- Email: [email protected]
Commercial heap biooxidation of refractory gold ores – Revisiting Newmont’s successful deployment at Carlin
Minerals Engineering xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
Minerals Engineering journal homepage: www.elsevier.com/locate/mineng
Commercial heap biooxidation of refractory gold ores – Revisiting Newmont’s successful deployment at Carlin F.F. Roberto Newmont USA, Limited, Englewood, CO, USA
a r t i c l e
i n f o
Article history: Received 2 August 2016 Accepted 23 September 2016 Available online xxxx Keywords: Biooxidation Commercial heap leach Refractory gold ore Scale-up
a b s t r a c t Newmont Mining Corporation pioneered the investigation, development, and commercial-scale implementation of refractory gold whole-ore heap biooxidation, during a period spanning 1988–2009 at Carlin, Nevada. Basic and applied research and development from 1988 to 1999 included laboratory test work and increasingly larger pilot test heaps culminating in the full-scale implementation of a process that was estimated to contribute 120,000–180,000 oz/year to Carlin’s production between 2000 and 2005. Key parameters that influenced performance of the on-off heap biooxidation process, and factors that led to the discontinuation of the operation are described. Ó 2016 Published by Elsevier Ltd.
1. Introduction Newmont Mining Corporation pioneered the investigation, development, and commercial-scale implementation of refractory gold whole-ore heap biooxidation, during a period spanning 1988–2010 at Carlin, Nevada (Fig. 1 summarizes key developments and gold price during this period). Basic and applied research and development from 1988 to 1999 included laboratory test work and increasingly larger pilot test heaps leading to full-scale implementation of a process that was estimated to contribute 120,000– 180,000 oz/year (12.2 t total by May 2005; Logan et al., 2007) to Carlin’s gold production from 2000 to 2005. Even within Newmont, it is difficult to determine exactly when the closure of the commercial operation (internally referred to as the Refractory Leach Project) occurred, but by combining U.S. Securities and Exchange Commission annual 10-K reports and internal reports, it can be concluded that initial planning for project closure began in 2007, operating data exist for 2008 and 2009, and a 2011 final closure plan for the biosolution pond described cessation of operations in 2010. It is likely then that the commercial biooxidation plant contributed at least another 5 tonnes (ca. 160,000 oz) of gold to Carlin production before closure in 2010, after a decade of operation. Brierley et al. (1995) summarized the pilot heap tests that predated the Gold Quarry biooxidation demonstration facility (Schutey-McCann et al., 1997), a $13.5 million project that constructed and operated an 800,000 t/year large-scale pilot that integrated lessons learned from the early, smaller heap tests. This
E-mail address: [email protected]
facility was operated from 1994 to 1997, and the biooxidation and ammonium thiosulfate leach pads were ultimately enlisted to validate ‘‘bio-milling” at Carlin’s Mill 5 carbon-in-leach (CIL) facility in conjunction with the commercial biooxidation process. Initial bench-scale work to test the feasibility of biooxidizing Carlin low-grade refractory gold ores began in 1988 at Newmont’s central metallurgical services laboratory, then located in Salt Lake City. Carlin-type ore is typified by microscopic or dissolved gold disseminated in pyrite, arsenopyrite, or arsenian pyrite hosted within carbonaceous sediments. Concerns regarding increasing sulfide content in Gold Quarry ore at that time justified an examination of biooxidation as an alternative processing strategy for low-grade refractory ore.
2. Small pilot heap test summary A Refractory Leach Test Facility to construct the first small pilot heap (476 t) was authorized by Newmont management in 1990, and constructed adjacent to the Gold Quarry pit at Newmont’s Carlin Mine. By 1994, a series of successively large heaps were operated in different areas of the Carlin mine (North and South Areas), ranging from 476 t to 25,900 t. Many concepts that have become accepted practice in commercial copper bioleaching and were used during operation of Newmont’s commercial whole-ore biooxidation pads were tested during this period, including use of drip emitters, rather than wobbler, or impact head sprinklers, use of Wilfley diffusers in bacterial production bioreactors, the beneficial effect of rest periods on bacterial activity (as measured by accumulated ferric iron and increased Eh in leach solution),
http://dx.doi.org/10.1016/j.mineng.2016.09.017 0892-6875/Ó 2016 Published by Elsevier Ltd.
Please cite this article in press as: Roberto, F.F. Commercial heap biooxidation of refractory gold ores – Revisiting Newmont’s successful deployment at Carlin. Miner. Eng. (2016), http://dx.doi.org/10.1016/j.mineng.2016.09.017
2
F.F. Roberto / Minerals Engineering xxx (2016) xxx–xxx
Fig. 1. Timeline of developments leading to Commercial Refractory Leach Facility.
