Peter Whincup, Senior Project Manager, Cullen Mining Services. Marcus Binks, Senior Metallurgist, Perseverance Corporation Fosterville Mine

Transcription

1 A CASE STUDY OF THE DEVELOPMENT, ENGINEERING AND CONSTRUCTION OF A BIOLOGICAL OXIDATION PLANT FOR A REFRACTORY PYRITE- ARSENOPYRITE GOLD ORE Peter Whincup, Senior Project Manager, Cullen Mining Services. Marcus Binks, Senior Metallurgist, Perseverance Corporation Fosterville Mine Peter Whincup Cullen Mining Services Locked Bag 1003 Sandringham Vic 3191 Telephone (03) whincupp@cullenmining.com.au Bac-Min 2004 Conference

2 A CASE STUDY OF THE DEVELOPMENT, ENGINEERING AND CONSTRUCTION OF A BIOLOGICAL OXIDATION PLANT FOR A REFRACTORY PYRITE- ARSENOPYRITE GOLD ORE Peter Whincup, Senior Project Manager Cullen Mining Services. Locked Bag 1003 Sandringham Vic 3191 Telephone (03) whincupp@cullenmining.com.au Marcus Binks, Senior Metallurgist Perseverance Corporation Fosterville Mine McCormicks Rd., Fosterville Vic 3557 Telephone (03) MarcusB@perseverance.com.au

3 A CASE STUDY OF THE DEVELOPMENT, ENGINEERING AND CONSTRUCTION OF A BIOLOGICAL OXIDATION PLANT FOR A REFRACTORY PYRITE- ARSENOPYRITE GOLD ORE Peter Whincup, Senior Project Manager, Cullen Mining Services. Marcus Binks, Senior Metallurgist, Perseverance Corporation Fosterville Mine Abstract This paper presents a case study that describes the selection, development, engineering and construction of a process plant for the biological oxidative pretreatment of a refractory pyrite-arsenopyrite gold ore flotation concentrate. Prefeasibility and feasibility study resource modeling, metallurgical testwork, and environmental and engineering studies are outlined. Options considered for oxidation of the ore, which examine cost, risk, operability, recovery, engineering and environmental aspects are compared, with particular reference to pressure and biological oxidation. The BIOX biological oxidation process was selected as the preferred oxidation process and the flowsheet for this process is presented. Selection of the BOIX process resulted in special requirements for the feed preparation circuit, residue treatment and disposal, materials of construction and engineering. Use of a biological process in ore treatment also introduces particular process and engineering risks the management of which are discussed.

4 Table of Contents Introduction...1 Resource Base...2 Treatment Options...4 Roasting...6 Activox...6 High Temperature Redox...6 Albion Process...6 Bacterial Oxidation...6 Pressure Oxidation vs BIOX...7 Metallurgical Testwork and Process Design...9 Comminution...10 Oxidation...11 Cyanide Leach...18 Continuous Pilot Scale Testing...22 Flowsheet...22 Inoculum Build-Up...24 Plant Design and Engineering Considerations...25 Environmental...25 Water Supply and Quality...26 Provision for Future Expansion...26 Corrosion...27 Blending...27 BIOX Plant Materials of Construction...27 i

5 Plant Control System...28 BIOX Cooling and Aeration Requirements...29 Process and Engineering Risk Management...29 Water Requirements...29 Flotation Performance...29 Sulphur Oxidation and BIOX Plant Capacity...30 Power Supply Interruption...30 ii

6 INTRODUCTION This paper presents a case study of the development and engineering of a process including biological oxidation for treatment of refractory gold ore from Perseverance Corporation s Fosterville Gold Mine. The project is located some 20 km east of Bendigo in Central Victoria on the historic Ellesmere Goldfield and as shown on Figure 1. Figure 1: Project Location Mining at Fosterville has taken place intermittently since Contemporary exploration and heap leach operations commenced in the 1980 s and Perseverance produced a total of 7.5 t of gold from these operations. By December 2001 open pit oxide ore reserves were exhausted and mining operations ceased. During 2001 and 2002 the main activities were drilling of the primary deposits and leaching of existing heap leach pads. 1

7 Investigations into mining and processing of deeper primary ore commenced in 1992 and in 1997 feasibility proposals based on open pit mining and processing of 500,000 t/y of refractory primary ore using bacterial oxidation were completed. Following completion of deeper drilling in 2001 and 2002 a further detailed feasibility study was undertaken in 2003 (Perseverance Corporation Limited, August 2003) based on mining and processing of refractory primary open pit and underground ore at a rate of 800,000 t/y and which forms the basis of this paper. Engineering of the ore processing plant by Aker Kvaerner Australia commenced in November Plant construction commenced in March 2004 and production will commence early 2005 and it is anticipated production over the initial years of the project will underpin development of additional orebodies and an expanded operation. Cullen Mining Services was appointed Project Manager for engineering, procurement and construction or the ore processing plant This paper presents a study that describes the selection, development, engineering and construction of a process plant with emphasis on the biological oxidative pretreatment of a refractory pyrite-arsenopyrite gold ore flotation concentrate. RESOURCE BASE Resources on which the feasibility study was based are contained within several shoots off the central part of the Fosterville Fault zone and are shown in long section on Figure 2. Figure 2: Long Section of the Central 2.5 km Area of Resource and Exploration Targets (Perseverance Corporation 2003 Annual Report) 2

8 Ore will be won by open pit and underground mining methods and the feasibility study mining reserves are summarised in Table 1. Table 1: Central Zone Reserves (Perseverance Corporation 2003 Annual Report) Reserve Tonnes Grade Contained Gold % S % Fe % As g/t Au Oz Open Pit 1,748, ,000 Underground 4,348, ,000 Total 6,096, ,000 Ore reserves shown in Table 1 represent reserves contained only in the upper shoots off the Central Zone of the Fosterville Fault and it is expected that deeper drilling of this area and drilling of previous shallow sulphide intersections along strike will add to reserves. Gold mineralisation is represented visually as disseminated arsenopyrite and pyrite forming a halo to veins in a quartz carbonate veinlet stockwork, which is in turn controlled by late brittle faults. The arsenopyrite occurs as fine-grained (0.05 mm to 1 mm), acicular needles to 6 mm in length with no preferred orientation. The disseminated pyrite associated with gold mineralisation occurs as crystalline pyritohedrons 0.1 mm to 2 mm in size. Mineralisation is readily amenable to concentration by sulphide flotation with sulfur recoveries on average in excess of 98%. Concentrate mineralogy calculated from assays is shown in Table 2. Table 2: Calculated Concentrate Mineralogy Mineral % Weight Pyrite 30.5 Arsenopyrite 13.8 Stibnite 1.0 Total Sulphides 45.4 Gangue 54.6 Dense media separation and electron microprobe analyses of flotation concentrate indicated that the arsenopyrite contains 100 ppm Au to 1000 ppm Au and the auriferous pyrite 10 ppm Au to 100 ppm Au. 3

