Environmental Contamination (Studies in Environmental Science)

ENVIR 0NME NTAL C 0NTAMINAT10N

A selection ofpaperspresented at the 5th international Conference on Environmental Contamination, Morges, Switzerland, 29 September- 1 October 1992

Studies in Environmental Science 55

ENVIRONMENTAL C 0NTA MINAT I0N

Edited by

J.-P. Vernet Institute E-A. Forel University of Geneva Versoix, Switzerland

ELSEVIER Amsterdam

-

London

-

N e w York - Tokyo 1993

ELSEVIER SCIENCE PUBLISHERS B.V Molenwerf 1 P.O. Box211,1000AEAmsterdam,The Netherlands

L i b r a r y o f Congress C a t a l o g i n g - i n - P u b l i c a t

on D a t a

I n t e r n a t i o n a l C o n f e r e n c e on E n v i r o n m e n t a l C o n t a m i n a i o n ( 5 t h : 1 9 9 2 : Morges. S w i t z e r l a n d ) Environmental contamination : a s e l e c t i o n o f papers presented a t t h e 5 t h I n t e r n a t i o n a l C o n f e r e n c e on E n v i r o n m e n t a l C o n t a m i n a t i o n , M o r g e s , S w i t z e r l a n d , 29 S e p t e m b e r - 1 O c t o b e r 1992 / e d i t e d b y J . - P . Vernet. p. cm. -- ( S t u d i e s i n e n v i r o n m e n t a l s c i e n c e ; 55) I n c l u d e s b i b l i o g r a p h i c a l r e f e r e n c e s and i n d e x . ISBN 0-444-89868-9 I. V e r n e t . J . - P . ( J e a n - P i e r r e ) 1. P o l l u t i o n - - C o n g r e s s e s . 11. T i t l e . 111. S e r i e s . TD172.5.1545 1992 363.77--dc20 93-3600 1 CIP ISBN: 0-444-89868-9

0 1993 Elsevier Science Publishers B.V. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system or transmitted i n any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior written permission of the publisher, Elsevier Science Publishers B.V., Copyright & Permissions Department, P.O. Box 521,1000 A M Amsterdam,The Netherlands. Special regulations for readers in the USA - This publication has been registered with the Copyright Clearance Center Inc. (CCC),Salem, Massachusetts. Information can be obtained f r o m theCCCaboutconditionsunderwhich photocopiesofpartsofthispublication may be made inthe USA. All other copyright questions, includ-ing photocopying outside of the USA, should be referred to the copyright owner, Elsevier Science Publishers B.V., unless otherwise specified. No responsibility isassumed bythe publisher forany injury and/ordamageto personsor property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein. This book is printed on acid-free paper Printed i n The Netherlands

Studies in Environmental Science Other volumes in this series 1 2 3 4 5 6

7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33

Atmospheric Pollution 1978 edited by M.M. Benarie Air Pollution Reference Measurement Methods and Systems edited by T. Schneider, H.W. de Koning and L.J. Brasser Biogeochemical Cycling of Mineral-Forming Elements edited by P.A. Trudinger and D.J. Swaine Potential Industrial Carcinogens and Mutagens by L. Fishbein Industrial Waste Management by S.E. Jmrgensen Trade and Environment: A Theoretical Enquiry b y H. Siebert, J. Eichberger, R . Gronych and R. Pethig Field Worker Exposure during Pesticide Application edited by W.F. Tordoir and E.A.H. van Heemstra-Lequin Atmospheric Pollution 1980 edited by M.M. Benarie Energetics and Technology of Biological Elimination of Wastes edited by G. Milazzo Bioengineering, Thermal Physiology and Comfort edited by K. Cena and J.A. Clark Atmospheric Chemistry. Fundamental Aspects by E. Meszaros Water Supply and Health edited by H. van Lelyveld and B.C.J. Zoeteman Man under Vibration. Suffering and Protection edited by G. Bianchi, K.V. Frolov and A. Oledzki Principles of Environmental Science and Technology by S.E. Jmrgensen and I. Johnsen Disposal of Radioactive Wastes by Z. Dlou hy Mankind and Energy edited by A. Blanc-Lapierre Quality of Groundwater edited by W. van Duijvenbooden, P. Glasbergen and H. van Lelyveld Education and Safe Handling in Pesticide Application edited by E.A.H. van Heemstra-Lequin and W.F. Tordoir Physicochemical Methods for Water and Wastewater Treatment edited by L. Pawlowski Atmospheric Pollution 1982 edited by M.M. Benarie Air Pollution b y Nitrogen Oxides edited by T.Schneider and L. Grant Environmental Radioanalysis by H.A. Das, A. Faanhof and H.A. van der Sloot Chemistryfor Protection of the Environment edited by L. Pawlowski, A.J. Verdier and W.J. Lacy Determination and Assessment of Pesticide Exposure edited by M. Siewierski The Biosphere: Problems and Solutions edited by T.N. Veziroglu Chemical Events in t h e Atmosphere and their Impact o n t h e Environment edited by G.B. Marini-Bettolo Fluoride Research 1985 edited by H. Tsunoda and Ming-Ho Yu Algal Biofouling edited by L.V. Evans and K.D. Hoagland Chemistry for Protection of the Environment 1985 edited by L. Pawlowski, G. Alaerts and W.J. Lacy Acidification and its Policy Implications edited by T. Schneider Teratogens: Chemicals which Cause Birth Defects edited by V. Kolb Meyers Pesticide Chemistry by G. Matolcsy, M . Nadasy and Y. Andriska Principles of Environmental Science and Technology (second revised edition) by S.E. Jmrgensen and I. Johnsen

34 35 36 37 38 39 40 41 42 43 44 45 46 47 48

49 50 51 52 53 54

Chemistry for Protection of the Environment 1987 edited by L. Pawlowski, E. Mentasti, W.J. Lacy and C. Sarzanini Atmospheric Ozone Research and its Policy Implications edited by T. Schneider, S.D. Lee, G.J.R. Wolters and L.D. Grant Valuation Methods and Policy Making in Environmental Economics edited by H. Folmer and E.van lerland Asbestos in Natural Environment by H. Schreier How t o Conquer Air Pollution. A Japanese Experience edited by H. Nishimura Aquatic Bioenvironmental Studies: The Hanford Experience, 1944-1984 by C.D. Becker Radon in the Environment by M. Wilkening Evaluation of Environmental Data for Regulatory and Impact Assessment by S.Ramamoorthy and E. Baddaloo Environmental Biotechnology edited by A. Blazej and V. Privarova Applied Isotope Hydrogeology by F.J. Pearson Jr., W. Balderer, H.H. Loosli, B.E. Lehmann, A. Matter, Tj. Peters, H. Schmassmann and A. Gautschi Highway Pollution edited by R.S. Hamilton and R.M. Harrison Freight Transport and the Environment edited by M. Kroon, R. Smit and J. van Ham Acidification Research in The Netherlands edited by G.J. Heij and T. Schneider Handbook of Radioactive Contamination and Decontamination by J. Severa and J. Bar Waste Materials in Construction edited b y J.J.J.M. Goumans, H.A. van der Sloot andTh.G. Aalbers Statistical Methods in Water Resources by D.R. Helsel and R.M. Hirsch Acidification Research: Evaluation and Policy Applications edited by T. Schneider Biotechniquesfor Air Pollution Abatement and Odour Control Policies edited byA.J. Dragt and J. van Ham Environmental Science Theory. Concepts and Methods i n a One-World, Problem-Oriented Paradigm by W.T. de Groot Chemistry and Biology of Water, Air and Soil. Environmental Aspects edited by J. Tolgyessy The Removal of Nitrogen Compounds from Wastewater by B. Halling-S~rensenand S.E. J ~ r g e n s e n

vii

P R E F A C E

It is becoming increasingly necessary that we face the problems ofglobal environmental contamination, taking into account interactions between the different contaminants and natural environments such as the atmosphere, soil, water and sediment, as well as flora and fauna. With fauna, the threat comes from man, his survival and his unhealthy need to refuse change and evolution. For him global warming is an aggression, even though it is intrinsic to the quaternary period that drastic and repeated changes in climate occur without any human interference. Those same variations have already allowed the moderate climate of the regions in the northern hemisphere to go through alternating glacial and interglacial periods. A new sociological dimension should thus be added to the simple observation of the natural environment and its contamination by man. This new dimension has resulted in making every gesture towards the environment a highly politicized act, as the environment has become a primordial dimension of our society, almost a religion for a growing number of Westerners. Scientific research is hindercd rather than helped by these political interferences.

In this field, fundamental and aiplied research tend to become mixed up. This sometimes goes too far, as in the case of atmospheric pollution. Before scientists elucidate the real weight of human activity on global warming, a dangerous amalgam has been made, giving man responsibility for climatic evolution and the hasty implementation ofenergy taxes. Such taxes are not, as declared, incitations to energy saving but an opportunity for political circles to increase state revenue - the heavy taxes on fuel have demonstrated the failure ofsuch a policy well. The environmental problems have become necessary for the society that generates them as a source ofincome. The growing importance of ecology has attracted men hungry for power, who can redirect our system’s dynamics to the benefit oftheir own ambitions. Others have a fundamentally different vision since, for them, ecology has become a new means ofslowing down our society’sdevelopment, or even to destroy it, causing a retrograde step. Finally, in an early developmental phase, industrial circles found themselves held back by this excessive ecological intervention. Later, a “green”industry developed in western countries that satisfied the new values and contented industrials for a while. Future development will take place in countries less concerned about the, often irrational, requirements of their ecological and political circles that penalize the industry in the western countries. For the above reasons, scientists will have to leave their ivory tower and communicate with the political and economical circles if this ecological drifting is to be limited. A worldwide extension of norms is necessary if we are to have a successful world economy. It is certain,

nevertheless that some measures are wasteful in terms of finance and energy.

...

Vlll

In the light of the above, consideration must be given to scientists from different fields. Economists, chemists, geologists, biologists, pedologists and sanitary engineers can now voice their opinions.

The first field considered in this work is atmospheric pollution. This has become a major problem of our time, and could soon develop into a scandal through the overestimation of the impact of green house gases such as ozone, NOxor SO, Which allows their use as political instruments. Another growing field of importance covers the fundamental questions posed by the problem of soils and their interaction with waste from human activities, for example, sterile mining waste, sewage sludge, and contaminated dredged material. How do these different elements react and how tolerant can our system be towards them? Continental aquatic systems are still matters ofprime importance, and pose such questions as what part do sediments play in the evolution of eutrophication and how do bacteria affect drinking water or bathing? A volume such as this, containing a selection of the most important work presented at the 5th

International Conference on Environmental Contamination, reflects some aspects of the present state ofresearch and our society’s dominant preoccupations. It can be noted that atmospheric pollution and the use by agriculture of wastes such as sewage sludge and contaminated dredged material, provide a major concern. We cannot conclude, however, without seizing this opportunity of synthesizing our knowledge on elements as important as mercury or selenium. Through their efforts and expertise, the different authors have contributed to the success of this book. The Editor would like to thank all the authors for their important contributions and Elsevier Science Publishers, Amsterdam, for their most appreciated assistance in bringing this to fruition.

Jean-Pierre Vernet

1 July 1993

F.-A. Fore1 Institute University ofGeneva Switzerland

IX

CONTENTS Preface

vii

List of Contributors

xiii

Chapter 1

Aspects of environmental contamination Ethical aspects ofenvironmental protection. B. B ~ G E N M E I E R .

Chapter 2

1

Atmospheric pollution Effects of air pollutants on man, animals, plants and buildings: mechanisms and dose-response effects.

S. HIPPELI and

E.F.

ELSTNER.

13

Effects ofair pollution on the condition of sessiIe oak forests in Hungary.

I. Mf?SZAROS, I. M6DY and M. MARSCHALL.

23

Long term effects of acid deposition: Implications on the performance

of high level nuclear waste repositories. J. NEBOT and J. BRUNO.

35

On character of ejection of radionuclides out of earth surface. I.V. MELIKHOV, 2. VUKOVIC and V. SIPKA. Chapter 3

65

Soils and contamination from mining uses Trace elements dynamics in soils and aquifers of western Switzerland.

0.ATEM, J.C. VEDY and A. PARRIAUX.

19

Leaching behaviour of granulated nonferrous metal slags. F.M.G. TACK, P.H. MASSCHELEYN and M.G. VERLOO.

103

Environmental impact of mining activities on the Hermioni area, Greece.

S.P. VARNAVAS,

KRITSOTAKIS.

A.G. PANAGOS

and

K.G.

119

x

Chapter4

Human impacts on soils by wastes and contaminated dredged material uses Beneficial and toxic effects of chromium in plants: solution culture, pot and field studies. J. BARCEL6, Ch. POSCHENRIEDER, M.D. V k Q v E z and B. GUNSE.

147

Interactive effects of the application of different Cd forms and an acidibing agent on plant available metals and postharvest soil extractability. R. NOGALES, D. HERVLS, J. SOT0 and F. GALLARDO-LARA.

173

Evolution of heavy metal species in leachates and in the solid phase during composting of municipal solid wastes. P. PRUDENT, C. MASSIANI and 0. THOMAS.

187

Effect ofseveral industrial wastes on soil respiratory activity. E. MARTI, R. CRUmAS, M.A. GARAU, E. DE MIGUEL and M.T. FELIP~.

217

Long-term evaluation of plants and animals colonizing contaminated dredged material placed in upland and wetland environments. D.L. BRANDON, C.R. LEE, J.W. SIMMERS, J.G. SKOGERBOE and G.S. WILHELM.

23 1

Contaminated aquatic sediments and waste sites as toxic chemical time bombs. U. FORSTNER.

259

The importance of biological testing in the assessment of metal contamination and site remediation. C.R. LEE, J.W. SIMMERS, D.L. BRANDON, L.J. ONEIL, M.J. CULLINANE and J.M. ROBERTSON. Chapter 5

293

Inland waters Recovery from eutrophic to oligotrophic states in lakes: role of sediments. D. SPAN, V. COPPEE, J. DOMINIK, G. BALVAY, F. BERTHIER, C. MARTIN and J.-P. VERNET.

303

xi

Dynamics of the autochthonous and contaminant bacterial colonization of lakes (lake of Cadagno and lake of Lugano as model systems). R. PEDUZZI, A. DEMARTA and M. TONOLLLZ.

323

The role of the bacterial community in the radionuclide transfers in freshwater ecosystems. F. HAMBUCKERS-BENIN, A. HAMBUCKERS and J. REMACLE.

337

Effects ofplants on the accumulation of Zn, Pb, Cu and Cd in sediments of the Tagus estuary salt marshes, Portugal. I. CAGADOR, C. VALE and F. CATARINO.

355

Studies on heavy metals of periphyton and its host plant / Phragmites australis (Cav.) Trien ex Steudel / in shallow lakes. G. LAKATOS.

365

Origin and pathways of Cadmium contamination in the Gironde Y. estuary, Garonne river and tributaries. J.M. JOUA"EAU, LAPAQUELLERlE and

Chapter 6

C.LATOUCHE.

373

Synthesisand methods Mercury pollution and cycling in aquatic systems. F.M. D I W .

391

Analysis of Selenium. M. SAGER.

403

Selenium occurence and ecology. M.SAGER.

459

AUTIIOR INDEX

475

SUBJECT INDEX

477

This Page Intentionally Left Blank

...

XI11

LIST OF CONTRIBUTORS

AlTEIA 0.

ATE-Pedology, Swiss Federal Institute ofTechnology (EPF), Lausanne, Switzerland.

BALVAY G.

Institut National de la Recherche Agronomique, Thonon lcs Bains, France.

BARCELO J.

Laboratorio de Fisiologia Vegetal, Facultad de Ciencias, Universidad Aut6noma de Barcelona, Bellatema, Spain.

BERTHIER F.

Syndicat Intercommunal do Lac d'Annecy (SILA), CranGevrier, France.

BRANDON D. L.

Environmental Laboratory, US Army Engineer Waterways Experiment Station, Vicksburg, Mississippi, U.S.A.

BRUNO J.

MBT Tecnologia Ambiental, Parc Tecnolbgic del Valles, Cerdanyola, Spain.

BURGENMEIER J.

Department of Economics, University ofGeneva, Geneva, Switzerland.

CACADOR I.

Departamento de Biologia Vegetal, Universidade dc Lisboa, Lisboa, Portugal.

CATARINO F.

Departamento de Biologia Vegetal, Universidade de Lisboa, Lisboa, Portugal.

COPPEE V.

Institute F.-A. Forel, University ofGeneva, Versoix, Switzerland.

CRURAS R.

Laboratori d'Edafologia, Facultat de Farmacia, Universitat de Barcelona, Barcelona, Spain.

CULLINANE M. J.

Environmental Laboratory, US Army Engineer Waterways Experiment Station, Vicksburg, Mississippi, U.S.A.

DEMARTA A.

Laboratory ofmicrobial Ecology, University ofGeneva, and Cantonal Institute of Bacteriology, Lugano, Switzerland.

D'ITRI F. M.

Institute of Water Research, Michigan State University, East Lansing, Michigan, U.S.A.

DOMXNXK J.

Institute F.-A. Forel, University ofGeneva, Versoix, Switzerland.

ELSTNER E. F.

Institut fur Botanik und Mikrobiologie, Biochemisches Labor, Technische Universitat Miinchen, Miinchen, Germany.

FELIPO M.T.

Laboratori dEdafologia, Facultat de Farmlcia, Universitat de Barcelona, Barcelona, Spain.

FURSTNER u.

Technische Universitat Hamburg-IIarburg, Arbeitsbereich Umweltschutztechnik, Hamburg, Germany.

xiv

GALLARDO-LARA F.

Estacion Experimental del Zaidin, Granada, Spain.

GARAU M.A.

Laboratori d’Eddologia, Facultat de Farmacia, Universitat de Barcelona, Barcelona, Spain.

GUNSE B.

Laboratorio de Fisiologia Vegetal, Facultad de Ciencias, Universidad Autonoma de Barcelona, Bellaterra, Spain.

HAMBUCKERS A.

Microbial Ecology, Department of Botany, University of Liege, Liege, Belgium.

HAMBUCKERS-BERHIN F.

Microbial Ecology, Department of Botany, University of Litge, Liege, Belgium.

H E R V ~ SD.

Estacion Experimental del Zaidin, Granada, Spain.

HIPPELI S.

Institut fir Botanik und Mikrobiologie, Biochemisches Labor, Technische Universitat Miinchen, Miinchen, Germany.

JOUANNEAU J. M.

DCpartement de GCologie et OcCanographie, UniversitC de Bordeaux I, Talence, France.

KRITSOTAKIS K. G.

Department of Mineralogy,University of Mainz, Mainz, Germany.

LAKATOS G.

Department of Ecology, Lajos Kossuth University, Debrecen, Hungary.

LAPPAQUELLERIE Y.

Dkpartement de Gkologie et OcCanographie,UniversitC de Bordeaux I, Talence, France.

LATOUCHE C.

DCpartement de GCologie et OcCanographie,UniversitC de Bordeaux I, Talence, France.

LEE C.R.

Environmental Laboratory, US Army Engineer Waterways Experiment Station, Vicksburg, Mississippi, U.S.A.

MARSCHALL M.

Botanical Department, Lajos Kossuth University, Debrecen, Hungary.

MARTI E.

Laboratori dEdafologia, Facultat de Farmacia, Universitat de Barcelona, Barcelona, Spain.

MARTIN C.

Institute F.-A. Forel, University ofGeneva, Versoix, Switzerland

MASSCHELEYN P. H.

Laboratory for Analytical Chemistry and Agrochemistry, University ofGent, Gent, Belgium.

MASSIANI C.

UniversitC de Provence, Laboratoire de Chimie et Environnement, Marseille, France.

MELIKIIOV I. V

Lomonosov State University, Moskow, Russia.

M k Z h O S I.

Botanical Department, Lajos Kossuth University, Debrecen, Hungary.

DEMIGUEL E.

Laboratori dEdafologia, Facultat de Farmacia, Universitat de Barcelona, Barcelona, Spain.

xv

MODY I.

Botanical Department, Lajos Kossuth University, Debrecen, Hungary.

NEBOT J.

MBT Tecnologia Ambiental, Parc Tecnolbgic del Vallts, Cerdanyola, Spain.

NOGALES R.

Estacion Experimental del Zaidin, Granada, Spain.

O’NEIL L. J.

Environmental Laboratory, US A m y Engineer Waterways Experiment Station, Vicksburg, Mississippi, U.S.A.

PANAGOS A. G.

Department ofApplied Geology, Technical University of Athens, Athens, Greece.

PARRIAUX A.

GEOLEP, Swiss Federal Institute ofTechnology (EPF), Lausanne, Suisse.

PEDUZZI R.

Laboratory of microbial Ecology, University ofGeneva, and Cantonal Institute of Bacteriology, Lugano, Switzerland.

POSCHENRIEDER Ch.

Laboratorio de Fisiologia Vegetal, Facultad de Ciencias, Universidad Autonoma de Barcelona, Bellaterra, Spain.

PRUDENT P.

Universite de Provence, Laboratoire de Chimie et Environnement, Marseille, France.

REMACLE J.

Microbial Ecology, Department ofBotany, University of Litge, Liege, Belgium.

ROBERTSON J. M.

Ware & Freidenrich, Attorneys at Law, Palo Alto, California, U.S.A.

SAGER M.

Landwirtschaftlich-Chemische Bundesanstalt Wien, Wien, Austria.

SIMMERS J. W.

Environmental Laboratory, US Army Engineer Waterways Experiment Station, Vicksburg, Mississippi, U.S.A.

v.

Institute of Nuclear Sciences VinEa, Belgrade, Yugoslavia.

SIPKA

SKOGERBOE J. G .

Environmental Laboratory, US Army Engineer Waterways Experiment Station, Vicksburg, Mississippi, U.S.A.

SOT0 J.

Estacion Experimental del Zaidin, Granada, Spain.

SPAN D.

Institute F.-A. Forel, University ofGeneva, Versoix, Switzerland.

TACK F. M. G.

Laboratory for Analytical Chemistry and Agrochemistry, University ofGent, Gent, Belgium.

THOMAS 0.

Universite de Provence, Laboratoire de Chimie et Environnement, Marseille, France.

TONOLLA M.

Laboratory of microbial Ecology, University of Geneva, and Cantonal Institute ofBacteriology, Lugano, Switzerland.

VALE C.

Instituto Nacional de InvestigaCIo das Pascas, Lisboa, Portugal.

xvi

VARNAVAS S. P.

Department of Geology, University of Patras, Patras, Greece.

V ~ Q U E ZM. D.

Laboratorio de Fisiologia Vegetal, Facultad de Ciencias, Universidad Autonoma de Barcelona, Bellatem, Spain.

d D Y J.-C.

IATE-Pedology,Swiss Federal Institute ofTechnology(EPF), Lausanne, Switzerland.

VERLOO M.G.

Laboratory for Analytical Chemistry and Agrochemistry, University ofGent, Gent, Belgium.

VERNET J.-P.

Institute F.-A. Forel, University ofGeneva, Versoix, Switzerland.

WKOVIC

2.

WILHELM G . S.

Institute of Nuclear Sciences Vin&, Belgrade, Yugoslavia. The Morton Arboretum, Lisle, Illinois, U.S.A.

CHAPTER 1 Aspects of environmental contamination

This Page Intentionally Left Blank

1

ETHICAL ASPECTS OF ENVIRONMENTAL P R O T E ~ I O N

B. Biirgenmeier' * Department of Economics, University of Geneva, 102 Bd Carl-Vogt, 1211

Geneva 11, Switzerland

INTRODUCTION I n the actual public debate about how the environment can be most effectively protected, there is increasing emphasis on economic instruments (Tietenberg, 1989). The pre-eminence of economics in social relations, is also extended to environment protection. The best results are supposedly achieved through optimum allocation of productive resources by means of the market mechanism. According to this view, scarcity can best be managed by market forces, backed u p by measures which are likewise market-oriented. This has consequently become the preferred environmental strategy i n the OECD countries (OECD, 1991a). The measures taken in order to protect the environment so far went to the opposite way. Direct control has taken over market incentives. The antagonism between norms and incentives still overshadows actual policy recommendations and takes it as granted tha t environment protection should be exclusively submitted to economic constraints. However, the need for a n ethical yardstick is clearly expressed (Etzioni, 1991). Why does this claim persist? This contribution seeks to give a n answer to this question. I t is divided into three parts. The first discusses the distinction between the positive and normative aspects of economics. The second part recalls the ethical content of the market as sole reference of the functionning of our society. Finally, the third part examines the issue of environmental protection in the light of both market and policy failures.

VALUE JUDGMENTS AND ECONOMIC SCIENCE The current school of thought claims to take ethics into account by portraying economics as both positive and normative. However, this interpretation is inadequate. The positive aspect of economics, reflecting a n approach common i n the natural sciences, is based on observable facts. I t highlights causal relationships

2

and seeks to analyse economic mechanisms on the basis of available information. However, the nature of this information, which is frequently unquantifiable and historically unique, has led to an emphasis on purely deductive methods, which form the basis for many of the conclusions in mainstream economics. This current of thought is typified by the neoclassical school, which takes as its main premise the much-disputed hypothesis of economic rationality. Human behaviour is deemed to be rational when consumers maximize their satisfaction and producers their profits, subject t o constraints such as differing incomes and production costs respectively. Optimal strategies for both consumers and producers can best be determined by the mechanism of market forces. This leads t o the important conclusion that maximum economic welfare is the result of individual strategies, and that society does not exist as a separate entity but is merely the sum total of individuals behaving rationally. The key issue here is efficiency. The normative aspect of economics involves value judgements. It is therefore based on a subjective portrayal of the facts and considers the economy as it should be, using criteria which can only be identified by the collective decision-making mechanism of political choice. Economic theory acknowledges that each economic agent is perfectly free t o express value judgements, and thus there may conceivably be as many different points of view as there are agents. If one point of view is to prevail, it can only be the majority view, identified as such by a democratic process. Here the key issue is equity, embodied in laws and regulations which represent an institutional codification of society’s moral and ethical judgements. This rough outline of the positive and normative aspects of economics provides a conceptual link between efficiency and economics on the one hand, and equity and politics on the other. Seen from this point of view, the market economy is indissociable from democracy - neither is conceivable without the other. Accordingly, strict application of the principles of the market economy and political democracy should eliminate the need for a separate ethical approach to economics (Novak, 1987); ethical issues can then be taken into account simply by sustaining both of these collective decision-making mechanisms. Despite this essential conclusion, which is central t o the functioning of Western industrial society, recent years have seen a rapidly growing need to introduce ethical considerations into economics. This is because both the boundary between the normative and positive aspects of economics and the distinction between efficiency and equity are unclear, for the following reasons (Biirgenmeier, 1992): - The distinction between the normative and positive aspects of economics and their linking with politics and economics respectively are, of course, gross oversimplifications. Political logic displays certain rational traits, and value judgements play an undoubted part in economic behaviour. A great deal of rhetoric - some would say ideological bias - has gone into presenting positivism as the only rational standard for human economic behaviour. Not only is the

3

definition of economic rationality a tautology, but human behaviour and institutions quite definitely influence one another. The definition of economic behaviour is tautological in that a deductive approach is used to obtain results compatible with the economic rationality on which much of economic policy is based. This is not to say that the hypothesis of rational economic behaviour is wrong, or to deny that it has profoundly influenced our understanding of how society functions. However, it does mean that economic models based on this hypothesis are bound to treat social issues as mere problems of constrained optimization. Not only is this likely to result in circular reasoning, but inherent in the market model are value judgements which are not separate from the economic sphere but specific to it. We can then no longer relegate the normative aspects of economics to the political sphere and treat economics as a scientific discipline in the pure positivist tradition. Were ethcs nothing more than the expression of value judgements, there would still be a place for it in economics. As for the influence of economic behaviour on institutions and vice versa, it must be acknowledged that institutional criticism of the market model has not succeeded in diverting the mainstream of economic thought. Mainstream economics is eager t o prove the existence of economic laws which are independent of place and time. In contrast, a n apparent common thread in the institutional criticism, and the basis for socio-political approaches to the economy, is the idea that economics is culturally determined and cannot be viewed in isolation from Western cultural history. Once again, were ethics nothing more than a cultural phenomenon, there would still be a place for it in economics. - The notion of the market as an efficient allocator of factors of production is tending to be extended to areas other than economics. Ceaseless intellectual effort has gone into attempts to apply economic reasoning t o such fields as law, politics, sociology and medicine, on the principle that efficiency must be achieved before there can be any question of equity. Such attempts suggest a wish to see production constraints take priority over all other social, o r indeed ethical, considerations.

THE ETHICAL REFERENCE OF THE MARKET The market is merely one of a number of collective decision-making mechanisms. The others include the democratic process and the civil service. Comparing these mechanisms, one can observe a tendency for decisions to be reached not by market forces or by the democratic process, but by private or public bureaucracies. This shift towards a more administrative form of collective

4

decision-making has caused a change of attitude. Not only are there new groups of voters, but society’s image of the economy has changed, transferring the emphasis from efficiency to equity. The market mechanism has begun to make way for more political decision-making mechanisms in which problems of organization have supplanted problems of trade. This development is made easier by the neoclassical misapprehension of the role of the State, and by the problems of redistribution of income and wealth that are inherent in the functioning of any market economy. Neoclassical theory sees growth as the answer to problems of equity, and suggests that optimal growth will be achieved by the operation of individualism in a harshly competitive market. Politics, on the other hand, proposes to deal with problems of equity by creating conditions conducive to the development of solidarity. In a society characterized by numerous market failures and a lack of solidarity, the emergence of bureaucratic mechanisms and the need for an ethical frame of reference are thus inevitable. There is thus a place in economics for an ethical component which reflects particular value judgements or a particular culture and is backed up by the various collective decision-making mechanisms society has established. All that is necessary is to acknowledge that economic theory is essentially normative and that it is merely one part of the social whole. Should ethics then be assigned a universal value and also be approached in positivist terms? An example of such an approach is the demonstration of the universal nature of human rights. If ethics is indeed an inherent feature of human nature, then an extra dimension needs to be added to the hypothesis of rational behaviour by economic agents. Failing this, a positivist approach t o values will exclude them from the scope of economics, relegating ethical standards t o the field of theology or political philosophy. This must be avoided a t all costs. Instead of attempting as mainstream economic theory does - t o show that economic rationality is the only predominant human behaviour that can be generalized in a manner which is independent of history, culture and institutional development, one may quite legitimately suppose that ethics is a feature which is inherent in and exclusive t o man. Such an approach would treat ethics independently of economics and would place its study within the field of philosophy. This would be the consequence of excluding all value judgements from economic theory, and would lead to the creation of a specialized branch of study for ethics. This would be in keeping with the way in which the transmission of knowledge is currently organized, namely via an increasingly impenetrable web of specialization. In order t o ensure ethics a lasting place in economic reasoning, then, the hypothesis of economic rationality should not be rejected (since the strategies consumers and producers use to maximize their utility and profits respectively reflect essential motives in human behaviour), but it should be developed further. Economic agents act emotionally as well as rationally. Emotional intensity thus adds a moral dimension t o economic and social actions and is a t variance with scientific interpretations based on the principle of causality,

5

which reduce emotionality to the level of what has been termed "hallucinatory reality". To quote Claude Olievenstein (1988, p. 191): "What is unspoken, particularly with regard to morality or ethics, cannot be written off as hallucinatory - human memory is too powerful for that. We thus have the astonishing situation in which part of reality is kept, as it were, under cover. Things which exist but cannot be fitted into a logical interpretation are either censored or, at best, contemptuously relegated to the nether world of the social sciences...". Attempting to rethink contemporary economic theory along these lines doubtless involves a n unprecedented intellectual challenge; yet such a n approach may well be the only way to help economics out of the positivist impasse it has got into. The methodological obstacles are enormous. Economics has only managed to bypass the normative issues by adopting a n abstract conceptual approach to society, based on deductive reasoning. For a complete view of social reality, economics must open itself up t o other disciplines - yet such a n interdisciplinary approach is inevitably more inductive. Despite the fact th at all problems are essentially interdisciplinary, knowledge has hitherto progressed through specialization. Our education system can scarcely be said to have encouraged diversity, and it is worth recalling Plato's story (Blumenberg, 1987) in which the philosopher Thales fell down a well while gazing a t the heavens. A Thracian servant girl made fun of him for having his head in the clouds and failing to see things th a t were right under his nose (and feet). Although nowadays Thales is famous and the servant girl is not, both of them had a point. An ethical approach to economic problems necessitates a n inductive review of contemporary economic theory in the light of knowledge acquired in other disciplines. Such is the challenge posed by contemporary problems, and in particular by environmental pollution, th at there is really no other choice.

PROTECTION OF THE ENVIRONMENT AND THE FAILURE OF COLLECTIVE DECISION-MAKING MECHANISMS Since the market is merely one of a number of collective decision-making mechanisms, environmental policy must be seen on various levels, including not only the economic, but also the political and administrative spheres. Accordingly, i t is prone to the failures observed in each of these spheres, and must come t o terms with the various collective decision-making mechanisms, which may supplement, oppose or reinforce one another. Thus the implementation of a n environment policy is a f a r more normative process than economic theories on the internalization of external costs are prepared to admit.

6

The following table sets out the various defects inherent in each of the collective decision-making mechanisms. The economic sphere is represented by the market and the political sphere by democracy, while the State is represented purely by its administrative functions. Yet the functions of the state are more complex than this, and impinge on the environment in numerous ways. We need only recall that the State is not only itself a producer and consumer of goods and services, but also intervenes in environmental economics as the supervisor of all manner of constraints. It thus guides individual choices, selects alternatives and regulates - a t times repressively - the social costs associated with environmental damage. It is therefore an oversimplification to reduce the State to its purely administrative dimension; however, this happens to be the discretionary space used for all collective decisions which the other mechanisms have failed to reach, and the table clearly shows that, when faced with the problem of implementing environmental policies, society must reconcile the failures that occur in the various collective decision-making mechanisms. These failures have some common features and therefore lend themselves to a tabular presentation, which will be discussed from an ethical point of view.

The main failures observed in the various collective decision-making mechanisms

Market

Democracy

Civil service and legal authorities

Redistribution of income and wealth

Intensity of preferences is not expressed

Increase o f public spending

1 Imperfect competition

Results may be logically inconsistent (Condorcet's paradox)

4

Public goods

Increase in discretionary space available to civil service

5 No incentive to keep oneself informed

7

3

2

6

Civil service seen as a public good 9

This table indicates, column by column, the main failures observed in the various collective decision-making mechanisms (Frey, 19811, but expressly omits collective decisions made by special interest groups such as umbrella organizations and occupational associations or trade unions. Since the latter operate simultaneously in all three of the spheres mentioned, it is not felt necessary to assign them a separate collective decision-making process. If the table is read horizontally, it is possible - admittedly, with some effort - to discern links between the various collective decision-making mechanisms. The first objection to the market economy (1)was so considerable that it encouraged the growth of an alternative economic system based on central planning. This first market failure thus not only profoundly shaped the history of social emancipation in the countries which espoused the market economy, but was also to result in unprecedented disillusionment in the countries which adopted the central-planning model. Accordingly, no workable environmental policy can be conceived of outside this frame of reference. However, the implementation of any such policy in our society remains subject - even though such matters lie within the "unspoken" realm (Olievenstein, 1988) - to the issue of redistribution of income and wealth, a problem which market mechanisms fail to deal with. Variations in market prices and in relative prices between market sectors (between "non-polluting" and environmentally harmful goods) lead to instances of market exclusion which are scarcely acceptable from the point of view of equity. Mainstream economists (Arrow, 1977) readily admit this, but insist on maintaining a clear distinction between positive and normative reasoning. They claim that economics is there to analyse how the market works, and that it is up to politics to settle matters of social justice. However, if we turn t o the collective decision-making mechanism known as democracy, we are forced t o admit that this is no answer (2). The democratic principle of "one man, one vote" does not allow intensity of preferences to be expressed. There is little point in calling for internalization of external effects if, in budgetary terms, one is not prepared to accept the resulting price rises. This leads to the pursuit of alliances and agreements which enable such objectives to be achieved: in other words, environmental protection for oneself, but at other people's expense. The consequences of this paradox are not only that a minority succeeds in imposing its will by democratic means, but that problems of environmental pollution are often settled by appealing to a third collective decision-making mechanism, namely the civil service (3). Economists who seek to avoid value judgements by appealing to the democratic mechanism thus help to strengthen the role of the civil service, whose measures (such as technical regulations) supposedly run counter to market forces (Baumol and Oates, 1979). In order to soften the blow, public (State) expenditure tends to increase (4). Accordingly, the State's share in changes in GNP is following the same trend (OECD, 1991b), suggesting - perhaps fallaciously - a n increase in the non-commercial sector of the economy. The top line of the table makes it clear that a positivist approach t o the economy cannot be separated from a normative approach expressed in political

8

terms. Such separation entails costs which, in turn, have economic consequences. The increase in public spending can thus be seen as a logical consequence of the separation between the economic and political spheres. Let us assume for the time being that economic models of how the economy runs represent an ideal (like any model). In that case, Todorov's comment is surely relevant (1991, p. 109): "What is ideal is only effective if it remains in touch with what is real; however, this does not mean that it should be dragged down to an "accessible" level, but rather that it should not become separated from the realm ofinquiry. Instead of neutral scientists and technicians on the one hand and moralists who disregard human reality on the other, we would then have researchers who are aware of the ethical dimension of their research, and men of action who keep abreast of achievements in the field of inquiry." This conclusion - that only acknowledgement of the ethical dimension can reconcile the positive and normative aspects of policies in the field of environmental economics - is further reinforced by the second line of the above table. A market economy modified by environmental incentives can only function if a number of hypotheses are satisfied; one such hypothesis is that there is a large number of market agents, without which competition cannot be guaranteed as an effective allocating mechanism (4). In practice, the frequent occurrence of imperfect competition is evidence of a second market failure. This lack of competition means that environmental incentives remain ineffective. Not only does the competitive oligopoly model produce a 'stand-off on prices which thus - within certain limits - remain unresponsive to incentives of this kind, but duopolies of the Nash-Cournot type also lead to situations which are less than optimal. Each situation which justifies technical regulations to protect the environment then becomes a situation in which strategic games ( 5 ) are acted out in the democratic sphere, thus helping t d increase the discretionary space available to the civil service departments (6) whose duty it is t o implement the regulations and monitor their effectiveness. Finally, the third line of the table reminds us that the economic sphere not only displays price flexibility, but also includes a significant non-commercial sector. According to economic theGry, this sector comprises public goods, whose prices cannot be determined by market forces (7). The environment is part of this, although it has an intrinsic value and thus enjoys a special status. If we choose to consider it external to the market, we return to what has become a classic pattern: the idea that the market and the environment on the one hand, and the economy and the State on the other, are diametrically opposed. In that case, the environment only acquires an economic value once it is internalized into a market. This process necessitates monetarizing the environment so that it can be analysed by a method which has been known t o us for nearly 40 years, namely the internalization of external effects. Thus the environment undergoes the same mental portrayal as the whole of society: the predominance of economic considerations over all others, even ethical ones.

9

The notion of the environment as a public good has one further consequence: a dissociation between collective costs and individual benefits. This is used to justify State intervention. Since it is in no-one’s interest to bear the external costs, the State must do so in order to remedy this third market failure. This lack of individual incentive is again encountered in the democratic collective decision-making mechanism. Since the aim of voting is to reflect the main underswells in society, a single vote in isolation is meaningless. Accordingly, there is no obligation on individual voters to keep themselves informed (8). If voters do not believe their votes will be decisive, they will be likely to vote without bothering t o find out exactly what they are voting on. As a result, collective preferences regarding environmental pollution will not be correctly expressed. This has two major consequences. Firstly, political propaganda attempts t o take advantage of the lack ofinformation. Transparency - a vital condition for any collective decision-making mechanism - is limited. Add to this a considerable lack of certainty in o u r scientific knowledge regarding complex environmental processes, and the result is inevitably that voters will be seriously ill-informed. In the absence of incentives to verify the information disseminated concerning the environment, the results of democratic voting procedures may well be biased. The second consequence concerns the very nature of social organization. If economic rationality is bounded by lack of information, markets cannot function optimally. Finding their capability limited, agents then seek to evade the mechanism of competition and t o protect themselves by means of alliances among the various special interest groups. Individual interests are no longer expressed through the market, but through the highly organized network which society has become. The administrative collective decision-making mechanism takes advantage of this lack of transparency to re-justify the increase in its own discretionary power. Yet, in turn, it is subject to the same influences. The civil service, being a public service (91, possesses the very characteristics which previously led to market failure. As a result, it too is influenced by the interplay of alliances and clashes between the various pressure groups, and thus becomes a n integral part of the organizational network of society. In the field of environmental policy, the civil service therefore tends, by virtue of the regulations which it issues, to relinquish its independent, symbolic function as the embodiment of the public interest and, instead, to reflect the failures that have affected the other collective decision-making mechanisms. Once we see the problem of policy in these terms, therefore, consensus will be more costly and governments tend to gratify established pressure groups. However, established alliances in matters of economic policy must increasingly acknowledge the presence of new pressure groups which as yet have no institutional basis. The established alliances have acted as a channel for public opinion (which governments do not always perceive correctly) and have thereby helped the latter in determining their policy. A t the same time, governments have relied on such alliances to spread information, especially in

10

order to popularize policies. The emergence of new and often short-lived pressure groups has therefore proved most disconcerting. As long as such groups are without an institutional basis, governments lack a policy guide, and are therefore tempted to hedge their decisions with extra precautions. As a result, expert reports proliferate and refuge is sought in environmental impact statements. The end result of this refusal to discuss values is an increasingly pettifogging, bureaucratic approach to environmental problems.

CONCLUSION Social sciences which study the decision-making process show the importance of values in the implementation of effective environmental policies. The underlying reference becames crucial. The value system on which public policy is based is bound to change. In order t o legitimate any public action in order to protect the environment, the ethical references will change too. This raises the issue of what society's perception of the state actually is. The need to protect the environment emphasizes the profound distinction between general and individual interest and between the public and private spheres. Marking out the boundaries between these spheres remains the essential task of politics and as to be exposed to the criterion of transparency. But communication and information systems which are supposed to keep transparent the boudaries between private and public interests are not a public but an economic good: Information is produced according t o economic principles. It is therefore not astonishing that environmental protection has to be submitted to the same logic too. In this perspective, ethics is only seen as the expression of moral philosophy. As a matter of fact, ethics has to become integrated part of ecological economics.

REFERENCES Arrow, K.J. (1977), "The Organization of Economic Activity, Issues pertinent to the Choice of Market versus non-market Allocation",in Public Expenditure and Policy Analysis, R.H. Haveman and J. Margolis, eds., Rand MacNally College Publishing Company, Chicago. Baumol, W.J. and Oates, W.E. (1979),Economics, Environmental Policy and the Quality of Life, Prentice Hall, Inc., Englewood Cliffs, N.J. Blumenberg, H. (1987), Das Lachen der Thrakerin. Eine Urgeschichte der Theorie, Suhrkamp, Frankfurt am Main.

Biirgenmeier, B. (1992), Socio-Economics: an Interdisciplinarv Approach: Ethics, Institutions and Markets, Kluwer Academic Publishers, Boston, Dordrecht, London. Etzioni, A., Lawrence, P.R. (19911, Socio-Economics: Towards a New Synthesis, M.E. Sharpe, Inc., h o n k , New York, London. Frey, B. (1981), Theorie demokratischer Wirtschaftspolitik, Franz Vahlen Verlag, Munich. Novak, M. (1987), Une Qthique 6conomique, Les valeurs de 1'Cconomie de marchb, Les Qditions du Cerf, Institut La Bobtie, Paris (translated from the English "The Spirit of Democratic Capitalism", 1982). OECD (1991a), "Recent Developments in the Use of Economic Instruments for Environmental Protection in OECD Countries", OECD Environment Monomaph, No. 11. OECD (1991b), OECD Economic Studies, International Comparisons, Paris. Olievenstein, C. (1988), Le non-dit des Qmotions, Editions Odile Jacob, Paris. Thomas Jr., W.L. e t al. (1956), Man's Role in Chankng the Face of the Earth, University of Chicago Press, Chicago. Tietenberg, T.H. (1989), "Marketable Permits in the U.S.: A Decade of Experience", in Karl W. Roskamp, ed., Public Finance and the Performance of Enterprises, Wayne State University Press, Detroit. Todorov, T. (1991), Les morales de l'histoire, Bernard Grasset, Paris.

This Page Intentionally Left Blank

CHAPTER 2 Atmospheric pollution

This Page Intentionally Left Blank

13

Effects of air pollutants on man, animals, plants and buildings: mechanisms and doseresponse effects S . Hippeli and E.F.Elstner

Institut fur Botanik und Mikrobiologie, Biochemisches Labor, Technische Universiat Miinchen, Arcisstr. 21, 8000 Miinchen 2, Germany

ABSTRACT Air pollution has become a major public concern since the beginning of industrialization, including motor car exhaust since the past three decades. Besides direct effects on living organisms, effects on buildings as well as on climate have to be considered. For living organisms, SO NO2, ozone, certain hydrocarbons and particles may have toxic effects whileqor buildings acids and their anhydrides (SO , N02) exhibit destructive power. Epidemiological studies together with clinical triafs and experiments in exposition chambers, including biochemical model reactions, contribute to our knowledge about potential dangers and increase the understanding of corresponding mechanisms and dose-response effects. Comparism of the almost daily appearing threatening reports in the press with the digest of more than 800 relevant original scientific publications allows the statement that impacts of ozone and NOx on health and performance of plants and animals are widely overestimated and appear to be used as political instruments. In contrast, combinations of SO2 with soot particles may represent an underestimated toxic potential. New experimental results clearly demonstrate the generally detoxifying power of the 3-way catalytic converter of Otto motor engines.

INTRODUCTION Air pollution by industry, traffic and agriculture is a growing threat all over the world. In the last few years especially automobil exhaust has moved into the focus of public and political discussions. Besides SO2 from coal-burning power plants, NO and hydrocarbons from gasoline-driven engines and petrol industry, together with their secondary products such as ozone and different peroxides have been identified as responsible for adverse effects on man, plants and buildings. Airborne soot particles stemming both from industrial incinerators and diesel engines envisage growing attention as potential carcinogens and inducers of respiratory diseases. The combinations of SO2 and soot from coal burning have been epidmiologically shown to be responsible for thousands of deaths during several severe smog episodes both in London and in New York not too long ago. In this review a digest of epidemiological data, clinical studies, measurements in exposition chambers as well as ex vivo or in vitro experiments on the effects of trace gases and particulate matter on animals, plants and buildings is reported. More than 800 original reports, revies and official statements have been incorporated into this documentation [I] thus allowing certain conclusions as to the relevance of continuously appearing reports on the increasing danger of air pollution.

14

EFFECTS OF AIR POLLUTANTS

Dose responses and reaction mechanisms Most trace gases or particles in the air are not directly toxic per se in ambient concentrations. Toxic effects on living cells are in most cases obtained indirectly via the formation of peroxides and other reactive oxygen species (ROS) [2, 31. This holds for N02, ozone, particles and partially also for S02. Since peroxides and other ROS are strongly under metabolic control in the living aerobic cell, destructive reactions are only observed after regulatory processes or reconstructive potentials have become limiting or inactivated; these complex processes connecting oxidative damage to biomolecules with cellular antioxidant properties and repair processes have recently been reviewed [4-61. General biochemical reactivities of potentially toxic trace gases are summarized in table 1 .

Table 1 Biochemical reactions of potentially toxic trace gases (nogtr = not of general toxicological relevance) ~