and observation that mesophilic iron-oxidizing bacteria such as Leptospirillum and moderately-thermophilic bacteria such as Sulfobacillus emerge without inoculation and are presumably enriched from the environment as sulfide minerals are oxidized. Newmont evaluated heaps that contained not only siliceous, sulfidic, refractory (SSR) ore, but also carbonaceous, sulfidic, refractory (CSR) ores that possess significant ability to remove cyanidecomplexed gold (so-called ‘‘preg-robbing” ore), generally thought to be a function of non acid-soluble, organic carbon in the ore. The test heaps included variations in sulfide sulfur, clay, pregrobbing carbon, carbonate, biooxidation cycle, and particle size. Belt agglomeration with a substantial (106–107 cells/mL) quantity of microbial inoculum (Acidithiobacillus ferrooxidans) reduced to practice what would become Newmont’s patented BIOPROÒ technology, with 4 related patents (Brierley and Hill, 1993, 1994, 1998, 2002, 2004). Heaps were stacked with either a front-end loader or radial stacker (larger heaps). The acidic conditions of biooxidation created within the heaps precluded direct leaching with cyanide or other gold complexing agents. Ore was off-loaded from the heaps with a front-end loader and partially neutralized with cement, caustic, or lime addition to achieve alkaline pH. Gold extraction with cyanide, thiourea, or ammonium thiosulfate (ATS) was tested. It was concluded that thiourea was ineffective perhaps due to the large particle size of the ore, while ATS yielded results equivalent to, and in several cases better than cyanide. Newmont patented the ATS process for gold leaching from preg-robbing ores (Wan et al., 1994) and tested this approach at larger scale in the subsequent Gold Quarry demonstration. The coupling of this lixiviant to biooxidation of preg-robbing ores was initially planned to be a key feature of the commercial whole-ore biooxidation facility. Barrick Goldstrike recently achieved commercial production of gold using the Total Carbonaceous Material (TCM) leach process based on calcium thiosulfate in the 3rd quarter of 2015 (Barrick Gold Corporation, 2016). Other significant parameters important to operation of biooxidation heaps for refractory gold ore that emerged from these tests included: (a) defining a lower limit for sulfide-S content of 0.2– 0.4%, (b) crushing to a smaller particle size is not always better for biooxidation within the heap, particularly with the presence of clays, and (c) carbonate content up to 2.2% could be accommodated in spite of the requirement of low pH conditions to promote growth of the iron-oxidizing bacteria, although considerable sulfuric acid addition was necessary to achieve pH control. It is interesting that active aeration through the use of a blower and air distribution lines at the base of the heap was tested during this
time as well, but the benefit was not clear, although it was noted that heap residues appeared to indicate uniform biooxidation had occurred throughout the heaps and might be attributed to aeration. A record low temperature for the region ( 26.1 °C) was experienced in December 1992 and caused concern, although the heap internal temperatures were observed to remain about 10 °C. A solution heater was introduced to prevent freezing of the recirculating solution but did not appear to have measurable benefit on the process within the heap. Finally, the cost for biooxidation of refractory gold ores in heaps was estimated at that time to be within the range of $4–6/t. The commercial plant costs were later found to be at the low end of this range.
3. Gold Quarry demonstration The scale-up of facilities to process up to 800,000 t/year of lowgrade (1–3 g/t Au) siliceous sulfidic refractory (SSR) and carbonaceous sulfidic refractory (CSR) ores on five 136,000 t leach pads required a bacterial production capability that incorporated six 200 m3 (1200 m3 total) tanks started in batch mode, ultimately producing 800 m3/day in continuous mode flowing into two solution storage ponds: a conditioning pond with a capacity of 800 m3, and an operating pond of 2377 m3 volume. An acid neutralization plant was part of the initial design for the facility, but no data could be located in internal documents or published reports that described details of operation or performance. Crushing utilized the existing Gold Quarry crushing circuit to generate nominal 19 mm material after passage through a gyratory crusher and cone crushers. Target ore composition included 5–15% clay, 1.5–2.5% sulfide-S, and 0.07–0.5% acid-soluble carbon (the CSR material contained 1% organic carbon versus 0.07% in the SSR material). Two methods of heap stacking were evaluated, comparing conveyor stacking with truck-dumped ore. Ore was stacked to a height of 10 m. Ore was inoculated at a rate of 24 L inoculum/t at the loadout conveyor for the haul trucks in the case of truck-stacked ore, while it was sprayed over the crushed ore at a portable feed hopper dropping ore onto the conveyor-stacker train at 18 L/t for conveyor-stacked heaps. Five different heaps were constructed to test liner designs including clay and HDPE, as well as combinations with clay base, and drain rock. No details were available regarding this evaluation, although it is presumed that the results were considered in the final commercial plant design.
Please cite this article in press as: Roberto, F.F. Commercial heap biooxidation of refractory gold ores – Revisiting Newmont’s successful deployment at Carlin. Miner. Eng. (2016), http://dx.doi.org/10.1016/j.mineng.2016.09.017
F.F. Roberto / Minerals Engineering xxx (2016) xxx–xxx
Substantial differences were noted in the hydraulic conductivity of truck-stacked SSR ore (2.8 10 4 cm/s) and conveyorstacked SSR (6.7 10 2 cm/s) and CSR (1.3 x 10 1 cm/s) ore. Dramatic reduction in daily sulfide-S oxidation rates – 0.125%/day for the truck-dumped SSR compared to 0.22%/day and 0.175%/day for the conveyor-stacked SSR and CSR ores was attributed to the reduced conductivity of solution. Schutey-McCann et al. (1997) concluded their summary of the Gold Quarry demonstration stating that ‘‘. . .to achieve the optimum Global Oxidation rate Newmont will conveyor stack all bioleach ore” which sadly, was trumped by fiscal considerations in the operation of the commercial biooxidation facility. Extremely high peak temperatures approaching 80 °C were observed for the conveyor-stacked SSR ore heap at about 85 days, while conveyor-stacked CSR ore heap internal temperatures increased slowly to about 60 °C over 160 days, and temperatures in the truck-dumped SSR heap declined from an initial temperature of 60 °C to about 30 °C after 160 days. Other chemistry extremes were also observed, including peak total iron levels >40 g/L, and Fe2+ over 20 g/L. This high ferrous iron level was later seen to cause problems during preparation of the biooxidized ore for cyanide or ATS leaching for Au recovery. Rinsing of the ore prior to neutralization was attempted, but this led to localized iron precipitation and production of sludge build up in the solution pond, and was discontinued. Best results were obtained by direct lime addition to the biooxidized ore in a pug mill. In spite of failing to achieve a target sulfide-S oxidation (55%) in any of the test heaps, the overall results of this much larger-scale pilot were considered a success, and a prefeasibility study commissioned in 1995 used results of the demonstration to make a case for a full-scale commercial biooxidation facility and justify the reclassification of 42 Mt of waste rock as reserves containing 62 t (2 million oz) Au.