9 Flotation concentrate was shown to contain very little liberated native gold with the bulk of the gold contained within pyrite and arsenopyrite (Amdel, 1994). Although pyrite was about twice as abundant as arsenopyrite it contained less of the gold than arsenopyrite. Typical gold analysis were: Pyrite 35 ppm Arsenopyrite 290 ppm That is, the arsenopyrite contained some 80% of the gold. Work is continuing on preparation of detailed resource and mining models that include, as well as metal value and cost data, mineralogy, sulfur, arsenic, iron, organic and carbonate carbon, and a range of metallurgical parameters. TREATMENT OPTIONS Fosterville primary ore is highly refractory generally exhibiting less than 10% cyanide leach extraction after grinding to a size of P 80 = 45 micron and requiring oxidative pretreatment to release gold from pyrite and arsenopyrite. The grade and flotation response of Fosterville ore indicated oxidative pretreatment of a flotation concentrate rather than whole ore. Oxidative pretreatment options considered were: BIOX bacterial oxidation of flotation concentrate. BacTech bacterial oxidation of flotation concentrate. Pressure oxidation of flotation concentrate. Roasting. High Temperature Redox. Activox. Table 3 summarizes key aspects of each of these processes. 4

10 Process Process Summary Advantages/Disadvantages BIOX Bacterial oxidation of flotation concentrate by Moderate capital and operating costs. acidithiobacillus ferroxidans, At.thiooxidans and Established technology. Leptospirillum ferrooxidans at controlled temperature and Relatively simple and low cost management of arsenic bearing ph. Goldfields Ltd (GFL) owns process technology. residue. More difficult to achieve complete S oxidation relative to roasting and pressure oxidation. BacTech Roasting Pressure Oxidation High Temperature Redox Activox Bacterial oxidation of flotation concentrate by thermophilic bacteria similar to those above. BacTech Mining Corporation owns process technology. Oxidation of sulphides contained in whole ore or flotation concentrate at elevated temperature in the presence of excess air. Aqueous acid oxidation of sulphides contained in flotation concentrate or whole ore using oxygen at elevated temperature and pressure. Pressure oxidation using nitric acid as a catalyst at elevated temperatures. Fine grinding of concentrate followed by oxidation with oxygen under pressure at low temperature. Western Minerals Technology owns process technology. Requires water of certain quality. Moderate capital and operating costs. Relatively simple and low cost management of arsenic bearing residue. More difficult to achieve complete S oxidation relative to roasting and pressure oxidation. Requires water of certain quality. Low capital and operating costs. Established technology. Complete sulphur oxidation. Suitable for ore containing detrimental carbonaceous material. Difficulties in environmental management and disposal of sulphur and arsenic residues. High capital and operating costs. Established technology. Relatively simple and low cost management of arsenic bearing residue. More complete S oxidation relative biological oxidation. Rapid reaction with good S oxidation. High capital and operating costs. Not established technology. Low capital and operating costs. Not established technology. Table 3: Oxidative Pre-Treatment Options 5

11 Roasting Use of roasting for oxidative pretreatment of refractory sulphide ores, particularly those containing arsenopyrite, has become increasingly less acceptable over the pat 20 years due to both environmental and human health concerns. Environmental considerations including management of sulfur dioxide and arsenic oxides precluded serious consideration of the use of roasting at Fosterville. Activox A series of Activox tests on flotation concentrate sample showed that high sulphide oxidation, <90%, was achievable but gold recoveries were poor, with a best result of 63.6%. For this reason and that introduction of relatively commercially unproven technology would expose Perseverance to an unacceptable risk there was no further consideration of the Activox process. High Temperature Redox A sample of flotation concentrate was successfully oxidised using a batch Redox process. In excess of 99% of the sulphidic sulfur was oxidised with 98.9% of the gold subsequently extracted by cyanide leach. However, an oxidised residue from a test that lost pressure and nitric acid gave only 59% gold extraction, which indicated the probability of substantial gold loss following upset process conditions. Issues of disposal of excess nitric acid also contributed to the decision not to proceed with the process. Albion Process A further process that was considered but for which no testwork was undertaken was the Albion process owned by Xstrata Plc., which uses fine grinding of concentrate followed by oxidative leaching at atmospheric pressure and elevated temperature. Claimed advantages include lower power requirements and lower environmental risk when compared to bacterial oxidation. However, as with other non-established technologies, its use was considered too high a risk for the project without extensive pilot plant testing. Bacterial Oxidation Early testwork using both BIOX and BacTech technologies indicated Fosterville ore was amenable to both processes. However, results of early cyanide leach testwork on both BIOX and BacTech oxidised flotation concentrate indicated better recovery from the BIOX product. In addition issues relating to the operation of the Beaconsfield plant that used 6

12 BacTech technology were taken into account in making the decision that BIOX technology would be the preferred biological oxidation technology for Fosterville ore. Pressure Oxidation vs BIOX Testwork also indicated that relative to BIOX, pressure oxidation (POX) could yield a 2% increase in Carbon-in-Leach (CIL) cyanide leach gold recovery for the same percentage sulfur oxidation and for this reason a comparative study of BIOX and POX oxidation was undertaken prior to completion of the feasibility study. The study involved developing comparative nominal +30% capital and operating cost estimates and comparing the two processes using a range of other criteria. Comparative capital and operating cost estimates are respectively shown in Tables 4 and 5. Readers should note these are comparative costs and do not reflect currently projected capital and operating costs. Cost Area Table 4: Summary of Comparative Capital Costs BIOX $M POX $M Direct Costs: Oxidation and CIL Capital Other Capital Subtotal - Direct Costs Indirect Costs: EPCM (15%) Contingency (15%) Plant Total Table 5: Summary of Comparative Operating Costs Cost Area BIOX $/t ore POX $/t ore Labour Consumables Power Maintenance Laboratory Total $/t ore In addition to estimating capital and operating costs and incremental recovery differential, Perseverance staff visited both POX and BIOX plants to asses operating criteria, the more significant of which included: 7