~~

~~~

~

binds to heme groups thus reversibly blocking respiration; only a problem in high concentrations; ("garage effect", suicides); nogtr acts as free radical; binds to heme groups; is rapidly oxidized to NOg; is physiologically "known" as a regula tor of blood vessel tonus; nogtr acts as free radica1,forming nitro compounds and peroxides; precursor of strong acids; is physiologically "known" but of potential toxicity strong oxidant forming peroxides; splits double bonds and oxidizes amino acids in proteins; toxic; (LA-type smog indicator) precursor of strong acids;forme free radicals and rapidly activates peroxides producing extremely reactive alkoxyl- or hydroxyl- radicals; (London-type smog indicator) contain aromatic hydrocarbons, nitroaromats, quinones and chelated transition metals thus initiating toxic redox cycling; catalyse the autoxidation of SH-groups and aecorbate; oxidize SO2 and NO? yielding sulfuric and nitric acids; (London-type smog indicator).

15

Biochemical effects may not always directly reflect physiological impacts since complex metabolic interactions may mask primary molecular events and thus interfere with the expected straight forward symptom expression. In concentrations up to the highest daily means, all the above mentioned gases are metabolized by several general mechanisms or constitutive pathways. Detoxification pathways in plants and animals may follow different routes. Some physiological effects of important air pollutants are listed in table 2.

Table 2 Physiological and ecological effects of relevant air pollutants effects on respiratory parameters. acts as plant fert i 1 izer ; 03

influences membrane equilibria (transport blocker); effects on respiratory parameters; may reduce resistance against viral infections; may reduce plant growth;

so2

irritates eyes and airways; nectrotizes plant tissues; monocausal agent of forrest die-back ( "Waldsterben") ;

soot

mutagenic,carcinogenic and allergizing;

par-

ticles

+ acute toxicity at high concentrations; soot provoke leukocytes (alveolar macrophages) but parimpair their bactericidal properties; potential ticlss inducer of chronic respiratory dieseases; SO2

Clinical and epidemiological studies and field observations Except epidmiological studies and field observations, clinical studies or chamber- and model-reactions have mostly been performed with overdoses of pollutants or doses in the upper region of occasionally observable gas- or particle-concentrations. It is thus not surprising that toxic effects or metabolic limitations (growth, development, performance) have been reported. Data from clinical and epidemiological studies under relevant conditions clearly indicate that a) gaseous components such as NO2 or ozone have no acute or chronic effects besides provoking regulatory systems (respiratory parameters: decrease of air flow in the range of 4-7%) or growth retardations (up to 10%)in some sensitive plants; b) air borne particles, especially those from diesel engine driven cars (trucks), in cooperation with S 0 2 , may most probably be involved in induction of sensitivities towards certain allergens (pollen, house dust mites) and also in the asthmatic complex;

16

c) acid anhydrides such as SO2 and NO2 in cooperation with soot particles as

carriers and catalysts are assumed to be mainly responsible for stone weathering and thus the decay of buildings. Tables 3 and 4 show some dose-response-effectsrelations for NO2 and ozone, as the key trace gases for photodynamic smog situations. Table 3 Dose-response data for NO2 Concentrations up to 850 pg/m 3 (1/2 h means) may be measured in urban areas Effects on no effects in healthy persons in concentrations up to 1900 pg/m3; asthmatics, especially under man and an ima1s strain, may recognize respiratory restrictions under 950 pg/m3; increase of respiratory infections at concentrations over 5000 pg/m3 were reported for animals; no indication for lung cancer after chronic exposition Effects on no negative effects under 4000 pg/m3; cooperation plants with NH4+ as fertilizer; under acid conditions soil losses of Mg++ and Ca++-ions; changes in ecological equilibria due to altered competitive behaviours, especially in N-intolerant ecotypes such as boglands

Table 4 Dose-response data for ozone Concentrations up to 450 pg/m3 (1/2 h means) may be measured close to, but outside of urban or industrial centers; Effects on sensitive persons (not identical to asthmatics!) man and may undergo reversibel respiratory restrictions anima 1s (ca. 4-8% reduction at 200-400 p g / m 3 ) ; during strain respiratory functions change already at concentrations above 160 pg/m3; adaptations after repeated expositions have been observed;Headache, coughing and throat irritations were individually reported at concentrations above 200 pg/m3; in mice, increased susceptibilities for viral infections ( > 160 pg/m 3 ) were observed; no indication for carcinogenic potentials

L

Effects on losses in crop yields and growth reductions up to plants 10% in sensitive plants ( 7 h up to 70 pg/m 3 ) were reported.

17

Motor exhaust condensates Certain components of motor car exhaust undergo condensation or dissolve as soon as they get in contact with cold or water containing surfaces. In order to compare the condensates of different motor concepts (Diesel, Otto-engine without catalytic converter (Bok) or Otto engine equipped with a regulated 3-way catalysator (Kat)) we [7] installed a condensation trap onto the exhaust tube of the vehicle on the test platform (figure 1) and followed the official FTP75 test procedure (federal test program).

Figure 1. Arrangement for trapping aqueous motor exhaust condensates during the "federal test program" (FTP75)

The water soluble condensates of this motor operation cycle, representing a mean city tour, were tested for their activities to modify biological molecules such as a. low molecular weight antioxidants (ascorbate, cysteine and glutathione) and proteins, such as BSA; b. crocin, a water soluble model substance representing polyene structures (typical for biomenbranes or poly unsaturated fatty acids) and

c. enzymes (glycerinaldehyde-3-phosphatedehydrogenase or xanthine oxidase).

Ascorbate oxidation. A strong stimulation of ascorbate oxidation, measured as oxygen uptake, is observed in the presence of BoK, while in the presence of Diesel only weak oxidation and, in the presence of KAT, no oxidation is observed . Fig. 2 shows the oxidation of ascorbate, treated with 100 p1 of the different condensates.

18

Data calculated as oxygen uptake/100 pl are compared to the same data calculated on the basis of the "FTP-cycle". (During the "FTP-cycle" the three motortypes tested emit different volumes of condensates due to their different fuel consumption (diesel) or the catalyzed water consumption (KAT). The corresponding volumes of the aqueous condensates were as follows: BoK: 647 ml/l FTP-cycle KAT: 1001 ml/l FTP-cycle Diesel: 439 ml/ 1 FTP-cycle )

0

Ia

5c

150

.*

e

2:

d

100

X u U

a

0

BoK

&KAT

0

Figure 2. Oxygen consumption during ascorbate oxidation, calculated on the basis of volume (100 pI) and on the basis of the FTP cycle.

Oxidation of sulfhydryl compounds. SH-groups play an important role in metabolic redox chemistry and generally also exhibit antioxidative properties in living cells. As corresponding model compounds we analyzed the effects of the different condensates on the SH-status of cysteine, glutathione and serum albumine. Cysteine oxidation is strongly enhanced by both BoK and Diesel whereas KAT has no influence. Glutathione oxidation is only slightly enhanced by Diesel and more strongly by BoK while SH-groups in serum albumine are exclusively oxidized by BoK. Fig. 3 shows the amount of oxidized SH-groups in cysteine calculated on the basis of the used volume (100 pl) and calculated on the basis of the "FTP-cycle". The oxidizing activity of the diesel-condensate is less pronounced under the condition of the "FTPcycle" due to the lower fuel consumption and thus condensate production.

19

-

100

I

Ir.

=:

9

=: x

X

v

v

m

P

z

y1

n

2

k

100 c. a

0

50

i%

mI

50

I v)

U N u ."

U

.-w

'0

9

0

0

0

0

Figure 3. Loss of free SH groups in cysteine.

Influence on enzymic activities. Glycerinaldehyde-3-phosphatedehydrogenase (GAP-DH) is a key enzyme in the glykolytic pathway and thus of predominating importance for intermediary metabolism. This SH-containing enzyme is sensitive towards oxidation and is rapidly inactivated by peroxide [8]. Increasing amounts of BoK or Diesel increasingly inhibit enzyme activities; KATcondensates are much less active. In comparison with GAP-DH, Xanthine oxidase (XOD) is an enzyme which is not oxygen sensitive; in contrast, it produces reactive oxygen species. Addition of BoK to the XOD test system results in a 68% inhibition of oxygen uptake after 30 min incubation; Diesel yields 20%inhibition under the same conditions while KAT is without any effect. Tab. 5 represents relative activities of the individual aqueous condensates as far as their influence on the enzymes GAP-DH and XOD is concerned. These data were calculated on the basis of comparable condensate volumes as well as on "FTP-cycle". The inhibitory activity of the BoK-condensate was set as 1 for each of the enzymes. The other activities were calculated in relation to this standard value. The KATcondensate showed the lowest, the BoK-condensate the strongest inhibitory capacity.

20

Table 5 Calculation of the relative desactivation factors of GAP-DH and xanthine oxidase in the presence of corresponding amounts of aqueous condensates from different motor concepts

QAP-DH

XOD

100 pl

FTP

100 p1

FTP

BoK

1.00

1.50

1.00

1.50

Kat

0.11

0.25

0.00

0.00

Diesel

1.31

1.31

0.29

0.29

The results clearly document the loss of destructive (oxidative) power of Otto-engine exhaust after passing the regulated 3-way catalytic converter.

Particles and cooperative effects. Soot particles have been shown to contain carcinogenic properties in certain animal models [9] epidemiological studies concerning indoor coal burning in China [lo].

0' 0

50 100 Diesel s o o t particles (vg)

Figure 4.Enhancement of chemiluminescence of butylhydro-peroxide in the presence of 0.5 m M sulfite and increasing amounts of diesel soot particles.

21

There is also growing evidence that respiratory diseases and allergic reactions may be induced and/or enhanced by particle inhalation. Model experiments have shown that diesel soot particles, enhanced by aqueous SO solutions (bisulfite), exhibit a considerable destructuive potential concerning vita? biomolecules such as SHcompounds, polyenes and certain enzymic properties [l 11. Specially the peroxide activating potential has to be addressed in this context since activated leucocytes, which excrete peroxide, exhibit a change in their physiological properties (decrease of superoxide formation and increase of phagocytosis) after soot particle contact (incubation) in the presence of bisulfite [ 121. As shown in figure 4, light emission from the decay of a peroxide (t-butyl hydroperoxide) in the presence of bisulfite is enhanced by diesel particles where the detection limit of particles is in the range of 5-10 pg/ml. The effects on biomolecules can impressively be documented by the oxidative destruction of the biological dye crocin (a water soluble carotenoid from Crocus spec.) which can be suppressed by the addition of superoxide dismutase indicating the importance of the superoxide radical anion in this bleaching reaction (fig. 5 ) .

Crocin Crocin/Sulf i t Crocin/SulFit/Diesel

soat/SOD

Crocin/Sulfit/Diesel

soot

x

0

0

5

10

15

20 25 30

Time (min)

Figure 5. Bleaching of the biological carotenoid "crocin" by 400 pg diesel soot particles and 0.5 mM sulfite: effect of superoxide dismutase (SOD)

CONCLUSIONS From our studies, partially documented in this report, we wish to draw the following conclusions:

22

1) Impacts of ozone and NOx during typical summer "smog" episodes on humans and plants are widely overestimated 2) Diesel soot particles in cooperation with SO2 immissions have to be considered as inducers of chronic respiratory diseases and/or allergic sensibilization 3) It has been documented with the aid of several independent indicator reactions that the regulated catalytic converter prevents the emission of condensable factors initiating oxidative destructions of biomolecules; the responsible oxidants may also be involved in the provocation of inflammatory processes in the respiratory tract and in the "corrosion" of plant cuticles finally yielding loss of structural resistances against certain fungal pathogens.

REFERENCES 1 Elstner EF, Hippeli S . BI Wissenschaftsverlag (in preparation). 2 Menzel DB. In: Pryor WA, ed. Free radicals in biology, vol 2. Academic, New York 1976, 181-202.

3 Mudd JB. In: Pryor WA, ed. Free radicals in biology, vol 2. Academic, New York 1976, 159-180 4 Hippeli S, Elstner EF. In: Sies H., ed. Oxidative Stress Academic, New York 1991, 3-55. 5 Elstner EF, ORwald W. Free Rad Res Comms 1991: 12/13: 795-807. 6 Elstner EF. Der Sauerstoff - Biochemie, Biologie, Medizin, BI-Wissenschaftsverlag. Mannheim, Wien Zurich 1990, 53Opp.

7 Blaurock B, Hippeli S , Metz M, Elstner EF. Arch Toxicol (in press). 8 Hyslop PA, Hinshaw DB, Halsey WA Jr,Schraufstiitter IU et al. J Biol Chem 1988: 263: 1665-1675 9 Heinrich U , Pott F, Rittinghausen S . In: Ishinishi N et al. eds. Carcinogenic and mutagenic effects of diesel engine exhaust. Elsevier, Amsterdam 1986, 441-455. 10 Mumford JL, He XZ, Chapman RS, Cao SR et al. Science 1987: 235: 217-220. 11 Hippeli S , Elstner EF. Z Naturforsch 1989: 44c: 514-523. 12 Hippeli S , Elstner EF. Free Rad Res Comms 1990: 11: 29-35.

23

Effects of air pollution on the condition of sessile oak forests in Hungary I. MCszBros, I. Mody, and M. Marschall Botanical Department, Lajos Kossuth University, Debrecen, Hungary H-4010

INTRODUCTION It is widely recognized that many European forests suffer from the new-type decline and show various signs of degradation [ 1,2,3,4]. There is considerable evidence that the phenomenon of forest decline in Europe cannot be explained by a single cause but it is considered a complex multifactoral process involving the natural and anthropogenic stresses. Industrial air pollution has been implicated as an important cause which may alter directly and indirect1 the healthy state of forests [I, 51. Beside inducing serious direct damages in the tree oliage they also result in the deterioriation of many soil processes and thereby influence the conditions of plant mineral nutrition. Air pollution can cause soil acidification and in this way affects the availability of nutrients and microbiological activity [4, 61. This may lead to a reduction in the rates of biochemical processes and a considerable decrease in the overall biological activity of the soil environment. As a consequence of the direct and indirects impacts of air pollution trees can be weakened and predisposed to injuries from naturally occurring stresses as summer drought [ 1, 4, 7, 8, 91. During the past two decades an accelerated deterioration of natural forests has also been observed in Hungary which has affected mainly the natural stands of sessile oak (Quercus petraea) on the mountainous and hilly regions [ l , 6, 10, 111. The decline of sessile oak appeared first in the North Hungarian Central Range, but from the middle of the 1980s this new dieback process has been observed in the western Transdanubian part of the country, too [ 121. The most serious decline of sessile oak stands has appeared in the industrial districts where the percentage of the decline of Quercus petraea may exceed 45% [ 11. Repeated surveys indicated that significant soil acidification took place in the Hungarian forest stands during the past 25 years [ 1, 131. This paper presents the results of studies aiming to reveal the alteration of the condition of sessile oak stands as consequence of imission loads. Three different aspects of oak forest degradation were investigated. Firstly emphasis was laid on the soil compartment and its main chemical properties and microbial activity were examined. Secondly the responses of underground vegetation layers were studied with respect to the species composition and the leaf NR activity of some dominant shrub species. The third aspect of this work was the investigation of physiological state of healthy and declining trees of Quercus petraea in forest stands with different imission loads [4, 8, 9, 141.

r

OUTLINE OF STUDY AND METHODS

Five years ago we started extensive comparative studies o f sessile oak forests in the nothern industrial region of the country [l, 111.

24

S i t e 1 (Kazincbniclka)

A

S i t e 2 (Fonagysag ) Site 3 ( S i k f A k u t )

Three forest sites were selected for the present experiments situated in the eastern part of the Hungarian Central Range (Fig. 1). Site I (Kazincbarcika) is in close vicinity of a chemical plant which emits a variety of pollutants including large amounts of nitrogen dioxide, chlorine and ammonia. The main products of this chemical plant involve different N and P fertilizers [15]. Site 2 (Fhnagysrlg) and Site 3 (Sikfiikut) are located a long way from the industry under less pollution influences. Soil and plant samplings were performed during the summer months. Soil samples were taken from the upper 10 cm layer under the litter. Soil unolyses involved the measurements of pH (in suspension with 1 M KCI), the available inorganic nitrogen forms (NH4N, NO N) and phosphorus and the ex&angable K, Mg and Ca using standard methods of soil chemistry [ 161.

Figure 1. Location of sampling areas There were also measurements on the levels of organic micro-pollutants (chlorinated pesticides and polychlorinated biphenyls) in the soil with gas chromatography after extraction by hexane. T o describe the soil biological activity at sites with different pollution degrees the microbial carbon dioxide ( C 0 2 ) production and phosphatase activity [4, 17, 18, 191 were measured. Microbial CO release was measured in fresh soil samples (at 50% WCmax) by determining the C d 2 release during 12 day-incubation at 27 ' C . During the experiment C 0 2 was trapped in 2n NaOH solution and estimated after measurement of the unneutralised alkali. Soil phosphatase activity, with some modifications, was determined at the original pH value of the soil according to KrBmer and Erdei [22]. The soil samples for phosphatase assay were air-dried, disodium phenylphosphate was used as substrate, and the liberated phenol was measured cholorimetrically. Since the underground vegetation may indicate well the levels of environmental loads [9] the herb and moss layers were also surveyed in ten 4x4 m plots in forests Site 1 and Site 2. All species appearing in each plot were recorded. While the surveys of mosses covering the bark of trees and the soil surface focused only o n the presence/absence data in case of herbs the individuals were also counted. Leczfnifrute reductuse activity was used as a possible indicator of the external nitrogen input into the forest sites [21, 22, 231. The activity of the enzyme was measured in leaves of some dominant shrub species with in vivo methods as described by Pizelle and Thiery

25

[24]. Leaf samples of approx. 0.2 g were cut into pieces and then were vxiium infiltrated for 30 min in buffer containing 0.02M K N O solved in O.1M NaK-phosphate, pH 7.5. Afterwards the samples were incubated for in the dark at 30 "C. Triton X-100 was added to the incubation medium to prevcnt i t from becoming turbid i n the presence of the acidic sulphanilamide solution used for measuring NO?. The nitrite formed by NRA was assa ed cholorirnetrically after adding 1 % (w/v) sulphanilaniide in 3 M HCI and NED (N-I-naphtyl-ethylenediamine-~liliydrochlori~le)solution to the 0.01%

fh

6,")

sample solution. For the estimation of the physiological state of sessile oak at the selected sites apparently healthy and declining trees were chosen o n the basis of the visible symtoms in their foliage. As in numerous studies the measurement of kuf pliotosyntheiic pignlerzt compositoiz was used as an assay for air pollutant effects.[25, 26, 271. Photosynthetic pigment - especially chorophyll - formation is known to be very sensitive t o almost any factor which disturbs metabolic processes, for example light intensity, temperature, mineral nutrient status, water stress and waterlogging as well as many pathovens, and last but not least air pollutants [26, 28, 29, 301. For the investigation of the pigment composition leaves were collected from the upper (shade) and lower (sun) parts of healthy and declining sessile oak trees. The extraction and determination of pigments were carried o u l following Czuhajowska and Przybylski [3I]. Concentrations of chlorophylls (a and 13) a n d carotenoids were determined in the same extract made with 100 o/o acetone. The absorbance of extracts were measured at wavelengths of 410.5, 044 and 662 nm. Care was taken i n measuring the turbidity at 750nm [32]. Concentrations of pigments were calculated by Holm's formulae [33]also used by Watts and Eley [33].

RESULTS AND DISCUSSION Chemical properties of soils The soil of each sampling site helongs to the brown forest soil type. The results of the soil chemical and biological analysis are summarized in Table 1, Figure 3, 3, 4, and 5. Regarding all the analyzed soil parameters there are suhstantial differences between the investigated areas.

Table 1 Mean values of soil chemical analysis (0-20 cm)

PH (KCV NH4-N (nig/lOOg) N03-N (mg/lOOg) P (mg/1OOg) K ( m d 1OOg) Mg (mgllOOg) Ca (md1OOg)

Site I

Site 2

Site 3

2.95 3.72

3.70 1.52 0.00 0.33 12.23 9.33 114.23

5.18 1.20 0.10 0.59 7.94 5.30 293.8 1

3.96

1.85 8.19

5.20 59.20

26

On the basis of pH values it can be stated that the degree of the soil acidity is the highest in the forest stand close to the industrial point-sourcewhich may be attributed to the acid load caused by the emitted NO and SO .The soil acidity tends to decrease with the distance from the industry (Table 1f In correlation with the low soil pH at site 1near the industry the base saturation of the soil is reduced and it is especially poor in calcium. Inorgmc nitrogen forms (Table 1) and phosphorus exist in elevated uantities in the soil of the polluted area (Site 1) that can be the result of the obviously 'gher external de osition of these elements. %he level of the total tested organic chemicals, chlorinated pesticides and PCBs (Fig. 2) indicates also external load into the soil of the forest near the industrial lant where the contamination by their total quantity amounts to 178%, compared wit the more distant site (Fig. 3)

R

K

200

/L

150-

100

50 -

total pesticides

0Slte 2

-

Pas

0

total

Oestlcides

PCBs

Slte 1

(undisturbed) (disturbed)

Figure 2. Concentrations of the main pesticides and PCBs in the soil

Fipre 3. Degree of the load of organic mcropollutants into the polluted area as expressed in the percentage of values at unpolluted site

As it has already been reported [34] among the chlorinated pesticides hexachlorcyclohexanshave been shown to accumulate most significantly in the soil of the polluted site.

Microbial activity of soils The biological activity of soil - the capacity of soil microorganisms to perform the biolo 'cal transformations of organic matter and thereby determining the soil fertility and t e processes of the nutrient cycles react very sensitively to the disturbances of the habitat [4, 171. As resented in Figure 4 and 5 the activity of soil microflora has shown differences accor&y to the sites and has reflected well the alterations in chemical properties and external oads near and further from the industry (Table 1).

I?

-

27

The microbial CO roduction, one of the frequently used measures of the soil biological activity [4, 35f h ave been shown to be of very low intensity in the industrial area, near Kazincbarcika where a considerable soil acidity was also established (Fig. 4). This might be caused by the effects of acidification since the naturally occurinp microor anisms were driven out due to the pH conditions changed [4,6,13]. Moreover it takes a lpong time while acidophilic communities can be developed and replace the role of existing species [4]. mg g-1 day-1

O

Z

i

5

g-1 h-1

1

1.5

0.1

"

1

1

Site 1 (disturbed)

"

Site 3 (undisturbed)

Figure 4. C 0 2 production of soil (0-20 cm)

Site 1 (disturbed)

Site 3 (undisturbed)

Figure 5. Soil phosphatase activity (0-20 cm)

Many reactions of soil matter transformations may be catalyzed by enzymes existing outside the microorganisms and plant root s stems, therefore soil enzymes may also indicate the biological activity of soils [ 17, 181. is generally accepted that hosphorus i s taken up by plant roots as inor anic hosphate. The large proportion of however, in many soils is organically boun , so t e rmneralization rate of its organic fractions is of great importance. Soil phosphatase lays a major role in this mineralization process [17]. The phosphatase activity of soil &ig. 5 ) has shown good positive correlation with the chan es of pH and microbial CO production of the soil at the sites investigated. It was sixfoh lower in samples from h e industrial area [19]. This together with the low insensi of microbial C02 production reflect slow decomposition processes of organic materi s in the soil of the polluted area [4,61. Lowerin of the rate of decomposition may lead to considerable accumulation of litter at pollute sites [ 11. Our work suggests that the level of microbial CO roduction and the phosphatase enzyme activity can be used as good indicators for the3eration in the biological activity of soils due to the external anthropogenicloads.

C K

s

it

$

li

Species composition of herb and moss layers In the disturbed forest site near the industry the herb layer is poor in species, it consists of only 11 species as compared to that of the control site (Site 2) where the number of the appearing species is much higher (31 species). The herb layer surveys were evaluated on the basis of the nature conservation values introduced by Simon [36]. It was stated that the native herbaceous species (values 1-6) are repressed and the

28

B

roportion of disturbance tolerant and weed s ecies (values 7-10) is higher than in the Forest Site 1 than in Site 2, further from the in ustry (Fig. 6). Similar floristic alterations in the herb layer of disturbed forest sites have been presented by MolnAr e t al. [lS]. Table 2 Species composition of in forest stands near (Site 1) and furt er (Site 2) from the emission source

U"

60

life strategy

Sites 1 2

40

20

a

Slte 1

(disturbed)

Slte 2

(undtstur bed)

Lophocolea heterophylla P + + Hypnumcupressiforme P + + Brachytheciumrutabulum P + + B. velutinum P + + Plagiothecium denticulatum P + + Amblystegium serpens P + + Pylaisia polyantha P + + P + + Atrichum undulatum Isothecium mnyurum P Plagiomnium affine Ls + Bryum flaccidum C P perennial species LS:long-life shuttle species

+ +

Figure 6. Results of herb layer surveys

Ccolonist

There have been also important changes in the individual number of species nearer the in try. The herb layer of Site 2 is rich in individuals which have amounted in total to 109 m- .How er, in Site 1in close vicinity to the industry the individual number has been only 3 m- . The resence of mosses [37] also indicates some signs of degradation in the polluted site (Tab, 2). Although there are not large differences regarding the strategic spectrum of the occurring moss species [38,39] the decrease of species richness and disappearance of some species can be observed near the industry.

9

7

Leaf NR activity There were significant differences in the leaf nitrate activity (NRA) of dominant species (Fig. 7) between the sites. In the polluted area (Site 1) the species had much higher NRA (0106-3.81 pmol g-' d.w. than in the forest (Site 3) far from the industry (0.02-0.57 pmol g' d.w.). d e s e findings are in accordance with other results in the literature [22,23,24,40]. It has 4enerally observed that in forests having significant NO3- deposition and nitrification plants are found to have considerable nitrate reductase activity in root or leaf tissue or both, so NRA may therefore be useful as an indication of the soil status including availability of NO3 [22,23]. It is well-known that the absorption of atmospheric NO enhances the NRA of plants [24] and the rate of NR induction is controlled by the 3 0 3 - uptake of the plant [41]. These reactions are generally subjected to the regulation by the same process: induced

-

29

by nitrate flux into the plant [42]. Our measurements indicate that the imission load into the forest near the chemical industrial plant is due to nitrogenous compounds.

Acer campestre Acer tat ar i cum Euonymus europaeus

E u o n y m u s verrucosus

b

3.8 1

Sambucus nigra

0

02

04

06

08

1

umol NO, g-1 (d w )

RE8 Site

Site I

ElSite

2

3

Figure 7. Nitrate reductase activity of leaves Physiological state of sessile oak When the study aimed at revealing the physiological damages of sessile oak in the samplin sites it was found that the sli ht visible discolouration of leaves is connected with a gastic (40-60%) degradation o the photosynthetic pigments (Fig. 8, 9, 10, 11) 1431.

B

ma a-1

/L

H D Sire1

H D Site2

=sun

H D Site3 @shade

Fi re 8. Total igment concentration in eaves of hea thy (H) and damaged (D) trees

H"

P

H D Site 1

I

H D Site 2 =sun

H D Site 3 =shade

Figure 9. Concentration of chlorophyll a in leaves of healthy (H) and damaged (D) trees

30

Figure 10. Concentration of chlorophyll b in leaves of healthy (H) and damaged (D) trees

Figure 11. Concentration of carotenoids in leaves of healthy (H) and darnaged (D) trees

This was also shown in a number of surveys around point sources, the degree of visible injury was related to the loss of chlorophyll attributed to gaseous pollutants [26]. The reduction of photosynthetic pigments has appeared typically in larger extent in the upper part of the damaged trees. In case of sun leaves the deterioration of total pigments (Fig. 8) was very similar in all sites and amounts to 50-55%. Considering the shade leaves, however, there were significant differences in the degradation degree of pigments among the sites. In Site 1 (Kazincbarcika), close to the point pollution source the total pigment content (Fig. 8) and the concentrations of individual components (Fig. 9, 10, 1I ) in the shade leaves of damaged trees has changed slightly compared with the respective crown layer of healthy trees and 80-90% of the pigment content still remained. At the same time in areas located in larger distances from the industry (Site 2:FbnagysBg and Site 3:Sikfo”klit) the degree of total pigment deterioration in the shade leaves of damaged trees is more similar to the sun leaves and it has reached 55%. These results suggest that in close vicinity to the pollution source the direct effects of gaseous pollutants on the foliage of the sessile oak may be more pronounced than the indirect influences [lo, 111. In spite of large differences in the total leaf pigment content between crown layers and sites the pigment ratios changed in relatively smaller degree (Table 3). The content of total carotenoids (Fig. 11) decreased less in the damaged trees than the chloro ylls (Fig. 9, 10) and depending on the crown layer its level was 6 5 8 5 % of that of the heaEhy trees. It is generally accepted that carotenes are the second most abundant pigments in leaf tissues after chlorophylls [26] therefore Arndt suggested [44] the use of beta-carotene concentration as a reliable early indication of injury from several fumigation including SO2, H F and HCI.

31

Table 3 Ratio of the main pigment components

Site 1

healthy trees damaged trees

Site 2

healthy trees damaged trees

Site 3

healthy trees damaged trees

sun leaves shade leaves sun leaves shade leaves sun leaves shade leaves sun leaves shade leaves sun leaves shade leaves sun leaves shade leaves

chla/b

chl/car

3.39 2.9 1 3.46 2.98 3.83 3.11 4.3 1 3.39 2.5 1 2.88 2.79 2.85

3.7 1 4.4s 2.8 1 4.34 3.79 4.32 3.22 4.0 1 4.60 4.74 3.40 3.82

Our studies suggest that beside the concentrations of individual pigment constituents their ratios are also very informative concerning the state of the photosynthetic pigment systems and may he widely used for the indication of the changes induced by the disturbing effects involving the air pollutants. As shown in Table 3 in o u r studies there was a decrease in the ratio of chlorophylls to carotenoids with the decline which was much pronounced in the sun leaves.

ACKNOWLEDGEMENTS

This study was performed partly in the framework of the scientific cooperation between the L. Kossuth University, Debrecen, Hungary and the Institute of Chemical Ecology, G S F Neuherberg, Germany. The authors are very grateful to G. L6rinci and I.Gebefugi for their help in performing the analysis of the PCBs and pesticides.

REFERENCES

1 2 3 4

Jakucs P. (1988): Ecological Approach to forest decay in Hungary. AMBIO 16:267-274. Kubikovi I. (1991): Forest dieback i n Chechoslovakia. Vegetatio 93: 101-108. Schulze E.D., Lange O.L., Oren R. (eds.)(lY80): Forest Decline and Air Pollution. Ecological Studies 77:475. Springer Verlag, Berlin. Smith W.M. (19x1): Air pollution and forests. Interactions between air contaminants and forest ecosystems. Springer Verlag, New York.

32

5 6 7

8 9 10 11

12 13 14

15

16 17 18 19

20 21 22 23 24

Villanueva V.R. and Santerre A. (1989): On the mechanism of adaptive metbolism of healthy-resistant trees from forest polluted areas. Water, Air, and Soil Pollution 4859-75. Jakucs P., Meszriros I., Papp B.L. and T6th J.A. (1986): Acidification of soil and decay of sessile oak in the Sikf6klit Project area (N-Hungary). Act. Bot. Hung. 32~303-322. Amundson R.G., Alschner R.G., Fellows S., Rubin G., Fincher I., Van Leuken P., Weinstein L.H. (1991): Seasonal changes in the pigments, carbohydrates and growth of red spruce as affected by exposure to ozone for two growing seasons. New Phytol. 1183127-137. Langebartels C., Ernst D., Heller W., Sandermann H. (1991): Reactions of trees to air pollutants. Proceedings of the Ecological Approaches of Environmental Chemicals. 15-17. April, 1991. Debrecen, Hungary. pp.274-285. Lewitt J. (1980): Responses of plants to Environmental Stresses. Academic Press, New York. Meszriros I. (1990): Comparative study of healthy and diseased sessile oak trees. Proceedings of the International Symposium on Forest Decline in Bavaria and Eastern Europe, Passau Mesziros I. (1991): Effects of soil acidification on the physiology of sessile oak trees. Proceedings of the International Symposium on Ecological Approaches of Environmental Chemicals. 15- 17. April, 1991. Debrecen, Hungary. pp.306-313. Pagony H. (1992): VitalitBtsabschwBchung der Traubeneichenbestiinde. Allgemeine Forst Zeitschrift 2:98-99. Stefanovits P. (1986): Some new data on the soil acidification. Magyar TudomBny 31 :339-341. Lange O.L., Hebber U., Schulze E.D., Ziegler H. (1989): Atmospheric Pollutants and Plant Metabolism. In: Shulze E.D., Lange O.L., Oren R. (eds.) Ecological studies 77:238-273. Springer Verlag, Berlin. Molnar E., Szabb M., Szab6 J. (1991): Case study o n the pollution status of soils, vegetation and groundwater in the Sajo valley, Hungary. Proceedings of the International East-West Symposium o n Contaminated Areas in Eastern Europe, 25-27. November, 1991. Gosen, Germany. Papers Session 2125-27. Allen S.E. (ed.) (1974): Chemical analysis of ecological materials. Blackwell Scientific Publications, Oxford. pp. 1 19-234. Burns R.G. (ed.)(1978): Soil Enzymes. Academic Press, New York. Fauvel B., Rouquerol T. (1970): The phosphatase test considered as an index of soil activity and evolution. Review Ecol. Biol. Soil 7:393-406. Meszaros I., M6dy I. (1991): Applicability of microbial C 0 2 production and phosphatase enzyme activity as indicators of pollution loads into brown forest soils. In: Teller A., Mathy P., Jeffers J.N.R. (eds.) Responses of forest ecosystems to environmental changes. Elsevier Applied Science, London and New York. pp.835-836. Kranier M. and Erdei G. (19%): Investigation of the phosphatase activity of soils with the use of disodium phenylphosphate. Agrokem. Talajt. 7:361-366. A d a m M.A., Attiwill P.M. ( 1982): Nitrogen mineralization and nitrate reduction in forests. Soil Biol. Biocheni. 14: 197-202. Hiigberg P., Granstriim A., Johansson T., Lunmark-Thelin A., NBsholm T. (1986): Plant nitrate reductase activity as an indicator of availability of nitrate in forest soils. Can. J. For. Res. Vo1.16, 6:1165-1169. Thoene B., Schroder P., Papen H., Egger A. and Rennenberg H. (1991): Absorption of atmospheric NO by spruce (Picea abies L. Karst.) trees I. NO2 influx and its correlation with nizate reduction. New Phytol. 117575585. Pizelle G., Thiery G. (1986): Reduction of nitrate in the perennial tissues of aerial parts of Alnus glutinosa. Physiol. Plant. 68:347-352.

33

25 20

27 28

20 30

31

32 33 34

35 36

37 38 30

40 41 42 43 44

Baillon F., Dalschaert X., Grassi S., Geiss F. (1988): Spruce photosynthesis: possibility of early damage diagnosis due to exposure to magnesium o r potassium deficiency. Trees 2: 173-179. Darral N.M., Jiger H.J. (1084): Biochemical diagnostic tests for the effect of air pollution on plants. In: Koziol M.J., Whatley F.R. (eds.) Gaseous Air Pollutants and Plant Metabolism. Buttenvorths. London. pp.333-349. Howell R.K. and Kremer D.F. (1973). The chemistry of physiology of pigmentation in leaves injured by air pollution. J. Env. Qual. 2:434-438. Heher U., Yin Z.H., Dittrich A,, Pfanz H., Lange O.L. (1990): The response of leaves t o potentially acidic gases. In: Ulrich B. (ed.) Proceedings of the International Congress on Forest Decline Research: State of Knowledge and Perspectives. Friedrichshafen, Germany. pp.499-5 16. Rabe E. and Kreeb K.H. (1980). Bioindication of air pollution by chlorophyll destruction in plant leaves. Oikos 34: 163-367. Tschaplinski J., Timothy J., Norby, Richard J. (199 1): Physiological indicators of nitrogen response in ;I short rotation sycamore plantation 1. CO assimilation, photosynthetic pigments and soluble carbohydrates. Physiol. Plant. $2: 117-120. Czuchajowska Z., Przybylski T.( 1978): The seasonal changes of chlorophylls and carotenoids in unpolluted and polluted needles of Pinus silvestris. Bull. Acad. Polon. Sci. Skr. Sci. Biol. XXVI:30L)-37h. Watts D.F., Eley J.H. (1081): Changes in the chlorophyll a:h ratio during autumn coloration of Populus sargentii. Bull. Torrey Bot. Club 108:379-382. Holm G. (1954): Chlorophyll mutations in barley. Acta Agric. Scand. 4:457-47 1. Meszaros I., Liirinci G., Gebefiigi I. (lY91): Accumulation of pesticides and chlorinated hydrocarbons in brown forest soils. Proceedings of the International Symposium on Ecological Approaches of Environmental Chemicals. 15- 17. April, l901. Debrecen. Hungary. pp. 314-318. Lett1 A. (1990): Influence of industrial SO2 emissions on microorganisms of forest soils. Ekol6gia (CFR) VoI.0, 3:315-330. Simon T. (1988): Nature conservation ranks of the Hungarian vascular flora. Abstr. Bot. 12:1-23. Vajda L., Orbin S. ( 1983): Magyarorszrig mohatl6rrijrinak kezikijnyve. Akademiai Kjadh, Budapest. During H.I. ( 1979): Life strategies of bryophytes: a preliminary review. Lindbergia 5:2-18. Orhfin S. (1984): A mohik strategirii 6s TWR ertekei. Separatum Act. Acad. Paed. Agrensis-nova series tom XVII. pp.755-765. Campbell W.H. (1988): Nitrate retluctase and its role in nitrate assimilation in plants. Physiol. Plant. 74:214-219. Sinha S.K., Nicholas D.J.D. (1981): Nitrate reductase. In: Paleg L.G., Aspinall D. (eds.): The physiology and biochemistry of drought resistance in plants. Acadeniic Press, New York. pp.145-171. Stewart G.R. (1Y9l): The comparative ecophysiology of plant nitrogen metabolism. In: Porter J.R., Lawlor D.W. (ecls.): Plant growth: interactions with nutrition and environment. Cambridge University Press, Cambridge. pp.55-8 1. Ricks G.R. and Williams R.J.H. (1975): Effects of atmospheric air pollution on deciduous woodland. Part 3. Effects on photosynthetic pigments of leaves o f Quercus petraea (Mattuschka Leibl.). Env. Poll. 8:97-106. Arndt U. ( 1971): Konzentratioiisanclerungen hei Blattfarbstoffen unter dem Einfluss von Luftveruntreinigungen, Ein Disskussionsbeitrag zur Pignientanalyse. Environmental Pollution 237-48.

This Page Intentionally Left Blank

35

Long term effects of Acid deposition: Implications on the performance of high level nuclear waste repositories J. Nebot and J. Bruno MBT Tecnologia Ambiental, Parc Tecnologic del Valles, 08290 Cerdanyola, Spain.

1. INTRODUCTION The Earth's global environment can be considered with regard to the mass transfer as an (almost) closed system. The atmosphere's outer layer, the stratosphere, acts as an interface between the system and the surroundings, allowing at the same time energy transfer. The different compartments (atmosphere, geosphere and litosphere) forming our environment, the biosphere, can globally be considered to be in a steady state with respect to the redox reactions or to the proton and electron balances. Obviously, this global stationary state is locally and regionally disturbed. In fact, naturally occurring processes can cause the above-mentioned decoupling between proton producing and proton consuming reactions. One of this natural processes responsible for local proton imbalances is the aggradation of vegetation. In terrestrial ecosystems, accumulation of biomass occurs in inmature forests until the establishment of the definitive population of the stand. In addition, other processes can cause the proton balance to be upset, i.e. the human activities. At the present moment, one of the major causes of these local and/or regional imbalances is the anthropogenic input of protons to the environment originating from the combustion of fossil fuels for energy. The natural cycling of some principal elements, such as carbon, nitrogen and sulfur has been altered by the combustion of fossil fuels. This activity has caused the rate of the oxidation reactions to increase, while the reduction rates remain unaltered. Accordingly the concentration of oxidized compounds of the above mentioned elements has resulted globally increased in the case of CO,, and locally or regionally increased in the case of sulfur and nitrogen oxides. This increase was first noticed in the troposphere due to the smaller volume and the relatively short mixing time of this reservoir. Once in the atmosphere, sulfur and nitrogen oxides are transformed into the corresponding "strong acids", H,SO, and HNO,, which are quickly removed from the atmosphere and deposited in the litosphere. Several processes in the soil, such as denitrification, sulphate reduction and chemical weathering can neutralize the human-enhanced deposition of acids. In the long-term, the quantitatively most important one is the chemical weathering of minerals. Carbonate minerals are widespread around the earth's crust. The carbonate rocks are easily weatherable minerals, exhibiting a fast kinetics of dissolution. Hence, they provide an enormous potential for neutralizing the acid load into the soil. However, in some areas of North America and Fennoscandia, the dominant type of minerals forming the bedrock are crystalline rocks, such as granite, gneiss and quartz,

36

which are only slowly weathered. In these regions, nor the shallow and base-poor soils deriving from the crystalline rocks neither the weathering of the bedrock itself, can keep pace with the acid deposition. This results in a transfer of this acidity to the downstream compartment in the acid flux: the hydrosphere. Within the hydrosphere, the subcompartment first showing the acidification effects is the surface water, namely lakes and streams. Groundwater can also be affected over long exposition to acid deposition. The element responsible for the acid flux to the surface waters is mainly aluminum, which has been exchanged into the soil solution by protons. The first signs of local acidification were early noticed in areas where ore smelting was a current practice. Linne, in 1734, described that around a 500 years-old smelter at Falun (Sweden) no herbs could grow. However, the first major unification of knowledge about acid precipitation was achieved by Oden, a soil scientist of Uppsala (Sweden). In 1961, Oden's studies showed that acid precipitation is a large scale phenomenon and hypothesized that probable ecological consequences of the acid precipitation would be changes in surface water chemistry, decline of fish populations, leaching of toxic metals from soils, decreased forest growth and accelerated damage to materials. Unfortunately, at the present moment, Oden's hypothesis have been widely confirmed and this study itself is aimed at assessing the long-term consequences of the acid deposition on mineral weathering rates and soil erosion rates. The objective of this report is to make a prospective study of the effects of the ongoing acidification of soils, surface waters and groundwater on the host rock of high level waste repositories in granitic formations. Two main impacts on the performance of such repositories can be expected: I/ on the geological stability of the granitic formations through increased rates of chemical weathering and erosion and, 21 on the chemical composition of groundwater. This report will focus on the impact of acid deposition on the rates of weathering and erosion of granitic bedrock. Three different scenarios, based on the atmospheric emissions of CO, and SO,, are explored up to next Ice Age, which is "scheduled" to happen in some 58.000 years (1). A simple conceptual model is developed in order to asses the consequences of the environmental acidification for deep bedrock disposal.

2. SCENARIOS The main cause of acid deposition is the release of SO, and NO, into the atmosphere. The origin of the emissions of SO, is from natural sources, including both, geological activity (i.e. volcanoes and some hot springs) and biogenic activity. Anthropogenic emissions are mainly originated from the combustion of sulfur-containing fossil fuels. This is the quantitatively most important source of SO,. More than 90% of SO, emissions in Eastern Europe and industrial areas of North America are the result of the burning of fossil fuels (2). The anthropogenic total emissions of SO, account approximately for 150 millions tons per year (3). SO, emissions have nearly tripled in Europe since 1900, with the largest increase since the World War II (2).

37

NO, released into the atmosphere is also originated from the combustion of fossil fuels and from the biomass burning. Some 40 millions tons of NO, are yearly emitted into the atmosphere (3). Fossil fuels (coal, oil and gas) supply at the present 88% of global energy requirements and nuclear energy provides most of the rest (4). As fossil fuel combustion is the main source of SO, and NO,, a central scenario related to this process is proposed. It is assumed the ratio of fossil fuel energetic supply to the nuclear one will be constant until depletion of fossil fuels, although this is very uncertain under present conditions. Sulfur dioxide concentrations over Europe in last years were reviewed and projected 100 years ahead by Graedel et al. (3) (Figure 2-1). These projections assume that population and energy consumption will grow and that the firing of coal for energy (a major source of SO,) will increase. Taking into account that the world coal reserves total about 950 billion metric cubic tons (4), at the today's production rates reserves will last for almost 300 years. Hence, in our calculations we have assumed by the year 2300 the atmospheric acidity will peak. Furthermore, three different sub-scenarios are studied depending on how stringent emission controls might be during next centuries: mild control (scenario A), moderate control (scenario B) or severe control (scenario C). These scenarios are based on the same criterium (SO, emissions) as in the RAINS model developed by the IIASA (5). It is considered that the build-up of atmospheric SO, up to depletion of fossil fuels will continue at the rate estimated by Graedel et al. (3) for next 100 years. A linear increase of varying slope is thought to be the most realistic approach to model the evolution of the atmospheric SO, concentrations (Figure 2-l), although from the mathematical point of view the best fitting for the measured and 100 years projected SO, concentrations is an exponential curve. As a consequence, Eq 2-1, 2-2 and 2-3 are used to describe the build-up of atmospheric SO, (ppb) up to the depletion of fossil fuels. SOzAyear = 14 + 0.149*(year-1970) SOZByear = 14 + 0.445*(year-1970) SO,Cyear = 14 + 1.324'(year-1970)

l i i U

2,""

2"10

ZiOa

2150

2 x 0

2250

2x0

Is.,

Figure 2-1. Measured and projected atmospheric concentrations of S 1, (3). The linear fitting is taken in order to model the atmospheric SO, concentrations up to the depletion of fossil fuels (2300).

38

Results of the calculations based on Eq 2-1, 2-2 and 2-3 for the expected sulfur dioxide concentrations in the atmosphere for the explored scenarios by the exhaustion of fossil fuels are shown in Table 2.1.

CONTROL EMISSIONS

RATE OF INCREASE (PPb*Y-l)

scenario A scenario B scenario C

severe moderate mild

0.149 0.445 1.324

[~OzI,,, (PPb) 63 161 45 1

Table 2.1. Rate of increase and expected concentrations of atmospheric SO, over Europe at the depletion of fossil fuels (2300) depending on the explored scenario. It is well known that CO, concentration is also increasing in the atmosphere due to the same processes as SO, and NO, (consumption of fossil fuels) as well as deforestation. This accounts for perhaps half of the build-up since the year 1800 and 20% of current emissions (6). The current average concentration is around 350 ppm, while the background concentration is estimated to be some 290 ppm (6-8). The proposed scenario assumes a moderate increase of CO, during next centuries. When making long-term predictions (more than 300 years) a moderate rate of increase of 0.8 ppm per year (7) is more realistic than the ones assumed in shortterm predictions i.e. 1.3 ppm per year as projected by Wayne (8) or CO, concentration by 2000 to be 21% higher than the 1977 concentration (9). Hence, we have studied in the various scenarios the increase in CO, concentration and consequently, the acidity and the buffering capacity of the C0,-HC03- system. The evolution of atmospheric concentration of CO, (pprn) from present to year 2300 is described by Eq 2-4. CO, yeat = 334 + 0.8*(year-1970) (Eq 2-4) It has to be noticed that the evolution of the NO, concentrations in the atmosphere is not considered in the scenarios. Two main reasons justify this assumption. The first one is the fact that reactions of oxidation of SO, in the atmosphere yield a sulfur residence time of several days; this corresponds to a transport distance of hundreds to a thousand kilometers. The formation of HNO, is more rapid, this results in a shorter travel distance, thus it can be assumed that nitrogen deposition accounts only for local acidification processes which are not considered in our regional model. The second reason is that nitrogen is a growth-limiting nutrient in many terrestrial ecosystems in Europe and North America. Thus, a moderate additional input of nitrogen will usually lead to an increased primary production. A growing forest needs 0.5 to 0.8

39

g.m-2.y-1of nitrogen. This is partly confirmed by the fact that the analytical data from 12 Swedish stations performed by Rodhe et al. (10) showed no further increase of NO,- in precipitation between 1972 and 1984. The scenario developed considers that after depletion of fossil fuels (2300) the remainder of SO, is washed out from the atmosphere until its preindustrial level. The half-life of SO, in the atmosphere is assumed to be 10 days (0.027 years) and its natural background concentration 5 ppb. Assuming a first order kinetics, the removal of SO, from the atmosphere is described by Eq 2-5. InSO, year = InSO, 2300 - (25.3*(year-2300)) (Eq 2-5) The CO, concentration in the atmosphere after depletion of fossil fuels will decrease slowly due to the fact that its average residence time in the atmosphere is estimated to be some 100 years. This is much longer than the average residence time of sulfur dioxide (a few days). The decrease of atmospheric CO, is also considered to follow a first order kinetics, assuming 290 ppm as a final background pre-industrial concentration (this concentration corresponds to 1860 based upon data of Schneider, 1989) and a half-life time of 100 years, the decay of atmospheric CO, after depletion of fossil fuels is described by Eq 2-6 (Table 2.2 shows the half-life time and rate constants used to model the kinetics of removal of atmospheric SO, and CO, after depletion of fossil fuels). InCO, year = InCO, 2300 - (0.007*(year-2300)) (Eq 2-61 After recovery of SO, and CO, background concentration, the scenarios developed are kept to cover a time span of some 58.000 years in which SO, and CO, levels remain constant until next Ice Age.

so* scenario A scenario B scenario C CO,

HALF-LIFE TIME

RATE CONSTANT

(year)

(year-1

0.027

25.3

100

0.007

TIME REQUIRED RESTORE CONDITIONS (year)

0.100 0.137 0.178 106

Table 2.2. Half-life time, rate constants and computed time required to restore background concentrations of atmospheric SO2 and C02 from 2300 on.

40

3. DEFORESTING EFFECTS OF ACID PRECIPITATION Although an increase in forest damage has been recorded in Europe and North America since the early seventies (11,12), no consensus on its causes and mechanisms has yet emerged. Among the hypothesized causal agents are virtually all the pollutants species present in industrialized environments. Our interest focuses on tree damage due to acid deposition, including both wet and dry deposition effects. Basically, the suggested mechanisms of damage can be separated in two main groups: - Direct effect of the gaseous pollutants on the foliage. - Indirect effect on the root system through acidic deposition on soils. The first mechanism is mainly exhibited by gaseous sulfur dioxide. In addition to adsorption on leaf surfaces, SO, enters the leaves through the open stomata and can cause abnormalities in cell structure and altered metabolic activity. In wet surfaces, SO, and dry deposited sulphate particles dissotve in water and the H,SO, produced enhances the erosion of the leaf surface, which protect the leaf from the water losses and prevents leaching of nutrients from the foliage . Furthermore, a mechanism of co-deposition of SO, and NH, in coniferous trees has been suggested (13). This mechanism may cause an enhanced deposition of nitrogen compounds over large areas of Europe. Acid deposition may cause soil changes detrimental to forest vegetation, either by stripping nutrients from the soils or by mobilizing phytotoxic elements. It is still matter of discussion the way that acidic deposition on soils can cause the reported mortality of fine roots of declining trees (12). An increase of aluminum in soil solution was noted as was a calcium deficiency in roots in other declining stands. It has also been suggested that the ratio AI:Ca is responsible for the forest dieback (14 and references therein). It has been described that not only the SO, concentration, but also the period of exposure determines the degree of forest damage. For that reason the threshold concept can not be applied to tree damage caused by SO,. The IUFRO Air Pollution Section, at this 1982 meeting in Oulu, Finland, decided that no safe limit could be set for SO, ambient concentration to protect forest trees (1 5). In view of the variety of pathways to damage the trees as well as that there is no safe threshold, it seems quite reasonable to conclude that the effect of SO, on the forest trees is related to a physiological stress which makes the trees much more sensitive to other factors i.e. diseases and climate (16). The forest decline induced by acid deposition might eventually lead to a large areas of deforested land around Central and Northern Europe. The main consequences of such deforestation processes concerning the geochemical stability of granitic formations are: A/ Medium to poor forest soils are formed into podzols (the predominant soil type in Sweden). Podzolic soils exhibit a pronounced stratification. The litter accumulates on the top of the soil and it is slowly decomposed into humus by bacteria and mainly by fungi. The raw humus layer often exhibits a pH below 4. Some of the processes occurring in forested podzols concerning the acidic flux are:

41

- The regeneration of the cation store in humus layer is supplied by the decomposition of the litter. Tree roots act as a base cation pump through the A, and B horizons of the podzols (17). The base cations originated by weathering in the B horizon are translocated in forested soils to the humus layer, by root uptake, litterfall and litter decomposition. - Complex bacterial populations are supported in forested ecosystems, carrying out among others processes, the reduction of sulphate and denitrification. Both processes are a sink for hydrogen ions. - The acid load into the soil due to nitrogen compounds is considered to be of lesser importance in front of sulphate deposition due to the uptake of such compounds during biomass growth (see 2. Scenarios). As deforestation proceeds, this sink of nitrogen compounds will disappear, thus, an increase in the acidity of the soil solution can be expected. Hence, deforestation could result in a decrease of base saturation status in the top soil and in a decreased capacity of sinking protons and nitrogen compounds. This is likely to cause an increasing rate of acidification on such deforested soils. It has been well established from laboratory studies that mineral weathering rates are dependent on hydrogen ion activity (18 and references therein). A field study comparing two forested basins in Central Europe representing strongly acidified due to high deposition of SO, and less acidified environments of industrial and rural countryside, indicated increased weathering rates of bedrock and depletion of exchangeable cations from soils due to such an acidification (19). Therefore, from lab and field studies, it seems reasonable to conclude that an increase in proton activity in soils due i.e. to deforestation can account for a faster weathering rate. BI It is well known that afforestation is the best way to prevent soil denudation. Denudation is a complex process involving detachment and downslope transport under the influence of both gravity (mass movement) and water in motion (slope erosion) (20). Weathering and soil formation are dependent upon several factors i.e. temperature, particle size, surface effects, biotic effects (21 and references therein) and variation in these factors causes the processes to proceed at different rates. If the rate of soil formation is defined as the rate at which rock is converted into soil, then weathering and soil formation are closely related processes, particularly in granite. On relatively pure limestone, however, a large depth of rock may weather giving only soluble species and leaving only a shallow soil. Soil formation implies a loss of mass due to several processes i.e. dissolution of mobile elements. Although the bulk density of the soil formed is lower than the parent material bulk density (soil formation from bedrock implies an increase in the porosity), there is an overall reduction in the total volume of material, i.e. the volume of soil formed is less than the volume of rock weathered to form it. Hence, the process itself of bedrock weathering and soil formation implies a net surface lowering or landscape reduction, even without taking into account additional processes like denudation. As an example, the rates of weathering and soil formation on granite were studied in two areas of Rhodesia using small watersheds (22,23). The rate of granite weathering was calculated using Barth's equation (24) which links the rate of weathering to the amount of an element removed in solution per unit of time, its concentration on the rock and its concentration on the weathered product. Results of these field studies indicated

42

rates of granite weathering of 15.4 and 11.O mm per 1000 years for the higher and lower rainfall areas, respectively. The soil production rates were 11.0 and 4.1 mm per 1000 years. Thus, the net surface lowering rates were 4.4 and 1.7 mm per 1000 years, even in areas where soil production is high. An estimation of the rate of surface lowering or landscape reduction under deforestation is a key point in this study. The maximum expected surface lowering rate corresponds to a situation where the soil is continuously removed while forming as it could happen in deforested areas. In this case, the rate of surface lowering is the same as the rate of bedrock Weathering. In southern Fennoscandia where thin soils are often found, deforestation could be responsible in short periods of time for total losses of existing soil due to denudation. In addition, no further recovery of soil mantle can be expected in deforested areas. A deep mature soil over the bedrock protects it from weathering due to its buffer capacity. As the soil mantle thickens, the weathering zone is farther removed from the surface being decreased the available surface and consequently the weathering rate. An example of the role played by the soil preventing the chemical weathering of the bedrock is the field study conducted in upland forested areas in New England (Hubbard Brook Forest). As mentioned by Johnson et al. (25) only a moderate rate of weathering has been measured in this area in spite the acidification of their waters. The cause of this slow weathering rate is not clear. However, the maturity of the soil was suggested as a contributing agent. On the other hand, in shallow or non-existing soils, fresh rock surfaces are constantly being created. These surfaces are completely exposed to physical, chemical and biological weathering agents. Among the physical agents are precipitation and changes of temperature (freezing and melting processes) causing rock breakdown. Biological processes involve the direct colonization of bedrock by microflora and microfauna exerting a mechanical effect on rock as well as a chemical one releasing i.e. organic acids. The final result of all these effects is again an increased rate of weathering of the bedrock. The effects of deforestation discussed in A and B cannot be quantified separately. However, as it has been shown by Schnoor (26) weathering rates measured in the field are always 1-2 orders of magnitude lower than the ones measured in the laboratory at the same pH conditions (Figure 3-1). This has been rationalized by taking into account that the hydrological condition in the field is not the same as in the lab. Under ideal laboratory conditions, weathering proceeds much faster than in the field because of the larger availability of wetted surface. -1 1

c -12 m

’?

E

2

13

-14

c a

8

-15 -16

; -7

-6

5

4

3

2

1

lcg Fbwraie I Mass Soil (L day.’ g-’ )

Figure 3-1. Dissolved silica release rate (weathering rate) vs flow-rate:mass ratio (26).

43

A similar phenomena could be thought to be the final stage of deforestation, where maximum surface and porosity are available due to the disruption of the soil and “fresh” surfaces are continuously formed as a result of the increased erosion. Hence, in our model the worst case scenario is considered to be the one causing total deforestation. This leads to a situation where weathering rates are large as the ones measured in laboratory studies.

4. DESCRIPTION OF THE MODEL A conceptual model was designed to study the environmental acidification processes. The total sulfur concentration (as total sulphate) is considered the main driving variable of such processes. A compartment structure (box model) was found to be a good approach to simulate the physically different reservoirs involved in the process of acidification. Such compartments or reservoirs are assumed to be connected by the transfer of sulfur compounds. Four reservoirs were considered to be relevant in the natural flow of sulfur, namely atmosphere (troposphere), surface waters, soil-bedrock and groundwater. A schematic representation of the compartment system is shown in Figure 4-1. A more detailed description of the designed compartments and assumptions is given in next subchapters. The model developed in this report refers to the atmosphere, soil and bedrock compartments The modelling of surface and groundwater compartments is not considered in detail in this work. A regional model of the hydrological conditions of the region is under development. The consequences of the acidification processes in these compartments will be studied when a complete hydrological model for surface and groundwater mixing will be available.

water

Lqran1te

I

I

weathering] I

water

Figure 4-1. Schematic representation of the acid flux through the compartments involved in the developed model.

44

4.1. Atmosphere As mentioned above, the main driving variable for the environmental acidification processes is the total deposition of sulfur. In the developed model the atmospheric compartment is the origin of the acid flux, being sulphate the assumed species responsible for the acidic deposition. Sulfur is emitted as a gas, sulfur dioxide, but it is transformed into fine sulphate particles once in the atmosphere (8,27). A process which, among others, gives sulphate particle is direct SO, photochemical oxidation, schematically shown as: 2S0, + 0, + hv ---> 2S03 where SO, is further hydrolyzed into H,SO,:

SO, + H,O + M ---> H2S04+ M The model assumes all the SO, is finally washed out from the troposphere in the form of H,SO,. No difference is considered in the deposition mechanism (dry or wet) on how the acidity reaches the soil reservoir. The average SO, atmospheric concentration is used as the main variable in order to describe the scenarios. The sulphate concentration is considered to drive the acidification of soils. Hence, a linkage between SO, atmospheric concentrations and sulphate deposition has to be defined. A direct correlation is assumed between SO, atmospheric values and sulphate concentration in precipitation on Southern Sweden. In order to obtain an average value for sulphate concentration in rainwater in this area, we relied on data taken from the European Air Chemistry Network. The area considered comprises from 5 9 N to 60QNlatitude and from 1OQEto 20GE longitude (basically, Central and Southern Sweden). Around this region 44 stations were monitored for rainwater chemical data between 1955 and 1979 by the Swedish University of Agricultural Sciences of Uppsala and the International Meteorological Institute of Stockholm (28). We used selected data sets (see Appendix A) to calculate the average sulphate concentration in rainwater in 1970. Median values of each station have been used to estimate average sulphate concentration in precipitation in the concerned area (28,29). The computed value was assigned to 1970 as the origin of our calculations. The available sulphate values (28) referred to excess sulphate (i.e. the non-seasalt fraction). Since in ocean islands and coastal areas sulphate originating from seaspray contributes very significantly to the sulphate concentration in precipitation samples, the seasalt fraction of sulphate was deducted in this study based on the sodium concentration according to the following formula: [S0,2-]ex= [S042-]- 0.25 “a+] In order to calculate the total sulphate concentration (seasalt + non-seasalt), the sodium average concentration in precipitation has to be known. This value was calculated from the data .of the sodium content in precipitation taken from Soderlund et al. (30) by using median values of each station. The average sodium content in precipitation was calculated to be 47 ~ m o l e . d m - ~ . The computed average excess sulphate content of the rainfall was also 47 ~ m o l e . d m -Hence, ~. the calculated total concentration of sulphate was 58 p m ~ l e . d m - ~ .

45

This concentration was taken as a base level corresponding to 1970 for next calculations. All calculations were made by taking as average for the region considered an annual precipitation of 1000 mm.m-* (31 and references therein). The linear relationship between time and concentration of sulphate (mole.dm-3) in precipitation is described for each scenario by Eq 4-1, 4-2 and 4-3. S0,2-Ayear = 5.8'1 0 ~ +5 0.497*1O-6*(year-l970) (Eq 4-1 1 S0d2-Byear= 5.8*10-5 + 1.403*1O-6*(year-l970) (Eq 4-21 S0d2-Cyear= 5.8'1 0-5+ 5.067'1 O-6*(year-1970) (Eq 4-31 In order to validate the reliability of such equations to predict the total sulphate concentrations of the rainfall, a check against the measured concentrations during the eighties was made. Our results are in good agreement with the measured data of the Co-operative Program for the Monitoring and Evaluation of the Long Range Transmission of Air Pollutants in Europe (EMEP) (32). These data are presented as isolines of volume-weighted average of sulphate in precipitation on Europe (1 978-1982). Southern Scandinavia falls between the 1.5 and 4.5 mg S042--S.dm-3isolines. Assuming the scenario B (moderate control of emissions), the one actually operating in Scandinavia, we estimate a sulphate concentration of 2.3 mg S042--S.dm3. Base cations can neutralize in the atmosphere a substantial amount of acid deposition, this has to be considered in order to calculate the acid load entering the soil compartment. Alkalies are generated in the atmosphere as the carbonates of windblown dust, generally of natural origin, and from seaspray in coastal areas. It is known from precipitation measurements (33) that most of the base cation contain is in the form of calcium and magnesium. It is difficult to quantify base cation deposition. As noticed by Kamari (cited in 34) the ratio base deposition/sulfur deposition is fairly constant. The combined Ca2+ and Mg2+deposition was estimated to neutralize an average of 33% of the sulfuric acid in bulk precipitation, within a range of 12% to 44%. Hence, we assume in the model that one third of the sulfuric acid is readily neutralized in the atmosphere before reaching the soil. The computed total acidity (Htot) of rainwater as a function of SO,2- concentration - ~ each scenario is described by Eq 4-4. and dissolved CO, (H,CO,(aq)) in m ~ l e . d mfor HtOtA,B,Cyear = 1.33*S042-A,B,Cyear+ 2'H2C03(aq)year (Eq 4-4) where H,CO,(aq)year is calculated from CO, Henry's constant as shown in Eq 4-5. H2C03(aq)year= 1.132*10-5+ 2.71 1'10-8'(year-1970) (Eq 4-51 Due to the composition of rainwater, the total and free acidities are almost similar and H+ concentration approximates the Htot (36). The expected pH values of the rain solution in southern Sweden are shown for the three studied scenarios in the period 1970-2300 in Figure 4-2.

46

42

1

1970

20m

2033

2060

2090

21a

219

2180

2210

2240

2270

2300

V ,..

Figure 4-2.Evolution of the rainwater pH up to depletion of fossil fuels (1970-2300). 4.2 Soil-Bedrock The soil is considered in this model as an interface between the atmosphere and the bedrock, for this reason they are not treated as a separate reservoirs. This chapter is devoted to the acidification processes which occur in forested soils. Since agricultural soils are intensively managed with lime and other chemicals, acidification is not expected to proceed in these soils. Furthermore, one of the objectives of this study is to assess the degree of deforestation due to acid deposition. 4.2.1.Processes of acidification in forested soils

Soil acidification has been defined as a decrease in the soil acid neutralizing capacity (alkalinity) accomplished by removal of alkaline earth cationic components or, to a lesser extent, by addition of acidic components. This occurs by an irreversible flux of protons to the soil. The proton sources include: - atmospheric inputs of acidic or potentially acidic substances - net assimilation of cations by vegetation - net mineralization of anions from organic matter - deprotonation of weak acids - oxidation reactions - precipitation of cations as a consequence of secondary phase formation - mineral weathering of anionic components We will only consider the first point which represents the anthropogenic input of protons to the soil system (acid deposition), whilst the rest are taken as background processes which do not upset the global balances. Protons entering or produced in the soil are removed by: - export in drainage water - net mineralization of cations in organic matter - net assimilation of anions by vegetation - protonation of anions - reduction reactions - weathering of cationic components - precipitation of anions Once the acid input to the soil is completely neutralized, cations are removed by vegetation uptake and/or export into drainage water, resulting in soil acidification and in an increase in the alkalinity of percolating water.

47

Different reactions are responsible for the soil buffer capacity. The concept of buffer capacity corresponds to the above mentioned concept of alkalinity or acid neutralizing capacity. It is defined as the total reservoir of basic components in the soil. Ulrich (cited in 37) classified the inorganic buffering reactions in the following buffer ranges: - Carbonate buffer range: the buffering capacity in this range is supplied by CaCO, (Eq 4-6), hence only calcareous soils are within this range. The amount and the rate of dissolution of CaCO, in such soils is usually enough to buffer moderately large amounts of acidic deposition. The soil solution pH within this range is higher than 6.2. CaCO, + H+ = HCO; + Ca2+ (Eq 4-61 - Silicate buffer range: this corresponds to the buffering range of feldspar weathering reactions which can be generalized as described by Eq 4-7. KAISi,O, + 7 H,O + H+ ---> AI(OH),(s) + 3 H,SiO, + K+ (Eq 4-71 These are kinetically slow reactions and cannot be described as an equilibrium buffer capacity. Hence a silicate buffering rate capacity is defined. This corresponds to the rate of reaction 4-7. The pH of the soil solution in this range is between 5.0 and 6.2. - Cation exchange buffer range: base cations such as Ca2+, Mgz+, Na+ and K+ exist in the soil mainly complexed by the organic matter (humic and fulvic substances) and/or bound to clay particles. The buffer capacity given by exchange reactions with humic substances can be described by a general reaction (Eq 4-8) as follows:

HUM-Ca + 2 H+ = HUM-H, + Cazi (Eq 4-61 The buffer capacity given by surface reactions or: clay particles may be generalized in reactions of the type: >SiOK + H += >SOH + K+ (Eq 4-91 where the symbol > denotes a surface process. These two processes are kinetically fast and they constitute the more readily available buffering capacity of non-calcareous soils. The number of reactive sites available is approximated to the Cation Exchange Capacity (CEC). Base saturation (0) is the fraction of CEC consisting in base cations. Buffer capacity of the cation exchange range (OCE) is defined as the product of CEC and 0. If the total acid load overalls the cation exchange buffer capacity, this buffer becomes depleted and silicate weathering becomes the only neutralizing process available. The pH of the soil in this buffer range is between 5.0 and 4.2. - Aluminum hydroxide buffer range: as the alkalinity of the soil solution is depleted and as a consequence of feldspar weathering, AI(III) hydroxide becomes abundant in the soil solution. At the low AI(III) concentrations and pH range of this buffer (pH= 4.2 to 3.0),the main equilibrium is given by :

AI(OH),(S) + 3 H+ = ~ 1 3 ++ 3 H,O for poorly crystallized gibbsite this is a fast reaction.

(Eq 4-10)

48 If this buffer capacity is exhausted, other solid phases dissolution and precipitation reactions can take over. For instance, Fe(OH),(s) dissolution has been considered to buffer the soil solution at very low pH values. 4.2.2 Main assumptions of the model developed

In our effort to model the process of soil acidification in Southern Fennoscandia, we assume, based on data taken from the FAO-UNESCO Soil Map of the World (1974) (in 34), that the type of forest soil predominant in the area to be studied is orthic podzol. The orthic podzol is a non-calcareous soil deriving from granitic or base-poor material. Podzols in Scandinavia are often shallow soils, exhibiting a pronounced stratification. On the top of the soil it is found a humus layer originating from the slow decomposition of litter. The pH of this layer is often below 4 in medium to poor podzols. A bleached horizon (A2), usually grayish in color due to intense leaching, follows below the raw humus layer. Below this layer, there is an enrichment zone, the B-horizon, reddish to brownish. Iron leached from the bleached horizon precipitates here in trivalent form as goethite (17). In order to simplify the calculations, the model assumes that the soil layer is a homogeneous box 50 cm deep. In addition, we deal with yearly variations in soil acidification in order to avoid seasonal fluctuations. This seasonal variations are mainly caused by biological activity and tend to be internally compensated in the ecosystem over the year. Since podzols are non-calcareous soils, it is postulated that the carbonate buffer range has been surpassed. Therefore, the model assumes that at the origin of our calculations (1970) the soils considered here are at the stage at which cation exchange is the dominant buffer reaction. Cation exchange buffer capacity (OCE) includes in our model the ion exchange capacity of the organic substances as well as the cation exchange from silicate interlayer positions by a slow diffusion phenomenum reported in young soils of Scandinavia by Graunstein (cited in 17). It is postulated that the soil exhibits some capacity of recovery in front of acid deposition due to the organic matter, represented in our model by an initial term (OCEO). This recovering capacity is assumed to be linearly depleted with time at the same rate as acid load increases as computed in Eq 4-1 1, 4-12 and 4-1 3. (Eq 4-1 1) DCEOAyear = 2.0 - (7.182’1 O-4*(year-l970)) (Eq 4-1 2 ) BCEOByear = 2.0 - (1.924*10-3*(year-I970)) DCEOCyear = 2.0 - (6.809*1O-3*(year-l970)) (Eq 4-13) Silicate weathering rate (see section 4.2.4) is assumed to be only slightly dependent on the soil solution pH. If we assume that the weathering rate (brSi) is proportional to [H+Io5,a decrease of pH in the soil solution of 2 units results in a ten-told increase in the rate of weathering. In addition, the rate of release of base cations is assumed not to depend on the soil pH. The modifying effect of the forest canopy (forest filtering) on the deposition is also taken into account in the model. There are two ways in which the canopy is modifying precipitation input to soil. One way is dry deposition. From an ecological point of view, any vegetation canopy behaves like a filter or a sink for the fluxes of matter passing along its surface, the filter efficiency being very dependent upon its physical and

chemical properties (38). Hence, forest filtering can cause forested soils to receive more sulfur deposition than adjacent cropland or pastureland. Another way of modifying soil input is by leaching substances previously taken up by the roots and translocated to the upper parts of the tree. Thus, when they get to the soil surface the total precipitation input to soil is a mixture of substances coming from outside the ecosystem, i.e from the atmosphere, and other compounds which are merely completing an internal cycling (38). In addition, it has been shown by lvens (cited in 34) that forest filtering is quantitatively different in coniferous stands or in deciduous stands. The total surface of the needles of coniferous trees is bigger than the total surface of broad-leaved trees. Furthermore, in wintertime deciduous trees loose their leaves, while coniferous needles are kept through the year. Both factors cause the coniferous stands to intercept much more acid deposition than the deciduous stands. It is difficult to quantify the concentrating effect of coniferous canopies. Our model assumes that soil under coniferous trees receives an acid load due to sulfur compounds 1.6 times greater that soil under deciduous trees (35). In our model the net surface lowering is defined as the depth of bedrock weathered minus the depth of soil subsequently formed. A direct correlation between the assumed silicate weathering rate and bedrock surface lowering rate is assumed in order to calculate the net surface lowering. One of the critical assumptions of the model is that once acidification proceeds further to the AI(OH), buffer range, deforestation proceeds for the lifetime of the forest stand (50 years). Deforestation causes the total lost of soil mantle due to the increased effects of erosion and denudation. In these conditions the dissapearence of the soil mantle is irreversible and there is no longer soil formation. As a consequence, due to the disruption of the soil structure, the bedrock surface becomes more available and the weathering rates get closer to the ones measured in ideal laboratory conditions. 4.2.3 Assignment of values to soil parameters

A key point in the development of any regional model is the correct quantification of the parameters involved in the model such as, in our case, cation exchange capacity and base saturation. Ideally, field measurements of the relevant soil parameters should be preliminary to modelling and forecasting. In our study, since direct field measurements of such soil parameters were not accessible, literature values had to be relied upon. Following an extensive literature search it was decided to use the data elaborated by Kauppi et a1.(34) based on data reported in the Appendix of the FAO-UNESCO Soil Map of the World (1974). In our model, initial conditions (1970) are always set to the less favorable conditions. The CEC and 0 values are initially set to 20 moles.m-* and 0.1 respectively, giving a cation exchange buffer capacity of 2.0 mo1es.m-2. Silicate weathering rates were also taken from Kauppi et al. (34 and references therein). These weathering rates were calculated from the data reported in the Geological Map of Europe and the Mediterranean Region. The initially assigned silicate weathering rate is 0.05 mole.m-2.y-'. The weathering rates measured across Europe range from 0.02 to 0.2 mole.m~z.y-i(cited in 34). We assume that the above mentioned effect of soil disruption caused by extensive

50 acidification over the Al(OH), buffer range, increases 1.5 orders of magnitude the silicate weathering rates. This is based on the observations and discussions of section 3 regarding the differences between measured weathering rates in laboratory and field conditions. Under these conditions, the silicate weathering rate (brSiAl) is assumed to be 1.58 moles.m-2.y-1. Lerman (39) reported rates of bedrock surface lowering, elaborated from chemical weathering field measurements (40) in different geological environments, ranging from 0.006 to 0.14 mm.y-'. An intermediate bedrock lowering rate of 0.075 mm.y-' is used in our model. The assumed rate of soil formation from granite is 9.10-3mm.y-' (20). A conversion coefficient (f) of 1.5*10.3 is used in our model to relate silicate weathering rate (mo1e.m-2.y-I)and bedrock surface lowering rate (m.y-') based upon the assigned values to bedrock surface lowering and chemical weathering rates. 4.2.4 Model development

The main features of the developed model can be summarized as follows: - calculation of the annual acid load entering the soil - determination of the dominant buffering range - calculation of the annual soil pH - calculation of the net surface lowering at the corresponding year The model computations are separated into the following steps: 1. The acid load entering the soil per year (HA,B,Cyear) is calculated from total acidity of rainwater (Eq 4-4) assuming an annual precipitation of 1000 mm.m-2. 2. The annual cation exchange buffer capacity (RCEyear) is calculated by subtracting the silicate weathering rate (brSi) from the acid input into the soil. This result is then subtracted from the annual initial buffer capacity of the cation exchange range (DCEO). These calculations can be represented as: (Eq 4-14) OCEyear = DCEOyear - (Hyear - brSi) This is an annual iterative step while soil KE>O (the cation exchange buffer capacity of the soil is enough to neutralize the annual acid load). If OCE=O, it is considered that the soil drops into the AI(OH), buffer range. 3. The soil pH is calculated within the silicate, cation exchange and upper aluminum buffering ranges f D>O) according to a non-linear relationship between base saturation and pH (41) as follows:

(Eq 4-15) PHyear 4.0 + 1.6 * (OCEyear/CEC)o.75 4. The silicate weathering rate (brSiyear) is recalculated as a function of pH in periods where acid load entering the soil is high (1970-2300). The weathering rate constant (ksi) is calculated assuming a 0.5 order dependence on free proton activity. 5. The total amount of silicate weathered (WSi) during any period is calculated by integrating the weathering rate equations within that period. 6. The net surface lowering (NSLow) is calculated by establishing a direct correlation between the amount of silicate weathered (WSi) during the corresponding period, the bedrock surface lowering rate and the soil formation rate (SoF) as follows: NSLowyear = (WSiyear * f) - (SOF* nyears)

(Eq 4-16)

51

where f is the conversion factor from mole.m-2to m deep and nyears is the time span considered. The driving variable of the developed model is acid load (HA, HB, HC) which depends on the SO, and CO, atmospheric concentrations for each scenario considered. Based on the acid load entering the soil compartment three different periods of time can be distinguished (see section 2.): - a period of 330 years, from 1970 to 2300 (estimated exhaustion of fossil fuels) during which the soils are receiving an increasing acid load. - a short period of time (a few months for SO2 and some 100 years for CO2) for the recovery of the background concentration of such gases. The soil Compartment receives a decreasing acid load. - finally, a time span of some 58.000 years up to next Ice Age, where the acid load entering the soil is constant and corresponds to the acidity originated from natural sources. The model computations are performed separately for each of the above mentioned periods of time.

Increasing acid load: 1970-2300 The annual acid load (moles. m-2. year-') for each scenario is calculated according to Eq 4-17 as follows: (Eq 4-17) HA,B,Cyear = a'l 03*S0,2-A,B,Cyear + 2*1O3'H,CO3(aq)year where a= 1.33 in soils supporting deciduous stands and a= 2.13 (1.33'1.6) in soils supporting coniferous stands (see Eq 4-4 and section 4.2.2). The evolution of the annual acid load in forested podzols along this time-span for the three scenarios studied is depicted in Figures 4-3 and 4-4.

Im

an,

2 x

2%

mo

m

year

Figure 4-3. Acid load entering soils supporting deciduous forests from 1970 to depletion of fossil fuels (2300) for the three studied scenarios.

Figure 4-4. Acid load entering soils supporting coniferous forests from 1970 to depletion fossil fuels (2300) for the three studied scenarios.

52

The total acid load received by the soil during the period 1970-2300 depending on the scenario studied and the type of afforestation is shown below in Table 4-1.

Total Acid Load (moles.m-*) scenario B scenario C

scenario A Deciduous stands

72.0

137.8

403.8

Coniferous stands

109.0

214.2

639.8

Table 4-1. Total acid load received by podzols in Southern Sweden depending on the scenario studied and the type of afforestation during the period 1970-2300. The variation of the cation exchange buffer capacity with time is calculated according to Eq 4-18. OCEA,B,Cyear = OCEOA,B,Cyear - (HA,B,Cyear - brSi) (Eq 4-18) The evolution of cation exchange buffer capacity and base saturation of the soil during this period of time depending on the afforestation is shown in Figures 4-5, 4-6, 4-7 and 4-8. I

1w

am

m

891

mo

m

vear

Figure 4-5. Evolution of soil cation exchange buffer capacity indeciduous forests from 1970 to depletion of fossil fuels (2300) for the three studied scenarios.

Figure 4-6. Evolution of soil cation exchange buffer capacity in coniferous forests from 1970 to depletion of fossil fuels (2300) for the three studied scenarios.

53

Figure 4-7. Evolution of the base saturation of soils supporting deciduous stands from 1970 to depletion of fossil fuels (2300) for the three studied scenarios.

Figure 4-8. Evolution of the base saturation of soils supporting coniferous stands from 1970 to depletion of fossil fuels (2300) for the three studied scenarios.

The predicted soil pH values under different conditions of afforestation during the period 1970-2300 are shown in Figures 4-9 and 4-10.

Figure 4-9. Evolution of the pH of soils supporting deciduous forests from 1970 to depletion of fossil fuels (2300) for the three studied scenarios.

Figure 4-10. Evolution of the pH of soils supporting coniferous forests from 1970 to depietion of fossil fuels (2300) for the three studied scenarios.

As mentioned above, the silicate weathering rate was kept constant through the computations. Once the pH of the soil was calculated, a check of the reliability of this assumption was made. The silicate weathering rate constant (ksi) was calculated by assuming that the initial weathering rate (0.05 mole.m-2.y-1)was measured at pH= 4.33. By taking into

54

account a 0.5 order kinetics of the silicate weathering rate (1 8) ksi is calculated as:

- brSi * (10-4.33)-0.5 kSI. (Eq 4-19) Therefore, the dependence of silicate weathering rate upon time (or soil acidity) for each scenario can be written as a rate equation as follows: (Eq 4-20) brSiA,B,Cyear = ksi * (HA,B,Cyear)O.' The variation of the silicate weathering rate with time (or soil acidity) is depicted in Figures 4-1 1 and 4-12. The total amount of silicate weathered (molem2) can be easily calculated by integrating the weathering rate equation (Eq 4-16) within 1970 and 2300, except for the scenario C. In this worst case scenario, the soil cation exchange buffer capacity is exhausted by 2144 in the case of soils supporting deciduous stands, and by 2078 in soils under coniferous forests. The rate equation developed can only be applied within this buffer range and the amount of weathered silicate is computed by this method up to those limits. During next 50 years (see section 4.2.2.) the weathering rate is assumed to remain constant and finally from 2128 (coniferous stands) and 2194 (deciduous stands) the silicate weathering rate is assumed to be the one which applies for the AI(OH), buffer range. Table 4-2 shows the weathered silicate during this time span depending on afforestation and scenario.

Figure 4-11. Variation of the weathering rate with time (or soil acidity) from 1970 to depletion of fossil fuels (2300) on soils supporting deciduous forests depending on the studied scenario.

Figure 4-12. Variation of the weathering rate with time (or soil acidity) from 1970 to depletion of fossil fuels (2300) on soils supporting coniferous forests depending on the studied scenario.

The net surface lowering is calculated for the scenarios A,

B as shown in Eq 4-21.

NSLowA,B1g70-2300= (f * WSiA,B1970-2300) - (SoF (2300-1970)) (Eq 4-21)

55

where f= 1.5'10-3, being the conversion coefficient from silicate weathering rate (mo1e.m2.y-1)to surface lowering rate (m.y-l)(see sections 4.2.2 and 4.2.3).

scenario A

Weathered Silicate (mole.m-2) scenario B scenario C

Constant weathering rate: coniferous and deciduous stands

16.5

16.5

16.5

Weathering rate dependent on the pH: deciduous stands

18.0

19.0

227.1

Weathering rate dependent on the pH: coniferous stands

18.4

(8.7+3.5+214.9)

19.9

281.O (5.6+3.6+271.8)

Table 4-2. Total amount of silicate weathered (in molem2) in forested soils during the period 1970-2300 considering a constant weathering rate (0.05 mole.m-2.y-1)and a weathering rate dependent on the soil pH. In the case of the scenario C, the three terms correspond respectively to the silicate weathered during the period of dominance of the cation exchange buffer range, to the intermediate period during which deforestation proceeds and to the time span where AI(OH), dissolution is the dominant buffer reaction.

The net surface lowering exploring the scenario C is computed according to Eq 422 in three steps depending on the weathering rate: NSLowC1~ ~ o - ~ ~ o o = N S L g70-year+(50'brSiyear)+((2250-year)*brSi~~) OWC~ (Eq 4-22) where year = 2144 in deciduous stands and year = 2078 in coniferous stands. The calculated net surface lowering by 2300 in Swedish podzols depending on the type of afforestation is shown in Table 4-3.

Net Surface Lowering (rn) scenario C

scenario A

scenario 6

Deciduous stands

0.024 (0.022)

0.026 (0.022)

0.34

Coniferous stands

0.024 (0.022)

0.027 (0.022)

0.42

Table 4-3. Expected net surface lowering by 2300 for the studied scenarios depending on the forest stand assuming a weathering rate dependent on soil pH. Figures in brackets denote the expected surface lowering assuming a constant weathering rate.

56

Decreasing acid load: 2300-2406 After exhaustion of fossil fuels SO, atmospheric concentration returns very rapidly to the pre-industrial background level (1-2 months) (see section 4.1). In the developed model it is considered that SO,, as sulphate, accounts during this period for a constant acid load into the soil corresponding to the background sulfur cycling. Eq 2-6 and 4-5 describing the decay of atmospheric CO, from 2300 on and the solubility of CO, in rainwater are used to calculate the annual concentration of dissolved CO, (H,CO,(aq)) according to Eq 4-23. (Eq 4-23) H,CO,(aq)year = H,CO,( aq)2300*e-O007(year-2300) Hence, the acid load received by the soil can be calculated as indicated by Eq 424. (Eq 4-24) Hyear = a * l O3*SO,'-backg + 2*1O3*H&0,(aq)year where a = 1.33 in deforested soils (scenario C) or soils supporting deciduous stands and a = 2.13 in soils supporting coniferous stands. The evolution of the annual acid load entering the soil during this time span is shown in Figure 4-13. The soil cation exchange buffer capacity in the scenarios A and B is computed as in the previous period, by using the Eq 4-25: (Eq 4-25)

I

005,

I

Figure 4-13. Evolution of the acid load entering the soil in the scenarios A and B depending on the type of afforestation during the period of decay of CO, atmospheric concentration (2300-2406).

The variation of the soil cation exchange buffer capacity and base saturation with time within the scenarios A and B is depicted in Figures 4-14 and 4-15. The pH of the soil for the scenarios A and B is again calculated from the above mentioned Reuss' non-linear relationship (Eq 4-15). The evolution of the soil pH for the scenarios A and B during this period is depicted in Figure 4-16. Since it has been previously shown that for the scenarios A and B, only a slight increase on the silicate weathering rate can be expected as a consequence of the increased activity of hydrogen ions in the soil solution, the total amount of silicate

57

weathered during the decay period is calculated assuming the weathering rate to be constant (brSi). In the worst case scenario (C), the amount of weathered silicate is computed by using the weathering rate corresponding to the aluminum buffer range (brSiAl). The net surface lowering corresponding to this period is calculated for the scenarios A and B as was from 1970 to 2300. For the scenario C, it is computed according to Eq 4-26, assuming no formation of soil subsequent to weathering of silicate minerals (see section 4.2.2). (Eq 4-26) NSLowC2300-2406 = wSic2300-2406 The total amount of silicate weathered and the corresponding net surface lowering from 2300 to 2406 for the three explored scenarios is shown in Table 4-4.

Figure 4-14. Evolution of the soil cation exchange buffer capacity in the scenarios A and B depending on the type of afforestation during the period of decay of CO, atmospheric concentration (2300-2406)

Figure 4-16. Evolution of the soil pH in the Figure 4-15. Evolution of soil base scenarios A and B depending on the type saturation in the scenarios A and B depending on the type of afforestation during of afforestation during the period of decay of CO, atmospheric concentration the period of decay of CO, atmospheric concentration (2300-2406). (2300-2406).

58

scenarios A,B

Acid Load (molem2) Weathered Silicate (mole.m-2) Net Surface Lowering (m)

scenario C 3.0 167.5 0.4

3.0 5.3 0.007

Table 4-4. Acid load, amount of weathered silicate and net surface lowering from exhaustion of fossil fuels to the recovery of CO, background concentration for the three studied scenarios.

Constant acid load from natural sources: 2406-60.000 (Next Ice Age) The acid input to the soil-bedrock system during this time-span is assumed to be constant and in our model corresponds to 1860 atmospheric concentrations of SO, and CO,. The annual acid load is computed according to Eq 4-27 giving a result of 0.024 mo1e.m-2.y-1. (Eq 4-27) Hyear = 1.33'1 03*S0,2-backg + 2'1 03*H2C03(aq)backg The annual cation exchange buffer capacity is computed for the scenarios A and B as in the preceding periods according to Eq 4-9. The evolution of the soil cation exchange buffer capacity and soil base saturation in the most favorable scenarios (A and B) are shown in Figures 4-17 and 4-18.

12.

10.

0 5-

8-

severe conlrol

bane

CEBC 6.

..1ur.t,on

0 3.

imdema 4

Ol-

2.

0,

01

m

m

m

m

m

m

m

yes,

Figure 4-17, Evolution of soil cation exchange buffer capacity from the recovery of background atmospheric concentrations of CO, on for the scenarios A and B.

2400

2450

2500

2550

2600

2650

27M

yen

Figure 4-18. Evolution of the base saturation of the soil from the recovery of background atmospheric concentrations of CO, on for the scenarios A and B.

In these favorable scenarios, the soil pH is calculated as in previous periods (Eq 415) and its evolution with time is depicted in Figure 4-19.Table 4-5 shows the amount of

59

weathered silicate and the net surface lowering from recovery of background concentrations up to next Ice Age depending on the scenario considered. The global results of acid load, weathered silicate and net surface lowering (1970 to next Ice Age) of the developed model for each of the studied scenarios are shown in Table 4-6.

Figure 4-19. Evolution of the pH of the soil from the recovery of background atmospheric concentrations of CO, on for the scenarios A and B.

scenarios A,B

Acid Load (mo1e.m-2) Weathered Silicate (mole.m-2) Net Surface Lowering (m)

1.38*103 2.88*103 3.8

scenario C 1.38*103 9.1'1 O4 136.5

Tabley-5. Acid load received by Swedish podzols from 2406 up to next Ice Age. Amount of weathered silicate and corresponding expected surface lowering depending on the scenario explored.

scenario A

scenario

B

scenario C

__

Total Acid Load (mo1e.m-2) deciduous forests 1.46'1 03 1.49*103 coniferous forests Total Weathered Silicate (mole.m-2) 2.9'103 Total Net Surface Lowering (m) 3.8 Soil buffer range by the cation recovery natural levels Cop,SO2 exchange

1.52*103

I .60*103 2.9*103 3.8 cation exchange

1.79*103 2.02*103 9.147 O4 137.1 AVH)3

Table 4-6. Total acid load received by forested podzols from 1970 to next Ice Age depending on the explored scenario. Estimated amount of silicate weathered, expected

60 surface lowering and soil dominant buffer reaction after recovery of background concentrations of atmospheric SO, and CO, .

5. SUMMARY AND CONCLUSIONS As mentioned above, there are two ways in which environmental acidification could have an impact on the planned HLNW repositories. In this report we have focussed on the effects on the geological stability of the bedrock. A central scenario related to the combustion of fossil fuels for energy is explored. Sulfur dioxide emissions resulting from the use of these fuels are considered to be responsible for the acidification processes. It is also taken into account the expected global build-up of CO, concentration in the atmosphere. Three different sub-scenarios are studied depending on how stringent SO, emission controls might be during next centuries: mild control (scenario A), moderate control (scenario B) or severe control (scenario C). A linear increase of varying slope is thought to be the most realistic approach to model the evolution of the atmospheric SO, and CO, concentrations. For the scenarios A and 8 the depth of weathered bedrock is calculated by using the silicate weathering rate corresponding to the cation exchange buffer range. In the worst case scenario (C), the amount of weathered host rock is computed by using the silicate weathering rate corresponding to the aluminum buffer range. The main conclusions of the model calculations performed along the previous chapter are: - Due to the short residence time in the atmosphere of SO,, the time required to restore atmospheric background concentrations is short. This means that any global reduction on the emissions has an immediate effect on the various reservoirs. The behavior of CO, is rather different due to its longer residence time. The appropriate reduction measures would have effect only one century later. - The average atmospheric concentration for the scenario C is respectively 7 and 2.5 times larger than for scenarios A and B. However, the resulting acid toad introduced into the soil up to next Ice Age is only a 20% larger in the worst case scenario (C). This is mainly due to the fact that the anthropogenic perturbation lasts for only 0.5% of the total time covered in these calculations. - There is a well defined threshold for soil acidification, when the cation exchange capacity is depleted. In our calculations this would only happen in the worst case scenario (C). However, for the average case scenario (the one actually operating in Scandinavia), the degree of acidification has a dependence on the type of afforestation. At the end of the "fossil fuel age" (2300), for deciduous forests the pH of the soil is expected to decrease to regional average values around 4.2. In the case of soils under coniferous stands, the pH is expected to decrease down to 4.05. What is more important, the remaining cation exchange buffer capacity ranges from 5% for deciduous forests to only 1Yofor the coniferous ones. These are extremely low values for a buffer capacity to be effective and irreversible damage could be expected even in the average case scenario, particularly for podzolic

61

soils supporting coniferous forests. As a matter of fact, larger sensitivity to acidification has been observed in coniferous forests (12). - In the scenarios A and B the total weathering is only slightly affected by the differences in soil acidity. Over the acid load threshold value (total exhaustion of the cation exchange buffer capacity), the weathering rates are largely increased and so it is the resulting total weathering (case C). This a consequence of a major disruption in the soil structure and the larger availability of wetted surface. - The net surface lowering resulting from scenarios A and B is only around 4 meters up to next Ice Age. This amounts only to 1% of the depth of the repository (500 meters). In the worst case scenario, the computed net surface lowering amounts to some 140 meters (between 25 and 30% of the total repository depth). As we have already discussed, this is the result of a large acidification, extensive deforestation and consequently increased weathering rates. The net result of this scenario would be that the Scandinavian ecosystem would become hardly habitable and the performance of the HLNW repository a lesser problem for the biosphere. However, the increased weathering results in general in a larger penetration of the acidified surface waters into the undisturbed groundwater system. This could possibly have larger implications in the performance of the HLNW repository. These implications will be discussed in a forthcoming report in this series. The main consequence of this study is that an expanded usage of fossil fuels without the pertinent emissions control could affect the already stressed Swedish ecosystem. This has to be kept in mind when balancing different energy alternatives and their related cycles.

6. REFERENCES 1 Ahlbom K, Aikas T, Eriksson, LO. SKBnVO Ice Age Scenario.1990;

SKB Technical report. Semb A. Atmos. Environ. 1978; 12: 455-460. Graedel TE, Crutzen PJ. Scient. Am. 1989;261:(3), 28-36. Gibbons JH, Blair PD, Gwin HL. Scient. Am. 1989; 261 :(3), 86-93. Alcamo J, Shaw R , Hordijk L. (eds.) The rains model of acidification.1989; Kluwer Academic Publishers, Dordrecht. 6 Schneider SH. Scient. Am. 1989;261:(3), 38-47 7 Raiswell RW, Brimblecombe P, Dent DL, Liss PS. Quimica Ambiental. Ediciones Omega S.A., Barcelona:1983. 8 Wayne RP. Chemistry of Atmospheres. Oxford University Press, New York:l985 9 Sekihara K. Possible climatic changes from carbon dioxide increase in the atmosphere. In: J. O'M. Bockris (ed.) Environmental Chemistry. Plenum Press, 1. New York: 1 9 7 7 ; ~285-31 . 10 Rodhe H, Rood MJ. Nature 1986;. 321: 762-764. 11 Jacobson JS. Experimental studies on the phytotoxicity of acidic precipitation .In:T.C. Hutchinson, M. Havas (eds.). NATO Conference on effects of acid precipitation on vegetation and soils .Toronto: 1978;.Plenum Press, p. 151-160.

2 3 4 5

62

12 Johnson AH, Siccama TG. Environ. Sci. Technol. 1983;17:(7), 294-305. 13 McLeod AR, Holland MR, Shaw PJA, Sutherland PM, Darrall MN, Skeffington RA. Nature 347, 277-279. 14 Wolt JD. Effects of acidic deposition on the chemical form and bioavailability of soil aluminum and manganese. In: A. Lucier and S.Haynes (eds.) Mechanisms of forest response to acidic deposition. Springer-Verlag, New York:l990 15 Tomlison II GH. Environ. Sci. Technol 1983;17:(6), 246-256. 16 Makela A, Schopp W. Regional-scale SO2 forest impact calculations. In: J. Alcamo, R. Shaw and L. Hordijk (eds.) The Rains model of acidification. Kluwer Academic Publishers, Dordrecht, 1989; p.263- 296. 17 Jacks G, Knutsson G., Maxe L., Fylkner A. Effect of acid rain on soil and groundwater in Sweden. In: 6. Yaron, G. Dagan and J. Goldshmid eds. Pollutants in Porous Media. Springer-Verlag,l984 ; p. 94-1 14. 18 Stumm W, Wollast R. Coordination chemistry of the weathering. Reviews of Geophysics, 1990. 19 Paces TJ. Geol SOCLondon:1986; 143: 673-677. 20 Slaymaker 0. Slope erosion and mass movement in to weathering in geochemical cycles. In: A. Lerman and M. Meybeck (eds.) Physical and chemical weathering in geochemical cycles.1988; NATO AS1 Series. 21 White GN, Feldman SB, Zelazny LW. Nutrient release by mineral weathering. In: A. Lucier and S. Haynes (eds.) Mechanisms of forest response to acidic deposition. Springer-Verlag, New York. 1990 22 Owens LB, Watson JP. Soil Sci. SOC.Am. J. 1979a; 43,160-166. 23 Owens LB, Watson JP. Geology 1979b; 7, 281-284. 24 Barth TFW. Geochim. Cosmochim. Acta 1961; 23, 1-8. 25 Johson NM, Reynolds RC. Science 1972; 117, 514-516. 26 Schnoor JL. Kinetics of chemical weathering: a comparison of laboratory and field weathering rates. In: W. Stumm (ed.) Aquatic chemical kinetics. Wiley & sons Inc. 1990 27 Bricard J. Aerosol production in the atmosphere. In: J. O'M. Bockris (ed.) Environmental Chemistry. Plenum Press, New York: 1977; p. 313-330. 28 Rodhe H, Granat L, Soderlund R. Report CM-64,Dept. of Meteorology of University of Stockholm and International Meteorological Institute of Stockholm.1984a 29 Rodhe H, Granat L. Atmos Environ 1984b; 18 (12), p. 2627-2639. 30 Soderlund R, Granat L. Report CM-54, Dept. of Meteorology of Univ. of Stockholm and lntl. Meteorological Institute of Stockholm.1981; 31 Tanke M, Gulik J v The global climate (Atlas). Mirage Publishing, Amsterdam: 1989 32 Hordijk L, Shaw R, Alcamo J. Background to acidification in Europe. In: J. Alcamo R. Shaw and L. Hordijk (eds.). The Rains model of acidification. Kluwer Academic Publishers, Dordrecht: 1989; p. 31-60. 33 Likens GE, Wright RF, Galloway JN, Butler TJ. Acid rain. Scient. Am. 1979; 241(4), 39-47. 34 Kauppi P, Alcamo J. Modelling soil acidification in Europe. In: J. Alcamo, R. Shaw and L. Hordijk (eds.) The Rains model of acidification. Kluwer Academic Publishers, Dordrecht, 1989; p. 179-221. 35 Kauppi P, Alcamo J. Linkages in the Rains model. In: J. Alcamo, RShaw and L. Hordijk (eds.) The Rains model of acidification. Kluwer Academic Publishers, Dordrecht, 1989; p. 297-317.

63

36 Johson AH, Sigg L. Acidity of rain and fog: conceptual definitions and practical measurements. Chimia 1985; 39, 59-61 37 Berden M, lngvar Nilsson S, Rosen K, Tyler G. Soil Acidification,extent, causes and consequences. Report 3292, National Swedish Environment Protection Board.1987. 38 Mayer R , Ulrich B. Input to soil, especially the influence of vegetation in intercepting and modifying inputs - A review. In: T.C. Hutchinson, M. Havas (eds.). NATO Conference on effects of acid precipitation on vegetation and soils, Toronto, 1978;. Plenum Press, p. 173-182. 39 Lerman A. Weathering rates and major transport processes. An introduction. In: A. Lerman and M. Meybeck (eds.) Physical and chemical weathering in geochemical cycles. NATO AS1 Series. Kluwer Academic Publishers, Dordrecht,l988; p. 1-10, 40 Wright RF. Influence of acid rain on weathering rates. In: A. Lerman and M. Meybeck (eds.) Physical and chemical weathering in geochemical cycles. NATO AS1 Series. Kluwer Academic Publishers, Dordrecht,l988; p. 181-196. 41 Reuss JO. Implications of the calcium-aluminum exchange system for the effect of acid precipitation on soils. J. Environ. Qual.1983; 12:(4), 591-595.

7 . ACKNOWLEDGEMENTS This work has been financed by SKB (Swedish Nuclear Fuel and Waste Management). We are very indebted to Dr. Peter Wikberg for his encouragement and support.

64

APPENDIX A. Selected stations to calculate the average total sulphate content in precipitation in Central and Southern Sweden. Median excess sulphate and sodium concentrations (pmole.dm-3) (Rodhe et al., 1984a and Soderlund et al, 1981). STATION

STATION

s042-

Na+

NUMBER

NAME

median value

median velw

98 1

20 6

16

KVARNTOR

21

FIAHULT

65 7

23 5

23

PLONNiNG

1090

79 6

32

SKURUP

1200

55 1

38

GOTEBORG

160 0

131 0

39

STOCKHOLM

1760

33 4

40

BOHUSMAL

97 2

336 0

42

RYDAKUNG

76 6

124

43

UPPSAIAN

96 8

193

44

KLUNKHYT

92 1

174

46

FARNABRU

91 7

129 14 1

47

BJORSUND

83 4

48

TARNA

90 0

14 5

51

AS

76 6

24 6

80

SJOANGEN

73 0

170

124

ARUP

98 I

23 5

126

GRANAN

81 6

28 0

128

KOMOSSE

72 8

42 7

131

SODERARM

1160

1250

133

GRIMSO

77 4

11 1

134

KALIANDS

69 5

150

135

AKERSHUS

84 7

23 3

1%

TORSO

66 2

10 1

137

VASE

83 0

18 9

65

ON CHARACTER OF EJECTION OF RADIONUCLIDES OUT OF EARTH SURFACE l.V.Melikhova, i.Vukovikb and V.Sipkah 'Lomonosov State University, Moskow bInstitute of Nuclear Sciences VinEa, Belgrade SUMMARY It has been established that the concentration of radionuclides at each point of atmosphere varies deviating from mean values in an order of magnitude or even more [l-41. Concentration variance are so significant that a problem of establishing the regularities of this change is recognized. In order to ascertain those regularities the results of many years rneasurernents of radioactivity of the ground layer of the air, carried out in The Institute of Nuclear Sciences in VinCa near Belgrade, have been analyzed. Those measurements which were done using standard methodology [S], characterize a particular region of Serbia, but may also be typical for any other region. The analysis has shown that the change of content of radionuclide in the ground layer of the air represents random ergodic process characterized by the probability of appearance of different meaning of concentration which change in accordance to the Fokker-Planck equation. Concentration of radionuclides has been changing impulsively showing minor dependance upon meteorological conditions, probably due to the seismic variations of radionuclides ejection froni the earth surface.

CHARACTER OF THE USED DATA The results (data) of radioactivity of the ground layer of the air were analyzed during the period of 1963-1991, at three checkpoints, located in a line towards dominating wind direction. Distance between first and second point was 2 km and between second and third point S km. At each checkpoint ground air was constantly filtered through a tube of 100 cm2 cross section at a height of I m froni the ground totalling 550 m2/24hr. The air was filtered through paper filters which retained more than 80% of radionuclides in molecular or aerosol state [5]. Filter radioactivity was measured either constantly or periodically. During measurements in situ filter was nioved at constant speed through a camera of proportional 2 7r counter [ 5 ] . During periodical ineasurements filter was immobile. Filter was measured at proportional 27r counter or spectrometer with NaI(T1) detector [6]. Each filter was measured several times in order to carry out measurement of contribution of short-lived (with disintegration constant (A > 2 10 sec-') and long-lived radionuclides. Filter radioactivity was compared with standards of'"K, '"Sr and '"U of appropriate activity. During measurements in situ instantaneous concentration was determined according to formula C = ( v x .At,,)-' where: v=500 m3/24h/,x=0.8 efficiency of retain of radionuclides, At<, time of filter presence in measurement chamber, +-radioactivity of filter. Part of the results of air radioactivity ineasureiiients during 1963-1991 period are shown in Figures 1-2 and Tables 1-4.

-

+

+

66 The average months concentration B(AtJ(Bq/m3)on the measured site. The results are given as some monthly average for continuous measurement in the period of 1963 t o 1975 y. (At =30x12 days)

Table 1.

Month

B(At)

Coefficient of variation

I

18.2+0,2

0,05

II

15.6 +0,5

0.11

Ill

17.4+0,4

0.07

IV

17,8+0,6

0.11

V

18.5k0.4

0.08

VI

18,O+0,7

0,13

VII

20.4+ 0.7

0,ll

Vlll

28,O +0,7

0,08

IX

28,3+0,9

0.11

X

35,3+ 2.0

0,19

XI

26.4 +0,8

0,lO

XI1

20,2 *0,5

0,08

Table 2.

I

Season Year

The average seasonal concentrations B(At) long lived radionuclides (mBq/m3) on the measured site II (At = 3 months) Winter

Spring

Summer

Autumn

1982

0.60

0,70

0,80

0,80

1983

0,40

0,50

0.65

0.66

1984

0,90

0,80

0,46

0,60

1985

0.67

0.67

0,58

0,95

1986

0,87

0.88

6,65

1,62

1987

1,50

1 ,oo

1,25

1,20

1988

0,90

0,70

0,80

1 ,oo

~

~

67

A.BqIm

3

I

I

I

I

1

2

3 t , days

Figure 1 . In sity air radioactivity Measurement start 21.1.1974.Impulses time A-6 min, B-7 h, '2-21 h, D-46 h.

1

3

Bq I m 40

- 1

---2

30 20

10

0

4

12

20

28

t , days

Figure 2. Succesive average daily radioactivity concentration in air ( 1 ) checkpoint I , measurement in situ, start 01.01.75. (2) checkpoint II, periodical measurement of daily samples, start 01.1 1.89. for long lived radionuclides.

68 The average years concentration B(At)(Bq/m3)of short lived radionuclides on the measured site I on the results of continuous measurement (At = 1 year)

Table 3.

Table 4.

Year

B(At)

Year

B(At)

1963

18.3

1969

23,l

1964

18,7

1970

15,3

1965

25,9

1971

20,2

1966

24,6

1972

20,5

1967

27,4

1973

24.9

1968

27,O

1974

16.9

The average years concentrations B(At)(m3q/m3)of long lived radionuclides ( A t = l year)

Measured site Year

I

II

Ill

1982

0.60

0,72

0.57

1983

0,60

0.55

0,61

1984

0.66

0,69

0,76

1985

0.67

0.71

0.74

1986

3.08

2.50

1,93

1987

1.08

1.24

1.16

1988

0,84

0,85

0,79

1989

0,60

0.67

0,66

1990

0,651

0,71

0,72

DATA PROCESSING PROCEDURE During the analyses of measurement results each increase and following decrease of C concentration was treated as an impulse of radionuclides concentration. For short duration impulse w e took each increase and decrease of C between the nearest adequate minimums, as shown on Fig.2. We considered as medium duration impulses those with the increase and decrease between deeper relative minimums, located at smooth curve, as shown on Fugures 2 and 3.Each impulse was characterized by time interval ramong the moments of minimum K

appearance and "instantaneous" amplitude C , in accordance to the condition

C = i =I

where

Ci

K is general number of impulses appearing in given moment (of time). Also, medium

69

n

I.oJ

0-4

+-2 11-3 0-4

1.5-

0-5 0-6

--7

V-8

1 % . 1

0

2

3

4

X

Figure 3. Integral distribution of amplitude and duration of impulses Measurement in situ at checkpoint I from 01.02.1963. t o 31.03.1963. (1-3):

Duration of impulses r(h):0,1-2,0 (1,4); 2-1 4(2,5) and 1 4 - 3 0

(3.6). (7~-calculationaccording t o formula (4). (8)"Sr in troposphere according to [91. p = 1 , W,=O,88,

A , = 1.91.

chit

amplitude of impulse

Ai

Ci dt

= 5-l

was calculated, where th represented the

Ch

moment of beginning of the impulse. /all impulses appearing during given rather long period 8 were treated as totality disregarding the fact whether they lean at each other or follow each other. In such totality number N(r)of impulses was calculated whose duration was less at arbitrary amplitude, and number N(A) of impulses where mean amplitude was less than

70

Ai at arbitrary duration or at duration within given interval r close t o arbitrary meaning. Periods during which the change of C concentration remained within the confidential interval levels were treated as periods of stationary background and all remaining time - as periods of higher radioactivity. During higher radioactivity periods amplitudes of impulses which appeared successively every i impulse and next (i+ 1) impulse of the same type appearing at certain measurement checkpoint during control period, were compared. Then, fraction n(AA) of impulses where difference of consecutive impulses was less AA, then density of probability (p(AA) = dn(AA)/d(AA) of successive appearance of impulses with different amplitudes was calculated. In order to compare impulses which appeared simultaneously at different phases w e calculated difference AA, = A,-Ai, of impulse with amplitude A, from the checkpoint I and amplitude A, of the impulse of the same type which appeared at checkpoint II or Ill. On the basis of the obtained data regarding the differences AA,, for all impulses a fraction of n(AAl) impulses was found where the difference surpassed AAl, and density of probability cp(AA,) =dn(AAl Id(AAl) deviation of amplitude of impulses which appeared simultaneously at different checkpoints. In order t o clarify the connection between values B (At) and meteorological conditions, w e analyzed all days in 1990 during which mean daily concentrations had remained within the levels, as shown on Table 5. From that group of days those ones with the stability class from A t o F, determined according t o standard methodology 17) were selected, and w e estimated time contribution, for the given classes for each interval B(AT), during their stay in atmosphere in a stable phase. Contribution of rainy days was calculated analogically, mean module

of wind speed at distance of 1 m from the earth surface and coefficient of its

variation H, during period with different mean daily concentrations B(At). Coefficient H, was calculated according to formula

H,

=

[$

(Vl/T-l)’

r2

(K-1)

, where K was number

of days during which B(At) remained within given interval;, V,-modules of mean daily wind speed for particular days, V-mean modules of wind speed in all K days. Table 5.

The period of varions stages of atmospheric stability with different concentration of long lived radionuclides in 1990 y on measured site I. A

B

C

D

E

F

0,2-0‘5

0,015

0,14

0,17

0‘4 1

0,025

0‘24

0,5-1,O

0,Ol

0.14

0,19

0,38

0,02

0,26

1,O-1.5

0,03

0,21

0.20

0.18

0.03

0,35

2.5-3.2

0.02

0,31

0,20

0,02

0

0.46

A,mBq/m’

RESULTS OF PROCESSING Concentration of radionuclides in ground layer of the air changes impulsively (Fifures 1 & 2). During 5% of impulse time they had small amplitude and practically were not noticeable, i.e. they were in stationary condition of radionuclides background. On the basis of the results obtained in situ we were able t o measure amplitude and duration most of impulses. Calculation of N(r) and N(A) based on data collected in situ for the whole measuring period (Fig.3). showed that impulses were distributed according duration and amplitude in regard t o empiric formula.

71

where &,-ISfrequency of impulses appearance, j m , and p-distribution parameters, shown in Table 5, R = 1 + A/Ai.Mean length of impulse in regard t o data obtained in situ did not show tendency of changing during monitoring. It was particularly obvious after averaging of the amplitude of impulses which used to appear in the very same months during monitoring period (Table 1). Impulses of duration r = 14-30 hr used t o appear one after the other in groups, thus mean daily concentration B(At), as a rule, kept increasing for several days, then during 1 - 4 days kept decreasing (Fig.2). Groups contained from 2 up to 10 impulses. Duration of impulses K

group

t i , (where impulse duration

tk =

r, refers t o the group of K-impulses) was not

i =1

explicitly dependant upon amplitude of each impulse and mean amplitude in a group. Distribution N ( r ) and

3

of impulses

N(%) of a group of impulses according t o duration and

mean amplitude are shown on Fig.4. They are described by formulas ( 1 ) a t frequency of appearance B , = 0 , O l h.', m = p = 2, r, = 8 0 hr, W, = 0,018. Groups of impulses appeared as a result of 24 hours measurements, both for short-lived and long-lived radionuclides (Fig.2). As far as the long-lived radionuclides were concerned the impulses did not show tendency towards seasonal changes (Table 2 ) . Concentration B(At) of the long-lived radionuclides was high for three years after Chernobyl catastrophe, and decreased afterwards t o the initial level, which was equal for all check-points (Table 4). Amplitude and duration of individual impulses did not depend upon the character of previous impulses. As a rule, in groups of impulses amplitude increased where the first members of the group were concerned and decreased with the last ones (Fig.l,2). Distribution n(AA) of successive impulses in relation to the differences of their amplitudes is shown on Fig.5. Distribution is close t o normal. Distribution of n(AA,) impulses which used t o appear simultaneously at different check-points, in relation t o the differences between their amplitudes differed from the normal one (Fig.5). Concentration dependance B(At) upon meteorological conditions are shown in Tables 4, 6, 7 and on Fig.6. Distributions n(B) of value B(At) are obtained at all three check-points for different time interval of averaging t, are shown on Fig.7.

Table 6.

The distribution parameters of impulses according characteristic

II

II

1

0,2-2

9421

0,36*0,03

1

10

1

2,0*0,05

2

2-14

8.Ok0,l

4,0?0,1

1

1

1

1,8rtO,l

3

14-30

2,5+0,1

21.7k0.3

6

0,027

1,5

1,8+0,1

72

n

1.o

0.5

1

0

X

2

Figure 4. Integral distribution function of impulses group vs amplitude and duration at checkpoint I.

(1)- H = T # ~ / ~ = N ( T J/ N 0 , 7 , = 8 3 , 1h,i0=1130 impulses in the period 1 9 6 3 - 1 9 7 5 ,

( 2 ) - H = Z / A ~ , ~ = N ( ;/~N T ), , A ~ = ~ ~z , ~ ~

in the first half of year for each year, A , = 2 7 , 5 Bq/m3 for 6-average second half of year, each year, (3)-H = WA,, n =“BIN,, monthly concentration of 13’Cs in 1966-1 9 6 7 period, No= 2 2 , A, = 0,3 mBq/m3.

All

-1

s

0

1

E

Figure 5. Appearance probability of differences for successive impulses a t checkpoint I ( 1 ) and simultaneous fluctuation of amplitudes impulses at checkpoint I and Ill ( 2 ) for long lived radionuclides, (l)-rp=cp(AA), E=AA/A,, A , = 0 , 7 0 mBq/m3, (2)-cp=cp(AA,), E = AA,/A,. A, =0,75 mBq/m3. Measured in time period 1 1 . 1 9 8 9 - 10 . 1 9 9 0 .

/

~

73

I

Figure 6. Part Q of rainy period during the day, where concentration of longlived radionuclides is B(At).

I

1.o

0.5

0

5

0 6 - 7

1.o

2.0

3.0

B/E

I

Figure 7. Distribution n(Bj for long lived radionuclides

-

B - mean value

at measured spot (1,4,5,6) checkpoint I, (2) checkpoint I I , (3) checkpoint Ill. (7) calculation by eq.(3)-(10).

14

Table 7.

The wind speed at near surface layer in periods with different concentrations B(At) of long lived radionuclides (At = 1 day)

0,2 - 0.5

1,26 f 0'50

0,40

0,5 - 1.0

1.10 f 0.46

0,42

1,0 - 1,5

1,Ol f 0,54

0,53

1.5 - 2,o

1,35

* 1.17

0,87

2.0 - 2,5

1,02 f 1.09

1,07

2.5 - 3 , O

1.00 f 0,085

0.85

> 3.0

1.31 f 1,3

1,Ol

DISCUSSION OF RESULTS Obtained results may be explained on the basis of the following models. Concentration of each radionuclide in the ground layer of the air is a result composition of concentration impulses, that reflects unhomogeneous transmission of radionuclides from the atmosphere and their pulsate ejection from the earth. In monitored region five types of impulses appear. First four types, i.e. minute Cj= l ) , hourly (J= 2), daily (j=3) and monthly ( j = 4 ) are responsible for the composition of formulas (1 1. Characteristics of those impulses are stated in Table 5 . The fifth type appears in the presence of the groups of daily impulses (Fig.2), where each group may be treated as one impulse with parameters

and r,. Each region

of ground layer of the air is characterized by its own function of distribution of impulses in accordance with amplitude and duration.

where n-is contribution of impulses whose amplitudes and duration are less than A and r respectively. Impulses not included in the groups are treated as independent ones. They can lean at each other, which do not alter their characteristics. Mean amplitude and duration of impulses change with time and it becomes noticeable in time 8 , which is longer than duration of most of impulses. Under such conditions change represents slow Markov's process, which can be described by Fokker-Planck equation [81.

where: G,-mean in relation to collective of impulses, change of A and r. D, and D, coefficients of fluctuation of those changes. Amplitude and duration of impulses change independently, so

Under condition (4) equation no.3 disintegrates into t w o relations

a t a“ at

---

During stationary distribution of impulses

-0 the solution of equation (3)- (5)

is

r,

Amplitude of impulses reduce due to radionuclides desintegration. Besides, both amplitude and duration are reduced by all factors which homogenize the air. Due to this at approximation by power

where

index

0 indicates characteristics

of

impulses

under

R A = l + k A / G A n . Inserting (7)into (6) we obtain

where

w,

= G;,/DAnk, W, = C , , T ~ / D ~ ,.

A t n, =q, =0, n,sO and q,=O

relations (6) - (8)lead to formulas

A

~ ( A =I S c p A ~ ‘ i - e x p [ - ~ A ( ~ : - n ’ - i ) ]=, n ( r ) 0

A t n, = 1 and n2= O the solution is:

standard

conditions,

16

Formulas (9) describes all data regarding radionuclides concentration in monitored region. A t n, = N(A)/&,B, n(r)= N ( r ) / D , B ,A, = G,,/A, T, = roW;lim, p = 1 -n, m = 1 -qz these formulas agree with relation (I) Fig.(3). A t n,=q,=I,

( D , , * s o / G z , ) ’ / z = 8 0 , A A j / G 1 , = 6 , WA=O, 0 1 8

formulas (8) and (9)describe appearance of the group of impulses w i t h concentration of short-lived radionuclides,and at

n,=

-1, GX,/h=O, 0 3 mBq/m3,D A , / G A , = 4 ,3 m B q / m 3

data on values B(At) a t all check-points. Thus, we may conclude that the monitored region possesses its o w n stable distribution of possibilities of impulses appearances. In accordance t o relations ( 1 143) that distribution would be

Individual impulses and their groups appear independently which is proved by distribution “A), close t o normal (Fig.5). from each check-point. Portion of impulses appear simultaneously at t w o or three check-point. It is demonstrated by distribution n(AA,) which differs from the normal one towards the increased probability of small deviation AA,, which characterize simultaneous appearance of impulses of equal amplitude a t different checkpoints. It is obvious that a part of the impulse overcomes a larger space of ground layer, including t w o or three check -points at once. Function (1 1) and stationary background Co determine value: T

5

B ( A ~ =) C,

F j ( a t ) oj

+ j=1

0

1 M

sds

A * @ j ( A , s dA )

0

where f;(At) - is a random function which determines time series of appearance of impulses of j-type. Impulses with frequency appearance of &,))At-’ follow f;(At)+l. Those impulses do not cause hesitations of value B(At) at successive measurements with equal interval At. When A t = 1 month it relates to minutes, hourly and daily impulses. Short-lived radionuclides characterize amplitude Aj of those impulses within three seasonal changes in 1 year period which could be seen from Table 1. But, value Aj changes so slowly that practically does not effect the conditions of stationariness of impulses distribution according their characteristics (features). Seasonal changes of Aj cannot be noticed in long-lived radionuclides, (Table 2 ) . A t t > 1 month fluctuations B(At), (Tables 2 and 4) basically are connected t o the appearance of the groups of daily impulses and monthly impulses. With the increase of At those fluctuations tend t o average. A t three check-points function (1 1 ) possesses identical character but different meanings of Aj. It is apparent through identity of functions D(B) which are standardized to mean meaning BlAt) at all three check points (Fig.7.).

77 REFERENCES 1 Junge CE, Air Chemistry and Radioactivity, Mainz Academic Press 1963. 2 Stein L, Radiochimica Acta 1983, 32: 163-17 1 . 3 Minato S, J.Nuclear Sci.Technology 1980, 17: 451 -469. 4 Kataoka T, at al. J.Nuclear Sci. Technology 1982, 18: 831 -836. 5 SmiljaniC R, PatiC D, Atmosphere Protection (Belgrade) 1979, 15: 29-32. 6 SmiljaniC R , Sipka V, Radiation Protection Report IBK-VinEa, Belgrade 1990. 7 Jain RK, Urvan LV, Stacey GS, Environmental Impact Analysis, New York: Van Nostarand Reinhold Comp. 1977. 8 Krasovskii AA, Phase Space and Statistic Theory of Dynamic Systems (in Russian)M, Nauka, 1974. 9 Data on Environmental Radioactivity Collected in Italy. Cornitato Nazionale Energia Nucleare, B10/26/62, 1962, B10/03/63, 1963.

This Page Intentionally Left Blank

CHAPTER 3 contamination from mining uses

This Page Intentionally Left Blank

19

Trace elements dynamics i n soils and aquifers of western Switzerland. 0. Atteiaa, J.-C. VCdp, A. Parriauxb. BIATE-Pedology, Swiss Federal Institute of Technology (EPF), Ecublens CH1015 Lausanne, Switzerland.

bGEOLEP, Swiss Federal Institute of Technology (EPF), Ecublens CH- 1015 Lausanne, Switzerland.

INTRODUCTION Along with increasing use of trace metals in industrial processes, many areas presently show abnormally high levels of these trace elements. This contamination, reinforced by the spreading of organic substances, poses a significant problem for the quality of drinking water. A better understanding of trace elements transfer in natural media is needed to manage water resources. The soil, and to a larger extent the pedosphere, supports and regulates many abiotic and biotic processes. By acting as a dynamic geoderma or skin of the earth, it protects, for example, the lithosphere from impacts of man’s activities. The chemical composition of rain is altered when it comes in contact with the vegetation cover, the soil surface, and percolates through the body of the soil cover. Soil represents a high capacity buffer medium: due to its porosity, to the presence of mineral and organic components, the pedosphere serves as source, filter and sink of fluxes of matter during atmosphere-pedosphere-hydrosphere interaction. Trace elements studies in earth sciences mainly originate in mining. In this field, trace elements are concentrated in specific minerals or rocks. Interactions between these minerals and solutions have been studied by geochemistry [ll. Soil geochemistry is well studied concerning major elements [21, but few papers deal with the geochemistry of trace elements in soils [3,41. On the other hand, soil scientist began trace elements studies with the advent of soil contamination and acid rain studies [5,6]. Numerous papers during the last two decades deal with pollutants adsorption equilibrium on clay o r organic matter [71. In biogeochemical studies of particular element, mass balance studies try to attach fluxes to biological cycling or weathering [8,9,10]. Some of these studies have proven t o be very useful to understand the biogeochemical cycles, but it remains difficult to have access to the real occurring processes. The key soil and aquifer properties such as pH, organic matter content, clay minerals content, oxidation and reduction processes are pre-eminent factors controlling the chemical behaviour of metals in ecosystems. The soil cover controls the concentrations of metal ions and complexes in the soil solution and

80

thus exerts a major influence on the quality of deep water. The behaviour of trace elements in natural media have been described using the atomic characteristics of the elements and the Goldschmidt rules. Taking into account the actual soil fbnctioning could be helpful in continuing this first classification. In this paper we try to specify some limits of trace elements fluxes in natural, low contaminated, ecosystems. This is done by comparing fluxes and concentrations of elements in rain, snow, soil solution and aquifer outlet. We first describe trace element behaviour at the general level, for the whole ecosystem. The role of soil vs. that of the aquifer, which is a n important task in hydrogeology, is defined using the behaviour of some trace elements. Taking in account concentrations of elements in soils and soil solutions, it is possible to precise the role of some phenomena involved in trace elements fluxes, mainly organic matter adsorption, biological cycling, weathering, soil acidification, etc.

SITES, MATERIALS AND RlETHODS Sites: A network studying aquifer typology is presently implemented in western Switzerland. This network includes a long term survey of 17 sites on a transect representing the principal alpine lithologies [ill. It is completed by a short term network analysing the chemical composition of numerous springs on some specific lithologies. The soil types existing in the network vary significantly; major types are typical of temperate climate zones [121: calcic cambisol, acid cambisol, Podzol (FA0 soil classification). For the soil study, we selected 7 sites in the whole network (fig.1).

Figure 1.Location of the studied sites.

81

The selection of the site is described in another paper [131. Here, we present the results of the three most important sites: Jura, Lutry and Argentihres. The Jura site is located in the Swiss Jura and concerns two calcic cambisol, one under forest and the other under extensive pasture; this site is called JUR. These soils are both thin with high organic matter content and nearly neutral pH. The Lutry site stands 20 km north of the Lausanne city. It is a small watershed under spruce forest with a deep acid cambisol lying on a fairly homogeneous sandy burdigalian molasse; this site will be referred as LRY. The Argentihres site is in France near Chamonix a t 2000 m altitude. The soil is totally covered by rhododendron and few short larches are spread near the site. This soil is a typical podzol developed on a recent gneiss moraine 1141; this site is referred to as ARG. Main site characteristics are summarised in Table 1. Table 1. Main characteristics of the sites. Altitude: altitude of the soil station/ spring: altitude of the spring/ Prec.: precipitation amount/ Int.: interception by forest canopy/ ET: Evapotranspiration/ size: estimated size of the watershed. Site

Altitude Spring Prec.

Int.

ET

Size Underlyingrock

m

m

mdy mdy mdy

LRY

885

880

1200

150

420

ARG JUR

2000

-1200

1700

0

350

?

Gneissicmoraine

1330

970

1600

300

250

-40

MalmLimestone

km2

0.02 Burdigalianmolasse

Soil survey is used to test the homogeneity of the watershed and to define the number of sites to be installed on each watershed according to the major soil units. Only JUR is equipped with two sites, one under forest and another under pasture, these land uses corresponding statistically to different soil types [ 151. Hydrological limits of the watersheds are appreciated with different techniques like structural analysis, tracers and isotopes [16]. The behaviour of trace elements from atmosphere to soil and spring is analysed on the three main sites only. Chemistry of the hundreds of springs of the short term network is used to test the validity of our results.

Material: To collect soil solution, two types of lysimeters (tension or free tension one) have been placed at the contact between the soil and geological layers [131. In the ARG and LRY soils, another lysimeter series is installed below the A2 and B1 layer respectively. The choice of the lysimeter type is made according to soil texture and structure. Only the LRY site is equipped with porous cups. On this

82

site tensiometric measurement allow the determination of the suction to apply to porous cups and the validation of a simple hydraulic model 1171. Chemistry of the extracts obtained with porous cups are compared using two types of lysimeters [ 151. Water balance has been verified using cumulative chloride fluxes analysis in rainfall and soil solution 1151. Bulk precipitation is collected with polyethylene (PE) f h n e l s and throughfall under forest is obtained with three PE gutters. All the materials have been tested and proved no trace elements contamination. Prior t o installation, all materials were rinsed with nitric acid and deionized water. After every sample collection, collectors are washed with distilled water. Spring water is collected directly from springs in PE bottles.

Analysis: Collected water (rain, snow, soil solutions and spring water) is analysed for anions, major cations and trace elements. Before analysis all the samples were filtered at 0.45 pm. Samples are partly stored at -2OOC for anion analysis and partly acidified with 0.5% of suprapure HN03 and stored at 4°C. Major cations are determined with a DCP-AES (ARL,Spectraspan 11). Sulphates are analysed with ICP (Jobin-Yvon 38) in S form and with ion chromatography (Sykam). The results agree well with a difference less than 10 %. Other anions are determined by automated colorimeter technique (Technicon).Trace elements are analysed with ICP-MS [181. For total soil content of major elements, soil is grounded and fused at 1200°C with a melter product (Sr metaborate) and then dissolved with HNO3 2%. For total trace elements content, X-ray fluorescence is performed on pastilles of grounded soil. For extractable elements we used two techniques: (i) Acetate Ammonium-EDTA (NH4EDTA) at pH 4.65 with a 1 : l O soil-solution ratio for the determination of "mobile" elements [ 191 and (ii) oxalate-oxalic acid (Tamm reactant) at pH 3 with a 1:40 soil-solution ratio for the determination of amorphous oxides [20,211. The role of the extraction procedure is analysed elsewhere [El. On the LRY soil samples, simple water extract with a 1 : l O soil to solution ratio has been tested. Fluxes calculation: Precise calculation of water fluxes in soil requires an important device in the field and good computer codes [221. In trace element budget studies fluxes are commonly calculated with precise concentrations of elements but with very poor estimates of water fluxes. Considering that the hydraulic functioning of the lysimeters and rain gauge are consistent around the year, we can consider the average concentration, weighted with volume, as a good estimate of yearly element concentration. To allow comparison of element concentrations at different levels of an ecosystem, evapotranspiration must be taken into account because it modifies concentration of elements in solution. Concentrations of the elements in the atmospheric water were calculated by removing the amount of estimated evapotranspiration from the total amount of precipitation. This calculation is also done between the two lysimeter depths when they exist. No evapotranspiration is assumed between soil bottom and spring. Evapotranspiration calculations are detailed elsewhere [151.

83

To appreciate the storing or releasing capacity of soil for each element, we calculated ratios of input concentrations divided by output ones. If this ratio is higher than 1 the element is accumulated in the system, o r budget is positive, it is the inverse when ratio is Iower than 1.Two of these indicator ratios have been calculated: from atmosphere to spring to test the whole ecosystem response and from atmosphere to soil bottom to test soil effects. Hydrological processes To be able to compare concentrations of elements at different levels of the ecosystem we have assumed that all the spring water comes from the atmosphere and passes through soil. Nevertheless this assumption is rarely entirely verified in mountainous aquifers. Figure 2 presents the major types of circulation of water existing in aquifers. The first case (1) is the one considered in our hypothesis: continuous flow between atmosphere and spring. In mountains, important zones are bare rock (2) and therefore soil effect acts on a small part of the water entering the aquifer. If soil bedrock is different from that of the aquifer (31, these two media can be totally disconnected. In aquifers, water,can flow through discontinuities, such as faults, joints and layering, (4) rather than in the whole rock. In this case water is in contact with a material eventually different from the bulk rock considered as the aquifer. Presently, it is known that variation in flows can modify the chemical characteristics of water such as pH, redox potential, temperature.

Figure 2: Diagram of the flows in soil-aquifer systems in mountain. In real conditions, it is very difficult to quantify the role of each of the previous hydrologic passes. This is why ratios of element concentrations obtained in each medium will be used as indicators, but the discussion will be focused on phenomena.

84

RESULTS Atmospheric deposition Study of atmospheric deposition of trace elements is detailed elsewhere [231. Average values are presented in Table 2;major features can be described as the following: - Bulk precipitation is fairly homogeneous over the whole region for most elements. - Only Cu, Pb, Zn and B clearly originate from human activities. - Fe, Ni, V, Rb and C r show low levels of concentrations and mainly come from earth eolian dust. - Sr and Ba show variable concentrations which are influenced by the existence of local eolian transport from outcrops surrounding the site. Table 2. Median and mean concentrations of trace elements (pgA)in rain and snow on the six sites. pgA B median 3.1 mean

V

Cr

Fe

0.3

0

3

Mn

Ni

Cu

Zn

Rb

Sr

Ba

Pb

6 0 1.9 21 0.5 2 2.6 1.8 5.6 0.45 0.3 8.04 5.63 0.34 2.75 22.3 0.37 2.45 6.8 2.17

The effect of forest on trace element concentration is described in another paper [24]. Concentrations of most heavy metals in throughfall are similar to that of bulk precipitation. Sr is enriched by roughly 30% and Ba by more than 100%; Fe and B are slightly retained by vegetation. Mn and Rb show important biological cycling, throughfall concentrations being an order of magnitude higher than those of precipitation. Hence, we won't focus on variability of atmospheric input in the following discussion,but forest effects will be considered for some elements.

Spring water composition Trace element contents of spring water are presented in fig.3 and 4 for the total network (AQUITYF') and for three different types of aquifers. Presently Evaporites of the Trias, Flyschs of the Niesen nappe, and Cristalline rocks of the Mont Blanc massif have been obtained. Similar studies on carbonates and molasse have been started; see table 3 for a description of the whole network. Strontium is considered as trace element, even if its concentration can reach 8 to 10 mgA because it has been less studied than major elements and its concentration in atmosphere and soil solution is at trace level. For other trace elements, Ba presents the highest concentration (from 10 to 120 pgA) followed by B, Fe and Li which concentrations vary from 0 to 60 MA. Among this group, Ba is more concentrated in Flysch aquifers and Li in Evaporites.

85

Table 3. Spring of the global AQUITYP network (geological substratum followed by spring symbol). LIO Alluvial gravels LUC Aquitanian CHE Malm Jura molasse MAL Delta gravels CHA Burdigalian LRY Triasic BOR molasse poz limestones Fluvio-glacial MOR Niesen Flysch LLI Helvetic Malm SAR gravels limestones Fluvio-glacial DIZ Gurnigel nappe ALL Alpine NOC gravels under THI Evaporitic Trias BLE moraine subAlpine COR Cretaceous TIL Cristalline Mont BOR molasse Jura Blanc

'267 75

+AOUITY P (1 7) 4 Evaporitic (86)

Flvschs 1206)

+ Cristalline (1 10)

50

25

0k-b

Sr

0

L As

I

U

Mo

I

W

Figure 3. Composition of spring water (symbols represent the median, bars the 10 and 90% percentiles, number of springs in parentheses).

V, Co, Rb, Ni, Cu, Cr, Mn and Zn concentrations in spring water are quite low having median values lower than 1pgA, the range of variation is also quite small. Among this group, Evaporite aquifers present statistically higher Rb, Ni, Cu and Mn concentrations, the highest difference existing for Rb. In Cristalline aquifers Rb concentrations show a larger dispersion than in the general network. Concentrations of U in Cristalline massif can reach very important values (up t o 1 mgA) and become the first trace element at some places. High concentrations of W, Mo and As are also noticed in these aquifers.

86

Concentrations of Mo and W are lower than pgA in the other aquifers of the short term network. We do not present here the results of I and Br analysis, even if they exist at concentrations higher than 1 pgA, for different reasons: (i) they have not been analysed everywhere, (ii) the detection limit is higher than for other elements and (iii) their behaviour is different from the other elements as they are anions in water. Numerous other trace metals and rare earth elements have been detected and analysed in the semi-quantitative mode. Their concentrations lie between 0.01 pg/l and 1 pgA. Further analyses are being developed t o quantify the concentrations of these elements in the range of ngA. PLll

P@ 120

l8

-8-

AQUlTYP(17)

++

Evaporitic (86) 16

100

+

Flyschs (206)

80

+

Cristalline (110) 12

?

!) 29.3

23

14 10

60

a

40

2oi 0

I

Ib

lb

Jk n

Zn

Figure 4. Composition of spring water (symbols represent the median, bars the 10 and 90% percentiles, number of springs in parentheses).

Trace elements budgets at the ecosystem level We call budget at the ecosystem level the behaviour of the elements from the inlet t o the outlet of the watershed. In this first approach, soil + aquifer is considered as a big black box. We analyse the ratio, R1, of average weighted concentration of one element in rain and snow divided by its concentration in spring water. Therefore, if the ratio of the element is higher than 1,the element is accumulated in the whole ecosystem, the budget being positive; in the opposite case the element is released during the water course with a negative budget. Values of R1 for each of the sites and for the numerous springs of the short term network are reported in fig.5. We can separate 4 groups: - Mn, Zn, Cu, Pb, and to a lesser extent Fe, are accumulated in the ecosystem (R1=8 to 30).

87

- Ni, Rb, B, V and Ba concentrations are almost unchanged during the ecosystem crossing. - Sr is highly dissolved in every ecosystem. - Cr shows important differences between sites. - U, Li and Co are dissolved in some aquifers. Atmosphere I springwater

1

I.

JUR

0

ARG

.

0

.

0

*

0

0

1 .

LRY AQUITW

0

~

0

0.01

1

.

Zn Mn Fe Pb Cr Cu Rb Ni

B Ba

V

-T/

Sr Co

U

Li

Figure 5. Value of R1 ratio of trace element concentrations in meteoric water to spring water. (error bars for the global network present 10 and 90% percentiles). Difference of ratios between sites is generally important. The use of ratios can give considerable importance to slight differences in spring water concentrations if these concentrations are low. Therefore, these ratios must be used only for description of general behaviour of trace elements.

Soil effect The average concentrations of trace elements in the different sites and different media are gathered in Table 4. Major elements are presented as reference to define the whole soil behaviour. Statistically different groups obtained by an Analysis of Variance (ANOVA)are differentiated. The elements can be divided into three groups: (i) Mn, Zn and Fe that can reach concentrations greater than 50 pgA in some sites, (ii) Ba, Sr, B and Cu whose concentrations are between 1 and 20 pgA and (iii) Ni, Pb, V, Cr and Rb generally lying around 1 pgA. Among these elements Ba, Rb, Pb and Cr do not present any statistical differentiation between sites. No differentiation between

88

sites is valid for all the elements, but trace element content is generally higher at LRY and lower at ARG. Table 4. Average weighted concentrations of elements in soil solutions. Period of March 1990 to November 1991. Concentrations are weighted with volume of water. For trace elements, letters show statistically different groups of data. mg/l pH Ca Si Mg K Na Al C1 N-NO3 S-SO4 ARG 4.88 0.59 1.75 0.11 0.17 0.37 0.34 0.43 0.21 0.46 JUR 7.66 33.7 5.86 0.38 0.16 0.49 0.09 1.42 1.71 0.93 LRY 5.62 2.79 4.94 1.59 0.37 0.93 0.41 1.41 0.33 2.85 pg/l ARG JUR LRY

Mn Zn Fe Ba 4.62a 11.P 69.4b 12.la 1.09a 13.9 17.28 18.8a 158b 130b 20.6a 168

Sr B 1.85a 3.39 11.9b 9.64b 12.7b 12.4b

Cu Ni Rb Pb V Cr 1.48a 1.08a 1.08a 0.88a 0.08a 0.02a 2.42a 0 . 7 9 0.4a 0.25a 0.72b 0.238 5.68b 2.04b 0.69a 0.83a 1.11b 0.21a

Table 5 shows soil solution concentrations of some heavy metals in different ecosystem studies. Table 5. Average concentrations of some heavy metals obtained in other studies: Solling, Spanbeck (Germany): [251 apart for Ni and Cr: [81, Varsjo and Gardsjon (Sweden): [lo], soil layers 1 , 2 , 3 are respectively A layer: 15 cm, B1 layer: 35 cm and B2 layer: 55 cm. PLgfl

Solling Spanbeck Varsjol 2 'I

3

Gardsjonl " 2 "

3

Zn

Cu

Ni

Pb

Cr

520

9

15

2

0.6-1.2

540

9.6

34

4

0.9

1.2

65

2

2.2

1.7

151

1.4

4.1

0.8

38

4.1

1.8

19.5

2.3

55

0.7

2.5

2.5

1.2

75

0.7

3.6

1

1.3

4

89

Concentrations a t the Solling site are higher than those existing in each of our sites. But Solling site is known t o be highly polluted and highly acidified. The two Swedish sites present soil solution concentrations of Cu, Zn and Ni similar to LRY levels and even slightly lower. These concentrations are higher than the one existing in JUR and ARG. Concerning Pb and Cr, concentrations in all our sites are lower than those of Swedish soils. For Cu,KeZZer [261 finds results lying between LRY and JUR values for one cambisol and one podzol in the same region of Switzerland. The studied sites can be qualified as slightly contaminated in heavy metals for ARG and JUR whereas LRY seems t o show intermediate levels. Soil effects on atmospheric water composition is tested by using R2, i.e. the ratio of concentrations in soil solution to those in atmospheric water. This ratio has been plotted in fig.6 with the same representation as for R1 in the previous section. Values of R2 are spread on both sides of 1 for many elements, the exceptions being Cu and Pb which are retained in all soils and Ba, and t o a lesser extent Sr, Ni and Fe, which are released by all soils. Soil solution I Atmosphere 0

0

8

0

-.-

0

0

0

~

w

H

'

H I

0 -

I

Pb Cu Cr Mn Zn V

B Rb Sr Ni Fe Ba

Figure 6 . Value of R2 ratio of trace element concentrations in meteoric water t o soil solution. (Mn and Rb ratios have been calculated using concentrations in precipitation to exclude biological recycling effect). Some elements (V, B) present R2 values close to 1 on all sites. This behaviour suggests that soil has no effect on the concentrations of these elements. By contrast, values of R2 for Fe, Mn and Zn show important differentiation between soil types. Therefore, few trace elements found in soil solutions are influenced by pedogenetic processes.

90

Elements concentration on their way through the ecosystem

6 ,4 Q,

32 0

-

.4 Q,

='2

0 1251 - 100 75

4

.

.

-

Q,

2 2

=L

0

-

- - I

Ni

BP

Tf

A2/B1 Bs/B2 Sp

50

25 0 125 1004

Zn

:75-1

BP

Tf

A2/B1 BdB2 Sp

Figure 7. Average concentrations (pg/l) of trace elements in bulk precipitation (BP), Throughfall (TO, soil solution (A2,Bs, B1, B2) and spring (Sp) for the three main sites.

91

Fig.7 describes the concentrations of each element for the different levels of the ecosystem (atmosphere, soil and spring). Used concentrations are the average concentrations weighted by volume and corrected for evapotranspiration or interception effects as stated in the fluxes calculation section. The ARG spring composition is a n average of water coming from different fractures of one tunnel, therefore it has to be considered with care. Concentrations of Cu and Pb decrease from atmosphere to soil solution, the decrease being dramatic for Pb, concentration in soil top layers is low and conservative until spring. Copper shows smoother decrease in each compartment. For V, Ni and Rb, levels of concentration in top soil are close to atmospheric ones. Between top and deep soil layers, besides Ni concentrations that increase at the LRY site, all elements concentrations remain constant. Ba concentrations always increase between atmosphere and top soil. While this increase of concentration remains at the LRY site in deep soil, concentrations decrease a t the two other sites. Fe levels are quite stable between atmosphere and soil solution and become null in spring water. The major exception to this scheme being the A2 layer of ARG which releases a large amount of Fe compared to other sites. Nevertheless Fe concentrations remain at trace level. A t the JUR and ARG sites, Zn concentrations decrease along the water way and Mn levels are always very low. At the LRY site the scheme is totally different with a net release of Mn and Zn in the B layer. Mn pattern at LRY is different from Zn one: concentrations abruptly increase between precipitation and throughfall and Mn is retained in the top soil. This pattern is characteristic of biological cycling as cited by different authors [24,27,28]. The site showing the more specific features is LRY: not only levels but also vertical evolution of concentrations are different from the other sites. This fact is surprising because the most similar sites, JUR and ARG, are the most different ones from the pedogenesis point of view.

DISCUSSION

Trace elements budgets for the whole ecosystem One can find the sequence of ecosystem budget in different kind of studies concerning the general behaviour of trace elements [291. For instance Tardy [3] worked on the concept of mobility which is a ratio of element concentration in solution to its concentration in solid phase. Hence accumulated elements ( R b 1 ) will show a lower mobility than elements dissolved during water course. Study of solids and solutions in the Massif Central (France) leads to the following mobility sequence: Sr > Ni, Rb, V, Cu, Pb > Ba, Cr > Fe, Mn. Groups and order between groups are the same as in our study except for Cu and Pb. The difference existing for these elements can be explained by the high concentration of these elements in atmospheric deposition that was not taken in account in Tardy study. Nevertheless the validity of the previous scheme is of little practical impIication for many elements because of the important differences existing between sites (fig.3).

92

Trace elements retained by the ecosystem (Cu,Pb) Only Cu and Pb are retained by soils and aquifers. Their concentrations decrease along the entire water course while their total concentration in solid increase from the bottom to the top soil layers. This pattern is specific of elements retained in the topsoil, and has also been observed by different authors [10,251. Concerning Pb, we show elsewhere 1151 that this element is not submitted to biological cycling by forest. This is also concluded by Heinrichs [8].In both acid soils, Pb is almost totally extracted by T a m reactant (cf. Table 6). Moreover, in these soils total concentration of Pb in soil is directly related to the total organic matter content (fig.8). The difference of slope of Pb - organic matter correlation between the two sites is due to higher organic matter content at ARG site than at the LRY one, whereas total amount of Pb is similar (cf. Table 5). In soil solutions of low pH, such as in acid sites, the concentration of dissolved Pb is equal t o the one existing in the carbonated site. In other studies [301, i t can be seen that even in more acidic soils Pb is retained with organic matter in the top of mineral layers. In the studied soils Pb seem to remain bounded to organic matter and only moves when this one moves. This is also described by Tyler [311 and Berggren [32]. It is important t o study origin and movement of organic matter and particularly its relationship with hydrological factors like quick snowmelt on podzols.

60,

1

I

Regression lines

LRY Pb = 1.99 * %O.M. -3.42,

R2 = 0.97 ARG: Pb =O. 47* %O.M. + 9.91, R2 = 0.98

0

20

40 60 % Organic Matter

80

Figure 8. Relationship between total soil content of Pb and organic matter.

93

Table 6. Total soil content (ppm or mg/kg), % extractable with T a m reactant (Ox.)and Acetate-NH4-EDTA (Ac.) of some trace elements. for JUR,f: forest, p: pasture. For JUR and LRY numbers are median depths of the layer, for ARG layers are referenced by names. B: bedrock composition. Mn Ox Ac Pb Ox Ac Cr Ox Ac Ni Ox Ac V OxAc ppm % % pprn % % ppm % % ppm % % ppm % % JURf3 1010 39 19 29 5 5 110 1 0 57 14 1 137 3 0 JURfl5 1030 42 13 24 3 3 118 1 0 63 13 0 146 3 0 91 2 0 JURf25 740 45 6 8 7 4 44 18 0 103 4 0 JURp5 1120 46 0 31 8 0 111 1 0 46 23 0 141 4 0 JURpl5 1130 45 9 16 13 4 110 1 0 46 22 0 134 4 0 JURp25 1210 43 9 21 4 3 112 1 0 46 21 0 142 3 0 J U R B 110 6 26 9 4 LRY4 LRY14 LRY30 LRY5O LRY70 LRY9O LRYllO LRY130 LRY150 LRY B

440 660 940 840 750 680 620 500 790 570

ARGaO ARGal ARGe ARGbh ARGbhs ARGbs ARGB

50 110 190 250

38 53 57 50 52 38 38 29 48

15 2 0 100 340 100 420 1 420

12 4 2 3

2 1 1 1

5 8 1 0

45 30 2

46 10 4 6 5 3 4 5 1. 1.

83 42 100 41 100 51 100 19 100 16 100 15 100 10 88 7

77 100 94 106 99 118 105 121 110 122

51 3 14 20 12 12 13

21 62 100 100 100 39 100 19 100 16 72 15

18 0 0 18 1 0 9 2 0 30 14 1 32 13 2 22 5 0 26

6 5 5 3 4 3 3 3 1

2 0 0 0 0 0 0 0 0

15 100 0 20 100 7 21 98 2 26 52 1 30 54 1 36 24 0 41 18 0 42 16 0 35 14 0 55

44 55 54 5 6 5 9 59 55 55 5 1 53

10 92 73 2 100 100 0 100 100

9 11 6 28 2 1 29 4 0 92 15 1 56 8 0 41 2 0 43

n

n 6 100 17 7

10 7 7 5 5 3 3 3 2

1 0 0 0 0 0 0 0 0

Concerning Cu, Xray fluorescence detection limits (approx. 4 mg/kg) are too high to measure Cu levels in the different soils. Cu concentrations in the soil solution are similar in the soils with pH varying from 4 to 8. This behaviour could be explained by Cu strong complexing with organic matter a s cited by Berggren 1321, B u f f e [33] or KeZZer [261. The average ratio of element in soil solution (pg/l) vs. “mobile” amount (NH4EDTA in mg/l) is 0.42 for Pb and 26 for Cu which shows that Pb is much more strongly bounded to organic matter than cu.

94

Trace elements showing similar behaviour in all soils (B, V,Cr, Ni, Ba) Average B concentrations in soil solution are close to the existing one i n atmospheric deposition and almost not influenced by soil type. B and Ba concentrations, as Si ones, increase in summer and decrease in winter. This evolution can be related linearly to a drought factor on the ARG and JUR sites (fig.9) and water suction at LRY (Not shown). This drought factor can be expressed as the inverse of a rain intensity factor or in month / mm of rain. Concentrations increase when soil receives less water, this increase being higher than pure evaporation effects. Keren [34,35] showed that B concentrations depend on the soil to solution ratio and the clay type existing in the soil. Co-evolution of B and Si concentrations can be expIained by similar chemical properties: they are both metalloids, undissociated weak acids being hydroxides, at the soil solution pH. As the interlayers are generally filled with hydroxides [361, the relationship between concentration and drought could be explained by clay shrinkage which excludes B hydroxides from their interlayers.

40 -

B = 2.31*A + 4.25 R2 = .8

I

Ba = 6.75*A + 9.53,

30-

2 d 60-

I

2 20F9

.d

0

.

10

0

0 0

8 101214 A (month I mm of water) 2

4

6

0

2 4 6 8 1012 A (month I mm of water)

4

Figure 9. Relationship between B and Ba concentrations in soil solution and a drought factor (A) in month/ mm of rainfall. Concentrations of V, Rb,Cr in soil solutions are low and similar in different soils. The role of the soil is difficult to specify because of tiny differences between atmospheric and soil solution concentrations. On the LRY and ARG sites we have calculated the total amount of one element lost in the upper 80 cm (or 50 cm for ARG). Soil concentrations have been corrected for density effects by assuming that Zr is invariant. The depletion of one element is calculated as the difference between corrected soil

95

concentration and bedrock concentration. This depletion is compared t o the annual fluxes lost in soil solution at this depth. This is equivalent to the time t o deplete the soil layer with a constant flux equal to the currently measured one. For major elements, table 7 shows that this simple calculation leads to a time of the same order of magnitude as the soil age (8 t o 10000 years). This can be interpreted as a constant dissolution rate for these elements, this rate being independent of soil pH and organic matter content evolution. Table 7. Calculated time (thousands of years) to deplete the soil layer with the current flux LRY

Al K Si Mg 13.5 13.4 4.6 -

ARG

6.6 3.8 3.5

Rb

Ni

V

Cr

Fe

Ba

6.2

9

-

100

28

1.2 0.01

3.6 4.9

-

13

1.8 0.01 0.01

11.6 24

Zn

Mn 0

Calculated time of depletion for Rb, Ni and V are similar t o the one calculated for major elements. Calculation has not been done at ARG for Ni because Ni lost by weathering can be biologically recycled and adsorbed to organic matter. For V at LRY, fluxes in solution are too close to fluxes in atmospheric deposition regarding the errors made on these measurements. These three elements can be regarded as originating from weathering in soils. The fluxes are different and decrease from Ni to Rb and V. This can be related to the extractable proportion of these elements (Table 6) and to their different solubilities. The low soil content of these elements and the slow speed of the weathering process justifies the low concentrations in solution. In minerals, Rb substitute K with a ratio of 160 to 300 [l] because of its similar radius and charge [37].A t the LRY site, K/Rb equals 228 in bedrock, 180 in soil top layer and 333 in soil solution. For ARG, this ratio equals 166 in bedrock, 105 in the soil top layer and 196 in soil solution. These ratios are fairly close which shows that Rb behaviour is close to K one. The ratio decreases from bottom t o top of the soil and is higher in soil solution than in soil. This proves that Rb is more resistant to weathering than K. All ratios are higher for LRY than for ARG so the composition of soil solution from different soils reflects the difference of mineral composition. Rb concentrations or WRb values are then related to the composition of the original material and to the weathering rate. Comparison for Cr shows very long time for achieving current depletion. C r being in particulate form at this pH [381,soil solution concentrations do not indicate total annual flux. Release of Ba in soil solution has not been related to a particular chemical phenomenon though its presence in feldspars [l] can explain its high concentration in soil solutions. Similarities between B and Ba (fig.9) cannot be explained by classical chemical properties of these elements, which are very different. Ba behaviour needs to be studied more deeply.

96

Elements influenced by pedogenetic processes (Fe,Mn, Zn) Iron is an important metal in pedogenesis studies [361. In fact in this study Fe behaviour is very different among the sites (cf. fig.7). Profiles of total Fe in soil outline a depletion in the A2 layer of the Podzol and in the 50 top cm of the cambisol. These high depletions are not compatible with the current fluxes (cf. Table 7) because Fe, like Cr, is mainly in particulate form [391. Fe shows an original and well-known behaviour in Podzol [40,41,42] with a depletion in the A2 layer. Accumulation of Fe in amorphous form in the spodic layer leads t o a total content twice as high as that of the bedrock. Lateral movements of solutions in the top layers amplify this accumulation [14,15].V and Cr follow the same accumulation pattern as Fe in this soil type. Mn and Zn are only released in acid soils and particularly at LRY (Acid cambisol), concentrations at this site being 5 to 12 times higher than those a t ARG. Following the same calculation of total depletion time done in previous section, table 7 shows that current Mn and Zn fluxes are high compared t o the historical depletion. Release of Mn and Zn at the LRY site has recently increased tremendously. Moreover, fig.10 shows that concentrations of these elements in a water extract of LRY soil increase when soil pH drops below 5. Actual Mn and Zn release in soil solution seems t o be related to recent soil acidification. This behaviour is similar to that of A1 during soil acidification 1431,and has been forecasted by models of soil acidification [441, by laboratory experiments [3 11 and ecosystem studies [lo].

250

u I

Figure 10. Evolution with depth of H+, Mn and Zn concentrations in water extracts of LRY soil. At ARG, Mn and Zn concentrations are much lower than those at LRY while soil pH is slightly more acidic. The low value of Zn concentrations can be

97

explained by the almost total depletion of Zn in the A2 and Bh layers. Concerning Mn, redox potential role seems t o be insignificant because Mn concentration does not vary with the level of the water table. Moreover, ARG site is more affected by soil saturation than LRY site. Up t o 50% of Mn is extractable with Tamm reactant in LRY soil mineral layers whereas this value is less than 5% in ARG. This could be explained by the existence of Ferro-magnesians minerals a t the LRY site. SOpH is not the only factor governing Mn and Zn release; existing mineralogy plays a significant role. Fe, Mn and Zn exist at very low concentrations in spring water. This is due to the pH of the water. A t such a pH these elements precipitate in hydroxides.

Elements released by the aquifers Sr is the only trace element released from all soils and aquifers to solutions. In almost all solutions, Sr concentrations are correlated to Ca ones. In carbonated soils the Sr/Ca ratio is equal to that of the underlying rock proving that Sr originates from calcareous gravel existing in soils. Sr concentrations in these soils are lower than those of spring because equilibrium with calcite and Sr-CO3 is not reached in soils. This is due to the low contact time between calcareous gravel and soil solutions. Sr can be considered to be a major element in certain springs (concentration of several mg/l), this fact being characteristic of aquifers containing evaporites [451. These high concentrations come from the high solubility of celestine (SrS04). On other sites, Sr mainly comes from SrC03 dissolution, its behaviour being close to that of Ca. Numerous metals (V, Co, Mn, Fe, Zn, Cu, Pb, Ni) exist at low concentrations (app. 1 pgA) in springs compared to soil solutions. They are retained by different reactions: precipitation in carbonate and hydroxides forms, adsorption on clays existing in rock fissures, etc. Considering the p H of the spring water each of these reactions are likely to occur and we have no means to precise the prominent one. The relative importance of each reaction is determined by differences in reactions kinetic. The major exceptions to this scheme are the evaporite aquifers. In the spring water coming from these aquifers, concentrations of these metals are higher than 1 pgA or even than soil solution concentration. This is mainly explained by the association of these metals with highly soluble sulphur minerals lying in the aquifer. In evaporitic aquifers other traces such as Li and Rb are released in solution, these alcaline elements could be associated with halide minerals knowing their chemical structure. Behaviour of B and Ba differs highly among aquifers as they are stopped in numerous aquifers and liberated in others. B is released by evaporite and cristalline aquifer and Baby flyschs. More knowledge about the clay and feldspar content of these aquifers is required t o understand the behaviour of these elements. Cristalline aquifer is particular in the sense that trace elements existing in soils are absent in the spring water whereas totally different elements (U, Mo, W) are released by the aquifers. The conditions of rock forming can explain the high content of these elements, generally existing in reduced conditions. Cr present high concentrations at LRY spring, and considering the water pH, it can only be in Cr2072-form [39]. As in soil, Cr is in Cr3+ form [46] it has

98

undergone an oxidation between soil and spring. This phenomenon is rather surprising but it has been observed by Bartktt 1471 who associates this oxidation to Mn reduction at specific Redox potential. All elements typically released by aquifers have been found i n solutions of rocks powder mixed with CO:!acidified water. The origin of these elements can therefore be related directly to rock weathering in aquifers. Along the discussion we have compared concentrations at different levels of the ecosystem, but we have explained in methodology section that this comparison could be erroneous in mountainous aquifers. In fact cases must be differentiated: First (i), in small aquifers, such as LRY, where flow passes are known, the assumptions concerning water flow are almost verified. Second (ii), in large aquifers of highly soluble rocks, pH retains the metals existing in soil solutions and composition of the springs are quickly determined by the composition of the rocks. This is the case of calcareous and evaporite aquifers. Third (iii), in cristalline aquifers direct infiltration from atmosphere and fracture flow are important, but in that case most of the trace elements contained in spring are specific to aquifers and their origin cannot be attributed to soil. The only elements for which the roles played by soil and aquifers are not clear are Ca, Sr, B and Ba. For Ca and Sr the question is the proportion of Ca and Sr coming from calcareous soil’s gravels. We have determined a proportion of 2/3 in one site, but it vanes according to the microbial activity. Moreover, it is not possible to determine the proportion of Ca and Sr coming from bare parts of the catchment. As B and Ba can be released or stopped in aquifers, the proportion of B and Ba due to soil cannot be estimated if the proportion of the aquifer covered by soil is not known.

CONCLUSION: The first task of this study was to define the respective roles of atmosphere, soils and aquifers in the release of trace elements in springs. In a general thinking we can say that few trace elements (B) pass through the whole system without any influence. Most of the elements released in the soil are stopped in the aquifer due to their pH. New kind of mineral or chemical conditions induces the solubilisation of specific trace elements in aquifers. The liberated elements differ with lithologies of aquifers, therefore they can be used as tracers of specific rocks. This characterisation is the aim of the AQUImP project. If we look at the pollution aspect, studied springs are currently free of trace contamination. This is only valid for dissolved trace elements. Nevertheless, the three types of ecosystems will react differently if the current scenario of atmospheric inputs continues. - On carbonated rocks, all trace metals are fixed in the first meters of the aquifer, the size of the aquifer protects them against trace element’s contamination. This fact does not exclude nitrate or organic pollution - On Podzol, organic matter carries a lot of heavy metals but its fixation in spodic layers is a good trace stabiliser. A problem can arise when organic matter passes directly to spring or rivers as in the thin soil - aquifer system in Scandinavia. In this case, rivers and lakes can be highly polluted.

99

Acid cambisol soil releases high levels of pollutants (Mn, Zn) which mainly originate from soil acidification rather than atmospheric deposition. For the LRY example when the decarbonatation wave will reach the spring level, water composition will be above the law values. In similar sites, existence of preferential flow could lead to the contamination of aquifer even now or in the close hture. Concerning trace elements dynamics the most important and new point concerns B and Ba behaviour related to soil hydraulic status and apparently t o clay swelling. This original process deserves further research, particularly for Ba, which solubilisation mechanisms need to be clarified. The important features t o study trace elements dynamics are: origin of organic matter mobility, role of biological cycling, actual redox conditions in soils and aquifer and precise location of elements in the solid phase. Numerous minerals and chemical conditions found in aquifers induce important variability in the content of these elements in springs.

ACKNOWLEDGEMENTS We thank M. Bensimon for ICP-MS analysis, J.D. Dubois, Y. Mandia and P. Basabe for use of the material gathered in their PhD dissertation, and the Swiss National Research Fondation for the project financial support.

REFERENCES Wedepohl KH. Springer Verlag, Berlin-Heidelbrg-New York, 1968-1979. Drever J I (Ed). The chemistry of weathering D. Reidel publishing company (Holland), 1985. Tardy Y. MQmoires du service g6ol. d'Alsace et de Lorraine, n"31, 3 Strasbourg, 1969. 4 Mosser C. ULP Strasbourg, Institut de gbologie, MBmoire n"63, 1980, 227p. Oden S. Proceedings of the first international symposium on acid 5 precipitation and the forest ecosystem, Dochinger L.S., T.A. Seliga (Eds). USDA Forest service general technical report, 1976, 1-36. 6 Heinrichs H and R Mayer . J. of Envir. Quality, 1980; 9: 111-118. 7 Schmitt H W and Sticher H. In Metals and their componds the environment Merian E. (Ed), VCH Weinheim New York Basel Cambridge, 1990: 311-332. Heinrichs H and R Mayer. J . of Envir. Quality, 1977; 6: 402-407. 8 Turner RS. PHD, Univ. of Pennsylvania, Philadelphia PA. (Diss. Abstr. 839 160981, 1983. 10 Bergkvist B. Water, Air and Soil Poll., 1987; 33: 131-154. 11 Paniaux A, JD Dubois, Y Mandia, P Basabe and M Bensimon. Memoires of the 22nd congress of IAH,Paniaux Ed, Lausanne, 1990; : 254-262. 12 PQdroG. Revue gQogr.phys. gQol.dyn., X, 1968: 157-170. 1 2

100

13 Atteia 0,JC Vedy, A Parriaux, E Dambrine. Memoires of the 22nd congress of IAH,Parriaux Ed., Lausanne, 1990. 14 Dambrine E. PhD Diss. Univ. Paris VII, 265 p, 1985. 15 Atteia 0. EPFL Lausanne, PhD Diss. no 1031,1992,253pp. 16 Parriaux A and M Bensimon. Memoires of the 22nd congress of IAH, Parriaux Ed, Lausanne, 1990: 719-727. 17 Atteia 0.Submitted to Journal of Hydrology, 1993. 18 Bensimon M, Gabus JH,Parriaux A. J. Trace and Microprobe techniques, 1991;9: 81-93. 19 Lake DL, PWW Kirk and J N Lester. J. of Envir. Quality, 1984;13: 175-183. 20 Jeanroy E. PhD Diss. UniversitR de Nancy, 1983,157p. 21 Cavallaro N and MB McBride. Soil Sci. SOC. Am.J., 1984;48: 1050-1054. 22 DeCoursey DG. Proceedings of the NATO advanced study institute on recent advances the modeling of hydrologic systems Sintra, Portugal. Luwer Academic, 1988:35-48. 23 Atteia 0.Submitted to Atmospheric Environment, 1993. 24 Atteia 0.Annales ScientXques Forestibres. (in press). 25 Schultz. Vergleichende Betrachtung des schwertemetallhaushalts verschiedener waldokosystems, Waldsterben busgenweg 2,3400 Gottingen, 1987,245~. 26 Keller C. PhD Diss. EPFL Lausanne, 1991,171p 27 Hofken KD. Effects of Accumulation of air pollutants forest ecosystems B Ulrich, J Pankrrath Ed., 1986: 57-64. 28 Godt J , Schmidt M, R Mayer. Atmospheric pollutants forest areas, Georgii H.W. Ed., Reidel, 1986:263-274. 29 PBdro G and AB Delmas. Ann. Agron. 1970;21: 483-518. 30 Bourg ACM and J-C Vedy . Geoderma, 1986;38: 279-292. 31 Tyler G.Water, Air and Soil Poll., 1981;15:353-369. 32 Bergrren D.Intern. J. Environ. Anal.Chem., 1989; 35: 1-24. 33 Buffle J and RS Altmann . Aquatic surface chemistry Stumm W. (Ed). Wiley, New York, 1987:351-383. 34 Keren R and H Tdpaz. Soil Sci. Soc. Am.J., 1984;48:555-559. Am. J., 1981;45:478-482. 35 Keren R and RG Gast. Soil Sci. SOC. 36 Duchaufour P. PBdologie, tome 2: constituants et propriBt6s des sols. Masson Ed. Paris, 1983. 37 Mahan BH. Chimie, InterEditions, Paris, S.A., 1977. 38 Bergback B, S Anderberg and U Lohm. Water, Air and Soil Poll., 1989;48: 391-407. 39 Stumm W and JJ Morgan. John Wiley, 1970. 40 Tonkonogov VD,BP Gradusov, NY Rubilina et al. Pochvovedeniye, 1987;3: 68-81. 41 Sanborn Pand LM Lavkulich. Soil Sci. Soc. Am.J. 1989;53: 517-526. 42 Farmer VG. Soil. Sci. Plant Nutr. 1982;28: 571-578. 43 Driscoll CT,N Van Breemen, J D Mulder. Soil Sci. SOC. Am. J., 1985;49: 437-444.

101

44 De Vries W, M Posch and J Kamari. Water, Air and Soil Poll., 1989; 48: 349390. 45 Mandia Y PhD Diss, Lausanne EPFL, 1992,318 p. 46 Bartlett RJ and JM Kimble. J. of Envir. Quality, 1976a; 5: 379-383. 47 Bartlett RJ and B James. J. of Envir. Quality, 1979; 8: 31-34.

This Page Intentionally Left Blank

103

Leaching behaviour of granulated non-ferrous metal slags F.M.G. Tack, P.H. Masscheleyn & M.G. Verloo Laboratory for Analytical Chemistry and Agrochemistry, University of Gent, Coupure Links 653, B-9000 Gent, Belgium ABSTRACT The leaching behaviour of selected heavy metals from granulated non-ferrous metal slags was investigated. Experimental results and sorptiorddesorption data of selected soils were used to estimate metal migration from a hypothetical slag disposal site into the underlying soil. In a heavy textured soil, metals leaching from the slag site were predicted to accumulate mainly in the upper 10 cm. I n more sandy soils, the metals were dispersed over larger soil depths, resulting in lower accumulated levels. When comparing the modelled migration results with soil and soil water quality standards, only the migration of Zn was of practical importance and may limit reuse of the slags in bulk form. INTRODUCTION Metal slags are a by-product of non-ferrous metal processing and refining. These slags consist of an inert matrix, in which residual elements are included that can no longer be extracted in an economically feasible way. As opposed to slags resulting from iron and steel production, those from non-ferrous metallurgical industries do contain, besides high amounts of Fe and Ca, relatively high levels of several toxic elements (e.g. Cd, Cu, Pb, Zn, As, Ni, Cr) as compared to levels found in the environment. Formerly, those slags were produced as coarse blocks appearing like rocks. They were widely used to reinforce shores and as building material in dike construction. In the early eighties, new extraction technologies became available resulting in the production of slags with lower residual amounts of non-ferrous metals. Those slags, however, are granulated and other ways of reuse had to be considered. Because of their high bulk density, good and quick settling and resistance against frost, the granulated non-ferrous metal slags were found to be an excellent material for use in road construction. They were found suited to replace conventional materials in drains, foundations and capillary screens. Due to increased environmental awareness, investigations of possibilities for reuse also had to consider environmental aspects. Despite of the lower total toxic trace element content as compared to the old slags, the leachability of elements from the granulated slags is expected to be higher due to the finely granulated structure and resulting larger specific surface area. As slags may periodically or permanently be percolated by drain water or ground water, this aspect is critical when using the slags in road construction. Using leaching tests we assessed the leaching behaviour of selected elements including Cd, Zn, Pb, Cu, Cr, Ni and As from the granulated non-ferrous metal slags. The results were modelled to predict migration of metals from disposed slags

104

with time. Based on model estimations, the environmental restrictions of reusing the slags in bulk form were evaluated.

LEACHING BEHAVIOUR

Introduction To assess the leaching behaviour of the granulated non-ferrous metal slags we used the Standard Leaching Test of the Netherland Energy Research Center (SOSUV-test) [l]. This test, originally developed for investigation of combustion residues, allows to evaluate the leaching behaviour at different WS-ratios (Liquid to Solid ratios). The US-ratio is used as a relative time scale. Relating the relative time scale to a real time scale of a particular field situation allows for short, medium and long term leaching behaviour to be estimated. The velocity by which a particular US-ratio is reached depends on the porosity of the material and on the rate of advection of fresh leaching medium. A column percolation test is run to test leaching behaviour at cumulative USratios u p to 10. For higher cumulative L/S-ratios (20-loo), a batch equilibration cascade test is used. The advantage as compared to the column test is that high WSratios can be reached in shorter time spans. Distilled water acidified to pH 4 is proposed as standard leaching agent and simulates the most significant effects of acid rain. If a particular field situation is known, the actual leaching agent can be used instead. Materials and methods

Column percolation test Leaching agent was prepared by acidifying demineralized water with diluted ultrapure nitric acid to pH 4. A 30 cm long, 3.5 cm diameter polyethylene column was filled with 0.4 kg of the metal slags. At the bottom and top of the column a 0.45 pm membrane filter prevented fine particles from entering the leachate. The column was percolated from bottom to top at a rate of 10 mWhr. Fractions were collected up to a cumulative L/Sratio of 10. pH and conductivity were determined using standard equipment. Samples were acidified to pH 2 with nitric acid and kept in polyetylene bottles prior to analysis.

Butch equilibrufion test In 1 L polyethylene flasks, 40 g of the metal slags were shaken for 23 hrs with leaching agent (distilled water acidified to pH 4 with HNO,). After 15 minutes of settling, the liquid was decanted and filtered through a 0.45 pm membrane filter. The membrane filter and collected solids were added to the slags in the flask and, after addition of fresh leaching agent, the extraction was repeated until five fractions were collected. After determination of pH and conductivity, the samples were acidified to pH 2 with nitric acid and stored for analysis.

105

Amlysis Concentrations in the extracts were determined with flame atomic absorption (Varian AA-1475). Concentrations below the detection limit were measured with graphite furnace atomic absorption. Mercury was determined with flameless absorption using a Coleman Mercury Analyser. Calcium, K and Na were determined with flame emission (Eppendorf ELEX 6361) and anions with ion chromatography (Dionex 2000i/SP).

Results and discussion

Total contents Total contents and physical characteristics of the metal slags tested are given in Table 1. Zinc is, amounting to 4 %, a main constituent of the slags. Several elements (Pb, Mn, Ba, Cu, Sn and Cr) are found in the range of 0.1 to 0.5 %. Many other elements are present at the m d k g level. Mercury is not present.

Table 1 Total element contents and some physical characteristics of the metal slags Total element contents (mg/kg) Al As

B Be Cd

co Cr cu Fe Hg

17727 325 713 7 123 455 1175 2876 246320 0.001

Mn

M0 Ni Pb Sb Se Sn Ti TI

v

5840 223 89 5437 220 257 2590 760 26 108

Zn Ba Ca K Mg Na

s

41816 2920.0 74173 2775 15085 11920 11000

Physical characteristics Density Bulk density Granulometry (%) 0 - 0.2 mm 0.2 - 1 mm 1-2mm >2mm

3567 k d m 3 2200 k d m 3 5.7 36.4 46.8

11.1

As compared to levels accepted for soils (Table Z), the total concentration of several heavy metals is very high. However, total contents should not be the only criteriurn to decide on possibilities for reusing the slags. As the metals are bound in an inert matrix, leaching may be so low that, even in the long term, no negative environmental effects can result. In the latter case, beneficial reuse is preferred over

106

Table 2 Reference levels in soil, surface water and ground water used in the Netherlands [2] Soil quality (mg/kg) A B

Element Cr co Ni cu Zn As Cd Sn Ba

Hg Pb

*

50 + 2L 20 10 + L 15 + 0.6(L + H) 50 + 1.5(2L + H) 15 + 0.4(L + H) 0.4 + O.O07(L + 3H) 20 200 0.2 + O.O017(2L H) 50 L + H

+

+

250 50 100 100 500 30 5 50 400 2 150

Water quality (Ilg/L)

C

A

B

C

800 300

1 20 15 15 150 10 1.5 10 50 0.05 15

50 50 50 50 200 30 2.5 30 100 0.5 50

200 200 200 200 800 100 10 150 500 2 200

500 500 000 50 20 300 000 10 600

A background level; some levels are calculated depending on clay content L (fraction c 2 pm) and organic matter content H

B: warning level for (closer) investigation C: warning level for sanitation (investigation) confined disposal. Leachability and mobility of the contained metals should be considered in the evaluation of the environmental effects of reusing of the slags.

leaching behaviour At US-ratios of 0.1 to 1, concentrations of several elements (Table 3) are high, compared to ground water quality standards (Table 2). The levels rapidly decrease in the subsequent fractions. At the highest US-ratios, only Pb and Zn persist in the leachate. Their concentrations exceed the Dutch C-levels for surface and ground water quality, even at the highest US-ratios. Metal slags are a potentially continuous source of Zn and Pb release in the environment. Concentrations of leached Zn are an order of magnitude higher than the concentrations of other elements. This is due to its much higher total content but also to the higher solubility of Zn as compared to other metals [3]. The cumulative leaching behaviour of several elements is depicted in Figure 1. The leached quantities generally amount to 0.1 - 0.5 7% of the total content at the highest US-ratios. Up to 1 % of the total Zn is leached. Even when leached concentrations are higher than accepted concentrations in surface water or ground water, adverse environmental effects resulting from beneficial reuse can be unimportant compared to the environmental and economical drawbacks of confined disposal. Particular field situations have to be modelled in order to assess the extent of metal migration. From the results, potential adverse effects and risks can be identified. Only when those effects are considered acceptable, reuse of the slags in bulk form can be allowed.

I07

Table 3 Concentrations in leaching test fractions f o r the different cumulative WS-ratios (L/S cum.)

ws

cum.

Vol. (mL)

PH

EC S/cm

As

40 160 200 400 400 800 2000 800 800 800 800

20.2 0.9 0.4 0.2 0.1 0.1 0.0 0.1 0.0 0.0 0.0 0.0

630.0 170.0 40.0 20.0 20.0 11.5 9.5 11.7 3.3 1.2 1.4 1.7

880.0 250.0 60.0 37.0 32.5 0.8 12.5 23.1 6.8 3.8 3.5 2.7

Concentration (pg/L) MO Ni

Pb

Se

1540 450 180 120 100 30 27 51 15 9 9 8

1720 660 460 410 320 300 270 315 250 170 225 340

161 27

Concentration (mg/L) Mg Na

c1

so,

140.0 7 7.0 2.0 I .o 1.o 0.0 0.0 0.0 0.0 0.0 0.0 0.0

208.0 93.0 60.0 20.0 1.0 1.3 0.7 1.a 3.3 3.1 3.0 1.1

684.0 173.0 42.0 15.0 8.4 4.8 3.2 6.4 2.6 1.3 2.0 1.o

1200 980 90 62 62 42 37 37

800

5.42 6.09 6.15 6.52 6.22 6.22 6.51 6.44 6.40 6.17 6.47 6.45

cum.

Cr

cu

Mn

0.1 0.5 1 2 3 5 10 20 40 60 80

9 1 6 0 1 0 3 53 1 10 12 3

190.0 45.0 14.0 21.o 170.0 43.0 39.0 18.0 11.0 8.O 12.0 49.0

1010.0 220.0 80.0 70.0 65.0 101.0 58.0 12.6 7.1 11.3 9.9 10.9

cum.

Zn

Ca

K

0.1 0.5 1 2 3 5 70 20 40 60 80 100

3.7 4.5 6.3 3.9 9.2 7.1 7.0 8.5 5.9 4.2 3.6 2.7

73.0 9.0 1.o 1.o 1.o 0.0 1.o 0.0 0.0 0.0 0.0 0.0

12.0 1.0 0.0 0.0 0.0 7 .0 0.0 0.0 0.0 0.0 0.0 0.0

WS

100

US

CO

~-

~

~~

0.1 0.5 1 2 3 5 10 20 40 60 80 100

Concentration (p g/L) Be Cd

23

21 18 17

255.0 37.5 0.0 3.1 0.9 0.0 7.0 7.2 4.6 3.5 2.3 0.0

5 0 0 0 0 0 0 13 7 11 10

11

11.7 2.3 0.2 0.2 0.3 0.2 1.2 0.0 0.0 0.0 0.0 0.0

2 0 1 2 6 2 2 3 3 3

108

CUMULATIVELEACHING (%of total)

I1

/

I I

..a

I

m....m'

II

0.001 ,,,......'.

0.0001

0.1

II

Column Test I Shaking , , , ,,

,

,

, ,,

1

10

I TBSt,

, ,,

J

100

US-RATIO

Figure 1. Cumulative leaching behaviour of selected metals from the metal slags.

ESTIMATING METAL MIGRATION IN THE ENVIRONMENT Introduction Results of a leaching test can only readily be understood in terms of potential environmental effects when they are applied to field situations. Besides the observed leaching behaviour, the particular characteristics of a disposal site also determine the actual migration of metals. These characteristics include geometry and dimensions of the slag mass involved, sources of leachate (precipitation, surface water, ground water), percolate flows and rates, and element fate determining mechanisms involved (adsorption, convection, dilution, plant uptake). Once a field situation is described, the time scale of the leaching process can be related to the US-scale of the leaching test. Next, from the leaching test results, concentrations and flows of metals from the slags can be estimated in function of time. Finally, the resulting risks and potential hazards can be evaluated. When worst case assumptions are made, one can estimate the most unfavourable effects. Whether these effects are acceptable or not will determine the feasibility of reusing the slags in bulk form. It can however remain dificult, especially when migration is low, to decide whether the observed effects are acceptable or outweigh the disadvantages of having to dispose the slags in isolated sites as a waste material. When metal slags are disposed on the land, metals may migrate with run-off water and contaminate surrounding fields where they can be taken up by plants. They also may enter surface waters where they can be taken u p by aquatic

109

organisms, or sorb on suspended material and settle in sediments. When the soil permeability is sufficiently low, the percolating flow can penetrate into the underlying soil where sorption processes retard metal migration. Eventually, contamination of ground water may occur. TO provide a better understanding of the significance of the observed leaching test results, metal migration from a hypothetical slag disposal is estimated. The ecological significance of the resulting contamination is evaluated using Dutch reference levels for quality of soil, ground water and surface water.

The field situation Migration from a hypothetical metal slag disposal is calculated. The slags are accumulated 1 m high on an area of 1 ha. The yearly precipitation extends to 780 mm. The bulk density of the slags is 2200 kg/m?. The underlying soil is assumed to have a bulk density of 1600 kg/m3 and a soil moisture content of 0.2. Relating the L/S-ratio scale to a time scale Since 1 m2 of the slag disposal area or 2200 kg of slag material yearly receives 780 L of precipitation, an US-ratio of 780/2200 or 0.35 is reached after 1 year. Inversely, a US-ratio of 1 is reached after 2.82 years. In reality, evaporation o f precipitated water will occur which will delay the predicted effects. In Table 4,some US-ratios are given with the corresponding time scale (denoted as Time) for the proposed field situation. Metal concentrations in the leachate and in a receiving surface water If all of the leachate of the slag disposal was collected by means of an impermeable layer and discharged in a surface water, a percolate volume of 0.247 L per second and per ha would be produced. The evolution of the concentrations in this leachate with time are directly estimated f r o m the observed concentrations of the leaching test (Table 3). The B-level, the level above which pollution can be suspected, is 200 pg/L for Zn (Table 2). As the concentrations in the first percolate amount to 123700 pg/L (Table 3) a dilution of 618 is needed to lower the concentrations to 200 p@. If the volumetric rate of a receiving surface water is at least 0,247 x 618 = 153 L/s, the Znconcentration will not raise above the B-level. After 28 years (WS-ratio of lo), a volumetric rate of only 8.6 L/s is needed. The pollution thus would not result in observable effects, even at the long term. Resulting environmental effects would thus be negligible. However, upon generalized reuse of the slags in road construction, it would be necessary to prevent leachates from entering directly into surface waters. The presence of many of those weak contamination sources in the environment would cause a slow but continuous increase of natural background concentrations in surface waters and sediments.

110

Table 4 Calculated average penetration depth (x), metal concentration in solution (c) and adsorbed concentration (s) for different L/S ratios and three types of underlying soil (L = light sandy loam, M = light loam and H = heavy clay)

ws cum. 0.1 1 10

20 100

Time (Years)

Volume (Urn31

0.3 2.8 28 56 282

220 2200

22000 44OOo

m o o 0

Adsorbed quantities (s)

L

(mg/kg) M

H

L

Depth (x) (cm)

M

H

Cadmium

0.006

35.5 8.6 1.5 0.9 0.3

316.5 80.1 14.1 8.9 3.2

280.1 69.4 12.1 7.7 2.7

0.2 2.4 24.1 48.2 240.7

0.0 0.3 2.6 5.1 25.5

0.0 0.3 3.0 5.9 29.6

0.190 0.044 0.052 0.029 0.023

146 35 42 24 18

198 48 56 32 25

1377 404 468 276 215

0.0 0.2 1.7 3.4 16.9

0.0 0.1 1.3 2.5 12.6

0.0 0.0

0.1 1 10 20 100

1.720 0.666 0.335 0.322

1047

2464 1167 630 608 510

10593 6016 3542 3427 2927

0.0 0.2 2.0 4.0 20.0

0.0 0.1

0.0

0.1 1 10 20 100

123.7 34.3 10.7 9.2 5.5

2103 7 488 783 709 476

4843 4478 3635 3490 2894

1.4 7.6 59.9 117.8

0.8 3.2 18.7 35.8 158.5

0.1 1 10 20 100

0.630 0.151 0.026

0.017

Copper 0.1 1 10 20 100

0.2 0.3 1.4

Lead

0.267

442 229 220 184

0.7 1.5 7.2

0.0 0.1 0.3 1.3

Zn

1233 617 243 215 133

563.7

0.4 1.1 4.0 7.3

26.1

Ill

Migration in the underlying soil

Estimation of average penetration depth When the underground is permeable, a considerable amount of the leachate can penetrate. Migration of metals however will severely be retarded as compared to the migration velocity of the bulk liquid phase because of sorption processes [4,5]. When a column is percolated with a volume V (L) of a solution, containing a concentration c (m@) of an element, the average penetration depth x (mm) of the element can easily be estimated from [5, 61: x = V/(e

+ r.s/c)

(1)

where 0 is the volumetric soil moisture ratio, is the bulk density of the soil (g/cm’) and s is the adsorbed amount (mglkg), in equilibrium with the concentration c (m@) in solution. The equation may more readily be understood when rewritten as: x(0c +

rS) =

vc

The original amount of metals added (right side of the equation) is distributed over a depth x in the soil, both in the liquid (ec) and the solid phase (rs). The relation between c and s is provided by an adsorption isotherm. For our example, a Langniuir adsorption isotherm [5, 71 was used with experimentally determined constants for three soils (Table 5). s = K.c/(1

+ K.c/s,)

(2)

To apply equations (1) and (2) on the data from the leaching test, weighed average concentrations of the total volumes corresponding with each US-ratio are calculated (denoted as c in Table 4). For Zn f.i., the weighed average concentration of the total volume at a L/S-ratio of 1 is calculated as: 123740 x 0 . 1

+

34530 x 0.4 + 16320 x 0 . 5

(0.1 + 0 . 4 + 0 . 5 )

= 34346 pg/L

Those concentrations are used in equation (2) to yield the corresponding adsorbed amount. After substituting the total volume, the corresponding weighed average concentration and adsorbed concentration in equation (l), the average penetration depth after each period of time is estimated. The results for the three soils are given in Table 4.

Modelling the concentration profile A prediction of the actual shape of the concentration profile in function of depth is obtained by using a one dimensional transport model, which includes adsorption phenomena [8]. Since sorption and desorption mass transfers occur much faster than the transport processes, adsorption equilibrium can be assumed. Sorption-desorption processes in soils have been shown to occur in a few minutes [5]. At any time, the relation between the concentration in solution and the amount sorbed is given by the Langmuir isotherm (2). The transport equation is expressed as [3]:

112

Table 5 Langmuir-adsorption parameters for the light sandy loam (L), light loam (M) and heavy clay soil (H) (S, in nig/kg, K in kg/L) [7] Medium (M)

Light (L)

s,

K

Sm

K ~~

Cd cu Pb Zn

3333 2480 7665 2000

4540 3845 8290 2500

57 818 705 26

Heavy (H) Sm

K

~

6370 5000 20400 5000

540 1096 2039 107

465 10000 12811 1250

where R is termed the retardation factor given by: z

ds

@

dc

R = I+--=

1 +

z

K/@

(1 + Kc/s,)*

with: o = volumetric moisture ratio (L3/L3) c = pollutant concentration (M/M) D = coefficient of dispersion (L2/T), calculated as Do + E V D,, = molecular diffusion coefficient ( L V ) , macroscopic velocity (L/T) s = adsorbed quantity (WM) s, = maximal adsorbed quantity (M/M) z = bulk density of the solid phase (M/L') t = time

c

=

dispersivity (L) en v =

This model was solved numerically, using the finite difference technique. The same conditions were assumed as in the estimation of the average penetration depth. In addition, the following assumptions were made: the volumetric rate being 780 mndyear, the macroscopic pore water velocity is calculated to be 1.068 cm/day. With Do = 1 cm?/day and E = 5 cm [8], D becomes 6.34. The input of metals in function of time was given by the concentrations observed in the leaching test, the L/S-scale being related to the appropriate time scale as illustrated before. The concentration profiles of Cd, Cu, Pb and Zn in function of time are shown in Figures 2 and 3.

I13

SORBED CADMIUM PROFILES HEAW CLAY

LIGHT LOAM

j U / , + 90

1ooo

1

2

3

4

5

looo

, 1

,

, 2

,

, . , 3

4

,I :I,, 5

1ooo

1

. , . , . ,

2

3

4

ADSORBED AMOUNTS (mg/kg) ADSORBED AMOUNTS (mg/kg) ADSORBED AMOUNTS (mg/kg)

SORBED COPPER PROFILES LIGHT LOAM

5 0 ~ 10

20

30

HEAW CLAY

L L 5 0 ~ 20 30 2~ 30

5 0 ~ 10

ADSORBED AMOUNTS (mg/kg) ADSORBED AMOUNTS (mg/kg) ADSORBED AMOUNTS (mg/kg)

2.8 years (US= 1) 28 years (US= 10) 56 years (L/S = 20) 282 years (US= 100)

-----

------_---

Fib'ure 2. Concentration and adsorption fronts of Cd and Cu resulting from leaching of a slag disposal after different time spans in different underlying soils.

114

SORBED LEAD PROFILES

or

. HEAW CLAY r- ---_-<- .................... .....

- --

10

20

-

'

-

30 -

-

40

I

L

100 200 300 1 500 100 200 300 1 ) 500 200 300 ADSORBED AMOUNTS (mg/kg) ADSORBED AMOUNTS (mg/kg) ADSORBED AMOUNTS (mg/kg)

500

100

1

SORBED ZINC PROFILES

HEAW CLAY

0-

:

* , ,,

...._..

40 L .......

,,: .'.

, .

,

.

60*;-

60 7

80 -

80 :

100

d

20 i

,' ;. , ,,'

40;'

- -: j_.--r------

-

100:

120 7 1401).

,

,

,

,

, . ,

140

r

.

I

.

i

.

I

,

I

00 0 200 400 600 800' ADSORBED AMOUNTS (mg/kg) ADSORBED AMOUNTS (mg/kg) ADSORBED AMOUNTS (mg/kg)

(US= 10) 56 years (US= 100) _ _ - -(US - _ _=_ 20) _ _ 282 years ................... -----

2.8 years (US= 1) 28 years

Figure 3. Concentration and adsorption fronts of Pb and Zn resulting from leaching of a slag disposal after different time spans in different underlying soils.

I I5

Discussion When comparing the results of the calculations of the average penetration depth and accumulation (Table 4) with the more realistic sorption profiles resulting from modelling (Figure 2 and 3), it appears that the latter extend to approximately two times the average penetration depth. The average penetration depth approach thus proves to be useful to get a first approximatirm of the extent of metal migration and accumulation that can result from an observed leaching behaviour. During the first years, migration of metals is in the order of a few centimeter. For the heavy clay soil, migration extends to only a few mm. Migration velocity varies between 1 - 5 cm/year for Zn, 0.1 - 7 c d y e a r for Cd and 0.01 - 0.8 cnv'year for Cu and Pb. Under natural conditions, migration into the clay ground will be even lower because of the low permeability of this type o f soil, which was not account for in the assumptions. Cadmium and Zn are illustrated to be the more mobile elements in the soils. The effect of soil sorption characteristics on metal movement appears very clearly and is dominant in determining the extent of metal migration in the underlying soil. In the light textured soils, metals migrate more quickly and are dispersed over greater depths. This results in a dilution effect: accumulation is low, but occurs over a larger depth. O n the contrary, in heavy textured soils, metals are expected to be retained very strongly. Accumulations are higher but occur over a smaller depth as compared to the light textured soils. Depending on the amount of metals, disposal of the slags on heavy textured soils thus will assure complete retention of the metals, even over very long periods of time. On the other hand, when metal amounts are sufficiently low, migration occurring in light textured soils may result in low accumulations that are hardly detectable and of no ecological significance. The latter is clearly the case for Cd, Cu and Pb. Cadmium accumulation in the light sandy soil extends to more than 1 m after 56 years, but is not higher than 0.5 m@g, a level which equals the reference level for a clean soil with similar clay content and organic matter content (Table 6). At the long term (300 years), migration is predicted to be 2.4 m deep, but no accumulation could be obsewed in practice. In the medium and heavy textured soils, Cd accumulation is 2 m g k g after 170 years in the first 25 cm. This accumulation is far below the B reference level and the depth is of little practical importance. Quantities of Pb and Cu, expected to be released a t the long term, result in similar effects as described for Cd, except that retention is much stronger. Even in the lightest soil, migration is not expected to extent below 20 cm. Lead accumulations in the latter soil are higher than the B-level of the Dutch soil quality standards (Table 2). For Cu, accumulations at the long term are below the A levels for the respective soils. Zn concentrations in the leachate are an order of magnitude higher than Pb and Cu levels. The estimated migration is therefore much more important than for the other elements considered. At the short (10 y) and medium term (50 y), Zn is expected to be retained in the upper 50 cm of the medium textured soil. In the heavy textured soil, Zn migration remains even at the long term within 50 cni. This migration depth contrasts with the penetration depth of Cu and Pb which was restricted to only a few cm. At the long term ( > 100 y), an important migration of Zn is expected to occur both in the medium and light textured soils. In the latter soil, zinc migration is predicted

1 I6

Table 6 Soil characteristics [7] and calculated A reference values [2] for the studied soils Light (L)

Medium (M)

Heavy (H)

6.3 2.6 7.8 Light sandy loam

6.5 2.8 13.6 Light Loam

7.1 3.6 39.6 Heavy Clay

0.6 24.9 66.4 95.1

0.8 40.9 93.2 174.2

Soil characteristics pH-H20 % OM % clay Texture

Calculated A reference values (mdkg dry soil) Cd cu Pb Zn

0.5 21.2 60.4 77.3

to extend to over 1 meter at the medium term and to over 3 m at the long term. Adsorbed amounts remain below the B-level, but could readily be detected. For the medium textured soil, the penetration depth is more than 1 m at the long term and accumulations are close to the B-level. In the light textured soil, groundwater contamination is most likely to occur. In the medium textured soil, groundwater contamination becomes likely only at the medium term, but will be more persistent than in the sandy soil as higher amounts are accumulated. The heavy soil is shown to retain Zn sufficiently. Even after the long term (250 years) no risk for groundwater contamination results. The high amounts of Zn leached may thus restrict reuse of the slags in bulk form. Environmental effects originating from isolated sites may not appear even at the long term because of retention or dilution effects. However, a generalized use in road construction may result in an overall increase of Zn background levels in surface and ground waters. Reuse in bond form (e.g. concrete) remains a valuable alternative because leachability will be strongly reduced. SUMMARY AND CONCLUSIONS

The leaching behaviour of selected heavy metals from granulated non-ferrous metal slags was investigated using the Standard Leaching Test of the Netherland Energy Research Centre (SOSUV-test). Using the results and sorptioddesorption data of selected soils metal migration from a hypothetical slag disposal site (area and height = 1 ha and 1 m, respectively) into the underlying soil was estimated. At liquid/solid (L/S) ratio’s of 0.1 to 1 (short term leaching, up to 3 years), concentrations of several elements were found to be present in significant concentrations. At the highest L/S-ratios (20 to 100, corresponding with 60 to 300 years), only Zn and Pb concentrations persisted in the leachate.

117

An unconfined disposal of the metal slags may lead to surface water pollution. It is however not likely that significant environmental effects would occur, as the amounts are easily diluted to background levels. In a heavy textured soil, metals leaching from the slag site mainly accumulated in the upper 10 cm. In more sandy soils, the metals were dispersed over larger soil depths, resulting in lower accumulated levels. When comparing the modelled migration results with soil and soil water quality standards, only the migration of Zn was of practical importance and may limit reuse of the slags in bulk form. References 1 Van der Sloot HA, Piepers 0 & Kok A. A standard leaching test for combustion residues. Shell BEOP-31, PettedNetherlands, 1984 2 Leidraad Bodemsanering, afl. 4. VROM, 's-GravenhageAVetherlands, 1988 3 Lindsey LW. Chemical equilibria in soils. Wiley & Sons, New York, 1979 4 Bolt GH. SoiI chemistry. B. Physico-chemical models. Elsevier, Amsterdam, 1979 5 Bolt GH, Bruggenwert MGM. Soil chemistry. A. Basic elements. Elsevier, Amsterdam, 1976 6 Van Genuchten MT, Wierenga PJ. Mass Transfer Studies in Sorbing porous media I. Analytical solutions. Soil Sci Soc Am J 40,473-480. 7 Kiekens L. Ph.D. dissertation, University Ghent, Belgium, 1980 (In Dutch) 8 De Smedt F. Simulation of ion transport in porous media. In: Rondia D, ed. Belgian research on metal cycling in the environment, Scope Committee, Liege, Belgium.

This Page Intentionally Left Blank

I19

Environmental impact of mining activities on the I-iermioni area, Greece

S P. Varnavasa, A G Panagosb and K G Kritsotakisc aDepartment of Geology. University of Patras, Patras (Greece) bDepartment of Applied Geology, Technical University of Athens, Athens (Greece) CDepartment of Mineralogy, University of Mainz, Mainz, Germany

INTRODUCTlON In Greece mining and metallurgical activities are among the main sources responsible for the release of toxic metals in the environment. The exploitation and transportation of the ores, as well as the ore processing cause, in many areas, fluxes oftoxic metals at the land sea interface resulting in offshore metal pollution. Except for the metal pollution caused by the escape of metals during the above community activities a significant degree of metal pollution offshore is the result of dumping of metal-rich solid toxic waste onto the seafloor. In the Antikyra Bay, on the northern Gulf of Corinth, bauxitic red mud resulted from an alumina processing plant is discharged on the seafloor through a pipeline at a rate of500,OOO t/yr On average the red mud contains 3 1 3% Fe, 5.70’0 Al. 1045 ppm Ni, 130 ppm Pb, 71 ppni Co, 637 ppm Mn, 87 ppm Zn, 43 pprn Cu and 3 ppm Cd ( I ) It has been demonstrated that during the first ten years of operation of the alumina processing plant a seafloor area of 28 h i 2 was covered by 2,250,000 tons of metal-rich toxic waste, while about 41,000 tons were dispersed away The difference between the total dry weight of each metal released on the seafloor during the first ten years of operation of the alumina processing plant and that found in the main toxic waste body formed on the seafloor has been calculated It has been found that 23 7o/b of Ni. 9.9% of Fe, 7 4% of Co, 6 7% of Cu, 5 1% of Pb, 3 3% of Cr and 3 2% of Ti from the red mud which was discharged on the seafloor during this period was dispersed away from the main area of discharge (2) Similarly, in Northern Euboekos Bay a seafloor area of about 45 kin2 is covered by 7,600.000 tons of metal rich toxic waste produced in an Fe-Ni smelting plant and released from r ship on the seasurface. It has been demonstrated that the main body of the solid toxic waste present on the seafloor contains about 1,456,000 tons of Fe203, 50,000 tons of C q O j , 3,30(J tons ofNi, 222 tons of Co and 5,500 tons ofMn ( 3 ) . It is implied that the continuous toxic solid waste-seawater interaction results in continuous release of toxic metals from the solid to the liquid phase which may get into the food chain Sulphide ores containing significant amounts of toxic metals such as Pb, Hg, Cd, As, etc were exploited in the past from sites situated not far from the coasts, and were exported from ports constructed for this purpose The environmental impact of such community activity on the shallow water marine environment in Greece is still unknown. The aim of this paper is to investigate the environmental impact of mining activities in the Herniioni area, Greece, including the offshore area

120

Qranitic mlruiivea

Subhlde M.fk

mines

~llr"iI"e8

Ultrrmafic rocks

unglng

Aegean

Figure 1

-

0

0 5 km

Sea

Outline geological map of the Hermioni sulphide mine area, after Aronis, 195 1 (4) reprinted from A H F Robertson, S P Varnavas, A G Panagos, Sedimentary Geology, 53, 1-32, Elsevier, Amsterdam, 1987), showing the offshore area studied

GEOLOGICAL AND PHYSIOGRAPHIC SETTING The Hermioni area, located at the southern part of the Argolis Peninsula of Peloponnese, is covered by sedimentary rocks, while magmatic rocks are present to a lesser extent However, significant volumes of metal sulphides and manganese oxides are associated with magmatic ophiolitic rocks in this region (4-10) being the major sources of toxic metals released in the Hermioni coastal environment Loose unconsolidated ferromanganese sediments occurring at the periphery of the above deposits, a few kilometers from the coast contain up to 2000 ppm Cu. 800 ppm Zn, 200 ppm Pb, 255 ppm Co and 320 ppm Cr (6) Exploitation and transportation of metal sulphides in the Hermioni region began in the end of last century and the ore was exported through a small port situated east of the town of Hermioni. The average annual production was

121

25.000 tons of FeS2 and 7.50 tons of Cu, with a total production of 1,000,000 tons of ore

The ofishore area investigated is characterized by two narrow embayments, the Ntartiza and Thermisi embayment and the cape Thermisi between The water depths vary between a few meters and about 40 meters, the seafloor being characterized by smooth topography In the Thermisi embayment a lagoon is formed ofan ellipsoidal shape, with its long axis oriented almost parallel to the coastline A major geological formation present on the

Figure 2

Detailed geological maps of Herniioni mine areas simplified and modified after Aronis, I95 I (4). reprinted from A H F Robertson, S P Varnavas, A G Panagos, Sedimentary Geology, 5 3 , 1 -j2, Elsevier, Amsterdam, 1987 (a) Roros-Cambrorosso area

adjacent land is the shale-sandstone formation (flysch) in which mafic and ultramafic rocks and sulphide and ferroinanganese ore bodies are encountered Two major streams flowing in the Ntartiza embayment transport weathering products from the ultramafic rocks, the flysch and tlie quaternary deposits occuring in the vicinity of the enibayment Similarly, a major stream flows in tlie Thermisi embaynient transporting weathering products from the flysch and the inatic rocks The Cape Thermisi is composed of limestones It is noted that the sulphide ore mines, including open pits, are situated only a few kilometers

I22

from the coastline In the mine area large volumes of "gossans" (the oxidation products of the sulphide ore) are exposed on the ground, being continuously leached by the rainwater

Figure 2 (continued) (b) Karakasi area The rainwater sulphide-ore interaction leads to the formation of H2SO4 which reacts further with the sulphide ore, the gossans, the ferrornanganese ores and the associated rocks and sediments releasing metals in the environment

COMPOSlTlONAL CHARACTERIZATION OF THE SULPHIDE AND MANGANESE ORES Manganese arid ferromanganese deposits Chemical analysis of sediments associated with diabase lavas in the Hermioni mine area showed significant enrichments in Fe, Mn, Zn, Cu and As relative to normal pelagic clal:; Marked geographic compositional variations within the Hermioni area were observed, the highes' metal enrichments found at Baroutospelia (Table 1 ) Mineralogically the Fe-Mn sediments contain hematite, quartz and minor amounts of illite, smectites and felspars The Fe-rich Mn poor

I23

sediments contain abundant hematite and smaller amounts of quartz, calcite and smectites At Baroutospelia, except for the disseminated Fe-Mn sediments an open pit of Mn-ore exists which is characterized by significant Zn and Cu enrichments Moreover it contains higher amounts of Fe, Zn and Al compared with all subpelagonian and Pindos Greek manganese deposits The chemical composition of the Hermioni manganese and ferromanganese deposits is shown in Table l(8). Table 1 Chemical composition of Hermioni manganese and ferromanganese deposits 1 Average of I 5 samples of manganese ore samples from Baroutospelia, Ermioni; 2 and 3 representative analyses of Fe-Mn sediments from Baroutospelia, Ermioni, 4. representative analyses of Fe-Mn Sediments from Kapsospiti, Ermioni; 5. and 6. representative analyses of Fe-Mn sediments from Canibrorosso, Erniioni, 7 representative analyses of Fe-rich, Mn-poor sediments from Roros, Ermioni Fe (Yo) 1 2 3

4 5 6

7

Mn Ni Co Pb Zn Cu Cr Al (90) (ppm) (ppni) (ppni) (ppm) (ppm) (ppm) (%)

2 3 7 4447 223 1285 I93 I 1 59 100 986 I 52 9 I4 I30 41 90 0 13

I 5 83

199 187 167 194 172 170 37

100 75 59

58 54 56

255

32 60 SO 40 40 40 200

265 347 287 232 233 267 800

118 1027 81 170 100 71 2275

45 70 70 90 90 80 320

222 670 730 780 770

890 140

Ca

Fe/M

(YO) I36 I60 200 320 120 130 430

0053 7 10 669 I I 59 649 703 2205

Reprinted from A H F Robertson, S P Varnavas, A G Panagos, Sedimentary Geology. 53, 1-32, Elsevier. Amsterdam, 1987

Siilpliide ores The chemical analysis of sulphide ore samples showed that the Herniioni ores are Cu-rich pyrites containing always greater amounts of C'u than Zn The Cu concentration levels are high at Roros, intermediate at Kapsospiti and low at Karakasi ore bodies By contrast, the Roros ore bodies are characterized by lower amounts of Zn compared with the Kapsospiti and Karakasi deposits(9). A greater thickness of the sulphide ore bodies was observed at Karakasi and Kapsospiti compared with those of Roros area (5) The concentrations of Si and Zn in the pyrite increase, whereas Cu decrease with increasing thickness of the ore body A positive correlation between Cu and Fe and a negative correlation between Cu and Si was found (9) At Karakasi and Kapsospiti a greater size of pyrite crystals than at Roros was observed (5) The Herniioni sulphide deposits have comparable chemical composition with that of modern sulphide deposits formed on the Galapagos Rift at 86O W (9,l I ) On the basis of their geochemical and geological features it was implied that they were formed from high temperature submarine hydrothermal exhalations resulted from seawater-rock interaction, in a process similar to that giving rise to the formation of modern sulphide deposits on mid-ocean ridges. The

124

chemical composition of the Hermioni sulphide deposits is shown in Table 2 (9) Table 2. Average chemical composition of Hermioni sulfide deposits compared with that of modern submarine sulfide deposits. Kapsospiti Karakasi Red Sea* ( I 1) Galapagos EPR*** Roros Rift** (12) 21 N (13) 86 W Fe(%) 40.68 34.05 39.61 15.74 31.2 12 6 1.44 0.74 2.22 8.0 1 16 Cu(%) 3.20 0.31 0.3 1 10.93 0.7 49.70 Zn(%) 0.22 0.13 1.11 0.01 Mn(%) 0.40 6 g/t 120 ppm 16 pprn 290ppm Ag 20 g/t 0.2 s/t Au 0.4 .s/t Si02(%) 4.12 10.02 13.4 2. I4 AI?O?(%) 0.22 0.19 2.81 0.6 1 (Al) Reprinted from S.P Varnavas and A G Panagos, Chemie der Erde, 49, 81-90, VEB, Gustav Fischer Verlag, Jena, 1989

MATERIAL AND METHODS The seafloor sediment samples were collected from a fishing boat using a stainless steel Van Veen grab. Subsamples were dried in dessicators and ground to tine powder using an agate mortar and pestel and stored in plastic vials. Weights of 0.2 g of each sample were disolved by a mixture of HF/HNO3/H2O2 (5:l. I ) in high pressure vesells at 180OC. For the Mercury determination 0,25 g sample weight was leached using the acids HCI/HNOj (3:l) in closed teflon vessels at 6OoC. After filtration the solutions were diluted with water to 25 ml. Analytical grade acids and double distilled water were used for the sample disolution and dilution. The concentrations of the major and trace elements in the sample solution, along with those of synthetic standard solutions of similar matrix composition were determined by inductively plasma emission spectrometry (ICP-OES) at optimized analytical measuring conditions. Instrumental precision and long term stability (drift) were tested using a synthetic multielement solution. The accuracy of the analytical results was determined by the analysis of the USGS soil standards GXR-5 and GXR-4. Analytical precision was better than k l , % with the exception ofNi (k3.1 %), As (kl.6 %) and Pb (k1.2 %); accuracy was better than 10 %, except for Be (+I4 %), Ni (+I3 %), Sb (+I0 %), Pb (+I2 %) and Hg (114 %). A11 concentrations are in ppm, except for Hg and Ag which are given in ppb. The locations of the sediment samples analysed are shown in Figure 3.

Table 3 Results from the chemical anal)ses of Hermioni offshore sediments S102 A120;, Fe2O3, Ti02, MnO, CaO and P205 in 70,Hg and Ag in ppb.

e

N

m

Table 4 Maximum concentrations of elements in the Hermioni offshore sediments compared with their maximum concentrations in other Bays Bays/Ref Pb Mo Zn Co Cu Ni Cd Cr V Mn Fe pprn pp m ppm ppm pp m pp m pp m pp m ppm (%> (“A) Hermionr Bays*

1064

152

2539

341

7323

2500

140

364

228

299

20

37

160

40

430

33

101

168

Saronikos Gulf163i7 Thermaikos Gulf17 Patraikos Bayi8 Navarino

59

Kalamata Bay2()

40

86

193

352

29

66

151

355

43

56

173

Itha’ci Bay21

106

106

359

28

55

150

Argostoli Bay22

206

43

I4 18

85

20

10

984

131

0 16(Mn0)

1604(Fe2O3)

Method total

250

0 46

05NHCI

107

185

0 5 N HCI

210

06

total

414

39

233

total

418

0 36

6 09

total

383

0 19

4 38

total

005

2.90

total

132

24

46

106

205

Pagassitikos Gulf17

34

74

14

32

114

395

N Euboicos Bay23

27

58

336

28

41000

106

2 N HCI

2.33

2 N HCI

908

510

227

278

2 N HCI

Lesbos Island l 7

39

43

12

247

2 N HCI

Shallow Water24

22

Kavala Bay’’

*This study

1

92

13

56

35

60

145

0.085

6.5

total

Table 5

e

N

4

128

10 m In

a13

c

01.5

II

3

u12

ui 4

X

018 YMlN = -.27

Figure 3.

Sketch map of the area studied showing the location of the sediments analysed

RESULTS The results from the bulk chemical analysis of the Hermioni offshore sediments are showtl in Table 3 Metal Concentration Levels It is seen that significant enrichments in Hg, Ag, Pb, Cd, Cu, Zn and Fe relative t o normal nearshore sediments occur in the region studied. Mercury reaches the value of 2082 ppb, this value being much higher than that of the average shale, mean crust and mean sediment (14). The maximum Hg vatue is also greater than that reported for Navarino Bay sediments ( I 5). Silver reaches the value of 76 ppm being significantly enriched relative to the average shale, mean crust, mean sediment and the Navarino Bay sediments. The maximum value of Pb is 1604 ppm, being 73 times higher than the average concentration of Pb in shallow water sediments. The above value is greater than the maximum Pb values reported for all Greek Bays Studied (Table 4). Cadmium varies between 0.5 ppm and 14 ppm its maximum value being 64 times higher than its average value in shales. The above value is also greater than the maximum Cd values reported for all Greek Bays studied. Copper varies between 40 ppm and 7323 ppm, showing a 130-fold enrichment relative to normal shallow water sediments. Significant, but lower Cu enrichments in shallow water sediments were reported from the Vassilikos Bay, Cyprus (25). A 28-fold enrichment in Zn was found relative to normal shallow water sediments, its maximum value being 2540 ppm. This value is comparable with the maximum concentration of Zn reported for the heavily polluted Saronikos Bay sediments. Generally, cobalt is not among the metals markedly enriched in the metal polluted coastal sediments However, in the sediments

I29

508

F i g u r e 4 . V a r i a t i o n s of Fe2O3, Cu and Zn r e l a t i v e t o t h e i r average v a l u e s

I30

Figure 5. Variations of Hg, Co and P b relative to their average values

131

Figure 6. Variations of Cd and Be relative to their average values

I32

Figure 7. Variations of Mo, Ba and V relative to their average values

133

Figure 8. Variations of M n and Ag relative to their average values

134

E-

Figure 9 . Variations of Cr, Ni and P2O5 relative to their average values

135

studied a 26-fold Co enrichment is observed, this enrichment being greater than that observed in Northern Euboekos Bay metal polluted sediments (16-26) Figures 4 to 9 demonstrate the variations in the element concentration levels relative to their average values in the area studied. On the X axis of the above figures the sample coding numbers are shown in order of increasing number. On the Y axis the 0- level represents the average concentration value which appears on the right margin, while the positive and negative values represent the positive or negative deviations from the average value. The average concentration of Fe203 is 4.8% with one major geochemical anormaly at station 4 and a minor one at station 9 At these two stations similar geochemical anomalies were observed for Cu, Zn, Hg and Co, while Pb and Cd exhibit marked anomalies only at station 4 . By contrast Be shows a major anomaly at station 4 and minor anomalies at stations 9, 16 and 20 Station 4 is characterized by a depletion in Al, Mo, Ni and Ti Manganese exhibits a major anomaly at station 7 and a minor one at station IS Silver shows a distinct peak at station 5 . Cr, Ni and P show no significant variations except for a peak at station 23 Sr exhibits two major anomalies, one at station 3 and another at station 22. Geostatistical Analysis In order to investigate the geochemical behaviour ofthe elements studied and their mode of incorporation in the sediments the correlation coeffrcients among their concentrations were calculated The results are shown i n Table 5 The application of factor analysis on the geochemical data showed that Factor I accounts for 41.1% of the data variance and shows strong loadings of Fe, Zn, Pb, Cd, Hg, Cu, Co and Be. This factor represents the sulphide-bearing fraction of the sediments and demonstrates the strong input of sulphide related metals to the marine environment. Factor 2 accounts for 23. I % of the data variance, it shows strong loadings of Al, Zn, V, Ba, Mo and Ag and represents the clay fraction of the sediments. Elements like Ag which originally were associated with the sulphide minerals may be adsorbed after their oxidation on the clays. Factor 3 accounts for 15.0% of the data variance and shows strong loadings of P. Cr and Ni and may represent weathering products of magmatic rocks Factor 4 accounts for 9 4% of the data variance It shows strong loadings of Ca and Sr representing the carbonate fraction of the sediments. Factor 5 accounts for 5 9% of the data variance and shows strong loading only of Mn, representing the manganese oxide fraction of the sediments The fact that no loadings of trace metals are observed in this factor supports the different geochemical behaviour of Mn.

~

~~

I -0 0 10 -0 200 0 953 -0 287 0 19.3 0 164 0 I20 0 0-39 0 974 0 390 -0 188 0 060 -0 727 0 005 0 078 -0 0 5 I 0 967 0044 0 088 -0 174 -0 1 I 5 TI I