4. Commercial operations (Refractory Leach Project) The qualified success of the Gold Quarry demonstration project increased confidence that Newmont now had another process for pretreatment of low-grade refractory gold ore, and that coupled with ATS leaching, could also address preg-robbing, carbonaceous ores. Funds were authorized to proceed with construction of a commercial plant in the South Area of the Carlin mine in 1997. This plant would have included twelve 147 m 305 m aerated on/off leach pads with an annual throughput of 11.7 Mt/year (32,000 t/ d) producing 8.4 t Au/year. Ore was to be crushed through 3 stages to achieve a P80 of 10 mm, agglomerated with microbial inoculum in a drum agglomerator, and stacked to a final height of 10 m by conveyor-stacker, due to the clear benefits shown in the Gold Quarry demonstration with respect to increased air permeability and hydraulic conductivity when compared to truck-dumping. A lime precipitation water treatment circuit was to be included in the plant design to overcome solution chemistry problems that plagued many of the pilot heap tests, and minimize the carryover of excessive iron and other metals to the gold leach circuit, thus reducing cyanide consumption or interference with ATS. Biooxidized ore was to be moved from the biooxidation leach pads to the gold leach pads using a bucket wheel excavator. Circulating biosolution was to be contained in a 113,500 m3 pond. However, the gold industry experienced a dramatic decline in gold price below $300/oz between 1997 and 2000 that led to the delay and substantial reduction in the project budget that significantly changed the as-built project (see Fig. 1 timeline). The final budget was less than 50% of the original proposed amount, and major revisions to the project design included a reduction in the
3
number of biooxidation pads from 12 to 3, a larger, crushed particle size target P80 of 19 mm, elimination of the agglomeration drum, conveyor-stacker, water treatment circuit, bucket wheel excavator, and instead of gold recovery by heap leaching, introduction of the biooxidized ore directly into the Carlin Mill 5 CIL plant (this coupled process was referred to as ‘‘bio-milling”). Fig. 2 depicts the operating facility, with the 3 pads in various stages of loading, oxidizing, or unloading. As will be discussed later in this paper, the decision to direct biooxidized ore to Mill 5 might have sealed the fate of commercial whole-ore biooxidation at Carlin, since changes in the Mill 5 operating philosophy and design ultimately reduced the value of biooxidized ore to overall Au production at Carlin. First-year operating results were summarized by Bhakta and Arthur (2002) and described the bio-milling process at Mill 5, where a semi-autogenous grinding (SAG) mill and ball mill produced a P65 = 75 lm leach feed to an 8-tank CIL circuit with 8– 10 h retention time. The biooxidized ore (160 days of biooxidation) was neutralized after off-loading by addition of pebble lime (7.5– 15 kg/t) at the SAG mill feed belt. Initial operating results with this feed noted low dissolved oxygen levels and increased cyanide consumption, and lower than expected gold recovery. It was discovered that ferrous iron, copper, zinc, and nickel were introduced by the biooxidized ore and likely accounted for the oxygen depletion and additional cyanide utilization. These effects were mitigated somewhat by the addition of air to the slurry in the surge tanks upstream of the CIL circuit, and reduction in the slurry loading, but it was concluded that a larger water treatment facility was needed for the biooxidation circuit to remove metals before biomilling. This was never built, although various options were explored. 2.4 million t of ore were processed in the first year, about 70% of the target throughput. Tempel (2003) revisited the performance of the commercial biooxidation process in 2003, and described how creative operations overcame some of the design compromises that were necessitated by a reduced project budget. Coarser particle size of the crushed ore somewhat compensated for the expected reduction in hydraulic conductivity and air permeability resulting from truck-dumped stacking of the heaps. In addition, stacking of the heaps to an initial height of 13.7 m, bulldozing of approximately 0.9 m of this ore to a height of 12.8 m, and finally, ripping the pad surface to a depth of approximately 1.5 m was performed to eliminate much of the estimated 2.4 m deep compaction layer created by 140-t haul trucks. Off-loading of the pads was mentioned as another significant problem, since the reduced budget only permitted changing the pad base air lines after 3 cycles. One meter of ore was retained after bulldozing the bulk of the biooxidized ore from the pads, but damage still occurred to air lines. Cementation of dried leach precipitates was also observed around the air lines on occasion, and noted to be extremely hard, with attempts to break up the material often leading to additional air line damage. Observation of the biooxidation pads with a thermal imaging camera revealed cold spots apparently experiencing reduced biooxidation. It was determined that higher-carbonate content ore from the North Area was present. Attempts to blend this ore with Gold Quarry ore with reduced carbonate content were unsuccessful in improving biooxidation rates, and actually depressed the sulfide-S oxidation of the lower carbonate ore. The temperature extremes of the Carlin site (cold winters and hot summers) posed a problem for controlling agglomeration moisture, as the initial 4% moisture target led to saturated ore that caused sloughing on the heaps and localized ponding. The heaps themselves achieved internal temperatures well beyond the design limits of the liners (rated to 60 °C), sometimes reaching 81 °C on the first commercial pad. Attempts to cool the heaps through
Please cite this article in press as: Roberto, F.F. Commercial heap biooxidation of refractory gold ores – Revisiting Newmont’s successful deployment at Carlin. Miner. Eng. (2016), http://dx.doi.org/10.1016/j.mineng.2016.09.017
4
F.F. Roberto / Minerals Engineering xxx (2016) xxx–xxx
Fig. 2. Refractory Leach Facility, 2001.
