supervisor, Private Communication, Drew Refining Co., Glendale, CA, 1983.
CHAPTER VIII
PHYSICO-CHEMICAL ASPECTS OF BIOLOGICAL TREATMENT
This Page Intentionally Left Blank
537
MICROBIAL SULPHATE REDUCTION
W. M. DREW
Water & Materials Science Division, State Rivers and Water Supply Commission, Victoria, Australia G . H. A. HOLDER Monash University, Victoria, Australia
ABSTRACT The mechanisms of sulphide attack o n effluent disposal pipelines are discussed. In particular, results are presented arising from bench scale experimental work using a mixed culture of micro-organisms containing sulphate reducing bacteria. The results reported are derived from experiments conducted at near optimum conditions (i.e. optimums for temperature, pH and ORP were all determined by separate experiments). The kinetic parameters derived by fitting the data to Monod and MichaelisMenten mathemathical models are reported and include the maximum specific growth rate, yield and half velocity constants. The results were also analysed to yield an expression sinlilar to that derived independently from field collected data o n sulphide generation rates. An example is given of how a knowledge of sulphate reduction mechanisms and rates can be useful in practical pipeline design.
1. INTRODUCTION
Sulphate reducing bacteria, present in effluent disposal pipelines under favourable environmental conditions, can be responsible for corrosion of the pipelines by a mechanism generally known as sulphate attack. The sulphate reducing bacteria are obligate anaerobes and use sulphate ion as a terminal electron acceptor to produce hydrogen sulphide as an end product of anaerobic respiration. The utilization is similar t o the aerobic respiration of oxygen by aerobic bacteria to give water as the reduction product or the use of nitrate ion by facultative bacteria in the denitrification process to give nitrite, nitrous oxide OF nitrogen gas as reduced products. Sulphate is present in most natural waters and in varying concentrations in industrial wastewaters. Sulphate reducing bacteria have been found in environments ranging over hot brines, thermal springs, soils, estuarines, aircraft fuel tanks and human faeces [ 1,2]. This ubiquity means that given appropriate environmental conditions it is difficult to prevent the seeding of effluents by these bacteria.
538 Control can be effected by such measures as chlorination, direct oxygen injection or ozonation. However, in the first place it must be recognized that a problem could exist with a new effluent pipeline and that remedial action may be necessary during its life. The assumption being that it is difficult to entirely eliminate the possibility of such a problem at some time in the future of an installation and it is prudent to consider the problem at the planning stage rather than to ignore it. An example of where such consideration was given in the planning stages of a pipeline is a 54 km, 610 mm O.D. diameter, line recently installed to convey saline ash sluicing water from the Loy Yang Power Station (under construction) in the Latrobe Valley, Victoria, to Bass Strait. Some further discussion of this project is given later in the paper. Essentially this paper discusses some experimental results from laboratory bench scale studies on mixed cultures of micro-organisms containing sulphate reducing bacteria and looks at the usefulness of such data to full scale problems.
2. CHkRACTERISTICS OF SULPHATE REDUCING BACTERIA
The presence of sulphate reducing bacteria is generally evidenced by their ability to reduce sulphate to sulphide. This ”respiratory sulphate reduction” by sulphate reducers is often referred to as ’dissimilatory’ (i.e. opposite to assimilatory) and pertains to their ability to reduce more sulphate than required for normal metabolism. In 1965 and 1966 Postgate and Campbell [3, 41 published two joint papers on the classification of sulphate reducing bacteria. Their classification contains two genera, Desulfovibrio and Desulfotomaculum, and eight species. Generally the Desulfovibrio organism characteristics are : (1) Absence or sporulation (2) Presence of characteristic pigment, desulfoviridine, as an electron carrier (3) Motile by polar flagellae (4) Presence of a cytochrome, C3 (5) Growth characteristics in certain carbon sources including pyruvate, malate, lactate, oxamate, acetate, proprionate and butyrate (6) Vibrio shaped cells (i.e. curved rods) (7) Gram negative, pleomorphic (8) Organisms are anaerobic and require low 0 - R potential and must have iron for its cytochrome . Organic materials are dehydrogenated and the hydrogen is transferred to sulfite, sulfates and thiosulfates which are reduced to H,S. The Desulfotomaculum organisms have the following characteristics : (1) Presence of sporulation (2) Absence of the cytochrome, C3 (3) ’Twisting and turning’ motility (4) Rod shaped spore forming cells (5) Gram negative. Additionally, dissimilatory sulphate reducing bacteria are strict anaerobes and, in
539 general, growth media must be poised at a low oxidation/reducing potential (0 to -100 mV) for initiation of growth to occur. Reports that the organisms were autotrophic have been discounted by Postgate [S] who points out that the reasons for suspecting autotrophicity were due to trace organic impurities in the inorganic salts used to formulate certain test media. For the remainder of this paper the two genera of sulphate reducing bacteria are referred to as 'sulphate reducers' for convenience.
3. PHYSICOCHEMICAL CONSIDERATIONS
Reduction of sulphate reducers can be represented by the following equation in which lactate is used as the organic substrate: 2CH3. CHOH . COONa + H2S04-+2CH3. COONa + H,S + 2C0, + H,O Sodium Lactate
Sodium Acetate
(1)
Some sulfate reducers can use molecular hydrogen in the reduction of sulfate.
Depending on the pH of the medium in which the reaction occurs the sulphide species evident will be either HS-, H,S or S'. Fig. 1 is an equilibrium diagram which illustrates the relative concentration of each for the pH range 1 to 14. It is evident from the figure that at pH 7 approximately equimolar concentrations of HS- and H,S aq exist whilst there is a negligible amount of S'. Brown et al. [6] indicates that S'is inhibitory to activity hence alkali conditions are unfavourable to the activity of sulphate reducers. Equation (1) indicates that 3.1 mg of alkalinity as CaCO, is produced by the reduction of 1 mg of SOX expressed as S. In a static environment the activity of sulphate reducers should be self regulating. The combined effects of oxidation-reduction potential, pH and ionic constituents can be assessed by first considering the general electro-chemical equation : RT (activity of oxidized form) E = Eo + - In nF (activity of reduced form)
(3)
in which: E oxidation-reduction potential for the system, V E, standard hydrogen electrode potential, V R ideal gas constant T absolute temperature, K n number of electrons F Faraday number Substituting appropriate values into equation (3) and assuming a constant temperature of 25°C yields the following equation :
540
1.0 DISSOCIATION
CONSTANTS
x
2. 9.1
10-8
I. 7.1
x
10-15
2 . 1.2
x
10-15
3 . 1.0 X 10'13 REFERENCE DREW
I.
-
JA- I' TS
I 0
2
4
6
8
-
10
12
14
PH Fig. 1 . Equilibrium diagram for sulphide species in solution
E = Eo
LOX] + 0.059 -log n
(4)
[RED]
in which [OX] and [RED] are the concentrations of oxidants and reductants respectively. An estimate for E, can be found for a system in which sulphate is being reduced in the presence of sulphate reducers with lactate as an electron donor by firstly considering the following equations. 1/12CH3CHOHCOO-+1/3 H,O+ 1 / 6 C 0 2 + 1 / 1 2 H C O ~ + H ' + e 118 SOX+ 19/16 H + + e + 1/16 H,S
1/8 CO,
+ 1/16 H S - + 112 H,O
+ 1/8 HCO; + H' + e + 1/8 CH, COO-+ 3/8 H 2 0
541 Using literature values [7,8] for the free energy AGO of the above reactants and products the free energies for the total reactions can be determined. Having determined these values the AGO for the total reaction on a single electron basis is found. By definition the relationship between the standard electrode potential Eo and the free energy, AGO is: RTlnK Eo=-=nF
-AGO
nF
Therefore from this equation Eo is found (i.e. 290 mV for the example under consideration). Returning to equation ( 3 ) the following equation can be written which gives the oxidation-reduction potential (ORP) in terms of the oxidants and reductants involved: E = 0.290
+ 0.059 [(7/24 log (CO,)) + 5/24 log (HCO;) + 118 log (SOX)
- 1/8 log (CH$OO-) - 1/12 log (CH3CHOH COO-) - 1/16 log (HS-) - 1/16 log (H,S)]
- 0.19 pH
(9)
Equation ( 9 ) at first sight is somewhat complex, however, it is quantitatively useful when one wishes to predict what will happen when sulphate reduction occurs in an organic rich environment. Essentially the equztion indicates that as pH increases ORP decreases. Further, as the electron donor is used up then the ORP increases (i.e. becomes more positive) although this trend is counteracted by the appearance of volatile acids ( e g acetic acid). The appearance of H,S and HS- would also decrease the O W , however, as H,S is often lost from the system the net effect pH shift is marginal. The formation of CO, and HCO; have the effect of increasing ORP whdst removal of SOX decreases ORP. By pursuing the above analysis it is not proposed to advance ORP as control measure for systems involving sulphate reducers. Rather it was performed to illustrate the factors that must be considered if sulphate reduction is to be prevented in situations where it would ultimately result in unwanted gas build-up and/or corrosion of structures. 4. KINETIC CONSIDERATIONS
The rates of substrate utilization or product formation by micro-organisms can be modelled using numerous mathematical expressions. The models used in these studies were : Monod Model:
Michaelis-Menten Model:
dS dt
-
lXpmS
YK,+S
dS VmS - =-dt
Km+S
Set point mV recorder
ORP meter
0 ?1400mv
--
0-20mg/i
Sporging gas entry tube
Over f l o w trap 250ml
Magnetic stirrer bar
H,S
obsarption flask
250ml
Biatec m,nifermenter Pump
lndicotor I m p s : Diogrom I S
E
schemotx only
Wotson Morlowe penstolt,c pump
1N HCI
Fig. 2. Arrangement of apparatus for batch fermentations
’Simple’ rate expression:
r = K C C:~
in which : C, S substrate concentration, ML” Y yield constant pni maximum specific growth rate, T-’ X microbial cell concentration, M K, half velocity constant, ML” Km Michaelis-Menten constant, ML-3 K rate constant.
5. EXPERIMENTAL APPARATUS CONDITIONS AND PROCEDURES
A schematic diagram of the apparatus used is shown in Fig. 2. The central item of equipment is a modified Biotec Model FE007, 1 litre, mini-fennenter. The fermenter was modified by adding larger ports to the stainless steel head. The additional ports were of sufficient size to accept pH, dissolved oxygen, and O W electrodes. The original ports were used for sampling and pH control. The apparatus included a magnetic strirrer, heater and temperature controller. Stirrer speed was adjustable in the range 100-200 RPM and the temperature from ambient to 60°C. pH control was achieved using a peristaltic.pump feed controlled by a pH meter and set point recorder. HCl solution (1 N) was used to adjust the pH.
543 I00
10
1.0
I
Topt.
-
42.25OC
I
0.
3.00
3.05
3.10
3.15
3.20
3'25
3-30
I ~ O K 103
Fig. 3. Effect of temperature on H I S production
Nitrogen was used to sparge the system and H,S in tlie exit gas was absorbed in zinc acetate solution which was periodically removed and analysed by titration after acidification with HCl. The titrant used was 0.025N sodium thiosulphate and tlie end point determined using a starch indicator. A gas chromatograph was used on occasions to analyse the exit gas mixture for CO, and H,S. The growth medium was the same as that used by Postgate et al. [4] in their classification studies of sulphate reducers. The principal electron donor used in the medium was sodium lactate at a concentration of 3.5 gL-'. A sample of sediment and water containing sulphate reducers was obtained from a polluted creek near Melbourne. This sample was used to initially inoculate Postgate
544
'. 4 2
Ot
Fig.4. Constant initial lactic acid concentration with various initial suplhate concentration experiments, Runs 412 and 413
medium held at room temperature. Subsequent sub-culturing yielded an active stock culture. The activity of sulphate reducers is highly temperature dependent, hence experiments were conducted to optimise this factor. The procedure used was to measure sulphide production rates at a range of temperatures with all other conditions constant. A plot of production rates versus temperature indicates that the optimum temperature for the mixed culture was 42°C (Fig. 3). All aubsequent kinetic studies were conducted at the optimum temperature, at pH 7, with fured gas sparging rate (20 Llday), and constant stirrer speed (150 RPM). The results reported are for batch runs only.
545
.4
I
3
II \
i
i
I
Fig. 5 . Plot of H i S/CO data for run 413
6. RESULTS AND DISCUSSION 6.1. CO,/H,S Production
Measurements made using a gas chromatograph for CO, and HzS were of interest The results obtained show that, during fermentation in which sulphate reducers are active, it is possible to measure a ratio for H,S to CO, of up to approximately 0.5. This indicates that at this point (i.e. H,S/CO, 1 0.5) the fermentation was proceeding at maximum efficiency with 1 mole of H,S produced per 2 mole of COz produced (see Figs. 4 and 5). Generally, the elapsed time to attain a ratio of 0.5 coincided with the maximum rate of product formation but lagged the maximum rate of substrate utilization by several hours. The phenomenon of storage of substrate has been observed by a number of researchers working with activated sludge [9], however, it would appear that this is the first time it has been reported for sulphate reducers. 6.2. Kinetic Model Parameters 6.2.1. Monod
The data collected for a series of batch runs was examined using the integrated Monod equation and the technique proposed by Gates and Marlar [lo]. This technique generates values for P m , Ks and Xo/Y.Table 1 summarized the values for four experimental runs. These values can be compared to those obtained by other workers [6, 11, 121.
546 Tab. 1. Monod parameters as determined from lactate uptake data Run number Parameter
1
2
3
4
Sulphate concentration, mgL-' Lactate concentration, mgL-' Maximum specific growth rate, fim, h-' Saturation constant, Ks, mgL-' a = Y/xo xo/Y, mgL? Y, gcells/gsubs Doubling time, h
2100 500 0.57 140 0.22 4.5 0.44 1.2
2100 1500 0.17 98 0.01 88 0.01 4.1
2000 4700 0.32 4524 0.005 200 0.0043 2.2
2000 4000 0.303 1579 0.0046 217 0.005 2.3
Parameters determined by using following expression: XOIY
l / t In (S/So) = (1
+s
+ so - s
xo/Y
+1 In (
xo/Y
KS
so
i-
-1
-Pmt(
KS
xo/Y
which is an integrated form of the Monod equation, i.e. : ds/dt = - xo/Y (-
PmS KstS
1
Tab. 2. Comparison of Monod model parameters with literature data ~
~~
Source of data ~~~
Parameter
Hallberg [11] Brownet al. 161 Middleton&Lawrence [12] Experimental
Pmax, h-'
0.08 1 result 8.6 not reported
0.08 to 0.28 4 results 8.6 to 2.5 not reported
0.33 (max. temp. 31°C) 5.7,92 and 250
0.17 to 0.57 (opt. temp. 42°C) 1.2 to 4.1 98 to 4524
not reported
not reported
0.065
0.005 to 0.44
Doubling time td, h Ks mgL-'
Y-
g cells g subs
The key points to note regarding the values given in Table 2 are that Hallberg [ 111 and Middleton et al. [6] results were obtained from batch studies. Further, both Hallberg and Brown used pure cultures in their studies whilst Middleton used a seed taken from a municipal treatment plant. Middleton's carbon source was acetic acid. Other points to note are that Brown Achieved his highest growth rate of 0.28 h-' at a pH of 6.95 in asparged fermenter. His lower results were produced at a pH of 8.0. Therefore, as discussed earlier, it is possible that product inhibition occurred at the higher pH (i.e. S = present). The comment by Hallberg that he added alkali during the experiment is contrary to what would be expected from theoretical considerations and supports and theory that product inhibition could have been responsible for the lower observed growth rates.
400
.-
01
.E
9
so-
427 mg/l
2 111 T1 Y)
t
Simplified Monod Equation
LP- V
moxS
K,+
dt
5
x a 0 0
P .c a
-3 n
Constants derived for line shown
V max-28.7 rng1-lh-l Km = 98.3 mg/l Standord Deviations V max =%.7 rngl-lh-’
K, Note
=&79 Run
rngl-lh-’
4 2 used
1
Fig. 6 . H , S production rate data fitted using Michaelis-Mcnten Model
As a general comment it is evident that whilst the experimental results obtained by all workers agree to some extent, more data are required before design studies could be undertaken using the Monod or similar expressions. 6.2.2. Michaelis-Menten
Using the integrated form of the Michaelis-Menten model and a computer solution based on an ’interval halving’ technique, values for Vm and Km in the Michaelis-Menten model were found. Typical values obtained for Vm and Km are shown in Figure 6. No comparable literature values have been found. 6.2.3. Simple rate expressions
A large number of regression analyses were carried out using the general rate expression in the log form. However, the expression which related rate of H,S production to lactic acid and sulphate concentration was of most interest. The resultant expression was:
1/X d(G:2s)
= 0.0218 [S0,]o.44
where : X cell concentration, gL-’ GH,S hydrogen sulphide produced, mg [SO,] sulphate concentration, mgL-’ [LAC] lactate concentration, mgL-’
548
Fig. 7. Locality plan showing Latrobe Valley region of Victoria, Australia
Equation (13) can be compared to one quoted by Thistlethwayte [ 131 which has the form
G , = 3 2 . 3 X 1 0 - 6 X V, [BODs]0*8[S04]Oa4 1.139(foC-20)
(14)
where : effective average velocity, ft s-l V, BOD, 5 day Biochemical oxygen demand, mgL-' Thistlethwayte's equation was the result of a detailed study undertaken by the major sewerage authorities in Australia. The data used in the study was derived from investigations into sulphide build-up in pressure (rising) mains of diameters varying from 305 mm to 1220 mm with velocities from 0.3 m/s to 0.9 m/s. Temperatures varied from 19°C upwards. Sewages varied in BOD, strengths from 90 to 800 mgL-' and sulphate concentrations from 4 . 2 to 660 mgL-'. A cursory examination of equations (13) and (14) will reveal a close resemblance between the two, in particular the close agreement in powers for sulphate concentrations and carbon sources (i.e. BODSor Lactate).