13 Safety. Environmental acceptability. Capital and operating cost risk. Ease of expansion. Technical risk. Implementation and ramp-up time. Operability, maintainability and employee skill requirements. Water quality sensitivity. Spares holding. Sensitivity to variation in feed sulfur grade. These and other criteria were evaluated using a semi quantitative analysis process results of which indicated with reasonable certainty that BIOX would be the more suitable of the two technologies for processing a concentrate produced from Fosterville ore at a rate of 800,000 t/y. This should not be taken as an endorsement of biological oxidation over pressure oxidation for all situations. This review of the two options also indicated that POX could possibly be more competitive for a larger or longer life project. The assessment showed the following to be the most significant advantages of BIOX and POX with respect to the Fosterville Project: BIOX POX Shorter implementation time. Shorter ramp-up time. Better maintainability and operability. Lower technical risk. Ease of expansion. Lower water quality sensitivity. 8

14 METALLURGICAL TESTWORK AND PROCESS DESIGN The metallurgical test program undertaken on ore samples, excluding testwork on carbonaceous ore is summarised in Figure 3. In addition to this work comminution and neutralization testwork was undertaken on limestone samples. Figure 3: Test Program Whole Core Samples Comminution UCS Impact Work Index Abrasion Index Rod Mill Work Index Ball Mill Work Index JK SAG Parameters Batch Flotation Optimum grind Au S Fe As recovery, mass recovery & concentrate grade Reagent regime & consumption Residence times Concentrate and tails rheology Concentrate and tails settling data including flocculant testing Variability Mineralogical Analysis Chemical analysis Optical mineralogy XRD Batch BIOX Process water amenability Ore amenability Oxidation rate & extent of oxidation Lime & limestone consumption Solid and liquid products geochemistry Product rheology Oxidised product settling data including flocculant testing Batch Cyanide Leach Gold extraction Leach kinetics Carbon loading & kinetics Oxygen uptake Reagent consumption Residue geochemistry Residue rheology Batch Detoxification Process evaluation Residence times Reagent consumptions Residue geochemistry 9

15 Comminution Comminution testwork was undertaken on samples from all ore zones and comprised: Crushing Work index. Rod mill Work Index. Ball mill work Index JK Drop Weight Test Parameters. Ore SG. Abrasion Index. The comminution data were evaluated and it was concluded that Fosterville ores would be amenable to milling in any one of the following three circuits: Single stage, semi autogenous grinding (SAG) milling. SAG / ball mill combination. Three stage crushing and ball milling. Selection of the comminution circuit for the plant considered the following: Capital and operating costs. Need to optimize grind size to meet size requirements for BIOX feed, nominally P 80 =45 micron, and minimise overgrinding of sulphide minerals, in particular arsenopyrite. Risk of critical size build-up in a single stage SAG milling circuit with some ores. Possibility of future plant capacity expansion. Single stage crushing followed by a single stage SAG mill closed by hydrocyclones was selected as the preferred circuit configuration. The top ball size and ball charge specified for the SAG mill is expected to eliminate any likelihood of critical size build-up in the mill, however provision was made in the plant layout for a recycle crusher should grinding of some ores result in a significant critical size build-up. 10

16 Flotation Average results of single stage rougher flotation testwork on a suite of 12 samples representing the expected range of ore types and shown in Table 6 were used for process design. Table 6: Cumulative Average Flotation Concentrate Data Flotation Concentrate Cumulative Laboratory Flotation Time min Concentrate Mass Pull % Gold Au g/t Cumulative Flotation Concentrate Assays Sulphur S % Arsenic As % Gold Au % Cumulative Recovery Sulphur S % Arsenic As % Concentrate Concentrate Concentrate Concentrate Selected Design Point Average iron recovery was 57% reflecting the presence of a quantity of non-sulphide iron minerals. The flotation circuit comprises a conditioning tank followed by one roughing and two scavenging stages The combined rougher/scavenger concentrate will be directed to a single stage of cleaning with the cleaner tails returned to the rougher feed. Provision is made to direct rougher concentrate to the final concentrate stream. Flotation concentrate will be pumped to a 48 h capacity BIOX feed surge tank and, as necessary, to a 7-day capacity concentrate storage pond. Flotation residue will be pumped to a conventional residue storage structure. Oxidation BIOX testwork was undertaken intermittently over a period of some 7 years on a wide range of flotation concentrate samples and under a variety of conditions. Results of early testwork indicated: The BIOX culture adapted easily, grew actively, and exhibited negligible lag period for commencement of bacterial growth and subsequent sulphide oxidation. Water toxicity tests indicated that proposed plant water sources were not inhibitory to the activity of the BIOX culture. 11

17 Greater than 95% sulphur oxidation of the concentrate was achieved after 18 days in a batch BIOX reactor operating under the following conditions: ph 1.5 to 1.6. Feed Size not greater than P 80 of 75 microns. Temperature Range 40 C to 42 C. Liquid to solid mass ratio of 4:1 or 20% solids by weight. The oxidation rate of concentrate is shown in Figure 4. Final BIOX liquors had soluble iron to arsenic (Fe:As) molar ratios exceeding 3:1 and generated stable basic ferric arsenate precipitates on neutralization. A gold cyanide leach extraction of only 88% was achieved in one test and it believed this was due to gold associating with carbonaceous material as calcining of the final products resulted in a gold extraction of 96%. CIL cyanide leaching of a subsequent BIOX residue from a sample free of carbonaceous material yielded a gold extraction of 94%. The concentrate was acid generating, thus a continuous BIOX circuit would require no acid for ph control and would require a source of limestone for low cost ph control and BIOX liquor neutralisation. Figure 4: Sulphur Oxidation Rate Fosterville Sulphide Concentrate Sulphur Oxidation Rate % Sulphur Oxidation BIOX Oxidation Time 12