increased irrigation resulted in saturation of the heap which inhibited bacterial oxidation. The high internal temperatures in the heaps led to the decision to inoculate with archaeal species 6 months after the facility began operation, and these microbes were maintained at cell densities between 106 and 108 cells/mL, and appeared to benefit the process. This addition to the microbial inoculum was noted as another operating change that resulted from modifications to the initial project plan. Large amounts of ferrous iron were introduced into the biosolution pond with breakthrough of irrigation in the biooxidation pads. This observation was interpreted to mean that the microbial activity in the heaps was insufficient to regenerate ferric iron for additional pyrite oxidation. Therefore irrigation was suspended to increase Eh and apparent microbial activity. Increased heap temperatures also accompanied cessation of irrigation. At the same time, oxygen depletion was noted in the biosolution pond, along with a drop in Eh below 450 mV. Use of a commercial aeration device designed for municipal wastewater treatment plants increased the dissolved oxygen concentration and improved the Fe3+/Fe2+ ratio. Excessively high total iron concentrations had been seen in some of the pilot heaps and the Gold Quarry demonstration, and while the average levels in the commercial pads ranged from 8 to 26 g/L, levels as high as 59 g/L were observed at start of ore leaching. Not for the first time a comment was made about the ‘‘. . .need to treat the biosolution to lower metals concentration. . .” The soluble metals carried over to the bio-milling process at Mill 5 were considered at least in part, to contributing to the observed 3-fold increase in cyanide consumption at the mill compared to oxide ores. Estimation of overall sulfide-S oxidation after a biooxidation cycle was noted to be a problem, as the results of tail assays indicated sulfide-S levels higher than the head assays taken from belt samples during truck loading. It was determined that these higher tail assays resulted from artifacts of residual sulfate contributing to the apparent sulfide-S quantified by the LECOÒ carbon/sulfur roast method in use at the time. In spite of the inaccurate sulfide-S determinations, it was acknowledged that the biooxidation process was not achieving the target oxidation (which ultimately reduced Au recovery), and that the ability to perform a longer biooxidation cycle could be accomplished if more heaps were available. This was noted to also provide some flexibility for different ore types. Unfortunately, it is apparent that a request to increase the number of biooxidation pads was not approved. Logan et al. (2007) provide the most comprehensive overview of the commercial biooxidation facility design and operation, through
mid-2005. Readers are directed to that excellent book chapter, coauthored by Jim Brierley, for detailed operational commentary including a microbiologist’s perspective. By that time, 12 batches totaling 8.8 Mt of ore had been processed, resulting in the production of 12.2 t (379,500 oz) Au. Optimal solution chemistry parameters included a pH range of 1.3–2.2, Eh of >550 mV, Fe3+ concentration between 5 and 25 g/L, and dissolved oxygen concentration >2 ppm. Some details that were not described in the previous operating summaries (Schutey-McCann et al., 1997; Tempel, 2003) discussed above included heap irrigation rates (4 L/m2 h) through drip emitters spaced on 760 mm centers, and liner construction (1 mm HDPE liner on 150 mm clay base), and placement of aeration lines in 1 m of crushed rock below the actual ore base. They note that truck dumping of the ore may have reduced sulfide-S oxidation rates by as much as 50%, and that when the heaps were off-loaded, dry areas that had not received irrigation were estimated to comprise as much as 10% of the heap volume. The dry areas could have resulted from solution short-circuiting, or excessive drying by aeration (the low humidity in the Carlin area, less than 30%), or a combination of the two. These observations, when considered with apparent cold regions of the heaps using thermal imagery made by Tempel, suggest that reduced solution and air permeability, along with higher localized regions of carbonate within the heaps where biooxidation was inhibited, could account for the disappointing overall global oxidation of sulfide-S of 21.9%, compared to the target 30%. In 2004, a froth flotation circuit was added to Mill 5, in order to increase recovery of gold from biooxidized ore (Au that remained locked in unoxidized sulfides). An additional 10% gold recovery was attributed to this ‘‘bio-float” modification to Mill 5 in 2005.
5. Mill 5 and the demise of the Refractory Leach Project For several years, the acidic biooxidized ores were presumed to be a benefit to processing high-carbonate ores through the modified Mill 5, as less acid was required to drop the pH of the latter ores to permit flotation of contained sulfides. Use of acidic effluent from biooxidized ore to neutralize carbonate-containing ore led to another Newmont patent issued in 2009 (Brierley and Sawyer, 2009). The resulting flotation concentrates could then be used as fuel in the Carlin Mill 6 roaster, or sent to the Sage Mill autoclaves at Newmont’s Twin Creeks mine near Winnemucca, Nevada. However, as operating experience with the flotation circuit increased, and performance was improved, biooxidized ore only contributed 10–20% of the material processed through Mill 5 by 2007. In mid-2007, a thorough internal assessment was begun to consider
Please cite this article in press as: Roberto, F.F. Commercial heap biooxidation of refractory gold ores – Revisiting Newmont’s successful deployment at Carlin. Miner. Eng. (2016), http://dx.doi.org/10.1016/j.mineng.2016.09.017
F.F. Roberto / Minerals Engineering xxx (2016) xxx–xxx
the value of biooxidized ore to Mill 5 gold production (apparently sulfide-S oxidation rates had not improved). The improved Au recovery attributable to biooxidation when compared to flotation alone was only an incremental 7.2%. Handling of biooxidized ore also added cost and complexity in terms of crushing, rehandling, corrosion, scaling, and material properties already mentioned. When coupled to the relatively high cost to use biooxidized ore to neutralize high-carbonate ores compared to sulfuric acid, and the reduced sulfide-S content of biooxidized ore, which reduced the sulfide content of Mill 5 concentrate available for fuel to the Mill 6 roaster or the Sage Mill autoclaves, it was concluded that the biooxidation process should be suspended and held as an option for future use if excess sulfide-S became an issue in utilization of ores and concentrates across Newmont’s facilities in Nevada. To their credit, Newmont staff considered the option of improvements to the biooxidation process, such as a return to the original heap stacking using conveyor-stackers. This additional cost was viewed to make the case worse for continued operation of the biooxidation heaps. The assessment also considered that at some point, the real estate occupied by the idled biooxidation pads might become too valuable to remain vacant.
6. Conclusions Over a period of 12 years, Newmont microbiologists, metallurgists, and engineers took promising bench-scale work through a series of progressively larger pilot tests that led to the full-scale commercial implementation of a biooxidation process for pretreatment of refractory gold ores, including the ability to tackle preg-robbing refractory ores when followed by ammonium thiosulfate leaching. The project may be considered by some as a failure, but when examined in hindsight, considering the price of gold between the years 1999–2008, and significant technical compromises that were made due to the economic environment when the project was implemented, as well as complex, integrated considerations that are the reality of ore and concentrate management at Newmont’s Nevada facilities, it can be seen that whole-ore biooxidation of refractory gold ores was a viable and value-generating process that might still be operating today under different circumstances.