7. APPLICATION OF STUDIES TO WASTEWATER DISPOSAL PROBLEMS
An example of where knowledge of the activity of sulphate reducers, their ubiquity and their symbiotic interaction with sulphur oxidizing bacteria (i.e. Thiobacillus concretivorus) is useful, was in the design of a 54 km,610 mm O.D. diameter saline wastewater pipeline for a 4000 MW base load powers station which is under construction at Loy Yang in the Latrobe Valley in Victoria (Fig. 7). The pipeline, now completed, is required to convey an ash pond solution having the following estimated range of ionic concentrations:
549 Na' Ca++ Mg" c1-
so,
Total Soluble Salts PH
1800- 7000 mgL-' 350- 400 50- 1000 750- 2500 2900-1 5000 6000-25000 8.5-10.5
The peak volume of saline water to be conveyed to the sea is 35 ML/day which includes similar effluents to Loy Yang power station from two other nearby stations (i.e. Yallourn and Hazelwood). Cement mortar lining for the mild steel pipeline could be susceptible to chemical degradation because of the sulphate and magnesium concentrations, and to degradation by biologically generated sulphuric acid, by e.g. Desulfovibrio desulfuricans and Thiobacillus concretivorus mediating sulphur cycle reactions. Nevertheless a protective lining over the stell is necessary, not only because the salinity and chloride levels could cause corrosion of the steel despite alkaline conditions but also because (a) sulphate-reducing bacteria could produce iron sulphide deposits, strongly cathodic to the steel and potential cover for shielded corrosion pits, (b) iron-oxidizing bacteria could mediate sulphuric acid production from reduced products of the anaerobic bacteria, given the translation of these products t o more aerated sites along the pipeline. Other planned or potential developments based on the region's 35,000 million tonnes of easily winnable brown coal reserves included several additional base load stations of the Loy Yang scale and oil from coal conversion plants. Extensive paper milling, wood chipping and other industries add to the complexity of the industrial picture. Each of these operations can, or already does, produce substantial volumes of wastewater (10 to 20 ML/d for coal conversion plants) which require discharge to the ocean. Land disposal is already practiced in the region but the capacity of the farm irrigation system is limited, and it is intended that it should be reserved principally for domestic waste disposal in the future. Thus there is a strong possibility that additional pipeline systems will be required in the future. The volume of waste liquors with sulphate concentrations greater than a few mg/L-' is seen to be potentially very significant. There is also a definite likelihood that these same liquors will contain suitable carbon substrate for utilization by sulphate reducers. Preliminary evaluation work could conceivably involve bench scale studies designed to establish the activity of sulphate reducers in the waste liquor in question. The range of environmental conditions that might be expected should then be estimated and suitable control measures established. Such control could be achieved through one or more of the following design or operational means: Design - Minimisation of slime growth by constructing the pipeline with sufficient slope to create aequate shear to scout the l i m
550
- Provision of good ventilation. - Use of suitable corrosion resistant linings and coatings in critical locations where it is impossible to avoid undesirable pipeline configuration.
- Provision for the injection of oxygen, peroxide of chlorine to control anaerobic conditions and for injection of caustic chemicals to control pH. Operational - Monitor waste liquors to limit sulphur, sulphide and sulphate contents. - Similarly, monitor waste liquors to limit organic content if high sulphur content liquors are unavoidable. - Use selective injections of chemicals such as oxygen, peroxide, chlorine and caustic to modify the environmental conditions so as to inhibit microbial activity. - Avoid shutdown situations where the pipeline might be left half full, thus creating conditions suited to H,S generation.
8. CONCLUSIONS
The results reported in this paper relate to a broad study where the batch kinetic parameters were used to model continuous flow bench scale experiments. However, the knowledge gained from these and earlier studies regarding the nature of sulphate reducing bacteria has been beneficial when consideration has been given to full scale projects. The close agreement of the experimentally derived gas production expression (13) to that derived empirically by Thistlethwayte [13 J suggest that the characteristics of the mixed culture used in the studies were reasonably representative of 'real world' mixed cultures. The information on growth rate, yield, optimum temperature and half velocity constant give an impression of the activity of the anaerobic organisms. The physiochemical considerations including pH, O W and ionic species give an insight into the growth control mechanicsms that might be employed.
REFERENCES
1 W. M . Drew, Kinetics of Microbial Sulphate Reduction, Thesis (Ph. D.) Monash University, Victoria, Australia, 1976. 2 H. Leclerc, C. Odger and Beerens, Occurrences of sulphate reducing bacteria in the human intestinal flora and in the aquatic environment, Water Research, Vol. No. 14, No. 3 (1980), 253-256. 3 L. L. Campbell and J. R. Postgate, Classification of the Spore-forming Sulphate-reducing Bacteria, Bacteriological Reviews, Vol. 29, No. 3 (1965), 359--363. 4 J. R. Postgate and L. L. Campbell, Classification of Desulfovibrio species, the non sporulating sulphate-reducing bacteria, Bacteriological Reviews, Vol. 30, No. 4 (1966), 732-736. 5 J. R. Postgate, Versatile medium for the enumeration-of sulphate reducing bacteria, Applied Microbiology, Vol. vii, No. 3 (1963), 265-267. 6 D. E. Brown, G. R. Groves and J . D. Miller, pH and Eh control of sulphate reducing bacteria, Journal Applied Chemistry and Bacteriology, Vol. 23 (1973), 141-149. 7 Handbook of Chemistry and Physics, 44th Edition, Cleveland, OHIO, Chemical Rubbcr Publishing Co., 1962. 8 P. L. McCarty, Kinetics of waste assimilation in anaerobic treatment, in Developments in Industrial Microbiology, No. 7 (1966), 144-155.
55 1 9 T. H. Venkitachalam and D. K. Koh, Activated sludge composition and carbonaceous substrate storage, Fourth Australian Biotechnology Conference, Melbourne, 1980, pp. 29-35. 10 W. E. Gates and J. T. Marla, Graphical analysis of batch data using the Monod Expression, Water Pollution Control Federation Journal, Vol. 40, No. 11, Part 2 (1968), R469-R475. .I 1 R. 0. Hallberg, An apparatus for the continuous cultivation of sulphate reducing bacteria and its application to Geomicrobiological purposes, Antonie van Leeuwenhoek, Vol. 36 ( 1 970), 24 1-254. 12 A. C. Middleton and A. W. Lawrence, Kinetics of microbial sulphate reduction, Water Pollution Control Federation Journal, Vol. 49, No. 7 (1977), 1659-1669. 13 D. K. B. Thistlethwayte (Ed.), The control of sulphide in sewerage systems, Melbourne, Butterworths, 1972.
This Page Intentionally Left Blank
553
EFFECT OF SOME PHYSICAL PARAMETERS ON COMPOSTING RATE AND YIELD
J. Y. PLAT, D. SAYAG and L. ANDRE
Laboratoire de Pkdologie: Physique et Chimie du Sol Ecole Nationale Supkrieure Agronomique, Institut National Polytechnique 145, avenue de Muret, 31076 Touluse Ckdex, France
ABSTRACT The effects of temperature, aeration and free air space on composting rate hnd yield were investigated o n a001 industry wastes, after the incorporation of suitable agricultural wastes promoting aerobic fermentation, under adiabatic conditions, in an insulated and unheated reactor, during a single composting nin. The bimodal CO, production rate - temperature profile reflects the change in microflora as well as stages in the depletion and conversion of the organic waste. The sigmoidal CO, production rate - aeration profile shows the occurrence of different ranges of microbial activity and the adaptation of microorganisms to restricted aeration. The parabolic yield-free air space profile corroborates earlier literature data on optimum composting conditions.
1. INTRODUCTION
Considerable interest has developed in composting as a means of solid waste disposal. Composting can provide a rational approach to solving the solid wastes dilemma without polluting the air, land or water resources. In addition to the concern for environmental problems, the increases in costs of fertilizers and the need for soil conditioners are to be taken into account. The laboratory scale investigations reported here were aimed at determining optimal composting conditions. The available nutrient content (3-4% nitrogen, 0.5-1 % potassium) of wool industry sludges can be valorised by the incorporation of suitable agricultural wastes and various additives. The effects of the hydration rate, fat, cellulose and lignin contents, as well as those of pH, particle size distribution, structure and porosity, had been investigated previously [ I ] so as to assess the optimum conditions promoting aerobic fermentation, first in a 2.5 1 reactor, then on a practical scale. 2. EFFECT OF TEMPERATURE ON THE RATE OF COMPOSTING
Temperature is a parameter frequently used to assess the efficiency of composting processes. Initially, the material is at the same temperature as the surrounding air; then, as
554 microorganisms multiply in the composting mass, a rapid warming occurs. The time elapsed before temperature began to rise rapidly was 16 hours. The temperature rose from 27 to 47OC during the following 18 hours (mesophilic stage), the remained nearly constant for the next 10 hours. It increased further to stabilize around 6OoC for 14 hours (thermophilic stage), then, it gradually and slowly cooled down. Microbial activity can be assessed by measuring oxygen consumption, carbon dioxide production, or chemical and physical changes in the composting mass. The carbon dioxide production lagged behind the temperature by eight to ten hours and followed the general pattern of change shown by the temperature. Figure 1 shows the plot of COz production rate against temperature. The hourly CO,
0
' 0 0 0
0
/ /
I
0
30
40
0 0
c
9 P
1
Fig. 1. CO, production rate vs temperature.
50 60 70 Temperature ('C
-
555 production per kg dry matter exhibits a first peak at 3S°C, followed by a trough at 42OC and a second peak at 60-65OC. The slopes of the linear ascending parts of the plot reflect the metabolic activity which decreases during the thermophilic stage to approximately 50% of that observed during the mesophilic stage, as a result of the change in microflora. The substrate has then been depleted of its more readily soluble and available constituents and microbial activity becomes less intense as the thermophilic microorganisms attack the lipid, protein, hemicellulose, cellulose and even lignin fractions [2]. The above data can be compared to those obtained by Schulze [3] who measured the compost temperature and the oxygen uptake rate at various times during a single composting run under adiabatic (i.e. insulated and unheated) conditions. Although the experimental conditions closely parallel those used in our investigations, the poor linear correlation found between the two measurements was not interpreted. A scrutiny of the oxygen uptake rate - temperature profile clearly shows that the data can be actually rationalized more satisfactorily in terms of a bimodal pattern which is analogous to that decsribed above for the carbon dioxide production rate - temperature profile.
3. EFFECT OF AERATION ON THE RATE OF COMPOSTING
Sufficient oxygen is also very important. Since the microorganisms are aerobic, they require oxygen for their growth. The earlier investigation of the effect of the air flow rate on the composting yield and on the maximum temperature reached had clearly shown [ I ] the highest efficiency of aeration rates of about 100 cm3/(min. kg dry matter). As there were only minor differences in yield between air flow rates of 50 and 120 cm3/min., the adverse effects of excessive aeration proved to be practically more important than the oxygen requirements of the microorganisms. It thus appeared that the minimum value (5%) reported in the literature [4] for the residual oxygen rate in the exit air was actually to be considered as a maximum permissible value. The effect of aeration on microbial activity was assessed by comparing the maximum theoretical CO, production from air oxygen (CO, theor.) to the CO, production observed over 48 hours during maximum biological activity ( C 0 2 obsd), as shown in Figure 2. The slope at the inflection point of the sigmoidal curve obtained for the plot of CO, obsd against the air flow rate is the same as that of the straight line corresponding to CO, theor. Three ranges of microbial activity can thus be distinguished: -When the air flow rate is below 80 cm3/(min. kg dry matter), i.e. under poorly aerobic conditions, CO, obsd was surprisingly greater than CO, theor. The observed CO, production does not correspond to the available air oxygen when the residual oxygen rate is below 2%. The adaptation of the microorganisms to restricted aeration results in increased microbial breakdown within the composting medium which thus provides add it ional oxygen. As the air flow rate rises from 80 to 120 cm3/min., the oxygen supply is thoroughly used for CO, production: CO, obsd = CO, theor. The residual oxygen rate at the outlet of the reactor was then 2%. As oxygen is then a limiting factor, the microbial activity markedly depends on the aeration rate. The highest compost temperature (75'C) was observed for an air flow rate of 120 cm3/min. These observations therefore corro-
CO, obsd.
Air flow
rate
c m 3 / (min.kg d r y m a t t e r )
Fig. 2. CO, production rate vs air flow rate.
borate the earlier remark on the interpretation of the 5% value in terms of a maximum permissible value. - A s the air flow rate is increased further, the observed CO, production levels off and a plateau is reached: CO, obsd < CO, theor. The limiting factors are then the temperature and the substrate whose effects are reflected in the plateau value. These observations can be advantageously used for the control of the composting process, particularly for the control of the rise in temperature. The dissociation of the different biological activity stages then affords more detailed investigations.
4. EFFECT OF FREE AIR SPACE ON COMPOSTING YIELDS
The optimum moisture requirements of organic wastes for successful composting vary widely as they depend markedly on the particle size and on the physical structure. The free air space concept originates from soil science and relates the specific gravity, bulk weight and dry mass of the waste to the air-filled voids within the material [5]. The free air space parameter, which depends on the moisture content, the particle size and the physical structure of the organic waste, can afford a further insight into the mechanisms of composting as it reflects the air permeabiility of the organic mass. The composting yields were expressed as the carbon weight loss per kg dry matter and plotted against free air space (Figure 3). The parabolic curve obtained shows that optimum conditions correspond to 30% free air space. Below this maximum, aeration is re-
557
50
40
0
L
W c c
E"
30
>r L
'13
," 20 \
o l
V
10.
I
I
10
20
1
30 Free a i r s p a c e
c ,
I
40
50
(%I
Fig. 3. Composting yield vs free air space.
stricted by compaction and excessive moisture, whereas, above 35% free air space, the yields are decreased by water deficiency. These data therefore corroborate those previously reported for the oxygen consumption rate in the literature IS], since 30-35% free air space is required to obtain optimum composting for a wide variety of materials.
5. SUMMARY
The results of these investigations have been used for the development of composting procedures that are being applied successfully to the full-scale operational composting of wool industry wastes as well as of slaughterhouse wastes. The most significant points are the fast and nearly complete disappearance of fats and the breakdown, of ligno celluloses to humic matter characterized by a high ion-exchange capacity. In spite of the relatively high conductivity of the aqueous extracts, the composts thus obtained do not induce any phytotoxicity. Their ion-exchange, buffering and water-holding capacities as well as their mechanical properties compare favourably with those of commercial peats. The land application of these organic fertilizers will therefore be highly beneficial since the average organic matter content of soils in South-western France is low (<2%).
558 REFERENCES 1 J. Y. Plat, TliPse D o c t e u r - I n g ~ i ~ i e 1981, ~ r , 1N.P. Tolouse. 2 R. P. Poincelot, Conn. Agr. Exp. Stat. Bull. No. 754, 1975. 3 K . L. Schulzc, ‘Aerobic decomposition of organic \vaste materials’, Final Rcport, Project RG-4 180 (C5R1), National Institute of Health, 1961. 4 I<. L. Schnlze, Appl. Microbiol. lO(1962) 108-122. 5 J. S. Jcris and R . W. R q a n , Compost Sci. 14 (2), 1973, 8-15.
559
TREATMENT OF HIGH STRENGTH WASTEWATERS BY AN ANAEROBIC FLUIDISED BED PROCESS
D. BARNES, P. J. BLISS, R. B. GRAUER, C. H. KUO and K. ROBINS
School of Civil Engineering, University of New South Wales, Kensington, 2033, Australia
ABSTRACT Fixed film reactors arc onc of the oldest methods of wastewater treatment. In the last few years scvcral processcs have been developed which have led to a resurgcnce of intercst. An anaerobic fluidised bed-contact reactor system has been developed for the pretreatment of high BOD wastewaters and the system has becn tcstcd using a synthetic wastewater of molasses and yeast cxtract at approximately 3000 mg/L BOD. Trcatment efficiencies in excess of 90% soluble BOD reduction and 75% total BOD reduction wcrc possible with loading rates of up to 20 kg n1-3 d-' BOD. Sludge production was approximately 0.1 kg . kg-' BOD removcd and the system demonstrated considerable process stability.
1 . INTRODUCTION
Fixed film reactors are arguably the oldest form of wastewater treatment process and currently are the systems which are offering the most exciting prospects for improved wastewater treatment. Traditional trickling filters were developed as an engineered version of wastewater irrigation at a sewage farm. If wastewater is irrigated over soil, bacteria attached t o the soil particles degrade the organic components in the wastewater. The porosity of soils is not high and irrigation systems tend to pond as the soil .blinds, thus resulting in standing areas of wastewater with the associated insect and odour problems. Trickling filters overcome these problems by growing the microbial culture on large particles (50 mni diameter), such that there is a large void space in the bed and flow is not restricted. It is instructive to note that there is now a renewed interest in the use of land treatment systems particularly for the treatment of industrial wastewaters. With a more detailed understanding of soil properties it is possible to irrigate land with wastewaters without adversely affecting the soil structure. Land treatment systems are not discussed in this paper, but they represent an optirnisation of the broad irrigation methods from a consideration of soil structure and wastewater properties. The conventional single pass trickling filter was a popular method of wastewater treatment during the first half of the twentieth century and was improved by modifications
560 Tab. 1. Comparison of activated sludge and trickling filter plants Factor
Activated sludge Trickling filter
Difference
Capital cost Operating cost
Low High
High Low
Land area
Low
High
5 0-1 00% Filter can consume zero electricity 4 X biological 2 X total plant
Technical control Climate
Much Problems in dry summer months Not suited
Little Best at higher temperatures Suited
Low S . S . High NHi-N Much Low
High S.S. Nitrified Little High
Industrial and strong wastcwaters Effluent quality Mechanical equipment Head Loss
Little heat loss from filter High biomass and solids retention
10 cm and 2 m respectively
such as recirculation and alternating double filtration. It was gradually supplanted by activated sludge processes such that by the 1970's large and medium sized municipal and industrial plants tended to use the activated sludge method. Table 1 gives a comparison of the properties and performance of trickling filters and activated sludge plants. Trickling filters by their ability to retain a high biomass concentration have particular advantages for the treatment of shock loads. Their major disadvantages are the high capital cost of construction and the large land area required for the biological reactor. The filters require a relatively low level of technical operation, although it is difficult to assess the electromechanical maintenance requirement. The majority of operating trickling filters are greater than 10 years old and many include old heavy constructional materials such as cast iron, so that comparison with modern activated sludge plants are inappropriate. Several modification have been made in which the support media are altered from the conventional rock to a range of shapes and geometries, to combine the advantages of fixed film and suspended growth processes. These modifications include: 1. The replacement of rock media with plastic media of high specific surface area. The plastic media can be random packed mouldings o r interconnected sheets designed to give a good distribution of liquid. Both have found application for the treatment of industrial wastewaters. The filter media are self supporting and can be arranged in units several metres high, and high hydraulic loadings (10-20 m3m-2 d-' ) and high organic loadings (1-2 kg BOD m - 3 d - ' ) are used. The media are more expensive than traditional rock media. 2. The construction of a rotating biological contactor in which the media are rotated through a trough of wastewater. The most common form of this type is a series of discs (2-3 m diameter) mounted on a ahsft, and rotated axially through the wastewater. For small plants (less than 500 persons) a disc BOD loading of 6-8 gm-2 d-' is recommended, while for larger plants design curves are available. 3. The traditional trickling filters have been improved by the use of modern materials of construction and the inclusion of flow balancing and recirculation.