18 Sample From the sulfur oxidation results determined from the batch laboratory testwork GFL determined a design residence time of 5 days. Concentrates derived from composites samples representing the major orebodies and ore types and which contained estimated mining dilution were tested as part of the feasibility study. Head assays of these samples are shown in Table 7. Table 7: Flotation Concentrate Head Assays of Orebody and Variability Composites Gold g/t Au Silver g/t Ag Antimony g/t Sb Total Iron % Fe TOT Total Sulphur % S TOT Sulphide Sulphur % S 2- Arsenic % As Total Carbon % C TOT Organic Carbon % C ORG Carbonate % CO , , , , , The head assays indicated the following: Gold grades varied significantly and reflect variation in head grade of ore samples used to generate the variability composites. Silver grades were typically low with a gold to silver ratio typically greater than 25:1. However the sample 8 concentrate had a gold to silver ratio of less than 4:1. Consideration was therefore given to the impact of high silver ores on the elution circuit capacity. Concentrates 5, 7 and 9 had a high antimony concentration. Antimony in sufficient concentration in BIOX feed has been known be detrimental to bacteria activity. Testwork however showed little or no effect on the BIOX process. Some concentrate samples, particularly 3, 5 and 7, showed a significant difference between total sulphur and sulphide sulphur levels, which may indicate some degree of oxidation of sulphide minerals, either in situ or as the result of sample oxidation. 13

19 BIOX oxidation results are summarised in Table 8. Sample ID Table 8: Summary of BIOX Orebody and Variability Testwork Biological Oxidation Time days Mass Gain % Reagent Consumption Lime Net Acid kg/t BIOX kg/t BIOX Feed Feed Sulphide Oxidation % Arsenic Dissolution % Average The following comments relate to the BIOX testwork on the Fosterville variability samples: A positive mass gain was seen in most BIOX tests, averaging 10% but as high as 35%. Some variance was seen between the mass gain for duplicate with different leach times samples possibly indicating variations in leaching conditions between the two samples. Lime consumption varied between 100 kg/t and 200 kg/t of concentrate (BIOX feed) with an average consumption of 140 kg/t of concentrate. All concentrate samples tested were net acid producing. The degree of sulphur oxidation ranged between 71.1% and 99.9% averaging 96.6%. Eliminating the lowest sulphur oxidation result (71.1% for sample 8) narrows the range of sulphur oxidation to 91.5% to 99.9% and increases the average to 98.1%. 14

20 There appeared no mineralogical reason for the poor extent of oxidation seen for the first sample 8, as a duplicate test on the sample achieved 99.9% sulphur oxidation. Arsenic dissolution was highly variable between the concentrate samples, most likely due to varying extents of arsenic re-precipitation during the BIOX tests. Duplicate BIOX tests on samples 3 and 5, which are at the upper and lower levels of the major Phoenix orebody, achieved 96.8%, 97.6%, 98.7% and 99.3% sulphur oxidation respectively indicating minimal variation with depth in the deposit. Duplicate BIOX tests on sample 7, which represented the highest underground antimony mineralisation in the Phoenix orebody, achieved 99.2% and 98.9% sulphur oxidation indicating little or no detrimental impact of the antimony in the concentrate sample on the BIOX process. 15

21 BIOX circuit design parameters are shown in Table 9. Table 9: Summary of BIOX Circuit Design Criteria Parameter Units Value BIOX Circuit Nominal Feed Rate t/h 9 BIOX Residence Time Primary Reactors d 2.5 Secondary Reactors d 2.5 Number of Reactors Primary Reactors No. 3 in parallel Secondary Reactors No. 3 in series BIOX Pulp Density Liquid:Solids 4:1 Operating Temperature C 42 Mass Gain % 11 Sulphur Oxidation Design % 100 Nutrients K As potassium kg/t conc 0.5 N As nitrogen kg/t conc 1.0 P As phosphorous kg/t conc 0.2 Limestone As 100% CaCO 3 kg/t conc 415 CCD Circuit Specific Settling Rate (Flux) BIOX Residue t/m 2 h 0.18 Underflow Solids BIOX Residue % w/w 33 solids Number of Stages No. 3 Wash Ratio Water : Solids m 3 :t 7.8 Liquor Neutralisation Neutralisation Residence Time h 6 Limestone As 100% CaCO 3 kg/t conc 213 Hydrated Lime As 100% Ca(OH) 2 kg/t conc 26.5 The oxidation circuit is a typical BIOX plant comprising the following: Primary Oxidation The bulk of the oxidation will take place in three parallel agitated, aerated tanks each of nominally 1,000 m 3 capacity and 2.5 days residence time. Ground limestone slurry will be used to control ph and cooling water circulated through the tank baffles will control temperature. Provision was made to take individual tanks offline. Nutrient solution from a nutrient mixing facility will be metered to each tank, which together with carbon dioxide liberated from the added limestone will provide for bacterial growth. During oxidation sulfur, iron and arsenic will be solubilised. 16

22 Secondary Oxidation Oxidation will be completed over an additional 2.5 days in three 1000 m 3 agitated, aerated tanks in series with and also supplied with ground limestone, cooling water and nutrients. Provision was made to bypass individual tanks. Separation Of Oxidised Product And Liquor Slurry from the last secondary reactor will gravitate to the first of a series of three counter current decantation (CCD) thickeners where the bulk of the solubilised sulfur, iron and arsenic will be removed in a separate solution stream. A portion of the BIOX product will be side streamed to a separate small thickener, the overflow of which will be used for CIL residue detoxification. Underflow from this thickener will be pumped to the first CCD thickener. Underflow from the last thickener will be pumped to a ph adjustment tank where the ph will be raised to neutral using lime slurry ahead of a cyanide leach circuit. Neutralization During the BIOX process iron, sulfur and arsenic will be solubilised to Fe 3+ (as ferric 2-3- sulphate), SO 4 (as sulphuric acid) and AsO 4 (as arsenic acid) respectively. The neutralization circuit will consist of six aerated and agitated 100 m³ tanks in series and bypass launders will individual allow tanks to be taken off line for cleaning and maintenance. The BIOX liquor will be neutralized in two stages. In the first stage ground limestone slurry will be added to the first 2 or 3 tanks to raise the ph to 4 to 5 where iron and arsenic will be co-precipitated as ferric hydroxide and stable basic ferric arsenate (FeAsO 4. xfe(oh) 3 ). In the second stage the ph will be raised to 6 to 8 in the remaining tanks using limestone and lime slurries to precipitate the sulphuric acid as calcium sulphate (CaSO 4.2H 2 O). The two-stage process will provide for: Precipitation of basic ferric arsenate under controlled ph conditions for optimum stability. Optimal of the use of limestone and the more expensive lime. Minimizing formation of excessive jarosites and other undesirable precipitates that may be hard to settle. The neutralized effluent will be combined with flotation residue. 17