5
Acknowledgements The author is grateful to the Newmont pioneers who took whole-ore biooxidation of refractory gold ores from the laboratory to commercial implementation, particularly Dr. James Brierley and Mrs. Karen Tempel, who reviewed my internal retrospective on this work (Roberto, 2015), and provided valuable comments and corrections based on their experiences. References Barrick Gold Corporation, Thiosulphate Project,
Please cite this article in press as: Roberto, F.F. Commercial heap biooxidation of refractory gold ores – Revisiting Newmont’s successful deployment at Carlin. Miner. Eng. (2016), http://dx.doi.org/10.1016/j.mineng.2016.09.017
Contents lists available at ScienceDirect
Minerals Engineering journal homepage: www.elsevier.com/locate/mineng
Commercial heap biooxidation of refractory gold ores – Revisiting Newmont’s successful deployment at Carlin F.F. Roberto Newmont USA, Limited, Englewood, CO, USA
a r t i c l e
i n f o
Article history: Received 2 August 2016 Accepted 23 September 2016 Available online xxxx Keywords: Biooxidation Commercial heap leach Refractory gold ore Scale-up
a b s t r a c t Newmont Mining Corporation pioneered the investigation, development, and commercial-scale implementation of refractory gold whole-ore heap biooxidation, during a period spanning 1988–2009 at Carlin, Nevada. Basic and applied research and development from 1988 to 1999 included laboratory test work and increasingly larger pilot test heaps culminating in the full-scale implementation of a process that was estimated to contribute 120,000–180,000 oz/year to Carlin’s production between 2000 and 2005. Key parameters that influenced performance of the on-off heap biooxidation process, and factors that led to the discontinuation of the operation are described. Ó 2016 Published by Elsevier Ltd.
1. Introduction Newmont Mining Corporation pioneered the investigation, development, and commercial-scale implementation of refractory gold whole-ore heap biooxidation, during a period spanning 1988–2010 at Carlin, Nevada (Fig. 1 summarizes key developments and gold price during this period). Basic and applied research and development from 1988 to 1999 included laboratory test work and increasingly larger pilot test heaps leading to full-scale implementation of a process that was estimated to contribute 120,000– 180,000 oz/year (12.2 t total by May 2005; Logan et al., 2007) to Carlin’s gold production from 2000 to 2005. Even within Newmont, it is difficult to determine exactly when the closure of the commercial operation (internally referred to as the Refractory Leach Project) occurred, but by combining U.S. Securities and Exchange Commission annual 10-K reports and internal reports, it can be concluded that initial planning for project closure began in 2007, operating data exist for 2008 and 2009, and a 2011 final closure plan for the biosolution pond described cessation of operations in 2010. It is likely then that the commercial biooxidation plant contributed at least another 5 tonnes (ca. 160,000 oz) of gold to Carlin production before closure in 2010, after a decade of operation. Brierley et al. (1995) summarized the pilot heap tests that predated the Gold Quarry biooxidation demonstration facility (Schutey-McCann et al., 1997), a $13.5 million project that constructed and operated an 800,000 t/year large-scale pilot that integrated lessons learned from the early, smaller heap tests. This
E-mail address: [email protected]
facility was operated from 1994 to 1997, and the biooxidation and ammonium thiosulfate leach pads were ultimately enlisted to validate ‘‘bio-milling” at Carlin’s Mill 5 carbon-in-leach (CIL) facility in conjunction with the commercial biooxidation process. Initial bench-scale work to test the feasibility of biooxidizing Carlin low-grade refractory gold ores began in 1988 at Newmont’s central metallurgical services laboratory, then located in Salt Lake City. Carlin-type ore is typified by microscopic or dissolved gold disseminated in pyrite, arsenopyrite, or arsenian pyrite hosted within carbonaceous sediments. Concerns regarding increasing sulfide content in Gold Quarry ore at that time justified an examination of biooxidation as an alternative processing strategy for low-grade refractory ore.
2. Small pilot heap test summary A Refractory Leach Test Facility to construct the first small pilot heap (476 t) was authorized by Newmont management in 1990, and constructed adjacent to the Gold Quarry pit at Newmont’s Carlin Mine. By 1994, a series of successively large heaps were operated in different areas of the Carlin mine (North and South Areas), ranging from 476 t to 25,900 t. Many concepts that have become accepted practice in commercial copper bioleaching and were used during operation of Newmont’s commercial whole-ore biooxidation pads were tested during this period, including use of drip emitters, rather than wobbler, or impact head sprinklers, use of Wilfley diffusers in bacterial production bioreactors, the beneficial effect of rest periods on bacterial activity (as measured by accumulated ferric iron and increased Eh in leach solution),
http://dx.doi.org/10.1016/j.mineng.2016.09.017 0892-6875/Ó 2016 Published by Elsevier Ltd.
Please cite this article in press as: Roberto, F.F. Commercial heap biooxidation of refractory gold ores – Revisiting Newmont’s successful deployment at Carlin. Miner. Eng. (2016), http://dx.doi.org/10.1016/j.mineng.2016.09.017
2
F.F. Roberto / Minerals Engineering xxx (2016) xxx–xxx
Fig. 1. Timeline of developments leading to Commercial Refractory Leach Facility.