56 1 4. Combination of trickling filter and activated sludge units by direct integration of the two processes, referred to as activated biofiltration. 5. A series of new processes which use small particle sized media 1-50 mm diameter but maintain flow distribution. These systems are described under the packed and fluidised bed headings. 6 . The resurgence of interest in fixed film reactors has led to theroretical studies, particularly on the diffusion of reactants and products within the biomass.
Several industries produce high BOD wastewaters which contain a high concentration of soluble and colloidal organic material. Such wastewaters have to be treated'either by aerobic or anaerobic biological processes. Anaerobic processes are more appropriate for high strength wastewaters because of the low sludge production rates, no requirement to provide oxygen as an electron accepter and the generation of methane which can be used as a fuel. However traditional suspended biomass digesters tend to provide only slow rates of biochemical conversion and hence require long hydraulic retention times with consequent high capital cost. More intensive anaerobic systems such as the upflow sludge blanket [ 11 and expanded bed digesters [2], offer more cost effecrive method of treatment. Conventional municipal anaerobic digesters [3] have volumetric volatile solids loadings of 0.5-1.5 kg m-3 d-' with high rate digesters loaded at 2.5-4 kg m-3 d-' . The intensive anaerobic systems [ 1, 21 can operate at loadings an order of magnitude greater than conventional digesters and hence at loadings similar to or in excess of aerobic systems.
2. FLUIDIZED BED SYSTEMS
Biological fluidised bed processes have been developed to provide a compact wastewater treatment option [4, 51. The biomass is grown on a small particulate support medium which is suspended by the upward flow of wastewater. This achieves a very large biomass concentration and permits high organic loadings. The fluidised biological bed concept has been applied mainly to anoxic and aerobic processes with relatively little reported study on anaerobic biological fluidised beds [S]. Fluidised bed reactors [7, 81 are operated by passing a flow of wastewater up through a bed of sand. The upflow rate is sufficient to separate the sand particles and to expand and fluidise the bed. Biomass can grow on the sand medium and this metabolises the components of the wastewater. The sand provides a large specific surface area (3,300 m2m-3) while the fluidisation prevents clogging. Figure 1 illustrates the aerobic version of the process [7]. The low solubility of oxygen makes it necessary to dissolve pure oxygen under pressure and to include facilities for recycle. Treated effluent can be removed from the top of the column reactor. Excess biomass can be removed from the sand in a high shear unit, so eliminating the need for conventional clarifiers. The sand is returned to the reactor, while the sludge passes for treatment and disposal. The aerobic process [4, 51 has been applied to settled sewage (BOD 'L 100 mg/L) where it reduced the BOD by 93% with a treatment time of only 16 minutes. A similar system has been used to nitrify secondary effluent (NHf-N 20 mg/L) when 99% ammonia removal was observed with only 11 minutes retention at 24°C. For nitrification this
562 Reycl e tank Eff b e n t v
r'
I Hyd roclone
--
S t u d gc for P d i sposa 1
Fluidized
Effluent revcie
Waste-water
~
g
I 1
~ P~e a qg r a v~e l ~ 1
q
D i s t r i b u t i o n pLate
Oxygen
Fig. 1 . Fluidised bed aerobic treatment system
represents an ammonia nitrogen loading of T, 900 nig m-'d-' at a hydraulic load of 140 m3m-3. d - ' . If oxygen is eliminated from the reactors an anoxic culture can be established which wilI denitrify effluents. For municipal wastewaters a retention time of 6.5 minutes achieved 99% NO3 -N removal. The aerobic and anoxic beds offer a compact method for BOD removal, ammonia oxidation or nitrogen removal. The high surface area of the media provides high rates of treatment. These reactors are likely to be used to extend plants to achieve nitrification or denitrification and for the aerobic treatment of soluble industrial effluent [4, 51. One major application of anaerobic fluidised beds should be to the pretreatment of industrial effluents prior to sewer discharge. The standards required for an industrial effluent to be discharged to the sewer usually attempt to ensure that pollutional load is similar to that of domestic wastewaters [6, 71. In many cases the critical conditions are that the five day BOD and the suspended solids concentration are each less than 600 mg/L. Wastewaters from the food and beverage industries usually contain relatively high concentrations of soluble and colloidal degradable organic material, for example those from intensive animal husbandry, dairies, fermentation industries and abattoirs. The BOD of these wastewaters often exceeds 2000 mg/L and is well suited to treatment by an anaerobic fluidise d bed process. An anaerobic fluidised bed system has been established specifically as a pretreatment process. The overall performance of the system treating a well defined synthetic wastewater is reported in this paper. Subsequent papers will report the treatment of specific industrial wastewaters, full scale treatment, transient load responses of the system and aspects of the microbiology of the system.
3. DIGESTER DESIGN AND OPERATION
A pilot scale system was constructed which included a 40 L fluidised bed reactor in
563 Biogas
4
,Tern pe r ot ur e probe
I
,Gas
vent Dilution * water
Feed :oncen t r a t e
L=EIl Ref rige r a tcr Ternp e rature controller
4
er but
Fig. 2. Pilot plant anaerobic fluidized bed
series with a contact reactor (Figure 2). Liquid was pumped from the contact reactor to the base of the fluidised bed reactor by a close coupled centrifugal pump. The fluidised bed reactor was constructed of 150 mm diameter clear persex with a concentric 200 mm diameter clear persex chamber in the upper section of this reactor. The contact reactor was a modified mild steel drum fitted with a conical base and an internal baffle. All connections were made with clear polyethylene hose using brass or galvanised fittings. The fluidised bed reactor was partially filled with screened and washed beach sand and was fluidised by the upflow of liquid pumped from the base of the contact reactor at a superficial velocity of approximately 0.39 m * min-' . The system was seeded with anaerobic sludge from a domestic wastewater treatment plant. The reactor system was fed with a non limiting substrate based upon molasses and yeast extract, Table 2. The substrate concentrate was stored under refrigeration and added to the reactors via a chemical dosing pump. Water from a separate ambient storage tank was added via a separate dosing pump and admixed with the substrate concentrate to produce a wastewater feed with a BOD of 2500-3000 mg/L. Antifoam agent was included in the con-
564 Tab. 2. Composition of Substrate Concentrate Component
Concentration mg/L
Molasses Yeast extract Sodium hydrogen phosphate Sodium carbonate Antifoam agent
65000 92000 29000 36000 500 36000 39000 2300
BOD Total nitrogen Total phosphorus
centrate to minimise foaming as the substrate concentrate was prepared and to facilitate gas removal from the reactors. Biogas was collected from a conical collection device at the top of the fluidised bed reactor and from the head. space of the contact reactor. The system was maintained at 36 f l°C by controlled heating of the contact reactor with an external coil containing hot water. In the set of experiments reported in this paper the reactor system was seeded by batch wise feeding until a reliable biomass was observed. Continuous feeding was used in a series of experiments which progressively lowered the total hydraulic retention time by increasing the rate of addition of both substrate concentrate and water while maintaining an approximately constant feed strength. Analyses of the influent and effluent from the system were carried out over a period of approximately 12 months using the conventional analytical techniques [8], modified where appropriate for specific determinations [9].
4. RESULTS AND DISCUSSION
The performance of the reactor system is shown in Table 3 and Figure 3. As the hydraulic retention time was reduced from 44 hours, there was little deterioration in overall performance until the hydraulic retention time was reduced to 4.6 hours. Significant changes in both effluent quality and volatile fatty acid concentration (Figure 4) were observed by day 320. This corresponds to a hydraulic retention time of 3 hours and an organic (BOD) loading of approximately 30 kg m-3 d-' (Figure 5). The reactor system appears to have become unstable at a retention time of 4.6 hours (Figures 3, 4 and 5). On day 290 there was a slight deterioration in effluent quality and an increase in volatile fatty acid concentration. However the reactor appears to have recovered by day 305. These changes were more pronounced as the load was increased (and the loading experiment was terminated) and is reflected in the changes in performance, volatile acids and pH. The variations in volatile fatty acid (VFA) concentration (Figure 4) over the 100-200 day period were relatively small. The 10 day average concentration was between,200 and 300 mg/L acetic acid. During the start-up period there was a higher volatile fatty acid concentration, presumably as the anaerobic biomass stabilised. During the Christmas-New Year period feeding of substrate was restricted and volatile acid concentrations reduced to less than 100 mg/L. The system recovered by day 280 during the 4.6 hour loading. Figures 3 and 5 illustrate the difference between soluble and total effluent BOD, this difference represents the BOD of solids wasted from the reactor system. The difference
565
-s
100
-
I
x
c" 1 5 %
Y
*-
'c w
Z
c
0
50-
c
Solubte treatment efficiency
A Tofal treatment
effiiiency
U
'u
44 hr.
I=
HRT
25
22 hr. 14 hr. i H R T , HRT
I
8 hr. HRT
4-6hr. 3 hr. H R T , HRT
1
Fig. 3. Influence of hydraulic retention time on BOD removal, influent BOD 2500-3000 mg/L
z.
-1-
I
v
Fig. 4. Variation of volatile fatty acid concentration
f
E"
50-
0
w ~ u b ktreatment efficiency A totol treatment efficiency
Q
10
M
Average Organic Load kg BOD m3 day-'
Fig. 5. Influence of organic load on treatment efficiency
30
566 Tab. 3. Performance of Anaerobic Fluidished Bed System BOD (mg/L)
Day
48 56 63 92 98 107 113 118 125 127 134 139 147 155 162 188 202 210 216 225 230 25 8 265 280 293
HRT hours 44
mean 22
mean 14
mean 8
mean 4.6
Loading rate* kg BOD . m - 3 - d-'
2.9 2.6 2.4 2.3 2.3 2.4 4.1 5.2 6.8 5.0 4.9 4.5 6.0 5.2 6.8 6.1 6.9 6.6 6.6 8.5 13.7 15.8 10.8 12.3 12.2 9.7 11.9 22.9 21.8
VFA mg/L
480 370 370 300 240 350 250 250 240 290 280 330 310 260 230 300 290 270 280 200 230 310 260 600 60 230 270 250 680
300
15.7
-
307
21.1 20.6 28.8 29.8 29.3
4 10 45 0 770 860 820
320 323
mcan 3 mcan
Treatment efficiency (%)
effluent feed
total
2320 2480 2280 2150 2190 2280 25 00 3150 41 30 3030 3000 2730 370 3180 2640 2370 2670 2550 2560 1890 3050 3500 2400 2730 2700 2150 2630 2920 2780 201 0 2820 2630 2400 2480 2440
600 470 320 460 330 440 440 320 340 270 270 470 -
350 420 610 610 580 560 530 750 550 480 640 750 380 580 250 740 730 460 540 1520 1520 1520
soluble
320 170 190 220 83 200 120 60 93 52 52 120 100 89 86 130 200 1770 150 100 88 66 100 110 250 100 120 120 440 310 310 300 1350 1540 1440
total
74 88 86 79 85 82 83 90 92 91 91 83 -
88 84 74 77 77 78 72 75 84 80 77 72 82 77 91 73 64 84 78 37 39 38
soluble
86 93 92 90 97 92 95 98 98 98 98 96 97 97 97 94 93 93 94 94 97 98 96 96 91 95 95 96 84
85 89 88 44 38 41
* Calculated for a digcstic volume of 0.040 m3 does not vary greatly with loading. For a volatile solids concentration in the fluidised bed reactor of 25.4 g/L in an active bed volume of 26 L and assuming a biomass composition of C 5 H 9 0 3 N 1 the solids residence time of the biomass in the fluidised bed reactor"is approximately 30 days. Under constant loading conditions (hydraulic retention time (HRT) of 8 hours) the rate of sludge production was similar to that reported for other anaerobic processes, approximately 0.1 kg * kg-' of BOD removed. Similarly the biogas composition is comparable to that reported for other anaerobic systems, approximately a 2:1 ratio of methane to carbon dioxide, however in this nitrogen rich feed a high proportion of nitrogen gas is present in the biogas. A range of anaerobic bacteria has been
567
Fig. 6. Electronmicrographof anaerobic biomass attached to support media
identified in the biomass, although the relative proportions are still to be confirmed. The structure of the biomass can be seen from the interlocking matrix observed by electron microscopy, Figure 6. The mixture of microbial shapes, rods, filaments, cocci, is similar for different substrates but the proportion of each type varies both with substrate during start up and vertically within the reactor.
5. CONCLUSION
This study, for which the overall loading data have been reported, indicates that the anaerobic fluidised bed process is capable of treating high concentrations of soluble BOD. The systems appears to have some significant advantages when compared with competitive processes for pretreatment of industrial wastewater. These include: 1. The reactions are rapid and very high BOD loadings of 10-20 kg m-3 d-' are possible, which are competitive with aerobic processes. Conventional activated sludge type reactors usually cannot exceed 5 kg m-3 d-' . Hence the area and volume required for the fluidised bed system are relatively small. 2. There is only one major mechanical moving unit, the pump (Figure 2). Pumps can be selected which are very reliable and require little maintenance. 3. As a consequence of the previous points, the capital cost of an anaerobic fluidised bed system is likely to be 40-80% less than a conventional aerobic system, and the installed energy requirements only 20-70% of those of the aerobic system.
568 4. The sludge production is similar to that for other anaerobic processes so it is almost an order of magnitude less than for conventional aerobic processes. For a typical wastewater of 3000 mg/L soluble BOD, at 90%reduction of soluble BOD and sludge production of 0.1 kg * kg-' BOD removed, the effluent will contain approximately 300 mg/L soluble BOD and 300 mg/L suspended solids. This will meet a 600/600 discharge standard without further treatment. 5. The methane gas is generated from relatively small reactors and can be collected and used if required. 6. The system is enclosed to exclude air and any odorous or noxious gases are removed with the methane.
REFERENCES
1 G. Lettinga, A. F. M. von Velsen, W. de Zeeuw and S. W. Hobma, The application of anaerobic digestion to industrial pollution treatment, Proc. Int. Symp. Anaerobic Djgestion, Cardiff, U.K. 1979. 2 M. S. Switzenbaum and W. S. Jewell, J. Water Pollut. Cont. Fed. 52, 1980, 1953-1965. 3 D. Barnes, P. J. Bliss, B. W. Gould, H. R. Vallentine, Water and Wastewater Engineering Systems, Pitmans, London, Melbourne, Marshfield 1981. 4 J. S. Jeris, C. Beer and J. A. Mueller, J. Water Pollut. Cont. Fed. 46, 1979, 2118-2128. 5 P. F. Cooper and B. Atkinson (eds), Biological Fluidised Bed Treatment of Water and Wastewater, Ellis, Horwood, Chichester 1981. 6 J. M. Sidwick, J. Inst Water Eng. Sci, 30, 1976, 116-123. 7 L. J. Brack and D. W. Lynch, In Costs and Benefits of Environmental Protection, Aust. Gov. Publ. Serv. Canberra, 1981, pp. 157-187. 8 Am. Publ. Hlth Assoc., Standard Methods for the Examination of Water and Wastewater, Am. Publ. Hlth Assoc., 14th ed., Washington, U.S. 1975. 9 R. Dilallo and 0. E. Albutson, J. Water Pollut Cont. Fed. 33, 1961, 356-365.
569
A RESPIROMETRIC STUDY OF THE INFLUENCE OF ALIPHATIC ALCOHOLS ON ACTIVATED SLUDGES
P. LE CALVE and N. THERIEN
Facultk des sciencies appliqukes Universitk de Sherbrooke, Sherbrooke, Qudbec, Canada J I K 2RI
ABSTRACT The influences of primary aliphatic alcohols o n oxygen consumption of activated sludges in endogenous states were measured using a laboratory respirometer. The alcohols studied were n-propanol to n-octanol. Using beef extract as a reference substrate, results demonstrated a two-stage action for all alcohols. It was found that the alcohols enhanced the assimilation capacity of the biomass when they were below critical concentrations, but had inhibiting effects above these concentrations. Hypotheses and possible mechanisms of action were proposed to explain these results. An analogy with detergents was made.
