Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The biooxidation and cyanidation of a refractory arsenical gold concentrate: Jones, Lyn 1998

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_1998-0489.pdf [ 5.97MB ]
JSON: 831-1.0078588.json
JSON-LD: 831-1.0078588-ld.json
RDF/XML (Pretty): 831-1.0078588-rdf.xml
RDF/JSON: 831-1.0078588-rdf.json
Turtle: 831-1.0078588-turtle.txt
N-Triples: 831-1.0078588-rdf-ntriples.txt
Original Record: 831-1.0078588-source.json
Full Text

Full Text

THE BIOOXIDATION AND CYANIDATION OF A REFRACTORY ARSENICAL GOLD CONCENTRATE: FACTORS INFLUENCING CYANIDE CONSUMPTION by LYN JONES B.A.Sc. University of British Columbia, 1996 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE in THE FACULTY OF GRADUATE STUDIES Department of Metals and Materials Engineering We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA August 1998 ©Lyn Jones, 1998 In p resent ing this thesis in part ial fu l f i lment o f t h e requ i rements fo r an advanced degree at the Univers i ty o f Brit ish C o l u m b i a , I agree that the Library shall make it f reely available f o r re ference and s tudy. I fu r ther agree that permiss ion fo r extens ive c o p y i n g o f this thesis f o r scholar ly pu rposes may b e g ran ted by the head of my d e p a r t m e n t or by his o r her representat ives. It is u n d e r s t o o d that c o p y i n g or pub l i ca t i on of this thesis fo r f inancial gain shall n o t be a l l owed w i t h o u t m y w r i t t e n permiss ion . D e p a r t m e n t of f^}iA<, t v * \ J {jjitr^j / The Univers i ty of Brit ish C o l u m b i a Vancouver , Canada DE-6 (2/88) Abstract A significant drawback to the application of biooxidation as a pretreatment for refractory gold concentrates is the problem of high cyanide consumption. The presence of cyanicides in biooxidized residues are commonly referred to in the literature, but specific mechanisms are not proposed. In this study, a four-stage continuous bench scale bioreactor apparatus was set up to generate a representative biooxidized product suitable for an extensive cyanidation study. In all, four biooxidation steady states were reached. Residence times ranged from 2 to 5 days, while temperature and pulp density remained constant at 35 °C and 20%, respectively. With the first two steady states, the primary stage pH was controlled at 1.8, whereas for the second two the pH was controlled at 1.5. Sulfide oxidation ranged from 70% to 93%. Cyanidation experiments were conducted on the final products, as well as on samples taken from the intermediate stages of the reactor. The solids were leached under varying conditions of pulp density, NaCN concentration, pH, temperature, and stirring speed. Gold extractions ranged from 46.1% to 95.6%. The sodium cyanide consumption varied from 2.70 to 12.07 kg/tonne-concentrate. Results indicated that the high cyanide consumption associated with this concentrate is the result of two separate reaction mechanisms: one forming thiocyanate (SCN") possibly from a reactive form of elemental sulfur, and the other forming hexacyanoferrates (Fe(CN)63" and Fe(CN)54') by decomposition of basic ferric sulfate precipitates. Cyanide use can be reduced by employing higher pulp densities, primary stage aeration, and shorter overall leach times. ii Table of Contents Abstract ii List of Tables v List of Figures vi Nomenclature viii Acknowledgements ix 1. Introduction 1 2. Literature Review 5 2.1 Cyanidation 5 2.1.1 Chemistry 5 2.1.2 Cyanicides 9 Cyanate 9 Thiocyanate 10 Metal Cyanides 12 2.1.5 Cyanidation of Refractory Ores 15 Telluride Ores 15 Carbonaceous Ores 16 Sulfide Ores 16 2.1.6 Modified Cyanidation 17 2.1.7 Oxidation Pre-treatment 18 2.2 Bacterial Oxidation of Refractory Gold Ores 20 2.2.1 Leaching Bacteria 20 2.2.2 Leaching Mechanisms 20 2.2.3 Cell Attachment... 22 2.2.4 The Organic Capsule 24 Sulfur Pathways 25 Sulfur Granules 27 Sulfur Globules 29 2.2.5 Bacterial Metabolism 30 Iron Oxidation 31 Sulfur Oxidation 34 2.2.6 Summary 38 2.3 Plant Practice 41 2.3.1 Operating Parameters 41 Temperature 41 pH 42 Nutrients 42 Gas Concentrations 43 Toxic Compounds 43 2.3.2 Design Considerations 44 Tank Size and Configuration 45 Residence Time 46 Pulp Density 46 iii 2.3.3 Applications 47 Fairview 48 SaoBento 49 Wiluna 50 Youanmi 50 Ashanti 51 2.4 Biooxidation and Cyanide Consumption 52 2.5 Summary 56 3. Methods 57 3.1 The Concentrate 57 3.2 Bacterial Leaching Experiments 60 3.3 Cyanidation 63 3.3.1 Standard Cyanidation Tests 64 3.3.2 Cyanide Consumption Tests 66 4. Results and Discussion 68 4.1 Bioleaching Results 68 4.1.1 Steady State #1 73 4.1.2 Steady State # 2 75 4.1.3 Steady State #3 77 4.1.4 Steady State #4 79 4.2 Analysis of the Four Mass Balance Periods 80 4.2.1 Lime Consumption 87 4.2.2 Leaching Kinetics 89 4.3 Cyanide Mass Balance 92 4.4 Effect of Leach Parameters on Cyanide Consumption 96 4.4.1 Pulp Density 96 4.4.2 Temperature 99 4.4.3 pH 102 4.4.4 Cyanide Concentration 103 4.4.5 Stirring Speed 105 4.5 Sample Treatments 107 4.6 Sulfur Balance 115 5. Conclusions and Recommendations 121 5.1 Recommendations 123 Works Cited 125 APPENDIX A - Analytical Methods 133 APPENDIX B - Thermodynamic Data 137 APPENDIX C - Sample Calculations 138 APPENDIX D - Sample Data 141 APPENDIX E - Statistical Analysis 148 iv List of Tables Table 2-1. Some common metal-cyanide complexes and their equilibrium constants 12 Table 2-2. Cyanide consumption following bacterial oxidation pre-treatment 52 Table 3-1. Reagents used in the production of the concentrate 59 Table 3-2. Results of chemical analysis of the concentrate 59 Table 3-3. Nutrient media used during continuous operation 62 Table 4-1. Operating conditions at Steady State #1 74 Table 4-2. Bioleach data summary for Steady State #1 74 Table 4-3. Cyanide leach data for Steady State #1 75 Table 4-4. Operating conditions at Steady State #2 76 Table 4-5. Bioleach data summary for Steady State #2 76 Table 4-6. Cyanide leach data for Steady State #2 77 Table 4-7. Operating conditions at Steady State #3 77 Table 4-8. Bioleach data summary for Steady State #3 78 Table 4-9. Cyanide leach data for Steady State #3 78 Table 4-10. Operating conditions at Steady State #4 79 Table 4-11. Bioleach data summary for Steady State #4 79 Table 4-12. Cyanide leach data for Steady State #4 80 Table 4-13. Lime consumption for each of the four steady states 87 Table 4-14. Possible cyanide consuming species, the most probable complex formed, and the mass ratio of sodium cyanide complexed to species present 92 Table 4-15. Mass balance results for CN-4-P3 93 Table 4-16. Mass balance for the sealed reactor leach 95 Table 4-17. Comparison of cyanide consumption and % of cyanide consumed as thiocyanate and hexacyanoferrate for three different pulp densities 98 Table 4-18. Comparison of gold extraction, cyanide consumption, and % of cyanide consumed as thiocyanate for three different leach temperatures 101 Table 4-19. Comparison cyanide consumption and % of cyanide consumed as thiocyanate and hexacyanoferrate at varying pH 103 Table 4-20. Gold extraction and cyanide consumption for varying sodium cyanide concentrations 104 Table 4-21. Comparison of cyanide consumption and % of cyanide consumed as thiocyanate and hexacyanoferrate for two sample treatments 109 Table 4-22. Results of the S7SCN" balance 117 Table 4-23. Results of the S27SCN" balance 119 Table 4-24. Results of the NaOH leach 120 List of Figures Figure 2-1. Eh-pH diagram for the Au-CN-H20 system at 25 °C 8 Figure 2-2. Eh-pH diagram for the Fe-CN-H20 system at 25 °C 13 Figure 2-3. Transmission electron micrograph of bacterial cell attached to pyrite 24 Figure 2-4. Sulfur oxidation pathway proposed Lorbach et al. (1993) 25 Figure 2-5. Two possible arrangements for sulfur micelles proposed by Stuedel (1989) 29 Figure 2-6. Possible structure of the polythionate membrane as proposed by Stuedel (1989) 30 Figure 2-7. Oxidation of ferrous ions in the periplasmic space 32 Figure 2-8. Possible mechanism of electron transfer from ferrous iron to oxygen in T. Ferrooxidans 32 Figure 2-9. Eh-pH diagram for the Fe-S-H20 system at 25 °C 34 Figure 2-10. Electron transport chain for sulfur in T. ferrooxidans 36 Figure 2-11. Possible pathways of sulfur at the bacteria-mineral interface 40 Figure 2-12. Typical biooxidation flowsheet 45 Figure 3-1. Schematic diagram of the pilot flotation circuit 58 Figure 3-2. Schematic of the continuous reactor 61 Figure 3-3. Schematic of the cyanidation reactor 64 Figure 3-4. Schematic of the sealed cyanidation reactor 66 Figure 4-1. Eh and pH measurements for the four-stage reactor system 69 Figure 4-2. Change in iron concentration over the continuous run 71 Figure 4-3. Change in arsenic concentration over the continuous run 72 Figure 4-4. Change in pulp density between stages for the four steady states 81 Figure 4-5. Comparison of sulfide oxidation, and iron and arsenic extraction, for each steady state 82 Figure 4-6. Gold extraction vs. sulfide oxidation 84 Figure 4-7. Cyanide consumption as a function of sulfide oxidation 85 Figure 4-8. The percentage of NaCN consumed as SCN" plotted against sulfide oxidation 86 Figure 4-9. Lime consumption as a function of sulfide oxidation for the cyanidation experiments 88 Figure 4-10. Gold extraction vs time for cyanide leach of steady state 4 product 89 Figure 4-11. Consumption of cyanide during a 24 hour leach test of Steady State #4 product ...91 Figure 4-12. Effect of pulp density on sodium cyanide consumption measured in kg/t 97 Figure 4-13. Effect of pulp density on sodium cyanide consumption measured in ppm 98 Figure 4-14. Relative thiocyanate formation 99 Figure 4-15. Effect of temperature on sodium cyanide consumption 100 Figure 4-16. Effect of temperature on relative SCN" formation 101 Figure 4-17. Effect of pH on sodium cyanide consumption 102 Figure 4-18. The effect of cyanide concentration on consumption 104 Figure 4-19. Effect of stirring speed on sodium cyanide consumption 105 Figure 4-20. Thiocyanate and hexacyanoferrate concentrations in cyanide leach solutions at three stirring speeds 106 Figure 4-21. Sodium cyanide consumption for a slurry sample treated at 110 °C for 3 hours... 108 Figure 4-22. The effect of aeration on sodium cyanide consumption. 109 vi Figure 4-23. The distribution of iron-cyanide and sulfur-cyanide species for tests conducted at varying pH 112 Figure 4-24. Eh-pH diagram for the Iron-Sulfur-Cyanide-Water system at 25 °C 114 Figure 4-25. Eh-pH diagram for the Sulfur-Cyanide-Water system at 25 °C 116 vii Nomenclature Eh: potential [mV] SHE: standard hydrogen electrode reference P80: mesh or grind size that results in 80% passing tpd: tonnes per day Lpm: litres per minute viii Acknowledgements I would like to express my utmost gratitude to my supervisor, Dr. Ralph Hackl, for his advice and guidance during the past two years, and for his patience over the last two months. During the experimental phase of the program, a number of people in the biohydro and hydro labs contributed their expertise to this project. In particular, I would like to thank Chris Pasetka and Masoud Aftaita for teaching me everything I know about bench-scale continuous bioleaching. Also, the analytical work of Simon Jupp, the advice of Be Wassink, and the proofreading expertise of Mike Timmins and Mike Noble, are greatly appreciated. I would like to thank my parents for their support of my decision to return to school seven years ago, and for all the encouragement they have provided along the way. Finally, this thesis would not have been possible without the generous support of the Natural Sciences and Engineering Research Council and Placer Dome Inc. ix 1. Introduction Refractory gold ores are hydrothermal deposits containing microscopic gold particles encapsulated within a gangue mineral, or adsorbed on carbonaceous matter. In many instances the matrix consists of sulfide minerals such as pyrite or arsenopyrite. Typically, these deposits contain less than 10 g of gold per tonne of ore. Conventional methods of gold recovery such as cyanidation are ineffective at extracting the gold. Even after grinding to 70-80% minus 74 microns cyanidation will yield recoveries anywhere from 6% to 40% (Haines and van Aswegen, 1990). For many years, the solution to refractory sulfide ores has been roasting. The high sulfur concentration of the ore renders the overall reaction exothermic, making it economically viable. Unfortunately, the roasting of sulfide ores produces both sulfur dioxide, which contributes to acid rain, as well as the toxic compound arsenic trioxide. Although new technology exists to convert S02 to sulfuric acid and to strip arsenic trioxide from flue gas emissions, increasingly stringent environmental regulations make this a rather tenuous option. The need for an alternative process has spawned a number of "second generation" treatment methods. These include pressure oxidation, biooxidation, the Arseno Process, the Nitrox Process, and chlorination. Of these, only pressure oxidation and biooxidation have been successfully implemented on a commercial scale. 1 Despite the initial success of pressure oxidation, the use of bacteria to improve gold recovery from refractory ores has gained increasing attention since the late 1970's (van Aswegen, 1993). Since that time, six plants have been commissioned. The process involves the breakdown of the sulfide minerals by a mixed culture of iron- and sulfur-oxidizing microorganisms. Leaching takes place in large tanks under conditions of controlled pH, temperature, and oxygen concentration. As with pressure oxidation, the result is the production of sulfate and arsenate, and the liberation of the finely disseminated gold particles. Unfortunately, biooxidation of refractory ores has not yet found a niche in the North American metal processing industry, but successful pilot operations such as those at the Dickenson and the Salmita Yellowknife Mines have proven the process to be technically viable (Lawrence, 1994; Hackl, 1996). Lawrence (1994) suggests that North American operators are hesitant to adopt bioleaching because of perceived technical and economic risks associated with the complex nature of the process. Perhaps foremost among these concerns is the ability of a bioleaching system to tolerate various process upsets such as fluctuations in temperature and dissolved oxygen concentration, changing feed compositions, and the influence of toxic compounds such as mercury, arsenic, chloride, and even flotation reagents. Under most conditions, bioleaching has proven to be remarkably robust: capable of adapting rapidly to changes in process parameters. However, the stability of biooxidized residues, particularly with respect to arsenic, are also of concern. 2 One specific characteristic of bio/xidized refractory gold concentrates that may prevent the application of this technology to lower grade feeds is high cyanide consumption. Indeed, reagent costs for cyanidation are many times greater for biooxidized solids than for their pressure-oxidized counterparts. This added cost has evidently proven acceptable for the high grade material fed to the five commercial-scale plants currently operating, but it is unlikely to be profitable in North America, where concentrate gold grades are lower. While much research has focused upon the mechanisms, kinetics, and process control of bacterial oxidation, there has been very little attention paid to the factors affecting cyanide consumption. As a result, this thesis endeavors to better understand this phenomenon. To accomplish this goal, the project has three principal objectives: 1. To generate a representative biooxidized product using a multi-stage, continuous, bench-scale rector. 2. To investigate the influence of process parameters such as pH, temperature, and cyanide concentration on cyanide consumption. 3. To identify the mechanisms responsible for the consumption of cyanide, and to propose ways to mitigate this effect. 3 This thesis consists of four principal sections. The first contains a review of the current literature pertaining to both bacterial oxidation and cyanidation. Secondly, a detailed description of the experimental program is included. The results obtained are presented and discussed in section three. Finally, some conclusions are drawn, and recommendations made. 4 2. Literature Review 2.1 Cyanidation The use of cyanidation for the recovery of gold and silver has gained widespread acceptance in industry since it was patented in 1889. This method has replaced more traditional practices such as mercury amalgamation and smelting. Ten of the world's largest gold producing plants employ a cyanidation circuit (Jha, 1987). Cyanidation offers a simple, low-cost method of dissolving gold particles to produce a soluble, gold-cyanide complex. Dissolution takes place in heaps or in agitated leach tanks, using potassium or sodium cyanide at concentrations of less than 0.3 %. Under optimum conditions, gold particles will dissolve at a rate of 3.25 mg/cm2/h (Fleming, 1992). In other words, a spherical particle passing 325-mesh will take thirteen hours to dissolve. However, the influence of gangue materials in the leach, as well as oxide coatings and films, will tend to slow the rate of dissolution. As a result, a gravity concentration step may be employed to remove any coarse gold particles from the leach prior to cyanidation. 2.1.1 Chemistry The dissolution of gold in a cyanide solution proceeds according to Eisner's equation: 5 4Au + SCN~ + 02 + 2H20 -> 4Au(CN)~2 + 40H' (2-1) However, the reaction probably has an intermediate step involving hydrogen peroxide (Habashi, 1991). Thus, the equations are as follows: The reactions indicate the electrochemical nature of the cyanide leaching process with the gold undergoing anodic dissolution. In turn, the cathodic reaction involves the reduction of oxygen. Similar equations can be written for silver. The cyanide ion consists of a triply bonded carbon and nitrogen atom. One of these is a sigma-bond, while the other two are pi-bonds. The stability of the metal cyanide complexes is due to pi-bonding between the d-orbital electrons of the metal and the pi-electrons of the cyanide ion. The rate of gold dissolution has been shown to be principally dependent upon the concentration of cyanide and oxygen in the leach solution (Kudryk and Kellog, 1994). In an air saturated solution the rate of dissolution increases dramatically with cyanide concentration until a maximum is reached. After that point, between 0.1 and 0.2 percent, the rate slowly decreases. Similarly, higher oxygen partial pressures will also improve the rate of dissolution. 2Au + 4CN- + 02 + 2H20 -> 2Au(CN)2 + 20H' + H202 (2-2) 2Au + 4CN~ + H202 2Au(CN)2 + 40H~ (2-3) 6 Hydrogen cyanide can be evolved according to the following reaction: CN~ + H+ -> HCN1~ (2-4) Since the formation of hydrogen cyanide produces a toxic gas and results in the loss of reagent from the system, steps are taken to prevent HCN evolution. Usually, this means maintaining the pH of the cyanide circuit well above 9.21, the pKa of the reaction. Commonly, the pH of the leach solution is raised above 9.21 through the addition of Ca(OH)2. This practice can, however, generate calcium peroxide when H202 (from equation 2-2) is present in solution, as follows: The Ca02 can retard dissolution by forming a passive layer on the surface of the metal. This passivation becomes particularly evident as the pH rises above 11. Conversely, this phenomenon is not observed when the pH is controlled with sodium hydroxide (Habashi, 1991). However, gold dissolution slows down at pH's greater than 12, possibly the result of oxidation of CN" to CNO" (cyanate). Ca(OH)2 +2H202 -> Ca02 +2H20 (2-5) Sparrow and Woodcock (1995) cite several advantages for cyanide leaching at moderate pH's: 9.5-10.0. These include lower lime consumption, a more favorable pH for CIP absorption, and 7 the ability to use magnesium containing make-up water. Gold dissolution slows below pH 8, however, as the concentration of CN" declines. The Eh-pH diagram for the Au-CN-H20 system is shown in Figure 2-1. The diagram illustrates the stability of the aurocyanide complex even at moderate oxidizing potentials. In addition, the typical leaching pH (10-11) also favors complexation over formation of an oxide or a passive film. P H Figure 2-1. Eh-pH diagram for the Au-CN-H 2 0 system at 25 °C. Activity of cyanide 10"2 M . activity of other aqueous species: 10"3M. Note: Au(CN) 2 ( S ) not included, (drawn using the CSIRO Thermochemistry Program, see Appendix B for data). 8 From Eisner's equation it can be seen that oxygen is required during the anodic dissolution of gold. Since 02 is expensive to produce, and is only sparingly soluble in water, a number of other oxidants have been used in attempts to increase the leaching kinetics. These oxidants include hydrogen peroxide, potassium permanganate, bromine, and chlorine (Jha, 1987). However, these additional reagents add further expense to the process, and offer little advantage over a conventional, well aerated system. The rate of dissolution has also been shown to be temperature dependent. At about 85 °C the reaction reaches a balance point between kinetics and oxygen solubility. However, the cost of heating large volumes of slurry, coupled with higher cyanide consumption, make this option impractical in most situations. 2.1.2 Cyanicides The cyanide process may be adversely affected by the presence of cyanide consuming compounds. These cyanicides include natural acids, soluble salts, sulfur compounds, and oxides of copper, zinc, and antimony (Bhappu, 1990). Cyanate The oxidation of cyanide to cyanate (CNO") can occur in the presence of oxidants such as hydrogen peroxide and ozone. The reaction with H202 is kinetically very slow, but the reaction 9 is catalyzed by the presence of Cu2+, and also proceeds at an observable rate in the presence of thiosulfate (Flynn and Haslem, 1995). A similar conversion to cyanate is effected with 02 in the presence of S02 in what is known as the "Inco S02/air" process for cyanide destruction (Devuyst et al., 1989). Here, the oxygen and sulfur dioxide combine to form peroxomonosulphate (HOOSOy), which in turn oxidizes the cyanide. The oxidation of cyanide can also be accomplished by oxygen according to the reaction: 2CN~ + <92 2CNCT (2-6) Although thermodynamically favorable under typical leaching conditions, the kinetics of this reaction are reported to be very slow (Adams, 1990). The rate in this case may be controlled by several factors including temperature and the presence of impurities (Flynn and Haslem, 1995). In addition, the reaction has been shown to be catalyzed by activated carbon and copper. Thiocyanate Free cyanide can also be lost from the leach solution by conversion to thiocyanate. The formation of SCN" is attributed to the presence of sulfur compounds such as sulfide, elemental sulfur, or thiosulfate. For example, feed material containing only 1% elemental sulfur by weight is capable of consuming over 15 kg/t of NaCN (Hackl, 1989). A general form of the reaction of elemental sulfur with cyanide is as follows: CN~ + | 5 g -+SCN~ (2-7) 10 Luthy and Bruce (1979) report that aqueous sulfur species such as polysulfide and thiosulfate also react with free cyanide, as follows: CN' + S2X~ -> SCN~ + S2" (2-8) CN~ +S2 O2" -> SCN' + S02f (2-9) Thiocyanate, in turn, can be oxidized to sulfate and cyanate by oxygen, ozone, and hydrogen peroxide, as discussed earlier. Conversely, thiocyanate may be oxidized back to CN" by ozone A novel process developed by Soto et al. (1994) uses 1 kg of 03 per kg of SCN" under mildly acidic conditions to promote the following reactions, SCN~ + 303+20H~ -> CN' + S024~ +302 + H20 (2-10) SCN~ +03 + H20^> CN~ + H2SO, (2-11) thus recovering the cyanide. The low pH is effective in slowing the subsequent oxidation of cyanide to cyanate because the cyanide is present primarily as the slow to oxidize HCN rather than CN". 