and observation that mesophilic iron-oxidizing bacteria such as Leptospirillum and moderately-thermophilic bacteria such as Sulfobacillus emerge without inoculation and are presumably enriched from the environment as sulfide minerals are oxidized. Newmont evaluated heaps that contained not only siliceous, sulfidic, refractory (SSR) ore, but also carbonaceous, sulfidic, refractory (CSR) ores that possess significant ability to remove cyanidecomplexed gold (so-called ‘‘preg-robbing” ore), generally thought to be a function of non acid-soluble, organic carbon in the ore. The test heaps included variations in sulfide sulfur, clay, pregrobbing carbon, carbonate, biooxidation cycle, and particle size. Belt agglomeration with a substantial (106–107 cells/mL) quantity of microbial inoculum (Acidithiobacillus ferrooxidans) reduced to practice what would become Newmont’s patented BIOPROÒ technology, with 4 related patents (Brierley and Hill, 1993, 1994, 1998, 2002, 2004). Heaps were stacked with either a front-end loader or radial stacker (larger heaps). The acidic conditions of biooxidation created within the heaps precluded direct leaching with cyanide or other gold complexing agents. Ore was off-loaded from the heaps with a front-end loader and partially neutralized with cement, caustic, or lime addition to achieve alkaline pH. Gold extraction with cyanide, thiourea, or ammonium thiosulfate (ATS) was tested. It was concluded that thiourea was ineffective perhaps due to the large particle size of the ore, while ATS yielded results equivalent to, and in several cases better than cyanide. Newmont patented the ATS process for gold leaching from preg-robbing ores (Wan et al., 1994) and tested this approach at larger scale in the subsequent Gold Quarry demonstration. The coupling of this lixiviant to biooxidation of preg-robbing ores was initially planned to be a key feature of the commercial whole-ore biooxidation facility. Barrick Goldstrike recently achieved commercial production of gold using the Total Carbonaceous Material (TCM) leach process based on calcium thiosulfate in the 3rd quarter of 2015 (Barrick Gold Corporation, 2016). Other significant parameters important to operation of biooxidation heaps for refractory gold ore that emerged from these tests included: (a) defining a lower limit for sulfide-S content of 0.2– 0.4%, (b) crushing to a smaller particle size is not always better for biooxidation within the heap, particularly with the presence of clays, and (c) carbonate content up to 2.2% could be accommodated in spite of the requirement of low pH conditions to promote growth of the iron-oxidizing bacteria, although considerable sulfuric acid addition was necessary to achieve pH control. It is interesting that active aeration through the use of a blower and air distribution lines at the base of the heap was tested during this
time as well, but the benefit was not clear, although it was noted that heap residues appeared to indicate uniform biooxidation had occurred throughout the heaps and might be attributed to aeration. A record low temperature for the region ( 26.1 °C) was experienced in December 1992 and caused concern, although the heap internal temperatures were observed to remain about 10 °C. A solution heater was introduced to prevent freezing of the recirculating solution but did not appear to have measurable benefit on the process within the heap. Finally, the cost for biooxidation of refractory gold ores in heaps was estimated at that time to be within the range of $4–6/t. The commercial plant costs were later found to be at the low end of this range.
3. Gold Quarry demonstration The scale-up of facilities to process up to 800,000 t/year of lowgrade (1–3 g/t Au) siliceous sulfidic refractory (SSR) and carbonaceous sulfidic refractory (CSR) ores on five 136,000 t leach pads required a bacterial production capability that incorporated six 200 m3 (1200 m3 total) tanks started in batch mode, ultimately producing 800 m3/day in continuous mode flowing into two solution storage ponds: a conditioning pond with a capacity of 800 m3, and an operating pond of 2377 m3 volume. An acid neutralization plant was part of the initial design for the facility, but no data could be located in internal documents or published reports that described details of operation or performance. Crushing utilized the existing Gold Quarry crushing circuit to generate nominal 19 mm material after passage through a gyratory crusher and cone crushers. Target ore composition included 5–15% clay, 1.5–2.5% sulfide-S, and 0.07–0.5% acid-soluble carbon (the CSR material contained 1% organic carbon versus 0.07% in the SSR material). Two methods of heap stacking were evaluated, comparing conveyor stacking with truck-dumped ore. Ore was stacked to a height of 10 m. Ore was inoculated at a rate of 24 L inoculum/t at the loadout conveyor for the haul trucks in the case of truck-stacked ore, while it was sprayed over the crushed ore at a portable feed hopper dropping ore onto the conveyor-stacker train at 18 L/t for conveyor-stacked heaps. Five different heaps were constructed to test liner designs including clay and HDPE, as well as combinations with clay base, and drain rock. No details were available regarding this evaluation, although it is presumed that the results were considered in the final commercial plant design.
Please cite this article in press as: Roberto, F.F. Commercial heap biooxidation of refractory gold ores – Revisiting Newmont’s successful deployment at Carlin. Miner. Eng. (2016), http://dx.doi.org/10.1016/j.mineng.2016.09.017
F.F. Roberto / Minerals Engineering xxx (2016) xxx–xxx
Substantial differences were noted in the hydraulic conductivity of truck-stacked SSR ore (2.8 10 4 cm/s) and conveyorstacked SSR (6.7 10 2 cm/s) and CSR (1.3 x 10 1 cm/s) ore. Dramatic reduction in daily sulfide-S oxidation rates – 0.125%/day for the truck-dumped SSR compared to 0.22%/day and 0.175%/day for the conveyor-stacked SSR and CSR ores was attributed to the reduced conductivity of solution. Schutey-McCann et al. (1997) concluded their summary of the Gold Quarry demonstration stating that ‘‘. . .to achieve the optimum Global Oxidation rate Newmont will conveyor stack all bioleach ore” which sadly, was trumped by fiscal considerations in the operation of the commercial biooxidation facility. Extremely high peak temperatures approaching 80 °C were observed for the conveyor-stacked SSR ore heap at about 85 days, while conveyor-stacked CSR ore heap internal temperatures increased slowly to about 60 °C over 160 days, and temperatures in the truck-dumped SSR heap declined from an initial temperature of 60 °C to about 30 °C after 160 days. Other chemistry extremes were also observed, including peak total iron levels >40 g/L, and Fe2+ over 20 g/L. This high ferrous iron level was later seen to cause problems during preparation of the biooxidized ore for cyanide or ATS leaching for Au recovery. Rinsing of the ore prior to neutralization was attempted, but this led to localized iron precipitation and production of sludge build up in the solution pond, and was discontinued. Best results were obtained by direct lime addition to the biooxidized ore in a pug mill. In spite of failing to achieve a target sulfide-S oxidation (55%) in any of the test heaps, the overall results of this much larger-scale pilot were considered a success, and a prefeasibility study commissioned in 1995 used results of the demonstration to make a case for a full-scale commercial biooxidation facility and justify the reclassification of 42 Mt of waste rock as reserves containing 62 t (2 million oz) Au.