~~~

~

~

2 -~

~

-0.184 0 914 0.086 0 854 0 029 -0 318 -0 177 -0 552 -0 177 0 882 -0 I 9 4 -0 I93 0 236 03 5 -0 1.33 0 788 -0 22.3 -0 159

12; 0 908 0 575 13 I -0

~~~~

~~

-0005

-0 969 -0 075 0 027 -0 1 1 1 0 069 0 859 -0 143 0 746 0 084 0 001 -0 173 0 064 0 109 -0 023 0 088 -0 I 6 6 0 055 -0 006 0 048

-0 01 I 0 152 15 0

-0 I 3 6 04

0 105 -0 234 -0 071 0 248 -0 014 0 066 0 957 -0 174 -0 042 0 004 0 933 -0 0.39 0 91s -0 250 -0 0 1 5 -0 5 1 3 -0024 -0003

~

-0 006 0 088 0 227 -0 167 0 962 0 19s -0 0 10 -0 140 0 069 0 260 -0 060 0 072 0 052 0 027 0 I10 -0 080 -0 012 -0 I07 0 107 -0 052 -0 167 59

-0 038 ~~

~~

5

A

)

~~~~~

~~

Figures I0 and I I denionstrate the dilt'ei-ent geochemical behaviour of Ag and Flg in thc area s1tidied

Ile~iierrtGeographic I)istributiori 111 order t o investigate the eleriient geographic variations, the geochemical niap of each zleirient was tlrawii tising ii specilic coniputing piogi-ainrne lie, ZII,1'11, Cd, Ilg, <:II, c'o :iritl lle T h e sulpliidc associated elements s h o w similar- geographic variaiions 'rheil- Iiigliesi concentrations iii-e encouritcred in the eastern inner part of the N t a i - t i n enibayment f?om wherc the 01-e w a s loatled o n t o !lie ships 'l'liere is a tendency for this group of elements 10 decrease w 1 w i r d iii the eiiibayiiieiit, while Ilieii. lowest viiltics ai-e observed in the castei-n par! ol' the stud! Rl-ea

137

I

Figure 10 Relationship between Ag and A1203

Figure 1 1 Relationship between Hg and F q O j

138

Al, V, Ba, Mo, Ag In the western part of the study area the highest concentrations of Al, V, Ba, Mo and Ag are observed mainly away from the coastline, as a result of the transportation of the fine material Increased concentrations of these elements observed near the coast at the eastern part of the study area are probably associated with a coarser detrital fraction of the sediments. Cr, P,Ni The highest concentrations of Cr, P and Ni are observed in the eastern part of the study area Elevated values are also observed in the western part, while in the area between, intermediate values are found Relatively low values are found in the sediments enriched in sulphide related elements Mafic and ultramafic rocks are the sources of Cr, P and Ni

Mn The areal distribution pattern of Mn is completely different from that of all other elements studied. The easter and western parts of the study area, which are enriched in Ni and Cr are characterized by very low concentrations of Mn. Although Mn does not follow Fe, Cu, Pb, Zn, Cd, Hg and Co in its areal distribution the area enriched in the above elements is also enriched in Mn. It is therefore, implied that Mn is at least partly associated with the sulphide ore disseminated on the seafloor.

DISCUSSION The results of field and laboratory work carried out in this study reveal that the Hermioni area has been greatly affected by the mining activities. The field observations made in different seasons revealed that a considerable amount of sulphide ore and gossams are exposed in the area of the old mines, being a constant source of toxic metals released in the surrounding environment. The rate of oxidation of the exposed sulphide ore increases during the winter due to its leaching by the rain and the formation of acid waters. The acid waters not only accelerate the release of toxic metals from the sulphide ore itself but in addition they react with the Fe-rich and/or Mn rich metalliferous sediments associated with the sulphide ore increasing the availability of metals in the environment. At sites with Mn-rich outcrop their weathering has reached to a stage where it has produced Mn-rich soil now present in the olive tree fields. Since the Mn-rich outcrops are also rich in Cu and Zn an investigation is necessary to be carried to examine the possibility of the uptake of Mn, Cu and Zn by the trees Similarly, it is necessary to examine the quality of the ground waters of the area in regard with the presence of Mn, Fe, Cu, Zn, Cd, Ag and Hg. The enrichment of these metals in the sulphide and/or manganese ores implies their presence in their weathering products and the produced acid waters. The significant enrichments of Hg, Ag, Pb, Cd, Cu, Zn and Fe found in the surface sediments offshore Hermioni demonstrate that the mining activities which took place for some time in the region had an important environmental impact also on the marine environment. Although 23 years have been passed since the mining operations ceased in the area, the metal enrichment found in the seafloor sediments suggest the presence of significant amounts of sulphide andlor Fe-oxide phases. The strong impact of mining activity on the marine environment is hrther demonstrated by

139

m

"?

u)

II

3

X

YMlN = -.27

( w )

FeO

049

078

025

049

.

088 061

087

029

YMlN =

- 27

YMlN = -.27

Pb ( P P d Figure 12. Areal distribution of Fe, Zn and Pb

140

F i g u r e 13. A r e a l distribution of Cd, Hg and Cu

141

YMlN = -.27

co ( P P d

YMlN = -.27

MnO (

%

)

F i g u r e 1 4 . Areal d i s t r i b u t i o n of C o , B e a n d Mn

142

YMlN

3

-.27

Figure 15. Areal distribution of V, Mo and Ag

I43

YMlN = -.27

01

-

Lc! (0

II

3 X

YMlN = -.27

Ni ( P P m ) F i g u r e 1 6 . Areal d i s t r i b u t i o n of C r ,

P and N i

144

the high percentage of the total data variance which accounts for factor 1 in the results of factor analysis applied on the geochemical data. This factor shows high loadings of sulphide related metals such as Fe, Cu, Zn, Pb, Cd, Hg and Co. Except for the strong association of Cu, Zn, Pb, and Co with Fe in the seafloor sediments and their unusual enrichments, their ratios against Fe are comparable with those of the Hermioni sulphide ores. It is therefore implied that the source of the above metals is the Hermioni sulphide ores. However, it is not possible from the data of this study to deduce the mode of transportation of the metals to the seafloor? Has sulphide ore been dumped in the Ntartiza embayment during its loading on ships, Have the metals been transported to the sea through the stream or both ways are responsible for their enrichment? Further investigation needs to be carried out in the area, including stream sediment analysis, in order to answer the above questions. Sedimentation rate of non polluted sediment in the shallow water environment is an important factor controlling the final metal concentration levels in the seafloor sediments. The extremely high concentration of Fe, Cu, Zn, Pb, Cd and Hg found in the Ntartiza embayment would suggest that in this area the sedimentation rate is either very low or the metal accumulation rates are very high. Thus, before any suggestions in regard with the reclamation of the area are made, it is of great importance to determine the present metal accumulation rates, the way of their input into the sea, as well as the sedimentation rates. No data on Hg and Cd concentrations in the Hermioni sulphide ores are available in the literature. The strong association of Hg and Cd with the sulphide related elements (i.e. Fe, Cu. Zn, Pb) in the seafloor sediments suggest their common source. It is therefore revealed from the data of this study that two very toxic metals, Hg and Cd, are among the other metals which are released in the area of the old mines. The results obtained from the calculation of the correlation coefficients among the elements analysed show that Mn exhibits no relationship with the sulphide related elements, while in the results obtained from the factor analysis i t , is seen that it forms a separate factor. This geochemical behaviour of Mn in the Hermioni offshore sediments could be assigned to a different source than that of the sulphide ore. However, an examination of the chemical composition of the Hermioni sulphide ores shows that they contain significant amounts of Mn, especially the Roros ore bodies, which are situated more close to the coastline than any other sulphide ore body. Another alternative which could explain the fractionation of Mn from the sulphide related elements is its diagenetic remobilization and reprecipitation. This is a very usual process in the shallow water marine environment where reducing conditions occur in the subsurface sediments. It is therefore suggested that Mn has the same source as Fe, Cu, Zn, Pb, Cd, Hg and Co but its different geochemical behaviour is due to its remobilization and redistribution, The fact that Mn exhibits its highest concentrations at stations occurring in the vicinity of the sites with the highest values of Fe, Cu, Zn, Pb, Cd, Hg and Co is an additional evidence of their common origin. * In the diagenetic remobilization of Mn it is probable that a part of the remobilized Mn and associated elements like Ni, Co and Zn remain in solution being dispersed away. Silver like Mn behaves differently from Fe, Cu, Zn, Co, Pb, Cd and Hg. It shows strong affiliation with A1 although its correlation with Al is not as strong as that of V, Mo and Ba. In addition, its highest values are found at sites very close to the sites with the highest concentration of the sulphide related metals. Considering the Ag enrichment in the Hermioni

145 sulphide ore bodies it is suggested that disseminated sulphide ore in the shallow water should be at least partly, a source of Ag. Its present association with Al would suggest its removal from the sulphide ore and its reincorporation probably in the form of adsorbed ion on the clay minerals. Such a process would result in the release of part of Ag in the seawater increasing the chance of its uptake by marine organisms.

CONCLUSIONS It is concluded that the exploitation and transportation of sulphide ores which took place for a number of years in the Hermioni area (Greece) caused environmental contamination offshore. The continuous chemical weathering of sulphide minerals and metal-rich gossans exposed at the mining sites led to the release of considerable amounts of toxic metals in the surrounding environment A portion of these metals reached the marine environment and it still remains at the sediment-seawater interface It is implied that reniobilization and redistribution of certain metals takes place as process which increases their availability in the seawater and the possibility of their uptake by marine organisms and in getting into the food chain

REFERENCES 1. Varnavas, S.P., Ferentinos, G and Collins, M., Mar Geol. 1986, 70 21 1-222. 2 Varnavas, S.P. and Papatheodorou, G., Marine Mining, 1987; 6 37-70 3 . Voutsinou, F and Varnavas, S P., Marine Mining, 1987, 6 259-290 4 Aronis, G , Dept. of Subsurface Research Institute, Athens, 195 1 5. Aranitis, S., Ann Geol. Pays Hell 14:211-323, 1963. 6. Varnavas, S.P., Panagos, A . G , Chem Geol., 1984,42:227. 7. Varnavas, S.P., Panagos, A.G., Phillipakis, S.. Geol. Zb Geol. Carpathica, 1985,36 21 8. Robertson, A.H F , Varnavas, S.P , Panagos, A.G., Sedin-. Geol , 1987;53:1. 9. Varnavas, S.P , Panagos, A.G., Chem Erde, 1989;49.81 10 Mousoulos, L., Anal. Geol. Pays Hell 1958, 119-164. 1 1 Cronan, D.S., Underwater minerals, Academic Press, London 362 pp.. 1980. 12. Malahoff, A., Emberly, R , Cronan, D and Shirrow, R., Marine Mining, 1983; 4: 123-137. 13 Hekinian, B , Fevrier, M , Bischoff, J L., Picot, P and Shanks, W C , Science, 1980, 207 1433-1444 14. Salomons, W and Former, V., Metals in the hydrocycle, Springer Verlag, Berlin, Heidelberg, 1984. 15 Varnavas, S.P , Panagos, A.G. and Laios, G., Proc. Intern. Conf Heavy Metals in the Environment CEP Consultants Ltd., Edinburgh 1985, 41 5-417. 16 Grimanis, A.P , Vassilaki-Grimani, M. and Griggs, G.B , Journ. Radio Anal. Chem., 1977;37:761 17 Voutsinou-Taliadouri, F , VIIes Journees Etud. Polutions, Lucerne, C.I.E.S.M., 1984,25 I 18 Varnavas, S.P and Ferentinos, G VI Journees Etud Pollutions, Cannes, (2.1 E.S.M. 1983 19 Varnavas, S.P , Panagos, A.G. and Laios, G., Environ. Geology and Water Sci 1987;lO:159

146

20. Varnavas, S.P., Panagos, A.G., and Laios, G., Vlls Jour Etud. Poll. Lucerne C.I.E.S.M., 1984;267. 21. Varnavas, S.P., Panagos, A.G., Laios, G. and Alexandropoulou, S., Proc. 6th Intern. Conf Heavy Metals in the Environment, New Orleans, U.S.A., 1987;203. 22. Panagos, A.G.,Alexandropoulou, S. Agiorgitis, G. and Varnavas, S.P., Thalassographica, 1983 (in press). 23. Voutsinou-Taliadouri, F., and Varnavas, S.P., Proc. 5th Intern. Conf. Heavy metals in the Environment, Athens, Greece, 1985;336. 24. Wedepohl, K.H., Handbook of Geochemistry, Springer Verlag, Berlin, Heidelberg, 1979. 25. Varnavas, S.P., Proceedings of Conference Environmental Science and Technology. Lesvos, Greece, 1991 ;1. 26. Scoulos, M.J., Mar. Chem., 1980;18:249.

CHAPTER 4 Human impacts on soils by wastes and contaminated dredged material uses

This Page Intentionally Left Blank

I47

BENEFICIAL AND TOXIC EFFECTS OF CHROMIUM IN PLANTS: SOLUTION CULTURE, POT AND FIELD STUDIES. J u a n Barcel6, C h a r l o t t e Poschenrieder, Marla Dolores Vazquez 8 Benet Cluns6 Laboratorio de Fisiologia Vegetal, F a c u l t a d d e Ciencias, Universidad Aut6noma de Barcelona. E-08193 Bellaterra, Spain. INTRODUCTION

In a d d i t i o n t o t h e depositions from t h e atmosphere, t h e extended use of sewage sludge, municipal waste compost a n d fertilizers in agriculture, results in increasing concentrations of heavy metals in crop soils [1,23. Consequently, t h e necessity t o adopt guidelines f o r maximum p e r m i s s i b l e metal c o n c e n t r a t i o n s i n sludges, composts a n d soils f o r a g r i c u l t u r a l use is i n d i c a t e d . Although t h e maximum p e r m i s s i b l e c o n c e n t r a t i o n s f o r t o x i c metals given by t h e d i f f e r e n t governments or o r g a n i z a t i o n s a r e d i s c r e p a n t , a n d not only s c i e n t i f i c c r i t e r i o n s , b u t also economic a n d social i n p u t s h a v e d e t e r m i n e d t h e f i n a l l y e s t a b l i s h e d v a l u e s [l], t h e r e is a consensus a b o u t t h e importance of r e s t r i c t i n g t h e c o n c e n t r a t i o n s of t h o s e metals which a r e e i t h e r o r b o t h highly toxic a n d mobile, i n o r d e r t o p r e v e n t b o t h t h e t r a n s f e r of t o x i c metals i n t h e food c h a i n a n d t h e i r l e a c h i n g t o groundwater. Moreover, phytotoxic e f f e c t s a r e t o be avoided. I n this sense, maximum permissible c o n c e n t r a t i o n s f o r C d , Zn, Pb, a n d Cu a r e generally considered. The e s t a b l i s h m e n t of maximum permissible c o n c e n t r a t i o n s f o r C r t u r n s o u t t o be especially c o n f l i c t i n g . I n t h e e n v i r o n m e n t , t h e metal can occur e i t h e r i n t h e form of C r I11 o r C r V I . Under t h e p r e v a i l i n g e n v i r o n m e n t a l c o n d i t i o n s , t h e t r i v a l e n t s t a t e is t h e most s t a b l e form. A s C r I11 is poorly toxic a n d according t o s e v e r a l s t u d i e s most soils e a s i l y immobilize C r 111, t h e necessity to restrict Cr concentration in soils for a g r i c u l t u r a l use is n o t u n i v e r s a l l y recognized. I n t h e p r e s e n t Work, a f t e r a s h o r t review considering t h e emission of C r t o t h e environment, t h e toxicity of C r f o r

148

men and plants, t h e e s s e n t i a l role of t h e metal i n men, i t s beneficial influence on plants, a n d t h e recommended maximum permissible C r c o n c e n t r a t i o n s i n soils, sludges a n d compost f o r a g r i c u l t u r a l ' use, we p r e s e n t a summary of r e s u l t s on t h e influence of C r i n plants, using n u t r i e n t solutions w i t h C r 111 o r C r V I salts o r soils amended w i t h Cr-containing t a n n e r y sludge. Our r e s u l t s show t h e complex responses of p l a n t s t o Cr a n d t h e need f o r f u r t h e r i n v e s t i g a t i o n s , especially concerning speciation of C r i n soils. Input of C h r o m i u m i n the e n v i r o n m e n t Chromium is a r e l a t i v e l y a b u n d a n t element w i t h multiple i n d u s t r i a l uses. Most of t h e a b o u t 9106 t of chromite t h a t a r e yearly mined a r e employed i n s t a i n l e s s steel production. But d i f f e r e n t Cr-compounds are also r e q u i r e d f o r manufacturing dyes a n d pigments, t a n n i n g p r o d u c t s , wood p r e s e r v a t i v e s , corrosion i n h i b i t o r s f o r cooling water, a n d magnetic C r V I used i n t h e f a b r i c a t i o n of audio, video a n d computer tapes ( r e f . i n [3-5]). Estimations of t h e worldwide emissions of t r a c e elements i n d i c a t e a n a n n u a l i n p u t of 4134.10~t o 1,309403 t of C r i n t o soils, 45403 t o 239403 t i n t o a q u a t i c ecosystems a n d 7.3403 t o 53.6403 t i n t o t h e atmosphere [6]. For Europe, a y e a r l y emission of 18.9.i03 t C r i n t o t h e a t m o s p h e r e has been estimated [TI, The main anthropogenic source for cr i n t h e a t m o s p h e r e is b y f a r t h e i r o n a n d s t e e l m a n u f a c t o r i n g i n d u s t r y , While conventional t h e r m a l power p l a n t s , i n d u s t r i a l and r e s l d e n t i a l boilers and cement production only c o n t r i b u t e t o t h e t o t a l C r emission 8, 10 and 4% respectively 171. On a global scale t h e main sources f o r C r i n p u t i n t o s o i l s a r e coal and bottom f l y ash a n d the wastage of commercial products CS]. Tannery waste usually h a s a h i g h C r concentration and t h e mean contamination c h a r g e f o r 1 t of t a n n e d l e a t h e r has been estimated i n 3 Kg C r [8]. Taking i n t o account t h e a n n u a l p r o d u c t i o n of t a n n e d leather i n S p a i n (19,300 t ) [81, a n a n n u a l emission of 5'7.9 t of C r due t o tanneries i n Spain can be estimated. T h i s is a b o u t t h e same o r d e r of t h e t o t a l

I49

a t m o s p h e r i c emission d u e t o r e f use i n c i n e r a t i o n i n whole E u r o p e [7]. According t o F A 0 s t a t i s t i c s , a b o u t 5,600 lo6 s q u a r e f e e t of t a n n e d l e a t h e r a r e a n n u a l l y produced i n t h e developed c o u n t r i e s . Assuming t h a t b o t h a b o u t 75% of t h e l e a t h e r i s Cr-tanned a n d t h e mean contamination charge is t h e same as i n Spain, a y e a r l y emission of about 1,105 t of C r due t o t a n n i n g a c t i v i t i e s i n developed c o u n t r i e s can be roughly estimated. Although, on a global scale t h i s emission of C r by the t a n n i n g i n d u s t r y is r e l a t i v e l y small, i n Comparison t o the t o t a l emission of C r , and t h e C r i n t a n n e r y waste usually i s i n t h e t r i v a l e n t s t a t e , disposal of t a n n e r y waste may become a problem locally, especially i n developing c o u n t r i e s w i t h poor Cr-managing technology. T o x i c i t y of c h r o m i u m

Doubtlessly, chromium i n t h e form of C r V I is h i g h l y The danger of chromium-containing compounds f o r human health is known by occupational medicine f o r more t h a n 150 years. A selection of some key f i n d i n g s i n t h e h i s t o r y of chromium t o x i c i t y is shown i n Table 1. Already i n 1827, u l c e r a t i o n of s k i n b y chromium was described by Cumin a n d s i m i l a r o b s e r v a t i o n s were r e p o r t e d i n t h e following decades by o t h e r s ( r e f . i n [ 9 ] ) . A t t h e end of t h e last c e n t u r y a n d d u r i n g t h e f i r s t decades of t h e present, adenocarcinoma i n t h e nose, p e r f o r a t i o n of n a s a l septum, c o n t a c t d e r m a t i t i s a n d lung c a n c e r h a v e been f o u n d i n i n d u s t r i a l workers i n contact w i t h chromium ( r e f . i n [lo]). Besides the carcinogenic effects, the mutagenic and t e r a t o g e n i c p r o p e r t i e s of c e r t a i n Cr-containing compounds have deserved much r e s e a r c h i n t e r e s t ( r e f . i n [11-14]). The p h y t o t o x i c e f f e c t s of chromium h a v e been a l r e a d y described more t h a n 80 y e a r s ago [15] and, since t h e n , t h e toxic e f f e c t s of C r ions h a v e been demonstrated i n many plant species mainly i n solution c u l t u r e s t u d i e s or i n soils t r e a t e d w i t h Cr-salts 116-231. Besides o t h e r u n f a v o r a b l e f a c t o r s , t h e h i g h c o n c e n t r a t i o n of C r may p l a y a r o l e i n t h e poor performance of p l a n t s on s e r p e n t i n e soils [24, 251. Phytotoxic toxic.

150

Table i Some h i s t o r i c a l d a t a on chromium toxicity 1827 Chrome ulceration of s k i n (Cumin) 1890 Adenocarcinoma i n nose and p e r f o r a t i o n of nasal septum (Newman) 1910 Phytotoxicity of chromium salts 1925 Contact d e r m a t i t i s by chromium compounds (ParKhurst) 1932 Lung cancer i n i n d u s t r i a l workers (Alwers e t al.) as occupational d i s e a s e i n 1936 Lung cancer recognized chromate workers i n aermany

e f f e c t s of

Cr

from atmospheric emissions [26, 271 a n d t a n n e r y

e f f l u e n t s [28] have been occasionally reported. Chromium I11 is t h o u g h t t o be a b o u t 100 t o 1000 times l e s s t o x i c t h a n C r V I [14]. This low t o x i c i t y seems principally due t o t h e low capacity of C r I11 t o p e n e t r a t e biological membranes. If t h i s d i f f i c u l t y i s experimentally overcome, genotoxic e f f e c t s can a l s o be observed f o r C r I11 compounds [I,?]. Essentiality and beneficial ef fecto of chromium Chromium I11 is considered a n essential element f o r

men and animals, since e a r l y worK from Mertz and Schwartz [29] has proposed C r I11 a s t h e a c t i v e component of t h e glucose tolerance f a c t o r . Deficiency of C r i n malnourished c h i l d s and a f t e r prolonged p a r e n t a l n u t r i t i o n has been observed ( r e f . i n C301). Chromium also seems t o play a role i n l i p i d metabolism [31]. Although, t h e role of C r I11 i n glucose tolerance has been questioned [32] and t h e possibility t h a t C r I11 a c t s by b i n d i n g i n s u l i n t o s p e c i f i c r e c e p t o r s h a s been dismissed, r e c e n t s t u d i e s suggest t h a t C r I11 may be implicated i n intermolecular i n t e r a c t i o n s i n i n s u l i n , causing m a j o r changes i n t h e conformation and aggregation i n insulin and influencing t h e a c t i v i t y of t h e complex [33]. A beneficial influence of C r also be r e l a t e d t o its protective e f f e c t a g a i n s t oxidative damage by acting as a oxygen r a d i c a l scavenger [34], Chromium is not considered as a n essential n u t r i e n t f o r p l a n t s [35]. N e v e r t h e l e s s , b e n e f i c i a l e f f e c t s of low C r may

151

Table 2 Maximum p e r m i s s i b l e c o n c e n t r a t i o n s (mg k g - l ) of C r i n s o i l s , sludge and compost proposed i n d i f f e r e n t c o u n t r i e s [42-471. s o 11 50 + 2 x X c l a y ( N L ) 75 ( C H ) 100 ( D ) 100-200 (CEC) 100 (pH<7) ( E ) 150 (pH>7) ( E l 600 ( E & W) 2000 C r I11 ( I ) 15 C r V I ( I )

Sludge 100 200 500 600 1000 1500 1000

(DK)

( N , F) ( A , B, NL) (G) (pH<7) (pH>7)( E ) (CH, FL, F, S ) 1200 ( D ) 1000-1750 (CEC) 2000 (F)

Compost 150 (CH) 200-300 ( D ) 750 ( E ) 500 C r 111 ( I ) 10 C r V I ( I )

A, A u s t r a l i a ; B, Belgium; CEC, Commission European Community; CHI Switzerland; D, Germany; DK, Denmark, El Spain, E & W, England 8 Wales; F, France; FL, Finnland; G I Greece; I , I t a l y ; N, Norway; NL, Netherlands; S, Sweden.

concentrations have been found i n b o t h s o i l and solution grown The mechanism of this plants (ref. i n [36]; [37-391). beneficial e f f e c t is not clearly e s t a b l i s h e d . Improvement of mineral nutrition, water relations and assimilate translocation, as well a s t h e a n t i f u n g a l p r o p e r t i e s of C r have been made responsible f o r g r o w t h stimulation by C r ( r e f . i n [39-411). Regulations concerning Chromium Because of t h e p o t e n t i a l danger of c e r t a i n Cr-containing compounds f o r human h e a l t h , legal r e s t r i c t i o n s o r guidelines f o r C r concentrations i n t h e ambient a i r of worKing places, a s well as f o r d r i n k i n g w a t e r and i n d u s t r i a l e f f l u e n t s have been now adopted i n many c o u n t r i e s . The occupational e x p o s u r e l i m i t s f o r C r i n ambient a i r and d r i n k i n g water s t a n d a r d s o r guidelines f o r Cr generally a r e q u i t e s i m i l a r i n t h e d i f f e r e n t European c o u n t r i e s a n d t h e USA 1421. Contrastingly, still a t p r e s e n t t h e environmental r i s k of t h e deposition of C r containing residues i n s o i l s and t h e maximum allowable C r

152

Table 3 Arguments a g a i n s t environmental t o x i c i t y of chromium 1.

The l e s s t o x i c Cr I11 form i s t h e main o x i d a t i o n

2.

C r i n soils. Chromium V I

3. 4. 5. 6.

state

of

added t o s o i l s i s g e n e r a l l y reduced t o Cr I11 w i t h i n s h o r t p e r i o d of time S ever al s t u d i e s d i d n o t observe a d v e r s e e f f e c t s of Cr i n p l a n t s a t s o i l c o n c e n t r a t i o n s up t o 500 or 1000 mg K g - 1 . The low c a p a c i t y of p l a n t s t o t r a n s l o c a t e C r t o upper or other edible parts. The f a c t t h a t C r is an e s s e n t i a l element for men Cr concentrations The beneficial effects of low o c c a s i o n a l l y observed i n c r o p p l a n t s .

c o ncentr ations i n farm l an d a r e subjected t o controversy. Table 2 shows a s e l e c t i o n of t h e d i f f e r e n t l i m i t a t i o n s of C r c o n c e n t r a t i o n s i n sewage sludges, composts a n d s o i l s f o r a g r i c u l t u r a l u s e t h a t h a v e been a d o p t e d i n d i f f e r e n t countries. The r a n g e of p e r m i s s i b l e l i m i t s is c o n s i d e r a b l e a n d generally u n r e l a t e d t o t h e soil t y p e a n d t h e Cr concentrations found i n t h e soils of t h e d i f f e r e n t c o u n t r i e s (compare w i t h values given in [48]. The large d i f f e r e n c e s between permissible co n ce n t ra t i o n s not only r e f l e c t t h e i n c l u s i o n of nonscientific interests, but also the poor scientific knowledge abo u t t h e long-term environmental behavior of Cr i n soils. Cr iticis m on these r e s t r i c t i o n s mainly a r i s e from t h e leather industry. Although a main aim of the last I n t e r n a t i o n a l Congress of t h e leather a n d t a n n i n g i n d u s t r y held in B a r c e l o n a (1991) w a s t h e presentation of new technologies f o r minimizing t h e environmental impact of t h e l e a t h e r i n d u s t r y [49], d u r i n g discussions it was claimed b y some s e c t o r s t h a t t h e h i g h C r concentrations i n t a n n e r y wastes a r e harmless, and t h a t t a n n ery sludge may be used without harm as nitr ogen source f o r c r o p p l an t s. T h e arguments t h a t most frequently are brought up against an environmental s i gnif icance of Cr-containing wastes a r e l i s t e d i n t a b l e 3.

I53

P oints 5 a n d 6 i n t h i s list, r e f e r r i n g t o e s s e n t i a l i t y a n d benef icial e f f e c t s of C r , h a v e been p r o v e d by s e v e r a l i n v e s t i g a t i o n s , as i n d i c a t e d above, b u t t h e y a r e e x t r e m e l y weak ar gument s a g a i n s t t h e environmental s i g n i f i c a n c e of C r contamination. Neither essentiality nor beneficial effects allow t o draw an y conclusion a b o u t t h e toxic c h a r a c t e r of a n element. Copper, manganese, cobalt, nickel o r z i n c a r e some r e p r e s e n t a t i v e examples. Among t h e o t h e r f o u r p o i n t s , t h e low environmental t o x i c i t y of C r ,111 i n comparison t o C r V I a n d t h e s t r o n g tendency of C r V I t o be reduced t o C r I11 u n d e r most of t h e p r e v a i l i n g environmental conditions a r e by f a r t h e most important and sound arguments. Taking i n t o account t h e p E values of s o i l s a n d t h e pE values f o r r edu ct i o n / o x i d a t i o n change of C r VI/CrIII given by Sposito [50], t h e o x i d at i o n of C r I11 t o C r VI is possible i n oxidized soils. Doubts a b o u t t h e innocuousness of waste C r i n soils also a r i s e from i n v e s t i g a t i o n s which h a v e shown t h a t u n d e r c e r t a i n circumstances C r I11 can be oxidized t o C r V I i n soils [51, 521 an d t h a t , a t l e a s t i n C r s e n s i t i v e p l a n t s s u c h a s bean, t a n n e r y s l u d g e may c aus e p h y t o t o x i c i t y [53, 541. Moreover, t h e assessment of t h e d a n g e r of C r f o r t h e environment i s hampered by t h e lack of s t a n d a r d i z e d methods f o r t h e d e t e r m i n a t i o n of C r I11 a n d C r VI i n s o i l s o l u t i o n s a n d for es timat i o n of t h e v'plant-available*' C r (reviewed i n r e f . 145, 55]), a s well a s by t h e need f o r long-term s t u d i e s on C r s peciat i o n i n soils t r e a t e d w i t h waste cr. S u r e l y a b e t t e r knowledge of both t h e environmental chemistry of C r and t h e mechanisms of C r e f f e c t s i n p l a n t s a r e n e c e s s a r y f o r e s t a b l i s h i n g environmentally u s e fu l guidelines f o r permissible c o ncentr ations of C r i n soils. I n o r d e r t o i l l u s t r a t e t h e complex response of p l a n t s t o C r u n d e r d i f f e r e n t experimental c o n d i t i o n s , i n t h e p r e s e n t s t u d y w e r e p o r t a summary of r e s u l t s from v a r i o u s of o u r i n v e s t i g a t i o n s p erfo rm ed w i t h C r s a l t s o r t a n n e r y s l u d g e e i t h e r i n t h e laboratory o r i n t h e field.

I54 Table 4 S e l e c t e d c h a r a c t e r i s t i c s of s o i l s and sludge used i n pot and f i e l d experiments. If not otherwise s t a t e d , values a r e f o r a i r d r i e d s o i l s and oven-dried sludge. S o i l 1 (pot study)

PH (H20) o r g an i c matter organic C carbonates CEC (meq/ 1 OOg ) d r y matter C r (HF/HC 104) (mg/Kg 1 C r VI

8.05

2. 26% 1. 31% 35% 8 . 42

-

60

n. d

Soil 2 ( f i e l d study)

Sludge

7. 98 1. 6% 0.93% 21% 12. 0 3

-

45 n. d.

65. 06% 32. 53%

17% 16000

n. d.

n.d. not d e t e c t a b l e

WTEilIhLs AND =HODS

Three types of experiments were performed using e i t h e r bean ( P h a s e o l u s v u l g a r i s L.) o r maize ( Z e a m a y s L.) plants, grown on s o i l w i t h Cr-containing tannery sludge o r i n n u t r i e n t solution w i t h C r VI o r C r I11 salts.

Solution culture experiments: Bean ( P h a s e o l u s v u l g a r i s L.) o r maize ( Z e a m a y s L.) p l a n t s were grown on n u t r i e n t solutions w i t h s u f f i c i e n t o r d e f i c i e n t Fe supply and w i t h o r without C r i n t h e f o r m of C r V I or C r I11 salts. arowth, chlorophyll content and structural and u l t r a s t r u c t u r a l aspects were considered (methods i n r e f . [21, i.

23,

38,

39)).

Potted soil: Bean plants were grown on potted soil ( v e r t i c u s t o c h r e p t , pH 8.1) amended w i t h 9, 17 or 34 K g / t t a n n e r y sludge (Cr conc. 0.27% f.w.). Different Cr-fractions i n t h e s o i l and plant growth were analyzed (methods i n r e f . [53)). 2.

155

140 120

%

4 I

CONTROL

80 60 40 20 0

ROOT

PL

DRY WEIGHT Fig.

TL

PL

TL

CHLOROPHYLL

1

Toxic e f f e c t of 1 p M C r V I on d r y weight and c h l o r o p h y l l c o n t e n t of bean p l a n t s grown i n n u t r i e n t s o l u t i o n s ( r e l a t i v e Z of c o n t r o l s without C r ) . PL = Primary l e a v e s ; TL = values, Trifoliolate leaf.

experiments: Maize p l a n t s were grown i n t h e f i e l d on soil a t p H 7.9 (calcixerollic x e r o c h r e p t ) amended w i t h 0, 25 o r 50 t / h a of t a n n e r y sludge ( C r conc. 0.27 f.w.). C r a n d Fe f r a c t i o n s i n t h e soil a n d Cr a n d Fe conc. i n t h e 4 t h leaf were analyzed. 3.

Field

I561.

Some c h a r a c t e r i s t i c s of t h e soils a n d t h e sludge used i n pot a n d f i e l d s t u d i e s a r e indicated i n table 4. RESULTS AND DISCUSSION

Solution culture studies Bean p l a n t s exposed i n n u t r i e n t solution t o 1 pM C r V I showed s i g n i f i c a n t growth r e d u c t i o n (Fig. 1) a n d Cr-toxicity 1.

156

260

%

rl

200

......... :.:.:.:.:.:.:.

10uM Fe

Crlli

+

Crlll

- Fe

............... .:.:.:.:.:.:.:. ........ ............... ...............

160

100

60

0

Root

PL

DRY WE I GHT

TL

PL

TL

CHLOROPHYLL

Fig. 2 Dry w e i g h t and c h l o r o p h y l l c o n c e n t r a t i o n i n bean p l a n t s grown in n u t r i e n t s o l u t i o n w i t h s u f f i c i e n t ( b l a c k columns) or d e f i c i e n t Fe (shaded columns) c o n c e n t r a t i o n s ( r e l a t i v e v a l u e s ; of c o n t r o l s w i t h s u f f i c i e n t or d e f i c i e n t Fe s u p p l y ) . PL = P r i m a r y l e a v e s ; TL Trifoliolate leaf.

symptoms i n t h e f o r m of c h l o r o s i s i n t h e young t r i f o l i o l a t e l e a v e s ; t h e c h l o r o p h y l l c o n c e n t r a t i o n s i n t h e s e l e a v e s was s e v e r e l y r e d u c e d (Fig. I). S c a n n i n g e l e c t r o n microscopy of r o o t s , showed s u r f a c e damage d u e t o C r V I ( P l a t e I A ) . These corrosive lesions s u r e l y h a m p e r a b s o r p t i o n processes i n p l a n t r o o t s . A l t h o u g h , C r t r a n s l o c a t i o n t o u p p e r p l a n t p a r t s is r e l a t i v e l y low [21, 221, transmission e l e c t r o n microscopy revealed damage of c h l o r o p l a s t s d u e t o C r V I s u p p l y ( P l a t e I B). Poor g r a n a stacking, dilated thylakoid membranes and abundant p l a s t o g l o b u l i were o b s e r v e d , These e f f e c t s may be caused i n d i r e c t l y b y t h e t o x i c e f f e c t of C r i n r o o t s [23].

157

PRIMARY LEAF

TRIFOLIOLATE LEAF

140156

120

i

70

DRY WEIQHT

CHLOROPHYLL DRY WEIGHT

10

20

CHLOROPHYLL

Fig. 3 Dry weight and chlorophyll content in leaves of bean plants grown in nutrient solution with 10, 20 or 70 pM Cr 111. (relative values x of control). Bean plants exposed to 1 pM Cr I11 in nutrient solution did not reveal ultrastructural damage. Not a decrease, but a significant increase of dry weight within both roots and leaves was observed (Fig 2). The chlorophyll concentration of young leaves was not significantly influenced. The beneficial effect of the low Cr I11 concentration was even more pronounced in plants cultivated in Fe deficient nutrient solutions. Under those conditions both dry weight and chlorophyll concentrations were significantly increased in comparison to the Fe-deficient controls without Cr supply (Fig. 2). Electron microscopy observations of chloroplasts from Fedeficient bean plants without Cr I11 supply (Plate 1 C ) and from Fe-deficient plants grown with 1 pM Cr I11 in the

158

CONTROL

8.7

:x. 3%

RES. 66%

17,4

34.7 . 23% E.R.1% EX, 1%

RES. 48%

RES. 44%

Fig. 4: D i s t r i b u t i o n of C r i n the s o i l ( v e r t i c u s t o c h r e p t ) used i n p o t s t u d i e s w i t h bean p l a n t s . Control s o i l , without sludge had a t o t a l Cr c o n c e n t r a t i o n of 60 mg/Kg d . w . ; s o i l amended w i t h t a n n e r y sludge a t d i f f e r e n t r a t e s c o n t a i n e d 89, 112 and 160 mg/kg d.w. of t o t a l Cr. (modified from [53]). EX = Exchangeable; E. R. = E a s i l y r e d u c i b l e ; M. R. z Moderately r e d u c i b l e ; ORG. = Organic; RES. = Resiudual.

n u t r i e n t s o l u t i o n (Plate 1 D) clearly demonstrate t h e beneficial influence of C r on t h e chloroplast u l t r a s t r u c t u r e . Chloroplasts f r o m t h e Fe-deficient p l a n t s e x h i b i t e d a d a r k s t r o m a , a d i s o r g a n i z e d thylaKOid s y s t e m a n d a b u n d a n t C h l o r o p l a s t s f r o m Fe-deficient p l a s t o g l o b u l i ( P l a t e 1 C). p l a n t s w i t h C r 111 supply had a b e t t e r organized membrane system and much less plastoglobuli (Plate 1 D). Considerably h i g h e r C r I11 concentrations were r e q u i r e d for inducing growth i n h i b i t i o n i n hydroponically grown bean p l a n t s (Fig 3). When exposed t o solutions w i t h a d e q u a t e Fe concentration, t h e supply of 20 V M C r or h i g h e r concentrations s i g n i f i c a n t l y decreased d r y m a t t e r p r o d u c t i o n , while t h e

159

% loo[

I-

T

€

80

r

2

=

0.95

80

t n

a

40

LL

6

w

20

0 20

0

40

t

t

t

Q

0

cr

60

17

34

Sludge (Wt)

Fig 5: Negative correlation between leaf dry weight and acid oxalate extractable Cr from potted soil amended with tannery sludge Table 5 Chromium (pg concentrations exposed for 14 solution with 0

d. w. ) , chlorophyll and carotenoid (mg g - l f.w.) in the 4th leaf of maize plants d to 0 or 1 pM Cr I 1 1 in 10% Hoagland nutrient or 10 pM Fe as Fe-EDTA. g-l

Treatment Fe

Cr I 1 1

no no

1 crM

chlorophyll

carotenoids

5. 14a 6.26a

0.303a 0.264a

0.070a

no

a. 2aa

1 pM

6.34a

0.330a 0.652b

0.139b

-10 pM 10 pM

no

Cr

values within a column followed by the significantly different (p > 0.05)

0.077a 0.0a2a

same letter

are not

160

A

CONTROL

E.R. 3% EX. 4%

M.R. E.R. I% EX. 1%

Fig. 6: D i s t r i b u t i o n of C r i n t h e s o i l ( c a l c i x e r o l l i c xerochrept) u s e d i n f i e l d s t u d i e s w i t h maize p l a n t s . C = c o n t r o l s o i l (without s l u d g e ) ; A = Control s o i l p l u s 2 5 t / h a of t a n n e r y sludge; B = Control s o i l p l u s 50 t/ha o f t a n n e r y Sludge. EX = Exchangeable; E. R. = E a s i l y r e d u c i b l e ; M. R. = Moderately r e d u c i b l e ; ORG. = Organic; RES. = Residual

in young leaves was not chlorophyll concentration s i g n i f i c a n t l y a f f e c t e d by t h e s e h i g h C r I11 concentrations. The beneficial e f f e c t of low C r I11 c o n c e n t r a t i o n s i n n u t r i e n t solution on p l a n t p e r f o r m a n c e u n d e r Fe-def i c i e n c y conditions is not a s p e c i f i c response of bean plants. Similar r e s u l t s were obtained i n maize p l a n t s grown i n hydroponics w i t h moderate Fe d e f i c i e n c y (10 pM), b u t n o t i n s o l u t i o n s w i t h o u t Fe supply (Table 5). Transmission electron micrographs showed t h e b e n e f i c i a l e f f e c t of low C r I11 c o n c e n t r a t i o n s on t h e u l t r a s t r u c t u r e of chloroplasts from maize leaves (Plate 2 A a n d B). Chloroplasts from maize p l a n t s grown w i t h a d e f i c i e n t Fe c o n c e n t r a t i o n

161

100

Sludge (t/ha)

0

26

140

180

60

Fig. 7: P o s i t i v e c o r r e l a t i o n between s l u d g e s u p p l y and a c i d o x a l a t e e x t r a c t a b l e Cr from s o i l amended w i t h t a n n e r y s l u d g e s i n f i e l d study.

e x h i b i t e d poorly organized a n d d i l a t e d thylaKOid membranes (Plate 2 A). In t h e presence of I pH Cr 111, a b u n d a n t g r a n a s t a c k s were fo u n d (Plate 2 B). The b e n e f i c i a l e f f e c t of Cr I11 w a s als o e v i d e n t fro m l i g h t microscopy o b s e r v a t i o n s . The leaves from p l a n t s grown u n d e r Fe-deficiency w i t h o u t C r I11 had less cells a n d less c h l o r o p l a s t s t h a n t h o s e (Plate 2 C) from p l a n t s w i t h Cr I11 supply (Pl a t e 2 I)). Our hydroponic c u l t u r e s t u d i e s c l e a r l y d e m o n s t r a t e t h e d i f f e r e n t degree of p h y t o t o x i c i t y of Cr I11 a n d Cr VI. A t t h e lowest Cr c o n c e n t r a t i o n assayed, 1 pM, Cr V I h a d t o x i c e ff ects , a f f e c t i n g r o o t growth a n d c h l o r o p l a s t u l t r a s t r u c t u r e . The benef icia l e f f e c t of Cr I11 a t t h e same c o n c e n t r a t i o n seemed t o be r e l a t e d t o improved i r o n n u t r i t i o n , y e t t h e

162

stimulating Fe-def i c i e n c y

effects were conditions.

2. Studies w i t h tannery

especially

enhanced

under

sludge-amended s o i l s Sequential e x t r a c t i o n of C r from potted soil which was amended w i t h C r - r i c h t a n n e r y sludge showed t h a t sludge C r w a s i n c o r p o r a t e d mainly i n t h e moderately r e d u c i b l e f r a c t i o n , e x t r a c t a b l e w i t h a c i d oxalate (ammonium oxalate + 0.2M oxalic acid). A s u b s t a n t i a l i n c r e a s e of C r c o n c e n t r a t i o n w a s a l s o observed i n t h e organic f r a c t i o n , e x t r a c t e d w i t h h o t 30% H202 + 1M ammonium a c e t a t e a t pH 2.5 ( F i g . 4). No s i g n i f i c a n t i n c r e a s e of C r e x t r a c t a b l e w i t h d i s t i l l e d w a t e r , p h o s p h a t e b u f f e r , D T P A o r a c e t i c a c i d was detected [53]. Bean p l a n t s c u l t i v a t e d on t h e sludged s o i l showed s i g n i f i c a n t growth reduction. Decrease of leaf d r y weight was significantly correlated t o the acid oxalate e x t r a c t a b l e C r concentration (Fig. 5). Bean p l a n t s showed C r t o x i c i t y symptoms i n t h e form of chlorosis i n young leaves which were q u i t e similar t o t h o s e observed i n n u t r i e n t solution s t u d i e s w i t h C r V I salts. A t r a n s i e n t oxidation of C r I11 t o C r V I i n loamy clay soil amended w i t h chrome l e a t h e r f i b e r s h a s been observed. Addition of e i t h e r MnS04 o r lime (CaC03) increased t h e oxidation r a t e [57]. Our experimental soil h a d a loamy clay t e x t u r e a n d a v e r y h i g h c a r b o n a t e content. We f a i l e d t o d e t e c t a n y C r VI; b u t as bulk soil was analysed, we cannot exclude t h a t C r V I w a s formed i n t h e rhizosphere soil causing growth r e d u c t i o n and t h e severe phytotoxicity symptoms. The d i s t r i b u t i o n of C r from t a n n e r y sludge i n alKaline s o i l u n d e r field conditions, f o u r w e e K s a f t e r sludge s u p p l y is shown i n f i g u r e 6. A s i n t h e pot study, C r was incorporated mainly i n t h e moderately reducible f r a c t i o n e x t r a c t a b l e w i t h a c i d oxalate. While t h e C r concentration i n t h e exchangeabie, easily reducible a n d organic f r a c t i o n s w a s n o t s i g n i f i c a n t l y changed. A s i g n i f i c a n t l i n e a r c o r r e l a t i o n between sludge supply o r t o t a l soil C r concentration and acid oxalate e x t r a c t a b l e C r w a s observed ( F i g 7). T h e C r c o n c e n t r a t i o n analysed i n t h e f o u r t h a n d the youngest leaves of t h e maize

163

Table 6 Chromium c o n c e n t r a t i o n (mg/Kg d . w . ) o f l e a v e s and g r a i n s of maize p l a n t s grown on s o i l amended w i t h t a n n e r y s l u d g e ) 4th l e a f

Contro 1 25 t / h a 50 t / h a

1.48 2 0.37 2 . 49 f 0 . 26 3.09 ! 1 . 4 5

youngest l e a f

grain

1.15 t 0.43 1. 32 f 0. 4 2 1.49 ! 0. 2 4

0. 2 4 ! 0. 17 0. 2 0 ! 0 . 0 6 0. 15 ! 0. 0 5

p l a n t s s l i g h t l y i n c r e a s e d w i t h t h e sludge s u p p l y (Table 5 ) . But differences were hardly significant. Chromium c o n c e n t r a t i o n i n t h e g r a i n was u n a f f e c t e d by sludge supply ( T a b l e 6). The supply of 25 t / h a t a n n e r y sludge s l i g h t l y i n c r e a s e d t h e leaf d r y weight of t n e p l a n t s , while a small decrease of growth was observed f o r t h e h i g h e s t sludge supply i f compared t o p l a n t s grown on s o i l w i t h o u t sludge b u t a n e q u i v a l e n t amount of nitrogen supply i n t h e form of ammonium s u l f a t e ( F i g 8). Nevertheless, t h e d i f f e r e n c e s between t r e a t m e n t s were n o t s t a t i s t i c a l l y s i g n i f i c a n t . A s i g n i f i c a n t c o r r e l a t i o n between g r a i n p r o d u c t i v i t y a n d leaf i r o n s t a t u s was observed (Fig 9). P l a n t s grown on soil amended w i t h 25 t / h a sludge e x h i b i t e d a s i g n i f i c a n t l y h i g h e r leaf i r o n c o n c e n t r a t i o n t h a n b o t h c o n t r o l p l a n t s ( w i t h o u t sludge supply) a n d p l a n t s grown w i t h 50 t / h a s 1u d g e . Our r e s u l t s on maize grown on t a n n e r y sludge amended soil a r e i n l i n e w i t h worKs from o t h e r s which d i d not detect toxic e f f e c t s of t a n n e r y waste i n f i e l d grown p l a n t s [54,58,59]. Grain p r o d u c t i v i t y on experimental p l o t s amended w i t h sludge w a s n o t s i g n i f i c a n t l y d i f f e r e n t f r o m t h a t achieved on plots receiving e q u a l amounts of n i t r o g e n i n t h e form of ammonium s u l f a t e . The p o s i t i v e e f f e c t o f t h e 25 t / h a sludge t r e a t m e n t , which was c o r r e l a t e d t o leaf i r o n Concentration, confirms o u r p r e v i o u s r e s u l t s w i t h maize a n d bean p l a n t s grown i n Fed e f i c i e n t s o l u t i o n s w i t h low C r I11 c o n c e n t r a t i o n s . Our f i e l d d a t a s u p p o r t t h e h y p o t h e s i s t h a t t h e a v a i l a b i l i t y o f low C r I11 c o n c e n t r a t i o n s improve t h e i r o n n u t r i t i o n of p l a n t s .

164

120

% I

0

26

60

Sludge ( t /ha)

Fig. 8: Leaf d r y weight of maize p l a n t s grown i n t h e f i e l d s t u d y on s o i l amended w i t h 0, 25 or 50 t / h a of sewage sludge ( r e l a t i v e values Z of control).

Nevertheless, t h i s positive e f f e c t on observed w i t h t h e h i g h sludge dose.

Fe

In

n u t r i t i o n w a s not case, leaf C r

this

concentration was increased and probably a n i n h i b i t i o n of Fe translocation occurred. In p o t s t u d i e s w i t h beans we proposed t o e s t a b l i s h c r i t i c a l t o x i c i t y concentrations f o r waste Cr based on t h e acid oxalate e x t r a c t a b l e s o i l Cr concentration [53]. Under t h e f i e l d conditions used i n t h i s s t u d y , Cr f r o m t a n n e r y sludge was also incorporated mainly i n t o t h e moderately r e d u c i b l e soil f r a c t i o n . W i t h t h e h i g h e s t sludge dose, 147 vg g-l Cr w a s found i n this s o i l fraction. But as t h e decrease o f p r o d u c t i v i t y was n o t s t a t i s t i c a l l y s i g n i f i c a n t no c r i t i c a l toxicity concentration can be fixed.

165

t /ha

0

100

200

300

400

600

Fe (ug/g)

Fig. 9 : Correlation between grain production c o n c e n t r a t i o n ( 4 t h l e a f ) i n maize p l a n t s s l u d g e amended s o 1 1.

and grown

leaf iron on t a n n e r y

CONCLUSIONS

as a h a z a r d f o r e n v i r o n m e n t a l h e a l t h is n o t c l e a r l y e s t a b l i s h e d . Our r e s u l t s show t h a t i n s o l u t i o n c u l t u r e C r V I , a t c o n c e n t r a t i o n s as low as 1 pM, can get s e v e r e l y t o x i c t o bean p l a n t s . Twenty o r more times h i g h e r c o n c e n t r a t i o n s of C r i n t h e t r i v a l e n t s t a t e a r e r e q u i r e d t o induce growth r e d u c t i o n i n t h i s species. Low C r I11 c o n c e n t r a t i o n s h a v e b e n e f i c i a l e f f e c t s on plants. The g r o w t h p r o m o t i o n b y C r I11 seems r e l a t e d t o improved Fe- n u tr it i o n Using p o t t e d soil, C r I11 from t a n n e r y waste can c a u s e s e v e r e t o x i c i t y i n t h e C r - s e n s i t i v e species P h a s e o l u s v u l g a r i s The

significance

of

Cr

166

even a t t o t a l soil Cr-concentrations t h a t a r e w i t h i n t h e range of t h e values proposed by t h e EC a n d w i t h o u t d e t e c t a b l e amounts of C r YI. Contrastingly, u n d e r field c o n d i t i o n s t h e amendment of s o il w i t h t a n n e r y s l u d g e t o t o t a l s o i l C r c o n c e n t r a t i o n s w i thin t h e range of t h e EC proposal does not seem t o cause a n y harm t o corn plants. Both u n d e r f i e l d co n d i t i o n s and i n potted soil, C r from t a n n e r y sludge w a s mainly i n c o r p o r a t e d i n t o t h e moderately r e ducible soil f r a c t i o n ( a c i d o x a l a t e e x t r a c t ) . A s t h i s C r f r a c t i o n was s i g n i f i c a n t l y c o r r e l a t e d t o t h e amount of t h e su pplied waste material a n d , moreover, i n t h e p o t t e d s o i l s t u d y a s i g n i f i c a n t c o r r e l a t i o n between growth r e d u c t i o n a n d t h e a c i d o x a l a t e e x t r a c t a b l e C r could b e e s t a b l i s h e d , we propose t h e use of t h i s C r f r a c t i o n f o r distinguishing between t h e n a t u r a l l y o cc u rri n g soil C r (mainly i n t h e s t r o n g a c i d e x t r a c t a b l e f r a c t i o n ) an d t h e C r due t o contamination from waste

dis pos a l . F u r t h e r i n v e s t i g a t i o n s on t h e long-term b e h a v i o r o f d i f f e r e n t C r species i n d i f f e r e n t soil t y p e s considering n o t only t h e bulK soil, b u t also t h e r h i z o s p h e r e f r a c t i o n h a v e t o be performed, so t h a t environmentally u s e f u l g u i d e l i n e s f o r C r i n s oils a n d Sludges f o r a g r i c u l t u r a l use can arise from a b e t t e r Knowledge of t h e behavior of C r i n soils.

P a r t of t h e experimental work was supported by DGICYT (PB PB 91-0668)

88-0234 a n d

REFgRIMCES T i l l e r KG. In : BA S t e w a r t , ed. Advances i n S o i l Science, vol 9. S p r i n g e r V e r l a g , New YorK, 1989; 113-142. 2 Haygarth PM, J o n e s KC. I n : DC Adriano, ed. Biogeo-chemistry of Trace Metals. L e w i s Publ, Boca Raton, 1992; 249-276 3 D a r r i n M. Chromium, vol I , Chemistry of chromium and i t s compounds. Reinhold P u b l i s h i n g Corp., New YorK, 1956. 4 P os chenr ied er C h , Barcel6 J, Guns&,B. C i r c u l a r F a r m a c i u t i c a (Barcelona) 1986; 290: 23-38 5 Gauglhofer J. In : E Merian, ed. M e t a l l e i n d e r Umwelt. 1

167

6 7

8

9

Verlag Chemie Welnheim, 1984; 409-421. Nriagu JO, Pacyna JM. Nature 1988; 333: 134-139. Pacyna JM, Munch J, Axenfeld F. In: J. -P. Vernet, ed. Heavy Metals in the Environment. Trace Metals in the Environment I. Elsevier, Amsterdam. 1991; 1-20, Spanish Ministry for Industry and Energy. Tecnologias biisicas aplicables a la depuracibn de 10s efluentes lxquidos de la industria del curtido de pieles y cueros. Madrid, 1984. Burrows D. In: D Burrows, ed. Chromium: Metabolism and Toxicity. CRC Press, Inc. Boca Raton, Florida. 1 9 8 3 ; . 137-163.

Langard S. In: D Burrows, ed. Chromium: Metabolism and Toxicity. CRC Press, Inc. Boca Raton, Florida. 1983; 13-30. Norseth T. In: L. Friberg, G.F. Norberg, V.B. 11 Langard S , V o w , eds. HandbooK on the Toxicology of Metals. vol. 11. Elsevier/North Holland Biomedical Press, Amsterdam. 1986;

10

185-210.

Bianchi V, Levis AG. Sci Total Environm 1988; 7 1 : 3 5 1 - 3 5 5 . Gebhart E, Rossman G. In: E. Merian, ed. Metals in the Environment. Occurrence, Analysis and Biological Relevance. Verlag Chemie, Weinheim. 1991; 617-640. 14 Gauglhofer J, Bianchi V. In: E. Merian, ed. Metals and Their Compounds in the Environment. Occurrence, Analysis and Biological Relevance. Verlag Chemie Weinheim. 1991; 12 13

853-878. 15 16 17

Konig P. Landwirt. Jahrb. 1910; 39: 775-916. Hunter TG, Vergnano 0. Ann App Biol 1953; 40: 761-777. Anderson AJ, Meyer DR, Mayer FK. Aust. J. Agric. 1973;

24:

Res.

557-571.

18 19 20

Austenfeld FA. Photosynthetica 1979;13: 434-438. Misra SG, Jaiswal PC. J Plant Nutr 1982; 5 : 755-760. Zaccheo P, Cocucci M, Cocucci S . Plant Cell Environm.

21

Barcelb J, Guns6 B, Poschenrieder Ch. Photosynthetica 1986;

22

Barcelb J, Poschenrieder Ch, Guns6 B. J Plant Nutr 1985; 8:

23

VBzquez MD,

1985;

8: 721-726.

20: 249-255. 211 -217. 1987;

59:

Poschenrieder

Ch, Barcelb J. Annals of Botany

427-428.

24 25

Proctor J, Woodell SRJ. Adv. Ecol. Res. 1975; 9 : 2 5 5 - 3 6 6 . Kinzel H, Weber M. In: H Kinzel, ed. PflanzenoKologie und Mineralstoffwechsel. Verlag Eugen Ulmer, Stuttgart, 1982;

26

Desbaumes

381-410

P,

Ramaciotti

D.

Pollut.

Atmos.

1968;

10:

224-226. 27 28 29 30 31

MatsouKa C, Kitanura TI Tahara I, et al. J. Japan. SOC. Air Pul lut. 1969; 4: 38-48 Rajamani S , Hilda A. Leather Science 1987; 34: 111-114. Mertz W, Schwartz K. Am J Physiol 1959; 1 9 6 : 614-618. Baetjer AN, Bingham DJ, Enterline PE, et al. Panel on chromium. Nat Acad Sci, Washington D.C. 1974. Riales R. In; D Shapcott, J Hubert, eds. Proceedings Conference Chromium in Nutrition and Metabolism Vol 2. Elsevier/North Holland Biochem Press, Amsterdam, 1979;

168

199-212. 32

Haylock SJ, BucKley

PD, BlacKwell

LF.

J

Inorg

Biochem

1983; 1 9 : 105-117. 33

Govindaraju K, Ramasami 1989;

34

35:

Tezuka M, 1991;42:

35 36 37

T, Ramaswamy D.

J Inorg Biochem

137-147.

Momiyama

K,

Edano

T, et al. J. Inorg Biochem

1-8.

Huffman EWD, Allaway WH. Plant Physiol 1973; 52: 72-75. Pratt PF. In: HD Chapman, ed. Diagnostic Criteria for Plants and Soils. Univ Calif, Riverside, 1966; 136-141. Bertrand D, de Wolf A. Comp Rend Acad Sci ser D 1968; 1494-1495.

38 Bonet 14: 39

Poschenrieder

Ch, Barcel6 J. J Plant Nutr

Poschenrieder Ch, Vizquez 1991;

40

A,

1991;

403-414.

MD, Bonet A et al. J Plant Nutr

14: 415-428.

Barcelb J, Poschenrieder Ch,

Guns5 B.

J Exp Bot 1986; 37:

178- 187.

Barcelb J, Poschenrieder Ch, Guns6 B. Circular FarmacEutica (Barcelona) 1987; 293: 31-48. 42 Ewers U. In: E Merian, ed. Metals in the Environment. Occurrence, Analysis and Biological Relevance. Verlag Chemie Weinheim, 1991; 687-711. 43 Hani H. In: E Merian, ed. Metals in the Environment. Occurrence , Analysis and Biological Relevance. Verlag Chemie Weinheim, 1991; 357-368. 4 4 Hofstede HI Ho G. In: JP Vernet, ed. Heavy Metals in the Environment. Trace Metals in the Environment 1. El sevier, Amsterdam, 1991; 67-94. 45 Williams JH. Chromium in sewage sludge applied to agricultural Land. Final Report. Commission of the European Communi t ie s, Luxembourg , 19 8 8. 46 EEC. Proposal for modification o f the Council Directive 8 6 / 2 7 8 / EEC with respect to chromium in sewage sludge used in agriculture. Corn Eur Communities COM ( 8 8 ) 624 final,

41

1988. 47

Real Decreto

1310/1990.

Boletin Oficial del Estado, Madrid

1. 11. 1990. 48 49

50 51 52

53

54

Angelone M, Bini C. In: DC Adriano, ed. Biogeochemistry of Trace Metals. Lewis Publ, Boca Raton, 1992; 1 9 - 6 0 Ylla-Catala A. In: J Cot, ed. Compendium o f Advanced Topics on Leather Technology. Associacibn Quimica EspaAola de la Industria del Cuero, Barcelona, 1991; 4 Sposito G. Thermodinamics of Soil Solutions. Clarendon Press, Oxford 1981. Barlett RJ, James B. J Environ Qua1 1979; 8: 31-35. Ceccanti €3, Silva S. In: A Frigerio, ed. Rischi e tossicita dell’inquinamento da metalli pesanti: cromo e mercurio. vol 1, Diesseti, Milano, 1980; 5 1 - 5 4 Guns6 B, Poschenrieder Ch, Barcel6 J. In: ML van Beusichem, ed. Plant Nutrition, Physiology and Applications. Kluwer Acad Publ, Dordrecht, 1990; 307-312. Sykes RL,, Corning DR, Earl NJ. J Am Leather Chem Assoc (JALCA) 1981; 7 6 : 102-125.

169

55 56 57 58 59

Otabbong E. Acta Agric Scand 1989; 39: 119-129. Poschenrieder Ch, Vdzquez MD, Bonet A. et al.Beneficia1 effects of low chromium 111 concentrations in iron deficient maize plants. Suelo y Planta 1992; 2: 605-615. Gauglhofer J. J SOC Leather Techn Chem 1985; 70: 11-13. Kick H, Braun B. Landwirtsch Forsch 1977; 30: 160-173. Shivas SAJ. J Am Leather Chem Assoc 1980; 70: 288-298.

170

Plate 1 Influence of Cr on ultrastructure of bean plants. A:SEM from the surface of a root exposed to 1 pM Cr VI. B: TEM showing chloroplast from a plant damaged by Cr VI. C: TEM showing chloroplast from a Fe-deficient plant. D: TEM showing chloroplast from a Fe-deficient Plant receiving i pM Cr 111.

I71

Plate 2 Beneficial influence of Cr I 1 1 on Fe-deficient maize plants. A: TEA showing chloroplast from a Fe-deficient plant. B: Chloroplast from a Fe-deficient plant treated with 1 ,LIH Cr 1 1 1 . Note the better developed thylaltoid system. C 8 D: Light micrographs showing transversal sections of leaves from Fedeficient maize without Cr I 1 1 ( C ) and with 1 V M Cr 1 1 1 .

This Page Intentionally Left Blank

173

INTERACTIVE EFFECTS OF THE APPLICATION OF DIFFERENT Cd FORMS AND AN ACIDIFYING AGENT ON PLANT AVAILABLE METALS AND POSTHARVEST SOIL EXTRACTABILITY R . Nogales, D. Heivis, J . Soto and F. Gallardo-Lm

Estacion Experimental del Zaidin. C.S.I.C. Box: 41 9, 18080-Granada, Spain.

1. INTRODUCTION Although tlie transfer of heavy metals to tlie soil-plant system by tlie application of sewage sludge has been investigated for some time, there are still inany unanswered questions with regard to this process, which currently continues to attract interest (1,2). One issue awaiting resolution is Cd contaminatiori, a worrisome problem because of the toxicity of this element to plants (31, which readily absorb, :iccumulate, and transmit it to the food chain. When consumed by Iiuinans, Cd shows affinity for certain organs, where it accumulates and may cause health problems (4-6).

A review of the main aspects of Cd contaminatiorr studied to date shows that the relation between increasing doses of sewage sludge and parallel increases in Cd concentration in plant tissues has been well tiocuinented, i n plants grown 011 both farinlands (7-9) and drastically disttirbed lands (10). However, few studies have examiiied the distribution of Cd uptake by different parts of the plant, and w e were tin~ibleto find any studies dealing specifically with wheat ( 1 1-13). Considerable research efforts have bee11 devored to elucidating the effects of sewage sludge on Cd extractilbility in soils; most such studies provide results obtained with extractant solutions containing DTPA (14-16). The effects of sewage sludge on Cd movement and dishibution in the soil profile have also been studied (17-19). However, little information is ;iv:\ilable on specific effects of the addition of Cd-enriched sewage sludge on diitei-ent soil and plant painineters. Street et al. (20) examined the interactive effects of the application of phosphorus anti Cd-enriched sewage sludge on Cd uptake by maize, and on the solubility of soil Cd. Mahler et 31. (21) added Cd-enriched sewage sludge to acid and alkiiline soils aiid then determined the uptake of Cd and other heavy metals by different crops; this study also milyzed these elements in soils, using both water and DTPA as extractants. 'There appear to be not studies of the effects of the simultaneo~isapplication of Cd-enriched s e ~ a g sludge e and an acidifying agent, which would hypothetically induced the mobilization of heavy mctds, thus affecting their distribution in the plant and their extractability from tlie soil. The present study w;is desigiied t o investigiite the interactive effects of the application of Cd-enriched sewage sludge or and inorganic Ccl salt, together with an acidifying agent, on the distribution of Cci aiid otlier liravy met:ilh i n wheat plants, and on the partial or total extractability of these elements in soil.

174

2. MATERIAL AND METHODS

2.1. Soil and wastes The soil used in this study was collected from the surface layer (0-15 cm) of a calcaric Fluvisol. Some chemical and physical characteristics of the soil were as follows: sand: 53.5%, silt: 43.4%, clay: 14%, pH=7.04, CE,,= 0.31 dS/m, organic matter= 3.32%, N= 0.32%, P available: 44.8 mg kg-', CEC: 10.7 meq IOOg-'. Sewage sludge was collected froiii a drying bed at the Orgiva (Granada), wastewater treatment plant. The sludge was dry when collected, and was crushed to pass a 2 mm sieve. Selected properties of this sludge were pH= 7.08, Organic C= 24.04%, N= 2.5%, C/N= 9.6, P= 0.78%, K= 1.34%, Ca= 20%, Mg= 2.6%. The concentrations of total (4M NOjM-extractable), AAA-EDTA-extractable and DTPAextractable metals in the soil and the sludge are given in Table 1.

Table 1 Concentrations of metals in soil and sludge 4M HN03-extractable mg kg-' (Total)

Soil Sludge

Cd

cu

Zn

Mn

Ni

Pb

0.29

17

68

284

13

41

4

65

810

165

19

1104

AAA-EDTA-extractable mg kg-'

Soil Sludge

0.15

6

22

176

3

34

2

37

166

39

7

517

DTPA-extractable mg kg-' Soil

0.13

4

7

48

1.5

20

Sludge

0.33

6

166

9

1.9

43

The acidifying agent was watewuter from olive oil processing (alpechin) obtained from the Sari Rogelio Cooperative Plant in Illora (Granada). Its main physicochemical characteristics were as follows= pH= 4.8, dry residue (105")= 90 g l-', N= 126 mg l-', P= 24 mg I-', K= 560 mg 1-', Mn= 0.07 ing I-', Cu= 0.1 1 mg 1-' and Zn= 0.26 mg I-'.

175

2.2 Experimental layout and procedure To obtain a high concentration of 10 mg kg-l soil, cadinitmi, as Cd2C1.2.5H,0, was added to soil in two different forms: Direct addition to tlie solution and incubation of the soil sample at 28°C for I ino (Salt-Cd), or previous mixing and incubation ( I moj with sewage sludge ( S l u d ~ - C d ) The . amount of sewiige sludge applied to the soil was the equivalent to 200 Mg ha- . Two further treatiiients including cadliiium and the acidifying agent were prepared: Salt-Crl-A and Sludge-Cd-A. Wastewater wiis added to bring the moisture content up to 25% water holding capacity of [lie soil. 'Two treatments including soil alone (Soil) or soil mixed with sewage sludge at rate of 200 Mg l i i ' (Sludge) were used as controls. Each treatment was replicated three times. Winter wlieiit (Triric/mi ~ ~ ~ . s t i iLv. ~ m Mr.si/) w a s grown in a greenhouse i n 0.5 liter pots. Each pot contained 400 g of soil alone or treated previously with cadmium, sewage sludge and the acidifying agent, depending of tliz treatinent assayed. Before sowing, chemical fertilizers were applied to the soil at rates suplying 130, 1x0 antl 2x0 kg ha-' of N, P205 and K,O respectively. Additional N (70 kg Iia-') w a s adtled to tlie soil 25 and 50 days after sowing. Twenty wintei- wheat seeds were placeti in eacli pot, and 7 days after ennergence the stand was thinned to five plants per pot. A l l tlie pots were subirrigated daily to field capacity with distilled water. Wheat plants were liarvested 75 days after gerinination when tlie grain had matured. Straw and wheat grain were washed sequt'riti;illy in distilled water acidified with HCI, distilled water and finally iri deionized wtiter. Plant inaterial was then oven-dried at 6OoC, weighed and g~olllldi n a stainless-steel blender. After plant mateiial was harvested, soil samples were collected from eacli pot, iiir-drizd antl Ilomogenized. ('11

2.3. Laboratory analyses Plant inaterial (straw and grain) were digested with HN03-HCI0, (1:l j until clear. All samples were analyzed for Cd. Cu. Zii. Mil. Ni and Pb using ;in 1L Model 357 atomic absorption spectophotometer. Metals were extracted from tlie iiiitial soil, sewage sludge and soil samples after harvesting using three different extract;liits:;\) 0.005M DTPA-0.01 M CaCI2-O.1MTEA, pH 7.3 (DTPA) (22), b) 0.5N NH4Ac-0.5N acetic iicicl-O.02M EDTA (AAA-EDTA) (23), c) Digestion with 4M HNO,, 80°C overnight (Totali (74).

3. RESULTS AND DISCUSSION 3.1 Plant production The addition of sewage sludge to soil (treatinent Sludge) increased dry weight of straw and wheat grain and nuinber of grains pcr pot i n cornparison with treatment (Soil) (Table 2). The sewage sludge used in this experiineiit, with its high N concentration and low C/N ratio,

176

would be expected to increase soil fertility and plant production, especially when applied at a high dose. Reductions of up to 8% were obseived in dry weight of straw and wheat grain when soil was contaminated with inorganic cadmium (Salt-Cd) with respect to dry matter obtained with (Soil). A siniilar effect was obseived after treatinent (Sludge-Cd) in comparison with treatment (Sludge). These results are in agreement with those reported by (25), who found a 25% decrease in wheat yield when the soil was contaminated with 50 mgCd kg-'. The reduction in wheat yield caused by inorganic Cd was greater when the acidifying agent (SaltCd-A) was added to soil. However, if the acidifying agent was applied to soil together the cadmium-sludge mixture (Sludge-Cd-A), dry weight of straw and wheat grain and number of grains per pot increased appreciably (Table 2).

Table 2 Dry matter yield of straw, grdin and number of graiixpof' of wheat. Dry weight (1ng.Kg.l)

Straw

Grain

crains. pot"

Soil

I .43c

1.122

48c

Sludge

2.11b

1.64b

72a

Salt-Cd

1.3Icd

1.03~

44cd

Sludge-Cd

1.94b

I .56b

60b

Salt-Cd-A

1.13d

0.89c

35d

Sludge-Cd-A

2.423

2.25~

80a

Means within the same column followed by the same letter are not significantly different at the 0.05 level

3.2. Cadmium in plant and soil

Concentrations of Cd in straw arid wheat grain were below the reliable detection limit (0.01 mg kg") of the analytical procedure used i n treatments that included soil or soil + sludge. Cadmium applied directly to soil directly (Salt-Cd) or mixed previously with sewage sludge (Sludge-Cd) significantly increased the coiiceiitration of this element in straw and wheat grain (Table 3). I n straw, this effect was more pronounced when the acidifying agent was applied to soil. Although 110 eviderice of toxicity symptoms (eg, chlorosis) was observed in the plant, the concentration of cadmium i n straw and wheat was higher than normal levels (26, 27). The high level of cadinium found i n the edible part of wheat represents a health hazard if the crop is consurned (28). The addition of cadrniLiin mixed with sewage sludge, supplemented or not with the acidifying agent, significantly increased Cd uptake by the aerial part of wheat, in comparison with treatments that included inorganic cadmium (TabIe 3).

177

These results can be traced to the increase iri plant prodiiction due to the application of sewage sludge (Table 2).

C011 cent 1.at i oti Straw

Graiti

Uptake

Soil

l1.d.

l1.d.

I1.d.

S1 iidge

n.tt.

1i.d.

t1.d.

Salt-Cd

2XAb

I1;1

48C

Sludge-Cd

220

142

64b

Salt - C (1-A

31b

l3;1

47c

Sludge-Cd-A

27b

13a

9

Means within the suiie column followurl by the same letter are not significantly different at the 0.05 1evr.l 11.d.: 110 cletectrti. Detection limit: 0.01 nip Lg-'

Table 4 Concentration (Ing kg-l) of cadmium eutracttci witli DTPA, AAAEDTA and lM HNO, (Total) fruiii the soil :ifrer lxirvesting.

DTPA

A A A - E DTA

Total

Soil

0.12d

0.15b

0.31b

SI iidge

0. 17ti

0.3-017

0.60b

Sa I t -Ctl

1 Ob

I3;1

15a

Sludge-Ctl

Xu

1 .3;1

153

Sa It -Cd - A

1la

l3a

153

Sludge-Ctl-A

lob

1-32

I5a

Means withill the same columii followrd by [lie same letter are not significantly different at the 0.(15 Isvzl.

178

After wheat was harvested, more of 50% of the Cd added directly or mixed with sewage sludge remained in the soil as available forins (50-70% as DTPAextractable, 80% as AAA-EDTA-extractable). The percentage of DTPA-extractable Cd from the soil was similar to that reported by (29) in a field experiment with a soil pH of 6.7. In coiitr;1st, (30) found than most of the Cd added to a calcareous soil (about 70%) was immobilized i n unavailable forin after 100 days of incubation.

10-

80

1

0 Sludge

Cd

Salt-Cd

60

Sludge-Cd

ti?

0 Salt-Cd-A

40

Sludge-Cd-A

20 0

DTPA

AAA-EDTA

Figure 1. Percentage of total soil cadmium extracted with DTPA and AAA-EDTA after harvesting.

The concentr;itioii of DTPA-extractable Cd and the percentage of Cd extracted with DTPA of total soil Cd were higher i n treatments with inorganic cadmium than in treatments with cadnunm mixed with sewage sludge (Table 4, Figure 1). These differences may be due to the presence in the sewage sludge of inorganic components, including the carbonaceous fraction (31), which may favor the precipitation of part of the cadmium added; when this sludge-enriched Cd is applied to the soil, less cadmium may be extracted with DTPA than when this metal is added as salt to the soil. This effect was confirmed when Cd was extracted from the soil with AAA-EDTA (Table 4). This extractant can dissolve, in sludges and soils, the fraction of Cd precipitated with carbonates (32), hence more Cd was extracted using AAAEDTA, and no differences were observed between treatments. Finally, the amounts of cadmiuni extracted by 4M NO,H were siinilar in all treatments, the amounts of Cd removed being approximately equal to the ainoiint added to the soil.

3.3. Zinc and copper in plant and soil As expected, the application of sewage sltidge significantly increased Cd and Zn concentr;ltion and Cu and Zn uptake by the plant, compared with the unamended treatment (Table 5). After harvesting, DTPA-extractable, AAA-EDTA-extractable and total Cu and Zn

179

(Table 6), and tlie percentage of total Cu mid Zn extracted with DTPA or AAA-EDTA of total (Figure 2) were also increased. Tlir: higher Cu and Zii levels i n plant and soil after the application of sewage sludge were due to the sludge supplying these metals directly, rather than the solubilization of Cu and Zn from tlie soil by the sludgc (9, 14, 33). As a result, the Zn concentration in wheat and DTPA a n t i AAA-EDTA-extractable Zn in the soil were above the levels considered phytotoxic (34-36); whereas Cu levels can be considered adequate.

Table 5 Concentration (Ing kg-') and uptake (pg pot-') of zinc and copper by wheat

Cu colicen tration

211concentration Straw

Grain

Zn tlptake

Straw

(;rain

Cu uptake

Soil

31f

(,Id

1 1i d

9b

2%

41c

Sludge

73c

103'1

325b

I8;1

312

903

Salt-Cd

551

xxc

I O?c

Xb

26b

38c

sIt1dge-Cd

95b

92 bc

32Sb

llb

25b

60b

Salt-Cd-A

43e

96Jb

I33cd

9b

26b

32c

Sludge-Cd-A

10 1J

99h

4hlLl

1 Ob

21c

71b

Means within the saine column followed by the ~ the 0.05 level.

i letter ~ LIY e not significantly different at

Regardless of the fomi i n which Cd was iitliled. Zn concentratiori i n straw and 2 n uptake by the plant increased significantly with respect t o txatiiients (Soil) and (Sludge) (Table 5). These increases were greater when the Cd-einiclxtl x w a g e sludge was applied together with tlie acidifying agent. I n the soil, tre;itiiients that includzd inorganic or organic Cd yielded lower values of AAA-EDTA-Zn (Table 6) and perceiit;ige of Cd extracted with AAA-EDTA (Figure 2 ) in comparison with treatinents (Soil) ;ind (Sludge). However, total Zn levels were higher (Table 6). The increase i n crop Zii content obseived in this study may be due to a concentration phenomenon catised by tlir rediicted yield with treatments that included inorganic Cd (Salt-Cd) or Cd mixed with sewage sludge (Sludge-Cci) (Table 3). This effect was also noted by Turner (37) i n different veget;iblzs grown on solution cultures. This author offered the following explanations for this el'fect: a) Cadiiiuni caused root damage that subsequently enh;iiicztl Zn uptake, b) Redistribution of Zn betweeii roots and tops and c) Cadrniurn stimtiluted 211u p t k e by the plant. I n coiitr;i\t. ;in antagonic interaction between Cd and Zn has been observed in other studies: interaction this was dependent mainly on the soil charilcteristics and the type of plant cultivated (2 I , 3%). The addition to soil of inorganic Cd, either alone o r suppleinented with the acidifying agent, scarcely affected concentration and Cu tiptake by wheat as compared with the unanmended soil (Table 5 ) . However, Cd mixed with sew;ige sludge significantly decreased crop Cu content in relation to treatment (sluclge). After Iiiiivesting, differences in DTPA-Cu,

180

AAA-EDTA-Cu and total-Cu in the soil were caused only by the addition of sewage sludge, and not by Cd or the acidifying agent applied to the soil (Table 6).

Table 6 Concentration (mg kg-') of zinc and copper extracted with DTPA, AAA-EDTA and 4M HNO, (total) from the soil after harvesting ~

~~

cu

Zll

DTPA

AAAEDTA

Total

DTPA

AAAEDTA

Total

Soil

6b

27c

63e

2b

6b

19b

Sludge

273

6 3;1

102b

6a

1Oa

24a

Salt-Cd

6b

20d

71d

2b

6b

18b

Sludge-Cd

26a

54b

104b

6a

1oa

24a

Salt-Cd-A

7b

19d

77c

3b

6b

20b

Sludge-Cd-A

2%

57b

124a

5a

I Oa

25a

Means within the same column followed by the same letter are not significantly different at the 0.05 level.

75

n

60

cu

45

0 Sludge Salt-Cd Sludge-Cd

8

Salt-Cd-A

30

Sludge-Cd-A 15 0

DTPA

AA-EDTA DTPA IAAA-EDTA

Figure 2. Percentage of total soil zinc and copper extracted with DTPA and AAA-EDTA after harvesting.

181

3.4. Manganese in plant and soil

~~

Concent ration St 1'3 \v

Grain

Uptake

Soil

41c

43b

106c

Sludge

3 Id

14c

XXcd

Salt-Cd

47 b

15c

7Xd

S1 ud ge-Ctl

7 It1

22c

U4cd

Salt-Cd-A

93'1

5 (I

J

157a

SI u d ~ e - C t lA-

3sc

I oc

127b

Means within the saiiie column fi~llowedb ~ the , same letter arc no[ signific;inily differtnt ;I[ the 0.05 Icvrl.