11 Metal Cyanides In addition to silver and gold, a number of base metal ions also form cyanide complexes. In this case, their effect is considered deleterious as they reduce the free cyanide concentration in the leach solution. Table 2-1 lists some metal ion cyanide consumers and the species they form. Table 2-1. Some common metal-cyanide complexes and their equilibrium constants (Flynn and Haslem, 1995). Metal Ion Cyanide Complex LogK cobalt Co3+ Co(CN)63" 73.0 copper Cu+ CuCN(s)/Cu(CN)32- 10.5/27.0 gold Au+ Au(CN)2' 37.0 iron Fe27Fe3+ Fe(CN)647Fe(CN)63- 35.4/42.3 nickel Ni2+ Ni(CN)42" 30.2 silver Ag+ Ag(CN)2" 20.5 zinc Zn2+ Zn(CN)42" 19.6 Of particular interest during cyanidation of pyrite and arsenopyrite containing ores is the influence of iron. Thermodynamic data has been summarized by several authors (Zhang et al., 1997; Flynn and Haslem, 1995) and an Eh-pH diagram for the iron-cyanide water system is shown in Figure 2-2. 12 PLOT LABELS Temp. = 298.15 K [CN] = 0.01 H IFel = 0.01 H STRBLE AREAS Fe [C N16 <3-> Io*) F 9 [C N16 <«-> let) Fa 0 0 H Fe Fa <2+> (n«t) Fa <3+> (oq| H Fe IC N)6 <3-> (oq) LICRND AREAS Af) HCN (PQ) BB CN <-> IRQ) H20 3TRB1L1TY LIMITS 1 OXTCEN 2 HYDROGEN Figure 2-2. Eh-pH diagram for the Fe-CN-H20 system at 25 °C. Activity of cyanide: 10"2M; activity of aqueous iron species: 10"3 M. (drawn using the CSIRO Thermochemistry Program, see Appendix B for data). Over the near-neutral to alkaline pH range the stable compounds hexacyanoferrate (II) (Fe(CN)64") and hexacyanoferrate (III) (Fe(CN)63") are formed (although the former is more prevalent under typical leach conditions (Osseo-Asare et al., 1984)). Zhang et al. (1997) proposed the cyanide dissolution of pyrite according to the following mechanism: 13 FeS2 (s) + 6CN- - » Fe(CN)l~ +S22 2- (2-12) Under the oxidizing conditions in the leach, the sulfide is further converted, as follows: However, many cyanidation plants treating pyrite containing ores or concentrates experience low sodium cyanide consumptions (i.e. 1 kg/t or less), so reaction 2-12 is likely to be kinetically slow. Other workers have suggested that jarosites might be the source of the iron (Komitsas and Pooley, 1989). Jarosite compounds have the form AFe3(S04)2(OH)6 where A represents certain univalent ions such as H30+, K+, Na+, Ag+. Precipitation of these compounds is favored in a number of metallurgical processes as a means of iron, sulfate, or alkali control (Dutrizac and Jambor, 1987; Dutrizac and Hardy, 1997). Optimum temperatures for precipitation of crystallized jarosite is found to occur at temperatures in excess of 90°C, however a chemically similar amorphous form is associated with the more moderate temperatures maintained during bioleaching. A further consequence of iron cyanide complexation is the formation of Prussian Blue in neutral or acidic solutions (Burns et al., 1980). A pale blue precipitate forms when ferric ions react with hexacyanoferrate (II): 5 22 _ S" -> S202~ -> SO2' -> SO] ,2- (2-13) 14 K+ + Fei+ + Fe{CN)\- -» KFe[Fe(CN)6] (2-14) Conversely, the same reaction involving hexacyanoferrate (III) gives a brownish color, but does not precipitate. 2.1.5 Cyanidation of Refractory Ores For many decades, the cyanidation process was applied only to free-milling ores. Ores with poor cyanidation characteristics were labeled as refractory and usually separated and sent to the dump. As the high grade, free-milling ore deposits were exhausted more focus was placed on treating these refractory ores. From the perspective of cyanidation, refractory deposits can be grouped into three main categories: telluride ores, carbonaceous ores, and sulfide ores. In some cases, the ore may contain elements of more than one of these categories. Telluride Ores Gold tellurides require an oxidation step before cyanidation. Without this step dissolution is extremely slow or impossible. At the Emperor Mine in Fiji a chemical oxidation step is employed to improve gold and silver recovery. In this case, NaOH or Na2C03 is used with Ca(OCl)2(Jha, 1987). 15 Carbonaceous Ores Carbonaceous ores contain organic carbon compounds. They are said to be refractory for two principal reasons. Firstly, the ores can adsorb the aurocyanide complex soon after it has formed, and as a result are considered "preg-robbing". Secondly, the gold may possibly be chemically combined with the carbonaceous material. As with the tellurides, the solution is to oxidize or destroy the carbon prior to cyanidation. Sulfide Ores The largest and most important category of refractory ores is the sulfides. These usually consist of pyrite, arsenopyrite, and pyrrhotite (Pinches et al, 1994). In some cases, fine grinding or pre-treatment with lime is sufficient to improve recoveries. However, usually it is necessary to remove the sulfide matrix surrounding the finely disseminated gold particles. Again, the solution is oxidation. The resistance of sulfide minerals to cyanidation is thought to be the result of three principal mechanisms. The first of these mechanisms is physical. The gold particles are finely disseminated, often sub-micron in size, within the sulfide matrix. In turn, the sulfides may be contained as small particles within other gangue material, such as quartz. Grinding and flotation may be employed to isolate the sulfide mineral, but does little to permit the access of cyanide to the precious metals. 16 A second mechanism of resistance to cyanidation is in the chemical composition of the mineral: some base metal sulfides react with cyanide in solution. For example, pyrrhotite (FeS) combines with cyanide to form an iron-cyanide complex via the reaction: FeS + 0.5O2+6CN- + H20 ->[Fe(CN)6]4- + S° +20H- (2-15) For each mole of pyrrhotite leached there are six moles of cyanide and a half mole of oxygen consumed (Kondos et al., 1995). Furthermore, the sulfur can combine with cyanide to form thiocyanate (SCN'). Finally, cyanidation can be affected by an electrochemical mechanism. If the gold is in contact with a conductive mineral, passivation of the more noble gold can occur (Jha, 1987). Several methods to overcome the poor cyanidation characteristics of refractory ores have been proposed. These methods have been classified by Jha (1987) as falling into one of two categories: modified cyanidation and oxidation pre-treatment. 2.1.6 Modified Cyanidation Modified cyanidation refers to changes made in treatment methods to improve recoveries from refractory ores. This can include lengthening the time of leaching, increasing the concentration of cyanide, raising the temperature of the slurry, or increasing the oxygen partial pressure. Generally, these methods are effective at improving recoveries only from mildly refractory ores. 17 For preg-robbing ores, a further modification can be made by employing a carbon-in-leach (CIL) cyanidation circuit. By adding the activated carbon to the leach tanks, the aurocyanide complex is quickly adsorbed on the activated carbon before being affected by the preg-robbing organic material. 2.1.7 Oxidation Pre-treatment Unfortunately, modified cyanidation is usually not able to overcome the problems associated with gold encased in sulfide ores. Here the only solution is to oxidize the mineral and expose the precious metals. Currently, three very different processes are competing to achieve this conversion: roasting, pressure oxidation, and bio-oxidation. Roasting was the initial solution to refractory ores more than sixty years ago. Since that time, hearth roasters have evolved into a fluidized-bed design and represent the most common method of treating pyrite and arsenopyrite deposits. In this process most of the sulfur and arsenic is removed from the flue gas as sulfuric acid (H2S04) and arsenic trioxide (As203). However, increasing concerns about the impact of flue gas emissions on the environment have led to the development of other oxidation technologies. At present, pressure oxidation is the most popular treatment method for new refractory gold operations. In this process, a flotation concentrate of the sulfide mineral is fed into an autoclave. 18 There, in the presence of pure oxygen gas, high temperature and pressure, and a ferric catalyst, the sulfide minerals are oxidized to sulfate. In addition, it is claimed that arsenic can be removed from the system in the form of a stable ferric arsenate precipitate (Lawrence, 1994). Bioleaching employs naturally occurring bacteria to expose the gold particles. The microorganisms use sulfide minerals as an energy source and produce sulfate as a waste product. This process is much slower than pressure oxidation or roasting, but is also simpler to operate. Despite the differences in the processes, roasting, pressure oxidation, and bacterial oxidation can all yield metal recoveries in excess of 90%. Selection of the method is dependent largely upon economic considerations and process limitations, as no one option holds an advantage under all circumstances. For example, roasting offers significant economies of scale for plants processing in excess of 1000 tonnes of concentrate per day. Conversely, for smaller operations (<300 tpd), bacterial oxidation requires the lowest capital investment (Carter, 1991). In addition, the type of ore will greatly influence process selection. For example, as Lawrence (1994) points out, neither bacterial leaching nor pressure oxidation are favorable for treating alkaline or carbonaceous ores. Conversely, high arsenic levels can render roasting uneconomic, but may favor pressure oxidation or biooxidation. 19 2.2 Bacterial Oxidation of Refractory Gold Ores 2.2.1 Leaching Bacteria Leaching bacteria usually consist of a mixed culture of acidophilic chemolithotrophic bacteria, the most important of which are the strains Thiobacillus ferrooxidans, Thiobacillus thiooxidans, and Leptospirillum ferrooxidans. The latter derives energy from the oxidation of ferrous iron, while T. thiooxidans oxidizes reduced sulfur compounds. As the name implies, T. ferrooxidans is capable of using both of these energy sources. All of these bacteria thrive under conditions of low pH (0.5-3.0) and moderate temperature (30-40°C). Research is continuing into the use of moderate and even extreme thermophiles such as Sulfolobus, which operate at temperatures as high as 90°C. Despite the potential advantages of improved kinetics and reduced cooling requirements, these strains have so far met with limited success. The higher temperature organisms may not be as resistant to toxic compounds and shear stress as their more moderate counterparts (Morin, 1995). 2.2.2 Leaching Mechanisms The oxidation of pyrite and arsenopyrite by T. ferrooxidans is thought to be the result of two principal mechanisms. These mechanisms are referred to as direct and indirect leaching. In each case, the bacteria oxidize a reduced compound and use the released electron to generate adenosine tri-phosphate (ATP) via the electron transport chain. 20 Direct leaching describes the oxidation of the sulfide mineral to sulfate, for pyrite and arsenopyrite the reactions are as follows: FeS2 + 1202 +H20 " >FeS04+H2S04 (2-16) IFeAsS + 702 + 2 H20 + H2S04 bac,trta >2HiAs04 + 2Fe2(504)3 (2-17) The high oxygen requirement is evident. As a result, the process requires significant oxygen £ liquid mass transfer to operate efficiently, which is achieved by air sparging. Also, direct leaching can generate both elemental sulfur (S°) and ferrous (Fe2+) iron which can be further oxidized by the bacteria: S"+i02+H20 * " >H2S04 (2-18) Fe2++}02+H+ hacleria >Fei++iH20 (2-19) In the form of ferric iron, the species can act as an electron acceptor and oxidize the sulfide mineral on its own: FeS2 + 14Fe3+ + 8H20-> \5Fe2+ + 2S02' + 16H+ (2-20) 2FeAsS + 4Fe3+ +^02+3H20^2H3As03 +6Fe2+ +2S" (2-21) 21 This second pathway is known as indirect leaching. As with direct leaching, the oxygen requirements are considerable. Several authors suggest that the indirect method is the principal, and possibly only, mechanism responsible for leaching of sulfide minerals (Pinches et al., 1994; Sand etal., 1995). An important consequence of the indirect mechanism is the selective oxidation of arsenopyrite (Miller and Hansford, 1992). This result is of particular interest when the gold is closely associated with the arsenopyrite. Thus high gold extractions do not always require complete sulfide mineral oxidation. While the leaching mechanisms associated with T. ferrooxidans may seem straightforward, the formation and pathways of the intermediate sulfur compounds is much more complex. This investigation requires an understanding of the physical, chemical, and biological characteristics of the process at the microscopic level. To accomplish this goal, this review will first consider the mechanism by which the bacteria become attached to the mineral surface. 2.2.3 Cell Attachment During bioleaching, T. ferrooxidans is not only found attached to the mineral particles, but swimming freely in the solution as well. Adhesion to the particle surface is achieved by certain exopolymeric compounds secreted by the cell (Sand et al, 1995). These compounds act like a kind of glue that strongly binds the cell to the mineral. While the exact composition of the 22 substance is unknown, it is thought to contain metal cations. These ions provide the exopolymeric layer with a net positive charge. In doing so, the cell can easily bind to the negatively charged mineral surface. Without this layer, attached growth is not possible. Experiments which removed the extracellular polymers by centrifugation showed that attachment took several hours longer than untreated cells (Sand et al., 1995). From this result, the bacteria appear to have to generate a new layer before attached growth can occur. Similarly, attachment also may be dependent on the presence of metal ions in solution. Finally, the location of attachment is not randomly selected. Thiobacillus ferrooxidans contain the genes necessary for chemotaxis. As a result, they are capable of sensing, and moving into, a gradient of increasing substrate concentration. Specifically, the bacteria have been shown to be sensitive to thiosulfate in solution. Thus, they are attracted to sites on the mineral where electrochemical dissolution processes are already taking place. 23 „ „ , | Figure 2-3: Transmission electron micrograph of bacterial cell attached to pyrite. (Rojas-Chapana et al., 1996). 2.2.4 The Organic Capsule Figure 2-3 shows a transmission electron microscope image of a bacterial cell attached to a pyrite grain. The picture clearly illustrates a thick layer of extracellular material referred to as the organic capsule. The capsule includes the exopolymeric compounds responsible for attachment that were discussed in the previous section. In addition, the layer also acts as a reactive medium for sulfur metabolism, and as storage for colloidal sulfur particles that form during bacterial oxidation. These particles are thought to be stored for use as an energy source when all other sources are depleted. 24 Sulfur Pathways The formation of an organic capsule and the attachment of that capsule to the surface of the sulfide mineral provides the setting for bacterial oxidation. Several theories exist about the specific pathways followed. This section examines some possible direct and indirect mechanisms. The direct oxidation of sulfur compounds was investigated by Lorbach et al.(1993). The authors compared kinetic data collected from experiments using elemental sulfur, sulfite, thiosulfate, and tetrathionate. Figure 2-4 illustrates the sulfur pathway proposed by their results. S2Os Figure 2-4: Sulfur oxidation pathway proposed Lorbach et al. (1993). The S-S bond in thiosulfate can be cleaved by the enzyme Rhodanese. In this diagram, the [S] state is intended to represent an "activated" sulfur compound. This compound is thought to be a more readily usable form for the bacteria, particularly at a pH of around 6. As a result, the sulfur pathway appears to branch giving rise to a more acidic route 25 going from sulfur to sulfite near the mineral side of the capsule, versus the "activated" pathway occurring closer to the neutral environment in the cell. Interestingly, thiosulfate is shown as the precursor to both pathways. In comparison, a plausible mechanism for indirect attack has been suggested by Sand et al. (1995). In this instance, the ferric iron hexahydrate attacks the pyrite surface in a series of steps that ultimately leads to the formation of sulfonic acid (see equation 2-22, p] represents pyrite). The sulfonic acid is then hydrolyzed to ferrous iron and thiosulfate (equation 2-23. Note: equation is unbalanced in the literature). The ferrous iron is quickly oxidized to ferric iron by the bacteria, and the cycle begins again. The thiosulfate, on the other hand, is unstable at low pH and is transformed into elemental sulfur or polythionates. p] - Fe - S~ - S' + 6Fe(H20)l+ + 3H20 -> p]- Fe - S~ - SO; + 6Fe(H2O)26+ + 6H+ (2-22) p] - Fe - S~ - SO; + 2H+ ->/?] + Fe2+ + S2 0]~ (2-23) overall: FeS2 + 6Fe(H2 O) f +3H20-> Fe2+ + S2 O2' + 6Fe(H2 O) \+ + 6H+ (2-24) The importance of thiosulfate as an intermediate in sulfide mineral oxidation has been further confirmed by the work of Shrihari et al.(1992). In their experiments the growth of unattached cells were investigated. The authors proposed that the attached cells produced a soluble sulfur intermediate which was then being used as an energy source by the unattached cells. In an 26 attempt to identify and measure this compound, a reactor was constructed that separated attached and unattached cells by a 0.22 um membrane. The free swimming bacteria were confined to the side without any sulfur substrate, surviving only upon substances which could pass through the membrane. Results of this investigation indicated that cell densities increased on both sides of the reactor initially, with the non-sulfur side reaching a stationary phase first. This appears to support the idea of a soluble intermediate compound. The authors then employed a series of methods including iodometric titrations, UV spectrum, and differential pulse polarography, to identify the substance. The analysis indicated levels of thiosulfate and sulfite of 14 and 0.3 ppm, respectively. These findings were considered sufficient to explain the growth of unattached cells in the sulfur-free medium. The formation of intermediate sulfur compounds during the direct and indirect attack of sulfide minerals plays a key role in determining the composition of the organic capsule. The next section examines how the cell organizes and stores sulfur in this outer layer. Sulfur Granules The organic capsule shown in Figure 2-3 clearly illustrates that the layer is filled with colloidal particles. Rojas-Chapana (1996) used energy-dispersive X-ray microanalysis to determine that these colloids contain sulfur. The source of this sulfur is believed to be pyrite and the colloids are thought to serve as energy storage for the bacteria. While elemental sulfur is the most stable of the intermediate compounds, and thus is easily identified in the capsule, these granules are more likely to be composed of long-chain, 27 polythionate species. Stuedel (1989) suggests a mechanism by which polythionates are generated from thiosulfate under acid conditions, as follows: S20]~ + H+ -> HS2 O; (2-25) HS2 0~ + S2 O]- -> HS3O; + SO]' (2-26) 2HS30- -> H2S + S5026~ (2-27) In this way, the stable polythionate species is formed. This mechanism also generates sulfite which is then oxidized to sulfate by T. ferrooxidans. The sulfur granules can be generated from polythionates by orienting their negatively charged, hydrophilic ends towards the outside of the structure, while the non-polar portion of the chain stays in the centre. Two such configurations, known as micelles, are shown in Figure 2-5. In Figure 2-5A the polythionate strands can stretch from one side of the micelle to the other. The circular compounds within represent hydrophobic elemental sulfur. The diameter of the micelle is about 6 nanometres. In Figure 2-5B, a larger, heterogeneous, micelle is shown. This structure consists of an outer layer of folded polythionate chains and an inner core of densely packed, sulfur ring molecules. 28 Figure 2-5: Two possible arrangements for sulfur micelles proposed by Stuedel (1989). A) a homogenous configuration; B) polythionates organized at the surface with core of sulfur ring molecules. (Stuedel, 1989) Sulfur Globules The sulfur granules discussed in the previous section can also combine together to form larger elements known as globules. These globules can be found adjacent to the cell wall, floating . freely within the organic capsule, or even outside the capsule itself. Apparently these larger structures serve to further organize the sulfur intermediates for storage and possible future use. The globules are typically between 0.1 and 0.5 um in diameter (Stuedel, 1989). Their shape is maintained by a polythionate membrane similar to that found in the individual micelles. Figure 2-6 illustrates the orientation of the hydrophilic "heads" towards the 29 outside of the membrane surface. In this case, water is found inside and outside the globule, but it cannot pass through the membrane. Figure 2-6: Possible structure of the polythionate membrane as proposed by Stuedel (1989). 