23 Effluent will be drawn from the third or fourth tank by effluent recycle pumps and recycled to the head of the neutralization tank train to provide seeding for precipitation. This will increase particle size and improve the settling characteristics of precipitates. Cyanide Leach Oxidised residues from the nine orebody and variability composite samples we subject to cyanide leach testwork under standard, non-optimised CIL conditions with high excess cyanide levels to maximise gold extraction. Gold extraction from direct cyanide leach of the as received concentrate samples ranged from less than 1% to 38% with most samples in the range of 1% to 5%. This confirmed the refractory nature of Fosterville primary ore. The following comments relate to the cyanide leaching of the duplicate residues from the BIOX testwork: Lime consumption varied between 1 kg/t and 53 kg/t of concentrate (BIOX feed) with an average consumption of 9.9 kg/t of concentrate. At the high non-optimised cyanide addition rates used in the cyanide leach testwork cyanide consumption varied between 12 kg/t and 28 kg/t of concentrate (BIOX feed) with an average consumption of 17 kg/t of concentrate. Gold extraction ranged between 88% and 96% averaging 92.7%. Eliminating a high antimony sample narrowed the range in gold extraction to 90% to 96% and increases the average to 93.1%. This value was adopted as the design gold extraction in the CIL circuit. Duplicate cyanide leach tests on samples which represented the highest underground antimony mineralisation in the Phoenix orebody, achieved the lowest gold extraction at 88.5%, 91.7%, 89.3% and 89.7%, and the highest lime and cyanide consumptions. This indicated that blending of high antimony ore might be required. The duplicate BIOX tests on samples from the upper and lower levels of the major Phoenix orebody, achieved 94.1%, 90.3%, 95.8% and 94.7% gold extraction respectively indicating minimal variation in metallurgical response with depth in the deposit. The averages of results of non-optimised CIL cyanide leach testwork are shown in Table 10. Table 10: Average Cyanide Leach of BIOX Orebody and Variability Residues 18

24 Testwork Method Reagent Consumption Lime kg/t BIOX Feed Sodium Cyanide kg/t BIOX Feed Residual Cyanide CN FREE ppm Gold Extraction % CIL , Subsequent cyanide optimisation testwork resulted in determination of an optimum cyanide consumption of 9.7 kg/t CIL feed and this was adopted for design. Leach rate testwork indicated that the leach curve flattened at about 30 h, however for design 48 h leach/adsorb capacity was provided in six CIL stages for the following reasons: Provision of some inherent over capacity to allow for increased plant throughput. Provision of 6 stages was considered prudent in view of the high CIL feed metal content. It is proposed to direct spent elution liquor to CIL circuit that will result in periods of less than design pulp density and hence less residence time per unit volume. This circuit will need be operated at an atypically low pulp density due to the high pulp viscosity resulting from of the BIOX process. There is a risk that in practice it may not be possible to operate at design density due to, for example, inter-stage screen capacity and the pulp density may need to be reduced. On the basis of CIL testwork at varying pulp density and GFL experience on other projects 33% w/w solids pulp density was selected as the design CIL pulp density. Other testwork included carbon loading and loading kinetics, oxygen uptake, and viscosity. 19

25 CIL circuit design parameters are shown in Table 11. Table 11: CIL Design Parameters Parameter Unit Value Circuit Configuration CIL Nominal Feed Rate t/h 10 Number of Stages 6 Pulp Density % w/w solids 33 Leach Retention Time h 48 Dissolved Oxygen Levels mg/l Carbon Concentration g/l Initial Cyanide Concentration mg/l NaCN 1,500 Residual Free Cyanide mg/l NaCN 1,000 Pulp ph Lime Addition (90% Ca(OH) 2 ) kg/t CIL Feed 6.0 Cyanide Consumption kg/t CIL Feed 9.7 The cyanide leach and metal recovery circuits will comprise a conventional six stage CIL circuit, carbon acid washing, pressure Zadra elution, electrowinning onto steel wool and smelting. Detoxification The CIL residue will contain atypically high levels of free and weak acid dissociable (WAD) cyanide, thiocyanate and arsenic in solution (as As V ) when compared to plants treating oxide gold ores. Design discharge limits for the CIL residue for WAD cyanide and soluble arsenic were set at 50 ppm and 2 ppm respectively. Preliminary testwork was undertaken on BIOX detoxification methods; CIL residue samples using several SO 2 /Air using sodium metabisulphite and copper sulphate catalyst. Hydrogen peroxide. Caro s Acid. BIOX product liquor. Testwork using sodium metabisulphite resulted in a reduction in WAD cyanide to below design limits at a dosage rate of up to 14 kg/t residue solids. Dosage rates of hydrogen peroxide of up to 6.7 kg/t residue solids were required to reduce WAD cyanide to below the design limit. 20

26 Results of testwork using the equivalent of 15 kg/t residue solids of Caro s acid failed to reduce the WAD cyanide to below 50 ppm, probably due to the very high thiocyanate level of 2,500 ppm in the CIL solution. Use of SO 2 /Air, peroxide or Caro s acid circuit to remove cyanide would have two significant shortcomings; cost and inability to remove soluble arsenic, requiring further processing of the residue. Several series of tests were undertaken in which BIOX CIL residue samples were reacted with BIOX oxidation testwork liquor and the residue subsequently neutralized to neutral ph in a two stage process using ground limestone and lime respectively. The testwork was aimed at demonstrating that BIOX liquor could be used to remove free and WAD cyanide by formation of iron ferricyanides with subsequent of soluble arsenic by precipitation as basic ferric arsenate. Test results demonstrated that a residue WAD cyanide level of < 50 ppm and a soluble arsenic level of < 0.1 ppm could be achieved. Arsenic Toxicity Characteristic Leach Procedure (TCLP) testing of the detoxified residue showed xxx (Results coming). Use of BIOX liquor was selected as the preferred detoxification process on the basis of demonstrated ability to meet design WAD cyanide and soluble arsenic CIL residue discharge levels and significantly lower operating cost. The process developed for detoxification of the CIL residue comprises two stages for consecutive removal of free and WAD cyanide and soluble arsenic. In the first stage CIL residue will be reacted with solution separated from the BIOX oxidised product. This will take place in two agitated tanks operated in series. Provision is made to individually bypass tanks for maintenance. Cyanide will be largely precipitated as iron ferricyanides. The second stage involves a two-step neutralization process whereby ground limestone slurry is used to precipitate arsenic as basic ferric arsenate by raising the ph to 4 and lime slurry to raise ph to neutral precipitating calcium sulphate. This will take place using three agitated tanks in series with provision to bypass individual tanks for maintenance. All tanks will be vented to a scrubber for hydrogen cyanide removal using sodium hydroxide. Detoxified residue will be stored in a lined storage structure. 21

