Copyright is owned by the Author of the thesis. Permission is given for a copy to be downloaded by an individual for the purpose of research and private study only. The thesis may not be reproduced elsewhere without the permission of the Author. Practical Aspects of Phytoextraction A thesis presented in partial fulfilment of the requirements for the degree of Doctor of Philosophy in Earth Science Christopher William Noel Anderson 2000 Massey University Palmerston North New Zealand OGS 10 MASSEY UNIVERSITY APPLICATION FOR APPROVAL OF REQUEST TO EMBARGO A THESIS (Pursuant to AC 98/168 (Revised 2), Approved by Academic Board 16.02.99) Name of Candidate: Christopher W. N. Anderson ID Number: 93000196 Degree: PhD Dept/Insti tute/School: Institute of Natural Resources Thesis Title: Practical Aspects of Phytoextraction Name of Chief Supervisor: Prof. R. R. Brooks Telephone Extn: As author of the above named thesis, I request that my thesis be embargoed from public access until (date) 1 sI of April 2002 for the following reasons: � Thesis contains commercially sensitive information. o Thesis contains information which is personal or private and/or which was given on the basis that it not be disclosed. o Immediate disclosure of thesis contents would not allow the author a reasonable opportunity to publish all or part of the thesis. o Other (specify): Please explain here why you think this request is justified: This thesis contains unpublished information that describes in detail the steps necessary to effect the phvtoextraction of gold using induced hyperaccumulation. Ongoing research is investigating the economic potential of this technology. The information contained within this thesis will remain commerciallv sensitive until this period of research has been completed. Signed (Candidate): Endorsed (Chief Supervisor): Approved/N:et Appt t:l ,c{>,d (Representative of VC): Date: Date: Date: I � ! � /00 I I 12/2. / uO I I Note: Copies of this form, once approved by the representative of the Vice-Chancellor, must be bound into every copy of the thesis. Practical Aspects of Phytoextraction Abstract Phytoextraction for heavy metals is an emergmg technology that has potential application for the remediation of many contaminated sites around the world. The technology has similar application to the mining of low-grade ore bodies. Several practical aspects of the technology are addressed in this thesis. Natural and induced-uptake phytoextraction trials have been conducted on two contaminated substrates: an area of industrial pollution in northern France, where base metals are present in an oxide and carbonate mineral phase, and an area of mine tailings in New Zealand, where base metals are present in a sulphide or sulphate mineral phase. The uptake response of several hyperaccumulator and non-accumulator plant species is described. Geochemical models are then presented that explain the observed metal uptake as a function of the predominant chemical form of metal present in the soil. Natural uptake is dependent upon the form of metal. It appears that the relative efficacy of various hyperaccumulator species to accumulate metals is also dependent upon site­ specific geochemistry. The efficacy of chelating agents, in particular EDTA, to induce uptake is similarly dependent upon the chemical form of metals in the soil. A field trial for cadmium phytoextraction was conducted on an area of pastoral land contaminated with this metal due to the application of cadmium to soil through superphosphate fertilisation. Natural uptake at this site by the hyperaccumulator species Thlaspi caerulescens could remove the equivalent of 1 7 years of annual cadmium application in one harvest. The chelating agent EDT A (ethylenediaminetetraacetic acid) did not induce significant uptake by the non-accumulator Brassica species. Instead, the action of this chemical was to redistribute 14% of the cadmium initially present in the 0-5 cm soil depth to the 5-10 cm depth, and to leach approximately 4% of the cadmium initially present at the site to below 10 cm in the soil profile, as shown by mass balance calculations. Phytoextraction effected by Tcaerulescens i s proposed as a management tool for cadmium in the pastoral environment. Practical Aspects of Phytoextraction Phytoextraction for nickel has been investigated at a field site in the central North Island of New Zealand. Hyperaccumulation was effected by two A lyssum species and by Berkheya coddii. However, the biomass of the harvested plant material was below that reported in the literature. The conclusion from this trial is that substrate modification of ultramafic soil may be necessary before phytoextraction for nickel could be implemented. A significant obstacle hindering the practical application of phytoextraction in some environments, is the paucity of hyperaccumulator species that are native to some parts of the world. Western Australia has many sites that may benefit from phytoextraction for nickel. However, only one hyperaccumulator species is native to this region, Hybanthus floribundus, a species that has in the past been difficult to genninate from seed. This thesis describes a successful approach to gennination, involving the use of one-year-old seeds, treated with 'Regen 2000 smoke water' and genninated under dark conditions, that may overcome this practical aspect (a limitation) of phytoextraction technology. The most recent advance of induced phytoextraction technology has been the thioligand-induced uptake of gold by plants . The initial discovery and the geochemical rationale behind the induced uptake of gold is described. The maximum gold uptake presented is accumulation of 57 mg/kg dry weight gold by Brassica juncea and it is proposed that this level of uptake could make the phytomining of gold from tailings areas an economic proposition. The conclusion of this thesis is that potential for the implementation of phytoextraction is large. Globally, the technology could offer an environmentally and economically friendly alternative to the traditional decontamination of metals from some sites. There is also potential for the phytomining of metals from low-grade ores . The social implications of phytoextraction technology in third-world countries could also be large. Phytoextraction for gold, for example, from auriferous tailings in Africa and South America, has the potential to improve both the environment and the standard of living of the local communities who live off contaminated land. 11 Practical Aspects of Phytoextraction Acknowledgements I would like to express my sincere thanks to my supervisors. Firstly to Professor Robert Brooks, under whose guidance I feel privileged to have had the opportunity to study metal uptake by plants. Also to Bob Stewart and to Robyn Simcock for keeping me honest and on track to finish within time, and for pulling in the reins when I would slip out of control! I would like to express my thanks to the Agricultural and Marketing Research and Development Trust (AGMARDT) for funding this project through the award of a doctoral scholarship. I thank Professor Daniel Petit from the University of Lille in northern France who made possible my time researching metal uptake in Europe. I also express my deepest thanks to Annabelle Deram and Valerie Bert from Lille whose friendship made my time there so enjoyable. I acknowledge the support of the Western Mining Company Ltd. of Australia, in particular Colin Woolard, Tara Read, Bonny Nicholson and Frances Mills for making possible the Australian part of the project and for showing me the 'wild' ways of Western Australia. I am greatly indebted to other agencies for their financial support of this project. To the New Zealand Vice Chancellors' Committee for the award of a Claude McCarthy Fellowship which enabled me to attend the ICOBTE conference in Vienna and the PACRIM conference in Bali during 1999. To the graduate research fund for making travel and field trials possible. To the Massey University Ag-hort faculty for the award of a Helen E. Akers scholarship in 1998. I thank most sincerely Mr. Don Adams from Blairlogie, Wairarapa whose assistance made possible the Wairarapa cadmium field trial. I also thank Mr. Peter Strongman from Rorison Mineral Developments Ltd. of Piopio whose assistance made possible the Piopio nickel field trial. I say a big thank you to all the staff and students, past and present, who have supported me through my years in the Soil Science Department. Five years ago I swore I was never going to complete Honours let alone a PhD! The value of the advice, encouragement and assistance that goes with being part of a good department cannot be measured. I would like to thank all my friends outside of Massey, who have put up with me through good times and bad over the course of my studies. Finally, I would like to acknowledge and express my deepest thanks to my family for their encouragement and support as I became interested in science from an early age. ll1 Practical Aspects of Phytoextraction Table of Contents Abstract ....................................................................................................................................................... i Acknowledgements ................................................................................................................................... iii Table of Contents ...................................................................................................................................... iv List of Figures .......................................................................................................................................... viii List of Tables ............................................................................................................................................. xi Chapter 1 - Introduction and Overview of the Current Study .............................................................. 1 Chapter 2 - Phytoextraction: a Genera) Introduction ............................................................................ 6 2.1 Contamination vs pollution; sources of heavy metal in soil ..................................................... 6 2.2 Phytoextraction: phytoremediation and phytomining ............................................................... 8 2.3 Natural hyperaccumulation ....................................................................................................... 9 Hyperaccumulation of cadmium ...................................................................................................... 11 Hyperaccumulation of copper and cobalt. ........................................................................................ 12 Hyperaccumulation of lead .............................................................................................................. 13 Hyperaccumulation of manganese ................................................................................................... 13 Hyperaccumulation of nickel ........................................................................................................... 13 Hyperaccumulation of selenium ....................................................................................................... 14 Hyperaccumulation of thalium ......................................................................................................... 14 Hyperaccumu1ation of zinc .............................................................................................................. 14 2.4 Induced hyperaccumulation .................................................................................................... 15 2.5 The distribution of hyperaccumulators and reason for the phenomenon ................................ 17 2.6 The mechanisms for hyperaccumulation ................................................................................ 20 2.7 Bioavailability of metals in the soil ........................................................................................ 24 SECTION A - PHYTOEXTRACTION OF CADMIUM, LEAD AND ZINC: OBSERVED AND MODELLED UPTAKE ........................................................ 26 Chapter 3 - Trials on Contaminated Substrates ................................................................................... 30 3.1 Introduction ............................................................................................................................. 30 3.2 Field trials in northern France ................................................................................................. 31 3.3 Trials on tailings from the Tui base-metal mine ..................................................................... 34 Experimental Design ........................................................................................................................ 36 3.4 Hyperaccumulation trials using Berkheya coddii ............................................................... .... 37 3.5 Hyperaccumu[ation trials using Cardaminopsis halleri ......................................................... 37 3.6 Hyperaccumulation trials using other species ......................................................................... 39 3.7 Problems with metal uptake .................................................................................................... 39 Terminology for Chapters 4, 5 and 6 ............................................................................................... 41 Chapter 4 - Geochemical Model for Lead Uptake ................................................................................ 42 4.1 Introduction ............................................................................................................................. 42 4.2 Experimental Design ............................................................................................................... 42 4.3 Results: Brassicajuncea . ........................................................................................................ 43 4.4 Results: Thlaspi caerulescens ..................... . .................................................................... ....... 45 4.5 Results: total soil lead ............................................................................................................. 47 4.6 Results: plant-available lead ................................................................................................... 47 4.7 Discussion - a model for lead uptake ...................................................................................... 48 4.8 Conclusion .............................................................................................................................. 49 IV Practical Aspects of Phytoextraction Chapter 5 - Geochemical Model for Cadmium Uptake ........................................................................ 50 5.1 Introduction ........... .... ..... . .................. ........... ......... . ... ......... ..... ..................... ........................... 50 5.2 Experimental design ..... ............................... ................................ ...................... ..................... 50 5.3 Results: Brassicajuncea . . .... . ....................... ........................................................ ................... 51 5.4 Results: Cardaminopsis halleri ........... ................................... .... ........... .... ............................ . 54 5.5 Results: Thlaspi caerulescens ........ . ........................... ...... ........... . ............... ........................ .. .. 55 5.6 Results: total soil cadmium .......................... .............. ..... ........ ..... . .............. ......... .................. . 56 Effect of metal concentration on uptake .................................................................................... ....... 56 5.7 Results: plant-available cadmium ... ............ . ........................... ..... ....... .................................. .. 57 5.8 Discussion - a model for cadmium uptake ... .................. ......................................................... 58 Relationship between ammonium acetate and the metal phase ........ '" ............................................. 60 5.9 Conclusion ............. ....................... ..................... ................................................. ........... ......... 63 Chapter 6 - Geochemical Model for Zinc Uptake ................................................................................. 64 6.1 Introduction .... ......... . ........ ... ..................................................... ........ ................... .................. . . 64 6 .2 Experimental design ................................................ .... , ..................................................... , .... 64 6.3 Results: Cardaminopsis halleri ................... . ........................................................ .... .............. 65 6.4 Results: Thlaspi caerulescens ...................... ........................................................................... 68 6.5 Results: total soil zinc .............................. .... ........................................................................... 69 6.6 Results: plant-available zinc ............................. ...................................................................... 69 6.7 Discussion - a model for zinc uptake ........................................ ....... ..................... .................. 70 Relationship between ammonium acetate and the metal phase .............. ....................................... '" 71 6.8 Conclusion .......... ......................................... ..................................................... ...................... 73 Chapter 7 -An Integrated Geochemical Model for Cadmium, Lead and Zinc Uptake .................... 75 7.1 Practical aspects of phytoextraction for cadmium, lead and zinc ........... ................................ 75 7.2 Application of the geochemical model to results from pot and field trials ............. ... .. ........... 79 Auby - northern France ..................................................... ....... . .......... ......... ................... . .... ............ 79 Tui Mine tailings - New Zealand ., .......................................... ......................................................... 81 7.3 Conclusion .......... ................ ............. ................................................................ '" .................... 82 Chapter 8 - Phytoremediation: a possible management solution for New Zealand pastoral soils ... 85 8.1 Introduction and literature review of the issue of cadmium in soils ....................................... 85 Sources of cadmium contamination .................................................................................. ..... .......... 86 Phosphatic fertilisers ................................................ ...... '" .. , ............... ...................................... ....... 86 Cadmium in agricultural land " ..... . . .......................... ................................................................ ........ 86 8.2 Materials and methods ............................................... ................. ............................ .......... ...... 87 Experimental Design ................................................. ................ ................ .................. ................. . ... 87 Analysis of plant samples .................................................................................. . .............................. 91 Analysis of soil samples .................. .......................... ................................................................. .... .. 91 8.3 Results ..................................................................................................................................... 93 8.4 Discussion ................................................ ......... .... .................................................................. 94 Plant metal uptake ................... .......... . , ................................ ....................... '" ................................... 94 Effect of soil cadmium concentration on metal uptake ..................................... ............. .. ................ 97 Cadmium redistribution within the soil ........................................................................ .................... 97 Soil cadmium mass balance calculations ................................ ......... ................................................. 99 Zinc accumulation by the hyperaccumulator species ................ ....................................................... 99 8.5 Summary - a practical application for phytoextraction ........................................................ 101 8.6 Conclusion ....................................................................... .... ............................................. .... 102 CONCLUSION TO SECTION A ..................................................................... 105 v Practical Aspects of Phytoextraction SECTION B - PHYTOEXTRACTION FOR NICKEL AND GOLD .................. 1 08 Chapter 9 - A New Zealand Field Trial for Nickel Phytoextraction ................................................. 110 9.1 Introduction .................................. ............................... , ................. . . ........... ........................... 111 The initial study of Nicks and Chambers ..................................................................... .................. III Subsequent studies ......................................................................................................................... 112 The Ultramafic Belt of New Zealand ............................................................................................. 113 Geology of the Piopio serpentine exposure .................................................................................... 113 9.2 Materials and Methods .......................................................................................................... 114 Design of the experimental area ........... ........................ ................................... ............................... 114 Analytical procedure ...................................................................................................................... 116 9.3 Results . .................................................................................................... .............................. 117 Alyssum malacitanum .................................................................................................................... 117 Alyssum bertolonii ......................................................................................................................... 1 18 Berkheya coddii .............................................................................................................................. 119 Native species ................................................................................................................................. 122 Soil samples .................................................................................................................................... 123 9.4 Discussion ........................................................................ ................... .................................. 123 Reasons for the poor species performance ..................................................................................... 123 9.5 Conclusion ............................................................................................................................ 125 Chapter 10 - Hybanthusjloribundus, a Native Australian Nickel Hyperaccumulator .................... 127 1 0.1 Introduction ............. ............ .................................................................................................. 127 Hybanthus floribundus .............................. ..................................................................................... 127 Importance of 'fIre' to promote seed germination .......................................................................... 130 10.3 Methods and Materials .................................................................. ................... ..................... 131 10.4 Results and Discussion ......................................................................................................... 132 10.5 Conclusion ........................................... ................................................................................. 134 Chapter 11 - Phytoextraction for Gold ................................................................................................ 135 11.1 Introduction and review of the solution geochemistry of gold ............................................. 135 Solution geochemistry of gold ................................................................... .................................... 135 Economic mineralisation within a weathering profile ....................... ............................................. 136 Geochemical mobility of gold ........................................................................................................ 137 The biogeochemical pathway of gold ............................................................................................. 141 11.2 Analytical methodology ...................... ............ ...................................................................... 142 Analysis of substrate samples ......................................................................................................... 142 Analysis of plant samples ............................................................................................................... 144 Contamination ................................................................................................................................ 145 11.3 Induced uptake of gold ...... ................................................................................................... 146 Initial discovery ......................................................................................... '" .................................. 146 Subsequent work ............................................................................................................................ 147 Artificial gold experiment .............................. ................................................................................ 149 11.4 Model for induced uptake of gold ............................................................... .......................... 151 Thiocyanate-induced solubility ........ .............................................................................................. 151 Thiosulphate-induced solubility .......................................... ........................................................... 156 Application ofthe model to induced-uptake results from Waihi gold ore ..................................... 159 11.5 Conclusion: the choice of thiocyanate or thiosulphate ......................................................... 161 Chapter 12 - Practical Scenarios for Nickel and Gold Phytoextraction ........................................... 162 12.1 Introduction ............. .................................................. ......... ....................... ............................ 162 12.2 Nickel .................................................................................................................................... 162 Phytorernediation scenario ............................................................................................................. 162 Phytomining scenario .................................................................................... ................................. 165 Environmental concerns ............................................................................................. ................. '" 166 Sustainability of a nickel phytomining operation ........................................................................... 167 12.3 Gold ...................................................................................................................................... 168 Phytomining scenario ....... ....................................................................... ....................................... 168 Environmental concerns ................................................................................................................. 170 Vl Practical Aspects of Phytoextraction Practical Scenario 1 - gold tailings ................................................................................................. 172 Practical scenario 2 - artisanal mining ............................................................................................ 173 Summary - growing a crop of gold ................................................................................................ 175 SECTION C: CONCLUSiON .......................................................................... 1 77 Cbapter 13 - Practical Aspects of Pbytoextraction: a General Conclnsion ...................................... 177 13.1 Conclusions from this research ............................................................................................. 177 13.2 Future research ...................................................................................................................... 180 13.3 Concluding remarks .............................................................................................................. 180 References ............................................................................................................................................... 182 Appendices .............................................................................................................................................. 197 V11 Practical Aspects of Phytoextraction List of Figures Chapter 2 Figure 2.1. Sources of metal in soils and sediment. .......... . . . . . . ......................... ..................... . . . ... 7 Figure 2.2. Diagrammatic representation of the phytoextraction operation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 0 Figure 2.3. Thlaspi caerulescens, a hyperaccumulator of cadmium and zinc g rowing on base- metal mine tai l ings, Massey University, New Zealand . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 2 Figure 2.4. Map showing the locations of the majority of hyperaccumulators and associated metal l iferous soils . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 9 Figure 2.5. Theoretical uptake response of plants to heavy metal in soils. After Baker (1 981 ). 2 1 Figure 2.6. Theoretical uptake response of plants to heavy metals in soi l . After McGrath (2000) ............. . ............... .... . . ...................................... . . . ........... . .................. . . . ............ ...... ...... . . . 22 Figure 2.1. Modified theoretical uptake response of plants to metals in the soil . . . . . . . . . . . . . . . . . . . . . . . . 22 Chapter 3 Figure 3.1. Map of Central Europe. The red shaded region of France is the Nord Pas de Calais . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 Figure 3.2. Field-trial site at Auby, northern France . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 Figure 3.3. Map of the North Island of New Zealand showing in detai l the location of the Tui mine tailings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 Figure 3.4. The Tui mine tailings on the flanks of Mt. Te Aroha, looking northeast across the rich farmland of the Bay of Plenty . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 Figure 3 . 5 . EDTA-Induced uptake of cadmium and lead by Berkheya coddii growing on Tui mine tai l ings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 Figure 3.6. Thiocyanate-induced uptake (SCN) of Cd, Pb and Zn by Brassica juncea from Tui mine tai l ings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 Chapter 4 Figure 4.1 Natural uptake, and acetic acid-, citric acid- and EDTA-induced uptake of lead by Brassica juncea (Bi) and Thlaspi caerulescens (Tc) growing on artificial 1 % lead soils of d ifferent metal phases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 Figure 4.2. Summary: efficacy of EDTA-induced lead uptake by Brassica juncea as a function of the metal phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 Chapter 5 Figure 5.1. Natural uptake and EDTA- and citric acid-induced uptake of cadmium by Brassica juncea (Bi), Cardaminopsis ha/leri (Ch) and Thlaspi caerulescens (Tc) g rowing on artificial 200 mg/kg cadmium soils, of d ifferent metal phases . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52 Figure 5.2. Summary: efficacy of EDTA-induced cadmium uptake by Brassica juncea as a function of metal phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 53 Figure 5.3 Summary: efficacy of natural cadmium uptake by Cardaminopsis halleri as a function of metal phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 Figure 5.4. Summary: efficacy of natural cadmium uptake by Thlaspi caerulescens as a function of metal phase . . . . . . . . . . . . . . . . . . . . . . . . . . ... . . . . . . . . . . . . . . . . . ... . . ..... . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56 Figure 5.5. Brassica juncea growing on an artificial , carbonate phase, cadmium-contaminated soil shortly before harvesting . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 VllI Practical Aspects of Phytoextraction Figure 5.6. Plot of the cadmium concentration in Brassica juncea as a function of the plant- available cadmium concentration in the soil for each metal phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 Figure 5.7 . Plot of the cadmium concentration in Cardaminopsis halleri as a function of the plant-available cadmium concentration in the soil for each metal phase . . . . . . . . . . . . . . . . . . . . . . . 6 1 Figure 5.8. Plot of the cadmium concentration in Thlaspi caerulescens as a function of the plant-available cadmium concentration in the soil for each metal phase . . . . . . . . . . . . . . . . . . . . . . . 62 Capter 6 Figure 6.1. Natural uptake and EDTA-induced uptake of zinc by Cardaminopsis halleri (Ch) and Thlaspi caerulescens (Tc) growing on artificial 0.2% Zn soils of different metal phases . . .... . . . ..... ........ . . ... .. ........... . ... . . . . .......... ..... . ... . . . . . . ...... . . . . . . .. . . .... ......... . ... .... . . .......... .. ......... . . .. 66 Figure 6.2. Summary: efficacy of natural zinc uptake by Cardaminopsis halleri as a function of metal phase . . . , . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 Figure 6.3. Summary: efficacy of natural zinc uptake by Thlaspi caerulescens as a function of metal phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . 68 Figure 6.4. Plot of the zinc concentration of Thlasp; caerulescens as a function of the plant- available zinc concentration in the soil for each metal phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 1 Figure 6.5. Plot of the zinc concentration of Thlaspi caerulescens as a function of the plant- available zinc concentration in the soil for each metal phase . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 Chapter 8 Figure 8.1 . Map of the North Island of New Zealand , showing in detail the location of the cadmium phytoextraction field-trial site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... . . . . . . . . . . . 88 Figure 8.2. Schematic plan of the experimental-trial set up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ... . . . . . . . . . . . 89 Figure 8.3. View of two Brassica species in flower, growing at the Wairarapa experimental trial site shortly before treatment with EDT A . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 Figure 8.4. View of Thlapsi caerulescens (T) and Cardaminopsis halleri (C) growing at the Wairarapa trial site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 Figure 8.5. Comparison of two methods to determine the cadmium concentration of soi l . . . . . . . . 92 Figure 8.6. Surface plot (0-5 cm) showing the change in soil cadmium concentration effected by the trial . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 95 Figure 8.7. Surface plot (5- 1 0 cm) showing the change in soil cadmium concentration effected by the trial. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 96 Figure 8.8. Plot of the cadmium concentration in Thalaspi caerulescens as a function of the cadmium concentration in the soil. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 Figure 8 .9. Zinc accumulation by T.caerulescens and C.halleri from the Wairarapa trial site. 1 00 Figure 8.10. Phytoremediation of cadmium from New Zealand soils: a poster summary of a practical appl ication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 03 Chapter 9 Figure 9.1. Map of the North Island of New Zealand showing in detai l the location of the nickel phytoextraction tria l site . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1 0 Figure 9.2. North facing perspective of the Piopio serpentine quarry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 14 Figure 9.3. Schematic plan of the Piopio experimental tria l set up . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1 5 Figure 9.4. Photograph of the Piopio trial site before harvest of B.coddii i n May 1 999, looking towards the northeastern corner of the site . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1 6 Figure 9.5. Flowering specimen of Alyssum berlolonii growing at the Piopio trial site . . .. . . . . . . . . 1 1 8 Figure 9.6 . N ickel concentration in Alyssum berlolonii as a function of the soi l fertil iser treatments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 1 9 Figure 9.7. The author with a flowering stem of Berkheya coddii, growing at the Piopio site . . 1 20 IX Practical Aspects of Phytoextraction Figure 9.S. N ickel concentration in Berkheya coddii as a function of the soil fertil iser treatments .............. ......... ......... . . .. . .. ...... . . . ...... ......... ........ . . . . . ......................................... 12 1 Figure 9.9. N ickel concentration in Berkheya coddii as a function o f the plant nitrogen concentration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 2 1 Figure 9 . 1 0. Plot of the cobalt concentration i n Berkheya coddii as a function of the nickel concentration . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 22 Chapter 10 Figure 1 0.1 . Hybanthus floribundus, F. Muel l . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 28 Figure 1 0.2. Germination experiment for Hybanthus floribundus showing a final germination rate of 56% for replicate 1 and 47% for replicate 2 . .. . .... . . . . . ... . ......... ............ . . .. . . . . . . .. . . . . 1 33 Chapter 11 Figure 1 1 .1 . Thiocyanate-induced uptake of gold by Brassica juncea ( Indian mustard) from finely disseminated and native 5 mg/kg synthetic gold ores . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 50 Figure 1 1 .2. Eh-pH diagram of the Au-SCN-H20 system at 25°C and 1 bar pressure . . . . . . . . . . . . 1 54 Figure 1 1 .3. Effect of increasing pH , through the addition of incremental amounts of l ime, on the thiocyanate-induced solubilty of gold . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 55 Figure 1 1 .4. Thiosulphate-induced solubil ity of gold . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 57 Figure 1 1 .5. Eh-pH diagram of the Au-S-H20 system at 25°C and 1 bar pressure . . . . . . . . . . . . . . . . . 1 58 Chapter 12 Figure 1 2. 1 . N ickel yields (kg/ha) of successive crops of a theoretical hyperaccumulator plant growing over various u ltramafic soils . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 67 Figure 1 2.2. The possible economic value of a phytomined crop of gold as a function of the concentration in a plant with a biomass of 20 tlha . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 69 Figure 1 2.3. Mine tailngs at the now closed Paris gold mine, on the mining lease of the Western Mining Company Ltd . , near Kambalda, Western Austral ia . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 72 Figure 1 2.4. The author, Robert Brooks, and a local security guard surveying gold-mine tailings in the suburb of Germinston, Johannesburg, South Africa . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 74 Figure 1 1 .5. Growing a Crop of Gold: summary poster describing the steps involved with the phytoextraction for gold . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 75 x Practica l Aspects of Phytoextraction List of Tables Chapter 2 Table 2.1 . The relative abundance of a selection of biological ly significant trace elements . ....... 6 Table 2.2. Normal elemental concentration, hyperaccumulation criterion concentration and number of known representative hyperaccumulator species . . . ........................................ 1 1 Section A Table A.1 . Composition of lead aerosols originating from different from pollutant sources . ...... 26 Table A.2 . Relative abundance of the chemical forms of airbourne lead collected from an urban site . .......................... . . ........... ....... ......... ............ ........ . .. . ... .......................... . . . . .................... 27 Chapter 3 Table 3.1. Selected geochemical properties of the Auby soil used in this study . ....................... 31 Table 3.2. Metal concentration (DW) in field plants 21 days after EDTA treatment. . ........ . . . ...... 33 Table 3.3. Selected geochem ical properties of the Tui tailings material used in this study . ...... 36 Table 3.4. Natural and induced accumulation of cadmium, lead, and zinc by Cardaminopsis halleri growing on Tui mine tailings . .............. .. ..... .................................................... ......... 38 Table 3.5. The dominant chemical form of metal present at 3 contaminated sites . .. ................. 41 Chapter 4 Table 4.1. Total soil lead and pH for the control treatment soi ls of each metal phase ... ............ 47 Table 4.2. Plant-available lead for the control treatment soils of each metal phase .......... ........ 47 Chapter 5 Table 5.1. Total soil cadmium and pH for the control treatment soils of each metal phase ....... 56 Table 5.2. Percentage total cadmium that is plant-available from each metal phase ................ 57 Chapter 6 Table 6.1. Total soil zinc and pH for the control treatment soils of each metal phase . .............. 69 Table 6.2. Percentage zinc that is plant-avai lable from each metal phase ................................ 70 Chapter 7 Table 7.1. Summary of the mean plant-available metal concentrations (ammonium acetate) extracted from each metal phase . ..................................................................................... 79 Xl Practical Aspects of Phytoextraction Chapter 8 Table 8.1 . Summary of the metal concentrations found in the harvested plant material. .......... 93 Table 8.2. Biomass and extrapolated uptake figures over a one-hectare area for the p lant species used in this tria l . . . ...... .. . . . . .. ... ... . ............ ...... .... . . . ...... ..... ....................... . . . . ...... ... . . . . 97 Table 8.3. Average soil-cadmium concentrations across plot area .. .. . ......... . .............. . ........... .. . 98 Table 8.4. Average total-soil cadmium levels for plot area . .......... . . ............. . . . .. . . .... .................... 98 Chapter 9 Table 9.1 . Piopio wet-serp geochemical properties ................................. . ........................ . . . ..... 1 1 7 Table 9.2. Mean Alyssum bertolonii nickel concentration for each ferti l iser treatment. ........... 1 1 8 Table 9.3. Mean 8erkheya coddii nickel concentration for each fertiliser treatment. . ......... ..... 1 20 Chapter 10 Table 1 0.1 . Mean nickel concentrations (mg/kg DW) of species and subspecies of Hybanthus together with their locations in Western Austral ia ............................. .. ........... ................. 1 30 Table 1 0.2. End of experiment results of germination test 7 on seeds of Hybanthus floribundus ... . .. ........... . . . . . .................... .. .. . .. . . .... .................. ............. . . . .................... ........................... 1 32 Chapter 11 Table 1 1 . 1 . Gold concentrations (mg/kg DW) in Impatiens balsamina and I.holstii immersed for 48 hours in gold solutions of different anionic composition . ...... . . . ..................... . ............ 1 42 Table 1 1 .2. Digestion methodology test for replicate subsamples of an adopted herbage standard .. . . ......................................... .......................... . .. .. .............................................. 145 Table 1 1 .3. Induced uptake of gold by 8rassica juncea grown on Waihi gold ore . .................. 146 Table 1 1 .4. Selected summary of thiocyanate (SCN')-induced uptake of gold for various plant/substrate combinations . .......... .. . . . ....................... . ................... ............................... 1 48 Table 1 1 .5. Comparison of Tui tai l ings samples collected from three locations within the TSF i l lustrating the broad range of SCN-extractable gold values that can be observed ........ 1 52 Table 1 1 .6. Comparison of the total and extractable gold concentrations for Tui JWM with selected substrates on which plant trials have been conducted .... .. ............................... 1 52 Table 1 1 .7. pH and micronutrient concentrations of bulk and rhizosphere soi l of white l upin (Lupinus a/bus) grown in a phosphorus deficient soi l . ... ................................................. 1 58 Chapter 12 Table 1 2.1 . Number of crops of 8erkheya coddii required to reduce the nickel contamination in soi l to the European Union guideline of 75 mg/kg . .............................. ................ . . . ........ 1 65 Table 1 2.2. Comparative toxicity of 4 sodium salts that can solubi l ise gold and may be used for induced hyperaccumulation . .. . ................. .. ... ............ . . .. . ................. . . . .............. . . .......... . . . 17 1 XlI Chapter 1 Chapter 1 - I ntroduction and Overview of the Current Study Botanical exploration has identified many plant species that naturally accumulate very high concentrations of elements, including the group commonly referred to as 'heavy metals' . The enormous potential application for the use of some of these species to extract metals from soil (phytoextraction) is being slowly realised. Phytoextraction is a relatively new scientific term, and came about due to pioneering research conducted during the late 1 970s that was led by scientists from New Zealand (e.g. Brooks et al. 1 974, 1 977a, 1 977b, 1 979). The 1 990s saw an explosion in the interest surrounding phytoextraction due to growing concerns over 'heavy metal' contamination of soils around the world. A large volume of scientific literature has appeared during this timeframe, but the technology has not yet advanced to the point where large-scale implementation of working operations has ensued. This statement may be somewhat generalistic as examples of small-scale operations exist, one such example is discussed in Chapter 12, but is generally true. Implementation has been slower than some would have expected. The aim of the present study was to investigate the potential for phytoextraction in several different environments, and to examine various practical aspects of its implementation as they became apparent; practical aspects that may hinder or promote development of the technology. Research was conducted in parallel for two different groupings of metal and is described in two separate sections. Section A describes the phytoextraction of the metals cadmium, lead and zinc. Section B describes the phytoextraction of the metals nickel and gold. Chapter 1 serves as an overview of the study, and over the next several pages, the reasoning behind the various practical aspects of phytoextraction studied and described by this thesis are presented. 1 Chapter 1 Chapter 2 - Phytoextraction: a general introduction Before any practical aspects of phytoextraction can be described, the discussion must be prefaced with a review of the relevant literature that defines the current state of phytoextraction technology. Chapter 2 reviews the development of phytoextraction: the reason why studies into phytoextraction have been persued, the discovery of the plants that could effect the necessary concentrations of metal uptake, the reason why these plants accumulate metal, and the practical application of these plants to a contaminated environment. Section A - Phytoextraction of cadmium, lead and zinc: observed and modelled uptake Chapter 3 - Trials on contaminated substrates Two sites contaminated with the metals Cd, Pb and Zn were investigated for this study, two sites that differ in the chemical form of metal that contaminates each environment. Phytoextraction trials conducted on substrates from each site are described in Chapter 3 . Phytoextraction i s not proposed as a viable method of remediation at either o f the described locations, due to the high metal loadings. However, each does al 10w for the testing of uptake mechanisms and techniques. The metal response of several hyperaccumulator and non-accumulator species to both natural and induced uptake is described. Chapters 4, 5 and 6 - Geochemical models for lead, cadmium and zinc uptake The results obtained from the trials described in Chapter 3 were surprising when compared to the evidence for Cd, Pb and Zn accumulation found in the literature. To further examine the possible reasons for discrepancies that existed between observed and literature reported uptake, experiments were designed to test the effect of different chemical forms of metal present in a soil on both metal bioavailability and subsequent plant uptake. The mixing of Pb, Cd and Zn mineral salts with a commercial seed-raising 2 Chapter 1 mIX generated artificially contaminated soils. Chapters 4, 5 and 6 present, in turn, geochemical models for Cd, Pb and Zn uptake that explain the uptake of each metal as a function of the chemical form of that metal in a contaminated soil. Chapter 7 - An integrated geochemical model for Cd, Ph and Zn uptake In chapter 7, the results from the previous three experiments are integrated to generate a geochemical model that explains bioavailability and plant uptake of each of these metals as a function of the chemical form, and thus the source of metal contamination, in the soi l . This model is then used to explain the metal uptake patterns observed for experiments conducted on the natural substrates of Chapter 3 . Chapter 8 - Phytoremediation: a possible management solution for New Zealand pastoral soils Chapter 8 describes a phytoextraction field trial conducted on an area of agricultural land contaminated with cadmium. The report of this trial is prefaced with a review of the literature describing the problem of cadmium accumulation in New Zealand pastoral soils . McGrath ( 1 998) has previously suggested that pastoral land in Australia, lightly contaminated with cadmium or zinc, could be remediated using hyperaccumulator species. The aim of the trial described in Chapter 8 was to test this theory on a New Zealand site. The results show that natural hyperaccumulation by Thlaspi caerulescens effected phytoremediation. Mass balance calculations have been used to show that EDT A application to pastoral land may not induce increased metal uptake, but leach metals down the soil profile. The results for cadmium uptake described for Chapter 8 are subsequently related to the geochemical model of Chapter 7 in the conclusion to Section A. 3 Chapter 1 Section B - Phytoextraction for nickel and gold Chapter 9 - A New Zealand field trial for nickel phytoextraction Chapter 9 describes a field trial where the phytoextraction potential of several known nickel hyperaccumulators was tested on an area of nickeliferous soil . The aim of this trial was to ascertain the efficacy of nickel phytoextract ion, for both phytoremediation and phytomining, in an environment foreign to the hyperaccumulator species used. No ultramafic flora was endemic to this site. Nickel phytoextraction has been proposed for areas of ultramafic soil based upon data generated from controlled pot experiments. However, little work has translated this potential to the field . Chapter 10 - Hybanthlls floribllndus. a native Australian nickel hyperacclfm7l1ator At the outset of this research programme, one of the aims of the thesis was to test the phytoextraction potential of several nickel hyperaccumulator species on mine tailings in Western Australia. This part of the project was limited by the requirement that only native Australian species could be used, narrowing the choice of hyperaccumulators to one, Hybanthus floribundus. However, this species has in the past proved difficult to germinate from seed. The requirement that only native species may be used hampers the implementation of phytoextraction technology in some environments. Chapter 1 0 addresses this limitation and describes the approaches that I used to overcome the difficulties of seed germination by H.floribundus. Chapter 1 1 - Phytoextraction for gold Phytoextraction for gold i s the most recent advance in the technology of induced phytoextraction. Chapter 1 1 is prefaced with a review of the literature relevant to the geochemical mobility of gold, before a detailed description of the induced uptake of this metal is presented. My discovery that plants could be induced to accumulate gold was serendipitous : cadmium, lead and zinc induced-uptake experiments on auriferous rock generated unexpectedly high uptake results for this precious metal . The geochemical focus of this aspect of phytoextraction technology is described in Chapter I I . 4 Chapter 1 Chapter 12 -Practical scenarios for nickel and gold phytoextraction Chapter 1 2 discusses several scenarios where the practical implementation of phytoextraction for nickel and gold may prove viable. Of all the metals for which phytoextraction is possible, nickel and gold hold the greatest promise for phytomining due to a combination of value, and in the case of nickel, known hyperaccumulator species of high biomass . However, sites where phytoextraction of nickel may prove viable are less well documented in the literature than those for Cd, Pb and Zn. The potential for gold phytoextraction has never previously been described. Section C: Conclusion Chapter 13 - Practical aspects of phytoextraction: a general conclusion Conclusions from the previous 1 2 chapters are integrated in Chapter 1 3 , and used to review the practical aspects of phytoextraction described by this thesis . The range of phytoextraction applications described by the 1 3 chapters of this thesis are broad, but reflect the diverse range of environments in which phytoextraction may prove viable. A note on concentration, content and uptake Debate surrounds the correct use of the three words concentration, content and uptake. The definition of concentration is simple and in this thesis is expressed as the mass of metal (mg) per unit dry weight (kg) of plant. Content is the total mass of metal (mg) in the plant and is a function of both the metal concentration and the mass (biomass) of the plant. Uptake, however, does not specifically refer to either concentration or content and may be represented by either word. In this thesis I have chosen to express heavy metal uptake by plants using the concentration of metal accumulated by the plant, and hence in this thesis uptake refers directly to metal concentration rather than metal content. 5 Chapter 2 Cha pte r 2 - P hytoextracti on : a General I ntro d u cti o n 2.1 Contamination vs pol lution; sources of heavy metal in soil 'Heavy metal ' is a loose and ill-defined term used to describe a group of metals that are generally associated with pollution and toxicity (Alloway, 1 990). The list includes all the alkali metals, alkaline earth metals and aluminium, as well as the 'metal loids' arsemc, antimony and selenium (Striet and Stumm, 1 993 ) . Heavy metals exist naturally, i n all parts of the environment, but the concentration of metal observed varies dramatically. Soils are the weathering products of rocks, and thus the metal loading of a soil will be a function of the parent bedrock. Table 1 . 1 illustrates this by comparing several heavy metals and the range of these metals that can be found 'naturally' in soil . Soils at the low end of the range are formed from parent bedrock devoid of metal, while anomalous values at the high end of the range owe their origin to metalliferous bedrock. For example, a high nickel concentration can be due to soil development from serpentine rock (see Chapter 1 0), while a high cadmium concentration can be due to soil development from marine black shales (peterson and A1loway, 1 979) or from sulphide ores (Nriagu, 1 978a). Table 2.1 . The relative abundance of a selection of bio log ical ly sign ificant trace elements. Element Estimated crustal abundance (mg/kg) Global mean soil concentration (mg/kg) Range in non­ polluted soils mg/kg) EU l imit for soils· Arsenic 1 .8 Cadmium 0.2 Cobaij 25 Copper 55 Lead 1 2 .7 Nickel 75 Zinc 70 After Peterson and Alloway (1 979). 6 0 .06 8 20 1 0 40 50 0. 1 -40 0.01 -30 1 -40 2-300 2-200 1 0-1 000 1 0-300 3 1 40 300 75 300 • Maximum concentration al lowed in agricu ltura l soi ls receiving sewage sludge. From CEC (1 986), Official Journal of the European Community No L 1 81 , 6- 12 as cited by McGrath et al. (2000). 6 Chapter 2 Anthropogenic sources of contamination have dramatically exacerbated the heavy metal loading of soils over recent history. The sources for this metal loading are varied (Fig. 2 . 1 ) . The temporal increase of metal in the environment is particularly well highlighted by analysing chronological layers of ice from the polar ice caps. The concentration of lead in layers of ice sampled from Greenland and dated at 800 BC was 0.0005 J..Lglkg (Ppb). This rose to 0.2 J..Lglkg for ice dated from 1 965 (Murozumi et al. , 1 969). Sources of metal in soils and sediment I I I 1 I Natural J Anthropogenic I I I I Rock - Stationary Mobile Chemical weathering sources sources applications rl Fire I I Mining Automobiles Waste - e.g. tailings - I-- disposal Airborne dumps - dust U Industry Trains, Fertilisers � Volcanic i I e.g. a ir-fall """- planes etc - I I contamination , I Pest '-- control Figure 2.1 . Sources of metal in soils and sediment. After Nriagu (1 978a). The presence of heavy metals in soil does not necessarily represent a problem to the flora and fauna supported by that land. The two terms pollution and contamination are in this sense often confused. Where metal is present in soil at a concentration above average background (Table 2. 1 ) then that land can be considered as contaminated. Such contamination may be either natural or anthropogenic in origin, although it must be mentioned that no fixed 'threshold criterion' for contamination can be quantified due to the variabi lity of metal concentrations in soils around the world. Where the metal concentration negatively affects the health of flora and fauna living off this land, then the land can be considered as polluted. The two terms should be used carefully and 7 Chapter 2 should not be interchanged. Polluted land is contaminated land, however contamination does not necessarily constitute pollution. Thousands of sites exist worldwide that can be considered polluted, and as such, means are actively sought for their effective decontamination. A 1 993 report tabled by the United States Environmental Protection Agency estimated that more than 300 billion dollars would be needed to remediate 1 235 continental USA ' Superfund' sites contaminated by human activities (USEPA, 1 993). Traditional remediation techniques involve, for the most part, excavation and either leaching or disposal of the bulk material. These traditional techniques are costly and energy intensive, and generally simply relocate the mode and occurrence of the metal problem to a more contained site. Many new and emerging plant-based techniques for the removal and management of heavy metals are beginning to appear. The phytoextraction (plant extraction) technologies of phytoremediation and phytomining have captured public and scientific interest, as they potentially offer a 'green' and environmentally friendly alternative to the traditional decontamination and management of some contaminated sites. 2.2 Phytoextraction : phytoremediation and phytomining During the 1 980s, a practical application for plants that accumulate very high levels of heavy metal was recognised, and phytoremediation became defined as the in situ remediation of an area of land contaminated with heavy metals using hyperaccumulator plants (Chaney, 1983; Baker and Brooks, 1 989). During each growing cycle, these plants accumulate metals in their aerial parts . Harvesting and burning the foliage concentrates the polluting metal into a small volume of ash, called a 'bio-ore ' , that can be safely disposed of as landfill at a contained site (Kumar et al. , 1 995) . Phytomining is a more recent advance on this phytoremediation technology. Here the target metal is of sufficient economic value to warrant recovery of the metal from the plant ash. Phytoextraction is a term that describes both phytoremediation and phytomining - the use of plants to remove metals from either an area of contaminated land or a low-grade ore body. 8 Chapter 2 The concept of phytoextraction IS thus relatively simply to define (Fig. 2 .2) and involves 3 steps: 1 . growth of plants that can accumulate metals (natural or induced hyperaccumulation) from an area of metalliferous land, 2 . harvesting and burning of the metal-rich biomass, and 3 . smelting (phytomining) or disposal (phytoremediation) of the ash. There is a significant financial incentive for the successful use of phytoextraction. Cunningham and Berti (2000) modelled a reduction of the soil-lead concentration at a US contaminated site from 0. 1 4% to 0 .04% at a cost of $279 thousandlha for phytoextraction but almost $2.5 millionlha for conventional technology. Clearly, if these figures are true, phytoextraction may represent a cost-effective option relative to traditional means of soil decontamination. Phytoremediation is well established as a commercial enterprise, offered by various small phytotechnology companies for a limited range of metals, and is regarded by many as a potential growth market in the field of biotechnology. Phytomining, however, has yet to be proven on a large scale. 2.3 Natural hyperaccumulation The ability of some plant species to accumulate inordinately high levels of one or more heavy metals was first reported by Baumann ( 1 885) for zinc accumulation by Thlaspi calaminare growing near Aachen, Germany. The first quantitative record for any other metal was made in Italy by Minguzzi and Vergnano ( 1 948), with regard to nickel accumulation by the small perennial shrub Alyssum bertolonii. Minguzzi and Vergnano reported a nickel content of 0.79% (7 900 mg/kg) in the dry leaves of plants growing in soil containing only 0.42% of this metal. It was not until the late 1 970s, however, that the 'ability' of certain plants to actively accumulate metals was formally named. Brooks et al. ( 1 977a) used the term hyperaccumulation to describe this unusual plant character. At the time, the threshold concentration was set at 1 000 mg/kg (0. 1 %) dry weight (DW) for most metals, with the 9 Chapter 2 The phytoextraction operation Hyperaccumu lator P lant material burnt B io-ore smelted (phytomin ing) i. hem ica ls dded Bio-ore landfi l led (phytoremed iation ) Figure 2.2. Diagrammatic representation of the phytoextraction operation. 1 0 Chapter 2 exception of zinc, for which the threshold was set at 1 0 000 mglkg ( l %). The modem and more accepted threshold criterion is accumulation 1 00 times greater than in non­ accumulator plants growing in the same environment (Table 2 .2). To date there are more than 400 known hyperaccumulator species, of which three-quarters are hyperaccumulators of nickel. The first plant species to be given formal hyperaccumulation status was Alyssum bertolonii. It is important to note that the list of species reported in Table 2 .2 is in no way restrictive and many more species and metals will undoubtedly be added in the future. Rather, the list reflects the current focus of research into the trait of hyperaccumulation. Table 2.2. Normal elemental concentration, hyperaccumulation criterion concentration and number of known representative hyperaccumulator species. Metal Normal range (mg/kg OW) Criterion (mg/kg OW) No. of species Cadmium 0.03 - 20 1 00 2 Cobalt 0.05 - 50 1 000 26 Copper 1 - 1 00 1 000 24 Manganese 5 - 2000 1 0 000 8 Nickel 0 .2 - 1 00 1 000 3 15 Selenium 0.01 - 1 0 1 000 1 9 Thallium 0 - 0 . 1 500 2 Zinc 5 - 2000 1 0 000 1 8 Source: Baker et al. (2000), Brooks (1 998), Leblanc et al. ( 1 999) and Reeves et al. (1 995). The following paragraphs discuss in more detail the reported evidence and importance of hyperaccumulation for the metals presented in Table 2 .2 . Hyperaccumulation oJ cadmium A review of the literature yields only one species that can hyperaccumulate cadmium, the herbaceous biennial species Thlaspi caerulescens (Fig 2 .3), which can accumulate up to 1 000 mglkg cadmium DW (Baker et al., 2000). In this thesis the herbaceous metallophyte Cardaminopsis halleri is also reported as a hyperaccumulator of cadmium, based upon field observations in northern France (Chapter 3) . Near the city of Lille, Cardaminopsis halleri is part of a metalliferous floral association indicative of land contaminated by industrial activity. Hyperaccumulation for C.halleri i s contrary to the observations of McGrath ( 1 998), but is agreement with a paper currently in press (Dahmani-Muller et al., 2000) . 1 1 Chapter 2 Figure 2.3. rh/asp; caeru/escens, a hyperaccumulator of cadmium and zinc growing on base-metal mine tail ings, Massey University, New Zealand. Hyperaccumulatiol1 of copper and cobalt S o m e degree of un certainty su rro unds the hyperaccu mulation of copper and cobal t . Brooks a n d Malaisse ( 1 98 5 ) surveyed t h e Sh aba P rovence i n Zaire (now t h e Democratic Republic of Congo) and reported 2 4 hyperaccumulator species of copper and 26 hyperaccumulator species o f cobalt ( 9 speci es accumulated both metal s) . However, i n subsequ ent studies t h ese species raised from seed have a l l fai l ed to hyperaccumulate either metal (e.g. B ennett, 1 9 9 8 ) , and very little evi dence fo r copper and cobalt hyperaccu mulation h a s been observed outside t h e Democrat i c Republic of Congo . Reports of high copper accu m u l ati o n by Millota myosotidijo!ia (2 400 m glkg - B l i ssett, b 1 996) and by Minuarta vema ( 1 070 mg/kg - Ernst, 1 974) are exceptional . Confi rmation of the African hyperaccu mulation would be desirable but d ifficult due to the politi cal climate in central Africa. 1 2 Chapter 2 Hyperaccumulation of lead Debate surrounds the existence of natural hyperaccumulators of lead. Many authors report species that can accumulate large levels of this metal (e.g. Reeves and Brooks, 1 983), however, all such reports are based upon field-collected samples where the chance of contamination by an adjacent metal smelter or other airborne sources of lead was high. Replication of this field hyperaccumulation for lead has not been reported under controlled laboratory or greenhouse conditions. For this reason, reported hyperaccumulators of lead are not included in Table 2 .2 . Hyperaccumulation of manganese Eight hyperaccumulators of manganese are reported in Table 2 .2 , a number which is probably in no way indicative of the number of species in nature that hyperaccumulate this metal. Little attention has focussed on manganese, as little environmental or economic emphasis attracts the exploitation of manganese hyperaccumulators for phytoextraction. However, my own correspondence with mining companies in South East Asia has identified waste rock and tailings areas where high concentrations of manganese inhibit the revegetation of these substrates with the native, non-tolerant species, perhaps suggesting a reason for the further study of this metal. Hyperaccumulation of nickel Considerably more species of hyperaccumulator are known for nickel than for any other metal. This is perhaps due to two reasons. Firstly, the ultramafic soils on which nickel hyperaccumulators are found have a very specific 'ultramafic ' flora, and secondly, because of a very effective field test for nickel hyperaccumulation that involves the chemical dimethylglyoxime (DMG). Nickel forms a diagnostic red colour with DMG at concentrations above I 000 mglkg DW in the plant (the hyperaccumulation criterion), a property that allows rapid field screening of large numbers of plant species. Ultramafic soils are characterised as being very rich in Cr, Co, Fe, Mg and Ni, and deficient in the essential nutrients Ca, Mo, N, P and K (Brooks, 1 987). Ultramafic floras have thus adapted and evolved to withstand harsh environmental conditions. Brooks et al. ( 1 995) 1 3 Chapter 2 reported 1 88 hyperaccumulators of nickel, although with the recent discovery of additional species in Cuba (Reeves et aI., 1 996) the total is now 3 1 5 (Table 2.2). Hyperaccumulation of selenium Nineteen hyperaccumulators of selenium have been reviewed by Brooks ( 1 998). With the exception of two species native to Queensland, Australia, all selenium hyperaccumulator species are endemic to seleniferous soils of the USA. An exciting new advancement in the phytoextraction for this metal is the subsequent volatilisation of selenium metal to dimethyselenide vapour by some hyperaccumulator species. Loss of the metal burden to the atmosphere would overcome the problem of disposal of a metal-rich biomass. Hyperaccumulation of thalium Discovery of plant specIes that can accumulate inordinately high concentrations of thallium is a relatively recent event (Leblanc et aI., 1 997). Hyperacumulation of thalium has been defined by Leblanc et al. ( 1 999) as accumulation greater than 500 mg/kg DW. The authors of the 1 999 study reported a maximum metal content of 0 .28% metal in the dry leaves of Iberis intermedia growing on mine tailings in the south of France. Thallium is a rare metal in the environment, but is very toxic. Concentrations of 20 mg/kg DW in vegetables and 40 mg/kg for Brassica napus have been reported for crops growing in France (Tremel, 1996; Tremel and Mench, 1 997; Tremel et al. 1 997). Concern is mounting over the presence of such high concentrations of thallium in the food chain. Thallium is also a relatively valuable metal (ca. US$300/kg) and this combination of value and toxicity make it attractive for phytoextraction studies. LaCoste et al. (1 999) discussed the phytoremediation potential of Iberis species, while Anderson et al. (2000) have discussed the potential for the phytornining of this metal. Hyperaccumulation of zinc The metal that has attracted the second largest focus of attention for phytoextraction studies (after nickel) is zinc, a phytotoxic metal that will inhibit crop yields before harmful effects are manifest upon the food chain (Shen et ai. , 1 997; Ebbs and Kochian, 1 4 Chapter 2 1 998; McGrath, 1 998; Zhao et a/. , 1 998). Sixteen species are known that hyperaccumulate zinc, and two, in particular, have been studied in detail : Thlaspi caerulescens and Carciaminopsis halleri, species that also hyperaccumulate cadmium. Zinc hyperaccumulator species are in all cases species of low biomass, a factor which limits their effectiveness for phytoextraction. However, this is compensated for, to some degree, by the high concentrations of zinc accumulated (up to 3 . 5% for Tcaerulescens). McGrath et a/. ( 1 993) and McGrath ( 1 998) discussed the phytoextraction potential of the two hyperaccumulator species mentioned here for the remediation of zinc and cadmium concentrations elevated in European pastoral soils as a result of the application of sewage sludge to land. These authors concluded that phytoextraction of zinc (and cadmium) could be entirely feasible using these two species, but only for relatively low levels of soil contamination (e.g. 500 mg/kg zinc and 5 mg/kg cadmium - the European limits for these metals in soil are 3 00 and 3 mg/kg respectively). 2.4 Induced hyperaccumulation Uptake of metals that are not 'naturally' accumulated by any recognised plant species can still be effected using phytoextraction technology. If a chemical amendment added to soil 'targets ' certain insoluble metals, these metals can be ' induced' into the soil solution. Plants often accumulate this soluble metal complex passively through transpiration - induced hyperaccumulation, although the exact mechanisms for induced-metal uptake are poorly understood and a matter of some debate. The first successful report of induced-hyperaccumulation technology was for lead (Huang and Cunningham, 1 996). Zea mays (maize) growing on contaminated soil was induced to accumulate over 1 % lead in its dry-weight foliage through amendment of the soil with a protonated form of ethylenediaminetetraacetic acid (EDT A). Subsequent studies showed that increased uptake of the metals Cd, Cu, Ni and Zn as well as lead by Brassica juncea could be induced using EDTA (Blaylock et al. , 1 997; Huang et al. , 1 997). Conflicting reports, however, show that for some plant species, metal combinations, EDT A may actually effect a decrease in metal uptake by the plant. Robinson et a/. ( 1 997b, 1 999a) showed that EDT A reduced the uptake of cobalt and nickel by the hyperaccumulator species Berkheya coddii. 1 5 Chapter 2 EDT A is a well-known chelating ligand and has been used in agriculture since the 1 960s as a commercial soil amendment to improve micronutrient availability to plants (Li and Shuman, 1 996). In particular Mn-EDTA and Na2Zn-EDTA are used as effective fertilisers (Kabata-Pendias and Pendias, 1984) . There are , however, inherent problems surrounding the widespread use of EDT A to increase the uptake of less-desirable metals. The chemical is biodegraded in soil, but relatively slowly (Means et aI. , 1980). Persistence in the soil, in combination with its efficacy for complexing a wide range of metals, may lead to the leaching of soluble metal complexes into an underlying water table and the subsequent contamination of groundwater. Indeed the principle of leaching for soil remediation was reported by Kobayashi et al. ( 1 974 - as cited by Kabata­ Pendias and Pendias, 1 984) who decreased the cadmium concentration of a surface soil from 28 mg/kg to 14 mg/kg by successive treatment of the soil with EDTA. Tejowulun and Hendershot ( 1 998) reported a similar but more contemporary example. This aspect, or 'environmental concern' regarding the secondary effects of EDTA in soil, hampers the practical application of EDTA-induced phytoextraction. The only report in the literature of a field trial to test EDTA-induced phytoextraction for lead was conducted by the company 'Phytotech' (now known as Edenspace) in the USA. Blaylock (2000) reports effective soil decontamination of lead-contaminated soil, using a combination of EDTA and Brassica juncea, based upon the change in soil lead concentrations observed in soil cores. However, the author does not support his conclusion with figures for the metal concentration within the plants, and more importantly, does not calculate soil mass balances to account for all of the 'remediated' metal. Blaylock's report does not counter the criticism that EDTA may have leached lead to below the soil sampling depth, thus allowing for the appearance of soil decontamination. Another aspect of induced phytoextraction that is poorly addressed in the literature is the economic cost involved - the ch elates used are expensive. In a recent exercise, Rufus Chaney (pers. commn to R.R.Brooks, 2000) calculated that an application rate of 1 0 rnmol EDTA per kg soil (3 .52 glkg) would require 7 040 kg of chelate per hectare. At a unit cost of $4.30 a kg (quoted by Dow Chemicals) the cost of EDTA-induced phytoextraction of lead would be US$30 272 a hectare. An application rate of 3 .52 g/kg 1 6 Chapter 2 has been reported by the Phytotech group as the rate necessary to effectively induce lead uptake. EDT A is not the only chemical that has been used to effect induced-metal uptake. Huang et al. ( 1 997) reported the use of EGTA (ethylenebis[ oxyethylenetrinitrilo] tetraacetic acid), DTP A (diethylenetrinotrilopentaacetic acid) and HEDTA (N-(2- hydroxyethyl)ethylenetriacetic acid - a protonated form of EDTA) as well EDT A to effect lead uptake, but EDTA was the most effective. Blaylock et al. ( 1 997) included CDTA (trans- l ,2-cyclohexylenedinitrilotetraacetic acid) and citric acid in their studies, while Cooper et al. ( 1 999) report the use ofNTA (nitrilotriacetic acid). Inherent problems notwithstanding, induced hyperaccumulation remains a potentially powerful tool. Natural hyperaccumulator species are generally specific for only one or two metals and in many cases are species of low biomass. Hyperaccumulators of the 2 metals zinc and cadmium, for example, have a biomass of approximately 2 tlha (Thlaspi caerulescens and Cardaminopsis halleri) . If a high biomass species (e.g. maize with 30 tlha) could be induced to hyperaccumulate the same concentration of these two metals from a contaminated site, then metal uptake would be increased by a factor of 1 5 . Induced hyperaccumulation also has the potential to effect the phytoextraction of metals for which no natural hyperaccumulators have been recognised. The most recent example of induced hyperaccumulation was that reported for gold (Anderson et al., 1 998; Chapter 1 1 ), but work is also focus sing on induced uptake for the platinum group metals and mercury. 2.5 The distribution of hyperaccumulators and reason for the phenomenon Some metals are essential for living processes and deficiency can lead to poor growth or even death. However, excessive metal accumulation by living tissues is always toxic (Ensley et aI. , 1 997). In contrast, hyperaccumulator species appear to have the unique ability to tolerate a higher concentration of metal than is necessary to perform metabolic functions. 1 7 Chapter 2 All known hyperaccumulator species occur naturally on metalliferous soils. Robinson ( 1997) summarised 4 types of soil environment where hyperaccumulation has been observed: 1 . serpentine - distributed throughout the world, rich in Cr, Mn and Ni, 2. base-metal (calamine) - distributed throughout the world, rich in Cd, Pb, Zn and TI, 3 . copper-cobalt - found only in the Democratic Republic o f Congo, rich in Co and Cu, and 4. seleniferous soils - found only in the USA and Australia and rich in Se. The presence of a metalliferous soil does not necessarily predicate the presence of hyperaccumulator species. Hyperaccumulators have only been discovered on metalliferous soils present in tropical and temperate regions (Fig. 2.4). This worldwide species distribution indicates that evolution has played an important role in the development of the trait, and that this evolution has occurred over a relatively long timescale. Quaternary glaciations have cleared all record of hyperaccumulator species that may have existed in environments outside contemporary tropical and temperate regions. Hyperaccumulation has subsequently not re-evolved in these recently ( 10 000 years before present) glaciated environments (Brooks, 1 987). Why p lants have evolved the hyperaccumulation trait is poorly understood. Boyd and Martins ( 1992) summarised a number of possible explanations for hyperaccumulation: 1 . tolerance or disposal of metal from the plant, 2 . drought resistance, 3 . interference with neighbouring plants, 4. inadvertent uptake, and 5 . defence against herbivores and/or soil pathogens. Of these hypotheses, p erhaps the most notable is the interference hypothesis. Baker and Brooks ( 1 989) suggested that the hyperaccumulation of nickel might be a ' survival or defence strategy against competition from other plant species' . Boyd ( 1998) further 1 8 Chapter 2 described this hypothesis as ' elemental aIIeIopathy' . Metal accumulated by a species, sequested in leaves and then returned to the ground through leaf abscission, will poison the surrounding soil to non-tolerant species. Boyd and Jaffre ( 1999) reported a higher nickel concentration in the leaf litter and soil beneath the canopy of the New Caledonian nickel hyperaccumulator tree Sebertia accuminata than for non-accumulator species, and proposed that this was elemental allelopathy in action. Support for the defence hypothesis was reported by Pollard and Baker ( 1 997), who found that the high foliar­ zinc concentration of Thlaspi caerulescens acted as a natural deterrent to herbivores . No advancement beyond the theory stage has yet been reported for the possible explanations of drought resistance and inadvertant uptake. Figure 2.4. Map showing the locations of the majority of hyperaccumulators and associated metalliferous soils. Shaded areas indicate the extent of world-wide glaciation at the end of the last Ice Age. After Brooks (1 987). The link between metal tolerance and accumulation in hyperaccumulator species IS unclear. Baker et al. ( 1 994) and Chaney et al. ( 1 997) have suggested that the two are inextricably linked and that hyperaccumulation necessarily requires tolerance. However Macnair et al. ( 1 999) showed genetic segregation of tolerance and hyperaccumulation for zinc, exhibited by the F2 generation of Cardaminopsis halleri crossed with 1 9 Chapter 2 c.petraea. Macnair (pers. cornmn. 