2.2.5 Bacterial Metabolism Chemolithotrophs use inorganic compounds to generate the energy and reducing power necessary for survival. The energy is generated through the electron transport chain and is stored as ATP. The ATP can then be converted to back to ADP (adenosine di-phosphate) releasing 26.4 kJ/mol of energy. This energy is used during various cellular processes including motility and biosynthesis. In addition, ATP can be used via the reverse electron transport chain (an energy consuming set of reactions) to generate nicotinamide adenosine di-phosphate (NADPH): reducing power. 30 Autotrophic organisms use stored energy and reducing power (ATP and NADPH) to convert carbon dioxide into fructose using the dark reactions of photosynthesis. Fructose is converted to pyruvate which then enters the tricarboxylic acid cycle (Brock, 1994). From this point, all of the amino acids needed for biosynthesis can be generated. For the chemoautolithotroph T. ferrooxidans the energy required for these reactions is generated by the oxidation of reduced iron and sulfur compounds. Iron Oxidation The bacterial oxidation of ferrous iron is a relatively straightforward process. The ferrous ions can be pictured as floating freely within the organic capsule. The outer membrane of the bacteria contains porins which are large enough to allow the passage of ions into the periplasmic space (see Figure 2-7). Once inside, the electron is transferred along a series of protein intermediates before reaching the terminal electron acceptor, oxygen, within the cell. 31 pH6-5 inside cell outside cell membrane cell ATP 2 F e 2+ 2Fe 3+ pH2-3 Figure 2-7: Oxidation of ferrous ions in the periplasmic space. (Leach, 1992). cell wall/periplasm membrane cytoplasm Fe (II) RCu cy t a o Figure 2-8: Possible mechanism of electron transfer from ferrous iron to oxygen in T. Ferrooxidans. (Blake et al., 1992) 32 A recently proposed pathway, by Blake et al.(1992), is shown in Figure 2-8. Here, the ferrous ion passes an electron to the acid-stable iron:rusticyanin oxidoreductase (cyt c). From there, the electron is passed to rusticyanin: an acid-stable, soluble, blue copper protein that the authors claim may constitute up to 5% of the total protein synthesis during autotrophic growth on iron. The electron is then transferred to a cytochrome-a containing oxidase embedded in the cell membrane. The final step involves the arrival of the electron at the terminal electron acceptor, and the subsequent reduction of oxygen. Ferric ions are also likely to precipitate as jarosites and ferric arsenate (FeAs04). Pinches et al. (1994) report the precipitation of a basic ferric arsenate in cases where the molar ratio of Fe3+:As5+ is 4:1 or more. This compound has a solubility that is 2 or 3 orders of magnitude lower than ferric arsenate formed at lower molar ratios. The Eh-pH diagram for the iron-sulfur-water system is shown in Figure 2-9. The stability of hydronium jarosite under the conditions of low pH (<1) and high Eh (>600 mVshe) is clearly evident. Ahonen and Tuovinen (1992; 1993) report the precipitation of jarosites in column leaching experiments, and suggest that the concentration of iron in solution is controlled by the solubility of these precipitates. In an effort to identify a more accurate value for iron and arsenic oxidation in bioleached residues, some researchers have used a weak HC1 wash in order to dissolve these precipitates prior to analysis (Zhang et al., 1994; Xiangetal., 1994). 33 Figure 2-9. Eh-pH diagram for the Fe-S-H20 system at 25 °C. Activity of aqueous species: 1 M. (drawn using the CSIRO Thermochemistry Program, see Appendix B for data). Sulfur Oxidation In contrast to the mechanism of iron oxidation, the mechanisms associated with the oxidation of sulfur are extremely complex. In this instance, the electron donor takes several forms, from the relatively soluble polythionate and thiosulfate, to the very insoluble elemental sulfur. The 34 oxidation of these compounds has been suggested to be the result of two main pathways: the reaction of sulfur with enzymes excreted by the cell, or, a reaction occurring between components at the cellular surface (Roy and Trudinger, 1970). Since we have already seen that the bacteria is surrounded by an organic capsule, contact between the cell surface and the sulfide mineral seems unlikely. Thus a direct attack mechanism would require an enzyme to be secreted into the organic capsule. However, oxidation of the intermediate sulfur compounds would probably be best accomplished closer to the cell membrane, where the protein catalysts are less susceptible to loss to, or damage from, the acidic environment. Consequently, the actual mechanism is likely to be a combination of reactions occurring in both the capsule and the periplasmic space. A simplified electron transport chain for sulfur compounds is shown in Figure 2-10. The sulfide electron is picked up by a flavoprotein, while the thiosulfate and sulfur molecules are shown to interact with cytochrome-c. Cytochrome-b has also been detected during sulfide oxidation and was found to be absent in cells grown on Fe(II) (Arkesteyn, 1979). As with the iron oxidation, the terminal electron acceptor is oxygen. 35 N A D P H ^ ,1/2 0 : N A D P ^ — • F P . ^ - ^ Q ^ • C y t • C y t — • C y t a a 3 S2otor S8 e lec t ron f low re leases ene rgy reverse e lec t ron f low g e n e r a t e s reduc ing p o w e r Figure 2-10: Electron transport chain for sulfur in T. ferrooxidans. (Brock et al., 1994). Although the specific pathways remain unclear, some of the important enzymes in the oxidation of sulfur compounds have been identified. Of these, APS-reductase, ADP-sulfurase, glutathionate reductase, and rhodanese are of particular interest. Both APS-reductase and ADP-sulfiirase are involved in the terminal step of sulfur oxidation: the conversion of sulfite to sulfate. Karavaiko et al.(1977) suggest the following two step mechanism, SO]~ + AMP -> APSreductase -> APS + 2e' (2-28) APC + Pt -> ADPsulfurase -> ADP + S0 24~ (2-29) The first reaction forms adenosinephosposulphate (APS), which is involved in the second reaction, and two electrons that enter the electron transport chain, as described earlier. The second equation illustrates the enzyme mediated conversion of APS and a phosphate group to 36 adenosinediphosphate (ADP) and sulfate. A third step involves the reaction of two ADP to form an AMP (to be used in the first equation) and an ATP, the energy unit within the cell. How reduced sulfur compounds are oxidized to sulfite by Thiobacillus has been the subject of much research. In one theory, elemental sulfur can be reduced to sulfide possibly by way of a membrane bound thiol group (Trudinger and Roy, 1970). A likely candidate for this role is the nonprotein thiol glutathione (GSH), which is found in high concentrations in aerobic bacteria (Huxtable, 1986). GSH has also been shown to be a catalyst in the conversion of sulfur and sulfide to sulfite, and finally thiosulfate. Oxidized GSH is subsequently regenerated by the enzyme glutathionate reductase. Another enzyme that might play a part in bacterial sulfur metabolism is rhodanese. This enzyme catalyses the formation of thiocyanate from cyanide and thiosulfate as follows, CAT + S202~ -> SCN~ + SO2- (2-30) Rhodanese is found in mammalian liver cells where it removes cyanide from the bloodstream. Humans ingest up to three times the lethal dosage of cyanide daily in foods such as apples, almonds and tapioca (Whatley, 1980; Huxtable, 1986). However, the release is slow and they can be easily detoxified before harmful quantities can accumulate. Kelly (1980) reasons that rhodanese could be responsible for thiosulfate metabolism. However, there is still the question of the acceptor for the sulfane-sulfur of the thiosulfate. In general, 37 Thiobacilli have high rhodanese activity, but low cyanide concentrations, indicating that another acceptor is dominant. 2.2.6 Summary Based on the information presented here, a possible sulfur pathway, from pyrite to sulfate, begins to take shape. The path is not a direct route, but more like a series of chemical reservoirs and intermediates diffusing through a complex aqueous medium. Figure 2-11 attempts to illustrate some of the transformations taking place in this system. Some of the key elements of the pathway outlined here are as follows: 1. The initial oxidation of pyrite at the mineral surface. This can occur by direct oxidation of the sulfide catalyzed by glutathione leading to sulfite and thiosulfate. Or, it can occur by an indirect mechanism involving the ferrous iron hexahydrate. Again, the first intermediate is thiosulfate. 2. The acid-unstable thiosulfate is converted to polythionate and sulfite. The polythionates combine to form the sulfur granules (micelles) that fill the organic capsule. The soluble sulfite diffuses towards the cell membrane. 38 3. Some of the soluble capsule components are lost to the outside environment. These include waste products such as sulfate, as well as thiosulfate and sulfite which can serve as substrate for unattached cells. 4. Ferrous ions diffuse into the periplasmic space where they are met by cytochrome-c: the initial electron acceptor. The electron is then passed along the chain generating ATP. 5. Sulfite ions also diffuse into the periplasmic space where they combine with AMP, and, under the influence of APS-reductase and AMP-sulfurase, go on to form ATP and sulfate. Subsequently, the ATP is used as energy and the sulfate diffuses out of the cell as a waste product. 6. Finally, the capsule also contains numerous compounds necessary for metabolism. Of these, nutrients such as nitrogen and phosphorous, and oxygen for use as the terminal electron acceptor, are of principal importance. 39 Figure 2-11: Possible pathways of sulfur at the bacteria-mineral interface. (Adapted from Rojas-Chapana et al., 1996). This pathway focuses specifically on the oxidation of pyrite by T. ferrooxidans as a result of the overwhelming amount of research that has been done on this particular system. Quite possibly 40 similar mechanisms exist for other sulfides, such as pyrrhotite and arsenopyrite, and other bacteria such as T. thiooxidans and L. ferrooxidans. However, not enough information currently exists to confirm this assertion. 2.3 Plant Practice The bacterial oxidation of refractory gold ores is presently accomplished by treating a flotation concentrate of the sulfide minerals in a series of stirred tank bioreactors. The biooxidized residue is then treated by the conventional recovery process of cyanidation. Presently, four plants are operating using the "BIOX" process licensed by Genmin, while a fifth employs BacTech's moderately thermophilic culture. 2.3.1 Operating Parameters The objective of stirred tank biooxidation is to maintain the bacterial biomass in a state of maximum growth rate. Thus, their energy and sulfide mineral consumption will be optimized. To achieve this goal requires careful control of key parameters such as temperature, pH, nutrients, carbon dioxide concentration, and oxygen concentration. Temperature With the exception of BacTech's moderate thermophiles, which operate in a temperature range 41 of 40 to 50°C, the mixed culture used in bacterial leaching is characterized as mesophilic. Generally, mesophiles show a preference for temperatures in the 30-35°C range. However, cultures can be gradually acclimatized to temperatures as high as 45°C. pH While iron and sulfide mineral oxidizing bacteria occur over a wide pH range in nature, commercial biooxidation tanks operate only at the bottom end of this scale. Typically, the pH is maintained between 1 and 2 through the addition of sulfuric acid, lime, and/or limestone. Concentrates that contain significant quantities of carbonate minerals will be acid consumers, while those without will tend to produce acid in excess. The target range is low enough to limit the formation of iron precipitates such as ferric hydroxide or jarosites, yet high enough not to inhibit metabolism. Nutrients The presence of nutrients in solution during biooxidation is essential for cell growth. Principal among these are nitrogen and phosphorous along with trace elements such as potassium, magnesium, and calcium. For laboratory study and experimentation the 9K medium developed by Silverman and Lundgren (1959) has gained universal acceptance. A modified version, OK, which does not contain ferrous sulfate, is sufficient for cultures grown on sulfide minerals. In addition, commercial applications may change the nutrient recipe if the concentrate already contains appreciable amounts of a particular mineral, or if a less costly substitute is available 42 (Morin, 1994). One such recipe, used in the MinBac process, provides only nitrogen and phosphorous to the leach (Pinches et al., 1994). Gas Concentrations As illustrated earlier, the breakdown of the sulfide minerals requires significant amounts of oxygen. Unfortunately, for a solution at 40°C the saturation concentration of dissolved oxygen is less than seven parts per million. As a result, the tanks must be continuously sparged with air to provide the stoichiometric oxygen requirement. At the same time, the availability of carbon dioxide is also a consideration. The autotrophic leaching bacteria obtain the carbon necessary for growth by fixing C02 using the dark reactions of photosynthesis. When carbonate minerals are present in the leach, their reaction with the acid produces carbon dioxide: CO]' + 2H+ -> C02 + H2 O (2-31) If, however, little carbonate mineral is present, and no limestone is added for pH control, the air sparged into the tanks can be enriched to 1% C02 to meet the requirement. r Toxic Compounds The bacterial dissolution of the sulfide minerals can also lead to the release of toxic compounds 43 into the leach solution. These compounds can include heavy metals such as lead, mercury and copper, as well as other inhibitors like arsenic. De et al. (1997) have reported that concentrations of mercury, silver, lead, or cadmium that exceed 10 mg/1 are inhibitory to the iron oxidizing ability of Thiobacillus ferrooxidans. The inhibition was found to be uncompetitive in that the metals bind to the enzyme substrate complex and not specifically to the enzyme itself. Conversely, similar concentrations of tin, cobalt, copper, and zinc had no deleterious effect. The presence of arsenic has also been linked to the inhibition of iron oxidation (Mandl et al., 1996). In this case, however, the affect is different depending on whether the trivalent (As(III)) or pentavalent (As(V)) form is present: the toxicity of As(III) has been suggested to be three times greater than that of As(V). In comparison, other research has indicated that both compounds result in a reduced rate of bacterial oxidation with the pentavalent form also inducing a lag phase (Breed et al., 1997). 2.3.2 Design Considerations The design of a bioleaching system requires an understanding of the fundamental parameters influencing the process. Of these, the tank size and configuration, residence time, and pulp density are of key importance. 44 Tank Size and Configuration The configuration and size of the tanks is dependent upon the characteristics of the feed stream. As recently as 1993, the economics of aeration and agitation suggest a maximum tank size of 880m3 (van Aswegen, 1993). However, Briggs and Millard (1997) propose much larger reactors: up to 1380 m3. Figure 2-12 illustrates a flowsheet for a typical plant. In this case, three primary oxidation tanks are operated in parallel followed by another three secondary tanks in series. Afterwards, the slurry goes to liquid-solid separation and then to the cyanidation circuit. Figure 2-12. Typical biooxidation flowsheet, (van Aswegen, 1993) 45 The leaching circuit operates as a four-stage reactor with the first stage being three times the volume of the other three. This configuration is common for an autocatalytic process and has the effect of yielding a high rate of conversion in the first stage. Residence Time The mean residence time of the leaching reactor is dependent upon both the concentrate and the bacterial biomass. If the dilution rate (the inverse of the residence time) is greater than the maximum specific growth rate of the organism, then washout will occur. Generally, this is not a significant concern in bioleaching as longer retention times (2-5 days) are employed to achieve maximum gold recovery (Morin, 1995). As a result, the length of time the slurry spends in the leach reactor is more likely to be related to the mineral being leached and the effectiveness of the biomass to release the gold particles from the sulfide matrix. Furthermore, certain sulfide ores are preferentially leached. Spencer and Budden (1992) showed that bacterial oxidation can selectively attack the more reactive sulfide minerals. Thus, the valuable metals can be recovered without having to completely oxidize the mineral assemblages. For example, an ore containing refractory gold that is associated mainly with arsenopyrite can be selectively leached without oxidizing a significant quantity of the pyrite that may be present. Pulp Density The solids concentration in the reactor is determined by the competing factors of economics and 46 the welfare of the biomass. In this respect, a higher pulp density reduces the size of the reactor and, in turn, the overall cost of the operation. Research indicates that the optimum bacterial kinetics are found at a density of 10-15% (Morin, 1995). Most plants operate at 20%, however, to take advantage of higher throughputs. Apparently, above this concentration the slurry becomes too viscous to permit efficient gas transfer. Also, higher pulp densities may result in substrate or end-product inhibition, as well as shear damage to the bacteria (Pinches et al., 1994). 2.3.3 Applications The application of bacterial oxidation to the treatment of refractory gold ores is dependent on the process being an economically viable alternative to roasting or pressure oxidation. In this respect, biooxidation offers significant savings in capital and operating costs for smaller operations because of the simplicity of the process. As Carter (1991) points out, bacterial oxidation is likely to be cheaper than pressure oxidation for plants handling less than 1200 tpd of concentrate, and cheaper than roasting for less than 1700 tpd. These results concur with the findings of Haines (1986) that suggested that biooxidation would hold a significant capital cost advantage over the other two processes for a plant treating 240 tpd of concentrate. This advantage is realized because of high costs associated with the pollution control systems of the roaster and the oxygen plant required for pressure oxidation. However, as plant size increases, bacterial oxidation has difficulty competing with the other two methods due to upper limitations in reactor size. 47 Small-scale biooxidation plants hold a similar advantage in operating costs. This is due to the simplicity of the process: highly skilled operators are not required. Much of the operating costs are associated with power requirements (for aeration and agitation) and reagent consumption. As a result, the relative costs for biooxidation decrease slowly as plant size increases. In comparison, costs associated with pressure oxidation and roasting decrease more rapidly; with roasting holding an advantage over biooxidation over 400 tpd. While there are no biooxidation plants currently operating in North America, the technology is gaining acceptance overseas. Of the 11 refractory gold treatment processes operating in other parts of the world, 5 have employed either the "BIOX" or "BacTech" process (Lawrence, 1994). What follows is a brief summary of these applications. Fairview The initial inspiration for Genrnin to develop their BIOX process came from the need to find an alternative to roasting at their Fairview Mine in Eastern Transvaal, South Africa. The ore processed at the mine is highly refractory. As a result, conventional cyanidation yields a recovery of only 35%. A flotation concentrate of the pyrite/arsenopyrite feed is produced which grades 120 g/tonne Au and contains 24% sulfur (van Aswegen, 1993). A continuous pilot plant was commissioned in 1984 to treat 750 kg per day of the Fairview 48 concentrate. Based on this experience a demonstration plant was constructed two years later to process 40% of the 35 tpd of concentrate, with the remaining feed going to the roasters. In 1991, the BIOX plant was expanded to treat all of the concentrate, and the roasters were subsequently decommissioned. The plant consistently achieves gold recoveries in excess of 90%. Sao Bento A second BIOX plant was commissioned in 1990 at the Sao Bento mine in Brazil. In this application, bacterial leaching has been implemented as a pre-treatment step to pressure oxidation in order to improve autoclave efficiency. The ore processed by the mine contains the sulfide mineral pyrrhotite which is converted to elemental sulfur at high temperature and pressure. The elemental sulfur has a tendency to agglomerate and thus reduces oxidation and recovery rates. At the same time, the ore also contains siderite, a carbonaceous mineral, that reacts with the acid to form carbon dioxide. In turn, the partial pressure of oxygen in the autoclave is lowered, along with the leaching efficiency. The BIOX pre-treatment step has been used to bacterially oxidize the pyrrhotite and dissolve the siderite before the pulp reaches the autoclave. A single leach tank has been implemented to provide the conversion. The BIOX module fit easily into the existing flowsheet as the autoclave effluent is used to repulp the concentrate prior to biooxidation. Also, the effluent from the module is fed to the previously existing neutralization circuit. As a result, the addition of the pre-treatment stage required only the 49 capital expense of the BIOX system and did not upset the plant's water balance. Wiluna A second Australian plant was commissioned at the Asarco Wiluna mine in 1994 (The first one, at Harbour Lights, was shut down in 1996 due to depleted ore reserves). This plant was designed to treat 115 tpd of flotation concentrate. The concentrate grades 80-90 g/t Au. Gold recoveries of up to 98% are achieved using a first stage consisting of three tanks in parallel, followed by three tanks in series. Most of the gold is associated with the arsenopyrite, but enough is found in pyrite and stibnite to require a sulfide oxidation of between 96 and 98%). Youanmi The use of Bactech's moderately thermophilic culture for bioleaching refractory gold ores has been demonstrated at the Youanmi mine in Western Australia. The plant was commissioned in September of 1994 and processes 93 tpd of flotation concentrate. Originally, the plant was designed to process 120 tpd with a four-tank primary stage and a two-tank secondary stage. However, this configuration has been modified