27 Continuous Pilot Scale Testing With the exception of limited continuous flotation tests designed primarily for preparation of BIOX test samples, no continuous pilot scale testwork was undertaken. Although all circuits were able to be designed from batch bench-scale testwork, which resulted in cost and time savings, difficulties were experienced in undertaking cyanide leach and detoxification testwork due to the time taken to prepare BIOX product. If for no other reason than ensuring sufficient sample material, continuous pilot scale comminution, flotation and BIOX testing would be recommended for future such projects. Flowsheet The process developed for Fosterville ore is shown in Figure 5. Not shown is a separate limestone grinding circuit. 22

28 Figure 5: Fosterville Plant Process Route 23

29 The plant is to operate on a continuous 24 h/d, 7-d/week basis. Mining and ore treatment schedules for the project were developed within the constraints of not only the physical capacity of the crushing and milling circuits but the sulfur oxidation and mass flow capacities of the BIOX circuit. The ore treatment schedule includes treatment rates of up to 960,000 t/y or 120 t/h, depending on feed sulfur grade. Inoculum Build-Up Inoculum build-up is the process whereby some 15 L of BIOX starter culture is grown progressively with flotation concentrate to about the 5,500 m 3 needed to fill the BIOX reactors prior to commencing continuous plant operation. Several options were examined for the build-up process: Completion of the Fosterville grinding and flotation circuits and undertaking build-up using Fosterville concentrate in three purpose built reactors to produce sufficient active material to inoculate the first primary reactor. Inoculum build-up would then continue by successive build-up in the remainder of the primary and then secondary reactors. Procurement of sufficient bacterially active material from another BIOX operation to commence build-up in the first primary reactor with Fosterville concentrate. About 100 m 3 of active material would be required. Commencing the build-up onsite using suitable concentrate sourced from another operation using small purpose built vessels followed by successive build-up in the BIOX nutrient mixing tank followed by one neutralization tank. Material from the neutralization tank would then be used to inoculate the first primary reactor and the build-up continued using Fosterville concentrate. The first option would require the construction and commissioning of the grinding, flotation and services sections of the plant to be complete some 12 weeks ahead of the remainder of the plant. This was not practical and adoption of this option would have effectively meant a delay in commencement of operations of some 12 weeks. Production of concentrate on site using a pilot scale milling and flotation circuit was considered but this would have been relatively more expensive than other options. Evaluation of commercial and logistics aspects of using concentrate and active material resulted in adoption of a build-up process using concentrate purchased from another operation. 24

30 Build-up commenced in a 75 L purpose built heated agitated vessel followed successively by a 500 L vessel, the nutrient mixing tank and one neutralization tank. For build-up the nutrient mixing tank and neutralization tank will be fitted with cooling/heating coils and heavier duty agitators and supplied with temporary power and air from a portable air compressor. PLANT DESIGN AND ENGINEERING CONSIDERATIONS Engineering design for the processing plant and associated residue storages commenced in November 2003 and was scheduled so as to deliver equipment specifications and design packages that allowed early commencement fabrication of long lead items and on site construction. The Fosterville environment and use of the BIOX process required consideration of a number of particular aspects during design and engineering of the plant. Environmental The resource geochemistry, nature of the process and plant location required particular consideration of some environmental aspects, in particular: Noise. Acid, arsenic and cyanide management. The location of nearby residences required the crushing plant to be set as low as possible and operated day shift only, provision was made for sound attenuating enclosures around the SAG mill and BIOX air compressors and fitting of rubber liners to the SAG and limestone grinding mills. Limestone imported from southwest Victoria will be ground on site and used to neutralize the bulk of the acid and stabilise arsenic generated in the BIOX process. Some 50,000 t of limestone will be required annually. Arsenic required incorporation of processes to fix arsenic as the stable basic ferric arsenate, a form of arsenic that has been recognised by the US EPA as the Best Demonstrated Available Technology (BDAT) for treatment of arsenic bearing mine wastes. As mentioned before, the WAD cyanide of the CIL residue is to be reduced to less than 50 ppm before entering the CIL residue storage. 25

31 Water Supply and Quality Water requirements for the Fosterville plant will be high relative to oxide gold cyanide leach plants and results from: Evaporative losses due to the exothermic oxidation of sulfur. In addition there is a high cooling tower blow-down loss as a result of the high chloride content of make-up water. Inability to recycle water from the cyanide leach residue as it will contain cyanide species toxic to the BIOX bacteria and which dictates separate storage of CIL residue. High evaporative losses from the main residue storage due the location of the project. Early ore processing studies were based on supply of water to the plant from the nearby Campaspe River, a significant but limited Central Victorian water resource. However during the last series of studies undertaken in 2003 it became apparent that an additional water source would be required. Desalination of local groundwater was considered however two factors precluded this option: Cost. Inability to effectively dispose of the resultant brine stream. Following a series of engineering studies agreement was reached with the authority responsible for operation of the City of Bendigo waste water treatment facility to take sufficient treated effluent to meet the total ore processing requirements. This involves construction of a pumping station, 20 km pipeline and associated storage tanks. Both River water and treated effluent contain chloride levels at times in excess of 250 ppm and chloride will concentrate within the process plant due to evaporation and return of water from the flotation residue storage. Restrictions on the design chloride level in the BIOX section of the plant arising from considerations of tolerance of the BIOX bacteria to chloride, and corrosion required careful design of the plant water system to limit the chloride level in the BIOX circuit to about 1,000 ppm. Provision for Future Expansion An integral component of Perseverance growth strategy is further deeper underground exploration of the Fosterville lease and it is reasonably expected that additional resources 26

