CHARACTERIZATION, LEACHABlLlTY AND AClD MINE DRAINAGE POTENTIAL OF GEOTHERMkL SOLID RESIDUES

Genandrialine Laquian Peralta

A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy Department of Chemical Engineering and Applied Chemistry University of Toronto

Q Copyright by G. L. Peralta. 1997 C\U~Y~~.IYIIY Y. mw ..--, ------Bibliographic Setvices services bibliographiques 395 Wellington Street 395, nie Wellington Ottawa ON KIA ON4 Ottawa ON K1 A ON4 Canada Canada

The author has granted a non- L'auteur a accordé une licence non exclusive licence allowing the exclusive permettant a la National Library of Canada to Bibliothèque nationale du Canada de reproduce, loan, distribute or sel1 reproduire, prêter, distribuer ou copies of this thesis in microform, vendre des copies de cette thèse sous paper or electronic formats. la forme de microfiche/film, de reproduction sur papier ou sur format électronique.

The author retains ownership of the L'auteur conserve la propriété du copyright in this thesis. Neither the droit d'auteur qui protège cette thèse. thesis nor substantial extracts from it Ni la thèse ni des extraits substantiels may be printed or otherwise de celle-ci ne doivent être imprimés reproduced without the author' s ou autrement reproduits sans son permission. autorisation. Geothermal Solid Residues PhD 1997 Genandrialine Laquian Peralta Department of Chernical Engineering & Applied Chemistry University of Toronto

ABSTRACT

A solid waste is classified as hazardous if it contains sufficient leachable components to contaminate the groundwater and the environment if disposed in a landfill. Solid residues from three geothermal fields (Bulalo, Philippines, Cerro Prieto, Mexico, and Dixie Valley, USA) containing S, Cu, Zn, and Pb at levels above earth's crustal abundance were studied for their leachability. Several procedures were used to assess the potential mobility of the elernents - protocol leaching tests, sequential chemical extraction, and acid mine drainage potential test. In addition, whole rock analysis, optical and electron microscopy, X-ray diffraction, radioactivity counting, and toxicity testing were also performed. The geothermal residues are mostly silica (70%) with trace elements and have varying crystalline and amorphous character. All the samples tested can be classified as nonhazardous since their leachate quality is below the regulatory limits. Toxicity tests were negative and radioactivity counts were within norm. Sequential extraction indicated the potential for metal release into the environment but only under extreme conditions (pHs 2, 85-175 OC). Batch kinetic tests identified that leaching of Pb in the presence of oxygen is diffusion-controlled with a rate equation, r = 4.6 x 10" a, t "" + 1.1 x 1O4 a, t''" mol.L'' .h-' . A batch reactor technique (AMDP test) using the iron and suifur oxidizing bacteria, Thiobaci//us ferrooxidans, was developed for geothenal wastes to predict their acid mine drainage and bioleaching potential. It was observecf that almost 100% of Cu and Zn in the Mexican scale and less than 2% in the Philippine scale and sludge were released. Despite a significant Pb content, only <6% leached from the Mexican scale. Geochemical thermodynamic modelling using MINTEQAZ showed that much heavy metal content must be inaccessible to the leach medium. The hazard and risk involved from geothermal residues were assessed to be very low for the Mexican sludge and drilling mud and the American scale. However, a low but manageable risk was attributed to the Philippine scale and sludge. The fine sized Mexican scale was found to have medium risk that will require special handling prior to landfill disposal. I grareruiiy acrtnowieage rne Toiiowing wno nave conrriDutea ro my rnusruaies: My supervisor Prof. Donald W. Kirk who was ever wise, helpful, patient, and friendly. I was fortunate for having done my PhD under his supervision. My Reading Cornmittee - Prof. Robert E. Jeivis, Prof. Vladirniros G. Papangelakis, and Prof. Patricia L. Seyfried with the other examinen: Prof. D. Grant Allen, Prof. Grant Ferris and Prof. Kostas Fytas (Universite' Laval, Quebec) for their comments and advice. Professon, colleagues and friends who provided valuable contribution to my thesis in tens of equipment, comments, and services: Prof. Greg Evans and Dr. Sandu Sonoc for radioactivity counting, Prof. Patricia L. Seyfried and Ms. Sheree Yin for the toxicity tests, Dr. John W. Graydon for the photomicrographs, XRAL Laboratory for elemental and bulk analyses, Dr. Srebri Petrov for X-ray diffraction, Mr. Battista for transmission electron microscopy, Mr. Fred Neub for assistance with light microscope and videotaping, Prof. Grant Allen for the incubatorlshaker, Dr. Dmitri Rubisov for his suggestion on particle sire analysis and leaching experiments, Dr. Karen Liu for support in optical microscopy with image analysis, Mr. Durga Prasad for assistance with the autoclave and initial bacteria culture, Mr. Jeff Bain and Prof. Charles Jia for assistance in geochemical modelling, Dr. Martin H. Birley for introducing Endnote reference software, Dr. Taylor Eighmy for introducing MINTEQA2, and Mr. Rene C. Peralta for computer support and maintenance. O The World Bank and the University of the Philippines (UP) for my scholarship especially Dr. Francisco L. Viray and Dr. Reynaldo Vea of UP as well as Dr. Estrella F. Alabastro, Ms. Lydia Tansinsin, and Ms. Teody Dayoan of DOSTESEP. The geothermal community especially Dr. Marcelo Lippmann of Lawrence Berkeley Laboratoiy, California for his advice, networking and assistance. Geothermal companies who provided samples for this study: Philippine Geothermal Inc., Oxbow Geothermal Corp. Nevada, USA, and Gerencia de Proyectos Geotermoelectricos, CFE of Mexico. Colleagues and friends in the laboratory notably Dr. John Graydon, Mr. Cam Nhan, and Mr. Chris Chan for being there to listen and iend assistance. The administrative staff of our department and the International Student Centre especially Liz Paterson. My dear friends whom I cannot al1 mention here but are listed in my Christmas card directory. Some friends who through e-mail have sent technical and moral support especially Ms. Jane Y. Gerardo, Dr. Efren F. Abaya, Dr. Martin H. Birley, Mr. Robert Bos, Mr. Florencio Ballesteros, Mrs. Dionisia Ali, Dr. Keryn Lian, and Dr. Michael Gattrell. Relatives particularly rny parents Antonio and Gloria Laquian and parents-in-law Paterno and Remedios Peralta krtheir prayers and full support. My wonderful family - husband Gil Renato (Rene), my two sons, Kevin and Patrick - for their invaluable support, patience, understanding, affection, massages, and share of household chores. I dedicate this thesis to them. Abstract ii Acknowledgments iii Table of Contents iv List of Figures viii List of Tables x Nomenclature xi CHAPTER 1 INTRODUCTION 1 1.1 Background 1 1.2 Objectives 2 1.3 Thesis Overview 2 CHAPTER 2 BACKGROUND AND LITERATURE REVIEW 3 2.1 Environmental Impacts of Geotherrnal Residues 3 2.1.1 Geothermal Energy and its Environmental Impacts 3 2.1.2 Treatment and Disposal Practices of Geothermal Residues 6 2.2 Guidelines for Waste Classification and Regulation 7 2.2.1 Average Crustal Abundance of Elements 7 2.2.2 .Leachate Quality Criteria 8 2.2.3 Permissible Heavy Metal Concentrations for Agricultural Use 9 2.2.4 Naturally Occurring Radioactive Materials IO 2.3 Techniques for Waste Characterization IO 2.3.1 Chemical Analysis IO 2.3.2 X-ray Diffraction 11 2.3.3 Radioactivity Counting 12 2.3.4 Optical Microscopy 12 2.3.5 Evaluation of Toxicity 12 2.3.6 Weathering Tendency 13 2.4 Leaching Protocol Tests 13 2.4.1 Principles of Leaching 13 2.4.2 Batch Versus Column Leaching 16 2.4.3 Agitated Extraction Procedures 16 2.4.4 Sequential Chernical Extraction 2.5 Microbial Leaching 2.5.1 Acid Mine Drainage 2.5.1.1 Occurrence 2.5.1.2 Role of Microorganisms 2.5.2 Mechanism of Bacterial Leaching 2.5.2.1 Direct Method 2.5.2.2 Indirect Method 2.5.3 Factors Affecting Bacterial Leaching 2.5.3.1 Composition of Leaching Medium 2.5.3.2 Oxidation-Reduction Potential 2.5.3.3 Temperature 2.5.3.4 Hydrogen Ion Concentration 2.5.3.5 Agitation and Oxygen Transfer 2.5.3.6 Particle Size and Substrate Concentration 2.6 Prediction of Acid Mine Drainage Potential 2.6.7 Prediction Procedures 2.6.1.1 Static (Initial) Tests 2.6.1.2 Kinetic (Confirmation) Tests 2.6.2 Laboratory Scale Bioleaching Techniques 2.6.2.1 Stationary Flask Technique 2.6.2.2 Shake Flask Technique 2.7 Geochemical Equilibrium Modeling 2.7.1 Different Thermodynamic Models 2.7.2 Applications CHAPTER 3 METHODS AND PROCEDURES 3.1 Waste Characterization 3.1.1 C hemical Analysis 3.1.2 Radioactivity Counting 3.1.3 X-Ray Diffraction 3.1.4 Optical Microscopy 3.2 Toxicity Testing 3.3 Sequential Chemical Extraction 3.4 Accelerated Weathering Test 3.5 Protocol Leach Tests 3.5.1 Leachate Extraction Procedure (LEP) 3.5.2 Toxicity Characteristic Leaching Procedure (TCLP) 3.6 Extended Leach Tests 3.6.1 Oxic Conditions 3.6.2 Anoxic Conditions 3.7 Preliminary Acid Mine Drainage Potential Test 3.8 Acid Mine Drainage Confirmation Test 3.8.1 Bacteria Culture and Medium 3.8.2 Acclimation of Inoculum 3.8.3 Acid Mine Drainage Potential Test 3.9 Batch Kinetic Experiments 3.9.1 Effects of Agitation, Temperature, and Sterilization 3.9.2 Monitoring and Sampling 3.10 Microstructural Analysis 3.10.1 Light Microscopy with Image Analysis 3.10.2 Transmission Electron Microscopy 3.1 1 Geochemical Modeling CHAPTER 4 RESULTS AND DISCUSSION 4.1 Waste C haracterization 4.1 .1 Chemical Analysis 4.1.2 Radioactivity 4.1.3 X-ray Diffraction 4.1.4 Optical Microscopy 4.1.5 Toxicity Testing 4.1.6 Weathering test 4.2 Protocol Leaching Tests 4.3 Extended Leach Tests 4.3.1 Oxic Conditions 4.3.2 Anoxic Conditions sequential Chemicai txtraction Preliminary Acid Mine Drainage Potential Test Confirmation of Acid Mine Drainage Potential 4.6.1 Bacterial Growth and Acclimation 4.6.2 Acid Mine Drainage Potential (AMDP) Test Batch Kinetic Experiments 4.7.7 Effect of Sterilization 4.7.2 Effect of Agitation and Temperature 4.7.3 Effect of Bacteria Reaction Rates and Mechanisms 4.8.1 Chemical Leaching Kinetics 4.8.2 Metal Solubilization Rate Cornparison Between Acid Leaching and Bioleaching 4.9.1 Between TCLP and LEP 4.9.2 Between TCLP and AMDP 4.9.3 Evaluation of the AMDP Procedure Geochemical Modeling Sumrnary Risk Assessrnent of Geothermal Residues CHAPTER 5 CONCLUSIONS CHAPTER 6 PROPOSED FUTURE STUDIES REFERENCES APPENDICES A -Trial Experiments Prior to Procedure Development B - Bacterial Density Estimation for Thiobacillus ferrooxidans C - Calculation for the Preliminary Acid Mine Drainage Potential Results D - About the Geochemical Model MINTEQAZ E - Input Data Derivation for Geochemical Model F - Sample Output of Geochemical Modelling G - Uoff Acid Mine Drainage Potential Test H - Results of Toxicity Testing 1 - Additional Dissolution Kinetics Data Fig. 2.1 Schematic model of a hot-water geothermal system Fig. 2.2 Schematic drawing showing the leaching of an ore particle Fig. 4.1 Photomicrographs of Philippine scale, PSC (1OOx, 250x) Fig. 4.2 Photomicrograph of Philippine sludge, PSL (50x) Fig. 4.3 Photomicrograph of American scale, ASC (50x) Fig. 4.4 Photomicrograph of Mexican drilling mud, MDM (50x) Fig. 4.5 Photomicrograph of Mexican scale, MSC (1000~) Fig. 4.6 Photomicrograph of Mexican sludge, MSL (1OOx, 50x) Fig. 4.7 Comparison of extent of leaching between LEP and TCLP Fig. 4.8 Comparison of extended TCLP leaching of Pb coarse and fine Fig. 4.9 Dissolution behavior in the oxic extended TCLP Fig. 4.1 O Dissolution behavior in the anoxic extended TCLP Fig. 4.1 1 Sequential extraction results for PSC and PSL Fig. 4.12 Sequential extraction results for ASC and MDM Fig. 4.13 Sequential extraction results for MSC and MSL Fig. 4.14a Photomicrograph of T. ferooxidans : high density (800x) Fig. 4.14b Photomicrograph of T. ferooxidans : medium density (800x) Fig. 4.14~ Photomicrograph of T. fenooxidans : low density (800x) Fig. 4.14d Photomicrograph of T. ferooxidans with a cloud-like capsule Fig. 4.15 TEM photomicrograph of various sizes of T. ferrooxidans (60K x) Fig. 4.16 TEM photomicrograph of T. fenooxidans about to partition (99K x) Fig. 4.17 pH-Eh change over tirne in AMD experiments Fig. 4.18 Leaching in agitated AMD experiments : PSC, PSL and MSC Fig. 4.19 Leaching in stationary AMD experiments : PSC, PSL and MSC Fig. 4.20 Variations in pH over time for batch kinetic AMD experiments Fig. 4.21 Variations in Eh over time for batch kinetic AMD experiments Fig. 4.22 Effect of sterilization on metal bioleaching : PSC Fig. 4.23 Effect of sterilization on metal bioleaching : PSL Fig. 4.24 Effect of sterilization on metal bioleaching : MSC Fig. 4.25 Effect of agitation and temperature on metal bioleaching: PSC Fig. 4.26 EfFect of agitation and temperature on metal bioleaching : PSL Fig. 4.27 Effect of agitation and temperature on metal bioleaching : MSC 97 Fig. 4.28 Effect of bacteria on metal bioleaching : PSC 98 Fig. 4.29 Effect of bacteria on metal bioleaching : PSL 99 Fig. 4.30 Effect of bacteria on metal bioleaching : MSC 1O0 Fig. 4.31 Dissolution kinetics of Pb in Mexican scale 102 Fig. 4.32 Overall metal bioleaching, percent over time for MSC 103 Fig. 4.33 Solubilization rate for Cu, Zn, and Pb in the Mexican scale 104 Table 2-1 Average abundance of the elements in crustal rocks Table 2-2 Leachate quality criteria Table 2-3 Limits of heavy metal concentration in sludges for agriculture Table 2-4 Allowable values for metals in the European Union Table 2-5 Summary of sorne static AMD test methods Table 2-6 Summary of some kinetic AMD test methods Table 3-1 Summary procedure for sequential chemical extraction Table 3-2 Comparison of protocol leaching tests Table 3-3 Culture medium for Thiobaci//us fenooxidans Table 3-4 Input data for modeling protocol leach tests Table 4-1 Chernical analysis of selected geothermal samples Table 4-2 Crustal abundance ratio of selected geothermal residues Table 4-3 Concentrations of radionuclides in the geothermal residues Table 4-4 X-ray diffraction data of selected geothermal residues Table 4-5 Preliminary acid mine drainage potential test results Table 4-6 Maximum solubilization rate in batch process Table 4-7 Comparison between BC Research Confirmation and AMDP Tests Table 4-8 Sumrnary of results from geochemical modeling of the Mexican scale 109 Table 4-9 Hazard and risk rating for Mexican geothermal residues 114 Table 4-10 Hazard and risk rating for American and Philippine geothermal residues 715 Seothemal samples :

PSC - Philippine scale PSL - Philippine sludge ASC - American scale MRM - Mexican drilling mud MSC - Mexican scale MSL - Mexican sludge

Terms and Procedures :

ABA - acid base accounting AC - acid consumption AMD - acid mine drainage AMDP - acid mine drainage potential APP - acid production potential ARD - acid rock drainage BCRIT - British Columbia Research initial test BCRCT - British Columbia Research confirmation test DMSO - dimethylsulfoxide DSTP - direct sediment testing procedure ICP - inductively coupled plasma mass spectrometry LEP - leachate extraction procedure LOI - loss on ignition NORM - naturally occuring radioactive materials ORP - oxidation - reduction potential SCE - sequential chernical extraction TCLP - toxicity characteristic leaching procedure TEM - transmission electron microscopy XRF - X-ray fluorescence XRD - X-ray diffraction

Institutions :

AECB - Atomic Energy Control Board ATCC - American Type Culture Collection APHA - American Public Health Association BNL - Brookhaven National Laboratory CANMET - Canada Centre for Mineral and Energy Technology EU - European Union (formerly European Community) ICRP - International Commission on Radiological Protection UNSCEAR - United Nations Scientific Cornmittee on the Effects of lonizing Radiation Uoff - University of Toronto USEPA - United States Environmental Protection Agency WHO - World Health Organization WC - Wastewater Technology Centre Hac - acetic acid Ac - acetate ion Co- original concentration (mol. L1) k, - rate constant for coane fraction (mo1.L-'. h -%) 4 - rate constant for fine fraction (rno1.L-'. h -') a, - coarse fraction ' a, - fine fraction (1 -a,) t - leaching time, h Bq - Becquerel Ci - curie pCi - picocurie pSv - microsievert pm - micron ppm - part per million

Particle size :

-125 pm = less than 125 pm -4mm = less than 4 mm -9.5 + 6 mm = between 6 - 9.5 mm

Conversion:

1 Bq = 27 pCi 1 pCi = 1 x 10''2Ci 1 pm = IO4m 1 ppm = 1 mg/L (in dilute solution) CHAPTER 1 INTRODUCTION

1.5 Background Development of appropriate waste management methods requires fundamental understanding of the physicochemical properties and leaching behavior of the waste material. Multidisciplinary techniques from chernical engineering, metallurgy, hydrometallurgy, geology, microbiology, process mineralogy, biohydrometallurgy and environmental engineering can be utilized to develop an acceptable waste characterization program. Time and money can be saved in waste management if there was a thorough knowledge of the characteristics of the waste, the long-term leaching behavior and an assessment of the risk involved. The same principles are applicable to geothermal wastes. Geothermal energy has received increasing attention as an attractive alternative to fossil-fueled energy sources since it is economical and creates less environmental pollution. It is widely used in almost 25 countries. The exploration and utilization of geothermal resources generate solid residues such as scale, sludge, and drilling mud. In this study, geothermal residues were studied in liquid-dominated geothermal systems in the Philippines (scale and sludge) and Mexico (scale, sludge, and drilling mud). Scale only was obtained from the US as sludge was not generated in that vapor-dominated geothermal system. Since these solids contain Fe, Cu, Zn, and Pb at levels above normal soils, they have been labelled as hazardous and require special treatment and disposal. There is very little information on geothemal residues with respect to their characteristics, leaching behavior, and environmental hazard. This study will contribute largely to the understanding of their true nature in order to be able to recommend appropriate measures for their management. In disposing geothermal residues in a landfill, the most serious threat to the environment is leaching of the toxic components (such as heavy metals) to groundwater. There is concern that the leachate will contaminate the aquifer which will eventually affect the human population through ingestion if such an aquifer was used for drinking water or for irrigation of agricultural lands. Further investigation of conditions and mechanisms under which metals might be eventually released would clarify any potential environmental contamination. 1.2 Objectives The main objective of this thesis was to understand the leaching behavior of geothemal residues in a landfill environment. It will consider microbial action in evaluating the environrnental impact of disposing these wastes on land. Specifically, this work will predict and estimate the possible environmental effects of geothermal wastes in a landfill through the study of their (a) physico-chemical characteristics, (b) leachability, and ( c) acid mine drainage potential.

1.3 Thesis Overview The thesis is composed of six chapters and an appendix. The first chapter is the introduction which is essentially a thematic foreword about the research along with the objectives and expected output. The second chapter is a brief background of the study as well as a review of related literature specifically about geotherrnal energy and its residues, environrnental impacts, acid and microbial leaching, protocol tests. regulatory limits, acid mine drainage and geochemical modelling. The third chapter describes of the experimental work, computer modelling, materials used, methods, measurements, and analytical techniques. In Chapter 4, under results and discussion, experimental and computational results are given along with a discussion of their implications in relation to the objectives. The conclusions arising from the study are listed in Chapter 5 while the proposed future studies are in Chapter 6. At the end of the six chapters are nine appendices supporting the main body of the thesis. CHAPTER 2 BACKGROUND AND LITERATURE REVIEW

2.1 Environmental Impacts of Geothermal Residues The following section is an introduction to geothermal energy and some of its important environmental concerns. Past and present waste management practices are also summarized.

2.1.7 Geothennal Energy and its Environmental Impacts Geothermal energy for power generation has received increasing attention as an attractive alternate energy source both due to its environmental and economic advantages. Where it is abundant and economical to exploit, geothenal power has been used for a number of years to commercially generate electricity. Geothermal power dates back to 1913 in ltaly and has since spread over the Pacific Rim [Il.It is now an important source of power in more than 25 countries worldwide with potential in 40 countries [l, 21. The environmental impact of any electric power production system is reflected in the number and complexity of the steps in the fuel and production cycle. Since geothermal power plants use naturally occurring stearn, there is no need for the complex stearn- generating systems or extensive mining, processing, storage, or transportation facilities that are required for other thermal power plants. The creation of geothermal resources begins with a source of heat - hot or molten rock, lying close to the earth's surface as shown in Figure 2.1 [3]. The high temperature rock zone is overlain by a permeable rock formation containing water from precipitation which rises upward as it is heated by the rock below. Generally the flashed steam process is used for electricity generation from hot-water systems. In this process, as the hot water under a very high pressure, is pumped out of the reservoir by wells and as it nears the surface and the pressure decreases, about 20% of the fluid boils and "flashes" into steam. Separators are used to separate the steam from the water and the former is directed to turbines for power generation. Two types of geothermal system exist : liquid-dominated (hot water) and vapor-dominated (steam) with the former more common worldwide and having greater environmental concerns. The exploration and utilization of geothermal resources generate residues such as scale, sludge, and drilling mud. Scale is deposited in steam gathering systems, wellbores, separaton, and turbine blades and is manually removed during preventive maintenance shutdown. The water leaving the separators is available for further processing depending on its mineral content. It is woled and allowed to partially evaporate in a cooling or thermal pond. In this cooling pond at atmospheric pressure and lower temperature, silica precipitates and settles at the bottom of the pond to form a residual sludge. The supernatant liquid is either reinjected or discharged to a body of water. Drilling mud is a by-product of drilling operations during the exploration and development of the geotherrnal well field.

Figure 2.1 Schernatic model of a hot-water geothermal system, adapted from Muffler and White (1978). There is a widespread belief that geothemal resources represent a relatively "clean" non-polluting energy source, hence the increased public interest in geotherrnal developrnent. While it is true that geotherrnal resources offer significant environmental advantages over fossil and nuclear energy, there are also adverse local impacts on land, water, and air as summarized by Brown and several geothermal practitioners and El- Hinnawi [4,5]. It was reported that CO,, H,S, As, B, and Hg are species of most concern in geothermal power plants in New Zealand (61 and in the Philippines [7]. Bowen and Axtrnann had made two of the early studies on environrnental impact of a geothermal power plant which recognized emissions to air (CO, and H, S) and water (thermal discharges) as the main problems [8,9]. In the US, Hahn had identified exposure to heat and noise, H,S, NH,, hazardous chemicals and wastes as the major occupational health hazards associated with geotherrnal energy development from drilling to power production [IO). Relatively unknown prior to development and an aftermath of power generation is the formation of solid residues (especially scale and sludge) from several processes. These residues have not been widely studied but were reported to contain elevated amounts of trace elernents notably, arsenic, copper, zinc and lead which may be the result of the wastes being enhanced with metals from the rock formation or salts from the reservoir fluids [Il,121. For a 433 MW geothermal power plant at lrnperial Valley, California, an estirnated 50,000 tonslyear of production wastes were generated [12]. There is a scarcity of reports on the characteristics of scale and sludge precipitates in geothermal brines. There is hardly any literature about drilling mud. Dissolved silica is a major constituent in these geothermal effluents which are currently managed by means of holding ponds before the aqueous component is discharged to natural waterways or reinjected to geothermal wells [13]. Significant research effort on the chemistry of scales has been carried out and reported [14-171. Wong and Shugarman studied a silica-rich sludge containing high levels of lead, copper and zinc and reported ways of reducing their concentrations to acceptable levels [15]. Hickman reported the presence of arsenic, sulfur, and lead accumulating as cooiing tower basin sludge [16] which is quite different from the cooling pond sludge. The sludges produced by geothermal operations have been called by different narnes : geothermal residues, residual sludge, silica slurry, sump sludge, holding pond sludge and cooling tower sludge. The fint five terms refer to the same sludge, mainly silica containing various inorganic elements, which precipitates in the holding pond while the geothermal brine is cooling. The cooling tower sludge originates in the cooling tower basin. Several researchers have studied the characteristics of scale and sludge. Webster and Kukacka found large amounts of arsenic, lead, cadmium and çhromium in geothemal residues in the US [18]. Karabelas et al determined the characteristics of scales from a geothermal power plant in Milos, Greece and found them enriched with silicates as well as heavy metal sulfides of lead, zinc, copper and iron [19]. Gallup and Reiff characterized geothermal scale deposits from Salton Sea, California and Tiwi, Philippines, using Mossbauer spectroscopy and X-ray diffraction [20]. It was found that the bulk of the mineral phases present were microcrystalline, poorly crystalline or glassy. He concluded that the iron silicate deposited in scale was derived pnrnarily from geothermal brine with co- deposition of poorly-crystalline steel corrosion products. Premuzic et al have studied sludges for almost eight years but mostly for treatrnent purposes with little effort on characterization [Il, 21-25]. Most of these studies only examined the physicochernical characteristics without investigating the long term stability or mobility of the heavy metals under changing environmental conditions.

2.1.2 Treafrnent and Disposal Practices of Geothermal Residues Most geothermal residues have posed disposal problerns to geothermal operators since some have been reported to contain toxic elements at elevated concentrations above normal soils [12, 16, 20, 281. Landfilling is widely practiced as a disposal option since it is inexpensive and perceived to have low environrnental impact. Most effluents of liquid-dominated geothermal systems are generally elevated in dissolved salts, boron, ammonia, arsenic, and heavy metals 1291. Such effluents are currently managed by means of holding ponds which discharge to natural waterways or are reinjected back into suitable geothermal wells [13]. While reinjection is generally favored, it is very expensive (30% of total capital cost of steam-water gathering system) and can affect the energy potential of the geothermal resources through lowering of reservoir temperature [5] and plugging of the formation adjacent to the reinjection wells [8, 301. The popular disposal option practiced by rnany geothermal operators is landfilling onsite. Technologies such as solidification1stabilization [18], bioleaching of heavy metals 126, 271, and entombmenvlandfilling 1311 have been investigated to address the disposal problem of geothermal residues. Premuzic at al used biotechnology for thermophilic metal leaching using mixed cultures of Thiobacillus femooxidans and Thiobacillus thiooxidans at 55 OC using an agitated bioreactor [25]. Field operators in the Philippines are currently studying alternative methods of handling and permanently isolating the wastes [32]. In 1985 in the USA, Dobryn reported that an estimated one million dollars per year was being spent at a typical 50 MW geothermal power plant for disposal of 175 kglday of solid residues [33]. He also found the cost of the biological waste treatment plant (0.17- 0.23centslkwh) was the sarne as the cost of hauling the solid waste to a hazardous disposal site ($1,022,000 per year or 0.23 centslkwh). Either cost accounts for about 5% of the cost of generating electricity from geothermal power [33]. In addition, as the cost of disposat increases, the long term liabilities also increase at even greater proportion with lirnited space available for landfilling. The disposal cost for regulated wastes was five times the cost of non-regulated wastes and had doubled during the period 1985 to 1991 [25]. Brookhaven National Laboratoiy (BNL) paid US$ 550 per metric ton for disposal of geothermal sludge in 1991 while the corresponding non-regulated waste cost only US$ 110 per metric ton. Haulage or shipping cost for sludge having chromium, lead and radium was US$ 1400 per m3while treated sludge (without toxic components) was only US$270 per m3, providing a five-fold saving. It was reported that BNL obtained a 60070% savings in disposal cost using the biochemical waste detoxification technology but which nevertheless require intensive power supply and capital investment [26,271.