1. INTRODUCTION
It is well known that detergents and tensio-active materials present in wastewaters have important effects on the efficiency of activated sludge treatment plants [ 1-31. This is also true for domestic wastes due to the use of laundry detergents, as well as wastewaters from textile plants. On the other hand, research on a controlled activated sludge process demonstrated the importance of aeration on the dissolved oxygen level in the aeration basin and generated new models based on oxygen consumption [4]. It is thus seen as important to be able to predict oxygen consumption as a function of inflowing wastewater characteristics. In the course of the study of an existing biological treatment plant, a sudden drop in dissolved oxygen in the aeration basin was noted in response to a significant inflow of detergent. This study [S] demonstrated that such a drop could not be adequately explained only by a change in the oxygen transfer rate at the liquid-gas interface, the detergents being strongly absorbed by the biomass. Elsewhere, studies using a continuous microcalorimeter demonstrated the influence of primary aliphatic alcohols on the metabolic heat production of microorganisms [ 6 ] , a parameter directly related to oxygen consumption [ 7 ] . These alcohols are non-ionic tensio-active agents for which the active group is the hydroxyl radical, and have an effect on oxygen transfer analogous to that of synthetic 'detergents [8]. We were thus drawn
570 to evaluate the influences of these compounds on the oxygen consumption of activated sludges using respirometric techniques.
2. THEORY AND PRACTICE OF RESPIROMETRY
Respirometry is a technique for continuous measurement in a closed system of the oxygen consumption of a population of microorganisms when exposed to a particular substrate. It has found many applications in the biological treatment of wastewaters, notably in studies of biodegradability, toxicity, and short-term biological oxygeh demand (BOD,) of wastewaters [9-lo]. and allows the laboratory-scale reproduction of conditions found in aeration basins. The scheme of the respirometer we used (Tech-Line Instruments) is shown in Fig. 1. It is composed of a 1.0 litre aeration column, a 0.5 Umin air-circulation pump, a diffuser, a CO, absorber using a potassium hydroxide solution and a manometer for pressure measurements. The system is closed. The air pump permits continuous aeration of an activated sludge sample and recirculation through the absorber. The fine-bubble diffuser assures efficient oxygenation and good mixing of the liquor. Oxygen consumption due to microorganism respiration is accompanied by a equimolar production of CO, which is absorbed by the potassium hydroxide solution. The system being sealed, there is a corresponding drop in pressure measured by the manometer. The pressure transducer is calibrated to indicate the volume of oxygen consumed over time, which is recorded. This volume can be treanslated into mass. The operating conditions are maintained such that neither the aeration rate nor the amount of dissolved oxygen can limit biological activity. Our experiments consisted of injecting a certain quantity of the material under investigation into a sample of activated sludges containing a known concentration of microorganisms, and subsequently measuring the oxygen consumption over time. The activated sludge samples were taken from the aeration basin of the wastewater treatment plant of
r "
1
f I
1 2 3 4 5 6 7
reactor d i f f u s e r stone a i r pump CO2 scrubber manometer pressure transducer recorder
Fig. 1. Laboratory wastewater respirometer
57 1 the Centre Hospitalier Universitaire de Sherbrooke (CHUS). The sludges were aerated sufficiently long (12 to 24 hours) to stabilize the oxygen consumption rates, and were thus considered to be at the endogenous respiration state. After injection of a certain quantity of organic material into an activated sludge sample in the endogenous respiration state, the measured respiration rate RT will be sum of the endogenous respiration rate RE and the substrate respiration rate Rs:
RT = RS + RE
(mg 02/min)
Since this respiration rate is proportional to the quantity of microorganisms present, we have reported the specific oxygen uptake rate as:
SOUR = RT / VSS
(mg 0 2 / g VSS min)
where VSS is the concentration of volatile suspended solids as measured using standard methods [ 111. - Tests were run for 30 minute periods. The integrated value of oxygen consumption RT is calculated as:
R T = t =.fwO
R ~ d t
(mg 0 2 )
Specific oxygen consumption
RT* is then calculated as:
More details on the experimental procedure and data analysis are reported elsewhere [12].
3. EXPERIMENTAL PROCEDURE
In this study we were primarily interested in the effects of primary aliphatic alcohols on activated sludge oxygen consumption. We examined the homologous series of primary aliphatic alcohols from n-propanol to n-octanol. Consistent with work of Beaubien [6], alcohol concentration was expressed in millimoles per litre (mmoles/e) and biomass concentration in grams of volatile suspended solids per litre (g VSS/e). The ratio alcohol/ microorganisms C* thus obtained (mmoles/g VSS) allowed us to make comparison among the alcohols. The activated sludge samples we used contained from 1.5 to 3.0 g VSS/!2, which for a 1.O litre sample assured an oxygen consumption of 5 to 50 mQover a 30 minute period well within the operating constraints and precision of the measuring system. The alcohol concentration ranges varied with the length of the carbon chains. At high concentrations we were limited either by excessive foaming or, in the case of octanol, by solubility in water. In a first step, we examined the effects of the alcohols on the microorganisms in a endogenous respiration state by varying C* (alcohol/biomass). Each test lasted 30 mi-
572 16 I
.-.
. vl
ul w
\
m
N
0
m E
\/
n- butanol
x\
n- hexanol
c
Id+
n- heptanol 2 t
0
I
I
I
I
I
I
I
2
4
6
8
10
12
c*(mmol e s / g VSS) Fig. 2. Specific oxygen consumption as a function of the ratio of alcohol/microorganisms for different alcohols.
nutes, and we examined both the specific oxygen consumption RT*and variations in the specific respiration rate SOUR during the test. After these experiments we decided to examine the effect of one alcohol, pentanol-1 , on biodegradation efficiency of the biomass. This was done by measuring the evolution of the respiration rate after contacting the activated sludges with a rapidly-assimilated substrate. We chose beef extract as the substrate due to the amount of information available on this substance [ 101. A control with no added pentanol-1 was compared with three levels of pentanol-1 dosing, and the experiments were carried out for up to 4 hours to obtain the long-term evolutions of the respiration rates. Finally, two series of 30 minute experiments were carried out. In the first series the dose of pentanol-1 was fixed and the concentrations of beef extract were varied, while in the second series the doses of pentanol-1 were varied for a fixed concentration of beef extract. We examined the effects of these two parameters on the integrated specific oxygen consumption &*. The sludges were considered to be subjected to conditions representative of those occuring in the aeration basin of a wastewater treatment plant receiving such effluents.
4. RESULTS
The integrated specific oxygen consumption RT*as a function of C * (alcohol/biomass) is shown in Fig. 2 for the alcohols examined. It should be noted that for butanol the doses went as high as 98.0 mmoles/g VSS with a corresponding drop in RT* to 2.8 mg O2/g VSS; this part of the curve was omitted for clarity. The curve has the same general form for all alcohols, with two different zones as a function of dosage: - below a critical dosage Cg, oxygen consumption increases rapidly with increasing dosage C *, up to a maximum at C * = C,*
573
2
3
4
5
6
7
0
9
Carbon c h a i n l e n g t h Fig. 3. Critical concentration as a function of carbon chain length.
-above a dosage of C,*, oxygen consumption decreases with increasing dosage C*, possibly going below the endogenous respiration. As can be seen in Fig. 3, the critical dosage C,* decreases with increasing carbon chain length of the alcohol. If in place of RT* one examines the maximum specific oxygen uptake rate SOUR,,, which occurs during the first couple of minutes of the experiment, one finds the same two-zone curve. These results, shown for n-pentanol in Fig. 4, compare well with the maximum metabolic energy release found by Beaubien [ 6 ] , shown in Fig. 5. A positive relationship with carbon chain length is also found, as shown in Fig. 6 . The examinations of long-term respiration rate trends help distinguish the physiological effects of doses below and above the critical point. 1) - for dosage C: below the critical dosage C,* (Fig. 7), the respiration rate eventually returns to the endogenous level, analogous to a typical situation with a completely biodegradable substrate [ 101
574 0 .t
0.1
h
2 >
0.4
cn \ S .r
E
\ N
0
0.2
0-l
E
v
X
fa E
5
0.2
0 v,
0 .I
0
10
20
c*( mmol e s / g
30
40
VSS)
Fig. 4. Maximum specific oxygen uptake rate as a function of the ratio of alcohol/microorganismsfor pentanol-1.
2) - for a dosage C: above the critical dosage C,* (Fig. 7), the respiration rate falls below the endogenous respiration rate RE after 150 minutes, eventually falling almost to zero. In two tests, beef extract was injected after the sludges had been in contact with this dosage. Moreover, the specific respiration rate of these sludges in response to an addition of beef extract (Fig. 8B) is essentially the same as the one obtained with sludges not having been exposed to the alcohol (Fig. 8A) -after 150 min, when RT = RE, the reaction to beef extract is less than 50% of normal (Fig. 8C) -after 210 min, when RT = RE, the reaction of RT* to beef extra-.t is less than endogenous respiration and RT falls rapidly toward zero (Fig. 8D). The results show the inhibiting effect of high dosages of alcohol, but also suggest that at dosages below the critical dosage C,*, the alcohols might act as a biodegradable sub-
575
90 80
70
60 h
2
E
v
50
x
2
40
Ic
c,
a
I
30 20 10
0
1 20
I 40
I 60
P e n t a n o l (rnrnol e s / g
1 80
I
100
vss j
Fig. 5. Maximum released thermal energy as a function of the ratio of alcohol/microorganisms for pentanol-1 (from Beaubien [ 61).
strate. If this hypothesis was correct, alcohol would be a competing substrate with beef extract. However, let us consider the effect of small dosages of pentanol on activated sludges contacted with beeg extract. In the absence of alcohol, the specific consumption of oxygen RT*is proportional to the substrate concentration up to a certain point. At still higher concentrations, the assimilation of substrate is limited by the biomass present and RT* approaches an asymptotic value, as shown in Fig. 9. This asymptote is a measure of the maximum substrate that can be assimilated during a given time period, and is a function of the biomass present. Also shown in Fig. 9 is the same situation except with a small dosage of pentanol (66 X 1 O - j mmoles/g VSS). Oxygen consumption reaches a notably higher level than in the absence of a dosage. This is in direct contradiction with the hypothesis of competing substrates, under which the two curves should be almost the same. Microorganism activation by the alcohol is thus indicated. Finally, consider the case of sludges saturated with beef extract (i.e. near the asymptote for oxygen consumption). When alcohol is added in increasing dosages there is a further increase in oxygen consumption RT*above the asymptotic value obtained with beef extract alone (C* = 0) toward a new asymptote, as shown in Fig. 10. This curve
576
h
C .r
E
v, v,
>
-. 01
N
0
cn E
v
X 5
E
m
2 0
v,
0.3 3
4
5
6
7
8
Carbon c h a i n l e n g t h Fig. 6. Maximum specific oxygen uptake rate SOUR,,
as a function of carbon chain length.
indicates the asymptotic behavior of microorganism activation produced by increasing dosage of alcohol.
5. DISCUSSION
When one microorganisms in the endogenous respiration state and submit them to a dosage of alcohol, their oxygen consumption in 30 minutes RT increases with increasing dosage up to a critical dosage C,. Above this critical dosage, the oxygen consumption ET as well as the maximum specific oxygen uptake rate SOUR decrease with increasing dosage. There are therefore two dosages C1 and C, for which ET is equal, but C1 < C, < C,. The difference between the responses to these two dosages is essentially in the respiration rate RT, which in the second case falls over time almost to zero. Also, microorganisms subjected to a dosage of alcohol above C,* are no longer able to normally metabolize the substrate: their biodegradation efficiency is irreversibly affected.
577
0.4 I
..-.
C1* = 0 . 2 mmole/g VSS
m
> v)
C7* = 9.2 rnnmle/g VSS
\
I
3 0
.
Endogeneous respiration rate 60
\-I20
180
T i m e (minutes)
Fig. 7. Specific respiration rate over time for two different doses of pentanol-1, one above and one below the critical concentration C$.
The action of primary aliphatic alcohols on the biological functions of the cytoplasmic membrane of bacteria has been the object of several studies [13-161. The effects are frequently biphasic, i.e. stimulatory at low concentrations, inhibitory at high concentrations. This phenomenon is due to a modification of the structure of the cell membrane. The major barrier to transport of solutes across the membrane is the double layer of lipids, whose rigid structure and orientation result in low diffusion ocefficients in the absence of additives. In low concentrations the alcohols dissolved in the liquid state augment the fluidity and reduce the density of the lipids, thus allowing the diffusion of solutes to increase. Additional molecules of substrate can then become available for assimilation, resulting in increased respiration and a higher level of saturation. This can also explain the observed effect on microorganisms in the endogenous state, where the substrate is considered to be exhausted, but in fact many substances remain that are normally difficult to assimilate. The addition of alcohol would allow assimilation of some of these substances, with a consequent increase in respiration. However, this does not
578 0.4
c E cn
0.3
.C
\
c u
0 . 2 mmole/g 210 m i n
C9.2 mmole/g 159 m i n
\ -
0 0,
L Control no p e n t a n o l )
0.2
E
R*,
v
RT* 8.09
lA
wl >
4.22
\
+
&
0.1
------
RE
A 0
10
20
min
30 0
10
20
min
30 0
10
20
33 0
10
min
20
30
min
Fig. 8. Specific respiration rates for beef extract with three different does of pentanol and a control (no pentanol).
exclude the possibility of an activation of enzymatic hydrolysis of certain substrates due to a modification of substrate aggregates by the alcohol. At higher concentrations of alcohol, the increase in fluidity and decrease in density of the lipids leads to a loss of optimal interaction among the lipids or between the lipids and the proteins, which alters inversibly the activity of membrane enzymes. Finally, for the alcohols studied (n-butanol to n-octanol), the critical concentration falls with increasing size of the alcohol molecule, in agreement with microcalorimetric measurements of Beaubien [ 6 ] .This indicates that the determining character of the action of the alcohols is their hydrophobic nature; the longer the carbon chain of the alcohol, the more soluble it is in lipids and the greater its ability to modify the membrane structure and the enzyme activity.
6. CONCLUSIONS
Our experiments have allowed us to elucidate the mechanisms of the action of aliphatic alcohols on activated sludges. The increased oxygen consumption of microorganisms caused by these agents is thought to be due to their action on the cell membrane, essentially via their liposolubility (i.e. the hydrophobic parts of their molecules). Due to their hydroxyl radical alcohols are tensio-active molecules and the results obtained here may be indicative of the action of detergents on activated sludges. In effect these compounds could act on the cellular functions in a fashion more or less similar depending on their
579
vss
= 3.47 g/e
-
ye
e-
0
Beef e x t r a c t Beef e x t r a c t
+
66 x mmoles pentanol/g VSS
I 500
I 1000
I 1500
I 2000
2500
S*(mg o*/g V S S ) Fig. 9. Specific oxygen consumption over 30 minutes as a function of beef extract doses, with and without alcohol.
hydrophobic parts regardless of whether their hydrophilic parts are anionic, cationic or non-ionic. We are however in a position to affirm that certain compounds that may contribute little to the pollutant charge, at least in the normal sense of biochemical oxygen demand, may nonetheless have profound effects on the rate of oxygen consumption in an aeration basin, and thus the level of dissolved oxygen. Since these two parameters can be used to control an activated sludge process, it is important to be aware of such an interference, in addition to the effect of such tensio-active agents on oxygen transfer at the gas-liquid interface, the principal effect considered up until now. It is these conclusions that have drawn us to undertake an additional study on the influence of synthetic detergents on activated sludges, a study currently near completion.
580 20
VSS = 2.72
g/l
t v, Ln
> cn \ cu 0
cn E
* I-
v
ICL
0
0.I
0.2
0.3
0.4
0.5
C*(mmoles/g VSS) Fig. 10. Specific oxygen consumption over 30 minutes for a constant limiting dose of beef extract as a function of alcohol dose.
REFERENCES
1 K. H. Mancy and W. E. Barlage, Mechanism of Interference of Surface Active Agents, in Aeration Systems, in E. F. Gloyna and W. W. Eckenfelder (Eds.), Adv. in Water Quality Improvement, Univ. of Texas Press, 1968, 262-286. 2 N. I. McClelland, The Effect of Surface Active Agents on Substrate Utilization in an Experimental Activated Sludge Systems, Ph. D. Thesis. Univ. of Michigan, 1968, p. 97. 3 H. Roques, Fondements thkoriques du traitement biologiques des eaux, Volume 1, Technique et documentation, Paris, 1979, p. 871. 4 R. C. Clifft and J. F. Andrews, Predicting the Dynamic of Oxygen Utilization in the Activated Sludge Process, J. Water Poll. Cont. Fed., 53 (71, 1981, 1219-1232. 5 M. Bacquerot, Influence des solides en suspension et des matieres tensio-actives sur le transfert d’oxyghe dans un bassin de boues activies, Mkmoire de Mailrise k s Sciences Appliqukes, Sherbrooke, 1982.
58 1 6 A. Beaubien, Etude microcalorimhtrique des phhnomhes d’intoxication dans les processus d’6puration biologique, MBmoire de Maitrise 6s Sciences, Sherbrooke, 1982. 7 L. Yerushalmi, B. Volesky and J. H. T. Long, Metabolic Heat Relationships for Aerobic Yeast Respiration and Fermentation, Proc. 2nd World Congress of Chemical Engineering, MontrBal, Canada, October 4-9,1981, Vol. 1, paper 16.5.29. 8 W. W. Eckenfelder, and E. L. Barnhart, The Effect of Organic Substances o n the Transfer of Oxygen from Air Bubbles in Water, A.1.Ch.E. Journal, 7 (4), 1961, 631-634. 9 G. Veits, Moglichkeiten der Respirationsmessung, GWF/Wasser-Abwasser, 20 (S), 1979, 21 1-215. 10 N. Therien et F. Ilhan, Relating BOD, with Oxygen Uptake Rate Measurements Using Automatic Respirometers in View of Process Monitoring and Control, in R. Arthur (Ed.), Proc. 2nd Annual Activated Sludge Process Conference, Chicago, Illinois, Nov. 3-4, 1982, Ann Arbor Science. 1 1 Standard Methods for the Examination of Water and Wastewater, APHA-AWWA-WPCF (Eds), 15th edition, New York, 1981,92-94. 12 P. Le Calve, MBmoire de Maitrise &sSciences AppliquBes, Sherbrooke, 1983. 13 B. Fourcans and M. K. Jain, Role of Phospholipids in Transport and Enzymatic Reactions, Adv. Lipid Res., 12 (1974), 147-227. 14 M. K. Jain, D. G. Toussaint and E. H. Cordes, Kinetics of Water Penetration into Unsonicated Liposomes: Effects of n-Akanols and Cholesterol, j. Membrane Biol., 14 (1973), 1-16. 15 G. Lenaz and G. Curatola, Perturbations of Membrane Fluidity, J. Bioenergetics, 7 (1975), 223-229. 16 M. K. Jain and E. H. Cordes, Effect of n-Alkanols on the Rate of Hydrolysis of Egg Phosphatidylcholine, J . MembraneBiol., 14 (1973), 101-118.