Table 8 Concentriition (ing kg-') i r l M i 1 esti.acttd w i t h DTPA, AAA-EDTA and 4M HNOl (total) from tlir \oil Lifter' 1xtrvestin.E

UTPA

AAA-EDTA

Total

Soil

3lc

165b

2XSbc

Sludge

20b

1 6%

283bc

Salt-Cd

IZd

170b

2XXbc

SI ud ge-Cd

1Sd

16%

27%

Salt-Ctl-A

-lOa

1XXLI

29%

Sludge-Cd-A

42d

IX9d

31 l a

Means within the s:aiie coluiiiii followed ,)JI [lie same letter are not significantly different ;it the 0.05 luvel.

The addition of s e w q e sludge to soil significantly drcrrased tlir concentration and uptake of Mn in plant, in cornparison wiili the results i n uiitrcated soil (Table 7). This finding was most likely due to thc low Mn coiiceiitratioii i n the sluclge (Table I ) , which was ever lower tlian the innate concentration in the soil. 'I'lius iipplicxtion of sludge to the soil may have

182

diluted this metal i n the plant, as a consequence of the higher yield obtained with this treatment.

-57 601

soil

0 Sludge

Mn

Salt-Cd Sludge-Cd

ts 45!

Salt-Cd-A Sludge-Cd-A

15 0

DTPA

AM-EDTA

Figure 3. Percentage of total soil manganese extracted with DTPA and AAA-EDTA after halvesting. Manganese concentration and uptake by wheat decreased in soil contaminated with either forin of Cd in cornparison to tlie ainouiits found in treeatments (Soil) and (Sludge) (Table 7). This effect is consistent with previous studies, which reported a decrease in Mn content in plants grown on solution cultures (39, 40) or calcareous or limed soils (21, 41) when Cd was added at high concentrations. In the soil, DTPA-extractable Mn and the percentage of total soil Mn extracted with DTPA were drastically decreased, but no differences were found between the amounts of Mn extracted with AAA-EDTA or 4M HN03 (Table 8, Figure 3). As reported in (21). Cd may compete with Mn i n the soil, and not necessarily in tlie plant or at the root surfaces. The addition of the acidifying agent to soil (Salt-Cd-A, Sludge-Cd-A) increased Mn uptake by plant (Table 7), Mn extracted from the soil by the three extractants (Table 8) and the percentage of total soil Mn extracted with DTPA or AAA-EDTA (Figure 3). Increases in DTPA-extractable Mn were also noted by (42) in an incubation experiment using a similar wastewater. This effect is most likely due to the decrease i n soil pH (data not shown) induced by the acidifying agent, and the subsequent increase i n the availability of this inetal(43). Finally, Mn concentrations i n plant and soil encountered in all treatments assayed were far below the levels considered phytotoxic (34-36).

183

3.5. Nickel and l e ~ c lin plant and soil Nickel arid lead were not detected (iletection liinirs of the unalytical procedure employed: Ni= 0.06 m g kg", Pb= 0.1 mg kg-') i n straw or wheat grain i i i the different treatments.

Table 9 Concentration (mg kg-') of nickel and lead extracted with DTPA. AAA-EDTA and 4M HNO, (total) i n soil after harvesting.

Ni DTPA

AAA-

'rotill

DTPA

AAAEDTA

Total

EDTA Soil

1 .osc

3a

I4a

2lb

30b

48b

Sludge

I .S?b

3 ;I

I4a

392

S1a

70a

Salt-Cd

1.03c

3a

I4a

24b

29b

46b

Sltldge-Cd

1.50b

3il

1 ?a

34a

51a

6%

Salt-Od-A

1.1oc

32

142

25 b

30b

48b

Sludge-Cd-A

1,953

3;I

I5a

38a

51a

70a

Means within the sanie column followed by the s;iint: Ielter are not significantly different at the 0.05 level

0 Sludge Salt-Cd Sludge-Cd

0

Salt-Cd-A

DTPA AAA-EDTA DTPA AAA-EDTA

Figure 4. Percentage of toL:il soil nickel and lead extracted with DTPA and AAA-EDTA after Iwvcsting.

184

The concentmtion of Ni extracted fron tlie soil with the three different extractants was little affected by tlie addition of sewage sludge, cadmium and the acidifying agent (Table 9). Only DTPA-extractable Ni were increased when sewage sludge was applied to soil, this effect being more pronounced when i t was added together with the acidifying agent. No interactions between Cd and Ni were observed in the soil. Sewage sludge application at ;I rate of 700 Mg ha-' increased DTPA-extractable, AAAEDTA-extractable and total Pb i n the soil i n comparison with unarnended soil (Table 9). This effect, also observed for Cu and Zn, was probably due to the sludge supplying this metal directly to soil. The addition of Cd tended to decrease DTPA-extractable Pb, but no appreciable changes were observed on the ainounts of AAA-EDTA Pb and total Pb. The percentage of Pb extracted with DTPA ruiged from 48% to 62% in the different treatments. These figures were higher than those found for the other metals: Zn (8% to 23%, Figure 2), Cu ( 1 1 % to 25%, Figure 2 ) , Mn ( 7 7 ~to 14%, Figure 3) ond Ni (8% to 13%, Figure 4).

4. S U M M A R Y A N D CONCLUSIONS A pot experiment WLIS designed untler greenhouse to investigate the effects of the addition of different f o r m of cadinium (inorganic as salt or previously mixed with sewage sludge) and an acidifying agent on Cd, Zn, Cu, Mn, Ni and Pb concentrations and uptake by wheat. After harvesting, soil metals extracted with DTPA, AAA-EDTA and 4M HNO, (total) were also detemiined. The data obtained lead to the following conclusions: 1) Dry weight of straw a i d wlie;it grain contaminated with cadinruin.

WAS

reductcd

tip

to 8% when the soil was

2) Cadmium levels found i n tlie edible part of wheat may represent a health hazard if the crop is consuiiied.

3 ) After harvesting, more than 50% of Cd iiddetl rem:iined i n the soil as available forms. 4) Cudmium mixed with sewage sludge led to a higher increase i n Cd uptake by wheat than the inorganic Cd. 111 contrast, after halvesting, DTPA-extractable Cd was higher when inorganic Cd was added to the soil.

5) The addition of both Cd foniis significantly increased Zn concentration and uptake i n wheat. I n contrast, a decrease in plant Mil content was observed. The amounts of others metals (Cu, Ni and Pb) were little nffected i n plant and soil. 6) The addition of the acidifying agent (wastewater from olive oil processing) increased plant and soil Mn content. Cacimium, Zn and Cu were affected to ;i lesser extent.

7 ) 111 a11 treatments assayed, tlie percentage of total soil Pb extracted with DTPA was greater than the percentages of Zii, Cu, Mn and Ni extracted.

185

5. ACKNOWLEDGEMENTS The authors are grateful to the CICYT for financing this study through project I+D AGR890500. D. Heivas and J . Soto thanks 10 tlie ICI-CSIC-UNESCO for finding their stay at the Estacion Experimental del Zaidin. CSIC, Granada,Spain. We would also like to express our appreciation to Ms Karen Sliashok foi- assihtiiig i n tile translation of tlie original manuscript into English.

6. REFERENCES 1 2

3 4 5

6 7 8 9 10 11 12 13 14 15

16 17 18 19 20 21 22 23 24 25 26 27 2X

29 30 31 32

Jones KC, Johnston AE. Environ Poll 19x9; 57: 199-210. Petruzzelli G. Agric Ecosys Enviroii 1989; 27: 493-593. Miles LJ, Parker GR. J Enviroii Qua1 1079; X: 229-232. Cole JF, Volpe R. Ecotoxicol Eiiviron 19x3, 7: 1.51-159. Steiness E. Toxicol Envjron Clieiii 1980; I ‘I: 130- 145. Alloway BJ, Juckson AP, h4org:tii tl. Sci Total Environ 1990; 0 I : 223-236. Joiies RL. Hinesly TD, Zirgler EL. J Environ Qua1 1973: 3 : 351-353. Zwarich MA, hilills JG. C h i J Soil Sci 1982; 62: 243-247. Hues NV, Silva JA, Arifiii R . J E i i v i r w Qua1 19SX; 17: 384-390. Subey BR, Pentileton RL, Webb BL. J Eiiviron Qu:il 1990; 19: 580-586. Higgiiis AJ. J Environ Qu:il. 19x4: 13: 441-448. Keefer RF. Singti RN, Horvath DJ. J Environ Qua1 19x0; 15: 146-152. Singli RN, Keefei. RF. Agi-ic Ecosys Eiiviron 1989; 25: 27-38. Sheaft’er CC, Decker AM, Chuney RL, Duuglas LW. J Environ Qunl. 1979; 8:455-459. Bidwell AM, Dowdy RIH J Enviroii (.his1 19x7; l b : 438-442. Lercli KN. Barbarick K A , Westfall DG, Foller I I H , McRride TM,Owen WF. J Prod Agric 1900: 3: 00-05. Coinertord N B , Fiskell JGA. Soil Crop Sci Soc Fla Proc 1983; 42: 176-180. Williaiiis DE. Vlaniis J , Pukite A H , Corey IE. Soil Sci 1984: 137:351-359. Davis RD, C;trl~on-SmithCI-I, Stark J H , Campbell JA. Eiiviron Poll 1988; 49-1 1.5. Street JJ, Sabey HR, Lindsay WL. .I Enviruii Qu~il 1078; 7: 286-290. Maliler R J , Bingham FT, Page AL, R y a n J A . J Eiiviron Qua1 1082; 11: 694-700. Lindsay, WL, Noi.wr11, WA. Soil Sci Soc Am J 1978; 42: 421-428. Lmkniien E. Ervio K . A c t a Agr. Fenn 1971; 123: 223-232. Sposito C, Lurid LJ, Cliaiig AC. S o i l Sci Soc .41nJ 1987; 46: 260-264. Binghsm FT. Environ Health P t - r ~ p1970; 28: 30-43. K l o k e A . Mi~l.opolluiaiiis it1 ilir Etivii~ctiiiiieittC‘oiii~ereiicc. BI-tissels:I.A.W.P.R., 1981; 1-12. Singli SP. ‘Tlikkar PN. N a y y a r VK.Intern J Eiivii-on Studies 1989; 33: 59-66. World Health Organization. WHO Tech Rep Srr no 505. Geneva, 1972. Baker DE, Amaclirr MCI, Leach RM. Environ I-lealtli Persp 1979; 28: 45-49. Baca MT, De Nobili M. Leita L. Navarro A , Nogalcs R. Intrrnationaal Symposium on Solid am1 Liqtiid Wastes: Their Best Dcslinatiuii. ‘renerilc: ANQUE, 1091; 2 : 13-21. Stover RC, Sominers LE, Sil\Iiera DJ. .I Water Pollut Cuiitrol F d 1976; 21652168. Beckett PHT. Adv Soil S c i 19x9; 0: 143-176.

186

33 34

35 36 37 38 39 40 41 42 43

Fresquez PR, Francis RE, Dennis GL. J Environ Qua1 1990; 19: 324-329. Viets FG, Lindsay WL. In: Walsh LM, Beaton JD, eds. Soil Testing and Plant Analysis. Madison: SSSA. 1973; 153-172. Sillampaa M. F A 0 Soil Bulletin 48. Rome, 1982. Benton Jones J, Wolf B, Mills HA. Plant Analysis Handbook. Athens, Georgia: MicroMacro Publishing, 1991; 1-186. Turner MA. J Environ Qua1 1972; 2: 118-119. Bingharn I T,Page AL, Mahler RJ, Ganje TG. J Euviron Qua1 1975; 4:207-211. Iwai I, Hala T, Sonoda Y. Soil Sci Plant Nutr 1975; 21: 37-46. Borges AC, Wolliun AG. J Environ Qua1 1980; 9: 420-423. Wallace A, Romney EM, Alexander GV, Soufi SM, Patel PM. Agron J 1977; 69:1820. Perez JD, Nogales R , Gallardo-Lara F. International Symposium Humus et Plant IX. Prague:CSVTS, 1988; 172. Adriano DC. Trace Elements i n the Terrestrial Environment. New York:Springer-Verlag, New York, 19x6; 1-533.

I87

EVOLUTION OF HEAVY METAL SPECIES IN LEACHATES AND IN THE SOLID PHASE DURING COMPOSTING OF MUNICIPAL SOLID WASTES. P. PRUDENT, C. MASSIANI, 0. THOMAS. [Jniversite de Provence, Laboratoire de Chimie et Environnement, case 29, 3 place Victor Hugo, 13331 MARSEILLE - FRANCE.

ABSTRACT Reconstituted municipal solid wastes are used for the determination of metal origins. The speciation of Cd, Cu and Pb in waste components, in fresh and mature compost, is assessed using a chemical sequential extraction procedure. Chemical speciation of metals in the composts is largely dependent on their origins in wastes. In order to characterize organic-metal complexes, an analytical method is developed showing the molecular weight distribution of copper, cadmium and lead species : Gel Chromatography is coupled with 1J.V. detection for the semi-determination of molecular species and Flameless Atomic Absorption Spectrometry for the analysis of the metals. The rank method is used to improve the whole LJ.V. spectra. The evolution of the distribution of molecular weight of organic compounds during the maturation process is revealed. Polycondensation reactions take place during composting and characteristic compounds with different specific affinities to metals are isolated.

INTRODUCTION Increasing production of Municipal Solid Wastes (M.S.W.) and the presence of heavy metals in by-products cause major problems for their elimination or valorization. Because of their persistence and toxicity at low concentrations, trace metals are of concern as contaminants of terrestrial and aquatic ecosystems [ 13. The risk of environmental contamination and of accumulation in the food chain is the major limit to the application of urban composts on land. The difficulties reside in the evaluation of risks when organic wastes are used as soil amendments. Today, there is considerable evidence that the environmental impact of trace metals is markedly influenced by the physico-chemical forms in

188

which they are present [2]. Mobility, environmental diffusion and bioavailibility will depend on their chemical speciation [ 3 , 41. However speciation of metals in composts is largely dependent on their origins and on the modifications induced by the biological treatment of the organic wastes. The aims, therefore, of this work are as follows : - to determinate the contribution of the different components of M.S.W. to the copper, cadmium and lead loads, and evaluate their influence on the chemical speciations of these metals in the compost, - to study, in correlation with the evolution of the organic matter, the modifications induced in metal speciations by the biological treatment of wastes.

Chemical sequential extraction procedures are widely used in literature [S, 61 and there is no need to demonstrate their interest ; but it is necessary nevertheless to develop experiments of a finer nature. In order to analyse in greater detail the criteria of risk assessement, an attempt is made to obtain a better understanding of the role of organic matter and of its evolution on metal complexation. For this purpose, an analytical method is developed i.e. gel chromatography coupled with 1J.V. spectrometric and absorption atomic detection. In order to obtain maximum information, the whole range of U.V. wavelengths is used. This is the main originality of the method.

MATERIAL AND METHODS Material Municipal Solid Wastes The composition of reconstituted M.S.W. is based on the results of ROUSSEAUX [7] (figure 1). The contributions to metal loads of each component are determined, and some examples will be presented to point out the relations between the origin of metals and their speciation in compost.

Compost The compost is produced in an industrial plant near Cavaillon in France. The first step is an accelerated fermentation and a dilaceration

I89

in a cylinder for a period of three days. Then the wastes undergo sieving aiid magnetic, densimetric and bnllistic sorting. Afterwards, the maturation of the sorted wastes takes place in the open air. 'The threeday-old compost is called "fresh compost" and nfter six months' maturation it is referred to as "mature compost".

Others 3 %

Wood, leather, rubber 3% Textiles 3 % Metallics 4%

Glasses 7 % Plastics 10 % Small sized particles 14 % Organic materials 2 4 % Paper-cardboard 32 % Figure 1. Composition of reconstituted M.S.W.(% humid-weight) [ 7 ] . Chemical speciation procedure

This analytical method is used to find out the combination states of metals in each constituent of M.S.W. and in the composts. Metal speciation in M.S.W. and i n compost is assessed using the extraction procedures experimented by TESSIEK 151 and DUDKA IS], and adapted by DEL FAVA [4]. Different extracting reagents are used and the resulting solutions correspond to the following fractions F 1 to F5. F 1 : slightly acid medium pH=5, acetate buffer 1 M, 20°C and stirring for 5 hours.

F2 : reductor medium pH=2, NH20H-HC10.1 M, 20°C. stirring for 35 minutes.

i 90

F3 : complexing agent pH=9.5, K4P207 0.1 M, 20°C,stirring for 24 hours. F4 : strong reductor medium, NH20H-HC10.04 M in CH3COOH 25%, G O T , stirring for 6 hours.

F 5 : strong mineral acids, HN03-HC1 (1/3 : aqua regia), 1 2 hours at 20°C then 3 hours at 105°C.

Fractionation of organic metal complexes For the isolation of organic complexes, the two composts are treated with deionized water and with K4P207 0.1 M. The interest of using these two extractants lies in the nature of the organic complexes each of them extracts. Deionized water is used to accede to easily leachable organic matter, and potassium pyrophosphate, which corresponds to a basic and strong complexing medium, to accede to more stable organic complexes such as humic compounds [!I]. The optimal duration of extraction is determined by following the evolution of U.V. spectra until they no longer vary. The methodology used for the fractionation of organic complexes (figure 2) is based on the adaptation of various methods experimented by GRANET et al. [lo] and by BERAIL et al. [ 111. Extraction : H 2 0 or K4P207 for 10 hours Compost (humid weight) / extractant volume : 15 g / 300 ml Centrifugation : 2,800 rpm Filtration : 0.45 pm GEL CHROMATOGRAPHY : Sephadex G 75 gel : Volume of extract injected : 5 ml Dimension of the column : 86 cm high-1.6 cm internal diameter Dead volume : 54 ml (corresponding to the fraction 12) Total volume : 225 ml (fraction 50)

Sephadex G25 gel : Volume of extract injected : 3 ml Dimension of the column : 65 cm high-1.6 cm internal diameter Dead volume : 58 ml (corresponding to the fraction 13)

191

Total volume : 244 ml (fraction 54) Eluant : 75.4% w/w KH2PO4 - 24.6% w/w NaHPO4 (0.0667 M each) solution in order to fix ai ionic strength of 0.1 M and a pH of 6.4 Flowrate : 21.7 ml/h Volume of fractions : 4.5 nil

COMPOSTS : fresh or mature

/ EXTRACTION

deionized water easil~Icachablc frac-lion

\ K4P207 orgaaicxll> bound traction

U.V. analysis

CHROMATOGRAPHIC COLUMN Sephadex G75 (3,000 to 70,000 Da) Sephadex G25 (1,000 to 6,000 Da) with eluant : 0.1 M ionic strength pH= 6.4

FRACTIONS

Cd, Cu, Pb concentrations analysis of organic matter

200 to 350 nm pathlength 10 cm

Figure 2. Methodology of the fractionation of organic-metal complexes.

192

The Sephadex G75 column is calibrated with Dextran blue (2,000,000 Da) (corresponding to the dead volume), Albumin (66,000 Da), Carbonic Anhydrase (29,000 Da), Cytochrom C (12,400 Da), and phenol red for the determination of the total volume. The U.V. absorbance spectra are obtained from filtered diluted leachates and from fractions with a U.V. Spectrophotometer (SECOMAM SlOOOPC). The pathlength of the quartz cell is 10 cm and the spectrum is drawn between 200 and 350 nm. Cadmium, copper and lead concentrations are obtained by Flameless Atomic Absorption Spectrophotometer (Perkin Elmer 1100B, HGA 700).

Examination of U.V. spectra The examination of U.V. spectra of waters and wastewaters was reviewed in the papers of GALLOT and THOMAS [12, 131. One of the simplest ways of interpreting U.V. spectra is the rank method the theory of which is presented by GALLOT and THOMAS [14]. Its application for 3D chromatograms is of great interest because of the existence of -and the difficulty of studying- a set of numerous spectra, and the need to compare the results. THOMAS and id. [15] select with the rank method the relevant spectra automatically extracted from the 3D chromatogram. These spectra, called reference spectra, are independent and any spectrum of the chromatogram may be restored by a linear combination of the reference spectra. In connection with a classical chromatographic system with a rapid scanning spectrophotometer (diode array detector for example), a computer is used for the acquisition of absorbance values of the whole U.V. spectrum, and later for computation. Firstly, the acquisition of data continues until the end of elution and then, the program gives the number and the corresponding spectra of the relevant fractions. If another run is needed, for a different sample for example, the procedure allows the possibility to check if the detection signal is independent of the previous spectra. Then interpretation may begin, with the following procedure, for the identification of the relevant fractions i.e. the independent spectra of the 3D chromatogram. This method is general for the study of mixture composition and the comparison of absorption spectra [ 121 :

A=

I93

matrix A is built from the absorbency data of the n fraction files (for m wavelengths h 1, -

- the rank of matrix A is computed either with a simple procedure based 011 a variant of the Gaussian method, or with another method using singular value decomposition. The rank of the matrix is equal to the number of independent rows corresponding lo the reference spectra. Any spectrum of the chromatogram can be restored from the reference spectra. The coefficients of the linear combination are computed from the following relation with least squares method : P AB(hi)

=

C

aj.SRj(hi)

j=l

where AB(hi) is the absorbance of a fraction, for a wavelength hi, and aj the coefficient of the jth reference spectrum whose absorbance for hi is SRj( hi ) . A t the end of the procedure, the analyst is able to determine easily the qualitative nature of the organic material of any fraction, on the basis of the knowledge of all the coefficients aj. Therefore, this quick method of interpretation of the 3 D chromatographic response gives the relevant information of the set of spectra for each chromatographic assay, and allows the comparison of the results within different samples. This proposed method must also be simple and easy to apply for any analytical chemist. It is not only useful for the interpretation of chromatograms of complex mixtures such as environmental samples but also for the development of chromatographic methods : mathematical separation must be considered as an aid for chemical separation. It can also be directly applied, for example, for the comparison and interpretation of the U.V. spectra of waters and wastewaters [13, 151.

RESULTS AND DISCUSSION

Origin and evolution of heavy metal species in compost. The results of the sequential extraction procedure show the differences between metal speciations in fresh and mature composts (figure 3 ) .

194

c19/9 5

4

3 2 1

0

frcsh

mature

compost

fresh

mature

compost

fresh

mature

compost

Figure 3. Metal speciation in composts. If a comparison is made of the distribution of metal species in fresh and

mature composts in relation to different extracting reagents, no significant macroscopic evolution is observed, except a slight decrease of fraction F 1 (easily leachable metals) at the end of composting. This last observation could result from lixiviation by rain during the maturation process in open-air windrow. However, the distribution among the fractions varies from one metal to another. Thus for copper, the main part of metal extracted is that obtained with strongly acidic and oxidant medium (fraction F5). It corresponds to residual metals which are leachable only in certain specific conditions. We can also observe the relatively large proportion of the organically bound fraction ( F 3 ) . This last result is close to the well-known complexation of copper by the organic fractions in sewage sludge or soils [16, 171.

The lead speciation diagram is characterised by the fact that the largest proportion is bound to organic matter and by a significantly large pH 5 soluble fraction. Cadmium is spread out over the five Practions, also with a noticeable part of easily leachable metals. These results can be explained by the determination of the origins and the chemical forms of metals in the various components of wastes. These origins are obtained from metal speciation in each constituent of the reconstituted M.S.W.. The major load of copper comes from scrap-metal. Copper is also associated, to a lesser extent, with organic materials such as paper (ink),

195

leather, wood and plastics (additive material) (figure 4). In every case copper is mainly extracted with strong reactive mixtures. The main environmental impact (mobility, biodisponibility) will be a long term one. 116 GO 50

10

scnpmctal leather

paper cardboard \ v o o d

phstic-s

constituents Figure 4. Distribution of Cu in M.S.W. constituents. The pH 5 soluble part and the F3 organic fraction of Pb compounds come from additives such as fungicide and pesticide in wood, or pigments in leather (figure 5 ) . They also come from fine particles of dusts [ 7 ] .In compost the two fractions F 1 and F3 which are considered as interacting more rapidly with the soil-plant system than the F5 fraction, represent about 70% of the total Pb content. Leather and wood, with pigment and protection agents (paint, fungicide, pesticide), are also at the origin of cadmium contamination (figure 6). Cadmium is also contained in plastics in the form of pigments and stabilising agents, but liberation of these cadmium compounds needs a strong oxidizing medium to attack the plastic matrix.

196

x 0.2 I

120

a

2 a

80

n n

-40

n wood

plastics

leathcr textile crockery

paper

constituents Figure 5. Distribution of P b in “M.S.W. constituents. 19 8

6 \

g-4

3

2

0

lcather

wood

plastics rubber

textile

cardboard

consti tuen ts Figure 6. Distribution of Cd in M.S.W. constituents.

I97

The results confirm that the speciation of metals in compost is largely dependent on their origins in wastes. Metal distribution is unlike that observed in composted sewage sludge [18, 19, 201. The need for studies on the evolution of metal speciation during biological treatments or in compost amended soils becomes obvious. Information cannot be obtained by extrapolation of the results of studies on composted sewage sludge or on composted sludge/soils interactions. For urban composts, the liberation of metals (and in consequence, environmental risks) depends not only on the evolution of metal binding in soils but also on the rate of attack of the matrix in which the metal is contained. Evaluation of risks becomes a hard task. Macroscopic selec live extraction procedures give operational information but do not show if there is a recombination of metal inside a given fraction, as for example in the organic fraction F3, the evolution of which during coniposting is wellknown [2 1, 22, 231. Therefore, following the evolution of metal-binding in correlation with the evolution of organic matter, during the maturation process of compost, appears to be an interesting way of research. Organometallic complexes are of prime importance in the transfer to vegetal matter [17]. Fractionation of organometallic complexes U.V. characterization o f H 2 0 and KqP207 extracts of fresh arid mature

conipos ts The extraction time for compost is optimal after ten hours. The examination of the U.V. spectra, figures 7 and 8, shows that the quantity of organic matter extracted with water or with potassium pyrophosphate from mature compost is approximately twice as large as that extracted from fresh compost. A comparison of figures 7 and 8 shows that much more organic matter is extracted with potassium pyrophosphate than with water. Water has, in particular, the capacity to extract low molecular weight compounds and few humic compounds, whereas potassium pyrophosphate has the property of extracting humic and fulvic acids [l, 16, 17, 23, 251. These spectra are monotonous and resemble those of wastewaters or landfill leachates [12, 141, mainly with two parts. The first is located at the beginning of the spectra around 200-230 nm, and the second is a shoulder around 250-280 nm. Both for fresh and mature composts, the first part of the spectra of water extracts is distinctly different from that of potassium pyrophosphate extracts. These comparisons between the forms of spectra corroborate the fact that organic matters extracted with H 2 0 or with K4P207 are different.

198

1

mature compost (dil. 10 times) fresh compost (dil. 10 limes)

-. 200

2 50 300 wavelength in nm

350

Figure 7 . Absorbance spectra of leachates of fresh and mature composts (extraction with H20). 3

. 8

fresh compost (dil. 1 0 times) mature compost (dil. 2 0 times)

2

9

e 2

n 4

1 I

I 250 300 350

0 200

wavelength in nm

Figure 8. Absorbance spectra of leachates of fresh and mature composts (extraction with K4P207).

I99

In comparison to the fresh compost spectrum profile, that of mature compost is characterized by a modification of the shoulder around 2 8 0 nm which is not so marked. This probably corresponds to the presence of a larger number of organic molecules and to a greater complexity of their nature. The concentrations of copper, lead and cadmium in raw extracts are presented in Table 1. Table 1. Metal concentrations in raw extracts of fresh and mature composts (pg/g dry-weight). ~~