2.2 Guidelines for Waste Classification and Regulation Several known regulatory limits and criteria are presented in this section. These will be used later in the Results and Discussion for comparison purposes. Normal values for certain elements and radioactive materials in soi1 are also given.

2.2.7 Average Cnistal Abundance of Eiements The earth's crust consists almost entirely of oxygen compounds, especially silicates of aluminum, calcium, magnesium, sodium, potassium and iron [34]. In Table 2-1 are presented the data on the average abundance of the elements in the earth's crust that are of importance in environmental analyses. Table 2-1 Average Abundance of the Elements in Crustal Rocks

Major Crustal Minor Crustal Elements Average, % Elernents Average, ppm

2.2.2 Leacha te Quality Cntena The leachate quality criteria in Table 2-2 were derived by rnultiplying 100 times the WHO Drinking Water Standards [35,361. This is to account for dilution and attenuation effects in the groundwater if the waste is disposed in a landfill and the elements leach out. Several leaching protocol tests which will be discussed in Section 2.4 use these criteria to evaluate whether a certain waste is hazardous or not. If the concentrations in the extracted leachate are below these lirnits, then the waste is classified as nonhazardous and can be disposed as a non-regulated waste. Severai countries worldwide including the US, Canada, Philippines and Mexico use these criteria for regulatory purposes. Table 2-2 Leachate Quality Criteria, mglL

Element Concentration Element Concentration

2.2.3 Permissible Hea vy Meta1 Concentrations for Agricultural Use Sewage sludges may be disposed on agricultural lands provided that they meet the heavy metal concentrations shown in Table 2-3 which was reported by Weber et al and cited by Tyagi [37]. Only the values for Canada and the median for several European Union (EU) countries are listed here although limits are also given for Co, Mn, Mo, and Ni.

Table 2-3 Maximum Permissible Heavy Metal Concentration in Sludges Considered to be Acceptable for Agricultural Lands, mglkg dry weight

Element Canada EU Median 1 Elernent Canada EU Median As 75 10 Cc! 20 7 Cr - 1000 Cu - 1100

Moreover, in Table 2-4 is presented a summary of the norms providing guidelines to the European Union of concentration limits for soils and sludges as reported by Davis and cited by Tyagi [37]. Davis also observed that more than half of the sludges disposed did not conforrn to these guidelines. However, only metals in the dissoived form are available for plant uptake. Therefore, in addition to knowing the elemental composition of sludges, speciation through X-ray diffraction or sequential extraction is important in determining potential harm to the environment. Table 24 Lirnit Values for Metals in the European Union

2.2.4 Naturally Occumng Radioactive Materials In the Salton Sea, California, Gallup and Featherstone studied the control of naturally occurring radioactive materials (NORM) that precipitate from geothermal brines [38]. Their treatment efficiency was based on a reduction of radium concentrations in geothermal scales and sludges below 0.2 Bqlg (5 pCilg), an anticipated NORM regulation for solid wastes. Ra-226 and Ra-228 were present in concentrations ranging from 9 -15 Bqlg (250-400 pCi1g). Conversion units used are: 1 becquerel (Bq) is equivalent to 27 picocuries (pCi) while one picocurie is equivalent to 1 x 1012 Ci. The same limit of 5 pCi$ was used by Prernuzic et al in their work which removed NORM from geothermal sludges using bioleaching technology (26,271. In Canada, the Atomic Energy Control Board has proposed radioisotope release concentrations (C-123) but these have not been approved up to this writing [39].Maximum permissible release concentrations for Ra-226 to the atmosphere, to sewer, and to landfill or incinerator are 0.007 Bqlm3, 10 BqlL, and 0.3 Bqlkg, respectively. They are based on the annual dose criterion of a maximum of 50 pSvIyr which is only a srnall fraction of both the average annual dose received by members of the general public in Canada from natural background radiation and the regulatory dose limit of 5 mSvIyr for the public [39].

2.3 Techniques for Waste Characterization 2.3.7 Chernical Analysis It is vital to know the chemical composition of a waste sample at the outset since it will be an important basis for subse~uentwaste characterization steps. Chernical analysis is perfonned as whole rock analysis after total digestion with aqua regia and hydrofluoric acid. This is a method commonly employed for determining the chemical composition of geochemical samples. X-ray fluorescence (XRF) spectrometry is used for analyzing the major species and inductively coupled plasma emission (ICP) spectrometry for the trace elements.

2.3.2 X-ray Diffraction X-ray diffraction (XRD) is a unique technique that identifies the compounds of crystalline materials [40]. X-rays that impinge upon atomic layers of a material cause the atoms to vibrate and emit energy of the same wavelength as that of the incident X-ray. The diffraction pattern contains information about the crystallinity, phase composition, orientation and lattice stresses of the samples. The peak positions and intensity in the diffraction patterns of crystalline materials provide information about crystal structure and lattice parameters. A typical diffraction pattern for an amorphous material is a broad spectrum with no prominent sharp peaks relating to long range periodicity. In a diffractometer, the signal intensities recorded by the X-ray detector as it slowly traverses around the circle, can be plotted on a chart. The values obtained from this chart, with its abscissa showing 20 values and its ordinate showing signal intensities, can be used to identify minerals by comparing signal intensities and the appropriate d-spacing values against standard values contained in powder diffraction data files [41]. The &values can be calculated using the Bragg equation nh = 2dsin8 where n is a small integer (usually 1) A is the wavelength of the incident beam d is the distance between adjacent atomic planes, and 0 is the angle between the incident beam and the reflecting crystal plane.

XRD has a number of limitations [42]. With multi-phase samples there is a significant possibility of line overlap and the three strongest lines may not be due to the same substance. It is not possible to identify noncrystalline or amorphous substances since these do not register normal diffraction patterns. Components in a mixture occurring below 1 to 2% by weight are not detected since these are insuffident quantities of the materials to give rneasurable diffraction lines.

2.3.3 Radioactivity Counting Gamma radiation is detected and measured using various methods [40]. A common equipment is a gamma spectrometer equipped with a high purity germanium detector. The sample is contained in a small via1 and placed in a cylindrical hole with Pb shield. The principle is similar to pulse-height analysers which consist of one or more pulse-height selectors that are configured to provide energy spectra. A single channel analyser typically has a voltage range of perhaps 10V or more with a window of 0.1 to 0.5 V. Multichannel analysers typically contain several hundred separate channels, each of which acts as a single channel that is set for a different voltage window. The signal from each channel is then fed to a separate counting circuit, thus permitting simultaneous counting and recording of an entire spectrum.

2.3.4 Optical Microscopy For light microscopy, powdered samples are mounted in epoxy resin on glass and polished into thin sections of 1 cm diarneter to reveal the interna1 structure and morphology of the particles. Inthe reflected light microscope, the light from a high intensity source enters the instrument through a side tube, and is then reflected perpendicuiarly by a system of rnirrors on to a polished specimen held on the microscope stage 1411. A photornicrographic equipment is mounted on the microscope ta be able to take photographs of the specimen under investigation. A video camera feeding to a computer may be attached to the microscope to transfon a light microscope to an image analysis system. The images can be manipulated and enhanced by viewing the computer screen instead of the eyepiece. It is also possible to Save images in cornputer format and to do size measurement, particle counting, and video recording.

2.3.5 Evalua fion of Toxicity Two microbial colorimetric bioassays: SOS-Chromotest and Toxi-Chromotest have been developed recently to detect genotoxic and toxic activities of chemicals, phanaceuticals, and food stuffs. The tests have also been applied to environmental samples such as water, sewage and sediments [43]. The Toxi-Chromotest toxicity bioassay is based on the ability of toxicants to inhibit the de novo synthesis of an inducible enzyme, B-galactosidase. The arnount of de novo synthesized enzyme is determined by a colorimetric reaction. The SOS-Chromotest is a qualitative method of detecting the presence of genotoxicants [44, 451. A genotoxicant or genotoxin is any DNA-damaging agent, e.g., mutagen which attacks the genome (DNA) part of the ceIl. A genetically engineered strain of E. coli PQ37 developed at the lnstituit Pasteur, France produces 8-galactosidase in response to genotoxins. Genotoxins affect living cells by altering or creating lesions in the DNA structure causing mutations through faulty base pairing during the excision and repair pathway. The effect of toxins on living cells is more rapid and simply causes cell death. On the other hand, the Toxi-Chromotest involves a different strain of lyophilized E. coli K12 OR85 bacteria (EBPI Canada) produced by arbitrary bombardment with UV light.

2.3.6 Weathering Tendency Accelerated weathering experiments can provide further insights into long-term behaviour of waste samples in a landfill environment. Chemical weathering is related to such factors as climate, topography, parent material, and time with temperature and moisture flux as the major environmental variables affecting weathering rates [46,47]. This can be simulated in the laboratory by agitating the samples continuously for months in a gyratory incubator-shaker at elevated temperature thereby promoting physical and chemical changes which usually are very slow processes. Weathering caused by min, wind and Sun can, over time, release the heavy metals that are found in rock or soi1 samples.

2.4 Leaching Protocol Tests 2.4.7 Principles of Leaching Leaching for environmental purposes has been derived from hydrornetallurgy principles in which metals are recovered from low-grade and submarginal ores. Through leaching of the ore, precious metals are solubilized and subsequently recovered by processing the aqueous stream. In the environmental field, leaching tests have been used to determined whether the elements of environmental concern, now called contarninants, will solubilize and pollute the environment, e.g. surface water or groundwater. Leaching, in a hydrometallurgical sense, is a typical heterogenous process, in which a solid phase (ore particle), a liquid phase (leach solutions), and a gaseous phase (0, and sometimes CO,) are al1 involved. It is depicted schematically by the ore leaching process in Figure 2.2. According to this rnodel, the oxidant diffuses into the ore particle and reacts with the mineral grains. As the leaching front moves into the particle, some mineral grains present in veinlets or as discreet disseminated particles, may only be partially leached and the inner core of the particle may remain unleached (top figure) [146]. As time increases, the reaction zone moves further inward and a diffusion mechanism becomes the rate- controlling or slowest step as shown in the lower figure which was adapted from Rossi [48]. For environmental purposes, leaching tests involve contacting the waste material with a liquid to detemine which components dissolve. Prior to contact with the waste, the liquid is called the leachant; after contact, it becomes the leachate. Various types of leach tests have been developed that differ in parameters such as leachant composition, method of contact, liquid-to-solid ratio, agitation, contact time, and temperature, in order to investigate the chemical and mass-transport phenornena involved in leaching. Leaching tests are undertaken with several objectives but the most important ones are identification of leachable constituents, classification of hazardous wastes, and risk assessment of land disposal. There are more than 26 protocols, either as standards or accepted research tools that are available [49-511 but only three will be used in this study. It is recognized that laboratory leaching procedures which attempt to simulate field conditions are never adequate to predict long-term leachability [49]. This is because leaching typically occurs very slowly and the test would only represent a time period equivalent to the test duration. Thus short tests are actually accelerated procedures that cause significant matrix alteration, erosion and eliminate mitigating effects such as redox potential which is controlled by biological activity. These accelerated tests produce results that would never be observed under field conditions. Protocol leaching tests, which because of their nature must use accelerated conditions are often an exaggeration of the real environment with a very large safety margin. Nevertheless, they are quite useful for regulatory purposes in spite of their shortcomings because they provide quick information about the wastes to policy and decision-makers. Leached rim

Reaction zone

structure

Figure 2.2 Schernatic drawing of leaching of an ore particle containing a disseminated metal sulfide and the reaction zone (top) with a plot of concentration versus time showing a diffusion-controlled leaching mechanism (bottom), adapted from Rossi [48]. 2.4.2 Batch Versus Co/umn Leaching Both column and batch leaching methods can generate useful leachates to evaluate potentially hazardous wastes. However. the relative accuracy of predicting levels leached from landfilled wastes remains uncertain. Reproducibility in experimental data is a factor of two better for batch methods than column rnethods. The disctepancy was attributed to channelling in the column during leaching [52]. The batch extraction method is simple and more reproducible with controlled agitation while the column method is more realistic in simulating leaching processes under field conditions. According to Jackson [52],the batch method is more aggressive than the colurnn method in extracting elernents of concern presumably because of the better contact of the waste with the leachant during rotary agitation. Batch tests can be set up and used routinely by laboratory personnel more easily than the column method. Better reproducibility associated with the batch method facilitates more satisfactory interlaboratory cornparisons required in regulatory extraction procedures.

2.4.3 Agitafed Extraction Procedures Nurnerous protocol leaching tests are available but the rnost widely used are Toxicity Characteristic Leaching Procedure (TCLP) in the US [36]and the Leachate Extraction Procedure (LEP) in Ontario, Canada [35]. These procedures assume that the wastes are destined for a sanitary landfill which is dominated by municipal solid waste. Both methods use a 20:l liquid to solid ratio, the rotary extraction mechanisrn and acetic acid extracting solution. The main differences are that extraction for the Canadian LEP is six hours longer with intermediate pH adjustments and at slower rotation speed (10 rpm against the 30 rprn of TCLP). These procedures are described more in detail under experimental work. These tests are single measurements which assume that equilibrium was attained in the laboratory within one day. Batch kinetic tests could provide kinetic information over a period longer than one day but are not currently used as protocol tests.

2.4.4 Sequential Chernical Extraction Sequential chernical extraction (SCE) is not a protocol procedure but is a more aggressive test for leachability and provides speciation information. It involves several extractions using successively stronger reagents and tem peratures. each intended to remove one phase from the sample. The basis of the sequential extraction procedure is the existence of specific fractions in a solid material which can be selectively extracted by the corresponding reagents [53]. Most SCE procedures were adapted for environmental samples such as contaminated soils, sediments, dust, and fly ash 153-611. In this study, the procedure used is a version of the most widely used scheme of Tessier et al [55] which was based on the work by Gupta and Chen [54]. Sequential extraction has been popular, though tirne consuming to perform, since it can provide detailed information on the origin, mode of occurrence, biological and physicochemical availability, mobilization, and transport of trace metals [54, 55. 62, 631. The Tessier procedure involves five extraction stages, designated A to E, used to leach metals associated with different chernical phases in the residues [55]. The various fractions have been classified as Fraction A represents the extraction of elements that are soluble in or are in an exchangeable form with water; Fraction B represents the extraction of elements from carbonate minera1 phases susceptible to mild acid conditions; Fraction C represents the extraction of elements bound with iron and rnanganese oxides; Fraction D represents the elements bound with sulfides which may be decomposed under oxidizing conditions; and lastly, Fraction E represents the elements that are associated with unreactive minerais, mainly silicates, which can incorporate metals within their crystal structures. This lasi fraction may contain metals that are not expected to be released in solution to the environment under normal conditions, even over a long time.

2.5 Microbial Leaching Geothermal solid residues have been found to have similar components (silica and metal sulfides) to mine tailings [28,921 and therefore experience from the mining industry may be applicable to the geothermal industry. The following section will deal with the occurrence of acid mine drainage and the various mechanisms of bacterial leaching.

2.5.1 Acid Mine Drainage 2.5.i .1 Occurrence Acid mine drainage (AMD) is a problem commonly found in coal and copper mines whereby sulfide materials rejected during the mining of coal and metal mines and deposited in mine tailings or heaps are oxidized to sulfates. This in tum releases runoff with high acidity and heavy rnetals causing pollution to the environment over a long period of time [64-661. Acidic drainage also causes severe corrosion problems to mining and ancillary equipment [67].In western USA, the Forest Service estirnated that between 20,000 and 50,000 sites (including abandoned and operating mines) are currently generating acid on forest lands and that drainage from these mines is affecting between 8,000 and 16,000 km of streams [68].The annual volume of acid-generating waste rock or tailings produced by the Canadian mineral industry is estimated at 140,000 dry tonneslyear [69]. The sulfide oxidation is microbially enhanced by the presence of iron and sulfur oxidizing bacteria such as Thiobacillus ferrooxidans, that cm survive at low pH (< 3.5) and high temperature (up to 60 OC). Since geothermal residues possess characteristics resembling mine or rock tailings and have also been traditionally disposed in open dumps, the investigation of possible AMD potential was most prudent.

2.5.1.2 Role of Microorganisms The reaction rate causing AMD is greatly accelerated by the presence of T. ferrooxidans to as much as IO6 - fold [70, 711. These bacteria promote indirect oxidation of pyrite and other sulfides through the catalysis of the oxidation of ferrous ion to ferric ion which is an effective oxidant. However, they may also catalyze direct oxidation of pyrite by oxygen. These organisms act only as redox catalysts; they do not oxidize substrates or reduce oxygen but mediate the reaction or electron transfer. In doing so, they obtain a source of energy from these energy-yielding redox reactions for their metabolic needs. Thiobacillus ferrooxidans tends to live in environments such as hot springs, volcanic fissures, and sulfide deposits as well as oil brines, coal water, mine wastes, peat soil, concrete and building stone [48, 721. Thiobacillus fenooxidans is a chemoautotroph in the genus of Thiobacilli and can derive metabolic energy from oxidation of iron and sulfur compounds [73-771.It requires oxygen and carbon dioxide for its metabolism. The bacteria can draw energy from the oxidation of ferrous ions to ferric ions as the sole energy source. The bacteria have been observed to survive under anaerobic conditions by oxidizing sulfur with ferric ion serving as the oxidant. This was reported by Sugio et al 1791,Corbett et al and Goodman et al as cited by Tyagi [37] and most recently by Pronk [78]. Thiobacillus ferrooxidans is morphologically similar to Thiobacillus thiooxidans, a sulfur oxidizing bacterium, and both have been found in heap and dump leaching of minerals where acid mine drainage is suspected. T. femoxidans is highly polymorphic. The shape of its cells Vary frorn large rods with rounded ends, from 2 pm in length and 0.5 pm in diameter, to spheres, ovoids and rods from 0.5 to 0.7 pm in length and 0.3 to 0.4 pm in diameter. White they cm occur singly or in pairs, chains consisting of six to seven cells have also been observed. Two modes of cell division have been recognized : constriction, the most common multiplication mode, and partition. It was reported that the combination of T. femoxidans and T. Thiooxidans is more effective in bioleaching [22, 26, 371. The most robust leaching microorganisms are the thermophilic (60°C) acidophilic species of the genus Sulfolobus. Of these, Thiobacillus femoxidans is the easiest bacteria to cuiture in the laboratory since it requires less energy at room temperature (25-35 OC).

2.5.2 Mechanism of Bacterial Leaching Due to biochemical reactions, largely insoluble metal sulfides can be degraded to soluble metal sulfates by direct and indirect methods of bacterial metabolism [37,70, 801.

2.5.2.1 Direct Method In the direct mechanism, the metal sulfide is oxidized to metal sulfate : bacteria Mes + 2 0, -+ MeSO, where Me is a bivalent metal. The heavy metal sulfides such as NiS, ZnS, COS, PbS, and CuS are generally insoluble in aqueous acid leach media, while their sulfates have solubility with the exception for lead sulfate which is sparingly soluble (K,, of 1.6 x 109. This is illustrated by pyrite oxidation which involves rapid oxidation of ferrous sulfate with the rnediation of Thiobacillus ferooxidans. Equation 2 is the overall reaction for Equations 2a and 2b.

bacteria 4 FeSO, + 0, + 2 H,SO, -+ 2 Fe,(SO,), + 2 H,O 2.5.2.2 Indirect Method In the indirect mode of bacterial action, the metal sulfide is oxidized by a ferric ion without the direct participation of bacteria as shown below in Equation 3. This reaction takes place geochemically under conditions of weathering and leaching. Ferrous ions can be reoxidized as in Equation 2 and again ferric ions can act as the oxidizing agent. In indirect leaching processes, Thiobacillus femoxidans catalyses ferrous ion oxidation which takes place very slowly under normal conditions (Equation 5).

The elemental sulfur that has been set free in Equation 3 will be oxidized to sulfuric acid mediated by bacteria in the following mode:

bacteria S* + 1.50, + H20 + H$O,

In the same manner, the ferrous ion is reoxidized, mediated by the microorganisms to ferric ion:

bacteria 2Fe2++ 0.50, + 2H+ 2Fe3++ H20 and the iron redox cycle is repeated. The production of sulfuric acid will decrease the pH, which will enhance further the solubilization of metals. It is also possible to form a yellow insoluble precipitate called jarosite, KFe,(SO,),(OH), which can hamper transport phenornena by coating the minerai. In bacterial leaching systems, it is desirable to prevent jarosite generation because of the formation of diffusion barriers on mineral surfaces and the scavenging of metal ions from the leach solution [81]. In the direct mode of bacterial oxidation, as shown in Equation 1, bacteria rnust remain close tu the surface to be adsorbed onto the solid substrate where dissociation takes place according to the solubility product of the metal sulfide : The released sulfide moiety of the metal sulfide is then immediately captured by the enzyrnic systern of bacteria and is oxidized to sulfate :

As the sulfide ion is oxidized to sulfate, further dissolution of MS as shown in Equation 6 will occur, shifting the reaction to the right. The prerequisite for bacterial oxidation of solid sulfides according to this explanation is the availability of the substrate in the soluble form. In contrast to the direct leaching of suifide minerals, the agent responsible for leaching (H,SO,) is produced by bacterial oxidation. Reactions for other sulfide minerals shown below were presented by Dutrizac and MacDonald [26]:

ChalcopvHte CuFeS, + 4 Fe '+

CuFeS, + 4 Fe3++ 2H20+ 30, -+ Cu 2++5 Fe2++2H2S04 (9)

bacteria 2 CuFeS, + 8.50, + H,S04 -+ 2 CuSO, + Fe,(SO,), + H,O (10)

bacteria

ZnS + 20, -+ ZnSO,

PbS + Fe,(SO,), -+ PbSO, + 2FeS0, + S0 bacteria

PbS+20, -+ PbSO,

2.5.3 Factors Affecthg Bacterial Leaching The effkiency of rnicrobial leaching processes depends mainly on the bacterial activity and on the chemical and mineralogical composition of the sample (80, 821. The bacterial activity is influenced by several environmental factors, such as composition of the leaching medium, Eh, pH, temperature, particle size, and oxygen availability. A maximum rate of metal extraction can be achieved if the environmental conditions correspond with bacterial growth conditions.

2.5.3.1 Composition of Leaching Medium Thiobacillus femoxidans is markedly affected by variation in ammonium sulfate and dipotassium hydrogen phosphate concentrations of the leach medium [81]. Other nutrients such as nitrate, calcium, potassium and chloride ions are present in sulfide-bearing substrates. Best concentrations for ammonium sulfate were 3 gIL, dipotassium hydrogen phosphate, 0.5 glL, sulfate at 2 glL, magnesium from 2 mglL to 12 mg1L and ammonium and phosphate 9.6 mglL [83]. There has been a wide variation of opinion and experience on the importance of the amount of ferrous ion in the bioleaching medium. Lakeview Research, Peterborough, Ontario has been including FeSO, in the leaching medium when performing the BC Research Confirmation test (see Section 2.6.1.2)with adapted bacterial cultures. McGoran et al [76]achieved the highest growth rates and minimum generation time of Thiobacillus ferooxidans when using Fe2+as substrate in 9.0 glL concentration. Barron and Lueking conducted a detailed study of the growth and maintenance of Thiobacillos ferrooxidans cells [84]. They found that bacterial growth was significantly influenced by the concentration of FeSO,, with maximal growth rates in the presence of 2 to 3 g of Fe2+per liter. However, in the BC Research Confirmation Test, FeSO, was withheld from the culture media since it was expected that the required iron can be obtained by the bacteria from the sample [67].This may be a critical issue for the survival of the bacteria throughout the duration of this test since there is no soluble iron present initially. The samples to be tested usually contain iron sulfides that are insoluble and need to be oxidized to more soluble sulfates through bacterial catalysis. However, without bacterial maintenance, this reaction may not occur.

2.5.3.2 Oxidation-Reduction Potential The ferrous-ferric system requires an oxidation-reduction potential (ORP) of 747 mV at 25 OC. The experimental Eh values from +220 mV to +540 mV have been observed during the oxidation of metal sulfides as substrate [75, 851. This is an important parameter to measure since it is an indicator of the progress of reaction within the system. An increasing positive Eh value is a sign of oxidation primarily the ferrous-ferric redox reaction. Highly effective leach solutions are produced when the Eh is maintained at around +750 mV [86].This ideal situation may be difficult to attain due to formation of iron precipitates and poor aeration.

2.5.3.3 Temperature Maximum leaching of metal sulfide ores and oxidation of ferrous ion by iron- oxidizing bacteria such as Thiobaciiius femoxidans has been deterrnined to occur between 25-35 OC while 55OC is the limiting temperature for biological oxidation. Only chernical oxidation occurs above this temperature [37]. The downside to using controlled temperature for bioleaching is the cost of the energy required to operate the systern. Boogerd et al as cited by Bos [87]reported that at around 30-35 OC, the contribution to the overall kinetics of the oxidation of metal sulfide by ferric ion (indirect method, Equation 3) is negligible. Only in thermophilic temperatures such as 45-70 OC will there be an increase in the overall kinetics [87].

2.5.3.4 Hydrogen Ion Concentration ldeal pH is between 1.O and 2.5 for the oxidation of ferrous ions and rnetal sulfides [37]. Thiobacillus ferrooxidans is active in the pH range of 1.5-5.0 but will survive up to a maximum of 6.0 and minimum of 1 .O without adaptation of culture. Better results on metal solubilization have been obtained by lowering the initial pH to 2.0 [37]and up to 3.5 [76, 771. Maximum bacterial activity can be attained at around pH 3.2 whereas minimum activity can be reached at approxirnately pH 1.5 and pH 5 [88]. Also, at lower pH, the formation of the orange brown precipitate jarosite or ferric hydroxide is rninimized.

2.5.3.5 Agitation and Oxygen Transfer The effects of mixing have been studied in bioleaching by several researchers [33, 481. Leaching rates were increased by agitation but only with gentle mixing (100-250 rpm). Rapid mixing (300 rpm and above) resulted in lowering the metal solubilization rates. It should be noted that Thiobacillus femoxidans requires a minimum oxygen concentration of 2.5% of saturation in coal mines as reported by Bos [87].However for sewage sludge as a substrate, a minimum of oxygen concentration of 35% of the saturation is required by Thiobacillus femoxidans. Lower oxygen levels resulted in reduced rnicrobial activity [37]. At 23 OC, the oxygen solubility in pure water is 8.56 mglL and will decrease to 6.60 mg11 in the presence of 45 ppt salinity (ORION oxygen probe manual) and at higher temperature [130]. The amount of oxygen that must be made available for a satisfactory bioleaching process can be roughly estimated using the simplified equation: bacteria MS+20, -t MSO, where M is a bivalent metal. According to the above equation, to convert 1 mole of the metal sulfide to sulfate, 2 moles of oxygen are consumed [48]. Liu et al 189) has compared the solubility values of oxygen in aqueous solutions and culture media used in bioleaching of metal sulfide ores and concluded that it is reasonable to use the saturation solubitity of oxygen in water (6.68 mglL) for the culture media in bioleaching process [89].

2.5.3.6 Particle Size and Substrate Concentration Several researchers have utilized various particle sizes when carrying out AMD potential studies. Lakefield Research, Peterborough, Ontario have been grinding samples with a mortar and pestle achieving particle sizes of 150-180 Fm for their BC Research Confirmation Test. University of Waterloo in their AMD kinetic studies used particle size fractions between 90 to 125 pm from mine tailings 1901. Lawrence et al used -75 pm as particle sample size in evaluating various AMD potential prediction procedures of mine tailings [91]. Although it is preferable to use -45 pm which is the average size of mine tailings, it is difficult to achieve this using mortar and pestle especially if the sample contains high amounts of silica. It is also not advisable to use the mechanical grinder especially for small quantity of sarnples since contamination from other users is bound to be a problem. Nicholson as cited by Ferguson and Erickson [66]gave a good explanation for why particle size is an important physical factor that affects the AMD process. Coarse-grained mining wastes allow greater oxygen advection and hence active acid generation can occur to a greater depth in a waste heap than fine-grained waste. In coane metal mine waste rock dumps, air convection is promoted by wind action, barometnc pressure changes, and interna1 dump heating from the exothermic oxidation reactions. Under these conditions, active acid generation rnay occur throughout the dump rather than being limited to the surface zone, as in fine-grained mining wastes such as tailings [66].With this assessrnent, it rnay be more advantageous not to have very fine grain particles for the AMD potential test.