32 will be identified that may justify an expansion in plant capacity. Provision has been left in the design of the plant for the following additions: Ball mill to allow greater grinding capacity by conversion of the circuit to SAG-Ball. Roughing flotation cell. BIOX tankage. CCD thickener train. BIOX air compressor. Corrosion At a ph of around 1.5 and chloride level of around 1,000 ppm BIOX liquor is corrosive and the layout of the BIOX section of the plant was therefore designed to minimise transfer pipework and pumping requirements. Blending Run of Mine (ROM) ore will be blended before feeding to the plant and will need to consider the sulfur, iron, arsenic and gold content of the ore. A constant sulfur grade is required for steady operation of the BIOX process, a minimum iron to arsenic ratio is critical for formation of a stable ferric arsenate precipitate and sufficient iron needs to be present to adequately remove free and WAD cyanide from the CIL residue. Ability to blend will be assisted by: A sufficiently large ROM pad area to accommodate blending stockpiles. Provision of a large flotation concentrate storage capacity in the form of both a 48 h agitated tank and a 7-day capacity pond. Location of the BIOX feed sampler ahead of concentrate storage rather than after. BIOX Plant Materials of Construction The corrosive nature of the BIOX liquors required detailed consideration of materials of construction for the following parts of the circuit including pumps and interconnecting pipework: BIOX reactors. 27

33 CCD thickeners. Neutralization reactors. CIL residue detoxification plant. This involved consultation with the owners of the BIOX technology, suppliers of specialist steels and operators of biological oxidation plants. Consideration was given to use of: Duplex stainless steel (SAF 2205). 316 L grade of stainless steel. Rubber covered mild steel (RCMS). Although significantly cheaper, RCMS has the disadvantage of being prone to rupture, which would be a particularly high risk in the neutralization reactors that will be subject to regular removal of built-up scale. At the temperatures, ph and chloride levels envisaged 316 L stainless steel would be at risk of significant rates of corrosion. SAF 2205 was therefore selected for all sections of the BIOX circuit and the cyanide removal reactors in the CIL detoxification plant. 316 L stainless was selected for the three neutralization reactors in the detoxification plant. Plant Control System The overall plant control system is relatively simple and generally typical for a plant involving crushing, grinding, flotation and cyanide leaching. The BIOX section of the plant requires few special process controls, which are limited to: Mass flow to the BIOX circuit is automatically controlled to an operator preset level. Nutrient dosing rate is controlled by the BIOX circuit feed rate. Automatic control of BIOX reactor temperature by control of cooling water flow by temperature sensors in each reactor. Airflow to each BIOX reactor is automatically controlled to a preset level. 28

34 Lime and limestone slurry additions to neutralization tanks are automatically controlled by ph probes located in the tanks. Limestone slurry additions to BIOX reactors is controlled manually. BIOX Cooling and Aeration Requirements BIOX cooling and aeration requirements are significant. The design total net heat rejection requirement for the BIOX reactors is 12,000 kw and will be met using a 3 module cooling tower with 16,000 kw total cooling capacity. Cooling tower evaporative losses are estimated be 20 m 3 /h with a further 7 m 3 /h water lost in cooling tower blow-down, depending on the chloride content of incoming water. A total of 36,000 Nm 3 /h air at 140 kpag will be required and will be supplied from three 11kV, 750 kw centrifugal air compressors in a two duty one standby configuration, each with a capacity of 22,000 Nm 3 /h. These compressors will also provide the air to the CIL and detoxification circuits. PROCESS AND ENGINEERING RISK MANAGEMENT The BIOX process, plant design and the need to be mindful of capital cost gave rise to a number of potential process and engineering risks outside of those usually encountered, which required consideration during design. Water Requirements Feasibility study estimates of water quantity and quality were subject to some uncertainty. During the past few years Central Victoria has recorded above average temperatures and evaporation rates and below average rainfall. If this trend is sustained use of historical climatological data could result in an underestimate of project water requirements. In addition there has also been a resultant trend upwards in the salt content of both project water sources. To accommodate these trends plant water requirements were modeled using reasonable worst-case estimates of precipitation and evaporation, and make-up water salt content. Flotation Performance Laboratory testwork indicated that Fosterville ores respond well to flotation and require only careful grinding and a single flotation stage to meet the competing requirements minimum 29

35 sulfur loss resulting from overgrinding and a fine BIOX feed sizing. The flotation circuit however was largely based on results of batch tests and therefore provides some risk. To provide some flexibility in operation of the rougher/scavenger circuit a cleaning circuit was added. In addition provision has been made in the plant layout for later installation of flotation equipment to treat SAG mill cyclone underflow and for regrinding flotation concentrate. Sulphur Oxidation and BIOX Plant Capacity The design capacity of the BIOX circuit is subject to uncertainties related to: Sulphur oxidation and mass flow capacities resulting from typical uncertainties related to the resource model and mining schedule. Sulphur oxidation rate or completeness. In addition to allowance of reasonable design margins in all sections of the plant, provision was made in the layout for later installation and operation of up to two BIOX reactors. Power Supply Interruption To maintain bacterial activity limited aeration and cooling of the BIOX required during a prolonged interruption to the project power supply. reactors will be The Fosterville project is supplied from a major 220 kv regional supply in both directions. Analysis of the availability of this supply confirmed its high reliability with only one prolonged (22 hours) unplanned outage in ten years and none in the past five. Notwithstanding the low risk of interruption to the project power supply, provision is being made to install onsite diesel generating equipment to maintain both bacterial activity and underground mine dewatering and ventilation requirements. 30

36 References Amdel Limited, Location of Gold in a Sulphide Flotation Concentrate, Amdel Report No. G854600G/95. Perseverance Corporation Limited, Perseverance 2003 Annual Report. Perseverance Corporation Limited, Fosterville Gold Project. Bankable Feasibility Study. 31

37 List of Tables Table 1: Central Zone Reserves (Perseverance Corporation 2003 Annual Report) Table 2: Calculated Concentrate Mineralogy Table 3: Oxidative Pre-Treatment Options Table 4: Summary of Comparative Capital Costs Table 5: Summary of Comparative Operating Costs Table 6: Cumulative Average Flotation Concentrate Data Table 7: Flotation Concentrate Head Assays of Orebody and Variability Composites Table 8: Summary of BIOX Orebody and Variability Testwork Table 9: Summary of BIOX Circuit Design Criteria Table 10: Average Cyanide Leach of BIOX Orebody and Variability Residues List of Figures Figure 1: Project Location Figure 2: Grade X Width contouring of Central Zone Figure 3: Test Program Figure 4: Sulphur Oxidation Rate Figure 5: Fosterville Plant Process Route 1