This Page Intentionally Left Blank
CHAPTER I X
AIR POLLUTION
This Page Intentionally Left Blank
585
AIRBORNE POLLUTION PROBLEMS IN SANTIAGO, CHILE
A. TRIER
Dpto. de Fisica, Universidad de Santiago de Chile
1. INTRODUCTION
The city of Santiago, population approximately 4 million, is located at 33'30' latitude South. Its urban spread is estimated at 1800 km2 [I]. The climate is semi-arid with annual rainfall below 400 mm. Conditions of thermal inversion prevail over Santiago most of the year, with the inversion layer at 300 to 450 m altitude [l]. The inversion is broken mainly a) during the summer months December and January and b) by rain-generating fronts in the winter months having a frequency of about every 8 to 1 0 days [l]. Air pollution levels as measured by total particulates and CO are exceedingly hi& [ 2 , 31. Particulate loadings in excess of 500 pg/m3 have been observed in high-volume sampling while loadings in excess of 300 pg/m3 are frequent [ 2 , 41.
2. SOME ESTIMATES OF Pb AND S EMISSIONS. POLLUTION LEVELS
Pollutants of particular concern in Santiago are Pb and S. Some estimates for 1981 emissions of these pollutants into the Santiago valley will now be presented. These estimates are based on the amount of petroleum-derived fuels marketed in Region Metropolitana (the administrative division in which the Santiago valley is located) and on average Pb and S contents of fuels [5, 61. I t must be kept in mind that actual Pb and S contents can vary markedly from one refinery batch to the next. On the basis described Pb emissions due to motor traffic can be estimated at 180 tonnes for 1981. Estimates for sulphur will be incomplete, not including data pertaining to LPG, coal, wood and heavy fuels which are not at hand. The 1981 sulphur emissions due to gasoline and diesel fuels in the Santiago valley can be estimated at 1450 tonnes. Kerosene, which is used widely for residential heating [7] and cooking, adds another 270 tonnes. Santiago city gas is derived from petroleum products; its contribution to sulphur emissions cannot be estimated with the data at hand. Several remarks are in order. Intrusions of sulphur pollutants from copper smelter operations outside the Santiago valley cannot be ruled out [8] and have been claimed by some private sources. The pollutant contribution by motor vehicles weighs heavily in a less industrialised country such as Chile: ground transportation (excluding railways)
used up 39.3% of the 1981 fuel consumption expressed in energy units [9]. Fuel energy consumption by this branch of the economy in fact exceeded by 23.4% that of all of industry and mining combined [9]. Sulphur content of fuels will depend on the source of the crude oil. Chilean crudes are relatively low-sulphur but imports amounting to 54% of total consumption [6] come mainly from high-sulphur Caribbean supplies [5]. The situation described above translates as follows. Lead concentrations of about 1 pg/m3 are easily observed in the autumn and winter months of April through June [4, 101 in atmospheric particulates. Average concentrations (1977 through 1980) as high as 55 pg/m3 SO, have been reported [2] in the autumn and winter months April through September. Ministerial authorities are looking into the possibility of eliminating lead additives to gasoline [5]. Sulphur emissions appear at present to be an unmanageable problem.
3. OBSERVING PARTICULATE POLLUTION IN SANTIAGO. SOME RESULTS
Observations on airborne particulates have been carried out at a fixed station on the Universidad de Santiago de Chile campus west of downtown Santiago since 1980. 3.1. Particulate fallout had been collected for 30-day periods for .13 months [ 111. Yield ranged from below 5 mg to above 100 mg. The water insoluble fraction of this fallout matter was analysed by X-ray fluorescence (XRF) and net elemental Ka or La intensities were extracted. Highly significant linear correlations between the Al, Si, K , Ca, Ti and Fe intensities were found. The data on the whole support the hypothesis that fallout matter is predominanthy of ‘natural’ origin. Size measurements on fallout particles will be referred to below. 3.2. Suspended particulates have been sampled at various rates ranging from 0.7 l/min to 10 m3/h. Both total and dichotomous sampling is being done. Ashless filter paper and polycarbonate or teflon membranes have been found useful as filtering media. Ashless filter paper requires smaller corrections for trace elements in XRF analytical work. When mass concentration data are required however the membrane filters are to be preferred. Sampling times range from 6 to 12 hours in daily sampling but have been as long as a week. Mass concentrations are deduced from net masses obtained by weighing. Samples are analysed by XRF and net elemental Ka or La intensities are extracted from the X-ray spectra. As a first approximation these net intensities can be regarded as proportional to the respective elemental concentrations. Work is in progress to make the analytical results quantitative. 3.3. Results are available for weekly samples covering a period of over 40 weeks. Total sampling at an average rate of about 0.7 l/min was done. Both ashless filter paper and polycarbonate membranes have been used in this work. Net collected mass determined on polycarbonate filters ranged from 0.25 to 15.41 mg.Sample thickness ranged from 1 8 to 11 10 pg/cm2. On all filters, both ashless paper and polycarbonate, Al, Si, S, Ca, K, Ti, Fe, Zn, Pb were detected by XRF. Bromine was detected on most filters. Si, S, Fe and Ca were always prominent in the X-ray spectra. Of the elements listed, Al, Si, Ca, K, Ti and Fe can be considered as being mainly of ‘natural’ origin [ l l , 121 while S, Zn and Pb are recognized as ‘man-made’ pollutants. Highly significant linear correlations are found between the net X-ray intensities observed on the samples for the elements listed. In parti-
587 cular highly significant Si/S, Pb/S and Zn/S are found. These results suggest that the cycle periods for the S, Pb and Zn pollutant concentrations are significantly shorter than one week [13]. 3.4. Results are available for dichotomous sampling carried out for a period of over one year. Total sampling rate averages approximately 1 m3/h. Nominal cut-point is 2.5 pm. Nominal inlet cut-point is 1 5 pm. Teflon membrane filters and ashless paper have been used in this work. Ashless paper has been found quite satisfactory when XRF analysis is the main concern. Sampling times have ranged from 12 hours down to 6 hours. Teflon filters in the fine fraction channel have been found to clog after 6 to 8 hours in the autumn and winter months April through June. Most of the samples have been taken in daytime hours and night sampling is now in progress. Net X-ray intensities for the elements of ‘natural’ origin correlate very well but no significant correlations have been found for pollutants S , Zn, Pb. This appears to be consistent with the conclusions of 3.3. Most the Al, Si, Ca and Fe is concentrated in the coarse fraction. Atmospheric particulate mass concentrations decrease by a factor as high as 7 after one or two days of rain but recover in 24 hours. Work is being pursued. 3.5. Size measurements by microscope according to Feret’s criterion have been carried out on atmospheric particulates [ 141. Observations have been made of fallout matter and on samples obtained at pumping rates of 0.7 l/min and 2.5 and 10 m3/h (figures are approximate). The size of fallout particles expressed as mean geometric diameter ranges widely. Particles collected by pumping average a mean geometric Feret’s diameter of 10.0 pm. This figure was found for instance in August 1980 and in August 1981. Runs of 300 measurements were made but 100 is now deemed adequate.
4. CONCLUSION
Some aspects only of atmospheric pollution in Santiago, Chile, and pertinent experimental work have been touched upon. A considerable integrated effort is still required to have a better graps of these problems and to be able to formulate reliable predictive models. In particular, estimates of the economic impact of air pollution are required since any effective measures to abate pollution will be costly to implement.
5. ACKNOWLEDGEMENTS
The author wishes to thank J . Ward, C. Rojas P. and J . Valdks L. for assistance with the XRF work and M. E. Cantillano, L. Melo and W. Fuentes for collecting samples. This work is supported by Direcci6n de Investigaciones Cientificas y Tecnol6gicas DICYT/USACH.
REFERENCES 1
2
R. Riesco, P. U. Catblica de Chile, Santiago, unpublished notes, 1982. Laboratorio de Contaminacibn Arnbiental, Ministerio de Salud, Annual Reports, Santiago, unpublished.
3 4 5 6 7 8 9 10 11 12 13 14
INTECCHILE, Corfo, unpublished report, Santiago, 1978. A. Trier et al., unpublished results. A. Villasante and S. J. Navarro, Empresa Nacional del Petr6leo/Chile, private communication, 1983. ENAP, Boletin Estadistico, Volumen No. 82, Santiago, 1982. E. Lissi et al., Bol. SOC.Chil. Quim. 27 (1982-01) 376. E. Lissi et al., Contribuciones Cientificas y Tecnol6gicas No. 57, 1983. USACH 1983, USACH. ENDESA, Producci6n y Consumo de Energfa en Chile 1981, Santiago, 1982. J. L. Ortiz et al., Bol. SOC.Chil. Quim 27 (1982-01) 286. A. Trier, Bol. SOC.Chil. Quim. 27 (1982-04) 25. M. M. P r h d ez and J . L. Ortiz, Bol. SOC.ChiL Quim. 27 (1982-01) 283. R . J. Bryan, in: Air Pollution, Third Edition, Volume 111, Academic Press, New York, 1976. A. Trier, Contribuciones Cientificas y Recnol6gicas USACH, No. 50, Santiago, September 1981 ; also No. 58, 1983
589
FLUE GAS DESULF'HURISATION USING LIME WASTE
E. M. BULEWICZ, C. JURYS, S. KANDEFER
Technical University of Cracow, Poland
ABSTRACT When coal is burned in fluidiscd beds, at 8O0-85O0C, over 70% dcsulphurisation of thc flue gases can be achieved by feeding in with thc coal a 2 t o 3 - fold excess of SO- sorbent, limestone or dolomite, and removing the CaSO, formed together with the ash. The drawbacks are a narrow optimum temperature range and large amounts of sorbent required. Using a laboratory - size fluid bed combustor it has been shown that lime wastc from C , H , production, consisting mostly of Ca(OH), , is considerably more reactive towards SO: and ovcr a wider temperature range, 720-920°C, than any natural sorbent. A numbcr of possible ways of utilising lime waste in fluid bed combustors is discussed.
1. INTRODUCTION
In dry high temperature desulphurisation under oxidising conditions sulphur dioxide can be captured as CaS04 using as sorbents natural Ca - containing minerals, such as limestone or dolomite (mainly CaC03 and MgCa(C03), respectively) [ l , 21. The CaC03 in these minerals first undergoes calcination to CaO, and then the overall process takes place CaO +SO,
+ 1/2 0, = CaS04
This reaction, however, does not go to completion. With limestone about 35% conversion appears to be the maximum attainable [3, 4, 51. With dolomites the degree of Ca utilisation is normally higher, but the MgO hardly reacts at all [3, 51. Moreover, the process is strongly temperature - ~ t m a x i m u mefficiency falling within the range 780-890'C [4, 61. At lower temperature the CaC03 in the sorbent cannot calcine [7], although there is evidence that a direct reaction between CaC03 and SO, is still possible [8]. The fall in the efficiency of SO, capture at higher temperatures is less readily explained, and a number of theories have been proposed to account for the experimental facts [3, 61. The required temperature regime and an oxidising atmosphere can be readily realised in fluid bed combustors (FBC) - desulphurisation can then take place in the fluid bed (FB) itself, if the sorbent, suitably crushed, is fed in with the coal and the CaS04 form-
590 ed removed with the ash [ I , 91. However, since on the average to achieve, say, 80% desulphurisation of the flue gases when coal containing 3% of sulphur is burned the amount of sorbent required is at least 1/4 of the weight of the coal - with approximately doubling of solid wastes, ash plus spent sorbent, much effort has gone in recent years into a search for better sorbents [5, 101 and into ways of increasing the reactivity of limestone [ I I , 12, 131.
2. EXPERIMENTAL
We have been considering the possibility of using Ca - containing materials other than natural minerals. One such material is lime waste (LW) from acetylene production via calcium carbide, consisting mainly of Ca(OH), . The composition of LW dried in air one month after production is shown in Table 1. CaC03 and CaS04 are present because of Tab. 1. Chemical composition of lime waste (LW) Compound
% b y wt.
Ca (Ot I) CaCO , CaSO, . 2 H :0 Moisture Impurities stable up to 1ooo"c (SiO:, Sic, etc.) Combustibles Total
10.7 11.7 8.0 4.0
3.1 0.4 97.9
Error limits % 1.5 0.7 t 1.5 0.7 f
f
*
f 0.7
+0.1
the capture of CO, and SO, from the air. Some impurities come from the original carbide and others can be described as accidental. The analysis has been carried out using thermogravimetric methods (Derivatograph, OD 102, System Paulik, Erdey, Hungary). A typical derivatogram of LW is shown in Fig. 1, which shows the decomposition temperature of the main constituents. One month - old LW contains about 60% water. When this is dried, there is practically no volume change and as a result the dry material is very porous and light. The total porosity is about 66%. The porosity of LW was studied using a mercury porosimeter (Carlo Erba Porosimeter 200, Italy), pressures up to 200 MPa. The distribution of pore size is illustrated in Fig. 2. Pores of about 1.O pm dlameter are responsible for most of the total pore volume. The reactivity of dried LW towards SO, was asessed using a laboratory - size FBC working atmospheric pressure. The combustor was 96 mm in diameter and 330 mm high. Its thermal power was about 10 kW. Crushed coal was used as fuel. The flue gases wete analysed continously for SO,, 0, , CO and CO, and contained typically 0.33% SO, and 4-5% 0,. A continuous record of the flue gas and FB temperatures was also abtained. The apparatus was designed so that selected parameters could be kept constant at the desired level, or could be programmed to change with time. The LW sorbent was dried, crushed and separated into narrow size fractions between 0.5 and 2.0 mm. The object of the experiments was to determine how much SO, the LW could take up under optimum
59 1
-1oc
Fig. 1. Derivatogram for LW
Fig. 2 . Pore size distribution for dried LW
592
conditions and t o check whether the temperature dependence of the process was the same as with natural sorbents. Two types of experiment were carried out; the LW was added batchwise and the amount of SO, taken up at constant temperature calculated from gas analysis records, o r it was fed in continuously so that the molar ratio of Ca (as Ca(OH), and CaC03) to S evolved was near unity, while the temperature of the FB was varied within the range 720-980°C. The temperature changes were stepwise and this produced abrupt changes in the gas analysis records. The effect of temperature on desulphurisation efficiency of LW is illustrated in Fig. 3. The capacity of LW for CO, capture is excellent - under FB conditions it can take up 0.60 g SO, per 1 g of the dried material compared t o about 0.18 g SO, per 1 g of average
8’
i I
0.4 700
1
1
,
800 1
1
1
1
1
900 ‘C 1 temperatur 1 1 1
-
Fig. 3. Desulphurisation efficiency - temperature dependence
natural limestone. Also, even with a thin fluidised layer, c.a. 150 mm, corresponding to time of contact of only 0.15 seconds, it is possible to achieve over 90% desulphurisation at a much lower Ca/S ratio than with any natural sorbent [S, 101. It is also characteristic of LW that its reactivity hardly depends o n granulation, at least for sizes between 0.5 ad 2.0 mm, which can be attributed to its great porosity, so that virtually the whole of a sorbent grain is available for reaction. It can be seen from Fig. 3 that between about 72O-92O0C the desulphurisation efficiency of LW is only very weakly temperature dependent. On the high temperature side the desulphurisation efficiency falls rapidly - at 980°C there is practically no SO, capture. Below 72OoC some of the drop in efficiency can probably be attributed t o the fact that the CaC03 contained in the LW (Fig. 1) can no longer calcine. Near 7OOoC coal combustion becomes incomplete and the CO concentration in the flue gases begins to rise rapidly. It is, however, highly probable that LW can capture SO, right down to room temperature. Thus LW is an excellent SO, sorbent for FBC. However, the dry product is obtained directly only with ‘dry’ acetylene production. It could be used either without any prior treatment, or after crushing, which would improve fluidisation, but could cause the fines to be elutriated from the FB, increasing the load on the cyclones. When acetylene is produced by the ‘wet’ method LW is obtained as a thin slurry -
593 this is a nuisance, since the slurry has t o spend a number of years in sedimentation reserVoixs to settle to a semi-solid consistency. LW taken from a sedimentation reservoir after settling contains 20-30% of water. Drying it would use up energy, but the semi-wet state may be an advantage here, since this makes hydraulic transport a possibility Extrusion of a paste under pressure can be an alternative to crushing, with the additional advantage of diminished elutriation. Other methods may also be considered. Dry LW or the semi-solid material after sedimentation have a definite commercial value for the building industry, but the fresh slurry is a burden and has to be fed into large sedimentation reservoirs where it tends to accumulate. Transport costs and/or drying immediately would be prohibitive. Another way out is, however, possible. The Ca(OH), in LW settles in several phases. The first phase in relatively rapid, taking only a few days, depending on the depth of the reservoir. The early setting material contains 35-40% of solids and over the next year or so this increases by only no more than 5%. Further concentration is possible only if water can be removed by slow filtration. With sedimentation in ponds, the solids content can be brought up to about 70% after several years. It should, however, be possible to consider the use of LW as a SO, sorbent after the first sedimentation stage. Although the possibility of direct addition of a slurry to FBC is excluded it could be sprayed over the coal, particularly if it is to be dried before use. Removal of excess water by filtration is possible, but not attractive. The fact that at the early stages LW still contains about 60% of water could, however, be turned to an advantage. The fines elutriated from a FBC usually contain an appreciable proportion of unburnt fuel enough to lower the overall thermal efficiency of a boiler installation to a significant degree. Most FBC designs make provision for returning the fines from the first cyclone back to the combustor [14], or a special carbon bum-up cell is envisaged [15]. With the first solution the unburnt fines tend to be elutriated again before combustion can be complete, the second is cumbersome and complicates the design. If the fines from the first cyclone were mixed with LW to a paste consistency and the paste extruded onto the FB, there could not be any immediate carry-over of unburnt material. The LW would serve a double purpose - binder for cyclone fines and SO, sorbent, or sorbent additional to limestone fed in with the coal. Another advantage would lie in the re-activation for any elutriated but not fully utilised limestone sorbent [16]. It can be easily shown thet the amount of heat necessary to evaporate the water acoompanying the LW is very substantially less than the heat of combustion of the material from the first cyclone. The various possible ways of utitising LW in FBC are represented in Fig. 4. If such suggestions are to be considered seriously, it is necessaryto take into account the hydraulic properties of LW after the first sedimentation stage. The viscosity of this material is of prime importance for hydraulic transport and sorbent preparation. Ca(OH), in water tends to exhibit flocculation and LW has the properties of a non-Newtonian liquid. The viscosity of LW was measured using the spinning cylinder method (Rheotest 2 instrument, DDR). Samples of the desired concentrations were prepared by adding water to LW after the first sedimantation stage. The measurements have shown that the viscosity increases with concentration, but decreases markedly with increasing velocity gradients. This is shown in Fig. 5. Suspensions with concentrations greater that 11-15% have the properties of a viscoelastic body, with resistance to shear stress 3-5 times higher that after passing into the liquid state. These properties are not constant with time. When the
594
Fig. 4. Possible ways of utilisin? LW in FBC installation
LW concentration 0
- 0.386
-0.296 x
- 0.175
Fig. 5. LW - hydraulic properties
material is agitated, the viscosity decreases, but the change is not fully reversible. This is probably associated with the physical destruction of filament - like Ca(OH), agglomerates. The hydraulic properties of LW, when properly taken into account, would not hinder transport and hydraulic methods.