2.6 Prediction of Acid Mine Drainage Potential In mineral mining, the prediction of acid mine drainage (AMD) is needed to find out if the quality of waters draining from a mine site will exceed environmental regulatory standards, and if sa, what mitigation measures have to be provided at the outset. Accurate prediction of AMD is required both to protect the environment and to ensure that resources are expended wisely to prevent or control AMD. The experience in the mining industry on prediction of acidification potential and metal release wiII be useful also to the geothermal industry where solid residues with components (silica and metal sulfides) similar with mine tailings are produced [28, 921.

2.6.1 Prediction Procedures Ten prediction techniques were evaluated by Lawrence et al [QI ] and Ferguson and Erickson [93]. Of these. four were static tests and six were kinetic tests. Calow et al [94] compared two common static tests: the BC Research Initial Test (BCRIT) from Canada and the Acid-Base Accounting (ABA) from the US. The BCRlT was favored over ABA after testing eight mine tailings samples [94]. The static and kinetic procedures are listed in Tables 2-5 and 2-6. The USEPA and CANMET defined 'static tests' as methods performed in a few hours or one day to determine initial acid producing potential [67,68]. On the other hand, 'kinetic tests' involve predicting the long-term weathering characteristics of a waste material as a function of time hence of longer duration from weeks to months and even years. Kinetic tests are usually carried out only if static test evaluation indicates AMD potential. The tenn 'static' was used since the tests do not consider the relative rates of acid production and consumption. Several studies have shown that AMD predictions using static and kinetic techniques correlated well with actual mine water quality [66, 91, 951. The uncertainty over AMD prediction results cm be overcome through verifkation of test results with field experience.

2.6.1.1 Static (Initial) Tests There is no standard AMD potential test but the most widely used especially in Canada is the 6.C Research Initial Test (BCRIT) which corresponds to the Acid-Base Accounting of the US [94]. It was found to be reproducible, less prone to operator error, conservative and more representative of the natural AMD problern. The BCRIT determines two parameters : the neutralization capacity of the sample and its acid producing potential. In cornparing the two values. if the acid production potential (APP) exceeds the acid consumption (AC) expressed in kg H2S04per tonne of material, the sample is classified as a potential acid producer and confirmation testing is recommended. The tests are chemical rather than biochemical in nature but correspond to the acidity contribution from H2S04as a result of its formation due to oxidation of the sulfides to sulfates. The acid production potential was calculated from the percent sulfur in the sample converted to kg H,SO, by a conversion factor (APP = 30.6 x %S) whereas the acid consumption was computed from the volume of acid used to reach the endpoint of pH 3.5 [67].This calculation is further explained in Appendix C.

2.6.1.2 Kinetic (Confirmation) Tests Several researchers have found that static tests were oh'en accurate in predicting drainage quality and were particularly valuable as screening tests to determine if more sophisticated procedures should be used [93, 941. Table 2-6 inciudes a summary of kinetic acid mine drainage prediction techniques as reported by CANMET (671, USEPA [68], Lawrence et al [91] and Ferguson and Erickson [93]. Of the many AMD kinetic tests. the B. C. Research Confirmation Test, is widely used at base rnetal and gold mines in Canada and even in USA [68,91, 93, 961. Positive results from this test are considered confirmation that the microbiologically catalysed reactions can become self-sustaining [97]. The B.C. Research Confirmation Test requires inoculation with T. ferrooxidans to stimulate the rapid stage of oxidation [91, 971. The sample (10-20 g depending on S content) is placed in 250 rnL Erlenmeyer flask with 70 mL nutrient media, 5-10 mL culture of T. ferooxidans at pH 2.2 - 2.5. The flask is placed on gyratory shaker at 35 OC in a CO, - enriched atmosphere, pH is monitored and additional sarnple provided. If the pH rises substantially, then the sample is considered nonacid producer. If the pH remains low, then sample is a potential acid producer [97]. The limitations of this procedure are : (a) there is no specified procedure to spawn an acclirnatized bacteria culture, (b) there is no assurance of bacterial growth as FeSO, was withheld from the culture media and bacterial viability was only checked at the onset and not periodically during the test, and (c) redox potential (Eh) and metal concentration of solution are not measured to indicate chemical reaction and rnetal release.

2.6.2 Laboratory Scale Bioleaching Techniques Two types of taboratory scale bioleaching methods were discussed thoroughly by Rossi [48]. The first type involves a qualitative or semiquantitative assessrnent of the amenability of a material (ore or residue) to biodegradation by a well specified bacterial strain. This class includes manometric, stationary flask and air sparging techniques. The second class, in contrast, provides quantitative evaluation of parameters in an analytical approach using kinetic simulation model. These methods are air-lift percolator, shake flask and pressure bioleaching. In this study, the two techniques representative of both classes are the stationary flask technique and the shake flask technique. Below is an almost verbatim description'based on Rossi [48].

2.6.2.1 Stationary Flask Technique This technique has been judged to be the simplest but rnost effective rnethod of microbial culture due to modest cost of equipment and experimental simplicity [48]. The experimental procedure is quite simple: culture medium, substrate and inoculum are introduced into Erlenmeyer or Florence flasks, which are then plugged with adsorbent cotton and placed in the cabinet or on the bench for the duration of the test. The flask is plugged with adsorbent cotton to allow air to enter while filtering out airborne contaminants. The area of air-liquid contact should be maximized to favor the diffusion of air into the liquid mass. This is important in the case of stationary flasks where the liquid surface is still and the rate of air transfer to the culture medium is controlled by Fick's law: Table 2-5 Summary of Static Test Methods, Costs, Advantages, and Disadvantages

Acid Base Accounting Modiiicd Acid Base BC RESEARCH Alkaljnc Production Net Acid Production II Accounting I Initial I Potcatiil: Sulfur I ACID PRODUCTION DETERMINATION Acid Producing Fotential = Acid Producing Potcntial = Total Acid Production = Total S used as indicator 300 mlH20, added to 30.6 Total S 30.6 Total S 30.6 Total S 5 g rock to directly

ON POTENTlAL DETEF 60 rnesh (240 urn) 60 mesh (240 um) 300 rnesh (380 um) 230 um sample paniclc size not sample sample sarnplc presentcd

add HCI as indicated by tin test. add HCI as indicated by tiu iitrate sarnple to pH 3.5 20 mL 0.1 N HCI to 0.4g acid produccd by iron sultide boil one minute test agilate for 23 with 1 .O N H2S0, solid for 2 hours at oxidation dissolves buffering then cool hou6 at room roorn temperature minenils temperature pH 1.4 - 2.0 required afier six hours agitation

titration endpt pH 7.0 titration endpt pH 8.3 titration endpt not iimtion endpt pH 4.0 titration endpt pH 7.0 applicable

duration: 1-2 hours duration: 24 hours luration: 5-8 hours duration: 2 houn duration: 2 houn cost: USS 34-1 10 cost: USS 34-1 10 OS^: USS 65-170 cost: US$ 34-1 10 cost: US$ 25-68 b ADVANl LGES AND DISADVANT, GES -simple and short time -simple -simple .simple -simple -no special equipment -fairly short time -shon time -short timc -short time -easy interpretation -no special equipment -no special equipment and -no special equipment -no special equipment many sarnples can be -easy interpretation -casy interpretation easy interpretation tested -rnany sarnples can bc tested

-does not relate to kineiic -does not relate io kinetic -assumes parallel acidl -moderate interpretation -1irnited reproducibility -assumes parallel acid -assumes parallel acidl alkaline releasc -uncemin if extent of alkaline release alkaline rctease different partide size not suIfide oxidation -if APP and NP are close. -ifAP and NP are close, reflected simulates that in tield hard to interpret hard to interpret -if APP and NP are close. -ditYerent particle size not -different particle sizc not hard to interpret 1 retlected reflected Source: This table was compiled from USEPA, 1994; CANMET, 1991; Lawrence, 1989; Ferguson, 1988; Ferguson, 1987; Bruynesteyn, 1984; and Sobek, 1978. Table 2-6 Summary of Sorne Kinetic Test Methods, Costs, Advantages, and Disadvantages

II HUMlDlTY CELLS 1 SOXHLET EXTRACTION I COLUMN TESTS - - - -- SUMMARY OF TEST METHOD 2.38 mm particle size panicle size not presented I variable particle sizc ZOOg of rock exposed to three days dry air. T=70°C and at T=2S OC colurnns containing mine waste are leached three days humidified air. and rinscd with water passed through sample is distilled wirh discrete volumes or recirculating 200 mL on day sevcn and recycled through sample solutions II duration: 3-8 days duration: 3-9 rnonths duration: 8- 10 weeks cost: US6 425-850 cost: USS 212-425 cost: USS2000 - 4000 ADVANTACES AND DISADVANTACES -models AP and NP well -simple -models AP and NP -rnodels weddry -results in shon time -modds cffect of different rock types -approximates field conditions and -assessrnent of interaction behveen AP and -models wetldry rate of acidity per unit of samplc NP -rnodels diffennt grain sizcs moderate to use -moderate to use -results take long time -need special equiprnent difficult interpretation -sorne special equipment -moderate interpreration -no1 practical for large numbcr of samples -moderate ease of interpretation in developmental stage and -large volume of sample -large data set generated relationship to naiural processes not clear -lots of data gcnenited -long timc -wtential oroblems: uneven leachate applicatibn, channelization

II BC RESWRCH CONFIRMATION 1 BATCH REACTORJSHAKE FUSKS 1 FIELD TESTS

iMETHOD II 400 merh particle rire l 200 mesh panicle size field scalc panicles 15-30g added to bacteriolly active solution sarnpleiwater slurry is agitated 800 to 1300 metric ton test piles at pH 2.2 to 2.5. T=35"C 200g.500 rnL consmcted on liners flow and water if pH increases. sample is non acid quality data collectcd producer tesu began in 1977 and are ongoing if pH decreases. Il2 original sarnple mass is added in each of two increments

duration: 3-4 weeks duration : 3 rnonths duration : at least 1 year cost: USSI70-330 COSI:USS425-850 COS!: usa 10000- 40000 ADVANTACES AND DISADVANTACES -simple to use -able to examine many samples -uses actual mine waste under -1ow cost sirnultaneously environmentiil conditions -assesses potential for biological leaching -relativeiy simple equipment can be used to determine drainage volume -moderate to use mitigation methods can be tested -longer time needed -subject to large sampling errors -expensive initial construction -some special equipment needed 4ack of precision -long time -difficult interpretation if pH change small -does not rnodel initial AP step

Source: This table was cornpiled from USEPA, 1994; CANMET, 1991; Lawrence, 1989; Ferguson, 1988; Ferguson, 1987; Bruynesteyn, 1984; and Sobek, 1978. a=-DA& dt dx where dQldt is the rate of transport across the liquid surface, D is the diffusion coefficient, dddx is the air concentration gradient across the liquid surface and A is the surface area of the liquid phase. This technique is useful in identifying the amenability of rninerals to bioleaching and the influence of physicochemical parameters in the process.

2.6.2.2 Shake Flask Technique The two main shortcomings of the stationary flask technique : the time-dependent heterogeneity of the suspension and the slowness of the gas diffusion (0, and CO,), are overcome in the shake flask technique [48]. Test equiprnent consists of Pyrex Erlenmeyer flask and a shaker consisting of a platform with several flask clamps moving in a reciprocating or rotary plane motion. The device is called a "reciprocating shaker" when the platform motion is reciprocating and a "rotary action shaker" when the motion is rotary. Both devices shake the suspension, ensuring thorough mixing and homogeneity as well as agitation of its surface, thereby enhancing the dissolution of atmospheric oxygen and carbon dioxide needed by the microorganisms for their metabolism. A rotary shaker is preferred over the reciprocating shaker since the suspension is uniformly agitated in al1 directions. Rossi had described this procedure thoroughly 1481. At the beginning of the test, after pH determination, the weight of flask and its content is measured. The flasks are clamped to the shaker and the apparatus is started up. At regular intervals, the agitation is interrupted and the solution allowed to rest, to measure flask weight and other parameters. The initial weight is restored by adding distilled water to compensate for evaporation estimated to be between 0.6 and 0.7 glday. A 1 mL aliquot from the supernatant is obtained for chemical analysis. The results are recorded and used to plot a metal leached vs time or pH vs time which usually exhibits a more or less pronounced "Snshape curve corresponding to the lag, exponential and asymptotic growth phases of bacteria. According to Rossi, the following operating conditions are considered appropriate: 250 mL Erlenmeyer fiasks, 1 to 10 g of ground sample, 1 mL of inoculum, 75-100 mL of solution (culture medium plus inoculum) and 200-300 rpm. In order to shorten testing times, the sample is finely ground from -40 pm to -200 pm. 2.7 Geochemical Equilibriurn Modeling A variety of mathematical models have evolved through the yean which attempt to predict the behavior of pollutants at equilibrium under various environmental settings. MINTEQA2, a geochemical equilibrium program developed by the US Environmental Protection Agency, is one of the most popular models [98]. The principle uses the "equilibriurn constant method" which is simultaneous solution of nonlinear mass action expressions and linear mass balance relationships. A brief description of MINTEQA2 is presented in Appendix D.

2.7.7 Different Thermodynamic Models Precursors of existing equilibrium rnodels such as MINTEQAZ are the EQ316, PHRQPITZ, SOLMNEQ, REDEQL. MINEQL, GEOCHEM [99-IOI]. EQ3NR and MICROQL 11[98], and PHREEQE [85] . They were al1 written in FORTRAN language and used a solution algorithm based on the Newton-Raphson technique. MINTEQA2 has several limitations. Firstly. to be able to simulate the actual system, al1 the complex solids that need to be modeled have to be known. Secondly, the MINTEQA2 thermodynamic database is not complete and may not contain al1 these solids and thirdly, it has a maximum iterations of 200 to reach equilibrium. However, several researchers have found MINTEQA2 to be less tedious than other thermodynamic models [98,102-1 051.

2.7.2 Applications Several researchers have used MINTEQA2 to sirnulate solid-phase dissolution from coal fly ash and incinerator residues as well as to compare with the controlling solids observed with experimental leaching methods 198, lO3,lO5]. Only a few like van der Sloot [IO51 had found reasonably good agreement between model-predicted equilibriurn aqueous phase concentrations and laboratory data. He recommended thermodynamic modelling to supplement regulatory protocol tests 11051. MINTEQA2 was used extensively by researchers from the University of Waterloo in Ontario, Canada under the acid mine drainage program [go, 106, 1071. To date, there is no reported application in modelling behavior of geothermal wastes. CHAPTER 3 METHODS AND PROCEDURES

The geothermal residues were obtained from three geotherrnal fields : (a) Bulalo, Philippines, (b) Cerro Prieto, Mexico, and (c) Dixie Valley, USA. They were al1 examined on an as-received basis since they were relatively dry with an average moisture content of less than 5% measured by drying overnight in a 105 OC oven. The samples were air dried at ambient temperature and stored in polyethylene bottles. Each sample was assigned the following acronyms to facilitate their processing and subsequently the presentation of results and discussion. The code is based on the first letter of the country of origin and the next two letters are descriptors of the samples: PSC - Philippine scale, PSL - Philippine sludge, ASC - American scale. MOM - Mexican drilling mud, MSC - Mexican scale, and MSL - Mexican sludge. PSC, PSL, MSL, and MDM have fine or flaky particles below 9.5 mm in size (15% were below 125 Pm, by weight) while MSC and ASC are hard and rock-like composed mostly of big particles ranging from 1 to 15 cm in size (2% were below 125 Mm, by weight). For the procedures requiring fine particles (-125 prn), the samples were ground in rnortar and pestle and sieved in Canadian Tyler standard screen (120 mesh). All chemicals, salts, acids, and pH buffers used were of analytical grade, while al1 solutions, standards, and dilutions were prepared using deionized water. The following laboratories performed some of the procedures and analyses reported in this work: at the University of Toronto: Centre for Nuclear Engineering for the radioactivity counting, Department of Microbiology for the toxicity testing, Department of Chemistry for the X-Ray diffraction, Faculty of Medicine for the transmission electron microscopy, and at XRAL Laboratories (SGS Canada) for the whole rock analysis and leachate analysis.

3.1 Waste C haracterization 3.1.1 Chernical Analysis Approximately 10 g each of the air-dried geothermal samples were used for multielement whole rock analysis, a method commonly employed for detenining the chemical composition of geochemical samples. The analytical techniques used were : Leco sulfur analyzer for sulfur, cold vapour spectrometiy for mercury, selective ion electrode for chlorine, X-Ray fluorescence spectrometry for the major species and inductively coupled plasma emission spectrometry for the trace elements. The loss on ignition (LOI) was determined at 950 OC. These analyses were perfomed by XRAL Laboratories.

3.1.2 Radioactivity Counting About 5 g per sample was used to detect radioactivity using a hyperpure germanium (HPGe) well-type detector (with cryostat well dimensions H=40 mm, D=15 mm), 10% relative efficiency, a full-width-half-maximum resolution of 2.60 keV at 1332 keV and a 5 cm lead shield giving a background of 8.1 pis in the energy range 35 keV-1780 keV. Measuring time was 60,000 s using as reference standard a soi1 sample with known values of U, Ra, and daughter's activities. Since MSC showed unusual radioactivity levels, four confirmation runs were carried out with measuring time between 70,000 to 413,000 seconds (1 to 5 days). Spectrum analysis and activity calculations were derived using equations from OSQIPIus Manual [108]. These analyses were carried out by the Centre for Nuclear Engineering, U of T.

3.1.3 X-Ray Diffraction Powder X-Ray diffraction (XRD) was used to identify minerals or crystalline compounds. Each sample was ground in acetone using a mortar and pestle and spread thinly on a glass slide. A Siemens MO00 diffractometer system having CuKa (A = 1.542A) radiation at 40 kV, 15 mA and scanning from 5' to 66 with a scan speed of 1 degree 28 per min was used. Phase identification was carried out manually using 1989 Hanawalt lndex of the Joint Cornmittee on Powder Diffraction Standards JCPDSIPDF-2 Data Set. These analyses were undertaken by the Department of Chemistry, PXRD Analyses and Services, U of T.

3.1.4 Optical Microscopy For light microscopy, the powders were mounted in resin and polished to reveal the interna1 structure and morphology of the particles. Photographs were taken using an Olympus Vanox C-35 camera at various magnifications. 3.2 Toxicity Testing Powder samples for toxicity tests were prepared using (a) solid extraction with 10% dimethylsulfoxide (DMSO) + 10% methanol and (b) direct sediment testing procedure (DSTP) [log]. Both bacteria, E. colistrain PQ37 and E. coli K12 OR85 were obtained from Environmental Bio-detection Products, Inc., Brampton, Ontario. These tests were carried out at the Department of Microbiology, U of T. The SOS-Chromotest was perfonned using 100 PL of an exponential growth phase culture of E. coli strain PQ37 in al1 the wells in a standard 96-well microtitration plate. Following a two-hour sample incubation at 37 OC, 100 pL of blue chromogen was added to the wells and reincubated for another hour. Genotoxic activity was noted by the presence of a distinctive blue colour in the wells. A relative measure of genotoxicity was determined by measuring the intensity of the blue color using a spectrophotometer. For the Toxi-Chromotest, serial two-fold dilutions of the samples were prepared in the microplate. A via1 of E. coli strain KI2 OR85 was rehydrated and mixed with the reaction mixture. To each well in the microplate, 100 PL of this mixture were added. Following a 90 min incubation at 37 OC, blue chrornogen was added and reincubated for another 90 min. Toxic activity was noted by the absence of blue color 11 IO].

3.3 Sequential Chemical Extraction A series of sequenüal extractions by Gupta [54] and Tessier 1551 as shown in Table 3-1 was used on the geothermal residues. Fractions A and B correspond to the Canadian LEP and the US EPA's TCLP described below. Since the samples were mainly inorganic in nature, the last fraction E, the residual phase, was revised by using HN03/HF/HCIfor digestion without HCIO,. The samples were pulverized with a mortar and pestle and 0.5 g each of the six air-dried samples were placed in 15 rnL polypropylene centrifuge tubes prior to extraction. Between each successive extraction, separation was effected by centrifuging for 5 min. The supernatant was removed with a pipet and transferred to a 50 mL centrifuge tube and diluted with deionized water and acidified to pH<2 with concentrated HNO, prior to analysis. The residue was washed with deionized water, centrifuged for 3 min and the washing was discarded. The five extraction steps were performed under a fumehood using full precaution specified in the material safety data sheets of the reagents. Table 3-1 Summary Procedure for Sequential Chernical Extraction

Fraction Extracted Procedure

A. Exchangeable 1 M sodium acetate (4 mL), pH 8.2, Ih, 20°C, continuous agitation B. Carbonate 1 M sodium acetate (4 mL), pH 5 (adjusted with HAc) , 5h, 20°C, continuous agitation C. Fe-Mn Oxides 0.04 M NH,OH.HCI in 25% HAc (10 ml), 6h, 96%, occasional agitation D. Sulfide 0.02 M HN03 (1.5 mL) and 30% H,O, (2.5 mL), pH 2, 2 h, 8S°C, occasional agitation; further 30% H,O, (1.5 mL), pH 2, 3h, 85'C, occasional agitation; then 3.2 M NH,OAC in 20% HNO, (2.5 mL), 0.5 h, Ca. 20°C. continuous agitation E. Residual 70% HNO, (6 mL), 40% HF (4 mL) to near dryness; further 40% HF (4 mL), 175OC; residue dissolved in 12 N HCI (5 mL) and diluted to 25 mL

3.4 Accelerated Weathering Test In a 250 mL Erlenmeyer flask containing 70 rnL bacteria nutrient media [lII], a 10 g pulverized sample (100 mesh) was shaken continuously inside an incubator-shaker (Lab-Line Instruments) for 3 months at 150 rpm and 35 OC 167, 961. The pH was monitored and the flasks and contents were weighed weekly. Deionized water was added to make up for weight loss due to evaporation. At the end of the shaking, a 15 mL aliquot was obtained from each flask, centrifuged, filtered, and analyzed by inductively coupled plasma spectometry (ICP).

3.5 Protocol Leach Tests Out of several leaching tests available, hoprotocol tests that are widely used in Canada and USA have been selected in this study (491. These are the Leachate Extraction Procedure of Ontario, Canada and the Toxicity Characteristics Leaching Procedure of the USEPA which are compared in Table 3-2.

3.5.1 Leachate Extraction Procedure (LEP) The LEP (Ontario, Canada) was used to investigate the leaching potential of toxic cornponents to the environment by extraction with an acidic medium [35]. A 50 g dry sample (9.5 mm or less particle size) was tumbled continuously for 24 hours in a 1 L polyethylene bottle containing 800 mL of deionized water in a rotary extractor (KBU Environmental Technologies VS309) at room temperature. A pH of 5î0.2 was maintained throughout the extraction by adding 0.5N acetic acid at 1, 3, 6, 22 h from starting time. No more than 200 mL of acetic acid may be added. After completion of the extraction, the slurry was centrifuged and the supernatant leachate was vacuum filtered through 0.45 pm cellulose acetate filter paper. The chemical composition of the leachate was analyzed using a Fisons ARL 3560 inductively coupled plasma atomic emission spectrometer (ICP- AES). The effect of particle size on metal leaching was investigated by carrying out the LEP test on the Mexican scale using three different particle sizes : -125 pm , - 4 mm and -9.5 + 6 mm. MSC was of particular interest since chemical analysis showed it had the highest levels of Cu, Zn, and Pb among the samples which were associated with particles less than 100 prn in size.

3.5.2 Toxicity Characteristic Leaching Procedure (TCLP) Similar to the LEP, the TCLP maintains a 20:1 liquid to solid ratio and requires particle size of 9.5 mm or less [36].Extraction was carried out continuously in a rotary extractor at 30 rpm for 18 hours. The choice of extraction fluid was dependent on the initial pH of the sample (taken after 5 minutes of magnetic stirring in deionized water). If the pH was <5, extraction fluid #1, composed of a buffer at pH 4.7 of dilute acetic acid and 1N NaOH, was used. On the other hand, if the pH was >5,extraction fiuid #2, composed mainly of dilute acetic acid (pH 2.8) was used instead. After the 18-hour extraction, the supernatant was filtered with a 0.45 pm membrane filter and analyzed for 33 elements using ICP-AES. Table 3-2 Comparison of Protocol Leaching Tests

Leachate Extraction Procedure (LEP) Toxicity Characteristic Leaching of Ontario, Canada Procedure (TCLP) of USA - 20: 1 liquid to solid ratio, 50 glL - 20:l liquid to solid ratio, 50 g1L - leachant : 800 rnL deionized water - leachant : buffered acetic acid - pH of solution : 510.2 if pH of sampIe6, use pH = 4.7 - particle size : 5 9.5 mm if pH of sarnple>b, use pH = 2.8 - extraction time : 24 hours with - particle size : s 9.5 mm pH adjustment @ 1, 3, 6, 22 h with - extraction time : 18 h, without pH 0.5N acetic acid (lirnit of 200 mL) adjustment - room temperature - room temperature - rotary extraction at IOrpm - rotary extraction at 30 rpm

Extraction fluid #1 is prepared by mixing 5.7 mL glacial acetic acid into 500 mL of deionized water and then adding 64.3 mL of IN NaOH (dissolve 40 g NaOH in 1 L deionized water) and diluted to 1 L. When correctly prepared the pH of this solution should be 4.93 I0.05. Extraction fluid #2 is prepared by adding 5.7 rnL of glacial acetic acid to deionized water and.diluting to 1 L. When correctly prepared. the pH of this duid should be 2.88 & 0.05. As in LEP, the effect of particle size on metal leaching was studied by using the TCLP test on MSC using three different particle sizes: -125 Pm, 4 mm, and -9.5 + 6 mm.

3.6 Extended Leach Tests The terms oxic and anoxic as used in this study conform to the definition by Berner [27]and adopted by Appelo and Postma [85] referring to groundwater environment. A distinction was made between oxic and anoxic conditions based on the rneasureable amounts of dissolved 0,(2 IO6 mollL). An anoxic environment will have a dissolved O, of 0.032 mglL or less. This value is very much lower compared to O, solubility of 8.56 mglL in pure water at 23 OC where oxic conditions prevail. 3.6.7 Oxic Conditions A two-part experiment using TCLP described above was carried out for the Mexican scale and Philippine sludge and scale by extending the duration of extraction from 18 h to 96h. Sampling of aliquots for metaf analysis was done every 1,2, 3,6,9, 18, 24, 48, 72 and 96 h. For the oxic TCLP, an aliquot of 5 mL was obtained and replaced with the extracting fluid at the specified monitoring intervals. The aliquots were centrifuged for 10 min, diluted to 10 mL, and acidified to pW2 with concentrated HNO, prior to analysis.

3.6.2 Anoxic Conditions For the anoxic TCLP. the extracting fluid in a polyethylene container was sparged with N, gas overnight at a flowrate of 225 mumin. In duplicate, 10 bottles were prepared to correspond to each sampling time. About 5 g of sodium sulfite was also added to each bottle as anti-oxidant. Based on the results of the one-day TCLP test where only Pb leaching occurred, the fine particles of the Mexican scale (-125 pm) were used with a duplicate for the coarse size (-9.5 + 6 mm). The pH was checked at the end of the extraction before taking aliquots. A 15-ml aliquot was obtained every 1, 2, 3, 6, 9, 18, 24, 48, 72 and 96 h, centrifuged for 10 min and acidified to pH*2.

3.7 Preliminary Acid Mine Drainage Potenüal Test The B.C. Research Initial Test (BCRIT) was used to rneasure acid-consuming and acid-producing components of the residues [67]. In a 250 mL beaker containing 100 mL of deionized water, a 10 g pulverized sample (100 mesh) was stirred continuously with a magnetic stirrer. The natural pH was measured after 15 minutes. While stirring, the sample slurry was titrated with IN H2S04to an endpoint of pH 3.5. Acid was introduced slowly from a titration pipette until the acid addition over a 4 hour period was 0.1 mL or less. The volume of acid consumed was noted and used for calculation of acid production potential. The sulfur content of the sample must be known from the chernical analysis to be used in estimating the acid potential of the sample. The choice of the endpoint of pH 3.5 is based on the assurnption that this represents the limit above which iron and sulfur oxidizing bacteria such as Thiobacillus feffooxidans are no longer active. Therefore, if the theoretical acid production is not sufficient to lower the pH to below 3.5, then biochemical oxidation of the wastes will not occur and the formation of acid mine drainage is unlikely. 3.8 Acld Mine Drainage Confirmation Test In addition to confirrning the acid mine drainage potential of geothermal residues, a series of experiments was carried out also to detenine (a) the best growth environment and medium for the Thiobacillus ferrooxidans, (b) the appropriate procedure for the geothermal wastes, and (c) kinetic performance of the AMD procedure. The B.C. Confirmation Test [67,97] which was found deficient in its method and monitoring scheme as discussed in Section 2.6.1, was modified and an acid mine drainage potential (AMDP) test for geothermal residues was developed. The effect of agitation, temperature, and sterilization on metal leaching and bacterial growth was investigated using this AMDP test. Supplementary techniques such as transmission electron microscopy (TEM) and light microscopy with image analysis were also performed in relation to bacterial analysis.