by removing one tank from each stage. The extra tanks are now used to store concentrate, and to maintain a backup culture in case of bacterial loss. According to Brierley and Winby (1996), the plant encountered no significant difficulty in start-up. Furthermore, because the reactors operate at a temperature of between 50 and 52°C, this 50 process has the advantage of reduced cooling requirements when compared to BIOX. Also, the higher temperature may lead to faster reaction kinetics. However, the retention time at Youanmi is five days: the same as at Wiluna. About 92% of the gold in the concentrate is associated with arsenopyrite, which in turn represents about 30% of the total sulfides. The plant oxidizes between 85 and 95% of the arsenopyrite but only 28% of the pyrite. Consequently, the percentage of total sulfide oxidation is only 1/3 of that at Wiluna (Briggs, 1996). Cyanide leaching of the oxidized concentrate is successful at recovering 90 to 95% of the gold. Ashanti The fifth, and by far the largest, BIOX plant was commissioned in 1994 at the Ashanti mine in Ghana. Bacterial oxidation was selected over other options such as the Nitrox process, roasting, or pressure oxidation. The BIOX process was found to offer the highest internal rate of return, required a reasonable capital investment, was well suited to the variances in mineralogy of the feed, and provided ease of operation and maintenance in a remote location (Nicholson et al., 1994). The circuit processes a flotation concentrate that grades 75.9 g/t Au, 11.8% sulfur, and 9.4% . arsenic. The principal sulfide constituents are arsenopyrite, pyrrhotite, and pyrite. In addition, a semi-oxidized concentrate will be produced during the first five years of mine operation consisting of weathered surface material. This concentrate contains more pyrite, but less 51 arsenopyrite and no pyrrhotite. The bioleaching circuit consists of four separate modules run in parallel. Each module has six 900 m3 tanks. The first three tanks (primary stage) are operated in parallel, whereas the last three (secondary stage) are operated in series. Each stage has a retention time of two days. The current throughput of the plant is 1100 tpd. 2.4 Biooxidation and Cyanide Consumption With the exception of Sao Bento, which uses a pressure oxidation finishing step, all of the operations described here involve cyanidation of a biooxidized product. In each case consumption of cyanide is excessive when compared to the products of pressure oxidation. Table 2-2 illustrates the cyanide consumption rates found in these plants. The figures indicate that consumption rates are many times higher when bacterial oxidation is used. Table 2-2. Cyanide consumption following bacterial oxidation pre-treatment. Data for Fairview is not presently available. Plant (commissioned) Process Capacity L I P l ) | " NaCN consumption [Kg tj Fairview (1986) BIOX 40 NA Wiluna(1993) BIOX 115 30 Youanmi (1994) BacTech 90 7.5 Ashanti (1994) BIOX 720 15 pressure oxidation - - <2.0 52 After the Fairview plant was constructed, Haines and van Aswegen (1990) suggested that the high cyanide consumption experienced was the result of the incomplete oxidation of the sulfide mineral. This led to the presence of elemental sulfur in the biooxidized product. A number of techniques were reportedly being investigated to rectify this problem, but no significant reduction in cyanide consumption has been realized for any of the plants built subsequently. Furthermore, cyanide consumption at the Youanmi mine is fifteen times greater for the bioleached flotation concentrate than it is for the flotation tails (Brierley and Winby, 1996). Komnitsas and Pooley (1989) have investigated the cyanide consumption of bacterially oxidized refractory gold concentrate from Olympias, Greece. The authors used air-stirred pachuca reactors to leach the pyrite/arsenopyrite mineral. Cyanidation tests were carried out at various pulp densities and initial cyanide concentrations. The results indicated that cyanide consumption was higher at low pulp densities. Indeed, at a density of 5% virtually all the cyanide added was consumed by the slurry; to a maximum of over 156 kg/t. Conversely, at a pulp density of 30% the amount of NaCN consumed was greatly reduced. However, gold recoveries also declined: from ~85%> to less than 62%. Both the lower recoveries and the lower rates of reagent consumption were likely to be the result of mass transfer limitations of the pachuca reactors at higher pulp densities. Specifically, the availability of oxygen may not have been adequate in all trials. Similar tests were conducted by the authors to correlate the gold recovery and cyanide 53 consumption to the percentage of sulfide mineral oxidized. In all cases, the pulp density was 5% and the duration of the test was 18 hours. Gold and silver recoveries increased with the percentage of mineral oxidized, and with the amount of cyanide added. Also, the consumption of cyanide increased with the percentage of mineral oxidized; with virtually all being consumed when the arsenopyrite oxidation reached -85%. While this investigation sheds some light on the factors influencing cyanide consumption by biooxidized residues, the work does not discuss any specific mechanisms involved. Komnitsas and Pooley suggested that high consumption is related to the formation and precipitation of iron and silver jarosites and iron arsenates, particularly at low pulp densities. However, no data is offered to support this conclusion. In addition, no consideration is given to the interaction between cyanide and the sulfide mineral, or any intermediate sulfur compounds formed during the leach. More recent work by Lawson (1997) has focused attention upon the role of the enzyme rhodanese in cyanide consumption. As discussed previously (equation 2-30), rhodanese can catalyze the reaction of cyanide and thiosulfate to form thiocyanate. The author suggests that inactivation of this enzyme will result in reduced reagent consumption. Limited success was achieved by boiling the slurry prior to cyanidation. This pre-treatment resulted in a 20% decrease in thiocyanate formation and a 33% decrease in cyanide consumption. However, the use of the chemical inhibitors sodium sulfite and iodoacetamide were found not to have appreciable influence. Thus, thiocyanate formation might not be catalyzed by rhodanese, and the effect of boiling may then be explained by changes to sulfur intermediates or other compounds 54 in the residue. Further effort has been made to link cyanide consumption to the presence of sulfur species in the feed to cyanidation (Schrader and Su, 1997; Hackl and Jones, 1997). By the reaction of elemental sulfur with cyanide shown in equation 2-31, a 1% sulfur content can be calculated to yield a maximum sodium cyanide consumption of 15.4 kg/tonne. Schrader and Su (1997) propose that consumption of cyanide increases with S° content. However, this conclusion is based solely on the initial concentration of elemental sulfur and the amount of cyanide consumed during the leach. No mention is made of how much sulfur remains in the residue, or what portion of the cyanide is consumed as thiocyanate. The work of Hackl and Jones (1997) attempted to balance the loss of cyanide with the formation of SCN" as well as complexes of copper, iron, and zinc. These species accounted for anywhere between 37% and 59% of the total cyanide consumed by the oxidized concentrate; with the unaccounted for fraction possibly being lost from the system as HCN gas. Thiocyanate was responsible for consuming between 9% and 62% of the cyanide. However, assays of the leached residues indicated that in three of the four tests elemental sulfur alone could not be responsible for SCN" formation. The authors also propose an aerated lime pre-leach as a method to reduce cyanide consumption. They speculate that elemental sulfur reacts with lime according to the reaction: (2 + 2x)S + 3Ca(OH)2 -> 2CaSx + CaS2 <93 + 3H2 O (2-32) 55 where, for dilute polysulfide solutions, x will be equal to 1. The treatment resulted in a 68% reduction in SCN" formation. Interestingly, the pre-leach also resulted in a 74% decrease in the formation of hexacyanoferrate. A similar treatment has been proposed by Xiang et al. (1994). Here, the solids are heated to 80-90 °C for one hour in a sodium hydroxide solution. In this case, a 50% reduction in cyanide consumption was realized. 2.5 Summary Clearly, the problem of high cyanide consumption represents a significant disadvantage in biooxidizing refractory gold concentrates, both from an economic and a process control perspective. Furthermore, the uncertainty surrounding the specific pathways and intermediate products associated with bacterial sulfide dissolution contribute to what Lawrence (1993) refers to as the "perceived technical risk" of biological pre-treatment. As a result, despite successful pilot plant trials, biooxidation has yet to gain acceptance among North American operators. To date, published research in the area of cyanide consumption by biooxidized solids has failed to identify the specific causes of the problem. Bioleaching tests done in shake flasks or batch reactors cannot accurately reflect conditions found in actual practice. Furthermore, little work has been done to identify specifically where the cyanide goes. Finally, a better understanding of the influence of cyanidation parameters is also required. A detailed description of the experimental setup designed to address these concerns now follows. 56 3. Methods This section describes the experimental methods and procedures followed during the production of the concentrate, the operation of the continuous reactor system, and the cyanidation tests. 3.1 The Concentrate A pilot plant run was conducted at Placer Dome's research centre in Vancouver, in order to produce a rougher flotation concentrate to serve as feed for bioleaching. A schematic of the pilot flotation circuit is shown in Figure 3-1. Feed to the circuit consisted of a low grade refractory arsenical gold ore, crushed to -1/4". 57 crushed ore CL" water oversize material back to grinding ball mill #1 underflow frother and collector tailings tic )tatic n ce lis overflow ball mill #2 rougher flotation concentrate Figure 3-1. Schematic diagram of the pilot flotation circuit. The dry ore was repulped to a density of 50% before entering the first ball mill. A spiral classifier was employed to return the oversize material to the mill. Frother, collector, and a promoter (to float any free gold) were added to the fine fraction and the slurry was sent to the flotation tanks via a sala pump. A summary of the reagents used is found in Table 3-1. Screen analysis of the feed to the flotation circuit revealed a P80 (80% passing size) of 100 microns. 58 Table 3-1. Reagents used in the production of the concentrate. Function Reagent Concentration [g/t-dry ore] frother Dowfroth 250 36 collector potassium amyl zanthate 120 (KAX51.KAX55) promoter Aerofloat 208 6 The flotation setup consisted of 7 cells in series. Froth collected from each cell was sent to the second ball mill for regrinding. Conversely, the underflow went to tailings. Since no scavenger flotation circuit was employed, the final product is said to be a rougher flotation concentrate. Results of the analysis of the concentrate are shown in Table 3-2. Assuming that the sulfide minerals consist exclusively of FeS2 and FeAsS, the feed grades 25.3% pyrite and 6.2% arsenopyrite. The concentrate is refractory as cyanidation of the untreated material resulted in only 15%) gold recovery. Table 3-2. Results of chemical analysis of the concentrate. Au [g/t] Ag [g/t] Cu [%] Fex [%] As [%] 20.25 8.7 0.021 13.27 2.87 S T [%] s2- [%] C x [%] Qnorg [ 0 / ° ] acid insol. [%] 13.36 13.06 6.01 0.85 60.57 59 Approximately 80 kg (dry weight basis) of the concentrate was shipped to UBC to serve as feed for the biooxidation tests. The pulp density of this material as received was 43.6%. Microtrac particle size analysis revealed the concentrate had a P80 of 30 microns. 3.2 Bacterial Leaching Experiments Continuous bioleaching df the concentrate was accomplished using a 4-stage system. A mixed stock culture of Thiobacillus ferrooxidans, Thiobacillus thiooxidans, and Leptospirillum ferrooxidans was chosen as inoculum for the leaching experiments. The culture had already been growing on a pyrite/arsenopyrite ore in the Biohydrometallurgy Lab at UBC. A 200 ml sample was used to inoculate a 2 litre tank containing OK media (Table 3-3) and the concentrate at 20% pulp density. After 8 days, the contents of this tank served as inoculum for the 4-stage system. A further three days of batch mode operation were required before continuous feeding commenced on August 5th. The system consisted of a 4 litre primary stage followed by three, 2 litre tanks in series (see Figure 3-2). Thus the total working volume of the system was 10 litres. Each tank was kept under constant agitation and aeration to ensure that the dissolved oxygen concentration remained above 1 ppm. In addition, the sparging air was enriched with C02 to yield a 1% C02-air mix to provide sufficient carbon for rapid bacterial growth. 60 pH Control Figure 3-2. Schematic of the continuous reactor. The concentrate was repulped to a density of 20% and nutrients were added. For the first few weeks these nutrients consisted of OK media. However, this addition was later reduced to that recommended for the MINBAC process. Table 3-3 lists the composition of each medium. 61 Table 3-3. Nutrient media used during continuous operation. All figures in kg/tonne of concentrate, (from Silverman and Lundgren,1959; Pinches et al., 1994) Compound OK MINBAC (NH4)2S04 12.00 4.28 KH2P04 2.00 0.47 MgSCy7H20 2.00 -KC1 0.40 -Ca(N03)2-4H20 0.04 -The temperature within each tank was maintained at 35°C by a 50 Watt immersion heater. The heaters were controlled manually and adjusted each day, as necessary. The primary stage was operated under pH control using a Cole Parmer pH/ORP controller (model #5652-10). Under the conditions tested, the concentrate proved to be acid producing. As a result, the controller activated a peristaltic pump that added a 20% lime slurry to the reactor. In this manner, the pH was maintained within 0.1 of the set point. The pH controller was calibrated each day using pH buffers 7 and 2. For the first few weeks the base used consisted of calcium hydroxide (Ca(OH)2). However, this was later replaced by limestone (CaC03), without significant difficulty. The pH in the other three tanks was monitored, but not controlled. The tanks were connected in cascade fashion, as shown in Figure 3-2. The working volume of each tank was taken to require that the slurry level be maintained at the overflow height. However, transfer solely by overflow was found to result in solids build up (Weston, 1996). Furthermore, since the flow rate between tanks was relatively small -80 to 165 ml per hour— 62 continuous transfer was not deemed practical. Instead, a timer was used to activate four separate peristaltic pumps for 1 minute out of every 30 minutes. Each day, the pH, redox potential, and temperature of each stage were recorded to provide an indication of the "steadiness" of the system. In addition, flow rates into and out of the reactor were monitored. Periodically, slurry samples were taken from each tank to determine pulp density as well as iron and arsenic extraction. When the system reached steady state, a mass balance period was initiated. During this time the product was collected in a separate container. Slurry samples were again taken to determine pulp density and metal extraction. The solid residues of these samples were sent to International Plasma Laboratories (IPL) of Vancouver to be analyzed for iron, arsenic, sulfide, and total sulfur. Analytical procedures are described in Appendix A. At the completion of the run, large slurry samples (1.0L) were taken from the intermediate stages of the leach. These samples, along with the product collected during the run, were filtered, washed, and repulped to serve as feed for the cyanidation experiments. 3.3 C y a n i d a t i o n Batch cyanidation experiments were carried out in order to study the cyanidation behavior of the biooxidized solid products. Figure 3-3 illustrates the setup used to conduct these tests. 63 pH Probe Rushton Turbine LaJ Temperature Probe Pyrex Beaker Figure 3-3. Schematic of the cyanidation reactor. 3.3.1 Standard Cyanidation Tests A standardized test was employed in order to compare the amenability of the various intermediate and final products of bioleaching to cyanidation. For each test, a product sample with a dry solids weight of 100 grams was re-pulped with de-ionized water to a pulp density of 30%. The slurry was then cyanide leached for 24 hours under vigorous agitation. All of the experiments were conducted in a standard 600 mL Pyrex beaker (Figure 3-3). Prior to cyanidation, the pH of the slurry was raised (to a range of 10.4-11.0) by the addition of calcium hydroxide. This pH level was maintained throughout the leach to prevent the loss of 64 cyanide as HCN gas. The leach tests were carried out under ambient temperature, which ranged from 20.4 to 22.6 °C. For all experiments, the initial concentration of sodium cyanide was 0.5 g/L. The concentration of free cyanide was measured at 1, 3, 6, and 12 hours, and additional cyanide was added to bring the concentration of NaCN back to 0.5 g/L. This determination was accomplished by titration with silver nitrate (Harris, 1991). In this case, a 20 to 30 mL slurry sample was taken from the reactor and centrifuged. A 10 mL sample of the supernatant was subsequently titrated using 0.05M AgN03. The remaining solution and solids were returned to the system along with 10 mL of deionized water. It is important to note that for all tests the concentration of cyanide in the leach declined significantly between the sampling at 12 hours and the termination of the test, because the test was not monitored during this period. After 24 hours the leached slurry was filtered. The filtrate was analyzed for free cyanide and thiocyanate. The latter determination consisted of acidifying the samples and sparging them with nitrogen gas to drive off the free cyanide followed by silver nitrate titration with ferric nitrate as the indicator. Finally, a solution sample was sent to IPL for gold analysis by fire assay. The filtered solids were washed, air dried, and also sent to IPL for gold analysis by fire assay. 65 3.3.2 Cyanide Consumption Tests In order to determine the parameters influencing cyanide consumption, various modifications were made to the test setup previously described. For example, to determine the influence of cyanide concentration tests were run at 0.25, 0.5, and 1.0 g/L NaCN. Similar adjustments were made to evaluate temperature, pulp density, and stirring speed. To better understand the role of pH, an alternate reactor was used (see Figure 3-4). In this case, cyanidation tests were conducted in a sealed system to prevent the loss of cyanide as HCN gas during runs where the pH ranged between 4 and 10. Furthermore, the headspace was filled with oxygen gas in an attempt to ensure that the concentration of oxygen was not a limiting factor. Gas in Eh Probe Sealed Lid Stir Bar Gas Out Sample Tube PH Probe Pyrex Flask Stir Plate Figure 3-4. Schematic of the sealed cyanidation reactor. 66 In later runs this setup was modified to permit the sparging of 02 gas from the headspace of the reactor back into the slurry. This was accomplished by adding a stainless steel tube with a rubber sparge ring suspended below the slurry surface. Finally, this setup was also used to determine the loss of cyanide gas (HCN) and ammonia (NH3) from the leach solution. In this case, air was pumped into the sealed headspace or beneath the slurry surface and the off-gas sparged through a dilute NaOH (for HCN determination) or H2S04 (for NH3) solution. 67 4. Results and Discussion The results of the experimental program are now presented and discussed. The first two sections summarize the data collected during the bioleaching and cyanidation tests. Attempts to balance cyanide inputs and outputs are discussed in Section 4.3, while Section 4.4 examines the influence of leach parameters on cyanide consumption. A fifth section looks at several pre-treatment trials, and the last section looks at the role of sulfur in cyanide consumption. 4.1 Bioleaching Results The four-stage reactor system was started in batch mode on July 30th, 1997. The reactor was inoculated with a mixed culture from the biohydrometallurgy lab at UBC. The culture was thought to contain the chemolithotrophs Thiobacillus ferrooxidans, ThiobacUlus thiooxidans, and Leptospirtilum ferrooxidans. Initial commissioning required the addition of 6M H2S04 to each reactor in order to lower the slurry pH to 1.8. After six days, the pH and Eh measurements indicated that the bacterial culture had become established. Continuous operation commenced on August 5th, 1998. Figure 4-1 summarizes the Eh and pH data collected during the 117 days that the system was run. 68 3 5 residence time (days) 4~ 3.75 Figure 4-1. Eh and pH measurements for the four-stage reactor system. Upper lines represent Eh; lower represent pH. SSI = steady state #1; T l = tank #1. Overall reactor residence time is indicated by the scale at the bottom. The pH in Tank #1 was controlled at 1.8 until day 92, when the pH was lowered to 1.5. During batch mode operation the Eh in all tanks rose rapidly to levels above 800 mV (SHE). As expected, the Eh increased consistently between stages. Conversely, pH showed a decrease between stages. The pH in the first tank was controlled at about 1.8 until day 92 when this parameter was lowered to 1.5. As the chart indicates, pH control was not always successful. In particular, the addition of lime as either calcium hydroxide or limestone proved problematic. The reagent became clogged in the input lines on several occasions resulting in a pH well below the set point. In contrast, at times the pH exceeded the target range; possibly due to drift in the calibration of the monitoring probe or precipitates forming on the surface of the electrode. Steps 69 taken to rectify these concerns included daily flushing of the reagent lines with deionized water, rinsing the pH electrode with a weak HC1 solution, and frequent re-calibration of the probe. The four mass balance periods achieved during the run are indicated in Figure 4-1. These periods were determined based on the apparent steadiness of the system in the three days preceding them. The parameters used to make this decision included pH, Eh, pulp density, temperature, and the flow rate between stages. Steady State #1 ended prematurely as a result of two separate disruptions: (1) excessive lime addition resulted in high pH's, and, subsequently, a significant drop in Eh, and (2) a mercury thermometer broke in the fourth stage. Since, mercury is toxic to bacteria and also poses a safety concern in the lab, the decision was made to terminate the mass balance period and remove the fourth stage from the system. After thorough cleaning, the tank was refilled with fresh slurry and inoculated. The second mass balance period was the longest of the four, lasting for 10 days. Halfway through this period, the fourth tank was permanently removed from the system. This was done to reduce the overall residence time of the system without shortening the residence time of the primary stage. The remaining two steady states used only three tanks in series. Steady State #3 was the shortest mass balance period, lasting only 72 hours. Also, large slurry samples (used for the cyanidation study) were not taken from the tanks, as it was thought that this might delay reaching the fourth steady state. As figure 4-1 indicates, the residence time had 70 decreased dramatically by that point, so, despite their shorter times, large product samples were generated by the last two steady states. Finally, at the completion of the fourth steady state, the continuous run was terminated. Figures 4-2 and 4-3 illustrate the change in iron and arsenic concentration between stages over time. Iron concentrations are consistently higher in successive tanks, as expected. The arsenic results are less clear than for iron, possibly due to difficulties encountered during analysis by Atomic Absorption. Furthermore, the apparent inconsistency between tanks 1 and 2 on day 36 (see Figure 4-3) is likely due to a sample mix up. "Si 20.0 c _o S 15.0 s U c g 10.0 5.0 65 75 85 Time [days] Figure 4-2. Change in iron concentration over the continuous run. T l = tank #1. Numbers at top indicate mass balance periods. Tank #4 was removed from the system on day 85. 