595 CONCLUSIONS
1. LW in the dry state is an excellent sorbent for SO, under FBC conditions. Its capacity for combining with SO, is at least twice that of an average limestone, and its reactivity reactivity remains high over a wider temperature range. 2. LW cannot be recommended as a universal SO, sorbent for FBC applications because its availability is limited and transport over long distances would not be economical. 3. The greatest potential of LW lies in the possibility of its use as a binder for first cyclone material. The mixture, in the form of extruded paste, could be fed back to the FB, where the LW would act as SO, sorbent additional to limestone, improvinng the overall of desulphurisation. 4. LW at concentrations greater than 11-15% is a viscoelastic body, becoming a non-Newtonian liquid on agitation, but its hydraulic properties are such as to make hydraulic transport and handling possible. 5. The proposed use for LW is environmentally attractive, since a material whose accumulation causes problems is used to reduce the level of atmospheric pollution resulting from coal combustion. REFERENCES
1 D.G. Skinner, The Fluidised Combustion of Coal, National Coal Borad, Res. and Develop. Dept., London 1970. 2 R. Snyder, W. Wilson, I. Johnson, Thermochim. Acta, 26, 1978, 257-267. 3 N.H. Ulerich, W.G. Vaux, R.A. Newby, D.L. Keairns, Experimental Engineering Support for EPA‘s FBC Program: Final Report, Vol. 1, Sulfur Oxide Control, EPA - 600/7-80-015a, Jan. 1980. 4 N.H. Ulerich, E.P. O’Neill, M.A. Alvin, D.L. Keairns, Criteria for the Selection of SO, Sorbents for Atmospheric Pressure Fluidised Bed Combustors, Final Report, Vol. 2. Project 721-1, EPRI FP-1307, 1979. 5 N.H. Ulerich, R.A. Newby, D.L. Keairns, Thermochim. Acta, 36, 1980, 1-16. 6 E.P. O’Neill, D.L. Keairns, AIChE Symp. Ser., 73, 1976, 100-107. 7 E.P. O’Neill, D.L. Keairns, W.F. Kittle, Thermochim Acta, 14, 1976, 209-220. 8 G. Van Houte, L. Rodrique, M. Genet, B. Delmon, Environmental Science and Technology, 15, 1981, 327-332. 9 L. Yaverbaum, Fluidised Bed Combustion of Coal and Waste Materials, Noyes Data Corporation, Park Rodge, N.J., USA 1977. 10 R.A. Newby, D.L. Keairns, M.M. Ahmed, The Selection of Design and Operating Conditions for AFBC to Meet Environmental Constraints, Technical Report, 80-9E3-ENPRO-P3, March 1981. 11 J.A. Shearer, I. Johnson, C.B. Turner, Enviromental Science and Technology, 13, 1979, 1113-1118. 12 S. Erlich, Inst. of Fuel Symp. Series no 1, Fluidised Combustion Conference, Vol. 1, C4, 1-8, London 1975. 1 3 G . Van Houte, B. Delmon, J.C. Maon, Ph. Dumont, J. of the Air Pollution Control Association, 28,1978,1030-1033. 14 J. Makansi, B. Schwieger, Fluidised Bed Boilers, Power, Ang. 1982, SlLS15. 15 P.F. Fennelly, D.F. Durocher, H. Klemm, R.R. Hall, Preliminary Enviromental Assessment of Coal-Fired Fluidized-Bed Combustion System, Us Dept. of Commerce, NTIS, PB-269 556, May 1977. 16 J.A. Shearer, G.W. Smith, D.S. Moulton, C.B. Turner, K.M. Myles, J. Johnson, Proc. Intersoc. Energy Convers. Engineering Conf., Aug. 1980.
This Page Intentionally Left Blank
597
DESUL'PHURIZATION OF GAS BY SORPTION OF SO2 ON CUPRIC OXIDE DEPOSITED ON ALUMINA PARTICLES IN A FLUIDIZED BED REACTOR
D. BARRETEAU and C. LAGUERIE
Institut du G h i e Chimique L A . CNRS No, I92 - Chemin de la Loge, 31 078 Tolouse Cedex, France
ABSTRACT The paper deals with desulphurization of air, NO,, CO, and water mixtures by chemical sorption of SO, on cupric oxide deposited on porous alumina particles.
1. INTRODUCTION
Among the flue gas desulphurization processes presently studied or in operation, those using the reactions of SO, with solid metal oxides are of particular interest [l]. They are carried out at temperatures between 300°C and 4OO0C, which correspond to usual flue gas temperatures. However, the sulphates formed are sometimes difficult to decompose. This problem may be avoided by using manganese oxide, alkanized alumina, or cupric oxide, but the mechanical strength of the solid particles must be sufficient to prevent dust pollution and the gas pressure drop must be small. An example of this kind of processes is the Shell process [2] which uses two fixed bed reactors operating alternately in sorption and in regeneration at the same temperature (300-4OO0C). The reaction which takes place is the following: CUO + 1/2 0, + so, -+ cuso; which is slightly exothermic. The cupric oxide is deposited on porous alumina particles. It should be noted that if the copper concentration is not sufficient (less than 3%) other compounds resulting from the reaction of SO, with alumina appear and the activity of the solids decreases. The sulphate formed is reduced by methane or carbon monoxide and hydrogen mixtures, at the same temperature, to copper which is easily oxidized by oxygen: CuSO, + 112 CH, +. Cu + 112 CO, + H,O + SO,
598
A hopper
B solid v a l v e
C sol id preheat er
Y---
D reactor E gas p r e h e a t e r F rotameters G manometer
Fig. 1. Desulphurization unit
Tab. 1 . Physical properties of the sorbent Density Particle diameter Minimum fluidization velocity at 20°C Specific area of unreacted sorbent Specific area of SO, saturated sorbent Mean pore diameter Weight fraction of copper (%)
1600 kg m - 3 400 pm 0.30 m s-' 184 mzg-' 161 rn'g-' 55 A 3.17-4.12
The exit gas from the regeneration reactor contains about 30 to 35% of SO,. It can be either treated in a CLAUS unit to produce sulphur or oxidized to SO, on a catalyst and then transformed into sulphuric acid. The Shell process is not really continuous. Each cycle takes one hour. Therefore this reaction can be conveniently carried out in a fluidized bed reactor. The main advantage of t h s is to make the operation continuous and the exit SO, concentration constant. With one stage, Barreteau [3] obtained a 50% desulphurization yield. With model simulation, he predicted that sulphurization yields of more than 90% could be obtained using a four stage reactor. The present work has been undertaken in order to try to verify this
599 Tab. 2. Influence of gas flowrate, solid flowrate and height of downcomers o n desulphuration yields Run.No.
101
102
103
104
105
106
107
108
109
110
111
112
2 5 2 5 2 5 2 5 3.5 3.5 3.5 3.5 10 10 18 18 10 10 18 18 14 14 14 14 30 30 30 30 60 60 60 60 45 45 45 45 1950 1924 1924 1933 1927 1965 1973 1998 1903 1888 1922 1913 297 298 293 292 298 297 291 295 296 293 294 294 375 314 376 328 380 380 346 340 350 354 360 360 273 42 998 568 95 30 705 243 250 265 294 200 50 1135 558 641 636 140 1316 932 370 710 755 592 910 330 1494 1195 798 184 1465 1000 978 1038 1097 942 1381 902 1694 1557 1212 567 1706 1402 1355 1400 1450 134 303 300 299 300 298 305 300 303 306 306 305 307 298 297 302 302 295 292 302 293 304 304 301 302 301 300 304 303 300 292 303 300 306 304 303 304 306 306 305 306 308 305 308 306 310 306 306 307 84
84
86
86
125
140
131
146
93
93
93
93
Tab. 3. Influence of temperature and inlet SO, concentration Run.No.
201
202
1910 1924 326 35 1 330 435 83 40 348 189 643 444 1130 957 330 358 328 355 329 357 334 364 99
98
203
301
302
303
304
305
306
1924 398 480 31 73 207 650 395 389 394 406
495 290 347 10 30 68 158 301 301 302 304
986 289 367 40 112 242 480 308 302 300 302
1500 290 353 92 316 568 902 301 299 299 302
2908 291 348 558 1267 1692 2158 309 307 306 306
3963 297 350 913 2013 2600 3200 308 305 305 309
4850 294 340 1583 2950 3608 4133 309 308 307 307
96
98
101
97
101
105
101
Q s= 3.5 kg h - ' ;Qg = 14 m 3h - ' ; L
=
45 m m
prediction. The influence of water, carbon dioxide and nitrogen oxide concentrations on the desulphurization yield have also been studied.
2. APPARATUS
The apparatus is shown schematically in Fig. 1. It has been described in full detail elsewhere [4, 51. The reactor is made of refractory steel. Four stages, 150 mm high and 98 mm in diameter are stacked vertically. The distributor for each stage is a perforated plate. Solids flown from one stage to another through downcomers. Pressure gauges, thermocouples, gas and solids sampling valves are provided. Solids, added from a hopper,
600
300
330
3fO
T 'C
40,O
Fig. 2. Desulphurization yields versus temperature
Fig. 3. Desulphurization yields versus inlet SO, concentration
and gas mixtures are preheated in two fluidized beds before being fed into the reactor itself. Gas mixtures are analysed either by gas chromatography or by infrared spectrometry. 4 commutating valve selects the gas samples from different stages. Solids leaving the reactor are stored and regenerated noncontinuously in another reactor by a reaction with methane.
60 1 Tab. 4. Influence of NO,, CO, and water vapour concentration on desulphuration yield ~~~~~
% (vol) Run No.
CO, %
H,O %
ppm NO,
P,
11 17 11 17 11 17 11 17 14 14 14 14
2 2 10 10 2 2 10 10 6 6 6 6 0
100 100 100 100 500 5 00 500 500 300 300 300 300
P, P,
P, P, P, P,
P, P, PI, PI, PI,
s, s2
s3
s,
0 0 0 0
0 0 0 0
0 0 0
Ce ppm SO,
Tge "C
Tpe "C
2026 2172 2125 2121 2190 1913 2176 2106 1837 1990 1891 1972 2054 1990 1880 2050
298 299 303 300 301 300 304 302 304 300 299 298 301 299 298 299
299 300 298 304 308 301 302 301 299 306 295 307 305 309 300 299
"C
AP cm o f H,O
C, ppm SO,
%
296 296 299 301 302 299 301 299 305 306 302 301 301 301 302 300
2.3 2.3 2.4 2.3 2.2 2.3 2.4 2.3 2.3 2.3 2.3 2.3 2.2 2.3 2.2 2.3
1114 1279 1296 1379 1204 1088 1284 1326 1186 1278 1186 1195 1161 1240 1031 1177
45.0 41.1 39.0 35.0 45.0 43.1 41.0 37.0 35.4 35.8 37.3 39.4 43.5 37.7 45.2 42.6
T
YG
I
Ce ppm 1000
+ 2000
t
0
I
3000
50
I0 1
H20 I
20 I
Fig. 4. Influence of concentration of SO, and water vapor o n desulphurization yield
602 3. PREPARATION AND ANALYSIS OF THE SOLID
Alumina particles are first sieved and dried. They are then immersed in an aqueous copper sulphate solution (0.76 mole/l) and dried at 80°C. The sulphate is reduced by methane in the regeneration reactor and the product is oxidized in air at 300°C to give copper oxide. The physical properties of the solid are presented in Tab. 1 . All solids analysis has been carried out by "Societe Nationale ELF AQUITAINE" in order to determine the sulphur, sulphate and copper contents.
4. EXPERIMENTS -RESULTS
The sulphurization yield is defined as:
where Ce is the inlet SO, concentration in the gas and Ci its outlet concentration. Thus for stage i we have:
where Ci and Ci +1 are the outlet and inlet SO, concentrations, respectively, for the stage. These yields depend on the gas flowrate (QG) the solids flowrate (Qs) solids residence time in each stage, defined in terms of the height of the downcomer (L), temperature (T) and the water vapour, CO, and NO, Concentration. The first set of experiments was concerned with the influence of the first three parameters, at a temperature of 300°C,and with an inlet SO, concentration of 2000 ppm, on the desulphurization of air-SO, mixtures. The experiments were performed according to a Z3 factorial design. The experimental conditions lead to a mean residence time for solids between 6 and 30 minutes, gas velocity between 0.77 and 1.4 m/s. Thus contact times between solid and gas were between 0.08 and 0.3 second. In order to determine the experimental error variance, the experiment was repeated four times for the average values of the three parameters. The results are presented in Tab. 2. The yield was always above 48% and it could be as high as 98%. Gas pressure drop was less than 140 mm water, the outlet SO, concentration less then 1000 ppm and usually less than 300 ppm. It should be noted that in France, the maximum allowable SO, concentration is about 1400 ppm, while in other countries like USA, Japan or Germany, it is about 350 ppm. If the solids flowrate was suitably chosen the exit SO, concentration could fall down below 300 ppm. It may also be observed that the mechanical strength of the solid was good. This was confirmed by comparing the particle size analysis before and after 150 hours of operation. It seems that the presence of copper renders the alumina particles more resistant to attrition [3,4]. If we compare the results of the experiments, we can see that the desulphurization
603
yields are higher when the solids flowrate and solids hold-up are higher, and the gas flowrate lower. A statistical analysis of these results shows the same thing [4]. In a second set of experiments the influence of temperature between 300 and 400'C and of the inlet SO, concentration between 500 and 5000 ppm was determined. The results are presented in Tab. 3 and Figs. 2 and 3. The reaction rate depends on temperature [5], the desulphurization yield higher when the temperature is increased, as can be seen from Fig. 2. The value of the yield Y, obtained in the first stage at 400°C is certainly wrong because of measurement error in C1 and C, which amount in this case to less than 100 ppm. The overall desulphurization yield approaches 100% for temperatures above 330°C. Increasing inlet SO, concentration affects the desulphurization yield adversly, due to an increase in the sulphate concentration, which results in a decrease of the sorbent activity. It should be noted that in all the experiments the first stage yield is always higher because the solid which enters this stage does not yet contain any copper sulphate. Finally, the third set of experiments was performed to assess the effect of CO, NO, and H,O concentration. Note that the effect of the 0, concentrations was not considered though it can be thought that it would affect the reaction kinetics. Only one stage of the reactor was used so that the desulphurization yield was less than in the preceeding experiments. A factorial design was still chosen with CO, concentration between 11 and 17%, H,O between 2-10% and NO, between 100 and 150 ppm. The results were compared with those for experiments without CO, NO, and water. They are presented in Tab. 4. Nitric oxide does not have any influence on the desulphurisation yield. The presence of carbon dioxide leads to a slight decrease of this yield. The water vapor exhibits the strongest influence. A statistical analysis of the results [6] confirms this observation. CO, molecules are bigger than the others (N,, O,, H,O) and thus can slow down the diffusion of SO, in the pores of the solid and reduce the desulphurization yield. Other experiments were performed by using mixtures of air, SO, and water. The results are reported in Fig. 4. It can be noted that the yield passes through a maximum with water concentration between 6 to 10% except for SO, concentration of 1000 ppm. At low concentrations, water seems to promote the desulphurization, but for higher concentrations, an adverse effect is observed. Influence of water concentration can be explained in different ways: - as in the corrosion of metals, which is enhanced by the presence of water, the oxidation of SO, to SO,, is assisted by water vapor so that the sulphatation of copper oxide becomes faster, - water can interact with alumina and copper sulphate to give hydrated salts. The bluegreen colour of some particles tends to confirm this supposition. The hydration of the sorbent could reduce its activity. These two opposed effects could explain the observed variations of the yield. For experiments at the lowest SO, concentration, the second effect is more likely.
5. CONCLUSION
Desulphurization yields of more than 90% can be obtained by chemical sorption of SO, on copper oxide deposited on porous alumina particles in a fluidized bed reactor
604
working at about 300°C. The SOz content of the exit gas can be reduced to less than 300 ppm. The gas pressure drop is not prohibitive and the mechanical strength of the solids is sufficient to avoid dust pollution. The presence of the other components of the flue gas leads to a slight decrease in the desulphurization yield. This could be compensated by a better choice of the operating conditions. The process could then be used to control SO, pollution by flue gas.