3.8.1 Bacteria Culture and Medium The growth medium for the Thiobacillus femoxidans (ATCC 19859) was modified from the standard laboratory technique of the American Public Health Association [lII]. This was popularly known as the 9K medium developed by Silverman and Lundgen in 1959 [77]and adopted by the APHA. The two modifications made in this work were the reduction of the FeSO, content to half the original formula and use of membrane filtration (0.45 pm pore size cellulose acetate) for sterilization of solution instead of autoclaving. The detailed procedure is described in Appendix G. The modified medium had the following constituents as shown in Table 3-3 below. Add 5 mL of Thiobacillus femoxidans inoculum to a 100 mL fresh culture medium in a 250 mL Erlenmeyer flask. Initial pH should be 2.8 - 2.9, if not add 10N H,SO, until stable. The culture is allowed to grow at room temperature without agitation. Growth of the organism can be detected by a decrease in pH, increase in Eh, and an increase in the concentration of oxidized iron as orange- brown or deep amber color of solution. Bacterial viability was checked under a light microscope with at least 800x magnification. It is not advisable to withhold completely the FeSO, from the leaching medium since the bacteria require at least 2-3 glL to 9 g/L to survive [37, 73, 77, 78, 841. The modified medium below contained 4.5 glL which was observed to be providing good bacterial growth from various trial experiments listed in Appendix A. An excess FeSO, can trigger increased formation of jarosite and iron oxyhydroxides hence shouid be avoided. 3.8.2 Acclimation of lnoculum A critical stage of the AMDP procedure is the acclimatization of the pure bacteria culture to the specific samples to be tested. To prepare a viable culture as inoculum and which will survive throughout the duration of the test, a series of acclimation steps was designed at room temperature (23-25 OC) and without agitation. This acclimation procedure is described further in Appendix G. For each culture, 5 mL inoculum was used per 100 mL of the medium. At the onset, a medium shown in Table 3-3 but using 44.22 g/L of ferrous sulfate [lIl] was used on the pure culture (Bo) inoculum of Thiobacillus ferrooxidans (ATCC 19859). Afterwards, the resulting culture (B,) was used as inoculum to a 100 mL fresh medium with the addition of 2 g ground sample (120 mesh) to be tested to obtain an acclimatized culture (83.Finally, B2was used on a freshly made culture media containing 22.1 1 g/L ferrous sulfate and 2 g sample to produce 8, culture that is ready as inoculum for the AMDP test. Prior to this, several experiments were performed to obtain the best conditions for high bacterial density and motility. These experiments, listed in Appendix A were carried out with and without agitation, room temperature (23 OC) and inside incubator (35 OC), various amounts of ferrous sulfate in medium as well as several pulp densities (weight of sample over volume of solution). The success of each experiment was determined qualitatively through bacterial viability (density and motility).

Table 3-3 Culture Medium for Thiobaciilus fenooxidans

Basal salts: in a 1 L Erlenmever flask

Ammonium sulfate (NH,),SO, Potassium chloride KCI Dipotassium hydrogen phosphate K,HPO, Magnesium sulfate MgS0,.7H20 Calcium nitrate Ca(NO,), Sulfuric acid, 10 N H2S04 Distilled water

Enerav source: in a 500 mL Erlenmeyer flask

Ferrous sulfate FeS04.7H20 Distilled water 3.8.3 Acid Mine Drainage Potential Test After evaluation of available literature and preliminary testing, the following acid mine drainage potential (AMDP) procedure was designed to further study the geothermal residues' amenability to land disposal. Ail g lasswares were cleaned in detergent, rinsed with tap water two tirnes, soaked in 20% HNO, overnight, rinsed with tap water two times and finally rinsed with deionized water. Once dry, the Erlenmeyer flasks were covered with aluminurn foi1 prior to use. The bacteria culture medium was prepared as described in Appendix G. The dry samples were pulverized in a rnortar and pestle to pass a 120 mesh Tyler screen and stored in air tight bottles prior to use. To 2 g of ground sample in a labelled 250 mL Erlenmeyer flask, 100 mL of culture medium was poured slowly. The flask was plugged with nonadsorbent cotton wrapped with gauze. The flask was swirled manually and the pH was checked. If the pH was above 2.8, a few drops of ION H,SO, were added until stable. Once the pH was stable, the flask was inoculated with an active acclimatized culture of Thiobaciilus femoxidans prepared as in Appendix G. The weight of flask with its contents without the cotton plug was taken initially to be able to monitor weight loss due to evaporation. The flask was placed at room temperature (at least 23-25

OC) with adequate ventilation. The flask was manually shaken every determination. Prior to each measurement, the fiask and contents (without plug) were weighed and deionized water was added to replace loss by evaporation. Around 1 mL aliquot was obtained and centrifuged at 1200 rpm for 10 min to separate solid from the supernatant. The supernatant was removed with a pipet and transferred to another clean 15 mL centrifuge tube, diluted to 5 mL with deionized water, acidified to pHc2 with -0.05 mL conc HNO,, and stored at 4% while waiting to be analyzed. Meanwhile 1 mL deionized water was added to ail the flasks to replace the 1 mL aliquot sarnple. Monitoring and sampling schedules are similar to that discussed in Section 3.9.1 below. The parameters monitored regularly were pH, Eh, bacterial growth, motility and density, color of solution, dissolved oxygen, and metals in leachate. When oxidativelbacterial activity had ceased as observed from the microscope and a stable pH has formed, the test was terminated. If the pH is below 3.5 and metals in the leachate were above regulatory limits, the sample is classified as having acid mine drainage potential or potential for bioleaching treatment. This test can be completed within 34weeks following inoculation. 3.9 Batch Kinetic Experiments 3.9.7 Effects of Agitation, Temperature, and Sterilzation In order to obtain kinetic information about the acidification potential of the samples, another set of experiments using the AMDP test were undertaken. To observe the effects of agitation and increased temperature on the sarnples, the flasks and contents were placed inside an incubatorlshaker (Lab-line Instruments) which operated continuously at 175 rpm and 35 OC. To detemine the effect of sterilization, control samples were prepared whereby the flasks and dry samples were sterilized inside the oven at 120 OC for 1 day and aftenvards covered with aluminum foi1 and cooled completely before use. In total, there were five simultaneous batch tests with the following designation: (A) with agitation and bacteria, inside the incubatorlshaker at 35 OC and 175 rpm, (B) stationary and with bacteria placed on laboratory bench at room temperature (23-25 OC), (C) sterile conditions : similar to B but with the sample and fiask sterilized at 120 OC inside oven for 1 day, (D) similar to C with oven sterilized samples and flasks but without any bacteria, and (E) nonsterile conditions, unsterilized medium and sarnple inoculated with acclimatized bacteria. Experiments B to E were al1 stationary experiments at room temperature. These experiments were designed to determine proper environment to be able to carry out the AMD potential test, in particular, in a laboratory with limited equipment such as in minesite, field laboratories, plant sites or in laboratories found in developing countries.

3.9.2 Monitoring and Sampling Every 3 days, the following parameters were monitored : pH (Corning pH meter model 7), Eh (Fisher Accumet pHlEh meter model 820) , bacterial growth, motility and density (MEF3 Reichert-Jung Microscope with Image analysis Hitachi KP-MIU CCD Camera at 800x magnification), color of solution, and dissolved metals (inductively coupled plasma spectrometry). Dissolved oxygen (ORION oxygen meter model 860) was also rneasured randomly in the solution to see if adequate oxygen was available for oxidation (oxygen solubility at 23 OC is 8.5 mglL from the ORION oxygen probe manual). The pH rneter was calibrated with pH 4 and 7 standards and al1 Eh readings were verified with ZoBell's solution 111 11. Utmost care was taken to avoid contamination among the replicates from the various meter probes. Each probe was rinsed thoroughly with deionized water spray and wiped with clean paper towel before doing any measurernent. 3.1 0 Microstructural Analysis 3. IO. 1 Light Microscopy with Image Analysis For routine bacterial monitoring, one drop (-20 PL) of sample taken at the surface layer of the solution and another drop taken near the bottom of the flask were both placed side by side on a labelled microscope slide each with a 22 x22 mm cover glass. These were examined at 800x magnification using a MEF3 Reichert-Jung Microscope with a Hitachi KP-Ml U CCD Camera connected to a Sony 20" television for image enhancement and attached to a Panasonic video cassette recorder. The bacterial growth and characteristics were observed visually and qualitatively as required by the AMDP procedure and noted as very high, high, medium, or low to describe density and slow, fast, and very fast for motility. A video of the bacteria at various stages of their growth as seen through the light microscope was recorded showing their motility and density. The bacterial motility was noted as motile or nonmotile, fast or slow since it was difficult to measure. For bacterial count, one drop (-20 pl) aliquot was placed on a microscope slide with a 22 x 22 mm cover glass and examined under a light microscope at 1000x magnification. Three to four fields per sample were photographed and stored in computer format as an image file using an Olympus Vanox C-35 carnera with a CCD X-77 video camera attached to a Macintosh Quadra 650 computer with an Image Scion 1.SI software. Direct bacteria ceIl count from the images was carried out to calculate the total cell count. The calculation of bacterial density is presented in Appendix B.

3.10.2 Transmission Electron Microscopy Bacteria from the flask experiments were haivested by vortexing -1 0 mL of solution to loosen bacteria adhered to particles for 10 min, centrifuging at slow speed for another 10 min and finally centrifuging at high speed for 15 min to form a white pellet and fixing overnight in 2% gluteraldehyde (vlv). The samples were later embedded in Epon 812 resin with the addition of osmium tetroxide and uranyl acetate. Thin sections (-60 nm) were cut and mounted on carbon and Fornivar-coated Ni grids and were viewed on a Hitachi H7000 transmission electron microscope operating at 75 kV. Photomicrographs were taken using a range of 24,000 to 99,000~magnification. 3 1 Geochemical Modeling The geochemical themodynamic model MINTEQAZ (version 3.1 1) [98] was used to determine equilibrium conditions and solid phase dissolution behavior in the TCLP for the fi ne-sized Mexican scale (-125 pm). All the other samples did not provide significant leaching hence the modeling was focused on the Mexican scale. Table 3-4 below lists the input data used in modeling of a closed systern. Four major mineral phases (pyrite, chalcopyrite, galena, and sphalerite) identified by XRD and microscopy were inputed as concentrations of solid phases. Since this is a complex system, only the major species identified in the chernical analysis, XRD and microscopy were included. Essentially this involved ignoring the major complex silicate phases that are substantially inert and the minor species as they were not detected in the ieachate analysis. After the first few trials, covellite showed up as a supersaturated solid with a positive saturation index. It was included in subsequent initial input. Appendix E shows the calculation for the Table 34 Input Data for Modeling Protocol Leach Tests

------Parameters Values - -- -- Concentration of major rninerals: mol/L Pyrite, FeS, 0.0120 Chalcopyrite, CuFeS, 0.0035 Covellite, CuS O.0036 Galena, PbS 0.0028 Sphalerite, ZnS 0.0120

Concentration of acetic acid: moVL LEP (0.5 N) 0.0025 TCLP (0.1N) O. 1O00

concentrations of these minerals while Appendix F is a sample of the model output with the input data on the first page. The cornponents (cations) were included as aqueous species at very low concentrations (1 x 10-l6 molal) to increase degrees of freedom. The pH was not fixed but allowed to reach an equilibrium value and was compared with the experirnental pH. Precipitation of solids was allowed only for those specified in the input file and Davies equation was used to calculate ion activity coefficients. The calculated

concentrations of the major ionic species Fe '', Cu 2', Zn '+, and Pb 2+ were compared with the actual leachate concentrations observed in the laboratory. CHAPTER 4 RESULTS AND DISCUSSION

4.1 Waste C haracterization 4.1.7 Chernical Analysis Whole rock analysis in Table 4-1 revealed that the geothermal residues are composed mainly of silica ranging from 66-82% by weight. The Philippine samples (PSL and PSC) contain higher levels of iron and aluminum content compared to the Mexican samples (MSC and MDM). The American scale (ASC) is essentially an aluminosilicate due to its high levels of alumina (10%) and silica (67%) with ail the rest of the elements in trace quantities. MSL is predominantly silica (82%) with little contamination from trace elements. From the chemical analyses, ASC and MSL will be less of a concern while the rest contain above normal crustal levels of S, Cu, Zn, As, Ba, Hg and Pb. In particular, MSC is concentrated with Cu, Zn, and Pb with around 1% each. Both PSC and PSL have the most As content. MSC, PSC, and PSL have sulfur content similar to mine tailings as shown also midway in Table 4-1. These geothermal residues may also be susceptible to producing acid mine drainage. These values are comparable to the S content of some mine tailings in Canada such as Equity Silver (3.40%) and Noranda Bell (2.99%) in British Columbia as well as Elliot Lake Quirke (3.79%) and INCO (0.69) in Ontario [91, 1121. In Canada, about $3-6 billions will be required in the near future for remediation of mine sites where 500 million wet tonneslyear of acid generating tailings are produced. In Table 4-2 is shown the crustal abundance ratio of selected species in PSC, PSL and MSC which were calculated by dividing the values in Table 4-1 to the crustal averages in Table 2-1 to get an abundance ratio. In these three geothermal residues, the levels of S,Cu, Zn, As, Ba, Hg, and Pb are elevated compared to normal earth's crust which is why they have been subjected to environmental regulations. More importantly, Pb in MSC has the highest abundance ratio at 900 times the average crustal concentration. Several techniques in this study have examined the availability of these elements to leach out to the environment and determined their true waste category. Table 4-1 Chernical Analysis of Selected Geothermal Sarnples

Species Units PSC PSL ASC MDM MSC MSL

SiO, TiO,

Fe203 Mn0 Mg0 Ca0 Na,O KP

Cr203 S CI Co Ni Cu Zn As Sb Cd Ba Hg Pb LOI CO3 Table 4-2 Crustal Abundance Ratio of Selected Geothermal Residues

Elements PSC PSL MSC

4.1.2 Radioactivity The radionuclides detected from the geothermal residues were Th-230, Pb-210. Ra- 226, Ac-228, K-40 and total U as summarized in Table 4-3. The most important radionuclide to monitor is Ra-226 since it decays to radon which is toxic when inhaled. All the activities were in the range of NORM (naturally occurring radioactive materials) with the exception of Pb-210 (t,,, = 22 y) in the MSC sarnple. The validation counting for MSC at longer duration of up to 5 days gave an average measurement of 130,000 Bqlkg (3510 pCilg) for Pb-210 at 90% confidence level. This Pb-210 activity was equivalent to a radiation dose of 32.5 mSvly received via ingestion (40 Bqlg of Pb-210 -1 OpSvly). This was 14 times the total annual effective dose equivalent from al1 natural sources of 2.4 mSv [114]. Nevertheless, it was still lower than the current occupational dose limit of 50 mSv1y [113, 1141 but higher than the Canadian public regulatory dose limit of 5 mSvly [39]. Gallup and Featherstone reported 250-400 pCi/g in the Salton Sea geothermal brines in southeastern California where the anticipated NORM regulation for solid wastes was 5 pCi/g 1381. UNSCEAR and ICRP reported that there are regions in the world where outdoor terrestrial background radiation levels appreciably exceed the NORM at 2-6 times the average natural background of 1 mSv/y : Guarapari, Brazil; Kerala, India; and Yanjiang County, Guangdong, People's Republic of China. This was due to the presence of monazite sands with high levels of thorium, uranium and radium. The inhabitants in these areas were studied between 1970-1985 and it was obsewed that there was no increase in the frequency of cancer among the population [113, 114]. Table 4-3 Concentrations of radionuclides in the geothermal residues (in Bqikg)

Th-230 Pb-210 Ra-226 Ac-228 K-4O U estimate*

PSC ~750 990 120 110 I IO 42 I22 360 * 33 ~9.4 PSL ~390 300190 80110 83120 240,:lO ~5.1 ASC ~350 420 180 530 I20 550 * 73 310 * 10 ~4.8 MDM ~220 430 94i10 67111 310112 6.2 MSC ~3801300001400 50110 36111 300*12 ~4.8 MSL ~610 <450 190*50 28110 900*60 20.0

* ppm. from activity of Th-234 < MDA (minimum detectable activity) with 90% level of confidence

4.1.3 X-ray Diffraction Several dominant phases were detected by XRD in the samples as shown in Table 4-4. The mineral name and formula are listed in order of abundance. There may be other phases present but they were not detected if they were less than 1- 2% by weight hence will not give detectable diffraction peaks. Alrnost al1 of the samples, except MDM, contain an amorphous silicate phase with a broad maximum at around 4 A. ASC does not contain significant crystalline material; it was mostly amorphous silicate, possibly aluminosilicate. MSL has halite and sylvite while MSC contains the minerals galena, sphalerite, chalcopyrite, and cubanite. PSC contains the largest amount of amorphous material. 60th PSL and PSC contain quartz, magnetite, and hernatite. MDM is a complex multi-mineral sample with little amorphous material. The important phases such as sulfides in MSC will be used as input data in geochemical modelling. In terms of their possible environmental impacts, the natural minerals and layer silicates are relatively inert while halite and calcite can dissolve and sulfides may oxidize releasing heavy metals. Table 4-4 X-ray Diffraction Data of Selected Geothermal Residues

Sample Species'

PSC Amorphous material maximum at 4.10 A Quartz, SiO, Magnetite, Fe,O, Hematite, Fe203 Jarosite, KFe3(S04),(OH),

PSL Albite, NaAISi,O, Homblende, NaCaMg,FefitSi7O2,(OH) Quartz, SiO, Hematite, Fe203 Magnetite, Fe30, Gypsum, CaS0,.2H20 Kaolinite, AI,Si,O,(OH), Amorphous material maximum at 4.04 A

ASC Amorphous material maximum at 3.43 A

MDM Quartz, SiO, Calcite, CaCO, Halite, NaCl Albite, NaAISi,O, Microcline, KAISi,O, Pyrite, FeS, Mica, K(AI,(Si,AI)O,,(OH), Montmorillonite, KAI,Si1,O,,(OH), Chlorite, (Mg,Fe),(Si,AI),Olo(OH), Dolomite, CaMg(CO,), Monticellite, CaMgSiO, Diopside, CaMg(SiO,),

MSC Galena, PbS Sphalerite, ZnS C halcopyrite, CuFeS, Cubanite, CuFe2S3 Amorphous material maximum at 3.80 A

MSL Halite, NaCl Amorphous material maximum at 4.04 A Sylvite, KCI

'Listed in order of abundance 4.1.4 Optical Microscopy The scale, sludge and drilling rnud samples were physically and chernically cornplex containing mostly particles of silica and iron and aluminum oxides. Detailed descriptions and findings are discussed further below.

Phili~~ineScale (PSC) Most particles showed a banded, oriented texture characteristic of cyclic deposition frorn a passing fluid on to the surface of the scale. The contrasting composition of successive layers reflects changing chemical composition, reduction potential and temperature in the fluid. Some of the unoriented particles are agglornerations of smaller fragments cemented together within a silicate matrix. Example of fragments with a layered texture can be seen in Figure 4. i where there are alternating bands of magnetite and silicate, with larger masses of magnetite. The particle at the top is 830 Pm while the particle in lower photograph is 330 @m. Some of the cracks have been partially filled with deposited pyrite. Scale deposition could have started at the top (pipe wall) following the sequence magnetite, silicate, pyrite, iron silicate.

Philippine Sludae [PSL) The sludge shown in Figure 4.2 is composed of very srnall, often colloidal sized particles of precipitate - primarily silica but including iron oxides and sulphide. These particles easily agglomerate, due to their small size and the sticky nature of hydrous silica. The particles making up the agglomerates are mainly submicron and porous, although srnaIl solid fragments and particles can range in size up to 30 Pm.The small particles resemble fine grained silt. The agglomerates are rounded equant lumps, that are highly porous and quite friable, easily disintegrating into finer particles.

American Scale (ASC) This particle in Figure 4.3 has a width of about 2.7 mm and is mainly silica with bits of magnetite (black color) and a vermicullar form indicating high porosity. It is composed largely of amorphous rnaterial with no distinct crystalline pattern. Mexican Drillina Mud (MDM) Particles shown in Figure 4.4 had.a variety of sires, textures, morphologies and phases mostly of equant grains from 20 prn to 180 Pm. The black spots were iron oxides and the white spots were iron sulfides surrounded by silicates.

Mexican Scale (MSC) As shown in Figure 4.5, the important minerals were occluded inside the silicate matrix. The top particle (120 hm) contains three important phases : the yellow was chalcopyrite, the greyish green was sphalerite and the off white in cubic form was galena. The lower particle (50 pm) was another silicate with pyrite at the center.

Mexican Sludae (MSL) Figure 4.6 shows the vermicullar structure of the particles indicating high porosity. The black spots were magnetite which were more obvious in top photograph depicting its formation. This indicates that silica was fonned first and the rnagnetite developed later as an envelope to the nucleated silica. The lower photograph was an overview of these porous silica particles with milky white background.

4.7.5 Toxicity Testing Toxi-Chromotest and SOS-Chromotest can provide an indication of the sensitivity of bacteria to toxic and genotoxic elements. From the results and interpretation of these tests provided in Appendix H, the samples did not appear to exhibit any toxicity or genotoxicity. This may also indicate that the genotoxins and toxins were bound within the silicate matrix or were insoluble in the extractant which makes them inaccessible to the microorganisms. The results of leaching tests and sequential extractions to be discussed in the following sections had confirmed that most of the heavy metals were in the residual phases. The XRD results in Table 4-4 also indicated that Cu, Zn, and Pb especially in MSC were present as insoluble sulfides or bound with silicates. Therefore as long as the samples were disposed and maintained in a stable or inert state, they probably would not exhibit toxicity or genotoxicity. NOTE TO USERS

Page(s) missing in number only; text follows. Microfilmed âs received.

UMI Figure 4.1 This is a particle (830 pm) of Philippine scale (PSC) showing a layered structure detailed on the lower photograph (330 pm). Note each particle is surrounded by a silicate matrix (1OOx, 250x). Figure 4.2 This is the Philippine sludge (PSL) showing agglomerates of fine particles made up of iron oxide and sulfide. The white background resembles fine grained silt. The two particles on the left are about 500 pm while the bigger particle on the right is 830 pm (50x). Figure 4.3 The American scale (ASC) has a vermicullar structure indicating high porosity with no distinct crystalline pattern. The black lines are magnetite in a silica background (50x). NOTE TO USERS

Page(s) missing in num ber only; text follows. Microfilmed as received.

. . UMI Figure 4.4 The Mexican drilling mud (MDM) contains a variety of textures and phases with particle size from 20 pm to 180 ,m.The black spots are iron oxides and the white spots are iron sulfides surrounded by silicates (50x). Figure 4.5 The Mexican scale (MSC) has minerals inside a silicate matrix. The top particle (120pm) has chalcopyrite (yellow), sphalerite (greyish green) and galena (white cubic) while the bottom particle (50 pm) has pyrite at its center. Figure 4.6 The top photograph of the Mexican sludge (MSL) shows silica surrounded by black magnetite. The bottom particles have verrnicullar structure indicating high porosity with milky white silica background (1 00x and 50x, respectively). 4.1.6 Weatherfng test Several minerals with solubility in water which are expected to be easily weathered such as halite, gypsum, calcite, dolomite, hornblende, primary layer silicates and albite [46, 47, 85, 1281 were also found present in the geothermal samples by XRD (Table 44). Only about 5% of the original amount of Al from its silicates and 30% of the transition metals Mn, Fe, Cu and Zn (probably from their suifides) were found in the leachate after three months of continuous agitation. However, none of the TCLP or LEP regulated elements (As, Cr, Cd, Ba, Hg, and Pb) were detected in the leachate in any sample. This may indicate that agitation does not contribute significantly to leaching of geothermal residues as will be shown in subsequent test results involving agitation.

4.2 Protocol Leaching Tests With the protocol particle size of 9.5 mm or less in the LEP and TCLP tests, the regulated elernents in al1 the samples were below the limits for leachate quality criteria in Table 2-2. This may be an indication that the heavy metals exist as insoluble species or may be trapped in the silicate matrix and were unable to leach within the one-day tests. However, in a worst case scenario of a finer particle size of -125 pm for the Mexican scale, Pb leached out above regulatory limits of 5 ppm. With increased surface area exposed to oxidation, about 2% (12 ppm) of the total Pb leached out in the LEP while 18% (105 ppm) was found in the TCLP leachate as shown in Figure 4.7a. All values plotted in this figure were the averages of three duplicate experiments with less than 5% standard deviation. ihere was little leaching in both TCLP and LEP for the coarse sizes (-9.5 + 4 mm) but was greater for the fine sizes (-125 pm). From these data, the TCLP can be considered a more aggressive test perhaps due to the presence of more acetic acid in the extracting fluid. In the LEP, 0.0025 mol1L of acetic acid were required to bring the pH to 5 whereas in TCLP, the leachant had 40 times more acid concentration (0.1 moVL acetic acid) at pH 3. These two procedures will be compared in detail in Section 4.9.1. TCLP has three times greater rotation speed compared to LEP which can contribute to the dissolution enhancement. Stumm suggested that faster diffusion controlled dissolution can be achieved with increased flow velocities or increased stirring [128]. However, the extended -125 um -4 mm Particle size

-125 um 4 mm -9.5 + 6 mm Particle size

Figure 4.7 Cornparison of extent of Ieaching between LEP and TCLP for Pb and Zn in the Mexican scale at various particle sizes. agitation in the weathering tests (Section 4.1.6) did not show any significant increase in leaching. Hence the difference between the LEP and TCLP must be the acid content of the leachant. The same trend was shown for Zn in Figure 4.7b where leaching was observed to increase by a factor of 3 in TCLP. Cu data was not shown since only <2 ppm leached out in both coarse and fine fractions. Based on the protocol size of -9.5 mm, the geotherrnal residues passed both Canadian and Amencan regulatory leach tests indicating they can be disposed of at bulk sizes in a secure landfill. However, the fine sized Mexican scale (-125 pm) may require treatment using appropriate technology prior to disposal. Concern over the long-term fate of the residues had prompted the investigation of the extended leaching behavior of both coarse and fine fractions of the Philippine scale and sludge and the Mexican scale. The TCLP was used but the duration of the test was increased from 18 h to 96 h.

4.3 Extended Leach Tests 4.3.7 Oxic Conditions As in the one-day TCLP test, there was no significant leaching of trace metals frorn the Philippine scale and sludge in either the coarse and fine fractions even after 96 h of continuous extraction. Therefore subsequent discussions on extended leaching results will focus on the fine sized Mexican scale (MSC) where leachate concentrations were found above the regulatory limits. In Figure 4.8 is shown a comparison of the leaching behavior of Pb from the coarse and fine particles of MSC. Each value in this figure was an average of duplicate experiments with less than 5% standard deviation. The Pb leach curves of coarse (Pb-C) and fine samples (Pb-F) shown in Figure 4.8a probably indicate a diffusion- controlled dissolution of a phase that did not reach constant solubilization of metals within the duration of the experiment as the metal concentration in the leachate was increasing over time. This typifies the general dissolution behavior in plots of concentration versus time shown in Figure 2.2, also observed by other researchers in municipal solid waste fly ash [104, 1151 and more evidently in silicates [116-1181. The leaching kinetics of Pb will be investigated further in Section 4.8.1 with supporting data for Fe and Zn in Appendix 1. As shown in Figure 4.8a, it was indicated that only about 35% of the total Pb can be leached out within one week. The unleached fraction may be assumed to be Pb existing either as an insoluble phase or trapped inside the silicate rninerals. Figure 4.8 Cornparison of extended TCLP leaching of Pb for coane PM= (- 9.5 + 6 mm) and fine particles Pb4 (-125 pm). The sequential extraction tests to be discussed in Section 4.4 also suggested that 65% of the W was bound as residuals in the silicate rnatrix and only 35% of the total Pb can be considered available. Figure 4.9 presents the dissolution behavior of Cu, Zn, Pb, Fe and Na in MSC. At longer times, the dope of Pb appeared similar to the slope of Zn (and possibly Cu but cannot be detected in this graph) which indicates that they may be associated together and were thus released at the same tirne. From the photomicrograph in Figure 4.5, galena, chalcopyrite and sphalerite were shown to be bound together in the same silicate matrix.