1 999) has gone further to suggest that the criterion for hyperaccumulation should be re-assessed in light of this new evidence for the independence of tolerance and hyperaccumulation. Macnair suggests that hyperaccumulation should be better defined than simply accumulation 1 00 times greater than that for non-accumulator plants growing in the same environment. What is clear is that the reason for hyperaccumulation remains speCUlative. Further research needs to be conducted in this controversial area. 2.6 The mechanisms for hyperaccumulation For uptake to occur, metals need to be solubilised in the rhizosphere and then moved across the root cell plasma membrane for subsequent transport into the xylem (Robinson, 1 997). Species that naturally hyperaccumulate metals presumably excrete metal-binding compounds into the rhizosphere to effect this process of accumulation (Raskin et al. , 1 994). These compounds have been termed phytochelatins and metallothioneins (Grill et al. , 1 987). Hyperaccumulation is an active uptake process, as opposed to the paSSive uptake mechanism true for accumulator (indicator) species. Hence, while the concentration of metal in an accumulator species is a linear function of metal in the soil, the concentration of metal in a hyperaccumulator species is not. Species that exude compounds into the rhizosphere that act to immobilise metals through precipitation are excluders. Induced hyperaccumulation can be considered a passive uptake process - phytochelatins and metallothioneins are artificially created in the rhizosphere through the addition of soil amendments such as EDT A. Baker ( 1 98 1 ) presented a series of theoretical curves to explain behaviour of the three metal-response mechanisms described above (Figure 2.5) . The hyperaccumulator curve shows physiological control of metal uptake; the plant has the ability to regulate and block metal uptake above a certain concentration, while the excluder curve shows the opposite physiological response, blocking of metal uptake below a certain concentration, but no control over metal uptake at higher concentration. 20 +"' c: co a. (J.) ..c: ..... Hyperaccumulator I ndicator Excluder Metal in the soil Figure 2.5. Theoretical uptake response of plants to heavy metal in soils. After Baker ( 1981) . Chapter 2 Increased knowledge of the mechanisms and reasons behind hyperaccumulation has shown these curves to be unrealistic . McGrath et al. (2000) presented a set of three new curves, although as early as 1 997 McGrath had disputed the curves of Baker (Fig. 2 . 6) . Baker's curves give the impression that indicator and excluder species will grow in the same soils as hyperaccumulator species, and similarly, that hyperaccumulator species will grow in soils with low metal concentrations populated by indicator and excluder species. McGrath' s curves clearly show the shift of thinking with regard to hyperaccumulation and tolerance, as they show the 'death' of indicator and excluder species at a much lower metal concentration than hyperaccumulator species. McGrath's curves also show that hyperaccumulators will not grow in soils of low metal concentration. In support of this change is evidence showing that Thlaspi caerulescens has a high physiological requirement for zinc, precluding its growth in soils with a low zinc concentration (McGrath et aI. , 1 997). Evidence presented in this thesis contradicts this theory (Chapter 8). The work of Robins on et al. ( 1 997a,b) showed that McGrath's curves could be further modified (Fig. 2 .7) . Research into the response of the nickel accumulating species Alyssum bertolonii and Berkheya coddii to increasing levels of soil nickel indeed showed hyperaccumulator behaviour consistent with McGrath's curves. But Robinson et at. ( 1 997 a, b) noticed that the curve of the hyperaccumulator species eventually reached 2 1 Chapter 2 another point of inflection. Presumably this point represents the onset of phytotoxic levels of metal in soil solution, a subsequent breakdown of the physiological mechanisms controlling metal uptake and therefore a flood of nickel into the plant effecting necrosis. Hyperaccumulator Excluder Metal in the soil Figure 2.6. Theoretical uptake response of plants to heavy metals in soil. After McGrath (2000) ..... C CO a. a> .c ..... Hyperaccumulator Excluder Metal in the soil Figure 2.7. ModifJed theoretical uptake response of plants to metals in the soil. Hyperaccumulator species are not only peculiar for their ability to accumulate metals within living tissues, but also for their ability to transport metals from roots to shoots. 22 Chapter 2 Many authors have studied the biochemistry of fluid extracts sampled from hyperaccumulator species, particularly from nickel-accumulating plants. Lee et al. ( 1 977) found nickel to be complexed with citric acid in the sap of many New Caledonian hyperaccumulators. Pancaro et al. ( 1 978) showed the nickel in Alyssum bertolonii to be complexed with a 1 : 1 molar ratio of malic and malonic acids. Stockley ( 1 980) deduced that the biochemistry of the Zimbabwean nickel hyperacumulator Personia metallifera was more complicated than that of other species, although nickel still appeared to be associated with carboxylic acids. Homer ( 199 1 ) further confirmed the association of nickel with citric and malic (carboxylic) acids. Recent work (Homer et al. , 1 997; Kramer et al. , 1 996) has shown that nickel is also complexed with amino acids within the xylem. The limited studies on zinc hyperaccumulator species suggest an association of this metal with citric acid (Godbold et al. , 1 9 88). The picture beginning to emerge is that amino acid complexes are involved in the transport of metal within a plant, but that carboxylic acids are involved with metal storage (Brooks, 1 998). The question that is now posed is 'what happens to the metal complex?' as metals need to be sequested in a non-toxic form. Zinc accumulated by the hyperaccumulator species Thlaspi caerulescens is stored within the vacuoles of epidermal and subepidermal cells (Vazquez et al. , 1 992), while nickel in the South Africa hyperaccumulator Senecio coronatus, is stored in the epidermal regions of this species leaves, stems and roots (Mesjasz-Przybylowicz et al. , 1 994). The tissues where metals have been found to be sequested are physiologically inert. That is to say, storage in these regions would allow a plant's critical biological machinery to operate in relative isolation of toxins (Boyd, 1 998) . A mechanism for nickel hyperaccumulation was proposed by Morrison ( 1 980) based on the work of Still and Williams ( 1980). The first component of this model is a ' selector' (S) molecule, confined to the root membrane. Presumably the selector exudes phytochelatins and metallothiones into the rhizosphere that selectively bind metals in a soluble form, and thus facilitate their uptake (S-Ni). The selector-metal complex is then passed through the root membrane to the inner surface, where a ternary complex is formed with a ' transport' ligand (S-Ni-T). A breakdown of this ternary complex frees the transport-metal complex to travel through the xylem, and allows the selector ligand to pass back through the root membrane. Metal is exchanged between the transport 23 Chapter 2 ligand and an acceptor ligand at the boundary of the vacuole. This forms a terminal acceptor-metal (A-Ni) complex which is sequested within the physiologically inert vacuole, effectively detoxifying the metal. The transport ligand is then free to flow back through the xylem and the process is repeated. 2.7 Bioavai labil ity of metals in the soil The key limitation to metal uptake is solubility in the rhizosphere, since metals must be soluble to some degree in the soil solution for uptake to ensue. As has been discussed, solubility can be a natural function of the soil (passive uptake), effected by the plant (active uptake), or can be artificially induced (passive uptake). The fraction of soil-metal that is in a readily available form to both plants and animals is termed bioavailable. Plant availability infers solubility of metals in the soil solution; animal availability infers solubility of metals in the gastrointestinaI tract of animals (Thornton, 1 999). Heavy metals can enter the food chain through accumulation by plants or through the direct ingestion of soil and/or dust by animals . The 'toxicity' of soil at contaminated sites is thus dependent upon the bioavailability of the associated metals. Soil physical factors such as water availability, air availability and soil strength indirectly control metal uptake through constraining plant growth (Robinson, 1 997). Soil chemical factors such as pH, nutrient status, and toxin concentrations (Emst, 1 996), affect metal uptake more directly. Likewise, the specific action of individual plant species may also affect metal uptake through the exudation of chemicals into the rhizosphere. This is not only true for hyperaccumulator species, but also for accumulator and non-accumulator species that can adjust the conditions of the rhizosphere. For example, many plant species can change the rhizosphere pH by as much as two units (Marschner, 1 995). The only definitive way to assess the potential for metal availability in soil is to grow plants in situ and analyse these for accumulated metal. However, this method is time­ consuming. Many chemical-extractant solutions have been used to estimate the bioavailability of metals in soil. Solutions have included; distilled water (Sing and 24 Chapter 2 Narwal, 1 984), O.SM CaCh (Whitten and Ritchie, 1 99 1 ), I M N&OAc (Haq et al. , 1 980; Ernst, 1 996), I M N�N03 (McGrath et al. , 1 997), and EDTA and DTPA (Haq et al. , 1 980). EDTA and DTP A show chelate properties in the soil, while the other extractants are used to model natural plant exudates that increase metal solubility in the rhizosphere. Robinson ( 1 997) compared the efficacy of several chemical extractants to estimate the bioavailable fraction of metal for a Cd, Pb and Zn contaminated soil from northern France. The author of this study showed that ammonium acetate ( l M) (Ernst, 1 996) gave reproducible results that could be mathematically related to bioavailability using extractant concentration and pH, and was thus a useful extractant to compare the bioavailability of metals between soils. Subsequently, ammonium acetate has been adopted as a Massey University standard measure for plant-available metal in soil, and as such is used in this study. There is considerable debate between research groups as to the usefulness of the various chemical extractant systems to estimate bioavailabiIity. However, there is as yet no strong data to show that there is a better extractant to use than ammonium acetate. Modelling of plant-available metal is a useful tool, because, if accurate, such modelling will predict the potential danger of metal loadings in soil. If metal is not bioavailable to either plants or animals then it is of reduced concern. However, if bioavailable, this metal should be removed, by phytoextraction or more traditional techniques. 25 SECTION A SECTION A - PHYTOEXTRACTION OF CADMIUM, LEAD AND ZINC: OBSERVED AND MODELLED UPTAKE Publications arising from section A: Anderson, C .W.N . , Brooks, R.R , Stewart, R.B . and Simcock, R., 2000. Phytoremediation: a possible management solution for New Zealand pastoral soils. Australian Journal of Experimental Agriculture, in prep. Anderson, C . , Deram, A. , Petit, D., Brooks, R , Stewart, R and Simcock, R , 1 999. Induced hyperaccumulation: metal movement and problems. In Proceedings of Extended Abstracts: 5th International Conference on the Biogeochemistry of trace Elements (Eds W.W.Wenzel , D .C.Adriano, B.Alloway, H.E.Doner, C.Kelier, N .W.Lepp, M .Mench, RNaidu, G .M .Pierzynski) pp 1 22-123 (July 1 1 - 15 , 1 999, Vienna, Austria). Anderson, C., Deram, A. , Petit, D., Brooks, R, Stewart, R and Simcock, R, 2000. Induced hyperaccumulation : is lead uptake a function of mineral-phase geochemistry? In Symposium Volume ICOBTE 1999: Bioavailability, Fluxes and Transfer of Trace Elements in Soils and Soil Components (Ed 1 .K. lskander) in press (CRC press: Florida). Hyperaccumulation of heavy metals and the associated technology of phytoextraction may be an exciting area of environmental science, but a literature review on the subject reveals that little or no attention has been paid to the potential importance of the chemical form of metal present in the soil at a contaminated site. The chemical form of metal is a function of the original source of this metal contamination. Nriagu ( 1 978a), in a review of lead in the atmosphere, summarised the composition of particulate lead forms in aerosols originating from different sources of pollution (Table A. I ) . Table A.1. Composition of lead aerosols originating from d ifferent sources o f poll ution. Source Mining activities Base-metal smelting and refin ing Coal-fired power generation Composition of particulates PbS, PbC03, PbS04, Pbs(P04hCI, PbOx Pb, PbOx, PbS04., PbC03, surface complex in fly ash, PbOx, PbS04, Pb(N03h Cement manufacture PbC03, Pbs(P04hCI Fertil iser manufacture Pb5(P04hCI , PbOx, PbC03 Note. Ordering of lead forms indicates decreasing abundance. Only simple lead salts are presented. Source: Nriagu ( 1 978a). Using electron microprobe techniques, Tee Haar and Bayard ( 1 97 1) determined the following relative abundance of metal salts found in airborne lead particles at a rural site (Table A.2). The data of these authors support contamination of the environment 26 SECTION A with lead from urban sources, i .e., the metal possibly has its origin from metal refining and from fertiliser/cement manufacture. The relatively minor bromine-lead compounds described in Table A.2 can be attributed to the exhaust of petrol-powered motor engmes. Table A.2. Relative abundance of the chemical forms of airborne lead collected from an urban site. Metal form PbC03 (PbO)zPbC03 PbO PbCI2 PbOPbS04 Pb(OH)CI PbS04 PbBrCI (PbO)zPbCI2 (PbO)zPbBrCI PbBr2 % 30 28 2 1 5 .4 5.0 4.0 3.2 1 .6 1 .5 1 .0 0 . 1 Chemical fractionation methods are commonly used to understand the distribution of metals between various soil components . The bioavailability of metal in soil is strongly related to its 'chemical form' . There are many extraction methods used, all modifications on several standard schemes, two examples being the methods of Shuman ( 1 985) and Tessier, et al. ( 1979). The basis of a fractionation method is to use sequentially stronger 'extracting agents' to extract metals from the soil. A generalised description of these fractions is as follows (after Gommy, 1 997): 1 . the exchangeable fraction (weakly stable organic-metal complexes); 2. acid-soluble fraction (carbonates); 3. the reducible fraction (oxides); 4. oxidisable fraction (sulphides, sulphates and some organic-metal complexes); and 5. the residual fraction ( silicate-bound minerals). While it is known that the chemical fractionation of soil metals affects bioavailability and thus metal uptake, the idea that different forms of metal may be present at different sites appears not to have been integrated into models for phytoextraction. Little is 27 SECTION A known about the specific effect of the form of metal present at a site on metal uptake, both natural, and more particularly, induced. Chapter 3 presents data collected from field and pot trials conducted on two contaminated substrates typical of environments where anthropogenic metal loading to the soil is of concern. Cadmium, lead and zinc contaminate these environments and occur in different chemical forms. Several species, both hyperaccumulator and non­ accumulator, have been trialled for phytoextraction, and the metal response of these species to hyperaccumulation (natural and induced) is described. To explain the observed findings, geochemical models for Pb, Cd and Zn uptake are presented in Chapters 4 , 5 and 6. The aim of generating these models was to explain variations in metal uptake exhibited by Brassica juncea, Cardaminopsis halleri and Thlaspi caerulescens as a function of the chemical form of metal present in the substrate. In an attempt to predict the plant-availability of metals in each experimental soil, ammonium acetate ( l M, pH 7) was used as an extractant based upon the conclusions of Robinson ( 1 997). The weak salt ammonium acetate is often used in sequential extraction schemes to extract the readily exchangeable fraction of soil metal (e.g. Almas et al., 1 999). An integrated geochemical model is presented in chapter 7 that highlights the importance of the chemical form of metal present in the soil on the following practical aspects of phytoextraction for the metals Cd, Pb and Zn: 1 . the choice of plant species to be used for natural hyperaccumulation, 2. the choice of chemical agent to used for induced hyperaccumulation, 3 . the choice of plant species to be used to maximise induced phytoextraction, and 4 . the value of ammonium acetate ( l M) as an extractant to model plant-available metal. An additional aim of this section of work was to quantify the effect that different metal salts, used to generate artificially contaminated soils for phytoextraction studies, would have on metal uptake. For example, Blaylock et al. ( 1 997) reported the use of carbonate 28 SECTION A salts to generate contaminated soils. Epstein et al. ( 1999) and Robinson et al. (1997a,b) reported the use of nitrate salts for this purpose while Li and Shuman ( 1 996) used nitrate and chloride salts. Equilibration times for soils used in such experiments also vary dramatically, from days to months. How the use of these different metal salts will effect uptake (natural and induced) is poorly addressed in the literature. Two questions that can be raised are: 'How will relative uptake between experiments compare?' and 'Do these artificial soils accurately model natural contamination? ' In Chapter 8, a phytoextraction field trial conducted on an area of agricultural land is described. This trial represents a test case for the model generated in the previous chapters. This third substrate is pastoral land contaminated by the heavy metal cadmium, and represents an environment where phytoextraction could potentially be very successful. The results from the field trial in Chapter 8 are integrated with the geochemical model of Chapter 7 in the conclusion to Section A. 29 Chapter 3 Chapter 3 - Trials on Contaminated Substrates 3 .1 Introduction Environmental concentrations of Cd, Pb and Zn have increased exponentially since about AD 1 750, a date that heralded the onset of the industrial revolution in Europe (Nriagu, 1 978b). The metals Cd, Pb and Zn are all chalcophilic (have an affinity for sulphur) and are thus geochemicaIly similar. These three metals often occur naturally together in the rock ores from which they are mined. Human toxicological conditions arising from heavy metal exposure are not new. Some archaeologists even believe that lead poisoning contributed to the fall of the Roman Empire (Nriagu, 1 978b). Improved understanding of human physiology has lead to greater concerns over the presence of these metals in the environment, and hence technology to minimise and reduce potential problems is actively sought. Ingestion of large doses of cadmium can cause acute gastrointestinal disturbances; an aqueous cadmium concentration of 1 5 mgIL will induce vomiting (Lauwreys, 1 977). Longer-term exposure to lower doses of cadmium can impair the functions of the kidneys, liver and the lungs. There is also evidence that cadmium can contribute to hypertension in humans and that the metal may be carcinogenic (Lauwreys, 1 979). Toxicological affects induced by lead poisoning include detrimental affects on the nervous system, renal system and gastrointenstinal tract, cardiovascular system, reproductive system and the endocrine system. Lead has also been implicated as a carcinogen and a mutagen (New land and Dunn, 1 982). Zinc is slightly different to cadmium and lead in that it is an essential element to many plant and animal metabolic functions (Ainscough and Brodie, 1 976; Marschner, 1 995; Williams, 1 973). However, at levels greater than trace concentration, similar toxicological problems can be expected (Knight et al., 1 997). Two contaminated sites have been studied in detail for this chapter, pastoral land in northern France and mine tailings in New Zealand. Both sites are heavily contaminated. Metal pollution of farmland surrounding the field site in France has lead to increased deformities in the local human population (A.Deram, pers. commn . 1 998), while the 30 Chapter 3 metal leachate from the mine tailings in New Zealand has polluted local waterways. I do not propose that either of these sites could realistically be remediated by phytoextraction utilising the current state of technology. The metal loadings at each site would necessitate hundreds of years of cropping. However, heavily contaminated land does provide an ideal <1ro/ - 1 glkg EOTA .- - 2 glkg EDTA Cl -.:x: 3g/kgEDTA -- Cl E ......... ..... c: b b ro a. 800 c: .0 a.. Figure 3.5. EDTA-Induced uptake of cadmium and lead by Berkheya coddii growing on Tui mine tailings (mean + SE, n=5). Means with the same letter are not significantly different (AN OVA p>O.05). 38 Chapter 3 EDTA effected a significant increase in lead uptake (Table 3 .4). DTPA also increased the level of lead accumulation, although this increase was not as significant as that for EDT A. The other chemicals used; ammonium thiocyanate and sodium thiosulphate, did not result in increased lead accumulation. With respect to cadmium and zinc, the 4 chemicals did not induce an increase in metal uptake; there were no significant differences among the 5 treatments for either metal. The final concentration of cadmium and zinc observed in Cardaminopsis hallei was surprising as it indicated this species, a known cadmium and zinc hyperaccumulator, did not hyperaccumulate these metals from the Tui tailings. 3.6 Hyperaccumulation trials using other species The hyperaccumulator species Thlaspi caerulescens and non-accumulator species Brassica juncea were also tested for their natural and induced-uptake potential from Tui tailings. The results follow the same pattern as for other species: 1 ) increased uptake of lead through treatment with EDT A, but no significant differences between treatment levels, and 2) no EDTA-induced effect on the cadmium and zinc concentration accumulated by these plants. Thlaspi caerulescens, a known hyperaccumulator of cadmium and zinc, like Cardaminopsis halleri, failed to hyperaccumulate these metals from the Tui tailings by natural or induced means. Ammonium thiocyanate added to the Tui tailings similarly failed to induce the uptake of Cd, Pb and Zn by Bjuncea (Fig. 3 .6). 3.7 Problems with metal uptake Sections 3 .2 - 3 .6 have outlined trials conducted on two different substrates that are typical of two different polluted environments . They both represent areas that could benefit from phytoremediation, using either natural or induced hyperaccumulation. Results from induced hyperaccumulation experiments on these substrates have, however, been somewhat surprising. Enhanced accumulation of lead was observed in all species tested on the Tui mine tailings once EDT A was applied to the substrate. However, increased uptake of cadmium and zinc was never apparent. No induced 39 Chapter 3 uptake of any metal was observed in any of the species tested during the Auby field trial . The concentration levels observed in both of these environments fell short of the lead and cadmium values observed in the literature, for plants after treatment with EDT A (Huang and Cunningham, 1 994; Blaylock et al., 1 997). 300 l 0_00 0.20 GAG Ma O_BD SCN added (g/kg soil) SCN added (g/kg soil) '_00 0.00 020 1 OAO 0.60 0.80 SCN added (g/kg soil) Figure 3.6. Thiocyanate-induced uptake (SCN) of Cd , Pb and Zn by Brassica juncea from Tui mine tailings (mean ± SE). n (control) = 3, n (treatments) = 7. LOO Both of the sites can be characterised as being contaminated with heavy metals that exist in different chemical forms (Table 3 .5) . Another site is introduced in this table to illustrate a third possible environment where heavy metal contamination may be encountered. The site is an area of pastoral land adjacent to a superphosphate storage shed, contaminated with cadmium due to the high levels of this heavy metal found in some phosphatic fertilisers. The results from a field trial conducted at this site are presented in Chapter 8. 40 Chapter 3 Table 3.5. The dominant chemical form of metal present at 3 contaminated sites. Location Auby - northern France Tui mine tai l ings - New Zealand Wairarapa - New Zealand Dominant chemical form of the contaminating metal Oxide, carbonate Sulphide. sulphate Phosphate, carbonate Source of Contamination Industrial a i r-fal l M i ne tai l ings Superphosphate fertiliser A question raised by the summation presented in Table 3 .5 is - what effect do these different metal fonns have on a) the bioavailability of metals to the plant under conditions of natural uptake, and b) the bioavailability of metals to the plant under conditions of induced uptake? To answer and model this question, a series of greenhouse pot trials were carried out in conjunction with the 199811 999 field trials, in which p lants were grown in a commercial seed-raising mixture spiked with cadmium, lead and zinc salts of different mineral phases. The results were used to develop uptake models for each metal. Each model is discussed in turn through the next 3 chapters. Terminology for Chapters 4, 5 and 6 The tenn 'metaVmineral phase' is used in Chapters 4 , 5 and 6 to describe the chemical fonn of metal that was added to each soil, to model heavy metal contamination that could originate from different sources . The phrase is used in this sense to describe the 'primary metal phase ' that may exist in soils, and not the secondary phases that may exist due to the release of metals from the contaminant's mineral 1attice. Natural uptake refers to metal uptake effected by the plant, as opposed to induced uptake effected by chemicals added to the soil. 4 1 Chapter 4 Chapter 4 - Geochemical Model for Lead Uptake 4.1 Introduction To model the geochemical conditions prevalent in the Auby soil , commercial seed-raising mix was used as a base substrate to create an artificial, lead-contaminated ' soil' (C=20%, N=O. 6%). A substrate lead concentration of 1 % was used. To induce hyperaccumulation, a high concentration of acetic acid , citric acid or EDT A was added to each relevant pot. A high concentration was used to ensure that the chemical-inducing agent was present in the soil at an excess concentration, thus minimising the chance that differential rates of chemical degradation (function of the half life of the chemicals used) was a factor controlling metal uptake. The aim of this experiment was to determine the affect of the chemical form of lead on induced and natural lead uptake. I realise that the rate of chemical-inducing agent may have negatively affected plant health and thus metal uptake, however, this factor is independent of chemical form and was not considered in this experiment. Leaching of metal out of the various pots was similarly not considered. 4.2 Experimental Design S ix mineral salts were chosen to represent the initial form of lead contamination that could occur in a wide range of environments: carbonate, nitrate, oxide, phosphate, sulphate and sulphide. The soluble nitrate salt was chosen to model lead, after dissolution, as part of the soil organic phase, i .e . , after a period of days no lead would remain in the soil as part of the original crystalline phase. Each salt (as a solid) was added to commercial potting�mix to give a final soil-lead concentration of 1 % (w/w). Pots (250 mL - 280 in total) were planted, i n equal numbers, with either the non-accumulator speCIes Brassica juncea or the hyperaccumulator species Thlaspi caerulescens. A control substrate was used, where the two species were planted in pots containing potting-mix with no added lead. During the growing cycle, pot positions were randomly changed on a periodic basis to equalise light exposure. The ambient temperature of the 42 Chapter 4 greenhouse was set at 1 5-25°C with no humidity control. Overhead watering was carried out each day with a hand-held hose. After approximately 10 weeks growth, five replicates of each plant species, for each metal phase, were treated with one of 2 g/ki EDT A (disodium salt), 2 glkg citric acid, 2 g/kg acetic acid or water as a control . All treatments were applied as a solution (20 mL) and were randomly allocated to replicate pots. Two weeks after treatment, the above-ground portions were harvested3, dried at 60°C to constant weight and subsamples digested in concentrated nitric acid before analysis by F AAS. As replicate specimens of Brassica juncea had reached different stages of maturity, and thus showed different weight ratios of stems, leaves and flowers, only the leaves of this species were analysed to minimise the variation in results that could be attributed to an uneven distribution of organs for an individual plant. Substrate samples were 'cored' from the control pots of each metal phase at the time of plant harvest, ground using a porcelain mortar and pestle, and subsamples digested in aqua regia to give the total-metal concentration for each prepared ' soil ' . Ammonium acetate ( lM, pH 7) was used to estimate the concentration of plant-available lead (Ernst, 1996; Robinson, 1997) for each mineral phase, by overnight shaking at a soil:liquid ratio of 1 : 1 0. In each case, analysis of the filtrate was performed using F AAS. Measurement of the soil pH was conducted in water, using a 1 :2.5 soil: liquid ratio. All data were tested for normality, and analysed using ANOVA due to the observed normal distribution. 4.3 Results : Brassica juncea. For each metal phase, EDTA caused a significant increase in the concentration of lead accumulated by the plant (Fig. 4. 1 ). There was no significant difference between control 2 glkg is grams of chemical applied per kilogram fresh (moist) weight potting-mix 3 Plant health at the time of harvesting is reported in Appendix 2 43 1 0000 Lead Control 1000 100 1 0 Sj Tc Lead Nitrate 1 0000 - � b 1 000 a 1 00 Sj Tc Lead Phosphate 10000 1 000 1 00 10 Sj Tc Lead Sulphate 1 0000 1 000 1 00 10 Sj Tc 1 0000 1 000 100 10 1 0000 1000 100 10 1 0000 1000 100 1 0 Lead Carbonate Sj Tc Lead Oxide Sj Tc Lead Sulphide Sj Tc Treatment c=J control f:} }:: :l citric acid _ acetic acid � EDTA Figure 4.1 Natural uptake, and acetic acid-, citric acid- and EDTA-induced uptake of lead by Brassica juncea (Bj) and Thlaspi caerulescens (Tc) growing on artificial 1 % lead soils of different metal phases (mean + SE, n=5). Means for the same species and phase, with the same letter, are not significantly different (ANOVA p>O.05). Chapter 4 44 Chapter 4 and citric acid treatments for each lead salt. In the case of the carbonate and oxide salts, acetic acid also caused a significant increase in the plant lead concentration. The results clearly show that the relative efficacy of EDT A to induce lead hyperaccumulation is dependent upon the chemical form of the metal present in the soil (Fig. 4 .2). The suitability of these forms to EDTA-induced hyperaccumulation can be written as follows: control (a) < carbonate (bc) - n itrate (cd ) = oxide (cd) .... sulphate (d) < sulphide (e) < phosphate (f)4 This ordering is different from that of the plant-available lead concentration in the soil, estimated by extraction with ammonium acetate ( lM, pH 7): control (a) < sulphide (b) .... sulphate (bc) .... phosphate (c) < carbonate (d) = nitrate (d) = oxide (d) The difference is attributable to the chelation effect of EDT A on lead, i .e . , the effect of EDTA to induce uptake is independent of the plant-available or ' soluble' concentration of lead in the soil. Perhaps the most interesting point to note is the efficacy for EDT A­ induced uptake from a sulphide lead form. Lead associated with the sulphide salt had the lowest plant-available metal concentration of the soils. However, EOT A-induced uptake was very high, more so than for the soluble nitrate form. 4.4 Resu lts: Thlaspi caerulescens A metal uptake pattern is less obvious for Thlaspi caerulescens (Fig. 4 . 1 ) . With respect to the lead sulphate, sulphide and phosphate saits, EOT A caused a significant increase in lead uptake relative to the control treatment. In the case of the sulphate salt, acetic acid 4 Letters refer to statistical differences between the mean values presented in Figure 4 . 1 and Appendix 7. Means with the same letter are not statistically different, ANOVA p>O.OS . 45 Chapter 4 also caused a significant increase in lead uptake . For each of these three salts, citric acid caused no significant increase in lead uptake re lative to the control treatment. .--. ..... .c Cl ·m 3: � "0 Cl � - Cl E ........ ..... C rn c.. C .0 a.. 1 0000 f** Lead phase 1 000 c::J control carbonate nitrate oxide 1 00 sulphate sulphide phosphate 1 0 1 .....L.....--'--- Figure 4.2. Summary: efficacy of EDTA-induced lead uptake by Brassica juncea as a function of metal phase. Means with the same letter are not significantly different (ANOVA p>0.05). Note .p 8>A = 0.087 •• P F>E = 0.069. In the case of l ead carbonate, acetic acid and E DTA did not cause a sign i ficant increase in metal uptake relative to the contro l . However, c itric acid caused a signi ficant decrease in p lant lead. The mean lead concentration for control plants grown on the carbonate contaminated soi l was 1 1 75 mg/kg. Natural hyperaccumulation was observed for Thlaspi caerulescens in this particu lar model environment. Hyperaccumu lation was similarly observed for lead added as a nitrate salt (mean value of I 220 mg/kg) . After treatment with E DT A, the increase in lead uptake was significant and represented the greatest increase in metal concentration observed in this experiment. The only treatment to cause a signi ficant increase in lead uptake from the oxide salt was acetic acid. For the oxide form of lead the increase in uptake due to EDT A was not significant. 46 Chapter 4 4.5 Results : total soil lead Digestion by aqua regia, and subsequent analysis of subsamples of substrate taken :from every control treatment soil (n=5) , showed the final lead concentration of each metal phase to agree with the ' designed' concentration of 1%. Soil pH was independent of mineral salt used (Table 4. 1) . Table 4.1 . Total soil lead and pH for the control treatment soils of each metal phase. Mean values with the same letter are not significantly d ifferent (AN OVA p>0.05). Lead phase Control Sulphide Oxide Sulphate Carbonate Phosphate N itrate pH 4.4 4.3 4.8 4.2 4 .4 4.3 4.5 4.6 Results : plant-available lead Total Lead (%) 2.42 (mg/kg) ± 0.88 (a) 0.81 ± 0.28 (b) 0 .92 ± 0.24 (b) 1 .20 ± 0.22 (c) 1 .2 1 ± 0 .1 0 (c) 1 .30 ± 0.20 (c) 1 .30 ± 0.20 (c) Under the geochemical conditions of this study, the relative ordering of the plant­ available lead concentration, modelled by ammonium acetate (Table 4 .2), confirmed that bioavailability of lead was strongly dependent upon the chemical form of lead contamination present in the soil. Table 4.2. Plant-available (ammonium acetate) lead for the control treatment soils of each metal phase. Mean l ead concentrations with the same tetter are not statistically different (AN OVA p>0.05). Lead phase Control Sulphide Sulphate Phosphate Carbonate Nitrate Oxide NH40Ac-extractable lead (mg/L) 0.28 ± 0 .2 (a) 33.1 ± 5 .4 (b) 42.7 ± 4.1 (be) 55.2 ± 9.5 (c) 92.6 ± 34.5 (d) 1 06.6 ± 33.1 (d) 1 1 5.5 ± 26.4 (d) 47 Chapter 4 4.7 Discussion - a model for lead uptake Except for the high lead concentration observed in Thlaspi caerulescens from the carbonate and nitrate-contaminated soils, natural uptake for both plant species used was never more than 100 mg/kg. This was not modelled by ammonium acetate (Table 4 .2), which predicted that significant metal accumulation could be expected by both Brassica juncea and T caerulescens, due to the high values for plant-available meta15 . Hyperaccumulation of lead by Bjuncea is an induced phenomenon that is highly dependent upon the chemical form of lead present in the soil. In the absence of soil applied chelates Brassica juncea may 'exclude' lead uptake, but natural hyperaccumulation of lead by T. caerulescens appears to be realistic under some geochemical conditions. Ammonium acetate did not predict this hyperaccumulation. Phytoextraction of lead is maximised from each mineral phase through EDTA-induced hyperaccumulation by the plant species Brassica juncea. The best choice of chemical inducing agent for an environment contaminated with lead present as a phosphate or sulphide phase is, based on the results of this experiment, EDT A. The individual plant species, lead phase combination plots (Fig. 4. 1 ) show that for an environment contaminated with lead as an oxide or carbonate phase, acetic acid would be a better choice, as its relative efficacy compared with EDT A in each of these cases is nominally greater. It is interesting to note that EDTA-induced hyperaccumulation by Brassica juncea from the nitrate phase was relatively low. The nitrate salt is completely soluble, and hence, after dissolution, was likely to be present in the substrate as an organic phase or in equilibrium with the soil solution. Soluble metal salts have often be used to model polluted soils in pot trials. This result suggests that data obtained from such experiments may be misleading, and may underestimate the ability of EDT A to induce hyperaccumulation in Bjuncea relative to the less soluble forms of lead in soil. 5 The assumption true for a good extractant is that the concentration of plant-available metal will be directly proportional to the concentration of metal accumulated by the plant. 48 Chapter 4 The concentration of lead accumulated once EDT A is applied appears to be species­ independent. The final metal concentrations of both Brassica juncea and Thlaspi caerulescens once hyperaccumulation was induced were very similar. This is in agreement with trials on the Tui mine tailings, where all species used could be induced to accumulate a similar concentration of lead. An exception to this observation is the EDTA-induced uptake by T.caerulescens of lead added to the soil as a nitrate salt. The comparison of induced lead uptake between the two plant species is interesting, and may involve some physiological response of T.caerulescens to the lead-organic phase - EDT A interaction. The nature of this response will not be speculated on here. It is important to note, however, that again, pot trials conducted where lead is added as a soluble metal salt may be misleading and overestimate the induced metal uptake potential of T.caerulescens from contaminated soils. Where the metal-chelate complex effects no physiological response by the plant, the technology of induced hyperaccumulation for lead may well be equally effective for all plant species. 4.8 Conclusion Uptake of lead by both Brassica juncea and Thlaspi caerulescens is maximised by induced hyperaccumulation. The choice of chemical to be used to effect this uptake is dependent upon the chemical form of the lead that initially contaminated the soil. Natural hyperaccumulation of lead was not observed for B.juncea but may be true for T.caerulescens for some forms of lead in the environment. Ammonium acetate did not effectively model lead uptake. The high values of plant­ available lead did not translate to metal uptake. Hyperaccumulation of lead by Thlaspi caerulescens was not reflected by anomalies in the plant-available metal concentration. 