Cd

cu Pb

~~~

~

fresh compost

mature compost

H20 0.9 3.7 2.1

H20 0.2 8 1

K4P207 0.3 8.8 110

~

K4P207 0.7 7.4 130

The quantities of metal extracted by K4P207 in this one-step procedure is approximately the same as those obtained in the sequential extraction procedure (fraction F 3 ) , except for lead. Lead extracted by K 4 P 2 0 7 is greater in the one-step procedure, which is probably due to the fact that a part of the metal is normally extracted with fractions F 1 and F2 in the sequential extraction procedure (figure 3 ) . Quantities of metal extracted with R 2 0 are very low in comparison with those obtained in fraction F1 (figure 3). This is particularly noticeable for C d and Pb. The difference between F1 (pH = 5) and water removable metal is very likely due to the precipitation of metal carbonates. Few soluble metal salts remain even in fresh compost ( d u e to their precipitation as metal carbonates during the three days of accelerated fermentation). Quantities of metal removable by water are very low, and only a small portion could be bound to the water soluble organic matter. Therefore more extensive experiments will be conducted on the nature of extracts obtained with potassium pyrophosphate.

200

Fractionation on a Sephadex G75 column Water as extracranr After chromatographic separation of water extracts, the U.V. absorbance chromatograms recorded at 2 80 nm exhibit some differences (figure 9) between mature and fresh composts. The choice of this wavelength is based on the general form of the previous spectra. The mature compost chromatogram presents two peaks : a major one around fraction 39 (Molecular Weight close to 10,000 Da) and a small one around fraction 14 (aggregates of high weight molecules, greater than 70,000 Da, not separated by gel). But the fresh compost elution profile exhibits one peak around fraction 44 (molecular weight around 6,000 Da). These latter compounds may also be present in mature compost, but if such is the case, their corresponding peak might be concealed by the main peak whose maximum is around fraction 39. More organic matter is extracted from mature compost than from fresh compost, and the major peak is shifted toward some new compounds with a higher molecular weight. Thus, this chromatogram shows the emergence of new compounds with higher molecular weight, probably resulting from polycondensation and from organic polymers which are hydrolysed and become extractable after maturation.

-

015

074 E

8

fresh mature

0,3

:m:0,2 0,1 0 90 0

15

30 45 fraction number

GO

Figure 9. Sephadex G75 absorbance chromatograms of leachates of fresh and mature composts, extracted with H20.

20 1

These results also confirm that extractable matter is more complex i n mature than in fresh compost. The Cd, Pb a i d Cu elution profiles of fresh or mnture composl extracts are also determined (figures 10, 11, 1 2 ) .

-

80

60

__t_

5 40 f

0

20

0 0

15

so

45 fraction number

GO

Figure 10. Sephades G 7 5 Cu chromatograms of leachales of fresh and mature composts, extracted with H 2 0 . The Cd and Pb concentrations are very low and interpretation requires some precaution particularly with regard t o fresh compost extracts. Compounds with an Apparent Molecular Weight (ArVl\'V) of 6,000 Da, which seem characteristic of fresh compost, appear to have a high affinity for Cd and Pb, whereas almost no copper is bound to these molecules. Compounds with an AMW of 10,000 Da, whose concentration increases with the maturity of the compost, are able to complexe the three metals, but it seems that they have a lesser affinity for Cd and Pb than those of 6,000 Da AMW. Another characteristic of the evolution of the metal binding during the maturation process is shift of metal, particularly of copper, towards molecules with a higher molecular weight. Copper is mainly found associated with the compounds of 16,000 Da AMW (fraction 32, not clearly defined on the absorbance profile) and to a lesser extent with the higher AMW compounds which are probably humic substances [21, 221.

202

40

-

I

30 -

fresh mature

7

m

I 2 0n n

10 -

O l 0

I

I

I

I

15

30

45

GO

fraction number Figure 11. Sephadex G75 Pb chromatograms of leachates of fresh and mature composts, extracted with H 2 0 .

20 -

-

9 '0

0

fresh mature

10 -

0 :

0

I

15

I

I

30 45 fraction number

I

GO

Figure 12. Sephadex G75 Cd chromatograms of leachates of fresh and mature composts, extracted with H 2 0 .

203

The chromatograms show that small molecules or ionic forms are also present for Cd and Pb ( metal in the fractions above 5 0 ) . Therefore, during the maturation process, water removable organic matter increases and its nature changes with a general increase of the molecular weight of extracted compounds (organic matter corresponding to fraction 34 in fresh compost shifts towards fraction 33 in mature compost). Their affinity for metals differs also from one metal to another. K4P2O7 as extractant After Chromatographic separation of the L i p 2 0 7 extracts, the chromatograms of leachates of mature and fresh composts, recorded at 280 nm (figure 13), exhibit two peaks corresponding respectively to fraction 15 (aggregates of high weight molecules, nor separated by gel), and to fraction 42 (molecules with a molecular weight around 8,000 Da). The absorbance values observed confirm the fact that organic matter is more soluble with K4P2O7 than with H 2 0 .

1

21

ti1 ; II u re

Ircsh

0

0

15

30 45 fraction number

GO

Figure 13. Sephades G75 absorbance chromatograms of leachates of fresh and mature composts, extracted with K3P207. In the case of mature compost, we can observe a wider dispersion between fractions 20 and 38. Thus a s for water, this chromatogram shows the emergence of new compounds with a higher molecular

204

weight, probably resulting from polycondensation and from organic polymers which are hydrolysed and become extractable after maturation. Compounds with a molecular weight > 70,000 Da are formed. These results also confirm that extractable matter is more complex in mature than in fresh compost. The copper elution profiles (figure 14) follow the U.V. chromatogram, with two peaks around fractions 15 and 42. In fresh compost, metal is principally bound with the 8,000-Da organic compounds. In mature compost, copper mainly remains bound with these same compounds, but there is a shift towards the new compounds with a higher molecular weight : a shoulder is noticeable around fractions 38 and 39, as well as a dispersion from fractions 2 5 to 38. These results are close to the well-known complexation capacity of the organic fractions for copper in sewage sludge or soil [ l G , 171.

150 1

100

I

5

50

0 0

15

30 45 fraction number

GO

Figure 14. Sephadex G75 Cu chromatograms of leachates of fresh and mature composts, extracted with K4P207. For lead, no major difference is observed between the metal distribution in fresh and mature composts. Lead is associated with fraction 4 2 (organic compounds with a molecular weight of 8,000 Da), but is also eluted with fractions 48-50, corresponding to compounds with a very low molecular weight and probably to some ionic forms (figure 15).

20 5

1

I

600 T UI I

-

maturc I'rcsh

P

n

400

200

0

30 45 fraction number

15

0

GO

Figure 15. Sephadex G75 Pb chromatogrims of leachates of fresh and mature composts, extracted with KqP207.

-

20

5 -

U

0

m'ilurc l'rcsh

10 -

0

1

I

1

I

Figure 1 G . Sephadex G75 Cd chromatograms of leachates of fresh and mature composts, estracted with K4P207.

206

The high concentrations of lead confirm the affinity of organic matter for lead [17, 261. The cadmium elution profiles (figure 16) show that cadnlium is bound to the compounds eluted with fraction 38. T h s confirms the part played by the molecules with an AMW of 10,000 Da in metal retention. The latter are not distinguishable on the absorbance profile, whether it is because they absorb little of the 280 nm radiation or because they are present at low concentration. Cadmium is also found in fractions 48 to 55. During the biological treatment, modifications of the binding state differ from one metal to another. Lead and copper are bound to the characteristic compounds with an AMW of 8,000 Da both in fresh and in mature compost. However, whereas copper is also eluted with the fractions below 42, lead is eluted with the fractions above SO. New compounds with an AMW around 10,000 Da, not visible on the absorbance profile, show particular affinity for Cu and Cd. The molecules with the highest molecular weight, formed during composting, are mostly associated with copper. For the fractions above 50 (Total volume), the gel no longer separates efficiently. Therefore, Cd and P b eluted with fractions 48-55 are either bound to molecules with a molecular weight of less than 3,000 Da or are in ionic forms. The hypothesis of the ionic forms must not be eliminated because, for example in the case of lead, the quantity of metal extracted by K4P207 in this one step is slightly larger than that obtained by a sequential procedure. Thus there are not only organic complex forms in this first extract. Complementary information will be obtained using a Sephadex G25 gel chromatographic column (separation power between 6,000 and 1,000 Da).

Fractionation on a Sephadex G25 column The results obtained with a Sephadex G75 gel column gave interesting information on organic-metal binding, and pLuticularlywith K4P207 as extractant. The G75 gel allows the separc\uon of specific organic compounds, in the molecular weight range of 70,000 Da - 3,000 Da, which present a particular affinity for metals. However, &hepresence of specific molecules with a molecular weight lower than 3,000 Da or ionic forms are disclosed. In order to characterize ihese compounds more finely, it was therefore decided to study the K4P207 extracts on a Sephadex G25 gel column. The U.V. chromatograms recorded at 210 nm (figure 17) and at 280 nm (figure 18) on Sephadex G25 gel exhibit mainly three peaks corresponding respectively to fraction 14 (aggregates o f molecules not

207

separated by gel and with a molecular weight greater than (3,000 Da), and to fractions 2 0 and 25 (corresponding to compounds with a molecular weight of less than G,000 Da).

0

15

30 45 fraction number

60

Figure 17. Sephadex G25 absorbency chromatograms at 2 10 nm of fresh and mature compost leachates, extracted with Q P 2 0 7 .

0

15

30 45 fraction number

Ircsh mature

60

Figure 18. Sephades G25 absorbency chroniatograms at 280 nm of fresh and mature compost leachates, extracted with lQP207.

208

The absorbency of the molecules eluted with fractions 2 0 and 2 5 is lower for radiation at 2 80 nm than for radiation at 2 10 nm. Both chromatograms confirm that much more organic matter is extracted in mature than in fresh compost. After maturation Lie predominance of fraction 14 is yet another sign of the formation of high molecular weight compounds during the process. Information about these compounds is less accurate than with the G75 column. On the other hand the small molecules could be separated into two types. The compounds eluted with fraction 2 5 , whose concentration increases with maturation, can be considered characteristic of the maturation process. The part these molecules play in metal binding can be studied. Copper is mainly bound to fraction 14 in fresh compost and to fractions 7 and 14 in mature compost (figure 19). For both extracts (fresh and mature composts) metal peaks are present around fraction 7. This can be due to a slight separation between some very high molecular weight compounds on which copper is complexed. The affinity of compounds with an AMW higher than G,000 Da, for copper, is negligible. On the copper profile, only a slight shoulder around fraction 2 2 can be observed. Some copper is also eluted with low molecular weight molecules or as ionic species. 100 -

80 -

A

-

frcsh mature

5603.

3

40 20 O ! 0

I

15

I

1

30 45 fraction number

I

GO

Figure 19. Sephadex G 2 5 Cu chromatograms of fresh and mature compost leachates, extracted with KqP207.

209

The case of lead is clearly different. Almost all the lead which is not bound to the high molecular weight molecules (6,000 - 70,000 Da) is bound to the molecules eluted with fraction 25 (figure 2 0 ) , whose concentration increases with the maturation process.

3 00

I'rcs h maiure

T

P)

I

n

n

2 00

100

0 15

30 45 fraction number

60

Figure 20. Sephadex G 2 5 Pb chromatograms of fresh and mature coinpos t leachates, extracted with IQP2 0 7 .

100

-

80 60

Ired1 tnLiIurc

S.

'D

u

40

LO

0 0

15

30 45 fraction number

GO

figure 21. Sephadex G 2 5 Cd chromntograms of fresh and mature compost leachates, extracted with K4P207.

210

These compounds, eluted with fraction 25, appear to be specifically associated with lead as no copper or cadmium (figure 2 1 ) is found in fraction 25. Cadmium seems to be principally bound with very small molecules or to be in ionic forms. G25 gel chromatography results are a significant complement to the information provide by G 7 5 gel fractionation. The small molecules (molecular weight < 6,000 Da) can be separated into two types on a G 2 5 column and their concentrations are higher in mature compost. Moreover, their respective affinities for different metals are more marked. The results are resumed in table 2.

Table 2 : Affinity of organic matters to metals, according to their molecular weight (number of the fraction) f.c. = fresh compost and m.c. = mature compost. Metal <3 8

c u f.c. Cu m.c. Pb f.c. Pb m.c. Cd m.c.

++

G7 5 Fractions 38 42

++

-

*

++

+++ +++

+++ +++

22 << <<

G25 Fractions 25 ionic

*

+++ +++

<<

<<

++

It is clear that copper is mainly bound to the compounds with high molecular weight i.e. 8,000 Da in fresh compost and 8,000, 10,000 Da and more in mature compost. With maturation molecules with a high molecular weight are formed and copper is dispersed among these new molecules. Cd is mainly associated with the 10,000-Da compounds, whereas Pb is divided between the compouncis of 10,000 Da and those of less than 6,000 Da. However, it remains unknown wether the nature of the various separated molecules is the same in fresh and mature compost. Therefore, an attempt is made to answer this question studying the U.V. spectra of fractions and comparing them by means of the rank method described above.

21 I

Examination of the U.V. spectra of the fractions For each elution fraction ((225 Gel, and IQPLO~extracts) a 5pectrum is registered. The rank method is applied on the two series of spectra corresponding respectively to the elution of fresh and mature compost extracts. The rank method allows the selection of the most relevant spectra from which all the others can be restored by linear combination. These spectra are independent and characterize the main types of organic matter present in specific fractions. In each series four independent spectra are isolated. They are represented in figures 2 2 and 23. They correspond to fractions 12, 15, 2 2 and 26 for fresh compost and to fractions 10, 14, 2 1 and 26 for mature compost. The selected spectra correspond to the main peaks observed on the two chromatographic absorbency elution profiles. The spectra forms are monotonous with two main parts, the first around 200-230 nanometers and the second around 260-320 nanometers. 'The comparison of the forms of the spectra of the relevant fractions shows particular differences. This suggests the presence of fundamen tally different types of molecules and confirms the analysis of the chromatograms. Evolution of organic matter is clearly visible by comparing figures 22 and 23.

2 00

250 3 00 wavelength in n m

350

Figure 2 2. Independent spectra corresponding to relevant fractions, selected by the rank method for Sephadex G 2 5 fractions of fresh compost, extracted with I(qP207 ( 1 : fraction 2 2 - 2 : fraction 26 - 3 : fraction 15 - 4 : fraction 1 2 ) .

212

200

250 3 00 wavelength in nm

3 50

Figure 2 3. Independent spectra corresponding to relevant fractions, selected by the rank method for Sephadex G 2 5 fractions of mature compost, extracted with K 4 P 2 0 7 ( 1 : fraction 14 - 2 : fraction 2 1 - 3 : fraction 26 - 4 : fraction 10). For the comparison of the two sets of reference spectra, the rank method is then applied to the eight independent spectra of the previous determination. After computation the first series of eight spectra is reduced to four real reference spectra : three relevant to the mature compost, and one to the fresh compost. These selected spectra correspond respectively to fraction 22 of fresh compost (spectrum 1 of figure 22) and fractions 14, 26 and 10 of mature compost (spectra 1, 3 and 4 of figure 23). These results show that the spectrum of fraction 2 2 corresponds to a family of compounds which characterizes fresh compost, whereas the three others are characteristic of mature compost. The main advantage is that computation of the rank allows the conceiitratioii of information, as in the case above in which only four spectra are needed in order to explain all the others. It is now possible to apply a general method [12] for the examination of chromatograms of new samples, based on the selected reference spectra. First, for the acquisition of the U.V. spectrum of each new fraction, the file of absorbency values is added to the reference matrix (the p lines of which correspond to the absorbency values of the p reference spectra). Second, the rank of the new matrix is computed, and if the rank

213

increases we can conclude that the new spectrum is independent of the reference spectra and must be considered as a new type of organic material. On the other hand, if the rank is unchanged two conclusions are possible : - the new spectrum does not belong to the list of the reference spectra and its contribution can be computed (then we a-e able to identify the characteristics of the new fraction), - the list of reference spectra is modified by including the new spectrum. This means that the lar ter may be considered more relevant than another one, which is in this case removed [12, 151. These results lead to the conclusion that it will be possible to use the described methodology, with a simple estraction a i d an analysis of the U.V. spectrum of the total extract, in order to determine the state of maturation of composts. The carrying out of new experiments on composts in different state of maturation, should give a reliable set of characteristic spectra allowing a rapid and reliable interpretation of new U.V. spectra belonging to unknown compost.

CONCLUSION The comparison of metal speciation i n various components of wastes and in compost has shown that metal species in compost are largely dependent on their origins in M.S.W. components. This will be o f major importance with regard to metal release in the environmcnl with time. Thus the chemical extraction procedure gives precious information on the quantity of metal that can be rendered soluble in different chemical media. It may be considered an operative means of estimating the risk of spreading compost on land. However, this analytical procedure is not sufficient in order to follow the evolution of metal-binding together with the evolution of organic matter during the biological process. Gel permeation coupled wirh 17. V. and Atomic Absorption spectrometr)' seem to be well adapted. Complementary information is obtained using Sephades Gi 5 and G2S gel separations. I t is evident that maturation is characterized by an increase in removable organic matter (both for H 2 0 and Q I ' 2 0 7 as extractants). Observation of the LJ.V. chromatogranis and utilization of the rank method allow the determination of the compounds which are characteristic of the evolution of organic matter. Thus reactions of polycondensation. and probably of hj'drolyze, are observed with the apparition of high molecular weight compounds 121, 2 2 , 2.31, a s well as

214

with the apparition of specific organic compounds with AMW of less than 3,000 Da. These compounds show a particular affinity with some metals, for example high molecular weight molecules with copper or molecules < 3,000 DA with lead. The multiplication of experiments on various composts (different origins and ages) could give a set of specific compounds and specific spectra. The application of the rank method to raw extract spectra could allow a rapid characterization of the state of maturity of unknown composts. This methodology is comparable with the idea of JIMENEZ and GARCIA [27] of linking the polycondensation reactions to the state of maturity of the compost, but here the main advantage is facility and rapidity. The method, which enable us to follow the evolution of metal-binding during the biological process and to isolate specific compounds, is a first step in the determination of the nature of these compounds.

ACKNOWLEDGEMENTS We gratefully acknowledge the interest and the financial support of the A.D.E.M.E., and more especially Mile I. F E E .

REFERENCES BRUMMER GW. Heavy metal species, mobility and availability in soils. In : BERNHARD M, BRINCKMAN FE and SADLER PJ, edts. The importance of chemical speciation in Environmental processes. 1986; 1G9-192. CRAIG PJ. Chemical species in industrial discharges and effluents. In : BERNHARD M, BRINCKMAN FE and SADLER PJ, edts. The importance of chemical speciation in Environmental processes. 1986; 443-464. CHANEY RL. Metal speciation and interactions among elements affect trace element transfer in Agricultural and Environmental foodchain. In : KRAMER JR and ALLEN HE, edts. Metal speciation : theory, analysis and application. 1988; 219-260. DEL FAVA J. Speciation des metaux dans les Ordures Menageres et leur produi t de traitement par digestion anaerobie. These Universite de Montpellier 11, France. july 1992; 1-172. TESSIER A and CAMPBELL PGC. Partitioning of trace metal speciation. In : KRAMER JR and ALLEN HE, edts. Metal speciation : theory, analysis and application. 1988; 183-199.

215

6

7

8 9 10

11

12 13

14

15 16

17 18

GAMBLE DS. Interactions between natural organic polymers and metals in soil and fresh water systems. I n : BERNHARD M, BRINCKMAN FE and SADLER PJ, edts. The importance of chemical speciation in Environmental processes. 1986; 2 17-236. ROUSSEAUX P. Les metaux lourds d a i s les Ordures Menageres. Ministere de I'Environneinenr. A.N.R.E.D., edts, Paris France. 1988; 1-145. DUDM S and CHLOPECU A. Effect of solid-phase speciation on metal mobility and phytoavailabili ty in sludge-amended soil. Water Air Soil Pollut. 1990; 51: 153-160. LESCHBER R, DAVIS RD, L'HERMITE P, edts. Chemical methods for assessing bio-available metals in sludges and soils. 1985; 1-96. GRANET C, MILLOT N, WICKER A, NAVARRO A, VERON J. Application de la chromatographie de permearion sur gel aux lixiviats de decharge controlee. T.S.M. l'eau 1355; 5: 223-229. BERAIL G, PRUDENT P, MASSIANI C. Isolation of heavy metalbinding proteins from a brown seaweed Cystoseira Barbata f. repem cultivated in copper or cadmium enriched seawater. I n : MERIAN E and HAERDI W, edts. Metal compounds in environment a i d life,4, (interrelation between chemistry and biology). 1991; 55-62. THOMAS 0 aid GALLOT S . I_Jltravioletniultiwavelength absorptiometry examination of natural waters-part I : general considerations. Fres. J. Anal. Chem. 1390; 338: 231-240. GALLOT S a i d THOMAS 0. State of the a r t for the examination of U.V. spectra of waters and wastewaters. Int. J. Environ. And. Chem. (in press). GALLOT S and THOMAS 0. Fast and easy interpretation of a set of absorption spectra : theory a i d qualitative applications for 1J.V. examination of waters a i d wastewaters. Fres. J. A n d . Chem. ( i n press). THOMAS 0, hL4ZAS N , GALLO?' S, CLEMENT B. Fast interpretation of 3D chromatographic signals : application to the study of organic materials in wastewaters. Intern. J. En\,iron. A n d . Chem. (in press). KARAPANAGIOI'IS NK, STERRIT RM, LESTER JN. Heavy metal binding by the polymeric organic fractions of sewage sludges. Environ. Pollut. 1990; 67: 259-278. KAKAPANAGIOTIS NK, STERRIT R M , LESTER JN. Heavy metal complexation in sludge-mended soil-The role of organic matter in metal retention. Environ. Technol. 1991; 12: 1107-1116. LAKE DL, KIRK WW, LESTER JN. Fractionation, characterisation aid speciation of heavy metals in sewage sludge aid sludge-amended soils : a review. J. Environ. Qual. 1984; 13 ( 2 ) : 175-183.

216

19 LAKE DL, KIRK WW, LESTER JN. Heavy metal solids association in sewage sludges. Wat. Res. 1989; 23 (3): 285-291. 2 0 OTTAVIANI M, SANTARSIERO A, DE FULVIO S. Heavy metals in sewage sludge utilized in agriculture. Ann. 1st. Super. Sanita. 1989; 25 (3): 525-530. 2 1 VAN DE KERKHOVE JM. Evolution de la maturite de trois dechets urbaiiis en cours de compostage. These INPL, Nancy France. sept. 1990; 1-77. 2 2 MOREL JL, GUCKERT A, NICOLARDOT B, et al. Etude de I'evolution des carracteristiques physicochimiques et de la stabilite biologique des Ordures Menageres au cours du compostage. Agronomie 1986; 8: 693-701, 23 HARADA Y and INOKO A. Relationship between cation-exchange capacity and degree of maturity of city refuse composts. Soil Sci. Plant Nutr. 1980; 2 6 (3): 353-362. 2 3 GAMBLE DS, SCHNEZER M, KERNDORFF H, LANGFORD CH. Multiple metal ion exchange equilibria with humic acid. Geochim. Cosmochim. Acta 1983; 47: 13311-1323. 25 CASTETBON A, CORRALES M, ASTRUC M et al.Comparative study of heavy metal complexatioii by fulvic acid. Environ. Technol. Let. 1986; 7 : 495-500. 2G GREGSON SK and ALLOWAY BJ. Gel permeation chromatography studies on the speciation of lead in solutions of heavily polluted soils. J. Soil Sci. 1984; 35: 55-61. 27 JIMENEZ EI and GARCIA VP. Evaluation of city refuse compost maturity : a review. Biol. Wastes. 1989; 27: 115-142.

217

E f f e c t of several industrial respiratory activity E.

wastes

on

soil

Marti, R . Cruafias, M.A. Garau, E. de Miguel and M.T. Felip6

Laboratori d'Edafologia, Facultat de Farmacia, Universitat de Barcelona, Avinguda Diagonal 643, 08028 Barcelona, Spain. INTRODUCTION The waste generation from diverse sources is increasing during the last decades, most of the residues are eliminated by landfill disposal and this way is not always necessary because with previous valorization they can be reused for different purposes. The chemical complexity and the heterogeneous nature of many wastes makes often difficult their exhaustive characterization. Moreover, sometimes it can be more interesting to know the effects rather than the detailed chemical composition. Consequently, to assess their ecotoxicological properties it seems suitable to combine a basic physico-chemical characterization with sets of biological tests covering toxicity, biodegradation or bioaccumulation. Several ecotoxicological tests have been proposed since 1908 when the first one was devised by Kolwitz and Marston ( 1 ) to detect any kind of pollution. Some tests use single-species bioassays for estimating human and environmental risks. Different bioassays are already available for water pollution assessment. Diverse indicator organisms (bacteria, algae, protozoa, rotifers, insect larvae, cladocerans, fishes,etc.) are used (2). Toxicity to the fresh water crustacea Daphnia magna is also used as a criterion to qualify toxic wastes, and the same occurs with the luminescent properties of a population of the marine bacterium Photobacterium phosphoreum (Microtox bioassay). This last assay has been utilized (3) to predict the land treatability of hazardous organic wastes. The evaluation of hazardous organic materials by these bioassays have also been applied to monitoring subsurface leachates from hazardous waste sites and contaminated groundwater ( 4 ) . In the opinion of other authors ( 5 ) these tests that use single-species are unrealistic because: a) species do not live isolated, b) favored test species may not be indigenous to the receiving ecosystem, c) toxic chemicals may not remain unchanged in a particular ecosystem, and d) toxic chemicals rarely occur alone. They propose a multispecies test based on the colonization of artificial substrates by microbial species with cosmopolitan distribution. Dealing with waste recycling or disposal through terrestrial environment and taking into account that soil microbial community plays the major role in the cycling and degradative processes of natural and xenobiotic compounds, it seems that the effect of wastes on soil microbial activity may be a better indicator than other organisms living in fresh or marine waters. In fact the researches on soil microbial biomass and total microbial activities (respiration, dehydrogenase activity), in addition to nitrogen transformations have been extensively used for testing

218

the side effects of environmental chemicals. Either the release of CO, or the uptake of O2 have generally been used to determine the microbial activity in soil, as the result of their easy determination. There are two possibilities when soil respiration is determined. The so called basal respiration, which derives from the available organic material in soil, and the stimulated respiration which is induced by the addition of easily degradable organic substrates. Both can easily be determinated in soil under different experimental conditions. Tests on soil respiration and nitrogen transformations have been adopted by several regulatory authorities in order to assess side-effects of pesticides on the soil microflora (6). The mesurement of soil respiration in the presence of agrochemicals and other potentially toxic substances is also used to evaluate their possible damage to the physiological functions of soil. Laboratory studies have also been used to know the rates of recovery of soils damaged by toxic compounds as well as by waste disposal or mining practices (7). Consequently the aim of this work is to study the influence of several insdustrial wastes on soil microbial activity, under different experimental conditions, in order to take into account soil respiration as ecotoxicological test to assess the risk associated to land waste reuse. MATERIALS AND METHODS

Samples from A horizons of two different soils were used in this study : Soil I - Haploxeroll with 2.9 % organic matter, 0 % CO,'* and pH = 6.8 Soil 11- Xerorthent with 3.8 % organic matter, 3 2 . 7 % CO,-, and pH = 7.5 Wet samples were sieved ( 0 < =2mm) and stored at 4°C before the determination of CO, evolution rates. Several wastes from different origin produced in Catalonia (NE Spain) were studied. Figure 1 shows the kind of wastes and the references of samples used in this paper. In order to obtain samples homogeneous in size these wastes were air-dried and ground ( 0 < = 0.5 mm) before their characterization. The elemental and mineralogical composition were determined by X-ray fluorescence and diffraction respectively. The physicochemical characteristics (pH, EC) and fertilizing potential (oxidable organic-C, N, P, K, Ca, Mg, Na and carbonates) were obtained by means of the methods usually employed for soil samples (8). The contaminant potential of wastes was evaluated by determination of several potentially toxic elements in the extract obtained by the Extraction Procedure (EP) method established by EPA (9) to qualify a waste as toxic. Respirometric assays were performed with different soil-waste mixtures. They were incubated in the dark at 30°C during a 2 5 days period. Mixtures were well homogeneized and distilled water was added to obtain a water content equivalent to 50 % of soil

219

I

water holding capacity. The CO, released by microbial activity

C ,;

CERAMIC WASTES

Amorphous material, quartz, feldspar. Amorphous material, quartz, feldspar, hematite. (Smelting industry). Amorphous material (99%). (Power station). Amorphous material, Fespinel, calcite, lime. Quartz, calcite, gypsum, halite. Quartz, gypsum, anhidrite, calcite, halite. ( Contry rock ) . Calcite, barite, quartz, dolomite, sphalerite, chlorite, pirite, feldsDar. (Contry rock). Calcite, quartz, sphalerite, pirite, chlorite, feldspar. ( Sludges ) . Quartz , do 1omite , caolinite, biotite, muscovite. feldsuar. Quartz, mullite, Fespinel, amorphous material. Quartz, Fe-spinel, lime, calcite, anhydrite, amorphous material. Calcite, caolinite, feldspar, quartz, talc. (Industrial + domestic wastewater treatment). Gypsum, calcite, quartz, feldspars. (Old industrial dumping). Quartz, calcite- dolomite, feldspars. ~

ER SLAGS EC EM ASHES FROM MUNICIPAL SOLID WASTE INCINERATOR

ES

~

MV2

LEAD-MINIG WASTES

FA POWER STATION FLY ASHES FC

PAPER INDUSTRY WASTE

PR

INDUSTRIAL SEWAGE SLUDGE

LS

~

~~

CONTAMINATED SOIL

SS

Figure 1. Samples, references and mineralogical composition was trapped on 0.5 M NaOH, being the excess titrated with 0.05 M HC1, daily during the first week and then at several days intervals.

220

Under these general experimental conditions several different respiration experiments were done in order to know the more suitable conditions to evaluate the effect of wastes on soil respiratory activity, according to the pourpose of this study. The studied variables were: . Soil type (soil I and 11) Stimulated or basal respiration according to the addition or not of a nutrient source (C,N,P) . Waste loading rates (from 0 to 500 Mg ha-') . Waste pretreatments before incubation (dry oxidation or water washing ) Five different experiments were conducted in duplicate and in all of them a control sample (soil alone) was included. The specific conditions for each one are the following: - In .the first experiment different soil-waste mixtures were prepared by using soil I and every waste at the rate equivalent to 500 Mg ha-' (on a dry weight basis). Also, an additional source of nutrients : 0.3 % C-Glucose, 0.017 % N-(NaNO,) and 0.005 % P-(KH,PO,) was added. - Several waste samples (ES, FC, LS, MV2, PR and SS) were selected to test their influence on soil 11 microbial activity. This kind of experiment was performed with nutrients addition (2nd experiment) and without it (3rd experiment). - In order to study the influence o f waste loading rates on soil respiratory activity some (ES, FC and LS) of the wastes previously selected were assayed at 50, 100, 250 and 500 Mg ha-' on soil I1 without nutrients addition (4th experiment). - Finally (5th experiment) wastes were subjected to different treatments before incubation as dry oxidation of one sample rich in organic matter ( L S ) and water washing of two samples (FC and E S ) with high electrical conductivity. Treated waste samples were added to soil I1 without nutrients at the rate equivalent to 500 Mg ha-'. '

.

RESULTS AND DISCUSSION

Wastes characterization The wastes elemental composition is given in tables 1 and 2. In general the prevailing elements are Si, A l , Fe and Ca. Sulphur was only detected in fly ashes (FA: 0.5 % and FC: 3.7 % ) and in industrial waste (LS: 5%). The microelements content is extremely variable depending on the waste nature and on the element considered. Mine wastes and municipal solid waste incineration ashes have high content of Zn, Pb and Ba. The concentration of Zn is also important in PR sample and extremely high in LS. To prevent environmental risk it is necessary to investigate the available fraction of total heavy metals content. The mineralogical composition of the wastes with the samples references used in this paper are shown in Figure 1. The amorphous material is specially high (up to 99% in some cases) in samples obtained by thermal treatment (ceramics, slags and ashes). The cristalline phase is dominated by quartz and calcite-dolomite in many samples, other minerals are also present according to the origin of the samples.

22 1 All samples have pH values above neutrality, being slags and ashes strongly basic (pH ranges from 8 to near 1 3 ) . The electrical conductivity (Table 3 ) is very high in one fly fly

Table 1 X-Ray Fluorescence Analysis. M a j o r elements ( % ) SiO, CaO Ma0 Na,O K,O CR CB

ER EC EM ES MV2 MV4 MP 1

FA FC PR

LS

ss

67.00 68.93 62.42 56.21 48.36 46.88 28.28 40.07 62.59 43.80 40.07 6.34 11.63 32.21

0.40 5.33 20.73 11.62 20.27 20.90 21.59 26.37 3.95 5.79 28.66 37.60 5.38 10.37

0.20 1.49 0.83 6.08 3.26 3.27 0.89 0.88 2.44 0.94 0.14 3.50 3.19 1.84

2.20 0.17 2.39 0.76 1.23 1.25 0.86 1.86 0.10 0.35 0.64 <0.05 0.54 2.45

2.60 2.93 4.76 1.77 1.35 1.44 0.93 2.79 5.36 1.43 1.58 0.12 0.56 3.12

A1,0,

Fe,O,

26.00 12.73 7.86 10.03 7.37 7.61 7.88 8.78 12.98 24.87 17.71 10.77 6.54 7.03

0.80 6.46 CO.01 10.42 4.32 4.53 1.12 2.15 2.73 20.75 6.45 <0.01 7.80 2.39

Table 2 X-Ray Fluorescence Analysis. M i n o r elements (mg Kg-') SamDle Ba Cr cu Ni Pb CR CB 768 60 24 34 110 ER 744 (5 17 16 9 EC 984 130 49 120 <5 EM

ES MV2 MV4 MP1

FA FC

PR LS

ss

1216 1329 21 3582 1588 639 459 162 3078 585

140 156 58 19 19 186 91 120 6311 58

659 613 156 81 35

70 47 46 279 204

57 66 25 14 9 93 47 <5 99 44

2207 3008 1678 947 3125 69 27 18 398 292

Zn

127 39 98 3172 3845 1338 6506 2497 278 103 2480 22222 403

ash (FC) a n d also in both municipal solid waste incineration ashes. Most waste samples have low fertilizing potential concerning to organic matter, nitrogen, phosphorus and potassium, except for

222

samples LS and SS, as it is shown in table 4. Some wastes (EM, ES) are specially rich in exchangeable bases. Fly ashes (FA, FC) and samples PR, LS and SS have high calcium content. The potentially toxic elements content evaluated by EP is given in table 5. It is important to point out that only the mine Table 3 Some characteristics of the wastes studied. Sample PH E.C. CaCO, % dS rn-'

7.40 7.64 9.39 8.08 8.53 8.56 7.30 7.23 7.57 11.88 12.95 7.88 7.43 7.10

CR CB ER EC EM ES MV2 MV4 MP 1 FA FC PR LS

ss

0.12 0.17 0.15 0.20 9.98 8.75 0.40 0.16 0.89 1.72 9.39 1.31 4.71 4.63

organic C %

1.03 2.32 4.93 3.90 12.48 10.50 33.40 39.90 9 -30 0.03 0.09 47.30 7.53 9.11

0.01 0.07 0.26 0.29 1.53 1.73 0.07 0.11 0.50 0.03 0.02 12.53 14.23 9.83

Table 4 Fertilizing potential (mg Kg-I). Sample CR CB ER EC EM ES MV2 MV4 MP 1 FA FC PR LS

ss

N

p2°5

85 679 421 511 330 768 43 61 27 363 320 112 1937 4581

Na,O 218 85 57 477 11862 11456 34 47 117 20 244 584 6320 1184

K20 142 412 47 444 2564 2564 66 85 459 10 227 169 545 7325

CaO 1222 2702 5355 4771 33038 34718 9687 9021 8302 25432 184649 31016 17807 14951

MgO 1356 416 521 992 5353 6038 511 529 2279 896 7736 3452 1473 1861

wastes and municipal solid waste incineration ashes have a Pb

223 content higher than the threshold limit. The concentration of the other elements is always lower than the allowable limit in all samples. Therefore, samples MV2, MV4, EM and ES must be qualified as toxic according to EPA criteria. Table 5 Potentially toxic elements extracted by E.P. ( m g L - l ) Sample A9 As Ba (0.03
Cd <0.02 1

9,

ss EPA limit

<0.05 5.0

<0.25 5.0

0.12 100

1.01 0.64 0.17 0.22 0.56 0.02 0.04 (0.02
Table 5 (continued) Potentially toxic elements extracted by E.P. (mg L - ' ) -

Sample CR

CB ER EC EM ES MV2 MV4 MP 1 FA FC PR LS

Cr <0.1

<0.05
H9

< o .1

0.007 0.001 0.002 0.002

0.29
SS

<0.05

0.001 0.002 0.001 0.002 0.001 0.001 10.001 <0.001 <=o.001

EPA limit

5.0

0.2

<0.1

0.16

0.19


Pb

0.001

Se <0.5

,*

<0.25
,

< o .1 0.77 <0.2

1.45

5.0

<0.5


224 Respiratory activity First experiment. From figure 2 it can be observed that wastes with high organic matter content (LS and PR) enhance soil respiratory activity. Sample LS has high total Zn content, despite that it is not in an available form, as it could be observed by analysis of EPA leachates. Probably Zn remains

Figure 2. Cumulative CO, released at various soil-waste mixtures (soil I with nutrients addition).

immobilized, consequently, negative effects due to heavy metals considered by other authors are not observed in this case. The effect of other wastes with lower organic matter content was not so remarkable. Cumulative curves corresponding to mixtures containing samples CR, CB, ER, EC, MV2, FA and MP1 remained in the area delimited by the control and MV4 curves. Sample FC shows total inhibition of the process, probably due to high pH and electrical conductivity. These results are in agreement with those obtained by Wong and Wong (10). Second experiment. A general decrease in respiratory activity with regard to first experiment is observed (Figure 3 ) . Soil 11 releases only a 75% of carbon dioxide released by soil I, showing for this reason a better sensibility to detect adverse effects of wastes on respiratory activity. Samples LS and P R increased soil C02 production as in the first experiment, and the inhibition due to sample FC is again observed. The behaviour of samples with lower organic carbon content is very similar to the control. Third experiment. In this case the soil-waste mixtures were incubated without additional source of nutrients and the results are shown in figure 4. The increase of carbon dioxide released by samples LS and PR with regard to control was considerably lower than in previous experiments. Sample FC inhibites again

225

;2500i 3000

12000

Figure 3 . Cumulative CO, released at various soil- waste mixtures (soil I1 with nutrients addition).

Figure 4. Cumulative CO, evolved at various soil- waste mixtures (soil I1 without nutrient addition). soil microbial activity and samples with <2% organic carbon (MV2 and ES) showed a slight inhibition on CO, evolution. The two last samples have been qualified as toxic according to EPA criteria, due to the presence of available Pb. ES sample has also high electrical conductivity. Consequently it can be said that nutrient addition may act as

226

a counteracting agent of possible negative effects of residues. So no nutrient addition appears to be more suitable in order to evaluate the influence of wastes on soil respiration. Fourth experiment. Three selected wastes, LS, ES and FC were tested at different loading rates, and the results are shown in figures 5, 6 and 7. Sample LS logically enhances soil microbial activity

0

f-

-

INCUBATION TIME ( d a y s ) 50 100 250 -*- 500

+0

-+

Figure 5. Cumulative COz evolved at different loading rates. (soil I1 without nutrients addition, LS waste). proportionally to the applied rate for its high organic matter content. The behaviour of mixtures containing sample ES is close to the control. Sample FC produces an important inhibition even at the lower rate (50 Mg ha-'), it may be due to high pH and electrical conductivity, according with Pichtel (11) who observed inhibitions of soil respiration ranging from 93 up to 100% after addition of an alkaline fly ash at rates from 5 up to 20% respectively (equivalent to 1 5 0 - 6 0 0 Mg ha-'). Fifth experiment. Samples selected in the former experiment were subjected to different treatments before incubation in order to identify which waste characteristic is the resposible for soil activity changes. The treatments performed were: dry oxidation to eliminate organic matter for sample LS, and water washing to eliminate soluble salts for samples ES and FC. Figure 8 shows cumulative COz released by soil-waste mixtures prepared with treated and non-treated LS sample. The behaviour of treated sample is very close to the control soil. So, in this case organic matter content is the responsible agent of soil activity increase. The general trend of washed and non-washed ES sample is similar to the control, but a slight increase on COz released can be observed in the washed one, this demonstrates the influence of soluble salts on the microbial activity (Figure 9).

221 ..

3

Figure 6. Cumulative CO, released at different loading rates. (soil I1 without nutrients addition, ES sample).

Figure 7. Cumulative CO, released at different loading rates. (soil I1 without nutrients addition, FC sample). The electrical conductivity in FC sample remained constant with successive washings, and differences among washing and nonwashing on biological activity could not be observed (Figure 10). In this case it can be presumed that high electrical conductivity was not due to soluble salts but to alkaline hydrolisis of calcium oxide contained in the fly ash. This hydrolitic process

228

-

INCUBATION T I M E (days) Non t r e a t e d --t T r e a t e d

0

Control

-X

Figure 8. Influence of LS sample treatment before incubation on CO, evolution.

\

”

g

600-

400-

1 0

5 0

-Control

10 15 20 INCUBATION TIMf (days) -Nan t r e a t e d t- T r e a t e d

25

Figure 9. Influence of ES sample treatment before incubation on CO, evolution. generates p H values around 13, which can also contribute to microbial inhibition. In addition pH causes an overestimation of the inhibition, because carbon dioxide released is probably trapped by the equilibrium solution, neutralyzing alkalinity and forming calcium carbonate. Thus CO, involved in this process could not be quantified.

229

400 0

200

Z

0

r-

o

0

-Control

5

10

lNCUBATION ++ Non

15 T I t l E (daqs)

20

25

treated +Treated

Figure 10. Influence of FC sample treatment before incubation on C02 evolution.

CONCLUSIONS

From the obtained results it can be concluded that: ReSQirOmetriC method seems suitable €or assessing waste influence on soil microbial activity. To elucidate this influence is recommended the use of low biological activity soils and basal respiration quantification. - In these previously selected conditions soil respiration is both enhanced or inhibited, according to the waste nature. The effect is proportional to the loading rates. - Pretreatments of wastes can be useful to identify which waste characteristics have more influence on soil activity, and also to predict medium and long-term effects due to waste evolution in soil. Organic matter, pH, electrical conductivity and potentially toxic elements are in this case the most influencing waste characteristics. - When samples have high pH values or alkaline hydrolisis oxide content an overestimation of respiration inhibition could be observed.

-

ACKNOWLEDGEMENTS

The authors wish to express their appreciation to "Institut de Cihcies de la Terra Jaume Almera (CSIC)" for the X-Ray diffraction and fluorescence analysis and also to "Servei d'Espectroscbpia de la Universitat de Barcelona" €or the determination of potentially toxic elements. Partial financial support was provided by Spanish Government

230

through the following projects: PB-87-0463-C02-02 and NAT-911340. REFERENCES 1

2

3 4

5 6

7 8

9 10

11

Kolwitz R, Marson R. Int Rev Ges Hydrobiol Hydrogeol 1908; 2: 52. Couillard Y, Ross P, Pinel-Alloul B. Toxicity Assess 1989; 4:451-462. Mattews JE, Bulich AA. ASTM Committee on Solid Waste Treatment (ASTM Committee D-34) Washington 1984. Calleja A, Baldasana JM, Mulet A. Toxicity Assess 1986; 1:7383. Cairns J Jr., Pratt JR, Niederlehner BR, Mc Cormick PV. Environ Monit Assess 1986; 6: 207-220. Gerber HR, Anderson JPE, Bugel-Mogensen B, Castle D, Domsh KH, Malkomes HP, Somerville L, Arnold DJ, Van de Werf H, Verbeken R, Vonk JW. Toxicol Environ Chem 1991; 30: 249-261. Nannipieri P, Grego S, Ceccanti B in: S o i l Biochemistry: Vol. 6. Bollag JM, Stotzky G eds. Marcel Dekker Inc. New Kork, 1990. Page A L , Miller RH, Keeney DR, eds. Methods of soil analysis: Part 2. Chemical and microbiological properties. Agronomy 9. Am SOC Agron. Wisconsin : Madison, 1982. U.S.E.P.A. 1980. Federal Register 5-19-80, vol 45 , np 98. Appendix 11. EP Toxicity Tests Procedure. Washington D.C. Wong MH, Wong JWC. Environ Pollut (Series A) 1986; 40: 127144. Pichtel JR. Environ Pollut 1990; 63: 225-237.

23 I

LONG-TERM EVALUATION OF PLANTS AND ANIMALS COLONIZING CONTAMINATED DREDGED MATERIAL PLACED IN UPLAND AND WETLAND ENVIRONMENTS D. L. Brandonb, C. R. Lee', and G.S. Wilhelmb

J. W. Simmers", J. G. Skogerboe",

'Environmental Laboratory, US Army Engineer Waterways Experiment Station, Vicksburg, Mississippi, USA bThe Morton Arboretum, Lisle, Illinois, USA INTRODUCTION

Navigable waterways of the United States play a vital role in the nation's development. The Corps of Engineers (CE), in fulfilling it's mission to maintain, improve and extend these waterways, is responsible for the dredging and disposal of large quantities of sediment each year. Nationwide, the CE dredges about 300 million cubic yards annually at a cost of about $ 4 5 0 million. Approximately ten percent of this dredged material is considered contaminated. The CE has conducted extensive research and development in the field of dredged material management. Federal expenditures on dredged material research, monitoring, and management activities have cumulatively exceeded $100 million. Contaminated sediment was dredged from Black Rock Harbor, Connecticut, in October 1983 and placed in aquatic, upland, and wetland environments as part of a joint CE/Environmental Protection Agency Field Verification Program (FVP), 1981-1986 [l]. Prior to dredging, laboratory tests were conducted to evaluate potential contaminant mobility under each of the disposal alternatives. Upland tests (i.e., plant and earthworm bioassays), wetland tests (i.e., plant, sandworm, snail, and mussel bioassays), and aquatic tests (i.e., toxicity and bioaccumulation) were conducted. Upland and wetland laboratory test results were subsequently field verified at "Tongue Point", Bridgeport, CT (Figure 1). The results of the upland disposal and wetland creation portions of the FVP, and the changes occurring since the completion of the FVP are summarized herein. This chapter emphasizes the contaminant mobility of heavy metals using data collected between 1983 and 1989.

UPLAND ENVIRONMENT

Prior to construction, the field site was used as an open dump. The debris consisted mainly of urban building rubble (i.e., bricks, cement, and rebar). Both upland and wetland ecosystems were present as were assortments of flora and fauna characteristic to each. During the initial field survey

232

Figure 1. Location of the FVP field test site, Bridgeport, CT, conducted in July 1983, plants were collected along four transects that included both upland and wetland areas, Phrasmites australis (Common Reed), Solanum dulcamara (Nightshade), Robinia meudoacacia (Black Locust), and Pomlus deltoides (Cottonwood) were collected and the tissues analyzed for metals (Table 1). The observed concentrations of heavy metals were typical of any urban vacant lot containing building rubble in the northeast USA. Both plant species and the general rubble constituents were typical of the neighborhood adjacent to "Tongue Point". The elevated cadmium concentrations in PoDulus deltoides are indications of potential contaminant mobility. These trees have been shown to enrich significantly the cadmium concentrations in leaflitter, thereby increasing the cadmium concentrations in many soil invertebrate species [2]. The field survey described the abundance and diversity of the animal species on site. These include rats, insects, snakes, and several species of birds [31.

233

Table 1 Chemical Analysis of the most Abundant Field Survey Plant Species Plant Species

Zn*

Cd

cu

Ni

Cr

Pb

Hg

UDland Species Phrasmites australis

29.2 (0.6)

0.13 (0.01)

8.9 (1.8)

5.7 (3.2)

0.71 (1.0)

1.4 0.003 (0.01) (0.0)

Solanum dulcamara

46.8 (0.0)

1.33 (0.0)

19.6 (0.0)

30.8 (0.0)

2.7 (0.0)

6.4 (0.0)

0.018 (0.0)

3.7 (2.9)

0.03 (0.04)

4.3 (1.5)

0.01 (0.01)

Robinia pseudoacacia PODUlUS deltoides

38.6 (3.5) 688 (239)

0.48 (0.05) 4.85 (1.3)

10.7 (3.7) 15.6 (4.7)

20.5 (4.8) 27.3 (20)

2.17 (0.7)

0.04 (0.1)

Wetland Species Phracnnites** australis

22.5 (9.5)

0.17 (0.1)

3.6 (1.2)

5.6 (2.9)

1.1 (1.7)

0.003 (0.01)

178 (70)

1.29 (0.9)

14.0 (2.5)

16.0 (3.2)

1.4 (2.4)

0.0 (0.0)

Distichlis*** 3 6 . 2 spicata (0.0)

0.02 (0.0)

4.5 (0.0)

4.5 (0.0)

0.0

(0.0)

0.0 (0.0)

SDartina*** patens

0.18 (0.0)

6.1 (0.0)

3.4 (0.0)

6.7 (0.0)

0.0 (0.0)

Solidaqo** semervirens

* ** *** (

)

16.0 (0.0)

ug/g dry weight high marsh with freshwater influence low marsh with tidal influence standard deviation

All grading and dike construction was performed with conventional equipment. Total surface area was approximately 2 , 6 0 0 m2 (Figure 2 ) . A weir with adjustable risers was installed to control overflow. Additional information on site design and preconstruction evaluations is described elsewhere [l]. The characteristics of the Black Rock Harbor sediment are presented in Table 2 . Further discussion of upland construction is described elsewhere [ 4 ] .

234

UPLAND SITE

-REPLANTED PRECWSTRUClION PLANTS PLAN15 FRDM COMMERCIAL SUPPLIER

-

BRIDGEPORT REACH

Figure 2. Schematic of the FVP Field Test Site Table 2 Characterization of Black Rock Harbor Sediment Parameter Zn Cd cu Ni Cr Pb H9 As Fe PCB as Aroclor 1242 PCB as Aroclor 1254 PP-DDD Phenanthrene Fluoranthene Benzo(a)pyrene Sum of PAHs Oil and Grease Chemical Oxygen Demand Total Organic Carbon % Organic Matter % CaCO, equivalent % Total Sulfur Wet sediment pH Dry sediment pH % Sand % Silt % Clay Salinity ppt

Content* 1307.1 22.4 2728.4 178.8 1651.0 397.8 2.0 22.9 31000.0 5.5 9.3 0.9 5.0 6.3 3.9 142.0 17452.4 232880.0 54104.0 19.8 1.0 1.3 7.6 6.6 42.0 47.0 11.0 28

* contaminant concentration units are ug/g dry weight unless denoted otherwise

235

Physicochemical Changes in Dredged Material

Placement of Black Rock Harbor sediment in an upland disposal environment resulted in pronounced changes in some attributes of the contaminated sediment. Following upland disposal, the dredged material dried and oxidized. The pH of the dredged material dropped from 7.6 in 1983 to 3.2 by the end of the FVP in 1986 (Table 3). This substantial decrease in dredged material pH was presumably a result of the decomposition of organic matter, the oxidation of sulfide, limited pH buffering capacity in Black Rock Harbor sediment, low percent CaCO, equivalent, and the acid rainfall of pH 4 . 0 4.5 at the field site location [ 5 ] . This substantial decrease in the pH of the dredged material enhanced the solubility and availability of the toxic metals: zinc, cadmium, copper, nickel, chromium, and lead [ 4 ] . These harsh conditions of extreme acid pH, excess soluble toxic metals, and excess salinity, resulted in the death of all plant species planted. Concomitantly, there was no colonization of the site by animals. After placement in an upland environment, control plots of unamended dredged material have been barren of vegetation from 1985 through 1989 (Figures 3a and 3b) even after repeated attempts to establish acid tolerant, salt tolerant, and/or metal tolerant plant species. Table 3 Changes in Soil Conditions After Upland Disposal of Contaminated Estuarine Dredged Material Control Plots Unamended Parameter

Amended*

10/83 11/85 6/86 10/89

10/83 11/85 6/86 10/89

Soil pH

7.6

3.2

3.2

3.4

7.6

NS**

NS

4.4

Salinity ppt

28

29

13

<1

28

29

13

<2

19.5

NS

NS

7.7

19.5

NS

NS

8.5

Organic Matter

%

* Lime + Manure + Sand + Gravel * * No Sample

236

( a ) 1985

( b ) 1989 Figure 3. Unamended Black Rock Harbor Dredged Material

237 Upland P l a n t s

The FVP laboratory test predicted that SDartina alterniflora would not grow in unamended dredged material in the field. This was verified as no S. alterniflora plants survived in the unamended dredged material. The lab test also predicted plant species might grow in amended material. SDorobolus virsinicus survived in amended material in the lab as well as the field [ 4 ] . Field tests were conducted using soil amendments (Table 4 ) such as lime, manure, and sand plus limestone gravel to stabilize the dredged material, control erosion, and alter the dredged material for plant and animal colonization. These soil amendments were selected to counteract the extensive chemical and physical changes that occurred when Black Rock Harbor dredged material dried and oxidized. As dredged material organic matter oxidized on drying, organic acids were produced to lower pH and organic bound metals were released. As sulfides oxidized, sulfuric acid was produced to lower soil pH and sulfide bound metals were released as soluble sulfate metals [6]. Since sediment CaCO, content was low at 1.0 percent, limited buffering of sediment pH occurred and the pH dropped from 7.6 to 3.2. Lime was added to neutralize the acids generated upon drying and oxidation of the dredged material. Manure was added to replace the organic matter oxidized; to increase surface metal adsorption; to immobilize metals released when organic bound metals were oxidized; and to ameliorate the toxic effects of excess salinity. Before exposure to the contaminated dredged material, a layer of sand and gravel was selected to provide a non-toxic microhabitat, provide a substrate for germination of seeds and establishment of plants. Coarse limestone gravel was placed on the surface to neutralize acid rainfall that might impact the plot and also provide release of limestone neutralizing material over the longer term. The site was partitioned into plots with each amendment being assigned randomly to 4 plots (Figure 4 ) . The lime and lime + manure amendments were surface broadcast and then a surface layer of sand and limestone gravel was placed over the lime and manure. A selection of acid, salt, and metal tolerant plant species were planted on the amended dredged material in an attempt to establish vegetation. Table 4 and Figures 5, 6, and 7 show the plant and amendment combinations during four attempts to vegetate this disposal site. All plant species were broadcast as seeds except Paswalum vauinatum and Sporobolus virainicus. These species were plugged.

238

Table 4 Phases of Restoration Evaluated on an Contaminated Estuarine Dredged Material Placed in an Upland Disposal Environment

I

Phase 1: ADDllCallOn of Sol1 Amendments. SeDtember 1984 1. Control 2. Lime 3. Lime +sand

+ gravel

4. Lime + manure

5. Lime + manure

+ sand + gravel

No amendments 28.2 metric tons /hectare (mtfha) 28.2 mtlha, lime 13-cm surface layer, sand 6.6-cm surface layer, limestone gravel 28.2 mtfha, lime 112 mVha, horse manure 28.2 rntlha, lime 112 mtfha, horse manure 13-cm surface layer, sand 6.6-cm surface layer, limestone gravel

Phase 2: First Seedlng of Cool-Seeson Gross Specles, October 1984 1. Puccinellia distans (alkali grass) 2. Fesfuca elafior (tall fescue 'Houndog') 3. Agrosfis alba (redtop) + Pod cornpressa (Canada bluegrass) 4. Aaroovron elonaaturn ltall wheatarass 'Alkar')

1. fuccine/l/adfsfans (alkali grass) 2. Fesfuca elafior (tall fescue 'Alta') 3. Agrosfis alba (redtop)

Phase 4: First Seeding of Warm-Season Grass Species, June 1985 1 Puccinellia distans (alkali grass) 2. Fesfucarubra (red fescue) 3 Sporobolus virginicus 'Fine' Sporobolus virginicus 'Coarse' 4 Agrosfis fenuis (seaside bentgrass) Phase 5: Rototllllng (lo a depth of 15 cm) o n Selected Plots, June 1986

I

1. Rototilled one quarter of control unamended plots 2. Reapplied and rototilled lime (56.4 mtfha) on 114 of lime plot 3. Reamlied and rototilled lime (56.4 mtlhal on 114 of lime + manure olot Phase 6: Planted Rototllled Portions of Selected Plots, June 1986 1. Paspalurn vaginafum 'FS-3' 2. Sporobolus virginicus 'Fine'

I

I1

Seedlng RatesTToleranl(s) Puccinellia dislans (alkali grass) Festuca elafior (tall fescue 'Houndog') Agrosfis aiba (redtop) Poa compressa (Canada bluegrass) Agropyron elongalurn (tall wheatgrass 'Alkar') Festuca rubra (red fescue) Agrosfis fenuis (seaside bentgrass)

11 kgihaisalt 32 kglha /acid, salt 11 kglha lacid, salt 32 kg/ha/acid 23 kg/ha/salt 32 kglhaimetal 11 kgihaisalt

Pluqqing RatesTTolerant(%) Paspalum vaginafurn FS-3' Sporobolus virginrcus Flne Soorobolus virginicus Coarse

1 per sq ft'salt 1 par sq fVsalt 1 oer sq ftfsait

239 3 96 m 4

-

h

A

0 WEIR

[TCONTROL

UME SAND GRAVEL

=LIME

r/M E &F>jLIME MANURE MAhURE SAND GWYEL

Figure 4 . Schematic of the Upland Site Soil Amendments (Table 4: Phase 1 )

3 96 rn 4-b

0 WEIR

&?

17CONTROL

LIME. SAND. GRAVEL

LIME

LIME, MANURE

LIME. MANURE. SAND GRAVEL

Figure 5. Schematic of the Upland Site Soil Amendments and Cool Season Grass Species (Table 4: Phases 2 and 3 )

240

PLANTS ~~~

I.Alka11Grass

2. Red Fescue

3.Sparooolua

dig,"#cw 'Fen*'

'Coarse' 4. Seas'de

Bentgrass

LIME

/.<.']LIME SAND GRAVEL

LIME,MANURE

LIME MANURE SAND GRAVEL

Figure 6. Schematic of the Upland Site Soil Amendments and Warm Season Grass Species (Table 4: Phase 4)

3.96 rn 1

L

Figure 7. Schematic of the Upland Site Soil Amendments and Warm Season Grass Species (Table 4 : Phases 5 and 6 )

24 I

Amendments of lime, lime + sand + gravel, lime + manure, and lime + manure + sand + gravel resulted in vegetation becoming established on the dredged material (Table 5). Of the plant species seeded on the field plots, only Aarostis alba (Redtop) became permanently established. Phrasmites australis (Common Reed) had extended its rhizomes from the upland containment facility dike of construction rubble onto some of the plots. Other plant species observed on the plots in 1989 are typical Eurasian species and urban weeds that probably originated from seed sources in the Bridgeport area (Table 5). The best vegetative establishment was observed on the lime + manure + sand + gravel amended plots in 1985 (Figure 8a). In 1989, these plots were almost completely covered with vegetation (Figure 8b, Table 5). Apparently, the sand + gravel cover allowed rainfall to soak the surface applied lime and manure into the surface of the dredged material enhancing plant growth and establishment. The lime and lime + manure plots showed 51 and 28 percent cover, respectively, in 1989 (Table 5). However, subplots receiving an additional application of 56.4 mt/ha of lime and rototilling resulted in a soil pH of 4.4 and 4.1 and produced 96 and 99 percent vegetative cover in 1989 (Table 5). Mixing lime into the surface material appeared to have greatly improved plant growth and vegetative cover [7]. vegetative cover plays a significant role in improving surface runoff water quality 181. Vegetation also decreases wind-borne and water-borne erosion, aids in dewatering the site, prevents pools of standing water, and provides an aesthetically pleasing cover. Continued long-term evaluation of the FVP site will identify plant species which maximize the properties listed above and minimize contaminant mobility.

242

Table 5 Influence of Soil Amendments on the Growth, Diversity, and Percent Cover of Plant Species Colonizing Estuarine Dredged Material Placed in an Upland Disposal Environment for 1989 Plant Species %

Treatment

Cover AGREL AGRAL A R T W

Control Tilled*

DIGSA

PHRAV POACO OTHER

0 0

0 0

0 0

0 0

0 0

0 0

0 0

0 0

Lime Tilled

51 96

0 0

36 85

1 1

6 4

4 1

1 1

3 4

Lime+Manure Tilled

28 99

0 0

2 94

0 5

0 0

25 0

0 0

1 0

Lime+Sand+ Gravel

50

0

45

0

2

3

0

0

Lime+Manure+ 97 Sand+Gravel

4

58

1

9

22

2

1

* In 1986, 1/4 of the plot was rototilled and received an additional application of lime. AGREL AGRAL ARTW DIGSA PHRAV POACO

= AqrODYrOn elonqatum (Tall Wheatgrass) = Agrostis alba (Redtop) = Artemisia vulsaris (Common Sage) =

Diqitaria sansuinalis (Hairy Crabgrass)

= Phrasmites australis (Common Reed) = Pea compressa (Canada Bluegrass)

OTHER = Aster pilosus (Hairy Aster) = Daucus carota (Wild Carrot) = EDilobium coloratum (Cinnamon Willow Herb) = Erechtites hieracifolia (Fire Weed) = Lactuca scariola (Pricky Lettuce) = Linaria vulsaris (Butter’n Eggs) = Oenothera biennis (Evening Primose) = Phleum Dratense (Timothy Grass) = Puccinellia distans (Alkali Grass) = Rumex crisws (Curly Dock) = Solidaso altissima (Tall Goldenrod) = Solidago (Seaside Goldenrod) = Verbascum thaDsus (Woolly Mullein)

243

Comparison to FV? upland Plant Results FVP laboratory and field upland plant bioassay results are shown in Table 6 and summarized below. When mean concentrations are statistically different, it can be concluded that one is higher than the other (i.e,, elevated). FVP Sworobolus virsinicus lab test results indicated that plant contents of zinc, cadmium, and chromium would be elevated; field results showed they were elevated. Lab tests also indicated that nickel, and copper would be elevated; however, field test showed they were not elevated. Lab tests indicated low lead contents; field-grown plants, however, had higher lead contents [4]. Sporobolus viruinicus only survived one growing season in the field. It would have been instructive to compare FVP results with current trends using S . virsinicus as an index plant. However, the plant species which survived on site are different from those evaluated in the FVP. Plant tissues from the dominant species were collected in 1988 and 1989.

Table 6 Plant Content (ug/g) of Selected Metals in Leaf Tissue of S. virainicus Grown in Sediment from Black Rock Harbor Greenhouse Metal

Field

Orisinal Sediment Washed Sediment Flooded UrJland Upland

Wetland I Flooded I

Upland

No Survival

66.Oab

Zn

26.2c*

40.lb

86.3a

Cd

0.857b

0.684b

1.34ab

2.22a

cu

10.7b

24.3ab

34.5a

19.8b

Ni

6.82bc

22.8a

13.4b

5.38~

Cr


8.64a

0.506b

7.64a

Pb

<0.013b

<0.013b

<0.013b

1.56a

* Means within a row followed by the same letter are not significantly different at P = 0.10 using the Least Significant Difference method. Reprinted from Table 13, [ 4 ] . The chemical analyses of Aarostis alba (Redtop) tissues are presented in Table 7. Redtop zinc, cadmium, and chromium mean concentrations have the same general magnitude as the FVP results. Hence, one might designated these concentrations as "elevated". The nickel concentrations have the same general magnitude as those designated "elevated" during FVP. The copper concentrations don't have the same relative magnitude from treatment to treatment or from year to year. Therefore,

244

Table 7 Tissue Contaminant Contents of Asrostis i&&i the FVP Upland Disposal Site Contaminant

1988

1989

LIMF Zn* Cd cu N i

Cr Pb Hg

N=4 138.7 (40.1) 1.5 (0.9) 92.2 (80.0) 11.9 (5.1) 26.5 ( 9 . 2 ) 6.2 (4.0) 0 . 1 (0.01)

N= 3 85.4 ( 7 6 . 3 ) 1.1 ( 0 . 1 ) 15.0 (1.8) 13.4 ( 4 . 0 ) 6.7 (0.8) 3.1 (1.7) 0.1 (0.03)

LIME + MANURE Zn Cd cu N i

Cr Pb Hg

N=3 105.0 (27.5) 1.0 (0.3) (8.0) 21.6 12.8 (4.0) 11.2 ( 3 . 3 ) 4.6 (1.4) 0.1 (0.01)

LIME Zn Cd cu N i

Cr Pb Hg

N= 110.1 1.7 23.6 13.7 11.1 6.6 0.2

LIME

+

SAND

N i

Cr Pb Hg

* (

)

(0.0) (0.0) (0.0) (0.0) (0.0) (0.0) (0.0)

ug/g dry weight standard deviation

(32.2) (0.8) (5.4) (4.4) (4.0) (3.7) (0.01)

GRAVEL N=4

(0.0) (0.0) (0.0) (0.0) (0.0) (0.0) (0.0)

iMANURE

111.0 1.1 140.9 13.7 24.4 8.6 0.1

+

1

N= 1

Zn Cd cu

N=3 124.9 1.1 15.3 15.2 10.6 6.4 0.05

128.5 1.5 101.0 13.5 34.8 11.0 0.07

+

SAND

(57.8) (0.8) (77.0) (5.3) (27.6) (7.9) (0.03)

+ GRAVEL

N=2 102.5 (26.0) 1.4 (0.06) 79.9 (73.1) 11.9 (1.4) (28.5) 36.0 6.5 (6.6) 0.1 (0.02)

Plants Growing on

245

classifying the copper concentrations would be impractical. Redtop lead concentrations are higher than the FVP field results and should be designated "elevated". Differences between FVP results, 1988, and 1989 concentration trends could be due to plant physiology and/or site conditions. Contamination of Plants Lee et al. [ 6 ] lists plant tissue information from a number

of sources to indicate demonstrated effects of contaminants on plants (Table 8). Using these data as guidance, Asrostis alba tissue contents of zinc are within the normal range of 15-150 ug/g found in agricultural crops. Plant cadmium concentrations are equal to or slightly above the normal range and substantially below the critical content level of 8 ug/g (Table 8). Copper concentrations appear to be either in the normal range or slightly elevated above phytotoxic levels in some samples. However, there is considerable variability in the analyses of the elevated samples. Future sampling and evaluation is needed to produce more precise results. Nickel concentrations were equal to or slightly above the critical content level. Chromium concentrations were above normal in 1988 and three of the four amendments showed tissue content above phytotoxic levels in 1989. Lead concentrations were equal to or slightly above the normal range. The only mercury reference tissue concentration available was 1.0 ug/g in wheat kernels as an action level for human foodstuff. Mercury contents of &. alba were approximately one tenth this action level, and therefore should not be of concern. Upland Animals

The upland animal bioassay predicted that earthworms would not survive in this dredged material under oxidized conditions. Through 1986, earthworms could not survive on this dredged material [ 4 ] . Presently, control plots devoid of vegetation contain few animals (Table 9). In particular, no soft bodied animals (i.e., slugs) were observed or collected from this environment. Those animals that were collected or observed were transient foraging arthropod species. In contrast, establishment of vegetation on the amended dredged material enhanced the abundance and diversity of animals present. There are numerous species of macroinvertebrates associated with the plant cover and the leaf litter layer of the soil. The animals observed (Table 9), while relatively abundant, have provided too little biomass per species for chemical analysis. Consequently, no data are available to evaluate contaminant uptake.

246

Table 8 Demonstrated Effects of Contaminants on Plants Plant Growth Effect-Contaminant Content, mq/ku leaves Critical 10% Yield 25% Yield Contaminant Normal’ Content’ Reduction‘ Reduction’ Phvtotoxic‘

---

--

3-10

775

---

--

75

Cadmium

0.1-1

8

15

Varies

5-700

Cobalt

0.01-0.3

---

--

--

25-100

--

--

20

Arsenic Boron

0.1-1

Chromium 0.1-1 (111), oxides Copper

3-20

20

20

20-40

1-5

---

--

25-40

---

Iron

30-300

Manganese

15-150

----

Molybdenum

0.1-3.0

--

---

0.1-5

11

26

50-100

2-5

--

--

--

--

--

100

--

--

10

290

500

Fluorine

Nickel

Selenium

0.1-2

Vanadium

0.1-1

----

15-150

200

Lead

Zinc

’ ’

-500

--

400-2000 100 500-1000

500-1500

From Chaney (1983) From Davis, Beckett, and Wollan (1978); Davis and Beckett (1978); Beckett and Davis (1977) From Chaney et al. (1978)

Reprinted from Table C-7, [6].

241

Table 9 Influence of Soil Amendments on the Abundance and Diversity of Invertebrate Animals Colonizing Estuarine Dredged Material Placed in an Upland Disposal Environment for 1 9 8 9 . Treatment

GAS

ORT

DIP

Animal Taxa+ HYM COL COP

OTHER TOTAL

ARA

Control

0

4

10

9

1

2

6

0

32

Lime

42

24

22

27

61

13

13

2

204

Lime+Manure

35

26

9

24

43

18

20

0

175

Lime+Sand+ Gravel

26

9

9

15

51

15

24

2

151

12

4

124f

10

35

29

3

223

Lime+Manure+ 6 Sand+Gravel

+ GAS

= Mollusca, Class Gastropoda (slugs, and snails) ORT = Insecta, Order Orthroptera (crickets, and grasshoppers) DIP = Insecta, Order Diptera (flys) HYM = Insecta, Order Hymenoptera (wasps, bees, and ants) COL = Insecta, Order Collembola (springtails) COP = Insecta, Order Coleoptera (beetles) ARA = Arthropoda, Class Arachnida (spiders, ticks, and harvestmen)

OTHER

=

Crustacea (woodlice) Insecta, Hemiptera (bugs) Insecta, Lepidoptera (butterflies, and moths) Insecta, Homoptera (leafhoppers)

* One trap was near an ant hill

248

( a ) 1985

(b) 1989

Figure 8. Black Rock Harbor Dredged Material Amended with Lime + Manure + Sand + Gravel

249 WETLAND ENVIRONMENT

The initial field survey of "Tongue Point" also included wetland areas. Plants were collected along several transects. The species collected included Phraamites australis (Common Reed), Solidauo sempervirens (Seaside Goldenrod), Distichlis spicata (Spikegrass, Saltgrass), Juncus aerardii (Blackgrass), Limonium carolinianum (Sea Lavender), Spartina Datens (Saltmeadow Grass), and Spartina alterniflora (Smooth Cordgrass) (Table 1). These 3 . alterniflora metal contents were in the range of northeast United States natural marsh Spartina alterniflora plants [9]. The field survey also listed animals which inhabited the stagnant and intertidal ponds. These include fish, shrimp, sandworms, snails, mussels, clams, and several species of crabs [3]. Prior to construction, SDartina alterniflora was collected from 650 m2 of the wetland. The construction involved the excavation of material to achieve the desired elevation. The total surface area was approximately 550 m2. A weir was installed and allowed an interchange of tidal flow with tidal pools within the "Tongue Point". At high tide, the water level within the site reaches a depth of approximately 0.3 meters [lo]. Further discussion of wetland construction is provided elsewhere [lo]. Wetland Plants

Spartina alterniflora and SDorobolus virginicus were used in laboratory and field bioassays. Laboratory tests indicated that the contaminated sediment was not toxic to the saltmarsh plants SDartina alterniflora or SDorobolus virainicus when placed in a wetland environment. 5. alterniflora survived well in the field test. However, S . virainicus did not survive in the field [lo]. Creation of a saltmarsh wetland with Black Rock Harbor dredged material has been successful. One half of the created FVP wetland was planted with Spartina alterniflora supplied by Environmental Concern, St. Michaels, Maryland (Figure 2 ) . Initial growth of Environmental Concern's transplants on the FVP field site appeared to be slow up to 1986, then in 1987, 1988, and 1989, the vegetation on the created wetland gradually expanded until the side planted with Environmental Concern's transplants was covered by a dense stand of S . alterniflora (Figures 9a and 9b). The highest biomass production was observed in 1987 (Table 10). The other half of the wetland was planted with native S . alterniflora collected prior to construction of the dredged material created wetland. These transplants were slower to grow in 1986 and 1987 but exceeded the Environmental Concern's transplants in 1988 and 1989 (Table 10). However, the native plants did not cover a large portion of the wetland (Figure 9b, left side).

250

(a) 1983

(b) 1989 Figure 9. Wetland Created with Black Rock Harbor Dredged Material

25 I

Table 10 Biomass Production (g/m2)of SDartina alterniflora in the FVP Wetland Created with Black Rock Harbor Dredged Material 1986

1987

1988

1989

Environmental Concern

511

798

226

297

Native Transplants

311

535

337

468

* The natural marsh biomass production was

627, [9].

As the wetland extended across the marsh creation site the most robust plant growth was observed in the vicinity of the outer edge of the marsh as it expanded to vegetate the more open areas. Consequently, the 1988 and 1989 sample data reflects an area of wetland behind the advancing edge of the marsh. This was especially true for the Environmental Concern’s transplanted wetland. These biomass yields have been in the range of naturally occurring saltmarshes in the northeast United States [9]. These naturally occurring saltmarsh plant samples included robust plants on the edge of SDartina alterniflora saltmarshes when such plants were present.

Comparison t o FVP Wetland Plant Results

FVP laboratory and field tests with Spartina alterniflora are presented in Table 11 and summarized below. When mean concentrations are statistically different, it can be concluded that one is higher than the other (i.e., elevated). Zinc contents of lab grown plants were not different from that of the 1985 field grown plants, but were significantly lower than the 1986 field grown plants; Cadmium contents of lab grown plants were significantly greater than the 1985 and 1986 field grown plants; Copper and nickel contents of lab plants were not significantly different from the 1985 and 1986 field plants. Chromium and lead contents of lab grown plants were significantly lower than the 1985 field grown plants, but not significantly different from the 1986 field grown plants [lo]. Using the FVP results (i.e., laboratory, 1985, and 1986 field data) to indicate elevated concentration levels provides a method to evaluate the 1988 and 1989 S . alterniflora data (Table 12). This evaluation is presented in Table 13 and summarized herein.

252

Table 11 Leaf Tissue Content of Selected Heavy Metals in S . alterniflora Grown in the Laboratory, in the Field, and Predicted from DTPA Sediment Extraction Data using the Equations of Lee, Folsom, and Bates (1983) Concentrations, uu/u Field-Grown Heavy Metal

Laboratory (N = 4 )

1985 (N = 7)

Predicted by DTPA (N = 3)

1986

(N = 71

ZN

12.1*

c**

13.5

BC

19.2

Cd

0.041

B

0.021

c

<0.0025 C

cu

4.02

A

5.65

A

7.48

Ni

0.954

A

4.23

A

0.743

Cr

0.274

B

10.4

A

6.17

AB

1.63

Pb

0.237

B

3.45

A

0.945

AB

0.70 AB

41.7

A

0.196

A

A

2.70

A

A

0.346

A

B+

AB

* Mean * * Letters in a row indicate statistical groupings of means (square root transformation, Waller-Duncan k-ratio t-test)

+ Elevated mean concentrations are underlined Reprinted from Table 4, [lo]

253

Table 1 2 Tissue Contaminant Contents (ug/g) of Spartina alterniflora Grown on a Contaminated Estuarine Dredged Material from Black Rock Harbor, CT Heavy Metal

Natural Marsh**

Field Collected Prior***

1985

1986

1988

N=7 N=7 N=8 Concentration, uu/u

N= 7

zn

44.3 (24.8)

22.5 (9.5)

13.5 (5.0)

1 9 * 2+ (7.1)

21.1 (4.3)

Cd

0.20 (0.19)

0.17 (0.11)

0.02 (0.05)

cO.003 (0.0)

0.25 (0.05)

cu

7.16 (2.16)

3.62 (1.18)

5.65 (1.74)

7.48 (5.55)

16.5 (8.9)

14.0 (13.1)

Ni

2.47 (1.76)

5.64 (2.90)

4.23 (6.13)

0.74 (0.68)

1.1 (0.4)

1.7 (0.9)

Cr

3.41 (1.8)

1.11 (1.70)

6.17

(8.2)

(5.5)

5.7 (3.4)

6.3 (3.9)

4.85 (6.5)

2.17 (0.80)

(4.9)

0.95 (0.9)

3.4 (2.0)

3.8 (3.0)

0.027 (0.02)

0.003 (0.01)

Hg

10.4 3.45

--

--

0.02 (0.003)

*

N

+

[lo] Elevated mean concentrations are underlined standard deviation

=

Number of samples collected and analyzed.

* * from [ 9 ] * * * samples collected prior to construction )

N=9

N=2 0

Pb

(

1989

20.3 (8.2) 0.23 (0.03)

0.02 (0.007)

254

Table 13 Evaluation of 1988 and 1989 SDartina data using the FVP Laboratory and Field Results ~-

Laboratorv*

~

Concentrationsl-us/s Field-Grown

Heavy Metal

(N = 4 )

1985 ( N = 7)

1986 IN = 7 1

Zn

not elevated

possibly elevated

elevated

elevated elevated

Cd

elevated

not elevated

not elevated

elevated elevated

cu

not elevated

not elevated

not elevated

---

---

Ni

not elevated

not elevated

not elevated

---

---

Cr

not elevated

elevated

possibly elevated

not possibly elevated elevated

Pb

not elevated

elevated

possibly elevated

possibly elevated elevated

1988** IN = 8)

1989** IN = 9 )

* not elevated implies Spartina concentrations from this wetland test aren't statistically greater than concentrations from any remaining wetland test; elevated implies SDartina concentrations from this wetland test are statistically greater than concentrations from at least one remaining wetland test; possibly elevated implies SDartina concentrations from this wetland test aren't statistically different from concentrations from the two remaining wetland tests. However, concentrations from the two remaining tests are statistically different (i.e., one is elevated and the other isn't).

**

---

1988 and 1989 SDartina concentrations are evaluated using the criteria above (i.e., in Table 11, the 1986 mean zinc concentration 19.2 ug/g is "elevated". Since the 1988 mean zinc concentration exceeds this value, this concentration is given the same designation as the 1986 mean zinc concentration).

cannot be determined

255

In Table 11, the 1986 field mean zinc concentration is 19.2 ug/g. It is statistically different (i.e., elevated) from the laboratory concentration 12.1 ug/g. The 1986 mean zinc concentration is designated "elevated" in Table 13. The laboratory mean zinc concentration is designated "not elevated". The 1985 field mean zinc concentration 13.5 is not statistically different from 19.2 or 12.1. It is designated "possibly elevated". Since the 1988 and 1989 (Table 12) mean zinc concentrations exceeds 19.2, these concentrations are given the same designation as the 1986 mean zinc concentration (i.e. , "elevated") . The laboratory mean cadmium concentration 0.041 ug/g (Table 11) is designated as elevated in Table 13. Since the 1988 and 1989 (Table 12) mean cadmium concentrations exceed 0.041, these concentrations are also designated "elevated" in Table 13. The FVP copper and nickel mean concentrations (Table 13) weren't given a designation since none of tests were statistically different. The 1988 and 1989 copper concentrations appear to be higher than the laboratory, 1985, and 1986 field test. The 1988 and 1989 nickel concentrations appear to be higher than the laboratory and 1986 field test but lower than the 1985 field test. The 1986 mean chromium concentration 6.17 ug/g (Table 11) is designated "possibly elevated" in Table 13 (i.e., not statistically different from the elevated concentration 10.4 or the nonelevated concentration 0.274). Since the 1988 and 1989 (Table 12) mean chromium concentrations fail to exceed and exceed 6.17 respectively, these concentrations are designated "not elevated" and "possibly elevated". The 1986 mean lead concentration 0.945 ug/g (Table 11) is designated "possibly elevated" in Table 13 (i.e.,not statistically different from the elevated concentration 3.45 or the nonelevated concentration 0.237). Since the 1988 (Table 12) mean lead concentration exceeds 0.945, it is designated "possibly elevated". The 1989 (Table 12) mean lead concentration 3.8 is designated "elevated". Contamination of Saltmarsh Wetland Plants The 1988 and 1989 plant zinc, cadmium, nickel lead, and

mercury tissue concentrations are generally no greater than those measured in the naturally occurring SDartina alterniflora at "Tongue Point" prior to wetland creation or those measured in nearby naturally occurring saltmarshes (Table 12). Copper and chromium tissue concentrations tended to be higher than the natural marsh or preconstruction concentrations in 1985, 1986, 1988, and 1989. Wetland Animals

Animal bioassay results from static and tidal simulation tests indicated that tidal simulation procedures are superior to static test for measuring uptake by organisms in the intertidal wetland habitat. Comparison of FVP field-collected animal data with laboratory tidal bioassay suggests that tidal simulation bioassay procedures are overpredictive of PCB

256

congener, hexachlorobenzene, and DDE bioaccumulation. No clear pattern between laboratory and field tests emerged for metals [ l o ] . Native sandworms Nereis succinea colonized the wetland in 1 9 8 6 . Since 1 9 8 6 , fish, crabs, and snails have been observed in the FVP created wetland. Contamination of Saltmarsh Wetland Animals

Snails Ilvanassa (=Nassarius) obsoleta were collected in 1988 and 1 9 8 9 and have been analyzed for contaminant contents (Table 1 4 ) . The 1988 and 1989 copper, cadmium, and mercury

concentrations are less than the respective concentrations of FVP laboratory control snails. It was noted that I. obsoleta typically contained elevated levels of copper possibly due to the high copper concentration in the respiratory pigment, haemocyanin ( 2 0 0 atoms per mole). Zinc, nickel, chromium and lead concentrations were not measured in the FVP control animals. Since there are no FDA guidelines in the United States, Australian values are presented as a potentially useful guideline. Lead concentrations in snail tissues appear to be approaching the level ( 1 5 ppm) that Australia has established as FDA type guidelines f o r non-specific seafood [ 6 ] . Cadmium contents are below the 10 ppm level that Australia has established as FDA type guidelines for molluscs, Table 1 4 Tissue Contaminant Contents (ug/g) of Snails Ilvanassa obsoleta Exposed to a Contaminated Estuarine Dredged Material Control Parameter

Eu.8 N= 1

zn cu Cd Ni Cr Pb H9

* **

+

( )

NS+

2913 8.6

NS NS NS 0.26

878.18 1335.68 2.93 8.79 9.02 10.21 0.08

1989 N=2

675.2 1881.7 3.6 13.3 29.7 16.1 0.1

31.9) 574)

0.5) 1.6) 16.0) 5.3) 0.05)

from Table 9 , [ l o ] 1 composite sample No sample standard deviation

INTERIM CONCLUSIONS

Upland control plots of unamended dredged material were barren of vegetation from 1 9 8 5 to the present time. Six years after placement in an upland disposal environment, unamended estuarine dredged material decreased in salinity from a high of 2 9 ppt after drying out to < 1 ppt. However, soil pH

251

remained extremely acidic at 3 . 4 (Table 3 ) , keeping toxic metals soluble and available to plants that attempt to colonize the unamended dredged material. In contrast, amended dredged material became vegetated and soil and vegetationdwelling macroinvertebrates have begun to colonize these plots and food webs involving vertebrates have evolved. During the FVP, as predicted, Spartina alterniflora did not survive in an upland environment. Results from FVP SDorobolus virsinicus tests were compared to 1988 and 1989 Redtop tissue concentrations. Redtop zinc, cadmium, and chromium were designated "elevated". These designations were consistent with the FVP lab bioassay and field test results. Redtop nickel and lead concentrations were designated "elevated". The nickel designation was consistent with the FVP lab bioassay while the lead designation was consistent with the FVP upland field test. The Redtop copper concentration wasn't categorized since no consistent trends exist. upland animals didn't yield sufficient biomass to allow chemical analysis. Therefore, contaminant mobility into upland animals cannot be assessed at this time. Both the native and introduced Spartina alterniflora plants are currently thriving in the wetland environment. FVP S . alterniflora concentrations were compared to 1988 and 1989 S . slterniflora tissue concentrations (Table 13). The 1988 and 1989 zinc and cadmium concentrations were designated "elevated". The 1988 and 1989 copper and nickel concentrations weren't assigned a designation. The 1988 and 1989 chromium concentrations were designated "not elevated" and "possibly elevated", respectively. The 1988 and 1989 lead concentrations were designated "possibly elevated" and "elevated", respectively. Neither FVP lab bioassay or field tests predicted current concentration levels of copper and cadmium in wetland plants. Snails collected in the field in 1988 and 1989 had copper, cadmium and mercury concentrations lower than the FVP control animals. Zinc, nickel, chromium and lead concentrations were not measured in the FVP control animals. In both the upland disposal and wetland creation field sites, there are developing plant and animal interrelations. The extent of the populations and the species compositions of the ecosystems may require management procedures if unanticipated routes of contaminant mobility develop. Continued evaluation will better define the extent and nature of contaminant mobility at the FVP site where contaminated estuarine dredged material was placed simultaneously in an upland and a wetland environment. This evaluation should include the contaminant mobility of organics into plants and animals in both environments. REFERENCES 1 2

Peddicord, RK. 1988; Technical Report D-88-6, US Army Engineer Waterways Experiment Station, Vicksburg, MS. Stafford, EA, Simmers, JW, Rhett, RG, Brown, CP. 1991; Miscellaneous Paper D-91-17, US Army Engineer Waterways

258

Experiment Station, Vicksburg, MS. Stewart, LL, Moffat, D, Buchholz, K, Coon, M. 1983; The University of Connecticut, Groton, CT (unpublished manuscript). Folsom, BL Jr, Skogerboe, JG, Palermo, MR., Simmers, JW, et al. 1988; Technical Report D-88-7, US Army Engineer Waterways Experiment Station, Vicksburg, MS. Skogerboe, JG, Lee, CR, Price, RA, Brandon, DL, et al. 1987; Miscellaneous Paper D-87-1, US Army Engineer Waterways Experiment Station, Vicksburg, MS. Lee, CR, Tatem, HE, Brandon, DL, Kay, SH, et al. 1991; Miscellaneous Paper D-91-1, US Army Engineer Waterways Experiment Station, Vicksburg, MS. Brandon, DL, Lee, CR, Simmers, J W , Skogerboe, JG, et al. 1991; Miscellaneous Paper D-91-5, US Army Engineer Waterways Experiment Station, Vicksburg, MS. Skogerboe, JG, Lee, CR. 1987; EEDP-02-3, Environmental Laboratory, US Army Engineer Waterways Experiment Station, Vicksburg , MS. Simmers, JW, Folsom, BL Jr, Lee, CR, Bates, DJ. 1981; Technical Report EL-81-5, US Army Engineer Waterways ExDeriment Station. Vicksbura. MS. 10 Sikers, JW, Rhett; RG, Kay, SH, Folsorn, BL Jr. 1989; Technical Report D-89-2, US Army Engineer Waterways Experiment Station, Vicksburg, MS. -

r

259

CONTAMINATED AQUATIC SEDIMENTS AND WASTE SITES AS TOXIC CHEMICAL TIME BOMBS Ulrich Forstner Hamburg-Harburg University of Technology, Environmental Engineering Section, Box 901052 2 1071 Ham burg, Germany Abstract Long-term effects of contaminants will be the major subject of geochemistry in environmental protection technology. Within the conceptual perspectives of hture landfill operations "final storage quality indicates the ultimate goal of geochemical engineering. Final storage conditions can be reached either by incineration and post-treatment (municipal solid waste) or be incorporation into low-temperature mineral formations which remain stable over geological times. 1. Introduction

A chemical time bomb is a waste deposit or contaminated hotspot which initially appears to be relatively benign, but which can eventually have disastrous environmental effects as geochemical conditions change and toxic contaminants are released. The effects of these time bombs are non-linear and delayed (e.g., toxic metals can "break through" once the specific buffering capacity of a sediment or soil system has been surpassed [ 13). Consequently, these sites require proactive assessment and management.

This new and challenging field in environmental geochemistry is concerned with the impact of humans on element cycling in air, water, soil and on biota. "Geochemical engineering" [3, 41 (Figure 1) applies geochemical principles (such as concentration, stabilization, solidification, and other forms of long-term, self-containing barriers) to determine the mobilization and biological availability of critical pollutants.

Geochemical Engineering A

T

Atmosphere

t Water

t Sediment/Sail

t

Mineral/Ore

Waste Materials

Environmental Geochemistry

t Mineral Exploratlon

t

f ProdLlcts .) Production

+

Rau, Materials

Magma/Rock

Geochemistry

A

I

Applied Geochemistry

Figure 1 New Developments in Environmental Geochemistry [2]

"Technology' 'Economics'

260

Inclusion of the time factor moves beyond a traditional chemical approach. It also transcends the civil engineering approach in waste management, which usually devotes little attention to long-term emissions from waste disposal sites. "Because we have become accustomed to considering the filling period as the most important phase in landfill operation, we have forgotten that subsequent to the active working period there is the infinitely long time in which the site has to finction as a depository for all materials unwanted in the biosphere" [S]. In modem waste management, the fields of geochemically oriented environmental technology include:

the study of material fluxes within and between the anthroposphere and different "geospheres"; the optimization of elemental distribution at high-temperature processes (e.g. incineration of solid waste materials); 4

the selection of favourable milieu conditions for the deposition of large-volume wastes, such as dredged materials, the selection of additives for the solidification and stabilization of hazardous waste materials, the development of test procedures for long-term prognoses of pollutant behaviour in all kinds of waste depositories.

This presentation decribes characteristic "time bomb" mobilization processes of metals, following which a review will be given on remediation techniques which are based on geochemical experience. For example, one economic method for storing large volumes of dredged material is to incorporate metals into newly formed sulphides, which are candidates for "sub-sediment storage" due to their low solubility. Municipal solid waste can be treated to produce an earth crust-like material which renders future management unnecessary, by combining high temperature treatment, leaching of soluble constituents, and stabilization upon disposal.

2. Effects of Redox- and pH-Variations on the Mobility of Metals

For systems involving solutiodsolid interactions, "mobility" typically reflects the flux of metal species in a certain medium, which contains both accelerating and inhibiting factors and processes [6]. Accelerating factors comprise effects of pH-lowering, redox changes, inorganic and organic complexation, and microbially mediated species transformations such as biomethylation. Within the spectrum of "barriers", physical processes include adsorption, sedimentation, and filtration; chemical barriers comprise mechanisms such as complexation and precipitation; and biological barriers are often associated with membrane processes, which can limit translocation of metals (e.g., from plant roots to the shoots and h i t s ) . "Complexation" in its various forms can both inhibit and accelerate metal fluxes, particularly in biological systems consisting of different types of membranes.

26 1 2.1 Oxidation of Sulphides and Proton Release

Acidity is perhaps the most serious long-term threat from metal-bearing wastes. For decades, water seeping from mine refuse has delivered increased metal concentrations into receiving waters. The threat is especially great in waters with little buffer capacity (i.e., in carbonate-poor areas where dissolved metal pollution can be spread over great distances). The acidity production can develop many years after disposal, once the neutralizing or buffering capacity in a pyrite-containing waste is exceeded. The major process (see equations 1 to 3) affecting the lowering of pH-values (down to pH 2 to 3) is the exposure of pyrite (FeSJ and other sulphide minerals to atmospheric oxygen and moisture, whereby the sulphidic component is oxidized to sulphate and acidity (H--ions) is generated. Bacterial action can assist the oxidation of Fe"(aq) in the presence of dissolved oxygen. 4 FeS

+ 9 0, + 10 q 0 = 4 Fe(OH), + 4 SO+:

8 H', or

4 FeS, + 15 0, + 14 %O = 4 Fe(OH),+ 8 SO,"+ 16 H'

(1) (2)

Pyrite

The acidification of a sedimenvwater system begins after hydrogen ions are generated during the oxidation (e.g., during dredging or resuspension of mainly fine grained material containing less carbonate than needed for long-term neutralization [7,81. Primary emissions containing high metal concentration issue from waste rocks and tailings, while tailing ponds are primarily responsible for secondary effects on groundwater. An important and long-term source of metals are the sediments reworked from the floodplain, mainly by repeated oxidation and reduction processes [ 9 ] . High concentration factors were found in inland waters affected by acidic mine effluents [ 101.

2.2 The Concept of Acid-Producing Potential

The concept of acid-producing potential (APP) was initially developed in the prediction and calculation of acid mine drainage and waste tailings management ([ 1I]; as summarized by Ferguson and Erickson [12]). Our findings on the effects of periodical redox processes on both APP and metal mobility in estuarine sediments [13, 141 have hrther enhanced research interest in this field. Periodical redox processes can spur an increase or decrease in APP or pH in a sedimendwater system [IS]. In a closed system, periodical redox processes can lead to the change or transfer between APP(s) and APP(aq) but the total APP of the system does not change. The processes are reversible. The hydrogen ions produced in the oxidation will be consumed by the following reduction. Contrarily, in an open system, the total APP of the system will change depending on the properties of the system and the reaction processes. Under certain conditions total APP in the system increases, while under other conditions total APP in the system decreases. Some processes are irreversible. The components producing or consuming H" leave the system and cause the change in APP(s), APP(aq) and permanent ANC (acid neutralizing capacity). The following chemical changes affecting increases or decreasing of the ANC in aerobic and anaerobic samples have been found (Figure 2):

262

Ferrolysis

Fe-reduction 8c displacement of sulfate & nitrate adsorbed on solid phase

[so,] (-qq c"Z",jrG;i I f

r-_--

i ZH*JOq;

'-r--

.

fldwata)

O2

. (ZOOZ]. co* 1

i

ANC(S)increases

Volatilization effect

Figure 2: Diagram illustrating the chemical changes in periodical aerobic (left) an anaerobic (right) samples. Components in broken-line rectangles are mobile. Small squares denote exchangeable sites. M stands for a divalent cation (Ca,' or Mg2') and C&O refers to oxidizable organic matter [ 161 Change of APP by the "Split" of Sulphate Permanent acidification in alternating aerobic and anaerobic systems [ 171 is proceeding in two stages [ 181: The first stage is characterized by the reduction of sulphates, in particular in tidal flats or seabottom sediments. Most of the suIphide formed is fixed in the sediment as FeS or FeS,. It leads to the increase in APP(s) while ANC(aq) (HCO;) formed during sulphate reduction (equations 4 and 5) is removed by tidal turbulence or by diffision into the overlying water. As a result mobile ANC(aq) (HCO,') and immobile potential acidity (FeSJ are separated or "split". The increase in APP(s) leads to a permanent decrease in the ANC(aq) after next aeration and oxidation and results in extreme acidification of the system.

SO,?-+2 w =)I2s+ 2 0, CO, + &O = HCO,' + W

(4)

(5)

263

If during the reduced stage M(HCO,), were retained in the system (e.g. as precipitated carbonate) the acidity formed by sulphide oxidation would be neutralized exactly by the carbonate formed without permanent change in ANC and APP.A similar process involving formation of FeS in young non-acid marine clay sediments may also lead to a rapid acidification of the system ~91. Change in APP by Ferric Iron Reduction and Displacement Reactions Alkalinization by femc reduction and desorption or displacement of SO,'- from the solid surface has been reported (1 8). In the reduction process (Figure 2) the sulphates adsorbed on positive colloids, basic iron sulphates or aluminium sulphates (e.g., jarosite and jurbanite) serve as proton acceptors in the reduction of femc iron: F%O, + I/2 C q O + 4 J T = 2 Fe" + 1/2 CO,

+ 5/2 q 0

(6)

The product ferrous sulphate, not ferrous hydrogen carbonate (Figure 2), appears in solution [16]. In fact OH' adsorbed as anion displaces the sulphate group adsorbed on the surface of the positive colloids. On the other hand part of the ferrous sulphate is oxidized at the interface between the solids and the water, producing a ferric oxide coating on the solid surface and releasing sulphuric acid in the water. The ferrous sulphate concentration gradient caused by these processes promotes hrther transport of ferrous sulphate to the sediment or soil surface. These processes will lead to a decrease in APP(s) and a permanent increase in ANC(s) by removal of SO, (€&SO,) [16]. Change in APP by "Ferrolysis" in Periodicat Redox Processes Ferrolysis ("dissolution by iron" [20 - 221) in redox processes proceeds in two stages (Figure 2). When the system becomes reduced, part of the Fe2- becomes exchangeable and displaces other cations such as Ca2- and M$-. The displaced cations together with the anions that appear simultaneously with dissolved Fez- (mainly bicarbonate and some organic anions [23]) can be removed by percolation or by difhsion into the surface water followed by lateral flow. If the supply of cations by flood water, ground water, or mineral weathering is negligible, the surface complex may eventually be depleted of bases. The depletion is not immediately apparent in the reduced stage when the pH is high. During aeration of the system, however, exchangeable Fezis oxidized to essentially insoluble Fe" oxide and W takes the place of adsorbed Fez*to the extent that formerly adsorbed base cations are leached [24]. It is suggested [16] that ferrolysis is typical for sediments or soils of older river or marine terraces in monsoon climates which have a seasonal perched water table caused by submergence with rain water. In those conditions, hydrology favours either lateral or vertical drainage and removal of bases liberated from the exchange sites by ferrous iron. Change in APP by Volatilization of E&S A sedimenuwater system containing sulphate may become alkaline after volatilization of K-S from periodic sulphate reduction and retention of M(HCO,), in a saline alkaline lake or marsh (Figure 2). If the system reoxidizes, it will only re acid^ partly, i.e., to such extent that reduced sulphur has been retained in the system (e.g., as FeS) and is available for the formation of sulphuric acid. The volatilization of KS from reduced sediments causes a part of acidity leave to the system. The APP of the system will decrease. The escape of %S from a sedimenvwater

264

system depends on temperature, atmospheric pressure, vapour pressure of YS,and the state of the system ( e g , the disturbance of the water phase). Although the increase of ANC in the system is permanent due to volatilizaton of gaseous KS,alkalinization due to sulphate reduction would be temporary if %S sulphur is fixed as iron sulphide and becomes oxidized later [16]. 2.3 Changes of pH-Conditions in River Sediments

Direct assessment of the pH-changes resulting fiom the oxidation of anoxic sediment constituents can be performed by ventilation of sediment suspensions with air or oxygen and subsequent determination of the pH-difference between the original sample and oxidized material. As greater is this difference. as higher is the short-term mobilization potential of metals, e.g. during dredging, resuspension and other processes, by which anoxic sediments get into contact with oxygenated water or - following land deposition of dredged material - with atmospheric oxygen. A typical example demonstrating the temporal development of redox and pH-values in a sludge suspension from Hamburg harbour is presented in Figure 3. Results from titration experiments using 1 molar nitric acid on sediment suspensions of 100 g/L are presented in Figure 4. The titration curve of the Rhine River sediment exhibits a small plateau in the pH-range of 5 . 5 and 6, probably due to a certain fiaction of carbonate, which is consumed by addition of 80 mmol of acidity. On contrary, the titration curves of both Elbe River sediments are continously decreasing as a result of the low contents of carbonate in these sarnples. The sediment from the inland harbor basin of Harburg, originally sulphide-rich material which has been stored for 1 year in a closed bottle, has already reached an initial-pH of 4.3; this is probably due to the consumption of the low residual buffer capacity by oxidation of parts of the sulphide fraction. PH

-

9 -

'T

T

600 6oo

500

--

400 400

Eh

200

--

100

--

0

[mv

-1 00

0

5

10

15

20

25

30

35

time [d]

Figure 3: Influence of Redox Potential on pH During Oxidation of a Sludge Suspension from Hamburg Harbor [25]

265

Figure 4 Variation of pH-Values from Titration Curves of Suspensions (1 00 g L-') of Sediment Samples from Rhine and Elbe Rivers after Addition of 1 M Nitric Acid [26] 0

2

0

4

0

6

0

8

0

mmol(H+) 2.4 Redox Influences on Metal Mobility in Sediments

It can be expected that changes from reducing to oxidizing conditions, which involve transformations of sulphides and a shift to more acid conditions, increase the mobility of typical "B-" or "chalcophilic"elements, such as Hg, Zn,Pb, Cu, and Cd. On the other hand, the mobility is characteristically lowered for Mn and Fe under oxidizing conditions (Table 1). Table 1 Relative Mobilities of Elements in Wastes and Soils as a Function of Eh and pH (Plant and Raiswell[27]) Relative

Very low mobility Low mobility

I

Electror Activity

Proton Activity Acid Si

AI, Cr, Mo, V, U, Se. S. B, Hg, Cu, Cd, Pb Si, K, P, Ni,

Si, K, P, Pb

Zn, Co, Fe

Si, K, P, Pb, Fe, Zn, Cd

Medium mobility

Mn

Co, Ni, Hg, Cu, Mn

High mobility

Ca, Na, Mg, Sr

Ca, Na, Mg. Sr, Ca, Na, Mg, Cr, Mo, V, U, Se Zn, Cd, Hg

Very high mobility

CI, I, Br

C1, I, Br, B

CI,I, Br, S, B, Mo, V, U, Se

K, Fe(II1)

C1, I, Br, B

266

Processes affecting metal mobilization from sulphide oxidation are highly complex. Factors involved are not only the extent of protonation, but also exchange processes involving interactions with Fe" and earth alkalies as well as the buffering effect of organic substances (Figure 5 [28]). Part of the Fe3t formed according to equation 1 (see above) is hydrolysed, another portion is kept in solution due to the formation of organic complexes. The dominant process depends upon the oxygen potential as well as from the concentration and reactivity of organic matter [29]. The organic substance inhibits, among others, the oxidation of Fe(I1). Exchange reactions (e.g., for adsorbed manganese and trace elements) can proceed on different ways: Fe'- + Ads-Met = Ads-Fe2" + Me2+ 2 H30'+ Ads-Me' = Ads-&+ + Me2' + &O

Ca'+ + Ads-Me' = Ads-CaZt+ Me2t Considered the possible precipiation of iron hydroxide, Fe'+ + 6 &O

= Fe(OH),

+ 3 &O+

the oxidation reaction of FeS can, hypothetically, completed by two reaction pathways: FeS + 2.25 O2+ 4.5 %O = Fe(OH), + SO,'-+ 2 k$O' FeS i- 2.25 O,+ &O'

+ Ads-Me'

=Ads-Fez'

+ SO:

+ Me2'+ 1.5 KO

Reaction (9a) is an acid-producing process, in contrast to reaction (9b). (Whether the decrease of the pH-values during oxidation is caused by this reaction or is due to the initialisation of microbial dissimilative activity and respective production of CO, [28]) is still controversial.)

I

OZ

FeS .MeS

Figure 5 Model Conception on the Oxidation of Metal Sulphides and the Interactions Between Matrix and the Newly Formed Reaction Products (Peiffer [28])

261

Organic matter obviously pfays a major role as a buffer for acidity. Either acid producing Fe3' is directly taken up or there is a permanent titration of the organic matter with protons from the hydroxid precipitation of Fe(OH),. This titration again leads to the release of metals according to equation (7b) and the gross reactions can be given as: FeS + 2.25 O,+ 3.5 KO + Ads-Me" == Fe(OH),+ SO+:'

Ads-%'

+ Me2+

(10)

According to reaction (9b) and reaction (lo), the concentration of released metals should be be equal to the sulphate concentration.

Experimental investigations by PeiRer [28] on the long-term development of sewage sludge materials provide detailed insight into the sequence of processes taking place in the post-methanogenic stage of such deposits (Figure 6).

2 02-Input 3 pH-lowering 4 Addition of Earth Alkaline Salts

2

3

4

1

,

5 5 pH-Increase

1 AnoxicPhase

6 Anodc Phase

.o 0

500

1000

1500 2000 2500 3000 3500 4000

4500 h

Figure 6: Experiments on Cadmium Release from Solid Waste Material (from Peiffer [28]) During an initial phase, anoxic cadmium is bound to sulphides, resulting in very low metal concentrations in solution (phase 1). Aeration by addition of dissolved oxygen initiates release of cadmium from the solid substrate (phase 2); this process is enhanced by the production of acidity, which lowers the pH from 6.7 to approximately 6.4 (phase 3). An even stronger effect is observed from the addition of alkaline earth ions (phase 4). The pH increase, which may be induced from buffering components within the system, leads to a reduction of dissolved cadmium (phase 5). Formation of new sulphide ions from the degradation of organic matter brings the concentrations of dissolved cadmium back to its original, extreme low level (phase 6). The observed pH decrease seems to indicate that zinc and cadmium are being exchanged for protons, whereas lead and copper, because of their much stronger bonding to the solid substrate, do not.

2.5 Metal Mobilisation from Aquatic Sediments (Examples)

Release of potentially toxic metals from contaminated sediments pose problems both in aquatic systems and subsequent to land deposition of dredged materials. Examples are given by various authors, indicating the major factors, processes and rates of metal mobilisation:

268

Field evidence for changing cadmium mobilities was reported by Holmes et al. [30] from Corpus Christi Bay Harbor: During the summer period when the harbor water was stagnant, cadmium precipitated as CdS at the sedimenvwater interface. In the winter months, however, the increased flow of oxygen-rich water into the bay r w l t e d in a release of the precipitated metal. In the St. Lawrence Estuary, Gendron et al. [31] found evidence for different release mechanisms near the sediment-water interface - the profiles for cobalt resemble those for manganese and iron with increased levels downwards, suggesting a mobilization of these elements in the reducing zone and a reprecipitation at the surface of the sediment profile. On the other hand, cadmium appears to be released at the surface, probably as a result of the aerobic remobilization of organically-bound cadmium.

-

Biological activities are typically involved in these processes remobilization of trace metals has been explained by the removal of sulphide fiom pore waters via ventilation of the upper sediment layer with oxic overlying water, allowing the enrichment of dissolved cadmium that would otherwise exhibit very low concentrations due to the of formation of insoluble sulphides in reduced, KS-containing sediments. Emerson et al. [32] suggest a significant enhancement of metal fluxes to the bottom waters by these mechanisms. It was evidenced by Hines et al. [33] from tracer experiments that biological activity in surface sediments greatly enhances remobilization of metals by the input of oxidized water. These processes are more effective during spring and summer than during the winter months. 4

From enclosure experiments in Narragansett Bay, Hunt and Smith [34] estimated that by mechanisms such as oxidation of organic and sulphidic material, the anthropogenic proportion of cadmium in marine sediments is released to the water within approximately three years. For remobilization of copper and lead, approximately 40 and 400 years, respectively, is needed, according to these extrapolations. Prause et al. [35] studied the release of Pb and Cd from contaminated dredged material after dumping in a harbour environment. During an observation period up to 24 h no significant Cd or Pb release could be found fiom the dredged sludge, but during the long term experiments extensive Cd remobilisation was recognized. The contact of the polluted freshwater sediments with seawater finally cause the release of 1-2 mg Cd per kg solids. It is suggested that the reaction kinetics, mainly with respect to the initial mobilization of metals from solids, are controlled by microbial activity.

Metal release from tidal Elbe River sediments by a process of "oxidative remobilization" has been described by Kersten ([36] Figure 7). Short (30 cm) sediment cores were taken from a site, where diurnal inundation of the fine-grained fluvial deposits takes place. In the upper part of the sediment column, total particulate cadmium content was approximately 10 mg kg-', whereas in the deeper anoxic zone the total particulate concentration of Cd was 20 mg kg-'. Sequential extractions indicate that in the anoxic zone 60-80% of the Cd was associated with the sulphidic/organic fraction. In the upper - oxic and transition - zone the association of Cd in the carbonatic and exchangeable fractions simultaneously increase up to 40% of total Cd. This distribution suggests that the release of metals from particulate phases into the pore water and hither transfer into biota is controlled by the frequent downward flux of oxygenated surface water. From the observed concentrations, it would be expected that long-term transfer of up to 50% of the Cd from the sediment subsurface would take place either into the anoxic zone located hrther below the sediment-water interface or released into the open water.

269

sediment a c c w l a t i o n

-

1988

I0

15

1984

12 16

.5

20

B

24

26

0

5

20

0

5

10

15

20

0

Cadnium partitioning (rq/kg)

1984

0.5

Cd in porewater

1988

Exchangeable Cd

1

1988

69 Fe/m oxide bound Cd Sulfidic/organic bound Cd

Surface oxide and carbonate bound Cd

III]Residual

Cabniwn

Figure 7: Total Concentrations and Partitioning of Cadmium in a Tidal Flat Sediment Profile in the Heuckenlock Area Sampled in 1984 and 1988. Sedimentation Rates were Determined by the "'Cs-method. Cadmium Pore Water Profde was Termined at Low Tide (Kersten [36])

Metal

Channel Sediment Porewater (a)

River Water Cone. (b)

Effluent at An Artificially Created Marsh

ah

Expected Cone.

I

Measured Cone.

Hg (I@) Cd cu Zn

6.94 57.3

%Change

0.03

230

1.34

1.19

0.26

220

11.12

6.01

- 11 - 46

~

Mn Fe Ni Pb

I

0.054

0,001

54

0.01 1

0.035

+ 218

0.077

0.002

38

0.016

0.142

+ 788 + 144

3.2

0.26

12

0.82

2.0

0.009 0.012

0.001

9

0.0025

0.019

0.004

3

0.0055

0.05 1

0.12

0.052

2

0.065

5.30

+ 660 + 827 + 8069

210

Compared to the river water concentration, the channel sediment porewater is enriched by a factor of 200 for iron and manganese, 30-50 for nickel and lead, approx. 10 for cadmium and mercury, and 2-3 for copper and zinc. When the expected concentration of metals following hydraulic dredging, which were calculated from a rate of porewater to river water of about 1:4, were compared with the actual measurements at the pipe exiting the dredging device, negative deviations were found for iron and manganese, suggesting reprecipitation of Fe/Mn-oxide minerals; the positive deviations of zinc (factor SO), copper, lead and cadmium (factors 7-8)indicate that during dumping of the sludge-water mixture significant proportions of these elements were mobilized and transferred into the effluent water. Pore water data from dredged material from Hamburg Harbor indicate typical differences in the kinetics of proton release from organic and sulphidic sources (Table 3). Recent deposits are characterized by low concentrations of nitrate, cadmium and zinc; when these low-buffered sediments are oxidized during a time period of a few months to years, the concentrations of ammonia and iron in the pore water typically decrease, whereas those of cadmium and zinc increase (with the result that these metals are easily transferred into agricultural cropsl). Table 3: Mobilization of Metals and Nitrogen Compounds from Dredged Material after Land Deposition ([38] and other authors) Element or Comoound

I

Reduced Water

I

Oxidized Water

Ammonia

I

125mg/L

I

<3mg/L

Cd

B

......

E, low pH higl:

I

D

C .......................................

.....

\

oxidation lowering

Months Days Permanent Years

.......................................

oxidation stable 10 - 20 Years

/ oxidation lowering

Time ?

Figure 8: Schematic Diagram Illustrating Different Phases of Metal Release from Land-Disposed Dredged Material [39]

271

The different steps are schematically given in Figure 8. Oxidation of sulphides during stage B strongly increases the concentrations of cadmium and zinc in a relative short time. When acidity is consumed by buffer reactions (phase C), cadmium and zinc concentrations drop, but are still higher than in the original sulphidic system. In phase D, oxidation of organic matter again lowers pH-values and can induce a long-term mobilization of Zn and Cd. 2.6 Metal Release from Municipal Solid Waste Landfills

In municipal solid waste landfills, initial conditions are characterized by the presence of oxygen and pH-values between 7 and 8. During the subsequent "acidic anaerobic phase", the pH drops to a level as low as 5 because of the formation of organic acids in an increasingly reducing milieu; concentrations of organic substances in the leachate are high (Figure 9). In a transition time of one to two years, the chemistry of landfill changes from acetic to methanogenic conditions. Typically increased concentrations of metals have been found for iron, manganese and zinc in leachates during the acidic decomposition phase compared with the methanogenic phase.

There is not much experience with landfill evolution subsequent to the initial 30 years. What could happen is that the landfill is again oxidized. It has been inferred that oxidation of sulphidic minerals by intruding rainwater may mobilize trace metals. The impact on the underlying groundwater could be even higher if a chromatographic process, involving continuous dissoluion and reprecipitation during passage of oxidized water through the deposit, were to preconcentrate critical elements prior to final release with the leachate.

-7-'5 -6

5-

-

pH?

-S *--

Gas

solids

-- -

Oxygen

Oxygen?

Degradation of organic matter

Months

Months or years

ca.30- 100 years?

Figure 9: Scheme of Chemical Evolution of Municipal Solid Waste Landfills (Forstner et al. [401)

Comparison of inorganic groundwater constituents upstream and downstream of 33 waste disposal sites in Germany [41] indicates typical differences in pollutant mobilities, which may partly be related to releases during the acidic phase of the landfill development. High contamination factors ("contaminated mean"/"uncontaminated mean"; see Figure 10) have been found for boron, ammonia, and arsenic; heavy metals such as cadmium, chromium, lead and copper are significantly enriched in the leachates as well.

212

ip List of inorganic parameters

Figure 10: Influence of Waste Disposal on Ground Water Quality from 33 Sites in Germany (Ameth et al. [41])

I

Cora

CI

Zn

500

600

[m&l Mean leachate concentration after 50 years landfill operation

600

I

Cd [PLgnI 2

1

Hg

0.1

Mean concentration in the uncoptaminated groundwater ~~

Mean annual increase of concentration in groundwater

Mean annual increase in %

0.24 50%

0.2

7%

0.24

5%

0.008 4%

0.00016

3%

273

Mean Content

Parameter

In 1:lO - Suspension ya~ p~

Acid Producing Potential (APP)

1

~~

Organic Carbon

3.6%

Organic Nitrogen

0.27%

0.19 mM/g

0.019

1.71

Organic Sulphur

0.05%

0.03 M

g

0.003

2.82

FeSz

1Yo

0.33 M

g

0.033

1.50

Total

I

1

0.55Mg

I

0.055

I

1.28

The acid-producing potential not only is related to the oxidation of sulphides, but oxidation of organic matter must be considered as well. Table 5 indicates that the contribution of protons from organic-N and organic-S in a sample containing approximately 5% organic carbon is equivalent to the acid-producing potential of 1% FeS,. Studies of the long-term evolution and diagenesis in sewage sludge landfills and similar natural sediments (peat, organic soils) by Lichtensteiger et al. [45] suggest that the transformation of organic material will last for geological time scales (10’ to 10’ years).

3. Geochemical Concepts for Metallic Pollutants in Solid Wastes

Three geochemical concepts demonstrate the advantage of a long-term strategy compared to the traditional approaches in waste management: (i) the mobility concept, (ii) the concept of capacity controlling properties, and (iii) the concept of final storage quality, the first being relevant for process studies, the second for effect evaluations and the third for problem solutions.

3.1 Mobility Concept

Among the criteria to assess which element or elemental species, beside its toxic potential, may be of major concern in ecological evaluations, one question deserved primary attention [46]: “Is the element mobile in geochemical processes, mainly because of its volatility or solubility in water, so that the effect of geochemical perbations can propagate through the environment?”

As mentioned above, typically for systems involving solutiodsolid interactions, mobility is given by accelerating and retarding factors and processes. The composition of interstitial water is the most sensitive indicator of the types and the extent of reactions that take place between chemicals on solid substrates and the aqueous phase which contacts them. Particularly for finegrained material the large surface area related to the small volume of its entrapped interstitial ill be indicated by major changes on water ensures that minor reactions with the solid phase w the composition of the aqueous phase. With respect to the behaviour of toxic metals it has been stressed by Salomons [47] that from an impact point of view it is import to know whether the concentrations in the pore waters are determined by adsorptioddesorption processes or by precipitatioddissolution processes. If the latter is the case the concentrations in the pore waters of pollutants should be widely independent from the concentrations in the solid phase.

274

Including new experience from impact evaluations related to capacity controlling properties, the mobility concept of environmental geochemistry can be implemented into waste management practice by different ways of optimizing barrier systems. As shown from the examples of largemass wastes dredged material, mining residues and municipal solid waste, long-term immobilization of critical pollutants can be achieved by promoting less soluble chemical phases, i.e., by thermal and chemical treatment, or by providing respective milieu conditions. In general, microscale methods, e.g., formation of mineral precipitates in the pore space of a waste body, will be employed rather than using large-scale enclosure systems such as clay covers or wall constructions [48]. A common feature of geochemically-designed deposits, therefore, is their tendency to increase overall stability in time, due to the formation of more stable minerals and closure of pores, thereby reducing water permeation.

-

3.2 Storage Capacity Controlling Properties

This conceptual approach has been developed in the framework of the concept of "chemical time bombs". To make the scientific objectives clearer, it is useful to distinguish between two different mechanisms [49]: The first is direct saturation, by which the capacity of a soil or sediment for toxic chemicals becomes exhausted. The second way to "trigger" a time bomb is through a fundamental change in a chemical property of the substrate that reduces its capacity to adsorb (or keep adsorbed) toxic materials. Within a scientific perspective of the chemical time bomb concept the aspect of the storage capacity controlling properties (CCP's) of solid substrates will play a key role. In this respect, the potential of sediments, soils and waste materials to immobilize toxic chemicals is conceived as a first and in most cases preferential barrier against dispersion of these substances in both ground and surface waters or their transfer to terrestrial or aquatic biota. It is common for the three subdisciplines (sediment/soiVwaste) that priority questions at this stage of effect assessment are directed to the long-term behaviour of critical components under certain borderline conditions. Development of methodologies for solving these questions should be designed for assessing effects related to processes of "early diagenesis" [50],i.e., mechanisms and effects by which solids are changed in their chemical form, involving new equilibrium between solid and their dissolved species. Such evaluations include the type of dissolvedsolid interactions, transfer rates of contaminants between various substrates, degradation of toxic organic chemicals (incl. the fate of intermediate products) and in particular processes in interstitial waters (see above). 3.3 Final Storage Quality

The final storage approach is one way to develop and control landfills on a conceptual basis. It has been defined by the Swiss Federal Government in 1986 [51] and received wider attention by the book edited by Peter Baccini on Landfills - Reactor and Final Storage [52]: "Landfills with solids of final storage quality need no further treatment of emissions into air and water". Final storage properties can be achieved by using with typical geochemical engineering techniques such as: +

selection of favourable milieu conditions for the deposition of large-volume wastes such as dredged materials, selection of additives for the solidification and stabilization of hazardous waste materials, and

+

optimization of elemental distribution at high-temperature processes, e.g. incineration of solid waste materials.

215

I

.

1. Mobility Concepf Dispersion Velocity is lnbanced by Dissolution, Desorption, Complexation, Biometbyhtion efc., and by Elevded Permeability for Solutions 0 Elutriate Tests, Adsorptionsisotherms, Diffusion Tests

2. Concept o f Capacity Controlling Properties (CCPs! "Chemical fims Bomb" [Elon- linear, Delayed Processes] Capacify far Pollutant Upfake is Exbaustm'by DirJcf Safurafion or by Reduction o f Sorption-Active Substrates by External Influences

0 Determination o f Buffer Capacity and Acid Producing Potential * * * 3. Long-Term Stability [ II Final Storage Quality"-

Delete Reactive and Reaction-Mediating Components b Provide Long-Term Buffer Capacity or Neutralization Potential b Reduce Permeability b Favour Precipitation Above Sorption Processes b

Salomons [47]

>

I

'Solids b

Treatment with the Objective of Re-Utilization

216 4. Prognostic Tools for the Behaviour of Metals in Wastes

4.1 Acid Producing PotentiaVNeutralking Capacity

Long-term release of protons can be expected even from the relatively low organic carbon ashes and slags from municipal waste incineration, as has been suggested by Krebs et al. [53]. Microbial degradation of 1 - 2 % of residual organic carbon will produce approximately 1 mol R per kg bottom ash, which is about equivalent to the acid neutralizing capacity of this material (but several orders of magnitude higher than the P-input from acid precipitation). The acidneutralizing capacity is defined operationally by the amount of acid required t o reach pH 7; Figure 11 presents titration curves of typica1 bottom ashes from municipal solid waste incineration [54].

'.

' 4

0

0.5

I

1.5

2

Acid addition (mol

2.5

H -t kg

-'

3

3.5

4

bottom ash)

Figure 1 1 : Titration of Bottom Ash Samples. Full Signature: Extraction Time 48 h; Open Signature: 197 h [54] Laboratory experiments by Belevi et al. [54] suggest that non-metal fluxes by leachate (such as chloride, sulphur and DOC fluxes) would adversely impact the environment for years to decades after disposal. Heavy metal fluxes by leachate are expected to be compatible with the environment; however, additional laboratory and field studies are necessary to assess their behavior over longer time periods. At present, bottom ash cannot be considered as a material of final storage quality. It should either be disposed into monofills with leachate collection and treatment systems or be treated prior to disposal to achieve final storage quality 1541. 4.2 Experimental Design for Long-Term Metal Release

With an experimental approach, which was first used by Patrick et al. [55] and H e m s and Briimmer [56], metal mobility of industrial waste materials has been studied in a circulation apparatus by the controfled intensification of significant release parameters such as pH-value, redox-potential and temperature. Here, an ion-exchanger system is used for extracting and analyzing the released metals at an adequate frequency (Figure 12).

277

Ion-Exchanger

- (159 Chelex 100)

'O 20

1s 10

5

Control

Recorder

and Recorder

Unlt

Sample Mixed

Unlt

Qua Sand-

0

14

with Quartz Sand

Supporttng Surface Filter

Figure 12 (Left): Experimental Design for Long-Term Prognosis of Metal Release [57]. Figure 13 (Right): Release of Arsenic and Zinc &om Industrial Solid Waste in a Long-Term Circulation Experiment [58]

In a series of experiments, a number of industrial waste materials of different types, intended for co-disposal in borrow pits, were investigated with this method. In these experiments, special attention was given to the efficiency of individual components with respect to long-term behavior of critical trace elements in such mixed deposits, The kinetics of element release from "conditioned" waste material (i.e., treated with high-pH additives) is shown in Figure 13 for As, and Zn. By treatment with pH 5 solutions, mobilization of As is essentially completed after the initial five weeks of the experiment. Mobilization of Zn is strongly enhanced toward the end of the study period. Regarding the latter element, it can be expected that the cumulative percentage of release from the treated material would significantly be enhanced upon continuation of the experiment.

4.3 Chemical Characterization of Dredged Material

With regard to the selection of disposal options for dredged materials, the study of the water phase only with not be fully satisfling. While most of the actual situation would be reflected in these data, the potentialities of future adverse effects as well as the possible measures for reducing such hazards cannot be predicted. In this respect, extractability of pollutants with chemical agents of different strength will provide more reliable information of the potential release of these substances under typical environmental conditions. Elutriate Test. To estimate short-term chemical transformations, the interrelations between solid phases and water has been increasingly subjected to laboratory experimentation. The advantage of such experiments is that especially important parameters can be directly observed and particularly unfavourable conditions simulated. The Army Corps of Engineers and the US Environmental Protection Agency have developed an elutriate test that is designed to detect any significant release of chemical contaminants in dredged material. This test involves the mixing

278

of one volume of the dredged sediment with four volumes of the disposal site water for a 30min shaking period. If the soluble chemical constituent in the water exceeds 1.5 times the ambient concentration in the disposal site water, special conditions will govern the disposal of the dredged material [59]. Sequential Extraction. In connection with the problems arising from the disposal of solid wastes, particularly of dredged materials, chemical extraction sequences have been applied which are designed to differentiate between the exchangeable, carbonatic, reducible (hydrous Fe/Mn oxides), oxidizable (sulphides and organic phases) and residual fractions. The undisputed advantage of this approach with respect to the estimation of long-term effects on metal mobilities lies in the fact, that rearrangements of specific solid "phases" can be evaluated prior to the actual remobilisation of certain proportions of an element into the dissolved phase [60]. One of the widely applied extraction sequences of Tessier and co-workers [61J has been modified by various authors (Table 7).

Table 7 Sequential Extraction Scheme for Partitioning Sediment Samples [62, 631

I Extracted Component

I Extractant

Fraction ~

Exchangeable

1 MNH,OAc

Exchangeable Ions

Carbonatic

1 M NaOAc, pH 5 w/ HOAc

Carbonates

Easily Reducible

0.01 M W O H HCI w/ 0.01 M HNO,

Mn-Oxides

Moderately Reducible

0.1 M Oxalate Buffer pH 3

Amorphous Fe-Oxides

Sulphidic/Organic

30% KO,pH 2 w/ 0.02 MHNO, extr. w/ 1MNH,OAc - 6% HNO,

Sulfphides together with Organic Matter

I Hot HNO, conc.

Residual

I00

I LithogenicMaterial

[%I

90 80

70 60

I . 1exchangeable

50

0 carbonatic

40

30 20 10

0

A

B

C

D

-

easily reducible moderately reducible

sulfidJorg. residual

Figure 14: Partitioning of Cadmium in Anoxic Mud from Hamburg Harbour in Relation to the Pretreatment Procedure (Kersten et al. [64])

279

Partition of metals was determined in a sample from Hamburg harbor, which was pretreated in different ways [57]: (1) The EPA Standard Elutriate Test, 1.4 sedimentkite water for 30 min (see above); (2) freeze-dried sample; (3) oven-drying at 60°C. It was observed that oxidation had a great effect on regulating the chemical form of cadmium and other trace metals (Figure 14). Compared to the original sample (A), which was extracted under an argon atmosphere, there was a typical change from oxidizable phases, which were mainly Cd-sulphide, to easily reducible forms upon application of the shakinglbubbljng test (B). During freeze-drying - which is commonly assumed to represent a relative smooth mode of sample pretreatment, transformation to carbonatic and exchangeable forms takes place (C). This effect is krther enhanced during oven-drying at 60°C @).

In Table 8, an example is given for the possibilities of standardizing the data from elution experiments with respect to numerical evaluation. An “elution index” for sediment samples from various rivers in West Germany is based on the metal concentrations exchangeable with 1 N ammonium acetate at pH 7. These metal fractions are considered to be remobilizable from polluted sediments at a relative short term under more saline conditions, for example, in the estuarine mixing zone. Comparison of the release rates from oxic and anoxic sediments clearly indicates, that the oxidation of samples gives rise to a very sigdicant increase in the mobilization of the metals studied. This effect was particularly important for Cd. When proceeding krther in the extraction sequence, more long-term effects can be estimated, but generally with a reduction of prognostic accuracy (Table 8). Table 8: Elution-Index for Selected River Sediments, as Determined from Exchangeable Proportions (1 M Ammonium-Acetate). Calculated Relative to Background Data from Elbe River Sediments [ 2 6 ] . These Values are Multiplied by a Factor of 100

I Copper

NeckarR

I

Main R

I

0.2

RhineR

I

ElbeR

1

1

1

WeserR

Lead

1

1

2

1

1

Zinc

7

9

28

36

9

Cadmium

30

30

230

30

TotalOxic (Anoxic)

I

38 0.5

I

40

0.3

I

261

8

I

68 >4

I

10 4

4.4 Metal Transfer between Inorganic and Organic Substrates

With respect to the modeling metal partitioning between dissolved and particulate phases in a natural system, e.g, for estuarine sediments, the following reqirements have been listed by Luorna and Davis [65]: 6

6

the determination of binding intensities and capacities for important sediment components the determination of relative abundance of these components; the assessment of the effect of particle coatings and of multi-component aggregation on binding capacity of each substrate, the consideration of the effect of major competitors (Ca2’, Mg”, Na’, C1-), the evaluation of kinetics of metal redistribution among sediment components.

280

It seems that thermodynamic models are still restricted because of various reasons: (i) adsorption characteristics are related not only to the system conditions (i.e., solid types, concentrations and adsorbing species), but also to changes in the net system surface properties resulting from particldparticle interactions such as coagulation; (i) influences of organic ligands in the aqueous phase can rarely be predicted as yet; (iii) effects of competition between various sorption sites, and (iv) reaction kinetics of the individual constituents cannot be evaluated in a mixture of sedimentary components. At present, experimental studies on the dissolvedsolid interactions in such complex systems seem to be more promising. One approach uses a six-chambered device, where the individual components are separated by membranes, which still permit phase interactions via solute transport of the elements [66]; in this way, exchange reactions and biological uptake can be studied for individual phases under the influence of pH, redox, ionic strength, solid and solute concentration, and other parameters. The system is made of a central chamber connected with 6 external chambers and separated by membranes of 0.45 pm pore diameter (Figure 1Sa). The volume of the central chamber is litres and each of the external chambers contains 250 ml. Either solution or suspension can be inserted into the central chamber. In each external chamber the single solid components are kept in suspension by magnetic stirring. Redox, pH and other parameters may be controlled and adjusted in each chamber. In an experimental series on the effect of salinity (i.e., disposal of anoxic dredged mud into sea water) quantities of model components were chosen in analogy to an average sediment composition: 0.5 g algal cell walls (=5%), 3 g bentonite (=30%), 0.2 g manganese oxide (=2%), 0.5 g goethite (=5%), and 5 g quartz powder (=SO%). In the central chamber, 100 g of anoxic mud from Hamburg harbour was inserted; salts were added corresponding to the composition of sea water. After 3 weeks solid samples and filtered water samples were collected from each chamber and analysed.

'"

pg g-1

In 100s Sludge: 32200 ug

J.

Treatment wirh Seawater

.1

MobiiizPdCopper: 417 pg

ibrane '

(Dissolved: 156 pg

n

Algal

Cell Walls

Figure 15: Metal Transfer Between Sedimentary Components. 15a: Schematic View of the Multichamber Device. 15b: Transfer of Copper from Anoxic Harbor Mud into Different Model Substrates after Treatment with Artificial Seawater [66]

28 1

The effect of salinity on metal remobiliation from contaminated sediments is different for the individual elements. While approximately 16% and 9% of cadmium and zinc, respectively, in the dredged mud from Hamburg harbour is released, for metals such as copper the factor salinity increase seems to be less important in the transfer both among sediment substrates and to aquatic biota. This is, however, untrue as can be demonstrated from a mass balance for the element copper in Figure 1%: It is indicated that only 1.3 % of the inventory of Cu of the sludge sample is released when treating with seawater. Only one third stays in solution, equivalent to approx. 40 pgl-', and there is no sigmficant difference to the conditions before salt addition. Two thirds of the released copper is readsorbed at different aflinities to the model substrates. Slight enrichment of copper is observed in the iron hydroxide (approx. 80 ppm) and manganese oxide (1 00 ppm), whereas the cell walls - a minor component in the model sediment - has accumulated nearly 300 ppm of copper. The dominant role of organic substrates in the binding of metals such as Cd and Cu is of particular relevance for the transfer of these elements into biological systems. It can be expected that even at relatively small percentages of organic substrates these materials are primarily involved in metabolic processes and thus may constitute the major carriers by which metals are transferred within the food chain.