38 Table 1: Central Zone Reserves (Perseverance Corporation 2003 Annual Report) Grade Contained Reserve Tonnes Gold % S % Fe % As g/t Au Oz Open Pit 1,748, ,000 Underground 4,348, ,000 Total 6,096, ,000 Table 2: Calculated Concentrate Mineralogy Mineral % Weight Pyrite 30.5 Arsenopyrite 13.8 Stibnite 1.0 Total Sulphides 45.4 Gangue

39 Table 3: Oxidative Pre-Treatment Options Process Process Summary Advantages/Disadvantages BIOX Bacterial oxidation of flotation concentrate by acidithiobacillus Moderate capital and operating costs ferroxidans, At.thiooxidans and Leptospirillum ferrooxidans at Established technology controlled temperature and ph. Process technology is owned by Relatively simple and low cost management of arsenic bearing residue. Goldfields Ltd (GFL) More difficult to achieve complete S oxidation relative to roasting and pressure oxidation BacTech Roasting Bacterial oxidation of flotation concentrate by thermophilic bacteria similar to those above. Process technology is owned by BacTech Mining Corporation Oxidation of sulphides contained in whole ore or flotation concentrate at elevated temperature in the presence of excess air Pressure Oxidation Aqueous acid oxidation of sulphides contained in flotation concentrate or whole ore using oxygen at elevated temperature and pressure High Temperature Redox Activox Pressure oxidation using nitric acid as a catalyst at elevated temperatures. Fine grinding of concentrate followed by oxidation with oxygen under pressure at low temperature. Process technology is owned by Western Minerals Technology. Requires water of certain quality Moderate capital and operating costs Relatively simple and low cost management of arsenic bearing residue. More difficult to achieve complete S oxidation relative to roasting and pressure oxidation Requires water of certain quality Low capital and operating costs Established technology. Complete sulphur oxidation. Suitable for ore containing detrimental carbonaceous material. Difficulties in environmental management and disposal of sulphur and arsenic residues. High capital and operating costs Established technology Relatively simple and low cost management of arsenic bearing residue. More complete S oxidation relative biological oxidation Rapid reaction with good S oxidation High capital and operating costs Not established technology Low capital and operating costs Not established technology 2

40 Table 4: Summary of Comparative Capital Costs Cost Area BIOX $M POX $M Direct Costs: Oxidation and CIL Capital Other Capital Subtotal - Direct Costs Indirect Costs: EPCM (15%) Contingency (15%) Plant Total Table 5: Summary of Comparative Operating Costs Cost Area BIOX $/t ore POX $/t ore Labour Consumables Power Maintenance Laboratory Total $/t ore Table 6: Cumulative Average Flotation Concentrate Data Flotation Concentrate Cumulative Laboratory Flotation Time min Concentrate Mass Pull % Gold Au g/t Cumulative Flotation Concentrate Assays Sulphur S % Arsenic As % Gold Au % Cumulative Recovery Sulphur S % Arsenic As % Concentrate Concentrate Concentrate Concentrate Selected Design Point 3

41 Sample Table 7: Flotation Concentrate Head Assays of Orebody and Variability Composites Gold g/t Au Silver g/t Ag Antimony g/t Sb Total Iron % Fe TOT Total Sulphur % S TOT Sulphide Sulphur % S 2- Arsenic % As Total Carbon % C TOT Organic Carbon % C ORG Carbonate % CO , , , , , Sample ID Table 8: Summary of BIOX Orebody and Variability Testwork Biological Oxidation Time days Mass Gain % Reagent Consumption Lime Net Acid kg/t BIOX kg/t BIOX Feed Feed Sulphide Oxidation % Arsenic Dissolution % Average

42 Table 9: Summary of BIOX Circuit Design Criteria Parameter Units Value BIOX Circuit Nominal Feed Rate t/h 9 BIOX Residence Time Primary Reactors d 2.5 Secondary Reactors d 2.5 Number of Reactors Primary Reactors No. 3 in parallel Secondary Reactors No. 3 in series BIOX Pulp Density Liquid:Solids 4:1 Operating Temperature C 42 Mass Gain % 11 Sulphur Oxidation Design % 100 Nutrients K As potassium kg/t conc 0.5 N As nitrogen kg/t conc 1.0 P As phosphorous kg/t conc 0.2 Limestone As 100% CaCO 3 kg/t conc 415 CCD Circuit Specific Settling Rate (Flux) BIOX Residue t/m 2 h 0.18 Underflow Solids BIOX Residue % w/w 33 solids Number of Stages No. 3 Wash Ratio Water : Solids m 3 :t 7.8 Liquor Neutralisation Neutralisation Residence Time h 6 Limestone As 100% CaCO 3 kg/t conc 213 Hydrated Lime As 100% Ca(OH) 2 kg/t conc 26.5 Table 10: Average of Cyanide Leach of BIOX Orebody and Variability Residues Testwork Method Reagent Consumption Lime kg/t BIOX Feed Sodium Cyanide kg/t BIOX Feed Residual Cyanide CN FREE ppm Gold Extraction % CIL ,

43 Table 11: CIL Design Parameters Parameter Unit Value Circuit Configuration CIL Nominal Feed Rate t/h 10 Number of Stages 6 Pulp Density % w/w solids 33 Leach Retention Time h 48 Dissolved Oxygen Levels mg/l Carbon Concentration g/l Initial Cyanide Concentration mg/l NaCN 1,500 Residual Free Cyanide mg/l NaCN 1,000 Pulp ph Lime Addition (90% Ca(OH) 2 ) kg/t CIL Feed 6.0 Cyanide Consumption kg/t CIL Feed 9.7 6

44 Figure 1: Project Location Figure 2: Long Section of the Central 2.5 km Area of Resource and Exploration Targets (Perseverance Corporation 2003 Annual Report) 7

45 Figure 3: Test Program 8

46 Figure 4: Sulphur Oxidation Rate 9

47 Figure 5: Fosterville Plant Process Route 10