49 Chapter 5 Chapter 5 .. Geochemical Model for Cadmium Uptake 5.1 Introduction To model the geochemical conditions prevalent in the Auby soil, commercial seed-raising mix was used as a base substrate to create an artificial, cadmium-contaminated ' soil' (C=20%, N=0 .6%). A soil cadmium concentration of 200 mg/kg (0.02%) was used. To induce hyperaccumulation a high concentration of citric acid or EDTA (disodium salt) was added to each relevant pot. A high concentration was used to ensure that the chemical-inducing agent was present in the soil at an excess concentration, thus minimising the chance that differential rates of chemical-degradation (a function of the half life of the chemicals used) was a factor controlling metal uptake. The aim of this experiment was to determine the affect of metal phase on induced and natural cadmium uptake. I realise that the rate of chemical-inducing agent may have negatively affected plant health and thus metal uptake, however, this factor is independent of mineral phase and was not considered in this experiment. Leaching of metal out of the various pots was similarly not considered. 5.2 Experimental design Five mineral salts were chosen to represent the initial form of cadmium contamination that could occur in a wide range of environments: carbonate, nitrate, oxide, phosphate and sulphide. The soluble nitrate saIt was chosen to model cadmium, after dissolution, as part of the soil-organic phase, i . e. , after a period of days no cadmium would remain as part of the original crystalline phase. Each salt (as a solid) was added to commercial potting-mix to give a final soil-metal concentration of 200 mg/kg (w/v). Pots (250 mL - 3 1 5 in total) were planted, in equal number, with the non-accumulator species Brassica juncea, or the hyperaccumulator species Cardaminopsis halleri or Thlaspi caerulescens. A control substrate was used, where the three species were planted in pots containing potting-mix with no added cadmium. During the growing cycle, pot positions were randomly changed on a periodic basis to equalise light exposure. The ambient 50 Chapter 5 temperature of the greenhouse was set at 1 5-25°C with no humidity control. Overhead watering was carried out each day with a hand-held house. After approximately 1 0 weeks growth, 5 replicates of each plant species, for each cadmium phase, were treated with one of 2 glkg EDTA (disodium salt), 2 glkg citric acid or water as a control. All treatments were applied as a solution (20 mL) and were randomly allocated to replicate pots. Two weeks after treatment, the above-ground biomass was harvested6, dried at 60°C to constant weight and subsamples were digested in concentrated nitric acid before analysis by F AAS. As replicate specimens of Brassica juncea had reached different stages of maturity, and thus showed different weight ratios of stems, leaves and flowers, only the leaves of this species were analysed to minimise the variation in results that could be attributed to an uneven distribution of organs for an individual plant. Substrate samples were ' cored' from the control pots of every metal phase at the time of plant harvest, ground using a porcelain mortar and pestle, and subsamples digested in concentrated nitric acid to give the total metal concentration for each of the prepared soils. Ammonium acetate ( lM, pH 7) was used to estimate the concentration of plant­ available cadmium (Ernst, 1 996) for each metal phase by overnight shaking at a soil: liquid ratio of 1 : 10 . In each case, analysis of the filtrate was performed using F AAS. Measurement of the soil pH was conducted in water, using a 1 :2.5 soiI:liquid ratio. Al data were tested for normality and analysed using ANOV A due to the observed normal distribution. 5.3 Resu lts: Brassica juncea With the exception of the carbonate salt, EDT A caused a significant increase in cadmium uptake from every metal phase (Fig. 5 . 1 ). There was no significant difference between control and citric acid treatments. The concentration of metal accumulated through natural uptake was a function of the chemical fonn of metal in the soil . 6 Plant health at time of harvesting is reported in Appendix 2. 5 1 1600 1200 800 Cad m i u m Carbonate a a 8j Ch Te Cad m i u m Oxi d e a 8j Ch Te Cadmium Sulph ide a b 2000 1600 2000 1600 1200 BOO 400 0 Cad m i u m N itrate 8j b 8j Cad m i u m Phosphate Ch Tc Treatment control I. ' . . . . . / 1 citric acid _ EOTA Figure 5.1 . Natural uptake and EDTA- and citric acid-induced uptake of cadmium by Brassiea juncea (Bi), Cardaminopsis hailed (Ch) and Th/asp! caeruleseens (Tc) g rowing on artificial 200 mg/kg cadmium soils. of different metal phases (mean + SE, n=5). Means for the same species and phase, with the same letter, are not significantly different (ANOVA p>O.05). 5 52 Chapter 5 The relative suitabi lity of the 5 salts to natural uptake can be ordered as: sulphide (a) - phosphate (ab) - n itrate (ab) - oxide (be) - carbonate (c) 1 This is different from the ordering of plant-available cadmium, estimated by extraction of each control soi l with ammonium acetate ( l M , pH 7): su lphide (a) :: n itrate (a) :: oxide (a) < phosphate (b) :: carbonate (b) Summarising the E DTA treatment response of cadmium uptake by Brassica juncea ( F ig. 5 .2), allows comparison of the relative efficacy of EDTA to induce uptake from each metal phase . 2000 1600 1 200 600 400 0--1...-- b a Cadmium phase _ sulpnide _ carbonate _ nitrate _ oxide _ pnosphate Figure 5.2. Summary: efficacy of EDTA-induced cadmium uptake by Brassica juncea as a function of the metal phase. The ordering of induced uptake can be written as : sulphide (a) :: carbonate (a) :: nitrate (a) :: oxide (a) < phosphate (b) I Letters refer to statistical differences between the mean values presented in and Appendix 6. Means with the same letter are not statistically different, ANOV A p>O.05 . 5 3 Chapter 5 EDT A has a similar effect on each o f the tested metal salts, with the exception of cadmium phosphate. 5.4 Resu lts: Cardaminopsis halleri For every metal phase there was no d i fference in cadmium uptake between the three treatments (Fig. 5 . 1 ) . N either citric acid nor E DTA influenced the cadmium uptake potential o f Cardaminopsis halleri under the geochemical conditions of this study. Summarising natural uptake from each of the 5 c admium salts (Fig. 5 . 3 ) allows comparison o f the importance of the chemical form of cadmium on this species abil ity to effect natural hyperaccumulation . The ordering of natural uptake can be written as : su lphide (a) = oxide (a) = n itrate (a) = carbonate (a) < phosphate (b) --..... .r:: 0> '03 3: � "0 0> � -.... 0> E '-' ..... C m c.. c "0 () 2000 1 600 Cadmium phase 1 200 b - sulphide - oxide - nitrate - carbonate 800 - phosphate 400 Figure 5.3 Summary: efficacy of natural cadmium uptake by Cardaminopsis halleri as a function of the metal phase. 5 4 Chapter 5 With the exception of the phosphate salt, natural uptake is the same from each cadmium phase. This observation is inconsistent with the ordering of plant-available cadmium (ammonium acetate extractable), estimated from the control soils of each cadmium salt: su lphide (a) = carbonate (a) = nitrate (a) < oxide (b) < phosphate (c) 5.5 Results : Th/asp; caeru/escens The cadmium uptake response of Thlaspi caerulescens, effected by the three treatments on each metal phase, followed a less simple pattern than for the other two plant species (Fig. 5 . 1 ). For the carbonate, oxide and phosphate phases there was no significant difference among the three treatments. Replicate uptake values for each treatment on these three phases were highly variable; outliers were not removed from the data set before statistical analysis. For the nitrate phase, both citric acid and EDTA caused a significant increase in cadmium uptake. However, for the sulphide phase, EDTA caused a significant decrease in cadmium uptake. There was no significant citric acid-induced effect on uptake from the sulphide phase. The natural cadmium uptake potential of Thlaspi caerulescens from each mineral phase (Fig. 5 .4) can be written as follows: oxide (a) = nitrate (a) = carbonate (a) < su lphide (b) - phosphate (b) Again, natural uptake from the phosphate phase was the highest of all the salts. For Thlaspi caerulescens, natural uptake from the sulphide phase was also significantly greater than from the remaining three salts. The ordering of natural uptake for Thlaspi caerulescens is very different from the ordering of plant-available (ammonium acetate) cadmium estimated from the control soils of each mineral phase: sulphide (a) < oxide (b) - carbonate (be) - nitrate (c) < phosphate (d) 55 Chapter 5 ---.... .c Cl '0) :: >-� "0 Cl � - Cl E -- .... c: CO c.. c: "0 () 2000 1600 b Cadmium phase 1 200 - oxide - nitrate - carbonate b - sulphide BOO phosphate 400 0--'-- Figure 5.4. Summary: efficacy of natural cadmium uptake by Th/aspi caeru/escens as a function of the metal phase. 5.6 Results : tota l soil cadmium Digestion b y concentrated nitric ac id and subsequent analysis of subsamp les of substrate taken from each of the control treatment pots, showed signi ficant variation in the final cadmium concentration for each phase (Table 5 . 1 ) . The five metal salts did not affect the pH of the soi l . Table 5.1 . Total soi l cadmium and pH for the control treatment soils of each mineral phase. Mean values with the same letter are not statistically d ifferent (ANOVA p>0.05) . Mean (±sd) . Cadmium phase pH Total cadmium (mg/kg) Control 4 .9 Note Sulphide 4 .9 1 BO ± 20 (a) Carbonate 4 .B 220 ± BO (b,c) Nitrate 4 .B 240 ± 1 5 (c) Oxide 4 .8 240 ± 35 (c) Phosphate 4 .9 485 ± 50 (d) Note. The total cadmium concentration of the control soil was below detection l imits by FAAS and is not reported in this table. Effect of metal concentration on uptake The concentration of 2 00 mg/kg c admium was chosen for this experiment to model the level o f cadmium observed in Auby field so i ls, and to ensure that cadmium was present 5 6 Chapter S in the soil to excess concentration (i.e. excess metal relative to the potential for uptake). To test the effect of metal concentration on plant uptake, two cadmium phosphate soils were prepared; a soil with a final concentration of 485 mglkg cadmium (Table 5 . 1 ) and a soil prepared to have a final concentration of 200 mg/kg cadmium. For the control plants of each species, the concentration of cadmium accumulated by the p lant was the same for both soil concentrations. By ensuring that soil cadmium was present at an excess level, differences in the total cadmium concentration between each phase did not affect plant uptake. The important factor was the mineral phase itself, i .e . , uptake was independent of the total soil-metal concentration. 5.7 Results: plant-avai lable cadmium The ordering of plant-available cadmium, reported for each plant species, was based upon values determined from each of the control treatment soils for each species. These values do not account for differences in metal concentration between each set of soils, and are therefore only relevant to their respective plants. To accurately compare the b ioavailablility of each mineral phase, and hence the potential solubility of cadmium in the rhizosphere as a function of cadmium geochemistry, these data for plant-available metal must be corrected for total concentrations (Table 5 .2). Reporting the percentage of the total metal in the soil that is plant-available has made this correction. Table 5.2. Percentage of the total cadmium that is plant-avai lable ( 1 M ammonium acetate) from each metal phase using a 1 : 1 0 soi l : l iquid ratio . Mean cadmium values with the same letter a re not statistical ly different (ANOVA p>O.OS). Cadmium phase % extractable cadmium Control Note Sulphide 27.7 ± 2.6 (a) N itrate 32.4 ± 1 .0 (ab) Phosphate 34.3 ± 2.2 (b) Oxide 38.1 ± 1 0 . 1 (b) Carbonate 44.S ± 1 S .0 (c) Note. The plant-available cadmium concentration of the control soi l was below detection l imits by FAAS and is not reported in this table. Under the geochemical conditions of this study, the concentration of plant-available cadmium, estimated by ammonium acetate, confirms that bioavailability of cadmium is strongly dependent upon mineral phase. 57 5.8 Discussion - a model for cadmium uptake Chapter 5 Increased uptake of cadmium can be induced in the non-accumulator species Brassica juncea using the chemical EDT A. With the exception of a phosphate salt, a two- to three-fold increase in cadmium uptake was apparent. Uptake from cadmium phosphate was increased approximately five-fold. Induced uptake is dependent upon the phase of the metal present. In this experiment, natural uptake by Brassica juncea from each mineral phase was above the cadmium hyperaccumulation threshold. This was a surprising result, but the plants, while showing symptoms of metal stress, were all alive and relatively healthy (Fig. 5 .5 ; Appendix 2). Brassicajuncea appears to have the ability to uptake high levels of cadmium from soils with a geochemistry similar to that of the experimental substrate used in this study, a finding that is agreement with Blaylock et al. ( 1 997). These authors found that Bjuncea could ' naturally' accumulate 220 mg/kg cadmium DW when grown on a carbonate-contaminated soil (artificial) with a substrate cadmium concentration of 1 00 mg/kg. Cadmium itself rarely inhibits plant growth (McGrath, 1 998). However, cadmium uptake occurs naturally, contemporaneously with zinc which is considerably more phytotoxic . In the absence of zinc, it is possible that non-accumulator species may be able to 'hyperaccumulate' cadmium. In light of this conclusion, the question of 'what is hyperaccumulation?' can be raised. Brassica juncea should not be regarded as a hyperaccumulator species, as no evidence indicates that this species has evolved the perculiar trait to allow it to survive in metal-rich environments, although it may be an indicator species. The issue that the criteria for hyperaccumulation should be re­ examined has been raised (M.Macnair, pers. commn. 1 999 - Chapter 1 .5), and evidence such as this supports such a re-examination. Based upon the results of this experiment, there is no strong evidence that citric acid- or EDTA-induced uptake to a higher concentration of cadmium is possible for the hyperaccumulator species Cardaminopsis halleri and Thlaspi caerulescnes. This is in contradiction to the findings in Chapter 4 where EDTA-induced uptake of lead was apparent for T. caerulescens. The natural uptake potential for each of these species 58 Chapter 5 varied between each of the different mineral phase. Thlaspi caerulescens showed poor uptake from the oxide, nitrate and sulphide phases relative to C.halleri, but greater relative uptake from the carbonate and phosphate phases. The relationship between this uptake and the plant-available cadmium concentration is discussed below. Figure 5.5. Brassica juncea growing on an artificial. carbonate phase. cadmium-contaminated soil shortly before harvesting. The EDTA treated specimen shows signs of chlorosis, however, the control-treatment specimen to the left shows no sign of metal stress. 59 Chapter 5 Relationship between ammonium acetate and the metal phase 1. Brassica juncea The ordering of natural uptake from each metal phase, effected by Brassica juncea, was not the same as that modelled using ammonium acetate. This can be presented graphically (Fig. 5 . 6) . ..... c ro 0.. .S ,....... Cl .:s:. --Cl E -- c o +=l � ..... C ID U C o U -0 U 1 000 800 600 400 200 o Cadmium phase + carbonate • nitrate + • oxide • phosphate A sulphide R2=O.64 n itrate , / I oxide and sulphide . I phases / �I ______ -----' + + rf. R2:::0.01 al l phases - - -. a e . 1 0 20 + plant-avai lable Cd (mg/L) in soil + 30 Figure 5.6. Plot of the cadmium concentration in Brassica juncea as a function of the plant-available cadmium concentration in the soil for each metal phase. The sulphide, oxide and nitrate phases appear to plot on a linear relationship (R2=O. 64 - Fig. 5 . 6) . However, the phosphate and carbonate phases behave differently. The regression coefficient for all phases modelled together is very low (R2=O . 0 1 ) . This suggests that ammonium acetate may not model plant-available cadmium equally well for each metal phase, an observation that could explain the discrepancy between modelled uptake and the actual ordering of accumulated metal in the plant as a function of metal phase. 6 0 Chapter 5 2. Cardaminopsis haller; EDTA and citric acid did not affect the cadmium uptake of Cardaminopsis halleri in this experiment. The data show that these chemicals neither increase nor decrease the concentration of accumulated metal. The discrepancy between the observed ranking of metal uptake (natural) from each mineral phase and the modelled potential for uptake using ammonium acetate, can again be depicted graphically (Fig 5 . 7) . • 1 200 • R2=O.35 carbonate, nitrate, phosphate and sul phide phases ...... • C CO 0. c - R2=O. 1 7 al l phases 0> .::.c. "- 800 0> + E - • Cadmium phase C 0 ++�� / + carbonate :;::. CO '- • nitrate ...... C +� ' (]) • - . • oxide t) C :4' 0 • phosphate t) 400 /.: $ • • "0 ... sulphide () ... . • - o 1 0 20 30 plant-avai lable Cd (mg/L) in soi l Figure 5.7. Plot of the cadmium concentration i n Cardaminopsis halleri as a function o f the plant-available cadmium concentration in the soil for each metal phase. The sulphide, nitrate, carbonate and phosphate phases all appear to plot roughly on a linear relationship (R2=O . 3 5 - Fig. 5 . 7) . For Cardaminopsis halleri, the oxide phase behaves differently and effects a drop in the regression coefficient of the data set when analysed with the other 4 phases. This result supports that assertion that ammonium acetate as an extractant does not accurately model plant-available cadmium equally well for every mineral phase. 6 1 Chapter 5 3. Thlasp; caerulescens Thlaspi caerulescens responded differently to EDT A and citric acid treatment when compared with Cardaminopsis halleri. There was a significant increase in cadmium uptake from the organic (nitrate) phase using both of these chemicals. From the sulphide phase, EDT A caused a significant decrease in plant cadmium. Graphical representation of predicted verses modelled natural cadmium uptake clearly shows this discrepancy in uptake from the sulphide phase (Fig 5 . 8) . ....... c co a.. 2000 c 1 500 c o . .;::; � ....... c Q) u c o U "0 () o • • • • R2=O.48 carbonate, n i trate, oxide and phosphate phases / / / R2=O.2S all phases Cadmium phase + carbonate • nitrate • oxide • phosphate .. sulphide 1 0 20 plant-available Cd (mg/L) in soil 30 Figure 5.S. Plot of the cadmium concentration in Thlaspi caerulescens as a function of the plant-available cadmium concentration in the soil for each metal phase. The carbonate, nitrate, oxide and phosphate phases all appear to plot on a linear relationship (R2=0.48 - Fig. 5 . 8) . The sulphide phase plots independent of the other points and when analysed with the other 4 phases, effects a drop in the value of the regression coefficient for the data set. Again the explanation for this discrepancy could lie in the relative efficacy of ammonium acetate as an extractant for the various mineral salts as was shown true for Brassica jllncea and Cardaminopsis halleri. 62 5.9 Conclusion Chapter 5 Natural and induced uptake of cadmium by the non-accumulator species Brassica juncea i s dependent upon the metal phase of cadmium present in the soil, a function of the source of the original metal contamination. An increase in uptake was observed for all phases except the carbonate phase using EDT A. The phosphate phase had a significantly greater ' amenability' to induced uptake. Citric acid had no effect on any phase. The phytoextraction potential of Bjuncea under the geochemical conditions of this study is maximised using EDT A. Induced hyperaccumulation for cadmium is apparently ineffective usmg the hyperaccumulating species Thlaspi caerulescens and Cardaminopsis halleri. The uptake potential of T caerulescens was increased from the nitrate (organic) phase using both c itric acid and EDTA. However, uptake by C halleri was not increased for any phase using either chemical. The uptake mechanisms of these species appear different. Thlaspi caerulescens has a greater potential for metal uptake from the phosphate and sulphide phases, while C halleri has a greater potential for uptake from the carbonate, oxide and nitrate phases. The reason for this difference can only be speculated on here as it is beyond the scope of this study. However, the chemicals actively exuded by each species may in fact be different, and more suited to ' dissolving' metals from different minerals. The use of ammonium acetate as an extractant to estimate the plant-available fraction of cadmium in soil also appears to be dependent upon the mineral phase. If each phase was equally modelled by ammonium acetate, then a single curve describing the relationship would be expected. This is not the case when several phases are modelled in the same graph. Further work should be conducted to clarify the relationship between mineral phase and species-specific modelled plant availability. It appears that ammonium acetate may not be a suitable extractant to estimate plant-available cadmium for all sources of contamination, or for all plant species. 63 Chapte r S Chapter 6 - Geochemical Model for Zinc Uptake 6.1 I ntroduction To model the geochemical conditions prevalent in the Auby soil, commercial seed-raising mix was used as a base substrate to create an artificial, zinc-contaminated ' soil' (C=20%, N=O.6%). A substrate zinc concentration of 0 .2% was used. To induce hyperaccumulation a high concentration of EDTA (disodium salt) was added to each relevant pot. A high concentration was used to ensure that the chemical-inducing agent was present in the soil at an excess concentration, thus minimising the chance that differential rates of chemical degradation (a function of the half life of EDT A) was a factor controlling metal uptake. The aim of this experiment was to determine the affect of mineral phase on induced and natural zinc uptake. I realise that the rate of EDT A applied may have negatively affected plant health and thus metal uptake, however, this factor was independent of mineral phase and was considered in this experiment. Leaching of metal out of the various pots was similarly not considered. 6.2 Experimental deSign Five metal salts were chosen to represent zinc pollution that could occur in a wide range of environments : carbonate, oxide, phosphate, sulphate and sulphide. The soluble sulphate salt was chosen to model zinc, after dissolution, as part of the soil organic phase, i . e. , after a period of days no zinc would remain in the soil as a crystalline phase. Each salt (as a solid) was added to commercial potting mix to give a final soil-metal concentration of 2 000 mg/kg (w/v) . Pots (250 mL - 1 80 in total) were planted, in equal numbers, with either the non-accumulator species Brassica juncea, or the hyperaccumulating species Cardaminopsis halleri or rh/aspi caerulescens. A control substrate was used, where the three species were planted in pots containing potting-mix with no added zinc. During the growing cycle, pot positions were randomly changed on a periodic basis to equalise l ight exposure. The ambient temperature of the greenhouse was set at 1 5-25°C with no humidity control. Overhead watering was carried out daily using a hand-held house. 64 Chapter 6 After approximately 1 0 weeks growth, 5 repl icates of each plant species, for each mineral phase, were treated with either 2 glkg EDT A or water as a control . All treatments were applied as a solution (20 mL) and were randomly allocated to rep licate pots. Two weeks after treatment, the above-ground biomass was harvested8, dried at 60°C to constant weight and subsamples were digested in concentrated nitric acid before analysis by F AAS. Substrate samples were 'cored' from the control pots of every metal phase at the time of plant harvest, ground using a porcelain mortar and pestle, and subsamples were extracted with hydrochloric acid (5M) to give the total metal concentration for each of the prepared soils. Ammonium acetate ( lM, pH 7) was used to model the concentration of plant-available zinc (Ernst, 1 996; Robinson, 1 997) for each metal phase by overnight shaking at a soil: liquid ratio of 1 : 1 0. In each case, analysis of the filtrate was performed using F AAS. Measurements of soil pH were performed in water, using a 1 :2 .5 soi1 : liquid ratio . Data were tested for normality, and analysed using ANOVA due to the observed normal distribution. 6.3 Resu lts: Cardaminopsis halleri Zinc uptake was increased from the phosphate phase using EDT A, but decreased from the oxide phase relative to the control (Fig. 6 . 1 ). From the carbonate, sulphate and sulphide phases, no change in plant zinc was apparent after treatment. Summation of the natural uptake response of Cardaminopsis halleri to each of the 5 salts i l lustrates the importance of the chemical form of zinc present in the soil on this species ability to naturally accumulate zinc (Fig. 6 .2). The highest metal uptake observed in this experiment was for EDT A-induced uptake by C.halleri from the carbonate phase (2.8%). 8 Plant health at time of harvesting is reported in Appendix 2. 65 25000 Zinc carbonate Zinc oxide 20000 a 15000 p b*>a = 0.065 a a a a a 1 0000 5000 5000 0 0 Ch Te Ch Te 25000 Zinc phosphate Zinc sulphate a 15000 b a 10000 0 0 Ch Te Ch Te 25000 20000 Zinc sulphide Treatment I I control a - EDTA 5000 0 Ch Te Figure 6.1 . Natural uptake and EDTA-indueed uptake of zinc by Cardaminopsis halleri (Ch) and Thlaspi caerulescens (Te) growing on artificial 0.2% Zn soils of different metal phases (mean + SE. n=5). Means for the same species and phase, with the same letter, are not significantly different (ANOVA p>0.05). 6 66 Chapter 6 However, mean EDTA-induced uptake by this species from the carbonate salt was not significantly different to mean uptake for the control treatment, due to the anomalously high standard-error-bar for this spec ies, treatment combination (Fig. 6. 1 ). 16000 .......... ..... ..c. 0> 1 2000 ·ID b � � "0 0> Zinc phase � - 8000 - sulphide 0> E phosphate ........- ..... - oxide C -m sulphate c.. - carbonate C 4000 c N o Figure 6.2. Summary: efficacy of natural zinc uptake by Cardaminopsis halleri as a function of metal phase. The ordering of natural uptake can be written as : sulph ide (a) = phosphate (a) < oxide (b) = sulphate (b) = carbonate (b) This is different to the ordering of plant-avai lab le zinc for this species, estimated by extraction with ammonium acetate ( lM, pH 7): su lph ide (a) « phosphate (b) < carbonate (c) < oxide (d) = sulphate (d) There appears to be a large discrepancy between modelled and observed zinc uptake from the sulphide phase. 67 Chapter 6 6.4 Results : Thlaspi caerulescens F or a l l zinc phases except the carbonate phase, treatment with E DT A e ffected no change i n the zinc uptake of Thlaspi caerulescens relative to the control . E DTA caused a sign i ficant decrease in zinc uptake from the carbonate salt ( Fig. 6 . 1 ) . S ummary of the natural response of T caerulescens to each zinc salt allows qualification of the importance of mineral phase on natural uptake (Fig. 6 . 3 ) . 16000 -- C .... C ..c 12000 Cl ·m 3: � -0 Cl Zinc phase � 8000 - sulphide -Cl E - phosphate .... - oxide C - carbonate et! c.. - sulphate C 4000 C N o Figure 6.3. Summary: efficacy of natural zinc uptake by Thlaspi caerulescens as a function of metal phase. The ordering of natural uptake can be written as: sulphide (a) < phosphate (b) < oxide (c) = carbonate (c) = su lphate (c) The above is similar to the ordering of natural uptake for Cardaminopsis haileri. However, the ordering of uptake for Thlaspi caerulescens is more simi lar to the ordering of p lant-availabl e zinc for this spec ies, estimated using ammonium acetate, than was true for C haileri. sulphide (a) « phosphate (b) < carbonate (c) = oxide (c) < su lphate (d ) 68 6.5 Resu lts: total soil zinc Chapter 6 Subsamples of substrate taken from each of the control treatment pots were extracted with hydrochloric acid (5M) at a soil to acid ratio of 1 : 1 0. Analysis of the filtrate showed reasonable agreement of total zinc with the ' designed' concentration of 2 000 mg/kg, although significant differences in zinc concentration between phases were apparent (Table 6. 1 ). The pH of the substrate was independent of mineral phase. Table 6.1 . Total soil zinc and pH for the control treatment soils of each metal phase. Mean zinc values with the same letter are not significantly difference (ANOVA p>O.05). Mean (±sd). Zinc phase Carbonate Sulphate Oxide Phosphate Sulphide Control pH 5.2 4.9 5 .3 5.2 5.3 5.2 Total zinc (mg/kg) 2355 ± 1 76 (a) 251 7 ± 2 14 (ab) 2563 ± 1 90 (b) 2769 ± 1 74 (c) 3039 ± 1 96 (d) Note Note. The total zinc concentration of the control-phase soil was below detection l imits using FAAS and is not reported in this table. 6.6 Results: plant-avai lable zinc The ordering of plant-available zinc, reported for each of the two hyperaccumulator species, was based upon values determined from each of the control treatment soils for each species. These values do not account for differences in metal concentration between each set of soils, and are therefore only relevant to their respective plants. To accurately compare the bioavailablility of each metal phase, and hence the potential solublility of zinc in the rhizosphere as a function of zinc geochemistry, these data for plant-available metal must be corrected for total concentrations (Table 6 .2). Reporting the percentage of the total metal in the soil that is plant-available has made this correction Under the geochemical conditions of this study, the percentage concentration of plant­ available zinc, estimated by ammonium acetate, confirms that bioavailability of zinc is strongly dependent upon mineral phase. 69 Chapter 6 Table 6.2. Percentage zinc that is plant-available (ammonium acetate) from each metal phase. Mean values with the same letter are not statistically different (ANOVA p>0 .05) . Mean (±sd). Zinc phase % extractable zinc Sulph ide 0.7 ± 0 . 1 (a) Phosphate 1 8.0 ± 1 .6 (b) Carbonate 27.2 ± 3 . 0 (c) Oxide 27.8 ± 1 .7 (c) Sulphate 3 1 .0 ± 3 .5 (d) Note. The total and plant-available zinc concentrations of the control-phase soil were below detection l imits by FM and are not reported in this table. 6.7 Discussion - a model for zinc uptake Of the three metals studied in Section A, ZInC is the most phytotoxic. The non­ accumulating and non-tolerant species Brassica juncea would not grow in any of the zinc soils of this study, including the sulphide soil with a very low plant-available metal concentration (2.2 mgIL). This statement explains why no results for uptake by BJuncea have been presented. It appears that the phytotoxicity of zinc would preclude the use of non-accumulator species for phytoextraction, except where the concentration of zinc in the soil was very Iow. There was no increase of zinc uptake by Thlaspi caerulescens in response to EDT A treatment for any metal phase. However, EDT A caused a decrease in the plant zinc concentration taken up by this species growing on the carbonate phase soil . For Cardaminopsis halleri, EDT A effected no response in zinc uptake from the carbonate, sulphate and sulphide phases. EDT A caused an increase in zinc uptake from the phosphate phase, but a decrease in plant zinc from the oxide phase. Based upon the results of this experiment, there is no strong evidence that EDT A­ induced uptake of zinc to a higher concentration, is possible using the hyperaccumulator species Cardaminopsis halleri and Thlaspi caerulescnes. This contradicts the findings in Chapter 4 where EDTA-induced uptake of lead was apparent for T. caerulescens, but is in agreement with the findings of Chapter 5 , where EDT A-induced uptake of cadmium, to a higher concentration relative to natural uptake, was not observed for either hyperaccumulator species. The natural uptake potential for each of these two species varied between each of the different mineral phases. The relationship between this uptake and the plant-available zinc concentration is discussed below. 70 Chapter 6 Relationship between ammonium acetate and the metal phase 1 . Thlaspi caerulescens Natural uptake from each salt agrees well with the predicted ordering of uptake, estimated using ammonium acetate, as shown in Figure 6 . 4 . Each phase plots on a linear relationship (R2=O.82), indicating that for T.caerulescens, ammonium acetate models plant uptake of zinc equally well from each of the 5 metal phases. - C O.05). Before EDT A (mg/kg) After EDTA (mg/kg) Significance B.juncea 0.67±0.28 (0.28-1 .08) n=1 0 0 .62±0 .25 (0. 1 8-0.92) n=1 0 ns B.napus 0.34±0 . 1 5 (0.23-0.67) n=1 0 0 .55±0.23 (0 . 1 8-0.98) n=1 0 s Significance s ns T.caerulescens 84.0±55.2 (27.4-193.4) n=1 0 57. 1 ±44.5 (7. 1 -1 37.0) n=1 0 ns C.hal/eri 27.0±10 .6 ( 1 7.6-45.4) n=8 5 . 1 8±1 . 1 6 (4.2-7 . 1 0) n=4 S Significance s s Application of EDT A to the trial area significantly increased the cadmium uptake by Brassica n apus but not by Bjuncea. EDTA significantly decreased the level of cadmium uptake for Cardaminopsis halleri. EDT A did not adversely affect the health of the Brassica species. There was no significant change in biomass between the before and after treatment harvested material. However, the same cannot be said for the hyperaccumulator species. Application of EDT A to the site effected necrosis in C. halleri and Thlaspi caerulescens. This was an unexpected factor, and forced harvesting of the remaining hyperaccumulator species at the time of after-treatment harvesting of the Brassica species. 93 Chapter 8 Total-soil cadmium concentrations were analysed using the Golden Software graphics package, Surfer, and statistically ' kriged' to generate profile plots of soil cadmium concentrations before and after treatment of the trial area with EDT A, for both the 0-5 cm and 5 - 1 0 cm soil depths (Fig. 8 .6, 8 . 7) . There is a clear difference in the concentration and distribution of cadmium between the before-and-after trial 0-5 cm cadmium contour plots (Fig. 8.6) . Before the trial, the highest cadmium concentrations could be observed in the corner of the plot area away from the shed, on the airstrip side of the hill (cadmium concentration greater than 5 mg/kg) . This is where aircraft have in the past parked to receive a reload of superphosphate material . The affect of the trial was to reduce soil metal concentrations across the plot area to between 3 and 5 mg/kg, a 1 5% reduction in the mean metal concentration. There was, however, a 1 2% increase in the mean cadmium concentration of the plot area effected by the trial for the 5 - 1 0 cm soil depth (Fig. 8 . 7) . 8.4 Discussion Plant metal uptake Equivalent biomass figures per hectare for the Brassica speCIes were calculated by extrapolating the harvested weight recorded from each quadrat to a one-hectare area (Table 8.2). Cardaminopsis halleri and Thlaspi caerulescens were harvested well before maturity and thus realistic biomass figures cannot be extrapolated from the weights of this trial. The value of 2 t/ha for C.halleri has been determined from field observations in northern France. The biomass value of T.caerulescens (2 t/ha) is that reported by McGrath et al. ( 1 993) and is for field growth in the United Kingdom under similar climatic conditions to those experienced in New Zealand. This biomass figure for T.caerulescens is conservative; some researchers believe a biomass for this species of up to 6 t/ha is sustainable (C. Schwartz, pers. commn. 1 999) Based upon the weight of Thlaspi harvested at the end of the trial, a biomass of 1 0 kg/ha for each of the hyperaccumulator species would have been apparent. 94 '0 (/) c: c: o :;:. ro .... -c: Q) u c: o U -0 o .--. 0> � -.... 0> E -- '0 (/) .!: c: 0 :;:. ro .... -c: Q) u c: 0 U -0 0 Cadmium surface plot before tria l : 0-5 cm soil depth Mean Cd = 3.4 mg/kg Cadmium su rface plot after trial : 0-5 cm soi l depth Mean Cd = 2.9 mg/kg 4.0-1 3. �. Chapter 8 Figure 8.6. Surface plot (0-5 cm) showing the change in soil cadmium concentration effected by the trial. 95 Chapter 8 Cadmium surface plot before tria l : 5- 1 0 cm soil depth Mean Cd = 1 . 85 mg/kg - Cl � -- Cl .... # 5 '0 � tJ) � c c ..... <::§'> 0 � � � � - 3.0 c Q) � (J c 0 � (J "U () 2.0 ""<::> Cadmium su rface plot after tria l : 1 .0 5- 1 0 cm soi l depth Mean Cd = 2 . 1 0 mg/kg 0.5 Oi � -- Cl tJ) .!: � c 0 .. ..... <5'> co � � c � Q) (J c � 0 (J "U � () � .... <::§'> Figure 8.7. Surface plot (5-10 cm) showing the change in soil cadmium concentration effected by the trial. 96 Chapter 8 Table 8.2. Biomass and extrapolated uptake figures over a one�hectare area for the p lant species used in this trial . Biomass (tlha) Natural uptake (g I nduced uptake (g Cd/ha) Cd/ha) B.juncea 1 5 1 0 9 B.napus 7 2 4 T.caerulescens 2 1 68 1 1 4 C.halleri 2 54 1 0 Using these biomass figures, the annual cadmium removal potential a crop of each of these plants may offer can be predicted (Table 8 .2). Thlaspi caerulescens has the greatest natural potential for uptake, up to 1 70 g of cadmium per hectare per crop. This is equivalent to 17 years of annual cadmium application to the soil removed in one growing season. The Brassica species could accumulate only a limited amount of metal, approximately one year of annual cadmium application in each growing season, even after the application of EDT A. The annual application rate of cadmium to soils used for these calculations is 1 0 glha/yr (Fergusson and Stewart, 1 992). Effect of soil cadmium concentration on metal uptake To investigate the importance of the total-soil cadmium concentration on metal uptake exhibited by Thlaspi caerulescens, the four values for soil cadmium at the corner of each plot determined from soil samples collected at the start of the trial were averaged, to give a figure for the mean cadmium concentration within each plot. The plant cadmium concentration of T caerulescens was then plotted as a function of these values for soil metal (Fig. 8 .8) . The regression coefficient for the data set is very low (R2=0.00 1 ), thus the cadmium uptake potential of Thlaspi caerulescens is independent of the cadmium concentration in the soil. Cadmium redistribution within the soil The average soil-metal concentration across the plot area before and after the trial was calculated to quantify the change in the soil cadmium balance effected by this 97 Chapter 8 experiment (Table 8 .3) . Soil concentrations were used to calculate the total amount of cadmium within the plot area (Table 8.4), based on each 5 cm unit of soil having a mass of 900 kg over the plot area (20 m2) , and a bulk density (dry) of 1 .2 g/ cm3 • 200 • 1 80 en 160 • :1!: Cl 1 40 .§. 1 20 • -I: I'G 1 00 Q. .E 80 • E ::s 60 � 'E • • "0 • • I'G 40 • (.) • 20 o � 2 2.5 3 3.5 4 4.5 Cadm ium in soil (m g/kg) Figure S.S. Plot of the cadmium concentration in Tha/aspi caerulescens as a function of the cadmium concentration in the soil. Table 8 .3. Average soil-cadmium concentrations across the plot area. Soil depth Cd before trial (mg/kg) Cd after trial (mg/kg) Cd change (mg/kg) 0-5 cm 3.40 2 .93 - 0.47 (14 %) 5-1 0 cm 1 .85 2 . 1 0 + 0.25 (1 2%) Cadmium loss or gain from 0-1 0 cm soil profile 0.23 mg/kg loss Table 8.4. Average total-soi l cadmium levels for the plot area. Soil depth Cd before trial (mg) Cd after trial (mg) Cd change (m g) 0-5 cm 3670 3 1 90 - 480 ( 14%) 5-1 0 cm 2000 2270 · + 270 (1 2%) Cadmium loss or gain from 0-1 0 cm 5011 profile 210 mg net loss 98 Chapter 8 Soil cadmium m ass balance calculations Considering the maximum possible biomass of 2 tlha for Thlaspi caerulescens, and an average metal concentration within the plant of 84 mg/kg, 168 g of cadmium removed per hectare equates to a decrease in the soil cadmium concentration of only 0 .08 mglkg. The ThZaspi plants of this trial had a biomass only a fraction of 2 tlha ( 1 0 kglha) due to harvesting well before maturity, and thus changes in soil cadmium would be difficult to detect in this experiment. Harvesting before maturity was unfortunate, but necessary due to necrosis of the hyperaccumulator species apparent after the application of EDT A. It was initially envisaged that these species would remain at the site until flowering. The large differences between across-plot cadmium distribution for the 0-5 and 5- 1 0 cm soil profiles can therefore be attributed to EDTA redistribution of metal within the soil profile. Total plant-metal uptake across the plot area was 6.92 mg cadmium for the non­ accumulator Brassica species, and approx . 