4. Commercial operations (Refractory Leach Project) The qualified success of the Gold Quarry demonstration project increased confidence that Newmont now had another process for pretreatment of low-grade refractory gold ore, and that coupled with ATS leaching, could also address preg-robbing, carbonaceous ores. Funds were authorized to proceed with construction of a commercial plant in the South Area of the Carlin mine in 1997. This plant would have included twelve 147 m 305 m aerated on/off leach pads with an annual throughput of 11.7 Mt/year (32,000 t/ d) producing 8.4 t Au/year. Ore was to be crushed through 3 stages to achieve a P80 of 10 mm, agglomerated with microbial inoculum in a drum agglomerator, and stacked to a final height of 10 m by conveyor-stacker, due to the clear benefits shown in the Gold Quarry demonstration with respect to increased air permeability and hydraulic conductivity when compared to truck-dumping. A lime precipitation water treatment circuit was to be included in the plant design to overcome solution chemistry problems that plagued many of the pilot heap tests, and minimize the carryover of excessive iron and other metals to the gold leach circuit, thus reducing cyanide consumption or interference with ATS. Biooxidized ore was to be moved from the biooxidation leach pads to the gold leach pads using a bucket wheel excavator. Circulating biosolution was to be contained in a 113,500 m3 pond. However, the gold industry experienced a dramatic decline in gold price below $300/oz between 1997 and 2000 that led to the delay and substantial reduction in the project budget that significantly changed the as-built project (see Fig. 1 timeline). The final budget was less than 50% of the original proposed amount, and major revisions to the project design included a reduction in the
3
number of biooxidation pads from 12 to 3, a larger, crushed particle size target P80 of 19 mm, elimination of the agglomeration drum, conveyor-stacker, water treatment circuit, bucket wheel excavator, and instead of gold recovery by heap leaching, introduction of the biooxidized ore directly into the Carlin Mill 5 CIL plant (this coupled process was referred to as ‘‘bio-milling”). Fig. 2 depicts the operating facility, with the 3 pads in various stages of loading, oxidizing, or unloading. As will be discussed later in this paper, the decision to direct biooxidized ore to Mill 5 might have sealed the fate of commercial whole-ore biooxidation at Carlin, since changes in the Mill 5 operating philosophy and design ultimately reduced the value of biooxidized ore to overall Au production at Carlin. First-year operating results were summarized by Bhakta and Arthur (2002) and described the bio-milling process at Mill 5, where a semi-autogenous grinding (SAG) mill and ball mill produced a P65 = 75 lm leach feed to an 8-tank CIL circuit with 8– 10 h retention time. The biooxidized ore (160 days of biooxidation) was neutralized after off-loading by addition of pebble lime (7.5– 15 kg/t) at the SAG mill feed belt. Initial operating results with this feed noted low dissolved oxygen levels and increased cyanide consumption, and lower than expected gold recovery. It was discovered that ferrous iron, copper, zinc, and nickel were introduced by the biooxidized ore and likely accounted for the oxygen depletion and additional cyanide utilization. These effects were mitigated somewhat by the addition of air to the slurry in the surge tanks upstream of the CIL circuit, and reduction in the slurry loading, but it was concluded that a larger water treatment facility was needed for the biooxidation circuit to remove metals before biomilling. This was never built, although various options were explored. 2.4 million t of ore were processed in the first year, about 70% of the target throughput. Tempel (2003) revisited the performance of the commercial biooxidation process in 2003, and described how creative operations overcame some of the design compromises that were necessitated by a reduced project budget. Coarser particle size of the crushed ore somewhat compensated for the expected reduction in hydraulic conductivity and air permeability resulting from truck-dumped stacking of the heaps. In addition, stacking of the heaps to an initial height of 13.7 m, bulldozing of approximately 0.9 m of this ore to a height of 12.8 m, and finally, ripping the pad surface to a depth of approximately 1.5 m was performed to eliminate much of the estimated 2.4 m deep compaction layer created by 140-t haul trucks. Off-loading of the pads was mentioned as another significant problem, since the reduced budget only permitted changing the pad base air lines after 3 cycles. One meter of ore was retained after bulldozing the bulk of the biooxidized ore from the pads, but damage still occurred to air lines. Cementation of dried leach precipitates was also observed around the air lines on occasion, and noted to be extremely hard, with attempts to break up the material often leading to additional air line damage. Observation of the biooxidation pads with a thermal imaging camera revealed cold spots apparently experiencing reduced biooxidation. It was determined that higher-carbonate content ore from the North Area was present. Attempts to blend this ore with Gold Quarry ore with reduced carbonate content were unsuccessful in improving biooxidation rates, and actually depressed the sulfide-S oxidation of the lower carbonate ore. The temperature extremes of the Carlin site (cold winters and hot summers) posed a problem for controlling agglomeration moisture, as the initial 4% moisture target led to saturated ore that caused sloughing on the heaps and localized ponding. The heaps themselves achieved internal temperatures well beyond the design limits of the liners (rated to 60 °C), sometimes reaching 81 °C on the first commercial pad. Attempts to cool the heaps through
Please cite this article in press as: Roberto, F.F. Commercial heap biooxidation of refractory gold ores – Revisiting Newmont’s successful deployment at Carlin. Miner. Eng. (2016), http://dx.doi.org/10.1016/j.mineng.2016.09.017
4
F.F. Roberto / Minerals Engineering xxx (2016) xxx–xxx
Fig. 2. Refractory Leach Facility, 2001.