71 8.0 Time [days] Figure 4-3. Change in arsenic concentration over the continuous run. T l = tank #1. Numbers at top indicate mass balance periods. Tank #4 was removed from the system on day 85. In general, a trend can be seen in both arsenic and iron towards higher extraction as the continuous run progressed (although for iron this result begins to level off after about day 97). This is particularly interesting as the overall residence time was shortened considerably over the same period. Possible explanations for this result include the switch from lime to limestone on day 47 and the lowering of the primary stage pH on day 92. Both of these changes may have influenced the precipitation of basic ferric sulfates or ferric arsenate. The results of the four steady states are now summarized. These results represent average values for the measurements taken during the mass balance period. Temperature, Eh, and pH data were collected daily, as stated in the methods section. However, the calculations for sulfide oxidation, and iron and arsenic extraction, were based on only two measurements during Steady States #1, 72 #2, and #4 due to the expense of analysis. Steady State #3 involved only one such measurement by operator oversight. The results of this mass balance period are included in this report, but cannot be considered as reliable as the others. Appendix E contains a statistical interpretation of the data collected. 4.1.1 Steady State #1 The first steady state was reached on September 9th. Operating parameters during this period are listed in Table 4-1. Over the course of the run, pH, Eh, and temperature were measured each day. In addition, slurry samples were collected at 4 and 72 hours. These samples were filtered and analyzed to determine iron and arsenic extraction, and sulfide oxidation. The results of this analysis are summarized in Table 4-2. Cyanidation experiments were conducted on the final product as well as on large samples collected from the intermediate stages. The operating conditions and results of these runs are shown in Table 4-3. It should be noted, however, that actual gold extractions are slightly higher, as some of the gold was lost to intermediate sampling for free cyanide. 73 Table 4-1. Operating conditions at Steady State #1. Controlled Parameters Stage 1 Residence Time 2 days Total Residence Time 5 days Stage 1 target pH 1.8 pH control by Ca(OH)2 Nutrient addition OK Temperature 35 °C Feed Pulp Density 20% Run Time 96 hours Table 4-2. Bioleach data summary for Steady State #1. Measured parameters (averaged) Stage 1 Stage 2 Stage 3 Stage 4 pH 1.81 1.32 1.17 1.07 Eh [mV] 785 845 871 893 temperature [°C] 35.4 35.9 35.5 35.5 pulp density (%) 23.65 21.56 20.17 18.81 Ca(OH)2 cons.[kg/t-conc] 103 - - -Analytical results (averaged) Stage 1 Stage 2 Stage 3 Stage 4 sulfide oxidation [%] 61.9 76.8 86.7 93.2 iron extraction [%] 6.1 32.9 52.3 61.2 arsenic extraction [%] 4.7 41.9 75.7 82.6 residue analysis: Fe [%] 9.21 7.13 5.05 4.38 As [%] 2.00 1.25 0.60 0.53 S° [%] 0.23 0.29 0.29 0.29 S2' [%] 4.21 2.80 1.73 0.93 SS04 [%] 5.15 5.21 5.18 5.17 ST [%] 9.76 8.61 7.52 6.64 74 Table 4-3. Cyanide leach data for Steady State #1. Cyanidation parameters pulp density 30% NaCN cone* 0.5 g/L temperature 20.4-22.6 °C leach time 24 hours Cyanidation results Stage 1 Stage 2 Stage 3 Product Au extraction [%] 91.9 88.9 91.2 93.8 NaCN consumed [kg/t-res.] 3.56 2.90 3.01 3.76 NaCN cons.[kg/t-conc] 4.03 3.00 2.91 3.31 Ca(OH)2 consumed [kg/t-conc] 65.8 44.7 32.3 17.1 * concentration maintained for first 12 hours, then a lowed to decline 4.1.2 Steady State # 2 The second mass balance period commenced on October 24th, 1997. The operating conditions selected in this case are listed in Table 4-4. Changes in process consisted of a shorter residence time (3.75 days), a reduction in the nutrients used, and the replacement of calcium hydroxide with calcium carbonate (limestone). Steady State #2 lasted for 240 hours. However, after 120 hours the 4th stage was removed from the system, and product was collected from the third reactor. Data collection was the same as in Steady State #1 with the exception that slurry samples were taken at 4, 76, 146, and 240 hours, and solid samples were not assayed for S° and Ss04 in an effort to reduce analytical costs. The results are summarized in Table 4-5. Again, cyanide leach tests were conducted on the washed solids. Results are shown in Table 4-6. 75 Table 4-4. Operating conditions at Steady State #2. Controlled Parameters Stage 1 Residence Time 1.5 days Total Residence Time 3.75 days Stage 1 target pH 1.8 pH control by CaC03 Nutrient addition MinBac Temperature 35 °C Feed Pulp Density 20% Run Time 240 hours Table 4-5. Bioleach data summary for Steady State #2. Measured parameters (averaged) Stage 1 Stage 2 Stage 3 Stage 4 pH 1.79 1.35 1.17 1.07 Eh [mV] 793 828 859 879 temperature [°C] 35.4 35.1 34.8 34.5 pulp density (%) 19.64 17.60 16.91 16.58 CaC03 cons, [kg/t-conc] 54.8 - - -Analytical results (averaged) Stage 1 Stage 2 Stage 3 Stage 4 sulfide oxidation [%] 44.4 62.8 74.2 77.8 iron extraction [%] 24.7 55.5 69.0 76.4 arsenic extraction [%] 30.9 79.9 88.7 90.4 residue analysis: Fe [%] 9.44 6.18 4.64 3.68 As [%] 1.78 0.60 0.41 0.35 s2- [%] 6.97 5.21 3.78 3.27 ST [%] 10.44 8.41 7.20 5.51 76 Table 4-6. Cyanide leach data for Steady State #2. Cyanidation parameters pulp density 30% NaCN cone* 0.5 g/L temperature 20.9-21.9 °C leach time 24 hours Cyanidation results Stage 1 Stage 2 Stage 3 Product Au extraction [%] 87.3 90.7 89.4 89.4 NaCN cons.[kg/t-res] 4.26 4.27 4.45 4.28 NaCN cons.[kg/t-conc] 4.15 3.73 3.74 3.52 Ca(OH)2 consumed [kg/t-conc] 36.4 23.8 8.3 9.7 * concentration maintained for first 12 hours, then a lowed to decline 4.1.3 Steady State #3 After the second steady state, the pH was lowered to 1.5 in the first stage. Summaries of the operating conditions and results are found in Tables 4-7 and 4-8. During this steady state no intermediate samples were taken, thus cyanidation data is available only for the product. The results are summarized in Table 4-9. Table 4-7. Operating conditions at Steady State #3. Controlled Parameters Stage 1 Residence Time 1.5 days Total Residence Time 3.0 days Stage 1 target pH 1.5 pH control by CaC03 Nutrient addition MinBac Temperature 35 °C Feed Pulp Density 20% Run Time 72 hours 77 Table 4-8. Bioleach data summary for Steady State #3. Measured parameters (averaged) Stage 1 Stage 2 Stage 3 pH 1.48 1.26 1.11 Eh [mV] 799 836 864 temperature [°C] 34.7 34.7 34.6 pulp density (%) 17.86 17.30 16.32 CaC03 cons, [kg/t-conc] 43.7 - -Analytical results Stage 1 Stage 2 Stage 3 sulfide oxidation [%] 50.7 66.9 79.1 iron extraction [%] 49.2 64.0 80.5 arsenic extraction [%] 69.3 87.4 91.1 residue analysis: Fe [%] 7.04 5.30 2.89 As [%] 2.47 0.80 0.42 s2- [%] 6.47 4.43 2.97 ST [%] 9.05 7.51 6.03 Table 4-9. Cyanide leach data for Steady State #3. Cyanidation parameters pulp density 30% NaCN cone* 0.5 g/L temperature 21.4 °C leach time 24 hours Cyanidation results Product Au extraction [%] NaCN cons.[kg/t-res] NaCN cons. [kg/t-conc. ] 90.7 4.26 3.54 Ca(OH)2 consumed [kg/t-conc] 21.70 * concentration maintained for first 12 hours, then allowed to decline 78 4.1.4 Steady State #4 The fourth, and final, steady state period commenced on November 19th, 1997, and ran for 120 hours. In this case, the residence time was lowered to 2 days. Parameters and results are summarized in Tables 4-10 and 4-11. Cyanidation experiments were conducted on the final product collected, as well as the intermediate stages. Results are summarized in Table 4-12. Table 4-10. Operating conditions at Steady State #4. Controlled Parameters Stage 1 Residence Time 1 day Total Residence Time 2 days Stage 1 target pH 1.5 pH control by CaC03 Nutrient addition MinBac Temperature 35 °C Feed Pulp Density 20% Run Time 120 hours Table 4-11. Bioleach data summary for Steady State #4. Measured parameters (averaged) Stage 1 Stage 2 Stage 3 1.53 1.16 1.01 Eh [mV] 783 819 847 temperature [°C] 35.0 35.4 35.4 pulp density (%) 18.44 17.47 16.45 CaC03 cons, [kg/t-conc] 30.8 - -Analytical results (averaged) Stage 1 Stage 2 Stage 3 sulfide oxidation [%] 43.0 58.6 70.6 iron extraction [%] 34.7 53.3 66.2 arsenic extraction [%] 55.6 80.2 89.1 residue analysis: Fe [%] 9.20 6.87 5.13 As [%] 1.26 0.60 0.43 s2- [%] 8.37 6.36 4.84 ST [%] 10.27 8.62 7.18 79 Table 4-12. Cyanide leach data for Steady State #4. Cyanidation parameters pulp density 30% NaCN cone* 0.5 g/L temperature 20.6-20.2 °C leach time 24 hours Cyanidation results Stage 1 Stage 2 Product Au extraction [%] 87.4 83.1 91.0 NaCN cons.[kg/t-res] 3.16 3.14 4.32 NaCN cons.[kg/t-conc] 2.87** 2.70** 3.50 Ca(OH)2 consumed [kg/t-conc] 73.04 70.7 69.6 * concentration maintained for first 12 hours, then a lowed to decline **pH in these tests exceeded 11.0, possibly reducing cyanide consumption and gold extraction 4.2 Analysis of the Four Mass Balance Periods From the results presented in the previous section, several important observations can be made. The pulp density during Steady State #1 actually increased in the primary stage (see Figure 4-4). Initially this result was thought to be due to a solids build up problem caused by plugging of the transfer line between stages 1 and 2. However, lab analysis indicated that the concentration of acid insoluble material (i.e. inert gangue material) decreased during this stage. Thus, there was a net weight gain, which can only be attributed to the formation of an acid soluble precipitate. Since calcium hydroxide was added to this stage, and analysis showed more than 5% sulfate sulfur in the solids, this precipitate was thought to include gypsum. In addition, much of the leached iron precipitated, likely as basic ferric sulfates, as iron extractions were found to be 80 consistently lower than sulfide oxidation. Interestingly, this increase in pulp density was not as pronounced in the second mass balance period, probably because of the switch from lime to limestone after the first steady state, and not observed at all in the 3rd and 4th steady states where the pH in the first stage was lower. Thus, the precipitation of gypsum and basic ferric sulfate is greatly affected by changes in pH, and the use of lime apparently resulted in high localized pH's which intensified this problem. 25 T a •a 3 10 -5 + o 4 0 1 2 3 4 Tank # (0 = feed) Figure 4-4. Change in pulp density between stages for the four steady states. SS #1 SS#2 SS#3 •SS#4 A further comparison can be made between the steady state periods by looking at the extent of sulfide oxidation, and iron and arsenic extraction, in Figure 4-5. 81 100 -r Tank # ( 0 = feed) Figure 4-5. Comparison of sulfide oxidation, and iron and arsenic extraction, for each steady state. j 82 As expected, the 5 day residence time associated with the first mass balance period resulted in a sulfide oxidation exceeding 90%. This number decreased in subsequent steady states as the residence time was lowered. Indeed, the sulfide oxidation for steady state #4 was just over 70%. However, iron and arsenic extractions appear to have actually increased with lower residence times. In particular, the third steady state shows the highest numbers, possibly because it employs the longer of the two residence times at the lower initial pH. Higher pH in the first two stages of the system may have resulted in the hydrolysis and precipitation of Fe(III) and As(V). In addition, the selective leaching of arsenopyrite over pyrite is also evident. The third mass balance period resulted in 91.1% extraction of arsenic while only oxidizing 79.1% of the sulfides. Interestingly, the fourth steady state yielded a sulfide oxidation of only 9% less than that obtained in the third, despite a reduction in residence time of 33% (3 days to 2). Cyanidation experiments conducted on the biooxidized product were successful in achieving high gold extractions. Results ranged from a low of 83.1%, for stage 2 of Steady State #4, to 93.8%) for the final product of Steady State #1. The relationship between gold extraction and sulfide oxidation is shown in Figure 4-6. In this respect, very little difference between stages was observed. However, cyanidation of the untreated feed material liberated only 15.3% of the gold, indicating the enormous influence of even 43% sulfide oxidation. 83 10 20 30 40 50 60 sulphide oxidation [%] 70 80 90 100 Figure 4-6. Gold extraction vs. sulfide oxidation. All tests under standard conditions: ambient temperature; 30% pulp density; pH 10.4-11.0; [NaCN]= 0. g/L. Note: actual extractions are slightly higher as some gold was lost to intermediate solution sampling at 1, 3,6, and 12 hours. The consumption of cyanide during the leach ranged from 2.5 to 4.2 kg per tonne of feed material to the bioleaching system. Figure 4-7 plots cyanide consumption against sulfide oxidation. 84 z u • 2-1 2-3 • 2 - 2 % * 3-P • • 2-P 1-P • 1-1 • - 0 V l • 4-2 4 • 1-3 1 ! 1 1 ! 1-2 1 1 10 20 30 40 50 60 sulphide oxidation [%] 70 80 90 100 Figure 4-7. Cyanide consumption as a function of sulfide oxidation. Numbers indicate steady state - stage number (P = product). All tests under standard conditions: ambient temperature; 30% pulp density; pH 10.4-11.0; [NaCN]= 0.5 g/L. From the graph no clear trends are evident. Cyanide consumption appears to be relatively similar between stages of the same steady state, with the possible exception of 4-P. Thus, the influence of the primary stage (of biooxidation) pH or the extent of sulfide oxidation are not apparent. The cyanidation of the feed material was shown to generate a similar cyanide consumption as the bioleached residues. This fact brings into question whether the bioleaching step has any influence at all on the amount of reagent consumed. However, analysis of the leach solutions indicates that thiocyanate (SCN") accounts for only 4% of the cyanide consumed by the feed material, whereas this figure reaches as high as 77% for the bioleached solids. Therefore, although the level of consumption may be comparable, the mechanisms involved are different. 85 Figure 4-8 illustrates the relationship between the percentage of total cyanide consumed to form thiocyanate and sulfide oxidation. 0 10 20 30 40 50 60 70 80 90 100 sulfide oxidation [%| Figure 4-8. The percentage of NaCN consumed as SCN" plotted against sulfide oxidation. The line separates the results of steady states 1 and 2 (primary stage pH=1.8) from those obtained from steady states 3 and 4 (primary stage pH=1.5). Here a trend is more clearly evident. Thiocyanate accounts for a greater fraction of the cyanide consumed when more of the sulfide is oxidized. In addition, the lower pH associated with the third and fourth steady states has resulted in a noticeable increase in the percentage of reagent consumed in this manner. Residue assay of the composite product from Steady State #4 indicated an elemental sulfur content of 0.64%: double that found in the product of the first mass balance period. However, Table 4-2 indicates that S° formed in the early stages of the leach and then remained constant, while SCN formation during cyanidation of these residues increases. 86 4.2.1 Lime Consumption The consumption of lime varied greatly between the four mass balance periods. Table 4-13 compares the results by converting limestone to the equivalent weight of calcium hydroxide. Table 4-13. Lime consumption for each of the four steady states. Note that the first mass balance period used Ca(OH)2 while the others used CaC03. All figures in kilograms of reagent per tonne of concentrate feed. Steady State # Reagent Used Mass Consumed [kg/t] Equivalent Mass of Ca(OH)2 [kg/t] 1 Ca(OH)2 103 103 2 CaC03 54.8 40.6 3 CaC03 43.7 32.3 4 CaC03 30.8 22.8 The amount of lime added decreased considerably between the first and second steady states. While the lower residence time accounts for some of this difference, precipitation at the surface of the calcium hydroxide may have rendered some of the reagent unusable. Steady State #3 had the same residence time as the previous mass balance period, but a lower primary stage pH resulted in lower lime consumption. Finally, the fourth steady state employed this lower pH and had the shortest residence time (and the lowest sulfide oxidation extent), resulting in even lower lime requirement. Lime consumption was also monitored during the cyanidation step. In this case, calcium hydroxide was added to the repulped solids in order to bring the pH to between 10.4 and 11.0. 87 Reagent consumption ranged from a low of 4.40 kg/t for the unoxidized feed material, to 73.1 kg/t for the first stage of the fourth steady state. Figure 4-9 illustrates the relationship between sulfide oxidation and lime consumption. 10 20 30 40 50 60 sulfide oxidation [%] 70 80 90 100 Figure 4-9. L ime consumption as a function of sulfide oxidation for the cyanidation experiments. C a ( O H ) 2 was added to raise the p H of the slurry to between 10.4 and 11.0. While the graph appears to indicate lower lime consumption with more complete oxidation, the effect may also be due to leach tank pH. That is, cyanidation of tank 1 samples show higher consumption because of precipitation of basic ferric sulfates at higher pH. These precipitates may consume lime according to the reactions, H30Fei(S04)2(OH)6 + 40H ->3FeOOH+2SOk~ +5H2O ( 4 _ 1 ) H3OF&(SO*)a (OH)e + 40H -> 3Fe(OH)s + 2SoT + IHiO ( 4 " 2 ) 88 Subsequently, the dissolution of basic ferric sulfates at the lower pH's found in the downstream tanks would explain the lower lime requirements associated with these residues. 4.2.2 Leaching Kinetics The kinetics of gold dissolution during the cyanidation step were investigated using a product sample from steady state #4 (test conditions: ambient temperature, 24 hr leach, 30% pulp density, 0.5 g/L NaCN, pH = 10.4-11.0). In this case, intermediate solution samples were taken at 1, 3, 6, and 12 hours and analyzed for gold, free cyanide, thiocyanate, and ICP-metals. Gold extraction exceeded 92%, while the solids had a sodium cyanide consumption of 3.36 kg/tonne-concentrate. Figure 4-10 illustrates the extraction curve for the run. 100 T - , time [hours] Figure 4-10. Gold extraction vs time for cyanide leach of steady state 4 product. Ambient temperature, 24 hr leach, 30% pulp density, 0.5 g/L NaCN, pH = 10.4-11.0. 89 The graph clearly indicates the rapid dissolution of the gold particles under the conditions tested. Indeed, an extraction of over 93% is achieved after only 6 hours. Thus the liberated gold particles must have been quite fine. The slight decline in extraction at 24 hours is thought to be due to random experimental error associated with the analytical technique. Figure 4-11 illustrates the kinetics of cyanide consumption. Analysis of the leach solution indicated that most of the cyanide was consumed as thiocyanate and hexacyanoferrate. The graph also indicates the formation of these products during the leach, and represents them as a percentage of the total cyanide consumed. Interestingly, the iron is leached into solution rapidly during the first couple of hours and then levels off. In contrast, the formation of thiocyanate occurs steadily, throughout the leach. 90 time [hours] Figure 4-11. Consumption of cyanide during a 24 hour leach test of Steady State #4 product. Ambient temperature, 30% pulp density, 0.5 g/L NaCN, pH = 10.4-11.0. Based on the data presented in Figures 4-10 and 4-11, a simple method to reduce cyanide consumption can be proposed. By shortening the leach time to 6-8 hours the cyanide consumption can be cut in half, without affecting the gold extraction. For ease of comparison, the remaining cyanidation tests summarized here used the product material from steady state #4. 91 4.3 Cyanide Mass Balance In an effort to better understand what was happening to the free cyanide during the leach, attempts were made to accurately account for all of the cyanicides present. Based on the literature, possible consumers were thought to include metal complexes with gold, silver, copper, iron, nickel, and zinc, as well as thiocyanate and cyanate. Table 4-14 indicates the cyanicides, their complexes, and the stoichiometric amount of cyanide they complex. Table 4-14. Possible cyanide consuming species, the most probable complex formed, and the mass ratio of sodium cyanide complexed to species present. Cyanicide Compound Formed NaCN Complexed [g/g] Au Au(CN)2- 0.50 Ag Ag(CN)2" 0.91 Cu Cu(CN)32" 2.31 Fe Fe(CN)637 Fe(CN)64- 5.27 Ni Ni(CN)42" 3.34 Zn Zn(CN)42" 3.00 S° SCN" 1.53 02 OCN" 1.53 Mass balance results for one of the earlier experiments are shown in Table 4-15. 