4.5 Reaction Cell Experiments for Discrimination of Mobilization and Scavenging Processes of Metals at Sediment Resuspension

Resuspension of sediments from the Elbe River can significantly decrease pH values due to high acid producing potential and low neutralizing capacity. Metals can be released into the dissolved phase, but may subsequently be readsorbed or precipitated in part to solid phases. To evaluate a regional and long-term perspective experiments were undertaken at stable neutral pH-values, and both total release and scavenging rates were extrapolated from time series of measured net release values over 630 hours [67]. In laboratory simulation experiments, wet sediment material was transferred to a 180 ml reaction cell; 135 ml of artificial river water (2 mM CaCI, and 5 mh4 NaCI) was added to give a solid/liquid ratio of 1:64. The suspension was sampled at regular intervals and then centrifbged and filtered through a 0.2 pm membrane. The same volume of artificial river water used to wash the centrihgation tube was added into the reaction cell in order to keep the suspension volume constant. The experiments were carried out under exclusion of light. The example in Figure 16 shows that the initial lead release was high but then decreased continuously due to readsorption. After about 100 hours the relative equilibrium between release and readsorption was reached. Curves for copper indicate continuous release until about 350 hour, when scavenging rate became faster than the release rate. With respect to cadmium and zinc, release was found to be higher than the scavenging rate during the whole experiment. The order of total release from the sediment has been found to be Cd (5 %) > Zn (1.5 %) > Cu (1%) > Pb (0.3 YO). The percentage scavenging of released metals was in the order of Pb (85%) > Cu (53 YO)> Zn (35 %) > Cd (30 %). Dominant processes are adsorption on organic substances, adsorptiodcoprecipitation by fresh Fe-Mn oxides and precipitation of metal phosphates originating from the decomposition of organic matter. Based on experimental results and relevant literature data, a four stage interaction model has been developed for metals in anoxic sediments to describe the behaviour of heavy metals in the system subsequent to oxidation.

282

0.0

200.0

100.0

300.0

460.0

500.0

600.0

700.0

Figure 16 Lead Release from Elbe Sediments. Full Quadrangles: Net Release; Full Circles: Total Release; Open Circles: Scavenging (AdsorptiodCoprecipitation) The time at which the highest concentration (peak value) of metal release is found can be called "characteristic value" of metal release and represented by an index Omeghe in hours. When the oxidation time is longer than this value, the dominating process is scavenging or relative equilibrium; when the oxidation time is less than O m e g h , the dominating process is release of metals (Figure 17). It has been suggested that the latter situation in particular will pose long-term problems to the aquatic ecosystem.

I

.b',

Cu

i

t

! QZ*

@Cd

Time Figure 17: Schematic Diagram Showing Changes in Relative Concentrations of Metals During ResuspensiodOxidation [67]

283 5. Remediation Procedures

"Geochemical and biological engineering" emphasise the increasing efforts to use natural resources available at the disposal site for reducing the negative environmental effects of all types of waste material, in particular ofacid mine wastes [3, 68, 691. 5.1 Storage Under Permanent Anoxic Conditions

Regarding the various containment strategies is has been argued that upland containment (e.g., on heap-like deposits) could provide a more controlled management than containment in the marine environment. However, contaminants released either gradually from an imperfect impermeable barrier (also to groundwater) or catastrophically from failure of the barrier could produce substantial damage [70]. On the other hand, near-shore marine containment (e.g., in capped mound deposits, offers several advantages, particularly with respect to the protection of groundwater resources, since the underlying water is saline and inherent chemical processes are favourable for the irnmobilisation or degradation of priority pollutants. In a review of various marine disposal options, Kester et al. [70] suggested that the best strategy for disposing of contaminated sediments is to isolate them in a permanently reducing environment. Disposal in capped mound deposits above the prevailing sea-floor, disposal in subaqueous depressions, and capping deposits in depressions provide procedures for contaminated sediment [71]. In some instances it may be worthwhile to excavate a depression for the disposal site of contaminated sediment which can be capped with clean sediment. This type of waste deposition under stable anoxic conditions, where large masses of polluted materials are covered with inert sediment became known as "subsediment-deposit". The first example was planned for highly contaminated sludges from Stamford Harbour in the Central Long Island Sound following intensive discussions in the U.S. Congress 1721.

Figure 18 Solubility of Metal Sulphides and Metal Oxides (from Ehrenfeld and Bass [76]

284

Under sub-sediment conditions there is a particular low solubility of metal sulphides, compared to the respective carbonate, phosphate, and oxide compounds (Figure 18). One major prerequisite is the microbial reduction of sulphate. Thus, this process is particularly important in the marine environment, whereas in anoxic freshwaters milieu there is a tendency for enhancing metal mobility due to the formation of stable complexes with ligands from decomposing organic matter. Marine sulphidic conditions, in addition, seem to repress the formation of mono-methyl mercury, one of the most toxic substances in the aquatic environment, by a process of disproportionation into volatile dimethyl mercury and insoluble mercury sulphide [73].There are indications that degradation of highly toxic chlorinated hydrocarbons is enhanced in the sulphidic environment relative to oxic conditions [74,751. 5.2 Chemical Stabilisation by Additives

In general, solidificatiodstabilization technology is considered a last approach to the management of hazardous wastes. The aim of these techniques is a stronger fixation of contaminants to reduce the emission rate to the biosphere and to retard exchange processes. Most of the stabilization techniques aimed for the immobilization of metal-containing wastes are based on additions of cement, water glass (alkali silicate), coal fly ash, lime or gypsum [77-791. Laboratory studies on the evaluation and efficiency of stabilization processes were performed by Calmano et al. [80].As an example Figure 19 shows acid titration curves for Hamburg harbour mud without and after addition of limestone and cement/fly ash stabilizers. Best results are attained with calcium carbonate, since the pH-conditions are not changed significantly upon addition of CaCO,. Generally, maintainance of a pH of neutrality or slightly beyond favours adsorption or precipitation of soluble metals [81]. On the other hand it can be expected that both low and high pH-values will have unfavourable effects on the mobility of heavy metals.

\

-,

'.,

\

+ 10%cement

+IO%flyash

'\

+ 20% \

caicium carbonate

\ original sample 2

4

6

11

to

1

It

mmol [H+] / g solids Figure 19: Effect of Calcium Carbonate and CementLFly Ash Additives on Chemical Stabilization of Fine-Grained Sediment from Hamburg harbour (Calmano et al. [SO]) Experimental studies of the processes taking place with mixed residues fiom lignite coal incineration indicate favorable effects of incorporation of both chloride and heavy metals in newly formed minerals (Figure 20).

285

Figure 20: Schematic Representation of Mneral Formations in "Stabilisates" from Coal Residues [82] Ettringite (i.e., calcium-aluminate-monosulphate-hydratephases) in particular, can act as "storage minerals" for chloride and metal ions. The former may be incorporated at up to 4 kg CaC4 per m3of the mineral mixture. Calcium-silicate-hydrate phases may be formed in a subsequent process, and by filling hrther pore space these minerals can significantly reduce permeability of the waste body for percolating solutions. Experimental studies of the leachability of salts and trace elements from samples of "stabilisates", with a pressure-filtration method, indicate relative high rates of release for sulphate ions, but not for zinc and cadmium in the eluate [82]. An overview on various fields of environmental research and management to which mineralogical methods can be successfdly applied has recently been given by Bambauer [83].

5.3 Final Storage Quality of Municipal Solid Waste Materials

From a geochemical point of view the "reactor landfill" is characterized by labile conditions during the intial aerobic and acid anaerobic phases, the former mainly due to uncontrolled interactions with organic solutes (see above). Particular problems occur when leachate collection pipes are plugging during the acidic decomposition period. Sludge-only Iandfills have even lower permeabilities and interparticular porosities than municipal solid waste landfills.

On the other hand, solid residues with final storage quality should have properties very similar to the earth crust (natural sediments, rocks, ores, soil; Table 9). This can be achieved in several ways, e.g., by assortment or thermal, chemical and biological treatment. In most cases this standard is not attained by simple incineration of municipal waste, i.e., by reduction of organic fractions only. There is, in particular, the problem of easily soluble minerals such as sodium chloride.

286

Table 9: Comparison of Inventories of Chemical Components in the Two Landfill Alternatives and in the Earth Crust [84] Reaktor Landfill

I Final Storage

I Earth's Crust

Major solid constituents Solid "inert" waste

Silicates, oxides

Quartz, Fe-oxide, clay, carbonates

Putrefactive waste

[Gypsum, NaCI]'

(Gypsum, NaCI)

Grease trap waste

(Char)b

Keragenic compounds

Organic micropollutants

Organic micropollutants

-

Metals in reactive chemical forms

Metal-bearing minerals, mainly oxides

Metals mainly in inert forms

Protons, electrons

(Protons)

@H,acid rain)

Organic compounds

(Organic residues)

(Humic acids)

Dissolved salts

pissolved salts]'

(Dissolved salts)

Measures before incineration include the separate collection of (organic) kitchen and garden wastes (containing chlorine and sulhr), which can be transferred into compost; a major decrease of chlorine content, however, would require a significant reduction of PVC in municipal solid waste. After incineration washing of the residues can be performed either with neutral or acidified water. Future efforts should be aimed for optimizing the incineration process in a sense, that: critical components are concentrated in the filter ash and in the washing sludge, whereas the quality of the bottom ash is improved - by application of high-temperature procedures [85] in a way, that deposition is facilitated and even a reuse of this material is possible due to either the low concentrations or chemically inert bonding forms of metals [86]. Recent advances mainly concern the application of high temperature techniques for re-melting precipitator ashes and flue gas purification product. An example is the RedMelt-process, designed by Faulstich and colleagues [87], which - in a reductive milieu - afiects differentiation into three phases (Table 10): (1) metal-rich (Fe, Cu) bottom phase; (2) "condensate" phase with enrichment of zinc, lead and cadmium in the flue gas (from where metals can be recovered and transferred to a base metal smelter); and (3) relatively metal-poor silicate phase, which can easily be used for construction purposes, Table I 1 summarized available leaching data on different materials and trePtments municipal solid waste incineration products. All values relate to the pH-4 conditions as prescribed by the Swiss Guidelines in the Technical Advise on Solid Waste Materials (1991). While untreated and even washed electrostatic precipitator dusts do not meet the requirements given by the TVA-supplement, cement-solidified ashes are near the limit. Satisfylng results can only reached by using high-temperature processes, and best results are shown by the R e m e l t procedure. With regard to energy consumption, initial data suggest, that the average consumption is in the range of 1-2 kWh/kg residue, and this is nearly equivalent to the energy content of the original municipal solid waste material.

287

Element

Content (Mass-%) Input

Products Silicate

Metal

Condensate

3,4 02 0,6

64 03 03 1,4 0,3 0,03

Silicium Aluminium Calcium

22,o 9,o

26,2 6,8 10,6

Iron Copper Chromium

10,l 0,3 0,04

4s 0,03 0,04

85,O

Zinc Lead Cadmium

0,6

0,09 <0,01 <0,00001

0,08

Carbon

53

02 0,004

I

2

I

4.4 02


I

3E-6

14,s 63 0,1

I

0,3

-

______

Chlorine

43

0.3

0,03

23,s

(Total) Oxygen Total

59,4 40,6

56,4 43,6

98,7 1,3

81,3 18.7

100

100

100

100

~~

~

Lead [PPml

Cadmium [PPml

Zinc [PPml

Copper [PPml

Untreated Electro Precipitator Ash - MSWI Oberhausen

4,2

43

133

1,7

Washed Electro Precipitator Ash + Wet Washing Residues

0,77

0,94

57

0,26

1

0,1

5

0,s

Cement-Solidified Precip.Ash + WWR (g), Addit./Ash =1:2

0,14

0,08

5,l

0,05

Electro Precipitator Ash Treated by 3R-Process

0,1

0,02

0,7

0,08

Vitreous Residue from ABB(Asea-Brown-Boveri) Process

<0,040

<0,030

0.11

<0,040

Silicate Phase from RemeltProcess (Faulstich et al.,)

0,004

<0,001

<0,020

0,005

TVA-(Swiss)-Guidelines Leaching Values (pH = 4)

288

Outlook The enormous problems arising from the historic thoughtless dissipation of chemicals in the environment has become obvious in the last ten years, and especially with the opening of eastern Europe. The detection of extremely toxic chemicals at Love Canal, LeMrerkerk and Georgswerder are only a few of the harsh lesson. After solving a few spectacular cases with intensive efforts, it now seems that financial restrictions may inhibit the use of many newly developed remediation techniques (such as chemical extraction, high-temperature incineration, and some biological procedures) to sanitize contaminated land on a large scale (i.e., for areas of hundreds of square kilometers. Although this relates primarily to parts of Eastern Europe, the tens of thousands of old landfill sites which must be excavated, treated and recultivated will soon strain even more prosperous economies. It may well be, that economic considerations will strengthen the popularity of geochemically engineered solutions in these situations, as well as for large waste masses such as mine residues, dredged materials and filter ashes described here. But even if traditional engineering continues to dominate practical solutions the specific potential of geochemistry to provide instruments for long-term assessment of processes should be used and expanded.

Acknowledgements: My review was compiled on the occasion of the 1992 R.A. Vollenweider Lectureship in Aquatic Sciences, held at Burlington, Ontario, October 7, 1992. I wish to record my particular gratitude to my collaborators at the Section of Environmental Protection Technology, Technical University of Hamburg-Harburg. which have provided significant portions of the material for this review. With regard to the present subject, I received much inspiration by Dr. Wim Salomons during the last ten years and more recently by both Dr. Gemt Hekstra and Dr. William Stigliani as well as from numerous colleagues of the Chemical Time Bomb Project. I am particularly grateful to the National Water Research Institute at Burlington, especially to Dr. Rod Allan, who encouraged me to the present theme and to all, who were engaged in the preparation of this lectureship in honour to Professor Richard A. Vollenweider. References

[I] Stigliani WM (1991) Chemical Time Bombs: Definition, Concepts, and Examples. Executive Report 16 (CTB Basic Document). U S A Laxemburg/Austria, 23 p [2] Fdrstner U (1992) Geochemical concepts in solid waste management. In: J-P Vernet (ed) Proc 5th Int Conf Environmental Contamination, MorgedSwitzerland, 29 Sept - 3 Oct 1992. CEP Consultants Edinburgh (in press) [3] Salomons W, Fdrstner U (Eds)(1988b) Environmental Management of Solid Waste: Dredged Materials and Mine Tailings. Springer-Verlag, Berlin, 396 p [4] Schuiling RD (1990) Geochemical engineering - some thoughts on a new research field. Appl Geochem 5 : 25 1-262 [S] Stief K (1987) Zukiinflige Anfordemngen an die Deponietechnik und Konsequenzen fir die Sickerwasserbehandlung.In: Deponiesickerwasserbehandlung.UBA Materialien 1/87, pp 27-36. Erich Schmidt Verlag Berlin [6] Forstner U, Ahlf W, Calmano W, Kersten M, Schoer J (1990) Assessment of metal mobility in sludges and solid wastes. In: JAC Broekaert, S Giicer, F Adams (eds) Metal Speciation in the Environment. NATO AS1 Series vol G 23, Springer-VerlagBerlin Heidelberg, pp 1-41 [7] Drever .TI (1982) The Geochemistry of Natural Waters. Prentice-Hall, Englewood Cliffs N.Y.

289

[8] Breemen N van, Driscoll CT, Mulder J (1984) Acidic deposition and internal proton sources in acidification of soifs and water. Nature 307: 599-604 [9] Moore JN, Luoma SN (1990) Hazardous wastes from large-scale metal extraction. Environ Sci Techno124: 1278-1285 [lo] Forstner U (1981) Trace metals in fresh waters (with particular reference to mine effluents). In: KH Wolf (ed) Handbook of Strata-Bound and Stratiform Ore Deposits. Vol 9: 271-303 [ 1 1 1 Anonymous (1979) Suggested guidelines for method of operation in surface mining of

areas of potentially acid-producing materials. West Virginia Surface Mine Drainage Task Force. WV Dept Nat Resour. Charleston, WV [ 121 Ferguson KD, Erickson PM (1988) Pre-mine prediction of acid mine drainage. In: W Salomons, U FiSrstner (eds) Environmental Management of Solid Waste - Dredged Material and Mine Tailings. Springer-Verlag Berlin, pp 24-43 [ 131 Kersten M et al. (1985) Freisetzung von Metallen bei der Oxidation von Schltlmmen. Vom Wasser 65: 21-35. [ 141 Kersten M, Forstner U (1991) Geochemical characterization of the potential trace metal mobility in cherive sediment. Geo-Marine Letts 1 1 : 184-187 [ 151 Hong J, Forstner U, Calmano W (1993) Effects of redox processes on the acid-producing potential and metal mobility in sediments. In: J Hamelink et al (eds) A Mechanistic Understanding of Bioavailability, Proc. SEATC-Workshop held at Pellston/MI, August 17-22, 1992 (discussion initiation paper submitted) [I61 Breemen, N. van (1987), Effects of redox processes on soil acidity. Neth J Agric Sci 35: 27 1-279 [17] Akittrick JA, Fanning DS, Hossner LR (eds)(1982) Acid sulfate weathering, SSSA Special Publ 10, Soil Sci SOCAmer Madison, WI,p 234 [IS] Breemen N van, Mulder J, Driscoll CJ (1983) Acidification and alkalinization of soils. Plant Soil 1 5 : 283-308 [ 191 Breemen N van (1975) Acidification and deacidification of coastal plain soils as a result of periodic flooding. Soil Sci SOCAmer Proc 39: 1153-1 157 [20] Brinkman R (1970), Ferrolysis, a hydromorphic soil forming process. Geoderma 3: 199206 [21] Brinkman R (1979), Ferrolysis, a soil-fonning process in hydromorphic conditions. Agricultural Research Reports 887. PURDOC, Wageningen. [22] Eaqub M, Blume HP (1982) Genesis of a so-called ferrolysed soil of Bangladesh. Z Pflanzenerniihr Bodenkde. 145: 470-482 [23] Breemen N van (1988) Effects of seasonal redox processes involving iron on the chemistry

of periodically reduced soils. In: JW Stucki, BA Goodman, U. Schwertmann (eds) Iron in SoiIs and Clay Minerals. D Reidel Publishing Company, Dordrecht, The Netherlands, pp 197-809. [24] Breemen N van (1988) Long-term chemical, mineralogical, and morphological effects of

iron-redox processes in periodically flooded soils. In: J W Stucki, BA Goodman, U Schwertmann (eds) Iron in Soils and Clay Minerals. D Reidel Publ Co, Dordrecht, The Netherlands, pp 825-841 1251 Calmano W, Hong J, Forstner U (1992) E d u R von pH-Wert und Redoxpotential auf die Bindung und Mobilisiemng von Schwermetallen in kontaminierten Sedimenten. Vom Wasser 78: 245-257 [26] Fbrstner U et al. (1 990) Sediment criteria development - contributions from environmental geochemistry to water quality management. In:D Heling et al. (eds) Sediments and Environmental Geochemistry. Springer-VerlagBerlin, pp 3 11-338 [27] Plant JA, Raiswell R (1983) Principles of environmental geochemistry. In: I Thornton (ed) Applied environmentalgeochemistry. Academic Press London, pp 1-39

290

[28] Peif€er S (1989) Biogeochemische Regulation der Spurenmetallaslichkeit wi4hrend der anaeroben Zersetzung fester kommuler Abfillle. Dissertation Universitiit Bayreuth, 197 p [29] Stumm W, Morgan JJ (1981) Aquatic Chemistry. Wiley, New York [30] Holmes CW, Slade EA, McLerran CJ (1974) Migration and redistribution of zinc and cadmium in marine estuarine systems. Environ Sci Technol 8: 255-259 [31] Gendron A, Silverberg N, Sundby B, Lebel J (1986) Early diagenesis of cadmium and cobalt in sediments of the Laurentian Trough. Geochim Cosmochim Acta 50: 741-747 [32] Emerson S, Jahnke R, Heggie D (1984) Sediment-water exchange in shallow water estuarine sediments. J Mar Res 42: 709-730 [33] Hines ME, Lyons BWM, Armstrong PB, Orem WH, Spencer MJ, Gaudette HE (1984)

Seasonal metal remobilization in the sediments of Great Bay, New Hampshire. Mar Chem 15: 173-187 [34] Hunt CD, Smith DL (1983) Remobilization of metals from poIIuted marine sediments. Can J Fish Aquat Sci 40: 132-142 [35] Prause B, Rehm E, Schulz-Baldes M (1985) The mobilisation of Pb and Cd from contaminated dredge spoil after dumping in the marine environment. Environ Technol Letts 6: 261 -266 [36] Kersten M (1989) Mechanismus und Bilanz der Schwennetallfieisetzung aus einem SUB-

wasserwatt der Elbe. Dissertation Technische Universittit Hamburg-Harburg [37] Darby DA, Adams DD, Nivens WT (1986) Early sediment changes and element mobiliza-

[38]

[39] [40]

[41]

[42] [43] [44] [45] [46]

tion in a man-made estuarine marsh. In: Sly PG (ed) Sediment and water interactions. Springer-VerlagNew York, pp 343-351 M a d B, Miehlich G, GrBngroft A (1985) Untersuchungen zur Grundwassergefilhrdung durch Hafenschlick-Spiilfelder. II. Inhaltsstoffe in Spolfeldsedimenten und Porenwissern. Mitt Dtsch Bodenkundl Ges 43fi: 253-258. M a d B, Miehlich G (1988) Die Wirrkung des Redoxpotentials auf die Zusammensetzung der PorenlUsung in Hafenschlickspiilfeldern.Mitt Dtsch Bodenkunde Ges 56: 289-294 Forstner U, Colombi C, Kistler R (1991) Dumping of wastes. In: Merian E. (ed) Metals and Their Compounds in the Environment. Chapter I.7b, pp 333-355. VCH Verlag, Weinheim h e t h , J-D, Milde G,KerndorEH, Schleyer, R (1989) Waste deposit influences on ground water quality as a tool for waste type and site selection for final storage quality. In: The Landfill - Reactor and Final Storage, P Baccini (ed). Lecture Notes in Earth Sciences 20, Springer-Verlag, Berlin, pp 399-415 Baccini P, Brunner PH (1991) Metabolism of the Anthroposphere. Springer-VerlagBerlin, 157 p Baccini P, Belevi H, Lichtensteiger T (1992) Die Deponie in einer okologisch onentierten Volkswirtschaft. Gaia 1: 34-49 Swift RS (1977) Soil Organic Matter Studies. IAEA Vienna. pp 275-281 Lichtensteiger T, Brunner PH, Langmeier M (1988). Kltirschlamm in Deponien. EAWAG Project No. 30-681. EC-COST. h d r e a e MO et al (1984) Changing biogeochemical cycles - group report. In: Nriagu JO (ed) Changing Metal Cycles and Human Health. DahIem-Workshop. Springer Berlin pp

359-373 [47] Salomons W (1985) Sediment and water quality. Environ Technol Letts 6: 3 15-326 [48] Wiles CC, Barth E, de Percin P (1988) Status of solidificatiodstabilizationin the United

States and factors affecting its use. In: Wolf K, Van den Brink WJ, Colon FJ (eds) Contaminated Soil '88. Vol 1, pp 947-956. Kluwer Academic Publ, Dordrechflhe Netherlands [49] Stigliani WM (1992) Chemical time bombs, predicting the unpredictable. In: Chemical Time Bombs. European State-of-the-Art Conference on Delayed Effects of Chemicals in Soils and Sediments. Veldhoven, The Netherlands, Sept 2-5, p 12

29 1 [50] Salomons W, Forstner U (1984) Metals in the Hydrocycle. Springer-Verlag Berlin Heidelberg [511 Anonymous (1986): Leitbild fir die schweizerische Abfallwirtschaft (Guidelines for the Waste Management in Switzerland), Schriflenreihe Umweltschutz No 5 1. Eidgendssische Kommission f i r Abfallwirtscha!?, Bundesamt f i r Umweltschutz Bern [52] Baccini P (ed)( 1989) The landfill - reactor and final storage. Lecture Notes in Earth Sciences 20. Springer Berlin [53] Krebs J, Belevi H, Baccini P (1988) Long-term behavior of bottom ash landfills. Proc 5th Intern Solid Wastes Exhibition and Cod, ISWA 1988, Copenhagen [54] Belevi H, Sthpfli DM, Baccini P (1992) Chemical behaviour of municipal solid waste incinerator bottom ash in monofills. Waste Management Research 10: 153-167 [55] Patrick WH, Williams BG, Moraghan JT (1973) A simple system for controlling redox potential and pH in soil suspensions. Soil Sci SOCh e r Proc 37:331-332 [56] Herms U, Briimmer G (1978) Loslichkeit von Schwermetallen in Siedlungsabfallen und Boden in Abhtingigkeit von pH-Wert, Redoxbedingungen und Stoffbestand. Mitt Dt Bodenk Ges 27: 23-43 [57] Schoer J, Forstner U. (1987) Abschatzung der Langzeitbelastung von Grundwasser durch die Ablagerung metallhaltiger Feststoffe. Vom Wasser 69: 23-32 [58] Forstner U, Calmano W, G e m W (1991) Assessment of long-term metal mobility in heatprocessing wastes. Water Air Soil Pollut 57-58: 319-328 [59] Lee GF, Plumb RH (1974) Literature Review on Research Study for the Development of Dredged Material Disposal Criteria. US Army Corps of Engineers, Dredged Material Research Program, Report D-74-1. Vicksburg MS 1974: 145 p. [60] Forstner U (1985) Chemical forms and reactivities of metals in sediments, in: R Leschber, RD Davis, and P L'Hermite (eds) Chemical Methods for Assessing Bio-Available Metals in Sludges and Soils. Elsevier Applied Science London, pp 1-30 [61] Tessier A, Campbell PGC, Bisson M (1979) Sequential extraction procedure for the speciation of particulate trace metals. Anal Chem 5 1: 844-851 [62] Kersten M, Forstner U (1986) Chemical fractionation of heavy metals in anoxic estuarine and coastal sediments. Water Sci Techno1 18: 121-130 [63] Kersten M, Forstner U (1987) Effect of sample pretreatment on the reliability of solid speciation data of heavy metals - implications for the study of early diagenetic processes. Mar Chem 22: 299-312 [64] Kersten M et al. (1985) Freisetzung von Metden bei der Oxidation von Schlammen. Vom Wasser 65: 21-35. [65] Luoma SN, Davis JA (1983) Requirementsfor modeling trace metal partioning in oxidized estuarine sediments. Mar Chem 12: 159-181 [66] Calmano W, Ahlf W, Forstner U (1988) Study of metal sorptioddesorption processes on competing sediment components with a multi-chamber device. Environ Geol Water Sci 11: 77-84 [67] Calmano W, Forstner U, Hong J (1993) Mobilization and scavenging of heavy metals following resuspension of anoxic sediments fkom the Elbe River. In CN Npers, DW Blowes (eds) The Environmental Geochemistry of Sulfide Oxidation, Proc. ACS Geochemistry Division Symposium, Washington DC, August 25-27, 1992 (submitted) [68] Michaelis H von (1988) Integrated biological systems for effluent treatment from mine and mill tailings. In: W Salomons, U Forstner (eds) Environmental Management of Solid Waste - Dredged Material and Mine Tailings. Springer-VerlagBerlin, pp 99-1 13 [69] Kalin M, Everdingen RO van (1988) Ecological engineering: biological and geochemical aspects: phase I experiments, In: W Salomons, U FCIrstner (eds) Environmental Management of Solid Waste - Dredged Material and Mine Tailings. Springer-Verlag Berlin, pp 114-128

292 [70] Kester DR, Ketchum BH, Duedall IW,Park PK (eds)(1983) Wastes in the Ocean. Vol 2: Dredged-Material Disposal in the Ocean. Wiley New York, 299 p [71] Bokuniewicz HJ (1982) Submarine borrow pits as containments for dredged sediments. In: Kester DR, Ketchum BH, Duedall IW,Parks PK (eds) Dredged Material Disposal in the Ocean, John Wiley & Sons New York, pp 215-227 [72] Morton RW (1980) "Capping" procedures as an alternative technique to isolate contaminated dredged material in the marine environment. In: Dredge Spoil Disposal and PCB Contamination: Hearings before the Committee on Merchant Marine and Fisheries. House of Representatives, Ninety-sixth Congress, 2nd Session, on Exploring the Various Aspects of Dumping of Dredged Spoil Material in the Ocean and the PCB Contamination Issue, March 14, May 21, 1980. USGPO Ser No 96-43, Washington DC.: pp 623-652 [73] Craig PJ, Moreton PA (1984) The role of sulphide in the formation of dimethyl mercury in river and estuary sediments. Mar Pollut Bull 15: 406-408 [74] Sahm H, Brunner M, Schobert SM (1986) Anaerobic degradation of halogenated aromatic compounds. Microbial Ecol 12: 147-53 [75] Kersten M (1988) Geochemistry of priority pollutants in anoxic sludges: cadmium, arsenic, methyl mercury, and chlorinated organics, in: W Salomons, U FBrstner (eds) Environmental Management of Solid Waste - Dredged Material and Mine Tailings. Springer-Verlag Berlin: pp 170-213. [76] Ehrenfeld J, Bass J. (1983) Handbook for Evaluating Remedial Action Technology Plans. Municipal Environ Res Lab Cincinnati. EPA-600/2-83-076. August 1983 [77] Malone PG, Jones LW, Larson RJ (1982) Guide to the Disposal of Chemically Stabilized and Solidified Waste. Report SW-872, Office of Water and Waste Management. Washington DC: US Environmental Protection Agency [78] Wiedemann HU (1982) Verfahren zur Verfestigung von Sonderabfiillenund Stabilisierung von verunreinigtenBoden. Ber Umweltbundesamt 1/82. Erich Schmidt Verlag Berlin [79] Goumans JJJM, Van der Sloot HA, Aalbers ThG (eds)(1991) Waste materials in construction. Studies in Environmental Science 48. Elsevier Amsterdam 672 p [80] Calmano W et al. (1986) Behaviour of dredged mud after stabilization with different addiW Assink, WJ Van Den Brink (eds) Contaminated Soil. Martinus Nijhoff Pub1 tives, in: J Dordrecht: pp 737-746 [8l] Gambrel1 RP,Reddy CN, Khalid RA (1983) Characterization of trace and toxic materials in sediments of a lake being restored. J Water Pollut Control Fed 55: 1271-1279 [82] Bambauer HU (1992) Mineralogische Schadstohobilisierung in Deponaten - Beispiel: Riickstrinde aus Braunkohlenkraftwerken.BWK Umwelt-Special March 1992: S29-S34 [83] Bambauer HU (1991) The Application of Mineralogy to Environmental Management - An Overview. Int Con@ on Applied Mineralogy, Pretoria/RSA 2-4 Sept 1991. C.133 Volume I, Paper 3 [84] Forstner U, Kersten M, Wienberg R (1989) Geochemical processes in landfills. In: P Baccini (ed) The Landfill - Reactor and Final Storage. Lecture Notes in Earth Sciences 20, pp 39-8 1. Springer-Verlag, Berlin [85] Faulstich M (1989) Inertisierung fester Rucksthde aus der Abfallverbrennung. AbfallwirtschaflsJournal 1 (7/8): 20-56 [86] Brunner PH (1989) Die Herstellung von umweltvertraglichen Rohstoffen als neues Ziel der , Miillverbrennung. Mull und Abfall2 1 : 166- 180 [87] Faulstich M, Freudenberg A, Kocher P, Hey G (1992) Remelt-Verfahren zur Wertstoffgewinnung aus Rucksttinden der Abfallverbrennung. In: M Faulstich (ed) Riicksttinde aus der Mullverbrennung. EF Verlag f i r Energie und Umwelttechnik Berlin, pp 703-727 [$81 Anonymous (1 99 1) IMRA-Project. Immobilisierung von Rauchgasreinigungsrckstanden aus Kehrichtverbrennungsanlagen.MBT Umwelttechnik AG Zurich and others, 365 p

293

THE IMPORTANCE OF BIOLOGICAL TESTING IN THE ASSESSMENT OF METAL CONTAMINATION AND SITE REMEDIATION C. R. Lee', J. W. Simmers', D. L. Brandon", L. J. O'Neil' and M. J. Cullinane' and J. M. Robertsonb a Environmental Laboratory, U. S. Army Engineer Waterways Experiment Station, 3909 Halls Ferry Rd, Vicksburg, MS 39180.

Ware & Freidenrich, Attorneys at Law, 4 0 0 Hamilton Avenue, Palo Alto, CA 94301-1825 INTRODUCTION

The evaluation of the nature and extent of heavy metal contamination in the environment and the need for remediation has become one of the most perplexing environmental challenges for mankind in recent years. This paper describes a case study that used bioassays in conjunction with soil data to determine areas of contamination and the need for remediation. A Remedial Investigation was undertaken by the U. S. Department of the Navy at the Naval Weapons Station, Concord, California under the Navy Installation Restoration (IR) Program to meet requirements of the Comprehensive Environmental Response, Compensation, and Liability Act (CERCLA), as amended by the Superfund Amendments and Reauthorization Act (SARA). The investigation addressed the following questions: 1. What was the nature and extent of metal

contamination on site? 2. Was the contamination bioavailable and migrating

into foodwebs on the site? 3. What was the condition of the site for wildlife

habitat and how important was the site to the natural resources of an adjacent Bay? 4. Were remedial actions required? If so, where and what action? Contamination resulted from previous uncontrolled discharges of industrial waste into environments adjacent to chemical manufacturing plants. Chemical contamination was thought to have migrated from the sources either down a creek into wetlands, or down drainage ditches in the wetlands into tidal creeks that ultimately emptied into Suisun Bay. The extent of the migration was not known in 1984, when this investigation was initiated. Soil samples had shown elevated levels of arsenic, cadmium, copper, lead, zinc, and selenium in certain locations in the field. No other data, with the exception of some water data, were available at that time. The entire Remedial Investigation/ Feasibility Study and selection of remedial actions were completed in 1989.

294

EXPERIMENTAL METHODS

An experimental design for sampling the site was formulated based on previous soil and water data and the potential pathways for contaminant mobility from the sources to receptors [l]. The site was approximately 124 hectares including a freshwater creek that flowed from rolling hills down through pastures, along side a chemical processing plant, under railroad tracks, through other pastures into a freshwater wetland, under another railroad trestle and into a brackishwater wetland tidal creek that emptied into Suisun Bay. Other sources of contaminant discharges included another chemical manufacturing plant adjacent to the brackishwater wetland as shown in Figure 1. Two Reference areas were selected for the study, one upstream from waste discharges and the other on the shoreline of Suisun Bay, beyond the downstream path.

Figure 1. Aerial photograph of waste lagoon discharge (arrows) in 1967.

295

A total of 637 soil samples were collected from 414 locations across the site and 10 locations in the reference areas. At 38 locations, triplicate soil samples were collected for chemical analysis, and for plant and earthworm bioassay tests over a range of potential contamination levels, to enable a rigorous statistical analysis of the data. Locations were selected based on previous soil data, potential contaminant pathways for migration within the site, and to represent a range of contamination from the lowest to the highest levels present at the site. A total acid digestion was used to determine soil arsenic, cadmium, copper, lead, zinc, and selenium concentrations in all collected samples. Duplicate soil digestions were conducted for every 10th sample collected. Because of the diverse environments under investigation and the nonexistence of the same colonizing species of plant or soil invertebrate at each sample site, a plant and an earthworm bioassay were conducted in the laboratory on field collected soil as indicators of contaminant migration from soil into foodwebs. A total of 178 plant and earthworm bioassays were performed on the collected soil. Plant and earthworm bioassay procedures were described elsewhere [2 and 3, respectively]. The plant bioassay procedure as shown in Figure 2, exposed CvDerus esculentus under controlled greenhouse conditions to field collected soil samples for 45 days and observed toxicity as death of the plant. Live plants were harvested after the exposure period and analyzed for arsenic, cadmium, copper, lead, zinc, and selenium. The earthworm bioassay as shown in Figure 3, exposed Eisenia foetida under controlled laboratory conditions to field collected soil samples for 30 days. Toxicity as death of earthworms was observed. Live earthworms were collected, depurated for 24 hrs and analyzed for metals. Since aquatic animals of the same species were not observed across the creeks and wetland sites, triplicate clam bioassays were performed at 33 locations across the site and in the reference areas to indicate contaminant migration from surface waters into foodwebs. Locations of clam biomonitoring were selected based on previous data on sources of contamination and pathways of migration such as at a point of discharge and then at different points downstream of the point of discharge or down a drainage ditch from a source of contamination toward Suisun Bay. Clam bioassay procedures were described elsewhere [4]. Ten clams (Corbicula fluminea) were placed in each of three cages, as shown in Figure 4, and then placed at each sample location in existing surface water in the field as shown in Figure 5. After exposing clams to the surface water in the field for 30 days, the clams were retrieved, depurated for 24 hrs and analyzed for metal bioaccumulation. Clam biomonitoring has been successfully used extensively in Europe to detect surface water migration of metal contaminants and is similar to the world reknown "Mussel Watch". Consequently, clam biomonitoring was selected in place of other invertebrates. Had existing invertebrates been sampled across the site (as passive biomonitoring), it was questionable if

296

,

Cypsrus escuknfus YELLOW NUTSEOGE

& c

6-Inch PVC 340p Nyler Mesh

15-em Pleriglass 1

Dredged Malerial

Earlhworrns

Mop Nylcr Mesh 15-cm PVC

Water ReSeNOir

Figure 2. Plant Bioassay Test

Figure 3 . Earthworm Bioassay Test

sufficient biomass of one or more species would have been available for chemical analysis. In addition, data may have been collected on more than one species and the interpretation of any bioaccumulation among species would have been difficult to interpret across the site. Consequently, the same species of clam was used in an active biomonitoring effort to insure sufficient biomass for chemical analysis and sufficient data across the site to statistically analyze and interpret the results. Statistical analysis of the resulting data included analysis of variance (ANOVA) and Duncan's New Multiple Range Test. All significant differences were obtained at P<0.05. RESULTS

Soil contents (dry weight basis) ranged from 0 . 8 to 2,500 mg/kg for arsenic, from 0.3 to 8 9 mg/kg for cadmium, from 0.0 to 3,050 mg/kg for copper, from 0.0 to 7,600 mg/kg for lead, from 25 to 8 5 , 4 9 0 mg/kg for zinc and from 0.0 to 138 mg/kg for selenium. There were 6 and 5 soil sampling locations where soil samples were acutely toxic to plants and/or earthworms, respectively, as shown in Figure 6. Laboratory bioassay plants bioaccumulated arsenic, cadmium, copper, and zinc at certain soil sampling locations to concentrations

297

Figure 4 . Clams Placed in Cage

Figure 5 . Clam Cage Placed in Stream

statistically above those exposed to soil from the Reference areas. Earthworms bioaccumulated arsenic, cadmium, copper, lead, and selenium at certain soil sampling locations. Clams also bioaccumulated arsenic, cadmium, lead, and zinc from surface waters in localized areas on site when compared to tissue concentrations observed in clams exposed in the surface waters of the two Reference areas. Mice (Mus musculus). were trapped in the reference areas and in the contaminated areas using standard cages. Mice were found to have bioaccumulated cadmium in livers and kidneys and lead in femurs above that of reference area mice in some of the areas where laboratory bioassay tests indicated potential migration of contaminants into foodwebs. Areas of bioaccumulation and/or toxicity were delineated and are shown in Figure 6. Areas of contamination were delineated as those areas where one or more test results for either soil, plant, earthworm, or clam concentration data were statistically higher than test resuits at the Reference areas and included locations of plant and earthworm toxicity, as shown in Figure 7. This approach was interpreted to mean that there was a 9 5 % probability that at these locations, soil, plant, earthworm and/or clams would be observed to contain measurable increases in contaminants above those normally found in the Reference areas. Using this approach, approximately 18 hectares of the total 124 hectares was designated as contaminated and potentially requiring remediation. During small mammal trapping, numerous endangerea saltmarst harvest mice were captured and released. Other endangered species were sighted or thought to be present or, the site as showo. in Table 1.

298

Table 1 Listed and Candidate Endangered and Threatened Species Common Name

Scientific Name

Status

Salt Marsh Harvest Mouse

Reithrodontomvs Raviventris

Endangered

California Clapper Rail

Rallus Lonqirostris Obsoletus

Endangered

Bald Eagle

Haliaeetus Leucocephalus

Endangered

California Least Tern

Sterna Antillarum (=Albifrons) Browni

Endangered

California Black Rail

Laterallus Jamaicensis Coturniculus

Candidate

These observations indicated that the contaminated areas were very important to one or more endangered species and therefore to the overall resources associated with Suisun Bay, A toxicological evaluation was conducted using the data collected in the Remedial Investigation and available toxicological literature for the contaminants observed. The conclusion of the toxicological evaluation was that there was a need for remedial action in the areas of contamination identified in the Remedial Investigation. A Feasibility Study of remedial action alternatives was conducted and decision rules for the scope and extent of required remediation were developed [5]. These rules considered potential criteria such as soil metal content exceeding values statistically higher than Reference areas; soil metal concentrations that exceeded the California Total Threshold Limit Concentration (TTLC) or Soluble Threshold Limit Concentration (STLC); and soil pH below 5.0. Much of the soil data exceeding STLC appeared to agree with soil sample locations of observed plant, earthworm, and clam contaminant bioaccumulation or death of plants and earthworms. In addition to these criteria, modifying factors of topography, presence of wetlands, presence of endangered species, source of contamination to other areas, and precedence were considered in drawing the boundaries of areas to be remediated. In sensitive environmental areas, application of a stringent cleanup criteria can trigger the environmental impacts waiver [6 and 7 1 . In such cases, it is necessary to balance the short-term adverse impacts associated with the remediation process with the long-term potential benefits of remediation.

299

,

-0

SAMPLE LOCATIONS

V MOUSE CAPTURES

Figure 6. Area resulting in toxicity (death) and metal bioaccumulation in plants or earthworms or clams.

WASTE LAGOOh

* 0 SAMPLE LOCATIONS

I

\

\

,

Figure 7. Area of contamination requiring remediation

300

Accordingly, the site was divided into zones of active remediation, passive remediation and monitoring, as shown in Figure 8. Contaminated areas were divided into active and passive remediation zones. Active remediation resulted in a positive control/treatment of contamination such as excavation of contaminated soil or liming to raise soil pH to 7.0. Passive remediation resulted in leaving the contamination in place and conducting short- and long-term intense monitoring of soil, water, and biota in the affected area. The impact of actively remediating the entire contaminated area was thought to be more severe to endangered species and their important habitat adjacent to Suisun Bay than the impacts associated with leaving a portion of the contamination in place. Passive remediation was designed to identify problems, and if necessary, trigger future active remediation activities. A monitoring zone surrounding the contaminated area was established to evaluate the effectiveness of the remediation, and because contaminants may migrate into presently uncontaminated areas. Monitoring in this zone is less intensive than in the passive remediation zone and it is not considered a likely candidate for future active remediation. Earthworm bioassays and clam biomonitoring will be conducted periodically as needed as part of the monitoring plan for site remediation. Based on the results of the Remedial Investigation/ Feasibility Study, the total area of 124 hectares was divided into 8 hectares of active remediation, 10 hectares of passive remediation and the remaining 106 hectares were designated as the monitoring zone in Figure 8. CONCLUSION

This study supported litigation between the Department of the Navy and Responsible Parties that resulted in a clean up and site remediation at a cost of approximately 2 5 million, dollars. Laboratory plant and earthworm bioassays and field clam biomonitoring were important in identifying the areas of contamination and the area to be remediated. ACKNOWLEDGEMENT

The investigation resulting in the information presented herein was sponsored by the US Department of the Navy, Facilities Engineering Command, Western Division, San Bruno, CA. Mr. Carl Schwab and Ms. Beth Gibeau were the Department of the Navy project managers. Permission to publish this information was granted by the Chief of Engineers.

30 I

Figure 8. Areas of active and passive remediation (shaded) and monitoring (unshaded) REFERENCES 1 2 3

4 5 6 7

Lee, C . R. et a i . i986; Miscellaneous Paper-EL-86-2, U . S . Army Engineer iqaterways Experiment Station, Vicksburg, MS, p 191. Polsom, J r , . B . L. and Lee, C

R . 1381; In: Proceedings of Heavy !.letals i n the Environment, International Conference, Amsterdam, NL. Simmers, 3 . W., Lee, C. R . and Marquenie, 5. M . , 1984; In: Proceedings of the Third International Symposium on Interactions Between Sediments and Water, Geneva, Switzerland. Marquenie, J . !. 1981; in: Proceedings of Heavy Metals in the Environment, lcternational Conference, Amsterdam, NL. Cullinane, J r . ; t!. J . , Lee, C. R . , and O'Neil, L . J . , 1988; Mi.sce;Lanecus Faper EL-86-3, U . S . Army Engineer Waterways Experiment Station, Vicksburg, MS, p 878. U . S . Environmental Protection Agency, 1987; OSWER Directive 9234.0- 3, Washington, 13C. Cullinane, Jr., J. M., Lee, C . R., 3'Neil, L. 2 . and Robertson. J. M . , 1990; In: Proceedings of Superfund '88 Conference, Hazardous Materials Control Research Institute, Silver Springs, MD.

This Page Intentionally Left Blank

CHAPTER 5 Inland waters

This Page Intentionally Left Blank

303

RECOVERY FROM EUTROPHIC TO OLIGOTROPHIC STATES IN LAKES: ROLE OF SEDIMENTS

D. Span, V. Coppee, J. Dominik, G. Balvay" , F. Berthierb, C. Martin and J.-P. Vernet Institute F.-A. Forel, University of Geneva, Versoix, Switzerland Institut National de la Recherche Agronomique, Thonon les Bains, France SILA, Rue des Terrasses, F-74960 Cran- Gevrier, France

NATURAL OR CULTURAL EUTROPHICATION ? Eutrophication of lakes is not necessarily the direct consequence of anthropogenic influence. Initially, a deep lake with an oligotrophic state will receive the products of erosion from the catchment area. The result over a geological time scale (many thousand of years) is the accumulation of dissolved nutrients in a diminishing volume of water due to the in-filling of the basin.

&-o-

Geological \

time scale

1

oligotrophic state "in-filling"

I man-made cycle (few decades)&

of the basin natural eutrophication (many thousand years)

/

/

physical cuitural eutrophication

(;;;leath

I"

of the lake

\ Figure 1: The eutrophication problem: natural or cultural?

304

This natural fertilization of water may lead to an eutrophic state with slow changes for aquatic organisms. Finally the “death of the lake is physical. This evolution depends on the geological nature of the watershed, climatic and limnological conditions (4). However, if the same lake is placed in an industrialized and densely populated region, the activities of people can accelerate this natural process. Cultural eutrophication can reduce the ecosystem in only a few years. Since the 1940’sdeterioration processes associated with the excessive fertilization of natural waters have produced a new focus for the environmental studies of national and international organizations (4). Eutrophication problems (increase of algal biomass, reduced water quality, ...) directly affect the use of water by people. h a result, important efforts to reduce phosphorus loads in water have been undertaken. External (principally reduction of P loadings) or internal (artificial mixing, biomanipulation, O3-input, chemical treatment of bottom water ...) restoration measures have been undertaken in different lake settings (5, 6). Local or regional communities bordering lakes have taken the initiative to monitor the chemistry of lake water. These studies have been promoted, for instance, by the SILA (“Syndicat Intercommunal du Lac d’ Annecy”) for Lake Annecy since 1957 (l), by the CIPEL (“Commission Internationale pour la Protection des Eaux du Uman”) for Lake Geneva since 1957 (2)and by the CIPAIS (“CommissioneInternazionale per la Protezione delle Acque Italo- Svizzere”) for Lake Lugano (3). The studies have been focused firstly on chemical and biological investigations of water column, then on the sediment water interface and are now considering the whole lake system (watershed, water, rivers, sediments). The main aims of this paper are: -1) to provide a qualitative and quantitative cornparaison of three different trophic evolutions over the last few decades. -2)to contribute to a better understanding of the role of deep-lake sediments in Pcycle. -3)to present different lake restoration managements.

SAMPLING AND ANALYTICAL METHODS Figure 2 shows the location of the three lakes considered in this study. Lakes Geneva and Annecy are located in the Rh6ne River watershed. All sample sites are located in the deeper parts of each lake (Fig.3). These sites correspond to the reference stations choosen for monitoring of water quality by different official surveyers (1,2,3).

305 Sediment cores in Lake Annecy (about 40cm long) were collected in December 1991 by scuba-divers using a modified Phleger corer. The same device was used in Lake Lugano four times during one annual cycle between June 1989 and March 1990 by the manned submarine F.-A. Forel. In Lake Geneva, sediment cores were taken using a n Ambiilh gravity corer in May 1987.

Figure 2: Location of study lakes (G: Germany; F: France; I: Italy; Sp: Spain; Sw: Switzerland).

Interstitial water samples were collected using a n in-situ dialysis technique (7, 8). Dialysis pore water samplers were placed in the sediment near the coring operation with the aid of the divers or the submarine. For Lake Geneva, interstitial water was obtained from core sediment by centrifugation and filtration through a 0.45 pm membrane. Dissolved P was quantified by colorimetric methods (9). In sediments, Total

Fe and Mn were measured by ICP - AES after dlgestion with HC104 followed by HC1+ H F (10). Total Organic Carbon (TOC) was measured by a n oxydation technique (11) and Inorganic Carbon (CaCOd by volumetry. The forms of phosphorus in sediment were determined by colorimetry after a sequential leaching procedure modified after Williams (12, 13). This procedure is based on chemical extractions, separating P presents in organic and inorganic forms. organic DhosDhorus (0-P) associated with allochtonous and autochtonous organic matter: settling material consisting of dead or living planktonic organisms, excretion products, etc... Biologically this form is important, because of its availability for phytoplankton growth after bacterial decomposition. - inorganic DhosDhorus: may occur principally in the form of Fe-bound and Cabound phosphate. The procedure used distinguishes the two following forms: Non Apatite Inorganic Phosphorus (NAI-P) is mainly co-precipitated with Fe, Mn, Al hydroxides or may be adsorbed onto clay. This form seems easily bioavailable when redox or pH conditions change in the environment. Apatite Inorganic Phosphorus (AI-P) is included in apatite minerals (Ca,(PO,),OH). Ca-bound phosphate is probably not easily remobilizable under the existing conditions in surficial sediment.

-

306

RECENT TROPHIC EVOLUTION OF THE LAKES

The three lakes (Fig. 3) are very different in their depth, size, slope and trophic state. Their main morphological characteristics are listed in Table I. These lakes, like a majority of the other Alpine lakes were originally oligotrophic. During the fifties and the sixties, the deterioration of the water occured more or less rapidly, which modified the lake chemistry and biology. From the sixties, the politic of restoration have been different for each lake. However in most lakes, the P concentration in epilimnion began to decrease at the end of the seventies, depending to their water residence time.

Lake Annecy (France): Lake Annecy, divided in two sub-basins, is the smallest lake stuhed. The principal town on its shoreline is Annecy which is situated downstream, near the outlet of the lake. I t is a thermomonomictic lake stratified from June to October. Hypolimnic waters become depleted in oxygen during summer with a maximum depletion for the northern basin "Grand Lac". In the 1960's, it became increasingly meso-eutrophic (14). A difference between the two sub-basins in their maximum P contents is observed, with relatively high concentrations at the "Grand Lac" compared to the "Petit Lac" (Fig. 3a). Lake Annecy was a forerunner in lake restoration in France. An association of the communities bordering the lake was created in 1957 to supervise the restoration program. This organisation also ensures a continuous collection of scientific data of the chemical quality of water. The technical solution to restore lake quality was the construction of a peripheral pipe to collect domestic effluent around the lake. Construction began in 1962 and finished in 1976. Only one waste-water treatment plant was then necessary and discharges downstream of the lake. The total cost of the treatment plant with the network of main sewers (total length: 344 km) and the incineration plant reached 445 millions French Francs (90 millions $). This simple system to reduce nutrient influx into the lake is very efficient as Lake Annecy returned to an oligo-mesotrophicstate. It is now considered to be a "clean" lake.

surface area km2

watershed area km2

mean depth m

max. depth m

59 172 41

Lake Geneva "Grand lac" "Petit Lac"

582.0 503.0 79.0

7975

Lake Lugano northern basin southern basin

48.9 27.5 20.3

6 15 270 290

171 55

Lake Annecy "Petit lac"

26.5 6.2

278

42

volume lake km3

mean water residence Yr

trophic state

310 310 76

89.0 86.0 3.0

11.9

mesotmp hie

288 95

4.7 1.1

12.3 1.4

eutrophic eutropluc

1.1 0.2

3.2

54

number number of sewage of residents treatment plants 154

1,580,000

5

oligotrophic

Table 1 Principal morphometric and hydrological characteristics of the study lakes and their sub-basins.

200,000 1

116,500

308

a) Lake Annecy

Annecy

50 .

1960

..

1970

1980

1990

"Petit Lac"

. .. . '

Lake Geneva

*&

,

~

'

"

'

'

"

. : '

'

.''

'

'

'

'

'

'

'

~ ~ . . .

c

1960

Geneva

1970

1980

1990

From CIPEL (2,14)

c) Lake Lugano Vedeggio,

,Cassemte

2 o o ~ TotP I Pg/l 1 :

northern basin (0-284 m)

*uther,

150 100 507

southern bash A

sampling Site

*y *

principal treatment plant

.------

northern basin (0 -1 00 m)

--

basin

'

.,

From ClPAlS (3,16)

Figure 3: Temporal evolution of mean Total Phosphorus concentrations in the water column of the studied lakes.

309

Lake Geneva (Switzerland - France): Lake Geneva is the largest of the subalpine lakes. It is dwided into two principal basins: the "Grand Lac" the deeper basin and the "Petit Lac". The major tributaries are the alpine RhBne River and the Dranse River. The RhBne River is the largest source of P to Lake Geneva, with a n annual load of 1.100 tons of P (12). The population centres are located around the lake with important towns (Lausanne, Geneva, Thonon, Evian ...) and along the Alpine RhBne River. Compared to the Lake Annecy (in size of the lake and its watershed), the decentralization of the population and the higher efluent discharge required a different solution to the eutrophication problem. The principal treatment plants were constructed between 1972 and 1978. They dscharge into the upper part of the RhBne River or directly into the lake. The improvements in waste water treatment and the P ban in detergents since 1986 in Switzerland improved lake water quality. A recovery trend from eutrophic to mesotrophic state with a mean total P concentration of 55pgPll in 1990 (15) has been observed since 1980 (Fig. 3b).

Lake Lugano (Switzerland - Italy): Lake Lugano is subdivided into two distinct sub-basins with steep slopes. The northern basin is the deeper one. It is meromictic with a permanentaly anoxic hypolimnion between 80 to the bottom a t 285 m depth. The southern basin is shallower. Its morphology shows several sub-basins. I t is monomictic with strong seasonal variations of dissolved oxygen contents in the water column. The main characteristics are listed in Table 1 and additional information is reported in recent publications covering a variety of research undertaken on this lake (16,17,21,23). The deterioration of this lake oecured more rapidly than the two others. The high density of population and the complexity of hydrodynamic system of the lake were responsible for the rapid eutrophication. The comparison of the P evolution in the two principal basins shows a decrease initiated in 1976-78 in the upper 100 meters of the northern basin and since 1982 in the southern basin (Fig. 3c). The improvement observed in the epilimnion of the northern basin results from the shifting of the eMuent emissions from Lugano in 1976 from the northern to the southern basin, through the Vedeggio River (Fig. 3c). The significant inventory of available P due to the absence of mixing in the lower hypolimnion of the northern basin represents a serious problem.

310

Thus, despite the relative improvement observed, the two basins remain eutrophic. The P concentrations and the primary production are still high.

ROLE OF SEDIMENTS IN THE P CYCLE The three lakes have a large flat bottom area compared to their total surface area. Thus attention will be focused on the relationship between P cycle and the sedlment of these areas. In general, investigation of P content in the hypolimnion during a year, show strong seasonal variations. This evolution is well known and is related to the seasonal variation of oxygen content in the bottom water (18). After the turn-over (generally in February-March) the surficial sediments are oxydized and covered by a thin floc layer of Fe-Mn oxy-hydroxides with co-precipitated P. During this short time (few weeks), sediment behaves as a sink for P. Consequently the P diffusion from sediment is low as well as the P concentration in bottom water. During water stratification, microbial degradation of the organic matter consumes all oxidizing agents a t the sediment water interface. The surfieial sediment becomes reduced, and thus the oxyhydroxide barrier trapping P is gradually dissolved. The pool of dissolved P increases in the sediment and diffuses to the hypolimnic bottom water. This process is not always observed in the deeper part of the lakes studled. The mobility of P a t the sediment water interface depends not only on the redox conditions but also on other factors such as composition of the sediment, adsorption capacities of settling particles, sedimentation rates... The relationship between P and oxygen contents in the hypolimnion of the various lakes (Fig. 4) shows a net difference of behaviour. For the same period and especially during the period of oxygen depletion, the P-response in hypolimnic waters is particular to each basin. The clearest example is given by the difference between the two sub-basins of equal depth in Lake Lugano (Melide and Figino). At site Melide (Fig. 4), the mean concentration of P is two times higher than at site Figino, despite similar oxygen contents in bottom water. Other factors control the P-cycle, were as discussed in the following section. The figure 5 illustrates the relationship between dissolved P in bottom water and mean sedimentation rate. To the first approximation, the sedimentation rate relates to the amount of particles present in the water column close to the bottom. There is a negative correlation between the rate of sedimentation and the total P content of bottom water. In basins such as Lake Annecy, which have a relatively high sedimentation rate, there is a small pool of dissolved P in their hypolimnion. In

31 1

Tot P

Pg/l

*

500 -

3

400

3

-

during stratification

0

A

Lake Annecy (G.L.) LakeGeneva Lake Lugano (Melide) Lake Lugano (Figino)

c

z

during turn-over

100

2

0

4

8

6

10

Bottom water

12

0, mg/l

Figure 4: Mean Total P content at maximum and minimum oxygen concentrations in the bottom water of the three lakes.

300

b

c

E

200

-

100

-

A

m

0

0.05

0.10

0.15

0.20

sedimentation rate g/crn2.y

Figure 5: Mean Total P content evolution in bottom water compared to the sedimentation rate of the lakes (same symbols as in fig.4).

312

contrast, the other basin, which have a high pool of P in their bottom water, correlate with relatively low sedimentation rates. This relationship may be interpreted as indicating a strong capacity of particles to absorb P. The observations of the bottom water confirm the importance of the external parameters, e-g. climatic and hydrologic conditions. However information given by the sediment (lithology, interstitial water...), which contains significant enrichment of P, must be considered to understand the global P-cycle. Some major geochemical characteristics of the surficial sediments (0.5 cm) are given in figure 6. The data reflect a high heterogeneity between each site. The sediments of Lake Annecy are very calcite-rich, reflecting the calcareous nature of its watershed. Sedments in Lake Lugano are relatively rich in organic matter. The highest contents of Fe, Mn and P (5, 0.7 and 0.4 W respectively) are found at Figino (southern basin of Lake Lugano), which is close to a high solid input source (Vedeggio River). However relatively low values of Mn and P are measured at Gandria (northern basin of Lake Lugano) despite its classification as eutrophic. The permanently reducing conditions at the sediment-water interface prevent the accumulation of these mobile elements. The high concentration of Fe a t this site may be explained by a stable detrital form of this metal. 80 40

0 12

8 4

0

Lake Annecy (oligotrophic)

Lake Geneva (rnesotrophic)

LakeLugano (eutrophic)

Figure 6: Mean composition of surficial sediments (0-5 cm).

313

In general, the total P content in sediment shows a relative increase with the trophic state of the lake (Fig. 6 and 7). The depth distribution profiles of Total P in sediments show generally an enrichment at the interface (Fig.7 b,c,d,e). The high concentration in the surficial sediment results from an upward migration of &ssolved P from deeper anoxic zone. Dissolved P finally co-precipitates or is adsorbed in the oxidizing surficial sediment layer. The increased concentration observed a t the surface may also be due to increase P loadings. At depth, total P profiles can be disrupted by the presence of detrital layers alternating with the regular varved sediments. These layers, generally coarser, behave as a lithological "barrier" for the migration of the dissolved P towards the surface (19). At Gandria, sediments a t a depth of 27-30 cm, just below a thick turbidite series, have a large NAI-P content. This suggests that before deposition of these turbidites, estimated between 1955-1963 (20), oxic conditions prevailed in bottom water, allowing P storing at the former sedment water interface. Deposition of turbidite fossilised the ancient interface. At present conditions, similar deposition of turbid& do not lead to P trapping as there is no P concentration peak a t the sediment water interface (Fig. 7f).

Forms of phosphorus Total P can be divided into organic and inorganic forms. The vertical profiles of the different forms are presented in figure 7. In general, the Organic Phosphorus (0-P) shows a decrease with sediment depth, suggesting a delay in mineralization compared

to sedimentation. At site Melide only, the 0 - P decrease is very rapid, because of active microbial degradation of organic matter (21). This upper zone a t site Melide is subject to important seasonal redox changes (2 1) increasing the mineralization of the organic matter with rapid consumption of suitable oxidants. The Apatite Inorganic Phosphorus (AI-P) considered as inert P, shows an almost constant concentration and is not remobilized in the surficial sediments. Normally, the other inorganic form of P (NAI-P) decreases strongly with depth, due to the decreasing redox potential which favours dissolution of Fe and Mn oxyhydroxides in the deeper sediment layers. However, the vertical distribution is not so simple. NAI-P is also controlled by processes than adsorption or co-precipitation with Fe, Mn or N oxyhydroxides. A t depth, precipitation of other inorganic P mineral phases is thermodynamically possible in such reducing environments (22). For example, below the first few centimeters, interstitial waters are supersatured with respect to vivianite

314

Oligo - mesotrophic states

Eutrophic states

Petit Lac

a)

O F Tot P

0.5 g/h I 0

0 r

50

d)

56 100

0

I

southern basin (Figino) Tot P 2 4 g/kg o 50

100

20

40

I

AI-P

0-P NAI-P

I

Grand Lac

b)

t.

0

0.5

2

2 0

40 1

40

I

I

Lake inecy Central plain

2)

souther1 basin (Melide)

e)

50

fl

4

6

Lr

0

50

100

i

I

i

northern basin (Gandria)

0

I

10-

I

I

C

20

40

Lake Lugano

Figure 7: Vertical distribution of Total P (gP/kg) and its forms (%) in sediments of each study lake.

315

(Fe3(P04),.8H20) and reddingite (Mn3(P04).3H20). In P speciation procedure, these minerals may also be extrated during the NAI-P step (13). At site Figino, vertical P distribution is especially irregular. The frequent deposition of thin detrital layers at this site leads to successive P-trapping between the silty deposits (23). In comparison to varved sediment, the detrital layers present higher AI-P and lower NAI-P contents (19). The diagram of the percentage of P forms (Fig. 8) in the upper five centimeters of sediments shows the strong role of AI-P for the oligomesotrophic lakes (Lakes Annecy and Geneva) with 40 to 50 % of Total P. The AI-P percentage depends also on the petrography of the watershed rocks and the contribution of detrital fraction to the total sediment. Thus the most important part of P is irreversely trapped in sediment. At Gandria the site (eutrophic state) AI-P is abnormally high because under permanently anoxic conditions it is the only stable form of phosphorus in sediment. At the southern basin of Lake Lugano, quantities of NAI-P in surface sediments are very high, characterizing an eutrophic environment.

APATITE P 100 %

100 % NON-APATITE INORGANIC P LG: Lake Geneva LA 1 : Lake Annecy (Grand Lac) LA 2: Lake Annecy (Petit Lac)

100 % ORGANIC P

GA: Lake Lugano (northern basin) Me: Lake Lugano (southern basin) Fi: Lake Lugano (southern basin)

Figure 8: Terniary diagram. of the proportions of each P form in the sediments (0-5cm) of each lakes.

Overall, NAI-P shows significant correlation coefficients with Mn (r2= 0.98,n= 15) and Fe (r* = 0.57, n=15). The proportion of the different forms of P, especially inorganic, may reflect the trophic states of these lakes. AI-P is the prevailing form in sediments of oligo-mesotrophic lakes while the eutrophic lake sediments are characterized by high NAI-P content.

316

Phosphorus exchange at the sediment water interface EutroDhic environment (Lake Lugano) At site Melide (southern basin of Lake Lugano), the P behaviour in sediment under oxic or anoxic conditions is well characterized (Fig. 9). The comparison of the two extreme situations shows: -under oxic conditions (March 1990) a P- trapping in the upper centimeters of sediment. The P release is low with a vertical diffusive flux of 0.8 mgP.m-2.d". The flux is calculated according to Fick's first law in one dimension and simplified by Berner (24). -during the maximum of stratification (December 1989). the amorphous inorganic complexants for P (NAI-P) are progressively reduced and then dissolved, The P release is important (about 4.2 mgP.m-2.d-1) and increases the P content of the hypolimnion. The sediment acts as an internal source. The loss of P (NAI-P) between these two periods corresponds to 30 % of NAI-P initially present in the sehment. At site Figino (southern basin of Lake Lugano), subjected to the same seasonal redox changes of the bottom water, the P content in sediments does not vary significantly. The quantity of NAI-P is approximatively 5 times greater than at Melide. This very high P-sink capacity is related to the large reservoir of Fe and Mn in sediments. The metal-rich sediment seems to be able to act as a sink for P even under unfavorable conditions, though P loadings are important in this basin. This behaviour explains the lower concentrations of P during stratification in the bottom water observed at Figino than those at Melide (Fig. 4). Total Dissolved P (mgP/I) 0

2

4

6

0

0.5

1.o

2.0

-2 I

E

0

v

5P 0

0"

2

December 1989: 0.1 mg %/I

0

0.5

1.0

gP/kg

March 1990: 3.3 mg Q/l

Figure 9: Seasonal evolution of NAI-P and Dissolved Phosphorus in surficial sediment (0-2 cm) of Melide (southern basin of Lake Lugano). Dissolved oxygen data supplied by CIPAIS (16,17).

317

OlifzotroDhic environment (Lake Annecvl Only the deeper basin in Lake Annecy shows a P release fmm sediments during depletion of dissolved oxygen. In December (Fig.lO), the dissolved oxygen content reaches 0.3 mg 0,A. Even with this low concentration of oxygen, diffusion flux of P is very limited (0.6 mgP.m-2.d-1). I t is important to note that this flux is inferior to the lower diffusion flux observed in eutrophic basin of Melide. However, with a low external loadmg to Lake Annecy, the remobilization of P from s e h m e n t to hypolimnic waters is responsible for the increase of P observed over the last years in hypolimnion of the "Grand Lac" (Fig.3a and 10). In the southern basin which is not as deep, the surficial sediment remains oxidmd and the P flux in December, is insignificant and no seasonal variation of P in water column is observed.

0

-

0.5

Total Dissolved P (mgP/I) 1.0 0

0.5

1.o

10

A

E 30

5 a

50

0

-k 20

a-

Y

40 December 1991: 0.3 mg %/I

February 1992: 10.1 mg 041

Figure 10: Seasonal evolution of Total dissolved P in the water column and interstitial water of sediment from Lake Annecy ("Grand Lac"). MesotroDhic environment n a k e Geneva) The two periods (1973-78 and 1987-91) presented in figure 11, show a similar evolution of dissolved oxygen content in the bottom water of Lake Geneva. This phenomenon is due to the same succession of mild winter temperatures, leading to a n incomplete mixing of the hypolimnic waters. This results, during the years 1976-78, in a spectacular increase of P in bottom waters, which reaches, for instance, 290 pg P/l in 1978 (2). In parallel, the concentration of NAI-P in sediment is inferior to 100 mg P k g (Fig. 12). The deep sediments act as a significant internal source of P. During the second

318

period (1986-91), while a similar degradation of oxygen content is observed, the related P concentration increase in bottom water is much lower. Moreover seasonal variation of dissolved P are of lower amplitude than during the preceeding period. This is due to a marked decrease of the P peak value resulting from P release. The NAI-P contents in surficial sediment of the central plain of Lake Geneva show a significant evolution, both with oxygen content and with Total P in bottom water (Fig. 12). In surficial sediments, NAI-P content during the recent period (1987-89) remains high, at least 5 times higher than during the anoxic situation of 1978 (25). This evolution in surficial sediments may be related to the reduction of P- input into the lake. Indeed, the data from CIPEL (26) in hypolimnic waters reveal a recent diminution of Particulate Organic Carbon content with 0.10 mg Cfl for the period 1986-91against to 0.58 mgCfl for 1976-78. Tot P

O2

0

200

5

100 . . . _.. _

0

..

.... .).

.

0

Figure 11: Comparison of Total dissolved P and dissolved oxygen content between two periods in the bottom water of Lake Geneva (-308m). Data from CIPEL (2)

319

Consequently the consumption of dissolved oxygen responsible for organic decomposition is less important. The concentration of oxygen does not reach concentration lower than 1 mgOz A. This oxygen concentration can reduce or limit the "global" release of P from sediment. The chemical process in the hypolimnion is indirectly related to the decreased P loading in the epilimnion of Lake Geneva. In fact, even if the primary production remains high (26), the phytoplankton composition changes with a revival of small sized forms (27). This organic matter is more rapidly degraded in the photic zone and the quantities of settling organic material through the water column decrease and thus less oxygen is consumed in deep waters. Tot P PSP/I

200

150 -

5

zi 3

-

100

87

a

I

I

E 0 s

m O

a Mar 198

0

m

100

200

q

400

NAI-P mgP/kg

Sept 1987.

a

1

0

300

1

I

I

100

200

300

400

NAI-P

mg P/kg

Surficial sediments Figure 12: Variation of NAI-P content in surficial sediments (0-2cm) with respect to Total dissolved P and oxygen concentration in the bottom water of Lake Geneva (-308m). Data from CIPEL (2,25).

320

The relationship between dissolved oxygen and estimated diffusion fluxes of P from sediments is presented in figure 13. Apparently, diffusion flux gradients in different lakes occur at various thresholds of oxygen content. For instance, in Lake Geneva a threshold of about 3-4 mg 0, /l is observed whereas it decreases to about 0.8 mg 0,/l at site Melide in Lake Lugano. Each site has its particular behaviour of P release, controlled by: depth basin, exposed time of oxygen depletion, reservoir of available dissolved P in sediment and the chemical composition of these sediments especially in hssolved Fe and Mn contents.

$1 W E

8

-

3

G 4

I

1

1

LakeLugano IMelide Lake Geneva

Bottom water Figure 13: Comparison of diffusion fluxes at the sediment-water interface in each lake.

CONCLUSIONS The three case stuhes show that eutrophication may be reversed by a diminution of the external loading of P. This strategy consists of the introduction of the treatment of waste water eMuents to eliminate P and the substitution of polyphosphates in detergents. Restoration of a lake is determined by considering several factors: political and management interests, density and repartition of population, hydrology and morphology of the lake (residence time of water, degree of mixing of whole water...). After 20-years of nutrient control, the restoration of deep lakes appears to follow three steps (example of Lake Geneva):

32 I

1: a rapid decrease (few years) of P concentration in epilimnic waters due to reduced P inputs into the lake. The delay depends to the nature of P input eg point

s o u m s (treatment plants) or diffuse sources (agriculture). 2: a second step is the change of biomass composition with small sized species, which are easily recycled in the photic zone, dominating the phytoplankton. 3: the third consequence, after a delay-time of a few years, is a decrease in the consumption of dissolved oxygen close to the sediment water interface during stratification periods. In such conditions the P retention in deep lake sediment increases. Eventually, the lake returns to a more oligo-mesotrophic state. In the case of Lake Lugano, its partial restoration has started with P-decrease only in the epilimnic waters (step 1) and with some changes in phytoplankton population (step 2) (16). The northern basin of Lake Lugano, which is permanently anoxic at depth, remains a significant problem. Realistically, only additional internal measures, such as hypolimnic oxygenation may accelerate the recovery of this basin (5). In terms of the role of bottom sediments in P-cycle, we must distinguish between the northern basin of Lake Lugano, whose sehments may be considered "inactive", from the other sites. The P reservoir is presently the lower hypolimnion. The principal zone of nutrient exchange has been transfered from sediment water interface to the chemocline at 80-100 m in the column water. The sediments of the other sites, subjected to more or less regular seasonal redox variations, directly influence the quality of bottom waters. The main factors controlling the P behavior at the sedimentwater interface (release or retainment) are: oxygen content in bottom water, iron and manganese contents in the sediment, adsorptive processes, the origin and nature of particulate P, the rate of sedimentation and the possible deposition of detrital material.

ACKNOWLEDGEMENTS The authors would like to acknowledge the help and support of the staff a t the Institute F.-A. Forel, in particular P.Y. Favarger and B. Gallerini for the chemical analyses. We thank the "Banca Unione di Credito" of Lugano for sponsoring the dives in Lake Lugano and the two subaquaclubs ("La Coulee Douce" and the C.S.A.) for help in sampling in Lake Annecy. The manuscript benefited from corrections by Dr SM Hutchin.son and Dr J L Loizeau. This work was supported by the SILA (Lake Annecy), by the CIPEL (Lake Geneva) and by the Swiss National Science Foundation (projects 20-33569-92 and 20-33363-92) (Lake Lugano).

322

REFERENCES Documents related to the three lakes studied may be obtained from: 1 S.I.L.A.: Syndicat Intercommunal du Lac d'Annecy, Rue des Terrasses, F-74960 Cran

-

Gevrier, France. 2 C.I.P.E.L.: Commission Internationale pour la Protection des Eaux du U m a n , 23 av. de

Chailly case postale 80, CH-1000 Laueanne 12, Switzerland. 3 C.I.P.A.I.S.: Commissione Internazionale per la Protezione delle Acque Italo-Svizzere, Via Fabio Filzi 22,1-20124 Milano, Italia. 4 Meybeck M, Chapman DV, Helmer R. GEMS, WHO and UNEP, 1989; 306p. 5 Imboden DM. Aquatic Sciences, 54, 34, 1992 381-390. 6 Gachter R, Imboden D. In :Chem. Processes in lakes, John Wiley and Sons, 1985; 363-387. 7 Brandl H, Hanselmann KW. Aquat. Sci., 5311, 1991; 55-73.

8 Hesslein RL. Limnol. Oceanogr., 21, 1976 912-914. 9 H a r w d JE,Van Steenderen RA, Kuhn AL. Water Res., 3, 1969; 417-423. 10Favarger PY. Annal. Tech. In Env. Chemistry 2, Pergamon, 7, 1982; 371-376. 11Gaudette HE, Flight WR, Toner L, Folger DW. J. Sed. Petrology, 44, 1974; 249-253. 12 Burrus D, Thomas RL, Dominik B, Vernet JP,Dominik J. Hydrol. Proces., 4, 1990; 85-98. 13Williams JDH, Jaquet JM, Thomas RL. J. Fish. Ree. Board Can., 33, 1976; 413-429.

14Barroin G. Int. Symp. on Inland waters and Lake Restoration, Washington, 1980; 312-315. 15Blanc P, Corvi C, Rapin F. Report CIPEL, campaign 1990, 1991; 23-44. 16Barbieri A, Polli B. Aquat. Sci., 54, 314, 1992. 17 Barbieri A, Mosello R. Aquat. Sci., 54, 314, 1992; 219-237.

18Mortimer CH. J. Ecol., 29, 1941; 280-329. 19Span D, Dominik J, Loizeau JL, Belzile N, Vernet JP. Chem. Geol. , 102,1992; 73-82. 20Dominik J, Loizeau JL, Span D. Climate Dyn., 6, 1992; 145-152. 21 Lazzaretti MA, Hanselmann KW. Aquat. Sci. , 54, 314, 1992 285-299. 22 Nriagu JO, Dell CI. Am. Mineral., 59, 1974; 934-946. 23 Span D,Dominik J, Lazzaretti MA, Vernet JP. Aquat. Sci., 54, 314, 1992; 277-284.

24Berner RA.Princeton Univ. Press, N.J., 1980; 241P. 25 Span D, Arbouille D, Howa H, Vernet JP. Hydrobiologia, 207, 1990; 161-166.

26Pelletier JP. Report CIPEL, campaign 1991, 1992; 89-97. 27 Druart J C and Pelletier JP. Report CIPEL, campaign 1991, 1992; 77-87.

323

DYNAMICS OF THE AUTOCHTHONOUS AND CONTAMINANT BACTERIAL COLONIZATION OF LAKES (Lake of Cadagno and Lake of Lugano as model systems)

R. Peduzzi, A. Demarta and M. Tonolla Labotatory of microbial Ecology, University of Geneva, Cantonal Institute of Bacteriology, Via Ospedale 6, CH-6904Lugano.

INTRODUCTION In the aquatic environment the principal role of bacteria is the degradation and the remineralisation of the organic material produced [1,2]; it is therefore logical to deiennine the highest levels of bacteria correspondmg to the zones that also show the greatest concentrations of organic material, which constitutes their nutrient substrate. Numerous studies have established that in lake water the highest concentrations of bacteria are found in the trophogenous zone of algal production, ~ I Ithe water-sediment interface and in the first few centimetres of mud [3-51. The distribution of the bacteria in the water column can, however, appear different Gom that described in response to biotic and abiotic factors inside the aquatic environment considered, or to external inputs as for example allochthonous microorganisms. The condition of permanent stratification of meromictic lakes offers the possibility of studying microbial populations that are adapted to particular physicochemical conditions in a largely stable environment, compared with bodies of water that are subject to periodic circulation. Besides the hgh concentrations in the trophogenous zone and m the sediment, as found in takes that are not permanently stratified, the autochthonous heterotrophc bacteria of meromictic lakes also presents an additional peak corresponding to the chemocline [ 6 ] . For bacteria inhabitant of human and animal intestine and pathogens found in a lake, the principal source of contamination is domestic sewage. In the industrialised countries the majority of the raw sewage is transported to treatment plants whch greatly reduce the bacterial levels, but which are not able to eliminate them

324

completely if they are not equipped with appropriate disinfection systems (e.g. chlorination, UV disinfection). Therefore, the water leaving a purification plant still carries a certain number of intestinal and pathogenic bacteria. In fact, it has been demonstrated that for pathogenic bacteria of the genus Salmonella purification plants without means of disinfection are not only ineffective in eliminating them, but also that in certain cases the percentage of recovery of these bacteria in the effluent is greater than that in the affluent [7]. In regard to germs belonging to the genus Aeromonas, the concentrations are reduced between entering and leaving the trestment plant, from 105 CFU/ml to 104 CFUlml, respectively [8]. The autochthonous bacterial population of the body of water receiving contaminated water reacts to the input of extraneous microorganisms in a homeostatic manner, by trying to reestablish a balanced situation. The pathogenic microorganisms and the intestinal bacteria thus tend to decrease considerably. However, some of them are able to survive for a more or less long period, depending on the specific environmental conditions [9, 101. The present work is a contribution toward the understanding of the factors that determine the vertical distribution of bacteria, m a d y heterotrophc, in the water column.

BACTERIAL COLONIZATION OF A MEROMICTIC LAKE - THE LAKE OF CADAGNO

The Lake of Cadagno is a meromictic alpine lake situated at 1923 metres above sea level in the Saint Gotthard massif (Switzerland). The bed of the lake rests partly on crystalline rocks and partly on dolomitic rocks. Human influence on this body of water is limited to two summer months, with tourism, sports fislmg and alpine agricultural activity. The Lake of Cadagno is therefore suitable to be considered as a natural, uncontaminated model system. The state of crenogenic meromixis of its water is determined by the presence of a dolomitic vein in direct contact with the lake. Springs rising in the bed of the lake feed the monimolimnion with water originating from the dolomitic rock and carrying mineral salts in solution, such as calcium, magnesium, carbonates and sulphates. The mixolimnion, on the other hand, receives predominantly water that is poor in dissolved salts originating from

325

crystalline rocks. The intense biological activity of anaerobic mineralisation in the sediments causes high levels of trophogenous and toxic ions such as phosphates, ammonium and hydrogen sulphde in the monimolimnion. In the chemocline ideal conditions exist for the massive development of photosynthetic sulphur bacteria, which derive energy from sunlight and metabolise the trophogenous compounds originating from the monimolimnion [ 111 (Figure 1).

a)

Conductivity [IlS/cm]

too

0

200

1

Temperature "C

6

8

10

Turbidity [FTU]

10

300 0

I

4

b)

+

,

.

I

30 ,

,

I

40 .

,

0

2

4

6

60

50 '

Oxygen [mgn]

8

12

20 .

-

8

I

,

I

1

8

1 0 1 2

0 2 4

6 8

10 12 14 16

18 20 Depth [m]

i i'u 1

0

1

10

Sulfide [rngn]

20

30

+

Figure 1. Vertical profiles o f a) temperature (T), conductivity at 20OC (C) and b) oxygen (0),turbidity (TU) and sulfide (S) taken in the lake of Cadagno (September 17, 1992).

326

Besides the physicochemical parameters, the vertical distribution of bacteria in the water column was analysed by means of direct counting methods and indirect culture-plate methods (Figure 2).

m

6

2

4

[total cells/ml] x 1o 0

Depth [m]

sc

TC [ Ch. okenii celldml] 0

500 1000

log [CFU/ml]

0

2

4

6

Depth [ml

Figure 2. Vertical distribution of total bacterial counts (TBC), total count of Chromdtium okenii (TC) and saprophyte counts (SC) (September 17, 1992).

The direct counts with a fluorescence method (acridine orange staining), indicating the total number of bacteria present, showed concentrations of 2.0 x 105 cells/ml (k 0.2 x lo5 cells/ml) in the mixolimnion and high mean values of 2.3 x 106 c e W d

321

(* 0.4 x lo6 cells/ml) in the chemocline. The high mean counts of 2.6 x 106 cells/ml (5 0.2 x lo6 cells/ml) in the monimolimnion indicate a high rate of sedimentation. Using the same method the sulphur bacteria of the species Chromatium okenii, distinguishable morphologically because of their dimensions (ca. 10-15 pm), could be quantified separately. These are not present in the mixolimnion due to the absence of hydrogen sulphide and the presence of too high concentrations of oxygen. In the chemocline they are present in large amounts, with values around 4.8 x 104 cells/ml (* 1.3 x 104 cells/ml), imparting a purple colour to the water. At this depth they reach their maximum concentration and represent about 4% of the total number of bacteria present. In the monimolimnion, however, their concentrations remain lugh due to the effect of the sedimentation. The plate-culture method provides useful indications for estimation of the presence of heterotrophic bacteria capable of growing on not selective culture substrates. The saprophytic bacteria counts at 20°C vary from 1.4 x lo1 CFU/ml (+ 0.6 x lO'CFU/ml) to 9.9x 10' CFU/ml (& 6.3 x lo1 CFU/ml) in the mixolimnion and in the monimolimnion, respectively. The maximum values of 1.6 x 103 cells/ml (k 0.5 x 103 cells/ml) are found in the chemocline. It is interesting to note how a stratification exists in the chemocline between the heterotrophic and the phototropluc bacteria. The increase in the turbidity and in the saprophytic bacteria counts corresponding to the total oxygen consumption at a depth of 11 metres (Figure 2) is followed by the numerical proliferation of bacteria belonging to the species Chromatiurn okenir only towards 12 metres, where the hydrogen sulphide reaches concentrations around 0 mg/l. Although the intensities of light at the level of the bacterial zone are relativeIy weak, with values of about I vE/cm 2 x sec and exclusively of wavelengths between 530 and 600 nm [ 121, the photosynthetic sulphur bacteria contribute predominantly to the primary production of the lake. In fact, in summer periods of optimal light on the surface (ca. 1500 pE/m2 x sec) 3.5 mg C/m 3 x h are fixed in the volume of water between 10 and 12 metres (206'000 m3), while in the 1.8 million cubic metres of the mixolimnion, from 0 to 10 metres, 8 mg C/m3 x h are fixed by the action of the algae [ 131.

328

ALLOCHTHONOUS BACTERIAL CONTAMINATION OF EUTROPHIC LAKE - THE LAKE OF LUGANO (GULF OF AGNO)

AN

This study was carried out over a period of two years (1990-1991) with the aim of assessing the impact of the inflow of the waters of a river carrying allochthonous bacteria, including human pathogens, on coastal waters of the relatively shallow (maximum depth 93 m) and well delimited Gulf of Agno (Lake of Lugano, Switzerland) (Figure 3).

I

1 Km

Figure 3. Map of sampling sites in the gulf of Agno (Lake of Lugano-Southern basin).

329

The affluent river, with a mean daily flow of 3,820 m3/sec (mean flow for the period 1979-1991), receives the discharge from a treatment plant that treats the waste fiom a population- equivalent of 112,500 inhabitants, situated 1.5 km above its mouth. The mean annual microbial load carried by the river, which is calculated from a series of monthly samples taken at its mouth, is summarised in Table 1. Table 1 Bacterial counts (annual means, 12 samplings) in the Vedeggio river. ~

Year

1990 1991

~

~

Saprophyte count Fecal coliform (CFU/ml) (CFU/lOOml) 10'452 9'212

~

~

Salmonella (No of positives samples/total samples)

26'1 12 17'821

7/12 5/12

Four buoys, fixed to the bed of the gulf at a distance of 80-100 metres from the shore and at a maximum depth of 45 metres, were used to delimit the points of sampling and measurements on the water column. At seasonal intervals over the two-year period the abiotic parameters investigated in situ were oxygen, temperature, conductivity, pH and turbidity. Water samples were taken for bacteriological analysis in regard to saprophyte counts (PCA at 37"C, CFU/ml), faecal coliform (agar with lactose, TTC and Tergitol, at 44OC, CFU/100 ml) and Salmonella (presence/absence in 1 litre of water). The density of the water was calculated according to the following formula [ 141: D = dH20 + 0.65 x 10-6 x K20 dH20 = Density of distilled water

K20

= Standardised conductivity

at 20"

During the winter period (Figure 4) the greatest bacterial concentrations in the lake were found in the surface samples (0.5 m). At this season the lake is in a state of thermal uniformity, expressed as a relatively constant water density value on the 3 water column, of between 1 and 1.00003 g/cm . The river water shows a density of 3 1 g/cm and therefore when it flows into the lake it remains in the superficial layers, where the greatest concentrations of bacteria are found. Furthermore, under the

330

influence of the Coriolis force the river water tends to carry the bacterial load particularly towards the area indicated by buoy B. 1,00004

3000

1,00003

5

D

2000

1,00002

z

rn

I

0

1,00001

>

t v)

1000

Z W

SAMPLING STATIONS

1,00000

0

Vedeggi river

0,99999

Gulf of

-

2000

i?

0

1,00001

> t v)

1000

Z W

FECAL COLIFORMS [CFU/l OOrnl]

1.00000

SAPROPHYTE COUNTS [CFU/mI]

1 +

DENSITY [g/crn 3 ]

2000

s LL 0

4 1.00001 2 1000

/ 0

......... , ................................. .. ....................... .. .. .. . . . . . . . . . . . .

A

B C

,

1

L

W

1,00000 0

~

0,99999

D

SAMPLING STATIONS

Figure 4. Water density, saprophyte counts and fecal coliforms at the sampling sites in the gulf of Agno (February 18, 1991)

33 1

In summer, the water column is clearly stratified and the thermocline is located at a depth of about 7 m (Figures 5a and b).

a>

0

10

1

'

l

Turbidity [FTU] + 20 30 .

r

'

r

40 .

Conductivity bS/cm] 170 180 190

l

+

i50 ,

0

200

160 -

1

5

.

,

-

1

.

1

.

Temperature "C * 10 15 20 25

Density [g/cm3]

1

30

0,996 0,997 0,998 0,999 1,000 1,001 0

Depth [m] 12 14

16

Iepth 20 18 I [m] 1 SAMPLING STATIONS

- s - D Vedeggio river V

Figure 5 a). Vertical profiles of oxygen (0),turbidity (TU), pH, temperature (T) and conductivity (C) of the gulf of Agno (Sampling site A) and b) profiles of density at the different sampling stations in the gulf (August 19, 1991). The river water, with a density of 0.99675, flows into the body of water in the layer of algal production, but without touching the surface water. The high concentration of bacterial contaminants is therefore detectable particularly in the sample taken at a depth of 7 m (Figure 6 ) . In the spring and autumn periods, corresponding to the times of circulation of the waters of the gulf, intermediate situations are observed, but the allochthonous

332

bacteria transported by the river are anyhow found in the gulf according to the respective densities of the water.

1,00000

3000

V]

0,99900

2000

-

0,99800-€

20

P m

I

0,99700 1000

2

SAMPLING STATIONS

W

Vedeggi river

0,996000 0

r

B C V D

A

0,99500

1 0,99900 0,99800

> 0,99700 t

FECAL COLIFORMS [CFU/l OOml]

v,

2 W 0,99600 0

2oool t

0

1 {

0,99900

-

0,99800mE

+

SAPROPHYTE COUNTS [CFUlmI]

DENSITY [g/cm3]

s

120rn(B=12rn)DEPTH]

-

9

0,99700 $

1000

v,

- 0,99600 0

a

0,99500

B C

A

z

x

c_l

0,99500

A

B C

D

SAMPLING STATIONS

Figure 6 . Water density, saprophyte counts and fecal coliforms at the sampling sites in the gulf of Agno (August 19, 1991).

333

In Figure 7 we have presented in diagrammatic form the situation in summer, when the density of the surface water of the gulf is lower than that of the affluent. In this situation the contamination with allochthonous bacteria is located at a depth of 7 m, where the density of the lake water is the same as that of the water of the affluent with its high bacterial content. In this case the gradient of fecal colifoms measured in the water coumn between 0 and 20 meters varies from 250 CFU/lOOml at 0.5 m, with a maximum of 2400 CFU/100ml at 7 m, and decreasing again at 20 m (50 CFu/100mI).

Figure 7 . Density and distribution of fecal colifoms carried into the lake by the contaminated affluent. Diagrammatic presentation of a summer situation.

334

CONCLUSION

By means of determination of the presence of bacteria on the water column of a meromictic lake, with a permanent stratification, we were able to ascertain how the maximum bacterial concentrations are connected to the trophogenous zones. In fact it was possible to observe an abundance of heterotrophic bacteria at the level of the supplementary primary production due to bacterial photosynthesis. In the chemocline, the metabolites released from the trophogenous bacterial zone (Chromatium okenii) at a depth of 12 m and the sedimenting biomass allow a proliferation of heterotrophic bacteria, which we already located at 11 m. The original supply of oxygen in this layer adjacent to the anoxic zone allows a marked bacterial catabolism to take place, whch is demonstrated by the saprophyte counts, by the increased turbidity and, as a consequence of the bacterial metabolism, by the subsequent disappearance of the oxygen. The aerobic mineralisation activity is thus situated 1 m above the layer of water with a high hydrogen sulphide content. In the second lake ecosystem investigated it was possible to determine the vertical distribution of allochthonous bacteria of faecal origin canied into the lake by a contaminated affluent. These germs are distributed in the water column following the flow of a passive transport due to the stratification of the water of the affluent, which reaches the zone of the lake where the water is of equal density. These results demonstrate that the gradient formed by the amount of nutrient elements and organic material present is not the only factor which governs the distribution of bacteria in a body of water. Consequently, in the study of the impact of biological contaminants of faecal origin carried into lakes it seems essential to take into account the differences between the density of the water of the affluent and that of the body of water into whch it flows. The distribution of bacteria of faecal origin and of pathogens shows seasonal variations which are reflected in the public-health and sanitary assessments of the water destined for bathing or dmking, which are the most important uses for which water is treated and protected.

335

REFERENCES [ l ] Rheinheimer G. In:Rheinheimer G, eds. Wiley & Sons: New York, 1991; 18524 1 [2] Fenchel TM, Jorgensen BB. In: Alexander M,eds. Plenum Press. New York, 1977; VOI 1; 1-41. [3] Pointdexter JS. In: Fletcher M, Gray TRG and Jones JG, eds Cambridge University Press, 1987;283-3 17. [4] Cole JJ., Findlay S, Pace M.L. Mar Ecol Prog Ser 1988; 43: 1-10. [5] Peduzzi R, Demarta A, Tonolla M. Aquatic Sciences 1992; 54: 3/4: 33 1-337. [6] Kunetsov SI. In: Droop MR, Jannasch H W , eds. Academic Press. London, 1977; 1-45, [7] Leclerc H, Mossel DAA, In: Leclerc H, Mossel DAA,eds. Doin, Paris, 1989; 273-278. [8] Demarta A, Ph Thesis No 2290, University of Geneva, 1989. [9] Martin G, In: Martin G, eds. Lavoisier: Paris, 1982; 279-292. [lo] Geldreich E. In: Mitchell R,eds. Wiley & Sons. New York, 1972; 235-235. [ 111 Peduzzi R,Tonolla M, Demarta A, Del Don C, Hanselmann K, Bachofen R. In: Vemet J-P, eds. Troisitme Conference Internationale des Limnologue &Expression Franqaise. Morges, 199 1; 183-189. [ 121 Tonolla M, Diplom Work University of Zurich. [ 131 Fried1 C, unpublished data. [I41 Buhrer H, Ambuhl H. Schweiz. Z. Hydrol. 1975; 37: 347-369.

This Page Intentionally Left Blank

331

THE ROLE OF THE BACTERIAL COMMUNITY IN THE RADIONUCLIDE TRANSFERS IN FRESHWATER ECOSYSTEMS.

F.Hambuckers-Berhin, A.Hambuckers, J.Remacle

Microbial Ecology, Department of Botany B22, University ot Liege, B 4000 LIEGE, BELGIUM.

INTRODUCTION

During the last forty years extensive industrial use of radioactive and fissile materials has taken place in Europe and elsewhere. Large scale nuclear facilities, for civil and defence purposes, were developed some of which giving rise to significant radioactive contamination of the biosphere. Therefore several questions are to be asked concerning human health hazards following releases of radioactive substances in the environment either continuously or accidentally. Four nuclear plants are settled in two sites along the river Meuse, Chooz (1) and Tihange (3). Moreover two new plants are being build in the site of Chooz (France). The river Meuse flows through France, Belgium and Netherlands and play a great role for the human welfare of the populations living in these countries. The river Meuse supplies five millions people with drinking water. Eight hundred km2 (agricultural lands, meadows) are flooded with water pumped from the river. It must also be mentioned that seventy tons of fishes were caught in 1982 in the river and its tributaries. Our laboratory is involved in a multidisciplinary research programme dealing with the study of the impact of nuclear power plants on the aquatic ecosystem more especially on rivers [l-2-31. Two neutron activation products contained in the waste effluents of the above nuclear facilities are mainly studied 6oCo and 134Cs. For example, the nuclear plant, Tihange 2, released lo00 MBq of 334Cs in 1982 [3]. It can be supposed that the transfer and the immobilization of these radionuclides could occur along the aquatic food webs since: (i) Co ion is a trace metal required for normal growth of all living organisms, (ii) Cs competes with K and ammonium ions [4] which are important metabolites. Moreover they could effectively radiocontaminate the living components of the food web owing to their periods (60Co: 5.2 years; 134Cs: 2.05 years). The final aim of the programme is to elaborate a general model of radionuclide tranfers through the food webs. Unfortunately the role of bacteria was neglected up to the recent years which impeded the finalization of the model. It is indeed obvious that bacteria could take a significant part in the radionuclide cycle at three levels. Firstly, they are vectors of radionuclide transport because they are able to take up, store or immobilize radionuclides generally in organic molecules or to precipitate them as insoluble salts e.g. carbonates,

338

sulphides as observed for Cd and Zn ions 15-61, Besides under some conditions bacteria release the radionuclides. Secondly bacteria can be an important step in further intake of radionuclide by the other living components of the food web. Thirdly bacteria are responsible, through the biodegradation, of the release of radionuclides sequestered in the organic matter.

w I I

I

I

T

Ll -

I

BOM

A T E R

CRAZING 2 /' I

CATABOLISM

3

I

Figure 1. Conceptual scheme of the radionuclide fluxes by considering the interactions between bacteria and the organic matter. POM: Particulate Organic Matter ROM: Refractory Organic Matter UDOM: Utilizable Dissolved Organic Matter BOM: Bacterial Organic Matter

Radionuclides Organic matter transfers Radionuclide transfers

The radionuclide fluxes (Figure 1) can be conceived by considering the interactions between bacteria and the organic matter described by Billen and Fontigny 171. These fluxes occur in the water column and in the interstitial water of the sediments. The particulate organic matter or polymeric organic matter (POM) is hydrolyzed by bacterial

339

ectoenzymes and transformed in utilizable dissolved organic matter (UDOM) i. e. small molecules such as amino acids, organic acids, mono- and oligosaccharides which can be easily taken up by hcterotrophic bacteria. A part of POM, not or slowly metabolizable by bacteria, constitutes the refractory organic matter (ROM). A part of this durable organic matter arises from cell envelopes and walls of degraded bacteria. These components can remain in individual form or can be adsorbed on clays (specially smectite or kaolinite). In this case, the radionuclide-binding capacity of ROM would he reduced because the sites normally available to radionuclides-binding would be masked or chemically neutralized as shown for metal-binding capacity of walls- and cell envelopes-clays composites 18-91. During organic matter hydrolysis, radionuclides sequestered in POM are either transfered to the water phase, or complexed with UDOM. On the other hand, bacterial yields were evaluated in the river Meuse [lo] and averaged 30%. It means that 70% of consumed substrate are catabolyzed while the remaining 30% are transformed into bacterial biomass (BOM). It results that a part of sequestered radionuclides is soluhilized through bacterial activities. Finally bacterial biomass can be destroyed by cell lysis or grazed by aquatic predators. It leads to organic matter recycling and transfers of immobilized radionuclides either to the water column or to POM or to higher trophic levels of the food web. The density of the bacterial community is difficult to assess in river ecosystems. However it can be assumed that 30% of the primary production is consumed by planctonic bacteria [ 1 I]. In steady state conditions the bacterial concentration remains relatively unchanged due to grazing which is of the same order of magnitude as bacterial growth rate 0.1 h-l 110-121. Moreover, the bacterial density is controlled by environmental conditions. So, a pulse of UDOM provokes bacterial blooms [13], and the initial bacterial level will be reached again through predation. All the above processes exemplify the significant bacterial contribution to the radionuclide cycling. The first part of this paper deals with the comparison of the aerobic bacterial communities growing in the water column and the sediments in order to evaluate the homogeneity of bacterial colonization of the river and to know whether it is allowed to extrapolate the radionuclide flux kinetics to the two main compartments of the river. The second part is devoted to the study of 6oCo and 134Cs fluxes between bacteria and water in the river Meuse.

COMPARISON OF THE AEROBIC BACTERIAL COMMUNITIES COLONIZING THE WATER COLUMN AND THE SEDIMENTS Material and methods Water and sediments were collected at Hastiere along the river Meuse. The samples were aseptically taken, stored in sterile bottles until analysis. The bacterial strains were collected from the water column following the scheme depicted in Figure 2 .

340

100 ml Meuse water (5 aliquots)

100 ml Meuse water (5 aliquots)

/

\

filtration (0.22pm pore size)

/

\

filter on PCA

bacterial film dispersed in 15 ml

10 fold'dilution solution spread on PCA (3 replicates)

I

100 pl of dilution spread on PCA (3 replicates)

Figure 2. Analysis of the bacterial communities of the river Meuse. Isolation of bacterial strains from the water column (PCA = Plate Count Agar).

The bacterial strains were harvested from the sediments following the scheme of figure 3. The absence of contamination was checked by following the same schemes with sterile Meuse water and sterile distilled water. The Petri dishes were incubated at 30°C during maximum 5 days. The strains were then isolated and purified. Fifty strains were selected and characterized by their morphology, gram-stain and biochemical profile ( API strip, Bio-Merieux, Oxy-Ferm tube, Roche). The characters were coded 1 for positive or present, 0 for negative or absent and 1 for missing test. Strain similarity was estimated by the simple matching coefficient of Sokal and Mitchener [14]. Cluster analysis was carried out by using the average linkage method (procedure CLUSTER, UPGMA [15]). The strains showing the highest similarity coefficient are proximally gathered in a hierarchical structure of more and more large groups. The results of the cluster analysis are graphically represented by dendrograms using the TREE procedure [ 151.

+

34 1

sediments (250 g)in suspension in 400 ml of sterile Meuse water (5 aliquots) shaking 48 h, room temperature

/ / 100 PI of siispension

\

\

decantation

I

spread on PCA (3 replicates)

filtration of supernatant (20 ml, 0.22 pm pore size)

I

( 3 replicates)

/

100 pI on PCA

bacterial film dispersed in 15 ml of sterile Meuse water

I

10 fold dilution

(3 replicates) 100 pl bn PCA (3 replicates)

Figure 3. Analysis of the bacterial communities of the river Meuse. Isolation of the bacterial strains from the sediments (PCA = Plate Count Agar).

Results The dendrogram (Figure 4) is composed of all the strains (50) which are gathered by hierarchical level. The minimal distance between the strains is 0 which means full similarity between the strains, on the ground on the chosen tests. So, the dendrogram gives an evaluation of the similarity between the strains and each strain is characterized by its level of aggregation I 161. All the 50 strains were recovered in two main clusters. Cluster 1 which joined 70% of all the strains, is composed of 63 % of water column strains and 37 % of sediment strains. Cluster 2 contains the remaining 30 % with 1 strain from the water column and 14 strains from the sediments. The strains clustered in this group were different from the strains of cluster 1 by the ability to use carbohydrates such as saccharose, mannitol, sorbitol and rhamnose (Table 1). The cluster analysis shows that the aerobic bacterial community isolated from the sediment was constituted of two groups : one group presenting the same biochemical features as the

342

aerobic bacterial community isolated from the water column, and an other group composed of strains only present in the sediments.

Average d i s t a n c e between c u s t e r s 0.0 w U Y U

S

2

18 6 15 8

W

7

S

w

11

5

I

Q) 4-J

5

5

w

7

UY 3

W W

20 21

Y

2

-

w

8

0

Y W

3 4

w

5

w

la

S

21 23 6

S

S S U Y S

w S 9 S Y S S

c\l L

0

+ m 3 -

0

1.2

1.4

14 19 3 13 19

ao 23 22

15 17 26 24 16

S S

7 14

S S

25

S S

22 9

S

13

S

27 18

S

1 .o

12 1

S S

3

0.8

11

x Y

0.6

16

S

L

0.4

1 10 9 17

w

w

0.2

10

Figure 4. Dendrogram of the cluster analysis of 50 strains isolated from the water column (w) and the sediments (s). I t appears that a great proportion of the bacteria colonizing the sediment and the bacterial community of the water column share similar biochemical features. It could be due to the fact that a part of the bacteria of the water column settle down and colonize the interstitial water and the sediments.

343

'Table 1 Main characteritics of strains of clusters 1 and 2 (see Figure 4) CLUSTER 1 number of strains total from water from sediment..

CLUSTER 2

35 22

15

13

14

1

% of positive responses in each cluster

Gram + Gram-

34 66

20 80

oxidase catalase arginine dihydrolase indole ONPG~@)

63 86 29 0 14

20 80 53 20 100

degradation of gelatin

60

13

acetate malate citrate

0 0

0 0 60

glucose saccharose arabinose mannitol sorbitol melibiose rhamnose

43

utilisation of

?dl

51

0 6 0 0 6 0

ONPG, o-Nitrophenyl---galactopyranoside

100 67 93 87 80 80 87

344

Consequently the role of bacteria in the radionuclide transfers in aquatic ecosystem will be investigated by analyzing the aerobic bacterial community isolated from the sediments. It will be considered in a first step that the responses could be representative of the aerobic bacterial communities of the river.

KINETICS OF 6oC0 AND 134CS TRANSFERS MEDIATED BY BACTERIA. It was decided to study the radionuclide transfers in a bacterial community instead of in a monospecific bacterial culture in order to obtain more reliable results from an ecosystemic point of view. The inoculum of the bacterial community was isolated from sediments sampled in the river Meuse at Hastikre (Belgium). The sediment! were stored in sterile bottles at 4'C until the experiments. A new preculture was carried out before each experiment in order to obtain a bacterial community as close as possible to the natural community. It is well known indeed that several strains cannot develop in laboratory conditions so that subcultured bacterial communities lose a lot of strains with uncontroled changes of the specific composition.

a. Radianuclide immobilisation by the bacte rial community, Material and methods

Preculture and cultures conditions. Precultures are carried out by inoculating sterile Meuse water with sediments. The Meuse water is supplemented with starch (2 g.1-l) and bacto-peptone (10 g.1-l) in order to improve the bacterial growth. The preculture vessels are shaked and incubated at 20'C until the mid exponential phase of bacterial growth i.e. during 48 h. At this stage, the bacterial community of the sediment is harvested by centrifugation (10 min, 2,OOOxg). The pellet is discarded and the supernatant is again centrifugated 15 min at 15,200xg. The last pellet is used to inoculate the experimental cultures which are carried out in the same Meuse water medium as preculture. The initial O.D. of cultures are f0.06. The cultures are shaked 24h at 20'C. The bulk growth rate of the bacterial community rankes between 0.13 and 0.15 h'l which is of the same order of magnitude as mentioned in earlier observations [lo].

Radionuclide uptake by bacteria. The culture medium is contaminated with 6oCo and 134Cs (up to 2,000 Bq.1-l) before the inoculation. At regular time intervals, an aliquot of the culture is centrifugated

345

(3 min, 12,500xg). The dry weight of the bacterial biomass is determined and after ashing in nitric acid, the radioactivity level is detected by gamma spectrometry.

Results

The uptake of 6oCo and 334Cs by the bacterial communities collected from the Meuse river was investigated in presence of increasing radiocontamination (up to 2000 Bq ml-l) in a Meuse water medium. When the data were expressed in the form of a double-reciprocal Lineweaver-Burk plot, two distinct phases were observed for 6oCo. One phase characterized the low contamination levels betwen 24 and 90 Bq.ml-' whereas another phase occured in the high contamination levels between 90 and 2,000 Bq.rn1-l (Figure 5). Each phase followed the Michaelis-Menten kinetics: 1/V = (l/Vmax) + (Km/Vmax). (l/[S]) where, V was the rate of uptake (Bq.g-I.hl), [S] was the activity of the medium (Bq.ml-'), Vmax was the maximal rate and K, was the Michaelis constant. Tdbk 2 (Bq.g-I.h-I) values obtained from Lineweaver-Burk plots of the Km (Bq,ml-l) and V uptake of 6oCo and ly%s.

134cs

Radiocontamination range of w ter (Bq.ml-B)

Vmax

24 - 90 90 - 2,000

2,500 18,692

22 - 295

1,286

Km

63.70 1,145.79 240

Vmax and Km values were calculated from the Lineweaver-Burk plot (Table 2). Km value was significantly greater in the high contamination levels. The two different phases of 6oCo uptake were therefore assumed to reflect high and low affinity uptake systems. The level of radiocontamination respectively explained 58% and 98% of the variation of the uptake rates. The 134Cs uptake kinetic only showed one phase (Figure 6). 73% of the variation of the contamination rate were explained by the level of radiocontamination of the water column. The Km values of Table 2 show that the highest affinity transport system is observed for %3in the low radiocontamination levels. During the radiocontamination,

346

33% and 24% of the variations of the concentration factor could be explained by the contact time between the biomass and the radionuclides for 6oCo (from 22 to 2,000 Bq 60Co.m1-1) and 134Cs (from 16 to 300 Bq 134Cs.ml-1) respectively. At the end of the uptake experiment (24 h; bacterial biomass: 500 mg dw.l-I), the activity remaining in the water column ranked between 25% and 40% of the initial activity in the case of 6oCo (initial activity: 0 to 2000 Bq.ml-I) and between 45 and 95% in the case of 134Cs (initial activity: 0 to 300 Bq.ml-I). For the same radiocontamination levels e.g.fl50 Bq.ml-l, the activity level of the bacterial community was +2.5 times higher for 6oCo (+50,000 Bq.g-' d.w.) than for 334Cs(+20 OOO Bq.g-l d.w.). Thus, an important part of &Co and 134Cs can be stored by the aerobic bacterial community which constitutes therefore a pool of radionuclides in the river ecosystem.

Figure 5. Lineweaver-Burk plot of 6oCo uptake by the bacterial community. [S] : radiocontamination eve s of the medium (Bq.mi-l) V : rate of uptake (Bq.g' .h- ) : maximal uptake rate (fjq.g-l.h-l) ::?affinity constant (Bq.ml- ) Equations: (1) low radiocon 'nation lev sy = 0.4 loy+ 23.8 lo9 x (r = 0.7595) (2) high radioconta ination leve : y = 53.5 61.3 10-?fx (r = 0.9924) (r : correlation coefficient)

\ I

347

l/CSI .lo-3

Figure 6. Lineweaver-Burk plot of 134Cs uptake by the bacterial community. IS] : radiocontamination eve s of the medium-based Meuse water (Bq.ml-l) V : rate of uptake (Bq.g' .h ) Equation: y = 0.8 + 186.7 x (r = 0.8570) (r : correlation coefficient)

I -\

b. Radionuclide release by the bacterial community.

Material and methods The bacterial community is contaminated with 6oCo and 134Cs during the exponential growth phase. Bacteria are harversted by centrifugation (3 min, 12,5OOxg), washed and suspended in a dialysis tube with sterile Meuse water (Spectrapor 5 ; thickness 0.09 mm; molecular sieve 14,000 D). The dialysis tubes are incubated in running sterile Meuse water at different temperatures (13'C, 20'C) and pH (6.5;7.0;7.5;8.0;8.5;9.0). At time intervals, the radioactivity levels detected either in the bacterial biomass or in the Meuse water outside the dialyse tubes.

Results The decontamination of the bacterial community loaded with 6oCo and 134Cs was investigated in relation to two environmental parameters, the temperature and the pH. The

348

activity levels of bacterial biomass at time zero ranked between 200 Bq.g-' d.w.and 900 MBq.g-l d.w. for 6oCo and between 1 and 50 P4Bq.g-l d.w. for 134Cs. When the temperature of the water column wdS maintained at 20°C, the kinetic of the decontamination of the bacterial community was described by a double negative exponential model. It was fitted using the procedure NLIN of SAS 1151 following the DUD method of Raltson and Jennrich [17]: y = m e-ax + n e-bx, where y : 6oCo or 134Cs levels in bacteria (Bq.g-I d.w.) x : time (h) m,n : ordinate values for each exponential (Bq.g-l .d.w.) a : slope of the line with equation In y = In m + ax b : slope of the line with equation In y = In n + bx a and b parameters depended on the desorption rate and were considered as an estimation of the biological half-lives (Tbl and Tb2) calculated as follow: Tbl = In 2 I a Tb2 = In 21 b Examples of adjusted curves were given for 6oCo (Figure 7) and 134Cs (Figure 8).

Figure 7. Desorption of 6oCo by the aerobic bacterial community. Observed values: square; Predicted values: line; y = 69 e-350.347x + 165 e-o.046x (r = 0.9121) y : 6oCo levels in bacteria (Bq.gl d.w.) x : time (h)

349

A

1500

0,

\

D

m

x

1000

+.> .4

u

500

0

0

200

100

300

400

500

Time (h)

Figure 8. Desorption of 134Cs by the aerobic bacterial community. Observed values: square; Predicted values: line; y = 916 e-7.82x 879 e-o.018x r = 0.9504) y : 134Cs levels in bacteria (Bq.g' d.w.) x : time (h).

+

I

For both radionuclides, the biological half-lives Tbl were found to be extremely short, of the order of a few seconds or minutes whereas the biological half-lives Tb2 were longer (1Sh to 461h for 6oCo and 39h to 8,976h for 134Cs). Except for a high value of Tb2 observed with 134Cs (374 d) and a very small value observed with 6oCo (0.6 d), both corresponding with the lowest radiocontamination level of the medium during the radiocontamination of the bacteria communities, the values of biological half-lives were of the same magnitude as observed for a cyanobacterium Scenedesmus sp . [ 18- 191. When the temperature of the water column was maintained at 13OC (average temperature of the Meuse river), the results showed that radionuclides fixed by bacteria were not released. At 20°C, the decontamination of the bacterial community was followed as the pH increased from 6.5 to 9. The chosen criteria was the increase of radioactivity in the water column which was a consequence of the decontamination of bacterial biomass. In any cases, the data were fitted by the following mathematical relation: y = m(l-e-ax) + n(1-ebx), where y : 6oCo or the 134Cs levels in the water column (Bq.ml-l) x : time (h) m,n : ordinate values (Bq.g-l .d.w.) a : slope of the line with equation In y = In m + ax

350

b : slope ofthe line with equation In y = In n + bx This mathematical adjustment showed that the evolution of radiocontamination level of the water column was adequately described by a biphasique process with a first phase characterized by a rapid desorption of radioactivity from the bacterial biomass into the water column and a second phase during which the radioactivity level of the water column increased slowly. The parameters a and b depended on the contamination rate of the water column and allow to compare the biological half-lives of the bacterial biomass, Tbl and Th2 respectively. They were calculated as follow: Tbl = I n 2 l a Tb2 = In 21 b. In the case of 60Co,the longest Tbl and Tb2 were found at pH 8, the average pH of the Meuse river (Figures 9 and 10). For the other pH values of the water column, Tbl and Tb2 were shorter which means a more rapid desorption of the 6oCo by the bacterial biomass and consequently an increasin radioactivity level in the water column. The same observations were made in the case of f34Cs. Given that the mean temperature in the Meuse river is 13'C and pH averages 8 , bacteria are likely to act as radionuclide sinks for much of the time but, when the temperature increases in summer they could act as sources.

6.5

7.0

7.5

8.0

t

8.5

a v e r a g e Meuse value

9.0 PH

Figure 9. Biological half-lives Tbl (h) of the 6oCo desorption of the bacterial community vs the water column pH.

35 I

Figure 10. Biological half-lives Tb2 (h) of the "Co desorption of the bacterial community vs the water column pH values.

CONCLUSlON AND SUMMARY.

The radionuclide fluxes between the bacteria and the water in aquatic ecosystem were studied by examining the bulk transfers mediated by a bacterial community isolated From the river sediments. The comparison of the aerobic bacterial communities colonizing the sediments and the water column shows that the bacterial community of the sediments is composed of two sub-communities. The fist one is similar to the water column community by its biochemical feahires; the other one displays quite different characteristics and appears to be more representative of the sediments.

352

An important part of 6oCo and 134Cs can be immobilized by the bacterial biomass which therefore constitutes a pool of radionuclides, their transfers to the water column being controlled by tern erature and pH. The uptake of 68Co and 134Cs by bacteria can be described by the Michaelis Menten model. The uptake kinetics depend on the type of radionuclide and the level of radiocontarnination in the water column. The highest affinity uptake system is observed for 6oCo at low radiocontarnination levels. The decontamination of bacterial biomass develops in two phases. The first phase is characterized by a very short biological half-life, a few seconds or minutes while the second phase is longer, the biological half-lives reach between 15 h to 461 h for 6oCo and between 39 h to 8,976 h for 134Cs. It could be inferred from the study of the influence of temperature and pH that in the river Meuse where temperature averages 13’C and H is around 8, the rates of radionuclide releases are very low so that 6oCo and lY4Cs are effectively trapped by bacterial biomass. When the conditions change i.e. the temperature raises, the radionuclides immobilized by the bacterial biomass are released and radioactivity consequently increases in the water column. It appears thus that the environmental parameters play an important role in the radionuclide transfers mediated by bacteria and are to be more deeply investigated in the further.

REFERENCES

De Clercq-Versele H, Kirchmann R. L’impact des rejets de la centrale nuclkaire de Tihange (Belgique) sur l’kcosysteme Meuse: cinq annCes d’ktude in situ et d’approche expdrimentale (1976-1980). BLG 555, 1982. KirchmaM R. L’irnpact des rejets de la central nuclCaire de Tihange (Belgique) sur I’Ccosystkme Meuse: dtudes in situ et recherches expkrimentales durant la pCriode 1981-1984. BLG 573, 1985. Kirchrnann R, Vandecasteele CM, Foulquier L, Lambinon J., Sombre L. La radiokcologie des grands fleuves: des donnCes de sites et de l’expkrirnentation B la moddlisation (application la Meuse et au RhOne). BLG 635, 1992. Crerners A. Modelling the transport of radionuclides through the freshwater environment. CCE Project Bi7-008, Final Report, 1992. Rernacle J, Vercheval C. Can J Microbiol 1991: 37: 875-877. Remacle J, Muguruza I, Fransolet M. Wat Res 1992: 26: 923-926. Billen G, Servais P. in Bianchi M, Marty D, Bemand JC, Caumette P, Gauthier M eds Micro-organismes dans les Ccosystkmes ocdaniques. Paris: Masson, 1989; 2 19-245. Walker SG, Flemming CA, Ferris FG, Beveridge TJ, Bailey GW. Appl Environ Microb 1989: 55: 2976-2984.

353

9 10 11 12

13 14 15 16

17 18 19

Flemming CA, Ferris FG, Bevendge TJ, Bailey GW. Appl Environ Microb 1990: 56: 3191-3203, Servais P, Billen G, Hascoet MC. Wat Res 1987: 21 : 445-450. Azarn F, Fenchel T, Field JG, Gray JS, Meyer-Reil LA, Thingstad F. Mar Ecol Progr Ser 1983: 10: 257-263. Servais P. Etude de la degradation de la matikre organique par les bacteries hiterotrophes en rivkre. DCveloppement d’une demarche methodologique et application a la Meuse belge. Thkse de Doctorat. Universitk Libre de Bruxelles, 1986. Daurnas R, Bianchi M. Arch Hydrobiol Beih Ergebn Limnol 1984: 19: 289-294. Sokal RR, Michener CD. Univ Kansas Sci Bull 1958: 38: 1409-1458. SAS. User’s Guide: Statistics. Cary NC: Sas Institute Inc., 1985. Le Minor L, Viron M. Bacteriologie medicale. Paris: Flarnarion MedecineSciences, 1982. Ralston ML, Jennrich RI. Technometrics 1979: 1: 7-14 Nucho R, Baudin JP. Sci Eau 1986: 5: 361-376. Sombre’ L. Contribution h 1’Ctude du transfert du radiocesium (Cs- 134 et Cs- 137) dans une chaine alimentaire d’eau douce simplifiee: eau - algue verte (Scenedesmus obtiquw) - mollusque filtreur (Dreissena polymorpha) - Poisson (Barbus harhus). These de Doctorat. Universitk de Provence (Aix-Marseille I), 1987.

This Page Intentionally Left Blank

355

Effects of plants on the accumulation of Zn, Pb, Cu and Cd in sediments of the Tagus estuary salt marshes, Portugal. Isabel Caqador", Carlos Valeb and Fernando Catarino" Departamento de Biologia Vegetal, Universidade de Lisboa; Campo Grande, Bloco C2, 1700 Lisboa, Portugal

a

Instituto Nacional de InvestigaqSo das Pescas, Av. Brasilia 1900 Lisboa

ABSTRACT Vertical distribution of Zn, Pb, Cu and Cd concentrations, organic matter content, pH and Eh in sediment between roots and in intertidal non-vegetated sediments from two sites on the Tagus salt marshes are compared. A sub-surface metal enrichment was observed a t depths where roots are abundant. Increase of the non-available metals to plants in the salt marsh sediments indicates different sink capacities for metals in the Tagus salt marshes: Pb >> Zn > Cu > Cd.

INTRODUCTION Several studies have been carried out in wetlands suggesting these areas may act as natural sinks for metals (1,2). The recognition of the importance of this action led recently to the construction of artificial wetlands in order t o concentrate metals from industrial waste water and mine seepage (3). In fact, many salt marshes lie close to heavily industrialized areas in estuaries, facing important discharges of metals and other toxic wastes (4).Transported by the tidal currents, both in dissolved and particulate forms, metals are eventually incorporated on the surface sediment in salt marshes (5). Vegetation may act as sediment traps (6), playing thus an important role in the settling of suspended estuarine material arid their associated mctals. Available metals may be uptaken through the roots and subsequently incorporated in the plant tissues (7). Litter plant material enriched in heavy metals may return to the sediments. Incorporated in several organic and inorganic forms in sediments, metals may not be readily available for uptake by plant roots (8). The complexing reactions involving natural organic matter in the sediments may play a key role in controlling the metal bioavailability (9). Beyond organic matter input, emergent vegetation appears to have an important role in the metal cycling as root-induced changes in the rhizosphere may modify substantially potential redox, pH and root-associated microorganisms (7). For example, reddish-brown deposits on plant roots (10) and rhizoconcretion formation around the roots (11)on several salt mnrshtls and have t w n att,iibuted to the oxidative capacity of the roots. In this paper we report the concentrations of Zn, Pb, Cu and Cd in intertidal sediments and in salt marsh sediments of the Tagus estuary and assess the role of vascular plants on the metal enrichment in the sediment sub-surface.

356

THE TAGUS ESTUARY The Tagus estuary is one of the largest estuary on the Atlantic coast of Europe, covering an area of 300 km2a t low tide and 340 km2a t extreme high tide. The estuary is composed of a deep, straight and narrow inlet channel and a broad, shallow inner bay (Figure 1).The southern and eastern parts of the bay contain extensive inter-tidal mud-flat areas. A striking feature of these areas is the presence of halophytic vegetation. The Tagus river is the principal source of freshwater to the estuary, the average annual discharge being 400 m3.s1.The tides are semi-diurnal, with amplitudes that range from 1m at neap tide to 4 m a t spring tide and its effect reaches 80 km landward of Lisbon. Contrary to many european rivers crossing industrial regions, most pollutants are discharged directly into the estuary. The Tagus estuary receives the inflow of effluents from about 2.5 million inhabitants living in the Great Lisbon area, together with the effluents from its industries (chemicals, steelmaking and shipbuilding).

I

. 5 Krn

Figure 1. The Tagus estuary; location of the sampling sites.

357

MATERIALS AND METHODS Sediments were sampled in May 1991 from two sites on the Tagus estuary salt marsh with different degrees of metal contamination (Figure 1):Corroios (site 1)and Rosario (site 2). At both sites, three 60-cm long sediment cores were collected from intertidal non-vegets,cd zones and one from each of the areas on the salt marsh dominated by Halimione portulacoides, Spartina maritima and Arthrocnemum fruticosum. Sediment cores were sliced in situ a t the following depth layers: 0-5; 5-15; 15-25; 25-35; 35-45 and 45-55 cm. Redox potential (Eh) and pH of the sediment layers were measured using a Crison pWmV meter. The pH was measured after a short period of stabilization and the E h electrode was allowed to reach the equilibrium for 15 minutes. Both pH a n d E h measurements a t each layer were repeated four times. I n the laboratory, sediment samples were dried, roots removed and homogenized. The organic matter content (loss of ignition a t 600 'C for 2 hours, LOI), grain size distribution (by wet sieving) and concentrations of Zn, Cu, Cd and P b were determined. After homogenizing, sediments were extracted with the following solutions: 2.5 % acetic acid, 0.005 M DTPA i n CaC1, and 0.1 M triethanolamine a t pH=7.3, 1 M HNO, (9) and HNO,/HCl (12). For the acetic acid, DTPA and HNO, extractants, 10 nil of solution were added to 2 g dry sediment in polypropylene bottles and shaken on a rotating shaker for two hours a t room temperature. The suspension was then filtered. For the HNO,/HCl extraction 10 mL of HNO,/HC1(3:1 v/v) was added to 2 g of dry soil twice a t 130 degrees Centigrade (12). Heavy metal contents were determined by atomic absorption spectrometry. The strong HNO,/HCl digestion was considered to remove the total amount of Zn, Pb, Cu and Cd in the sediment samples. At both sites on the salt marsh, eight plants of each species were collected and transported to the laboratory in plastic bags. The plant roots were washed with demineralized water, diied and homogenized. For Zn, Cd, Pb a n d Cu analysis 10 ml HNOJHC10, were added to 100 mg of plant material according to the methodology described in (12). Standard additions and sludge plus vegetal reference materials were used for sediment and plant analysis, respectively.

RESULTS AND DISCUSSION All sediment samples were made up of silt and clay, containing less than 3 Redox potential and pH measurements were quite 1% sand (fraction ~ 6 um). variable. E h values recorded in sediment layers from non-vegetated sites and below roots were always negative, while in vegetated sediment they varied within a broader interval (from -290 to +210 mV). Positive values were recorded in upper sediment layers where larger amounts of roots are present. Values of pH were always below 7.0 and changed from 5.7 to 6.4 i n the upper 25 cm. E h a n d pH varied inversely (Figure 2) indicating that the rhizosphere is a more oxidative and acid environment than sediments where plants are absent or in layers below the root depth.

358

+++

0

T

:* + ** *** *#f

++

-100-

+

-200

-

t

-900

-

*t

-400 6.2

6.8

f* :#

*

8.0

8.4

0.8

1 2

PH

Figure 2 - Redox potential (mV) versus pH in sediment between roots (+) and in non-vegetated sediment (including sediment below roots) (*) a t sites 1 and 2 on the Tagus salt marsh.

Lot (%) 30 26

+

-

20 -

f

+ + + 4wt+

*

1610 -

6.2

6.6

6.0

6.4

6.8

7.2

PH

Figure 3 - Organic matter content (LOI, %) versus pH in sediment between roots (+) and in non-vegetated sediment (including sediment below roots) (*) a t sites 1 and 2 on the Tagus salt marshes.

359

Organic matter content (LOU was higher in Sediments colonised by plants: in non-vegetated sediment LO1 values remained below 15%, while in vegetated areas they varied from 12% to 30%. However, in the upper 25 cm where roots are present, displayed the highest organic matter content and the lowest pH, as shown by the LOI=pH relationship presented in Figure 3. This indicate8 the organic matter enrichment of' the salt marsh sediments as compared to the intertidal sediments. Vertical distributions of mean concentrations of total Zn, Pb, Cu and Cd, with standard deviations in sediments from sites 1 and 2 are shown in Figures 4. Concentration profiles in the vegetated sediment (x> were plotted separately from the non-vegetated sediment (0).In sediments from the non-vegetated areas a progressive decrease of metal concentrations with depth was observed in the upper 20 cm. Below that depth, metal levels remained rather constant. Metal enrichment has been observed in the surface sediments of the Tagus estuary and attributed to environmental contamination (13). Concentrations of Zn, Pb, Cu and Cd in the surface sediments a t site 2 were higher than at site 1, possibly due t o the proximity of metal pollution sources. Standard deviations obtained for the non-vegetated sediment samples were low, indicating that metal concentrations were comparatively uniform at each site. Profiles of total metal concentrations in the salt marsh sediments were quite different: sub-surface peak values were found and, in general, salt marsh sediments contained more metals than intertidal sediments (compare 0 and x profiles, in Figure 4). In particular, Zn concentration went up to 1150 ug.g-1 and Pb to 500 ug.g-1 a t site 2, while levels in intertidal surface sediments reached only 730 ug.g-1 and 200 ug.g-1, respectively. The increase of metal concentrations, and principally the sub-surface peak values, appear to be related to the presence and the activities of plant roots. Previous work on the Tagus salt marshes (14) have proved that roots are present in higher quantities in the sediments a t these depths. Standard deviation was found to be relatively high between 5 and 15 cm, suggesting that sub-surface enrichment does not have the same intensity all over the salt marsh. The percentages of metals removed by weak extractions (acetic acid, nitric acid and DTPA) in relation to the total concentrations vaiied with depth and commonly explained by the post-depositional processes. Ratios for Zn, Pb, Cu and Cd recorded in the 5-15 cm depth layer of non-vegetated sediments are compared values found a t the same depth with roots (Table 1). In sediments colonised by plants the percentages of metals removed by those solutions were smaller than in non-vegetated sediments. In spite of the high total concentrations of metals found at sediment between roots, the amount of metals removed is small, which means that the larger fraction is non-available for plants.

360

80

1

0

Po0

400

800

800

1000

I

1200

Zn (use-1)

v,

90

Pb

40

60

80 0 80

100

200

300

400

600rite 1600

0

100

200

300

400

600

I

Pb (ug.g-1)

I

600

so I 0

dtu 2 100

200

300

400

600

I

800

Pb (ug.g-1)

Figure 4 - Vegetal profiles of total Zn, Pb, Cu and Cd concentrations (ugg') in sediment between roots (x) and in intertidal sediment ( 0 )from sites 1 and 2; mean concentrations (full line) and standard deviation (dashed line).

36 1 0

fT

w

60-9

00

alb 2 80

Cd

Figure 4 - (cont)

-Dk

site 1

Cd

362

Table 1 - Percentages of Zn, Pb, Cu and Cd recovered with 2.5 % acetic acid, 1 M HNO,, and DTPA in relation to total concentrations in the 5-15 cm sediment depth layers of vegetated and non-vegetated areas of site 1. ~

~

Samples

Zn

Pb

cu

Cd

63 43 41

12 70 73

14 37 42

36 40 60

18 4 3

1 4 3

2 21 11

1 26 8

non-vegetated sediment HAdtotal HN03/total DTPNto tal vegetated sediment HAdtotal HN03/total DTPMtotal

'

Extraction with the strong chelating agent DTPA estimates metals bound to the organic matter, and the use of 1 M HNO, measures metal cations incorporated into carbonates, phosphates and partially organic matter. These forms may all be available to plant roots (9). The reduction of the percentages of those forms at sediment depths where plant roots are abundant and the observed increase of the forms easily exchangeable (acetic acid extraction) may result from several interactions between the plant and the sediment environment (7,9). Furthermore, metals are uptaken by the plants (4). Metal concentrations in the roots of the three most abundant vascular plants are presented in Table 2. Residues of Cd and Cu were lower than those of Zn and Pb. However, the amount of metals incorporated in the roots vaiied substantially from plant to plant. Table 2 - Concentrations of Pb, Cd, Cu and Zn (ug.g-1) in roots of Spartina maritima, Arthrocnemum fruticosum and Halimioneportulacoidesfrom the Tagus salt marshes. Areas

Plants

Zn

cu

Pb

Cd

Site 1

Spartina Arthrocnemum Halimione Spa rtina Arth rocnemu m Halimione

478 420 799 432 346 644

53 10 101 37 31 171

282 60 422 180 90 840

3 <1 4 2 1 3

Site 2

363

The quantities of Zn, Pb, Cu and Cd which were not removed either by the DTPA method or by the 1 M HNO, solution are normally considered the forms non-available to the plants (9). The fraction composed by those forms was calculated by subtracting the concentration of metals weakly bound to the sediment (1 M HNO, extraction) from total metal concentration. The residual fractions of Zn, Pb, Cu and Cd were calculated for each layer of vegetated and non-vegetated sediment cores from sites 1 and 2. In order t o minimize differences on metal contamination between sites 1 and 2 values were normalized. For each metal and all the sediment depths considered, the residual fraction (RF) of the vegetated sediment (sites 1 and 2) was divided by the residual fraction found a t the same depth in the intertidal sediments from these sites. The ratio is the enrichment factor (EF): EF = [RFI vegetokd Bedtment 1 [RFI non-vcgctnlcd scd,ment Values of EF higher than 1 mean that the corresponding salt marsh sediment layer contains higher residual fraction of the metal analyzed when compared to the sediments at the same depth from the non-vegetated intertidal areas nearby. Therefore, metals are bound in stronger ways to the sediments in the vegetated areas. Enrichment factors of Zn, Pb, Cu and Cd were calculated for the depth layers in the sediment cores from sites 1 and 2 (Table 3). Table 3 - Enrichment factors of Zn, Cu, Cd and Pb for the sediment cores from sites 1 and 2. Enrichment factor (EF) Site/depth (cm)