O 20 40 60 80 100 Time, houm

I pPb +Zn *Fe *Na +Cu

Figure 4.9 Dissolution behavior of major elements in the extended oxic TCLP for MSC fines (-125 pm). 4.3.2 Anoxic Conditions The anoxic extended TCLP leach test, perfomed on the Mexican scale fines, did not produce any signifiant leaching of the regulated elements including Pb. Likewise, the coarse samples leached out only a small fraction (4%) of Fe, Cu, Zn and Pb after 24 h of leaching (data not shown). As shown in Figure 4.10 below for the fine sized samples, Cu, Zn, and Pb leached out considerably less (-1%) during anoxic conditions. This somehow supports their association in the same silicate matrix. The low rnetal solubility of these sulfides will probably result in sequesteing of metals as long as reducing conditions prevail. Thus the leaching of Cu, Zn, and Pb may be attributed to oxidative dissolution and that, under anoxic conditions or where oxygen is depleted such as in a municipal solid waste landfill, the Mexican scale is likely to be stable.

Figure 4.10 Dissolution behavior of major elements in the extended anoxic TCLP for MSC fines (-125 pm). 4.4 Sequential Chemical Extraction The sequential extraction results for the geothermal residues are shown in Figures 4.1 1 to 4.13. They have been plotted in stack-bar graphs as the percent leached in each extraction stage (A to E) for the six regulated elements (Cr, As, Cd, Ba, Hg and Pb). Overall, little extraction was found in either exchangeable or carbonate phase (Fraction A and B) which were equivalent to LEP or TCLP protocol leach tests. The six elements were found associated with the phases involving Fe-Mn oxides and sulfides (Fractions C and D, respectively) with the highest values in Fraction El the residual silicate phase. Pb was the metal of importance that was associated with the non-residual phases in al1 samples. Below is a discussion of each element and their phase association.

Chromium In the Mexican drilling mud and scale, more than 60% of Cr was extracted as easily exchanged, also associated with carbonates and oxides. In the Mexican sludge, Philippine scale and sludge, less than 40% belonged to the non-residual phases (A-D). There was a significant fraction of this element associated with the silicate matrix of al1 the geothemal wastes. The proportion ranged from 95% of the American scale to 35% of the Mexican drilling mud.

Arsenic Only in Mexican scale and Philippine sludge was As found to be associated with the non-residual fractions with 40-60% released in the non-residual phases as Fractions A-C. All the other samples hold the As in their silicate lattice and hence were safely immobilized over time. Arsenic was generally bound in siliceous or carbonaceous material that was indigenous to the geothermal reservoir fluid [12].

Cadmium Present in small quantity, Cd was primarily solubilized in four of the six samples as easily exchanged or extracted where dissolution was greater than 80%. ASC had Cd associated with the carbonate phase (Fraction B) and the MSC with only the sulfide and silicate phases (Fractions D and E). PSC

Figure 4.11 Sequential extraction results for Philippine scale and sludge. ASC

MDM

Figure 4.12 Sequential extraction results for Amencan sale and Mexican drilling mud. MSC

MSL

Figure 4.13 Sequential extraction results for the Mexican scale and Mexican studge. Barium In al1 the samples, Ba was found mainly in the silicate phase and wiII probably not pose any danger of eventual release to the environment.

Mercury The Hg content of al1 of the geothermal wastes was generally low (~0.5ppm). It was found associated with the silicate phase in five samples and leachable in the carbonate phase in only one sarnple (Mexican sludge).

Lead More than half of the Pb content of al1 samples was associated with the silicate phase and will not be available to leach into the environment. No more than 40% of Pb was extracted in the Philippine scale and sludge and American scale while around 50% was found in the residual phases of Mexican sludge and drilling mud. In the Mexican scale, the leachable Pb (35%) was associated with carbonates, oxides, and suifides and 65% was lodged in the residual phase. There is agreement with protocol TCLP leaching results in Figure 4.8a where the maximum leaching for Pb reached 32% after one week and was increasing at a very slow rate thereafter.

4.5 Preliminary Acid Mine Drainage Potential Test The static test BC Research Initial Test provided a preliminary indication that 3 of the 6 samples had AMD potential. A calculation of the acid consumption (AC) and the acid production potential (APP) was illustrated in Appendix C. The APP was based on the % S which can be obtained from chemical analysis as in Table 4-1. The difference between the APP and AC in kg H2S04/tonne material, determines the AMD potential. If the value was positive then the waste was classified to have acidification potential which required further testing. If it was negative, it was classified as a nonacid producer. In Table 4-5 is shown that ASC, MSL and MDM can be considered as nonacid producer since they produced negative values of AMD potential. This may be due to their low S content and high pH with sufficient alkaline buffering capacity. On the other hand, MSC, PSC, and PSL were found to have acid mine drainage potential with positive values. Although these figures were low relative to mine heaps or dumps where AMD potential were in the order of 1O to 100 times more [66. QI],it was necessary to verify their long-tenn acidification potential involving microbial mediation of the iron and sulfur oxidizing bacteria, Thiobacillus ferrooxidans, present in the natural environment.

Table 4-5 Preliminary acid mine drainage potential test results

Sample PH %S APP* AC* APP - AC*

MSC 7.2 3.40 100.00 5.40 +98.0 PSC 3.5 1.70 51 .O0 0.98 +50.0 PSL 4.5 0.46 14.00 0.98 +13.0 MSL 7.8 0.01 0.31 2.90 -2.6 ASC 9.7 0.03 0.92 6.90 -5.9 MDM 8.3 0.10 3.10 86.00 -83.0

- .. * in kg H,SO, per tonne of sample

4.6 Confirmation of Acid Mine Drainage Potential 4.6.7 Bactehal GroW and Acclimation Based on the trial expariments found in Appendix A, the best Thiobacillus fenooxidans culture and medium providing highest bacterial density and motility were the pure culture ATCC 19859 and APHA media, respectivety. The pure culture was purchased from the Arnerican Type Culture Collection in Maryland, USA and the composition of the APHA medium is given in Table 3-3. These were used in acclimatization of the bacteria to the samples to be tested. A five-minute video (not submitted with this thesis) was produced which documented bacterial growth from the lag phase to the endogenous phase. In Figure 4.14a is shown a photomicrograph of viable acclimatized bacteria taken from a light microscope at 800x magnification. The bacteria were observed to be at the end of the logarithmic phase and entering the stationary phase after one week with a density of 2 x 10 cells per mL. A sample calculation for estimating bacterial density is shown in Appendix B. In the stationary phase as shown in Figure 4.14b, the bacterial population density remained steady at 1 to 2 x 106cellsImL for about two more weeks until they begin to die (Figure 4.14~).Also during this period, they were observed to have greater motility existing as single, in pairs or even short chains. These three figures represent high, medium and low bacteria density which can be used as a benchmark when monitoring bacterial growth using a light microscope. Precipitation and biofilm formation had been strongly linked with the tendency of Thiobacillus ferooxidans to grow on surfaces. Figure 4.14d was taken from a light microscope at 1OOOx magnification from a two-week old bacteria culture. It shows bacteria interacting with solid particles and the presence of an organic capsule surrounding the solid samples. According to Rojas-Chapana et al, this capsule may contain phosphorus and colloidal sulfur particles that act as a reactive medium for the sulfur metabolism [119]. Atkinson and Fowler as cited by Rossi [48] rnaintained that the biofilm or capsule which was pervious to air and nutrients for depths ranging from 50 to 150 pm, was a biochemically inert organic polymer rnatrix where inorganic material such as iron oxyhydroxides rnay also be trapped. The TEM photomicrographs (60,000~)in Figure 4.1 5 showed the intimate association of bacteria with amorphous and crystalline minerals and soma globules. The fine crystalline particles surrounding the bacteria could be ferric ion precipitates such as ferric oxyhydroxide or jarosite. The average size of each bacteria was 0.5 lm. There was an apparent shrinkage or distortion due to air drying and vacuum collapse of the cell wall as well as the absence of fiagellum since these bacteria were obtained at the end of a three-week leaching which was nearly the end of the stationary phase. A dark outline of the cell wall rnay indicate extracellular adsorption of heavy metals such as Cu, Zn or Pb onto the bacteria such as that reported by Davis et al [120]. However this couid not be confirmed by TEM-EDX since the concentrations were less than 1%. In Figure 4.16 is shown the Thiobacillus fenooxidans cell at higher TEM magnification (99,000~).Cell division of T. ferrooxidans is mostly by constriction but occasionally by partitioning [12 11 and this was captured in this photomicrograph. The celi about to divide into two cells is about 1.7 pm in length with a 0.5 pm diameter. There were two globules on each side of the bacteria each about 0.2 Pm in diameter. These globules, according to Rojas-Chapana et al could be colloidal sulfur, a source of energy for the bacteria, that act as energy reservoir for later use [Il91. However, these globules could also be cell remnants or section of dead cells. Figure 4.14a A p hotomicrograph of acclimatized Thiobacillus fenooxidans taken from a light microscope with an estimated high density of 2 x IO7 cellslmL (800~). Figure 4.14b A photomicrograph of acclimatized Thiobacillus ferrooxidans taken from a light microscope with an estimated medium density of 2 x 1o6 cellslmL (800~). Figure 4.14~ A photomicrograph of acclimatized Thiobacillus ferrooxidans taken from a light microscope with an estimated low density of 1.O x 1O6 cells/mL (800x). Figure 4.14d This is a photomicrograph of Thiobacillus ferooxidans taken from a light microscope showing a two-week old culture that developed a cloud-like capsule around the samples. Note the bacteria (white spots) inside the capsule and also on the surface of the solid samples (1000~). NOTE TO USERS

Page(s) m issing in num ber only; text follows. Microfilmed as received.

. . UMI

Figure 4.1 6 This is a TEM p hotornicrograph showing a Thiobacillus femoxidans cell (1.7 pm by 0.5 pm) about to partition. Note the two globules (G) on each side as well as the jarosite (J) precipitates (99,000~). Dissolved oxygen from ambient air was found sufficient in this study to sustain microbial growth of Thiobacillus femoxidans. The dissolved oxygen (DO) measurements near the surface of the liquid were in the range of 5-7 mglL throughout the experiment. This amount appeared to have supplied the oxygen requirement of the bacteria which is best maintained at close to oxygen saturation in water at 6.6 mglL 1891. (Refer to Section 2.5.3.5.) The DO was slightly lower in the agitated flasks inside the incubator at 35 OC presurnably due to the lower solubility of oxygen at higher temperature. In Figure 4.21 is shown that there was a progressive increase in redox potential (Eh) in al1 samples throughout the experiment, indicating that oxidation was probably taking place [77]. It was also seen in the videotape taken under the microscope that there was microagitation of bacteria under the glass slide due to regular air movement in the room such as opening or closing of the door. This may be related to possible agitation inside the stationary flasks. Jarosite and iron oxyhydroxide formation is not desired in bacteria culture and bioleaching as reported by several authors. The detrimental effects of jarosite formation are : (a) it diminishes the available femc ion in solution; (b) it limits the amount of biomass retention since ferric ion deposits occupy the bulk of the available space; and ( c) it creates a kinetic barrier because of the slow diffusion of reactants and products through the precipitation zone [73, 861. Jensen and Webb provided an excellent review of jarosite formation linked to wall growth and biofilms [73]. They cited a pH range of 2 - 2.5 when bacteria attain maximum growth and a range which inhibits ferric ion precipitation. Pesic and Kim reported that bacteria cells serve as nucleation sites for the formation of jarosite [122]. In this study, a yellow to orange brown solution color was observed after one week when the pH was 2.5-3.0 for stationary flask and pH 2 for agitated fiask, and a fine yellow precipitate was visually obsewed only on the 4th week towards the end of the experiments when the bacterial activity had declined. The yellow orange to orange brown precipitate was determined by X-ray diffraction as jarosite (K)Fe, (SO,),(OH),. Barron and Lueking reported that precipitates occurred at the stationary phase and color change was observed during logarithrnic phase with no precipitate [84]. There was more precipitation found in stationary flasks than in shake flasks. This was because the formation of ferric ion precipitates, especially jarosite, was dependent on pH as can be seen from Equations 2 and 5. In Figure 4.17 is shown the pH-Eh versus time graphs of the Philippine scale and sludge and the Mexican scale. These pH values had a standard deviation of k0.05 units taken from three replicates. The pH's of the left column (agitated flasks) had lower pH of 2 compared to pH 2.5 - 3 for the right column (stationary flasks). In agitated flasks, there was probably increased oxidation of ferrous ion and release of more acidity which lowered the pH. A pH of 2 was required to prevent precipitation [73].Also in Figure 4.17 is shown a sudden rise in pH at the onset which was

also observed by several researchers [48, 731. This could be due to the consumption of H,S04 during oxidation of Fe *' to Fe '+ catalyzed by Thiobacillus femooxidans (Equation

5). Afierwards the pH decreased as sulfur was oxidized to SO, 2' producing acidity (Equation 4). The Eh profiles also in Figure 4.17, especially for the stationary flasks increased steadily indicating progressive bacterial growth and metal solubilization [37,771.

4.6.2 Acid Mine Drainage Potential (AMDP) Test After 23 days of bioleaching using the AMDP test, Pb was not found in the leachate in any sample or experimental test conditions. For the Philippine scale and sludge, less than 2% leached out for both Cu and Zn with or without bacteria. On the other hand, for the Mexican scale, almost 700% of Cu and Zn were released in the leachate both for agitated and stationary experiments, respectiveîy. These results are shown in Figures 4.18 and 4.19. In general, the agitated flasks had leached out more metals by as much as 10% for MSC, 0.3% for PSL and 0.5% for PSC. Since Pb is the regulated element (and Cu and Zn are not), al1 the samples pass the regulatory requirement for leachate quality criteria in Table 2-2 and therefore should not be classified as hazardous wastes. Also, PSC and PSL leachate quality were within the limits for disposal to agricultural lands listed in Tables 2-3 and 2-4 while MSC requires special attention. As shown in Table 4-5. MSC, PSC and PSL were found to have AMD potential based on the preliminary static test. Further investigation using this kinetic test confirmed that they al1 had acidification potential since the final pH of al1 the solutions was below 3.5 as illustrated in Figure 4.20. However, in terrns of releasing heavy metals, only MSC has a confirmed acid mine drainage potential for Cu and Zn but not for Pb. In spite of the acid generation which was supposed to promote breakdown of the sulfide mineral lattice hence more dissolution of metals, the bioleachate only contained the soluble metal sulfates of Cu and Zn and not the insoluble lead sulfate. 1L ' 200 O 5 10 15 20 25 30 Time, dry8 Time, days

PSL

11 200 O 5 10 15 20 25 30 Time, dayr Time, dap MSC

11 J 200 O 5 10 15 20 25 30 O 5 10 15 20 25 30 Time, days Time, days

Figure 4.17 These graphs show the inverse relationship of pH-Eh change over tirne. The graphs on the left column (a, c, e) were agitated experiments at 35 OC while graphs on the right column (b, d, f) were stationary experirnents at 25 OC. PSC

Time, days

PSC

O 5 10 15 20 25 Time, days

100

80

O 60 C CuI 0 s- 4Q

20

O O S 10 15 20 25 Time, days

Figure 4.18 Agitated experiments at 35 OC: overview of metal leaching over time frorn PSC, PSL and MSC during acid mine drainage potential tests. PSC

O 10 15 20 25 Tims, dap

O 5 10 15 20 25 T ime, days

O S 10 15 20 25 Time, days r Cu + Zn t Pb

Figure 4.19 Stationary experiments at 25 OC: overview of metal leaching over time of PSC, PSL and MSC during acid mine drainage potential tests. Pb, however, which forrned an insoluble sulfate was not found in solution since it could have precipitated immediately and associated with the other precipitates such as ferric sulfates and oxyhydroxides. Silver and Tonna [144] and Tomizuka as cited by Rossi [48] detected lead sulfate (anglesite) through X-ray diffraction analysis of the insoluble residual rnatter from bioleaching of galena or lead sulfide concentrate containing 42% Pb and 30% S. Two methods as described in Section 2.5.2 can produce lead sulfate. In the presence of ferric ion. the oxidation can be expressed by Equation 13 as follows

PbS + Fe,(SO,), -+ PbSO, + 2FeS0, + S0 (13)

While with direct biological mediation, lead sulfate can be formed via Equation 14, with the Thiobacillus ferrooxidans also involved in the oxidation of elemental sulfur and ferrous ion generated in Equation 13.

bacteria PbS +20, -+ PbSO,

The low metal leaching in PSC and PSL can be due to a number of reasons. Firstly, MSC has a much higher metal concentration (which were identified as metal sulfides by XRD) than in PSC and PSL. On the average, MSC's original composition reported in Table 4-1 had 50x more Cu, 1OOx more Zn and 70x more Pb. Secondly, the heavy metals in PSC and PSL may not be in sulfide forrns which should be amenable to bioleaching. Thirdly, as shown in the photomicrographs in Figures 4.1 and 4.2, they were not accessible to the bacteria and to the leach solution since they were bound by strongly cemented silicates.

4.7 Batch Kinetic Experiments The original concentration of Cu, Zn, and Pb in MSC was about two orders of magnitude higher than in PSC and PSL as shown in Table 4-1. Thus the low levels of Cu and Zn in PSC and PSL rnay lower metal recovery cornpared to MSC. This will be reflected in the following discussion where the overall fraction extracted for Cu and Zn is almost 100% in MSC with less than 2% for PSC and PSL. Based on sequential extraction, rnost of the heavy metals were lodged in residual phases. According to X-ray diffraction. PSC and PSL contain silicates and crystalline particles that are not easily weathered as was reported also from the accelerated weathering test in Section 4.1 -6. The effects of sterilization, agitation and temperature, and bacteria on metal leaching are presented beloiiv. The results plotted in the graphs are averages of three replicates.

4.7.1 Effect of Sterilization There is no marked difference in the overall leaching efficiency of Cu, Zn, and Pb under both sterile and nonsterile conditions, as shown in Figures 4.22 - 4.24 for PSC, PSL and MSC, respectively. Without agitation, around 80% of Cu and Zn had leached out from MSC after 23 days. The leaching of Pb in MSC showed a decreasing pattern from 4% at the start to near zero at the end of the test. There is evidence of secondary precipitation of a Pb compound, possibly PbSO,, which is perhaps why it was not detected in the leachate. Both Cu and Zn had less than 2% leached out in PSC and PSL in both sterile and nonsterile conditions. These results could mean that sterilization of sample and the leaching medium was not critical as far as the growth of Thiobacillus ferrooxidans was concerned. This was also verified qualitatively under light microscope where the bacterial density and rnotility appeared to be the same for both conditions. The cotton plug on the Erlenmeyer flasks also served to filter any airborne contaminants and although spores and other microorganisms can enter the flask, they will not likely survive the low operating pH (c 3). The AMDP test can therefore be performed in nonsterile conditions.

4.7.2 Effect of Agitation and Temperature The intent of these experiments was to compare whether the AMDP test can be carried out using either stationary or shake flask technique. The results are shown in Figures 4.25 - 4.27 for PSC, PSL and MSC. In the Mexican scale in Figure 4.27, the maximum leaching efficiency (-100%) was attained for Cu and Zn within the first two weeks in agitated experiments. On the other hand for the Philippine scale and sludge. agitation and higher temperature did not appear to have any significant effect (~0.5%)on metal leaching. With these results, the duration of the test could probably be shortened if shaking technique was used at higher temperature (35 O C). The shake flask technique appeared to be more aggressive as it leached out approximately 20% more than the PSC

O 6 10 15 20 2S 30 Time, days

1 0 O 10 1s 20 26 30 Time, days

MSC

O 5 10 15 20 26 30 Tirne, days

Figure 4.20 These three graphs show variations in pH over tirne for the five sets of batch kinetic AM0 experiments. The agitateci experiments were performed at 35 OC while the rest at 25 OC. O 6 10 15 20 25 30 Time, days

PSL

00 O 5 10 15 20 PL 30 Time, days

MSC

O 5 10 16 20 2 30 Time, days

Figure 4.21 These three graphs show variations in Eh over tirne for the five sets of batch kinetic AMD experiments. The agitated experiments were performed at 35 OC while the rest at 25 OC. stationary flask in the MSC. Thus if there was no provision for an incubator-shaker in a laboratory, the stationary flask technique can still be used but the duration of the test has to be increased from three to four weeks. On the other hand, if the shake flask technique was used, then the test could provide reasonably good results within two weeks. Temperature may not be a concern in warm or tropical climates where the average ambient temperature is 35 OC or higher especially in geothermal areas. It was found also during bacterial acclimation that the culture of Thiobacillus femoxidans adapted well at room temperature yielding higher density than those shaken and incubated at 35 OC as shown in the bacteria growth curve in Appendix G. In addition, the particle size could be decreased also from the present -125 pm to as low as possible, perhaps -75 + 45 pm as suggested by several authors [48. 91, 971.

4.7.3 Effect of Bacteria The contribution of bacteria to the overall leaching was found more significant in MSC than PSC and PSL as shown in Figures 4.28-30. It can be observed from Figures 4.28 and 4.29 that the bacteria had no effect in the leaching of heavy metals Cu, Zn , and Pb from PSC and PSL. From XRD data in Table 4-4, Cu, Zn and Pb sulfides were identified to be present in the Mexican scale but not in PSC and PSL. From Figure 4.1, only the iron sulfide pyrite (FeS,) was found present in PSC interspersed with magnetite and silicate. As shown in Table 4-1, PSC and PSL contain smaller amounts of sulfur compared to MSC which may not al1 be in sulfide form amenable to bacterial leaching. With rnicrobial mediation in MSC, metal sulfides were oxidized either by oxygen (Equation 1) or ferric ion (Equation 3) to form metal sulfates which were generally more soluble except for PbSO, which is sparingly soluble (K,, = 1.6 x 'IO9). Due to the formation of insoluble Pb species, leaching of Pb ions to the environment may not occur. In addition, solubilized lead can precipitate with ferric hydroxide or jarosite (orange brown color) which were common indicators of acid mine drainage. Thus metal immobilization of Pb, the regulated element, may be a beneficial consequence of microbial mediation, as reported also by several authors 1123-1251. Both bacterial rnotility and density were found to have direct effect on metal leaching than bacterial density alone. In general, higher leaching efficiency was observed in agitated experiments which yielded more active bacteria though lower density than the stationary experiments. The bacterial motility was monitored visually as speed or activity was dimcult to measure [126]. It was noted as very fast, fast or slow and was captured vividly in videotape.

4.8 Reaction Rates and Mechanisms 4.8.7 Chernical Leaching Kinetics A reaction rate equation has been developed for Pb which can be used in predicting the rate of metal release over time if disposed in a landfill. Results reported in Section 4.3.1 on extended TCLP under oxic conditions have indicated that most of Pb extraction follow a dissolution mechanism that was controlled by diffusion, Le., transport of the reactants and products through a layer of the mineral surface as depicted in Figure 2.2. Initially, rapid leaching occurs which is surface-controlled reaction at the exposed surface area of the particles containing sulfides. As the inert or leached layer grows thicker, the distance of diffusion is increased which in turn slows the rate of leaching. An excellent reference on the theory and mathematics of diffusion is by Crank [145]. In the chernical leaching of the Mexican scale, the rate equation corresponds to a diffusion-controlled reaction known as the parabolic rate law. This was described by Stumm and Wollast [127-1291 as r = dC/dt = kt -Il2 mol. L-'. h" (1-1) where k is the reaction rate constant (mol.L''.h ""). By integration, the concentration in solution, C (mollL), increases with the square root of time as such

mol. 1-'

In Figure 4.31 is shown Pb concentration increasing linearly as a function of the square root of time. The plot for the coarse fraction fits linearly through zero. However for MSC fines, the intercept was displaced which indicates exposure of a more leachable Pb fraction or species at the initial period of the leaching. The intercept Cowas the number of moles rapidly exchanged between the mineral surface and the solution and in this case has a value of 145 pmollL from regression analysis. The grinding appears to have liberated a significant amount of highly leachable Pb at a rapid rate occurring within the first hour of the experiment. Once this first phase has been exhausted, a parabolic dissolution rate was observed which indicates a slow dissolution behavior of a less soluble or less accessible phase toward equilibriurn. This was also observed in Figure 4.9 for the other rnetals. Additional dissolution kinetics data for Fe and Zn may be found in Appendix 1. Cu was not plotted due to its very low recovery (~0.5%). The dope of the line represents k,the parabolic rate constant with values of k, .4.6 x IO6 mo1.L-'.h for coarse particles and k, = 1.1 x 1O4 mol.L".h -'"for fine particles as shown also in Figure 4.31. Thus to account for the dependence of the rate on particle size, the rate expression for Pb involving both coarse and fine particles should include a term a which represents the coarse fraction and fine fraction, a, and a, respectively with the relationship {aa,=1 - a, ). Equation 2-1 then becomes

and differentiating it gives

Thus the rate equation for the dissolution of Pb in the Mexican scale can be expressed as

The slower rate r, of the coarse fraction was obviousiy controlling the overall reaction rate. Equation 5-1 confirms the relationship between leaching and particle size: as particles are reduced ta finer sizes increasing the surface area, more leaching was expected to occur. Thus, eliminating or isolating the fines from the coarse fraction prior to disposal can reduce the leaching by a very wide margin. -- O 5 10 15 20 25 Time, days

O 5 10 15 20 25 Tirne, days

O 5 10 15 20 25 Time, days

Figure 4.22 The three graphs show the effect of sterilization on rnetal bioleaching in the Philippine scale for Cu, Zn and Pb, in stationary flasks at 25 OC. Tirne, days

Time, days

Figure 4.23 The three graphs show the effect of stenlization on metal bioleaching in the Philippine sludge for Cu, Zn and Pb, in stationary flasks at 25 OC. O 10 15 20 25 Time, days

O 5 IO 15 20 26 Timr, days

100

80 8 a0 a C = 40 8 20

h - v w 0- 0- ' = œ - O 5 10 iS 20 25 Time, days

Figure 4.24 The three graphs show the effect of sterilization on metal bioleaching in the Mexican scale for Cu, Zn and Pb. in stationary flasks at 25 OC. L

p 1*5 IL r

A I P 1 vA Av w -a# *ai I 0.5

O O 5 10 15 20 25 Time, days

2

O 1.5 rQ) Zr! P 1 0, I œ II- I--I m

A A ae 0.5 A7 .- T 'b/

O I O 5 10 15 20 25 Tirne, days

O O 5 10 15 20 25 Time, days

Figure 4.25 The three graphs show the effect of agitation and higher temperature on metal bioleaching in the Philippine scale for Cu, Zn and Pb. Tirne, days

O 5 10 15 20 25 Time, days

2

1.5 aa C I aIl C Pb 0 '0.5

I I 9 - I 0- 0- 9 I w - O 5 10 15 20 25 Time, days + Agitated, 35 OC &tationary, 25 OC

Figure 4.26 The three graphs show the effect of agitation and higher temperature on metal bioleaching in the Philippine sludge for Cu, Zn and Pb. Cs!