92 Table 4-15. Mass balance results for CN-4-P3: Steady State #4 product, ambient temperature, 10% pulp density, 0.5 g/L NaCN, pH = 10.4-11.0. metal cone, [ppm] conversion factor NaCN cons. [ppm] % of total Fe 6.62 5.27 34.85 3.74 Ni 0.22 3.34 0.73 0.08 A u 2.26 0.50 1.13 0.12 A g 0.91 0.91 0.83 0.09 Z n 0.15 3.00 0.45 0.05 C u 1.27 2.31 2.94 0.32 S C N - 460.13 0.84 388.35 41.64 O C N - 4.62 1.17 5.39 0.58 total = 434.67 46.61 From the table it is evident that a large portion of the cyanide consumed (-53%) is unaccounted for. Part of this discrepancy can be attributed to losses associated with intermediate sampling at 1, 3, 6, and 12 hours. However, assuming a 10 ml sample size and a leach solution volume of 280 ml, the maximum loss would be 13.5%. Initial experiments also indicated that cyanate was not a significant reservoir for cyanide. As a result, this species' rather lengthy assay procedure was dropped from subsequent trials. Further experiments yielded similar accounting problems. Results ranged from 24 to 93 % of the cyanide consumed being accounted for. In comparison, cyanidation of the unoxidized feed material resulted in only 7.6% of the cyanide being accounted for, with 4.6% of that reporting as SCN". Generally, experiments conducted at higher pulp densities, or those in the sealed reactor, showed a higher percentage of cyanide accounted for. 93 Several attempts were made to identify the source of the discrepancy. The possibility that a cyanide precipitate was forming was investigated using both the feed material as well as the oxidized product from Steady State #4. Wet samples of the leach residue were sent to Analytical Service Laboratories (ASL) of Vancouver for total cyanide analysis. While cyanide was detected in the oxidized sample, the solids accounted for a little more than 2% of the total consumed; leaving more than 50% unaccounted for. Analysis of the cyanide leached feed material found no detectable cyanide. The possibility that cyanate was degrading and being lost from the solution as ammonia gas was also investigated. In this case, the leach was carried out in a sealed reactor with air pumped into the headspace, and the outflow air being passed through a 0.1 M H2S04 strip solution. However, even after 24 hours, the concentration of NH3 in the strip solution was less than 1 ppm. Additional work considered the possible volatilization of HCN gas from the system. A similar setup to that used in the ammonia measurements was employed except that the strip solution consisted of 0.1 M NaOH. Experiments conducted by acidifying a standard cyanide solution in the apparatus indicated that the strip solution was capable of removing cyanide from the off-gas. However, a test conducted on the product material showed only trace concentrations of cyanide in the strip solution. 94 A final modification to this setup was introduced when the air inlet to the sealed reactor was moved to below the surface of the slurry. During this final run no intermediate samples were taken to eliminate the error described earlier. The mass balance is shown in Table 4-16. Table 4-16. Mass balance for the sealed reactor leach. Steady State #4 product, ambient temperature, 10% pulp density, pH = 10.4-ll.O.Initial NaCN concentration: l.Og/L (allowed to decline). No intermediate samples. C N c o n s u m p t i o n species cone, [ppm] conversion factor NaCN cons. [ppm] % of total F e 176.99 5.27 931.85 95.09 Ni 0.12 3.34 0.40 0.04 A u 2.15 0.50 1.07 0.11 A g 0.06 0.91 0.05 0.01 Z n 0.01 3.00 0.03 0.00 C u 0.35 2.31 0.81 0.08 S C N - 39.20 0.84 33.08 3.38 O C N - 9.50 1.17 11.09 1.13 total = 978.39 99.84 Analysis of the strip solution again showed only trace concentrations of NaCN (<0.01g/L). However, the results of the leach solution balance indicate a remarkably close agreement. Since this number is much closer to 100% than previously observed, this run might be considered an anomaly. Also of interest is the high concentration of iron reported. This result will be revisited in section 4.4.5. Thus, the objective of accounting for all of the cyanide during the leach was met with limited success. In general, experiments run at higher pulp densities were able to account for more of 95 the cyanide than those run at lower densities. This effect may be attributable to lower concentrations of iron and sulfur species at lower pulp densities. Reactions forming OCN" or HCN would, however, proceed at the same rate. Similarly, sealed reactor runs showed better accounting than those in an open vessel. The possibility that some losses occurred through HCN gas may still be realistic, although not evident under the conditions tested. Other possibilities include polycyanide complexes as well as complexes formed with proteins and other biological macromolecules generated during the bioleaching step. 4.4 Effect of Leach Parameters on Cyanide Consumption In addition to the bioleach and cyanidation tests summarized earlier, a number of experiments were conducted in order to better understand the influence of various parameters on cyanide consumption. For all of these experiments, the residue used was the product collected from the fourth steady state. Results pertaining to pulp density, temperature, pH, cyanide concentration, and stirring speed are now summarized. 4.4.1 Pulp Density The influence of pulp density on cyanide consumption was studied by leaching at three different densities: 10, 20, and 30%. Since cyanide concentration was maintained at 0.5 g/L in all tests, the lower densities showed higher consumption per tonne of solids (see Figure 4-12). The remaining tests in this series were carried out at 10% pulp density and as a result show higher cyanide consumptions than those reported in the previous section. 96 time [hours] Figure 4-12. Effect of pulp density on sodium cyanide consumption measured in kg/t. Ambient temperature, 24 hr leach, 0.5 g/L NaCN, pH = 10.4-11.0. Figure 4-13 considers the total cyanide consumed (in ppm) for the three different pulp densities. As expected, the higher densities consume more reagent. Thus, the result shown in the figure above may be attributed to losses that are independent of the solids (e.g. HCN and OCN") which would be constant at any density. 97 2000 time [hours] Figure 4-13. Effect of pulp density on sodium cyanide consumption measured in ppm. Ambient temperature, 24 hr leach, 0.5 g/L NaCN, pH = 10.4-11.0. Table 4-17. Comparison of cyanide consumption and % of cyanide consumed as thiocyanate and hexacyanoferrate for three different pulp densities. Pulp Density [%] NaCN Cons. [g/L] % NaCN as SCN % NaCN as Fe(CN)6 10 0.93 41.64 3.74 20 1.32 48.88 3.02 30 1.85 43.23 18.93 Table 4-17 illustrates that at higher densities more of the cyanide consumed can be accounted for, as described earlier. The distribution of iron- and sulfur-cyanide species yields no clear trend. Further study examined the relationship between the mass-of-SCN": mass-of-solids ratio and pulp density (see Figure 4-14). The graph indicates that cyanide is more readily consumed 98 in this form at lower pulp densities. A good linear fit is also evident, although more data is necessary to confirm this result. 5.0 -r 0.0 -I 1 1 1 1 1 1 5 10 15 20 25 30 35 pulp density [%] Figure 4-14. Relative thiocyanate formation. Ambient temperature, 24 hr leach, 0.5 g/L NaCN, pH = 10.4-11.0. 4.4.2 Temperature The consumption of cyanide over time at three temperatures ranging from 19.4 to 40.0 °C is plotted in Figure 4-15. As expected, higher temperatures resulted in increased reagent consumption. 99 time [hours] Figure 4-15. Effect of temperature on sodium cyanide consumption. 10% pulp density, 24 hr leach, 0.5 g/L NaCN, pH = 10.4-11.0. Table 4-18 compares the gold extraction, cyanide consumption, and percent of cyanide as thiocyanate for the three runs. As the temperature increased a greater portion of the cyanide consumed could be accounted for as thiocyanate. Similarly, gold extraction also increased: from 94.0% to 95.6%. 100 Table 4-18. Comparison of gold extraction, cyanide consumption, and % of cyanide consumed as thiocyanate for three different leach temperatures. Temperature [°C] Au extraction [%] NaCN Cons.[kg/trJ % NaCN as SCN 19.6 94.0 10.29 32.1 30.0 95.2 12.69 39.6 40.0 95.6 14.90 48.7 The concept of relative SCN" generation developed in the previous section was used again to further examine the influence of higher temperatures. Figure 4-16 plots the log of this parameter for the three runs. Again, a good fit is achieved, indicating that the reaction between cyanide and sulfur is temperature dependent. 0.2 | 0.1 --0.0 -I ! i 1 ! 1 1 j 10 15 20 25 30 35 40 45 50 temperature [C] Figure 4-16. Effect of temperature on relative SCN" formation. 10% pulp density, 24 hr leach, 0.5 g/L NaCN, pH = 10.4-11.0. 101 4.4.3 pH Sealed reactor experiments were carried out to measure the effect of low pH on cyanide consumption. It was speculated that this parameter might influence the dissolution of intermediate sulfur species, and, in turn, the consumption of cyanide as SCN". Figure 4-17 illustrates the results of first of these tests as well as the profile of a run done in the standard setup (pH=10.6). The lower pH's resulted in very rapid consumption of the cyanide, initially thought to be volatilization of HCN gas. However, ICP analysis of the leach solution indicated enough iron to account for most of the loss as hexacyanoferrates (see Table 4-19). 10 T time (hours] Figure 4-17. Effect of pH on sodium cyanide consumption. Ambient temperature, 24 hr leach, initial [NaCN] = 1.0 g/L (allowed to decline), 10% pulp density. 102 Table 4-19. Comparison cyanide consumption and % of cyanide consumed as thiocyanate and hexacyanoferrate at varying pH. (Note: virtually all cyanide was consumed in each run; the difference in NaCN consumption is the result of reagent lost during intermediate sampling.) pH NaCN Cons. [kg/tres] % NaCN as SCN % NaCN as Fe(CN)6 7.8 8.84 4.62 74.28 9.8 8.67 16.41 75.28 10.6 7.78 30.19 1.66 Although less cyanide was consumed as SCN" at lower pH, this result alone cannot substantiate a relationship. Since most of the cyanide was rapidly consumed by the iron in the first few hours, there may not have been sufficient opportunity for SCN" to form. 4.4.4 Cyanide Concentration Since commercial cyanidation processes operate over a range of reagent concentrations the effect of this parameter was investigated. In this case, the standard leach at 0.5 g/L NaCN was accompanied by runs at 1.0 and 0.25 g/L. The results of these trials are shown in Figure 4-18. As expected, higher cyanide concentrations resulted in higher consumption. 103 14 time [hours] Figure 4-18. The effect of cyanide concentration on consumption. Ambient temperature, 10% pulp density, 24 hr leach. Unfortunately, lower cyanide concentrations also resulted in lower gold extractions. As Table 4-20 shows, at 0.25 g/L the recovery dipped below 90%. At the same time, the percentage of cyanide consumed as SCN" also declined. Table 4-20. Gold extraction and cyanide consumption for varying sodium cyanide concentrations. NaCN Au extraction NaCN NaCN NaCN concentration [%] consumed consumed as consumed as [g/L] [kg/t] SCN" [%] Fe(CN)6[%] 0.25 89.9 7.50 26.4 7.82 0.5 94.0 10.29 32.1 1.62 1.0 95.2 13.33 34.8 1.68 104 4.4.5 Stirring Speed The influence of stirring speed was also considered. It was proposed that less agitation might slow the cyanide consuming reactions and thus limit reagent use. However, the reverse was found to be true. As shown in Figure 4-19, lower speeds were found to yield higher consumptions, particularly in the early stages of the leach. time [hours] Figure 4-19. Effect of stirring speed on sodium cyanide consumption. Ambient temperature, 10% pulp density, 24 hr leach, 0.5 g/L NaCN, pH = 10.4-11.0. Again, the explanation for this result seems to rest with iron. The analysis of the final leach solutions showed that iron the concentration steadily decreased as the stirring speed increased 105 (see Figure 4-20). In comparison, thiocyanate concentrations were almost identical for each run. This result, along with the profiles of the leach, suggest that some type of iron compound is dissolving in the early stages. Since lower stirring speeds also reduce oxygen concentration in solution, and the bioleach solids are likely to contain amorphous jarosites or other basic ferric sulfates, a reductive leach is suspected. Equation 4-3 proposes a mechanism by which the decomposition of jarosites might occur, MFe3(S04)2(OH)6 + 6H+ + 3e~ -> M+ + 3Fe2+ + 2S02" + 6H2O (4-3) where M might represent H30+, NH4+, or K+. Possible reductants for this reaction include reduced sulfur species such as sulfite or thiosulfate, Sx -» S2 02~ -» SO2' -> SO2' (4-4) stirring speed [rpm] Figure 4-20. Thiocyanate and hexacyanoferrate concentrations in cyanide leach solutions at three stirring speeds. Ambient temperature, 10% pulp density, 24 hr leach, 0.5 g/L NaCN, pH = 10.4-11.0. 106 Thus, the consumption of cyanide might be characterized as two distinct phases. Firstly, a kinetically fast step dependent on the presence of iron in solution which, in turn, is affected by Eh and ultimately, 02 concentration. And second, a slower, cyanide concentration dependent step resulting in the formation of SCN". From this perspective, the results obtained at low pH can now be explained. The sealed reactor system was agitated by a stir bar, and thus was probably operating under reducing conditions initially. Basic ferric sulfates were reductively decomposed, leached into solution, and consumed most of the cyanide within the first few hours. 4.5 Sample Treatments Based on the results obtained in the previous section, a number of sample treatments were proposed in order to better characterize the cyanide consuming reaction mechanisms. The first of these involved oven drying the slurry at a temperature of 110 °C for 3 hours prior to cyanidation. It was reasoned that this treatment might convert any polysulfides present to the stable orthorhombic form of sulfur and thus limit the formation of SCN". In addition, the high temperature would denature any viable molecules of the enzyme rhodanese. Figure 4-21 compares the results of this run with a control experiment. 107 1 2 time [hours] Figure 4-21. Sodium cyanide consumption for a slurry sample treated at 110 °C for 3 hours. Ambient temperature, 0.5 g/L NaCN, 10% pulp density, 24 hour leach, pH 10.4-11.0. The graph indicates that initially the cyanide consumption rate is not affected. This is consistent with the suggestion in the previous section that a period of rapid reagent loss at the start of the leach results from basic ferric sulfate leaching. Later in the leach, however, the treated sample proves to be less cyanide consuming. Furthermore, after 24 hours this sample had consumed only 2.12 kilograms of sodium cyanide per tonne of concentrate as thiocyanate, compared with 3.43 kg NaCN for the control (see Table 4-21). 108 Table 4-21. Comparison of cyanide consumption and % of cyanide consumed as thiocyanate and hexacyanoferrate for two sample treatments. NaCN Cons. [kg/t r e s] % NaCN as SCN % NaCN as Fe(CN)6 control 10.29 33.35 1.62 heat pre-treatment 9.51 22.27 0.67 aeration 9.91 26.41 1.10 Another trial investigated the influence of aeration on the leach. Since a reductive basic ferric sulfate leach is suspected, maintaining oxidizing conditions should result in lesser hexacyanoferrate formation. Sparging air into the slurry was employed to ensure that the concentration of dissolved oxygen was close to saturation. Figure 4-22 compares the results of this run with the control. 12 T 1 time [hours] Figure 4-22. The effect of aeration on sodium cyanide consumption. Ambient temperature, 10 % pulp density, 0.5 g/L NaCN, 24 hour leach, pH 10.4-11.0. 109 As expected, the early part of the leach was characterized by lower reagent consumption. Assay of the final filtrate showed that iron accounted for less than 2% of the cyanide consumed, while generation of SCN" was not appreciably affected. However, the control experiment also resulted in a low iron concentration in solution (see Table 4-21). Possibly an analytical error was responsible for this discrepancy, although this was not confirmed. The influence of aeration in controlling the loss of cyanide at low pH was investigated using the sealed reactor. No calcium hydroxide was added to the slurry and an initial cyanide concentration of 1.0 g/L was allowed to decline during the leach. Air was sparged from the headspace of the reactor back into the slurry. An Eh probe verified that oxidizing conditions were maintained throughout the leach with this parameter never falling below 445 mV (SHE). Despite the seemingly favorable conditions, most of the reagent was consumed in the first hour. Overall, 72% of the cyanide reported as Fe(CN)6. A similar run was conducted, this time with an average pH of 7.4, in which 75% of the cyanide was consumed as hexacyanoferrate. This result indicates that both pH and Eh influence jarosite dissolution. Equation 4-4 is proposed to account for jarosite dissolution under low pH, mildly oxidizing conditions: MFe3(SO,)2(OH)6+6H+ -> M++3Fei+ +2S02~ +6H20 (4-4) Again, M represents H30+, NH4+, or K+. 110 In an effort to confirm the source of the iron in the leach as resulting from jarosite decomposition, experiments were conducted on samples that had been pre-leached with hydrochloric acid to remove jarosite. Here, a 300 g slurry sample from the product of Steady State #4 was filtered and the wet solids were repulped with an equal mass of 4M HC1. Agitation for 15 minutes was accomplished with a magnetic stirrer after which time the slurry was filtered and washed, first with 4M HC1 and then deionized water. Finally, the solids were repulped with deionized water. The HC1 leach resulted in a sample mass loss of 13.7%. Experiments were conducted to determine the cyanide consumption at varying pH. Two of these tests were conducted in the sealed reactor as the pH was below the pKa of cyanide. Because of the difficulty of adding NaCN to the sealed reactor during the run, these tests consisted of adding cyanide to a concentration of 1 g/L at time zero and allowing the concentration to decline after that. Figure 4-23 compares the relationship between hexacyanoferrate, thiocyanate, and the total cyanide consumed during the leach. I l l 11 Figure 4-23. The distribution of iron-cyanide and sulfur-cyanide species for tests conducted at varying pH. All samples subjected to a 15 minute pre-leach in 4M HC1,10% pulp density, ambient temperature, 24 hour leach, initial NaCN concentration l.Og/L. As the figure indicates, the lowest cyanide consumption occurred in the moderate pH range of 8.4. The concentration of thiocyanate increases with pH, while the reverse is true for the iron cyanides. It is important to note that for all three runs much of the consumed cyanide was unaccounted for (72-76%). In addition, gold extraction was poor at the lower pH's: 62.6% and 87.3% for pH 6.8 and 8.4, respectively, while the other run yielded 91.4%. 112 Titration of intermediate solution samples taken during the leach at pH 6.8 showed that almost half of the free cyanide (0.47 g/L) was still in solution after 6 hours. However, poor recoveries experienced during this run can be attributed to most of this cyanide being present as HCN(aq). This supports the suggestion by Fleming (1992) that gold complexation slows considerably below pH 8. Also of interest is the increase in hexacyanoferrate at lower pH. This is probably the result of decomposition of basic ferric sulfates or iron hydroxides not eliminated in the pre-leach step. The fact that more iron is leached at lower pH seems to confirm the influence of this parameter on jarosite dissolution proposed earlier. Again, lower concentrations of SCN- are observed at lower pH, but this may also be due to competition for reagent with iron species. An Eh-pH diagram constructed for the iron-sulfur-cyanide-water system is shown in Figure 4-24, and can be compared to the iron-sulfur-water system shown in Figure 2-8. Here, hydronium jarosite, hematite, and magnetite have been included in the species present, and are represented in regions A, B, and G, respectively. Above pH 2, hexacyanoferrates are predominant within much of the H20 stability region. 113 Figure 4-24. Eh-pH diagram for the Iron-Sulfur-Cyanide-Water system at 25 °C. Activity of all species: 1 M. (Drawn using the CSIRO thermochemistry program, see Appendix D for thermodynamic data). The diagram offers thermodynamic support for the assertion that the jarosites and other basic ferric sulfates formed during the bioleaching step, as well as iron oxides that may form during the neutralization, are being leached into solution. However, the influence of Eh and pH observed in the leaching experiments is not substantiated here. Possibly these parameters have a greater affect on the kinetics of jarosite dissolution. Also, it should be noted that the free energy 114 data used in the construction of this diagram is specific to crystalline jarosite, and not the amorphous form suspected here. 4.6 Sulfur Balance In an effort to identify the source of the thiocyanate present in the leach solution, analyses for elemental sulfur, sulfide, and sulfate were performed on the leach residues of four of the cyanidation experiments. From the Eh-pH diagram presented in Figure 4-25, the stability of the sulfur species in the presence of a cyanide ligand can be seen. Area B indicates the stability region for thiocyanate. This species is shown to be stable in solution across the entire range of pH: from mildly oxidizing (low pH) to moderately reducing (high pH) conditions. 115 1.5 PLOT LABELS T«I»P = 298.15 K 1S1 = 0.01 M [CM = 0.01 M STR3LE RRERS H SOI <2-> IRQ) B H S <-> CRQ) C S <2-> IRQ) D S C N <-> IRQ) E H S OH <-> CRQ) F H2 S IRQI LIGHND BRERS Rfl HCN (RQ) BB CN <-> IRQ) H20 STRBIL1TY LIMITS 1 OXYGEN 2 HYDROGEN Figure 4-25. Eh-pH diagram for the Sulfur-Cyanide-Water system at 25 °C. Activity of aqueous species: 0.01 M. (Drawn using the CSIRO thermochemistry program, see Appendix D for thermodynamic data). In contrast, elemental sulfur is not thermodynamically stable in this system, and reacts with cyanide, as follows, 5° + C A T -> SCN' (4-5) (where for biooxodized residues, S° may include biologically generated, hydrophobic, reactive intermediate sulfur species). Using this reaction, and assuming all the consumed S° is converted to SCN", the total possible amount of thiocyanate generated was calculated for each of the four runs. The results are shown in Table 4-22. 