Cd

cu

Zn

Pb

0.3 2.3 0.5 0.7 0.5

0.1 2.3 0.7 0.5

1.0 4.1 2.1 2.5 1.0

0.1 34 51 6.0 0.1

1.7 1.9 0.3 0.8 0.9

0.8 2.4 6.2 5.2 0.9

0.2 5.4 1.3 0.5 1.3

2.9 27 7.7 41 0.1

Site 1 0-5 5-15 15-25 25-35 35-45

1.1

Site 2 0-5 5-15 15-25 25-35 35-45

364

Values of EF presented in this table evidence a sub-surface maximum for all metals. At the surface layer and a t deeper ones, E F values are close or less than 1, whereas below surface are always higher than 1. This means that the fraction of metals strongly bound to the sediment is higher at depths where large amount of roots are present (5-15 em). Otherwise, residual fractions observed in vegetated and non-vegetated sediments at other depths were similar. If we assume that EF gives a measurement of the retention capacity of the salt marsh sediments in relation to the intertidal sediments in nearby sediments, the zone where roots are more abundant is then acting more efficiently as a metal trap. This seems to be valid since the residual fractions are the metal forms not available to the plants, and plants in salt marshes are the dominant bioturbation factor of the sediments. Furthermore metals displayed distinct enrichment factors a t these depths. While EF was 2.3 for Cu and Cd for Zn it reached 4.1, and for Pb went up to 51 at site 1, that means one order of magnitude higher. Such a difference on the enrichment factor for Pb, indicates that the Tagus salt marshes have different sink capacities for the metals studied Pb >> Zn > Cu> Cd.

REFERENCES 1 2

3

4

5 6

7 8

9 10 11

12 13 14

Oenema 0, Steneker R, Reynders J . Hydrobiological Bull. 1988; 22: 21-30. Orson RA, SimpsonR L, Good RE. Estuar Coast Shelf Sci 1992; 34: 171186. Dunbabin JS, Bowmer KH. The Science of the Total Environment, 1992; 111; 151-168. Otte ML. In: Rozema J, Verkleij AC, eds. Ecological Responses t o Environmental Stresses. the Netherlands: Muwer Academic Publishers, 1991; 126-133. Bourg ACM. In: Salomons W,Forstner H, eds. Chemistry and Biology of Solid Waste, Dredge material and Mine Tailings. the Netherlands: Springer Verlag, 1987;3-32. Chenhall BE, Yassini I, Jones BG. The Science of the Total Environment 1992; 125: 203-225. Rozema J, Otte ML, Broekman R, Kamber G, Punte H. In Grank AJ, Benham PEM, eds. Spartina anglica - a research rewiew. London: HMSO, 1990; 63-68. Weimin Y, Batley GE, Ahsanullah M. The Science of the Total Environment 1992; 125: 67-84. Piccolo A. The Science of the Total Environment 1989; 81/82: 607-614. Otte ML, Rozema J, Koster L, Haarsma MS, Broekman RA. New Phytol. 1989; 111: 309-317. Vale C, Catarino FM, Cortesgo C, Caqador MI. The Science of the Total Environment 1990; 97/98: 617-626. Otte, ML. Amsterdam: Free Univ. 1991; 188p. Vale C. The Science of the Total Environment 1990; 97/98: 137-154. Broteriana 1981; 54: 387-403. Catarino FM, Caqador MI. Boletim da SOC.

365

Studies on hea metals of eriph ton and its host plant /Phragmites australis (Cav.) )iyrien ex Steu&l/ in &allow lakes G. Lakatos Department of Ecology, Lajos Kossuth University, H-4010 Debrecen, Hungary

INTRODUCTION

In Europe the most important plant communities in the littoral zone of lakes are the reed communities (Scirpo- Phragmitetuni) which play multifold role in point of view of nature conservation and environmental protection. They have significant role especially in the management of water quality. The effectivity of reed belt in improvement of water quality is enhanced by the periphyton (epiphytic biotecton) formed on the under-water parts of the reed. Through the habitat pecularities reed community offers appropriate biotop for many organisms. Although the significance of reed belt and its periphyton has been recognized, only some papers [l-61 deal with it in detail and with adequate deepness, especially in Hungary and in Austria (7-121. The phenomenon of reed decline was first observed more than 20 years ago in Hungary, the serious damages have been reported since the 80ies [ 13-15] and mostly from our large lakes (Lake Velencei and Lake Balaton). As the accelerated and large-scale reed decline appears as a great problem of environmental protection and nature conservation all over the Europe many research institutions and scientists [18-191 make efforts on revealing its main causes. The main aim of this paper are the followings: To corn are the heavy metals content of periphyton and its host plant in different shallow akes. To demonstrate the spatial changes in the heavy metals contents of periphyton in the reed stands of the same lakes. To estimate the possible reasons of common reed decline in eutrophic shallow lakes. To reveal the correlation among the contents of heavy metals in the periphyton and young reed, applying regression and path analysis.

P

MATERIAL AND METHODS

In the littoral zone of Lake Balaton the periphyton-reed complex has been studied in four transects in the bays at Paloznak, Bozsa, Szigliget and Keszthely for four years (Fig.1.).

366

Hungary

SZ

-

bay Szigliget

Figure 1. Sampling sites on the northern part of Lake Balaton. The reed stand in Paloznak and Bozsa bays can be considered healthy and undisturbed stands of the lake. The reed stands in the Bays at Keszthely and Szigliget are much richer in nutrient elements and heavy metals than in the other areas of the lake. In Bay Keszthely there has occured extensive eutrophication which offerred the reed stands negatively and the accelerated die-back of reed took place /lS/. The reed-periphyton samples were taken near the open water in every transect. In case of Lake Velencei 14 sampling points were selected at the open water side of reed stands in regions differing in water quality (Fig.2.). Unfortunately the reed-decline phenomenon can also be observed in this shoreline of this lake /20/.

367

Figure 2. The study area and sampling sites of Lake Velencei.

Lake Fertij (Neusiedlersee) is the second largest eutrophic shallow lake in Central Europe. The larger part of the lake belongs to Austria and Hungary has the smaller part of it. During the study of this lake reed-periphyton samples were taken in both Hungarian and Austrian parts of the lake. 34 sampling plots were selected on the eastern, northern, western and southern parts of the lake altogether (Fig.3.).Figure 4 shows the sketch of reed /Phragrnites/ and the samples which were analyzed.

8 sediment

Figure 3. Sampling sites of Lake Fertij /Neusiedlersee/

Figure 4. Sketch of reed and the samples which were analyzed.

368

Simultaneously with the reed and periphyton sampling, water samples were also taken for investigation. The wet mass of periphyton was estimated and after drying it at 105 Co, the dry mass was measured. The ash contents of reed and its periphyton samples were also determined. The concentration of 27 cations was measured from the periphyton, reed and water samples by ICP method. Concentration factors were also calculated. Linear regression analysis was used to reveal the relationships between the measured parameters. The direct and the indirect effects of the individual effective factors (independent variables) on the variance of the investigated parameters were estimated by means of path analysis.

RESULTS AND DISCUSSION

O n the basis of the ash content of periphyton samples from Lake Velencei and Ferto are grouped into the same category and can be described by the dominance of inorganic components. At the same time samples collected in Lake Balaton belong to the other category [21]. The ash content of green reed samples are smaller, than that of the old reed ones, only the green reed samples of Lake Ferto exceed the 10 percent. The lowest iron concentration was measured in the periphyton and green reed samples of Lake Balaton, and the highest one in case of Lake Velencei /Table 1./. In contrast to it the manganese content of the samples collected in Lake Balaton was about threefold higher than in the samples of the other lakes. Consequently the Fe/Mn ratio differed considerably in the periphyton of three lakes /Lake Balaton 4; Lake Ferto 20; Lake Velencei 391. Table 1 Heavy metal content of periphyton formed on green reeds in three shallow lakes (%) Ash

Fe

Mn

Zn

cu

42.6 (6.5) 0.71 (0.22)

0.178 (0.138) 0.066 (0.048) 0.024 (0.022)

Lake Velencei 69.0 (8.3) 1.85 (0.42)

0.047 (0.022) 0.028 (0.019) 0.023 (0.019)

Lake Ferto

0.049 (0.034) 0.01 1 (0.010) 0.005 (0.003)

Lake Balaton

71.2 (8.4) 1.01 (0.29)

Similarly to the manganese the highest concentration of zinc also occurred in case of Lake Balaton, which indicated twofold accumulation in comparison with the green reed samples and it is sixtimes higher than in the periphyton samples collected in Lake Ferto. Differences in the heavy metal content of periphyton appeared not only among the three lakes but among the different water regions of a given lake too. T h e heavy metal content of the periphyton was especially high in the water areas that were impacted by a large anthropogenic load or were situated in the neighborhood of sewage effluent. Considering Lake Velencei as an example Table 2 illustrates results to compare the heavy metal concentration of periphyton in the brown water region of the lake /southern part/ which is still in natural state and in the green water area /northern part/ where considerable planktonic eutrophication takes place due to the outer loading of plant nutrient and other inorganic and organic pollutants [22-231.

369

Table 2 Heavy metal content of periphyton formed on green reeds in two water regions of Lake Velencei (%) Mn

Zn

cu

Brown water region 1.66 (0.16)

0.039 (0.019)

0.023 (0.012)

0.018 (0.014)

Green water region 2.24 (0.26)

0.056 (0.003)

0.036 (0.010)

0.032 (0.028)

Fe

In case of Austrian part of Lake Ferto the periphyton samples collected in the southern part of the lake contain Mg in high concentration. The concentration of Mn, Cr and Pb was high in the western part of this lake which is more intensively used for recreation. These cations indicate the human impacts or anthropogenic loads for the water quality of lakes /Fig 51. The values of concentration factors which express the ratio give good evidence for the filtering, accumulating and eliminating role of the reed periphyton.

water

( IWI I ) Pb (pjm)

li (pPrn)

r

03

02

r----

-

0 , -

.... ~.

Figure 5. Concentration of Cr and Pb in water and periphyton. To reveal the ossible reasons of reed decline healthy reed and diseased reed stems were analyzed. d e samples were taken from the same time and site of the studied lakes. Consequent differences can not be stated for the concentration of macro cations, nitrogen, phosphorus and some heavy metals /Fe, Mn, Cd, Pb, e t d , but there are some

3 70

elements which show significant differences between the healthy and diseased reed stems. The date are in percentage of the healthy reed stems /Table 3/. Table 3 The percentage changes of element content of reed stems due to the disease

ash

Zn

cu

Mo

S

%

Lake Balaton

117.6

189.2

202.0

27.3

81.2

Lake Velencei

131.4

217.0

303.5

26.3

68.4

Lake Ferto

105.2

151.6

225.8

28.2

71.1

The results of stud of element content of reed did not help us to much in recognizing the direct causes ofYthe decline of reed community, but the higher concentration of ash, Zn and Cu and lower content of Mo and sulphur of the dieseased reed stems are interesting and warning. In case of iron and zinc positive correlation can be found between of the green reed and its periphyton /Fig.6./. For the periphyton the analysis showed positive correlation between the individual heavy metals and the magnesium which underlines the importance of the absorption processes.

BOZSA I bay p i p h y t on

Fe% 2,o

i -

J

o o

iron

contents of reed

I

Figure 6. Relationship between iron content of young reed and its periphyton

37 I

In case of zinc the variance of the concentration in the periphyton was directly determined by the concentrations in the green reed beside the significant influence of other factors lFig.7.1

Diagram 04 'Pa+h analyris ( Z L ) I

I

' ""

\young

J

+ I reed

"I'

' 0

1

Figure 7. Diagram of path analysis /Zn/ CONCLUSION

The most important plant communities in the littoral zone of shallow lakes are the reed communities. The effectivity of reed belt in improvement of water quality is enhanced by the periphyton. Differences in the heavy metal content of periphyton and its host plant appeared among the shallow lakes. The heavy metal content of periphyton was especially high where impacted by large anthropogenic load. On the basis of our results the biofilter role of periphyton-reed complex can be confirmed. Since the main reasons of reed decline have not been yet revealed and therefore the future research has to focus on studying the role of heavy metals and sulphur. ACKNOWLEDGEMENTS

This study was made within the framework of the Project No 1881 financed by OTKA (National Scientific Research Fund) and partly with the support from CEU (Central European University) which are greatly acknowledged. We are thankful to the Balaton Limnological Research Institute in Tihany and Kossuth University in Debrecen for providing us with good working facilities at their laboratories.

372

REFERENCES 1 2 3 4 5 6 7 8 9 10 11 12

13 14 15 16 17 18 19 20 21 22 23

Kornhrkovh J, Kornhrek J. Symp. Biol. Hung. 1975; 15: 77-95. Kowalczewski A. Ekol. Pol. 1975; 23: 509-543. Lakatos G, Voros L, Entz B. BFB-Bericht 1982; 43: 40-61. Meulemans JT. Arch. Hydrobiol. 1988; 112: 21-42. Szczepanska W. Pol. Arch. Hydrobiol. 1970; 17: 397-418. Wetzel RG. Oikos 1960; 11: 223-236. Buczk6 K. Studia Bot. Hung. 1986; 19: 63-71. Dokulil M. Verh. Internat. Verein. Limnol. 1975; 19:1295-1304. Lakatos G. BFB-Bericht 1989; 71: 125-134. Padishk J. BFB-Bericht 1982; 43: 95-105. Schiemer F, Prosser M. Aquat. Bot. 1976; 2: 289-307. Sommer U. Osterr. Akad. Wiss. Math.-naturw. Abt. I. 1977; 186: 219-246. Dinka M. Folia Geobot. Phytotaxon. 1986; 21: 65-84. Lakatos G. Proc. Int. Symp. Aquat. Macrophytes Nijmegen, 1983; 117-122. K o v h M. et al. Symp. Biol. Hung. 1989; 38: 461-471. Haslem SM. Ann. Bot. 1970; 34: 573-591. Dykyjova D. et al. Photosynthetica 1970; 4: 280-287. Den Hartog C . Aquat. Bot. 1989; 35: 1-4. Ostendorp W. Aquat. Bot. 1989; 35: 5-26. Lakatos G. BFB-Bericht 1989; 71: 157-164. Lakatos G, Bir6 P. BFB-Bericht 1991; 77: 157-164. Bartha Zs. Acta Bot. Acad. Sci. Hung. 1977; 23: 1-11. Lakatos G. Acta Biol. Debrecina 1978; 15: 147-168.

373

Origin and pathways of Cadmium Contamination in the Gironde estuary, Garonne river and tributaries. J.M. Jouanneau, Y. Lapaquellerie and C. Latouche Departement de Geologie et Oceanographie Universite de Bordeaux I 351, Cours de la Liberation 33405 Talence, France

INTRODUCTION The Gironde estuary, SW of France, (fig.1) is being to become a model for metallic pollution, due to industrial activities located far away, 300 km upstream of the estuary itself. Nevertheless, for a long time, the Gironde estuary had been considered as one of the rare example of an undisturbed estuary in the highly industrialized temperate climatic zone of the northern hemisphere. But, in 1979 "R.N.0."- the "Reseau National d'observation de la Qualite du Mlilieu Marin" (national network in charge of monitoring the quality of the marine environment) made known that there were high metallic element contents, particulary Zinc and Cadmium, in oysters of the lower part of the Estuary. These oysters, which developped mostly from natural beds, are not commercialized but only used for the spawn production which is later transplanted in the oyster farming areas. Given the very high rate of toxicity of cadmium - responsible for the "ltai - Itai" disease (in Nriagu, 1980) a plan was immediatly set out to study the distribution of this element in molluscs, water and sediments of the Gironde Estuary and invarious embayments of the southern part of Biscay coast. Following the different works carried out within the framework of the University of Bordeaux, (1-2-3-4-5), or research contracts (IFREMER, Ministry of the Environment 1984, 1985), IGBA studied the environmental distribution of cadmium in suspended matter and sediments within the estuary as well as within the adjacent marine zones : Ouest-Gironde and Marennes-Oleron Bay. The surveys were carried out either individually or in collaboration with national organizations in charge of the natural environments such as IFREMER (estuary and coastal zone) or the "Agence de I'Eau Adour-Garonne" (rivers). Hydrological and sedimentological characteristics of the Gironde estuary have been already discussed at great length (6). However, a brief review is given only of essential data for the interpretation of the results presented here. Below the junction between the Garonne and Dordogne rivers at the Bec d'Amb&s (fig.1) this estuary covers at high tide an area of 625 Km2. Total average river discharge of both rivers is about 766 m3s-1. The inlet tidal prism (marine water volume introduced into the estuary by current flow) is estimated at 2.109m3 for spring tide and at I .I .109m3 for neap tide. It decreases exponentially towards the upper estuary. The lower part of the 2 rivers which is subject to the dynamic tide but not to the saline one is considered as a fluvio-estuarine zone (Garonne upstream limit : La Reole ; Dordogne upstream limit : Pessac). In this zone, the limit of saline intrusion (salinity range 2 0.5 %o) reaches Bordeaux at low water level periods.

314

As in most estuaries there exists in the Gironde Estuary a residual circulation system, which is a major cause of the turbidity maximum (water masses loaded with suspended sediment : 0.1 to 10 9.1-1) and fluid mud (very turbid water lenses 100 to 300 g.1-l). The total suspended sediment reaches approximately 4.2 to 5.3.1 0%

'ertuis breton

Bellevue srtuis de Maurnusson

F'

France

Spain 0

f 20 km

Gironde estuary

Arcachon Bassin

Figure 1. Map of Gironde estuary and neighbouring nearshore Atlantic area.

375

Both turbidity maximum (1 .I to 2.0.106 t) and fluid mud (2.2 to 2 106 t) are subject to fluctuations, cyclical displacements (associated with tides) and especially seasonal effects (associated with river discharge). During flooding, turbidity maximum and fluid mud may go further than 80 km downstream Bordeaux, whereas at low water level they go as far as Cadillac (approximately 50 km upstream from Bordeaux). Further upstream, suspended matter becomes less important (about 10 mg.1-1) except during flood events (more than 100 mg.1-I). Sedimentological studies have shown that for suspended sediments ( 7 ) ,the downstream estuary supply by marine mineral material is minimal. Evidence of marine sand bed load transport has been reported, but only close to the estuary inlet.

SETTLEMENT AND CHARACTERISTICS GIRONDE ESTUARY AND TRIBUTARIES.

OF

CD

POLLUTION

IN

Estuarine fauna contamination The comparative studies made throughout specimen of oysters fields of the French coast between 1979 and 1988 (RNO, 1988) showed the exceptionally high level of Cd content in the Gironde oysters (20 to 1OOpg.g -1). Comparisons, for instance, with the Britanny and Arcachon oysters revealed that the Gironde oysters contained 10 times more Cd. Results obtained also showed that Cd content was a bit higher in the oysters of the Marennes-Oleron Bay than in the Arcachon oysters (fig.2).

Figure 2. Cd contents in oysters from the French coastal zone (from Boutier, 1987) Cd content in various components of the specific estuary fauna (shrimps, fishes) are compared (tabl. 1 ) with the mean value measured in great number of oyster specimens.

316

Table. 1 Cadmium in Gironde estuary animals

n

Mean contents dry weight pg.g-1

Crassostrea gigas

40

48.9

Crangon crangon

7

0.33

Pleuronectes platessa (muscle)

11

0,lO

Solea vulgaris (m uscle)

11

0,02

Platichthys flesus (muscle)

2

0.02

Limanda limanda (foie)

1

4,OO

SPECIES

Engraulis encrassicolus (total animal) 1 0,Ol R.N.O., 1979 B 1982 - n = number of measurements - data from Dumas, 1' 85. Mean results obtained from a great number of samples are given tabl. 1. It clearly appears that the oysters are, from far, the more affected by Cd pollution. A comparison (3) with several mean values measured in the same oyster species (Crassostrea gigas) all over the world (tabl. 2), shows, that the Cd contents of Gironde estuary oysters are the highest ones after those of the oysters of the Derwent Estuary (Australia) a well known very polluted area. Even in highly industrialized U.K. estuaries and harbours (Bristol Channel, Pool Harbour) oyster Cd contents are lower.

377

Table 2 Teneurs en Cadmium relevees chez Crassostrea gigas.

LOCATION

Cd conten

(pg.9-1dry weight)

Mean value

Extreme values

2,Oto 2,6

English Channe coast

REFERENCES

R.N.O. bull.

1979 to 1982 1,0a 116,5

Atlantic coast Gironde Estuary

48,9

until 116.5

Bristol Channel (R.U.)

23

4 to 43

Poole Harbour (R.U.)

4,6

BOYDEN et ROMERIL

I974 BOYDEN. 1975

4 to 6

THORTON, 1976

Saldanha Bay (South Africa)

9,o

WATLING et WATLING, (1 976 b.)

Knysna Estuary (South Africa)

3,7

WATLING et WATLING, (1976a.)

ramar Estuary (Australia)

33

22 to 45

AYLING, 1974

56,7

0 to 204

RATKOWSKY et al.,

Derwent Estuary and South coast of

1974 35. (a) : including Gironde estuary.

378

Cd in the estuarine environment

Regarding dissolved cadmium (fig. 3) the following distribution was observed : in the upper part of the fluvio-estuarine system, during low river discharge periods, either in Garonne or in the Dordogne branches of the estuary, the concentration are low (-20 ng.1-1) and similar to those usually found in fresh waters (8-9). Conversely, during high river discharge periods, the concentrations could reach 250 ng.1'' in the Garonne fluvio-estuarine part. Downstream, a large increase of dissolved cadmium concentration is always observed in relation with the turbidity maximum and for salinity range between 15 and 20 %o. In this part of the estuary the dissolved cadmium concentrations reach 400 ng.1-1, five times higher than normal (10). From the inlet of the estuary to the coastal waters, values decrease again from 50 ng.1-1 to 15 ng.1-1 that are normal values for coastal waters (1 1). Estuarine particulate Cd Compared to the mean world and to the natural regional background values (3), the Cd content in the surficial sediments of the lower part of Gironde Estuary (fig. 4) proved to be slighty higher than normal. By taking into consideration a reference value of 0,6 pg.g -1, it appears that the enrichment coefficient characterising the deposits of the lower part of the estuary reached a value slightly higher than 2. In contrast, Cd contents are quite normal both in the shelf sediments and in the deposits of Marennes-Oleron Bay.

Ia

C( 4 0 nq 4 0 (C

(

1-1

IOOnq I-'

1 3 c > loono

l

1-1

Figure 3. Dissolved Cd.

Figure 4. Cd contents in sediments.

319

The study of the sediment cores recovered in the oyster fields (4) allowed us to reconstruct the evolution in time of the Cd content in this part of the Gironde Estuary. For Le Verdon, La Coubre and above all Talmont sites, an important rise of Cd was reported in the deposits corresponding to about the last 30 years (fig. 5). Le V e r d o n

Ta Im on 1

Lo C o u b r e

Bellevue

Figure 5. Cd in cored sediments.

In the suspended matter S (J,l from the whole estuary (1) abnormal values were recorded in suspended matter of the lower part of the Garonne (La Reole) at the riverlestuary junction, especially during high water level periods when values attained 14 pg.g -1. Comparison of the Cd content in the suspended matter from La Reole and from dowstream to Bordeaux, showed that the wastes of the city, often thought to be the cause of pollution, cannot be considered as the main source of Cd for the Gironde estuary. Cd contents in sediments, SM and water of the drainage basin In order to precize the origin of Cd contamination of SM in the upper part of the estuary, surveys were carried out over the entire drainage basins of the two rivers flowing into the Gironde estuary, Garonne river and Dordogne river. Sampling of sediments in 70 selected stations allowed to carefully recognize the distribution of heavy metals (Cd, Zn, Cu, Ni) in bottom sediments of the various tributaries. Cd contents are presented fig. 6 . Bottom sediments from the Lot river -one of the main tributary- appear heavily contaminated by Cd. High level of several other elements (Zn, As, Ba, Cu) were also measured in the Lot fluvial deposits. The source of these high metallic concentrations was attributed to the drainage of the waste area of an

380

old industrial plant (Zn) in Decazeville, a town overflown by the Riou-Mort river, a small tributary of the Lot river located 250 km from the junction of Lot with Garonne rivers. Accidental metallic pollutions affecting Riou-Mort river have been often reported since 1979 (12). Sampling and analysis of water and SM confirmed the very high level of dissolved and particulate Cd and various other metals, in the Riou-Mort water. Conversely, lower -and often normal- dissolved and particulate Cd contents were measured in the other Garonne and Dordogne bottom sediment tributaries.

Figure 6. Particulate Cd content from the Lot drainage basin

PATHWAYS OF OYSTERS CD CONTAMINATION

In order to understand the series of processes responsible of the final oyster contamination, several phenomenons have been considered : - the Cd mobilization in the Riou-Mort and Lot river, - the Cd transfer into the estuary, - the kinetics for Cd accumulation in oysters.

38 I

The source of Cd and the riverine processes. As soon as the source of pollution was recognized, the waste area of the old plant has been isolated by building a dam. Since this construction, this waste is no longer a major source of pollution. But in spite of this work, the Lot river remains a source of Cd pollution. Indeed bottom deposits all along the Lot river contain very high contents of heavy metals, particularly Cd. These deposits are accumulated behind a series of dams which have been built all along the Lot river in order to regularize the water fluxes. The accumulation of sediment behind these major dams (fig. 7a) has been estimated by using seismic profiles (fig. 7b) and coring. Then, geochemical studies on cored sediments allowed to recognized the thickness of contaminated bottom sediments (fig. 7c). From these results, the global stock of Cd, in the bottom deposits of the whole Lot valley has been estimated to 200 T. The more recent studies (13) have for objectives to predict the risks of mobilization of this contaminated stock either by reworking of sediment or by diffusion of pore water at the water/sediment interface (fig. 8). Preliminary results show that resuspension of fine polluted sediments can occur during the river floods, when currents reach 30 cm.s-l (measurements at 1 rn above the bottom). In addition, major changes of bottom physico-chemical conditions (e.g. acidification andlor Eh increase) could have for consequence an increase of dissolved fluxes at the watedsediment interface. Preliminary calculations (13) applying Fick law to the pore water profiles (14) indicate that Cd molecular diffusion at the water sediment interface varies from 17.10 -6ng.cm2.s-1 (reducing environment) to 41.1O-6ng.cme.s-1 (oxydizing conditions) such fluxes have nevertheless to be precised through the study of a greater number of pore water profiles realised in the framework on the on going studies. The transfer of Cd, within the estuary Fluvial inputs of Cd within the estuary have been monitored since 1982. These inputs widely vary, according to the hydrological conditions of the rivers (7). For average riverine discharge conditions (Garonne 1 13,6109 m3.year-1 ; Dordogne : lO,2.lO9 m3.year-1) the fluvial contribution of Cd has been estimated to 23 t.year-1 for particulate Cd and 2 t.year-l for dissolved Cd. As for many metals (7) an abrupt upstream/downstream decrease of Cd content of SM (fig.9 [a]) has been observed along the estuary (2). Conversely, maximum dissolved Cd (fig. 9 [b]) concentrations have been observed for salinity of about 20 %o, i.e. in sites very close of the inlet. Far away towards the sea, at higher salinities, the Cd content decreases through dilution with the marine water. This repartition of particulate and dissolved Cd has been explained (15) by a solubilisation along the estuary of the Cd initially associated with fluvial SM.

382

a

I

) Molndornt

b)

C)

Selsmlc profile

Cojarc

d)

Core2 Sandy mud

I0 Muddy sand

20 24 Sand

32

30 40

Muddy sand

50 cm

Fig. 7. Cd in bottom sediments of Lot river (Cajarc).

crn

Colorc

Cd

content

383

o 10

,

'

"

"

3

1

20 30 40

50. cm

j_

Figure 8. Cd in pore water (Cajarc).

L

0 - $0 La Reole

10

b

'

30VeO S o l i n i t y

Oyster IbOKrn from Bordeoux area LeVerdon

Bordeaux

-GAAONNE RIVER-+ (estuorine port)

20

+--ESTUARY-PF

SHELF-

Figure 9. Upstream downstream evolution of particulate and dissolved Cd contents (low and high river dischage periods).

384

IGBA studies, by monitoring the behaviour of particulate and dissolved metallic elements along the estuary and by using radioactive zinc tracers previously showed that when metals such as Zn enter the Gironde estuary, they mainly occur in a particulate form but become soluble when they reach the estuary (16). As cadmium has similar chemical characteristics as zinc, it is logical to assume that it reacts in the same manner within the estuary and this would be the reason why, downstream, metal mainly occurs in a dissolved form. On the basis of various surveys devoted to the study of the behaviour of metals within the estuary and its outlet (7-l), it was possible to estimate the flow budget for various elements particularly for Cd. Estimated on the basis of hydrological conditions average year, the Cd budget (tabl. 3) showed that the lower part of the estuary receives an input of dissolved Cd exceeding 20 tons per year. Table 3 Annual budget of Cd in the Gironde estuary Ocean

Rivers

Estuary

<------

particulate Cd c------

It

3t

particulate Cd

23 t

I

dissolved Cd

<-----21 t

v

<------

dissolved Cd

I t 2t sedimentation

(after data from Jouanneau, 1982 ; Donard, 1983 ; Jouanneau et al. 1990). Average hydrological conditions (1980 to 1985)

-

Kinetics of cadmium accumulation i n oysters Processes of metal fixing in oysters were studied by implanting non-contamined specimens from the Arcachon Bay in the various natural oysters fields of the Gironde estuary (15). In summer, metal concentrations in transplanted oysters rose considerably from 4,2 pg.g-l in May to 72 pg.g-l in August. In winter, accumulation were not as high but, increased nonetheless from 4,3 pg.g-1 to 18 pg.g-1 within a 6 months period. Furthermore, (8) showed that there was, in general, a direct correlation between dissolved Cd contents in water and metal concentration in the mussel flesh. As for oysters, they were reported to fix three times more Cd than mussels. Given the dissolved Cd contents in the Gironde estuary mouth, values for oysters were thought to range between 90 and 115 p g . g - l ~thereby corresponding perfectly to the values measured in oysters.

385

EVOLUTION OF THE CD CONTAMINATION

In order to precize the evolution in time and space of the Cd contamination, monitoring of the fluxes of dissolved and particulate Cd is carried out in various parts of the drainage basins. Results obtained from the 4 last years showed an abrupt decrease of the flux of Cd since 1989. Nevertheless the Cd content of SM remains very high (13). Regarding oysters, the content of cadmium in the flesh of animals are also monitored since 1979 by the RNO. Results indicate a slight decrease of the Cd contamination since 3 or 4 years and this could be linked to a decrease of Cd fluvial inputs into the estuary (fig. 10). Nevertheless the Cd content remains very high, particularly in oysters from La Fosse.

C a d m i u m /Gironde Estuary (Bonne Anse) ( Oysters 1 L

c

79 80 81 82 83 84 85 86 87 88 89 90

Yeors

Figure 10. Evolution of the cadmium contents in oysters. In fact, it has to be highlighted that the evolution of Cd fluxes has to be considered in connexion with the hydrological conditions. Since 1989, in relation with the climatic conditions evolution, an abrupt decrease of the river discharge has been observed for the Garonne (fig. 11) and the Dordogne. The solid discharge itself has been considerably reduced, the maximum of transport of particles, being, as usual, always observed during the high river discharge periods (Jouanneau, 1982). Another important point to consider regarding the Cd contamination evolution is the behaviour of Cd at the inlet of the Gironde estuary particularly in the vicinity of Marennes-Oleron Bay (fig. 1) which is the largest site of French oyster farming.

386

Satellite photographs, as well as in situ measurements (16-17) suggest that part of the suspended estuarine material and water discharged into the ocean by the Gironde estuary (fig. 13) migrates towards the North-West and later, folds back to the South where it accumulates mostly in the "Vasibre Ouest-Gironde" a muddy area located 50 km west of the outlet of the estuary. Nevertheless a part of the suspended matter can also accumulate in the Marennes-Oleron Bay, in the north of the outlet. 900

700

- 500 In n

300

100 0 0 (0

0,

8

P

!?

E

0 0)

0,

Years

Figure 11. Mean river discharge in the Garonne river (1959-1991).

CONCLUSION To conclude, it is necessary to point out that the metal contamination of natural environments and organisms are tributary to various interdependent factors and multiple interactions : - first, the fact that the pollution is the result of the interaction of several kind of processes : the oysters contamination is not only due to biochemical processes characterizing the end of the chain, but, at an early stage, the result of sedimentary, dynamical and geochemical mechanisms inducing the influx of dissolved Cd in the lower part of the estuary ; more specifically, the sequences of successive exchanges between Cd linked with sedimentary particles and dissolved Cd have to be emphazised. This complexity of the mecanisms involved in the transfer of the cadmium pollution is summarized in the fig. 13.

387

Landsat 1 - View 03/09/76 Figure 12. Turbid plumes from the Gironde estuary (high river discharges). Riou Mort river Upper Lot river

adsorptionon pa

I D

- .

ID : Cd d6solved flux P : Cd particular flux

1

sdubilization

a% litr;

Lot river

- - -%.”

Estuarine solubiliration Bioaccumulationby phytoplankton

D . . - - - - - *

.......

*

;.-..

.-~

+ -.

- -p

Garonne river

River inputs Gironde Estuary (Garonne and Dordogne)

inner shelf

Figure 13. Schematic processes and patterns governing the cadmium behaviour in the studied environment.

To sum up, it has been shown that the cadmium is mainly introduce in the Riou-Mort river on a dissolved form. Very quickly, this cadmium is linked to the suspended matter. So, in the Lot river, cadmium is principaly on a particular form and settlewith particles, inducing all along the Lot river a new large source of contamination in the bottom deposits. This cadmium entrapped in the sediments of the river can escape and reach the Garonne river by two ways : first and principaly, on a particular form

388

during the flood events when the currents near the floor are strong enough to erode the sediments, secondly by geochemical diffusion at the sediment-water interface. These fluxes of cadmium from the Lot river then joins the Garonne ones and reach the estuarine system. There occur the estuarine processes ; conjunction of long residence time for suspended matter and increase of salinity towards the inlet induce a removal of the cadmium that goes on a dissolved form mainly in the dowstream edge of the turbidity maximum. Then, cadmium is principaly on a dissolved form at the inlet of the Gironde estuary, in the area of the natural oysters fields. However as early as the inlet, a new opposite mecanism occurs ; that is the assimilation of a large part of the dissolved cadmium by the phytoplankton. So that in the coastal waters close to the inlet of the Gironde estuary, dissolved cadmium contents decrease and a new increase of cadmium in the particular form is observed. - second, the very important distance (about 400 km) between the source of pollution and its relevant effects on estuarine and coastal fauna. From a pinpoint source, the whole area affected by the pollution is as large as the SW of France, taking into account that at least three french counties (Midi-Pyrenees - Aquitaine Poitou-Charente) are, from the administrative point of view, involved in the problem. - finally, given its complexity, the problem should be managed through a permanent dialogue between the various people and organizations involved in : not only scientists and technicians but also decision-makers, politicians and population representatives.

SUMMARY In this paper, characteristics, effects on biota, origin and transfer processes of Cd pollution in the Gironde estuary (SW France) are described. Cd mainly originates from a former industrial zone of the drainage basin localized about 400 km upstream to the inlet (mine of zinc, relevant waste, metallurgic plant, now out of work). At the estuary mouth, a major cadmium contamination of molluscs (wild oysters and mussels from natural beds) was first recognized in 1979 in the framework of the "National Observation Network", the office in charge of the monitoring of the quality of the french marine environment. Detailed studies were then undertaken on biota and on solid and dissolved media. Biota was studied by monitoring, populations of oysters and mussels transplanted from uncontaminated areas (Arcachon Bay) into the inlet of Gironde. Cd concentrations in oysters reached high level (up to 80 p9.g-l) in only 6 months. Kinetics of Cd accumulation appeared higher in summer than winter. Studies in the drainage basin enable to identify the source of cadmium. They showed that in the rivers, and till the limit with the estuary (150 km upstream to the natural oyster banks), Cd is mainly associated (up to 90 010 of the total flux) with suspended material transported by the river (Garonne). Downstream, SM undergoes major changes facilitated by a long residence time (2 years) of SM within the brackish waters of the estuary. Cd and many other associated metal pass, for a large part, into solution under the effects of macrotidal type estuarine processes. After solubilization, the metals are concentrated by

389

molluscs of the estuary mouth according to Kinetics of accumulation studied on the basis of an in situ experiment.

ACKNOWLEDGMENTS This work was partly supported by Agence de I'Eau Adour-Garonne and Association pour I'Amenagement de la Vallee du Lot. Gilbert Lavaux and Noele Maillet are acknowledged for assisting in this research and preparation of manuscript.

REFERENCES Donard 0, Contrat $etude n.88213. Ministere de I'Environnement. Rapport final 30 nov.1984, 25p, 22 fig. Donard 0, Latouche C, Bourg A.C.M, Vernet J.P, Int. Conf. on Heavy Metals, Abst. vol. Heidelberg, 1983; 960-963. Dumas P, DiplBme d'Etudes Superieures de Sc. Nat., Univ. Bordeaux I, 1985; 126p, 35 fig. Jouanneau J.M, Latouche C, Philips I, Rapport interne IGBA, 1986, 27p, 16 fig. Philipps I, Jouanneau J.M, Rapport interne IGBA, 1986, 24p, 14 fig. Jouanneau J.M, Latouche C, The gironde estuary. Contributions to sedimentology, H. Fuchtbauer, A.P. Lisitzyn, J.D. Milliman, E. Seibold, Stuttgart, 1981, 10: 115. Jouanneau J.M, These DoctAs-Sc., Univ. Bordeaux I, 1982: 732, 15Op, 62 fig., 31 tabl. Cossa D, these Doct.es.Sc., Univ. Paris VI, 1987: 374. Elbaz-Poulichet F. Huana W.W. Martin J.M, Zhu J.Y, Marine Chemistry, 1987: 22, 125-136. Boutier B, Chiffoleau J.F, Jouanneau J.M, Latouche C, Rapports scientifiques et techniques de I'lfremer, 1989: 14, 105p, 23 tabl, 32 fig. Bruland K.W, FranksR.P, In: Trace metals in Sea water, CS Wong et al. edit. Plenum Press, N.YorK, 1983. Roux M, Simonet F, Revue de I'Agence de Bassin "Adour-Garonne", hiver 1987: 34, 7-9. Lapaquellerie Y, Maillet N, contrat Adera n"421071, 3eme rapport d'avancement, 1992: 37p, 39 fig. Ly H, Gregory S, Geochim. cosmochim. Acta, 1974: 38, 703-714. Jouanneau J.M, Boutier 6, Chiffoleau J.F, Latouche C, The Science of the total Environment, 1990: 97/98, p.465-479. Jouanneau J.M, Etcheber H, Latouche C, Plenum Press, New-York, C.S.Wong edit, 1983: 245-263. Castaing P, These Doct.es-Sc., Univ. Bordeaux I , 1981: 701, 530p, 179 fig, 14 tabl. Castaing P, Philipps I, Weber 0, Oceanologica Acta, Paris, 1982: vo1.5, n o l , 8596.

-

10 11 12 13 14 15 16 17 18

This Page Intentionally Left Blank

CHAPTER 6 Synthesis and methods

This Page Intentionally Left Blank

39 1

MERCURY POLLUTION AND CYCLING IN AQUATIC SYSTEMS F.M. D’Itri Institute of Water Research, Michigan State University, East Lansing, MI 48824 INTRODUCTION The increasing quantities of mercury in various products and processes as well as the wider use of organomercurials raised serious concerns even before it was discovered that inorganic mercury could be methylated naturally in the environment. The modem phase of understanding environmental mercury contamination began with the discovery that human beings as well as birds and animals had been poisoned by eating fish and shellfish that had accumulated great quantities of methylmercury released into the bay from an acetaldehyde plant first at Minamata and later at Niigata, Japan [l]. While this is the most significant incident of methylmercury poisoning, the actual extent of human suffering is often underestimated. Most Minamata victims are acknowledged only after their symptoms are confirmed at autopsy, too late for direct compensation. Although the number of deaths has grown from 52 in late 1956 to 150 confirmed in 1960, by December 31, 1987, more than 17,000 individuals had been either directly or indirectly affected by methylmercury poisoning but only 2,840 were officially certified to have Minamata Disease and so were legally eligible for compensation. Of those, 999 (35.2%)had died (Table 1). Approximately 5,000people awaited medical evaluation and another 9,995 had been denied certification. So few people have been recognized as official Minamata victims because the certification process implemented by the Chisso Company with the consent of the Japanese government required each patient to be examined by 10 doctors. They must achieve consensus in their diagnosis for the patient to be certified as an official Minamata patient. Other victim’s lawsuits are working their way through a number of court systems in Japan, but they are even less likely to receive assistance. In early 1992 a Tokyo court declared the government to be free of legal responsibility, reinforcing a national policy of steadfastly resisting all pressure even to consider out-of-court payments to victims. This does not bode well for the appeal of a 1987 court ruling in Kumamoto prefecture where Minamata is located. In that case the lower court held national and prefectural government to be legally culpable [2]. While the Japanese government and industry have struggled to avoid responsibility and recompense, in North America and Europe discharges of inorganic mercury still did not generate concern as long as it was thought to be relatively inert in the environment. The extent of damage was first realized in the early and middle 1960s when fish and fish-eating birds from lakes and coastal waters of Scandinavia had elevated levels of methylmercury that primarily originated from inorganic discharges by industrial sources, especially the paper pulp and chloralkali industries. The same processes were observed in the United States and

Table 1 Summary of the Status of Minamata Disease Patients on December 31, 1987 [3]. ~~

~

~~

~

~~

Prefecture

Officially Certified Patients

Minamata Disease Deaths

Kumamoto

1038

677

4056

6842b

12,613

Kagoshima

3 14

121

75 1

233 1'

3,517

Niigata

489

20 1

15

822

1,527

TOTAL

1841

999

4822

9995

17,657

Certification Pending Rejected'

Total

As of June 30, 1988 [4]

1,470 individuals identified for Special Medical Care Status 513 individuals identified for Special Medical Care Status Canada in the late 1960s and early 1970s. In 1968 inorganic mercury was shown to be methylated by microorganisms [5-61. Then the industrial point discharges of inorganic mercury to waterways were halted, and the levels of methylmercury in fish initially decreased; but they remained higher than background, especially in Canada and Scandinavia. Then attention was turned to possible environmental effects from atmospheric cycling of emissions from natural and anthropogenic sources and the impact of acid rain on the process. ATMOSPHERIC CYCLING OF MERCURY The atmosphere is the major pathway for the global transport and deposition of naturally and anthropogenically derived mercury (Figure 1). It exists in the atmosphere primarily as a gas ( > 97 %). The rest is particulates [7]. Ionized compounds volatilize by three processes: 1) chemical reduction into the elemental form, 2) biological reduction into the elemental form through the activity of microbes, plants, or other living organisms, and 3) biotransformation into volatile organomercury compounds, mainly short-chain alkyl mercurials. Once it is released into the atmosphere, mercury is widely distributed by wind currents. The atmospheric concentration and residence times vary depending on factors such as wind speed and duration, temperature, and barometric pressure. While the background concentration in uncontaminated air over the Pacific Ocean appears to be about 1 ng/m3, the levels in the vicinity of urban centers usually average about 10 ng/m3 over a range between 2 and 50 ng/m3. Overall, the total global natural and anthropogenic emissions into the atmosphere have been estimated at 3,000 and 4,500 metric tons of mercury per year respectively [8]. As the anthropogenic inputs increased over the past century, they accounted for an estimated 25 to 30 percent of the total [9].

393

Figure 1. The emission, dispersal, transport, transformations and deposition of mercury in the environment. The atmospheric residence time has been estimated by Fitzgerald et al. [lo] at between 134 and 300 days and from several months to one or two years by Lindqvist and Rodhe [l 13; however, some estimates as high as three years have also been reported. Because of extended cycling with the winds, much of the mercury appears to mix rather uniformly throughout the troposphere over the northern hemisphere before it falls on watersheds or lakes, primarily in rain, snow, dust, or through gas exchanges with aquatic surfaces, sometimes long distances from the source. A fraction re-volatilizes after it reaches the ground or ocean before it can be incorporated into the soil or water. The extent of this emission, deposition, and re-emission process is unknown; therefore, this aspect of the cycle has not been quantified to any significant extent.

While only a very small fraction of the atmospheric cycle consists of particulates, they may be an important source of contamination, especially to local environments close to their release. The oxidized forms of airborne mercury are usually non-volatile and more water soluble, which allows them to be scrubbed from the atmosphere in relatively short time spans, often no more than a few days to a few weeks [12]. Consequently, they are normally transported over shorter distances, accumulating in the environment near the emission source. In combustion processes without flue gas treatment systems, between 20 and 60 percent of the mercury emissions can occur in the ionic form, much of which may be associated with particles [13]. WhiIe the residency time in the atmospheric cycle can be substantially reduced, the mercury may still be transported up to several thousand kilometers. The distance depends on the nature of the emitting source, the size and density of the

394

particles, and changes in their physical characteristics and/or chemical composition during transport, adsorption and solution, and meteorological conditions. Also, some fraction of the airborne elemental mercury is converted to ionic forms by oxidizing agents such as ozone and hydrogen peroxide, especially near the discharge point. Iverfeldt and Lindqvist [ 141 reported that the absorption of mercury in the water phase was experimentally increased by three orders of magnitude if ozone was present. They postulated that the atmospheric oxidation of elemental mercury by ozone in water can be significant if the ozone concentrations are high enough. Therefore, it appears that any anthropogenic activities that increase the concentration of oxidants in the atmosphere can, at least theoretically, enhance the mercury deposition rate. The reaction mechanisms have not yet been completely elucidated, and the relative importance of the process may vary as a function of solar radiation and the Occurrence of other airborne pollutants. Methylmercury has also been reported as a fraction of the total atmospheric mercury loading, however, not only is its source unknown, but the airborne concentrations have not been adequately quantified because its concentration is less than the limits of detection. It can, however, be measured in rainwater because methylmercury is so soluble in water; but the concentrations are usually less than 0.1 ng/l. The physical-chemical properties of inorganic and methylmercury, especially volatility, contribute to the ease by which these contaminants move among the lithosphere, hydrosphere and atmosphere. Of the species in the environment, elemental and dimethylmercury are the most volatile and, consequently, most likely to be re-emitted from the surfaces of soils and waters. They undergo solar radiation induced photochemical conversion to elemental or other non-volatile forms in hours or days [15-161. See Figure 2. Schroeder et al. [17] showed that the summer, total, vapor-phase mercury fluxes from a lake surface were diurnal and about six times greater (1.1f 0.4 ng Hg/m*/hr for soils vs. 6.3f 3.6 ng Hg/m*/hr for water) than over agricultural and forest soils. While the number of measurements was small, the data suggest that the rates at which volatile mercury species are re-emittedhe-deposited in an aquatic ecosystem are significant relative to the total deposition rate [17-191. MERCURY CONCENTRATIONS IN RAINWATER While the theoretical solubility of elemental mercury in rainwater at 25°C is estimated at 0.03 pM (6 pg/l), the actual background mercury concentration in an uncontaminated environment is usually about 1 or 2 ng/m3 [20]. Lindqvist et al. [21] suggested that the concentration in rainwater often exceeds the solubility of elemental mercury because water soluble mercuric ion is present. The reported mercury concentrations in precipitation range between 10 and 200 ng/l with some values as high as 1200 ng/l [22]. More recent studies utilizing ultra clean procedures indicate rainwater concentrations ranging from 3 to 40 ng/l in the United States [23-251 and from 8 to 37 ng/l in the Scandinavian countries [26].

395

CH,HgCH,

Aquatic Food Web

Water

Figure 2. The biogeochemical cycling of mercury in aquatic ecosystems [18]. MERCURY CONCENTRATIONS IN FRESH AND MARINE WATERS The concentrations of mercury at the surfaces of watercourses are a function of atmospheric deposition, air-water gas exchange, biological interactions, scavenging (washout) reactions and dilutions through water movements [27]. Gas phase air-water mercury exchange and biologically mediated reactions may significantly influence the mercury concentration and cycling in surface waters [lo, 27, 28-29]. Mercury concentrations in freshwater are presently being re-evaluated, especially in light of the ultraclean analytical protocols developed to minimize contamination during collection, storage, and analysis [30]. Earlier, mercury levels in rivers and lakes were reported at concentrations in the range of between 1 and 14 pg/l [7, 31-33]. The best current data developed with ultraclean analytical methods indicate that the concentrations for remote uncontaminated systems usually range between 1 and 5 ng/l [7, 34-35], but the concentrations in some headwater streams follow the hydrolic regime. They can vary greatly during the year and may exceed 20 ng/l. Mercury concentrations in seawater are small. When ultraclean analytical protocols are employed, they usually range between 1 and 10 pM, equivalent to 0.2-2 ng/l [lo, 36-39].

396

BIOTIC AND ABIOTIC TRANSFORMATION OF METHYLMERCURY In the mid 1960s the methylation of inorganic mercury by microorganisms in the environment was shown to occur [6]. Since then, it has been shown that mercury can be interconverted between elemental mercury, mercuric sulfide, methylmercury and dimethylmercury by biotic and abiotic processes in both aerobic and anerobic environments. Many of the transformation reactions are mediated by microorganisms. These include: 1) the enzymatic reduction of inorganic mercury to elemental mercury, 2) the precipitation of ionic mercury as mercuric sulfide by sulfur reducing bacteria and/or other hydrogen sulfide producers, 3) the degradation of methylmercury to elemental mercury and methane, and 4) the methylation of inorganic mercury to methylmercury by non-specific enzymatic pathways. Environmental factors such as microbial activity and speciation, availability of complexing ligands, light, temperature, and redox potential to affect the production rate and the stability of final products. The abiotic methylation of inorganic mercury by humic and fulvic acids in soils and sediments has also been demonstrated [40-441. While the quantity so produced is suspected to be small [42, 451 it may, nevertheless, represent a significant source of methylmercury in humic and peat materials of wetlands and/or catchment soils. In addition, exoenzymatic chemicals, released from dead bacteria and microorganisms, may also contribute to the production of methylmercury in the natural environment. While the exoenzyme methylation of mercuric ion with methylcobalamin has been demonstrated in the laboratory [5] it has not been shown to occur in situ in the natural environment. No matter how the mercury is methylated, the methylmercury is soluble and more toxic. Unlike inorganic mercury, it is also more completely absorbed by aquatic organisms and very poorly excreted; consequently, exposed organisms can concentrate high body burdens despite very low levels in the water. The chemical and biological transformations of mercury in the aquatic ecosystem are shown in Figure 2.

DEMETHYLATION While some microorganisms synthesize methyl and dimethylmercury in aquatic ecosystems, others demethylate these compounds. After demethylation by bacteria was demonstrated [46], the microbial degradation of methylmercury into elemental mercury and methane was observed in both lake and river sediments [47-491; and in the water column [50]. These methylation-demethylation interconversions establish an ecologically dynamic system of competing reactions that can result in steady-state concentrations of methylmercury in the sediments and water column. Whether or not the concentrations of methylmercury are increasing in the sediments depends on the difference between the rates of the two processses as well as the rates of transport to and from the system [49]. Whenever methylation is more efficient than demethylation, a net increase in methylmercury synthesis rate develops to produce a greater equilibrium concentration in sediments as well as a greater equilibrium rate of transport back

397

to the water column, which leads to the greater availability of mercury for accumulation by aquatic organisms. ACID LAKES AND MERCURY IN FISH Large areas of northeastern North America and Scandinavia have become increasingly acidified, and this appears to influence the rate of methylation by bacteria or the rate of methylmercury bioaccumulation by organisms, especially fish from softwater lakes. The emission, dispersion and transformation of mercury, sulfur dioxide and nitrogen oxides in the environment are presented in Figure 3. Many physical, chemical and biological characteristics of a lake such as pH, temperature, hardness, alkalinity, particulates, dissolved organic content and conductivity influence the mercury concentration in fish. Unproductive acidic softwater lakes, those with low algae and plant growth, i.e., low levels of particulates and dissolved organic carbon (DOC) tend to produce fish with high concentrations of mercury. On the other hand, fish with high body burdens of mercury from lakes with highly colored water (high DOC) and particulate matter may be the result of methylmercury production in the watershed followed by runoff into the lake. Oligotrophic lakes, especially if they are slightly acidic with well-oxygenated sediments, provide an aquatic ecosystem with less dissolved and particulate organic matter and also less dissolved inorganic bicarbonate ions with which the mercury can complex. This causes more reactive mercury to be methylated and available to fish and other aquatic organisms [51-521. The acidic conditions in the oligotrophic lakes allow any dimethylmercury, which normally would be lost through volatilization, to be converted into methylmercury and be available for fish to assimilate and bioaccumulate. The methylation rate in softwater lakes can increase not only in the sediments but also in the water column as the pH decreases [SO, 53-55]. Numerous studies have demonstrated that water bodies containing fish with elevated concentrations of mercury have five characteristics: 1) Low pH [54, 56-61], 2) Low alkalinity [51-52, 611, 3) Low calcium [54,611, 4) Low productivity [62], and 5 ) High dissolved organic content [55, 611. Fish from more productive lakes, on the other hand, usually contain smaller amounts of mercury for three reasons. First, the fish are well fed and actively growing; therefore, the assimilated mercury is spread throughout a larger biomass and, in effect, biologically diluted. Second, because of the high organic content of productive lakes, the mercury that has complexed with particulates settles into the sediments instead of remaining available for methylation in the water column. Third, the anoxic conditions often found in the sediments of eutrophic lakes facilitate the formation of mercuric sulfide, which is much less likely to methylate [62-631. Evaluating how pH affects the methylation rate in softwater lakes is difficult because of the confounding influences of small concentrations of calcium and varying alkalinities and dissolved organic levels. Waters with a fow pH and small concentrations of calcium often contain fish with high mercury levels. One hypothesis is that the uptake of mercury by fish in slightly acidic waters is affected by calcium mediated changes in gill permeability [a].

398

biilfur

dioxide

Combustion of Fossil Fuels

5% Other Acids

Figure 3. The emission, dispersion and transformation of mercury, sulfur dioxide and nitrogen oxides in the environment [MI. Another theory is that any methylmercury produced in the water column of acidified lakes is directly available for uptake by fish via the gill membranes [50]. MERCURY ACCUMULATION IN FISH As the concentrations increase along the food chain (Figure 4) they can range from natural background levels of 0.1 to 0.2 pg/g in fish living in areas remote from anthropogenic activities to between 25 and 35 pg/g in fish taken from contaminated areas [65-661. The bioaccumulation of mercury by aquatic organisms occurs by three processes: 1) from the water via respiration, e.g. over the gills, 2) by absorption of water from the body surface, and 3) by ingestion of food. Of these mechanisms the uptake of mercury through the food chain is the most important route of bioaccumulation. Because the long-lived, predatory fish like ocean tuna, swordfish, and freshwater fish such as lake trout, walleye, northern pike and muskellunge feed on prey that have already concentrated some methylmercury, they develop the highest levels over time. Because methylmercury readily complexes with sulfhydral groups in animal tissues to form very stable compounds, it is not easily metabolized and eliminated by aquatic organisms. Consequently, the biological half life in fish may be several years, depending on the species [67-681. These high levels are the result of very efficient uptake from food organisms coupled with a very low elimination rate [69].

The mercury levels in pike taken from Swedish waters showed the highest correlations with the following parameters: I) the content of the surficial sediments, 2) distance from the point source, 3) lake retention time, 4) watershed drainage area, and 5 ) lake surface area [70]. The importance of airborne contamination is manifested in the fact that concentrations

399

Concentralion 01 Organic Contarnfinantsin Food Chain

Figure 4. Bioaccumulation and biomagnification of mercury in the aquatic food chain. in 1 kg pike appear to be rising despite the significant decrease in mercury emissions from Swedish industries [70]. LOSSES IN FISHERIES RESOURCES FROM MERCURY CONTAMINATION

Elevated methylmercury levels have been observed in fish in remote oligotrophic lakes in Sweden, Canada, the United States and Finland. These have been correlated with the increased acidification and the high humic content of the lake water to further increase the methylmercury content in fish, zooplankton and algae [54-55, 57, 711. In Sweden, for example, it has been known for some time that fish with elevated mercury levels inhabit a large number of lakes, especially in areas where acidification has taken place. Most of these lakes do not have known point sources of mercury. Rather, it is deposited from the air, directly on the lake or on its surrounding watershed. As a consequence, fishery resources have decreased. In Sweden about 250 lakes have been blacklisted because of high mercury concentrations in fish while some experts [70, 721 have estimated that an additional 9,400 Swedish lakes should be blacklisted because they contain fish with more than 1 mg Hgikg. CONCLUSIONS Over the last two decades, scientists have begun to develop a better understanding of how mercury released into the environment can be transported in the atmosphere and deposited hundreds of kilometers away. In addition, more data have been generated relative to the acidification of the terrestrial and aquatic ecosystems and how it affects the cycling of mercury, not only making more of it available but also increasing the rate of its transformation into methylmercury. Since mercury first became recognized as an environmental contaminant posing a potential hazard to human health, public and government awareness has led to regulations and the elimination of most of the point discharge sources in the developed countries. These

400

measures were expected to decrease the levels of mercury in fish significantly, but this did not occur because of the emissions from natural and anthropogenic sources to the atmospheric cycle. They may be responsible for increasing the environmental mercury levels in the northern hemisphere. Even modest additional increases in atmospheric mercury loading can cause the body burdens in fish to increase [73]; and these, in turn, can concentrate up the food chain. Therefore, the transport and dispersal of mercury from anthropogenic sources to terrestial and aquatic ecosystems constitutes a serious enough environmental contamination problem, at least over the decades, that it is most appropriate that governments find additional ways to reduce anthropogenic environmental releases. Because mercury is transported in the atmosphere over long distances, both national and international cooperation is needed to implement an international remedial action plan. Whereas the objective of the 1970s and 1980s was to eliminate point discharges of wastes containing mercury, the goal now must be to reduce anthropogenic atmospheric deposition to a level below what is leached from the soil under natural conditions. To accomplish this nations need legislation to phase out uses of mercury in commerce, similar to what is being considered in Sweden [13]. One step would be to equip waste incincerator plants with flue gas treatment systems to reduce mercury emissions. Greater efforts also should be directed to improve the efficiency of mercury removal technology in other exhausts. Finally, although recent studies suggest that acid rain may not be the environmental culprit that it has been suggested [74], it is appropriate to consider how much acidity the environment can tolerate and develop innovative technologies to reduce the emission of sulfur and nitrogen oxides. REFERENCES 1.

2. 3. 4. 5. 6. 7.

8. 9.

D’Itri PA and D’Itri FM. Mercury Contamination: A Human Tragedy. New York: Wiley-Interscience, 1977. Schlesinger JM, The Wall Street Journal 1992; February 10, p A12 Japan Environment Agency. Minamata Disease, Quality of the Environment in Japan. Tokyo: Government Japan Printing Bureau, Ministry of Finance, 1988; 239245. Fujino T and Kozuma S. Proceedings of the International Forum on the Minamata Disease, November 7-8, 1988. Kumamoto: 1989. Wood JM,Kennedy FS and Rosen CG. Nature 1968; 220, 173. Jensen S and Jernelov A. Nature 1969; 223, 753. Lindqvist 0, Jemelov A, Johansson K and Rodhe R. Mercury in the Swedish Environment: Global and Local Sources, Report No. 1816. National Swedish Environmental Protection Board, Solna, 1984. Lindqvist 0, ed. Water Air Soil Pollut 1991; 55, 1. Andren AW and Nriagu JO. The Biogeochemistry of Mercury in the Environment. Amsterdam:Elsevier/North Holland Biomedical Press, 1979; p 1.

40 1

10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29.

30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42.

Fitzgerald WF, Gill GA and Hewitt AD. Trace Metals in Seawater, NATO Conference Series IV, vol. 9. New York:Plenum, 1983; pp 297-316. Lindqvist 0 and Rodhe H. Tellus 1985; 37B, 136. Bowen HJM. Proceedings of International Conference on Heavy Metals in the Environment. Toront0:University of Toronto Press, 1975; p 1. Hellner C, Hakanson L and Johansson K. Mercury in the Environment--Problems and Remedial Measures in Sweden. So1na:Swedish Environmental Protection Agency, 1991. Iverfeldt 0 and Lindqvist 0. Atmos Environ 1986; 20, 1567. Niki H, Marker PD, Savage CM and Breltenbach LP. J Phys Chem 1983; 87, 3722. Niki H, Marker PD, Savage CM and Breltenbach LP. J Phys Chem 1983; 87,4978. Schroeder WH, Munthe J and Lindqvist 0. Water Air Soil Pollut 1989; 48, 337. D’Itri FM. Water Report. MYU KK 1991; 1(1), 3. Lindqvist 0 and Schroeder WH. Control and Fate of Atmospheric Trace Metals. Dordrecht: Kluwer Academic Publishers, 1989; p 303. Sanemasa I. Bull Chem SOCJpn 1975; 48, 1795. Lindqvist 0. Waste Manage Res 1986; 4, 35. Matheson DH. The Biochemistry of Mercury in the Environment. Amsterdam: Elsevier/North Holland Biomedical Press, 1979; p 113. Glass GE, Leonard EN, Chan WH and Orr DB. J Great Lakes Res 1986; 12, 37. Glass GE, Sorensen JA, Schmidt KW and Rapp, Jr. GR. Environ Sci Techno1 1990; 24, 1059. Fogg RR and Fitzgerald WF. J Geophys Res 1979; 84, 6987. Iverveldt A and Rodhe HJ. Atmospheric Transport and Deposition of Mercury in the Nordic Countries. Goteborg: Swedish Environmental Research Institute, 1988. Fitzgerald WF. The Role of Air-Sea Exchange In Geochemical Cycling. Massachusetts: D. Reidel Publishing Co, 1986; p 368. Fitzgerald WF, Gill GA and Kim JP. Science 1984; 224, 597. Kim JP and Fitzgerald WF. Science 1986; 231, 1131. Fitzgerald WF. The Biochemistry of Mercury in the Environment. Amsterdam: Elsevier/North-Holland Biomedical Press, 1979; p 161. Matsunaga K. Nature 1975; 257, 49. Windom HL and Taylor FE. Deep S e a Res 1979; 26A, 283. Gill GA. On the Geochemistry of Mercury in Long Island Sound: An Analytical and Field Study (Master’s thesis, University of Connecticut, Storrs, CT, 1980). Iverfeldt A and Johansson K. Verh Int Verein Limnol 1988; 23, 1626. Mierle G. Environ Toxicol Chem 1990; 9, 843. Freimann P and Schmidt D. Fresenius Z Znal Chem 1982; 313, 200. Bloom NS and Crecelius EA. Mar Chem 1983; 14, 49. Olafsson J. Trace Metals in Seawater, NATO Series 4, vol 9. Plenum: 1983; p 475. Gill GA and Fitzgerald WF. Deep Sea Res 1985; 32, 287. Beckert WF, Moghissi AA, Au FHF, Bretthauer EW and McFarlane JC. Nature 1974; 249, 674. Nagase H, Ose Y, Sat0 T and Ishikawa T. Sci Total Environ 1982; 24, 133. Nagase H, Ose Y, Sato T and Ishikawa T. Sci Total Environ 1984; 32, 147.

402

43. 44. 45. 46.

Rogers RD. J Environ Qual 1976;5, 454. Rogers RD. J Environ Qual 1977;6, 463. Berman M and Bartha R. Bull Environ Contamin Toxicol 1986; 36, 401. Tonomura K, Maeda K, Futai F, Nakagami T and Yamada M. Nature 1968; 217,

644. 47.

Spangler WJ, Spigarelli JL, Rose JM, Flippen RS and Miller HM. Appl Microbiol

48. 49. 50. 51.

Spangler WJ, Spigarelli, JL, Rose JM and Miller HM. Science 1973; 180, 192. Billen G, Joins C and Wollast R. Water Res 1974;8, 219. Xun L, Campbell NER and Rudd JWM. Can J Fish Aquat Sci 1987;44,750. Scheider WA, Jefferies DS and Dillon PJ. J Great Lakes Res 1979; 5, 45. Akielaszek JJ and Haines TA. Bull Environ Contam Toxicol 1981;27, 201. Rudd JWM and Turner MA. Can J Fish Aquat Sci 1983;40,2218. Wren CD and MacCrimmon HR. Can J Fish Aquat Sci 1983;40, 1737. Verta M, Rekolaianen S and Kinnunen K. Pub1 Wat Res Inst (Helsinki) 1986; 65,

1973;25, 488.

52.

53. 54. 55.

44. 56. 57. 58.

59. 60. 61.

62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74.

Jernelov A, Landner L and Larsson T. J Wat Pollut Cont Fed 1975;47, 810. Hakanson L. Environ Pollut Ser B 1980; 1, 285. Stokes PM, Dreier SI, Farkas MO and McLean RAN. Environ Pollut 1983;5, 255. Wiener JG. Comparative Analysis of Fish Populations in Naturally Acidic and Circumneutral Lakes in Northern Wisconsin, Report FWS/OBS-80/40.16.West Virginia: US Fish and Wildlife Service, 1983. Wiener JG. Effect of Experimental Acification in Mercury Accumulation by Yearling Yellow Perch in Little Rock Lake, Wisconsin. Corvallis: US EPA, 1986. Helwig DD and Heiskary SA. Fish Mercury in NE Minnesota Lakes. MiRReSOta: Minnesota Pollution Control Agency, 1985. D’Itri FM, Annett CS and Fast AW. Mar Techno1 Soc J 1971;5, 10. Fagerstrom T and Jernelov A. Wat Res 1972;6, 1193. Rodgers DW and Beamish FWH. Can J Fish Aquat Sci 1983;40, 824. Kitamura S. Minamata Disease. Kumamoto, Japan: Kumamoto University, 1968, p 257. Annett CS, D’Itri FM, Ford JR, and Prince HH. J Environ Qual 1975;4, 219. Boudou A and Ribeyre F. Aquatic Ecotoxicology: Fundamental Concepts and Methodologies. Florida: CRS Press, 1990. Meili M. Water Air Soil Pollut 1991;56, 719. Bernhard M and Andreae MO. Changing Metal Cycles and Human Health. Berlin: Springer-Verlag, 1984;p 143. Hakanson L, Nilsson A and Anderson T. Environ Pollut 1988;49, 145. Wiener JG. Trans N Amer Wild1 Nat Resour Conf 1987;52, 645. Anderson T, Nilsson A, Hakanson L and Brydsten L. S N V Report 3291. Stockholm, 1987. Fitzgerald W. Mercury in Temperate Lakes, EPRI-Wisconsin Mercury Research Annual Progress Report. Storrs: The University of Connecticut, 1990. Roberts L. Science 1991;251, 1302.

403

ANALYSIS OF SELENIUM Sager, M. Landwirtschaftlich-Chemische Bundesanstalt Wien, Wien, Austria

1. DECOMPOSITION METHODS FOR THE DETERMINATION

O F TOTAL

CONTENTS OJSELENIUM

Most methods for separation and final determination start from a Se (IV) containing sample solution. Suitable decomposition procedures should yield such sample solutions from the various materials encountered, TO check the decomposition procedure, the following methods can be applied: Regain of inorganic selenium added to the sample solution just before final determination means that the determination from this solution is possible as such. The regain of inorganic selenium before starting the decomposition procedure means that this has not been volatilized; organically bound Se may behave in a different way. Labelling with 75-Se tracer simplifies the detection of loss and regain, whatever Se- containing compound has been formed. The behaviour of organically- bound S e has been examined by addition of selenocysteine, selenomethionine, and triniethylselenonium chloride, which are known to occur in selenium metabolism [1,2], in case of mineral oil, dilauryl selenide has been added as a test substance [3]. The most reliable, but also most laborious way appears to control the method with metabolized tracer-selenium Radioactive selenium compounds are injected or ingested by the test animals, which is metabolized in 1-2 days. The labelled tissues and body- fluids can be investigated as well as by radioactive counting as well as by wet- chemical methods (1,2,4,5]. Reagent blanks have been only explicitely given for chloric acid [ 6 ] ; sometimes, sulfuric acid may also contain some selenium Blanks of magnesium nitrate are usually sufficiently low. Running blank samples in parallel, however, is recommended.

1.1. SAMPLES FROM ORGANIC- BiOLOGICAL ORIGIN Organic compounds of the matrix very often interfere with common analytical methods, as can be easily shown by addition of selenite to incompletely mineralized samples: they block separation columns, make foam in course of hydride evolvation, make smoke in the graphite furnace for AAS, alter electrode surfaces etc. Therefore, it is recommendable to mineralize the sample as complete as possible without losses As selenate i s not detected by many determination methods, the selenate which may have been formed by excess of oxidants, has to be reduced to selenite again. In case of drying of plant and animal tissues as well as body fluids up to 120" or with microwaves, no loss of selenium was observed [2] Labelled selenium fed to rats and metabolized in blood, brain, lungs and muscles, was regained at more than 95% upon drying. Similarly, no losses were found for freeze- drying [7]. Losses, however, may occur during drying of sewage sludges (author's experience) because of reducing conditions.

404

1.1.1 DRY- ASHING PROCEDURES

Organic material is oxidized by the oxygen of air during heating in a muffle furnace. High sample weights lower the needs for sample homogeneity, which is not so easy to achieve for foods and plants. Addition of magnesium nitrate, (e.g. as 10% solution), careful drying of the mixture, and sufficient ventilation are essential to yield white ash and to keep the sample alkaline [5,8]. After charring at SO0-55O0, the ash is dissolved i n l i l - HCI and heated half an hour on the water bath to reduce possibly formed selenate [8,9,10]. Quartz or glass beakers are preferable; for Pt- dishes, losses are reported to occur [ 131. From nonalkaline dry-ashing, however, Se is volatile as oxide or halogenide [12,13]. The reliable destruction of organic matter yields an excellent sample solution for hydride-AAS determination, but it is difficult to employ graphite- furnace- AAS or atomic emission methods. In a stream of air, selenium is quantitatively volatilized from biological material at 600-700", and can be caught by soda lime [ 121. 1.1.2 DECOMPOSITION BY USE OF OXIDIZING ACIDS

I 1 2 1 Closed vessels Decomposition of organic samples with specially purified nitric acid at high pressure in vessels made from PTFE or glassy carbon [ 14,15,16] enables the complete destruction of micro- amounts of sdample [6] with minimum amounts of reagent and thus extremely low blanks The mineralization of metabolized Se has been controlled by gel electrophoresis [ 171 As the organic matrix yields gaseous pioducts during its reaction with nitric acid, the sample weight is, however, limited The resulting solution is saturated with nitrous acid, which interferes with some separation and determination methods of Se 1.1.2.2Open vessels After addition of the oxidant (usually nitric acid), it is better not to start with vigorous heating, but to wait some time, and to increase temperature slowly. Charring on the sand bath without temperature control leads to volatilization losses. The use of a reflux condenser, or at least long-necked vessels, is recommended. HNO, alone is not sufficient for complete destruction; it has been used for dissolution of blood and food prior to graphite- furnace- AAS [ 181 or prior to GC- determination with 5 - nitropiazselenol [ 17,191. The oxidation power of nitric acid can be increased by addition of H,O, (dropwise, because of gas evolvation) [5,20,2 I], and/or temperature increase by refluxing in admixture with perchloric and/or sulfuric acid, at least in a Kjeldahl- flask [ 1,3,4,1 1,22-381. Refluxing with nitric/sulfuric acid could even decompose coal samples without losses [35]. Gel chromatography of the decomposition solution containing labelled metabolites of selenium showed, that Se was completely mineralized from animal tissue within 1 5 min under reflux with HNO,/H,SO,, but a lot of other non- destructed material may interfere with further steps of analysis [39]. After sufficient refluxing to destruct most of the organics, the rest of nitric acid can be fumed off with perchloric or sulfuric acid; selenium is converted into selenate, and thus not volatile any more [4,5,6,20,22,23,26,28,30,32]. Finally, the ashing temperature can be raised to a maximum of 210" [4,23,30]. This is important for the decomposition of fat and

405 other resistant compounds [4,28] Selenium itself is hardly present in the lipoid phase [4,5], but in case of high fat contents, it is necessary to add a 10-15- fold excess of acid (by weight), to avoid reducing conditions [27] Tiimethylselenonium, the main metabolite in urine, is not volatile, but it needs about 20 min i n boiling HNO,/HCIO, = 1 1 to convert it to inorganic selenium ( 1 ) Decomposition with aqua regia, as usual for inorganic samples (including soils), has not been reported biological matrices so far If a mixture containing chloric acid IS used, the resulting solution can be even used for the determination of iodine [6] Open flames nearby have to be avoided, because of explosions If only H,O, is used as oxidant, little addition of acid (perchloric, sulfuric) is also necessary [40,41] After warming, the reaction is exotherm, and not sufficient in any case, it has been used for the mineralization of serum and animal tissues prior to coprecipitation with tellurium, which is interfered by nitric acid [42] Addition of K,Cr,O, [43] as well as catalysts like V 0, [ 3 I ] or Na-Moo, [29,44] to the acid decomposition mixture enables to lower the temperature for destruction, but the resulting sample solution is not well suitable for the analysis of other trace elements, not every selenium- organic compound is mineralized by chromate Vanadium acts as a visual indicator for the end of the decomposition, by changing its coloui from brown to green [311

1.2 DECOMPOSlTlON OF INORGANIC S A M P L E S

During the destruction procedure of biological materials i t is above all iiiiportaiit to mineralize selenium from its organic cleavages, and to oxidize a large amount of sample weight without losses and import of blanks In case of inorganic samples, however, its volatility as the fluoride, chloride, and bromide causes troubles, fluorination of silicates I S not recommendable Under reducing conditions, lihe during the dissolution of sulfides 01 metals, selenium can be lost as H,Se or left undissolved as elementary selenium

I.2.1 DECOMPOSITION I N CLOSED VESSELS

For rocks, sediments and soils, good experience has been made with HNO,/HF = 7 1 (author) Slags and dusts with matrix Cu 01 Pb have been decoinposed wlth HCI HNO, HCIO, = 3 3 1 at 110-150"[43], i n cast: of furnace dust HF HNO, = I I was used [45] For soils, regain of Se was complete after decomposition with HNO, only at 140" 111 the pressure bomb [46] 1.2.2 DECOMPOSITION IN OPEN VESSELS

Soils can be quantitatively extracted for selenium without losses with aqua regia. if reflux condensers are applied [46] Without reflux condensers, it is recommended to worh at the boiling water bath Zinc ore could be dissolved with HNO,/HCI = 3 I on the warer bath, after evaporation of HCI, HNO, could be fumed off with HCIO, [47] Cu and Cu- alloys can be dissolved in l i l HNO,, transferring also selenide and telluride into quadrivalent oxidational status HCl

406

must be avoided, since it produces gaseous hydrogen selenide and selenium chlorides [48]. After evaporation on the boiling water bath, the sample can be fumed off with HCIO, [49]. Steel, however, is resistant towards pure nitric acid; a mixture containing HCI:HNO,: H,O = l:l:2was used, which can also attack Se(0). Prior to the final determination, HCI and HNO, can be replaced by HCIO, /H,PO, by short heating [50]. Geological material has been decomposed in the classical way with NaOH/Na,O, at 600" after irradiation and addition of a carrier, with subsequent dissolution in hot water and filtration [ 10,591. Acid decompositions with activated material (rocks, soils) were done in open vessels after addition of inactive selenium carrier with HNO, /HCIO,IHF [52], with H2S0,/H20, [46], or with H,SO,/HNO, at 1 50°(air condenser)[53]. 1.2.3 DISTILLATION AS BROMIDE From acid solution, selenium can be distilled as SeBr, and adsorbed in cold water. HNO, interferes, and has to be evaporated or decomposed beforehand. The distillation has a low sample throughput, but it simultaneously offers a highly selective separation from nearly all elements, and it results in a solution of Se(IV), which is wanted for most final determinations. Soils containing selenite, only need addition of HBr [54]. From meteorites, SeBr, was distilled after reflux with sulfuric acid, SeBr, was distilled in a N,- stream with addition of HC1:HBr = 3.1; it is only accompanied by As, Sb, Sn, Re, and Hg [51]. Use of NH,Br instead of HBr avoids the carry-over of acid. Se (IV) and Se(V1) are volatilized from rocks and sediments by heating in admixture with ammonium bromide and phosphoric acid at 250", in a stream of air, and can be retained in weak acid solution. Selenide and elementary selenium do not react and can thus be discriminated. If KJO, has been added, all inorganic forms are distilled together [55,56,57]. 1.3 BURNING AND VOLATILIZATION IN A STREAM OF OXYGEN

Heating in a stream of oxygen can volatilize selenium from several inorganic and organic samples as SeO?, which is subsequently condensed on a cool finger, or adsorbed on a proper substrate. From the condenser, selenium can be dissolved with dilute nitric or acetic acids. If selenium is adsorbed on a filter cartridge made of quartz wool/ soda lime/silver wool, it can be directly counted after irradiation of the sample [9,10,12]. Beneath decomposition, a separation from many matrix elements as well as significant enrichment are achieved; the sample output, however, is moderate. Whereas at 1 1 OO", the volatilization of selenium is quantitative, it is only marginal in a mixture of N2 and H2 ~ 3 1 .

Ores, metals, minerals, rocks, soils, coals, ashes etc. are weighed into a sample boat and transferred into a quartz- tube, which can be put into a stream of oxygen and heated from outside [43,58]. For a big apparatus, a water-cooled condenser is sufficient; micro- amounts from pure substances were collected in a trap of liquid nitrogen [59]. In case of organic samples, volatilization of unreacted matrix compounds including selenium- organics can be faster than the reaction of burning. This can be prevented by a platinum gauze which is additionally fed with oxygen for catalytic oxidation [60]. For the combustion of micro- amounts of samples, a special apparatus, named Trace-0-Mat, has been constructed, where the sample is ignited electrically in a stream of oxygen. The products of burning are depleted upon a cool finger filled with liquid nitrogen.

407

Addition of cellulose facilitates the burning of inorganic matrices, addition of silica retains accompanying elements (e.g. Pb). Plant material can be pelletized with AI,O,. The Se is separated from the matrix together with Cd, Sb, TI and Bi only, and can be finally eluted with HNO,/HCI = 111, or with 2M acetic acid/20% H,Oz = 7+1 [61,62,63]. After neutron activation and burning of the sample in oxygen, selenium can also be retained in a cool trap, adsorbed upon active carbon in a stream of nitrogen, and finally counted [64]. The oxygen necessarry for burning can also be activated electrically under reduced pressure, which is called low-temperature ashing. Because the sample boat is not heated directly, selenium remains in the ash, which can be dissolved easily in mineral acid. The method has been tested for plants and soils, as well as for coal and coal gasification products [45,65].

408

2. METHODS FORSEPARATION _ANDSPECIATION O F S E L E N I U M 2.1 G A S CHROMATOGRAPHY 2.1.1 PIAZSELENOLS

Se(IV) slowly reacts with o-phenylenediamine and its substituted derivatives (CI, Br, NO,) to form piazselenols, i.e. 5- membered rings with two carbons, two nitrogens and one selenium, at pH 1-2.5 [ 5 9 ] , which can be extracted into toluene (benzene) and determined by GC.Substitution of o-phenylene-diamine in position 4 influences the extractability of the reaction product with Se(IV) into toluene in the order H
409

2.1.2 VOLATILE COMPOUNDS

By means of gas-chromatographic separation and detection, alkyl selenium compounds, phenyl selenium compounds, selenium fluoride, selenium hydride and selenium trimethylsilyl derivatives can be analyzed Trimethylsilylselenide can be obtained from selenate in diniethylformamide Selenium hydride and substitutes are the result of the reactions of selenite and selenonium compounds with NaBH,, finally detected by AAS (see "hydride- AAS") Alkyl selenium compounds are major metabolites i n biological systems GC offers the possibility of the direct determination of the most volatile of them, without derivatization Among them are dimethyl selenide, dimethyl diselenide, diethyl selenide, diethyl diselenide, di-n-propylselenide or diselenide, ethyl seleno cyanate, diphenylseleniurn, Se-acetophenone derivative, or carbon diselenide Element- specific detection with AAS is preferable, because the chromatogranis ale much simpler Quartz- tube atomization (in presence of some hydrogen), as well as graphite tube atomization have been used The deuterium background correction of the AAS at 196 I nm has sometimes difficulties with non-specific molecular absorption by various organic impurities, but addition of ca 10% hydrogen to the argon carrier gas circumvents this problem, and, moreover, doubles sensitivity [ 7 5 ] For the determination of volatile Se- species i n ambient air, pre- adsorption upon silicone oil on Chromasorb W [ 7 6 ] ,or cryogenic trapping with liquid nitrogen [75] can be used The stability of dimethyl diselenide decreases above 140". therefore, flash evaporation with pre-heated carrier gas was applied By GC with Se- specific detection and pre- adsorption, the formation of dimethyl selenide and dimethyldiselenide gases evolved by astralagus from given sodium selenate i n a greenhouse, could be monitored [76]

2.2 LIQUID CHROMATOGRAPHY

Within this chapter, all methods of separation and enrichment of Se- containing species from a liquid phase on a solid packed on a column, are treated Column methods are used both for enrichment and species separation The time needed for separations is rather long, but in many cases quantitative detection on line saves the time for the application of another method for final determination 2.2.1 INORGANIC SORBENTS

Aluminum oxide adsorbs both Se(V1) and Se(1V) from acidic solution I n presence of 1 Mphosphoric acid, only selenite I S adsorbed. together with Sc, Ta and the rare earths I n case of neutron activated samples, Ta interferes with 75-Se counting and has to be precipitated with inactive carrier prior to the load of the column [ 7 7 ]

410

2.2.2 USE OF CHELATING ANCHOR GROUPS Multielement reagents or rather specific groups are fixed on a solid, which is finally packed to a column to pick up the wanted species from a large sample volume, or from an unwanted matrix. In case of selenium, only few attempts of synthesizing home-made column packings are reported. From weakly acid solution, dithizone or dithiocarbamate on active carbon take 60-90% of Se [78]. Bismuthiol-11-sulfonic acid loaded on an anion exchange resin has been used for the enrichment of Se prior to HPLC [79], but it can be operated also in a simple column technique, with penicillamine/cysteine as eluent [SO]. On DEAE- cellulose, a weakly basic diethylaminoethylcellulose, selenate I S strongly adsorbed from dilute formic acid (0.1-IM), whereas selenite passes the column. Selenate can be eluted with 0.05-1MHNO, [81]. CI- decreases the adsorbability of selenate and should be less than 0.01M, which is important for the enrichment from sea water [81]. 2.2.3 REVERSED PHASE HPLC The detection limits in HPLC recently approach AAS and electrochemical techniques. 2.2.3.1 Derivatization To achieve either optimum sensitivity, or multielement capabilities, products with known reagents are chromatographed. As they commonly start from Se(IV), information about speciation depends on a prior separation. Various piazselenols, which result from reactions with selenite, and which are widely used for spectrophotometric, fluorimetric or gas-chromatographic determinations, can be separated from excess reagent on Bondapak C18 or Unisil 5C18 [82] by HPLC. Thus, the reaction product of diaminonaphtalene with selenite has been extracted with cyclohexane, chromatographed with acetonitrile as eluent, and finally detected fluorimetrically [82]. The naphtylpiazselenol can also be chromatographed in chloroform on Nucleosil 10, containing amino groups [83], or in pure methanol on C18. Fluorescence detection is more selective at 480nmheading at 580nm, because the reagent is not excited. A detection limit of 0.15ng in 100 p1 extract has been achieved [84]. The 5-nitropiazselenol and the 5-chloropiazselenol can be easily separated from excess reagent on Nucleosil I 0-C 18, with methanol/water as mobile phase. The Se- carbamates cannot be used for photometric measurement, because of co-extraction of other coloured compounds, and adsorption of the excess reagent. For the carbamates, an UVNIS detector, operated at 254nm, is sufficient [83]. On LiChrosorb RP-8 and Nucleosil 10-C 18 it is possible to separate various diethyldithiocarbamates of Se and other heavy metals, obtained from extraction at pH4 with chloroform. For Se, methanol/water = 7+3 is preferable over acetonitril/water= 6+4, because of overlap of Se with Pb in the latter eluent [85]. Chelates of Se with APDC (ammonium pyrrolidin dithiocarbamate) have been enriched from HWpH 1.2 at C18 bonded silica gel, subsequently eluted with methanol and finally determined by graphite furnace AAS [86]. Another promising derivative has been gained from the reaction of Se (IV) with penicillamine, which is selective in acid solution. The product can be chromatographed on

41 1

Capcell Pak C18 in acetonitnleiwaterlphosphoric acid = 400 600 1 and detected by UV, but much more sensitive by derivatization to a fluorophore by reaction with 7-fluoro-4-nitrobenz-2, I ,3-oxadiazole [87] 2.2.3.2 No derivatization Selenate is adsorbed on CIS reverse phase Altex Lichrosorb RP-I from 0.001M hexadecyltrimethylammoniumbromide pH9 5 and thus separated from selenite, which passes the column. Selenate can be eluted with methanol. Detection with Zeeman- AAS allows the separation of various low-molecular As and Se species simultaneously [88]. Similarly, As and Se species have been separated without derivatization in acetate buffer on nucleosil, and detected by sequential ICP. Selenate is strongly retained, and eluted with ammonium phophate buffer pH 6.9 [89]. Selenols RSeH, diselenides RSeSeR, and selenylsulfides RSeSR have been separated on a 3pm diameter particle size Biophase ODS column, in 0.05M NaH,PO,/ 5% acetonitrile/0.004% Na-octylsulfate. An electrochemical detector with dual HgiAu amalgam electrodes allows to distinguish between the 3 groups because of their different electrochemical reactivity. This enabled the determination of reduced and oxidized glutathione in blood plasma [92]. 2.2.4 ION EXCHANGE AND ION CHROMATOGRAPHY As selenium is anionic, it passes the Dowex 1x8 cation exchanger in dilute HCI, whereas e.g. Ag, Au, Cd, Hg,Zn are adsorbed [91] The cation- exchange resin 1RA-200, charged with 3M- HCIO,, has been used to separate Se from various cations prior to determination by anodic stripping voltammetry, because Se is not retained from > 0.05M- HC10, [92]. Selenite and selenate can be separated at many anionic exchangers. For selenite, the sorption generally increases from sulfate form to chloride form to hydroxide form, and for selenate it is just reverse. Versus pH, selenite has a steep sorption maximum at pH 3-4, weak sorption at pH>6, but none beyond pH2. Contrary to this, selenate is strongly adsorbed at pH<2 [93]. In dual column configuration, with separator and suppressor column, in carbonate buffer, selenate is slower than the common anions, even sulfate, but without interferences. It is also possible to separate selenite, which comes between phosphate and nitrate, by raising the pH with additional carbonate, but the time for each chronlatogram of 36 min seems too long for practical work. As the alternative, oxidation of selenite and elementary selenium with H,O, to selenate is proposed [94]. Among the 0x0-anions, however, selenate is fastest. In 0.003M- sodium carbonate, selenate elutes in only about 6 minutes, followed by tungstate, molybdate, arsenate and chromate, which can be all gained by treating the sample solution with hydrogen peroxide. In this case, an excess of As over Se of 1 : 1000 is tolerable, detection limit of 1 pg/l without column-preconcentration have been achieved

WI. In 0.002M- Na$03/0.002M- KOH, selenite is fairly resolved from neighbouring CI and nitrate, and selenate from neighbouring sulfate and phosphate [96]. With a single column device, in phtalic acid/formic acid pH 2.7 on Vydac 302, selenite as the slowest was separated from chloride, nitrite, nitrate and phosphate in soil extracts. At pH 4.5, however, chloride, phosphate and nitrite interfered with selenite. Sulfate and selenate did not elute from the column [97] at this pH When the column was eluted with KH-phtalate buffer pH6.5, the broadening of selenate could be depressed by use of the

412

sulfate form. Upon the conventional conductivity detector, the sensitivity of selenite was only 1/10 compared with nitrate [97,98]. For this reason, the eluate was converted to the hydride on-line and fed into an ICP via a gas- liquid separator. The conversion to the hydride was reported to be 100 times more sensitive than direct nebulization of the eluate

WI. For the detection of selenite in single column configuration, under optimum conditions, maximum excess of other anions up to 10000 is possible. It is, however, advisable to remove high chloride levels (analysis of sea water) by reaction with a Ag- saturated cation exchange resin [97]. For selenate, in phtalate buffer 4.6, the resolution selenate to sulfate is possible up to 80 mg/l sulfate, but the calibration of selenate is sulfate- dependent [98].

2.3 LIQUID-LIQUID EXTRACTION Nearly all procedures start from Se(IV), separating it from other matrix elements, or in a few cases from selenate. For the extraction of selenate into a non-aqueous phase, only dioctyltin-dinitrate in CHCIJTBP = 1+3 could be found, No method has been reported so far for the separation of organoselenium compounds, but from this review an idea about the possibilities for the formation of lipophile Se-containing compounds might be achieved. 2.3.1 EXTRACTION FROM HYDROCHLORIC ACID The extractability of Se into benzene, chlorobenzene, dichlorobenzene or nitrobenzene from HCI is generally poor, but high distribution ratios are observed with binary mixtures of 9M- sulfuric acid and acid halides [loo] Separation of many other elements is possible, S b is coextracted [ 1001 From 6M- HCI, Se is not extracted into diethylether or diisopropylether, and can thus be separated from Au,Fe, Ga, Mo, Sb(V) and TI(II1) [ l o l l Similarly, it does not move into MIBK from 7M- HC1/7M-LiCI, contrary to As, Au, Fe, Ga, Mo and TI(II1) [ l o l l From concentrated hydrochloric acid, Se can be quantitatively extracted into toluene [ 1021 In concentrated HCI, selenate is converted to selenite and cannot be detected per se 2.3.2 EXTRACTIONS FROM HYDROBROMIC ACID Toluene quantitatively extracts Se from > 4M- HBr, and chlorobenzene at >6M- HBr, but benzene takes only 60% at maximum, from concentrated HBr [ 100,102]. Quantitative separation from many elements can be, however, achieved from 9M- H,SO,/O.lM- HBr into dichlorobenzene [ 1001. Diethylether takes only 31% of Se from 6M- HBr, but none from dilute solutions (c3M). Thus, in IM-HBr, separation from Au and TI(III), and in 0. IM-HBr, separation from Hg can be achieved [101,103]. 2.3.3 EXTRACTIONS FROM HYDROCHLORIC ACID WITH THE AID OF UNSATURATED HYDROCARBONS Unsaturated hydrocarbons react with selenite to compounds of the type (RCHCICH2)2SeCI,, which are extractable into chloroform or methylene chloride [ 1041.

413 2.3.4 REAGENTS WITH N- CONTAINING GROUPS

0- phenylendiamine and its derivatives are widely used for the isolation of Se The extract can be subjected to gas chromatography or liquid chromatography (see chapters l,2), but also simply for extraction- photometry, or extraction and gamma-counting after neutron activation The reaction IS very slow, and requires I hour at room temperature, or 10 miri at 75" [ 1051 N- phenylbenzohydroxamic acid can extract Se from 7M- HCIO, into chloroform as a yellow complex, utilizable for spectrophotometry (345 nm) or AAS [ 1061 With tri-isooctylamine, however, Se is not extracted into 4M- HCI, and can thus be separated from Au,Ga,Os,Re, and Cd [ I0 I ] 2.3.5 REAGENTS WITH 0-CONTAINING GROUPS Selenite reacts with excess aldehydes or ketones to compounds, extractable into chloroform from > 4M- HCI solution [108,109] With 30% TCMA (trichloro-methylacetate), S e can be selectively extracted from 4 5MH,SO4I7M- HCI into xylene, it is accompanied only by Pu, Np(IV), Sb(1II) and Co, and used for radiochemical separation The extraction got enhanced with the addition of watermiscible organic solvents to the extraction system [ I091 2 3 6 REAGENTS WITH P- CONTAINING GROUPS The affinity to P-containing groups is low Into undiluted tributylphosphate. Se I S not extracted up to 6M- HCI With TBP, in IM-HCI, separations are obtainable from Au,Ta, and TI(III), and in 6M-HCI from Au,Fe, Ga, Nb,Pa,Sb,Ta,Te,In,Mo and TI[ 1011 With TOPO, separation of Se from 6M-HC117M- LiCl has been described [110] 2.3 7 REAGENTS WITH S- CONTAINING GROUPS With diethyldithiocarbamate ( - DDTC), extractable compounds are formed i n acid solution, but this competes with rapid decomposition of the reagent From 2M- HCI or 2M- H,SO,, probably a mixture of Se(DDTC), and Se(DDTC) is extracted [ I I I ] , together with Cu,Ag,Hg,TI(III), As(III), Sb(III), Bi,Te(IV), Pd and Pt Se is not extracted from pH7 into CCI, and thus separated from Cu,Ag,Zn.Cd,Hg,In,TI,Pb,Sb,Bi,Te,Mn,Fe,Co,Ni,Pd and Pt [ 1 1 I , I 121 From acetic acid pH 2 6, Se and A s are extracted with DDTC [ I 131 Se I S also extracted from acetate buffer pH4 with diethvldithiocarbamate into CCI,, the selectivity can be improved by masking acconipanving cations with EDTA To prevent decomposition of the reagent, DDTC is added to the neutral sample solution, and the buffer is added at last [ I 141 Corresponding to the final AAS- sgnal, MIBK takes much less Se than CCI, Selenate does not react [ 1 141 In 0 1 M- sulfuric acid and citrate buffer pH5, the reactivity with DDTC of Se is poor [ I IS] When ammonium pyrrolidin dithiocarbamate (APDC) is used instead of DDTC, the subsequent graphite-furnace-AAS signal is lowei. maybe because of volatilization from the tube. but APDC is less pH- dependent for Se [ 1 14, I I61 0- isopropyl-N-ethylthiocarbamate possesses very high selectivity for Ag and Hg from HNO,, H,SO,, HCIO, and HCI, whereas Se is not extracted into chloroform from samples containing no bromide [ 1 171

414 The reaction with dithizone is incomplete and cannot be used for the isolation of selenite; however, if a pH more than 1 is used, other cations can be removed from the selenium containing sample [ 114,1181. In graphite-furnace-AAS, the Se-dithizonate is volatile and yields nearly no signal [ 1141. With K- ethyl xanthate, Se(IV) is quantitatively extracted from 0.1-IOM-HC1, and Se(V1) above 7M;selenate may be converted to selenite in HCI. The separation is rather selective. From 2M- HCI into chloroform, Se is only accompanied by Cu,Ag,Au,Pd, As,Bi,Mo, and partially Pt,Sn,TI and Fe. From 1M- HCI, only Ag, Pd,As, Mo, beneath Se move quantitatively into the extract [ 119,120]. During the extraction of Mo-ethyl-xanthate into a mixture of CC14henzene or into chloroform from 1 .SM-HCI, Se is coextracted [121]. In these complexes, Se is presumably bivalent [122]. The co-extracted Mo does not interfere with the spectrophotometric determination with 3,3' diaminobenzidine [ 12 11. Xanthates allow separation of selenium from matrix Cu,Ni,Pb and Zn [121]. With K-butylxanthate, Se(1V) yields a 1.4 complex, which can be extracted into CC14 at pH 2.0-4.7 and used for spectrophotometric detection at 395nm [122]. If benzazoles contain a mercapto group, like 2-mercaptobenzthiazole or 2mercaptobenzoxazole, Se(IV) quantitatively passes into the organic phase from 4- 1OM- HCI (chloroform). The use of chloroform permits the separation Se-Te, the equilibrium is reached in 5-45 min, depending on the HCI- concentration. The solvent does not participate, because curves obtained with different solvents are analogous [ 1231. By means of 0.2M-di-n-butyldithiophosphoricacid/CCI,, Se is extracted in the range 0.03-9M- HCI and 0.01 5-2.5M- H,SO,. A total separation from Fe, Ga,Os and Te over the entire range is achieved. In 0.1M- HCI, there is no separation from Cd, In, Mo, Ni, Sn(II), TI(II1) and Zn; in 5M- HCI, for example, there is no separation from Ag,As,Au,Bi,Hg,In,Pb, Mo, Pd, and Sb [101]. 2.3.8 SEPARATIONS WITH ORGANOTINS In the range pH 2-6, compounds of e.g. triphenyltin, trioctyltin and others, with Se(1V) and Se(V1) move into chloroform, MIBK, octyl alcohol; if the non-polar solvents benzene, xylene or cyclohexane are used, TBP (tributylphosphate) or TOP0 (tri-octylphosphinoxide) are added. Optimum conditions have been achieved with 25% TBP in chloroform or o-xylene [124].

2.4 COPRECIPITATION AND SORPTION METHODS A solid collector is either added to the sample solution, or produced in the sample solution by precipitation of a carrier, and finally separated by filtration or centrifugation. This leads to a uniform solid matrix, which can be either dissolved further, or directly submitted to XRF or gamma-counting after activation. Concerning speciation, in many cases all forms of Se might be precipitated, which is an advantage, if total contents are wanted. Hydroxides, however, rarely adsorb Se(V1) [125,126] (compare: movement of Se in soils).

415 2.4.1 PRECIPITATION AS ELEMENT

Reductive coprecipitation with elemental tellurium or elemental arsenic is very selective, getting both selenite and selenate [127]. Huge excess of oxidants in the sample, e.g. nitrate or chlorate are better avoided, because they consume much reductant. To ensure the presence of excess reductant, iodide can be added as indicator; but in this case, Cu is coprecipitated as iodide and should be removed, when it is present in the sample above the normal geochemical level [ 1281. Se is coprecipitated with Aso from arsenite and hypophosphite in 6M- HCI during 20-30 min boiling [129,130,131]. Similarly, tellurium can be precipitated with sulfite from HCI/HBr [ 1321, with hydrazinium sulfate from 3M-HCI [42] or 4M-HCI [ 1271 or NaOH after the centrifugation of other hydroxides [ 1331. Active carbon has been used as a sorbent for selenium, either after reduction of selenite with ascorbic acid, or, to get total selenium, after reflux with thiourea [134,135]. Also, hydrogen selenide, obtained by reduction of the sample with Zn" or NaBH,, adsorbs on active carbon, whereas selenite and selenate do not [ 1361 Elemental Se is rapidly adsorbed upon pyrex glass and polyethylene [ 1341; precipation, centrifugation, decantation and dissolution without change of the vessels is therefore preferable. 2.4.2 COPRECIPITATION WITH HYDROXIDES

Coprecipitation with hydroxides separates Se together wit other trace elements from alkali, alkaline earths, or large excess of unwanted anions (e.g. nitrate, sulfate). The hydroxides of Al,In,Ga,Zn and Mn are not suited for the collection of Se [ 1261. Fe- hydroxide sorbs selenite in weakly acid solution (pH 5 [137]; pH 2.4 in presence of ammonium chloride [138]), but not in alkaline solution at pH 8 (bromothymol blue indicator) [135]; selenate is not coprecipitated with Fe- hydroxide [137]. As an alternative to filtration or centrifugation, Fe-hydroxide can be flotated with bubbling nitrogen at pH 3 . 5 - 5 . 3 in presence of Na- dodecylsulfate or Na- laurylsulfate [ 139,140]. S e is accompanied by Ge,As, and Sb. Like Fe-hydroxide, Zr- hydroxide at pH 8 is reported to coprecipitate selectively Se(IV) [141], but not Se(V1). Coprecipitation with La- hydroxide at pH 9-10 separates As,Se,Te,Sb,Bi, Pb,Fe and Sn from matrix Cu prior to hydride AAS [142,143]. 2.4.3 COPRECIPITATION WITH SULFIDES

Coprecipitationkorption with sulfides is only useful, when the precipitates need not be dissolved again, e g for final determination by XRF or NAA Se can be sorbed on thin layers of ZnS, MnS, CuS, or PbS at pH 3-6, but high salt loads interfere, like sulfite, thiosulfate, phosphate, citrate, tartrate [ 1441 Precipitation from homogenous solution with thioacetamide has been successfully applied to samples from soil decomposition, after hydroxide precipitation [ 1451

416

2.4.4 COPRECIPITATION WITH ORGANIC REAGENTS

In many cases, multielement preconcentration with common reagents I S currently used with subsequent determination by XRF, which requires a uniform solid and low-atomic weight matrix. Precipitation with dibenzylammonium-dibenzyldithiocarbamateat pH 5/pH 4 allows the separation of 12 elements from alkali, alkaline earths and lanthanides from natural water [146,147], prior to XRF. Selenite and selenate are not coprecipitated or sorbed quantitatively on active carbon together with dithiocarbamates, oxinates and dithizonates, [ 148,1491. After reduction with sulfite or thioglykolic acid, carbamates precipitate Se also [ l SO]. Boiling of a 3M-HCI sample solution with a poly-thioether yields a precipitate containing Se and Te; this separation is applicable to ore analysis [ 15 11. Precipitation of polyvinylpyrrolidon + thionalide at pH 4 has been also utilized for the separation of Se and other trace elements from alkali and alkaline earths from natural waters, prior to XRF [ 1521. Cellulose filters wlth immobilized 2,2’-diaminodiethylamine quantitatively sorb selenite and selenate from natural waters in the range pH 3-7, together with other 0x0-anions, like chromate, arsenate, vanadate, molybdate, and tungstate [ 1531. The collection efficiency is strongly depressed with salt (e.g. NaCI) concentrations above 0.01M.

417

3. METHODS F O R D E T E R M I N A T I O N O F S E L E N I U M SPECTROPHOTOMETRIC A N D FLUORIMETRIC DETERMINATION METHODS

3.1

In many cases the same reagents can be used either for spectrophotoinetric or fluorimetric determination of selenium In comparison with other determination methods, the detection limit is not much inferior at all, especially if larger sample weights are used Strong oxidants and reductants react with the reagents, and thus interfere To increase selectivity, solvent extraction as separation step IS often included If not stated otherwise, after wet decomposition with oxidizing acid mixtures, the sample has to be brought to Se(IV) with HCI 3.1 1 AROMATIC 0-DIAMINES Aromatic 0-diamines react with Se (IV) in acid solution to yield coloured and extractable compounds with a 5-membered ring containing one Se, two N and two C- atoms They are called piazselenols [ 1541 3.1.1.1 0- Phenylenediamine and derivatives The unsubsituted o-phenylenediainine [43,155] as well as derivatives substituted at position 4 and/or 5 with methyl-, chloro- [55,83,156] and nitro- groups [83]have been added to the sample i n 0.1-0.5% aqueous or 0. IM-HCI solution For the reaction velocity, there is a pH optimum, usually pH 1-3 [ 15.51. An overview of acidity constants of reagents and products has been given in [ 1541. Reaction conditions of pH I -3/50° [ 1551, pH 2/30 min/2O0(low salt contents)[43], 2h/20°[S5] and 10 miiii40" [ 1561 have been reported. The reaction product can be extracted into chloroform in the range pH 1 7-12.5, and moves only partially back into the aqueous phase with NaOH [ISS]. To avoid coextraction of the unconsuined reagent, however, extraction at pH 2 is preferable. Interfering cations can be masked with EDTA prior to the addition of the reagent [43], or Se is isolated from the matrix by volatilization of SeOZ [43] or SeBr, [55]. Photometric determination has been exerted of the toluene extract (335nm for the unsubstituted [43], or 34 I nm for the 4-chloroderivative [SS]. A clean-up of the chloroform- extract by HPLC prior to photometric determination at 340nm [83] or fluorimetric determination at 550nm [ 1561 on-line, very much improves the detection limit, because of the separation of the proper signal from the reagent and by-products [83] For HPLC, Nucleosil C 18/chloroform [83], and LiChrosorb RP-8imethanol-water 80-20 [ 1561, have been used. The piazselenol of the N-phenyl-substitute of o-phenylene-diamine, called 2aminodiphenylamine, is obtained by reaction 2h/2S0 or 1 h/40", and is extracted as the ion-association-complex with perchlorate (from 3M- perchlorate) into hexanol-chlorobenzene = 6.4. With respect to other phenylenediamines, the higher molar absorption coefficient (18000 l/mol.cm) allows to reach a detection limit of 0 5pg Se/lOml sample solution [ 1571.

418 3.1.1.2 2,3- Diaminonaphtalene This reagent has been introduced in the sixties [ 158,1591 and successfully used since then. Usually a 0.1% solution of the reagent in 0.1M- HCl is prepared, but also a 0.1% solution in 5N-H,SO, has been reported [ I 11. Stored in the dark under cyclo-hexane, the reagent is stable for two weeks [28], but most authors prepare the reagent solution daily. Before the reagent solution is added, excess oxidants (e.g. NO,) can be expelled by heating with formic acid [28,37,121], or addition of hydroxylamine [ 1 1,25,28,37]. Interfering ions, like Fe and Al, have to be masked with EDTA, EDTA/NaF [160], or EDTA/NH,F [34]. The adequate masking capacity of EDTA towards Fe is in the mole ratio 1 : 1 [28]. Separation from matrix Fe can be achieved by coprecipitation with Te [160], by ion exchange [80,126]. Matrix Sb can be masked with tartaric acid [34]. Matrix Sn and Cu still interfere [ 1621. As the reaction and subsequent extraction are rather strongly pH-dependent, the pH has to be adjusted exactly. This is done with ammoniahydrochloric acid [25,28,34,121], or sodium acetatehydrochloric acid [ 1601, and can be facilitated with glycine [27]. Optimum pH range is 1.8 - 2.0, but at pH 0.4 still 85% of the Se are extracted [161]. More troublesome seems to be the co-extraction of the reagent and others at higher pH; when pH 2.4 was chosen, the fluorescent peak had to be corrected for its background [27] The resulting piazselenol is extracted uniformly with cyclohexane or n-hexane [34]. After excitation at 365nm or 378nm, the fluorescence of the piazselenol is measured at about 520nm. As utmost detection limit, 2ng/g (for 1 g sample/ enrichment into 5ml cyclohexanel no aliquots) [34] has been reached. The detection limit can be significantly lowered to 0. I 5ng/20pI by HPLC- separation of the piazselenol on C 18 reverse-phase column in 100% methanol [84], or on Bondapak C I8 and Unisil 5C18 in acetonitrile [ 1631, with fluorimetric on-line detection. This removes several chemical species which rise the blank level of the fluorimetric determination The procedure, including reaction, extraction and measurement, could be successfully transferred to an automatically working segmented-flow system [ 1641 3.1 . I .3 3,3'-Diaminobenzidine Reaction and interferences are comparable to diaminonaphtalene and the o-phenylenediamines. The reagent is used in 0 2-0.5% freshly prepared aqueous solution, and reacted with the sample at pH2-3 [ 126, 155,165,1661, for 50min at 45' [ 1551; alternatively at pH 1.25 for 45 min at ambient temperature (1211. The pH is adjusted with ammonia, oxidants interfere. For the extraction of the resultant selenium compound into toluene or chloroform, pH has to be increased to pH 5-10, optimum is pH 6.5 [121,155]. Photometric measurement is done at 420nm, and fluorimetric measurement at 570nm. Large excess of Fe, Pb, Cu were removed by extraction as the xanthates into chloroform, or by ion-exchange, minor amounts can be masked with tartaric acid and EDTA [ 12 1 .I 261. 3.1.2 S- CONTAINING LIGANDS Many S-containing ligands react with Se, like diethyldithio-carbamates, dithiophosphates, toluene-3,4-dithiole, dithizone [ 1551 The selectivity, however, is poor, which often prevents their use for the analysis of real matrices 2-Mercaptobenzimidazole selectively and rapidly reacts with S e in 2M- H2S0,, and can be extracted as ion-associate into chloroform/isopentanol = 8 2, which is applicable to the

419 direct analysis of steel- decomposition solutions, -= 10400 I/mole cm at 33Snm Fe, Cr, Co,Ni,Co,Mo are not extractable without HCI, only Cu, Te and Bi interfere [162,167] 4,5-Diamino-2,6-dimercaptopyrimidine yields a pink colour with Se(1V) in 0 5-2M HCI, and does not need to be extracted into an organic phase As detection limit, 0 01 pg/ml could be achieved (3-fold of diaminobenzidine), and applicated for semiconductors and animal feed premix, Au, V, Fe, and Cu interfere [ I681 Methiomeprazine hydrochloride forms a blue-coloured species with Se(IV) in 9 9 - I 1 2MH,PO, in about 40 min, with a very high -=29830 Vmolecm at 644nm, from 23 ions investigated, only Au interferes [ 1621 3.2 DETERMINATION OF SELENIUM BY ATOMIC ABSORPTION METHODS 3.2.1 FLAME -AAS

The low levels of occurrence usually encountered, as well as the poor performance of the technique limit the application of flame-AAS to quality control of Se- containing products as well as sulphidic ores The main atomic line for the determination of selenruni is 196 03 nm At this short wavelength, absorption from molecules, and light scattering from oxygen as well as from carbon- containing gases and water vapour cause severe noise of the zero-line [ 169,1701 In the NIO/C,H, - flame, the detection limit is about 2 pgiml In the argon- or nitrogenentrained hydrogen-flame, the sensitivity is 2-2 5 fold higher than for the conventional acetylene /air flame [75,169,171,172], low gas-flow is favourable The Ar (or N,) is entered like air, and H2 like fuel to the burner system, and the auxiliary oxidant entrace I S closed Similar constructions have also been used successfully as element- specific detectors in gas- chromatography to detect gaseous Se- compounds after separation on column in a stream of inert gas [7S] The exit of the GC can also be arranged concentrically around a hydrogen flame, which burns within a quartz cuvette, mounted in the optical AAS- path (see further hydride methods/flame-in-tube techniques, [76]) Another possibility to achieve reasonably low background absorption, is to use the "Slotted tube atom trap" A quartz tube is mounted on the burner- head in that way, that the optical pathways lies within the tube axis, and the flame is led through two slots in this tube This lowers the detection limit for the determination of Se in flame by a factor 3-6, irrespective of the equipment and the geometry used [ 170,173,174] Use of an ultrasonic nebulizer and a heated mixing chamber instead of conventional pneumatic nebulization results in only minor improvement of the detection limit in the H,/air - flame [ 1751 Fe and phosphate (as P,-bands) can cause spectral interferences in real samples, resulting in a negative D2- compensation signal Separation from matrlx Fe, P and salt matrix, or use of Zeeman - background- compensation is thus recommended [ 1761 Direct aspiration of combustible organic solvents (esters, ketones) causes much more noise in their flames than from an aqueous phase, and only 20-30% of the analytical signal I S obtained [ 169,1771 Se could be selectively extracted from sample solutions from ores, Cu and Pb concentrates in HCI with 5% acetone into methylmethacrylate, and diiectly measured in the organic phase Only Au and Fe go along with the Se, and determinations in the pgig range are possible [ I 771

420

If the sample can be atomized directly from a Pt- loop rapidly inserted into the analytical flame, detection limits can be greatly improved by electrolytical deposition of the analyte upon this Pt- loop beforehand. Above all, this separates from a salt matrix, and especially from Fe and P, which are likely to yield non-specific signals. It is important to get the dried Pt- loop into a position of reproducible geometry, to cope with signals of the distortion of the flame itself. Thus, Se has been determined in sea water, brines, technical sulphuric acid, and biologcal samples [34,171,172,178,l79). If an AdH, - flame is used with the mentioned technique, a 10-fold improvement of the detection limit with respect to the C2H2-flame is achievable, but electrical pulse- heating of the Pt- loop is essential Because the electrolyte contents and coprecipitated Cu influence the electrolysis step andthus the analytical signal, standard addition is recommended. For biological samples, 0.05 pg/g has been achieved as the detection limit [31,171,172,179]. 3.2.2 HYDRIDE METHODS In hydride methods, selenite in an acid sample solution is reduced to H,Se and separated from most of the matrix via the gas phase (The hydride forming reaction may also cause the transformation of Se-containing organics to compounds of enhanced volatility, whlch may be of use for speciation studies of non- digested samples, in combination with gaschromatographic separation) The resulting H,Se can be atomized either in a11 H,- entrained flame, in a heated quartz tube, or a heated graphite furnace The Se- atoms are usually measured in the absorption mode, but plasma emission and atomic fluorescence methods have also been applicated

3.2.2.1 Hydride generation In acid solution, NaBH, decomposes within 10 msec, but the reaction to form hydrides is usually faster [180]. Generally, As, Sb, Bi, Se, Te, Ge, Sn and Pb yield volatile hydrides upon reduction, which are subsequently atomized by thermal decomposition. As only quadrivalent Se reacts to H2Se, Se- containing organic compounds have to be decomposed, and hexavalent Se has to be converted to yield all Se in the quadrivalent form Dimethylselenide and dimethyldiselenide are sufficiently volatile to be stripped from an aqueous sample with only He as the stripping gas, bot 0,-free conditions are necessary [181]. For determination of total Se, the H,Se formed is completely recovered from the reaction solution at pH < 0.7. The decline of the signal with increasing pH was less than calculated from the dissociation constant of H,Se. At pH > 6, no signal is obtained [ 1421. Interferences in the hydride generation step are either due to consumption of the NaBH, -reagent due to oxidants or catalytical decomposition, the reaction of the H,Se formed with other species in the reaction mixture, or just foaming, which prevents the hydrides to be swept out into the atomizer, The amount of interferences largely depends on the geometry of the device as well as the reagent and acid concentrations used Concomitant organics present after insufficient mineralization may cause intense foaming, which interferes in both modes of hydride generation; it flattens the peak and decreases the signal height. As H,Se is readily soluble in water, wet connections to the atomizer severely lower the signal.

42 I

3 2 2 1 I Hydride generation i n batch In batch methods, the acid sample solution is put into a reaction vessel, NaBH, reagent solution is added, and the hydrides together with the hydrogen formed are swept into an atomization device As the sample volume added into the reaction vessel can be varied widely (20 pl - 20ml), a dynamic range of 4 orders of magnitude is achieved Peak area measurement compensates for different speeds of hydride generation, caused by different acidities of the samples, oxidants, or foaming A1 powder [ 1831, SnCI,, and Zn powder [ I841 also yield H-Se, but stripping from the reaction solution is insufficient In 0 3M HCI, TiCI, reduces Se(V1) to Se(1V). but it does iiot react with Se (IV) further [1851 Nowadays, an about 3% NaBH, / I % NaOH aqueous solution is used for batch- hydride generation in most cases Stripping of H,Se from the reaction vessel into the gas phase sufficiently rapid Hydride generation is possible from 0 3 - 6M- HCI [6] Generally, the amount of tolerable interferents is higher at higher HCI- concentrations because of the formation of chloride complexes on the one hand, and more rapid generation and stripping off on the other Within the range I M - 6M HCI, the same peak area had been achieved, but peak height i n 6M- HCI dropped about 30% [ 1861 due to dilution with the H2 generated Sweeping the hydrides in a previously evacuated atomization device can improve the performance of Se in the peak-height evaluation, more than in case of Sb, Sn and As [ 1871 Hydride generation from H$O, IS possible within the range I-10M [ 1881, but S e blanks in H?SO, have to be considered [6] Hydride formation with 3% NaBH, in batch tolerates up to 2M HNO, [ I891 But i f only I%NaBH, is used, HNO, extincts the signal even as low as I 5%, Co, Cu, Ni, Ag and Sn [1901 Fe I S reduced to Fe(I1) and thus consumes reagent I t can be masked with SCN-, citrate, EDTA, I ,lo-phenanthroline and others Reduction of Fe(1ll) prevents the formation of elemental Ni, which catalyses the decomposition of H,Se [ I891 Co, Ni, and Cu catalyze the decomposition of NaBH, i n acid solution They have to be masked with halides or other stronger coniplexing agents KI increases hydride formation, decreases the NaBH, decomposition, and precipitates Cu as the iodide This leads to a significant increase of tolerable Cu level i n presence of k1 [I801 A s a proof for the catalytic action of NaBH, decomposition, and against H,Se rection, H,Se could not be absorbed in AgNO, or in Ni(NO,), - solution [ 1841 K l is preferably added to the NaBH,reagent, because it slowly reduces Se in acid solution [ 190,191] KI also masks excess Hg, which lowers the signal [I911 Cu can be masked with thiourea [47], and N i with a nearly saturated solution of citric acid [I921 Cu, Ni, and Co interferences can be overcome by addition of EDTA [I931 Other hydride forming elements consume NaBH, reagent, but they rather interfere in the atomization step rather than in the H2Se- formation itself H,Te is metastable and hardly got in the batch technique PbH, is got from the quadrivalent state, which I S usually not encountered in the samples, especially after conversion of Se(V1) to Se(1V) In batch method, addition of Te even improves the Se determination, because because CuTe I S precipitated, rather than H,Te metastable is evolved [ I941 Precipitation of elemental Te does not interfere with the Se- determination, but elemental Bi strongly does [ 1951

422

Bi, Ge, Sn and Pb matrices can be masked by addition of additon of EDTA. Citric and tartaric acid do not improve Se in presence of excess Sb and Tef1941. Pb and Hg at ambient levels are of no influence. Nitrite interferes in the determination via the hydride, because H,Se and NOz- react in acid solution. H,Se is formed, and further reacts with nitrite in the reaction solution before it can leave the hydride generation device. H,Se is absorbed in a nitrite solution at about 88%. At nitrite concentrations below 0.2 pM, standard addition is sufficient, otherwise nitrite can be masked by addition of a 2% solution of sulfanilamide [ 1961. Similarly, addition of an aqueous solution of NH,OH.HCI prior to the hydride forming reaction prevents interferences of nitrite and dissolved nitric oxides [ 1971, which are present after any decomposition with nitric acid. 3.2.2.1.2Continuous hydride generation In the continuous flow technique, low contact time of the generated hydrides with the reaction solution lower the interferences which are caused by secondary reaction of H,Se. In practical work, the standards have to be closely matched with respect of their acid- and Fe(II1)- contents to the samples analyzed. In the continuous mode, on-line separations are possible, like GC separations of organic substituted hydrides, or LC separation of the matrix prior to hydride generation. Even on-line reduction if Se(V1) by HCI- addition and heating has been exerted [ 1981. In continuous flow, Se evolution is not interfered up to 10% HNO, or H2S0, [199]. Hydride evolution from perchloric or nitric acid alone yields only low signals [38]. Hydride selenide generation from acetate/acetic acid buffer led to very unstable signals in continuous flow generation [ 1821. The Nal3H,- reagent has also been added via an anion exchanger in the tetrahydroborate form, generating the hydrides in a heterogenous reaction on line, and transferring the HzSe along with H, through a gas-permeable PTFE- membrane towards a flame -in-tube atomizer [200]. 3.2.2.2Atomization 3.2.2.2.1 Flame-in-tube technique The product of hydride generation is directly introduced into a H,/air flame or a H2/Ozflame by means of a N,- stream. The extremely soft diffusion flame has to be protected from room air currents by a burner shield equipped with quartz end windows and Pyrex side plates, which improves the precision two-fold. In the flame, H- radicals are formed in a cloud which atomize the H,Se. The shape of the cloud of H- radicals depends on the gas velocities. Interferents accelerate the consumption of H- radicals [24,186]. In the flame, not only H,Se, but also various Se-containing compounds, like dimethylselenide, dimethyldiselenide and others can be atomized in a similar way, which is important in speciation studies, Some Se- containing organicals can be reacted with NaBH, to yield volatile hydrides, which can be detected after separated by GC in a stream of He in an H,/air- flame [181]. After hydride generation in the batch- mode, it is possible to lead the reaction products directly into the N2/Hz/airflame [41,201], or into an Ar-H,/air flame via the nebulizer of an AAS device [202].

423 To improve the detection limit, the hydrides collected in a cool trap can be pulsed into an N,-HZ- flame with a stream of N,- carrier gas The composition of flame gases has to be optimized to get into the range, where the peak height is independent of the gas composition [ 1881 3 2 2 2 2 Heated quartz cell Hydrides together with H? are swept into a T-shaped quartz tube, positioned i n the optical path of the AAS- spectrometer, which is either mounted on the burner head and heated with gas, or within an electric furnace The inner surface of the quartz plays a decisive part in the destruction of the hydrides, because i t contains a catalytic film of H- radicals [203] Small amounts of oxygen enhance the signal. due to radical formation, but not due to reaction of 0 2 with the SeH, itself [204] 700-800" is regarded to be the optimum temperature for the decomposition reaction to yield S e atoms The Se, - molecule absorbs at 334 nm and can be measured with a D,-lamp Dimerisation takes place at high Selevels and lower temperatures [205,206] In analytical practice, optimum temperature is 850-950", to get reasonable temperature of the inlet part of the tube, too [207j At too short residence time, there is incomplete dissoziation, at too long residence time, there is dimerization At optimum gas flow, addition of 5% air to the carrier gas flow sharply loweres the shape of the signal because of consumption of H i n the system to yield H,O 0 I % H, is necessary in the decomposition reaction at 800" The residence time of the gas In the cuvette is 0 1-0 5 sec [ 180,20S] Water vapour does not decline the sensitivity [206] Some authors recommend background correction with D, [62], others get accurate results without 3 2 2 2 3 Graphite furnace atomization of hydrides The sensitivity for Se in graphite furnaces is considerably lower than in flame-in-tube atomizers The reproducibility, however, is excellent The on-line atomisation approach utilises a direct transfer of hydride from the generator to a furnace at atomization temperature The generated hydrides are introduced into the internal gas line of commercial furnaces To avoid contact with metal surfaces, however, hydride introduction through the injection hole of a graphite furnace via a sealed graphite tube is favourable [209] For the determination in excess of other hydrides (like As), the. HZSewas atomized from a pyrolytically coated tube a 2650" Because of interferences from volatile organics from the matrix, background correction was applied [2 101 Whereas a quartz- tube atomization device has to be constantly run at the atomization temperature, it is easy to use a graphite furnace for enrichment and pulse atomization The hydrides are trapped in a moderately heated graphite furnace, and subsequently atomized at high temperatures The trapping also eliminates effects of variable rates of hydride evolution Extraordinary rapid heating rates and thus high peak heights could be achieved at high heating rates, e g in a graphite- paper furnace [ 194,2091 Hydride trapping in a graphite furnace is greatly improved by addition of Pd to the furnace tube wall, which is added together with the surfactant Triton X-100, to get a homogenous distribution on the tube surface 121 I]

424 3.2.2.3 Intermediate hydride trapping For extreme trace analysis, enrichment of the hydrides in a cool- trap after generation in batch, and pulse atomization, together with simultaneous separation from H, and H,O vapour, greatly enhances the detection limit and lowers the background noise [6]. The time needed for the determination, however, increases at least 5-fold. Enrichment of H,Se also separates from excess H2 formed in the hydride generation reaction. Prior to cool trapping, the water has to be removed at first, because it clogs the cool trap and readily dissolves H,Se. Desiccants, like silicagel or Mg(CIO,), partially also adsorb H,Se [ 6 ] . CaCI, as desiccant also partially absorbs H,Se [ISI]. Therefore, the gases produced in the hydride generation vessel, are led through a cool trap with propanol/dry-ice or methanolldry-ice to freeze the water vapour, then the hydrides are caught in a cool trap with liquid nitrogen, and the excess hydrogen passes. The most difficult task is to separate CO, from H,Se, which is only essential, if glow-discharge atomization is used [212]. If a GC is placed between the stripping - trapping apparatus and the AAS detector, it is possible to discriminate methyl selenides from inorganic selenium, e.g. in the analysis of natural waters [ 1811. Alternatively to purely freezing techniques, after passing a glass fibre filter for removing liquid droplets, the hydrides can be adsorbed from a stream of He on Chromosorb W in a vitreous silica tube inside a liquid N- cooled steel tube[6,213]. For desorption of H,Se, the cool trap is rapidly heated, either electrically or by immersion into hot water, leading to an atomization pulse. Alternatively, collection of the hydridem, - mixture in a high- pressure vessel, which is suddenly opened to the atomizer, also resulted in an enrichment effect [205].

3.2.3 DETERMINATION OF SELENIUM BY ELECTROTHERMAL ATOMIZATION Matrix effects, spectral interferences, and volatility are major problems in the charring step. The analytical signal, however, is nearly independent of the speciation of Se. Full presence of matrix plus, spectral interference from Fe,and additional smoke from matrix modifiers necessitates Zeeman background compensation in many cases. 3.2.3.1. General atomization behaviour As analytical lines, 196.0 nm in the far UV, and the 4- times less sensitive line at 204 nm can be used. Optimized atomization temperatures found in the cited literature are medium, and vary from 2200 - 2700"; they depend on the tube material and the apparatus used. The differences between Se(IV) and Se(V1) - signals are within experimental error [63]. Similarly, no difference in the atomization signal between Se- methionine, selenite, and selenate at the 50 ng/ml level was observed [214]. Se-methionine in pure solutions, however, can be slightly lost even in the drying stage. For this reason, very slow drying of native serum samples is essential [215]. From 0.2% HNO,, in the absence of metal cations, losses from coated graphite tubes with platform occur even at 200°, but the presence of salts tends to stabilize the Se [63,216]. From acid solutions, Se may be also lost as H,Se [217].

425 Perchloric acid in admixture with sulfuric acid severely suppresses the signal, whereas HCIOJHNO, mixtures yield higher signals than the individual acid [218]. Thus, acid contents and matrix have at least to be closely matched [ 1901. Mass spectrometric sampling of gaseous species that evolve from graphite include Se-oxides, carbides, hydroxides, the dimer, and finally the free atoms, as a function of temperature and oxygen partial pressure. In case of vaporization of pure selenious acid in dilute nitric acid from the graphite tube wall, the maximum of SeO' appears at 250°C, Se02' at 3OO0C, and the dimer at 700'C [214]. Beyond lOOO'C, free Se atoms is the sole species in the gas phase. Hydroxides are never observed in vacuum vaporization. Some metal ions, like Pd, severely influence this atomization behaviour (see later)[2 191. Se can also be atomized from a W- ribbon, which offers the advantage of very high heating rates (6000°/sec). As application, however, only the determination of Se in natural waters with Zeeman background compensation has been given [220]. The variation of the modulation frequency of the Se- hollow cathode lamps did not significantly affect the relative analytical performance, but there was significant variation in the limits of detection for different lamp geometries. A boosted hollow cathode lampconfiguration provides lower limits of detection [ 2 2 11. 3 2 3 2 Limits of deuterium background conipensation 3 2 3 2 1 General For ETA-AAS with D2- compensation, the maximum tolerable level for one atomization cycle was found to be 0 5 pg chloride, 4 pg sulphate, and 5 pg phosphate, otherwise the signal gets too low because of overconipensation Fluoride, however, does not interfere [601 If a proper separation procedure is appplied, the level of interferents is lowered, and deuterium background compensation becomes possible 3.2.3.2.2 Interference of Fe In real samples, electrothermal AAS with a D,- compensation sytem is severely limited because of spectral interference of Fe. At 196.0 nm, there is a weak Fe- emission line within the spectral bandpass, leading to negative signals [222,223]. At high Se- levels, the line at 204 nm can be used, where there is no Fe interference, but spectral interference of Ni [223]. In case of overcorrection, standard addition gives too low results, because a part of the peak disappears in the overcompensated signal, but the added amount appears. If the atomization cycle can be timed very exactly, the slight delay of Se relative to Fe atomization offers the chance to cut away the wrong signals. Thus, a delay of 0.7 seconds made correct measurement in peak height mode possible. Mo- coating suppresses the negative overcompensation-peak up to 250ng Fe[2 141. Pd, Cu and Ni at elevated levels also produce over-compensation signals in H,SO,, but not in HNO, or HCI. This overcompensation can be overcome by charring with air at 300' [224]. Use of an extremely powerful "super-lamp" with significantly lower base- line noise made it possible to use large bandpass averages over a greater area by the background corrector, which increased the tolerance level of Fe [225].