increased irrigation resulted in saturation of the heap which inhibited bacterial oxidation. The high internal temperatures in the heaps led to the decision to inoculate with archaeal species 6 months after the facility began operation, and these microbes were maintained at cell densities between 106 and 108 cells/mL, and appeared to benefit the process. This addition to the microbial inoculum was noted as another operating change that resulted from modifications to the initial project plan. Large amounts of ferrous iron were introduced into the biosolution pond with breakthrough of irrigation in the biooxidation pads. This observation was interpreted to mean that the microbial activity in the heaps was insufficient to regenerate ferric iron for additional pyrite oxidation. Therefore irrigation was suspended to increase Eh and apparent microbial activity. Increased heap temperatures also accompanied cessation of irrigation. At the same time, oxygen depletion was noted in the biosolution pond, along with a drop in Eh below 450 mV. Use of a commercial aeration device designed for municipal wastewater treatment plants increased the dissolved oxygen concentration and improved the Fe3+/Fe2+ ratio. Excessively high total iron concentrations had been seen in some of the pilot heaps and the Gold Quarry demonstration, and while the average levels in the commercial pads ranged from 8 to 26 g/L, levels as high as 59 g/L were observed at start of ore leaching. Not for the first time a comment was made about the ‘‘. . .need to treat the biosolution to lower metals concentration. . .” The soluble metals carried over to the bio-milling process at Mill 5 were considered at least in part, to contributing to the observed 3-fold increase in cyanide consumption at the mill compared to oxide ores. Estimation of overall sulfide-S oxidation after a biooxidation cycle was noted to be a problem, as the results of tail assays indicated sulfide-S levels higher than the head assays taken from belt samples during truck loading. It was determined that these higher tail assays resulted from artifacts of residual sulfate contributing to the apparent sulfide-S quantified by the LECOÒ carbon/sulfur roast method in use at the time. In spite of the inaccurate sulfide-S determinations, it was acknowledged that the biooxidation process was not achieving the target oxidation (which ultimately reduced Au recovery), and that the ability to perform a longer biooxidation cycle could be accomplished if more heaps were available. This was noted to also provide some flexibility for different ore types. Unfortunately, it is apparent that a request to increase the number of biooxidation pads was not approved. Logan et al. (2007) provide the most comprehensive overview of the commercial biooxidation facility design and operation, through
mid-2005. Readers are directed to that excellent book chapter, coauthored by Jim Brierley, for detailed operational commentary including a microbiologist’s perspective. By that time, 12 batches totaling 8.8 Mt of ore had been processed, resulting in the production of 12.2 t (379,500 oz) Au. Optimal solution chemistry parameters included a pH range of 1.3–2.2, Eh of >550 mV, Fe3+ concentration between 5 and 25 g/L, and dissolved oxygen concentration >2 ppm. Some details that were not described in the previous operating summaries (Schutey-McCann et al., 1997; Tempel, 2003) discussed above included heap irrigation rates (4 L/m2 h) through drip emitters spaced on 760 mm centers, and liner construction (1 mm HDPE liner on 150 mm clay base), and placement of aeration lines in 1 m of crushed rock below the actual ore base. They note that truck dumping of the ore may have reduced sulfide-S oxidation rates by as much as 50%, and that when the heaps were off-loaded, dry areas that had not received irrigation were estimated to comprise as much as 10% of the heap volume. The dry areas could have resulted from solution short-circuiting, or excessive drying by aeration (the low humidity in the Carlin area, less than 30%), or a combination of the two. These observations, when considered with apparent cold regions of the heaps using thermal imagery made by Tempel, suggest that reduced solution and air permeability, along with higher localized regions of carbonate within the heaps where biooxidation was inhibited, could account for the disappointing overall global oxidation of sulfide-S of 21.9%, compared to the target 30%. In 2004, a froth flotation circuit was added to Mill 5, in order to increase recovery of gold from biooxidized ore (Au that remained locked in unoxidized sulfides). An additional 10% gold recovery was attributed to this ‘‘bio-float” modification to Mill 5 in 2005.
5. Mill 5 and the demise of the Refractory Leach Project For several years, the acidic biooxidized ores were presumed to be a benefit to processing high-carbonate ores through the modified Mill 5, as less acid was required to drop the pH of the latter ores to permit flotation of contained sulfides. Use of acidic effluent from biooxidized ore to neutralize carbonate-containing ore led to another Newmont patent issued in 2009 (Brierley and Sawyer, 2009). The resulting flotation concentrates could then be used as fuel in the Carlin Mill 6 roaster, or sent to the Sage Mill autoclaves at Newmont’s Twin Creeks mine near Winnemucca, Nevada. However, as operating experience with the flotation circuit increased, and performance was improved, biooxidized ore only contributed 10–20% of the material processed through Mill 5 by 2007. In mid-2007, a thorough internal assessment was begun to consider
Please cite this article in press as: Roberto, F.F. Commercial heap biooxidation of refractory gold ores – Revisiting Newmont’s successful deployment at Carlin. Miner. Eng. (2016), http://dx.doi.org/10.1016/j.mineng.2016.09.017
F.F. Roberto / Minerals Engineering xxx (2016) xxx–xxx
the value of biooxidized ore to Mill 5 gold production (apparently sulfide-S oxidation rates had not improved). The improved Au recovery attributable to biooxidation when compared to flotation alone was only an incremental 7.2%. Handling of biooxidized ore also added cost and complexity in terms of crushing, rehandling, corrosion, scaling, and material properties already mentioned. When coupled to the relatively high cost to use biooxidized ore to neutralize high-carbonate ores compared to sulfuric acid, and the reduced sulfide-S content of biooxidized ore, which reduced the sulfide content of Mill 5 concentrate available for fuel to the Mill 6 roaster or the Sage Mill autoclaves, it was concluded that the biooxidation process should be suspended and held as an option for future use if excess sulfide-S became an issue in utilization of ores and concentrates across Newmont’s facilities in Nevada. To their credit, Newmont staff considered the option of improvements to the biooxidation process, such as a return to the original heap stacking using conveyor-stackers. This additional cost was viewed to make the case worse for continued operation of the biooxidation heaps. The assessment also considered that at some point, the real estate occupied by the idled biooxidation pads might become too valuable to remain vacant.
6. Conclusions Over a period of 12 years, Newmont microbiologists, metallurgists, and engineers took promising bench-scale work through a series of progressively larger pilot tests that led to the full-scale commercial implementation of a biooxidation process for pretreatment of refractory gold ores, including the ability to tackle preg-robbing refractory ores when followed by ammonium thiosulfate leaching. The project may be considered by some as a failure, but when examined in hindsight, considering the price of gold between the years 1999–2008, and significant technical compromises that were made due to the economic environment when the project was implemented, as well as complex, integrated considerations that are the reality of ore and concentrate management at Newmont’s Nevada facilities, it can be seen that whole-ore biooxidation of refractory gold ores was a viable and value-generating process that might still be operating today under different circumstances.
5
Acknowledgements The author is grateful to the Newmont pioneers who took whole-ore biooxidation of refractory gold ores from the laboratory to commercial implementation, particularly Dr. James Brierley and Mrs. Karen Tempel, who reviewed my internal retrospective on this work (Roberto, 2015), and provided valuable comments and corrections based on their experiences. References Barrick Gold Corporation, Thiosulphate Project,
Please cite this article in press as: Roberto, F.F. Commercial heap biooxidation of refractory gold ores – Revisiting Newmont’s successful deployment at Carlin. Miner. Eng. (2016), http://dx.doi.org/10.1016/j.mineng.2016.09.017