116 Table 4-22. Results of the S7SCN" balance. All tests 24 hours at ambient temperature. Standard test: 0.5 g/L NaCN, pH~10.6, 30% pulp density. No lime: sealed reactor run with no lime addition, pH~4.0, initial NaCN concentration 1.0 g/L (allowed to decline), 10% pulp density. Falling CN: initial NaCN concentration 1.0 g/L (allowed to decline), 10% pulp density, pH~10.6. Low pH: initial NaCN concentration 1.0 g/L (allowed to decline), 10% pulp density, pH~7.4. test descriDtion: s tandard test no lime fall ina C N " low D H so l ids wt [g] 101.41 32.30 31.19 31.45 so l ids S° assay [%] 0.64 0.64 0.64 0.64 res idue wt [g] 103.25 32.30* 30.33 31.45* res idue S° assay [%] 0.47 0.64 0.44 0.61 max imum possible S C N " genera ted [ppm] 1253.22 0.00 426.89 60.36 actual S C N " assay fppml 834.12 17.42 370.41 20.33 intermediate slurry sampling during these runs resulted in lower final residue weight, in order to account for S° lost to sampling the initial weight is used here. With the exception of the second test, where a small error associated with the assay procedure might be at fault, a clear trend between the loss of S° and the formation of SCN" is evident. However, the amount of reactant consumed exceeds the product generated, particularly in the standard test. Three possibilities exist to explain this result. Firstly, the elemental sulfur may be involved in a second reaction which does not lead to thiocyanate. Conversely, the thiocyanate may be precipitating, and thus not reporting to the leach solution (this occurrence would also explain the difficulties in accounting for all of the CN" consumed during the leach). Finally, the S° might first be forming some other intermediate sulfur species, such as polysulfide (Sx2") which 117 then reacts to generate SCN". If the second reaction is slower than the first, then some of the sulfur may be tied up in this intermediate form. In addition, the possibility that sulfur intermediates may report as S° must also be considered. It is unclear where biologically formed sulfur granules, which contain both elemental sulfur and polythionates, would show up under assay. Since the granules have a hydrophilic surface they would not likely dissolve during boiling in perchloroethylene (the solvent used in the assays reported here). Thus the granules, including some elemental sulfur, might be counted as sulfide. From Figure 4-25 it can be seen that the formation of SCN" is thermodynamically favored over sulfide or bisulfide under normal leaching conditions. The reaction of polysulfide with cyanide is reported in the literature as follows (Flynn and Haslem, 1995), C A T + S2X~ -> SCN~ + S 2" (4-6) Table 4-23 evaluates the possibility that sulfide is responsible for consumption of cyanide as SCN". As with Table 4-22, the maximum thiocyanate that could be generated by the disappearance of S2" is compared with the assayed value. 118 Table 4-23. Results of the S27SCN balance. All tests 24 hours at ambient temperature. Standard test: 0.5 g/L NaCN, pH~10.6,30% pulp density. No lime: sealed reactor run with no lime addition, pH~4.0, initial NaCN concentration 1.0 g/L (allowed to decline), 10% pulp density. Falling CN: initial NaCN concentration 1.0 g/L (allowed to decline), 10% pulp density, pH~10.6. Low pH: initial NaCN concentration 1.0 g/L (allowed to decline), 10% pulp density, pH~7.4. test descr ipt ion: s tandard test no l ime fall ina C N low pH solids wt [g] 101.41 32.30 31.19 31.45 sol ids S 2 " assay [%] 4.80 4.80 4.80 4.80 residue w t [g] 103.25 32.30* 30.33 31.45* residue S 2 " assay [%] 4.30 4.91 4.81 4.48 m a x i m u m possible SCN" generated [ppm] 3275.08 -221.34 246.77 643.84 actual SCN" assay [ppml 834.12 17.42 370.41 20.33 intermediate slurry sampling during these runs resulted in lower final residue weight, in order to account for S2" lost to sampling the initial weight is used here. Unlike the results for the S7SCN" balance, no trend is particularly evident here. As a result, it is difficult to conclude by these methods what intermediate sulfur species might be present or how they influence cyanide consumption. A final experiment was designed to identify if a soluble sulfur species might play an intermediate role in SCN" formation. Three product samples were leached (10% pulp density, ambient temperature) for 12 hours at different pH's (8, 10, 12). No cyanide was added, and the pH was controlled with 1M NaOH. After the leach the slurry was filtered. Sodium cyanide was added to the filtrate to a concentration of 1 g/L. This solution was agitated for 1 hour, and then analyzed for free cyanide and thiocyanate. As only trace amounts of thiocyanate were detected (see Table 4-24), it appears unlikely that a soluble intermediate is involved. 119 Table 4-24. Results of the NaOH leach. Ambient temperature, 10% pulp density, 12 hour duration. Slurry was then filtered, sodium cyanide added to the filtrate, and agitated for 1 hour. Leach pH NaCN added [ppm] SCN" concentration [ppm] 8 1000 10.2 !0 1000 19.6 12 1000 67.5 Thus, it is reasoned that the cyanide must be reacting with an insoluble sulfur species. The identity of that species, however, is not clearly evident. The intermediate sulfur compounds predicted by the literature are difficult to identify by curre.t assay methods, so their presence, and whatever role they might play, is uncertain. 120 5. Conclusions and Recommendations Based on the results summarized in the previous section, a number of conclusions may be drawn. Firstly, bioleaching has been shown to be an effective means of pretreating the refractory gold concentrate used in this study. High gold extractions were observed for residence times as low as 2 days. Selective leaching of arsenopyrite over pyrite was observed. Since the gold was closely associated with the former mineral, high extractions were possible even when the extent of sulfide oxidation dropped below 50%. The sodium cyanide consumption associated with this ore was found to be 2.7-12.1 kg/t-concentrate under the conditions tested. Most of this reagent was accounted for as thiocyanate and hexacyanoferrate. The formation of hexacyanoferrate was found to account for between 2 and 95% of the cyanide consumed depending on the leach conditions. Possible sources for the iron include basic ferric sulfate precipitates or iron oxides, where the latter is in turn formed from the decomposition of jarosites under basic conditions. The reaction to generate Fe(CN)6 was observed to occur 121 rapidly, at the beginning of the leach. Furthermore, reducing conditions, as well as low pH, were shown to result in a higher proportion of the cyanide being consumed as hexacyanoferrate. Thiocyanate represented the largest single consumer of cyanide under most conditions: accounting for up to 77% of the reagent consumed. In contrast to iron, the reaction with sulfur occurred slowly, over the course of the leach. Key parameters for this reaction were found to be pulp density, temperature, and cyanide concentration. Since the kinetics of gold leaching were observed to be quite rapid, shorter leach times might be an effective method of reducing consumption in this manner. Finally, the source of the reactive sulfur was likely to be S° or an intermediate species closely associated with, and reporting under present assay methods as, elemental sulfur. A number of other cyanide consuming reactions were not found to occur in significant quantity for the concentrate tested. These include metal cyanide complexes of copper, zinc, and nickel. In addition, the formation of cyanate was not detected at anything more than trace levels, nor was ammonia found as a by-product of cyanate decomposition. The formation and subsequent loss from the. system of hydrogen cyanide was not observed under normal conditions of pH between 10.4 and 11.0. Mass balance results for cyanide indicated that in most cases a large fraction of the cyanide was unaccounted for. This discrepancy may be the result of problems encountered with the experimental setup, or formation of polycyanide complexes and other difficult to detect cyanide species. 122 5.1 Recommendations A number of recommendations for future work are also proposed: 1. Further study on other biooxidized concentrates, as well as pure minerals, to compare with the results obtained here. 2. Development of a standard cyanidation test apparatus. That is, a closed vessel system that will allow for adequate aeration and agitation, as well as monitoring of temperature, Eh, 02 concentration, and control of pH. 3. Development of a cyanidation test procedure specific to identifying key parameters. Trying to maintain constant NaCN concentrations is impractical (and usually impossible) for the residues in question. Rather, a procedure of adding a known concentration of cyanide at the start, and then monitoring the rate of consumption, may prove more useful. 4. The application of modeling to cyanide consumption. For biooxidized residues, however, this is likely to be difficult until their composition (particularly sulfur intermediates) can be accurately characterized. 123 5. The use of transmission electron microscopy (TEM) may prove useful in studying the oxidation products at the bacteria-mineral interface, if difficulties in sample preparation can be overcome. 6. For the specific concentrate studied here, the application of biooxidation can be pursued in a number of ways. Bioleaching with even shorter residence times and at lower pH's may be possible. Also, the use of a lime pre-treatment (Hackl and Jones, 1997) coupled with a shorter leach time, may result in a significant reduction in cyanide consumption. 124 Works Cited Abbruzzese, C, S. Ubaldini, F. Veglio, and L. Toro. "Preparatory Bioleaching to the Conventional Cyanidation of Arsenical Gold Ores." In: Minerals Engineering. 7(1) : 49-60, 1994. Adams, M.D. "The chemical behaviour of cyanide in the extraction of gold. 1. Kinetics of cyanide loss in the presence and absence of activated carbon." In: Journal of the South African Institute of Mining and Metallurgy. 90(2), 37-44, 1990. Adams, M.D. "The chemical behaviour of cyanide in the extraction of gold. 2. Mechanisms of cyanide loss in the carbon-in-leach process." In: Journal of the South African Institute of Mining and Metallurgy. 90(3), 67-73, 1990. Ahonen, Lasse, and Olli H. Tuovinen. "Alterations in Surfaces and Textures of Minerals During the Bacterial Leaching of a Complex Sulfide Ore." In: Geomicrobiology Journal. 10. 207-217. 1992. Anderson, John W. Sulphur in Biology. The Camelot Press: Southampton, 1978. APHA. Standard Methods for the Examination of Water and Wastewater. 16th ed. Washington, DC: American Public Health Association, 1985. Arkesteyn, G.J.M.W. "Pyrite Oxidation by Thiobacillus ferrooxidans With Special Reference to the Sulphur Moiety of the Mineral." In: Antonie van Leeuwenhoek. 45:423-435, 1979. Bard, Allen J., Roger Parsons, and Joseph Jordan. Standard Potentials in Aqueous Solution. 1st ed. New York: M. Dekker, 1985. Basaran, Ali H. and Olli H. Tuovinen. "Iron Pyrite Oxidation by Thiobacillus Ferrooxidans: Sulphur Intermediates, Soluble End Products, and Changes in Biomass" In: Coal Preparation. 5:39-55, 1987. Bath, Murray D. "Biohydrometallurgical Applications: Biological Oxidation of Refractory Gold Ores." In: Leaching Short Course. Mincorp Engineers and Constructors, 1995. Beck, M.T. "Critical Survey of Stability Constants of Cyano Complexes." In: Pure and Applied Chemistry. 59(12), 1703-1720, 1987. Bhappu, Roshan B. "Hydrometallurgical Processing of Precious Metal Ores." In: Mineral Processing and Extractive Metallurgy Review. 6: 67-80, 1990. 125 Blake II, Robert, Elizabeth A. Shute, James Waskovsky, and Arthur P. Harrison Jr. "Respiratory Components in Acidophilic Bacteria That Respire on Iron." In: Geomicrobiology Journal 10:173-192, 1992. Breed, A.W., A. Glatz, G.S. Hansford, and S.T.L. Harrison. "The Effect of As(III) and As(V) on the Batch Bioleaching of a Pyrite-Arsenopyrite Concentrate." In: Minerals Engineering. 9(12): 1235-1252, 1996. Brierley, Corale L. "Bacterial Leaching." In: CRC Critical Reviews in Microbiology. 3:207-262. 1978. Brierley, Corale L. and Richard Winby. "BacTech's Thermophilic Biooxidation Plant at Youanmi Mine: An Update On Performance and Cost." In: Randol Gold Forum - Perth '95. 1995. Briggs, Andy. "Practical Implementation of the Bacterial Oxidation Process for Refractory Gold Ores." Short Course Notes: Signet Engineering, 1996. Briggs, Andrew P., and Martin Millard. "Cobalt Recovery Using Bacterial Leaching at the Kasese Project, Uganda." In: International Biohydrometallurgy Symposium '97: Conference Proceedings. 4-6 A5gust, 1997. Brock, Thomas D., Michael T. Madigan, John M. Martinko, and Jack Parker. Biology of Microorganisms. Prentice Hall: Englewood Cliffs, NJ. 1996. Budden, Julia R. "Youanmi Bacterial Oxidation: Commissioning and Operation." In: Randol Gold Forum - Perth '95. 1995. Burns, D.T., A. Townshend, and A.H. Carter. Inorganic Reaction Chemistry: Volume 2. Ellis Horwood Limited: New York, 1980. Carter, A.J. "Economic Comparison of the Alternative Methods for the Recovery of Gold from Refractory Gold Ores." In: Colloquium on Bacterial Oxidation. South African Institute of Mining and Metallurgy; Johannnesburg. 1991. Crundwell, F.K. "An Amperometric Method for the Determination of Cyanide in Gold Process Streams and Slurries." Hydrometallurgy. 27:19-28, 1991. Crundwell, Frank K. and Sven A. Godorr. "A Mathematical Model of the Leaching of Gold in Cyanide Solutions." In: Hydrometallurgy. 44: 147-162, 1997. 126 Crundwell, F.K. and J.O. Jensen. "An Amperometric Instrument for the Determination of Cyanide on Gold Plants." Journal of the South African Institute of Minerals and Metallurgy. 94: 103-108, 1994. Curreli, L., G. Loi, R. Perreti, G. Rossi, P. Trois, and A. Zucca. "Gold Recovery Enhancement from Complex Sulphide Ores Through Combined Bioleaching and Cyanidation." Minerals Engineering. 10(6) : 567-576, 1997. Das, A. And A.K. Misra. "Role of Thiobacillus ferrooxidans and sulphur (sulphide)-dependent ferric-ion-reducing activity in the oxidation of sulphide minerals." In: Applied Microbiology and Biotechnology 45:377-382, 1996. De, G.C, D.J. Oliver, and B.M. Pesic. "Effect of Heavy Metals on the Ferrous Iron Oxidising Ability of Thiobacillus ferrooxidans." Hydrometallurgy. 44:53-63, 1997. Dutrizac, J.E. and D.J. Hardy. "The Behaviour of Thiocyanate and Cyanate During Jarosite Precipitation." In: Hydrometallurgy 45: 83-95, 1997. Dutrizac, J.E. and J.L. Jambor. "The Behaviour of Arsenic During Jarosite Precipitation: Arsenic Precipitation at 97 °C From Sulphate or Chloride Media." In: Canadian Metallurgical Quarterly 26(2): 91-101. 1987 Fernandez, Marcos G. Monroy, Christian Mustin, Phillipe de Donato, Odile Barres, Philipe Marion, and Jacques Berthelin. "Occurrences at the Mineral-Bacteria Interface During Oxidation of Arsenopyrite by Thiobacillus ferrooxidans." In: Biotechnology and Bioengineering. 46: 13-21, 1995. Flynn, Charles M. and Sandra McGill Haslem. Cyanide Chemistry—Precious Metals Processing and Waste Treatment. United States Bureau of Mines. 1993. Fleming, CA. "Hydrometallurgy of Precious Metals Recovery." In: Hydrometallurgy. 30: 127-162, 1992. Gehrke, T., R.Hallmann, and W. Sand. "Importance of Exopolymers from Thiobacillus ferrooxidans and Leptospirillum ferrooxidans for Bioleaching." In: Biohydrometallurgical Processing. 1995. Gomez, C, M. Figueroa, J. Munoz, A. Ballester, and M.L. Blazquez. "A Study of Bioleached Chalcopyrite Surfaces in the Presence of Ag(I) by Voltammetric Methods." Minerals Engineering. 10(1): 111-116, 1997. Habashi, Fathi. A Textbook of Hydrometallurgy. Metallurgie Extractive Quebec; Sainte Foy, Quebec. 1993. 127 Hackl, R.P. "What to be Aware of in Cyanidation of Bio-oxidized Products." In: Innovations in Gold and Silver Recovery: Randol Gold Conference Phase IV. Sacremento, Calfornia, NovenV'er 9-11, 1989. Hackl, R.P. and L. Jones. "Bacterial Sulfur Oxidation Pathways and Their Effect on the Cyanidation Characteristics of Biooxidized Refractory Gold Concentrates." In: International Biohydrometallurgy Symposium '97: Conference Proceedings. 4-6 August, 1997. Haines, A.K. "Factors Influencing the Choice of Technology for the Recovery of Gold from Refractory Arsenical Ores." Proceedings of the International Conference on Gold: Volume 2. SAIMM; Johannesburg. 1986. Haines, A.K. and P.C. van Aswegen. "Process and Engineering Challenges in the Treatment of Refractory Gold Ores." In: International Deep Mining Conference. Johannesburg: SAIMM, 1990. Harris, Daniel C, Quantitative Analytical Chemistry. W.H. Freeman and Company: New York, 1991. Hazeu, W., W.H. Batenburg-van der Vegte, P. Bos, R.K. van der Pas, and J.G. Kuenen. "The Production and Utilization of Intermediary Elemental Sulphur During The Oxidation of Reduced Sulphur Compounds by Thiobacillus ferrooxidans." In: Archives of Microbiology. 150:574-579,1988. Huxtable, Ryan J. Biochemistry of Sulphur. Plenum Press: New York, 1986. Jha, Mahesh C. "Refractoriness of Certain Gold Ores to Cyanidation: Probable Causes and Possible Solutions." In: Mineral Processing and Extractive Metallurgy Review. 2: 331-352, 1987. Karavaiko, G.I., S.I. Kuznetsov and A.I. Golonizik. The Bacterial Leaching of Metals from Ores. Technicopy Limited: Stonehouse, UK. 1977. Kelly, D.P. In: Sulphur in Biology Exerpta Medica: Amsterdam, 1980. Komnitsas, C. and F.D. Pooley. "Bacterial Oxidation of an Arsenical Gold Sulphide Concentrate from Olympias, Greece." Minerals Engineering. 3: 295-305. 1990. Kondos, P.D., G. Deschenes, R.M. Morrison. "Process Optimization Studies in Gold Cyanidation." In: Hydrometallurgy. 39:235-250, 1995. Kudryk, V., and H.H. Kellogg. "Mechanism and Rate-Controlling Factors in The Dissolution of Gold in Cyanide Solution." In: Journal of Metals. May, 1994. 128 Lawrence, R.W. — In: Impurity Control and Disposal in Hydrometallurgical Process. Canadian Institute of Mining, Metallurgy, and Petroleum: Montreal, Quebec. 1994. . Lawson, Ellen N. "The Role of the Enzyme Rhodanese in Cyanide Consumption for Gold Extraction From Bacterially Leached Refractory Ores." In: International Biohydrometallurgy Symposium '97: Conference Proceedings. 4-6 August, 1997. Leach, C.K. "Chemotrophy." In: Energy Sources for Cells. Butterworth-Heinemann: London, UK. 1992. Lenahen, W.C. and R. de L. Murray-Smith. Assay and Analytical Practice in the South African Mining Industry. The South African Institute of Mining and Metallurgy: Johannesburg. 1986. Lorbach, Stanley C, Jessup M. Shively and Vincenzo Bounfiglio. "Kinetics of Sulphur Oxidation by Thiobacillus ferrooxidans." In: Geomicrobiology Journal 10:219-226. 1992. Lowson, Richard T. "Aqueous Oxidation of Pyrite by Molecular Oxygen." In: Chemical Reviews. 82(5), 1982. Luthy, Richard G., and Samuel G. Bruce, Jr. "Kinetics of Reaction of Cyanide and Reduced Sulfur Species in Aqueous Solution." In: Environmental Science and Technology. 13(12), 1481-1487, 1979. Mandl, Martin, David Hrbac, and Hana Docekalova. "Inhibition of Iron(II) Oxidation by Arsenic(III,V) in Thiobacillus Ferrooxidans: Effects on Arsenopyrite Bioleaching." In: Biotechnology Letters. 18(3): 333-338, 1996. Marina, A., K.A. Natarajan, Preston Devasia, and G. Ramananda Rao. "Surface Chemistry of Thiobacillus ferrooxidans with Respect to Interaction with Sulphide Minerals." In: Mineral Processing: Recent Advances and Future Trends. Mustin, C, J. Berthelin, P. Marion, and P. de Donato. "Corrosion and Electrochemical Oxidation of a Pyrite by Thiobacillus ferrooxidans." In: Applied and Environmental Microbiology. Apr. 1992, pp. 1175-1182. Mustin, C, Ph. de Donato, J. Berthelin, and Ph. Marion. "Surface Sulphur as Promoting Agent of Pyrite Leaching by Thiobacillus ferrooxidans." In: FEMS Microbiology Reviews. 11:71-78,1993. 129 Osseo-Asare, K., T. Xue, and V.S.T. Ciminelli. "Solution Chemistry of Cyanide Leaching Systems." In: Precious Metals: Mining, Extraction, and Processing. AIME: Warrendale, PA, 1984. Pinches, A., R. Huberts, J.W. Neale, and P. Dempsey. "The MINBAC Bacterial-oxidation Process." In: Proceedings of the XVth CMMI Congress. Johannesburg, 1994. Riekkola-Vanhanen, Marja, and Seppo Heimala. "Electrochemical Control in the Biological Leaching of Sulfidic Ores." In: Biohydrometallurgical Technologies. The Minerals, Metals and Materials Society, 1993. Rojas, J., M. Giersig, and H. Tributsch. "Sulphur colloids as temporary energy reservoirs for Thiobacillus ferrooxidans during pyrite oxidation." In: Archives of Microbiology. 163: 352-356, 1995. Rojas-Chapana, Jose A., Michael Giersig and Helmut Tributsch. "The path of sulphur during the bio-oxidation of pyrite by Thiobacillus ferrooxidans" In: Fuel : 75(8), 1996. Rossi, Giovanni. Biohydrometallurgy. McGraw-Hill: Hamburg, 1990. Roy, A.B. and P.A. Trudinger. The Biochemistry of Inorganic Compounds of Sulphur. University Press: Cambridge, 1970. Rubisov, D.H., V.G. Papangelakis, and P.D. Kondos. "Fundamental Kinetic Models for Gold Ore Cyanide Leaching." In: Canadian Metallurgical Quarterly. 35(4), 353-361, 1996. Sand, W., T. Gerke, R. Hallman, A. Schippers. "Sulphur chemistry , biofilm, and the (in)direct attack mechanism - a critical evaluation of bacterial leaching." In: Applied Microbiology and Biotechnology 43 :961-966 1995. Schippers, A., H. Von Rege, and W. Sand. "Impact of Microbial Diversity and Sulphur Chemistry on Safeguarding Sulfidic Mine Waste." In: Minerals Engineering. 9(10): 1069-1079, 1996. Schippers, Axel, Peter-George Jozsa, and Wolfgang Sand. "Sulphur Chemistry in Bacterial Leaching of Pyrite." In: Applied and Environmental Microbiology. Sept. 1996 pp. 3424-3431. Shrader, Vandy J. and Sharon X. Su. "Factors Affecting Elemental Sulfur Formation in Biooxidized Samples: Preliminary Studies." In: International Biohydrometallurgy Symposium '97: Conference Proceedings. 4-6 August, 1997. 130 Soto, H., F. Nava, J. Leal, and J. Jara. "Regeneration of Cyanide by Ozone Oxidation of Thiocyanate in Cyanidation Tailings." In: Minerals Engineering. 7(4): 273-279, 1994. Sparrow, Graham J. and James T. Woodcock. "Cyanide and Other Lixiviant Leaching Systems for Gold with Some Practical Applications." In: Mineral Processing and Extractive Metallurgy Review 14 (pp. 193-247) 1995. Stefess, G.C., R.A.M. Torremans, R. de Schrijver, L.A. Robertson, and J.G. Kuenen. "Quantitative Measurement of Sulphur Formation by Steady-State and Transient-State Continuous Cultures of Thiobacillus Species" In: Applied Microbiology and Biotechnology 45: 169-175, 1996. Stuedel, Ralf. "Mechanism for the Formation of Elemental Sulfur from Aqueous Sulfide in Chemical and Microbiological Desulfurization Processes." In: Industrial Engineering Chemical Research. 35, 1417-1423. 