426 3.2.3.2.3 Interference of phosphorus P on the level present in biological matrices causes overcompensation effects due to molecule absorption and stray light. Especially Ni and Pd, but also Ce, Pt, W, and Zr favour the formation of P-atoms and delay of the background, resulting in lower absorption of P- molecules and thus less base-line distortion [226,227]. In the analysis of serum in presence of C u M g - nitrate modifier, interference from P and Fe could be overcome by a delay of 0.7 sec in the read period, using a platform [228]. Similarly, in platform atomization, 5 pg P could be tolerated with D, in presence of 100 pg Ni. With Zeeman compensation, however, no effect from 10 pg P + 50 pg Ni was observed [229]. 3.2.3.2.4Applications done with D,- systems D2- compensation was sufficient for the analysis of river water in presence of Ni matrix modifier [230], in acetic acid extracts from plants with Cu as matrix modifier [63], or in marine samples with low levels of P [229]. At the low Fe levels encountered, wall atomization from pyrolytically coated graphite tubes was preferred over platform atomization [63] or uncoated tubes [231]. The normal range of [Fe] in serum is 0.5 - 1.5 pg/ml, which was already too much for coated tubes/ wall atomization/ Cu-Mg modifier [223]. After dilution of the sample, an albuminiPd - modifier in pyrolytically coated tubes delayed interferents sufficiently for D,compensation [215], whereas other authors strongly recommend Zeeman compensation because of non- avoidable spectral Fe- interference [232]. 3.2.3.3.Combination with liquid-liquid extraction After liquid-liquid extraction for separation from interferents, the organic phase can be directly injected into the graphite tube. Difficulties may arise due to undissociated evaporation of Se-chelates. Using dithiocarbamates or similar reagents, direct injection of the extract with diethyldithiocarbamate from acetate buffer pH4 into CCl, yielded best results [ 1 141. As the reagents are not completely destructed below 600°, a modifier to stabilize Se has to be added. Ni matrix modifier and H,O, as'ashing aid have to be added as aqueous solution prior to atomization. Pd as a modifier, however, can be added before the extraction procedure, because it is coextracted with the Se, like Cu, whereas Ni remains in the aqueous phase [224]. Gas stop in the charring step at 700" increased the reproducibility in non-coated tubes [233]. From liver digests, Se was extracted with ammonium pyrrolidine dithiocarbamate (=APDC) into CHCI,, injecting the organic layer into the furnace, preferable into pyrolytically coated tubes [234]. A graphite furnace was also used as detector of liquid chromatography of APDC complexes at C18- bonded silica gel columns. after elution with methanol, the organic phase was not used directly, but the fractions were evaporated, and redissolved with HNO,/Ni [86]. The extract of various piazselenoles in e.g. toluene can also be directly used for electrothermal atomization. The substituents on the piazselenole molecule influence the volatility of the undissociated species. Best results were achieved with the nitro- derivative, and worst with the unsubstituted [235].

427 Extraction of the 4-chloro- piazselenole into toluene and injection of the organic phase was applicated for the analysis of biological matrices [236] If addition of Ni- matrix modifier I S wanted, it has to added as aqueous solution into the tube, because it is not coextracted 3.2.3 4. Stabilizing cations and matrix modifiers 3.2.3.4.1 General Losses of selenium from 0.2% HNO, solutions from coated graphite tubes with platform occur even at 200", as indicated with radiotracers. Se (VI) is less volatile than Se(I1) than Se(IV) [216]. Salts generally prevent volatilization of Se in the charring step Some of them act very effective (like Pd, Ni, Cu, Pt) and thus are added to various samples as matrix modifiers. The amount of stabilization, however, also depends on the tube surface and the kind of acid used. Additional smoke from the addition of these modifiers requires effective background compensation, which makes it necessary to use the Zeeman- effect, or Smith-Hieftje compensation. Rapid atomization from a platform inserted into the tube helps to atomize volatile elements before evaporation of the matrix, which is also applicable to atomization of Se [ 18,2281, In 0.2% HNO, and coated tubes with platform, even addition of 1% NaCl stabilizes all forms of Se in the charring step; selenate is completely stabilized up to 900". and selenite at 70-80% up to 800". If losses occur during the drying step, however, they cannot be prevented by salt addition, but only by very slow drying [216]. By means of 75-Se tracer, inorganic Se was quantitatively stabilized during charring from 3% HNO, solution on pyrolytically coated graphite tube walls by Sb, Cd, Mn, Mo. Ni, KJ, KJO,, Ag, Th, TI, Zn, and Zr. For charring of organoselenium compounds in urine under the same conditions, however, only Mo, Ni and Ag were effective. Contrary to other authors, Cu and Fe were found to stabilize inorganic Se only partially. Also, J als KJ enabled ashing up to 1000". In urine, KJ, Mn, Zn, Th, TI and Zr did not stabilize [237]. In a similar investigation using non-coated tubes and peak height evaluation, selenite in 5% HN0,- solution was stabilized during the charring step by Ni, Cu, Zr, Pt and Pd. To the contrary, Mg, Fe, Na, Al, Au, Ca, Mn, Cr were not effective [238]. 32342Pd Pd is partially maintained in the furnace up to 1700" [224] Mass spectrometric sampling of gaseous species that evolve in the atomization cycle from pyrolytically coated graphite tubes revealed, that Pd inhibits the Se-dimer formation i n all cases it is just introduced, but not thermally pretreated Simul-taneously, Pd inhibits the formation of hydroxides and lowers the amount of vapourized monoxide SeO? is inhibited by Pd only, if Pd has been thermally pretreated It traps SeO, from the gas phase to yield a PdlSelO compound, which turns to be Se-Pd at 1 200", and atomizes as such [2 191 In simple Se(IV) solutions, Pd stabilized Se during charring i n coated tubes t i l l 1200" P391 For the analysis of Se in serum, Pd modifier was pre-iiijected into the furnace, dried, and then the 111- diluted sample was added [239] Addition of 0 I % albumin together with Pd under certain circumstances doubled the signal of Se from serum samples in coated tubes The albumin was converted into graphite, proved by scanning electron microscopy Similarly, polyvinylalcohol was converted into graphite, which also improved Se- sensitivity in coated tubes to a lower extent 40 pglnil

428

Pd were found sufficient, and 150 pg/ml Pd caused the albumin to precipitate. At the 1.5% level of albumin, however, severe background absorptions were seen with D2 compensation [215]. In TaC- coated graphite tubes, peak heights and peak areas reached maximum values within the range of 1-10 pg Pd per atomization cycle. The TaC- coated tubes could not used without Pd, which yielded additional thermal stabilization. After extraction of Se with DDTC, 3 pg Pd were found to be optimum for charring of the DDTC- extract at 700" in the TaC- coated tubes [224]. For charring micro-amounts of biological material inside the tube, Mg- nitrate can be added as an oxidant, if it is possible to cope with the additional light scattering during the atomization by means of the Zeeman- effect. In the analysis of blood and serum, a combination of PdMg - nitrate allows the use of higher pyrolysis temperatures, delaying the background of e.g. CaHPO, matrix [227]. After separation as the piazselenol and stripping into the aqueous phase, Ni and Pd were equally sensitive as matrix modifiers. Platform atomization yielded greater sensitivity and better reproducibility than wall atomization from pyrolytically coated tubes [240] 3.2.3.4.3Ni Ni forms stable NiSe through the reaction between SeO, and Ni. NiSeO, is only stable to 300" in an 0,-free atmosphere, and NiO decomposes at 400". The presence of solid Ni is essential for Se- stabilization; in solid sampling, matrix modification with NiO was not effective. Glucose changed the nature of the surface and lead to low recovery of Se in spite of Ni presence. Ni as chloride was less effective as modifier because it was not reduced to metallic Ni fast enough, whereas NiNO, is easily reduceable by H2 also [217]. In 2.7% HNO,/ 0.1% Ni as matrix modifier, 1000 times excess As, Sb, and TI, 5000 fold excess Pb, 7500 fold excess B and 10000 fold excess Cd did not interfere [241], and D2compensation was sufficient. In the analysis of digests of geological material, however, variable signal depression lead to wrong results of standard addition technique with deuterium background compensation [233]. Therefore, separation from Fe, P and other possible interferents was done by volatilization of the analyte in a stream of oxygen [241]. Zeeman background compensation enabled measurement of Se in digests of geological material by high- speed atomization from a platform in a pyrolytically coated tube, in presence of IOOpg Ni per atomization cycle. Evaluation via calibration graph as well as standard addition were possible [222]. In the analysis of biological matrices, 100 pg Ni strongly accelerate the atomization of phosphate. 10 pg P were easily compensated with Zeeman- effect, whereas with D2 as modifier, upper limit of P was I pg P per atomization cycle [229]. 3.2.3.4.4Ag The efficiency of Ag as a matrix modifier is about like Ni, but Ag is presumed to precipitate in biological matrices as the chloride. When 20 pg Ag or Ni were present, loss of Se did not occur at 1000" charring. The atomization was improved to 140% [242,243].

429 3.2.3.4.5 Pt Pt in its elemental state is not volatilized up to 2400" [225]. Addition of Pt allows charring of Se up to IIOO", but it slightly delays the appearance of Se, and lowers the sensitivity, both with platform and with wall atomization. Significant Pt amounts stay in the tube and cover the surface with smll pellets of Pt- metal. As Pt does not stabilize organic compounds, it is often applied in admixture with Ni [226] ( 1 5 pgNi + 60-120 pg Pt) or Cu. Only Ni, Ag and Mo have been shown to reduce the volatility of organoselenium compounds. A pyrolytically coated tube with off the wall atomization gave better sensitivity than an uncoated tube and was less prone to complications from the unburnt residual carbon, which might arise in the platform [23 I]. In the analysis of blood and other biological fluids, addition of Pt/Ni matrix modifier favours the formation of gaseous decomposition products of phosphate and changes the absorption-time- profile of Fe. Ni and Pt together favour the formation of P atoms, which can be measured at 213.6 nm [226]. In a Mo- coated furnace, the use of a Pt/Ni matrix modifier resulted in complete separation of the appearance time of the Fe- signal and the Se signal [214]. 32346Hg Mixtures of HgCI,/PdCI, as a modifier were investigated in pyrolytically coated tubes up to 0 10% Each of them improved the signal, but they were superior in admixture Whereas Hg alone accelerated the peak appearance, Pd had nearly no effect, and the mixture even delayed, because Hg favours the atomization rate At 0 10% modifier and 1 100" charring, Al, Fe, Ca, Mg, K, and Na did not interfere when present in usual concentrations in digests of fly- ash samples, with SMITH-HIEFTJE background compensation [244,245] Among the anions, the appearance temperature increased from SO, to NO, to CI 32347Cu Cu as the nitrate has early been found to stabilize the charring of Se like Ni up to 1250" ~411 Its effect upon the absorbance signal from Se is much more pronounced for wall atomization at the 10 ppm and 100 ppm level of Cu, but equal to platform atomization at 1000 ppm Cu in the peak height mode After combustion of plant materials in oxygen, Cu as matrix modifier in dilute acetic acid was proposed (10 pg per atomization cycle) Acetic acid improved the atomization signal at > IM concentrations, but significant differences in performance between different batches of tubes were noted [63] 3.2.3.5. Tube material and tube design For wall atomization, pyrolytically coated graphite tubed are superior for the determination of Se. In the use of platforms, however, optimum conditions seem to depend on the kind of sample and the kind of apparatus used.

430

Soaking with W- solution lead to a change of the surface structure because of carbide formation, but to no improvement for the determination of Se from dilute nitric acid solutions[24 I]. Coating of graphite tubes with Nb or Ta resulted in better reproducibility and accuracy, at equal sensitivity for Se in presence of 50 pg Ni [233]. TaC - coated graphite tubes lead to additional thermal stabilization of Se in presence of excess Pd, but also caused some memory effects [224]. In boron-nitride coated tubes, retention of S e in HNO, was possible up to 500", but losses of Se(IV) from 0.1M- HCI occurred to 20-30% within the range of 200-500". The boronnitride -coating, however, was destroyed after 30-35 firings [216]. Coating with Mo was achieved by injecting 4 times 5 5 pI of 5% Mo solution and heating to 2500". The Mo- coating significantly suppressed the interference of Fe resulting from overcompensation of the D,- signal, due to delay of the Se in the peak height mode, in presence of P t N matrix modifier [214]. 3.2.3.6.Use of auxiliary gases Addition of 10% H, to the inert gas (Ar) reduced the peaks of organic molecules during graphite- furnace atomization (like pentane, hexane, methanol, ethanol, and chloroform; except aromates) and doubled the peaks of Se- containing organicals. Thus, D,compensation was sufficient to use a graphite furnace as a detector in gas-chromatographic separation of volatile selenium- containing species collected from ambient air [246]. 5% H? in Ar was used to reduce the Pt- modifier to its elemental state, which made i t far more effective [225]. Ashing of biological samples with O? inside the graphite tube may be risky, because Se forms volatile oxides [247]. Charring of dithiocarbamate- extracts at 300" in air doubled the signal of Se in TaCcoated tubes, because of improved destruction of the reagent [224]. CO was used as an auxiliary gas during charring up to 1000" at a flow rate of 0. I llmin to remove O? from the furnace. This prevented the formation and subsequent atomization of iron oxide, which interfered in the background correction of the Se- signal with D2 [225]. 3.2.3.7. Solid sampling and slurry atomization For Se, solid sample atomization is only reasonable since the introduction of powerful Zeeman-background compensation systems, and availability of sufficient homogenous solid standard material. As best and universal modifier, a mixture with graphite powder was recommended. For Se, however, few reference materials were available, and the detection limit was not sufficient in any case [248]. Introduction of micro-amounts of solid samples cannot only done rized about 60pg sample from a homogenized tablet of geological material, which was introduced into the tube together with the carrier gas [249]. For the analysis of Se in metal chips, about I mg of the alloy was inserted into the furnace and atomized. For Se, an atomization temperature from Ni-base alloys of 2600" was needed, and a detection limit of 0.2 pg/g for 1 mg sample could be reached only Standardization had to be made against doped alloys of similar composition [2SO], because the sample was not completely vapourized.

43 I In solid sampling of liver homogenate, the Ni/Ag- matrix- modifier was added to the sample before freeze drying and homogenization, to ensure sufficient mixing For Img sample weight of liver, a detection limit of 10 ppb could be achieved [247] In slurry atomization, the solid sample is mixed with a fluid for reasons of dilution, and pipetted into the graphite tube. For analysis of coal and coal fly ash samples, a solid slurry was prepared by ultrasonic mixing. Calibration was done against similar NBS- standard materials. Atomization was done from a platform inserted in a standards grooved tube. The solid was diluted with slurry was 5% H N 0 3 / 0.04% Triton X-100. A Zeeman compensation device was essential. Se was more difficult to determine than As, Pb, and TI. At 1900°, severe matrix absorbance appeared, which necessitated to atomize as low as 1 8 5 0 O . Pd as modifier sharpened the peak, but did not improve the accuracy [25 I]. For the analysis of milk powder, 0.5 g sample was mixed with 3 ml H,O, and 200 pl of this slurry added to lml RhiMg- matrix modifier (0.4% Rh(NO,), + 0.25% Mg(N0,)2.6Hz0 + 0.4% HNO, + 3% Triton x-100). Charring was done from a platform, Mg- nitrate acted as ashing agent. Evaluation was done via peak area and standard addition [252].

3.3 ATOMLC EMISSION SPECTROSCOPIC METHODS

3.3.1 INDUCTIVELY COUPLED PLASMA

- DIRECT ASPIRATION

The main atomic line of Se at 196.026 nm has a rather high excitation and ionization energy and therefore tends to peak rather high i n the plasma [253], nearly like an ionic line. The increase of observation height in the pure Ar plasma, however, lowers the signalibackground ratio, because of strongly increasing background Similarly, the signal/ background ratio is worse in the high power-range and upon addition of 5% N, to the plasma gas [254]. For Se, no improvement by means of switching the plasma excitation frequency from 50 MHz to 100 MHz has been achieved [255]. As far as spectral overlaps are concerned, the main emission line of Se at I96 026nm may be interfered by an Fe-line at 196.059nm, if the resolution of the spectrometer is insufficient [2 561. Detection limits of Se emission in the ICP using direct aspiration, are within the range 10-70 ngiml [255,257] and thus insufficient for the analysis of natural waters and digests of biological material [258]. Direct coupling of the ICP to a liquid chromatographic system with a flow rate matched to the usual nebulizer uptake, lead to a detection limit of about 0.1 pg Se at the insensitive line at 203.985 nm [89]. 3.3.2 DIRECT CURRENT PLASMA

- DIRECT ASPIRATION

In the DCP, matrix effects upon the emission signal of Se at 196 026nm by ionization enhancement are negligible The signal was regained in presence of up to 1 5 g/l Na, and Ig/l AI, Ca, Fe, Mg [259] The precision, however, decreases significantly at lower concentration levels [259] The detection limit of about 0 25 pg/ml [260] is insufficient for environmental samples

432 3.3.3 HYDFUDE METHODS The hydrides of interest, together with a constant amount of hydrogen, are produced in a continuous- flow system, and introduced into an analytical plasma without further nebulization. Evolution of a smooth stream of hydrogen without bumping and splashing is essential. Low spectral background and lack of nebulization losses improve the detection capability about two orders of magnitude, which is sufficient for many environmental samples. The multielement capability of the instrument is limited to the hydride-forming elements, only Se(IV) is monitored, and all interferences of hydride evolvation (see chapter 3.2, 3.8) have to be considered. The AES signal stabilized about 1 min after introduction of the sample to the hydride system [261]. The optimum range of acidity of the sample depends on the exact geometry of the device, and the carrier gas stream used. Starting from HNO, gave a slightly lower sensitivity compared to HC1 [262]. For the stabilization of the gas introduction into the plasma, in the simplest approach, the spray chamber of the conventional ICP- nebulizer was replaced by the mixing chamber of the hydride system [257]. Alternatively, the hydrides and hydrogen have been separated from the reaction mixture via a 2m or 5m long silicon rubber tube placed in the stream of Ar-plasma gas [263]. A long and narrow gas- liquid separator was filled with Pyrex beads, which provided smoother separation of liquid and gas [264]. If the hydrides are introduced into a DCP, half of the Ar flow is directed through the hydride generator to carry the evolved hydrides into the plasma, while the other half passes its normal way through the nebulizer system. The reaction gases (mainly hydrogen) are dried by passing CaCI,, and delayed in a suitable delay tube before nebulization, to provide a smooth gas stream. To obtain lower detectable concentrations, hydride evolution from a lOml sample batch is also possible, leading to a detection limit of 15 ng Se. Matrix Ni was successfully masked with phenanthroline [260,265]. 3.3.4 OTHER EXCITATION TECHNIQUES Excitation of electrically conducting solid samples in a DC-arc between crbon electrodes at 30 A suffer from the short wavelength of the Se- main emission line at 196.026 nm. A stream of argon shielding gas (7.5 l/min), long exposure time (65 sec) and optical grating of 1200 grooves/mm were necessary to obtain reasonable signals on the spectrographic plate. This method was used to determine Se in various sulfides after mixing with equal amounts of ultrapure graphite [266]. Analytical lines which can be measured without a vacuum are found in the visible region, but the energy required for their excitation is so high that they do not appear in arc and spark spectra. If the sample is dried upon an Al-cup, which is subsequently used as the target of a hollow cathode discharge of 240W, ionic lines at 444.62nm and 444.95nm appear, which can be easily detected spectrographically, with a detection limit of Ing/ml, achievable in human serum [267]. Contrary to other excitation sources, in a microwave induced plasma (MIP) within a flow of 1.1 I/min He, the Se- lines at 206.279nm and at 203.985nm emit signals as well, but applications for real matrices are not given [268]. Vaporization from a Ta- strip into an Ar-fed MIP allows to evaporate the solvent after solvent extraction, and subsequent pulsed vaporization of the analyte [ 161. After pressure

433 decomposition, the detection limit of the MIP, however, was insufficient for the determination of Se in fish and vegetables Molecular emission cavity analysis (MECA) uses the broad emission of SeO, Se2, and S e 0 2 in the range 330-530 nm, with a maximum at 413 nm Molecular eniission is achieved by injecting the sample into a steel cavity placed in an air-H1-N, flame, and cooled on the backside As many metal ions interfere, and H2Se is much more sensitively detected than Se- oxyanions, the sample is introduced after hydride formation For reasons of enrichment, the H2Se can be caught in a cool trap, and evolved by immersing the trap into hot water [269,270] 3.4 DETERMINATION OF SELENIUM BY MASS- SPECTROMETRIC METHODS

3.4.1 GENERAL Five selenium isotopes are naturally abundant: Isotope Natural for more details see 3.6.1 abundance 74-Se 0.87 YO 76-Se 9.02 YO 78-Se 23.52 % 80-Se 49.82 YO 82-Se 9.19 YO Mass spectrometric techniques have the advantage of large multielement capabilities, relatively simple spectra, and offer the possibility of isotope ratio measurements and isotope dilution analysis. 3.4.2 ICP

- MS

A horizontally mounted ICP - torch serves as sample introduction device for a quadrupol mass spectrometer [271]. Since this technique is rather new (first papers published about 1980), only few applications for the analysis of Se in real matrices can be found in the present literature. In case of Se, isobaric interferences of neighbouring elements hardly occur. Arsenic is monoisotopic at 75 m/z, where there is no S e isotope, and bromine is hardly found as a cation [272]. In matrices with high chloride content, or in case chloride containing acids have been used for sample dissolution, polyatomic ions containing CI are encountered in the mass spectrum [273]. Additionally, Se coincides with background peaks of the Ar-plasma itself, emanating from various Ar and C1 isotopes. 76 Se 36Ar - 40Ar 77 Se 40Ar - 37CI 78 Se 38Ar - 40Ar 80 Se 40Ar - 40Ar Only 82 Se is free from such interferences 12721. Spiking the sample solution with 10% propan-2-01 or introduction of 3% N2 into the nebulizer gas flow reduces these polyatomic signals; maybe there is competitive formation of Arc', ArO' and ArN' [273].

434 In ICP-MS of positively charged ions, without any matrix, a detection limit of only 0.8 pg/l could be reached within 10 sec single ion monitoring, according to its high ionization potential of 9.75 eV [274,275]. In multielement analysis of natural lake waters, the detection limit of only 3pg/1 Se was insufficient ( 2 sec integration) [275]. In acetate buffer leaches of soils, monitoring of 78-Se and 82-Se resulted in a detection limit of 5 pgil in a 20 sec total measurement time. Noteably, the sampling for the mass spectrometer was done at 22 mm above coil, which is unusually high, but reasons were not given. However, the precision of the recovery of spikes to the acetate buffer was only 128% [276]. Direct use of non-digested urine and serum samples into the ICP-MS yields polyatomic fragments. At mass 82, where there is no Ar-Ar or Ar-CI background like for the other Seisotopes, the results were anomalously high, in comparison with ICP-AES and AAS determinations. This may be due to the CCI" ion, which also appears in the spectrum of pure trichloro-acetic acid [277]. In the direct analysis of red blood cells, 74-Se coincided with FeO' [278]. In the negative ion scan, there are considerably fewer background peaks. The background peak at 78-Se disappears, but at 80-Se it is still present [279]. Introduction of hydrides into the ICP-torch reduces the number of interfering molecular ions in the mass spectrum, e.g. oxides and chlorides, but the start from Se(1V) and interferences in hydride generation have to be considered. A debubbler to remove excess air prior to the mixing zone greatly improves the results [278,280]. ICP-MS via the hydride was applicated in hunian metabolic studies employing stable isotope tracers [278]. 3.4.3 OTHER MASS SPECTROMETRIC METHODS

In spark source mass spectrometry, the electrical conductivity of selenium in most of its compounds as well as in most of the matrices of interest is insufficient. It needs to be electrolytically preconcentrated at gold- electrodes after suitable sample decomposition by wet chemical methods, or reduced by means of hypophosphoric acid after addition of gold chloride spike. The resulting gold containing Se, Te and some other trace metals, is directly sparked. For quantitation, isotope dilution with 78-Se [45] or with 82-Se [281] has been applied. Thus, analysis of Se in coal, in steel and Ni- based alloys down to 0 1 pgig is reported. Both selenite and selenate evaporate as negatively charged ions from a hot metallic ribbon, which has been used to determine Se- traces in natural waters ("negative thermionic MS"). Selenite and selenate could be discriminated after separation by anion exchange. Within the range pH 1-12, no isotope exchange between selenite and selenate could be detected [282]. 3.5 ELECTROCHEMICAL TECHNIQUES 3.5.1 PREFACE As selenium can occur in various oxidational states, For pure aqueous acids or alkalis, the normal potential of the respective redox reactions are [283]: 1M acid: H,Se <-0.40 V> Se <+0.74 V> H,SeO, <+1.15V> Se0,'I M alkali: Se'. <-0.92 V> Se ~ 0 . 3 V> 7 SeO,'. <+0.05 V>SeO,'-

435

Common to all techniques for real samples (except for detection of compounds i n HPLC) is to start from Se (IV) The formation of compounds with the electrode material on its surface can lead to shifts in the peaks obtained, as well result i n electrochemically inactive compounds Direct applications of electrochemical methods have been given only for saline waters, and for soil extracts obtained with neutral salts 3.5.2 DETERMINATION OF SELENITE IN AQUEOUS SOLUTIONS 3.5.2.1 Mercury electrodes 3.5.2.I . 1 Dropping mercury electrodes In 0.2M - HCI, the polarographic wave at -0.54V vs.SCE, deriving from the reduction of HgSe to Hg + H,Se, yielded a non- linear calibration graph Pb, Cu, and Fe influence both peak position and current, because of formation of selenides at the electrode surface [284] In 0.1M- HCIO, as well as in O.1M- KNO,, excess of Pb, Cd, and Cu decrease the Sepeak at -0.61 V vs. SCE, without change of its shape [285]. Even at pH=I, Pb can be masked with EDTA [286]. In weak acid ammonium sulfate, nitrate interferes, and has to be removed, eg by short warming the sample with ethanol [ 1291. In 1M NH,CI or NH,- acetate soil extracts, in'the range pH 5 4 - 9.0, an analytically useful peak could be obtained within the range - 1 . I V to - I .6V vs SCE. The best separation of the selenium current peak from that of the supporting electrolyte was obtained at pH 8. with peak potential of -1.34V versus SCE, and a detection limit of 5ng/ml. In this buffer solution, Fe, Pb, and Cu do not affect in 1000-fold excess, and Co, V, Te, Cr(lIl), Mo do not affect in 100 fold excess over Se. The interference from Zn is removed by addition of EDTA, which simultaneously causes e peak from the reduction of Pb-EDTA at - I .2V, but in real samples the Pb- concentration is usually too low to interfere Among the organics, at the potential and pH conditions used for the reduction, the most likely interferents are expected to be simple organic compounds containing carbonyl and carboxyl groups Only maleic acid was found to interfere, and could easily be removed by mild acid hydrolysis [284]. In acetate/borate/phosphate buffer pH 4.0, the second peak of the Se- reduction at - I .3 I V vs. SCE could be used in presence of large excess of Cu, Pb, and Cd [285]. In order to avoid precipitation of selenides, formation of selenosulfate from selenite was achieved by adding Na-sulfite to the acidified sample solution, with subsequent adjusting to pH 7-8 [ 129,2851. At pH 9-1 1 in ammonium sulfite solution, Se can be measured by reduction of selenosulfate at -0.95 V down to 5 ngiinl [129]. 3 5 2 1 2 Stationary Hg

- drop electrode

In a first step, Se is deposited and enriched by electrolytical reduction at the working electrode Cyclic voltammograms show, that a solid HgSe - film develops on the drop surface, which dissolves at a more cathodic polarization [287] In the presence of halide ions, the cyclic curve is modified, and the wave shifted to more negative values From selenous solutions containing halide ions, a selenium compound is deposited on the surface of the Hg drop, which transfers no electricity, but is cathodically reducible, termed HgSe

436

The observed shift in the peak potential is due to complex formation [287]. At increasing plating time, an additional peak at more negative potential appears in cathodic stripping, which is less at lower pH, and does not occur at pHc4.2. The sum of currents of the two peaks is constant, and equal to the intensity of the peak when only one is present. This is interpreted by the formation of a mixed deposit of Hg-Se", and the formation of Se" by reaction between H,SeO, and H,Se [288]. In presence of Cu, a selenide presumably containing Cu(1) is formed during the electroplating step, instead of the Hg-Se compound [289]. Anodic stripping voltammetry The position of the anodic stripping peak strongly depends on the pH of the solution. The determination is interfered from many elements deposited together with Se, and also from compound formation at the electrode surface [290]. In the analysis of milk powder, a matrix very low with respect to trace metals, the direct anodic stripping voltammogram of a decomposition solution with HCIO, yielded high background currents, which could be reduced by fuming with HCIO, [291]. Increased sensitivity could be reached by scanning in the differential pulse mode, with an amplitude o f f 20 to 50 mV. Cathodic stripping voltammetry Besides As, Se, and Te, also V, Cr(VI), Mo and W have been determined by this technique [292]. Cyclic voltammetric studies at a stationary hanging mercury drop electrode in acetate buffer in the range pH 3.85 - 5.55 shows three electrode reactions. At -0.3 V vs SCE, a broad and completely irreversible peak appears, which is due to the reduction of selenite to elemental selenium, and which is pH dependent and temperature - dependent. In cathodic stripping voltammetry, the two other peaks, at -0.75 V and at -0.90 V are utilizable. The third peak at -0.90 V is not observed at pH 4.2/30 sec plating time, but increases with increasing plating time. This is explained by both the formation of a mixed deposit og Hg-Se", and the formation of Se" by reaction of selenite with hydrogen selenide [288]. The selenium peak in the cathodic stripping voltammogram is shifted with pH, beause of participation of protons on the reduction to H,Se. The dissolution current is proportional to the electrode surface, which indicates, that the reduction takes only place at the electrode surface [293]. Increase of pH reduces the peak current, and shifts the dissolution peak to more negative values. At pH '8, no cathodic dissolution peak was observed [294]. The cathodic stripping voltammetry of Se in 0. IM HCI yielded a non- linear calibration graph for to 10"M Se. Variation of the deposition potential in the range of +0.05 to -0.30 V VS. SCE was of low influence [295]. The presence of metal ions may shift the stripping peak to a more negative potential. The determination of Se is interfered by Pb and Cd, which strongly suppress the peak at 1 mg/l already. The suppression is less in presence of some Cu, when deposition at -0.3V is used [36]. Zn and Cd make no peak themselves, but they severely suppress the Se- signal [294]. In dilute HNO,, 20 - 100 fold excess of selenide forming cations can be tolerated, but Pb and As interfere [296]. Arsenic forms a new peak at -0.3V vs. Ag/AgCl along with Se. In ammonium sulfate at pH 4.5, addition of Cu (up to 1 pg/ml) enhances the deposition of Se at Hg, and shifts the peak potential to more negative values. During cathodic

437

stripping, presumably a Cu-Se compound is dissolved within the range -0.8 1 to -0.84 V. Like Cu, also Bi, Ag, and Au shift the stripping peak potential to more negative values, with respect to the Hg-Se peak, which leads to peak enhancement and to peak splitting at deposition potentials more positive than -0.5 V. From the position of the peak potentials it can be concluded, that Cu-Se is the most stable compound investigated [290]. Whereas the cathodic dissolution peak of selenium in dilute H,SO, and H,PO, is found at -0.51 V vs. SCE, it moves to -0.704V in IM- ammonium sulfate/0.4M- EDTA/pH 4, and to -0.81 V in 0.3M KNa- tartrate/0.2M-EDTA/pH 6 [294]. Interference from Cd, Zn, As, Cr, Pb and W was minimized by utilizing ammonium sulfate/EDTA/pH 4 as supporting electrolyte. Cu, Sb, Ti and TI in excess still interfere. The surfactant Triton X-100 at > 1 O-'% completely suppresses the peak [294,295]. Cathodic stripping voltammetry in ammonium sulfate /EDTA at pH 2.3 in presence of 2 pg/ml Cu, enables the specific determination of selenium and tellurium in presence of large excess of each other [293]. Cu concentrations of 3 pg/l or more enhance the height of the cathodic stripping peak. In presence of Cu, peak height and peak potential largely depend on the deposition potential At deposition potentials more positive than the formation of Cu?Se, the stripping peak shifts to more positive potentials, and the peak height decreases [289,297]. Cathodic stripping of Cu-Se yields a narrower and higher peak than for Hg-Se. Optimum sensitivity has been achieved at pH 1.6 and 40 pM Cu [289]. In dilute HCI, Fe, Pb, Zn, Cd, and Te decrease the cathodic stripping peak; Zn and Pb can be masked with EDTA. After deposition at -0.35V vs.Ag/AgCI, the cathodic stripping peak is obtained by sweeping the potential of the working electrode to -0.9 Vwith a speed of 10 mV/sec [297]. Addition of Cd caused a decrease of the stripping peak of Cu,Se, while 2 new peaks were formed, due to Cd reduction, and the cathodic dissolution of CdSe. Cd can be partially masked with EDTA [297]. Cathodic stripping voltammetry at the hanging mercury drop electrode can be directly applicated to the analysis of drinking water. Ammonium sulfate as the supporting electrolyte, as well as EDTA to mask interfering cations, are added to the sample, the pH adjusted with sulfuric acid to pH 2.2, and 0.1 mg/l Cu to improve sensitivity. After 5 min plating at -0.254 vs Ag/AgCI, a detection limit of 1 pg/l at the stripping peak at -0.68 V could be reached. Sulphide strongly interferes [298]. For the determination of Se(IV) in sea water, only 1/1 HCI and Cu have to be added to adjust to pH 1.6. After 15 min of deposition at -0.4 V, down to 0 7 ng/l Se could be detected [289]. After volatilization of Se from most of the matrix in a stream of oxyge, and dissolution from the cool finger with dilute HCI, Se was determined by cathodic stripping voltammetry with ammonium sulfate/EDTA pH 4.5 as supporting electrolyte, and addition of some Cu [621. Direct use of the digestion solution from biological materials (liver, rapeseeds) resulted in no peaks in cathodic stripping voltammetry because of high background. Separation via extraction of a piazselenol with subsequent wet ashing of the extract lead to suitable sample solutions ready for the determination. Without the evaporation of the organic solvent, however, resulted in sensitivities less than 30 % with respect to the aqueous sample [26]. For the determination of selenium in biological materials, Se was separated and enriched from the acid digest by sorption upon an anion- exchange resin in the acetate form at pH 3, prior to either cathodic stripping or anodic stripping voltammetry [36].

438 3.5.2.2 Au- electrodes

In anodic stripping voltammetry of selenium from a Au-disk electrode in 1M- H,SO,, stripping peaks appear at +0.64V, +0.86V, and at -1.03 V. Hg interferes [299]. Similarly, in 0. IM-HCIO,, three anodic stripping peaks are observed for large quantities of deposited Se. At the beginning, and at low concentrations, approximately a monolayer is deposited. Thus, the anodic stripping of very small quantities of Se following deposition at a low flux yields a single anodic peak at 0.8V. Further electrolytic reduction of Se leads to irreversible diffusional transport of Se into the electrode, forming a Au-Se alloy of unknown stoichiometry [300]. In the anodic stripping voltammogram following deposition at high fluxes of Se on the Au- surface, two additional peaks, at +0.63V and at + l . 15V are obtained [300]. These peaks can be interpreted as due to bulk Se, adsorbed Se, and as intermetallic Au-Se compound of unknown stoichiometry. Only the adsorbed amount of Se is analytically usefull. To enable only formation of adsorbed Se upon the electrode surface, a deposition potential of 0.15V vs. Ag/AgCI is preferable over the deposition at more negative values. The optimum deposition potential at the Au-electrode is thus far more positive than the -0.35 V for the mercury drop electrode for the same sample solutions. This also reduces the extent of codeposition of interferents, such as Cu and Pb [36]. The Au- electrode was pretreated prior to each experiment by polishing the surface, and preconditioned in 0.2M- HCIO, by applying alternate cycles. However, difficulties in obtaining reproducible Au electrode surface area from one experiment to the other arose. The anodic stripping procedure at the rotating gold electrode equalled the sensitivity of the cathodic stripping procedure at the mercury electrode, but the reproducibility was worse [36]. To make the Au- surface continuously renewable, the Au can be plated upon glassy carbon prior to each run. Detection limits were 4 times less than for pure Au electrodes, but only I@'M Au was needed in solution, which made it rather cheap [301]. 3.5.2.3 Graphite and carbon electrodes At a graphite-pin-electrode, made a mixture of graphite and polyethylene- powder, the reduction of selenite gives two polarographic waves, within the range 0 to -0.2 V VS. SCE, and at -0.6 to -0.8V , which correspond to the reduction of selenite to elemental selenium, and of elemental selenium to selenide. At electrolysis in the range of the more negative maximum, the electrode surface is partially covered with red selenium, another part precipitates as colloid near the electrode in the solution. This amorphous Se is electrochemically incative. Its relative amount incrreases with pH, and is at maximum at a plating potential of -0.3 to -0.4V [302]. Increasing acid concentration increases both the anodic and the cathodic stripping peak of elemental hexagonal metallic selenium. During electrolysis in concentrated HCI, however, elemental chlorine formed at the counter electrode interferes by oxidation of selenite to electrochemically inactive selenate [302].

Anodic stripping voltammetry At a graphite disk, impregnated with wax polyethylene = 3 I , after electrolysis at -0 6 V vs SCE, anodic stripping voltametry can be performed in dilute HCI, which results in a peak at +O 22V for selenium Cu is added to increase the electroactivity of Se deposits because of formation of an intermetallic compound, but produces a second peak at -0 15V To obtain maximum sensitivity, Cu Se must exceed 30 1 [303] Acidic electrolytes give much better sensitivity for Se than neutral or basic ones After 1 min depositlon time a

439 detection limit of 0 1 pg/l was achieved for pure solutions [304] Upon the addition of some Cu, the peak of selenite ion increases and moves towards a more positive potential, which suggests the formation of an intermetallic compound At large excess of Cu, the anodic stripping peak for Cu appears at about 0 (vs AgIAgCI), which overlaps the Sepeak [304] Matrix silver delivers a peak at +O IZV, and has to be removed beforehand, whereas Pb increases the Cu-Se signal, and Fe and Mn are of no influence [303] Cathodic stripping voltammetry: At a graphitelwax electrode plated with Hg, contrary to the Hg drop-electrode, it is possible to determine Se by cathodic stripping in acid as well as also in alkaline sample solutions (> IM- NaOH), which masks Cu and other metals The sensitivity of the cathodic stripping peak is effected by the deposition potential and the time, the scan rate, and the thickness of the Hg- film. As optimum conditions, plating at -0.10 to -0.30 V, and potential sweep of 60 mV/s has been found [305]. For cathodic stripping voltammetry upon a rotating graphite disk electrode, Se is plated for 1-5 min in 0.4M H,SO,, in presence of Cu and chromate, and subsequently stripped in the pulse mode, to obtain a peak at -1.07V The method was applicated to the determination of Se in matrix Ga, after separation from the matrix and from nitrate, deriving from the dissolution procedure [306]. 3.5.3 DETERMINATION OF PIAZSELENOLS Isolation of selenium as a piazselenol, and the subsequent polarographic determination of the piazselenol itself enables separation and enrichment from interfering matrices (see 211,234,311) In 0 IM NH,CIO, i n formiate buffer at pH 2 5 , polarographic waves of the piazselenol from 3,3'diaminobenzidine at -0 1 1 V and at -0 63 V vs SCE are obtained at the dropping Hg- electrode The reagent itself gives reduction peaks at -0 4 I V and -0 97V A detection limit of 0 4 ngIml could be obtained [307] Similarly, at the hanging mercury drop electrode, adsorptive stripping voltammetry of the complex of Se with 3,3'dianiinobenzidiile in the differential pulse mode enabled the detection down to 0 2 pg/l [295] For the analysis of effluents from the mining industry, Se is extracted with o-phenylenediamine from ammonium perchlorate buffer pH=9 with toluene After addition of acetone and NH4CI/HCI - buffer pH1 3 , a homogenous solution with sufficient electrolytical conductivity is achieved to enable the polarography of the extract with a dropping Hg electrode Under conditions of anodic polarization, only one polarographic wave of Se is achieved, which yields niaximuin current at pH I 3 , leading to a detection limit of lygll in the original solution Prioi to the extraction, C u was removed by extraction with dithizone, and large excess of Fe was masked with EDTA [308]

3.5.4 ELECTROCHEMICAL DETECTORS IN LIQUID CHROMATOGRAPHY Liquid chromatographical separation on a cation exchange resin with on-line detection of Se at a tubular Au- electrode helps to cope with interferences of codeposited metals in anodic stripping voltammetry When Se is eluted from the column, the potential I S set to -0 30V, and the reduction current is monitored The electrode surface IS cleaned by rapid

440 cycling between -0.30 and +1.20 V, and kept inert during regeneration of the ion-exchange column at +1.30 V [92]. Selenols, diselenides, and selenenyl sulfides could be detected by group- selective electrochemical reactions after reversed phase liquid chromatography at two dual Hg/Au amalgam electrodes in series in the eluent stream. The upstream electrode is set at - 1.1OV VS. Ag/AgCI to reduce disulfides, diselenides and seleny sulfides, or at -0.5SV to reduce only the Se- compounds. The selenols are detected directly by facilitation of the oxidation of Hg from the downstream electrode, when it is set to +O. 15 V vs.Ag/AgCI. Diselenides and selenenyl sulfides are determined by first reducing them to the selenol and/or thio form at the upstream electrode followed by detection of the selenol and/or thiol at the downstream electrode [90].

3.6 RADIOCHEMICAL AND NUCLEAR METHODS

3.6.1 GENERAL As can be seen from table 1 , activation with neutrons leads to isotopes with overall soft b-lines. For their detection, the NaJ(TI)- detector is up to 10 times more sensitive, but lack

of selectivity requires chemical separation procedures. With the Ge(Li)- detector, direct counting after activation is possible in some cases [9]. --Tables 1 and 2

3.6.2 ACTIVATION WITH THERMAL NEUTRONS AND PURELY INSTRUMENTAL DETECTION OF 75-SE After sealing in polyethylene or quartz vials, 30h irradiation with thermal neutrons and 2-3 weeks storage, 75-Se could be determined by y- counting at 265 keV with a Ge(Li) detector. In particulate matter from riverine and marine waters [3 171, in atmospheric particulates [ 145,3181, coal, fuel oil and fly-ash [319], detection limits of 25 ng abs. [313] resp. 10 ng/g [145,3 181 were achieved. For biological materials, like tissues, blood plasma, serum and erythrocytes, even 5 days of irradiation, 6 weeks of cooling, and half an hour counting time with a Ge(Li)- detector were necessary [3 19,320,3211. At 264.6 kEv, 182-Ta at 264.6 keV cannot be discriminated from selenium because of its quite similar half- life [3 131, but this is not crucial for biological matrices. 3.6.3 ACTIVATION WITH EPITHERMAL NEUTRONS During activation with epithermal neutrons only, contrary to interfering concomitant elements, the cross section of 74-Se to yield the y- radiating isotope 75-Se, is not entirely lowered, which improves selectivity [32 I]. The production of 24-Na with epithermal neutrons is much smaller than with thermal ones, and the overall activity is about 20-fold less after Cd shielding [323]. Similarly, in epithermal neutron activation of silicate rocks and sediments, the interference of 181-Hf and 131-Ba at the 75-Se line at 136 keV is

44 I lowered, when epithermal neutrons are used [312]. However, at the 136 keV- line, the radiation from 99m-Tc at 140 keV, which is a daughter of Mo, is of equal sensitivity to the radiation of 75-Se at 136 keV [315]. The line at 265 keV is not observed in any sample because of insufficient detection limit [3 121. For the analysis of selenium in coal and fly ash, epithermal irradiation was found preferable, because of better precision, and equal sensitivity with respect to conventional y- irradiation method. Ash samples had to be irradiated for 1 day, and coal samples for 2 days. After 20 days of decay time, the 75-Se could be counted at 265 keV, achieving a detection limit of 0.1 pg/g [314]. In case of biological matrices, the improvement of using epithermal neutrons with respect to thermal ones was found to be negligible [321]. 3.6.4 METHODS OF ACTIVATION AND SUBSEQUENT CHEMICAL SEPARATION For counting the low-energy y- emission of 75-Se, chemical separation from interfering matrix elements allows to use the less selective, but more sensitive scintillation detector Addition of up to 1 OOmg inactive Se- carrier allows to do the clean-up on a macro- scale, and to control the regain of the analyte After coprecipitation or adsorption, active Se can easily be counted on the solid phase 3.6.4.1 Geological materials From acid digests of geological materials, Se was separated and counted on the solid after precipitation with sulfite [51,54,3 16,3241, with thioacetamide [325], with MnO, [IS], or sorbed on A&O, [326]. Interfering activated Ta was removed by coprecipitation with inactive Ta carrier from acid solution [326]. Alternatively, distillation as the bromide [54,57] as well as various extraction methods were utilized with activated samples [ 54,325,3271. Decomposition with alkaline fluxes after irradiation and addition of Se- carrier quantitatively yields soluble selenate, which is not precipitated along with hydroxides (Fe) or sulfides in alkaline sodium sulfide solution [325]. For large sample weights, Se and Te were extracted together with A u and Ag by fire assay in a flux consisting of lead oxide, soda, quartz, sugar, and borax, into newly formed metallic lead. Only this lead button was activated. The active material in matrix Pb is less hazardous to handle. However, 203-Pb emitted at 279 keV close to Se, and had to separated prior to counting. After dissolution in nitric acid, Se was reduced to the elemental state with hydroxylamine, and counted on filter [328,329]. 3.6.4.2 Biological samples After irradiation with thermal neutrons and 1-2 weeks of cooling, inactive carrier I S added If the samples are burnt in a stream of oxygen, Hg and Br come along with Se, which may interfere in radioactive counting This has to be also considered after distillation as bromide [309,333] Se was separated from Hg and Br by solvent extraction [327,330], adsorption on charcoal [64], or 82-Br was just allowed to decay within 20 days [I21 From acid digests of activated biological samples, Se was separated by liquid-liquid extraction as piazselenol [9,105], as carbamate [327], or as iodide into CC14 [53] Direct counting was possible after sorption as the dithizonate on carbon powder at pH=8, or after coprecipitation with Fe- hydroxide in presence of ascorbic acid [332]

442 3.6.4.3 Water samples Water samples were reduced with ascorbic acid at pH=2, and Se was adsorbed on active charcoal, which was finally activated and counted [ 1361. 3.6.5 USE OF THE SHORT-LIVED 77M-SE 76-Se can be activated to yield the short-lived 77m-Se with a half-life of only 17.5 seconds. No separations are possible, but nearly no handling is necessary. 28-AI, together with Ti, Mg, V, Mn and U, is also accessible to short time activation, decaying with a half-life of 2.3 min at 1779 keV [334]. In geological and coal matrices, short-lived Se can be interfered with 28-AI, causing high dead times and high background [313]; therefore, 77m-Se activation has been mainly used for biological and related matrices. Short activation of 77m-Se minimizes matrix activation. The main sources of error are the variation of the neutron flux [335], and exact timing. Biological samples were simultaneously treated with aqueous standards in cycles of 20 sec irradiation, 3 sec waiting, and 18 sec counting at 162 keV with a Ge(Li) detector The delay of 3 sec between irradiation and counting periods was necessary to decay 38m-CI with its half-life of 0.7 sec [336]. Food samples were irradiated for 20 sec, decayed for 20 sec, and counted for 20 sec at I62 keV for 77m-Se. The precision significantly improved by recycling the samples up to 4 times [337,338]. 200 mg of biological material was irradiated for only 2-4 seconds in a high flux of 10'' n/cm2.s, and measured I 5 sec after the end of the bombardment at 162 keV, thus achieving a detection limit of 5ng/g [335]. Similarly, for determination of Se in liver tissue, samples and liver reference standards were irradiated separately and in fixed sequence for 90 sec. The integrated value of the current neutron flux was determined by use of gold monitor samples [339]. In atmospheric particulates, a detection limit of I8 ng Se and a precision of 13% could be achieved after 60 sec irradiation with 5.10'' n/cm'.s, 5 sec delay and 30 sec counting [3 131. Alternatively to thermal neutrons, the 77m-Se could be also produced by 2 min photon activation from a 5-10 kCi co-60 source, but the detection limit of 1 mg Se in a pellet of 20g of animal food stuff was rather poor [340]. 3.6.6 PHOTON ACTIVATION With high energy photons of about 35 MeV , two nuclear reactions with Se- isotopes can occur [341,342,343,344]: 76-Se (y,n) 75-Se 120 d 265 keV/ 136 keV 82-Se (y,n) 8lm-Se 57.3 m 103 keV High energy photon activation has been applicated to determine Se in the multi-element analysis of river sediments, atmospheric particulates, and related materials. The samples were pelletized with Li,SO,, encapsulated in Al for neutron capture, and simultaneously irradiated with standards.After irradiation, the Al-foil had to be discarded because of to reduce background from 27-Al(n,y)24-Na [344]. Similarly, atmospheric particulates on polystyrene filters were pelletized along with elemental and flux monitors, and submitted to multielement photon activation with Ge(Li) detection [345]. Soil and fly-ash samples were irradiated, cooled 1 day, pelletized with cellulose powder, and finally measured with a planar intrinsic Ge-diode [343]. At 136 keV,

443 the limit of practical determination was estimated to be 0 2 pg/g, there was some interference by 57-co

3.7 X-RAY - SPECTROMETRIC METHODS FOR T H E DETERMINATION O F SELENIUM 3.7.1 GENERAL For analytical purposes of Se, the K a - emission lines are exclusively used. They hardly coincide with neighbouring elements, or with L - lines of heavy elements (see table 3). Table 3 Wavelength and energy of some X-ray lines [346] Element line wavelength energy keV As Ka2KL2 1 1799 10 508 KalKL3 11759 10 544 Pb LalL3M5 1 1750 10 552 Bi La2L3M4 I 1554 10 731 LalM3L5 1 1439 10 839 Ge KR3KM2 1 1294 10 978 KBlKM3 11289 10 982 Se Ka2KL2 1 1088 I I 181 KaIKL3 1 1048 I I222 As Kfi3KM2 I0578 I 1 720 KDlKM3 I0573 I I 726 Br Ka2KL2 1 1 878 10438 KaIKL3 10397 I1 924 In the wavelength dispersive mode, an overlap between the Se KOL2,3 line, the second order of the Hg L l M 2 and the Hg L2M4 line [347] is reported. 3.7.2 WAVELENGTH

- DISPERSIVE XRF

The best efficiency of Se-Ka line excitation is obtained with a Mo-target tube, which I S used in most cited cases, but continuous background of bremsstrahlung may be high, thus decreasing the accuracy Excitation with a Ag- target tube yielded nearly the same power of detection, but with much less background radiation [348] As a compromise for the determination of 12 elements in pharmaceuticals after coprecipitation enrichment, a Cr-tube has been used [147], which allows to detect S e by a factor of 2-3 less with respect to a Mo-tube The excitation spectrum is usually discriminated with a 100-LiF- crystal, at an angle 26 = 31 87" After excitation with a Mo-target tube, backgiound correction on both sides of the Se-peak was done at 3 I 10/32 60°[ 1441, or at 3 I 0132 74"[349] In case less than 10 mg of solid per cm' are available for final determination, a correction for non-infinite sample depth has to be applied [350] For water samples, Se can be collected by coprecipitation and sorption methods, and the resulting uniform solid finally counted on the membrane filter As collectors, ZnS or CdS freshly prepared on a membrane filter, were successfull at pH 3-4 Coprecipitation with

444 subsequent XRF- measurement has been done with polyvinylpyrrolidon/ thionalide at pH 4 [152], with diethyldithiocarbamate at pH 4.7-5.0 [150], or with combined dibenzylammonium and Na- dibenzyldithiocarbamates at pH 2.5 - 8.0 [ 1461. Similarly, from acid digests of biological materials and pharmazeutical preparations, Se was coprecipitated with the dibenzylaminosalt of dibenzyldithiocarbamic acid at pH 4 *I [147], or with Te [348,349]. Nitric acid interferes with the Te- reduction, and has to be reduced beforehand. Diethyldithio-carbamate as precipitating agent is not suitable in presence of large excess of Fe, which is also precipitated and can dominate the spectrum [349]. For the selective XRF- measurement of Se together with As and Sb, the corresponding hydrides were evolved from hydrochloric acid samples, and adsorbed on filter paper impregnated with AgNO,, which could be directly submitted to XRF measurement [347]. 3.7.3 ENERGY - DISPERSIVE XRF For Se, the line at 11.21 keV is used throughout, after Mo-K- excitation. Most authors prefer peak area evaluation, and background correction. Sample preparation procedures are not entirely different to wavelength dispersive methods. In soils and atmospheric particulates without chemical pretreatment, however, the detection limit of about 3 pg/g is not sufficient [350]. To reach the common geochemical level, wetchemical separation from the matrix and reductive coprecipitation with Te has been proposed [ 1281. Se in serum and whole blood could be directly counted after drying a 0.75 pI sample on a filter spot. The peak intensities were corrected according to the intensity of the back-scattered Mo K X-rays in the range 14-19 keV. As detection limit, 60 ng/ml in 100 sec was reached (normal value: 80-120 ng/m1)[351]. For the determination of Se in water samples, elementary Se was adsorbed on active carbon after reduction of selenite with ascorbic acid. Both Selenate and selenite were reduced by refluxing with thiourea in sulfuric acid. For excitation of Se in the active carbon, a W-target tube was used, and the detection limit could be extended to SO ng/l by an extraordinary long counting time [134]. 3.7.4 X-RAY SPECTROMETRY IN TOTAL-REFLECTION GEOMETRY Total reflection of the incident beam yields excitation only at the sample surface [353,358]. Aerosol samples could be excited directly in the total reflection geometry, but digestion or low-temperature plasma ashing was preferred to obtain a more representativ sample [359]. For serum, 30 pl samples were pipetted onto a Si X-ray mirror, spiked with Ge as internal standard, and air dried. A W- target tube was used to excite a Ni- secondary target, yielding a narrow band-pass of primary X-rays (K-edge at 12.66 KeV), to excite selective Se at 13.0 and 13.3 keV, while at the same time preventing the excitation of a relatively high concentration of Br (K edge at 13.475 keV) in the serum [358]. 3.7.5 PROTON - INDUCED X-RAY EMISSION (PIXE) The determination of Se by PIXE has only been reported for biological matrices yet. If biological specimen directly interact with the proton beam, they should be placed on pure backing thin targets backed with nuclear graphite, to avoid problems with the stability of

445

the shape of the sample. When the proton current exceeds 200 nA, evaporation of Se, Ca, As and K occurs [352]. The depth of penetration of exciting protons into the sample decreases with increasing average atom number of the sample, and ranges from 0.5 - 50 pm [353]. To reduce intense low-energy radiation, an Al- absorber of 78 pm thickness was placed in front of the detector [354]. By direct excitation with protons of 1.7 MeV, 0.08 pg/g Se could be detected [352] in a biological matrix Blood serum was dry ashed at 60°, ground, mixed with 20% graphite powder and Pd as internal standard, and formed to a thermally stable pellet. As the detection limit, 10 ng/g was found for 100 min counting after excitation with 1.8 MeV protons, or after 30 min counts after excitation with 4-MeV protons [ 3 5 5 ] . Alternatively, freeze- dried tissue and plant samples were converted into a fine powder, doped with Ag as internal standard, and fixed at the target frame by means of 1% solution of polystyrene in benzene [356]. From acid digests of biological matrices, Se can be selectively separated by coprecipitation with Te, and excited with 1.8 MeV protons, which leads to a detection limit of about 3 ng Se [42,354,357].

3.8 CONVERSION TO SE(IV)

In potable waters as well as after oxidational digestion, a substantial part of Se is present as selenate, which is advantageous because it is far less volatile and adsorbable on solids. Prior to most methods involving chemical reactions, Se has to be converted to the quadrivalent form. Besides the conversion reaction, excess nitrous oxides and chlorine have to be driven off, which interfere with subsequent hydride formation, colour reactions, reductive precipitation, electrochemical reactions etc.. In addition to interferents present in the sample, incomplete recovery is possible due to incomplete reduction of Se(VI), further reduction to Se", or volatilization. Losses due to volatilization or precipitation at vessel walls can be traced with 75- Se labelled compounds Most authors use reduction reaction of selenate with hydrochloric acid: H,SeO, + 2 HCI = H,SeO, + H,O + CI, Different optimum conditions found in the literature depend on the acid mixture used, the amount of sample to be oxidized, as well as on residual organics and cations (Fe) in the sample (see table 4). Remaining nitrite, which may interfere further, can be destroyed by addition of hydroxylamine, sulfanilamide,or amido-sulphuric acid [ 18 1,360,3611. Dry-ashing with Mg(NO,), and dissolution with 6M-HCI on the boiling water bath yields selenite and complete oxidation of organics, which renders a conversion step unnecessary. Besides heating with HCI, conversion was also achieved by UV- irradiation at pH> 7.5

W I

446

Table 4. Quantitative conversion to selenite in HCI min "C [HCII HCI only 180 20 3M [362] 30 105 4M [363] , 30 85 5M [363] I, 30 65 6M [363] potable water 5 boiling 4M (1811 , 6M [210] dil. KMn04 8 acid digests 30 95 5M [364] 30 SM [365] 90 15 95 SM [6,366] 10 80 6M [I871 15 80 6M 1181 4 boiling 4M * [205] 10 boiling 4M [I881 20 boiling 6M [368] 30 boiling 6M [I991 120 boiling water bath 6M [471 20 boiling water bath 6M [411 Cont.flow S heated coil 7.SM [I981 $9

REFERENCES 1 Janghorbani M, Ting BTG, Nahapetian A, Young VR, Anal Chem 1982; 54: I I88 2 Koh TS, Benson TH, Journal AOAC 1983; 66:9 I8 3 Walker HH, Runnels JH, Merrifield R, Anal Chem 1976; 48:20S6

4 Robberecht HJ, Van Grieken RE, Van den Bosch PA, Deelstra H, Van den Berghe H, Talanta 1982; 29:1025 5 Reamer DC, Veillon C, Anal Chem 1981; 53:1192 6 Piwonka J, Kaiser G, Tolg G, Fres Z Anal Chem 1985; 321:22S 7 Iyengar GV, Kasperek K, Feinendegen LE, Sci Tot Environm 1978; 10:1 8 Tam GKH, Lacroix G, Journal AOAC 1982; 65:647 9 Dermelj M, Byrne AR, Franko M, Srnodis B, Stegnar P, J Radioanal Nucl Chem Lett 1986; 10: 91 10 Dermelj M, Byrne AR, Franko M, Smodis M, Stegnar P, Vestn Slov Kern Drust 1986; 331443 11 Beal AR, J Fish Res Board Can 1975; 32:249 12 Byrne AR, Kosta L, Talanta 1974; 2 1 : 1083 13 Dams R, Billiet J, Block C, Demuynck M, Janssens M,Atmosph Environ 1975; 9: 1099 14 Kotz L, Kaiser G, Tschopel P, Tolg G, Fres Z Anal Chem 1972; 260:207 15 Gallorini M, Greenberg RR, Gills TE, Anal Chem 1978;'SO: 1479 16 Fricke FL, Rose 0, Caruso JA, Talanta 1976; 23: 3 I7 17 Cappon CJ, Smith JC, J Anal Toxicol 1978; 2: I14 18 Welz B, Melcher M, Schlemmer G, Fres Z Anal Chem 1983; 316:271 19 Shimoishi Y,Analyst 1976; 101:298 20 Haas F, Krivan V, Talanta 1984; 31:307

441 21 22 23 24 25 26 27 28 29 30 31 32 33 34

McClellan BE, Frazer KJ, Research Rep No 122, Univ Kentucky 1988 Siu KWM, Berman SS, Talanta 1984, 31-1010 Verlinden M, Talanta 1982; 29:875 Fiorino JA, Jones JW,Capar SG, Anal Chem 1976; 48 120 Chan CCY, Anal Chim Acta 1976; 82.2 I3 Blades MW, Dalziel JA, Elson CM,Journal AOAC 1976; 59: 1234 Michie ND, Dixon EJ, Bunton NG, Journal AOAC 1978, 61.48 Alfthan G, Anal Chim Acta 1984; 165. I87 Holynska B, Lipinska-Kalita K, Radiochem Radioanal Lett 1977; 30:241 Salisbury CD, Chan W, Journal AOAC 1985; 68:218 Bye R, Lund W, Fres Z Anal Chem 1985; 321:483 Watkinson JH, Anal.Chim.Acta 1979; 105:3 19 Briggs PH, Crock JG, USGS Report 1986-40 Nazarenko 11, Kislova IV, Gusejnov TM, Mkrtchjan MA, Kislov AM, Zh Anal Khim 1975; 30:733 35 Agterdenbos J, Van Elteren JT, Bax D, Ter Heege JP, Spectrochim Acta 1986; 41B:303 36 Adeloju SB, Bond AM, Briggs MH, Hughes HC, Anal Chem 1983, 55:2076 37 Harms TF, Ward FN, Geol Survey Bull 197.5; 1408:37 38 VlJan PN, Wood GR, Talanta 1976; 23:89 39 Yang IY, Yang MH, Liu SM, Anal Chem 1990; 62:146 40 Rail CD, Kidd DE, Hadley WM, Intern J Environm Anal Chem 1980; 8:79 41 Ng S, McSharry W, Journal AOAC 1975; 581987 42 Buso GP, Colautti P, Moschini G, Hu Xusheng, Stievano BM, Nucl Instrum Meth Phys Res 1984; B3: I77 43 Rozanska B, Rozowski J, Mikrochim Acta 1984 II:481 44 Cappon CJ, Smith JC, J Anal Toxicol 1982, 6: 10 45 Koppenaal DW, Lett RG, Brown FR, Anal Chem 1980; 2:44 46 Wolf A, Schramel P, Lill G, Hohn H Fres Z Anal Chem 1984; 317:512 47 Bye R, Fres Z Anal Chem 1984; 317.27 48 Bedrossian M, Anal Chem 1984; 56:3 1 1 49 Beary ES, Paulsen PJ, Lambert GM, Anal Chem 1988; 60.733 50 Akiba M, Shimoishi Y, Toei K, Analyst 1975; 100:648 5 I Hermann F, Wichtl M, in: Kiesl W, Malissa H jun (eds), Analyse extraterrestrischen Materials, Springer 1974 52 Tamari Y, Radioisotopes 1978; 28:3 53 Hoeda D, Massie R, Das HA, Radiochem Radioanal Lett 1975; 23:379 54 Zmijewska W, Semkowo T, Chem Anal 1978; 23:583 5 5 Terada K, Ooba T, Kiba T, Talanta 1975; 22:41 56 Rybakov AA, Ostroumov EA, Okeanologiya 1983; 23: 1039 57 Terada K, Okuda K, Maeda K, Kiba T, J Radioanal Cheni 1978; 46:217 58 Heinrichs H, Keltsch H, Anal Chem 1982; 54.121 I 59 Meyer A, Grallath E, Kaiser G, Tolg G, Fres Z Anal Chem 1976; 28 I :20 I 60 Meyer A, Hofer C, Knapp G, Tolg G , Fres Z Anal Chem 1981;

448 61 Forbes S, Bound GP, West TS, 305: 1 Talanta 1979; 26:473 62 Han HB, Kaiser G, Tolg G, Anal Chim Acta 1981; 128:9 63 Shand CA, Ure AM, J Anal At Spectr 1987; 2: 143 64 Muramatsu Y, Ogris R, Reichel F, Parr RM, J Radioanal Nucl Chem 1988; 125:175 65 Gleit CE, Holland WD, And Chem 1968; 34:1454 66 Al-Attar AF, Nickless G, J Chrom 1988; 440: 333 67 Dilli S, Sutikno I, J Chrom 1984; 298: 21 68 Kurahashi K, Inoue S, Yonekura S, Shimoishi Y, Toei K. Analyst 1980; 105: 690 69 Shimoishi Y, Toei K, Anal Chim Acta 1978; 100, 65 70 Flinn CG, Aue WA, J Chrom 1978; 153: 49 71 Olsen KB, Sklarew DS, Evans JC, Spectrochim Acta 1985; 40B:357 72 McCarthy TP, Brodie B, Milner JA, Bevill RF, J Chrom 1981; 225: 9 73 Uchida H, Shimoishi Y, Toei K, Analyst 1981; 106: 757 74 Dilli S, Sutikno I, J Chrom 1984; 300: 265 75 Jiang SG, Chakraborti D, Adams F, Anal Chim Acta 1987; 196: 271 76 Radziuk B, Van Loon J, Sci Tot Environm 1976; 6: 251 77 Knab D, Gladney E, Anal Chem 1980; 52: 825 78 Heuss E, Lieser KH, J Radioanal Chem 1979; 50: 289 79 Aoyama E, Nakagawa T et al, Bunseki Kagaku l987; 36: 801 80 Itoh K, Nakayama M, Chikuma M, Tanaka H, Fres Z Anal Chem 1985; 321: 56 81 Oguma K, Kuroda R, Fres 2 Anal Chem 1975; 277: 1 I I 82 Shibata Y, Morita M, Fuwa K, Anal Chem 1984; 56: 1527 83 Schwedt G, Fres Z Anal Chem 1977; 288: 50 84 Handelman GJ, Kosted P, Short S, Dratz EA, Anal Chem 1989; 61: 2244 85 Schwedt G, Chromatographia 1978; 11: 145-148 86 Sturgeon RE, Willie SN, Berman SS, Anal Chem 1985; 57: 6 87 Nakagawa T, Aoyama E, Hasegawa N, Kobayashi N, Tanaka H, Anal Chem 1989; 61: 233 88 Chakraborti D, Irgolic KJ, Proc Int Conf Heavy Met Environm Athens 1985: 484 89 McCarthy JP,Caruso JA, Fricke FL, J Chrom Sci 1983; 21: 389 90 Killa HMA, Rabenstein DL, Anal Chem 1988; 60: 2283 91 Haas HF, Krivan V, Fres Z Anal Chem 1986; 324: 13 92 Andrews RW, Johnson DC, Anal Chem 1976; 48: 1056 93 Dreipa EF, Pakholkov VS, Luk'yanov SA, Zh Prikl Khim 1980; 53: 54 94 Sayre WG, Constable DC, Proc 3 ann adv anal instrum conf , Toronto, 198 1. 95 Zolotov Yu A, Shpigun OA, Bubchikova LA, Fres Z Anal Chem 1983; 316: 8 96 Zolotov Yu A, Shpigun OA, Bubchikova LA, Sedel'nikova EA, Dokl Akad Nauk SSSR 1982; 263: 889 97 Karlson U, Frankenberger WT, Anal Chem 1986; 58: 2704 98 Karlson U, Frankenberger WT, J Chrom 1986; 368: 153

449 99 Nakata F, Sunahara H, Matsuo H, Kumamaru T, Bunseki Kagaku 1986, 35 439 100 Shabana R, Ruf H, J Radioanal Chem 1978, 47 17 101 Koch G, Koch-Dedic G, Handbuch der Spurenaiialyse Springer -Verlag 1974 102 Landsberger S, Boswell GGJ, Anal Chim Acta 1977, 89 28 I 103 Bock R, Kusche H, Bock E, Fres Z Anal Chem 1953, 138 167 104 Futekov L, Stojanov S, Specker H, Fres Z Anal Chem 1979. 295 7 I05 Dermelj M, Hancman G, Gosar A, Byrne AR, Kosta L, Vestn Slov Kern Drus 1985, 32 127 106 Agrawal YK, Desai KH, Anal Lett 1986, 19 25 107 Futekov L, Stojanov S, Specker H, Fres Z Anal Chem 1979, 294 125 108 Futekov L, Angelova G, Specker H, Fres Z Anal Chem 1979, 294 262 109 Shabana R, Radiochern Radioanal Lett 1978, 36 359 110 Heddur RB, Khopkar SM, Talanta 1988, 35 594 1 1 1 Bode H, Neumann F, Fres Z Anal Chem 1960, I72 I 112 Strel'nikova, Akad Nauk SSSR, Trudy kom anal khim 1963, 14 307 1 13 Safavi A, Townshend A. Anal Chim Acta 1984, 164 77 114 Kamada T, Shiraishi T, Yamamoto Y, Talanta 1978, 25 I5 1 I5 Wyttenbach A, BajO S, Anal Chem 1975, 47 I8 13 1 16 Subramanian KS, Meranger JC, Anal Chim Acta 198 I . I24 I3 1 117 Seyakova IV, Vorobiova GA, Glembotski AV, Zolotov Yu A, Anal Chim Acta 1975, 77 183 118 Stary J, Ruzicka J, Talanta 1968, I S 505 119 Donaldson EM, Talanta 1976, 23 41 I 120 Donaldson EM, Talanta 1976. 23 417 121 Donaldson EM, Talanta 1977, 24 441 122 Singh N, Garg AK, Analyst 1986, I I 1 247 123 Pangarova RB, Angelova V, Dokl Bolg Akad Nauk 1977. 30 863 124 Spivakov B Ya, Shkinev VM, Zolotov Yu A, Zh Anal Khim 1975, 30 2182 125 Kouimtzis Th A, Sofoniou MC, Papadoyannis I N , Anal Chim Acta 198 I , 123 3 15 126 Chau YK, Riley JP, Anal Chim Acta 1964, 3 3 36 127 Nakaguchi Y, Hiraki K et al, Anal Sci 1985, I 247 128 Wahlberg JS, Chem Geol 1981, 3 3 1 5 5 129 Pokrovskij EN, Kirilova LI, Sidorov VA, Zav Lab 1979 199 I30 Woo IH, Watanabe K, Hashimoto Y, Lee YK, Anal Sci 1987, 3 49 131 Drabaek I, Carlsen V, Just L, J Radioanal Nucl Chem 1986, 103 249 132 Brunfeldt HJ, Steinnes E, Geochim Cosmochini Acta 1967, 3 I 283 133 Artern ev 01, Stepanov VM, Zh Anal Khim 1976, 3 1 724 134 Robberecht HJ, Van Grieken R. Anal Chem 1980, 52 449 135 Raie RM, Smith H, Radiochem radioanal Lett 1977, 28 215 136 Massee R, Van der SIoot HA, Das HA, J Radioanal Chem 1977. 35 157 137 Siu KWM, Berman Sh S, Anal Chem 1984, 56 1806 138 Halicz L, Russel GM, Analyst 1986, 1 1 1 1 5 139 Tzeng Jau-Hwan, Zeitlin H, Anal Chim Acta 1978, 101 71 140 DeCarlo EH, Zeitlin H, Anal Chem 1981, 53 I104 141 Taniari Y, Hirai R, TSUJIH, Kusaka Y, Anal Sci 1987, 3 3 13 142 Bedard M, Kerbyson JD, Can J Spectrosc 1976, 2 I 64 143 Reichel W, Bleakley BG, Anal Chein 1974. 46 59 144 Disam A, Tschopel P, Tolg G, Fres Z Anal Chem 1979, 295 97

450

Kronborg OJ, Steinnes E, Analyst 1975; 100: 835 Moore RV, Anal Chem 1982; 54: 895 Linder HR, Seitner HD, Schreiber B, Anal Chem 1978; 50: 896 Ellis AT, Leyden DE, Wegscheider W, Jablonski BB, Bodnar WB, Anal Chim Acta 1982; 142: 89 149 Lieser KH, Burba P et al, Mikrochimica Acta 1980 11: 445 150 Scheubeck E, Jorrens Ch, Fres Z Anal Chem 1980; 303: 257 151 Nazarenko IJ, Kislova IV et al, Zh Anal Khim 1987; 42:1059 152 Panayappan R, Venezky DL, Gilfrich JV,Birks LS, Anal Chem 1978; 50: 1125 153 Smits J, Van Grieken R, Anal Chim Acta 1981; 123: 9 154 Neve J, Hanocq M, Molle L, Mikrochim.Acta 1980 I: 41 155 Bock R, Jacob D, Fres Z Anal Chem 1964; 200. 8 1 156 Schwedt G, Schwarz A, J Chrom 1978; 160: 309 157 Kasterka B, Mikrochim.Acta 1989 I: 337 158 Parker CA, Harvey LG, Analyst 1962; 87: 558 159 Parker CA, Harvey LG, Analyst 196 1 ; 86: 54 160 Itoh K, Nakayama M, Chikuma M, Tanaka H, Fres Z Anal Chem 1986; 325: 539 161 Koh TS, Benson TH, Journal AOAC 1983; 66: 9 18 162 Gowda HS, Achar BN, Indian J.Chem. 1980; 19A: 178 163 Shibata Y, Morita M, Fuwa K, Anal Chem 1984; 56: I527 164 Watkinson JH, Brown MW, Anal Chim Acta 1979; 105: 451 165 Dye WB, Bretthauer LG, Seim HJ, Blincoe C, Anal Chem 1963; 166 Watkinson JH, Anal Chem 1962; 32: 981 167 Uvarova KA, Dovbnya TV, Stand.Obrazcy -Chern.Metall 1979; 8: 51 168 Izquierdos A, Pret MD, Aragonis L, Analyst 1981 ; 106: 720 169 Chambers JC, McClellan BE, Anal Chern 1976; 48: 2061 170 Khaligie J, Ure AM, West TS, Anal Chim Acta 198 1; 13 1 : 27 171 Holen B, Bye R, Lund W, Anal Chim Acta 1981; 130: 257 172 Holen B, Bye R, Lund W, Anal Chim Acta 1981; 131: 37 173 Milner B, Int Lab Nov/Dec 1983, 30 174 Keil R, Fres Z Anal Chem 1984; 3 19: 39 1 175 Issaq HJ, Morgenthaler LP, Anal Chem 1975; 47. 1668 176 Fernandez FJ, Giddings R, Atom Spectrosc 1982, 3: 61 177 Futekov L, Paritschkova R, Specker H, Fres Z Anal Chem 1981; 306: 378 178 Lund W, Thomassen Y, DOvle P, Anal Chim Acta 1977; 93: 53 179 Lund W, Bye R, Anal Chim Acta 1979; 110: 279 180 Agterdenbos J, Bax D, Fres Z Anal Chem 1986; 323: 783 181 Cutter GA, Anal Chim Acta 1978; 98: 59 182 Agterdenbos J, Bussink RW, Bax D, Anal Chim Acta 1990; 232:405 183 Goulden PD, Brooksbank P, Anal Chem 1974; 46: 143 I 184 Branch CH, Hutchison D, Analyst 1985; 110: 163 185 Yu Mu-Qing, Gui-Qin Liu, Qinhan Jin, Talanta 1983; 30: 265 186 Dedina J, Anal Chem 1982; 54: 2097

145 146 147 148

35: 1687

45 1 187 Zhang Baogui, Tao Keyi, Feng Jianxing, Spectrochim Acta 1989, 44B 247 188 Brooks RR, Willis JA, Ltddle JR, Journ AOAC 1983, 66 130 189 Welz B, Melcher M, Analyst 1984, 109 577 190 Pierce FD, Brown HR, Anal Chem 1977, 49 I4 17 191 Sager M, ALVA-Nachrichten 1989. 19 33 192 Bye R, Analyst 1985, I10 85 193 De Kersabiec AM, Analusts 1980, 3 97 194 Dittrich K, Mandry R, Analyst 1986. 1 I I 269 195 Petrick K, Krivan V, Fres 2 Anal Chem 1987, 327 328 196 Cutter GA, Anal Chim Acta 1983, 149 391 197 Raptis S, Knapp G, Meyer A, Tolg G, Fres L Anal Chem 1980, 300 18 198 Vijm PN, Leung D, Anal Chtm Acta 1980, 120 141 199 Voth-Beach LM, Shrader DE, Spectroscopy I 200 Tesfalidet S, Irgum K, Fres J Anal Chem 1991, 341 532 201 Reamer DC, Veillon C, Anal Chem 198 I , 53 I 192 202 Fleming HD, Ide RG, Anal Chim Acta 1976, 83 67 203 Verlinden M, Anal Chim Acta 1982, 140 229 204 Agterdenbos J, Van Noort JPM, Peters FF, Bax D, Ter Heege JP, Spectrochim Acta 1985, 4 0 8 S O 1 205 Narasaki H, Ikeda M, Anal Chem 1984, 56 2059 206 Narasaki H, Fres Z Anal Chem 1985, 321 464 207 Lloyd B, Holt P, Delves HT, Analyst 1982, I07 927 208 Agterdenbos J. Van Noort JPM, Peters FF, Bax D, Spectrochtm Acta 1986, 41B 283 209 Willie SN, Sturgeon RE, Berman SS, Anal Chem 1986, 58 1140 210 Roden CR, Tallman DE, Anal Chem 1982, 54 307 2 I 1 Doidge PS, Sturman BT, Rettberg TM, J Anal At Spectr 1989, 4 2 5 1 212 Matsumoto K, Ishtwatart T, Fuwa I(, Anal Chein 1984, 56 I545 213 Alt F, Messerschtnidt J, Tolg G, Fres Z Anal Chem 1987,327 233 214 Deschuytere MA, Robberecht HJ, Deelstia H, Bull SOCChim Belg 1988, 97, 99 215 Teague- Ntshimura JE, Totninaga T, Katsura T, Matsumoto K, Anal Chem 1987, 59 1647 216 Krivan V, Kuckenwaitz M, Fres J Anal Chem 1992, 342 692 217 Dedina J, Frech W, Cedergren A, Lindberg I, Lundberg E, J Anal At Spectr 1987. 2 435 218 Ihnat M, Anal Chim Acta 1976, 82 293 219 Styris DL, Prell LJ, Redfteld DA, Holcombe JA, Bass DA, Majidi V, Anal Chem 1991, 63 508 220 Lam B, Zinger M, Research I%Development, Feb 1985 221 Rayson GD, Bass DA, Appl Spectr 1991. 4.5, 1049 222 Pruszkowska E, Barrett P, Spectrochim Acta 1984. 39B 485 223 Jowett PLH, Banton MI, Anal Lett 1986. I9 I243 224 Hocquellet P, Candtllier MP, Analyst I99 I, 1 16 5 0 5 225 Peile R, Gray R, Starek R, J Anal At Spectr 1989, 4 407 226 Bauslaugh J, Radziuk B, Saeed K, Thomassen Y, Anal Chim Acta 1984, 165 149 227 Welz B, Schlemnier G, Mudakavi JR. J Anal A t Spectrom 1988. 3 695

452

Sampson B, J Anal At Spectrom 1987; 2: 447 Maage A, Julshamn K, Andersen KJ, J Anal At Spectr 1991; 6:277 Manning DC, Slavin W, Appl Spectr 1983; 37: 1 Saeed K, J Anal At Spectr 1987; 2: I5 1 Saeed K, Thomassen Y, Anal Chim Acta 1981; 130: 281 Hocquellet P, Analusis 1978; 6: 426 Parsley DH, J Anal At Spectr 1991; 6: 289 Neve J, Hanocq M, Molle L, Anal Chim Acta 1980; 115 : 133 Neve J, Hanocq M, Anal Chim Acta 1977; 93: 85 Alexander J, Saeed K, Thomassen Y, Anal Chim Acta 1980;120:377 Robert RVD, Balaes G, NIM Report No. 2053 (1 980) Knowles MB, Brodie KG, J Anal At Spectr 1989; 4: 305 Donaldson EM, Talanta 1988; 35: 633 Meyer A, Hofer C, Tolg G, Fres Z Anal Chem 1978; 290: 292 Sanzolone RF, Chao TT, Analyst 1981; 106: 647 Sanzolone RF, Chao TT, Anal Chim Acta 1981; 128: 225 Garcia-Olalla C, Aller AJ, Anal Chim Acta 1992; 259: 295 Garcia-Olalla C, Aller AJ, Fres J anal Chem 1992; 342: 70 Jiang S, De Jonghe W, Adams F, Anal Chim Acta 1982; 136: 183 Aadland E, Aaseth J, Radziuk B, Saeed K, Thomassen Y, Fres Z Anal Chem 1987; 328: 362 248 De Kersabiec AM, Benedetti MF, Fres Z Anal Chem 1987; 328:342 249 Schron W, Bombach G, Beuge P, Spectrochim Acta 1983; 38B: 1269 250 Marks JY,Welcher GG, Spellman RJ, Appl Spectr 1977; 31.9 251 Bradshaw D, Slavin W, Spectrochim Acta 1989; 44B: 1245 252 Wagley D, Schmiedel G, Mainka E, Ache HJ, At Spectr 1989; 10: 106 253 Fernando LA, Kovacic N, Fres Z Anal Chem 1985;322: 547 254 Montaser A, Fassel VA, Zalewski J, Appl Spectr 1981;35:292 255 Boumans PW, Spectrochim Acta 1989; 44B: 1285 256 Botto RI,ICP Inf News1 1981; 6: 527 257 Schramel P, Li-Qiang Xu, Fres J Anal Chem 1991; 340: 41 258 Que Hee SS, Boyle JR, Anal Chem 1988; 60: 1033 259 Urasa IT, Anal Chem 1984; 56: 904 260 Coles SM, Plasmaline 1981; 2(2): 7 261 De Oliveira E, McLaren JW,Berman SS, Anal Chem 1983, 55: 2047 262 Pretorius L, Kempster PL, Van Vliet HR, Van Staden JF, Fres J Anal Chem 1992; 342: 391 263 Cave MR, Green KA, J Anal At Spectr 1989; 4: 223 264 Pyen GS, Long S, Browner RF, Appl Spectr 1986, 40: 246 265 Hayrynen H, Lajunen L, Peramaki P, At Spectr 1985;6(4):88 266 Heropoulos C, Seeley LS, Radtke AS, Appl Spectrl984;38:451 267 Szilvassy-Vamos Sz, Lazar J, Horvath M, Kertai-Simon A, J Anal At Spectr 1990; 5: 705 268 Timmins KJ, J Anal At Spectr 1987; 2: 251 269 Belcher R, Bogdanski SL, Henden E, Townshend A, Anal Chim Acta 1980; 1 13: 13 270 Bogdanski SL, Henden E, Townshend A, Anal Chim Acta 1980; 116: 93 271 Sansoni B, Brunner W, Wolff G, Ruppert H, Dittrich R, Fres Z Anal Chem 1988; 331: 154

228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247

453 272 Date AR, Hutchison D, Spectrochim Acta 1986, 41B 175 2 73 Evans EH, Ebdon L, J Anal At Spectr 1989, 4 299 214 Gray AL, Fres 2 Anal Chem 1986. 324 561 275 Henshaw JM, Heithmar EM, Hinners TA, Anal Chem 1989,61 335 2 76 Goergen MG, Murshak VF, Roettger P, Murshal I, Edelman D,At Spectr 1992, I3( I ) 11 277 Lyon ThB, G S Fell, Hutton RC, Eaton AN, J Anal At Spectr 1988, 3 265 278 Ting BTG, Mooers CS, Janghorbani M, Analyst 1989, I14 667 279 Vickers GH, Wilson DA, Hieftje GM, Anal Chem 1988,60 1808 280 Buckley WT, Budac JJ, Godfrey DV, Koenig KM, Anal Chem 1992, 64 724 28 1 Kingston HM, Paulsen PJ, Lambert G, Appl Spectr 1984, 38 385 282 Heumann KG, Grosser R, Fres Z Anal Chem 1989, 332 880 283 Zhdanov S Selenium, in Electrochemistry of the elements, Vol 4, JW Bard ed 284 Bound GP, Forbes S, Analyst 1978, 103 I76 285 Aydin H, Tan GH, Analyst 1991, I16 941 286 Shafiqul Alam AM, Vittori 0, Porthault M, Anal Chim Acta 1976, 87 437 287 Vajda F, Acta Chim Sci Hung 1970, 63 257 288 Campanella L, Fern T, J Electroanal Chern 1984, 165 241 289 Van den Berg CMG, Khan SH, Anal Chim Acta 1990, 23 1 22 1 290 Henze G, Monks P, Tolg G, Umland F, Wessling E, Fres Z Anal Chem 1979, 295 1 29 1 Kotz L, Henze G, Kaiser G, Pahlke S, Veber M, Tolg G. Talanta 1979, 36 681 292 Henze G, Mikrochim Acta 1981 11, 343 293 Ebhardt KB, Umland F, Fres 2 Anal Chem 1982, 3 10 406 294 Russu AT, Kopanskaya LS, Voznesenskii VA, Zh Anal Khim 1975, 30 I I53 295 Stara V, Kopanica M, Anal Chim Acta 1988, 208 231 296 Trivedi BV, Thakkar NV, Talanta 1989, 36 786 297 Baltensberger U, Hertz J, Anal Chim Acta 1985. 172 49 298 Arlt C, Naumann R, Fres 2 Anal t h e m 1976, 282 463 299 Lu Chung-Peng, Kwang-Kuo Hsieh, Hsien-Yin Tseng Fen Hso Hua Hsueh 1983, 8 223 300 Andrews RW, Johnson DC, Anal Chem 1975. 47 294 301 Posey RS, Andrews RW, Anal Chim Acta 1989. 124 107 302 Portyagina EO, Stromberg AG, Kaplin AA, Zh Anal Khim 1984, 39 493 303 Gornostaeva TD, Pronin VA, Zh Anal Khini 1981, 36 295 304 Aydin H, Somer G, Anal Sci 1989, 5 89 305 Aydin H, Yahaya AH, Analyst 1992. 117 43 306 Berengard IB, Kaplan B Ya, Zav Lab 1986, S2(3) I I 307 Howard AG, Gray MR, Waters AJ, Oroiniehie AR, Anal Chim Acta 1980, I18 87 308 Shchukin VD, Kozirov ID, Zav Lab 1978, 9 1057 309 Retief DH, Scanes S, Cleaton-Jones PE, Turkstra J, Smit HJ, Archs oral Biol 1974.19 517 310 Holzhey J Aktivierungsanalyse - Anwendung i n der Metallurgie, VEB Deutscher

Verlag fur Grundstoffiiidustrie, Leipzig I975 3 I I Vobecky M, Dedina J, Pavlik L, Valasek J, Radiocheni radioanal Lett 1979. 38 197 312 Baedecker PA, Rowe JJ, Steinnes E, J Radioanal Chem 1977.40 1 1 5 313 Landsberger S, Anal Chem 1988, 60 1842

454 314 315 316 317 318 3 19 320 321

Rowe JJ, Steinnes E, Jour Res US Geol Surv 1977; 5 : 397 Rowe JJ, Steinnes E, Talanta 1977; 24: 433 Kiesl W, Fres Z Anal Chem 1967; 227: 13 Boniforti R, Madaro M, Moauro A, J Radioanal Nucl Chem 1984; 84: 441 Kronborg OJ, Steinnes E, Radiochem Radioanal Lett 1975; 21: 379 Kut D, Sarikaya Y, J Radioanal Chem 198 1 ; 62: 16 I Krivan V, Geiger H, Franz HE, Fres Z Anal Chem 1981; 305: 399 Nakahara H, Tsukada M, Moriizumi A, Horiuchi K, Murakami Y, J Radioanal Chem

322 323 324 325

Czauderna M, Rochalska M, J Radioanal Nucl Chem 1989; 129: 425 Brune D, Bivered B, Anal Chim Acta 1976; 85: 41 1 Brunfeldt AO, Steinnes E, Talanta 1971; 18: I 197 Keays RR, Ganapathy R, Laul JC, Kriihenbuhl U, Morgan JW,Anal Chim Acta

1982; 72: 377

1974; 72: 1 326 Knab D, Gladney ES, Anal Chem 1980; 52: 825 327 Polkowska- Motrenko H, Dermelj M et al, Radiochem radioanal Lett 1982; 53: 319 328 Artem'ev 01, Aubakirova RB, Kiselev BG, Stepanov BM, Izv AN Kaz SSR 1975; Ser Fiz-Mat 6: 67 329 Artem'ev 01, Stepanov VM, Kovel'skaya GE, Zh Anal Khim 1978; 33: 493 330 Dermelj M, Stibilj V, Stekar J, Byrne AR, Fres J Anal Chem 1991; 340: 258 33 1 Sherif MK, Awadallah Rh4, Amrallah AH, J Radioanal Chem 1980; 57, 53 332 Raie RM, Smith H, Radiochem radioanal Lett 1970; 28: 2 I 5 333 Lievens P, Versieck J, Cornelis R, Hoste J, J Radioanal Chem 1977; 37: 483 334 Teherani DK, Altmann H, Michailov MI, Lazarova MS, Entschev DA, Radiochem Radioanal Lett 1981; 46: 221 335 Diksic M, McCrady MO, Radiochem Radioanal Lett 1976;26:89 336 Egan A, Kerr SA, Minski MJ, Radiochem Radioanal Lett 1977; 28: 369 337 McDowell LS, Giffena PR, Chatt A, J Radioanal Nucl Chem 1987; 110: 519 338 Woittiez JRW, Nieuwendijk BJT, J Radioanal Nucl Chem 1987; 110: 603 339 Turkstra J, Harthoorn AM, Beukes JL, Brits RJN, J Radioanal Chem 1977; 37: 473 340 Veres A, J Radioanal Chem 1977; 38: 126 341 Debrun JL, Albert P, Bull SOCChim France 1969: 3: 1020 342 Aras NK, Zoller WH, Gordon GE, Anal Chem 1973; 45: 1481 343 Segebade Ch, Fusban HU, Weise HP, J Radioanal Chem 1980; 59: 399 344 Berthelot Ch, Carraro G, Verdingh V, J Radioanal Chem 1980: 60.443 345 Olmez I, Aras NK, J Radioanal Chem 1977; 37: 671 346 Robinson JW,Handbook of Spectroscopy Vol 1, CRC Press Ohio 1974 347 Bohmer RG, Psotta PK, Fres Z Anal Chem 1990; 336: 226 348 Holynska B, Markowicz A, Radiochem Radioanal Lett 1977; 3 I . 165 349 Strausz KI, Purdham JT, Strausz OP, Anal Chem 1975;47:2032 350 Van Espen P, Nullens H, Adams FC, Fres Z Anal Chem 1977; 285: 2 I 5 351 Shenberg C, Mantel M, Izak-Biran T, Rachmiel 8, Biol Trace Elem Res 1988; 10: 87 352 Colautti P, Enzi G, Moschini G, Muzzio PC, Stievano BM, La Ricerca Clin Lab 1976; 6: 169 353 Garten RPH. Analytiker Taschenbuch IV, 259-286, Springer Verlag Berlin 1984 354 Liu QX, Shao HR, Moschini G, J Radioanal Nucl Chem Lett 1990; 144: 47 355 Berti M, Buso G et al, Anal Chem 1977; 49: 13 13

455 356 357 358 359 360 361 362 363 364

Maenhout W, de Reu L, Vandenhaute J, Nucl lnstr Meth Phys Res 1984; B3: 135 Agostoni A, Gerli GC et al, Clin Chim Acta 1983; 133: 153 Pella PA, Dobbyn RC, Anal Chem 1988; 60 684 Ketelsen P, Knochel A, Fres Z Anal Chem 1984; 317. 333 Raptis S, Dissertation, Graz 1979 Sinemus HW, Maier D, Fres Z Anal Chem 1985. 322: 440 Brimmer SP, Fawcett WR, Kulhavy KA, Anal Chem 1987; 59- 1470 Pettersson J, Olin A, Talanta 1991; 38 41 3 Negretti de Bratter, Bratter P, Tomiak A, J Trace Elem Electrolytes Health Dis

365 366 367 368 369

Welz B, Wolynetz MS, Verlinden M, Pure &: Appl Chem 1987; 59. 927 Raptis S, Kaiser G , Tolg G, Fres Z Anal Chem 1983;3 16: 105 Welz B, Melcher M, Neve J, Anal Chim Acta 1984; 165: I3 I Porter BL, Fuavao VA, Sneddon, J At Spectr 1983; 4 I85 Van den Berg CMG, Khan SH, Anal Chim Acta 1990, 231:221

1990;4:41

Table 1) NUCLEAR REACTIONS OF SELENIUM ISOTOPES [54,309,310] Isotope 74-Se 76-Se 78-Se

Natural abundance 0.87 YO 9.02 YO 23.52 YO

80-Se

49.82 %

82-Se

9.19 YO

thermal neutron cross section 30 barns 21 0.33 0.05 0.5 0.08 0.004 0.04

Some reactions with epithermal neutrons: 77-Se 7.65 % (n,p) 78-Se 23.52 YO (n,2n) 80-Se 49.82 YO 0.04 (n,cr) 82-Se 9.19 % 1.5 (n,2n)

product halfprincipal isotope life y-lines (keV) 75-Se 120.4 d 12 11136/265/279/401 77m-Se 17.5 s 161 79m-Se 3.91 m 96 weak 79-Se 65000 a no y 8 1-Se 18.6 m 280 weak 8 1m-Se 56.8 m 103 weak 25 m 360/520/830/13 10 83-Se 83m-Se 70 s 650/10 10/2020

77-As 77m-Se 77-Ge 81m-Se

39 h 17.5 s 11.3 h 56.8 m

Table 2) NUCLIDES INTERFERING WITH T H E y-RADIATION EMISSION SPECTRUM OF 75-SE 75-Se line keV 121.1

135 9

264 5

275 5 400 7

interferent line keV Eu 121 8 Ba 1 2 7 7 Hf 133 I

Yb Mo Ta Cd Hg

1307 1404 264 1 2609 2792

A u 411

resulting nuclide 152-Eu 131-Ba 181-Hf

half-I ife

169-Yb 99-TC 182-Ta 115-Cd 203-Hg

318d 66 7 h

l98-A~

27d

12.7 y 11 5 d 42 5 d

115 d

53 5 h 46 6 d

g Se/g interferent

Ref thermal n epith n 97 22 [3111 02 02 [3111 7 23 [3111 [3 121 [3131 24 17 [311] [3 I4,3 151 15 3 7 [311,313] 06 [3111 04 5 I 131 11 [3161 [91

This Page Intentionally Left Blank

459

SELENIUM OCCURRENCE AND ECOLOGY

1. SELENIUM IN T H E ATMOSPHERE For many urban areas in the Northern hemisphere, total selenium contents in the range of 0.3-20 ng/m' have been reported. In remote areas, Se goes down to values lower than 0.1ng/m3, the minimum has been measured on the South Pole (0 007 ng/m')[l-6]. The ChacaltaydBolivia has been taken as representative for remote locations in the southern hemisphere, the sampling station in the Canadian Northwest territories for the background of the northern hemisphere, and the Jungfraujoch/Switzerland for the background of Western Europe [7]. Notably, the arctic aerosol at Alert Bay /Northem Canada is somewhat higher in selenium than the aerosol at the South Pole [2]. An unusual high value has been reported for Ankara in winter 1976 [ l ] With respect to the average abundance in the earth crust of 50ng/g [8], Se in aerosols is generally enriched, even at its lowest values at the South Pole Enrichment of selenium is often estimated via the proportion over Fe in the atmosphere, with respect to SeRe of the average crustal abundance ("enrichment factor"). Enrichment factors of the aerosol in the range of 600- 10000 with respect to the surrounding soils are reported, which are quite high. In Ankara, in spite of the top values for the aerosols, the enrichment factor is lower, because Se in the soil is rather high [ I ] . Similarities of occurrence and chemical properties in coals and other fuels lead to the conclusion, that the output of SO,, which is responsible further for the acidification of rain, is accompanied by a parallel output of Se [6] (compare chapter 5. "Selenium in rocks"). Major sources of selenium input into the atmosphere are thus the combustion of fossil fuels (coal, oil), but also the processing of non ferrous metals (roasting of sulfide ores), as well as municipal incinerators [S]. In urban areas, distinct seasonal cycles can be observed, with maxima in December and minima in June, because of coal or fuel oil combustion in winter time. Rooted plants as well as soil microorganisms volatilize Se from soil (see chapter 7 ) In the background aerosols, however, there are no seasonal cycles [ 4 ] . Maximum to minimum concentration ratios are only at 3-4, whereas they are 30-40 for typically crustal elements [4] In Glasgow, however, a seasonal maximum in summer has been found, contrary to most other trace elements [3]. Selenium in western atlantic precipitation strongly correlates with acid concentration (0.959/n=15), and also with non sea-salt sulfate. Lack of correlation between Se and Na means, that it does not derive from marine sources [9] In the aerosol of St LouisiUSA, 75% of Se- contents can be statistically explained by a factor marked by sulfate as main component [lo]. Contrary to Pb, Br etc., traffic is no significant source of Se input into the atmosphere [ I ] . The residence time of aerosol particles depends on the aerodynamic diameter, which results from source, age and history of the particles [ 7 ] .High enrichment of Se in the atmosphere can be explained by the fact, that Se enters the atmosphere mainly as condensation aerosol and is thus present in very fine particles, or even gaseous During sampling on dry filter media, 19-35% of the selenium got lost, irrespective of the filter material, either because of volatility, or of small size of the particles [ 1 I ] This delays precipitation in comparison with "conservative" elements. like Fe or Al, which emanate mainly from soil abrasion. Enrichment of Se in the atmosphere is thus a result of increased volatility and slower precipitation. Se is removed from the atmosphere by processes such as wet deposition, through 11s association with sub-micrometre aerosols, and the efficient scavenging of these small

460 particles during precipitation events [9]. Removal by wet deposition predominates [ 9 ] , though the Se contents of rain water is lower than many surface waters Atmospheric deposition is an important input to oceanic surface waters [9]. Similarly, increasing precipitation correlates strongly positive with the selenium contents in humus layers of Norwegian forest humus soil or farmland soil [ 121. In Glasgow, however, the daily average weather conditions do not significantly influence the Se contents of the aerosol [3]. Only few papers deal with the speciation of selenium in aerosols. In rainwater samples from Japan and coastal California, selenite is the major Se- species. Rain and snow in Belgium contain variable quantities of selenite and selenate. In the precipitation of the Western Atlantic, however, the ratio of Se(IV)/Se(VI) lies within the narrow range of 1.26*0.95 [9]. Selenide and elemental selenium have been found in a laboratoy atmosphere [6], but they are negligible for outdoor sampling sites [9]. SeOz and Se" are released during fuel combustion, but also natural emissions of organic selenium compounds, mostly dimethylselenide, are known [9] (see also chapter 4. "Selenium in soils"). During the combustion of coal, the generated SO2 may partially reduce SeO, to elemental Se, which preferentially deposits on aerosol particles. On the other hand, oxidants on the troposphere, e.g. 0, or H,O,, may cause the production of selenate, which may serve as an indication reaction for strong oxidants [ 6 ] .

2. SELENIUM INNATURAL WATERS 2.1 TOTAL CONTENTS The selenium contents of natural waters can vary within a broad range, from <2ng/l [ 131 up to 300 pg/l [14]. For potable water, the upper tolerance limit in Austria is 10 pg/l (ONORh4 M 6250). High Se-levels are generally found in surface waters from Colorado, Wyoming, Utah, Western Australia, or Central Asia. These waters are all from arid regions, strongly oxidizing, and low in organic carbon [13,14]. Maximum Se- values are also observable in acid and neutral waters from sulfidic deposits, or in waters containing hydrogen sulfide ~41. The total dissolved selenium in the partially reducing fjord Saanich InletiCanada showed an increase in concentration with depth into the suboxic zone, then a general decrease within the deeper reducing and anoxic waters [ 151 In Lake Nasser, this is the reservoir of the Nile before the dam at Assuan, the total water contents of Se (dissolved + suspended) ranged between 0.25 and 5.3 pg/l. Layers of different contents of trace elements were observed, but Se did not depend on the depth or on the oxygen contents [ 161. In Italy, Se values tend to be very low in those rivers, which are not highly polluted [ 131. In general, Italian rivers show lower Se values than Canadian surface waters do [13]. The Se- supply to the oceans via the streams takes place mainly as dissolved species. Approximately 10% of the total load is adsorbed on suspended matter in the stream, but desorbed in the estuaries by the sea water [17]. In Nagoya harbour, selenium pollution attributable to industrial and domestic waste waters have been detected, because the amount in the rivers running into the sea at this site is very large. However, the Se- contents of the estuarine water and in the suspended particles were found to decrease in accordance with the distance of the sampling place from the coast [IS], which means steady sedimentation in the estuary.

46 I Dissolved selenium in ground- and surface waters does not primarily depend on the Secontents of the host rock, nor on the main components dissolved, but on the redox potential at its source, which can be explained from speciation studies [ 141 Enhanced levels can be expected either at >200mV (important for potable water resources), or at <-I00 mV (sulfidic waters) Other parameters, like Ca, Mg, HCO,, SO, - contents, are without significance Se does not seem to be mobilized during hydrothermal processes [ 13,141 Thermodynamic calculations lead to the prediction. that selenate should be the exclusive oxidation state in oxic sea water In anoxic reducing systems, on the other hand. selenide should become the stable dissolved form [ 151 2.2 SELENATE

Selenate does not form insoluble compounds and is hardly adsorbed It is thus very mobile in soils and ground waters, and overall comparable with sulfate (see chapter 4) Se i n ground waters of Italy and Western Australia, where a considerable proportion of selenium in selenate form IS assumed, does not correlate with other mobile anions in high oxidation states, like vanadate or uranyl [ 131 In Saanich Inlet, a reducing fjord at the Western coast of Canada, in the surface waters nearly all Se is present as selenate Selenate continuously decreases to values lower the detection limit with increasing water depth [ 151 In Nagoya harbour, the proportion of selenate varies widely, due to vaiious inputs [ 181, because selenate is possible in effluents, too In pore waters of the sediment, selenate is found under highly oxidizing conditions [ 191 In carstic ground waters (Rottenacker, BRD) as well as the Moscow tap water, selenium is preferably present as selenate [20,21] In ground waters near a surface coal mine, oxidation of Se- bearing pyrite lead to enhanced levels o f both selenate and sulfate [22], whereas dissolution of gypsum did not release significant S e Selenate was found to be slowly reduced to selenite by huinic acid in acidic solution (pH 3 24, [23], see chapter 4) Plant uptake of Se during the vegetation period causes an annual cycle of Se- contents in surface waters, exhibiting a maximum in May, and a minimum in mid- August While in summer the selenate dissolved in open streams is ieduced at only minor amounts, i t can be reduced to more than 90% in corresponding wetlands [24], and thus accumulated 2.3 SELENITE

Selenite has generally low geochemical mobility From sea-water, Fe- hydroxide coprecipitates 82% of selenite, but only 2 5% of selenate [25] Coprecipitation upon Alhydroxide, however, is much lower, I S takes 31% of selenite, and no selenate [25] Montmorillonite and illite adsorb appreciable amounts too, whereas adsorption on kaolinite, humic acid or silica is quite low [ 17,231 Both on Fe- hydroxide and on clay minerals, the absorbability of selenite decreases with increasing pH, with a particularly diastic deciease between pH 9 and 1 1 Above pH 1 I , it I S not adsorbed at all [ 13,251 At any pH, selenite adsorption exceeds the adsorption of selenate [26] Montmorillonite and kaolinite show different pH- dependencies as sorbing agents Selenite sorption by clay minerals IS affected to a greater extent by pH than by layer wpe

462 2.4 ORGANIC SELENIUM AND SELENIDE Organic selenium, maybe often termed as Se" because of lack of the identificatlon of the organic compounds, is hardly found in potable water [20]. In the water of Saanich Inlet/Canada, dissolved organic selenide begins at 70m water below surface, it is maximum fraction in suboxic and anoxid zones. At least 65% of organic Se have been found to be accounted for in the total amino acids fraction. The selenium is incorporated into many biota in the surface waters preferably as soluble selenite, and liberated in deeper layers, when dead biota or fecals dissolve [ 151. No dimethylselenide was found in anoxic waters, though transmethylation is a common process for the volatilization of selenium from soils (see chapter 4). Free selenide is hardly found, probably because of the stability of ferroselite FeSe, or coprecipitation wlth pyrite [ 151.

3. SELENIUM INSEDIMENTS 3.1 TOTAL CONTENTS

With respect to the global crustal abundance of Se, which is estimated to be 50ng/g [8], Se in both limnic and marine sediments is generally enriched. This can be explained by adsorption phenomena and geochemical affinities. The mean sediment Se- contents has been estimated to be 0.42 pg/g (like soil), and deep sea clay as 0.17 pg/g [27]. The bottom sediments of Kentucky and Barkley LakedUSA are in a rather low range (0.11-0.59 pg/g), with a mean of 0.32 p/g, The values were rather consistent and did not vary greatly from site to site [28]. Contrary to this, toxic amounts of Se (7.5 - 40.7 pglg) have been found in sediments of the Kesterson Reservoir (Merced County, California, USA) [29]. These sediments are alkaline (pH 8.1-8.5, saline, and the evaporation of the water body is large. Deep sediment from Lake Ontario with a contents certified as 1.02 pgig has been taken to produce standard reference material. It is moderate in organic carbon (2.95%) and calcium (2.07% CaO), but high in silica (61.4% S O 2 ) [30]. Another standard reference material, SGR-I Green River Shale, is quite high in Se with 3.4 pg/g [31]. Very high selenium contents has been found in the sediments of the Scheldt River in Belgium (9.1 and 9.3 pg/g at two sites)[32], which may be due to pollution. This Scheldt sediment is anoxic, heavily polluted with Ag, As, Cr and Zn, and in addition shows enhanced values of Cd, Pb, and Te [32]. Chinese river sediments, prepared as certified standard reference materials (GSD 1 to 8) cover the range from 0.1 1 to 1.06 pg/g. In the Gulf of St. Lawrence, the mean contents of surface sediments (up to 35 cm sediment depth) was 0.37 pg/g at the head of the estuary, and increased towards the open sea to I 28 pg/g. This is in accordance with the apparent precipitation of selenium inputs in the harbour of Nagoya (compare chapter 2). The depth profile shows rather constant concentration versus depth in this anoxic sediment, but there are shifts in speciation [34,35; see below). Deep sea sediments from the Pacific cover a range from 0.5 to 9.8 pg/g Se [36]. Dependency from the depth of the ocean may be possible, but the data are too scarce to conclude this.

46 3 With respect to sequential leaching, according to the Tessier sequence [34], oxalateleachable is the maximum desorbable fraction for Se, which in the depth profile of sediment cores is at maximum in the first 2cm, where conditions have not been totally anoxic In the pore water of sediments of the Gulf of St Lawrence, the concentration profile of soluble Se is strikingly similar to dissolved Fe and P, at a maximum of 10-12cm below the surface 1341 The total Se in the solid, however, is not correlated to total Fe [35] 3.2 SPECIATION

When selenate is added to freshwater sediments, it is partially reduced to Se(1V). Similarly, elementary selenium is autooxidized to Se(IV and then partially absorbed by the sediments [37], provided there is available oxygen. In spite of reducing conditions, in deep sea sediments from the Pacific, the sum of selenite and selenate has been found generally larger than reduced and organic forms of Se [36]. In aquatic environments, Se is depleted in the sediment together with biogenic material. When this biogenic material decomposes, Se is recycled, or it is captured by a sulfur mineral phase such as pyrite, or precipitated as ferroselite FeSZin case of low sulfur contents. As selenite, it is strongly adsorbed by iron hydroxide. Sorption of various Se- species on sediment components has not been done so far, but much can be concluded from respective experiments with soil components. Affinity and relationship to sulfur explain, why Se is often found together with S in metal deposits of Cu, Zn, Ag, and Pb, and is associated with Fe in both oxidizing and reducing environments [35] The affinity to sulfur is independent of the action of water, it is also noticeable in meteorites. In the St Lawrence estuary, at or near the sediment- water interface, adsorption of Se onto Fe- oxyhydroxide occurs In deeper layers, the reduction of Fe leads to a release of adsorbed selenite, which is removed by formation of ferroselite FeSe, at depth This explains the maximum occurrence of Se in the pore water at 12 cm, together with soluble Fe In sediment columns of Kesterson Reservoir (California, 4-1 7% org C), increasing the temperature from 15" to 35O promoted volatilization of Se about double Addition of organic amendments promoted volatilization, to yield dimethylselenide, dimethyldiselenlde and other gaseous compounds, at different rates, gluten, a wheat storage protein, had the most pronounced effect No Se alkylation occurred i n autoclaved soil, indicating a biological transformation [29]

&SELENIUM IN SOILS 4.1

OCCURRENCE

With respect to the average crustal abundance, Se in soils is enriched, the estimated average is 0 4 pg/g [27] For agriculture, soils with more than 5 pg/g are regarded to be too high, and soils with less than 0 03 pg/g to be too low for optimum crop produciion (ONORM L 1075) The median Se- contents of surface soils in the east of Austria is 0 23 pg/g [37,38] Soils of cornproduction were significantly lower in Se (0 20 pg/g) than soils in the greenland (0 29 pg/g) The lithological facies is also of significant influence on the actual Se-content in

464

non-contaminated areas. In the east of Austria, the soils from granites and quartz were lowest, and above schist they were highest in total Se [37] Canadian certified standard soils range from 0.04 to 1.0 pg/g Se. Vineyard soils at Ingelheim (Rhine province/FRG) are in the range 0.18-0.44 pg/g, with a mean of 0.24 pg/g [39]. In the Bavarian nature reservate at UnterschleiRheim, soil selenium contents has been found to be in a range of 0.57 - 2.0 pg/g Se. [40]. In China, it is reported that the soil of Se- deficiency regions has a selenium content of 0.09-0.17 pg/g, and toxicity regions have more than 1.70 pg/g [12]. For European conditions, this upper value, is however, far too low; european soil can have much more, and in spite of this, population suffers partially from Se- deficiencies. In Germany, no Se- toxicity symptoms have been recognized on a soil with 4. I pg/g [40], and in Norway with 2.0 pg/g [ 121. In Norway, some concern has been about the input of Se into the soils by atmospheric deposition ("acid rain"- problem). In the Se- contents of Norwegian soils, 2 distinctly statistical groups have been found, namely Inland soils with a mean of 0.2 pg/g Se, and Coastland soils with a mean of 0.9 pg/g Se. The comparatively high Se level in Norwegian coastal areas has been found to be associated with the amount and the composition of precipitation, as well as a high contents of organic matter and clays [ 121. In sewage sludge, Se contents is usually higher than in uncontaminated soil, compost and commercial fertilizers, the range given from literature data is 0.4 - 9.6 pg/g dry weight. Thus, fertilization with sludge leads to a slight increase in surface soil Se [41]. 4.2 ADSORPTION/DESORPTION In terrestrial, three major transformation mechanisms of ecological significance for Se are possible, which are ixidation/reduction, mineralization/incorporation, and volatilization [42]. The adsorption capability increases from selenate to selenomethionine to selenite. Whereas selenate is rapidly leached from soil columns, it is inicrobially reduced in soils selenite and organo- selenium compounds, and to a minor part volatilized via methylation [42], depending on the organic carbon contents and the oxygen in the pore gas. At elevated carbon contents (7%), transformation of added selenate preferably occurs to organoselenium compounds, which are more mobile than selenite. Selenate is only slowly reduced to selenite by humic acid under sterile conditions [23]. Without carbon, selenate is just reduced to selenite at a redox potential of 200-400 mV, and thus fixed [43]. Sorption of selenite chiefly takes place on FeiMn- hydroxides as well as on clay minerals (see also chapters about groundwater, and about sediments). Fe-bearing silicate minerals, like augite, hornblende and vermiculite can also adsorb substantial quantities of selenate, they may probably act via their weathering products, i.e. hydrated oxides, on their surfaces [44]. From calcite and apatite, selenate is easily desorbed. For the clay minerals, adsorptionidesorption behaviour depends on pH and on other salts dissolved in the soil pore water. The adsorption of selenite exceeds selenate at any pH and any model substance up to 20 times [45,46]. Cyclic irrigation reduces the leaching of selenium in a soil profile as compared to a steady-state irrigation regime [43]. In Japanese paddy soils, selenite is liberated from Fe(OH), during reduction, and moves down with percolating water.-, which leads to a distinct accumulation in the B- horizons., bound to fulvic acids. The proportion of Se in the humic acid fraction, however, decreased with increasing depth, and was thus not involved in the movement od Se [47]

465

The greater part of soil Se can be extracted repeatedly with dilute NaOH Not extractable are soil minerals and some organic fractions [47] Organically bound Se was found to be an appreciable part (20%) extractable with 0 IM-NaOHIO IM- Na,P20, Among the operationally defined organic compounds, Se was found in acidic hydrophobic fulvates, humates, and hydrophilic fulvates, but not in basic hydrophobic fulvates The hydrophilic fulvates containes about 20% of Se as selenomethionine [48] The selenate contents of soil is of some interest, because i t seems to be the only available form for the uptake by roots of higher plants In root systems of tomato, only negligible amounts are transported as selenite [49] In case of low Se- contents of soil, selenium compounds can be added together with the fertilizer as an essential trace element (e g in Finland) Fixation by the soil has therefore be of concern Three reasons might be, why added selenium is not effective for plant growth Se can be fixed too strong to be available to plants, it can be leached down below the rooting zone, or it gets lost by volatilization [SO] In alkaline soils, adsorption of both selenite and selenate was increased by organic carbon, clay content, calcite and cation exchange capacity, whereas high salt content of the pore water, high alkalinity and high pH act against [51] Addition of CaCI, results in significant increase in selenite adsorption above pH 5 [ 5 2 ] However, in alluvial soils, no discernible change in the amount of selenite adsorbed was found as a result of increased NaCI- concentration or through the addition of sodium sulfate[52] When Se was added to soil of Madras (India), a quarter was fixed within I 5 hours About half of it could be desorbed with CaCl, solution, nitrates and sulfates had no effect on this desorption The desorption of selenite showed no preference for sulfate or phosphate in most cases, but selenate was more efficiently desoibed by phosphate than by sulfate in all soils [ S 3 ] In highly organic soils, phosphate is more effective than sulfate i n displacing sorbed selenite, and i n saline soils it is reverse In Danish mineral soils, low in organic carbon, added Se is retained in the upper 7cni layer and is hardly leached by water [45] Liming increases the mobility of Se to a great extent, maybe because of pH- shifts [50], whereas organic matter decreases the movement The fixation of selenite by soils is a rather slow process. equilibrium takes a couple of days In 5 alluvial soils from California, no discernible change in the amount of selenite adsorbed was found as a result of increased chloride concentration, or through the addition of sodium sulfate The input of phosphate, e g as a fertilizer, can lead to a desorption of selenite because of competition [ 5 2 ] In peat, added selenate (e g from atmospheric deposition) moves downwards, accelerated by liming, but selenate is slowly reduced to selenite by humic acid [52] 4.3 VOLATILIZATION

Formation of volatile compounds can happen from any selenium species, it is most rapid with selenite, and slowest with elemental selenium [54] Bioniethylation occurs at 500mV, 200mV. and 0 mV, to yield dimethylselenide and similar compounds [ 191 Volatilization of spiked selenite from soil depends on the microbial activity temperature, moisture, water- soluble Se, and, last not least, on the season of the year, when the soil sample was collected Soils collected i n spring evolved most, from autoclaved soil, no volatilization took place Volatilization versus moisture content passed a maximum at 28%, which was far below saturation for the soil used [S5] Addition of selenate to fine-loamy,

466 calcareous, non- water- saturated soil led to volatilization of Se as dimethyl selenide or dimethyl diselenide to only 1.8 - 4.3% of input [42]. Microorganisms (like Aspergillus, Candida, Cephalosporium, Penicillium, and other fungi and bacteria) can reduce selenite to the volatile species dimethyl selenide, dimethyl diselenide, and dimethyl selenone. Fungi and bacteria contribute nearly equal. Therefor, the addition of organic matter to the soil increases the volatilization of Se, promoting the microbial activity. Thus, selenium is much more volatile (and also mobile) from peat than from normal soil [45]. Under aerobic conditions, dimethyl selenide is the main volatile compound, under anaerobic conditions it may also be H,Se [45,54].

LSELENIUM INROCKS A N D A L L O Y S In rocks, Se is a truly "chalkophilic" element in the term of Goldschmidt; that means, it occurs preferably in sulfidic phases. Selenide minerals can be found together with pyrite, chalkopyrite, sphalerite, galenite, argentite and other sulfide minerals of various ore geneses [56]. In northeastern Sibiria, a selenium containing polybasite (Ag-S-Se) with 4.80-8.20% Se, and naumannite (Ag-Se) with even 22.4-23.05% Se have been found 1561. Because Se and Te are important in the control of mechanical properties of alloys, ore processing tries to diminish their contents, Thus, in CCRMP reference ores Cu concentrate has 113/188 pg/g, Pb concentrate 28 pg/g, and Zn concentrate only 4.3 pg/g [57]. NBSstandard reference materials for steel and Fe- based alloys cover a range of 0.14 -237 pg/g Se [58]. In silicate rocks, Se is generally below its average crustal abundance, though even proper selenosilicate phases are known to exist. Se in basalt, diabase and granodiorite of North America is met at 0.060-0.120 pg/g [59,60,61], similar amounts (range 0.025 - 0.17 pg/g) occur in granites, leukogranites, granitoides etc. of the East Baikal region [62]. In limestone, the average Se contents is estimated minor 0.03 pg/g [8]. In the backfill material of a coal mine, main amounts of Se were water soluble and ionexchangeable (versus 2M- NH,- acetate), about 10-20% were in organic association (i e. volatile in high temperature ashing), only 6% in solid pyrite, but still 15% bound to silicates (fraction soluble in HF only) [22].

6. SELENIUM INMETEORITES The composition of meteorites indicates on the one hand cosmic abundances, and on the other it shows the affinities of the elements to one another, without the action of water or biota. In stony meteorites (chondrites), Se is dispersed in the silicate phases, and usually within the range of 4.8-26 pg/g [63,64,65]. These chondrites are regarded as only slightly metamorphosed aggregates of condensation products of the solar nebula [66]. In chondrites, the Se/S and S e n e ratios are rather constant [64,67]. In iron meteorites, Se prefers the troilite (FeS) phase (90-262 pgig), and to a lower extent graphite, whereas the metal and phosphide phases (Schreibersite) are nearly clean of Se (e.g. 5.5 ng/g in the metal phase of Cape York iron meteorite [68,69]. In lunar breccia (Apollo 17 mission), Se was found in the range 3.9-1 10 ng/g [70]. Extremely low Se- contents have been found in lunar glasses (0.74-4.1 ng/g)[71,72].

467

7. SELENIUM IN COAL AND COMBUSTION PROCESSES Se in coal generally occurs at elevated levels, probably because of its sulfide phase, the origin plays an important part [73,74] The Se- contents of Austrian lignites has been found to be 0 8 - 19 4 pgig, with a mean of 1 4 pgig [75] In 101 US- coals, Se is in the range of 045-7 7 pgig, with a mean of 2 08 p g i g [74] Turkish coals are reported to have 0 75-3 02 pgig Se [5] When the coal is burned, Se is easily volatilized (see also chapter 1 selenium in the atmosphere), the volatilized amount depends very much on the combustion process [5] Se is usually enriched in the fly ash (electrostatic precipitator), and is hardly found in slag-bottom ash [76,77] If the ash is alkaline, however, its Se contents is high [73] In the process of coal gasification, Se preferably moves to the filtered particulate matter, whereas the gasification char is lower than the original feed coal [78] In furnace waste dumps, 3 8-4 55 pgig Se have been found, only less than half of it was leachable with mineral acids, which means that Se is quite non-mobile from this matrix, when deposited in landfill sites [79] In fuel oils, used in Belgium and in Turkey in 1976, 2 4-7 5 p g i g has been determined, this is slighty more than In the respective coal for heating in homes [ S j

&SELENIUM N P L A N T S A N D A N I M A L S (non- food) In plants, the selenium contents varies over a rather broad range, and apparently depends on the selenium contents and speciation in the soil [3 1,79,80,81]. The tendency of plants to accumulate Se , partially corresponds with their sulfur requirement. Leaves and kernels usually have higher Se contents than roots and stalks [82]. Green plants (vegetables) take Na-selenate from soil solution more readily than native Se, which is mainly selenite [49,83] Selenium uptake by plants grown on kaolinite-textured soils was significantly higher than grown on sandy and clay soils. Under neutral, oxidizing and low- carbon conditions, Se- uptake is governed by drainage water rather than concentrartion in the soil

PI]. In simple drying, alfalva looses only 3% of its accumulated Se, whereas Astragalus racemosus releases up to 60% [83]. Some higher plants and mushrooms are known to accumulate Se up to 100 pg/g, e.g. astralagus, or boletus edulis [39]. NBS orchard leaves standard reference material has only 76 ng/g [59], which seems representative for green leaves Moss in the nature reserve at Unterschleifiheim /FRG, where pollution with heavy metals is generally low, contains 0.03-0.54 p g i g (mean 0.24 p g i g ) , which on the average is 116 of the surrounding soil Moss has a large surface, but no direct connection with the soil, it gets most of its traces by atmospheric deposition and is thus indicational [40]. At constant input into a riverine system, water plants in free flowing streams accumulate less Se than in the corresponding wetlands. Due to variable growth and consumption rates, the Se- contents of water plants undergoes a seasonal cycle, with a maximum concentration for stream plants in May, and for wetland plants in September/October [24]. Duckweed (lemna minor) had the highest accumulation of Se (max. I 1 5 pg/g) [24]. Marine plants and plankton are not higher in S e than terrestrial plants In pelvetia canaliculata (Atlantic ocean), 0.84 pgig have been found [25]. In plankton from Tsukumo BayiJapan, sampled at a depth of 10m, major amounts of Se are found metabolized [35].

468 For algae, because of different tolerance levels, speciation shifts in populations at a Se-level of 1-40 mg/l occur.Blue- green algae (e.g. anacystis nidulans, anabaena variabilis) are far more tolerant to selenate than to selenite, while diatoms are killed at even < I mg/l selenate. Contrary, when Se is present as selenite at the same level, blue-green algae are absent, and diatoms develop [84]. A survey of the trace element contents of migrant wading birds clearly shows, that dunlins accumulate pollutants, including Se, from their marine moulting and wintering grounds in Western Europe. Though Se is found in the kidney at rather high levels, after the birds< departure from the marine environment for the clean freshwater breeding areas in Northern Skandinavia, Se- levels decline rapidly [85]. Liver tissue of wild herbivorous mammals living in game reserves in South Africa (mountain goats, gazelles, impalas, kudus, buffalos) show seasonal variations of Se and Cu contents, with maxima in winter and minima in summer (93 samples). Se did not correlate with other trace elements determined [86].

9. SELENIUM IN FOOD

Selenium is an essential element The action of various Se- species in biota is quite complex and cannot be considered within this context Humans need about 50-200 pg Se/day, below 15 pg, deficiency symptoms develop For human and animal nutrition in Austria, upper tolerance levels for Se-contents in Food are within 0 5-1 0 pg/g dry weight Data about potable and mineral water are compiled in chapter 2 "Se in natural waters" 9.1 BEVERAGES

The Se- content of wine is usually more than in potable water from the same location. Data from various parts of grape-vines in Germany show, that S e is presumably taken only by the roots (that means, no uptake from atmospheric deposition - contrary to moss), and moves to the leaves, and partially into the stems. Especially in the grapes, the contents is very low. Red wines contain some more Se because of different ways of processing. Further, the soil composition is of some influence. Thus, German red wine from a soil of 0.24 pg/g has an average of 0.79 pgil, whereas another German red wine grown on calcareous soil with only 0.09 pg/g Se, has an average of 0.40 pg/l [39]. Belgian beers are reported to contain 0.2- 15.2 pg/I [39]. Much more Se than in wine is present in commercially available (FRG) fruit- and vegetable juices: 110-120 pgA [39]. For beverages, contamination of Se by vessels can only happen with red ruby glass, which is coloured with up to 0.4% Se; within I year, it can leave 1-2 pg/l to imbottled wine [39]. 9.2 CEREALS AND VEGETABLES

S e in cereals can be comparatively high. Wheat flour, rice flour, rape groats and soy groats are much higher than common dairy products [59]. The Se- contents of wheat grains from Colorado/USA ranged within 0.3-1.2 pg/g and was not affected by ammonium nitrate fertilization [87]. In oats grown on parachernosem soils in the East of Austria (median Se: 0.24 pg/g), 0.046 + 0.014 pg/g were found in the edible parts [38]. Maximum values for corn are reported from seleniferous soils in South Dakota, containing 29 pg/g Se [82].

469 Fertilization with ammonium nitrate did not affect the Se- contents of wheat grains in the drylands of Eastern Colorado The ammonium bicarbonate- DTPA- extractable level of soil-Se was a good indicator of Se in the grains, if soil samples 0-90 cm were used, the surface soil Se till 30 cm depth was a very poor indicator of Se- availability [87] Among vegetables, atragalus bisulcatus and broccoli strongly accumulate selenate from soil, whereas tomatoes react moderately Astragalus had the highest tissue selenium concetration and Se- volatilization rates, but in terms of S e removal from soil, broccoli was more effective [83] Some Se is methylated and thus volatilized, proportional to plant tissue concentration and competing with microbial acitivity [83] In water culture, corn and barley metabolize added selenite preferably to selenomethionine [41] Metabolization of Se in seleniferous cabbage lead to Se-methylselenocysteine, Se-methylselenocysteine selenoxide, and selenocystathionine, whereas selenocystine and selenopeptides were negligible 9.3 DAIRY PRODUCTS

Human milk (Dusseldorf/FRG) contains Se in a range of 98-155 ngig, probably i n a readily available form as selenoaminoacids (no speciation given) Commercially available milk formula baby food has only 12-63 ng/g, probably because cow milk is lower i n Se Thus, a formula- fed infant receives only about 3 5 pg Se per day, while a breast- fed infant gets about 13 3 pgiday [88] However, breast milk in Japan I S reported with only 23 ng/g, in spite of higher Se- plasma levels in Japan [89] 9.4 FISH AND M E A T

In the US, the maximum allowable level for fish food is 2 pgig, established by the National Health and Medical Research Council 1977 [28] Fish muscles cover a rather wide concentration range, probably dependent on feed habits Generally, marine samples (Northwest Atlantic) contain higher levels of total Se ( I 68-825 ng/g) than freshwater samples (Eastern USA, 143-576 ngig) Highest levels have been found in tuna fish Selenium upotake in freshwater fish seems to be species dependent, rather than on the sediments contents of the lake In Kentucky Lakes, Channel catfish, bass and white crappie were within the range of 30-1 50 ngig, whereas sometimes much more was found in drums For the drums, Zn was also enhanced, this may be related to th e feeding habit - drums feed much more on freshwater mussels than the other species Cd, Cr, and Cu did, however, not differ between the 4 fish species investigated in Kentucky lakes [28] In fish, 14-36% of Se is present as selenate, due to uptake from the water and aquatic organisms, the rest is selenite, and organically bound [90] There I S no reported evidence of in vitro or in vivo formation of selenate in animal tissues The main fraction of Se in fish muscle, as well as in selected human and animal tissues and chicken eggs, was water extractable (57% for fish on the average) In this extract, trichloro- acetic acid precipitate took about 10% of the Se Selenate was more extractable with water than other forms Contrary to this, molluscs and crustaceans contain Se mainly in a non-water-extractable form, which may be attributable to non-polar proteins or lipids 1911 In spite of some known interactions, Hg Se ratios, or CH,Hg Se ratios vary widely, no correlations were found [91]

470 Baking and broiling do not result in major Se losses from seafood [91]. No significant differences between animal oil- packed and water packed sea- food was noticed [91]. In reindeer muscle of Finland, Se- contents is due to local differences. Whereas Lapland reindeer are in the range 1.34-4.21 pg/g (mean 2.21 pgig), in Ostrobothnia only 0.17-0.71 pg/g (mean (0.34 pg/g) were found. Cows from both areas, however, were not different, and contained 0.17-0.95 pg/g (mean 0.46 pg/g) in their muscles [92].

10,SELENIUM =HUMANS For diagnostic and therapeutic reasons, many data about human serum, plasma, erythrocytes, and whole blood have been done. Though Se is highly controlled by metabolism, local differences occur, because of differences in soil contents and nutritional habits. In all countries, erythrocytes contain more selenium than whole blood; plasma and serum are on the same level. In terms of a world-wide comparison, the Se- level in the USA is highest, and in New Zealand it is lowest. Europe and South America are in the lowest third [88,89,93,94,95,96]. Se- levels of human whole blood and serum have been recently reviewed [97]. Se- contents in serum of children is age- dependent. It reaches a minimum in the first half year after birth and a maximum in the adult age, finally it declines at ages beyond 60 years [88,96]. In whole blood from healthy adults in Belgium, however, no significant differences due to sex and age were found (22-65 years of age). The mean values of the districts ranged from 96 to 129 pgA. In the districts of higher fish and pork consumption, whole blood Se was significantly higher [97]. Serum- Se depends also on the altitude of the habitat. Subjects living at high altitude in the Andeans in Peru (3800m), show significantly higher haemoglobin levels than in Lima at sea level; Fe and Cu serum levels were similar, but Se was lower in the Andean population [94]. During pregnancy, Se in plasma remains constant, though glutathionperoxidase activity is lowered [95]. In New Zealand, Se- levels in blood of cancer patients were no lower than those of elderly subjects and patients without cancer [98]. The average contents of blood samples from female patients sufferuing from Mamma carcinoma in Vienna was found to be 66 pg/l [38]. During blackfoot disease in Taiwan, Zn and Se are significantly lowered in plasma and urine, whereas As remains constant [96]. Horses in KentuckyLJSA have considerable lower Se-levels in blood plasma, than men i n the US- average [89,99]. Analyses of Se-contents of human hair I S doubtable, because it is greatly increased by Se containing anti- seborrheic shampoos (0.42 pg/g up to 9.8 pg/g in one treatment), which cannot be removed either with detergents, acids or complexation with DAN [ 1001. Probably uncontaminated human hair and finger nails contain 0.28-0.33 pg/g, and 0.55 pg/g, respectively [ 10 I]. Freshly extracted canine teeth from Bantus in the Witwatersrand area, a healthy population without caries, has a mean content of only 80 ng/g; the chemical form of Se in dental enamel is unknown [102].

47 I

REFERENCES 1 Olmez I, Aras NK, J Radioanal Chem 1977; 37: 671 2 Landsberger S, Anal Chem 1988; 60. I842 3 McDonald C, Duncan HJ, Atmospher Environm 1979; 13: 413 4 Adams F, Van Craen M, Van Espen P, Atmosph Environm 1980, 14: 879 5 Kut D, Sarikaya Y, J Radioanal Chem 1981; 62: 161 6 Muangnoicharoen S, Chiou Kuen-Yuan, Manuel OK, Talanta 1988; 35. 679 7 Adams FC, Van Craen MJ, Van Espen PJ, Env Sci Techno1 1980; 14: 1002 8 Morgan JW,Proc Lunar Sci Conf 7th, 1976; 2 189 9 Cutter GA, Church Th M, Nature 1986; 322: 720 10 Puxbaum H, Wopenka B, Fres Z Anal Chem 1984, 3 17: 278 11 Shendrikar AD, West PD, Anal Chim Acta 1977; 89. 403 12 Wu Xiaofu, Lag J, Acta Agric Scand 1988; 38. 271 13 Dall'Aglio M, Ghiara E, Proietti W, Rend SOCItal Min Petrol 1978; 34: 591 14 Krajnov SR, Gudz' ZG, Zakutin VP, Enikeev NI, Pribytkova SM, Geokhim 1983; 359 15 Cutter GA, Science 1982; 217: 829 16 Sherief MK, Awadallah RM, Grass F, J Radioanal Chem 1980; 60: 267 17 Kharkar DP, Turektan KK, Bertine KK, Geochim Cosmochim Acta 1968; 32: 285 18 Itoh K, Nakayama M, Chikuma M, Tanaka H, Fres Z Anal Chem 1985; 321: 56 19 Masscheleyn PH, Delaune RD, Patrick WH, J Environm Qua1 1991; 20:522 20 Nazarenko 11, Kislov IV, Zh Anal Khim 1978; 9: 1857 21 Heumann KG, Grosser R, Fres Z Anal Chem 1989, 332: 880 22 Dreher GB, Finkelman RB, Environ Geol Water SCI 1992;19:155 23 An Peng, Langqiu Xu, Sci Tot Environm 1987; 64. 89 24 Allen KN, J Environm Qua1 1991; 20:865 25 Chau YK, Riley Jp, Anal Chim Acta 1965; 33: 36 26 Bar-Yosef B, Meek D, Soil Sci 1987; 144: 1 I 27 Salomons W, Forstner U. Metals in the Hydrocycle, Springer Berlin- HeidelbergNew York 1983 28 McClellan BE, Frazer KJ, Univ Kentucky Research Report 1980; 122 29 Calderone SJ, Frankenberger WT, Parker DR, Karlsson U, Soil Biol Biochem 1990; 22:615 30 Cheam V, Chau ASY, Analyst 1986; 109. 775 3 1 Briggs PH, Crock JG, USGS Open- File Report 1986, 86-40 32 Berthelot C, Carraro G, Verdingh V, J Radioanal Chem 1980; 60: 443 33 Takanayagi K, Belzile N, Mar Chem 1988; 24: 307 34 Belzile N, Lebel J, Chem Geol 1988; 68: 99 35 Terada K, Ooba T, Kiba T, Talanta 197.5; 22. 41 36 Huang PM, Oscarson DW, Hammer UT, Lepinski NG, Liaw WK, Can Tech Rep Fish Aquat Sci 1982; 1163: 97 37 Danneberg 0, pers commun. 1993 38 Sager M, unpublished results 39 Eschnauer HR, Alt F, Messerschmidt J, Tolg G, Fres Z Anal Chem 1989; 332. 874 40 Wolf A, Schramel P, Lill G, Hohn H, Fres Z Anal Chem 1984; 317. 512 41 Cappon CJ, Sci Tot Environ 1991, 100.177 42 Alemi MH, Goidhamer DA, Nielsen DR, J Environm Qua1 1991. 20: 89 43 Neal RH, Sposito G , J Environm Qua1 1991; 20.808 44 Bird GW, Lopata VJ, Sci Bas Nucl Waste Man 1980, 2. 419

472 45 46 47 48 49 50 51

52 53 54 55

56 57 58 59 60 61

62 63 64 65 66 67 68 69 70

71 72 73 74 75 76 77 78 79 80 81

82 83 84 85 86 87 88

89 90

Ylaranta T, Ann Agric Fenn 1982; 21: 103 Cay EE, Gissel-Nielsen G, Soil SOCAmer Proc 1973;37: 590 Kang Y, Nozato N, Kyuma K, Yamada H, Soil Sci Plant Nutr 1991; 37: 477 Abrams MM, Burau RG, Comm Soil Sci Plant Anal 1989;20:221 Asher CJ, Butler GW, Peterson PJ, J Exp Botany 1977;28:279 Gissel-Nielsen G, Hamdy AA, Z Pflanzenern Bdk 1977;140: 193 Singh M, Singh N, Relan PS, Soil Sci 1981; 132: 134 Neal RH, Sposito G, Holtzclaw KM, Traina SJ, Soil Sci SOC Am J 1987; 5 1: 1161 Singh M, Singh N, Relan PS, Soil Sci 1981; 132: 134 Doran JW,Alexander M, Soil Sci SOCAm J 1977; 4 1 : 70 Zieve R, Peterson PJ, Sci Tot Environm I98 1 ; 19: 277 Oreshin V Yu, Vartanyan SS, Sov Geol 1981; 10: 41 Donaldson EM, Talanta 1988; 35: 633 Kingston HM, Paulsen PJ, Lambert G, Appl Spectr 1984; 38: 385 Han Heng-Bin, Kaiser G, Tolg G, Anal Chim Acta 1981;128: 9 Heinrichs H, Keltsch H, Anal Chem 1982; 54: 121 I Brunfelt AO, Steinnes E, Geochim Cosmoch Acta 1967;31:283 Troshin Yu P, et al, Geokhim 1988; 300: 1218 Kiesl W, Seitner H, Kluger F, Hecht F, Monatshefte fur Chemie 1967; 98: 972 Pelly IZ, Lipschutz ME. In: B Mason ed. Handbook of elemental abundances in meteorites. Gordon and Breach Science Publ., New York-Paris-London I971 Kiesl W, Grass F, Bock1 R, Ponta U, J Radioanal Chem 1970; 6: 447 Kiesl W. In: Brunfelt AO, Steinnes E eds. Activation analysis in geochemistry and cosmochemistry, Universitetsforlaget Oslo-Bergen- Tromso 1971 Weinke HH. Ricerche Spettroscopiche 1977; 3: 53 1 Weinke HH, Kiesl W, Kluger F, Koberl C, Proc 10th Symp Antarct Meteor 1985;198 Kiesl W, Weinke HH, Mikrochiin Acta 1970; 392 Chau YK, Riley JP, Anal Chim Acta 1965; 33: 36 Koberl C, Kiesl W, Weinke HH, J Non Cryst Sol 1984;67: 637 Koberl C, Ann Rev Earth Ptanet Sci 1986; 14: 323 Van Der Sloot HA, Hoede D, Klinkers TJL, Das HA, J Radioanal Chem 1982; 71: 463 Agterdenbos J, Van Elteren JT, Bax D, Ter Heege Jp. Spectrochim Acta 1986;41B:303 Augustin-Gyurits K, Schroll E, Mitt Ges Geol Bergbaustud Osterr. 1992; 38. I95 Rowe JJ, Steinnes E, Talanta 1977; 24: 433 Kaakinen JW,Jorden RM, Lawasani MH, West RE, Envir Sci Techno1 1975; 9:862 Koppenaal DW, Lett RG, Brown FR, Manahan SE, Anal Chem 1980; 52: 44 Zmijewska W, Semkow T, Chemia Analiticzna 1978; 23: 583 Forbes S, Bounds GP, West TS, Talanta 1978; 26: 473 Amy MP, Commun Soil Sci Plant Anal 1992; 23: 1397 Arthur MA, Rubin G et al, Environm Tox Chem 1992;11:541 Duckart EC, Waldron LJ, Doner HE, Soil Sci 1992;53:94 Patrick R, Am Scientist 1978; 66: 185 Goede AA, Nygard T, De Briun M, Steinnes E, Sci Tot Environm 1989; 78: 205 Turkstra J, Harthoorn AM, Beukes PJL, Brits RJN, J Radioanal Chem 1977; 37: 473 Soltanpour PN, Olsen SR, Goos RJ, Soil Sci Am J 1982: 46: 430 Lombeck I, Kasperek K, Harbisch HD, Feinendegen LE, Bremer HJ, Europ J Pediat 1977; 125: 81 Hojo Y, Sci Tot Environm 1987; 65: 85 Cappon CJ, Smith JC, Arch Environm Contam Toxicol 1981; 10:305

473 91 92 93 94 95 96 97

Cappon C, Smith JC, J Anal Toxicol 1982; 6: 10 Kurkela P, Kaantee E, Zentralbl Veterinarmed 1979; 26B. 169 Saeed K, J Anal At Spectrom 1987; 2: 151 Agostoni A, Gerli GC et al, Clin Chim Acta 1983; 133.153 Behne D, Wolters W, J Clin Chem Clin Biochem 1979; 17- 133 Lin SM, Chiang CH, Proc. XXIV CSI, Garmisch-Part. 1985 Van Cauwenbergh R, Robberecht H, Deelstra H, J Trace EIem Electrolytes Health Dis 1990; 4:215 98 Robinson MF, Godfrey PJ et al, Am J Clin Nutr 1979, 32: 1477 99 Sheliow JS, Jackson SG, Baker JP, Cantor A, J Animal Sci 1985; 6 I : 590 100 Alfthan G, Clin Chem 1985; 31: 500 101 T61g G. In: Bratter P, Schramel P eds. Trace element - analytical chemistry in medicine and biology, Vol 3, Walter de Gruyter Berlin-New York 1984 102 Retief DH, Scanes S, Cleaton-Jones PE, Turkstra J, Smit HJ, Archs oral Biol 1974; 19:s 17

This Page Intentionally Left Blank

415

AUTHOR INDEX Atteia 0.

79

Massiani C.

187

Balvay G.

303

Melikhov I.V.

65

Barcel6 J.

147

Mesz&ros I.

23

Berthier F.

303

De Miguel E.

Brandon D. L. Bruno J.

217

231,293

M6dy I.

23

35

Nebot J.

35

1

Nogales R.

173

CaCador 1.

355

ONeil L. J.

293

Catarino F.

355

Panagos A. G.

119

Coppee V.

303

Cruiias R. Cullinane M. J.

217

Paniaux A. Peduzzi R.

323

29 3

Poschenrieder Ch.

147

Demarta A.

323

Prudent P.

187

D'ltri F. M.

391

Remade J.

337

Burgenmeier J.

Dominik J.

303

Elstner E. F.

13

Felip6 M.T. Forstner U.

Robertson J. M.

79

293

Sager M.

403,459

217

Simmers J. W.

231,293

259

Sipka V.

Gallardo-Lara F.

173

Skogerboe J. G.

65 23 1

Garau M.A.

217

Soto J.

173

Guns6 B.

147

Span D.

303

Hambuckers A.

337

Tack F. M. G.

103

Hambuckers-Berhin F.

337

Thomas 0.

187

Hew& D.

173

Tonolla M.

323

13

Vale C.

355

Jouanneau J. M.

373

Varnavas S. P.

119

Kritsotakis K. G.

119

VBzquez M. D.

147

Lakatos G.

365

V6dy J.C.

Lappaquellerie Y.

373

Verloo M. G.

103

373

Vernet J.-P.

303

Vukovic? 2.

65

Hippeli S.

Latouche C. Lee C. R. Marschall M.

231,293 23

Marti E.

217

Martin C. Masscheleyn P. H.

303

103

Wilhelm G . S.

79

23 1

This Page Intentionally Left Blank

477

SUBJECT INDEX Page numbers are the first page ofthe paper in which the entry is discussed.

Acid deposition 35 rain 398 producing potential 261 Acidic mine 261 Acidifyngagent 173 Acidity 261 Adsorbability ofselenate 410 Aeromonas 324 Air pollution 13,23 Aquifer 79 Atmosphere 65 Atmospheric cycling 392 Atomic absorption 187 Bacterial colonization 323 community 337 Contamination 328 counts 326 Beneficial effects 150 Bioaccumulation 398 Bioassays 293 clam 295 earthworm 231,295 mussel 231 plant 231,295 sandworm 231 snail 231 Biogeochemical cycling 395 Biological activity 268 testing 293 Biomethylation 275 Buffering capacity 259 Bureaucracies 3 Cadmium 187,373 budget 384 enriched sewage sludge 173 molecular diffusion 38 1 salt 173 Capping deposits 283 Central planning 7 Changing impulsively 65 Charring 404 Checkpoint 65 Chemocline 323 Chemical time bomb 275 Chlorophyll 154 Chloroplasts 156 Chromatiurn okenii 326 Chromium 147 Civilsewice 3

Cluster analysis 340 Coastal environment 120 Collectivedecision-makingmechanisms 6 Combustion offossil fuels 37 Competition 8 Complexation 275 Compost 151,187 Compounds 214 Concentration ofradionuclides 65 Concept of capacity controllingproperties 273 final storage quality 273 Contaminant mobility 23 1 Contaminated sediment 23 1 Copper 187 Cross section 65 Daily impulses 76 Deforestation 40 Degradation 283 Demethylation 396 Democracy 6 Democratic process 2 Desorption 275,348 Direct control 1 Discretionary space 8 Dominatingwind direction 65 Dredged materials 260 Dredging 261 Earthsurface 65 Ecological economics 10 Economic rationality 2 Ecotoxicologicaltest 218 Efficiency 2 Eh 265,355 Ejection 65 Element mobilization 269 specific detector 408 Elutriate test 277 Emissions 148 Environmental effects 105 Epiphytic biotecton 365 Equity 2 Ethics 1 Eutrophication 303 Final storage quality 274 Fish 398 Fluxes 338 Fokker-Planckequation 65 Foodchain 398 Forest decline 23 Forms ofphosphorus 305

478 Freshwater ecosystem 337 Gel chromatography 187 Geochemicalengineering 259 Gironde estuary 373 Global transport 392 Ground layer 65 Grounwater 271 Guidelines 147 Heavymetal 105,147,173,187,231 content 365 Herb layers 24 Heterotrophic bacteria 334 High level nuclear waste repositories 35 Human behaviour 3 Immobilisation 283,344 Immobilize 274 Incineration 274 Industrial wastes 217 Informationsystems 10 Institutions 3 Interferingpeaks 408 Internalizationofexternal effects 7 Interstitial water 273 Intestinalbacteria 324 Iron 163 Kinetic 277,339 Labelled metabolites 404 tissues 403 Lake Annecy 304 Balaton 365 eutrophic 328 Fen6 (Neusiedlersee) 367 Geneva 304 Lugano 304 meromictic 323 restoration 304 Velencei 366 Landfills 271 Leachates 271 Leaching behaviour 103 Lead 187 Leafnitrate reductase activity 24 Leather 148 Littoral zone 365 Long-lived radionuclides 76 Market 6 failures 1 incentives 1 mechanism 1 Mercury 391 Metal 355 contamination 293 mobility 265 pollution 119 slags 103 transfer 280

Metallicpollution 373 Meteorologicalconditions 65 Methylation 396 Methylmercury 391 Microbial activity of soil 26 Migration 103 MinamataDisease 391 Mineralweathering rates 36 Mining activities 119 Mobilityconcept 273 Modelling 111,279 Monthly impulses 76 Moss layers 24 Motor car exhaust 13 Multielementcapabilities 410 Municipal solid wastes 187 Neoclassical school 2 Nitrite interferes 422 Nitropiazselenol 404 Nitrous acid 404 Normative aspect ofeconomics 2 NO, 13.418 Occupationalmedicine 149 OECD 1 Optimalgrowth 4 Optimum allocation 1 Organic matter 80,355 metal binding 206 metal complexes 187 micro-pollutants 24 substrates 281 Oxidation 264 Oyster 373 Ozone 13 Pathogenic bacteria 323 Periphyton 365 reed complex 365 Permeability 275 pH 260,347,355 Phaseolus vulgaris L. 154 Phosphorus 304 Photosynthetic pigment composition 25 sulphur bacteria 325 Phytotoxic effects 149 Policy failures 1 Porewater 270 Positive aspect ofeconomics 1 Pressure groups 9 Productive resources 1 Publicgoods 8 Radionuclide transfer 337 Rain 80 water 394 Random ergodic 65 Readsorption 281 Reagent blanks 403

479

Redistributionof income 4 wealth 4 Redox 260 Reed belt 365 communities 365 decline 365 Regulations 9 Release 338 Remedial investigation 293 Remobilisation oftrace metals 268 Roots 355 Salmonella 324 Saprophyte counts 326 Scarcity 1 Se7Stracer 403 Seasonal changes 76 Sediment 304,338,355,373 components 279 profile 269 Sequential extraction 162,278 Sessile oak 23 Short-livedradionuclides 76 Site remediation 293 Snow 80 SO, 13 Socio-politicalapproaches 3 Soil 79 acidification 80 erosion rates 36 extractability 173 phosphatase activity 24 respirastion 2 18 solution 80 Solid waste 260 Solidarity 4 Solidification 274 Soot particles 13 Sorbent for selenium 41 5 Sorption 111 Speciation 187 studies 422 Spectral interferences 424 Stabilization 274 State 10 intervention 9 Stationariness 76 Stationary distribution ofimpulses 75 Subsediment-deposit 283 Sulphide oxidation 263 related metals 135 Sulphidic environment 284 Switzerland 79 Tagussalt marshes 355 Tannery sludge 152 Temperature 347

Toxic effects 147 metals 119 Trace elements 79 metals 79 U.V. 187 Upland disposal 23 1 Uptake 344 Value judgments 1 Volatile hydrides 420 Volatilization losses 404 of selenium 406 Wastes characterization 220 Water 373 column 338 density 329 Weathering 80 Wetland creation 23 1 Wheat 173 Zen mays L. 154 Zinc 373

This Page Intentionally Left Blank