O 5 10 15 20 25 Time, days

O 5 IO 15 20 25 Tirne, days

100

80 'Lt 60 m s2 40 20

00 O 5 10 15 20 25 Time, days

+ Agitated, 35 OC + Stationary, 25 OC

Figure 4.27 The three graphs show the effect of agitation and higher temperature on metal bioleaching in the Mexican scale for Cu,Zn and Pb. 2

= 1.5 tQ, I 1 a9P I 0.5

O O 5 10 15 20 25 Time, days

O 5 10 15 20 25 Time, days

O 5 10 15 20 25 Time, days

+ With bacteria Without bacteria

Figure 4.28 Effect of bacteria on metal bioleaching in the Philippine scale for Cu, Zn and Pb in stationary experiments at 25 OC. O 5 10 15 20 25 Time, days

2

p IS r O a#crr 1 9 0.5 b O 1' O 5 10 15 20 25 Time, days

O O 5 10 15 20 25 Time, days

+ With bacteria Without bacteria

Figure 4.29 Effect of bacteria on metal bioleaching in the Philippine sludge for Cu, Zn and Pb in stationary experiments at 25 OC. Time, days

'O0 0

O 5 10 15 20 25 Time, days

O 5 10 15 20 2s Tirne, days

+ With bacteria Without bacteria

Figure 4.30 Effect of bacteria on metal bioleaching in the Mexican scale for Cu. Zn and Pb in stationary experiments at 25 OC. 4.8.2 Metal Solubilization Rate Leaching of Cu and Zn in MSC was almost at the same rate but ver- low or negligible for Pb. In Figure 4.32, the extent of leaching was compared for these three metals as a function of time. Of the five sets of experiments, the highest metal recovery was observed in the agitated flasks more than the stationary ffasks except in the case of Pb, as discussed earlier in Sections 4.6 and 4.7. Premuzic et al reported 85-90% recovery for Cu and Zn from bioleaching of geothermal sludges at 55 OC and pH 1-2 after only one day (21, 241. Similar recovery was obtained in this study for Cu and Zn but after one week and at 35 OC and pH 2-2.5. The rnetal dissolution rate for Cu, Zn, and Pb in the Mexican scale was at its maximum within the fint week of the AMDP test. As shown in Figure 4.33, the solubilization rate for Cu, Zn, and Pb was more enhanced in the agitated experiments at the beginning. However, both agitated and stationary flasks approached the same rate at the end of 23 days when the bacteria would have reached the end of their stationary phase. In Table 4-6 is presented the maximum solubilization rate that occurred within the first week especially with agitation. This indicates that as far as the AMDP test was concerned, the agitated test is probably more aggressive than the stationary test by as much as three tirnes.

Table 4-6 Maximum Solubilization Rate in Batch Process

Metal Ag itated, Stationary, Rate 35 OC, 25 OC, Difference mg/Ud mglUd Figure 4.31 Dissolution kinetics for Pb in the coarse (-9.5 + 6 mm) and fine (-125 pm) fractions of the Mexican scale. Data on concentrations over tirne were obtained from extended TCLP under oxic conditions. O 5 10 15 20 25 Time, days

O S 10 15 20 25 Time, days

Time, days + ~gitsted + Statîonsy + Stedle * Control -F Nonsterile Figure 4.32 Overall percent leaching over time of Cu, Zn and Pb for the five sets of AMD potential experirnents on the Mexican scale. The agitated experiments were performed at 35 OC while the rest at 25 OC. O 4 8 12 16 20 24 Time, days

O 5 10 15 20 25 Time, days

O 5 10 15 20 25 Tirne, days

r Agitated, 35 OC 4 Stationary, 25 OC

Figure 4.33 Soiubilization rate for Cu, Zn, and Pb in the Mexican scale during the AMD potential test for both agitated and stationary experiments. 4.9 Cornparison Between Acid Leaching and Bioleaching 4.9.1 Between TCLP and LEP TCLP appean to be more aggressive than LEP since it leached out more metals. A good example is illustrated for Zn and Pb in Figure 4.7. The concentration of Cu in the leachate was less than 2 ppm in al1 samples and was not reported in the graphs. The major component of the leachant was acetic acid (HAc) in both procedures. Of the three ions, Pb2+has the highest affinity for acetate ion followed by Zn2+andthen Cu2+[130].Therefore because the acetate concentrations were higher in the TCLP, it was most likely that lead acetate would be in abundance in the leachate especially since this was a Pb species that was highly soluble. The TCLP extracting fluid in the Mexican scale was 1 L of acetic acid solution (0.1 N HAc) at a pH of 3. The available {H'} and {Ac-} were 1.33 mmollL each. Whereas in the LEP, the extracting fluid was 1L of deionized water with periodic addition of 0.5N H Ac to the solution to maintain a pH <5 * 2. Total acid added for Mexican scale in LEP was typically 5 mL rnaking the available {H+} and {Ac} as 0.2 mmol/L. The amount of {Az} available in TCLP was therefore 7-fold more than in LEP. With this advantage of excess concentration of {H'} and (Ad), the TCLP solution was a better buffer and was able to maintain the H+ion demand during the test due to redox reactions or complexation. Since thete was greater supply of {Ac3 ions in the TCLP extracting solution, there was more acetate complexation with Cu2+,Zn2+, and pb2+which increased their solubility in solution [131]. However, since Cu2' has the least affinity for acetate ions, its leaching was not significant (c2 ppm) compared to Zn2+and Pb2+. As shown in Figure 4.7, Zn2+and Pb2+ were released more in the TCLP leachate than in LEP leachate. In addition, the agitation speed of the rotary extractor was three times more in the TCLP than in LEP (30 rpm against 10 rpm) which could prornote weathering. Thus, in the case of the geothermal residues, the American TCLP appears to be a more aggressive test for leachability than the Canadian LEP.

4.9.2 Between TCLP and A MDP In the case of the bioleaching, AMDP leached out more Cu and Zn than Pb, which was the reverse behavior obsewed in TCLP. Only the agitated procedure of the AMDP was used for cornparison with TCLP so that weathering of particles was a common process in both tests. In the case of the Mexican scale in Figure 4.18, it was shown that AMDP was a more aggressive test than TCLP (Figure 4.9a) since more metals were dissolved in the leachate. However, almost none of the regulated element (Pb)leached out. The major components in AMDP solution were H', SO, ", and Fe 24 with Cu ", Zn 2', and Pb 2+ from the sample. Of the three metal ions, Pb 2' has the highest affinity for sulfate ion followed by Zn2+and then Cu2' [130]. Therefore it is likely that PbSO, will precipitate first. PbSO, which is a sparingly soluble to insoluble compound (K,= 1.6 x lob) will probably be in the solid phase and hence will not be detected in the leachate. Though the AMDP test produced more Cu and Zn, the regulated element Pb was found to have been immobilized through bacterial action. From these results, it was not clear which was a better test in determining the hazard potential of the residues. It would be most prudent to use them conjunctively as they provide two different scenarios that may happen in field conditions.

4.9.3 Evaluation of the A MDP Procedure Several criteria were used to evaluate laboratory tests that were most appropriate for prediction of acid mine drainage potential. Most common evaluation criteria that were mentioned in Tables 2-5 and 2-6 were simplicity, time required, equipment required, cost, ease of interpretation, and correlation with field data [68,911. The AMDP procedure developed at the University of Toronto for geothermal residues can be described as low cost, simple, low technology, reproducible, requires little operator training, no specialized equipment needed, tolerant of sample variations, and can be performed in a nonsterile environment. Table 4-7 below shows a cornparison between the BC Research Confirmation and AMDP test. It shows the advantages of the AMDP procedure since it has the versatility to be used more widely in laboratories with limited equiprnent, in actual mine/wellfield sites, and especially in developing countries where use of geothermal energy was rapidly increasing. The major drawbacks of the BC Research Confirmation Test in relation to geothermal residues were: 1) the pulp density was too high 2) no indication of redox reactions 3) no indication of bacterial viability 4) no FeSO, in the leaching medium. Table 4-7 Cornparison between BC Research Confirmation and AMDP Tests

B.C. Research Uoff- Acid Mine Drainage Confirmation Test Potential Test - lnoculum no specific guideline reproducible method of bacteria culture Nutrient Media only basal salts, no FeSO, basal saits arrd 50% of the standard 9.0 glL Fe " Particle size 400 mesh 120 rnesh Solid concentration 15-30 gRO mL (amount depends on S content) Agitation With agitation, shaker With or without agitation, speed not specified (175 rpm if shaker available) Temperature lnside incubator at 35 OC Room temperature (between 23-25 OC) Sterilization Media to be autoclaved Media to be filtered with 0.45 pm ceilulose acetate Measurements pH only, metals are pH, Eh, bacterial growth, optional metals

There have been several studies to validate the results of AMD prediction tests and most of thern provided good correlation with field conditions. Around 44 static and kinetic tests were performed on 22 rock samples from seven rnetal mines in British Columbia and Yukon, Canada [66,93, 951. Samples of tailings and waste rock were obtained from both active and abandoned mine sites with sulfur tontent in the range of 0.1 3 to 49.2%. The static and kinetic tests correctly predicted the formation of AMD in al1 but six cases: three were incorrect and three were inconclusive due to inconsistent data. The Energy, Mines and Resources Canada carried out a comprehensive study to evaluate AMD prediction techniques used in Canada and USA [QI]. Eleven procedures such as those listed in Tables 2-5 and 2-6 were undertaken on 4 waste rock samples and 8 tailings samples for a total of 12 samples. The prediction of the AMD correlated well with field data in al1 but one tailings in static procedures (positive prediction but there was no AMD in the site) and two tailings in kinetic procedures (the reverse happened for both AMD predictions, one positive and the other negative). In the US, seven of the 56 mining-related sites were reviewed to determine if acid generation predictive tests were conducted at individual sites and if so. compare with actual AMD situation [68]. This study is currently going on. Although notning in literature has been reported on the occurrence of AMD from geothemal residues, these examples provide a higher confidence put on AMDP prediction techniques. In this study, the acid mine drainage potential of geothemal residues was evaluated using the AMDP test. It was found that the Mexican scale, Philippine scale and sludge have the potential for acidification since pH at the end of the test was below 3.5. However, on the basis of solubilizing metals and releasing them to the environment, only the Mexican scale has the potential to do so for Cu and Zn but not for Pb. These results require further verfication through testing of more samples and field observations.

4.1 0 Geochernical Modeling The use of MINTEQA2 in geochemical modeling can provide concentrations and speciation under equilibrium conditions. The interpretation of the results and comparison with experimental data can be challenging. Table 4- 8 below shows the results of modeling of the Mexican scale in TCLP under anoxic conditions, Le., closed system which closely fit the experimental conditions (the bottles were completely sealed throughout the test). All the other sarnples did not provide significant leaching hence modeling was focused on the MSC. The model outputs were verified by making material balance calculations and verification whether the Y, and K,, values were similar to those reported in literature. From Part 6 of Output File (Appendix F), the saturation indices of al1 input solids were zero which means that they were the controlling solids allowed to precipitate. Shown in Table 4-8 were the activities of each ions as obtained from Part 3 of Output File as Type I and II species. Type I are components as species in solution while Type II are complexes, free ions, adsorbed species in solution. There was a variation in the Ksp values from different sources [85,98, 130, 1431. The Ksp from Lindsay [143] appears to be close to the calculated Ksp especially for H,O, CuS, ZnS, and PbS. Table 4-8 Summary of Results from Geochemical Modeling of the Mexican Scale

Toxicity C haracteristic Leachate Procedure (TCLP)

Parameters Experimental Model Dissolved Species

Concn, mol/L Activity, molIL* Cu2+ 1.8E-05 1.838E-17 Cu2' (82.1 %), CuAcetate (1 5.8%) Fe2+ 1.1 E-03 1.176E-06 Fe2'(97.2%), FeAcetate (2.8%) Zn2+ 1.2E-04 4.834E-06 Zn2'(95.8%), ZnAcetate (4.1%) Pb2+ 8.7E-06 1.480E-09 Pb2'(53.2%). PbAcetate (45.8%)

'obtained from Part 3 of Output File (Type I components), Appendix F, y, activity coefficient : 0.84625

The detection limit for the elements of concern were 1.8 x IO8 mollL (0.1 ppm) for Fe, 1.5 x IO-' moVL (0.01 ppm) for Cu and Zn and 1.5 x 1ObmoVL (0.3 ppm) for Pb, based on the ICP analysis of the leachate. Any value below 1.O x IO'? was assumed to be below detection limit and was not significant except for theoretical calculations. The {H+) in the equilibrated mass distribution (Part 5 of Appendix F) does not correspond to the equilibrium pH. The equilibrium pH was treated by MlNTEQA2 as the hydrogen activity contributed by the H' species in solution (Part 3 of Output File, Type 1) whereas the equilibrated total {H') was the sum of both dissolved and adsorbed species (Part 4 of Output File). Any difference between the equilibrated concentration and the dissolved concentration represents the H' concentration bound in adsorbed species (Part 4 of Output File) as acetate. Note that the activities were used instead of concentrations since MINTEQA2 uses activity (a=y c where a is activity, yactivity coefficient and c molar concentration) in its calculations. Relative solubilities of metal sulfides can be predicted based on the kPvalues if they produce the same total number of ions during dissolution as in the case of CuS, PbS and ZnS which produce two ions each [130]. Thus among CUS (K,, = IO-^'), PbS (Y, = ) and ZnS (Ks, = 1 ), ZnS would be more soluble, followed by PbS and CUS, the least soluble. At equilibrium, this prediction was validated by MINTEQA2 where CuZ+was the least dissolved compared to Zn2' and pbZ'. In Table 4-8 is shown that equilibrium behavior was obviously not realized in the experimental tests since the concentrations obtained from the experiments did not correspond to the model results. At equilibrium, the pH was predicted to decrease from the experimental value of 6.5 to 2.8 since there will be more free H' and acetate ions in solution due to the precipitation of the metals. The dissolved metal species were in their ionic fons and as acetate compounds. In Sections 4.3 and 4.4, it was found that only 35% of the original concentrations of Pb in the Mexican scale would have been available for leaching, which was not reflected in the model. This is the kinetic constraint of thermodynamic models and without any cornplementary experimentation can lead to misinterpretation. The modeling results indicated that protocol leach tests do not represent long-term leaching behavior of the Mexican geothermal scale under natural environment. The inherent shortcoming of this modeling effort is the lack of accurate information on the other important species in the complex geothermal samples which may be controlling the solubility of the major species.

4.11 Summary Geotherrnal residues are composed rnainly of silica (-70% by weight) with iron and aluminum oxides and trace amounts of S, Cu, Zn, and Pb. Particte size varies from submicron to 1 mm and above. They are made up of heterogeneous particles with each particle having a complex microstructure and rnineralogy especially in the case of the scale. The elements of environmental concern are mostly trapped inside a hard silicate matrix. Sludges are porous and made up of agglomerates while drilling mud is a complex mixture of secondary silicate minerals. These geothermal residues were studied for their long-term leaching behavior in acidic medium and in the presence of bacteria.

4.71.1 Chernical Leaching Behavior in Protocol Tests The process of leaching is affected by various parameters such as type of leachant, dissolved oxygen, acidity, particle size, type of material, concentration of rninerals and to a certain extent, temperature and agitation. In the case of mineral leaching from geothenal residues, the following successive steps are postulated : a) mass transport of dissolved reactants such as oxygen and acid from bulk solution to the mineral surface, b) transfer of species through the reactive zone. c) chemical reactions of the sulfide minerals, d) transport of products to the surface, e) mass transport of species into the bulk of solution. The concentration vs time profile indirectly indicated that diffusion through the reaction zone is rate lirniting (Section 4.8.1). In the case of the Mexican scale where significant amounts of Fe,Zn and Pb were released, particle size, pH of the leachant, and oxygen availability were found to be most important factors. Laboratory experiments demonstrated that leaching was enhanced when particle size was reduced (Figures 4.7 & 4.81, oxygen was available (Figure 4.9), and pH was low (Figure 4.7). On the other hand, leaching was found minimal or even insignificant when oxygen was sparged from the system by nitrogen gas (Figure 4.10). The leachant, in this case, acetic acid solution, has a number of roles such as a) a carrier of dissolved oxygen and other reactants, b) provision of acetate ions for solubilizing metals. c) attack of mineral phases Thus with more acidity such as in TCLP vs LEP, more metals were found dissolved in the leachate. lncreased agitation could have contributed to a minor extent but in the long run when diffusion was the rate limiting or dominant reaction, agitation did not appear to have a marked effect on rnetal dissolution (Figure 4.9).

4.17.2 Bioleaching of Geothennal Residues Geothermal residues have sulfur content in the forrn of rnetal sulfides that indicate a potential in the long-terni for acid mine drainage (AMD). AMD is experienced in coal and metal mines whereby pyrites and other sulfide minerals are oxidized releasing acidity and metals to the environment. This is usually promoted by the presence of iron and sulfur oxidizing bacteria such as Thiobacillus ferrooxidans as discussed in Section 2.5.2. For the geothennal residues. a new test called Uoff's acid mine drainage potential test (Appendix G), was developed based on an existing procedure because the bacteria had to be acclirnatized to this substrate and ferrous ion had to be added to the culture media since the residues are low in Fe and S relative to mine tailings. From this test, only the Mexican scale was found to have a slight AMD patential since Cu and Zn leached out but not the regulated element Pb. Although less than 4% of Pb was in the leachate at the beginning, its disappearance was probably due to precipitation as insoluble PbSO, (Section 4. 6.2). Since the mineral sulfides are inside a silicate matrix and are relatively inaccessible to the bacteria, it is likely that enhanced leaching occurs primarily via indirect attack with ferric ion as the oxidant. More leaching was observed in the presence of bacteria (Figure 4.30). The results indicate that the geothermal residues will probably not pose a direct threat to the environment as they al1 passed the acid leaching protocol tests and they do not contain genotoxicants or radioactive materials beyond norm. In terms of acid mine drainage potential, the geotherrnal residues will presumably matea nuisance or aesthetic pollution due to the formation of iron precipitates which are visible as rusty orange brown solids. But since the geothermal residues have less iron content than mine tailings, there will likely be less iron precipitation cornpared to coal or copper mines. From the results of this work, it is likely that the geothermal residues can be safely disposed in a landfill environment.

4.12 Risk Assessment of Geothermal Residues This section will provide an overall risk assessment of landfill disposal of geothermal residues. Discussions on environmental risk assessment have been covered thoroughly by several authors [4, 132-1421. It was generally recognized that public perceptions of risks relating to waste disposal activities often reflect general societal anxieties and fears that were not necessarily supported by the results of a technical risk assessment. The technique used in this study is a qualitative rating (low, medium, high as against 1 in 106 risk) but based on quantitative figures from analytical and experirnental results. This study is a step further than existing desk top methodologies in that actual laboratory simulations were carried out. There are risks that are very low and acceptable (de minimus in legal terms - 'the law does not deal with trivialities') and are recognized to be true in certain cases [135] as in this study. A distinction can be made between hazard and risk. A hazard has the potential to cause harm while risk is the likelihood of a hazard causing harm. As there are great uncertainties, only a simple ranking or rating procedure can be used [133, 1351. In the case of geothenal residues, the hazard rating was based on its chemical content, radioactivity, and toxicity. On the other hand, the risk rating was derived from the leachability of the toxic constituents and the probability of producing acid mine drainage. The route of entry for these toxic elements to cause public health risk is ingestion via drinking water and food. The designations used in rating can be intenpreted from the text box below.

- - HAZARD AND RlSK RATING Very low : has low hazard, no significant risk, de minimus Low : has high hazard; within regulatory and normal levels Medium : has moderate to high hazard; regulatory and normal levels are partially met High : has high hazard; regulatory and normal ievels are exceeded, with significant risk.

A more realistic risk assessment must include the following analyses: a) site evaluation, b) release analysis, c) intermediate transfers, d) fate and transport analysis, e) exposure assessment, and finally, f) risk calculation 11391. Site evaluation alone will require about 50 input data about climate, hydrology, landfill design, and geology which were not offen easy to obtain if they were available. In this study, the scoping process had identified groundwater contamination as the most important threat to the environment through leachate production. It was also assumed from a practical point of view that the release mechanisms, exposure pathways and intake or dose rates were similar to al1 the geothenal wastes studied and that the landfill is secured with impervious liner. Thus what will Vary were the intrinsic characteristics of the wastes along with their mobility from one media to another. This simple but satisfactory approach can also be used in determining whether a certain waste is suitable for landfill disposal or not. A high risk perception was attributed to MSC based on its chemical content as shown in Table 4-9. However, as the level of screening becornes more sophisticated and detailed, the actual risk decreases to low and medium. A medium rating was assigned to MSC due to its low level radioactivity from Pb-210 and complete release of Cu and Zn during the AMDP test. However, it can be recalled from Sections 4.2 and 4.3, that only the powder sized paracles of -125 pm were leachable and that with 4rnm and above, the toxic etements were not mobile. Also. the regulated element Pb which could be released rapidly will form an insoluble precipitate at low pH and will not be available in solution. On the other hand, MDM and MSL, being basically silica, were practically harmless at a de minimus level thus the 'very low' rating.

Table 4-9 Hazard and risk rating for Mexican geothermal residues MDM MSC 1 MSL Parameter Very Low Low Medium High VeryLow Chernical content Radioactivity Toxicity Leachability AMD potential

PSC and PSL have been found to have similar rating as shown in Table 4-10. Their chemical content can provide significant hazard. However, closer examination will reveal that they will only pose low risk as leaching was not expected to occur even with bacterial mediation. A more cautious approach was to carry out treatability studies for PSC and PSL to reduce the hazard content prior to landfill disposal. The risk rating for ASC was practically the same for MSL and MDM as being very low and insignificant. Thus MSC appears to have greater risk cornpared to the other residues and would require special handling. Table 4-10 Hazard and risk rating for American and Philippine geothermal residues

- - PSC PSL ASC Parameter Low High Low High Very Low Chernical content J J Radioactivity Toxicity Leachability AMD potential

If more information was known about the wastes to be managed, then a reasonable risk assessrnent can be made based on sound predictions. There had been debates on the relative merit of static, kinetic and themodynamic approaches as well as errors made in contaminant prediction. However, the growing experience of successful predictions suggest that it can now be approached with confidence. Thus the prediction of risk may closely approximate actual risk. CHAPTER 5 CONCLUSIONS

Based on extensive characterization of the geothermal residues, experimental and analytical results, and geochemical modelling, the following conclusions can be drawn from the study: All the geothermal residues tested were mainly silica (66-82%) with trace elements SICu, Zn, and Pb at above earth's average crustal levels. They have varying crystalline and amorphous character. Scale samples were not chemically homogeneous but showed layered structures from the deposition process while the sludge samples were agglomerations of submicron particles forming a porous structure. The drilling mud was a complex mixture of secondary silicate rninerals. The radioactive levels of al1 samples were within acceptable levels for naturally occurring radioactive materials except for the Mexican scale which had an elevated Pb-210 content. However, the radioactivity was still lower than the occupational dose limit. Toxicity tests did not indicate the presence of toxins or genotoxins in any of the geothermal samples. The regulated elements (As, Ba, Cd, Cr, Hg, and Pb) were not found above regulatory limits in the leachate after a three-month weathering test (175 rpm, 35 OC). Results of the regulation leaching procedures do not classify the geothermal residues as hazardous since their leachate quality was below the regulatory limits. They can be safely disposed in a landfill since the heavy metal content in their leachate was sufficiently low to warrant an acceptable risk over a long period of time. However, particle size was important in the Mexican scale as leaching is increased considerably when the samples were ground from the protocol size of 9.5 mm to tess than 125 Pm. The powder sized particles, though present in small amounts in the raw sample, should be isolated and treated prior to disposal. Sequential chernical extraction indicated that As, Cd, and Pb from the Mexican scale and Philippine scale and sludge could be available to the environment but only under extreme conditions (pHs 2, 85-175 OC). Geochemical modeling and extended leaching experiments showed that leaching of Cu, Zn, and Pb were due to oxidative dissolution and that under anoxic conditions, only a very small amount was released. In the TCLP test, the rate mechanism for the chernical leaching of Pb in the Mexican scale follows a parabolic rate law for diffusion-controlled dissolution which is preceded by an initial, rapid leaching that is surface-controlled. The sulfur and iron oxidizing bacteria, Thiobacillus femoxidans can be cultured and acclimatized without agitation in nonsterile laboratory conditions. Of the geothermal samples, only the Mexican scale was found to have acid mine drainage potential and leaching of Cu and Zn was observed. The galena likely was attacked but Pb was not found in the leachate at the end of the test and could have precipitated as insoluble PbSO,. This may also indicate that Cu and Zn can be reclaimed through microbial leaching as in mineral recovery of metals. Results of the acid mine drainage potential (AMDP) test developed in this work indicated that none of the geothermal residues tested should have significant environmental impact in a landfill even with biological mediation. Although certain hazards exist in some samples, the associated risk due to leaching of the toxic cornponents to contaminate groundwater has been assessed to be low to insignificant based on a qualitative hazard and risk rating for evaluating landfill dis~osaiof aeothermal residues. CHAPTER 6

The following research areas are suggested for future worù: Column leaching on geothermal residues involving three types of leachant: a) with and without bacteria using distilled water b) without bacteria using weak acid c) with bacteria using AMDP culture medium. Rotary agitated protocol test at various pH (pH 3-10). lncreased testing time (longer than one week) for extended TCLP for both oxic and anoxic conditions. Use of acid mine drainage potential test with a mixed culture of Thiobacillus ferrooxidans and Thiobacillus thiooxidans. Verification of the AMDP prediction through comparison with occurrence of acid mine drainage in geothermal stockpile sites. Expanded geochemical modelling to consider a) simulation of leaching procedure with alt important species b) microbial reactions in acid mine drainage. More sampling and analysis to reflect temporal and spatial variation in characteristics of geothermal residues. Treatrnent of the Philippine scale and sludge and Mexican scale such as solidification, thermal treatment or microbial minera1 recovery. Waste disposal and utilization options for Mexican sludge and drilling mud and American scale. REFERENCES