0.40 mg of cadmium for the hyperaccumulator species. Table 4 shows a net loss of 2 1 0 mg cadmium from the 0- 1 0 cm soil depth, but only 7 .3 mg of this cadmium loss can b e accounted for by plant uptake. Assuming all of this 7 .3 mg of cadmium was removed from the top 5 cm of soil, then more than 470 mg of metal was leached from this depth. Redistribution of 270 mg of this cadmium can be accounted for in the 5- 1 0 cm depth. The remaining 200 mg of cadmium, or 4 % of the cadmium initially present within the top 1 0 cm of the trial area, was leached to below 1 0 cm soil depth. Zinc accumulation by the hyperaccumulator species Zinc was accumulated by the hyperaccumulator specIes Thlaspi caerulescens and Cardaminopsis halleri, but was not 'hyperaccumulated' (Fig. 8 .9). Mean accumulation by T caerulescens was 6 1 0 mg/kg ( 1 70 mg/kg - 1 900 mglkg range) before treatment of the site with EDTA, although no significant change in the foliar-zinc concentration was observed as a result of EDT A treatment. Similarly there was no significant difference in zinc accumulation between the two species. Both species are known hyperaccumulators of zinc. 99 Chapter 8 McGrath ( 1 998) stated that zmc hyperaccumulation by Thlaspi caerulescens is a 'physiological requirement' and that the minimum shoot concentration necessary for growth was in the order of 1 000 mglkg. However, the low level of accumulation by T. caerulescens in this trial was within the 'normal-range' of zinc accumulation by non­ accumulator species. The plants were very healthy (until treatment with EDTA) and successfully hyperaccumulated cadmium. The evidence from the Wairarapa trial is in contadiction to the work of McGrath, and suggests that yet again our understanding of the mechanisms underlying hyperaccumulation is relatively poor. � � C � .5 § � C � § rEi 1600 1200 800 400 o Treatment Before EDTA _ After EDTA Thlaspi caerulescens Cardaminopsis halleri Figure 8.9. Zinc accumulation by T.caerulescens and C.ha/leri from the Wairarapa trial site. The observation that Thlaspi caerulescens did not hyperaccumulate zinc is an important one. A concern sometimes voiced with regard to phytoremediation for pastoral cadmium, is that an operation using hyperaccumulator species would also remove trace concentrations of zinc from the soil, leading to a soil deficiency in the concentration of this important trace element. The results from this trial suggest that this is not an important concern. 1 00 8.5 Summary - a practical application for phytoextraction Chapter 8 Phytoremediation has been proposed as a solution for areas where anthropogenic pollution has contaminated natural environments. The example illustrated in this chapter is of elevated cadmium levels in New Zealand pastoral soils subjected to superphosphate fertilisation. A key drawback of phytoremediation technology is the long time frame over which decontamination will occur. In this trial a 2 tlha crop of Thlaspi caerulescens would only decrease soil metal levels by approximately 0.08 mglkg in each growing season. An alternative option is to view this technology as a management tool for areas with low levels of soil contamination. If cadmium is applied to soil at a rate of 1 0 glha/yr, then natural hyperaccumulation effected by a crop of Thlaspi caerulescens, which in this trial could remove up to 1 68 g of Cdlha in one season (assuming a biomass of 2 tlha is achieved), would remove the equivalent of 17 years of annual fertiliser application. Thlaspi caerulescens in this scenario could be planted periodically in a crop rotation cycle, to manage effectively an increasing cadmium load to the soil, and thus maintain cadmium at an arbitrarily set baseline level. At the trial site T caerulescens performed better than Cardaminopsis halleri. This was due to both superior cadmium uptake potential and the preference of this species for an open and dry soil environment, typified by the Wairarapa, in contrast to the sheltered and moist environment favoured by Cardaminopsis halleri (A.Deram, pers. commn . , 1 999). There was no relationship at this site between total soil cadmium and total plant cadmium. Total metal uptake for hyperaccumulator species appears to be independent of the metal concentration in the soil. It seems likely that Thlapsi caerulescens will hyperaccumulate cadmium equally well from areas of land that have a lower metal loading than the 'hot-spot' used in this trial, although this theory has yet to be tested. It also appears that Thlaspi caerulescens will not significantly and adversely decrease the zinc concentration of the surface soil, to the extent where zinc fertilisation would be necessary subsequent to removal of cadmium. This species did not hyperaccumulate ZInC. 1 0 1 Chapter 8 Hyperaccumulator species do not respond well to vegetative competition. If an area of land were to be managed in this way, an application of herbicide before seeding would be necessary. Once the crop reached maturity, harvesting would yield a low-volume of material, concentrated in cadmium, that could be disposed of or recycled. Growth of other species in subsequent seasons would effectively outcompete the foreign hyperaccumulator plants; the potential of hyperaccumulator species to become 'weeds' i s therefor low. Induced hyperaccumulation, effected by the chelating agent EDT A, did not result in high levels of cadmium uptake by the non-accumulating Brassica species. The amount of metal ' naturally' removed by the hyperaccumulator species remained higher. In this trial, the effect of EDTA was to l each approximately 4% of the cadmium initially present in the surface soil (0- 1 0 cm depth) to below 1 0 cm. EDT A did successfully lower the cadmium concentration of the surface soil and this in itself was a useful result. Soil leaching of cadmium using chelating agents has been suggested as a solution for soils with high cadmium surface concentrations. However, it is likely that other more important trace metals (e.g. Co, Cu, and Zn) will also have been complexed and thus leached down the soil profile, leading to a deficiency in the nutrient balance of the surface soil (Tejowulan and Hendershot, 1 998) . 8.6 Conclusion The loading of cadmium to agricultural land will remain a problem to both New Zealand and worldwide agriculture, as long as the heavy metal remains in phosphatic fertilisers. Removal of this cadmium is expensive, and although desirable, may not be feasible. Phytoremediation using hyperaccumulator species, in particular the species Th/aspi caeru/escens, may be an effective management tool for cadmium in the New Zealand pastoral environment. The equivalent of 1 7 years of annual cadmium application to agricultural land could be removed in one season assuming an attainable biomass of 2 t/ha. Using this technology, the cadmium concentration of New Zealand pastoral soils could be maintained at the current level which is safe for intensive agriculture, while still allowing for annual application of superphosphate fertiliser. 1 02 Chapter 8 A chapter summary is presented as Figure 8 . 1 0. Overleaf: Figure 8.10. Phytoremediation of cadmium from New Zealand soils: a poster summary of a practical application. 1 03 PHYTOREMEDIATION OF CADMIUM FROM NEW ZEALAND SOILS Massey University Chris Anderson , Robert Brooks , Bob Stewart and Robyn Simcock* Soil and Earth Sciences , Massey Un iversity, Palmerston North *Landcare Research Ltd . , Palmerston North Introduction I n New Zealand heavy metal contamination can occur in pastoral land due to the association of Cd with superphosphate fertiliser. Phytoremediation is a developing technology that uses plants to remove metals from an area of contaminated land. These plants take-up very high levels of metal, either naturally or with the assistance of chemicals applied to soil (eg EDTA). Objective To test the efficacy of both natural and induced hyperaccumulation to remove Cd from a contaminated field site. Materials and Method A small site contaminated with Cd ( 1 -6 mglkg), was identified on a farm 24 km SE of Masterton. The site was adjacent to a superphosphate fertiliser shed and loading zone for aerial top-dressing planes. Soil (0-1 5 cm) was sampled at regular spacing over the area before treatment A 2 by 1 0 metre area was divided into 20 plots, and each plot divided into 4 quadrats (Fig. 1 ). In November 1 998 quad rats were seeded with one of two non-accumulating species, Brassica juncea and Brassica napus (Fig. 2) or two Cd hyperaccumulating species, Cardaminopsis halleri and rhlaspi caerulescens. I n February 1 999, biomass from half of all quadrats was harvested. EDTA was then applied at the equivalent rate of 1 Vha to the trial area to induce additional Cd uptake. After 3 weeks the remaining biomass was harvested and soil sampled. Plant and soil samples were analysed for total Cd metal concentrations. Results and Discussion The two hyperaccumulating species took up high concentrations of Cd (Table 1 ). Tcaerulescens had a significantly higher concentration of Cd than Challeri. Non-hyperaccumulating species showed limited Cd uptake both before and after EDTA treatment B.juncea contained a significantly higher Cd concentration than B. napus. After the trial, Cd was either removed or mobilised deeper than 5an within the soil profile (Fig. 3). Natural uptllke EOT A induced Biomas. (tlhal Natural uptake (mglkg) uptake (mglkg) (g Cdlha) Bjuncea 0.7 . 0.3 (A) 0.6 . 0.3 (A) 15 10 8.nllpu5 0.3 .0.2 (A) 0.6 . 0.2 (8) 7 2 T.caerulescans 84 . 55 lA) 57 . 45 lA) 2' 166 C.halleri 27 . 1 1 lA) 5 . 1 (8) 2' 54 Fergusson and Stewart ( 1 992) report a Cd application rate of 1 0 g/halyr can be expected on a well managed farm. In this trial B.juncea had a mean biomass of 1 5 Vha and would remove about 10 g of Cd per hectare. Tcaerulescens has a biomass of about 2 t/ha (Baker et al. , 1 994). A crop of this hyperaccumulator could remove up to 168 g of Cd/ha. Conclusions The hyperaccumulating species Tcaerulescens removed the equivalent of more than 1 5 years of accumulated Cd i n three months. Phytoremediation could prove a useful management tool to reduce Cd contamination in pastoral soils. There is little point in inducing hyperaccumulation for Cd in plants of higher biomass, due to the associated costs and limited increase in uptake. References Baker, A J M . McGrath, S P , Sldoh, C M.D , and Reeves, R D . (1994) The POSSibility of In Situ heavy metal decontamination of polluted soils USing ClopS of metal�mulaMg plants Resources, ConseN8tjon and Recycling, 1 1 41..49 Fergusson. J E ,and Stewart, C . (1992) The transport of 81r-bOme trace elements c:opper, lead, cadmllJm. llnc and manganese from a CIty Inta rvraI areas SCIence of the Total Environment, 121 247269 ,c Cd sdrface plot before tnal , d � Cd 3urface plot lfter tnal Acknowledgments The aulhol'!!t gratefully acknowledge Mr Don Adams of BlaH1og.e Watrarapa far provIdIng a field tnal-slte and for hts assIstance ,n runnrng the (flal We also gratefully acXnowledge AGMAROT of New Zealand fOf the award at a OX;taral scholarship to support the sen.a autl'lOr SECTION A CONCLUSION TO SECTION A The integrated geochemical model for Cd, Pb and Zn presented in Chapter 7 described many of the results observed for metal uptake from the Tui and Auby contaminated substrates. The results from the Wairarapa field trial described in Chapter 8 present a test of this model in a third environment. Hyperaccumulator species Neither Thlaspi caerulescens nor Cardaminopsis halleri hyperaccumulated cadmium from the Wairarapa field site (using the criterion concentration of 1 00 mg/kg), although uptake was very high. The relative uptake efficacy of these two species from cadmium present in the soil as phosphate salt was in agreement with the cadmium-uptake model. Mean cadmium uptake by T.caerulescens was greater than that for C.halleri, and for both species, EDTA decreased the foliar-cadmium concentration. Uptake of zinc by these species from the Wairarapa site was in similar agreement with the zinc uptake model . Hyperaccumulation was not effected by either species from the zinc phosphate soil of Chapter 6, although induced-uptake of zinc to a higher concentration by C.halleri could have been expected. The plant-available (ammonium acetate) concentration of cadmium for the Wiararapa site was not reported in chapter 8, however, it is worth mentioning here. The mean concentration of cadmium across the plot area was 3 . 1 ng/mL which is lower than the concentration of plant available cadmium from the Tui substrate of Chapter 3 (9. 5 ng/mL). Yet uptake was higher from the Wairarapa substrate. The plot of the cadmium concentration in Thlaspi caerulescens as a function of the plant-available cadmium concentration in the soil, presented in Chapter 5 (Fig. 5 .8), showed anomalous uptake from the sulphide phase relative to the other forms of soil cadmium. Clearly, uptake of cadmium from the sulphide metal phase is not the same as uptake from a phosphate metal phase, and this inequality is not modelled by ammonium acetate. 1 05 SECTION A Non-accumulator species The geochemical model for cadmium uptake predicts that EDT A could effect a large increase in metal uptake from the phosphate phase by the non-accumulator species Brassica juncea. However, no such increase in uptake was observed for this species in the Wairarapa trial, although the cadmium concentration in B. napus was doubled. The soil mass balance presented in Chapter 8 shows clearly that EDT A induced a large increase in soluble cadmium that was subsequently leached below 5 cm soil depth. The question is 'why was this soluble cadmium not taken up by the plant?' Blaylock et al. ( 1 997) showed a strong pH dependency for EDTA-induced lead uptake. At pH 5 .0, lead-contaminated soil (600 mg/kg) supported EDTA-induced uptake to almost 4 000 mglkg by Brassica juncea. This uptake decreased to approximately 500 mg/kg for a pH of 7.0. The authors of this study justified the decrease in uptake through retention of lead within the root cell wall. EDT A in combination with low pH effectively prevents cell-wall retention of lead, freeing the metal for transport in the xylem to the shoots. Subsequent necrosis of these shoots shows that the non­ accumulator plant has no mechanism to sequester the metal in a non-toxic form, i .e . , it is not an accumulator species. Root samples were never analysed for the experiments described in Section A and thus this hypothesis cannot be tested. The design of experiments described in the previous chapters did not account for root uptake; the plant organs of interest were the aerial portions, representing biomass that could be easily harvested for a phytoextraction operation. However, if cadmium behaves similarly to lead, sequestering of metal in the roots of the Brassica species could help account for the discrepancy between modelled and observed data. Sequestering of cadmium in the roots may also account for some of the cadmium lost below 1 0 cm depth for the Wairarapa soil profile, although the conclusion that cadmium was leached as a function of EDTA treatment remains true. Any roots sampled with the soil were separated at the time of grinding. The above conclusion may also be valid for the poor EDT A-induced uptake results from the Tui tailings, and to some extent from the Auby soil. However, pH dependency does 1 06 SECTION A not explain why the high concentrations of plant-available lead modelled by ammonium acetate did not translate into high natural uptake. Active exclusion of lead uptake appears to be true for 'natural (non-induced) uptake ' . Could an active exclusion mechanism for a metal-EDT A complex be manifest at high pH? Maybe steric factors preclude the uptake of a bulky metal-chelate complex at this same high pH. Garvan ( 1964) showed the foHowing pH-dependency of the EDTA chelate: OH­ � (HEDTA) 3- pH 8. 1 (EDTA)4- pH 1 2.4 Presumably as the shape (conformation) of the chelate changes with the protonation or deprotonation of binding sites (a function of pH), the size of any chelate-metal complex will also change. Blaylock et al. ( 1997) and Vassil et af. ( 1998) suggested that the entire metal-EDT A complex is accumulated by a plant and transported to the shoots. Another theory is that the metal-EDT A complex is dissociated at the soil-root interface and the metal i s subsequently taken up as a free-metal ion. If the first of these two theories is correct, then steric factors should be an important aspect of uptake. Section A of this thesis has shown there is a complex interaction between geochemistry and species-specific plant uptake (natural and induced). At constant pH, the mineral phase strongly influences both the ammonium acetate concentration of plant-available metal i n the soil and the concentration of metal in the plant. The relationship between these two variables is not equal for every environment. I am unsure how variation of pH will, in practice, affect the geochemical uptake model of this study but the effect of pH remains an area for future research. The assumption, for example, that plant-available lead, estimated by ammonium acetate, can be removed from a contaminated soil using EDT A-induced hyperaccumulation, can be a dangerous one. Without specific knowledge of the soil geochemistry, the choice of plant species, the choice of extractant to model plant-available metal, and the choice of chemical to induce metal uptake may lead to less than optimal results for phytoextraction. 1 07 SECTION B SECTION B - PHYTOEXTRACTION FOR N ICKEL AND GOLD Publications a rising from Section B : Anderson, C .W.N. , Brooks, RR, Stewart, RB. and Simcock, R, 1 998. Harvesting a crop of gold i n plants. Nature 395: 553-554. Anderson, C.W.N . , Brooks, RR, Stewart, RB. and S imcock, R , 1 999. Gold uptake by plants. Gold Bulletin 32: 48-51 , 58. Anderson , C . , Brooks, R , Stewart, R, Simcock, R and Robinson B . , 1 999. The phytoremediation and phytomining of heavy metals. In Proceedings, PACRIM'99 Congress (Ed G .Weber) pp 1 27-1 35 (Australasian I nstitute of Mining and Metatlurgy, Victoria). Anderson, C.W.N. , Brooks, RR, Chiarucci , A., LaCoste,C.J . , Leblanc, M . , Robinson, B .H . , Simcock, R and Stewart, R , 2000. Phytomining for thallium, nickel and gold . Journal of Geochernical Exploration, in press. Anderson, C.W.N. , Brooks, RR, Stewart, RB. and Simcock, R, 2000. Seed germination in Hybanthus floribundus F. Muel l . Australian Journal of Botany, i n press. Brooks, R , Anderson, C . , Stewart, R and Robinson , B., 1 999. Phytomining: growing a crop of metal . Biologist 46: 201 -205. Phytoextraction for the metals nickel and gold, unlike that for cadmium, lead and zinc, has potential for both phytoremediation and phytomining. In fact, it is hard to conceive a scenario where gold would exist as a contaminant necessitating site remediation. Fewer detailed studies have been published on the phytoextraction of nickel than for the metals covered in section A, perhaps because of the economic sensitivity inherent to research into nickel uptake. Nickel phytoextraction technology has been patented for uptake effected by Alyssum species (Chaney et al. , 1 998), and the patentees believe they are close to commercial success (J.S .Angle, pers commn . 1 999). There are two reasons why phytoextraction for nickel holds such great potential. Firstly, more than 300 species are known to hyperaccumulate this metal (Chapter 1 ) , and secondly, some of these species yield a very large annual biomass. Berkheya coddii, a nickel hyperaccumulator native to South Africa, for example, has a maximum known metal concentration of over 1 % CDW) and a maximum possible biomass of 22 t/ha/year. 1 08 SECTION B Gold is a very recent addition to the list of metals for which phytoextraction is possible, and is effected by induced hyperaccumulation. This thesis reports the first detailed study of such precious metal uptake. Chapter 8 presents data from a field trial conducted on an area of ' serpentine soil' located in the central North Island of New Zealand. While phytoremediation of this site is not necessary, and any phytomining operation would be of such small scale that economic return seems unlikely, the area is a test case to examine the potential for phytoremediation in an environment foreign to the several hyperaccumulator species used. Chapter 9 describes a major constraint imposed on the successful implementation of phytoextraction technology. Phytoextraction for nickel holds great potential in many parts of Western Australia, however, only native species may be used. This excludes the use of A lyssum bertolonii and Berkheya coddii. Instead, the only nickel hyperaccumulator native to Western Australia, Hybanth us floribundus, must be used. Until the study described in Chapter 9 this species had proven difficult to germinate from seed. Phytoextraction of gold is described in Chapter 1 0, the mechanism for induced uptake and the geochemical rational behind this phenomenon. Finally, Chapter 1 1 presents various practical scenarios that illustrate environments where phytoextraction for nickel and/or gold may prove environmentally and/or economically viable. 109 Chapter 9 Chapter 9 - A New Zealand Field Trial for N ickel Phytoextraction The purpose of the project described in this chapter, was to field test the viabil ity o f both phytoremediation and phytomining on a n area o f ' quasi' serpentine soil , the terraced, ' wet-serp' byproduct material of a serpentine rock q uarry operation located near Piopio, southwest of Te Kuiti, central North I sland, New Zealand (Fig 9. 1 ). This substrate represents a foreign environment for the hyperaccumulator species that were used. The concept of phytoremediation is well accepted, but the hard question of ' w i l l the technology b e fi e l d effective i n many different metal-rich environments ' , has not been well addressed. If phytoextraction could be successfully effected in this environment, then the practical future of the technology for nickel could be better assured. Chapter 8 has not addressed the potential importance of specific n ickel mineral phases on plant uptake . As there is little reason to induce the hyperaccumu lation o f nickel due t o the large number o f high-biomass n ickel hyperaccumulator species, such a study is l ess relevant for this metal than for lead, zinc and cadmium. North I s land Wellington N + 4- SH 3 to New Plymouth 10 km Te Kuiti Piopio Mapiu \ Serpentine Quarry Figure 9.1 . Map of the North Island of New Zealand showing in detail the location of the nickel phytoextraction trial site. 1 1 0 9. 1 Introduction 9 Nickel is a heavy metal with the potential to have an adverse affect on life . While the metal is an essential trace ingredient in many of the living processes of higher plants (Marschner, 1 995), at higher concentration ( 1 0 - 47 mg!kg) , plant species can suffer from chlorosis, stunted root growth or necrosis (Uren, 1 992; Kriimer et al. , 1 997). Excessive exposure to nickel compounds can result in damage to the lungs, brain, l iver, kidneys, adrenal glands, spleen and pancreas of mammals. In extreme cases, ingestion of nickel compounds has been implicated in effecting fatal cardiac arrest (Knight et al. , 1 997). Nickel, like most other heavy metals, is a cumulative poison and thus low levels of anthropogenic contamination over a time frame of many years can result in a serious toxicity problem. Such anthropogenic contamination occurs in pastoral land to which sewage sludge has been applied. Nickel contamination of European soils above the EU maximum permissible guideline level of 75 mg!kg is common (McGrath and Smith, 1 990). Areas of anthropogenic nickel contamination are not as widespread as those for the more common heavy metals Cd, Pb and Zn. Except for the contamination described above, hot-spots for nickel are generally associated with the mining and processing of nickel ores. Nicks and Chambers, ( 1 994, 1 995, 1 998) and Robinson et al. ( 1 997a,b) have studied the nickel uptake potential of various plant species to assess their potential for phytoremediation and phytomining. The work of these two research groups is reviewed here. The initial study of Nicks and Chambers A pioneering study to examine the feasibility of applied nickel phytoextraction, was commissioned by the US Bureau of Mines (Reno, Nevada) in 1 993. The trial used a natural stand of the nickel hyperaccumulator Streptanthus polygaloides that was growing on ultramafic soil at Red Hills, California (Nicks and Cambers, 1 994, 1 995, 1 998). After harvesting, the plants contained, under optimal agronomic conditions, a 1 1 1 Chapter 9 mean nickel concentration of 5 300 mg/kg (0.53 %) DW while the soil contained only 0 .35 % nickel. The maximum known nickel concentration of this particular species is 1 .48 % (Reeves, et al., 1 98 1 ). Nicks and Chambers ( 1 998) proposed that a net return of US$5 1 3/ha to the grower could be achieved, assuming the following: • selective breeding of the genotypes to yield a nickel concentration of 1 % and biomass of 1 0 tlha DW, • a world nickel price of US$7.65/kg (the price at the time of the trial - although the price now, Feb 2000, is closer to US$ IOlkg), • the harnessing of 25 % of the energy produced from combustion of the biomass to produce electricity that could be sold to a local power grid, and • a return to the grower of 50 % of the gross metal yield plus the value of the energy yield. The possible metal recovery from a 1 0 tlha crop with a nickel content of 1 % is 1 00 kg. Incineration of this biomass would yield approximately 500 kg/ha of bio-ore containing 20 % nickel. Subsequent studies Subsequent to the Streptanthus polygaloides investigation, Robinson et al ( 1 997a) conducted experiments in Italy, using the hyperaccumulator A lyssum bertolonii, endemic to the ultramafic flora of the Tuscan Hills. This species can sustain, after moderate fertilisation, a biomass of 12 t/ha/yr with a mean nickel concentration of 0.8% dry matter. With these metal concentration and biomass figures, a nickel yield of 96 kg/ha can be realised, comparable to the yield of S.polygaloides in the above California project. Trials have also been conducted using the South African nickel hyperaccumulator Berkheya coddii (Robinson et aI, 1 997b). This plant in the wild has a maximum reported nickel content 1 .7% (Morrey et ai, 1 992). Based on New Zealand plot trials, 1 1 2 Chapter 9 Robinson et al. ( l 997b) reported a sustainable biomass for B. coddii, after moderate fertilisation, of up to 22 tlhalyr, and a conservative mean nickel concentration of 0 .5% dry weight. Using these figures, a crop of B. coddii would remove 1 1 0 kg of nickel in each hectare, worth $ 1 1 00 at a current world nickel price of approx US$ 1 0.00lkg. Taking into account the energy derived from combustion and a return of 50% of the nickel value to the grower, each crop would be worth approximately US$660 a hectare. The Ultramafic Belt of New Zealand The basement rock for part of the New Zealand continental landmass is an 'ultramafic ' igneous rock enriched in iron, nickel and magnesium relative to sodium, potassium and silica. The most common product resulting from the alteration of ultramafic rocks by heat and pressure in a hydrous environment is serpentine, and the name given the subsequent altered rock type is serpentinite (Gage, 1 980). Ultramafic rocks occur with a surficial exposure in Northwest Nelson for approximately 1 50 km north of the Alpine Fault, and then again south of the fault for another 1 85 km in Otago and Southland. Dextral movement on the alpine fault has displaced these two segments of what was once a continuous ultramafic assemblage some 500 km relative to each other. Gravity anomalies apparent on either side of the contemporary ultramafic exposure suggests that it extends hundreds of kilometres further, in both directions, which would explain the otherwise anomalous presence of serpentine rock in the North Island Tertiary sediments at Piopio and North Cape. At these two sites it appears that a s lice of ultramafic rock has been fault-emplaced into the younger strata from the ultramafic belt constituting the present day basement (Gage, 1 980). Geology of the Piopio serpentine exposure The surface exposure of serpentine rocks at Piopio is the result of serpentine basement intrusion up the plane of the Waipa fault. This fault separates the shallow, shelf-facies Mesozoic sediments from the deeper, marginal-facies Mesozoic strata. The intrusion of serpentinite has tipped the proximal Oligocene limestone to a near vertical orientation. 1 1 3 Chapter 9 At l east two periods of intrusion can be recognised due to the fact that the limestone contains serpentinite c1asts (Kear 1 960, 1967). Serpentinite rock is quarried at the Piopio site, crushed, and added to superphosphate fertiliser as a 'flowing agent' for aerial topdressing. Low-grade serpentinite material, mixed with the surrounding limestone and sandstone/siltstone sediments (papa) not suitable for crushing, is stockpiled as 'wet-serp ' on terraces adjacent to the quarry (Fig. 9 .2). Figure 9.2 . North facing perspective of the Piopio serpentine quarry. Serpentinite rock is the dark-coloured material, running through the centre of the photograph, fault emplaced adjacent to the pale-coloured limestone. The wet-serp terraces lie to the left of this photograph. 9.2 Materials and Methods Design of the experimental area An 1 1 metre square area on the top terrace of the 'wet serp ' stockpile was bul ldozed in September 1 997 and fenced to keep feral sheep, goats and rabbits out of the trial . Large 1 14 9 rubble clasts of limestone and serpentinite were removed by hand. The area was divided into 5 blocks, each of 2 metres by 1 0 metres, and each containing 20 individual 1 m2 plots (Fig. 9.3). N + / Limestone / J ! ! Serp�ntinite rock Trial site . x Each block divided into 20 plots Figure 9.3. Schematic plan of the Plopio experimental trial set up. X marks the location where the photograph in Figure 92 was taken from. One block was planted with seedlings of each of Alyssum bertolonii, Alyssum malacitanum and Berkheya coddii (Fig. 9.4). Due to problems of seed germination for an Australian hyperaccumulator species that was to be used for this trial (Chapter 10), the remaining 2 blocks were not planted during the trial , but have subsequently been planted with B.coddii ( 1 999). Each plot was planted with 5 individual plants. Two replicates of 1 0 fertiliser treatments were applied within each block: NoPo, NoPj , NoPz, N 1PO, NIP!, N 1Pz, N2PO, N2Pl , N2P2 and S. Nitrogen was applied in the form of calcium ammonium nitrate (27%N) where N I represents an application rate of 50 kg Nlha. Phosphorus was applied in the form of superphosphate (9% effective P) where P i represents an application rate of 20 kg Plha. Sulphur was applied a s elemental 'flowers 1 1 5 Chapter 9 of sulphur' at a rate of 40 kglha. Fertiliser was first applied to each block on the 1 6th September 1 997. This application was repeated on 14th March 1 998. A final application rate, double the concentration of that described above, was applied to the Berkheya coddii block only on 26th January 1 999. Figure 9.4. Photograph of the Piopio trial site before harvest of B.coddii in May 1 999, looking towards the northeastern corner of the site. Wooden pegs separate each plot. The block in the foreground is newly planted B.coddii. The block in the middle of the photograph shows the B.coddii plants analysed for this chapter. A small amount of the Piopio wet-serp was collected and brought back to Massey University. Two outside plots were filled with this material, each with dimensions of 1 by 1 by 0 . 1 5 metres. The first plot contained 'pure' wet-serp, the second a 1 : 1 mixture of wet-serp and sieved bark. Both plots were planted with Berkheya coddii. A nalytical procedure Plant samples were collected, rinsed in deionised water and dried at 60°C to constant weight. A lyssum samples (ca. 200 mg) were weighed directly into borosilicate test tubes, ashed at 550°C overnight and the residue taken up in hot hydrochloric acid ( 1 0 1 1 6 Chapter 9 mL of 2M) before analysis by F AAS . Berkheya coddii was first ground before subsamples (ca. 200 mg) were treated using the same procedure. Ground subsamples of Berkheya coddii were analysed for carbon and nitrogen using a Leco eNS furnace. Phosphorus was analysed using the Kjeldahl digest method of Blakemore et al. ( 1987). A soil:water suspension ratio of 1 :2 . 5 was used to measure pH. Three replicate cores (0-8cm) were sampled from each of the Berkheya coddii plots at the conclusion of the trial. These cores were then mixed to generate a composite sample from each plot, air dried and sieved with a nylon sieve « 500�m). Sieved subsamples ( l g) were digested in aqua regia before analysis by FAAS. Plant-available nickel was estimated by extraction with ammonium acetate ( 1 M) at a soil: liquid ratio of 1 : 1 0 (Table 9. 1 ) . Table 9.1 . Piopio wet-serp geochemical properties Mg (%) 23 Zn (mg/kg) 48 Fe (%) 4.8 Cu (mg/kg) 12 Mn (mg/kg) 850 Co (mg/kg) 96 C(%) 0.47 N (%) 0.02 Notes (1 ) metal analyses performed by FMS (2) Ext-Ni is extraction by 1 M NH40Ac at a 10 : 1 l iquid soil ratio (3) C and N analyses performed by Leco furnace. 9.3 Results A lyssum malacitanum Cr (mg/kg) 9 10 N i (mg/kg) 1 880 Ext-Ni (ng/L) 2.2 pH 8.4 This species did not grow well in the wet-serp environment and did not appear to increase in size from the initial planting. Samples collected during the trial showed an average nickel concentration of approximately 1 400 mg/kg across the block. However, due to the poor growth, no harvest was made at the end of trial. 1 17 Chapter 9 Alyssum bertolonii Samples of Alyssum bertolonii were harvested from each of the plots in January 1 999 and analysed for nickel (Table 9.2; Fig. 9 .5) . The mean concentration of plant nickel across the 20 plots was 4 1 60 mglkg. Table 9.2. Mean Alyssum berlolonii nickel concentration for each ferti liser treatment (n=2) . Fertil iser treatment NoPo NoP, NOP2 N , Po N 1P, N 1P2 N2PO N2P1 N2P2 S Mean nickel in plant (mg/kg) 4 1 30 (± 360) 2 980 (± 500) 4 660 (± 640) 3 580 (± 900) 5 095 (± 350) 4 900 (± 1 600) 2 690 (note) 4 220 (±1 0 1 0) 4 630 (± 290) 4 730 (± 990) Note: only one plot was harvested for the N2PO treatment. Figure 9.5. Flowering specimen of Alyssum berto/onii growing at the Piopio trial site. Photograph taken December 1 999. 1 1 8 Chapter 9 Application of N did not significantly change the nickel concentration m Alyssum bertolonii. However, there was evidence for a significant increase in plant nickel with increasing application of P (Fig. 9 .6) . There was no significant evidence for any interactive effect of N and P fertiliser on the nickel concentration in the plant. Application of sulphur did not result in higher metal uptake at the application rate used. 0> ...liI:: C» E ......... a a 6000- b ab a - 4000- � a; 4000-...liI:: a o 'c o 'c NO N 1 N2 PO P 1 P2 Figure 9.6. Nickel concentration in Alyssum bertolonii as a function o f the soil fertiliser treatments. Error bars represent standard error. Means with the same letter are not significantly different (ANOVA p>O.05). Berkheya coddii The above-ground biomass of Berkheya coddjj was harvested from each of the plots in May 1 999 (Fig. 9 .7), weighed and analysed for nickel (Table 9 . 3), cobalt, nitrogen, carbon and phosphorus. The mean concentration of plant nickel across the 20 plots was 1 400 mg/kg. There was a significant decrease in the nickel concentration in Berkheya coddii with addition ofN fertiliser, but no significant change in nickel through the addition ofP (Fig. 9 . 8) . Again, there was no significant evidence for any interactive N and P fertiliser affect on the plant nickel concentration. Application of sulphur to the plots did not result in higher metal uptake. 1 1 9 Chapter 9 Table 9.3. Mean 8erkheya coddii n ickel concentration for each fertil iser treatment. Fertil iser treatment NoPo NOP1 NOP2 N1PO N1P1 N 1P2 N2PO N2P1 N2P2 S mean nickel in plant (mg/kg) 2 2 1 0 (± 95) 1 200 (± 1 50) 1 780 (± 730) 1 1 35 (± 4) 1 520 (± 650) 900 (± 220) 1 305 (± 1 20) 1 205 (± 1 1 5) 8 1 0 (± 250) 1 590 (± 1 75) Figure 9.7. The author with a flowering stem of 8erkheya coddii. growing at the Piopio site. Photograph taken May 1 999. 1 20 a n I I NO b b Ni NZ a a a '-:g 8 a:i PO P1 P2 Figure 9.8. Nickel concentration in Berkheya coddii as a function of the soil fertiliser treatments. Error bars represent standard error. Means with the same letter are not significantly different (ANOVA p>O.05), To further elucidate any possible relationships between plant-nickel concentration and plant nutrient status, samples of Berkheya coddii were analysed for C, N and P. The plot of nickel as a function of nitrogen (Fig. 9.9) shows no relationship between these elements within the plant (R2=O.03). These data do not support a decrease in plant nickel with increasing application of N fertiliser. Similarly, there was no intraplant relationship between nickel and phosphorus (R2=O.009) and nickel and carbon (R2=O.007). 2.5 1 .5 • • • .. • • • • '* • • • • • • • 2 - � • • • c: 4' Cl R2=O.03 !; :le 0.5 o�------�--------�--------�-------+--------� o 500 1000 1500 2000 2500 Nickel (mglkg) Figure 9.9. Nickel concentration in Berkheya coddii as a function of the plant nitrogen concentration. 121 Chapter 9 Cobalt Samples of Berkheya coddii prepared for nickel analysis were also analysed for cobalt. The mean concentration of plant cobalt across the block was 1 1 0 mg/kg. Cobalt was correlated with nickel (Fig. 9 . 1 0) but was not, however, hyperaccumulated. The hyperaccumulation criterion for cobalt is accumulation greater than 1 000 mg/kg DW, and the range in 'normal' plants is 0 .05 - 50 mg/kg. 200 • 180 • 1 60 140 1 20 Oi "" 0, 1 00 .§. � 80 .Q 0 u 60 • • 40 20 0 0 500 1 000 1 500 2000 2500 Nickel (mglkg) Figure 9.10. Plot of the cobalt concentration in 8erkheya coddii as a function of the nickel concentration. Native species Several weed species have naturally colonised parts of the wet-serp terraces. Samples of these were collected from within the trial area and analysed for nickel uptake. The highest concentration of nickel observed in any of these species was 53 mglkg dry weight in a thistle plant. 1 22 Chapter 9 Soil samples Analysis of the substrate samples from each of the Berkheya coddii plots showed no significant difference in soil nickel concentration relative to control samples collected outside the plot area. 9.4 Discussion The three plant species used in this trial all failed to realise the high biomass yields reported by Robinson 1 997 and Robinson et al. ( 1997a,b). Biomass calculations were not made for the Alyssum species, although individual Berkheya coddii plants were weighed. The total dry mass of B. coddii plant material harvested from the plot area was 480 g, and represents an average mass of 20 g per individual plant. Robinson et al. ( 1 997b ) reported a fertilised biomass for individual B. coddii plants of 1 25 g (22 tlha 10) and an increase in biomass with increasing fertilisation. There was no evidence in this field trial of increased biomass for the plots with higher fertiliser loadings. The metal concentration of the harvested plants was also relatively low. Robinson et al. ( 1 997a) reported mean nickel concentrations of up to 0.77% for A lyssum bertolonii grown during a field trial in Tuscany, Italy (this species natural environment). Berkheya coddii is known to contain up to 1 .7% nickel in its leaves in the wild (Morrey et al. , 1 992), although a more conservative mean nickel concentration of 0 .5% was reported by Robinson et al. ( 1 997b). Reasons for the poor species performance The carbon and nutrient status of the Piopio substrate is very low: 0 .47% carbon and 0 .02 % nitrogen. This low nutrient status could explain the poor performance of control plants grown in plots where no fertiliser was added, but may not explain the low metal concentrations and low biomass in general. 