REFERENCES
1 G . Van Houte, “Soixante procedis de disulfuration”, Office International de la Librairie, Bruxelles 1973 2 F. M. Dautzenberg and J . E. Nader, Chem. Eng. Progr., 67 (1971) p. 8 3 D. Barreteau and H . Angelino, The Can. J. Of. Chem. Eng., vol. 56 (1978) p. 570 4 M. Vizcarra Mendoza and C. Laguerie, Environmental Technology Letters, vol. 2 (1981) pp. 215-224 5 R. J. Best and J. G . Yates, Ind. Eng. Chem. Proc. Des. Dev., 15 (1976) p. 2 6 D. Barreteau, H . Yi Duran, K. Iovtchev and C. Lagukrie, Submitted for publication to Environmental Technology Letters
NOTATIONS inlet SO, concentration (ppm) outlet SO, concentration for stage i (ppm) downcomer height (mm) gas flowrate (N m3 h - l ) solid flowrate (kg h - l ) reactor temperature c ) stage i temperature (‘c) inlet gas temperature (“0 inlet solid temperature (‘C) desulphuration yield of the stage i (%) desulphuration yield for the whole reactor (%)
e
YG
605
NITROGEN OXIDES EMMISION CONTROL CDL/VITOK ENHANCED ABSORPTION PROCESS
B. J. MAYLAND
Chenoweth Development Laboratories, Inc., Laouisville, Kentucky, USA L. D. ROLAND
Foster Wheeler Energy Limited Reading, England
ABSTRACT The CDL/Vitokenhanced absorption process provides an attractive solution to the problem of NOX emissions from chemical manufacturing plants. The technology was developed from pilot plants, full scale demonstration plants and detailed analyses of many operating absorbers. The process uses the princinple of scrubbing tail gas with nitric acid under conditions which reduce the nitrogen oxides to the desired level. Both physical absorption and stripping and chemical oxidation absorption are used. No chemicals other than water are required for the process. All the nitrogen oxides removed from the tail gas are converted to nitric acid at useful concentrations. The technology covers a wide range of operating conditions. A number of commercial installations of the process have been made including both the retrofitting of existing plants and new plants. These handle tail gases from the manufacturing of inorganics and nitration of organics as well as from nitric acid plants.
1. INTRODUCTION
Nitrogen oxide emissions from chemical manufacturing operations are a continiung problem for the industry. The technology is available for the effective reduction of the emissions and substantial progress has been made. More rapid progress has been impeded by the complicated balance between economic forces and the desire to minimize the pollution of the environment with esthetic considerations having a peripheral influence. This is especially true of operations that have been in service for a number of years and based on process design practices of some earlier period. The CDL/Vitok Process [ 13 was developed to cope with this situation. The technology covers a wide range of conditions that are encountered in various industries and the process is well suited for retrofit installations with minimum interruption of operations. Because of the range of conditions that can be handled, incorporation of the process in
606
the design of new installations can be attractive, allowing for greater flexibility in the choice of equipment and operating conditions for an optimum design. The versatility of the technology has shown the process to be compatible with many industrial operations having NOX emissions potentially harmful to the environment. Some of these are catalyst manufacturing, pickling operations in steel mills, nitration in explosives and organics, and inorganics manufacturing. Nitric acid manufacturing accounts for much of the NOX emissions. In the past few years many possible applications of the process to solve industrial operating problems have been investigated. Some of these have not been resolved and are still pending. Experiences with the successful application of the process are summarized in this paper along with some interesting prospects for the technology. For those not familiar with the CDL/Vitok Process some of the characteristics are as follows: a. The nitrogen oxides in gaseous effluents can be reduced to levels less than 1.5 kilograms of NOX as nitrogen dioxide per metric ton of nitric acid produced in nitric acid manufacturing or the equivalent in other types of manufacturing. b. Nitrogen oxides are recovered as useful nitric acid. c. Operation is stable and requires little operator attention and no additional operating manpower. d. Energy requirements are moderate. e. Equipment requirements and materials of construction are conventional. f. No additional chemicals or feed materials other than the normally present oxygen and water to from nitric acid are required. The economics of the process for any given situation depend upon a variety of factors but in many cases the process can show a positive pay-out because of the recovered nitric acid. Occasionally an increase in production capacity is possible with the incorporation of the process.
2. TECHNICAL DISCUSSION
The essential function of the process is to reduce the nitrogen oxide emissions from whatever source to useful nitric acid product. This involves oxidation, hydrolysis, and dissolution in aqueous nitric acid. The complexity of the chemistry, thermodynamic equilibria, reaction rates and mass transfer rates has been extensively covered in the literature [ 2 ] . The process is carried out in conventional absorption equipment. In the case of a nitric acid plant or other type of plant where absorption equipment is part of the system, the process takes the form of modifying conditions to improve the performance of the existing equipment and adding equipment as necessary. In other cases an absorption system needs to be included but the size of the equipment is minimized by using more efficient process design conditions. Process design criteria have been developed using operating data from a number of sources. Over seventy-five industrial NOX absorption systems have been investigated. Information on some of these is sufficiently complete that detailed analyses of in-
607
cremental sections of the absorption system have been made. The validity of the correlations has been verified in full scale test programs as well as pilot plant programs [3, 41. For convenience the enhanced absorption process can be divided into two modes, absorption by chemical oxidation and physical absorption.
2.1. Chemical Oxidation
The complexity of the phenomena involved presents a considerable challenge to the process designer. Various mathematical models have been proposed and calculation procedures have been developed for the design of absorption systems. With the advent of the ubiquitous computer more sophisticated procedures are feasible. For a given type of absorption equipment, the procedures are a useful tool in the process design. The success of the calculation procedures depends upon adjustments of the various rate constants to allow for changing conditions in the absorption system, In this respect the more sophisticated procedures are similar to simpler procedures. In either case the design depends upon a suitable choice of constants. Representing the oxidation absorption mode as a pseudo first order reaction, the reaction rate constant is given by k = (F/W) Ln(NOXl /NOX,) where F is the standard volumetric gas flow, W is the increment of superficial absorption volume and NOXl and NOX, are the nitrogen oxide concentrations in the gas entering and leaving the incremental absorption volume. The available data have been used to identify the important variables and determine quantitative effects on the rate constant. The variables that have been studied include the following: a. b. c. d. e. f.
Type of packing in packed absorbers Method of packing, whether tumbled or stacked Style of tray, whether bubble cap or perforated Tray spacing Pressure level Temperature level g. Nitrogen oxide concentration in the gas phase h. Residual oxygen concentration i. Cocurrent or countercurrent flow j. Liquid distribution in packed beds k. Liquid phase acid concentration 1. Bed depth in packed beds m. Liquid loading Pressure and temperature are the most significant process design variables in determining the value of the rate constant. For a pressure range of 1 to 10 atmospheres the rate constant may increase twenty-fold or more. Similarly, large increases are observed when the temperature is lowered from around 40 to below 5 centigrade.
608 The rate constant tends to drop off at very low values of NOX concentration as has been reported in other studies. This increases the difficulty of achieving very low levels of NOX emissions by oxidation absorption alone. Especially at low pressure levels, the size of the absorption equipment becomes prohibitive to achieve desired degrees of abatement. The residual oxygen level is not a major factor but the rate constant does decrease below some concentration depending upon the pressure of operation. Concentrations as low as 1% have been observed without seriously adverse results under some conditions. H g h emissions that have been reported with low oxygen concentration are usually related to concurrently h g h plant throughput which overloads the absorption system. The most efficient use of absorber volume is achieved with the highest wetted surface area per unit volume. Thus packings with favorable characteristics and provision for proper liquid loading and distribution give best results in packed columns. There is little difference between bubble cap trays and perforated trays if the hydraulics of the design avoids dry or partially dry trays. The dedication of extended space between trays €or gas phase oxidation is the least efficient use of column space. There is very little effect on the rate constant between countercurrent or cocurrent flow. However, where liquid recycle is used to improve the liquid loading of the packing or to control temperature, NOX bypassing can be appreciable with countercurrent flow. This is caused by the stripping from the liquid at the top of the absorption section of NOX physically absorbed by contact with gas richer in NOX at the bottom of the section.
2.2. Physical Absorption
The second mode of absorption relates to the dissolved nitrogen oxides that exist in the liquid phase in an incompletely oxidized state. This would include such species as NO, NO,, N,04, or HNO, . Which species or the relative amounts ordinarily are not determined by the analytical techniques but unique relationships can still be derived for process design purposes. In the liquid-gas thermodynamic equilibrium among nitrogen oxides, nitric acid, water and oxygen, the ratios among NO, NO,, and N,O4 in the gas phase are determined by the nitric acid concentration in the liquid phase for a given NOX level, the temperature and the pressure. In addition the liquid phase will contain an equilibrium amount of incompletely oxidized nitrogen oxides. The reactions involved in maintaining the equilibrium are readily reversible. These reversible reactions provide the basis for an absorption-stripping operation. The reactions are relatively fast compared to the oxidation reaction even at low levels of NOX. In combination with the oxidation absorption, physical absorption provides an attractive means of attaining low levels of NOX emissions with high recovery of NOX as nitric acid product. The usefulness of physical absorption is limited by the capacity of the liquid phase or the amount of dissolved nitrogen oxides at equilibrium. This can be increased many-fold by lowering the temperature of absorption. Curves have been developed relating NOX partial pressure, dissolved nitrogen oxides, temperature level and acid concentration. These are used in establishing operating conditions for the desired level of abatement.
609 To obtain nitric acid product at useful concentrations the water or weak acid feed to an absorption system must be limited in quantity to satisfy the material balance. The pinch point for the physical absorption step can be considerably improved by increasing the liquid to gas ratio. This is done by introducing an internal recycle taken from an appropriate point in the absorption system, stripped to low levels of dissolved nitrogen oxides and fed to the physical absorption section.
3. COMMERCIALIZATION OF THE PROCESS
The application of the technology to new plants is a relatively straightforward process design problem. Various design criteria can be postulated and the most economically attractive design determined by the usual methods. In the case of nitric acid plants, the technology tends towards a more viable low pressure design. The application of the technology to existing plants encounters constraints of equipment parameters as well as production schedules. Many existing nitric acid plants as well as other type of plants have been investigated to determine the feasibility of using the process to achieve an acceptable level of NOX abatement. The absorption equipment includes bubbler tanks, packed columns, bubble-cap tray and perforated tray columns. Single column and multi-bed and multi-column absorption systems are used. As would be expected deviations from predicted plant performance have been encountered because of equipment or operating problems. These usually arise from lack of technical supervision or insufficient plant maintenance. Typical equipment problems are as follows: a. Trays operating dry or partially dry because of leaking or off-level trays aggravated by a low liquid to gas ratio b. Dry packing because of channeling or poor liquid distribution c. Partial by-passing of the absorption system because of leaks in the heat exchange train In nitric acid plants low efficiency in the ammonia burning step reduces the feed-water requirement which can increase the importance of items 1 and 2 . The most frequent cause of emissions exceeding the original plant design is excess throughput. Cooling water problems, heat exchange fouling and excessive ambient temperatures often contribute to excessive emissions. Preceeding any retrofit design, an analysis of the operating performance of the plant is necessary to determine the condition of the equipment and the operating procedures. While the addition of the retrofit abatement package will improve the performance of most plants, obviously the first step is to bring the existing plant to an acceptable state. 4. TYPICAL RETROFIT DESIGN
There are many nitric acid plants designed a number of years ago that are still operating with nitrogen oxide emissions at an undersirable level. Some of the low pressure plants with absorption pressures of 2 to 4 kg/sq cm gage fall in this category. The absorption system already containts considerable active volume relative to the plant capacity. Multiple columns with multiple beds are used in the absorption train. Reducing the NOX
610
emissions by adding more absorption capacity becomes inordinately expensive and sometimes counter-productive because of pressure drop considerations. The enhanced absorption process is ideally suited for retrofitting this type of plant because of the flexibility of the multi-step absorption system. Reduction of the NOX emissions to low visibility or to levels below 200 ppm is feasible. In either case the enhanced absorption process would be incorporated in approximately the last third of the existing absorption train by adjusting operating conditions and modifying fluid flow patterns. 4uxiliary equipment would be added to accomplish the changes. The simplified flow sheet of a retrofitted low pressure nitric acid plant absorption train is shown in Figure 1. This plant uses multiple packed columns with twelve beds to produce nitric acid at 55% concentration. Only the last three beds are shown. The bulk of the enhanced physical absorption occurs in the penultimate bed and the bulk of the enhanced oxidative absorption occurs in the preceeding bed. There is some degree of overlapping.
Fig. 1. NOX abatement for low pressure multiple absorption train.
A recycle stream is picked up at the base of absorption bed Number 10. The stream is stripped by contact with air in an auxiliary column before returning to the absorption system at the top of bed Number 12. Air for stripping is obtained by by-passing part of the main bleach air from the plant. The recycle liquid acts as a refrigeration medium and improves the liquid to gas ratio for the physical absorption. Refrigeration is used to achieve absorption temperatures lower than those obtainable with cooling water. A package refrigeration machine is shown but in some plants considerable refrigeration is available from vaporization of ammonia process feed. The refrigeration load can be minimized by heat exchange with the optimum amount being dictated by economics. In the design shown, refrigeration is recovered from the tail gas in bed Number 12 and by exchange with the warm process gas in the lower part of bed Number 10.
A small stream o f nitric acid product is added to the cjrculating stream o f the ultimate tower to give a more favorable liquid phase concentration. The normal process water as
61 1 required by the material balance is also added at an appropriate point as shown. Heating of the recirculated acid stream for stripping is obtained by direct injection of steam. Indirect heating may be preferable in some cases to avoid contamination of the system by impurities in the steam.
5 . COMMERCIAL EXPERIENCE
One of the earlier units to be installed as a retrofit t o an operating plant was at an army ammunition plant. In the manufacture of tri-nitrotoluene, the recovery of unconverted nitric acid for recycle and the reduction of nitrogen oxides emissions is an important aspect of the operation. The normal recovery system consisted of multiple absorption columns and oxidation vessels analogous to the absorption system in a nitric acid plant. The absorption system operated at near atmospheric pressure and the system was relatively large and inefficient. Consequently the emissions were high. In this particular installation the emissions were reduced from over 10,000 ppm to less than 600 ppm. The plant was operated on an experimental and intermittent basis during 1974 and 1975. Based upon the experience described above, a nitrogen oxide recovery system was incorporated in a catalyst manufacturing operation. Nitric acid is used for the dissolution of metal and the resulting metal nitrate solution is used to impregnate a catalyst carrier. Nitrogen oxides are generated during the dissolution as well as during heat processing of the catalyst. The nitrogen oxide recovery system was designed to recover more than 99.2% of the nitric acid used as 35% acid which could be recycled. Again this installation operated at near atmospheric pressure and was designed to reduce emissions to levels equivalent to those considered desirable for nitric acid plants. Because of the recycle system used dilution air was added to the system at an appropriate point to reduce stack coloration which was also minimized by stack design. In a nitric acid-ammonium nitrate complex the potential refrigeration from the vaporization of ammonia is of such a temperature level and quantity that it provides a good match for the abatement process. Installations of this type have been made and satisfactory nitrogen oxide abatement levels have been achieved, With favorable values of such variables as ambient temperature, cooling water temperature, and plant through-put among others, emission levels of the desired 1.5 kg/metric ton can be achieved in the typical high pressure nitric acid plant. Conditions can be improved to provide more operating flexibility by addition of auxiliary refrigeration and refrigeration recovery. A new nitric acid plant designed for an NOX emission level of 2000 ppm presented a challenge for the abatement process. This plant was to produce 65% acid and operate at a single pressure level of 2.8 kg/sq cm gage with power recovery. The abatement system was incorporated in the original design with only minor changes t o the absorption vessels. The completed plant achieves NOX emissions of less than 200 ppm. The increased nitric acid production with the abatement system operating is approximately 1.8%. A characteristic of the process is the more efficient utilization of absorption equipment [ 5 , 61. In the case of multiple plants in one location a single abatement system can be used without excessive repiping of the plants. A single unit was installed as a retrofit
612
to handle four small nitric acid plants in one location. The unit was skid mounted and shop fabricated to minimize field work and interruption of production.
6. CONCLUSION
The development of the technology of the CDL/Vitok enhanced absorption process has provided the means for a practical solution to the problem of NOX emissions over a wide range of conditions. The study of a great many operating plants indicates that considerable improvement in the emission levels is possible with modest addition of equipment. The performance of the process has been demonstrated in a number of commercial operations, four of which are operating at the present time. Three of the plants were incorporated in new installations and the rest were retrofits. In most applications that have been studied the abatement process shows a positive payout. The operating costs and amortization of the equipment are covered by the value of the recovered nitric acid. The economic balance is influenced to a great extent by the initial level of NOX to be abated and the degree of abatement required. The latter, of course, is most often fixed by factors other than economic. Since lower residual oxygen in the tail gas is required by the process than normally used, appreciable increased production in the case of nitric acid plants is possible. Reducing the oxygen content of the tail gas by one percentage point will allow approximately 4% increase in nitric acid production. The operating difficulties have not been severe, especially in the case of new plant installations. In the case of retrofit installations peak performance is not reached immediately unless the existing plant is in a good state of maintenance and operator control. In either case the operating labor or supervision is not appreciably different than with the usual or original absorption system. Installation presents no problem in new plants. In the case of retrofitted plants preparations can be made during regular maintenance shut-downs for piping connections. The added equipment for the abatement system can be installed while the plant is operating. Actual down-time for connecting the abatement system to the main plant can be done in one day. The main plant can then continue to operate normally and the abatement system brought on stream when convenient.
REFERENCES 1 B. J. Mayland, U.S. Patent 4,081,517, G.B. Patent 1,544,106 and Can. Patent 1,069,675. 2 B. J. Mayland, Hydrocarbon Processing, p. 141, May 1972. 3 B. J. Mayland and R. C. Heinze, in Recent Advances in Air Pollution Control, R. W. Coughlin, R. D. Siege1 and Charanjit Rai (Eds.), AICHE Symp. Series, 137, p. 83, 1974. 4 B. J. Mayland and R. C. Heinze, Chemical Engineering Progress, Vol. 69, No. 5, p. 75, May 1973. 5 B. J. Mayland, in Proceedings o f Environmental Syrnp., The Fertilizer Institute, p. 143, Jan. 1976. 6 B. J. Mayland, in The Control of Gaseous Sulphur and Nitrogen Compound Emission, European Federation of Chemical Engineering, 2nd International Conference, Vol. 1, April 1976.