1996. Stuedel, Ralf. "On the Nature of the "Elemental Sulphur" (SO) Produced by Sulphur-oxidising Bacteria - a Model for SO Globules" In: Autotrophic Bacteria. Science Tech Publishers: Madison, WI. 1989. Stuedel, Ralf, Gabriel Holdt, Thomas Gobel, and William Hazeu. "Chromatographic Separation of Higher Polythionates Sn062- (n=3.. .22) and Their Detection in Cultures of Thiobacillus ferrooxidans; Molecular Composition of Bacterial Sulfur Secretions." In: Agnew. Chem. Int. Ed. Engl. 26(2), 1987. Tano, Tatsuo and Donald Lundgren. "Sulfide Oxidation by Spheroplasts of Thiobacillus ferrooxidans." In: Applied and Environmental Microbiology. June 1978, pp. 1198-1205. van Aswegen, Pieter C. "Commissioning and Operation of Bio-Oxidation Plants for the Treatment of Refractory Gold Ores." In: Hydrometallurgy Fundamentals, Technology, and Innovations. Society for Mining, Metallurgy, and Exploration, Inc: Littleton, Colorado. 1993. Vargas, Tomas, Angel Sanhueza, and Blanca Escobar. "Studies on the Electrochemical Mechanism of Bacterial Catalysis in Pyrite Dissolution." In: Biohydrometallurgical Technologies. The Minerals, Metals and Materials Society, 1993. Vargas, Tomas, Angel Sanhueza, and Blanca Escobar. "Electrochemical Studies on the Bacterial Leaching of Pyrite with Thiobacillus Ferrooxidans" In: EPD Congress.; The Minerals, Metals and Materials Society, 1992. 131 Wang, Xianghuai and K.S. Eric Forssberg. "The Chemistry of Cyanide-Metal Complexes in Relation to Hydrometallurgical Processes of Precious Metals." In: Mineral Processing and Extractive Metallurgy Review. 6: 81-125, 1990. Wei, Y., K. Zhong, E.V. Adamov, and R.W. Smith. "Semi-Continuous Biooxidation of the Chongyang Refractory Gold Ore." In: Minerals Engineering. 10(6) : 577-583, 1997. Weston, J.M. Continuous Biological Leaching of Copper From a Chalcocite Ore and Concentrate in a Saline Environment. Masters Thesis: University of British Columbia, 1996. Whatley, F.R. In: Sulphur in Biology Exerpta Medica: Amste2dam, 1980. White, David. The Physiology and Biochemistry of Prokaryotes. Oxford University Press: New York. 1995. Xiang, Lan, and Jiajun Ke. "Behaviour of Elemental Sulphur in Biooxidation of Gold-Bearing Sulphide Ores." In: Transactions ofNfsoc. 4(4), 42-46, 1994. Zhang, Yongzhu, Yiyuan Lu, Chuanfu Zhang, and Tinaeong Zhao. "Biooxidation of Two Arsenical Refractory Gold Concentrates and Gold Cyanidation." In: Transactions ofNfsoc. 4(3). 41-45. 1994. Zhang, Yu, Zhaoheng Fang, and Mamoun Muhammed. "On the Solution Chemistry of Gold and Silver Bearing Sulphide Ores. A Critical Evaluation of Thermodynamic Calculations." In: Hydrometallurgy. 46:251-269, 1997. 132 A P P E N D I X A - Analytical Methods Residue Analysis Sulfur Species Determination of elemental sulfur (S°), sulfate sulfur (SS04), sulfide sulfur (S2"), and total sulfur (ST) was accomplished by a standard wet chemistry method. The solids are washed in perchloroethelyne to remove S°. The solution is analyzed by precipitating the sulfur as barium sulfate (BaS04). SSQ4 is then be removed from the solids by selective dissolution in Na2C03 solution. Again, the sulfur is precipitated as BaS04. Finally, the solids are assayed for S2" the same method. Total sulfur determination is accomplished by the same method, except that the steps to remove S° and SSQ4 are not taken. Metals Analysis of solid residues for iron and arsenic were done by acid digestion and titration. Solution Analysis Metals Analysis for iron and arsenic in bioleach solutions were accomplished by atomic absorption using a Unicam 929 AA Spectrometer. To determine iron, copper, silver, zinc, and nickel, solution samples from the cyanidation experiments were sent to IPL for inductively coupled plasma (ICP) assay. 133 Gold Analysis of cyanide leach solutions for gold were carried out by IPL using a fire assay method. Free Cyanide Free cyanide is defined as that which is dissociated from the unstable cyanides under alkaline conditions and is available for complexation with metal species. The most common method for determining free cyanide is by titration with silver nitrate (AgN03). In this method, the end point can be determined turbidimetrically or by the use of an indicator. However, the turbidity of the solution, caused by the precipitation of dicyanoargentate, is not easily detected and therefore subject to bias (Lenahen and Murray-Smith, 1986). The indicators on the other hand, can be affected by the presence of heavy metals. Copper and iron cyanide complexes may interfere with the end point. For this work rhodanine was used as the indicator. An alternative method of free cyanide determination is by an amperometric technique (Crundwell, 1991; Crundwell and Jensen, 1994). This method employs measuring the current of a cyanide soluble electrode (e.g. gold, silver, copper) and comparing it against that of a calomel reference electrode. The rate of dissolution, and hence the current, varies linearly with the concentration of free cyanide. This technique is suitable for automated process control applications as it can provide continuous measurement of a slurry without the need for filtration. Also, there is little interference from other species such as halides and chlorides. However, the probe has to be calibrated, against silver nitrate titration, for sample tested. 134 Finally, the concentration of CN" ions can be determined indirectly using atomic adsorption. In this case, the solution sample is contacted with silver wool to form the silver cyanide species, Ag(CN)2". Then, the solution is analyzed for silver concentration, and compared with the original value. From this value the concentration of cyanide ion is calculated. Total Cyanide The total cyanide in solution includes the total available cyanide described previously as well as the metallocyanide complexes. The standard method for determining total cyanide involves phosphoric acid digestion of the sample to break down the complexes formed with metals such as silver, gold, copper, cadmium, mercury, nickel, lead, iron, and zinc. The sample is then distilled for two hours and the evolved hydrogen cyanide gas is absorbed into a solution of sodium hydroxide. Finally, the NaOH solution is then analyzed for CN" by the titration or ion selective electrode methods described previously. Cyanate Analysis of cyanate, CNO", is accomplished using an ammonia electrode (Standard Methods). First, the background ammonia level is recorded using the calibrated probe. Then a sample is acidified and heated in a water bath to 90oC. Hydrolysis of the cyanate ion proceeds as follows: INaCNO + H2S04 + 4H20^> (NH4)S04 + 2NaHC03 The concentration of ammonia in the sample can then be measured, the initial reading subtracted, and the concentration of cyanate back calculated. 135 Thiocyanate The concentration of cyanide is determined by titration with silver nitrate at low pH using ferric nitrate as an indicator. Interference from free cyanide can be eliminated by sparging the acidified solution with nitrogen gas prior to analysis (UBC lab standard method). 136 A P P E N D I X B - Thermodynamic Data All Eh-pH diagrams were generated using the CSIRO thermochemistry program. The use of molarity instead of activity for the construction of the diagrams was deemed acceptable as the concentration of aqueous species were generally quite low (<0.01 M). Note: the Fe-S-CN-H20 diagram uses activities of all species as 1 because the program was unable to generate a diagram at lower concentrations. Formula State Description AG0 [kJ/mol] Au c 0.0 Au(CN)2" aq 285.8 Au(CN)4" aq 804.3 Au033" aq -51.8 HAu032" aq -124.2 H2Au03" aq -218.3 Au(OH)3 aq -283.4 Au(OH)3 c -316.9 CN" aq 172.4 HCN aq 119.7 Fe c 0.0 Fe2+ aq -91.4 Fe3+ aq -16.7 FeOOH c goethite -487.4 Fe203 c hematite -742.2 Fe304 c magnetite -1015.5 Fe(OH)2 c -486.6 Fe(OH)3 c -696.5 H3Fe3(S04)2(OH)6 c jarosite -3230.1 FeS2 c pyrite -166.9 FeS c pyrrhotite -100.4 S c 0.0 s2- aq 86.3 HS" aq 12.1 H2S aq -27.9 S042" aq -744.6 HS04" aq -756.0 S032" aq -486.6 SCN" aq 92.68 references: Bard et al., 1985; Lowson, 1982 137 A P P E N D I X C - Sample Calculations Bioleaching pulp density mass of dry solids pulp density = — * 100% mass oj slurry for example, for Tank #1, Steady State #1, 4hr: 25.47e pulp density = 1 Q 9 1 2 g * 100% = 23.34% lime consumption total Ca(OH)2 added Ca(OH)2 consumed = total solids processed for example, for Tank #1, Steady State #1: Ca(OH) .consumed = ——f- = \03.\s/kg = \03.\kg/ 1.60kg iron extraction mass of iron in solution I litre of slurry iron extraction = — — * 100%) total mass oj iron(solutwn + solids) I litre of slurry where, mass of iron in solution pulpdensity slurry density — = solutionassayWA * (1 —— ) * — litre oj slurry 100 solutiondensity 138 mass of iron in solids solidsassay[%] * pulpdensity * slurrydensity litre of slurry 10 for example, for Tank #1, Steady State #1, 4hr: (2.028^*(1-^)*^) iron extraction = ] 0 0 9.05% 12334* 1.184 * 1 0 ° % = 6 A 6 % (2.028 8/L*Q- ) * ) + V V 100 J U2V 10 arsenic extraction Calculation of arsenic extraction followed the same format as that for iron extraction. sulfide oxidation 2_ (§solids/ \ %S in oxidized sample ^ / mlslurry/feed sulfide oxidation = ( 1 - ( o 2 _ — ) * ( 7 ) * 1 0 0 %S in feed ,gsolid*/ \ K /mlslurry J s 0 m p l e for example, for Tank #1, Steady State #1, 4hr: 4 . 4 0 % ( 2 8 ' % > 2 ) . sulfide oxidation = ( 1 - ^ * ^ 2 1 3 9 / ' ^ ^ 1 0 0 % = 5 6 . 9 % Cyanidation cyanide consumption 139 cyanide added - cyanide left - cyanide lost during sampling cyanide consumed 7 — — mass oj solids for example, for CN-1-1: (.120+.077+.060+.055+.081)g - ( . 0 3 /y. *0.240J - ((.18+.25+.27+.16)g/t*.010L NaCN cons.= - ; 0~\m = 3 - 6 6 percent cyanide consumed as SCN-SCN'assay[ppm] ^glmol NaCN ^_innn/ %NaCN consumed as SCN~ = —— , n J * — ; 7 7 ^ 7 " * 100% CN assay[ppm\ g I mo I SCN for example, for CN-1-1: 1441_ 49.01 , o / %NaCN consumed as SCN~ = , "P * n , / * 100% = 77.4% 140 A P P E N D I X D - Sample Data CONTINUOUS REACTOR LOG SS#1 September 9 to 1 3 , 1 9 9 7 T a n k l Date Time (hr) Temp (°C) PH Eh (mV) D02 Speed (rpm) Air Flow (Lpm) 9-Sep 0 35.0 1.87 791 839 1.46 10-Sep 24 35.5 1.91 789 4.41 840 1.51 11-Sep 48 35.1 1.72 791 3.87 897 1.48 12-Sep 72 33.7 1.78 792 899 1.38 13-Sep 96 37.9 1.78 762 900 1.53 avg 35.4 1.81 785 Tank 2 Date Time (hr) Temp (°C) pH Eh (mV) D02 Speed (rpm) Air Flow (Lpm) 9-Sep 0 35.0 1.24 844 837 0.97 10-Sep 24 36.0 1.38 842 3.77 826 0.87 11-Sep 48 36.3 1.33 844 4.52 885 1.02 12-Sep 72 36.6 1.33 850 887 1.05 13-Sep 96 35.5 1.32 843 895 1.08 avg 35.9 1.32 845 Tank 3 Date Time (hr) Temp (°C) PH Eh - corr (mV) D02 Speed (rpm) Air Flow (Lpm) 9-Sep 0 35.0 1.03 872 834 1.02 10-Sep 24 34.9 1.25 871 5.26 819 1.02 11-Sep 48 35.6 1.22 870 4.74 823 0.95 12-Sep 72 36.2 1.19 875 828 1.08 13-Sep 96 35.8 1.18 869 841 1.13 avg 35.5 1.17 871 Tank 4 Date Time (hr) Temp (°C) pH Eh - corr (mV) D02 Speed (rpm) Ai r Flow (Lpm) 9-Sep 0 34.0 0.89 893 883 1.00 10-Sep 24 36.2 1.10 893 6.11 875 1.05 11-Sep 48 36.1 1.11 890 5.88 881 1.02 12-Sep 72 35.3 1.13 896 883 1.05 13-Sep 96 36.0 1.10 895 1.05 avg 35.5 1.07 893 141 CONTINUOUS REACTOR LOG SS#2 O c t o b e r 20 to 2 4 , 1 9 9 7 T a n k l Date Time (hr) Temp (°C) PH Eh (mV) D02 Speed (rpm) Air Flow (Lpm) 20-Oc t 0 35.7 1.86 798 4.23 823 1.56 21 -Oc t 24 36.7 1.73 794 832 1.56 22 -Oc t 48 34.1 1.85 793 4.26 832 1.56 23-Oct 72 35.3 1.73 794 845 1.48 24-Oct 96 34.9 1.86 793 843 1.48 avg 35.3 1.81 794 Tank 2 Date Time (hr) Temp (°C) PH Eh (mV) D02 Speed (rpm) Air Flow (Lpm) 20-Oc t 0 35.2 1.29 824 2.24 847 0.95 21-Oct 24 35.0 1.37 831 849 0.9 22-Oct 48 33.9 1.33 835 1.95 844 0.9 23-Oct 72 34.8 1.33 830 856 0.95 24-Oct 96 34.8 1.35 827 840 0.97 avg 34.7 1.33 829 Tank 3 Date Time (hr) Temp (°C) PH Eh - corr (mV) D02 Speed (rpm) Air Flow (Lpm) 20-Oct 0 35.5 1.12 853 3.31 864 1.02 21-Oct 24 35.1 1.14 859 861 0.95 22-Oct 48 34.2 1.15 862 3.71 862 0.9 23-Oct 72 34.3 1.17 862 870 0.95 24-Oct 96 33.7 1.16 860 854 0.97 avg 34.6 1.15 859 Tank 4 Date Time (hr) Temp (°C) pH Eh - corr (mV) D02 Speed (rpm) Air Flow (Lpm) 20-Oct 0 34.3 1.03 876 4.13 862 1.02 21-Oct 24 35.1 1.01 880 865 1.00 22-Oct 48 33.4 1.06 883 4.53 864 1.00 23-Oct 72 34.0 1.06 880 874 0.97 24-Oct 96 35.6 1.04 878 857 1.00 avg 34.5 1.04 879 142 CONTINUOUS REACTOR LOG SS#2a O c t o b e r 25 t o 30, 1997 Tank 1 Date Time (hr) Temp(°C) PH Eh (mV) D02 Speed (rpm) Ai r Flow (Lpm) 25-Oct 0 34.0 1.73 797 834 1.53 26-Oct 24 35.3 1.71 796 3.74 847 1.51 27-Oct 48 35.9 1.80 790 836 1.51 28-Oct 72 35.5 1.78 790 3.49 842 1.61 29-Oct 96 34.9 1.77 791 848 1.61 30-Oct 120 36.0 1.83 794 838 1.56 avg 35.5 1.78 792 Tank 2 Date Time (hr) Temp(°C) PH Eh (mV) D02 Speed (rpm) Air Flow (Lpm) 25-Oct 0 34.3 1.32 826 835 1.00 26-Oct 24 34.2 1.35 827 1.94 852 0.97 27-Oct 48 35.8 1.39 830 849 0.97 28-Oct 72 35.1 1.37 826 1.77 848 1.05 29-Oct 96 35.7 1.32 834 871 1.02 30-Oct 120 36.6 1.39 823 853 1.05 avg 35.5 1.36 828 Tank 3 Date Time (hr) Temp(°C) PH Eh - corr (mV) D02 Speed (rpm) Ai r Flow (Lpm) 25-Oct 0 33.7 1.14 858 872 0.97 26-Oct 24 35.9 1.16 854 2.99 868 0.97 27-Oct 48 36.2 1.21 859 864 1.00 28 -Oc t 72 34.2 1.17 859 3.54 860 1.08 29-Oct 96 34.5 1.15 864 874 1.10 30-Oct 120 34.0 1.22 864 863 1.08 avg 35.0 1.18 860 143 CONTINUOUS REACTOR LOG SS#3 N o v e m b e r 5 t o 8 , 1 9 9 7 T a n k l Date Time (hr) Temp (°C) pH Eh (mV) D02 Speed (rpm) Air Flow (Lpm) 5-Nov 0 34.5 1.42 800 841 1.61 6-Nov 24 34.3 1.46 801 3.77 836 1.56 7-Nov 48 34.6 1.49 804 846 1.48 8-Nov 68 35.2 1.53 792 854 1.48 avg 34.7 1.48 799 Tank 2 Date Time (hr) Temp (°C) PH Eh (mV) D02 Speed (rpm) Ai r Flow (Lpm) 5-Nov 0 34.9 1.27 836 863 1.05 6-Nov 24 35.7 1.26 836 2.83 855 1.00 7-Nov 48 35.9 1.25 837 862 0.97 8-Nov 68 32.1 1.24 834 882 0.97 avg 34.7 1.26 836 Tank 3 Date Time (hr) Temp (°C) PH Eh - corr (mV) D02 Speed (rpm) Air Flow (Lpm) 5-Nov 0 34.8 1.13 865 865 1.10 6-Nov 24 35.6 1.12 865 4.89 861 1.10 7-Nov 48 35.5 1.11 865 864 0.97 8-Nov 68 32.6 1.07 862 882 1.00 avg 34.6 1.11 864 144 CONTINUOUS REACTOR LOG SS#4 N o v e m b e r 19 t o 2 4 , 1 9 9 7 T a n k l Date Time (hr) Temp (°C) PH Eh (mV) D02 Speed (rpm) Air Flow (Lpm) 19-Nov 0 38.3 1.56 781 845 1.71 20-Nov 24 34.9 1.46 781 2.86 841 1.63 21-Nov 48 34.1 1.47 785 847 1.66 22-Nov 72 33.8 1.63 782 3.68 841 1.68 23-Nov 96 34.1 1.47 785 848 1.68 24-Nov 120 34.9 1.58 783 3.48 855 1.51 avg 35.0 1.53 783 Tank 2 Date Time (hr) Temp (°C) PH Eh (mV) D02 Speed (rpm) Ai r Flow (Lpm) 19-Nov 0 36.8 1.11 816 860 1.10 20-Nov 24 35.1 1.18 814 2.09 852 1.10 21-Nov 48 34.7 1.08 820 852 1.10 22-Nov 72 33.8 1.27 820 1.43 856 1.10 23-Nov 96 35.4 1.17 817 857 1.10 24-Nov 120 36.5 1.15 826 2.71 863 1.10 avg 35.4 1.16 819 Tank 3 Date Time (hr) Temp (°C) PH Eh - corr (mV) D02 Speed (rpm) Air Flow (Lpm) 19-Nov 0 35.3 0.95 847 863 1.1 20-Nov 24 34.8 1.03 844 4.14 863 1.02 21-Nov 48 35.1 0.94 849 865 1.05 22-Nov 72 35 1.09 848 3.41 864 1.05 23-Nov 96 35.2 1.01 845 862 1.05 24-Nov 120 36.9 1.03 849 3.92 877 1.05 avg 35.4 1.01 847 145 C / ) ( / ) O T ( / ) O T t / ) C / > t O ( A C D C D C D C D C D C D C D C D C D -•—» -t—' -*—' -I—' -4—' 4—1 -*—' -*—1 4—I T 3 " 0 " O T 3 " 0 " 0 " 0 " 0 " 0 l _ k_ l _ I— L_ L_ k- L— 1— corocococororororo T 3 T 3 T 3 T 3 T D T 3 T 3 T 3 T 3 cocorocororocoroco —' 4—' -*—' 4—I 4—' -t—' 4—' -*—1 tntntntntntntntntn tn •O TJ T3 TJ p co T 3 tn tn a c a> <D •o " O a. a. ZJ ZJ a. C L cz o CD O c o o O o o cz 0 o cz o o O cn O o Z J 2 ro (D (D t 1 c .2 .cz A u extr Ca (OH) cons T - C 0 O 5 C 0 0 O C 0 00 0 O 0 O c o i ^ ^ T - o c o T - c N a ^ p c N c q c q c q i n C 0 a i 0 0 00 C 0 0 0 O 5 O ) 0 0 C 0 C 0 C 0 C f ) 0 0 O 5 C O I ^ O I ' - - l O a > C 0 O ' r - 0 0 0 0 0 0 C 0 C 0 C M C O r ; C 0 6 ^ 0 d c 6 c O M N M C \ i ' f ^ : ' 5 l : ( \ I O I > N N N N N T - T - T - I - I - T - ^ T -as Fe(CN) in o * n N o in N 00 CD m o CD CO CN CO 00 o r-00 00 OO 00 CD O h-^  ^ 00 . C O - 15 o as S C N total CO in CO •"3- 00 CO co o CM CD 00 m o 00 •51- CO ih O ) CO O CM CO 00 CD CM CN CO CN CM CD 00 co T— r-- CO in co in CM I-- CO CM CN CO CO CM NT CM If) CD in CD o oo m CO CD CO CN CM CM o O ) CM CO 00 00 o CO o co CO CO CO o CM CM CM CN T— T— CO CM 00 m co in 00 CO CO CO CM CO •sr 00 ro •sr in 00 O O CO CO m m in in in in m m in m m in m in m in m in CM m o * o in * o m o o o o o o o o o o o o o o o o o o o T _ o o m in CD CD CD r>- •t in i-- CN _^ 00 00 CO in oo ^ _ h- in co CO o o o o o o o o o o CM o o o o o o o o CO o o o o o o o o o o o o o o o o o o o o o o o o o 00 CO CO CO CO CO 00 CO CO 00 CO 00 00 CO CN m oo O ) O ) CO h- CM T— CD o r- 00 in co in CO r- CO m CO o o O ) co o o CO CO in CM o CN CO o o o o o o in CD in m CO CO •sr in CD •sr m in in in m c m CD co in 00 oo co o O ) CD co in 00 o o o o T -o CM o o O o o X— d o o o o o co o o o T— CM CM CM CM CM CM CM CM CN CM CM CM CM CN CM CM CM ^~ CM CN CN T CM CN CM CM CM CM CM CM CM CM CM CM CN CN CN CN CM CM CM CN CM CN CN CN cone. N a C N avg. pH tn i cu •4—' CD E co i ro pulp densi ty stirr ing speed tempera ture o durat ion co CO Q c o "co •g 'c CO >< O sample Q C L C L L U t- CM CO « ° - I -Q_ i i i co ro CM CM CM LUC/)C/)(/)C/)C/)C0(/)C/) C/JCOWCVDCOCOCOCO C N ^ C L ro i CM CO I— CM C L I C L C L C L I c o w t o c o w c o w c o w c o c o w c o c o w Test # Q LU LU ro .a o CM co co co • co ro CM CM CM I I I I I I I I I CM ro i CM co CM t- CM CO C L C L C L N 0 0 O ) C L C L C L ' ^ • ^ • ^ • • ^ • • ^ • • ^ • • ^ • s i - ' s r O T -C L C L I I I I 2 o o o o o o o o o o o o o o o 146 0 CL E 0 0 -4—' CD c _3 0 O E T3 0 0 73  Q. 0 ^ - 0 O C L z> CO "D i - O o co 0 J 0 1 I ^ 5 0 0 _- —-CL CL o o X X • o 0 73 73 CO 0 in CO _ o o cz to o X 0 X Q . E Q - 0 & & ° O O D l O B S 0 Z 0 5 p o g> - ° 2 S -£ Q . C L ^ o . !T2 _ W - 01 c O CN -5 CN ™ X O § ) Q -o A u extr. a> o Is-LO O) ^ <N CM Is- CX> O) CO O ^ W 1 -cd r-^  cb co co co •^ r Ca(OH) cons co n to q T -CO 00 00 CN Is-i r i o o CO LO I LO CN CD CN CO CD i n c o r -o oo r: CN CN as Fe(CN) CO 05 cn CN CO I s - CO CN CN O T - 1 i r i E 3 CO c o o z O co as S C N CD CD CN Is - 00 o 00 CD CO cxi CN T— CD 00 -— O CO NJ CN CN CN CN CN CN CN total a> CD •sl-00 00 00 O) CO 00 CX) 00 CO CO 00 o CN oo CO CN CN CN o 00 CD 00 CN Is-co T ° ° T ( X I LO LO T— h - CO o CN •<3-CO CO LO LO CN CN CO CN 00 LO 00 00 cone. N a C N avg. pH to i CD -4—' 0 E co i ro pulp density : L O ^ L O i O L O L O L O o b o o o ^ h h o - ^ o o o o o - - : - - : - - : o __: ,_: < 9 c o c o . 0 0 . o o c o . C D , " : ^ - o o o o N t L O ^ O O O O O T - O O ^ ( O ° N d o o o o o o o o o o o o o o o stirring speed C N C O C N - ^ - O O O T C O T -0 ) > C O ( N O ) O I - 0 TJ- C T - c o ^ r i O L O L O tempera ture o durat ion ro ro g ro ro c c LO c c o o LO oo o c o I s- I s- I s- CD 00 00 a> d CO ai o CN ai c i CN CN CN ai ai ai O CN C> CN ai ai CN CN CN CN CN CN CN CN CN CN CN CN •sr CN CM *t CN CN sample H Q . Q . Q . Q . Q . Q . H Q . Q . Q. Q_ D_ Q_ 0. i i i i i i i i i i i i i i i • ^ • • ^ • • ^ • ^ - ^ • ^ • • ^ - • ^ • ^ - • ^ ^ T 1^" ^ C 7 ) C O C O C O C / ) l 7 3 C 7 ) C O C 7 ) C O C 0 C 0 C 0 C 0 C 0 C O C O C O C / J C O C O C O C O C O C O c o c o c o c o c o Test # CO a. I • ^ r LO C D r - o o c o a. o_ i i • • z z o T -CN CN CN CN C L O O O O O O O O O O CO "fl" LO CO Is-CN CN CN CN CN 'CL.Q_Q_CL.CL. I I I I I 1 t t t l -z.-z.-z.-z.-z. o o o o o 147 A P P E N D I X E - Statistical Analysis This section examines the steady state assay data by evaluating the coefficient of variance for the measurements of Fe, As, S ° , S 2 " , S S 0 4 , and S t o t a | . where, s = V? C = (—)*100% X Residue Assays steady state sample average variance st. dev. coeff of var. jank numbe measurements [% X b a r S 2 S C [ % ] i ron 1-T1 9.05 9.36 9.21 0.048 0.219 2.38 1-T2 7.49 6.76 7.13 0.266 0.516 7.24 1-T3 5.03 5.06 5.05 0.000 0.021 0.42 1-T4 4.29 4.47 4.38 0.016 0.127 2.91 2-T1 9.50 9.37 9.44 0.008 0.092 0.97 2-T2 6.32 6.04 6.18 0.039 0.198 3.20 2-T3 5.10 4.18 4.64 0.423 0.651 14.02 2-T4 3.95 3.40 3.68 0.151 ,0.389 10.58 4-T1 9.34 9.05 9.20 0.042 0.205 2.23 4-T2 7.05 6.69 6.87 0.065 0.255 3.71 4-T3 5.35 4.90 5.13 0.101 0.318 6.21 arsenic 1-T1 2.04 1.96 2.00 0.003 0.057 2.83 1-T2 1.35 1.15 1.25 0.020 0.141 11.31 1-T3 0.60 0.60 0.60 0.000 0.000 0.00 1-T4 0.57 0.48 0.53 0.004 0.064 12.12 2-T1 1.84 1.72 1.78 0.007 0.085 4.77 2-T2 0.51 0.69 0.60 0.016 0.127 21.21 148 2-T3 0.39 0.42 0.41 0.000 0.021 5.24 2-T4 0.34 0.35 0.35 0.000 0.007 2.05 4-T1 1.43 1.08 1.26 0.061 0.247 19.72 4-T2 0.61 0.58 0.60 0.000 0.021 3.57 4-T3 0.40 0.45 0.43 0.001 0.035 8.32 1-T1 4.40 4.01 4.21 0.076 0.276 6.56 1-T2 3.06 2.53 2.80 0.140 0.375 13.41 1-T3 2.02 1.43 1.73 0.174 0.417 24.19 1-T4 1.00 0.86 0.93 0.010 0.099 10.64 2-T1 7.12 6.82 6.97 0.045 0.212 3.04 2-T2 5.78 4.64 5.21 0.650 0.806 15.47 2-T3 4.36 3.19 3.78 0.684 0.827 21.92 2-T4 3.39 3.15 3.27 0.029 0.170 5.19 4-T1 8.48 8.25 8.37 0.026 0.163 1.94 4-T2 6.39 6.33 6.36 0.002 0.042 0.67 4-T3 4.98 4.69 4.84 0.042 0.205 4.24 1-T1 9.82 9.70 9.76 0.007 0.085 0.87 1-T2 8.50 8.71 8.61 0.022 0.148 1.73 1-T3 7.41 7.62 7.52 0.022 0.148 1.98 1-T4 6.53 6.74 6.64 0.022 0.148 2.24 2-T1 9.93 10.94 10.44 0.510 0.714 6.84 2-T2 8.22 8.59 8.41 0.068 0.262 3.11 2-T3 7.35 7.04 7.20 0.048 0.219 3.05 2-T4 5.53 5.49 5.51 0.001 0.028 0.51 4-T1 10.43 10.11 10.27 0.051 0.226 2.20 4-T2 8.62 8.61 8.62 0.000 0.007 0.08 4-T3 7.36 6.99 7.18 0.068 0.262 3.65 S s 0 4 1-T1 5.03 5.26 5.15 0.026 0.163 3.16 1-T2 4.91 5.50 5.21 0.174 0.417 8.02 1-T3 4.78 5.58 5.18 0.320 0.566 10.92 1-T4 4.98 5.35 5.17 0.068 0.262 5.07 1-T1 0.24 0.22 0.23 0.000 0.014 6.15 1-T2 0.31 0.26 0.29 0.001 0.035 12.41 1-T3 0.30 0.28 0.29 0.000 0.014 4.88 1-T4 0.30 0.27 0.29 0.000 0.021 7.44 149 _5 o CO ro o • O o o > C D -O CO -f—* C O < D 1— * ro * c CO ro > C D ro ro S> X ro .OJ C O u c Q . C D E E ro < D co -C O ro C D E o o c o r S o o c N ^ - ^ c o o o c o o i n u ^ o o c n c o c \ i c o c \ i q - 5 f - r ^ O ^ O J C D O O L O ' N ' O J l O C O n i o s o o j o c o ^ i o o c o O O C O N t D t O T - O J O f f i N O O O O O O LO OJ Is -O O CO CN O •sr CN LO CN ° CN OJ CO o LO CN \ — CO CO LO CO LO d T— T — d d LO o OJ CO OJ o I s- LO CD co 00 CN LO h-' ^ — • CN CN CN CO CO oS LO CO -tf 00 oo t- CN 00 o h-CD CO CD CN LO CD CD d CO CO CN CN CN CT> OJ o o o CO o CN 00 00 LO O LO I s- CO o> Is -00 d CO CD d CN CN oo CN CO o 00 o CN LO LO CO CD I s -I s - o CN LO LO CN d CO 00 OJ 00 CN CD CN d d CN H -T2 -T3 ? H -T2 -T3 -T4 H -T2 -T3 CN CN CN CN 4 0) LO LO CO oo CD OJ Is- CD CD 00 LO CD CD LO o co OJ 00 LO d 1^  CO LO OJ LO LO oo CO O O O h - C O x - C N - ^ - O O f - C D O C N S O ^ C O t O ^ L O C N C N d d d d d d d d d d d O ^ O I O O ) o o -sr o T-d d d d d d o co cn ^ s T— CO T— CO o o ' d d d d d r- 00 CO OJ 00 •sj" CD CD 00 CD ^ CO c \ i i r i ( D f \ i i r i ( D N c 6 i o s CO 00 CO T -00 LO O CN r— LO CN O T— OJ O I s - CO CN LO 1^  I s - 00 LO CO 5" OJ T— 00 OJ CO CD o CO CN T— o CN CD CD un CN CO CN 00 I s-CN CD LO I s- LO r -O CN M" LO CO 00 / — \ CN CN I s- CN LO OJ I — ' CD LO LO I s- CO CN CO 00 LO I s-CD C D C N C O C D f o C N C O C N , * LO CN T - co — U J CN LO CD CN LO CD I s -oo LO T " C N CO ^ r -I— i— I— I— I— CM co -^ r I— I— I— to < T - CN CO I— I— I— C M CM C M C N ^ 150 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items