10 Planting dens ity for the trial described in this thesis was 5 plants/m2• The plan ting dens ity of Berkheya coddii for Robinson ( l997b) was 1 6 plants/m2• 123 Chapter 9 The lack of positive correlation for both the plant-nickel concentration and the harvested biomass with increasing fertiliser application was somewhat surprising. Robinson et al. ( 1 997b) showed a positive response of both of these variables with N fertiliser application. This evidence is in agreement with the observations of Kramer et al. ( 1 996), who showed an increase in the intraplant production of the amino acid histidine in response to plant-available nickel. Histidine acts to bind nickel and thus render the metal safe for intracellular storage (Chapter 2). Kramer et al. ( 1 996) suggested that the concentration of histidine in a plant (quantified by N%) could be directly proportional to the concentration of nickel accumulated by the aerial herbage. The negative response of plant nickel to the soil­ applied fertiliser for Berkheya coddii in this trial (Fig. 9 .3) , not supported by the intraplant nitrogen - nickel correlation (Fig 9.4), while significant, can be attributed to random variation across the block. The lack of organic material in the soil would have limited the number of 'capture sites' for N and P as these elements were released from the fertiliser granules. As a result, the residence time of the fertiliser in the rhizosphere would have been short. The form of N fertiliser used (calcium ammonium nitrate) is very water-soluble. Lack of nutrients and poor soil physical conditions could explain why no relationship existed between biomass and the plant nickel concentration with soil-applied fertiliser. All Berkheya coddii plants showed nitrogen and phosphorus deficiencies. Similarly, the lack of any effect of sulphur application on plant-nickel concentration could be attributed to the leaching of this element out of the rhizosphere before uptake (assuming the sulphur was released in the rhizosphere through biodegradation). Sulphur will lower the pH of soil, making metals more bioavailable (Robinson, 1997), but is also an integral component of many phytochelatins. An increased intraplant sulphur concentration through sulphur fertilisation could conceivably be correlated with an increased nickel concentration. Increasing the organic matter in the soil would lengthen the residence time of nutrients in the rhizosphere, but organic matter also binds metals into an insoluble form. Mean 1 24 9 metal uptake for B. coddii grown in outside plots of Piopio material at Massey was 1 700 mg/kg from the 1 00% wet-serp, but only 5 1 5 mg/kg from the serp/bark mixture. Both of these plots were fertilised with s low-release Osmocote fertiliser. There is clearly a trade off between organic material and metal bioavailability and hence uptake. Physical factors could have also contributed to the poor results yielded by the trial. The Piopio substrate has poor drainage due to compaction evident across the site. Compaction may have limited the rooting depth of individual plants. Several whole Berkheya coddii plants were excavated from the site, and it was noted that the root system of each of these was less prolific than that observed for specimens excavated from the outside plots at Massey University. Presumably, the potential for metal uptake is related to the root mass of an individual plant. Climatic conditions could have similarly affected plant growth. The summers of 97/98 and 98/99 saw low rainfall at the site and near drought conditions. The A lyssum species used have evolved to withstand dry conditions, however, the natural climate of Berkheya coddii in South Africa is characterised by a relatively higher rainfall. 9.5 Conclusion The Piopio trial showed that nickel pytoextraction could successfully be carried out in a hostile environment. The A lyssum species and Berkheya coddii all hyperaccumulated nickel, although the biomass production and metal accumulation were lower than could have been expected. Poor performance and yield in these factors explains why there was no significant decrease in soil nickel concentration effected by the trial. Nickel was removed from the site. Considering only the Berkheya coddii plants, a biomass of 480 g dry weight with an average metal concentration of 0. 1 4% equates to 0.67 g of metal removed from this block. However, with greater fertiliser loading and substrate amendments to improve the soil physical conditions and texture, factors that would increase the biomass of individual plants, as well as an increased 1 25 Chapter 9 hyperaccumulator species planting density, this figure could be significantly increased. Knowledge of these limitations and parameters are essential to ensure the successful implementation of phytoextraction technology. The trial has highlighted very clearly that extrapolation of pot and greenhouse trial data to the field can not easily be made. Physical conditions in the field are often not ideal for plant growth, a statement that is particularly true for ultramafic soils. Hyperaccumulator species grow well in their natural environment, where they have evolved to the prevalent environmental conditions. However, these same species may not 'perform' so well in a foreign environment. 126 Chapter 1 0 Chapter 10 - Hybanthus floribundus, a Native Australian Nickel Hyperaccumulator Hyperaccumulator species have a limited worldwide distribution (see Chapter 2) . This poses a serious constraint on the practical application of phytoextraction, for countries/sites where the use of exotic plant species is discouraged. Induced hyperaccumulation may allow the use of native non-accumulator species, but, as was discussed in section A, the technology has yet to advance to a level where induced hyperaccumulation can be guaranteed to work for all metal/plant combinations. Australia has many environments that may benefit from the use of hyperaccumulation technology. Decades of mining for base and precious metals has left overburden dumps and tailings dams contaminated with anthropogenic sources of metal. S ince October 1 997, part of my research has investigated the potential for the use of nickel hyperaccumulators to remediate areas of nickel contamination on the mining lease of the Western Mining Corporation Ltd. , near Kambalda, Western Australia. Three nickel hyperaccumulators exist in Australia, and one of these, Hybanthus floribundus, is native to the goldfields of West em Australia. Due to limitations on the use of exotic species in Australia, attention focused on the use of this particular plant. However, Hjloribundus had never been germinated from seed with any substantial degree success (no more than 2%). This chapter outlines my approach to overcome the inherent problems of germination associated with Hybanthus jloribundus, problems that limit the practical application of this species for phytoextraction. 1 0.1 I ntroduction Hybanthus jloribundus Hybanthus floribundus (Fig. 1 0. 1 ) is a small woody shrub found throughout southern Australia. Its most common occurrence is in the south of Western Australia and southeast of South Australia. Its distribution has been described by Bennett ( 1969) who 1 27 ssp. curv.:.L...,· ,""',us Form A capsule xS i\ V Figure 10 1 " . Hybanthus florib ssp. c:urv""-I" .. "" US Form B s s p . adpressu $ undus, F. MueJI. 128 Chapter 1 0 recognised several subspecies. These include subspecies jloribundus, curvifolius and adpressus, found mainly in the Eastern Goldfields region of Western Australia and extending from Leonora in the north to Esperance in the south. The present-day distribution of Hybanthus jloribundus and its subspecies is considered to be the relic of a much wider distribution that occurred before the onset of aridity in Australia. Hybanthus occurs in the warm-temperate zone where annual rainfall exceeds 250 mm and winter and summer isotherms are 1 2°C and 25°C respectively. Phytosociologically, this plant has been assigned by Ernst ( 1974), to the Hybanthion floribundii alliance. Together with Grevillea acuaria, it is a character species of this alliance. Severne and Brooks ( 1 972) and Cole ( 1 973) were the first to report nickel hyperaccumulation exhibited by Hybanthus jloribundus. This was only the third species reported as a hyperaccumulator of this metal. Even today, only two other hyperaccumulators of nickel is known from Australia (Stackhousia tryonii, Batianoff et al. , 1 990; Pimelia leptospermoides, S .Bidwell, pers. commn. 1 999). Nickel concentrations found in the various subspecies of Hybanthus jloribundus are high; > 1% dry mass in some individuals (Table 1 0. 1 ). This is unusual because the soils of Western Australia typically contain low nickel concentrations (ca. 1 000 mglkg) relative to the usual 5000 mg/kg found in ultramafic ('serpentine' ) soils. The plant/soil nickel concentration quotient is often > 1 0 and highlights the remarkable hyperaccumulating ability of this species. A plant/soil metal concentration quotient of 1 0 is 1 00 times greater than that for non-accumulator plants growing in the same environment (criterion of hyperaccumulation). Severne ( 1 972) concluded that Hybanthus jloribundus is confined to laterised ultramafic outcrops and creek beds, but Co le ( 1 973) proposed that this species is confined to soils derived from highly nickeliferous rocks although it does not necessarily indicate a sulphide orebody. It is interesting that the surficial expression of the first nickel orebody discovered on the Western Mining Corporation Ltd. lease in Kambalda, was expressed by the growth of Hjloribundus (R.Brooks, pers. commn. 1 997). 1 29 .n<:>f"\T<>r 1 0 Table 1 0. 1 . Mean n ickel concentrations (mg/kg DW) of species and subspecies of Hybanthus together with their locations in Western Austral ia. Values for the associated soi ls are also given . Seecies Location N i in elant Ni in soil Plant/soil N i H. fJoribundus Subsp. Curvifolius Marshal! Pool 7030 800 8.8 Form A Kurrajong 3 100 900 3.4 Form B Kambalda 3000 900 3.3 Southern Cross 451 0 1 400 3.2 Widgiemooltha 601 0 2000 3.0 Spargoville 740 970 0 .8 Subsp. FJoribundus Lake King 260 50 5.3 Southern Cross 1 020 50 20 Widgiemooltha 1 2 200 1 000 1 2 Dordie Rocks 1 3 800 1 000 1 4 Subsp. Adpressus Ravensthorpe 1 270 1 30 9.5 H. epacroides Subse. Bilobus Scadden 200 1 0 20 Source: Cole ( 1 973) and Severne (1 972). The practical application of phytoremediation or phytomining for nickel in Western Australia has been hampered by the requirement that only native species can be used. Hybanthus floribundus is therefore the only acceptable hyperaccumulator. There has been little or no success over recent years in inducing Hybanthus seeds to germinate. Plants have been raised either by tissue culture (S. BidweH, pers. commn. 1 999), or by striking cuttings. Importance of 'fire ' to promote seed germination Seeds have evolved dormancy strategies that prevent germination from occurring until conditions exist that are likely to promote growth (Goodwin et a!. , 1 995). In parts of Australia, Africa, California and Mexico fire has played an important role in the evolution of native flora. It appears that a fire event in these environments may trigger seed germination (Gill and Groves, 1 98 1 ) . Two factors are involved with fire: heat and smoke. The positive effect of heat on overcoming seed dormancy mechanisms has been recognised for some time (Beadle, 1 940; Boughton, 1 970), however other ' fire products' are now receiving greater attention. Dixon et al. ( 1 995) tested the effect of greenhouse exposure to cool smoke on the germination of 94 species of Australian natives. For 45 of these species exposure to 1 30 Chapter 1 0 smoke significantly enhanced gennination, and at least 23 of these had been totally unamenable to conventional methods of seed propagation. Hybanthus floribundus was included with 7 species for which gennination was not significantly improved relative to the control (ca. 2% gennination rate) . Smoke was generated by Dixon et al. ( 1 995) through the controlled burning of Banksia­ Eucalyptus woodland material in a large drum. This smoke was pumped through a water-cooled pipe before passage into a fumigation tent housing seed trays of the species being used. Commercially, smoke i s available in the fonn of ' smoke water' . The 'Kings Park smoke research group' has been instrumental in developing this technology to assist with the gennination of Australian native species. In addition to Kings Park smoke water, 'Regen 2000 Smokemaster' , a commercial smoke water produced by Tecnica Pty Ltd., Melbourne, was used as part of this study. Sales material accompanying this product reports that the genus Hybanthus is responsive to Regen 2000 under nursery conditions, although no specific species is mentioned. 1 0.3 Methods and Materials The following experiments were carried out to increase the gennination rate of Hybanthus jloribundus. Few experimental details are given for methods that were not successful. After treatment, seeds in tests 1 -4 were placed in moistened gennination blotters enclosed in Petri dishes sealed with ' Parafilm' . The seeds for experiments 1 -6 were ca. 6 months old. Those for experiment 7 were ca. one-year- old. 1 - Boiling seeds for periods of up to 1 0 sec. 2 - Scarification by placing seeds in a sealed round tin coated with sand paper and rotating the tin in an end-over-end shaker for periods of up to 7 days. 3 - Treating seeds with smoke generated from burning Australian wattle leaves. 4 - Treating seeds with smoke water supplied by the Perth Botanical Garden. 5 - Direct planting into ultramafic soil. 1 3 I Chapter 1 0 6 - Placing seeds on a 'Grant' Temperature Gradient Plate covered with damp K22 'Kimpac' sheets covered with three germination blotters. The temperature gradient was maintained from 7 to 33°C over a 43-day period. 7 - Placing about 35 one-year-old seeds on moistened germination blotters in Petri dishes covered with waterproof film. There were 8 different conditions in this final experiment (Table 10 .2). Germination was studied over a period of 35 days and the temperature was maintained in the range of 1 5°C - 2 1 °C during this period. Scarification was carried as in 2 above. For the water treatment, seeds (scarified and non-scarified) were soaked in either deionised water, or a 1 0 % solution of 'Regen 2000 Smokemaster' for 24 hours. Seeds were subsequently dried and used in the germination trials as per the manufacturer directions. Table 1 0.2. End of experiment results of germination test 7 on seeds of Hybanthus floribundus Smoke H2O RO H2O* Scarified Light Dark N %Germinated Yes No No No Yes 38 36.8 No Yes No No Yes 35 1 4 .3 Yes No No Yes No 37 1 0.8 No Yes No Yes No 33 3.0 Yes No Yes No Yes 33 3 .0 No Yes Yes No Yes 34 2.9 Yes No Yes Yes No 35 0 No Yes Yes Yes No 37 0 * RO H20 is water purified by reverse osmosis. Experiments 1 -6 had no positive results. The results from experiment 7 are discussed in the following section. 1 0.4 Results and Discussion Optimum germination occurred for the aged, smoke water (Regen 2000) treated seeds, germinated under dark conditions (Table 10.2). The germination rate of 37% was the highest recorded for this species. The use of one-year-old seeds appears to be of paramount importance, as the previous smoke water experiment conducted on younger seeds (expt. 4) was not successful. This suggests that seeds of Hybanthus floribundus may need to ripen within the seed coat after leaving the plant (E.Bennett, pers. comrnn. 1 32 Chapter 1 0 1 997). Scari fication appeared to hinder germination, perhaps because of tissue damage resulting from the long period o f abrasion. It appears that although darkness and smoke water are individually helpful i n promoting germination, together they have a synergistic e ffect. Even when smoke water was absent and the seeds were exposed to light, there was stil l a smal l degree o f germination (3%). Subsequent to the successful results for experiment 7, the optimal-germination conditions were repl icated to confirm these mechanisms for enhanced seed germination. S ixty , one-year-old seeds were soaked in a 1 0% solution of ' Regen 2 000 smokemaster' for 24 hours, dried and p laced on fi lter paper moistened with deionised water. S e eds were then p lace in the dark with a constant temperature range of 1 5°C - 2 1 ° C in P etri dishes covered by parafilm (n=2 ) . The germinate number was recorded over 45 days (Fig. 9 . 2 ) . Of the sixty seeds used, 30 had germinated after 6 weeks, with replicate germination rates of 56% and 47% (Fig. 1 0 .2 ) . - � 0 - .$ <0 ... c 0 .. <0 .= E ... Q) Cl 60 50 40 30 20 1 0 0 0 1 0 20 30 40 50 days s ince treatment Figure 1 0.2. Germination experiment for Hybanthus floribundus showing a final germination rate of 56% for replicate 1 (diamonds) and 47% for replicate 2 (squares). 1 3 3 Chapter 1 0 1 0.5 Conclusion This study highlights the importance of seed age, as well as the potential for the use of smoke or smoke water in combination with dark conditions to induce the germination of Hybanthus jloribundus. The importance of smoke has been shown previously for many other hard-to-germinate Australian species (Read 1 997; Dixon et al. , 1 995). Dark conditions in combination with the apparent necessity of embryo ripening within the seed coat are of paramount importance; it seems likely that Hybanthus seeds must be buried in the soil for germination to ensue. I am not sure to what extent the necessity of fire is practical for natural in-field germination of the species. I can find no recent record of a fire event in the W A Goldfields, even though natural populations of this species do exist. Scarification had no effect on germination, it is therefore probable that breaking of the seed coating is not one of the variables involved in germination of this species. The successful germination technique presented in this chapter overcomes the practical limitations hindering the use of Hybanthus floribundus for the phytoremediation and phytomining of nickel-contaminated soil; the plant has an appreciable biomass of around 1 0 t/ha. Even if neither of these possibilities are exploited, the plant should have potential for rehabilitation of mine dumps and tailings because of its high tolerance to adverse edaphic conditions. At many locations in the Eastern Goldfields Region it is the only species able to colonise such mineralised sites. Field trials have not yet been conducted using Hybanthus jloribundus, due to the initial unforeseen difficulties in seed germination hindering project advancement. However, implementation of such trials is now possible. 1 34 Chapter 1 1 Chapter 1 1 - Phytoextraction for Gold Phytoextraction for gold is the most recent advance in the growing technology of induced hyperaccumulation. Due to the extremely low solubility of gold in soil solution, no known plant will naturally accumulate the high levels of this metal necessary to successfully effect phytoextraction. Induced hyperaccumulation had been proven viable for lead, so the challenge was to extend this work to other metals. Gold seemed a very attractive target, as the significant uptake of gold by plants has long been the 'philosopher 's stone' of some scientists. This chapter reviews and outlines the background to gold mobility within soil, and hence the potential for uptake by plants. The initial experiments are then described, through which the ability to induce accumulation of gold was discovered. Finally, the current direction of research is presented, describing some of the inherent problems associated with this technology, and my approach to a better understanding of them. 1 1 .1 Introduction and review of the solution geochemistry of gold Solution geochemistry of gold The solution geochemistry of metal-ligand complexes can be described and predicted, to some degree, by the hard-soft acid-base theory of Pearson ( 1963). Hard acids are defined as small, slightly polarisable metal ions (e.g. A13+, Ti4+, and C03+) that preferentially complex with ligands through 'hard' binding sites such as P and 0 sites (e.g. H2EDTA2- and PO/-); hard bases. Soft acids are larger and more easily polarisable metal ions (e.g. Au+, Pd+ and Hg+) that preferentially complex with ligands through ' soft' binding sites such as S and N sites (e.g. SCN- and eN); soft basesl 1 . Using Pearson' s hard-soft acid-base theory, the metal gold i s classed as a soft acid in its cationic fonn, and hence in solution will have an affinity for soft bases. This theory has 1 1 A more detailed description of this theory and a more complete list of hard and soft acids and bases can be found in most inorganic chemistry textbooks. 1 3 5 Chapter 1 1 led to the attribution of some degree of the geochemical mobility of gold as thioligand complexes ; the sulphur atoms and overall electronic structure of these ligands imparting soft base chemistry. Bowell et al. ( 1 993) speculated that thiosulphate and thiocyanate ligands could be responsible for the mobility of gold within a weathering terrain in Ghana, although this was only modelled evidence to implicate thioligands in gold mobility. Kucha et al. ( 1 995) presented direct physical evidence for the deposition and transport of gold by thiosulphate in the Veitsch magnesite deposit, Austria. Gold mobility occurs in geological domains called weathering terrains. Examples of these terrains can be found in Western Australia, West Africa and South America, but the geochemical characteristics of the associated weathering fluids are poorly understood. A better understanding of the mobility of metals under weathering conditions is recognised by the minerals industry as being important. Such understanding will lead to the more effective use of 'pathfinder elements' (elements geologically associated with gold that are more abundant and hence more easily detected e.g. As), and ' geochemical haloes' (zones of metal enrichment that indicate the presence of an orebody) for defining and delineating orebodies. Likewise, at an environmental level, understanding the geochemical mobility of residual metals within a previously mined laterite terrain will effect maximisation of phytoremediation technology in the choice of plant needed to take up metals of interest. The geochemical mobility of metals and the biogeochemical pathways of plant metal cycling are inextricably linked in such a fashion. Economic mineralisation within a weathering profile Laterisation is a chemical weathering process that breaks down the primary fabric of the host bedrock. The resultant weathering profile has several component features. The surface expression of a complete laterite is a ferruginous zone, enriched in iron. Below the ferruginous zone is a clay-rich body of weathered rock called a mottled zone which grades into fresh, unweathered bedrock. The bedrock/mottled zone interface is termed a ' saprolite' while the term ' laterite' is used to describe the upper parts of the profile. Approximately 65% of the World's nickel deposits are hosted in lateritic weathering profiles. Of this total, 38% can be found in countries of the Pacific Rim. 1 36 Chapter 1 1 In a discussion of the geochemistry of gold in lateritic terrains, Gray et al. ( 1 992) described two regions that constitute a complete lateritic weathering profile. Economic mineralisation associated with the upper, laterite, zone is termed a ' lateritic supergene deposit ' . Economic mineralisation associated with the saprolite zone is termed a 'saprolitic supergene deposit ' . Supergene enrichment is defined as the secondary mobilisation and deposition of metals and enrichment through reprecipitation of any part of an orebody by the movement of geochemical fluids. The process of laterisation only occurs in humid, tropical, environments, characterised by a high water-flux rate. Much of Western Australia is today subject to more arid environmental conditions and the associated contemporary weathering is an overprint of the laterisation that occurred during the tropical conditions prevalent at various times since the Mesozoic Era (Mann, 1 984). The geological setting of Western Australia within a stable intracratonic basin, and associated low sedimentation rates, has led to the preservation of economic mineralisation within these relatively old lateritic profiles. Geochemical mobility of gold Over the past 50 years, many authors have presented evidence based on experiments conducted under laboratory conditions, to try to explain the movement of gold within a weathering profile. It seems likely that three sets of chemical species can be considered important, each having a domain dependent upon depth, pH and Eh (redox potential), where the individual species concerned can be considered as the primary species responsible for gold mobility. 1 . organic acids and inorganic components generated through the degradation of organic material within the surface layers of the weathering profile, 2 . free thiosulphate created through the weathering of mineral sulphides, and 3 . halogen species, in particular the cr ion, derived from the movement of often very saline waters through a weathering profile. 1 3 7 Chapter 1 1 1 . Degradation of organic material Humic acids Many authors have implicated humic acids in the supergene movement of gold (Baker, 1 973 ; Baker, 1 978; Mann, 1 984). Wood ( 1 995) goes further to attribute the potential mobility of gold as well as Pt, Pd, U, V, to individual organic acid components of the humic macromolecule. However, very little work has been carried out to determine directly the extractability of gold from an ore substrate by either humic acid or its more simple acid components. Baker ( 1 986) described a limited study on the effect of 500 mg/L humic substance solutions, showing significant levels of gold in solution after a 50- day extraction relative to H20 equilibrated with atmospheric CO2. Baker also reported the extractability of a range of other metals of economic importance from the mineral phase by both humic acids and the more simple organic constituent compounds. However, he did not report the extractability of gold by these organic acids. There is no report in the literature of physical evidence for the plant uptake of gold-humic or gold­ organic acid compounds. This is somewhat surprising, given the implication of these compounds in the biogeochemical pathway of gold, and represents a fruitful area for future work. Cyanogenic species Cyanide complexes have attracted much attention as solubilising agents for gold, due to the natural presence of these compounds in the environment. Natural cyanides exist due to the hydrolysis of cyanogenetic glycosides (Lakin et al. 1 974), and thiocyanates can be created through similar biochemical pathways. The dissolution of gold by cyanide and thiocyanate can be written as: Boyle ( 1 979) reported the various gold cyanide and thiocyanate complexes to be stable in mildly acid, neutral and alkaline conditions. This may be true for the gold cyanide 1 3 8 Chapter 1 1 complex, but this present study shows that more acidic conditions may be necessary to form stable, gold thiocyanate complexes. 2. Genesis of free thiosulphate Goldharber ( 1 983) stated that metastable12 sulphur oxyanions accumulate as intermediates in the pathway of pyrite oxidation over the pH range 6 to 9 : One would assume that other mineral sulphides, such a s chalcopyrite, would also release sulphur oxyanions under similar conditions. Goldharber goes further to state that these metastable species show a systematic pH dependence, with a more oxidised assemblage detected at lower pH. The transient nature of the S203 2- species under weakly acidic conditions was shown by Lyons and Nickless ( 1 968), who claimed that thiosulphate was readily oxidised to tetrathionate by weak oxidising agents: Under more strongly acid conditions, Davis ( 1 958) observed disproportionation to elemental sulphur and bisulphite: Goldharber inferred that the thiosulphate species was metastable under alkaline conditions as an intermediate in the sulphur oxidation pathway, along which further oxidation was somewhat arrested. However, Rolla and Chakrabarti ( 1 982) showed that thiosulphate is eventually oxidised to sulphate by dissolved oxygen in an alkaline medium: 12 Metastable ions are unstable on thermodynamic grounds but may exist in solution for long periods of time due to their kinetic inertness. 1 39 Chapter 1 1 This oxidative pathway is characterised by an induction period, the duration of which increases with increasing pH. Given the above equations, one would assume that gold mobility as a 52032- complex would be unlikely. This is due to the highly acid nature of supergene weathering fluids derived from the oxidation of ferrous iron (ferrolysis) within a lateritic profile (Webster, 1 986) . Instead we would expect the acid conditions to favour mobility in the fonn ofCr ion complexes, and/or humic acid complexes. However, if during laterisation sufficient carbonate is present, then the weathering pyrite will be buffered, leading to the release of thiosulphate in a neutral to moderately alkaline environment: Mann ( 1 984) calculated that 400 - 800 g of CaC03 would be sufficient for every 240g FeS2 to maintain a pH at which S20/- would remain in solution. This then, could justify the existence of supergene thiosulphate within a lateritic profile. Lintern et al. , ( 1 996) illustrated the relationship in gold values between the soil substrate and overlying vegetation for three goldfields in southern Western Australia which are characterised as having high levels of pedogenic carbonate. 3. Chloride ion complexes Gold chloride complexes are important in specific environments where the geochemical conditions are acidic (PH<4.0), very saline (CI>35 000 mgIL13) and highly oxidising (Eh>0.8 V): 1 3 This is approximately twice the salinity of seawater. 140 Chapter 1 1 Under these conditions, Gray et al. ( 1 992) theorised that gold concentrations in solution could reach 200 !J,g/L (ppb) . Such high values have never been observed in nature. These conditions are, however, realistic for arid environments such as much of Western Australia, but are not associated with tropical, humid environments undergoing laterisation. The biogeochemical pathway of gold Since the turn of the 20th century, there have been many reports of gold accumulation by plants, in particular trees (Lungwitz, 1 900). These reports have, however, often been discredited over time. Reliable evidence has come to light in the later half of the century, in particular due to the development of biogeochemical prospecting, an exploration technology that uses gold levels in plants to delineate underlying auriferous mineralisation. Work conducted in Canada showed that common conifers could accumulate up to 20 !J,g/kg gold dry weight over such mineralisation (Warren and Delavault, 1 950). Two reviews during the 1 980s have l isted over 1 50 references on this subject (Erdman and Olson, 1 985 ; Brooks, 1 992). In a more recent review, Dunn ( 1 995) reported a background level of gold in plants of only 0.2 !-lglkg dry weight, although values up to 1 00 !-lglkg could possibly be believed. Shacklette et al. ( 1 970) carried out a series of experiments in which they grew 2 species of Impatiens hydroponically in gold solutions prepared with several anionic species (Table 1 1 . 1 ). Included among these species were the thioligands thiosulphate and thiocyanate. This experiment showed that rooted plants and cuttings with their roots excised, could take up gold from solution. Clearly plants, or maybe only certain plants, can and will take up gold once the metal is in solution. The key point that limits plant uptake of gold is the limited solubility of this metal in solution. All relevant literature that describes the geochemical mobility of gold stresses this point . There are no known hyperaccumulators or gold and hence plant uptake is an induced phenomenon. My research to replicate and increase the naturally apparent levels of induced uptake necessary to effect phytoextraction, has focussed on the same chelates and ligands that perform this function in nature. 1 4 1 Chapter 1 1 Table 1 1 . 1 . Gold concentrations (mg/kg dry weight) in Impatiens ba/samina and I.ho/stii immersed for 48 hours in gold solutions of different anionic composition. Gold salt cone. PH Rooted Non-rooted Rooted Non-rooted (mg/L} I.balsamina I.balsamina I.holstii I.holstii Cyanide 29 1 1 .0 nd 32 1 60 320 7.7 nd 32 1 .7 4.2 6.5 78 260 1 80 1 30 Chloride 5.7 6.2 7.5 7.5 1 .0 7 Bromide 29 6.2 39 160 28 55 Iodide 29 6.0 2.7 45 64 33 Thiosulphate 0 . 1 6.2 0.9 <0.4 <1 <0.8 Thioc;tanate 2.2 6.2 3.3 28 1 .4 6.6 Source: Shacklette et al. , (1 970) . nd signifies no reported data. 1 1 .2 Analytical methodology Gold is generally present in geological samples at low concentration and hence graphite furnace atomic absorption spectroscopy must be employed (GF AAS). Interference of the absorption signal by iron is a common problem. Clarity of the signal is enhanced through selective extraction of a gold chloride complex into an organic phase. The solvent of choice for this research was methylisobutylketone (MIBK), after Brooks and Naidu ( 1 985). At low molality hydrochloric acid (ca. 2M), log Kt for gold in the organic phase is ca. 6, while iron is only very s lightly adsorbed from the aqueous phase (Kraus and Nelson, 1 956). Cook ( 1 998) established after exhaustive testing, the precision of this method using a GBC 3000 GFAAS instrument at Massey University. Analysis of substrate samples Total gold The bulk geochemistry of an auriferous substrate dictates the method that must be employed for the sample-digestion procedure. The method of choice for this research was digestion in aqua regia, giving a pseudo-total gold value. For some samples, however, this procedure would not remove gold from the substrate and an additional step of hydrofluoric acid (HF) digestion needed to be added. 1 42 Chapter 1 1 Samples that are oxidising, or samples that have completely oxidised are amenable to the psuedo-total, aqua regia, digestion procedure. Samples that are reduced are not amenable, as unweathered sulphide minerals within a silicate lattice may occlude the gold. This lattice needs to be destroyed with HF before the gold can be dissolved by aqua regia. Only one sample of this study, ore material from the Macraes Gold Mine in central Otago, New Zealand, needed to be treated in this way. Subsamples of sieved or crushed bulk sample were digested in the relevant acid mixture through heating on a hotplate to almost dryness. The residual liquid was subsequently diluted with hydrochloric acid (2M), filtered and adjusted to a final volume with deionised water. Aliquots were then extracted with MIBK and the organic phase analysed by GFAAS I4. Extractable gold The key variable for successful plant uptake of gold is dissolution of the metal within the soil solution. As a potential screening tool to assess the suitability of a substrate for induced solubility, a thiocyanate extraction system has been developed as part of this research. Thiocyanate was chosen due to the apparent efficacy of this ligand to induce the uptake of g01d (Shacklette et al. , 1 970). Crushed or sieved auriferous substrate (2g) was weighed into a 50 mL polypropylene centrifuge tube. Ammonium thiocyanate solution (20 mL of 0.2 g/L) was added and the tube rotated in an end-over-end shaker for approximately 20 hours. The solution was subsequently filtered and analysed by GF AAS . For some samples, the gold concentration in solution was high and clean enough to allow direct analysis of the aqueous phase. Where this was carried out, the results of the respective MIBK organic phase agreed within 5%. An equal volume of hydrochloric acid (2M) was used to acidify the filtrate before extraction with MIBK. A similar extraction is also possible using thiosulphate solution. An inherent problem with the use of thiosulphate is precipitation of elemental sUlphur following acidification 14 The GF AAS operating parameters for gold analysis are presented in Appendix 1 . 1 43 Chapter 1 1 of the solutionI5. Native sulphur will subsequently occlude any gold in solution, leading to low or no recovery of metal in the final MIBK analyte. Where the aqueous concentration of gold is high, this is not a problem as a 0.2% solution of ammonium thiosulphate gives a clear GF AAS gold signal. However, it must be noted that only ammonium thiosulphate can be used for direct aqueous analysis. The sodium salt of this ligand gives a large interference pattern. Where the gold concentration of the extractant solution is low, and thus not detectable in the aqueous phase, sulphur must be removed before analysis may proceed. The following experimental procedure was thus developed: 1 1 0 : 1 solution to substrate extraction ratio overnight in an end-over-end shaker, 2 filter solution through Whatman No 44 filter paper into a 1 00 rnL glass flask, 3 add 5 mL concentrated HN03 to precipitate native sulphur, 4 heat the solution on a hotplate. As the liquid boils the precipitated sulphur will form a small globule, 5 add dropwise sufficient liquid bromine to destroy the elemental sulphur, 6 evaporate the solution to near dryness. Care must be taken not to plate any gold on to the glass flask through heating to total dryness, 7 digest the residue in 1 0 rnL aqua regia to a final volume of ca. 2 rnL, 8 dilute the acid to 10 rnL with distilled water, 9 extract the final aqueous phase into MIBK (2 rnL) and analyse by GFAAS. This method is time consuming and a value for the percentage recovery of gold into the organic phase was not determined. However, the method has still been used as part of this research project to estimate the thiosulphate-extractable gold concentration of ore material from the Macraes mine in central Otago. A nalysis of plant samples Again the method of Brooks and Naidu ( 1 985) was employed. Dried subsamples of plant material were ashed at 550°C overnight and the ash digested in aqua regia to near 15 Gold will only extract into MIBK as an acidic chloro complex. 144 Chapter 1 1 dryness . The concentrated solution was diluted to 1 0 mL with hydrochloric acid (2M) and extracted into MIBK before analysis by GF AAS . A single Brassica juncea plant induced to accumulate gold from an artificial ore substrate (see section 10 .3) was finely ground and mixed with gold-free Brassicajuncea material to provide a test herbage-standard. Subsamples of this standard were analysed to examine alternative methods for the dissolution of gold from dry plant material (Table 1 1 .2) . Table 1 1 .2. Digestion methodology test for repl icate (3) subsamples of an adopted herbage standard. Mean gold concentrations with the same letter are not significantly d ifferent (AN OVA p>0.05). Method 1 . aqua regia d igestion of dry material 2. 6 hrs at 550°C, aqua regia ash 3. 6 hrs at 550°C, 2M HCI to hot ash 4 . 20 hrs at 550°C, aqua regia ash Mean gold concentration (J,.Lg/kg dry weight) 7 1 7 (a) 639 (a,b) 629 (a,b) 529 (b) This comparison shows that missing the ashing step gives a higher concentration of gold in the plant. This is in agreement with the often-quoted theory that some gold may be lost on ashing (Hall et al. , 1 99 1 ). The observation that ashing for 20 hours results in less gold in the plant than does ashing for 6 hours, further supports this theory. The most important observation is that digestion of the plant ash in aqua regia is not necessary. The quickest, cleanest and most accurate method appears to be ashing for a short period of time, followed by dissolution of the ash in hydrochloric acid (2M). Unfortunately, this experiment was conducted at the end of the section on gold work, and thus all analytical results in this thesis are based on a relatively long ashing time (ca. 20 hrs), followed by digestion in aqua regia. Further proving of this observation, for a range of plant species, will be the focus of future research. Contamination Contamination of laboratory equipment and analytical samples is a problem with geochemical analyses. Fortunately, the levels of gold in plants and soils of this study were above natural background. All glassware and non-disposable plasticware was 145 1 1 washed in dilute aqua regia and dried thoroughly before each new analysis, to minimise chances of sample carry-over between analyses. Blank samples were carried through with every analytical procedure. 1 1 .3 Induced uptake of gold Initial discovery The initial discovery (October 1 997) that plants could be induced to accumulate gold was serendipitous. Replicates of Brassica juncea seedlings had been transferred into a mixture of sieved Waihi gold ore materiaI l6. The aim of this experiment was to induce the uptake of the metals Cd, Pb and Zn from this particular substrate, one of few at my disposal at that time with sufficiently high levels of these metals to be of interest. Results from the common chelating agent EDT A were as predicted; induced accumulation of lead, but little change in the uptake of cadmium and zinc. However, in this particular experiment, I also decided to test the efficacy of ammonium thiocyanate and sodium thiosulphate as chelating agents to induce the uptake of heavy metals. The results for Cd, Pb and Zn were disappointing. There was no increase in the uptake of these metals by Brassica juncea through the addition of these chemicals. In an attempt to salvage something from this experiment, I decided to analyse the plants for gold, the theory being that this was an auriferous substrate and that thiocompounds were known to dissolve gold in geological terrains. The work of Shacklette et al. , ( 1 970) had shown that plants would take up gold thiocomplexes from solution. Subsamples of plant material were prepared for analysis and the gold concentrations in Brassica juncea determined (Table 1 1 .3 ) . 