CHAPTER X
PANEL DISCUSSION ON POSSIBLE DIRECTION OF RESEARCH AND DEVELOPMENT RELATING TO CHEMSTRY FOR THE PROTECTION OF THE ENVIRONMENT
This Page Intentionally Left Blank
615
POSSIBLE DIRECTION OF RESEARCH AND DEVELOPMENT RELATING TO CHEMISTRY FOR THE PROTECTION OF THE ENVIRONMENT
CHAIRMAN: D R W. J. LACY COCHAIRMAN: D R B. A. BOLT0 RAPPORTEURS: D R R. J . MARTIN, D R P. SEREICO
This Page Intentionally Left Blank
617
On Thursday afternoon of 22nd September 1983 in the Ecole Nationale Superieur de Chimie of Tolouse, a meeting was held to discuss the possible direction of research and development relating to chemistry for the protection of the environment. Following the presentation of Paul Sabatier Medals to Mr. W.J. Lacy and Dr. L. Pawlowski, Mr. L. Lacy took the chair. The Chairman told the international group assembled that those present were there as individuals rather than as representatives of particular organizations and therefore personal opinions, and not collective viewpoints, were to be a feature of the session. Fifteen delegates had been invited to speak. Each of these speakers was limited to a time of five minutes. The following represents a summary of the statement of each speaker.
(i) Dr. F. ElGohary (Egypt)
Dr. El-Gohary stated that although the last thirty years had seen extensive progress in the protection of water resources from pollution, we were entering an era in which conventional treatment processes were not fully adequate. Research and development in new chemical treatment processes together with a greater emphasis on treatment efficiency were considered necessary. Minimization of waste of raw materials, maximization of recycling of wastewater and optimization of energy usage were cited in the need for greater efficiency. (ii) Dr. L. P. Jackson (USA)
Dr. Jackson stated that long term problems in water and air pollution would result from the world’s accelerating hunger for energy unless industrial energy processes could be adequately controlled. Whilst those engineers in the energy business are concerned with the impact of harmful chemicals on the environment, improved communication between chemical engineers and environmental chemists would help to place that control on a more technical and objective basis. (iii) Dr. R . J. Martin (UK)
Dr. Martin stated that since the middle 1970’s, the world had witnessed an energy crisis, inflation and industrial recession. Perhaps neqer as before therefore, it must be shown that environmental standards in general, and the reasons leadhg to the formulation of these standards, haae a sound economic basis. Every penny must conunt and what is more, it should be seen to be counted. A sense of perspective must be maintained therefore in order to prevent the introduction of standards for which sound technical and economic support cannot be shown. (iv) Dr. P. Sereico (USA)
Dr. Sereico stated that concern in the USA and elsewhere in the world over the pro-
618
duction of trihalomethanes in the chlorination of water supply should lead to a logical evaluation of chlorination in both water and wastewater treatment. The benefits of chlorination should be contrasted with the long term health risks associated with chlorinated organic compounds. Alternatives to chlorinating agents would have to be found if those health risks were judged to be unacceptable. (v) Dr. B. A. Bolto (Australia)
Dr. Bolto stated that there was a need for new water and wastewater treatment processes. In developed countries, lower capital costs would characterize these new processes whereas in developing countries the emphasis would be on appropriate technology. Greater attention should be given to the recycling of products from wastewaters and to the removal of heavy metals from wastewater sludges. The significance of water supply for agriculture in Australia meant that methods of removal of salts and turbidity should be explored. (vi) Dr. A. Hamza (Egypt)
Dr. Hamza stated that the monitoring of wastes was a major concern worldwide and that in the future, there should be more attention paid to the monitoring of industrial and domestic wastewater treatment processes. Industry from developed countries was welcome in Egypt, but overseas aid, employment and commercial growth should not obscure responsibility over the impact of hazardous wastes on the environment. (vii) Dr. R. Ben Aim (France)
Dr. Ben Aim stated that wastewater treatment commonly involved the treatment of dilute aqueous systems. Research should consider the behaviour of micropollutants in dilute systems. Increasing use of physicochemical processes in wastewater treatment and increasing use of biological processes in water treatment has resulted in increasing similarity in water and wastewater treatment. Greater emphasis on the optimization of treatment plant operation should be of priority. (viii) Dr. L. Pawlowski (Poland)
Dr. Pawlowslu stated that treatment processes should become cheaper and more efficient. In order that these objectives could be met, a greater understating of the mechanisms by which these treatment processes removed particular pollutants was necessary. This understating could only result from greather chemical knowledge of the processes, the pollutants and their inter-relationships. (ix) Dr. G . N. Pandey (India)
Dr. Pandey stated that numerous research objectives were necessary in Iqdia. The significance of agriculture was stressed in that research should investigate the environmental impact of pesticides and fertilizers; a balance between energy use and environmental degradation should be sought in the production of food. Industrial waste treatment, corrosion research and the development of low energy technology were all of importance, as was the need to improve greater communication between environmental scientists and the medical profession so that the effects of organic pollutants on human system could be understood.
619 (x) Dr. L. H. Wang (Taiwan)
Dr. Wang highlighted areas of concern in Taiwan. The extremely high population density means that waste disposal is of major importance; disposal of solid wastes is hindered by shortage of land. Air pollution problems have been encountered; the presence of the highly toxic dioxin has been observed. The presence of nuclear energy plants has resulted in the discharge of nuclear wastes into seawater; research should consider the polluting effects of such wastes on seawater and its ecology. (xi) Dr. L. Liberti (Italy)
Dr. Liberti reported on two on-going research projects in Italy. One project is studying the interception of wastes from towns on the Adriatic coast; ammoniacal nitrogen and phosphates present in effluents are being converted to ammonium phosphate fertilizers. The benefits are two-fold: fertilizer production and clean-up of the Adriatic Sea. The second project is studying the re-use of wastewater for agriculture; large maturation ponds are used to collect and store effluents. Sunlight has been observed to inactivate viruses. (xii) Dr. U. Zoller (Israel)
Dr. Zoller felt that the essential issue was not what can be done in future years; it is what should be done. Decisions on what should be done depend on economy and various social constraints. The future should see greater emphasis on environmental education and ultimate disposal of pollutants; inadequate information is at present available on the final effects of pollutants on the world ecosystem.
(xiii) Dr. V. Soldatov (USSR)
Dr. Soldatov welcomed the conference and the fact that it had become a tradition for environmental scientists throughout the world since its conception in the mid-1970’s. However, the broadness of the conference was questioned; a more specific theme for each biennal conference was opined to be desirable. (xiv) Dr. H. Zimny (Poland)
Dr. Zimny stated that in the last thirty years, the emission of oxides of nitrogen and sulphur had significantly increased. Research was necessary to investigate the effects of such air pollution on soil and the plants and animals living in and on thnt soil. ( x v ) M r . L. Roland ( U K )
Mr. Roland reminded the assembled delegates that he, as a representative of industry, was in a minority surrounded by academic researchers. The economicl facts of life in industry meant that future research and development should be conducted with economic realism and without courting hysteria from pressure groups and the media. The Chairman then invited comments from other delegates present for the discussion session. Dr. A. L. Kowal (Poland) wanted to see more research on the utilization of wastewater carried out; re-use of water and recovery of products would be an investment for the future. Dr. A. Trier (Chde) pointed out the need for air pollution standards in de-
620
veloping countries. Various unidentified speakers stressed the need for the use of nontoxic products in agricultural applications, the significance of acid precipitation arising from air pollution and the importance of international co-operation in the fight against environmental pollution.
In his summing up, the Chairman concluded that the session had been of immense value; environmental problems likely to occur within the next five years had been identified by the internationally recognized experts assembled. The solution of these problems would be aided by communication and informed collaboration; the biennial conference would continue as a valuable forum for such communication involving the sharing of problems and courses of remedial action. The value of a conference of this sort is to present ideas and findings so as to avoid duplication of effort, add to the general knowledge while helping the other fellow solve his problem. This meeting had two purposes: 1. Formal presentations, and much more valuable 2. ”eyeball to eyeball” contact between individuals. There is no way to establish a dollar value for an idea and if we all left the conference with one more idea than we had the conference was a success.
621
AUTHOR INDEX
Alaerts G., 47 Andre L., 553 Baiocchi C., 381 Baleix A., 155 Barreteteau D., 597 Barcicki J., 5 Barnes D., 559 Ben Aim R., 618 Blasco D., 33 Bliss P. J., 559 Bolto B. A., 5 , 4 9 1 , 6 1 7 Bowron J. M., 109 Bulewicz E. M., 589 Caussade B., 155 Chae-Shik RHO., 65 Chen K. Y., 193 Chikuma M., 365 Dalmacija B., 245 Deryko A., 297 Dickson K. L., 119 Dixon D. R., 179 Dolejs P., 169 Drew W. M., 5 37 Dziubek A. M., 205
El. Gohary F., 6 17 Filipiak E., 437 Foldi-Poly& K., 373 Francois R. J.. 221 Gennaro M. C., 381 George J., 155 Gorzka Z., 437,451 Grauer R. B., 559 Gray K. A., 407 Grochulska-Segal E. M., 421 Gutkowski B., 285 HamzaA., 618 Haute A. A., 213,221 Hlavay J., 373 Holder G. H. A,, 537 Holmes R. F., 9 3 Hupka J., 253,269,285
Itoh K., 365 Jackson L. P., 617 Jaroniec M., 297 Jaroszyhska-Wolhkka J., 445 Jury; C., 589 Kandefer S., 589 Karlovic E., 245 Kaimierczak M., 437 Kolarik L. O., 179 Korte F., 457 Kotowski M., 491 Kowal A. L., 205, 343 Kowalska M., 285 Kozak Z., 5 Kuo C. H., 559 Lacy W. J., 2,93,615 Lafrance P., 3 1 3 Laguerie C., 597 Lamotte M., 133 Le Calve P., 569 Le Croirec P., 463 Liberti L., 513, 619 Lopez A., 5 1 3 Lore F., 235 Majewska-Nowak K., 387 Mansour M., 457 Marengo E., 381 Martin G., 463 Martin R. J., 329, 617 Masclet P., 133 Mathieu J., 155 Mayland B. J., 605 Mazet M., 313 Mch’eill R., 491 Mentasti E., 381 MikaCibaka A., 399 Miller J . D., 253 Miskovic D., 245 Mydlarczyk St., 285 Nakayama M., 365 Ng W. J., 329 Niekko J. Z., 485 Oblad A. G., 253
Inczedy J., 373
Palmisano N., 235 Pandey G . N., 618 Parlar H., 45 7 Passino R., 513 Pawkowski L., 2,5,491, 618 Pirnazari M., 525 Plat J. Y., 553 Poll0 I., 445 Poranek A., 399 Reynes A,, 155 Richardson M. L., 109 Robins K., 559 Rodgers J. H. jr., 119 Roland L. D., 605, 619 Ross J. W., 143 Saavedra F., 525 Sakurai H., 365 Saleh F. Y.,119 Sarzanini C., 381 Sayag D., 553 Sereico P., 617 Socha A., 451 Soldatov V. S., 353, 619 Sozahski M. M., 421 Stevens M . R., 525 Sugijanto J., 427 Surdia N. M., 427 Synnott J. C., 143 Tanaka H., 365 Tanaka T., 365 Therien N., 569 Tiravanti G., 235 Torres L., 155 Trier A., 585 Waite T. D., 407 Wang L. H., 619 Wang Sung-bh, 193 West S. J., 143 Wilms D. A., 21 3 Winnicki T., 387 Verdier A., 2 Villessot D.. 313 Zimny H., 79,619 Zoller U., 161,619
This Page Intentionally Left Blank
623
SUBJECT INDEX
Adsorption activated carbon, 155, 193, 297, 313, 329 BET surface, 155, 313 capacity, 173, 205, 291, 313, 329, 373 colloid, 169 isotherm, 155, 205,297, 313,329 mathematicalmodel, 155, 169, 297, 313, 329 organic material, 119, 155, 169,193, 205, 297, 313, 329 separation 155, 205, 297 synthetic adsorbent, 155, 169,179, 205,297, 329, 373 Air acid rain, 35, 617 pollution, 23, 585 purification, 24,445, 589, 597,605 Analytical method air, 133, 155 water, 109, 133, 143, 161,617 wastewater, 109, 133, 143, 161, 381,427, 617 Anion exchanger arsenic removal, 373 capacity, 21, 353, 365, 373, 381, 491, 513 chromate removal, 12, 15, 381,491 fouling, 179, 353, 365, 491 microbial carrier, 179, 553 nitrate removal, 18, 21, 513 phosphate, 513 resin selectivity, 353, 365, 363, 381,491 theory, 353
Chemistry impact on the environment, 29,33,119,617 Coagulant aluminium salts, 169, 213, 221, 373 dolomite, 205 ferric salts, 179 ferrous salts. 179 Coagulation auto-coagulation, 169, 213
mechanism, 169, 179, 193, 205, 213, 221 process, 169, 179, 193, 205, 221 Demineralization fouling, 21, 353,491 mixed bed, 491 operating capacity, 353, 431 water quality, 18, 20 Energy conservation, 26, 40, 253, 617 Environment generalproblems,5, 33, 47, 93, 119, 161 developing countries, 33,65, 618 Filtration caliform bacteria removal, 179 pilot plant, 179, 269, 285 pressure, 463 process, 179, 269, 285,463 wastewater, 179, 285, 463, 491 Flocculation bentonite, 179, 193, 253 bioflocculation, 179, 193 caliform bacteria removal, 169, 329 BOD removal, 169, 193, 213 COD removal, 169, 179, 193, 213 cost, 169, 193, 213 fulvic acid removal, 17 9 huinic acid removal, 169, 179 optimal dosing, 169, 179, 193, 213 mechanism, 169, 179, 193, 213, 221 phosphorus removal, 169, 179 process, 169, 179, 193, 213, 221 theory, 169, 179 turbidity removal, 169, 179, 193, 213 Ion exchange ammonia recovery, 18, 19, 21, 513 capacity, 353, 365, 373, 381, 513 chromate recovery, 15 clinoptilolitc, 20, 365, 5 13 condensate polishing system, 353
624 decolorizing, 353 deionization of water, 18, 19, 20, 365 demineralization, 18, 19, 20 heavy metal removal, 353, 381, 491 magnetic resin, 22, 179 nickel recovery, 16 nitraterecovery, 18, 21, 513 nitrate removal, 18, 21 phosphate recovery, 21,513 selenium recovery, 365 separation factor, 179, 353 theory, 353, 365 water recovery, 10, 11, 14, 18, 20, 353 zeolite, 365 quantitative predictions, 353, 365 Modeling general problem, 109, 119 Monitoring general. problem, 93,161 laboratory evaluation, 97 sampling methods, 98,161 Oxidation decolorization, 407,437, 451 process, 399,407,421,437,445,451,457, 491,525 Recovery acetic acid, 13 ammonia, 18, 19, 21,513 condensate, 18, 19, 20 nitrate, 13, 18, 21, 513, 617 phosphate, 21,513,617 pollutant, 18, 19,617 resource, 10, 11, 18, 19, 20, 21, 161, 485,491, 513,617 selenium, 365 water, 10, 11,16, 18, 19, 20, 161, 179, 253, 491,617 Recycling technique, 8, 10, 14, 16, 18, 19,20,253,485, 491,513,617 Reduction arsenic, 373 BOD, 179,399,407,451,553 COD, 179,399,407,451,553 colour, 179, 381 nitrate, 18, 21,513 organic matter, 179,407, 451 phatogen, 407
phosphate, 21, 513 sulphate, 537 turbidity, 179, 253 Removal ammonia, 18, 19, 20, 21, 513 arsenic, 373 BOD, 179, 193, 213,329, 399,407,451,457, 553 chlorobenzene, 329,45 1 COD, 179, 193, 205,213,329, 399,407,451, 457,553 coliform, 329 detergent, 437 dye, 179,193,329,387,451 heavy metal, 11, 179, 245 iron, 421 napthalene, 119 nitrate, 18, 21, 513 manganese, 421 oil, 193,253, 269, 285 organic compound, 179, 193, 205, 213, 329, 407,451,457 phenol, 21 3, 329 phosphorus, 21, 179, 513 silver, 525 slhdge, 213 turbidity, 179, 193, 205, 213, 327 Sludge disposal, 21, 65, 179 metal accumulation, 65 toxic pollutant, 65, 119 trace element, 119 Soil contamination, 80, 619
Sorbent activated carbon, 297, 313, 329 synthetic, 179, 297, 329 Water recovery, 8, 10, 14, 16, 18, 19,20, 161, 179, 253,491,617 treatment, 47, 65, 373,421 quality, 47 recycling, 50, 617 supply, 61, 619 Wastewater Treatment activated carbon, 297, 313, 329 activated sludge, 537, 559, 569 aeration, 169, 399,463, 553
625 biological treatment, 537, 553, 559 coagulation, 161, 179, 193, 205, 213, 221, 235, 25 3 disinfection, 407,421, 617 electrochemical method, 235,451,525 electrolysis, 235,451, 525 filtration, 253, 329 flocculation, 161,179,193,205,213, 221,235 flotation, 245, 253 heavy metal removal, 65, 161, 179, 245,525, 619
ion exchange, 8, 10, 11, 16, 18, 19, 20, 21, 353, 381,491,513 neutralization, 65, 205 nitrogen fertilizer industry, 18, 19, 20 oil removal, 161,179,245,253 recycling, 8, 10, 11, 16, 18, 19, 20, 263,491, 513,619 sorption, 179, 205, 278, 297, 313, 329 ultrafiltration, 353, 387
This Page Intentionally Left Blank