Henley, R.W. and A.J. Ellis, Geothennal systems ancient and modem : A geochemical review. Earth Science Reviews, 1983. 19:1-50. Dickson, M.H. and M. Fanelli, Geothemal Energy. 1995, Chichester: John Wiley 8 Sons. 214. Muffler, L.J.P. and D.E. White, Geothemal Energy, in Perspectives on Energy : Issues, Ideas, and Environmental Dilemmas, L.C. Ruedisili and M. W. Firebaughl Editor. 1978, Oxford University Press: New York. 483-489. Brown, K. L., Environmental Aspects of Geothemal Development. Vol. 4. 1995, International Geothermal Association: Auckland, NZ. 145. El-Hinnawi, E.E., The Environmental Impacts of Production and Use of Energy. Vol.1 Natural Resources and Environment Series. 1981 , London: Tycooly Press Ltd. 31 9. Davis, A., J. W. Drexler, M. V. Ruby, and A. Nicholson, Micromineralogy of mine wastes in relation to lead bioavailabilify, Bufte, Montana. Environmental Science & Technology, A 993. 27:141 5-1 425. Darby , D. C. En vironmental Management in Philippine Geothennal Projects. in International Confemnce on Geothennal Energy. 1982. Florence, Italy: BHRA Fluid Engineering. Bowen, R., Environmental Impact of Geothemal Development, in Geothemal Energy Resources, Production, Stimulation, P. Kruger and C. Otte, Editor. 1973, Stanford University Press: 197-215. Axtrnann, R.C., Environmental impact of a geothennalpowerplant. Science, 1975. l87(4179): 795-803. Hahn , J .L., Occupational hazards associaîed with geothemal energy. Geotherrnal Resources Council TRANSACTIONS, 1979. J(Septernber 1979): 283-286. Premuzic, E.T., M.S. Lin, and J.Z. Jin, Biochemical processing of geothemal bnnes and sludges: adap tability to multiple industriai applications. Geothermal Resou rces Council TRANSACTIONS, 1994.18(0ctober 1994): 427-1 31. Suess, R. E. and C. L. Wardlow, Geothermal waste issues in RCm reauthorization. Geothennal Resources Council TRANSACTIONS, 1993.17(0ctober 1993): 75-79. Harper, R.T., L.A. Thain, and J.H. Johnston, Towards the eficient utilization of geothemal tesources in New Zealand. Geothermics, 1 992. 21: 641-651. Quong, R. Scale and Solids Deposition at the Geothermal Loop Experimental Facility at Niland, California., Geothermal Resources Council TRANSACTIONS 1977.1: 249-250. Wong, M.M. and A.L. Shugatman, Process for Reducing the Concentration of Heavy Metals in Geothemal Bnne Sludge. 1987, U. S. Patent No. 4,7 10,367. Hickman, H.D. Geothennal Power Plant Waste Disposal at the Geysers :Past and Present Problems with fheir Solutions. in Geothemal Energy Symposium, 11th Annual Energy Source Technology Conference. 1988. New Orleans, Louisiana: 373-378. Kristmannsdottir, H. , Types of scaling accumng by geothemal utilization in lceland. Geothermics, 1989. l8(l/2): 183-190. Webster, R.P. and L.E. Kukacka. Stabilization of geothemal residues by encapsulation in polymer concrete and Portland cernent mortar composites. in Technical Symposium on Polymer Concrete, Annual Convention of the Amencan Concrete Institute. 1988. Orlando, Florida. Karabelas, A. J., N. Andritsos, A. Mouza, M. Mitrakas, F. Vrouzi, and K. Christanis, Characteristics of Scales from the Milos Geothemal Plant. Geothermics, 1989. 18: 169-174. Gallup, D. L. and W.M. Reiff, Charactenzation of geothermal scale deposits by Fe-57 Mossbauer spectroscopy and complementary X-ray diffraction and infrared studies. Geothermics, 1991. 20(4): 207-224. Premuzic. E.T., M. Lin, and L. Kukacka, Biological Solutions to Waste Management. 1988, US Department of Energy, Geothermal Program Review: BNL-41118. Premuzic, E.T., M.S. Lin, and K.K. Sun. Progress in Geothennal Waste Treatment Biotechnology. in Geothermal Resources Council TRANSACTIONS. 199 1. 15 : 149-154. Sparks, Nevada. Premuzic, E.T., M.S. Lin, and J.Z. Jin, Developments in geothemal waste treatment biotechnology. Geothermics, 1992. 21 (516):891 -899. Premuzic, E.T., M.S. Lin, and J.Z. Jin. Recent Developments in Geothemal Waste Treatment Biotechnology. in International Conference Heavy Metals in the Environment. 1993a. Toronto: CEP Consultants. Premuzic, E.T., M.S. Lin, and J.Z. Jin. Geothemal Waste Treatment Biotechnology. in Geothermal Progmm Review XI. 1993b. Berkeley, Calif.: Department of Energy. Dutrizac, J. E. and R. J. C. MacDonald, Femk ion as a leaching medium. Minerais Science Engineering, 1974. 6(2): 59-100. Berner, R. A., A new geochemical classification of sedimentas, environments. Journal of Sedimentary Petroiogy, 1981. 51(2): 359-365. Peralta, G. L., J. W. Graydon, and D.W. Kirk, Physicochemical charactefistics and leachability of scale and sludge from Bulalo Geothennal System, Philippines. Geothermics, 1996. 25(1): 17-35. White, D. E., Characteristics of Geothemal Resources, in Geothemal Energy Resources, Production, Stimulation, P. Kruger and C. Otte, Editors. 1973, Stanford University Press: California. 69-94. Hill, J.H., J.C.H. Otto, and C.J. Morris, Solids Control for High Salinity Geothemal Brines. Geothermal Resources Council TRANSACTIONS, 1977.1: 139-140. Gupta, H.K., Geothemal Resources :An Energy Alternative. 1980, Amsterdam: Elsevier Scientific. 227. Sussman, D., Persona1 communication. 1994, Philippine Geothermal Inc. Dobryn. DG., A. L. Brisson, C. M. Lee, and S. M. Roll, Bio-leaching of Toxic Metals from Geothermal Waste : A Prelirninary Engineedng Analysis. 1 986, Brookhaven National Laboratory: Massachusetts. 72. Mason, B., Pnnciples of Geochemistry. 3rd ed. 1982, New York: John Wiley & Sons, Inc. 329. Ontario, Government of, Regulation 347 (Revision of Regulation 309), Environmental Protection Act, General - Waste Management. 1990, Min istry of Environment: Toronto, Ontario. USEPA, Toxicity characteristic leaching procedure, Federal Register 40 CFR Parts 268, 271, and 302. 1990, United States Environmental Protection Agency: Washington D. C. Tyagi, R.D. and D. Couillard, Bactenal Leaching of Metals from Sludge, in Encyclopedia of En vironmental Control Technology, P.N . C heremisinoff, Ed itor. 1989, Gulf Publishing Company: Houston. 684. Gallup, D.L. and J.L. Featherstone, Control of NORM deposition from Salton Sea geothemal brines. Geothermal Science & Technology, 1995. 4(4): 215-226. AECB, Radioisotope Release concentrations - Proposed Policy Statement. 1995, Atornic Energy Control Board: Canada. Skoog, D.A. and J. J. Leary, Principles of lnstnrmental Analysis. 4th ed. 7992, Orlando, Florida: Harcourt Brace College Publishers. 812. Jones, M.P., Applied Mineralogy. 1987, London: Graham & Trotman. 259. Whiston, C., X-ray Methods. 1987, London: John Wiley 8 Sons. 80-1 13. Xu, H. and B.J. Dutka. A new rapid sensitive bacteda1 toxicity screening based on the detemination of A TP. Toxicity Assessrnent, 1987.2: 149-166. Quillardet, P. and M. Hofnung, The SOS Chromotest, a colonmetnc bactenal assay for genotoxins: procedures. Mutation Research, 1985. 147: 65-78. Quillardet, P. and M. Hofnung, The SOS Chromotest :a review. Mutation Research, 1993. 297: 235-279. B irkeland, P.W., Pedology, Weathenng, and Geomorphological Research. 1974, New York, USA: Oxford University Press. 285. Bohn, H.L., B.L. McNeal, and GA. O'Connor, Soi1 Chemistry. 2nd ed. 1985, New York, USA: John Wiley & Sons. Rossi, G., Biohydrometallurgy. 1990, Hamburg: McGraw-Hill GrnbH. 609. WC(Wastewater Treatment Centre), Compendium of Leaching Tests. 1 990, Environment Canada. Cote, P.L., T.R. Bridle, and A. Benedek (Editors), An Approach for Evaluating Long-tem Leachability from Measurement of lntrinsic Waste Properties. Hazardous and Industrial Solid Waste Testing : Sixth Volume, Editors. D. Lorenzen, et al. 1986a, American Society for Testing of Materials: Philadelphia. 63-78. Cote, P., Contaminant Leaching from Cernent-Based Waste Foms under Acidic Conditions. 1986b, McMaster University: PhD thesis. Jackson, D.R., B.C. Garrett, and T.A. Bishop. Companson of batch and column methods for assessing leachability of hazanlous waste. Environ. Sci. Technol., 1984. 18(9): 668-673. Jackson. M.L., Soi1 Chemical Analysis. 1958. Englewood Cliffs, N. J.: Prentice-Hall, Inc. 498. Gupta, S. K. and K.Y. C hen, Partitioning of trace metals in selective chemical fractions of nearshore sediments. Environmental Letters, 1975. 1O(2 ):1 29-1 58. Tessier, A., P.G.C. Campbell, and M. Bisson, Sequential Extraction Procedure for the Speciation of Particulate Trace Metals. Analytical Chemistry, 1979. 51(7): 844-850. Fraser, J.L. and K.R. Lum, Availabiiity of elements of environmental importance in incinerated sludge ash. Environmental Science and Technology, 1983. 17(1): 52-54. Wadge, A. and M. Hutton, The Leachability and chemical speciation of selected trace elements in fly ash coal combustion and refuse incineration. Environmental Pollution, 1987. 48: 85-99. Hirner, A.V., Trace element speciation in soils and sediments using sequential chemical extraction methods. 1 nternational Journal of Environmental Analytical Chemistry, 1992. 46: 77-85. Salornons, W., Adoption of common schemes for single and sequential extractions of trace metals in soils and sediments. International Journal of Environmental Analytical Chemistry, 1993. 51 : 3-4. Forstner, U ., Meta1 speciation - general concepts and applications. ln ternation al Journal of Environmental Analytical Chernistry, 1993. 51 : 5-23. Li, X., B. J. Coles, M. H. Ramsey, and 1. Thornton, Sequential extraction of soils for multielement analysis by ICP-AES. Chemical Geology, 1995. 124: 109-123. Jordao, C.P. and G. Nickless, Chemical associations of Zn, Cd, Pb, and Cu in soils and sediments determined by the sequential extraction technique. Environmental Technology Letters, 1989. IO: 743-752. Orsini, L. and A. Bermond, Application of a sequential extraction procedure to calcarnous soi1 samples :pre/iminary studies. International Journal of Environmental Analytical Chemistry, 1993. 51: 97-108. David, D.J. and R.V. Nicholson. Field Measurements For Detemining Rates of Sulphide Oxidation. in Sudbury '95 - Mining and the Envimnment. 1995. Sudbury, Ontario, Canada: CANMET, Ottawa. Atlas, R. M. and L.C. Park, Handbook of Micmbiological Media. 1993, Boca Raton: CRC Press. 1079. Ferguson, K.D. and P.M. Erickson, Pre-mine prediction of acid mine drainage, in Environmental Management of Solid Waste :Dredged Material and Mine Tailings, W. Salomons and U. Forstner, Editors. 1988, Springer-Verlag: 24-43. CANMET, Acid Rock Drainage Pmdiction Manual. ii9l. Canada Centre for Mineral and Energy Technology under the Mine Environment Neutral Drainage Program (MEND): Ottawa, Canada. USEPA, Acid Mine Drainage Prediction. 1994, US Environmental Protection Agency: Washington DC. MacDonald, R. J. C., P.D. Kondos, S. Crevier, P. Rubinsky, and M. Wasselauf. Generation of, and disposal options for Canadian mineral treatment industry etfluent sludges in ~iilin~sand Enluent Management. 1989. Halifax, Nova Scotia, Canada. Stumm, W. and J.J. Morgan, Aquatic Chemisty. 2nd ed. 1981, New York: Wiley-lnterscience. 780. Singer, P.C. and W. Stumm, Acidic mine drainage: the rate-determinhg step. Science, 1970. 167: 1121 -1 123. Brock, T.D., Biology of Microorganisms. 1970, Prentice-Hall, Inc. 737. Jensen, A.B. and C. Webb, Ferrous suiphate oxidation using Thiobacillus ferrooxidans: a review, Process Biochemistry, 1995. 30(3): 225-236. Sand, W., Fe& iron reduction by Thiobacilius femoxidans at extremeiy low pH values. Biogeochemistry, 1989.7: 195-201. Lundgren, D.G., J. R. Vestal, and F. R. Tabita, The iron-oxidizing bacteha, in Microbial lron Metabolism, J.b. Neilands, Editor. 1974, Acadernic Press: New York. 457-473. McGoran, C.J.M., D.W. Duncan, and C.C. Walden, Growth of Thiobacillus ferooxidans on various substrates. Canadian Journal of Microbiology, 1969. 15: 135-138. Silverman, M. P. and D.G. Lundgren, Studies on the chernoautotrophic iron bactew Ferrobacillus Femoxidans. Journal of Bacteriology, 1959. 77: 642-647. Pronk, T.J., J. C. de Bruyn, P. Bos, and J. G. Kuenen, Anaerobic growth of Thiobacillus fenooxidans. Applied and Environmental Microbiology, 1992. 58(7): 2227-2230. Sugio, T., C. Domatsu, O. Munakata, T. Tano, and K. Irnai, Role of femc ion-reducing system in sulfur oxidation of Thiobacillus ferooxidans. A pp lied a nd Environmental Microbiology, 1985. 49(6): 1401-1406. Bosecker, K., Biodegradation of sulfur minerais and its applications for metal recovery, in Sulfur, its Signifieance for the Geo-, Bio, and Cosmosphere and Technology, A. Muller and B. Krebs, Editors. 1984, Elsevier Science Publishers: Amsterdam. 331-348. Tuovinen, O. H., Biological fundamentals of mineral leaching processes, in Microbial Mineral Recovery, H.L. Ehrlich and C.L.Brierley, Editors. 1990, McGraw-Hill: New York. 55-77. Norris, P. R., Acidophilic bacteda and their activity h mineral sulfide oxidation, in Microbial Mineral Recovery, H.L. Ehrlich and C. L. Brierley, Editors. 1990, McGraw-Hill, Inc.: 3-27. McCready, R.G.L. A Review of the Physical, Chernical and Biological Measures to Prevent Acid Mine Drainage: An Application to the Pyritie Halifax Shales. in Acid Mine Drainage Seminar. 1987. Nova Scotia: Environment Canada. Barron, J.L. and 0.R. Lueking, Growth and maintenance of Thiobacillus fenooxidans cells. Applied and Environmental Microbiology, 1990. 56(9): 2801-2806. Appelo, C.A.J. and D. Postma, Geochemistry, Groundwater and Pollution. 1993, Rotterdam: A. A. Balkerna Publishers. 536. Guay, R. and M. Silver, Uranium biohydrometallurgy. Process Biochemistry, 198 1. l6(l): 8-1 1. Bos, P. and J. Gijs Kuenen, Microbial treatment of coal, in Micmbial Mineral Recovery, H.L. Ehrlich and C. L. Brierley, Editors. 1990, McGraw-Hill Publishing Company: New York. 343-377. Evangelou, V.P., PNte Oxidation and Its Control. 1995, Boca Raton, Florida: CRC Press, Inc. 293. Liu, M.S., R.M.R. 8ranion, and D.W. Duncan. Oxygen transfer to Thiobacillus Cultures. in International Symposium on Biohydrometallurgy. 1987. University of Wawick, UK: Science and Technology Letters. Kwong, E.C.M., J. M. Scharer, J. J. Byerley, and R. V. Nicholson, Prediction and Control of Bactenal Activity in Acid Mine Drainage. in Sudbury '95 - Mininy and the Environment. 1995. Sudbury. Ontario, Canada: CANMET, Ottawa. Lawrence, R.W., G.W. Poling, G. M. Ritcey, and P. B. Marchant, Assessment of predictive methods for the determination of A MD potential in mine tailings and waste rock. in Tailings and €Muent Management. 1989. Nova Scotia: Pergammon Press. Peralta, G.L., Characterization of Scale and Sludge from a Philippine Geothetmal Power Plant. 1994, University of Toronto: MASc thesis. Ferguson, K.D. and P.M. Erickson. Will It Generate AMD? An Overview of Methods to Predict Acid Mine Drainage. in Acid Mine Drainage Seminar. 1987. Nova Scotia: Environment Canada. Calow, R.W., D. Hevenor, and D.M. Stogran. Cornpanson of the B.C. Reseanh and EPA Acid Mine Drainage Predictive Static Tests. in Sudbury '95 - Mining and the Environment. 1995. Sudbury, Ontario, Canada: CANMET, Ottawa. Ferguson, K.D., Static and Kinetic Methods to Predict Acid Mine Drainage, in Fundamental and Applied Biohydrometallurgy : Process Metallurgy 4, R.W. Lawrence, R.M.R. Branion, and H.G. Ebner, Editors. 1986, Elsevier: 486-488. Sobek. A.A., W. A. Schuller, J. R. Freeman, and R. M. Smith, Field and laboratos, methods applicable to overburdens and minesoils. 1978, US Environmental Protection Agency: USA. Bruynesteyn, A. and R.P. Hackl, Evaluation of acid production potential of mining waste matenals. Minerals and the Environment, 1984. 4: 5-8. Allison, J.D., D.S. Brown, and K.J. Novo-Gradac, MINTEQA2PRODEFA2, A Geochemical Assessrnent Mode1 for Environmental Systems Version 3.17. 1993, Environmental Research Laboratory,US Environmental Protection Agency, Athens, Georg ia. Melchior, D.C. and R.L. Bassett, Editors. Chemical ModeIIhg of Aqueous Systems 11. 1990, ACS Symposium Series 416: Washington DC. 450. Dzornbak, D.A. and F.M. Morel, Surface Complexation Modelling - Hydrous Femc Oxide. 1990, New York: Wiley-lnterscience. 390. Alpers, C.N. and D.W. Blowes, Editon. Environmental Geochemistry of Sulfide Oxidation. ACS Symposium Series 550, 1994, American Chemical Society: Washington D.C. 681. Serkiz, S.M., J. D. Allison, E. M. Perdue, H. E. Allen, and D. S. Brown, Comecting errors in the themdynamic database for the equilibrium speciation mode1 MINTEQA2. Water Research, 1996. 30(8): 1930-1933. Eighmy, T.T., J. Eusden, J. E. Krzanowski, D. S. Domingo, D. Stampfli, J. R. Martin. and P. M. Erickson, Comprehensive approach toward understanding elernent speciation and leaching behavior in municipal solid waste incineration electrostatic precipitator ash. Environ. Sci. Technol., 1995. 29(3): 629-646. Kirby, C.S. and J.D. Rimstidt, Interaction of municipal solid waste ash with water. Environ. Sci. Technol., 1994.28(3): 443-451. van der Sloot, H.A., Leaching behaviour of waste and stabilized waste materials; charactenzation for environmental assessrnent putposes. Waste Management & Research, 1990. 8: 21 5-228. Jambor, J. L. and D. W. Blowes, Editors. Environmental Geochemistv of Sulfide Mine-Wastes. Short Coune Handbook, Vol. 22. 1994, Mineralogical Association of Canada: Waterloo, Ontario. 438. Bain, J.G., D.W. Blowes, and W.D. Robertson. The hydrogeochemistry of a sand aquifer affected by discharge from the Nickel Rim tailings, Sudbury, Ontario. in Sudbury '95 - Mining and the Environment. 1995. Sudbury, Ontario. Canada: CANMET, Ottawa. 108. APTEC, PC multichannel analyrerhvindows: basic display and acquisition software, OSQ+PlusManual Version 5.30/Release 1. 199 1, USA. 109. Kwan, K. K., Direct sediment toxicify testing procedute using Sediment-Chromatest Kit. Environmental Toxicology and Water Quality, 1995. 9: 193-196. 1IO. Kwan, K.K. and B.J. Dutka, A novel bioassay approach: direct application of the Toxi-Chromotest and the SOS-Chromotest to sediments. Environmental Toxicology and Water Quality, 1992. 7: 49-60. APHA (American Public Health Association), Standard Methods for the Examination of Water and Wastewater. 18th ed. 1992, Washington, D.C.: APHA-AWWA-WEF. 1500. CANM ET, investigation of Prediction Techniques far Acid Mine Drainage. 1989, Canada Centre for Mineral and Energy Technology: Ottawa, Canada. ICRP, 7990 Recommendations of the International Commission on Radiological Protection. 1991, Publication 60: Annex B. Annals of the ICRP, 21(1-3). UN S C EAR, United Nations Scientific Cornmittee on the Effects of lonizing Radiation: Sources and Effects of lonizing Radiation. 1993, U N General Assem bly, New York, USA. Egemen, E. and C. Y urteri, Regulatory leaching tests for fly ash: a case study. Waste Management & Research, 1996, 14: 43-50. Holdren, J., G. R. and R.A. Berner, Mechanism of feldspar weathering - 1. Expenmental studies. Geochimica et Cosmochimica Acta, 1979. 43: 1 161-1 1il. Luce. R.W., R.W. Bartlett, and G.A. Parks, Dissolution kinetics of magnesium silicates. Geochirnica et Cosmochimica Acta, 1972. 36: 35-50. Helgeson , H. C ., Kinetîcs of mass transfer among silicates and aqueoos solutions. Geochimica et Cosmochimica Acta, 1971. 35: 421-469. Rojas-Chapana, J.A., M. Giersig, and H. Tributsch, The path of sulfur dudng the bio-oxidation of pynte by Thiobacillus femoxidans. Fuel. 1996. 75(8): 923-930. Davis, B.S., D. Fortin, and T.J. Beveridge. Acidophilic bactena, acid mine drainage and Kidd Creek mine tailings. in Sudbury '95 - Mining and the Environment. 1995. Sudbury, Ontario, Canada: CANMET, Ottawa. Ehrlich, HL. and C.L. Brierley, Editon. Microbial Mineral Recovery. 1WOa, McGraw-Hill, Inc.: New York. 454. Pesic, B. and 1. Kim. Electrochemistry of 1. fenooxidans interactions with pyrite. in Minera1 Bioprocessing. 1991. Santa Barbara, California: Minerais, Metals & Materials Society, TMS. Brierley, C .L., Metal immobilization using bactena, in Microbial Minera1 Recovery, H.L. Ehrlich and C.L. Brierley, Editors. 1990, McGraw-Hill: New York. 303-323. Ferris, F.G ., Immobilization and mineralization of metallic ions by bacteria. Energy Sources, 1990. 12: 371-375. Beveridge, T. J., Role of Cellular Design in Bactenal Accumulation and Mineralization. Annual Review of Microbiology, 1989.43: 147-1il. Seyfried, P. L., Persona1 communication. 1996, University of Toronto. Stumm, W. and R. Wollast, Kinetics of the surface-controlled dissolution of oxide minerais. Reviews of Geophysics, 1990. 28(1): 53-69. Stumm, W., Chemistry of the Solid- Water Interface. 1992, New York: John Wiley & Sons, Inc. 428. Wollast, R., Kinetics of the alteration of K-feldspar in buffered solutions at low temperature. Geochimica et Cosmochimica Acta, 1967. 31: 635-648. Zumdahl, S.S., Chernical Phciples. 1992, Lexington, MA: D.C. Heath and Company. 1090. Cotton, F.A. and G. Wilkinson. Advanced lnorganic Chemistry. 5th ed. 1988. New York: John Wiley & Sons, Inc. 1455. Canter, L., Environmental lmpact Assessment. 2nd ed. 1996, New York: McGraw- Hill. 851. Birley, M. H., The Health lmpact Assessment of Development Projects. 1995a, London: HMSO. 241. Birley, M. H. and G.L. Peralta, Health lmpact Assessmen t of Development Projects, in Environmental and Social lmpact Assessment, F. Vanclay and D.A. Bronstein, Editors. 1995b, ~ohn~iley& Sons: Chichester, England. 153-170. Carpenter, R., Risk Assessment, in Envimnmental and Social Impact Assessment, F. Vanclay and DA. Bronstein, Editors. 1995, John Wiley & Sons Inc.: Chichester, England. 193-219. Ortolano, L. and A. Shepherd, Envimnmental lmpact Assessment, in Environmental and Social Impact Assessment, F. Vanclay and D.A. Btonstein, Editors. 1995, John Wiley & Sons Inc.: Chichester, England. 3-30. Petts, J. and S. Eduljee, Environmental Impact Assessment for Waste Treatment and Disposal Facilities. 1994, Chichester: John Wiley & Sons Ltd. 485. Spa1 ing , H. , Cumulative effects assessment: concepts and principles. lmpact Assessment, 1994.12(3): 231-251. Eduljee, G.H. Assessing the nsks of landfil activities. in 1992 Harwell Waste Management Symposium. 1992. Hatwell, Oxon, UK: Environmental Safety Centre, AEA. Ricci, P.F. and M. D. Rowe, Editors. Health and Environmental Risk Assessment. 1985, Pergammon Press: New York. 299. Whitney, J.B.R. and V.W. Maclaren (Editors), Environmental Impact Assessment: The Canadian Expenence. Vol. EM-5. 1 985, lnstitute of Environmental Studies, University of Toronto. 197. Whyte, A.V. and 1. Burton, Environmental Risk Assessment. 1980, New York: John Wiley & Sons. 220. Lindsay, W.L., Chernical Equilibria in Soils. 1979, New York: John Wiley & Sons. 449. Silver, M. and A. E. Tonna, Oxidation of metal sulfides by Thiobacillus ferooxidans grown on different substrates. Canadian Journal of Microbiology, 1974. 20: 141-1 47. Crank, J., The Mathematics of Ditfusion. 2nd ed. 1975, Oxford: Clarendon Press. 414. Braun, R. L., A.E. Lewis. and M. E. Wadsworth, In-place leaching of primary sulfide ores: laboratory leaching data and kinetics model. Metallu rgica l Transactions, 1974. 5: 1717-1 726. APPENDIX A -Trial Experiments Prior to Procedure Development To be able to select the appropriate AMD potential procedure for the geothermal residues from those listed in Section 2.6, several trial experiments were performed within the conditions and facilities available in a chemical/environrnental engineering laboratory. Based on the laboratory results and literature, the acid mine drainage potential (AMDP) test suitable for geothermal residues was developed. The success of each trial was based on bacterial viability in terms of mobility and density at 800x magnification. Table A-1 Series of Experiments Performed Prior to Test Development

Variables Experiments Various Sources of American Type Culture Collection (ATCC 19859) Thiobacillus ferrooxidans U of T Laboratories: culture -Mine Drainage from Dept of Microbiology -01d acclimatized culture from Dept of Civil Eng Growth Medium Standard Method [APHA, 19921 American Type Culture Collection (ATCC, 1989) Handbook of Microbiological Media [Atlas, 19931 Nutrient Concentration No FeSO, 50% of required FeSO, 100% of required FeSO, Solid Concentration 10-15 911 00 ml 5 g/100 mL 2 gI100 mL Agitation lnside incubator-shaker (agitated) On laboratory bench (stationary) Temperature lnside incubator-shaker (35 OC) Room temperature (23-25 OC) Sterilization Culture media autoclaved Culture media filtered Samples oven-sterilized Samptes not sterilized BC Research Confirmation 150 m particle size Test 20 g sample in 70 mL culture medium 150 rpm and 35 OC Medium without FeSO, 6 weeks duration APPENDIX B - Bacterial Density Estimation for Thlobacillus fenooxidans In estimating bacterial density using image analysis, the image was viewed at 800- 1000x magnification in a light microscope as described in Section 3.10.1. The image was captured in a computer format (1 image = 300 Kb) and saved with a graphics extension such as .xls for Microsoft Excel or .tif for any graphics viewer such as Graphics Workshop and ACDSee. Samples of such image are shown in Figure 4.14. The computer software was unable to discriminate the bacteria with the other particles in the sample hence manual counting from the images was resorted to by drawing gridlines on the image and counting the cells per grid. The various steps in estimating the bacterial density are briefly described below. The volume of one drop of liquid sample placed onto glass slide and covered with cover glass was around 0.02 rnL or 20 mm3- This was measured by counting the number of drops per known volume, Le. 1 mL and dividing it with the number of drops (50). Volume of one drop of liquid sample, V = 20 mm3 Area of image = (640)(480) pixel or 30.72 mm2 where 1 pixel = 0.01 mm Area of cover glass = (22)(22) or 484 mm2 Number of bacterial cellslimage = x Number of images/slide = 484130.72 = 15.75 Total cellslslide = 15.75~ Bacterial density, cells1mm3= 15.75~= 15.75~= 0.788~ v 20 t, thickness between slide and cover = 20 mm3 = 0.040 mm = 40 pm 484 mm2 From Figures 15 and 16, the Thiobacillus ferooxidans bacteria has a diameter of 0.4 to 0.5 Fm and length of 1.O to 1.5 pm for rods and 0.5 to 0.7 for spheres whereas the bacteria filled-liquid film under the cover glass has a thickness of 40 vm. It was therefore possible that the image analysis can sense only part of actual bacterial density than what was shown in the photograph. In practice, it was difficult to determine the number of viable cells since mutated cells and non-dividing cells are always present. Furthermore, in the case of Thiobacillus fenooxidans, other difficulties arise from the adhesion of a considerable number of cells to the solid substrate as shown in Figure 14d. The heat and light from the microscope could also make the bacteria retreat to a lower depth of the sample thus decreasing the density as captured in the image. It was therefore necessary to adjust the density to refiect actual bacterial count. From the TEM photos, the average area of each bacteria was computed to be about 0.5 !m2. The volume of a Thiobacillus femoxidans is around 0.23462 pm3 [48] hence the thickness was about 0.47 Fm. The adjustment can be calculated as thickness of the slide divided by the thickness of the bacteria : 4010.47 = 85. To account for the uncertainties, a factor of 100 was suggested. Table B-1 was a sample estimation for various bacteria culture and acclirnatization stage.