1 6 Wahi material was sieved to 4 grain-size fractions. Experimental substrate was a 1 : 1 x,y mixture of x<300 /lm and 710 /lm' 5 0 Kouaoua - New Caledonia Q) :Q Z 4 0 3 0 2 0 1 0 Qien - Spain __ -- Taafat - Morocco Rai - New Zealand Florence - Italy 0 �---r---'-- -'-- �=:::;::::==:;:=:;::::::::::; Puente - Spain ,- Dun - New Zealand 0 5 1 0 1 5 2 0 2 5 3 0 3 5 4 0 Estimated crop number Figure 1 2.1 . Nickel yields (kg/ha) of successive crops of a theoretical hyperaccumulator plant growing over various ultramafic soils. From Robinson et al. (2000b). The sustainability model shown in Figure 12 . 1 is a simplistic representation of a natural environment, but does show that a phytomining operation could be sustainable over a reasonable time frame. The nickel yield in successive crops would be expected to drop 1 67 Chapter 1 2 as the pool of plant-available nickel i s diminished. However, once a minimum desirable yield is reached, ploughing to transfer nickel-rich subsoil to the surface could provide a substrate amenable to subsequent harvests. It seems likely that unless the nickel is firmly bound within the silicate lattice, an equilibrium could be envisaged, where the pool of available nickel is replenished directly from metal-rich minerals as this nickel is removed through plant uptake (Robinson et ai, 1 999b). When the metal source for this replenishment becomes exhausted, the concentration of plant-available nickel in the soil solution will drop, leaving only the silicate-bound metal that is not available for uptake. The success of a phytomining operation is not governed by a ' timeframe' as is a phytoremediation operation. Phytomining, by definition, would be carried out on a site where the soil nickel levels pose no danger to health. In a phytomining scenario, the economic attractiveness of an operation would be proportional to its lifespan. Initially there would be a large capital outlay to establish the project if no incineration and/or ore-processing plant was available. However, if viable, subsequent years would see a profit return for little further expenditure. It is this conceptual difference that clearly distinguishes between phytoremediation and phytomining. 1 2.3 Gold Gold is not a toxic metal, and thus the phytoextraction of gold per se using induced hyperaccumulation would only be a phytomining exercise. Phytomining scenario Auriferous tailings usually contain residual gold in low concentrations. As extraction technologies have advanced over the years, the concentration that remains within the tailings has decreased from several thousands to several hundreds of ppb. This residual gold from both old and new tailings could be extracted using induced hyperaccumulation if the substrate were amenable to plant growth. Phytomining could also be used to extract the gold present in low-grade ores, stockpiled due to an unfavourable combination of metal yield and world gold price . 1 68 Chapter 1 2 Economics o f a crop of gold The feasibility of a gold phytomining operation using induced hyperaccumulation has been examined, assuming a crop with a biomass of 20 tlha per annum (Fig. 1 2.2). Such an operation is strongly dependent upon the world market price of this metal. At a unit cost of US$3Ikg, the cost per hectare of thiocyanate applied to a depth of 1 5 cm at a rate 0.64 glkg dry weight of substrate (the highest application rate of Fig 1 1 . 1 ) would be US$3627. At the current world price of approx US$300 an ounce, a gold concentration in the plants of around 17 mglkg would be necessary to cover the cost of the chemical. Concentrations of this order were achieved in several experiments where Brassica juncea was grown both on artificial ore, and on ore material from the Waihi gold mine in New Zealand (Chapter 1 1 ) . ..- c 2 c 8 -0 o CD 50 40 30 20 + / t / f / f • • • • • • / ... . / / . . � . .. / / ! / I . /. i/ � /. / / . / / / I • • Market price of gold (ounce) US$200 US$300 US$400 1 0 -1��"'_6. i --..- US$SOO o $3627 4000 8000 1 2000 16000 Value of the crop of gold (US$/ha) 20000 Figure 1 2.2. The possible economic value of a phytomined crop of gold as a function of the concentration in a plant with a biomass of 20 tlha. The break-even point to recoup the cost of the reagent applied at a rate of 0 .64 g/kg is shown by the line cutting the x-axis at $3627. 1 69 Chapter 1 2 Figure 1 2.2 indicates that gold phytomining might be feasible, although induced hyperaccumulation for precious metals has yet to be tested in the field. The choice of plant species to be used in an operation would be made according to the specific environment. It appears that most plants will take up metals that are in solution, providing the geochemical conditions are correct. The best choice would be a species, either exotic or native, that is hardy, withstanding extremes of temperature, water stress, acidity and salinity (where these factors were important), has a high biomass, and is fast growing. The area to be mined would first be amended to attain geochemical conditions that will effect optimal phytoextraction and plant growth, and then planted or seeded with the chosen species. The plants would be grown until the maximum biomass were realised. The area would subsequently be irrigated with the chosen solubilising chemical to induce hyperaccumulation. Once metal and/or chemical stress in the crop inhibits transpiration (metal uptake occurs during transpiration), the plants would be harvested, the biomass incinerated and the metal recovered. Environmental concerns When thiocyanate or other chemicals are applied to induce hyperaccumulation, other metals can also be made soluble and available for plant uptake. For some substrates thio-ligands are relatively selective for gold (Tui tailings, see Fig. 1 2 .5) . However, in other substrates these same thio-ligands are less selective (Paris tailings). It seems likely that this ' cocktail of metals' , as well as any inherent toxicity of the solubilising chemical itself, is responsible for the stress apparent in plants after hyperaccumulation has been induced. It is for toxicity reasons that it would never be a viable proposition to induce the hyperaccumulation of gold using sodium cyanide. It must be stressed, however, that thiocyanates themselves, are only very slightly toxic (Van Hoek, 1 995). Table 1 2.2 compares the indicative toxicity of the cyanide, thiocyanate and thiosulphate salts of sodium with that of sodium chloride. It is interesting to note that sodium thiocyante is only 4 times more toxic to rats than common salt, while sodium thiosulphate has a toxicity very similar to that of salt. 1 70 Chapter 1 2 Table 1 2.2. Comparative toxicity of 4 sodium salts that can solubil ise gold and may be used for induced hyperaccumulation. Chemical Criteria Toxicity NaCI LOw oral in rats 3000 mg/k� 1 NaCN L050 oral in rats 6.4 mg/kg NaSCN L050 oral i n rats 764 mg/kg 1 Na2S203 LOw iv in rabbits 4000 mg/kg2 Source - 1 Merck catalogue 1 999/2000, 2 - Merck Index (8 th edition) 1 968. The rate and volume of chemical application would have to be carefully controlled. Excess levels of a solubilising agent could lead to the loss of metals out of the rhizosphere zone for plant-metal uptake. This would occur if the amount of chemical applied, and hence the concentration of all metals introduced into solution, was above the level of metal in soil solution that could be effectively taken up by plants. Consideration would need to be given to climatic factors such as humidity, temperature and rainfall that would dictate the evapotranspiration rate of the plant crop and the degree of dissolution of chemical in soil. All of these factors would be important in minimising the degree of residual leachate that could potentially pollute adjacent water catchments. In the case of thiocyanate, the residence time of the chemical in the environment is relatively short, and the biodegradation pathways to ammonia, bicarbonate and sulfate have been well-studied (Hung and Pavlostathis, 1 997). It is not known, however, to what extent different substrates will affect these biodegradation pathways. It has been suggested that the use of transgenic plants expressing a bacterial thiocyanate-degrading system might be possible (Anderson et al., 1 999). Further research needs to be conducted to ensure that the application of chemicals during an induced hyperaccumulation operation will not create secondary environmental problems. The choice of the plant species to be used could also be of environmental concern. Unlike a natural hyperaccumulation operation, the plants used for induced hyperaccumulation do not have to be exotic species. A native plant that best suits the requisite features of high and rapid biomass production would be the ideal choice, minimising the risk of the uncontrollable colonisation of an unwanted exotic species in any one area. 1 7 1 Chapter 1 2 Practical Scenario 1 - gold tailings The P ar i s gold tailings of Western Austral i a (Fig. 1 2 .3) h ave b een di scussed previously. B ased o n the results o f the diagnost i c tool for induced solub i l i ty d escribed i n thi s thesis, they offer very good p otential fo r gold p hytoextract i o n . 1 00% of the gold is soluble in a weak solution of thio cyanate. The T S F covers an area 50 by 50 metres square (one quarter of a hectare) and is 6 . 6m high. It would be entirely feasible to l ower the elevation and expand the lateral dimensions to one hectare . The t ai l i ng s on average h ave a gold grade of 1 . 88 mg/kg but also have a copper concentration of 3 600 mg/kg . T h i s high copper concentration precludes the u se of conventio nal technol ogy to rep rocess the materi al . The pH of the T SF is 3 . 8 . Figure 1 2.3. Mine tailngs at the now closed Paris gold mine, on the mining lease of the Western Mining Company Ltd. , near Kambalda, Western Australia. The pale mound in the background of this photo is the TSF with the highest thiocyanate-extractable gold concentration discovered to date. Some degree of substrate modificati o n will be necessary before pl ants w i l l grow i n t h i s material - the low pH i s not o n l y inherently hostile, b u t al so promotes the leachi ng of phytotoxic leve l s of c o p p er into solut i o n . Trials are underway to i d entify the minimum 1 72 Chapter 1 2 pH at which plants will grow in this material. It may b e that this minimum pH will still support the formation of an Au-SCN complex but inhibit the leaching of copper. Alternatively, thiosulphate may be a better choice; at pH 7.0 the thiosulphate­ extractable gold concentration is very high (F.Msuya, unpublished data). Use of thiosulphate, however, would necessitate identification of a plant that did not exude acid. Initial indications are that Iberis intermedia may be a suitable candidate (S.Keeling, unpublished data). If future data suggest that geochemical parameters are attainable that support both plant growth and the formation of a stable gold-ligand complex, then a field trial should follow. The success of such a trial over a one-hectare area would offer some insight into the potential future commercial success for gold phytoextraction. Another geographical area rich in gold tailings is South Africa. In particular, the city of Johannesburg appears to be 'built on gold tailings. ' I have tested samples from one such Johannesburg tailings facility (Fig. 12 .4). The thiocyanate-extractable gold is low. However the potential for thiosulphate-induced gold uptake appears good. Again, further testing is necessary, but the Rotary Club of South Africa has expressed an interest in instigating phytomining of these tailings, as a way of employing and subsequently improving the living standards of the local communities. Practical scenario 2 - artisanal mining A significant portion of the tropical rainforest of the Amazon basin has been polluted with mercury due to the mining activities of artisanal gold miners. The underlying lateritic soils are rich in low-grade gold. Artisanal mining is a low-cost mining operation, and uses very simple, informal and unregulated technology that relies on the formation of an Au-Hg amalgam to recover the metal. The process is, however, very inefficient and leads to a waste of the gold resource. The associated hazards to human and environmental health are significant due to the large amounts of mercury lost to the environment, the siltation of drainage systems and the destruction of forest and landscape. Conservative figures estimate that a total of approximately 5 million people 1 73 Chapter 1 2 are involved with artisanal mining throughout Asia, Sub-Sahara Africa and South and Central America (H.Dahlberg et al. , pers commn . 1 999). Figure 12.4. The author. Robert Brooks, and a local security guard surveying gold-mine tailings in the suburb of Germinston, Johannesburg, South Africa. The solution geochemistry of mercury is similar to that of gold. This metal can also be made soluble with thioligands: with thiocyanate at low pH and thiosulphate at higher pH. The reasoning behind this is similar to that for the formation of gold complexes. It has been suggested that induced phytoextraction could remove both mercury and gold together from such a contaminated environment. The key target metal for the operation would be mercury. However, any gold recovered would pay some of the costs of the operation. The attraction of such a scenario is that the local communities could be involved or employed to carry out the phytoextraction process. Such an operation would be a mixture of both phytoremediation (mercury) and phytomining (gold). With the help of colleagues in the USA, Canada and Brazil this scenario may eventuate into a working trial . 1 74 Chapter 1 2 Summary -growing a crop of gold Figure 12 . 5 illustrates and summarises the complete cycle involved in growing a crop of gold, from identification of a suitable site to processing of the final ore. Each of the relevant steps of the cycle have been described in this section, with the exception of step 5, processing the bio-ore to yield the final product. Very little work on this step has been carried out to date. A small experimental test run, based on a total mass of 1 00 g dry weight of Brassica juncea material containing an average of 7 mg/kg gold, was processed according to the cycle depicted in step 5 of Figure 1 2. 5 . This test run was filmed by a BBC film crew, and screened as a segment on the BBC show 'Tomorrow's World' in England on the 14th April 1 999. On a large scale, the illustrated recovery process would never be economic. The chemical used to oxidise, and thus precipitate, the gold out of the organic phase was sodium borohydride, which is more costly than the gold recovered. However, this experiment showed that the gold in the plants was real, and could be recovered in a pure form. Future work needs to concentrate on devising an efficient and cost-effective process to recover the accumulated gold (and any other metals), from the plant material. Overleaf: Figure 12.5. GrOWing a C rop of Gold: summary poster describing the steps involved with the phytoextraction for gold. 1 75 ? Massey G ROWING A C ROP OF GOLD University Chris Anderson , Robert Brooks , Bob Stewart, F letcher Msuya and Hutham Sabti Soil and Earth Sciences, Institute of Natural Resources , Massey Un iversity, Palmerston North Step 1 • selection of a su itable site • Many thousands of hectares of gold-bearing rock exist throughout Asia, the Americas, Africa and Australasia (Fig. 1 ). • The gold content of this rock ranges from 0.1 to more than 2 g/ton, but is often unsuitable for conventional mining. Step 2 • growing a crop of plants • The selected area is seeded with a suitable plant species, or seedlings planted directly on site (Fig. 2). • A suitable plant species is one that has a high biomass, rapid growth rate and is tolerant to extremes of salin ity, aridity and heavy metal toxicity. Water extraction Step 3 • induced hyperaccumu lation Ammonium thiocyanate extraction 1 40000 ,- • Once the crop has reached its maximum biomass, a chemical is applied to the site to make the gold soluble. � �ooo-----------------,...... __ � � • Experiments at Massey University have used ammonium thiocyanate (SCN) and ammonium thiosulphate, both of which complex gold under different conditions. � 20000 -------- ------� � '0000 Cu Zn . • We have shown that ammonium thiocyanate is specific in solubilising gold (Fig. 3a,b) . , ·�$��;r���:�;;;��P�b u 0 _ .F 1 Da 5 Figure 3a COlumn expenmHnt w�re aCId nine 1<31 'ngs s • There is little risk of other dangerous heavy metals being leached into the groundwater. 'eacned.'f th wat�H for - days 'Cu Zn and Pb 119 Au · "19) Step 4 • uptake • Once the gold is soluble, plants act as a pump and hyperaccumulate gold inside living tissues (Fig. 4). • All organs of a plant accumulate gold - roots, shoots and leaves. o finely disseminated Au 0.'0 o native Au 0.01 +--,,.--r---,--r---,--r---,---, 0.0 02 0.' 0.' OA NH,SCN added (glkg dry weight subslrateJ Figure 4 - nduced uptake of Au by Brass/ca Juncea from two types of artifiCial ore After Anderson �t aJ 1998 Figure 6 • the final product gold recovered from plants Step 5 • harvesting and processing • Eventually the crop wi l l begin to die . This may take as little as a week, or as long as a month depending on the chemical used and concentration applied. • At this point the biomass is collected and the gold processed (Fig. 5). 1 - plant matenal is dried and burnt at 500OC. This Yields an ash, or bio-ore apprOlomately 1 / 1 5 the volume of the initial d biomass with 15 times more old. 2 - the bio-ore is dissolved in dilute acid. (j 3 - gold is selectively concentrated in a small volume of organic solvenl Only the gold would be soluble in the organic phase and hence would be separated out from the acid. Figure 5 • a pOSSible processing cycle for the recovery of gold from plants Step 6 • the final product • Figure 6 shows the amount of gold processed from 1 00 g of dried Brassica junc9a plant material. • There is only 0.7 mg of gold shown here, but when considered over a larger scale like a hectare, several hundred grams of gold could be recovered. • Assuming not all the gold initially present at the site was removed in the operation, the cycle could continue, and the phytomining process repeated. Reference Anderson C W N, Brooks R R, Stewart R B and Simcock R ( 1998) Harvesting a crop of gold in plants. Nature 395: 553-554. SECTION C: Chapter 1 3 SECTION C : CONCLUSION Chapter 1 3 - Practical Aspects of Phytoextraction : a General Conclusion 1 3.1 Conclusions from this research A general theme of this thesis has been the function of geochemistry on plant uptake, and how this factor affects phytoextraction. This geochemical focus on phytoextraction has been deliberate, because it represents an important part of the technology that has been poorly studied in the past. Several general conclusions can be drawn from the findings of this thesis; conclusions based on the broad range of environments and associated metals that constituted this study. I Natural uptake oJ metals is dependent upon geochemistry. To say that a hyperaccumulator of a metal will 'hyperaccumulate ' that metal when grown on contaminated soil containing that metal, is incorrect. Chapter 2 illustrates why. Both Cardaminopsis halleri and Thlaspi caerulescens, known hyperaccumulators of cadmium, failed to hyperaccumulate this metal from Tui mine tailings that contained 26 mglkg cadmium. Yet in Chapter 7, T.caerulescens accumulated close to 1 00 mglkg cadmium from a substrate containing less than 5 mglkg metal. The difference between the two sites was geochemical. Uptake (or lack of it) in Chapter 2 was from a sulphide metal phase in the substrate, but uptake in Chapter 7 was from metal contamination as a phosphate form. The data presented in Section A also show that certain hyperaccumulator plants may be better suited to specific geochemical environments than others. The greatest cadmium­ uptake potential from the phases used in this study was shown by Cardaminopsis halleri for carbonate, oxide and nitrate ( organic) phase contamination, but by Thlaspi caerulescens from the phosphate and sulphide phases. 1 77 SECTION C: Chapter 1 3 2 The modelling ofplant-available metal i.,>' dependent upon geochemistry. Modelling of the readily-soluble metal concentrat ion i n soi l i s a u seful tool . Metal i n soi l is n ot necessarily bad, but is a danger to the health of p lants and/or animal s when that metal loading i s bioavailable . The efficacy of ammonium acetate to model this fraction of a soi l ' s metal loading appears t o be dependent upon site-specific geochemistry. The data presented i n chapters 4, 5 and 6 show that the relationship between plant-available metal and metal accumulated by a plant varies between different chemical forms of metal that may be present in the soil. Caution should be observed when an extractant i s used to model plant-available metal i f no consideration i s given t o the geochemistry of the test substrate . Under some geochemical conditions, ammonium acetate, and presumably other extractants as well, may not accurately model the plant-avai lable fraction of soil metaL 3 The potential for induced hyperacclIl17lflatio" i.v large. Discovery that plants could be i nduced to accumulate gold has excited significant interest within the scientific community and especially among the general public, more so than any other aspect of phytoextraction. Yet gold uptake is nothing special . It i s simply an advancement of the d iscovery that EDT A could be used to i nduce l ead uptake. The key to i nduced u ptake is again geochemistry, and that poses the question of 'why stop there?' Mercury has a similar geochemistry to gold, as do the platinum group metals . Current research is focussing on these metals. My bel ief i s that induced uptake of any metal can be effected, if first, a suitable l igand can be found to i nduce solubili ty, and second, if rhizosphere conditions that faci litate uptake are promoted . This may be through species selection o r amendment of t he substrate . 4 The limitations (?!phytoexlrac/io!7 mllst he realised. An important aspect for the practical development of phytoextraction technology i s the establ ishment of achievable objectives . Hyperaccumulat ion wil l not be a viable answer to metal contamination in al l environments. This may simply be a function of a high level of metal contaminat ion (e.g. Auby) . It may be due to environmental conditions that 178 SECTION C : Chapter 1 3 preclude p lant growth (some areas of Western Australia), o r i t may be due to site­ specific geochemical conditions precluding economic l evels of metal uptake (e ,g , Piopio n ickel trial), It must also be remembered that the presence of a given metal is not necessarily undesirable environmentally, This point is particularly pertinent to induced hyperaccumulation, Removal of lead from a site effected by EDT A treatment may be an admirabl e goal, but if subsequent to l ead removal an underlying water table i s p olluted with soluble metal, or if trace nutrients are leached out of the soil, then the secondary environmental cost would be too high, Lead may be better left in the soil , or alternatively, removed using conventional t echnology , Care must be observed in the d evelopment of induced hyperaccumulat ion t echnology for gold , The publ ic' s perception of thiocyanate is generally bad, due in part to the similarity of the names ' cyanide' and ' thiocyanate' even though the latter is over 1 00 t imes less toxic than cyanide, The potential danger of these chemicals in the environment needs to be addressed , 5 Problems inherent in the practical development of phytoextracfion technology must be overcome. A key problem illustrated by thi s research is the limitation imposed on which hyperaccumulator species may be used in certain environments, In Western Australia there is no legal constraint on species use, so the easy option in Austral ia would be to ignore public opinion and use Alyssum species that could grow well in the Australian climate, However, in the interest of favourable public perception, mining companies stipulate that only native species may be used for revegetation, The problem of seed germination in Hybanthus jloriblllldus, existent during the early stages of this research project, appears now t o have been solved by my research, This potentially facilitates the use of this species for p hytoextraction , In environments where no suitable native species exist, exotics will have to be used if a phytoextraction operation i s to be implemented, Approaches to minimise potential environmental harm in such a scenario were presented in Chapter 1 0 ,2 , 1 79 SECTION C: Chapter 1 3 1 3.2 Future research The practical future of phytoextraction appears encouraging, but ongoing research is necessary to ensure a smooth transition from modelled to working systems. Implementation of laboratory results to the field environment is necessary for the 'real­ life ' potential of this technology to be realised. Continuing geochemical research into the importance of mineral phase as a controlling factor of species-specific bioavailability is e ssential. This will allow for extractant and plant species choices that optimise modelled and physical uptake, specific for a wide range of metals and environment. Development of induced hyperaccumulation should continue actively. There are practical scenarios where removal of metals cannot be effected using other means. The list of metals for which induced hyperaccumulation is practical should increase dramatically over the next few years. The potential for the economic recovery of precious metals using induced phytoextraction is particularly encouraging. The search for new natural-hyperaccumulator species should continue in many parts of the world that remain unexplored. As demand for the implementation of phytoextraction technology increases, the need for native hyperaccumulator species will also increase. Much of South East Asia holds promise for new discoveries . Better knowledge of the physiological mechanisms behind plant uptake i s necessary, mechanisms both inside and outside root systems. The role of soil bacteria and fungi in hyperaccumulation remains poorly understood. 1 3.3 Concluding remarks The potential for phytoextraction around the globe is enormous. Phytoremediation and phytomining are technologically simple processes. They are relatively inexpensive and could return large environmental and economic profits . In wealthy countries 1 80 SECTION C: Chapter 1 3 implementation should be a matter of time. The example of cadmium contamination in New Zealand pastoral soils, described in Chapter 7 , is one such environment where phytoremediation could be very close to practical implementation. To the mining industry, phytoextraction represents a potential useful tool. Various Australasian mining companies have already expressed interest in implementing the technology on various mine sites. Benefits from phytoextraction could be far greater for less affluent parts of the globe. Many third-world countries have large expanses of land contaminated by industrial activity, and the local communities that live on, or make a living from this land, represent a large and able workforce that could benefit from the technology. If profit could be made from 'phytomined' metal, this would be advantageous. However, the key aim of nickel, gold or other precious-metal return from the land would be to create a self-sustaining operation, worked by indigenous communities, to effect improved environmental conditions and a rise of living standards. This scenario is one of both phytoremediation and phytomining. The scope for future research into phytoextraction is as diverse as it is large. Phytoextraction is one of the areas of science that does not fall under one discipline. Botanical, physiological, agronomic and geochemical research approaches (to name but are few) are necessary to unlock the secrets of both natural and induced-metal uptake by plants. 1 8 1 References References A i n scou g h , E . W. and B rodie, A M . , 1 976. The rol e of metal ions in proteins a n d other b i o l og ical m o l ecu l e s . Journal of Chemical Education, 53: 1 56- 1 58 . Al l oway, B . J . ( Ed ) , 1 990. Heavy Metals in Soils ( J o h n Wi ley a n d Sons: N ew York ) . A l m a s , A. , S i n g h , B . R and S a l b u , B . , 1 99 9 . M o b i l ity o f cad m i u m - 1 0 9 and z i n c- 65 i n s o i l i nfl uenced by equ i l i b r i u m t i m e , temperature, a n d o rg a n i c matter. Journal of Environmental Quality, 28: 1 742- 1 7 50 . Anderso n , C . W . N . , Brooks, R R , C h i arucc i , A , LaCoste, C . , Leblanc, M . , R o b i n s o n , B . , S imcock, R a n d Stewart, R B . , 2000. Phytom i ni n g for n i ckel , tha l l i um a n d gOld . Journal of Geochemical Exploration ( i n press) . Anderso n , C . W. N . , Brooks, R R , Stewart, R B . a n d S i mcock, R , 1 998. H a rvesti n g a crop of g o l d in p l a nts . Nature, 395: 5 53-554. Anderso n , C . W. N . , Brooks, R R , Stewart, R B . and S i mcock, R , 1 999. G o l d u ptake by p lants. Gold Bulletin 32: 49-53. B aker, A J . M . , 1 98 1 . Accu m u l ators and excluders - strategies in the response of p la nts to heavy meta l s . Journal of plant Nutrition, 3: 643-654. B aker, A J . M . and B rooks, R R , 1 989. Terrestrial h i gher p l a nts wh i ch hyperaccum u l ate metal e lements - A revi ew of t h e i r d i stributi o n , eco l ogy a nd phytochemi stry. Biorecovery 1 : 8 1 - 1 26. B a ker A. J . M . , McGrat h , S . P . , Reev e s , R D . and S m i th , J . A. C . , 2000. Metal hyperaccum u l ator p lants: a review of the b i ol og ica l resource for poss i b l e exp l o itation i n t h e p h ytoremediation o f metal-po l l uted s o i l s . I n Phytoremediation of Contaminated Soil and Water ( Ed s N . Terry a n d G . B a n u e l o s ) pp 85 - 1 09 ( C R C P ress: F lorida ) . Baker, A J . M . , Reeves, R . D . and Hajar, A S . M . , 1 994. H eavy metal accum u l at i on and toleran ce in Brit ish populations of the metal lophyte Th/aspi cerulescens J . & C . Pre s l ( B rassicaceae). New Phyt% gist, 1 27: 6 1 -68 . B a ker, W. E . , 1 973. The role of h u m i c acids from Tasm a n i a n podzo l i c soi l s i n m i neral degradation a nd metal m o bi l ization. Geochimica et Cosmochimica Acta, 37: 268-28 1 . B aker, W. E . , 1 978. The rol e of h u m i c aci d i n the transport of gold . Geochimica et Cosmochimica Acta, 42: 645-649. 1 82 References B a ker, W. E . , 1 986. H u m ic substance s a n d thei r rol e i n the s o l u bi l ization and tra n s p ort of meta l s . In Mineral Exploration: Biological Systems and Organic Matter, Rubey Volume V, ( E d s D . C arl i s l e , W . L . Berry, I . R Ka p l an a n d I . R Watters o n ) p p 377-407 ( P rentice-Ha l l , E n glewood C l iffs: New J ersey ) . B a t i a noff, G . N . , S pecht, R L. and Reeves , R D . , 1 99 0 . Stackhousia tryonii B a i l ey: a n i ckel-accu m u l ating serpenti ne-endem i c species of C entral Queens l an d . Australian Journal of Botany, 38: 1 2 1 -1 30 . B a u m a n n , A , 1 885. D a s Verhalton von Z i nksa lzen gegen Pflanzen u n d i n B 6d e n . Landwirtsch Vers. -Stn . , 3 1 : 1 -5 3 . B ea d l e , N . C . W . , 1 940. Soi l temperature during forest fires a n d their effect on the survival of vegetati o n . Jounal of Ecology, 28: 1 80-1 92 . Bennett, E . M . , 1 969. The Genus Hybanthus Jacq u i n i n Austra l i a I n c l u d i n g the Cytology and Anatomy of the Western Austra l i a n S pecie s . M S c The s i s , U n i vers i ty of Western Austra l i a . B e nnett, F . A , 1 998. Studies on P l a nts that Hyperaccum u l ate Copper, Cobalt and N i ckel : their Potenti a l for use i n P hyto m i n i n g a n d P h ytoremediation. MSc Thesis, M assey U n i versity, N ew Zea l a n d , 77 pp. B lakemore, L . C . , Searle, P . L. and Daly, B . K. , 1 987. M ethods of Chem i ca l Ana l y s i s of S o i l s , New Zealand Soil Bureau Scientific Report 80, 1 03 p p . B l aylock, M . J . , 2000. F i e l d demon strations o f phytoremediation o f lead­ contam i n ated s o i l s . In Phytoremediation of Contaminated Soil and Water ( Eds N . Terry and G . B a n ue l os ) p p 1 - 1 2 ( C RC P ress: F lo r i d a ) . B l aylock, M . J . , S a l t , D . E . , Dushenkov, S . , Zakharova, 0. , Gussman, C . , Kap u l n i k, Y. , E n sley, B . a n d R a s ki n , I . , 1 99 7 . E n ha n ced accu m u l at i o n of P b i n I nd ian m u stard by s o i l -a p p l ied che lating a gents . Environmental Science and Technology, 31 : 860-865. B l i sset, A H . , 1 966. Copper-tolerant p l a nts from the Upakaringa Copper M i ne , Wi l l i amstown. Quarterly Geological Notes, Geological Society of South Australia, 1 8: 1 -4. B oughton, V . H . A. , 1 97 0 . A S u rvey of the L iterature Concern i n g the Effects of F i re o n the Forests of Austral i a . p p 1 0-1 3 ( H arvard P ress: P rymb le , Austra l i a ) . B owe l l , R J . , Foster, R P . a n d G ize, A P . 1 99 3 . The mobil ity of gold i n tropical ra i n forest soi l s . Economic Geology, 88: 999-1 0 1 6 . 1 83 References B o y d, R S . , 1 99 8 . Hyperaccu m ul ation as a plant defe nsive strategy. I n Plants that Hyperaccumulate Heavy Metals ( Ed R R B rooks) pp 1 8 1 -2 0 1 ( C A B I nternationa l : Wal l ingford) . B oyd , R S . a n d Jaffre, T . , 1 99 9 . P hytoenrichment of s o i l N i content b y Sebertia acuminata ( Sapotaceae) i n New Caledonia a n d the concept of e lemental a l le l opathy. In Conference Programme and Abstracts: Third International Conference on Serpentine Ecology pp 9 (22-28 March 1 99 9 , Kruger National P ark, South Afr i ca ) . B oyd , R S . and M a rtens , S . N . , 1 992. The rai s o n d'etre for metal hypperaccumulation by p lants. I n The Vegetation of Ultramafic (Serpentine) Soils ( Ed s A . J . M . Baker, J . P roctor, a n d R D . Reeves) p p 279-289 ( Intercept: Andover ) . B oyle , R W. , 1 97 9 . T h e geoch e m i stry of g o l d a n d i t s deposits. Geological Society of Canada Bulletin 280. B ra m l ey , R G .V. , 1 990. Cadm i u m i n N ew Zea l a n d Agricu lture. New Zealand Journal of Agricultural Research, 33: 505-5 1 9 . B rooks , R R , 1 98 7 . Serpentine and its Vegetation, 454 p p ( D i oscorides P ress: P ortland ) . Brooks, R R , 1 992. B iolog ica l methods of prospect i n g for g o l d . Journal of Geochemical Exploration, 1 7: 1 09-1 22. Brooks, R R ( Ed ) , 1 998. Plants that Hyperaccumulate Heavy Metals, 380 p p ( CAB I nternation a l : Wa l l ingford) . B rooks, R R , Dunn, C . E . and H a l l , G . E . M. , 1 995. Biological Systems in Mineral Exploration and Processing, 539 pp ( E l I i s H orwood: Hemel Hempstead) . B rooks , R R , Lee, J . a n d Jaffre, T. , 1 974. Some N ew Zea l a n d a n d N ew C aledon i an plant accumul ators of n i cke l . Journal of Ecology, 62: 493-499. B rooks , R R , Lee, J . , Reeves, R D . and Jaffre, T. , 1 977a. Detection of n i ckel iferous rocks by analysis of herbari u m speci men s of i n di cator p l a nts, Journal of Geochemical Exploration, 7 : 49-5 7 . Brooks , R R and M a l a isse, F. , 1 985. The Heavy Metal-tolerant Flora of Southcentral Africa, 1 99 p p ( B al kema : Rotterdam ) . Brooks, R R , McCleave, J .A. a n d Schofi e l d , E . K. , 1 97 7 b . Cobalt a n d n ickel u ptake by the Nyssaceae. Taxon, 26: 1 97-20 1 . Brooks, R R , Morrison, R S. , Reeves, R D . , D u d l ey, T. R a n d Akma n , Y. 1 97 9 . Hyperaccu m u l ation of n i ckel by Alyssum L i n n aeus ( C ru cifera e ) . Proceedings of the Royal Society of London, Series B, 203: 387-403. 1 84 References Brooks, R R , and Naidu, S . D . , 1 985. The determ i nation of g o l d in vegetation. Analytica Chimica Acta, 1 70, 325-329 . B rown , S . l . , C haney, R l. , Ang l e , J . S. a n d B aker, A J . M . , 1 99 4. P hytore m ed i ation potent i a l of Thlaspi caerulescens and b l adder com p a n i on for z inc- a n d cad m i u m-conta m i n ated soi l . Journal of Environmental Quality, 23: 1 1 5 1 - 1 1 57 . C ha ney, R l. , 1 983. P lant uptake o f inorg a n ic waste constitutes. I n Land Treatment of Hazardous Wastes ( E ds J . F. P arr, P . B . M arsh, and J . M . Kla) p p 50-76 ( Noyes Data Corp: Park R id g e ) . C haney, R l. , Ang le, J . S . , Baker, A J . M . and Yin-Lin g , l . , 1 99 8 . Method for phyto m i ni n g of n ic ke l , cobalt and other meta l s from s o i l . US P atent n o . 57 1 1 7 84 . C ha ney, R l. , Mal i k, M . , L i l , Y. M . , Brown , S . l. , Brewer, E . P . , Angle, J . S . and Baker, A J . M . , 1 99 7 . P hytoremed iation of s o i l meta l s . Current Opinions in Biotechnology, 8 : 2 79-284. C ochrane, R H . A , 1 96 9 Geology of Tu i M i n e , Mt Te Aroh a . M . S c thesis , U n i versity of Auckla nd , New Zea l a n d . C o l e , M . M. , 1 973. Geobotanica l and b i ogeochem ica l i nvesti gations i n the sclero p hy l l ous wood land and shrub associations of the Eastern G o l dfi e l d s o f Western Austra l i a with particu lar reference to the r o l e of Hybanthus floribundus ( L i nd l . ) F . M ue l l . as a n ickel i n d icator a n d accumulator p lant. Journal of Applied Ecology, 1 0: 269-320. Cook, A W. , 1 998. G o l d and Pal ladium as I n d i cators of a n Extraterrestrial Component in the CretaceouslTertiary B o undary Layer at Wood s i d e C reek a n d Chancet Q uarry, M arlborough, N ew Zealand , 75 p p . C ook, P . J . a n d Freney, R J . , 1 988. Sources o f cad m i u m i n agriculture . In Cadmium Accumulations in Australian Agriculture: National Symposium, Canberra 1 -2 March 1988 ( Ed s J . S i m pson and B . C urnow) pp 4-1 9 (Austra l i a n Govternm ent P u b l i c Serv i ce: C a nberra ) . C ooper, E . M . , S i m s , J .T . , Cunn i n g h a m , S . D . , H u a n g , J . W. a n d Berti , W. R , 1 999. C he l ate-assisted phytoextraction of lead from contam i n ated s o i l s . Journal of Environmental Quality, 28: 1 709-1 7 1 9. C o u rtney, S . F . , King, P. a nd Rodgers, K.A , 1 99 0 . A check l i st of m i nera l s from the Tui M i ne, Te Aroh a , N ew Zea l a n d . New Zealand Natural Sciences, 17 : 95-98. C rowdy, S . H . and Tanton , T.W. , 1 970. Water p athways in h igher p l a nts . Journal of Experimental Botany, 21 : 1 02-1 1 1 . 1 85 References C u n n i ng h a m , S. D . a n d B erti , W. D . , (2000). P hytoextract i o n and phytosta b i l i sation: tech n i ca l , e conomic, a n d regu l atory cons i derations of the s o i l -l ead i ss u e . In Phytoremediation of Contaminated Soil and Water ( E d s N . Terry a n d G . B a n u e l o s) p p 359 - 376 ( C R C Press: F lorida) . D a h m a n i M u l l er, H . , van Oort, F . , G e l i e , B . a n d Balabane, M . , 2000. Strategi e s of heavy met a l uptake by three p lant species growing near a metal smelter. Manus cr i pt in review. Davis, R E . , 1 958. D i sp l acem e n t reactions at the s u l phur atom . I . An i n terpretation of the deco m posit ion of acidified thi os u lphate. American Chemical Society Journal, 80: 3565-356 9 . D e Kock, P . C . and M itch e l l , R L . , 1 95 7 . U ptake of che l ated meta l s b y p lants. Soil Science, 84: 55-63. Dera m , A . , Petit, D . , Robinson, B . H . , B rooks , RR a n d G regg, P . E . H . , 2000. I n duced hyperaccum u lation of heavy meta l s by Arrhenatherum elatius: a possi b l e strategy for phytoremediation/phytom i n i n g . Communications in Soil Science and Plant Analysis, i n pre s s . Dixon, K. W . , Roche , S . and Pate, J . S . , 1 995 . The promotive effects of s moke derived from b urnt native vegetataion on seed germi nation of Western Austra l ia n plants. Oecologia, 1 0 1 : 1 85-1 9 2 . D u n n , C . E . , 1 995 . B i ogeochem ical prospecting for meta l s . I n Biological Systems in Mineral Exploration and Processing (Eds R R B rooks, C . E . D u n n , and G . E . M . H a l / ) p p 37 1 -426 ( E l / i s Horwoo d : H em e l H e mpstead ) . E b b s , S . D . a n d Koch i a n , L . V. , 1 998. Phytoextraction o f z i nc by oat (Avena satvia) , barley (Hordeum vulgare) and I n d ian m u stard (Brassica juncea) . Environmental Science and Technology, 32: 802-806. E l g awhary, S . M . , L i nd say, W. L . and Kemper, W. D . , 1 970. Effect of E D TA on the self-d iffu s i o n of z i nc in aqueous s o l ution and in soi l . Soil Science Society of America Proceedings 66-69. E n s ley, B . O . , Raski n , I . and Salt, D . E . , 1 997 . P hytoremediation app l i cati o n s for removi ng heavy metal conta m ination from s o i l and water. In Biotechnology in the Sustainable Environment ( E d S ayler et al. ) p p 59-63 ( P l e n u m press: N ew York ) . E pste i n , A L. , Gussman, C . D . , B l ayl ock, M . J . , Yerm i ya h u , U . , H ua n g , J . W. , Kap u l n i k , Y. and Orser, C . S . , 1 999. EDTA and P b- E DTA accu m u la t i o n i n Brassica juncea grown i n Pb-amm ended s o i l . Plant and Soi', 208: 87-94. 