Table 6-1 Typical Bacteria Counting and Calculation

-- Age of Filename Bacterial count, Bacterial Adjusted culture .xls/.tif cellslimage, X density, bacterial cells/rnL, density, çells/mL 788X = D 1000 6 d tf.xls 23 18000 1.8 x 10' bacteria.xls 40 32000 3.2 x 10' atcc2.xls 27 21O00 2.1 x 106 atccl .xls 13 d civl .xls civ2.xls 7 d asc2.xls mdm2.xls msc2.xls msl2.xls psc2.xls psl2.xls 7 d tfl .tif tf4.tif W.tif il d rb2.tif rb8.tif 22 d tf8. tif APPEMDlX C - Calculatioion for the Prelirninary Acid Mine Drainage Potential Results

Based on the data obtained from the BC Research Initial Test in Table 4-5, the acid * production potential (APP) and acid consumption (AC) were calculated as follows [67]:

Acid production potential = Percent sulfur x gB x IOOQ 32 IO0

APP= % S x 30.6 kg H,SO, per tanne

&id consumption = 98 x Vol acid. rpL x lN acid x ka11000 q 2 x sample weight, g x tonne 1000 g

AC = mL 1N H,SOI x 0.049 x 100Q kg H,SO, per tonne Sample weight in g

The acid production potential and acid consumption values are cornpared. If the APP exceeds the AC, the sample was classified as being a potential source of acid mine drainage. It was recommended to confirm the results using kinetic tests such as those listed in Table 2-6. APPENDIX D - About the Geochemical Model MlNTEQA2

The basic solution scheme used in MINTEQA2, a Geochemicai kssessment Model for Environmental Systems [Allison, 19931 is summarized as follows:

1. ldentify species of interest, choose a set of components, and set up a table. 2. Guess the concentration of each component. 3. Calculate the equilibrium composition of the system using the estimated component concentrations in the rnass law equations. 4. Calculate the error in the mole balance equation for each component. 5. Obtain improved estimates for the component concentrations using the multidimensional Newton-Raphson iteration technique on the mole balance errors. 6. Calculate a new equilibrium composition, the corresponding mole balance equation errors, and obtain irnproved estimates for the component concentrations. 7. Continue the iterative procedure until the errors in the mole balance equations are small. The mode1 has its own thermodynarnic database so the primary information that must be conveyed through the input file was the total dissolved concentration or fixed activity of each component of the system. Solids are identified to PRODEFA2 by specifying the component that represents the major cation and the main mineral group to which the solid belongs (e.g. carbonate, sulfide). Alternatively, one may specify the 7sligit ID number for any aqueous or solid species if it was known. Menus and prompts within PRODEFA2 allow al1 of these things to be done with relative ease. MINTEQA2 solves the equilibrium problem iteratively by computing mole balances from estimates of cornponent activities. PRODEFAP makes this guess automatically for every component as equal to the component total dissolved concentration but also provides the means for the user to change the guess. It was possible for the user to insist that certain conditions prevail at equilibrium for pH, pe, or gas partial pressure. There are four choices for units of concentration for the input data: 1) Molal (moVkg, same as molar for the dilute systems appropriate for MINTEQA2), 2) mgll, 3) ppm (parts per million), or 4) meqA (milliequivalents per liter). Regardless of the units chosen for input data, MINTEQA2 output data are always molal. APPENDIX E - Input Data Derivation for Geochemical Model The calculation of solid concentration for the important minerai species in the Mexican scale was undertaken with the following steps. Only the major species such as pyrite, chalcopyrite, galena, and sphalerite were considered in the modelling. Later on covellite was added since it showed as a controlling solid with a positive saturation index after the initial runs. To sirnplify calculations, amount for chalcopyrite was assumed to include covellite, and the C concentration was split into two. P = moles pyrite, FeS, C = moles chalcopyrite, CuFeS, G = moles galena, PbS S = moles sphalerite, ZnS v - moles covellite, CuS

Calculate the moles of the respective elements from Table 4-1:

moles S = 3.38 g1100g x 50 glL x 1 mole/32 g = 0.053 moles Cu = 9080 uglg x 1 g11O6 x 50 g/L x 1 mole163.5 g = 0.0071 moles Pb = 1 1,600 uglg x 1 911 O6 x 50 g/L x 1 mole1207.2 g -0.0028 moles Zn = 15,900 uglg x 1 g1106 x 50 g/L x 1 mole165.4 g = 0.0120

S balance : 2P + 2C + G + S = moles S - 0.053 Cu balance: C - moles Cu = 0.0071 Pb balance: G - moles Pb = 0.0028 Zn balance : S - moles Zn = 0.01 20

Solving the equations, the following values were obtained: 0.0120 - moles pyrite, FeS, O. 0035 - moles chalcopyrite, CuFeS, 0.0036 - moles covellite, CuS O. 0028 = moles galena, PbS 0.0120 - moles sphalerite, ZnS The moles of acetic acid added were calculated as follows: Leachate Extraction Procedure(LEP): 0.5N x 5 mL = 0.0025 mol/L Toxicity Characteristic Leaching Procedure (TCLP) :0.1 N x 1 L = 0.1 mollL Appendix F - Sample Output of Geochemical Modelling

PART 1 of OUTPUT FILE PCMINTEQA'v3.10 DATE OF CALCULATIONS: 20-NOV-96 TIME: 0:27:21

CALCULATE THE EQUILIBRIUM CONDITIONS BY ADDING 0.1N HAC TO MEXICAN SCÀLE IN TOXICITY CHARACTERISTIC LEACHING PROCEDURE ------Temperature (Celsius) : 25.00 Units of concentration: MOLAL Ionic strength to be computed. If specified, carbonate concentration represents total inorganic carbon. Do not automatically terminate if charge imbalance exceeds 30% Precipitation was allowed only for thrse solids specified as ALLOWED in the input file (if any) . The maximum number of iterations is: 200 The method used to compute activity coefficients is: Davies equation Intemediate output file ------330 1.000E-01 -1.00 H+ 992 1.000E-01 -1.00 Acetate 730 0.000E-01 -16.00 HS-1 1 0.000E-01 -16.00 E-1 280 0.000E-01 -16.00 Fe+2 600 0.000E-01 -16.00 Pb+2 950 0.000E-01 -16.00 Zn+2 231 0.000E-01 -16.00 Cu+2

H20 has been inserted as a COMPONENT 4 5 1028003 18.4790 -11.3000 1.200E-02 FeS, Pyrite 1023102 35.2700 -35.4800 3.5003-03 CuFeS, Chalcopyrite 1060001 15.1320 -19.4000 2.800E-03 PbS Galena 1095001 Il.6180 -8.2500 1.200E-02 ZnS Sphalerite 1023101 23.0380 -24.0100 3.6003-03 CuS Covellite

INPUT DATA BEFORE TYPE MODIFICATIONS

NAME ACTIVITY GUESS LOG GüESS ANAL TOTAL H+l 1.000E-01 -1.000 1.000E-O1 Acetate 1.000E-01 -1.000 1.000E-01 HS-1 1.000E-16 -16.000 0.000E-01 E-1 1.000E-16 -16.000 0.000E-01 Fe+2 1.000E-16 -16.000 0.000E-01 Pb+2 1.000E-16 -16.000 0.000E-01 Zn+2 1.000E-16 -16.000 0.000E-01 Cu+2 1.000E-16 -16.000 0.000E-01 H20 1.000E+00 o. O00 0.000E-01

Charge Balance : UNSPECIATED Sum of CATIONS= 1.000E-01 Sum of ANIONS = 1.000E-01 PERCENT DIFFERENCE = 0.000E-01 (ANIONS - CATIONS) / (ANIONS + CATIONS) PART 3 of OUTPUT FILE PC MINTEQAI! v3.10 DATE OF CALCULATIONS: 20-NOV-96 TIME: 0:27:22

PARAMETERS OF THE COMPONENT MOST OUT OF BALANCE:

ITER NAME TOTAL MOL O HS-1 0.000E-01 1 HS-1 0.000E-01 2 HS-1 0.000E-01 3 HS-1 0.000E-01 4 HS-1 0.000E-01 5 HS-1 0.000E-01 6 HS-1 o. 0003-01 7 HS-1 0.000E-01 8 HS-1 0.000E-01 9 HS-1 0.000E-01 10 HS-1 0.000E-01 11 HS-1 0.000E-01 12 HS-1 0.000E-01 13 HS-1 0.000E-01 14 HS-1 0.000E-01 15 HS-1 0.000E-01 16 HS-1 0.000E-01 17 HS-1 0.000E-01 18 HS-1 0.000E-01 19 HS-1 0.000E-01 20 HS-1 0.000E-01 21 HS-1 0.000E-01 22 HS-1 0.000E-01 23 HS-1 0.300E-Ci 25 HS-1 0.000E-01 26 HS-1 0.000E-01

NAME ANAL MOL CALC MOL LOG ACTVTY DIFF FXN H+l 1.000E-01 1.3583-03 -2.88524 2.9003-09 Acetate 1.OOOE-01 1.3723-03 -2.88063 3.1063-09 HS-1 0.000E-01 6.770E-10 -9.18752 4.386E-10 II20 0.000E-01 -l.256E-ll -0.00148 0.000E-01 E-1 0.000E-01 9.3033-01 -0.02764 0.000E-01 Fe+2 0.000E-01 1.389E-06 -5.92972 0.000E-01 Cu+2 0.000E-01 2.1723-17 -16.73572 0.000E-01 Zn+2 0.000E-01 5.7123-06 -5.31572 0.000E-01 Pb+2 0.000E-01 1.7493-09 -8.82 972 0.000E-01

Type 1 - COMPONENTS AS SPECIES IN SOLUTION

NAME CALC MOL ACTïVITY LOG ACTVTY GAbMA NEW LOGK H+l 1.3583-03 1.3023-03 -2.88524 0.95912 O. 018 Acetate 1.3723-O3 1.3163903 -2.08063 0.95912 O. O18 HS-1 6.770B-10 6.4943-10 -9.18752 0.95912 0.018 Cu+2 2.1723-17 1.8383-17 -16.73572 0.84625 O. 073 Fe+2 1.3893-06 1.1763-O6 -5.92972 0.84625 0.073 Pb+2 3.7493-09 1.48OE-09 -8.82972 0.84625 O. 073 Zn+2 5.7123-06 4.0343-06 -5.31572 0.84625 O. 073 ,------CI------Type II - OTHER SPECIES IN SOLUTION OR ADSORBED

NAME CALC MOL ACTIVITY NEW LOGK ZN ACETATE4 3.9293-16 3.3253-16 1.433 FeACETATE 4.053E-08 3.8879-08 1.418 OH- 8.0153-12 7.6873-12 -13.980 FeOH + 2.9663-13 2.8453-13 -9.482 FeOH3 -1 5.4913-29 5.2673-29 -30.982 ~80~2AQ 1.8523-21 1.8533-23 -20.570 Fe(HS)2 AQ 4.4173-16 4.4183-16 8.950 Fe(HS)3 - 3.2573-23 3.1243-23 11.005 Cu ACETATE 4.1863-18 4.015E-18 2.238 CuOH + 1.4663-22 1.406E-22 -7.982 Cu(0H)2 AQ 2.2473-25 2.2483-25 -13.680 Cu(0H) 3 - 1.0833-35 3.0393-35 -26.881 Cu(OH)4 -2 1.8703-45 1.5823-45 -39.527 Cu2 (OH)2+2 1.0233-38 8.6613-39 -10.286 Cu(HS)3 - 4.1583-19 3.9883-19 25.917 ZnOH + 4.2283-12 4.055E-12 -8.942 Zn(0H)2 AQ 3.5703-17 3.5713-37 -16.899 Zn(0H)3 - 9.0093-26 8.6413-2 6 -28.381 Zn(0H)4 -2 1.2383-35 1.048E-35 -41.126 Zn(HS) 2 A(;Z 1.7753-O9 1.775E-Og 14.940 Zn(HS)3 - 1.7373-37 1.6663-17 16.118 PbOH + 2.3023-14 2.2083-14 -7.692 Pb(0H)2 AQ 6.5713-21 6.5743-21 -17.120 Pb(0H) 3 - 6.0213-29 5.7753-29 -28.042 Pb20H +3 1.0653-21 7.3173-22 -6.197 Pb(HS)2 AQ 1.162E-12 1.1623-12 lS.270 Pb(HS)3 - 1.5703-20 1.5063-20 16.588 Pb3 (OH)4+2 1.7313-39 1-4653-39 -23.807 Pb(0H)4 -2 1.1993-37 1.0153-37 -39.626 H2S AQ 7.3903-06 7.3923-06 6.941 S -2 . 7.1163-20 6.0223-20 -12.845 H ACETATE 9.8633-02 9.866E-02 4.760 CuACETATE 2 1.3583-19 1.3583-19 3.630 CuACETATE 3 5.5023-23 5.2773-23 3.118 CuACE TATE 4 5.1793-26 4.3833-26 2.973 PBACETATE 1.5063-09 1.4443-09 2.888 PBACETATE 2 3.0823-11 3.0833-11 4.080 PBACETATE 3 1.3693-14 1.3133-14 3.608 PBACE TATE 4 1.3193-17 1.1163-17 3.473 ZN ACETATE 2.4653-07 2.3643-07 1.588 ZN ACETATE2 6.6513-10 6.6533-10 1.900 ZNACETATE3 4.2713-13 4.096E-13 1.588

Type III - SPECIES WITH FIXED ACTIVITY

ID NAME: CALC MOL LOG MOL NEW LOGK DH

2 H20 -1.256E-11 + -10.901 O. 001 O. 000

Type IV - FINITE SOLIDS (present at equilibrium)

ID NAME CALC MOL LOG MOL NEW LOGK DH 1028003 PYRITE 1.2OOE-O2 O. O00 18.479 -11.300 1023101 COVELLITE 3.6013-03 -5.845 23.038 -24.010 1023102 CHALCOPYRXTE 3.4993-03 -5.845 35.270 -55.480 1095001 SPWRITE 1.199E-02 -5.225 11.618 -8.250 1060001 GALENA 2.8003-03 -8.483 15.132 -19.400

Type VI - EXCLUDED SPECIES (not included in mole balance)

ID NAME CALC MOL LOG MOL NEW LOGK DH 3300021 02 (g) 0.000E-01 -71.471 -83.120 133.830 1 E-1 9.3833-01 -0.028 0.000 0.000

PART 4 of OUTPUT FILE PC MINTEQA2 v3.10 DATE OF CALCIJIATIONS: 20-NOV-96 TIME: 0:27:22

PERCENTAGE DISTRIBUTION OF COMPONENTS AMONG TYPE 1 and TYPE II (dissolved and adsorbed) species

PERCENT BOUND IN SPECIES # 330 H+1 PERCENT BOUND IN SPECIES #3309921 H ACETATE

Acetate PERCENT BOüND IN SPECIES # 992 Acetate PERCENT BOUND IN SPECIES W3309921 H ACETATE

PERCENT BOUND IN SPECIES W3307300

PERCENT BOUND IN SPECIES #3300020 OH- PERCENT BOUND IN SPECIES a2803300 FeOH+ PERCENT BOUND IN SPECIES U9503300 ZnOH+

PERCENT BOUND IN SPECIES # 280 PERCENT BOUND IN SPECIES #2809920

PERCENT BOUND IN SPECIES # 231 Cu+2 PERCENT BOUND IN SPECIES #2319921 Cu ACETATE PERCENT BOUND IN SPECIES #2317300 Cu(HS)3 -

PERCENTBOUNDINSPECIES# 950 Znt2 PERCENT BOUND IN SPECIES #9509921 ZN ACETATE

PERCENT BOUND IN SPECIES # 600 Pbt2 PERCENT BOUND IN SPECIES #6009921 PBACE TATE PART 5 of OUTPUT FILE PC MINTEQAZ v3.10 DATE OF CALCüIATIONS: 20-NOV-96 TIME: 0:27:23

IDX NAHE DISSOLVED SORBED PRECIPITATED MOL/KG PERCENT MOL/KG PERCENT MOL/KG PERCENT

H+1 Acetate HS-1 H20 E-1 Fe+2 Cu+2 Zn+2 Pb+2

Charge Balance : SPECIATED

Sum of CATIONS = 1.3723-03 Suni of ANIONS 1.3723-03 PERCENT DIFE'ERENCE = 2.3483-05 (ANIONS - CATIONS) /(ANIONS + CATIONS)

EQUILIBRIUM IONIC STRENGTH (m) = 1.3003-03

EQUILIBRIUM pH = 2.885

EQUILIBRIWM pe = 0.028 orEh = 1-64mv

DATE ID NUMSER: 961120 TIME ID NUMBER: 272304

PART 6 of OUTPUT FILE PC MINTEQA2 v3.10 DATE OF CALCULATIONS: 20-NOV-96 TLME:: 0:27:23

Saturation indices and stoichiometry of al1 minerals

ID # NAME Sat. Index Stoichiometry in [brackets] 1028000 FES PPT -8.317 [ -1.0001 330 [ 1.000] 280 1 1.0001 730

1028003 PYRITE

[ 2.0001 730 1023101 COVELLITE 0.000 [ -1 .OOO] 330 [ 1.0001 231 [ 1.000J 730 2023100 CU(0H)2 -19.608

2023101 TENORITE -18.507

1023102 CHALCOPYRITE 0.000

95000 ZN METAL 2095000 ZN(OH)2 (A)

2095001 ZN(OH)2 (C)

2095002 ZN(0H)2 (BI

2095003 ZN(0H)Z (G)

2095004 ZN(0H)2 (E)

2095005 ZNO(ACTIVE)

2095006 ZINCITE

1095000 ZNS (A)

1095001 SPHALERITE

1095002 WURTZITE

60000 PB METAL 2060000 MASSICOT

2060001 LITHARGE

2060002 PBO, .3H20

1060001 WNA

2060003 PLATTNERITE

3060001 MINIUM -77.048

2060004 PB(0H)2 (C) -11.212

2060005 PB20 (OH)2 -32.323

73100 SUL= -4.137

2028000 WUSTITE -11.533 APPENDIX G - UofT Acid Mine Drainage Potential Test

Introduction The proposed Uoff Acid Mine Drainage Potential test (AMDP) for geothermal residues is an irnprovernent of the BC Research Confirmation Test (BCRCT) which has been developed and widely used in Canada and the US for the last 15 years. It is a confirmation test to determine the acidification potential of a sample with biological mediation. The BC Research Confirmation Test had a number of shortcomings because of the nature of geothermal residues. The BCRCT was designed for mine tailings that are high in sulfur content (up to 40%) and have smaller particle size (-50 pm) making them more amenable to bacterial attack. Geothennal residues, on one hand, have less S content (-4%) and are bulkier with particle size from 1 mm and higher. Due to these reasons, the BCRCT was adjusted to the nature of geotherrnal residues. Barron and Lueking suggested methods on how to maintain Thiobacillus fhooxidans [84]. Bruynesteyn and Hackl [97] and CANMET 1671 were useful references on the evaluation of acid production potential of rnining waste materiais. The rnost important modifications are the use of a lower solid concentration (or pulp density), addition of FeSO, in culture media, regular measurement of redox potential (Eh), pH, metals as well as monitoring of bacterial growth and viability. The AMDP test can be performed in nonsterile conditions at room temperature even without continuous agitation. Table 4-7 provides a cornparison between the old and the new procedure. This procedure is a confirmation of acid mine drainage potential after a preliminary assessrnent has been made using the BC Research Initial Test described in Section 3.7. Sample collection, preparation and storage were undertaken using standard methods [IlIl.

GIA Short Version

1. Grind samples in a mortar and pestle to 120 mesh size (125 pm). 2. Have ready an acclimatized culture of Thiobacillus ferrooxidans. Refer to acclimatization of inoculum below (Section G1.3). 3. Prepare Thiobacillus femoxidans culture media using the formula in table beiow (Section G1.4). 4. If a shaker is available, apply agitation at 150-175 rpm. Otherwise place flasks on a laboratory bench or open shelf. Without agitation is acceptable but testing tirne will increase one week more. 5. If an incubator-shaker is available, se? tzmperature to 35 OC at 150-175 rpm. Otherwise place flasks on a laboratory bench or open shelf at room temperature of 23-25 OC. Monitor room temperature using a therrnometer. Lower temperature will increase testing time. 6. Use solid concentration or pulp density of 2 % (2 g sample in 100 mL media). 7. Prepare inoculum according to proposed new method (Section G?.3). Use 5 mL inoculum from logarithmic growth phase having a bacteria density of 2 x IO7 cellslmL for every 100 mL media (Section G1.7). 8. Monitoring schedule is at the begiiining and every 3 days for 3-4 weeks. 9. Monitor pH, Eh, bacterial density and rnotility (Sections G1.6 and G1.7), and metals (1 rnL aliquot diluted to 5 mL). Weigh flask and contents at the beginning and replace water due to evaporation.

1. Clean al1 glassware to be used in detergent, rinse three times, soak in 20% HNO, overnight, rinse with tap water three times and finally rinse with deionized water. Once dry, cover the 250 mL Erlenmeyer flasks with aluminum foi1 prior to use. 2. Pulverize the sample to pass a 120 mesh Tyler screen (approximate particle size -125 pm) and store in air tight bottles prior to use. 3. Prepare the bacteria culture media using the media specified in Section G1.4 below. The pH of the media must be 2.9. 4. In duplicate, weigh 2 g of ground sample into 250 mL Erlenmeyer flask. Label the flasks accordingly. Slowly add 100 mL of culture media and cover flask with a plug made of nonadsorbent cotton wrapped with gauze. Swirl manually and check pH. If the pH is above 2.9, add ION H2S04until stable at pH 2.9 I0.1. 5. Inoculate flasks with an active culture of Thiobacillus femoxidans prepared according to Section G1.3. Record weight of flask with its contents without the cotton plug. 6. Place the flask on a laboratory bench or open shelf at room temperature (at least 23-25 OC) with adequate ventilation. If an incubatorhhaker is available, place flask

on a shaker at 175 rpm and 35 OC. With agitation, testing time is shorter by one week. 7. Prior to each measurement every 3 days, weigh flask and contenis (without plug) and add deionized water to replace loss by evaporation. Obtain 1 mL aliquot, centrifuge for 10 min, and transfer supernatant to a clean 15 mL centrifuge tube. Dilute to 5 mlwith deionized water, acidify ta pH12 with -0.05 mL concentrated HNO,, and store at 4OC while waiting to be analyzed. Add ImL deionized water to ail the flasks to replace the 1 mL aliquot sample. 8. Monitor pH, Eh, bacterial growth (motility and density), color of solution, and dissolved metals every 3 days. Clean the pH and Eh probes at every measurements with water spray to avoid contamination from one flask to another. Manually shake flask at every determination. The color of solution will progressively change from light grey to yellow to deep amber or orange brown color which indicates iron oxidation. The solution will also change from a clear solution to slightly turbid which indicates bacterial growth and some precipitation. 9. Within the sampling period, monitor bacterial motility (Section G1.6) and density (Section G1.7) under a light microscope at 800x - 1000x magnification. When bacterial activity has ceased as observed from the microscope and a stable pH has been achieved, teninate the test. Analyze the aliquots for regulated metals. If the pH is below 3.5, and dissolved metals are present in the leachate above regulatory limits, the sample is classified as having acid mine drainage potential (or potential for bioleaching treatment). 10. This test can be completed within 3-4 weeks following inoculation. Gi.3 Acclimatization of lnoculum

Results of work on aeothermal residues A critical stage of the procedure is the acclimatization of the pure bacteria culture to the specific samples to be tested. A series of steps has been designed to assure that viable culture was ready as inoculum for the UOT- AMDP test. The trial experiments had shown that a half formula with 4.5 g/L Fe2+was providing the requirement for bacterial growth. It was not advisable to withhold completely the FeSO, from the leaching medium since the bacteria require from 2-3 g/L to 9 g/L Fe2+to survive [37,73, 77,78,84]. The BC Research Confirmation Test does not include FeSO, in its growth medium since its premise was that the bacteria will obtain FeSO, from the oxidation of pyrite from the mine tailings sample. However for geothermal residues, without the initial seed FeSO,, the bacteria cannot survive as some of the sulfides (including pyrite) are not readily accessible as they may be bound in silicate matrix. An excess FeSO, however should be avoided as it can increase the formation of a yellow precipitate called jarosite, KFe,(SO,),(OH), and iron oxyhydroxides, Fe(OH), and FeOOH. To begin the procedure. a full media containing 100% of the required FeS04.7H,0 was used to provide rapid bacterial growth. A full media on has 44.22 g/L FeSO4.7H,O while a half media has 22.1 1 g/L only. The bacteria growth medium is described below. Aftelwards, a second culture is grown with the addition of 2 g geothermal sarnple using the half media described below (Section G1.4) and the acclimatized bacteria to produce a culture that was ready as inoculum. A shorter step was required if an acclimatized culture already exists. Resultina inoculum Full media + pure culture BI Full media + pure culture (BI) + 2 g sarnple B2 Half media + acclimatized bacteria (82)+ 2 g sample 83 83 ready as inoculum for the AMDP test.

If a dormant B3 culture has to be revived, Full media + acclimatized culture (83) Half media + 84 + 2 g sample B5 is ready as inoculum for AMDP test. Gi.4 Bacteria Culture Medium Below is the proposed growth medium for the Thiobacillus femMxidans which was modified from the standard laboratory technique developed by the Amencan Public Health Association [Il11. The major differences are the reduction of the FeSO, content and use of membrane filtration for sterilization of solution instead of autoclaving.

Modified Growth Medium for Thiobacillus fen-ooxidans Basal salts: in a 1 L Erienmeyer flask Ammonium sulfate (NH4),S0, 3.0 g Potassium chloride KCI 0.10 g Dipotassiurn hydrogen phosphate K2HP0, 0.50 g Magnesium sulfate MgS04.7H,0 0.50 g Calcium nitrate Ca(NO,), 0.01 g Sulfuric acid, 10 N H,S04 1.0 mL Distilled water 700 mL

Energy source: in a 500 mL Erlenmeyer flask

Ferrous sulfate FeS04.7H,0 Distilled water

Separately filter using cellulose acetate (pore size 0.45 pm) the basal salts and energy source and combine after filtration. The medium will be opalescent and green and a precipitate will form (probably ferrous and ferric phosphates). The pH should be 2.9 with the solution containing 4500 mglL ferrous ion. The medium can be stored for at least 2 weeks in the refrigerator.

G1.5 Cultivation Add 5 mL of inoculum (American Type Culture Collection 19859) in a 250 mL Erlenmeyer flask containing 100 mL of fresh bacteria culture medium found in Section G1.4. Growth of the Thiobacillos femoxidans is manifested by a decrease in pH and an increase in the concentration of oxidized iron as orange-brown or deep amber color. Check under the microscope for bacterial motility and density with at least 800-1000x magnification. G1.6 Bacteria Moti!ity Bacteria motility is difficult to quantify but it can be described following the bacterial growth curve 148, 72, 1261. Thiobacillus femoxidans are very active and motile at the logarithmic and stationary phases. They can corne as single, pairs or short chahs. At the lag phase and death phase, they are dormant and nonmotile sometimes looking like white spots. The rnotility can be reported simply as motile (slow, medium, fast) or nonmotile.

G1.7 Bacteria Density Bacteria density can be estimated qualitatively by cornparhg what is observed in the light microscope with the photographs of bacteria in Figures 4. Ma to 4.14~which depict low, medium and high density (from IO6 to 10' cellslmL). Bacteria density can be calculated by following the steps in Appendix B. Below is a general bacteria cuwe for Thiobaci//usfemoxidans in both agitated (35 OC) and stationary (25 OC) experiments. This should serve as a guideline only. The y-axis was constructed with values from 1 to 5 corresponding to a range from low to high density with an F factor to cover the cell counts.

Bacteria Growth Cuwe for T.f.

6

O 5 10 15 20 25 30 35 f ime, days

Agitated, T=35 C + Stationary, T=2SC APPENDIX H - RESULTS OF TOXlClTY TESTING

For the Toxi-Chromotest, the blue color developrnent signifies that the E. Coli were alive and therefore the sample is non-toxic at the particular concentration. Conversely, if there was no color development, it rneans the bacteria were dead and the sample was toxic at the particular sample concentration. Non-toxic (NT) rneans the samples did not exhibit toxicity at every concentration. Five sample concentrations (%wIv) were used (50%. 25%. 12.5%, 6.25%. and 3.1 3%). In Table H1 below, a sample with a value of 3.13% was considered exhibiting toxicity since it required only a little amount to affect the bacteria, i.e., the lower the percent of substrate concentration (3.13% and below), the higher the degree of toxicity. A detailed explanation of this scheme can be found in several references [4345, IOM1 O]. The results in Table H1 showed toxicity only at higher concentrations : 50% for PSC and MDM as well as 25% for MSC while the rest (PSL. ASC, and MSL) were classified as non-toxic. These results can classify the samples as generally having negative toxicity.

Table H1 Toxi-Chromotest Results for Geothermal Samples

Sampfes Toxic at X % Samples Toxic at X % - - - - + Control 3.13 + Control 3.13 - Control Non-toxic (NT) - Control Non-toxic (NT) PSC 50 MDM NT 50 50 50 50 PSL NT MSC 50 NT 25 NT 25 ASC NT MSL NT

For the SOS-Chromotest, an indicator of genotoxicity is the presence of blue color in the chromopads (the opposite of Toxi-Chromotest above). All the samples did not produce blue color in the chromopads indicating they are negative for genotoxins. APPENDIX I - ADDITIONAL DISSOLUTION KINETICS DATA

Below are two graphs showing dissolution kinetics of Fe and Zn in the oxic TCLP test of the fine sized Mexican scale. Like Pb in Figure 4-31, there was an initial, rapid leaching of Zn and Fe with intercept values of 80 prnol and 670 pmol, respectively. This is presumed to be surface controlled followed by a slow diffusion reaction. A similar leaching pattern also was observed for Cu, Zn. and Pb in Figure 4.9, which indicates their common rate controlling mechanism. Fe could have leached out from chalcopyrite (CuFeS,) and pyrite (FeS,).

O 2 4 6 8 10 Time, h1l2

Time, hl" APPLJED 4 IWGE. Inc - 1653 East Main Street --.- Rochester. NY 14609 USA ------Phone: 71 6/482-0300 ------Fax: 7161288-5989

O 1993, Applled Image, Inc., All Righb Reserved