1 86 References Erdm a n , J . A. a n d O lson, J . C . , 1 985. The u se of p la nts i n prospect i n g for gold: a bri ef overview with a selected b i b l i og raphy and topic i ndex. Journal of Geochemical Exploration, 24: 2 8 1 -304. Ernst, W. , 1 974. Schwermeta l l vegetati o n der Erde. ( F i scher Verl a g : Stuttga rt). Ernst, W. H . O . , 1 996. B i oava i l a b i l ity of heavy meta l s and deconta m i nation of s o i l s by p l ants. Applied Geochemistry, 1 1 : 1 63-1 67 . F ergusson, J . E . a n d Stewart, C . , 1 992. The tran sport of a i rborne trace e lements copper, lead, cad m i u m , z i nc and manganese from a city to rural areas. The Science of the Total Environment, 12 1 : 247-269. Gage, M . , 1 980. Legends in the Rocks: an Outl i n e of N ew Zea l a n d Geology, pp 426 (Wh i tcou l l s : Chri stchu rch) . G arvan, F. L. , 1 964. Metal che lates o f ethyl e n e d i a m i netetraacetic acid and re lated s u bstances. In Che/ating Agents and Metal Chelates ( Eds F . P . Dwyer and D . P . Me l lor) p p 283-333 (Academic Press: New York). G i l l , A. M . and Groves, R . H . , 1 98 1 . F i re reg i mes i n heath lands and their p l ant - ecological effects. In Ecosystems of the World 98: Heathlands and Related Shrublands - Analytical Studies ( E d R . L . S pecht) pp 6 1 -84 ( E l sevier: Amsterdam) . G i l l i ngham, A. G . , Gregg, P . E . H . and N ewsome, P . , 1 998. Future fert i l i ser n utrient requirements of the New Zealand sheep and beef i nd ustry. In Long-term Nutrient Needs for New Zealand's Primary Industries: global supply, production requirements and environmental constraint ( Eds L . D . Currie and P . Loganath a n ) . Occasional report No. 1 1 . F erti l izer and Lime Research Centre, Massey U n i vers ity, P a l merston North . pp 35-55. Godbold, D . L . , Fritz, E . L . and H Otterm a n , A. , 1 988. Al u m i n i um toxicity and forest decl i ne. Proceedings of the National Academy of Sciences USA, 85: 3888-3892 . G o l d harber, M . B . , 1 983. Experi mental study of metastable s u l p h u r oxya n i on formation during pyrite oxi dation at pH 6-9 and 30°C. American Journal of Science, 283: 1 93-2 1 7 . Gom my, C . , 1 997. O pt i m i sation d ' u n schema de speciation des metaux P b , Zn, C d et Cu: app l i cation a des sols p o l l ues du N ord d e l a F rance. P h D Thes i s , U n ivers ite de Technologie d e Compiegne, France. Goodwi n , J . R. , Doescher, P . S . and Erd d l eman, L. E . , 1 995. After-ripen i n g in Festuca idahoensis seeds: adaptive dormancy and i m p l ication for restoration . Restoration Ecology 3: 1 37- 1 42 . 1 87 References Gray, C . W . , McLaren, R G . , Roberts, A. H . C . and Cond ron, L. M . , 1 999. C a d m i u m phytoava i l a b i l ity in some New Zealand s o i l s . Australian Journal of Soil Research, 37: 46 1 -477 . G ray, D . J . , B utt, C . R M . a n d Lawrence, L. M . , 1 992. Geoch e m i stry of gold i n l aterite terra i n s . I n Regolith Geochemistry in Tropical and Subtropical Terrains. Hand Book of Exploration Geochemistry V4 ( Ed s C . R M . B utt and H . Zeegers) pp 461 -482 ( C hapman H a l l : Londo n ) . G ri l l , E . , Winnacker, E. L . and Zen k, M . H . , 1 987. Phytoch e l atins, a class of heavy-metal-binding peptides from p lants, are funct i o n a l l y ana lagous to meta l l oth ione i n s . Proceedings of the National Academy of Sciences, 84: 439-443. H a l l , G . E . M . , Rencz, A. N . and Maclauri n , A. I . , 1 99 1 . C o mparison of analytica l resu lts for gold i n vegetation with and without h i g h-tem perature a s h i n g . Journal of Geochemical Exploration, 41 : 29 1 -307 . Haq, A. U . , B ates, T. E . and Soon , Y. K. , 1 980. Comparison of extractants for p l a nt-ava i l able z i nc, cad m i u m , n icke l and copper i n conta m i n ated soi l s . Soil Science Society of America Journal, 44: 772-7 7 7 . H awke, D . , H o ldaway, R , Ca user, J . a n d Ogden, S . , 1 99 8 . Contri butions to present day soi l fert i l ity by pre-E uropean sea b i rd bree d i n g . In Abstracts: New Zealand Soil Society Conference, Gisborne, November 1998 pp 53- 54. Hea ly, W. B . , 1 967. Ingestion of soi l by sheep. Proceedings of the New Zealand Society of Animal Production, 27: 1 09-1 20. H omer, F . A. , 1 991 . C hem ical Stud ies on Some P l ants that Hyperaccu m u l ate N icke l . P h D Thes is, M assey U n iversity, New Zea l a n d , 267 pp. H omer, F . A. , Reeves, R. D . and Brooks, R . R , 1 997. The pos s i b l e i nvolvement of a m i n o aci ds in n i ckel chelation in some n i ckel -accu m u lating p la nts. Current Topics in Phytochemistry, 14: 3 1 -33. H uang, J . W. , Chen , J . , Bert i , W.R and C u n n i n g h a m , S . D . , 1 997 . P hytoremed iation of lead-contam i nated so i l s : ro le of synthetic chelates i n lead p hytoextraction. Environmental Science and Technology, 31 : 800- 805 . Huang, J . W. and C un n i g h a m , S . D . , 1 996. Lead phytoextraction: species vari ation i n lead uptake and tra ns locat ion. New Phytologist, 1 34: 75-84 . H u n g , C h ien-Ho and Pavlostat h i s , S . G . , 1 997 . Aerobi c b i odegradation of thi ocyanate. Water Research, 3 1 : 276 1 -2770. 1 88 References Kabata-P e n d i s , A. a n d P en d i a , H . , 1 984. Trace Elements in Soils and Plants, 3 1 5 p p ( C RG Press: F lorida) . Kea r , 0 . , 1 960. Geo l o g i ca l M a p of N ew Zeal and 1 : 250 000, S h eet 4 - H a m ilton ( Department of Scientific and I n dustri a l Research : Wel l i ngto n ) . Kear, D . , 1 967. Economi c geol ogy o f t h e Wai kato . Earth Science Journal, 1 : 1 - 1 8. K ni g ht , C . , Kai ser, J . , L a l or, G . C . , R obotham, H . a n d Witter, J . V. , 1 997. Heavy meta l s i n surface water and stream sed im e nts i n J a m a i ca . Environmental Geochemistry and Health, 1 9 : 63-66 . Kob a ya sh i , J . , Mori i , F . And M u ramoto, S . , 1 974. Removal of cad m i u m from p o l l uted s o i l with the che lati n g agent E D T A In Trace Substances and Environmental Health Volume 8 ( E d D . D . H e m ph i l l ) 1 79 p p ( U n ivers ity of M i ssouri : C o l u m b i a ) . Kramer, U . , C otter-Howe l l s , J . D . , C h a rnock, J . M . , B a ker, A J . M. a n d S m ith, J . A C . , 1 996. Free h isti d i ne a s a metal chelator i n p l ants that accum u l ate n i cke l . Nature, 379: 635-638. Kra mer, U . , S m ith, R . D . , Wenzel , W.W. , Raski n , I . and Salt, D . E . , 1 997. The role of metal transport and tol e rance in n ickel hyperaccum u l ation by Thlaspi goesingense H a lacsy. Plant Physiology, 1 1 5: 1 64 1 - 1 650. Kra u s , K.A and N e l son, F . , 1 956. An i on exch a n g e stud i e s of the fission products. I n Proceedings of the International Conference on the Peaceful Uses of Atomic Energy: Volume 7, p 1 1 3 ( U n ited N ation s ) . Kucha, H . , P rohaska, W . and Stumpfl , E . F . , 1 995. D e p osit ion a n d transport of g o l d by thios u l phates, Veitsch, Austri a. Mineralogical Magazine, 59: 253- 258 . Kumar, N . P . B . A. , D ushenkov, V. , Motto, H . a n d Raski n , I . , 1 995. P h ytoextraction: the use of p l a nts to remove h eavy metals from soi l . Environmental Science and Technology, 29: 1 232-1 238. Labrune, L . and Douay, F . , 1 99 7 . Metaux l ourds ( Pb , Zn et Cd) et m icroflore du so l . In Programme de Recherches Concertees, etude d'un secteur pollue par les metaux - parlie 1, metaux pofluants des sols (Cd, Pb et Zn) et organismes vivants. U n p u b l i shed rep o rt com m i s i o ned by the Region N o rd P a s de C a l a i s , France. LaCoste, C . , Robinson, B . , B rooks, R . , Anderson , C . , C h i arucci , A . and Lebanc, M . , 1 99 9 . The phytore m e d i at i on potenti a l of thal l i um-contam i nated s o i l s u s i n g Iberis and Biscutella s pecies. International Journal of Phytoremediation, 1 : 327-338. 1 89 References Laki n , H . W. , Curtin , G . C . and H ubert, A E . , 1 97 4 . Geochemi stry of g o l d i n the weathering cycle. United States Geological SUlvey Bulletin 1 33 0 . L a u r i e , S . H . a nd Manthey, J .A , 1 994. T h e che m i stry and rol e o f metal i o n che l at i o n i n p lant u ptake processes. I n Biochemistry of Metal Micronutrients in the Rhizosphere ( Eds J . A Manthey, D . E . C rowly and D . G . L uster) p p 1 65 - 1 82 ( Lewi s P u b l i shers ) . Laurie, S . H . , Tancock, N . P . , McGrath, S . P . a n d Sanders, J . R , 1 99 1 . I nflu ence of comp l exation on the u ptake by p l ants of i ron , m a n g a nese, cop per a nd z i n c 1 . Effect of E D TA i n a m ulti-metal a n d com puter s i m u lation study. Journal of Experimental Botany, 42: 509-5 1 3 . L a uwreys , R , 1 977. E v a l uation of the i mpact of cadm i u m o n the health of m a n . Luxe m b ourg, C o m m ission o f t h e E u ropean C o m m u n i t i e s , pp 72-74. L a uwreys, R , 1 979. C a d m i u m in man. In The Chemistry, Biochemistry and Biology of Cadmium ( E d M . Webb) p p 433-455 ( E lsevier/North H o l l a nd B i o me d i ca l Press : Amsterdam ) . L e b l a nc, M . , P etit, D . , D eram, A , R ob i nson, B . H . and B rooks , R R , 1 999. The phytom i n i ng and envi ronmental s i g n ifi ca nce of hyperaccum u l at i o n of tha l l i u m by Iberis intermedia from southern France. Economic Geology, 94: 1 09-1 1 4. Leblanc, M . , Robinson, B . H . and B rooks, R R , 1 997. Hyperaccu m ulation de tha l l i u m par Iberis intermedia ( B rassica ceae) . Comptes Rends de I'Academie des Sciences, Paris, Serie D. Lee, J . , Reeves, R D . , Brooks, R R and J affre, T. , 1 977. Isolation and identifi cation of a citrato complex of n ickel from n ic kel-accum u la t i ng plants. Phytochemistry, 1 6: 1 503-1 505. L i , Z. and S h u man, L . M . , 1 996. Extracta b i l ity of zi nc, cadm i um and n i ck e l i n soi l s ammended with E DTA Soil Science, 1 6 1 : 226-2 4 1 . L i ntern, M . J . , B utt, C . R M . and S cott, K. M . , 1 996 . Gold i n vegetation a n d s o i l - three case stud i e s from the goldfi e l d s of southern Western Austral i a . Journal of Geochemical Exploration, 58: 1 - 1 4 . Loganatha n , P . and H e d l ey, M . J . , 1 997. Downward movement of cad m i u m and p h os phorus from phosphat i c fert i l i sers in a pasture s o i l i n N ew Zeal and. Environmental Pollution, 95: 3 1 9-324. Loganatha n , P . , Hedley, M . J . , G regg , P . E . H . and C urrie , L . D . , 1 997. Effect of phosph ate fert i l i ser type on the accumul ation and p l a nt ava i la b l ity of cadm i u m in grassl an d soi l s . Nutrient Cycling in Agroecosystems, 46: 1 69 - 1 78 . 1 90 References Loganathan, P . , Mackay, A. D . , Lee, J . and Hedley, M . J . , 1 995. Cadm i u m d i stri bution i n h i l l pastures a s influenced b y 20 years of phosph ate a p p l ication and sheep grazi n g . Australian Journal of Soil SCience, 33: 859-87 1 . Lungwitz, E . E . , 1 900. The l ixiviation of g o l d depos its by vegetat i o n . Engineering and Mining Journal, 69: 500-502. Lyo n s , D. and N i ckless, G . , 1 968. The l ower oxy-acid s of s u l p hur. I n Inorganic Sulfur Chemistry ( Ed G . N ickless) pp 79-84 ( E l sevi er: Amsterd a m ) . Macn a ir, M . R. , Bert, V . a n d Hu itson , S . , 1 999. Z i nc tolera nce and hyperaccu mulation i n Cardaminopsis halleri. In Conference Programme and Abstracts: Third International Conference on Serpentine Ecology pp 1 5 (22-28 March 1 999, Kruger Natio n a l Park, South Africa ) . M a n n , AW. , 1 984. M o b i l ity of g o l d a n d s i lver i n lateritic weatheri n g profi les: some observations from Western Au stra l i a . Economic Geology, 79: 38-49. M a rschner, H . , 1 995. The M i neral Nutrition of H i g her P l ants , 2nd E d ition (Academic P ress: Lond o n ) . McGrath, S . P . 1 998. Phytoextract ion for S o i l Remediat i o n . I n Plants that Hyperaccumulate Heavy Metals ( E d : R . R . B rooks) pp 2 6 1 - 287 ( CAB I nternational : Wa l l i ngford ) . McGrath, S . P . , Dunham, S . J . and C orre l l , R . L . , 2000. P otential for phytoextract i on of zinc and cad m i u m from soi l s u s i n g hypera ccu m u l ator p l a nts. In Phytoremediation of Contaminated Soil and Water ( Eds N . Terry a n d G . B a nuelos) pp 1 09-1 28 ( C R C P ress: F l ori d a ) . M cGrath , S . P . , S h e n , Z . G . a n d Zhao, F . J . , 1 997. Heavy metal u ptake a n d chemical changes i n t h e rh izosphere of Thlaspi caerulescens and Thlapsi ochroleucum grown in conam i nated s o i l s . Plant and Soil, 1 88: 1 53-1 59. McGrath, S . P . , S i d o l i , C . M . D . , B a ker, A J . M . and Reeves, R . D . , 1 993. The potent ial for the use of metal-accumulating p lants for the in-situ d econta m i nation of meta l-po l l uted s o i l s . In Integrated Soil and Sediment Research: A Basis for Proper Protection ( Ed s H . J . P . E ijsackers and T. Ha m mers) pp 673-676 ( Kl uwar Acad e m i c P u b l i s h ers: Dordrecht, H o l la nd ) . McGrat h , S . P . a n d S m it h , S . , 1 99 0 . Chromi u m a n d N i ckel . I n Heavy Metals in Soils, 2nd Edition ( Ed B . J . Al l oway) p p 1 52-1 78 ( B lackie: G l a sgow). McLa u g h l i n , M . J . , Smol ders, E. , M e rckx, R . and Maes, A, 1 997. P l a nt uptake of cadm i u m and zinc i n che l ator-buffered n utri ent solution d epends on l i g a nd type. In Plant Nutrition for Sustainable Food Production and 1 9 1 References Environment ( E d s TAndo, K Fuj ita, T Mae, H . Mats u m oto, S . Mori and J . S ekija) p p 1 1 3- 1 1 8 ( Kl uwar Academic P u b l i shers : Dordrecht) . McLaugh l i n , M . J . , T i l l er, K G . , Naidu, R and Stevens, D . P . , 1 996. Review: the behaviour and enviro n mental i mpact of contam i n a nts i n fert i l i zers. Australian Journal of Soil Research, 34: 1 -54. Mean s , J . L. , Kucak, T . a n d Crerar, D . A , 1 980. Relative d e g radation rates of NTA, E D TA a n d DTPA and e nv ironmental i m p l i cation s . Environmental Pollution, 1 : 45-60. Merck C ata logue 1 99 9/2000. 1 440 p p (Merck KGaA: Darm stadt) . Merck I ndex (8th E d i t i o n ) 1 968. ( Ed s P . G. Strecher, M . Wi nd holz, D . S . Lealy, D . M . B olton and L . G . E aton) 1 7 1 3 p p (Merck and Co: Rahway). M e sjasz-Prybytowicz, J . , B a l kwi l l , K , Prybytowicz, W. J . a n d Annegarn, H . J . , 1 994. P roton m i croprobe a n d X-ray fluorescence i nvestigations of n icke l d i stribution in serpentine flora from S outh Africa. Nuclear Instruments and Methods of Physical Research, B89: 208-2 1 0. M i nguzz i , C . and Vergnano, 0 . , 1 948. 1 1 conten uto d i n iche l nel l e ceneri d i Alyssum beriolonii. Atti del/a Societa Toscana di Scienze Naturale, 55: 49- 74. M o rre l l , W. J . M . , 1 99 8 . An a ssessm e nt of the revegetati on potent i a l of b ase­ metal ta i l ings from the Tui M i ne , Te Aroha, New Zea l an d . P h D Thesis, M as sey U n iversity, N ew Zea l a n d . Morre l l , W . J . M . , Stewart, R B . , Gregg , P . E . H . , B o l a n , N . S . and H orne , D . , 1 998. An assessment of s u l p h i de oxidation in a b andoned base-metal ta i l ings, Te Aroh a , New Zea l a n d . Environmental Pollution, 94: 2 1 7 -225. Morrey, D . R , B a l kwi l l , K. , B a l kwi l l , M . -J . and Wi " iamson, S . , 1 992 . A revi ew of some stud ies of the serpenti n e flora of S outh Afri ca . In The Vegetation of Ultra ma fic (Serpentine) Soils ( Ed A J . M . Baker, J . P roctor, and R D . Reeves) pp 1 47-1 5 7 ( Intercept: Andover) . Morrison, R S . , 1 98 0 . Aspects of Accum u l at i o n of Cobalt, Copper and N i ck e l by P l ants. P h D Thes is, Massey U n iversity, New Zea l a n d , 287 p p . M u rozum i , M . , Chow, T J . and Patterson, C . , 1 96 9 . C h e m i ca l concentrations of p o l l utant l ead aerosol s , terrestr i a l dusts a n d sea sal ts i n G reen land and Antarct i c snow strata . Geochimica et Cosmochimica Acta, 33: 1 247-1 2 94. N ewl a n d , L . W. and Dunn, K. A , 1 982. Lea d . In The Handbook of Environmental Geochemistry, Volume 3 Pari B: Anthropogenic Compounds ( Ed O . H utzi nger) pp 1 -26 ( S pringer-Verlag: Berl i n ) . 1 92 References N i ck s , L. a n d Chambers , M . F . , 1 994. N i cke l Farm i n g . Discover Magazine, 1 9 S eptember. N i cks, L . a n d C hambers, M. F . , 1 995. Farm ing for m eta l s . Mining Environmental Management, 1 5- 1 8 S e ptember. N i cks, L . a n d Chambers, M . F . , 1 99 8 . A p i oneer i n g study of the potenti a l of p h yto m i n i n g for n i ckel . I n Plants that Hyperaccumulate Heavy Metals ( E d R R Brooks) p p 3 1 3-326 ( CAB I nternational : Wal l i n gford) . N r i a gu , J . O . , 1 978a . Lead i n s o i l s , s e d iments a n d m ajor rock types . I n The Biogeochemistry of Lead in the Environment Part A ( E d J . O . N ri a g u ) p p 1 5-72 ( E l svier/N orth H o l l and B i omed i cal P ress: Amsterd a m ) . N r i ag u , J . O . , 1 978b. Propert i es and t h e b i ogeoch e m i c a l cycl e of l e a d . I n The Biogeochemistry of Lead in the Environment Part A ( Ed J . O . Nriagu) p p 1 - 9 ( E lsevier/North H o l l a n d B i om e d i ca l Press: Amsterd a m ) . N r i a g u , J . O . , 1 980. C a d m i u m i n t h e E nv i ronment, P a rt I : Ecolog i ca l Cycl i n g 682 pp ( J o h n Wiley and Sons: N ew york ) . N r i a g u , J . O. a n d Pacyna, J . M . , 1 98 8 . Q uantitati ve a ssessment o f worldwide contam i nation of a i r, water and soils by trace meta l s . Nature, 333: 1 34- 1 39 . P a n ca ro , L . , P e l os i , P . , Vergano G a m b i , O . and G a l o p pi n i , C . , 1 97 8 . F u rther contri bution on the relations h i p between n ickel and m a l i c and m a l o n i c acids i n Alyssum. Giornale Botanico Itafiano, 1 1 2: 282-283. Pearson , R G . , 1 963. Hard and soft acids and bases. Journal of the American Chemical Society, 85: 3 533-3539. Perdix, E . , Gommy; C . , Ga l l oo , J . C . and G u i l l ermo, R , 1 99 7 . D istribution du plomb, du zi nc, d u cad m i um et d u cuivre dans les s o l s par fracti onnement ch i m ique. I n Programme de Recherches Concertees, etude d'un secteur pol/ue par les metaux - partie 3, distribution et speciation des metaux. U n publ i shed report com m i sioned by the Region N ord Pas de C a l a i s , F rance. P eterso n , P . J . and AlIoway, B . J . , 1 97 9 . Cad m i u m i n s o i l s and vegetat i o n . In The Chemistry, Biochemistry and biology of Cadmium (Ed M . Webb) p p 45-92 ( E lsvi er/North hol l a n d B i omedical Press: Am sterdam ) . P o l l ard , A . J . a nd B aker A. J . M . , 1 99 7 . Deterrence of herbivory by z i nc hyperaccu mulation i n Thlaspi caerulescens ( B rassicaceae ) . New Phyt% gist, 1 35 : 655-658 . 1 93 References Rask i n , I . , Kumar, P . B .A N . , Dushenkov, S. and S alt, D . E . , 1 994 . B io co n ce ntration of heavy metals i n p l ants. Current Opinions in Biotechnology, 5: 2 85-290. Read , T. , 1 99 7 . Smoke, heat and fire effects o n s o i l seed bank germin atio n for the re-establ ishment of a native forest com m u nity in N ew S outh Wal e s . I n Proceedings o f the 2ZW Annual Environmental Workshop, p p 72-89 ( M i nerals Counci l of Austral i a , D ickson, ACT ) . Reeves, R D . , Baker, A J . M . , B orh i d i , A And Bereza i n , R , 1 996. N i ckel­ accum u l ating pants from the ancient soi l s of Cuba. New Phytologist, 1 33 : 2 1 7-224. Reeves, R D . , Baker, A J . M . and B rooks, R R , 1 995. Abnormal accu m u lation of trace m etal s by plants. Mining Environmental Management, 4-8 Septe m b er. Reeves, R D . and Brooks, R R , 1 983. Hyperaccumu lation of lead and zinc by two meta l lophytes from m i n i n g areas of central E u rope. Environmental Pollution, 31 : 277-285. Reeves, R 0, B rooks, R R and D u d ley, T. R, 1 98 1 . U ptake of n ickel by species of Alyssum, Bornmuellera and other g e n era of Old World Tribus Alysseae. Taxon, 32: 1 84-1 92. Roberts, A. H . C . , Longhurst, R D . and B rown , M . W. , 1 993. C a d m i u m acc u m u l ation in N ew Zeal an d pastoral agricu lture. I n Proceedings o f the XVII International Grass/ands Congress, pp 1 45 1 -1 452. Roberts, A. H . C . , Lon g h u rst, R D . and B rown , M . W. , 1 994. Cadm i u m status of soi l s , p l a nts, a n d grazing a n i m a l s in N ew Zealand. New Zealand Journal of Agricultural Research, 37: 1 1 9- 1 2 9 . Roberts, A. H . C . a n d M o rton , J . D . , 1 99 8 . H i n d s i g ht and fores ight: current a n d pred i cted future trends i n n utrient requ i rements for d ai ry i ng . I n Long-term Nutrient Needs for New Zealand's Primary Industries: global supply, production requirements and environmental constraint ( Eds L . D . C u rrie and P . Loganath a n ) . Occasional report N o . 1 1 . Fert i l izer a n d l ime Research Centre, Massey U n i versity, P a l m e rston N orth . p p 85-9 7 . Robinson, B . H . , 1 997. T h e P hytoextracti o n o f H eavy Meta ls from Meta l l ife rous Soi l s . PhD Thesi s , Massey U n iversity, N ew Zea land, p p 1 44. Robinson, B . H . , B rooks, RR a n d C l othier, B . , 1 999a. Soil amendments affecti ng n i ckel a n d coba lt uptake by Berkheya coddii: potenti a l use for phyto m i n ing and p hytoremed iation. Annals of Botany, 84: 689-694. 1 94 References Robinson, B . H . , B rooks, R R , G re g g , P . E . H . and Kirkman , J . H . , 1 999b. The n i ckel phytoextracti o n potential of som e ultramafic soi l s as determined by sequentia l extraction. Geoderma, 87: 2 93-304. Robi ns o n , B . H . , B rooks, R R , H owes, A W" Kirkman, J . H . a n d G re g g , P . E . H . , 1 997b. The potential of the high-biomass n i ckel hyperaccu m u lator Berkheya coddii for phytoremediation and p h ytomi n i ng . Journal of Geochemical Exploration, 60: 1 1 5- 1 26. Robi ns o n , B . H . , C hiarucci , A. , Brooks, R R , P etit , D . , Kirkman , J . H . , Gregg, P . E . H . and De D o m i nicis, V. , 1 997a. The n i ckel hyperaaccum u l ator p la nt Alyssum benolonii as a potential agent for phytoremed iation a n d phytomi n i ng o f n i cke l . Journal of Geochemical Exploration, 59: 7 5-86. Rol la , E. and Chakrabarti C . C . , 1 98 2 . Ki netics of decomp o sit ion of tetrath ionate, trithionate and t h i o s u l p h ate in a l ka l i ne m e d i a . Environmental Science and Technology, 1 6: 852-857 . Severne, B . C . , 1 972. B otan i ca l Methods for M in e ra l Expl oration i n Western Austra l i a , P h D Thesis, M assey U n ivers ity, N ew Zea land. S ev e rn e , B . C . and B rooks R R, 1 972. A n i ckel-accumu lating p lant from Western Austra l i a . Planta, 1 03 : 9 1 -94. Shack lette, H . T. , Lakin, H . W. , H u be rt, A E . , C u rt i n , G . C . , 1 970. Absorption of g o l d by p l ants. Contributions to geochemistry, Geological Survey Bulletin 1 3 1 4- B . S h e n , Z. G . , Khao, F . J . and McGrath, S . P . , 1 997. U ptake and transport o f z i n c i n t h e hyperaccum u l ator Thlaspi caerulescens and the n o n ­ hyperaccumu lator Thlaspi ochroleucum. Plant, Cell and Environment, 20: 898-906 . S h u m a n , L . M . , 1 985. F ractionation methods for so i l m icroelements. Soil Science, 1 40: 1 1 -22. Singh, B . R and Narwa l , R P . , 1 984. P lant ava i l a b i l ity of heavy m etal s in a s l udge treated soi l : 1 1 . Metal extractab i l ity compared with p l ant metal u ptake. Journal of Environmental Quality, 1 3: 344-349. S o mmers , L . E . and L i ndsay, W. L . , 1 979. Effect of pH and redox on predicted heavy meta l-chelate equ i l i br i a i n soi l s . Soil Science Society of America Journal, 43: 39-47 . Sti l l , E . R and W i l l iams, R J . P . , 1 980. P otential methods for select ive accumu lation of n i ckel 1 1 ion in p l ants . Journal of Inorganic Biochemistry, 1 3: 35-40. 1 95 References Stockley, E . , 1 980. B i ogeoch e m i cal Studies o n the N ickel C o m p l ex C onta ined in the N i ckel-accu m ulating Legume Personia metallifera from the Great Dyke Area, Zimbabwe. BSc ( H ons) The s i s , M a ssey U n i versity, N ew Zea l a n d . Striet , B . a n d Stu m m , W. , 1 993. C h e m i ca l properties o f metal s a n d the processes of b ioaccu m u l at i o n in terrestria l p lants. In Plants as Biomonitors: Indicators of Heavy Metals in the Evironment pp 3 1 -62 (VCH P u b l i s h ers: Wei n h e i m ) . Tanto n , T . W. a nd C rowdy, S . H . , 1 97 1 . T h e d i stri bution o f lead chelate i n the tra n s p i ration stream of h igher p l a nts. Pesticide Science, 2: 2 1 1 -2 1 3 . Tay, KA , 1 980. Geoche m istry a n d E nvironmental I mpact of D i scharge of H eavy M eta l s from the Tui M i n e and its Tai l ings. M S c thes i s , U n i versity of Wa i kato, New Zea land. Taylor, G . J . a n d Foy, C . D . , 1 985. D ifferent i a l u ptake and toxi city of ionic and che lated copper i n Triticum aestivum. Canadian Journal of Botany, 63: 1 27 1 - 1 27 5 . Taylor, M . D . a n d Perci v a l , H . J . , 1 99 2 . I nterim report on conta m ination o f N ew Zea l a n d soi l s by cad m i u m and sel ected other heavy meta l s . DSIR Land Resources Technical Record 1 16. Tee Haar, G . L . and Bayard, M . A , 1 97 1 . Composition of a i rborne l ead particles. Nature, 232: 553-554 . Tejowu l a n , R S . a n d H endershot, W. H . , 1 99 8 . Removal of trace meta l s from conta m i nated s ites using E DTA i n corporating resi n tra p p i n g tech n iques. Environmental Pollution, 1 03: 1 35-1 42. Tessier, A , C a m p be l l , P . G . C . and B is son , M . , 1 979. Sequential extraction proce d u re for the speciation of particulate trace meta l s . Analytical Chemistry, 5 1 : 844-85 1 . Thornto n , I . , 1 999. B i oava i l a b i l ity of trace elements i n the foodcha i n . In Proceedings of Extended Abstracts: 5th International Conference on the Biogeochemistry of trace Elements ( Eds W. W. We nzel , D . C . Adriano, B . Al l oway, H . E . D oner, C . Ke l l er, N . W . L epp, M . Mench, R N a i d u , G . M . P i erzynski ) p p XI-X IV ( J u ly 1 1 - 1 5 , 1 999, Vienna, Aust r i a ) . Ti l l er, K G . , O l i ver, D . P . , McLaughl i n , M . J . , Merry , R H . a n d N a i d u , R , 1 99 7 . Mana g i n g cad m i u m contam i nat i o n o f agricultural land. I n Remediation of Soils Contaminated with Metals, Science Reviews, Advances in Environmental Science ( Eds I . K l skander and D . C . Adriano) pp 225-255 ( Science Reviews , Northwood: U n ited Kingdom ) . 1 96 References Tre m e l , A , 1 996. Transfert d u tha l l i um d u sol vers l a plante. P h D thesi s , U n iversity of N ancy, France. Trem e l , A , Masso n , P . , Sterckema n , T. , Aize , D . and M e n ch , M . , 1 99 7 . Tha l l i u m i n French agrosystems - I . Tha l l i u m contents i n arable s o i l s . Environmental Pollution, 95: 293-302. Trem e l , A and M e n ch , M . , 1 997. Le thal l i u m d a n s le sol s et les vegetaux s uperieures. 1 1 . Le tha l l i u m dans le vegetaux s u perieures, pp 261 -26 9 , Agronomie v 1 7 . U nited States E nv i ronmental Protection Age ncy, 1 993. Toxics Release Inventory, U n ited States Government Printing Office, Was h i ngton 0 C U S E P AI7 45/R-93/003 . Ure n , N . , 1 99 2 . Form s , react i on s , a n d avai l a b i l ity of N i i n soi l s . Advances in Agronomy, 48: 1 4 1 - 1 95. Van H oek, Theo H. J . , 1 995. Thiocyanate , I n o rganic. In Ullmann's Encyclopedia of Industrial Chemistry, Vol. A 26 pp 759 - 766 (VC H Ver l g sgesell schaft, Wei n h e i m ) . Vass i l , A D . , Kap u l n i k , Y. , Raski n , I . and Sa lt, D . E . , 1 99 8 . The r o l e of E DTA i n l e a d transport a n d accum ulation b y I nd ia n m u stard . Plant Physiology, 1 1 7: 447-453 . Vazquez, M . D . , Barce l 6 , J . , P oschnri eder, C . , Mad i co , J . , Hatton , P . , B a ker, A J . M . and Cope, G . H . , 1 992. Loca l ization of zinc a nd cad m i u m i n Thlaspi caerulescens ( B rassicaceae ) . a metal l ophyte that can hyperaccum u l ate both meta l s . Journal of Plant Physiology, 1 40: 350-355. Wa l lace, A , Romney, E . M . , C h a , J . W. , S o ufi , S . M . and Chaudhry , F . M . , 1 97 7 . N ickel phytotoxicity i n relationsh i p t o s o i l p H m a n i p u lation a n d chelating a gents. Communications in Soil Science and Plant Analysis, 8: 757-764. Ward , N . ! . , B rooks, R R and Reeves, R D . , 1 976. C opper, cad m i u m , l ead a n d z i n c i n soi l , strea m sed i ments and vegetati on arou n d t h e Tui M i n e , Te Aroha . New Zealand Journal of Science, 1 9: 8 1 -89. Ward , N . L , B rooks, R R and R oberts, E . , 1 97 7 . Conta m i nation of a pasture by a N ew Zealand base-metal m i n e . New Zealand Journal of Science, 20: 4 1 3-4 1 9 . Warre n , H . V. a n d D e l avau lt, R E . , ( 1 950 ) . G o l d a n d s i lver content of some trees and horseta i l s i n B ritish C o l u m b i a . Geological Society of America Bulletin, 61 : 1 23 - 1 2 8 . 1 97 References Web b , M . ( Ed ) , 1 97 9 . The Chemistry, Biochemistry and Biology of Cadmium, Topics in Environmental Health Volume 2, 465 pp ( E l sevier/North-Hol land B i om e di cal P ress: Amsterd a m ) . Webster, J . G . , 1 986. The s o l u b i l ity o f gold and s i l ver i n t h e system Au-Ag-S- 02-H20 at 25°C a n d 1 atm. Geochemica et Cosmochemica Acta, 50: 1 83 7 - 1 843. Whitten, M . G. and R itchi e , G. S . P . , 1 99 1 . C a l c i u m chloride extractab l e cad m i um as a n estimate of cadm i u m u ptake by s ubterran e a n c lover. Australian Journal of Soil Research, 29: 2 1 5-22 1 . W i l l i a m s , D . R . , 1 97 3 . L ife' s e ssent i a l elements. Education in Chemistry, 1 0: 56- 80. Woo d , S . A. , 1 995. The rol e of humic substance s i n the transport and fixat i o n of meta l s of econo m i c i nterest (Au, Pt, P d , U, V ) . Ore Geology Reviews, 1 1 : 1 -3 1 . Zanders, J . M . Z . , 1 99 8 . Stu d i e s o n the Orig i n , D i stribution and M o b i l i ty of C a d m i u m in pastoral S o i l s . P h D Thes i s , M assey U n i ve rs i ty, N ew Zeal a n d , 259 pp. Zhao, F . J . , S hen, Z. G . and McGrath , S . P . , 1 998. S o l u b i l ity of zinc and i n te ractions between z i n c a n d phosphorus i n the hyperaccum u l ator Th/aspi caerulescens. Plant Cell and Environment, 21 : 1 08-1 1 4. 1 98 Appendices Appendices Appendix la : Methods for atomic absorption spectroscopy Element Lamp current (mA) Wavelength (nm) Slit Width (nm) Cadmium (Cd) Cobalt (Co) Gold (Au) Copper (Cu) Lead (Pb) Zinc (Zn) 3.0 6 .0 5.0 3.0 5.0 5.0 228.8 240.7 242.8 324 .7 21 7.0 21 3.9 0 .5 0 .2 0 .5 0 .5 1 .0 0 .5 Programme used for the determination of cadmium using GFAAS. For graphite furnace and phosphoric acid (1 000 mg/kg) as a modifier: Step number Final temp Ramp time Hold time Gas type Signal (0C) (s) (s) read 1 20 1 .0 0 .0 N2 2 95 20.0 20.0 N2 3 500 1 5 .0 1 0 .0 N2 4 1 500 1 .0 3.0 None 5 2000 1 .0 2.0 None Peak height Multiple loadings were dried to step number 2. Programme used for the determination of gold using GFAAS. For graphite furnace and MIBK extract: Step number Final temp Ramp time (oC) (s) 1 70 5.0 2 1 20 6 .0 3 1 80 5 .0 4 300 5.0 5 2500 1 . 1 Multiple loadings were dried to step number 2. Hold time (s) 2 .0 1 0 .0 3 .0 1 5 .0 2 .0 For graphite furnace and aqueous extraction analyte: Step number Final temp Ramp time Hold time (0C) (s) (s) 1 70 5.0 2 .0 2 1 20 6 .0 1 0 .0 3 1 80 5.0 3 .0 4 500 5 .0 1 5 .0 5 2500 1 . 1 2 .0 Multiple loadings were dried to step number 2 . Gas type Signal read Peak height Gas type Signal read N2 N2 N2 N2 None Peak height 1 99 Appendices Appendix Ib: Physical characteristics of experimental substrates Auby Soil The pedological characteristics of contaminated soils of the Auby region have been described in Perdix et al. ( 1 997). A summary is as follows: Depth Horizon Texture Pb Zn Cd Cu Fe Ca 0-8 cm A1ca A 1 -A1 s 1 030 995 1 9 44 45640 22100 1 0-25 cm A2ca A 1 -A1s 830 750 1 5 37 46550 21900 35-45 cm ACcag AL 18 75 3 9 50450 22650 55-80 cm I IC2cag L-LA <17 35 <2 5 23700 58800 Notes: 1 . Metal concentrations are in mg/kg. 2. Soils described are classified as colluviosol redoxisol . 3 . The characteristics summarised here are for field samples collected 1 km NE of the metal smelter. Auby soils on which field experiments for this thesis were conducted were located 1 00 m NE of the smelter, and hence show significantly greater contamination. Further details specific to the Auby site for this thesis can be found in Appendix 7. Commercial Seed Raising Mix The commercial seed raising mix used in Chapters 4, 5 and 6 was 'Liddle Wonder' mix with the fol lowing composition: 2/3 peat w/w 113 pumice w/w per m3 1 kg nutricote or osmocote 5-6 month release fertiliser 3 kg agricultural lime 2 kg dolomite 1 kg superphosphate 1 m3 of mixture weighs 600 kg. Wairarapa Soil The soil at the Wairarapa field site has been mapped and appears on the National Water and Soil Conservation Organisation - New Zealand Land Resource Inventory Worksheet number 1 5 8 (Masterton) . The specific soil present has been described by Heine ( 1 975) as Pirinoa hill soil (Atua silt loam, hill soil; soil set number 29H), a moderately leached intergrade between yelIow­ grey earths and yellow-brown earths of low P and Ca nutrient status. Further chemical characteristics of the soil may be found in Appendix 7. Reference Heine, le. 1975. Interim report on soils ofWairarapa Valley, New Zealand. Restricted Soil Bureau Internal Report, N.Z. Soil Bureau Record 40. 200 Appendices Appendix 2: Plant health at the time of harvest for soils contaminated with Cd, Pb and Zn (Chapters 3 - 5). Lead experiment 1 -Brassica juncea Lead salt Control Acetic acid Citric acid EDTA Control No signs of Treatment No change in No change in chlorosis caused necrosis plant health plant health relative to relative to control control Carbonate No sign of Treatment No change in Chlorosis caused chlorosis caused necrosis plant health by treatment, but relative to plants remained control healthy and turgid N itrate Very poor growth Treatment Treatment Treatment - stunted and caused necrosis effected necrosis effected necrosis unhealthy plants Oxide No sign of Treatment No change in Chlorosis caused ch lorosis caused necrosis plant health by treatment, but relative to plants remained control healthy and turgid Phosphate No sign of Treatment No change in Chlorosis caused chlorosis caused necrosis plant health by treatment, but relative to plants remained control healthy and turgid Su lphate No sign of Treatment No change i n Chlorosis caused ch lorosis caused necrosis plant health by treatment, but relative to plants remained control healthy and turgid Sulphide No sign of Treatment No change in Chlorosis caused chlorosis caused necrosis plant health by treatment, but relative to plants remained control healthy and turgid 2. Thlaspi caerulescens. Plants grew poorly in the nitrate salt. Treatment of the nitrate phase soil with acetic acid, citric acid or EDT A caused necrosis. For all remaining mineral phases the plants grew well, with no apparent sign of chlorosis. Acetic acid treatment caused necrosis of the plants but there was no visible change in plant health through treatment with citric acid or EDT A. 201 Appendices Cadmium experiment 1 . Brassica juncea Cadmium salt Control Citric acid EDTA Control No sign of chlorosis No change in plant No change in plant health relative to the health relative to the control control Carbonate No sign of chlorosis No change in plant Treatment caused health relative to the some leaf abscission control and a general increase of ch lorosis Nitrate Visible chlorosis of No change in plant Treatment caused the leaves, but plants health relative to the necrosis of the plants g rowing control Oxide No sign of chlorosis No change in plant Treatment caused an health relative to the increase in chlorosis control but no leaf abscission Phosphate Visible chlorosis of No change in plant Treatment caused the leaves slower health relative to the 50% leaf abscission growth relative to the control and a significant other metal phase general increase of soils chlorosis Sulphate No sign of chlorosis No change in plant No change in plant health relative to the hea lth relative to the control control 2. Cardaminopsis halleri This species grew well in all the soils for this experiment prior to treatment. There were no visible signs of chlorosis on the foliage. Citric acid caused no change in plant health. EDTA caused minor chlorosis and a loss of turgid pressure for plants growing on the carbonate and oxide phase soils. 3 . Thlaspi caerulescens This species grew weII in aII the soils for this experiment prior to treatment. There were no visible signs of chlorosis on the foliage. Citric acid effected no change in plant health. EDTA caused minor chlorosis and a loss of turgid pressure for plants growing on the sulphate phase soil. Zinc experiment Brassica juncea failed to grow in any of the zinc soils of this experiment. Cardaminopsis halleri and Thlaspi caerulescens showed signs of chlorosis and stunted growth on the sulphate phase soil . EDTA treatment caused necrosis of these plants. For the remaining phases, growth was unchanged by the form of cadmium present in the soil relative to the control both before and after EDT A treatment . 202 Appendices Appendix 3: Unsuccesful induced-gold uptake experiments described in Chapter 11. Plant species Substrate Treatment Notes used Chicory Waih i ore, SCN - 0, 0 .3 , 0.6, 1 .0 g/kg SCN treatment 1 998 batch - caused necrosis 2.5% l ime S203 - 0, 0 .3 , 0.6, 1 .0 g/kg Plants remained healthy EDT A - 0, 0.3, 0.6, 1 ,0 g/kg Plants remained healthy Brassica juncea Macraes ore SCN - 0, 0.2, 0 .5 , 0 .8 , 1 .0 g/kg SCN treatment caused necrosis S203 - 0 , 0.2, 0 .5 , 0 .8 , 1 .0 g/kg plant remained healthy Brassica juncea Macraes ore SCN - 0 . 1 g/kg citric acid with 2g or 4g citric acid treatement caused necrosis S203 - 0 . 1 g/kg citric acid treatment with 2g or 4g citric acid caused necrosis Cardaminopsis Tui tal ings - SCN - 0 .25 g/kg SCN treatment halleri 2.5 % l ime caused necrosis S203 - 1 g/kg plant remained healthy EDTA - 2 g/kg plant remained healthy DTPA - 2 g/kg plant remained health� 203 Appendices Appendix 4: Induced-gold uptake experiment for Brassicajuncea growing on limed Waihi ore, discussed in Chapter 11 Lime rate Treatment 0 (5 . 1 ) control SCN S203 SCNt S203 0 .5 (6 .4) control <5 SCN 25 9 S203 47 1 4 SCNI S203 30 1 .0 (7.3) control 1 8 SCN 1 2 34 S203 1 9 26 SCNI S203 45 1 .5 (7.3) control 1 6 SCN 21 S203 28 SCNt S203 1 5 2.0 (7.3) control SCN S203 SCNt Note: nd signifies that the plants for s treatment died before appl ication. The assistance of Anthony Kirk is gratefu l ly acknowledged in the preparation of this data set. Appendix 5: Data used to construct Figure 11.1. country Location Principal minerals Ni content pH (mg/kg) New Zealand Dun Mountain Quartz, goethite 2244 6.7 New Zealand Rai Val ley Antigorite, goethite, q uartz 2 1 09 7.5 New Caledonia Kouaoua (Mea M ine) Antigorite , goethite, q uartz 1 7 208 7.2 Spain Ojen Antigorite, quartz, ol ivine 1 800 7.8 Spain Puente Basadre Antigorite, quartz, talc 2 1 00 6.3 Italy Firenze (Tuscany) Antigorite, quartz 1 609 6.9 Morocco Taafat Antigorite, chlorite ,ol ivine 1 700 7.5 Source Robinson et al. ( 1 999). 204 Appendices Appendix 6: Plant references Agrostis tenius L. Alyssum bertolonii Desv. Alyssum malacitanum T. Dudley Arrhenatherum elatius L. Berkheya coddii RoessI. Brassica juncea (1...) Czern. Brassica napus L. Cardaminopsis halleri (L). Hayek Cardaminopsis petrea L. Grevillea acuaria F. Muell . Ex Benth. Hybanthus epacroides subsp. bilobus Bennett Hybanthus jloribundus (Lindl .) F . Muel!. subsp. adpressus Bennett subsp. curvifolius Bennett subsp . . flori bundus Bennett Iberis intermedia Guersent Impatiens balsamina L. Impatiens holstii Engler & Warb. Lupinus albus L. Millota myosotidifolia F. Muel l . Minuarta verna (L. ) Hiern. Personia metallifera Wild Pimelia leptospermoides F. Muel! . Sebertia accumuinata Pierre ex Baillon Senico coronatus (Thunb. ) Harv. Silene humilis L. Stackhousia tryonii Bailey Streptanthus polyxaloides Gray Thlaspi caeru lescens J .e. & R.Presl. Zea mays L. Appendix 7: the complete summary data set for this thesis i s saved on the accompanying disk, or (if the disk is not attached) can be obtained from: Soil and Earth Sciences Institute of Natural Resources Massey University Palmerston North New Zealand 205