SPECIAL ISSUE ON PROCESS DEVELOPMENT TESTING
Process development metallurgical studies for gold cyanidation process S. Acar
Consulting metallurgical engineer (D.E.Sc.), Highlands Ranch, CO, USA
Abstract
In recent years the gold mining industry has been challenged by declining metal prices as well as having to treat ores that have low grades and are refractory in nature. Many factors must be considered when metallurgical studies are designed to develop process design criteria for these type of ores. These factors are mostly related to the mineralogical and chemical compositions of the ores, the geological setting of the deposit and how the ore/waste will be mined, such as whether openpit or underground, the mine schedule, and the mine fleet type and capacity. The purpose of this paper is to provide an experience-based practical approach to how to select and characterize test samples as well as devise process development metallurgical studies for both oxide and refractory gold ores, and to provide insight into some of the matters that might go wrong during the course of the studies. This paper is intended as both a guide for someone new to the field and also as a reminder for those familiar with the field.
Minerals & Metallurgical Processing, 2016, Vol. 33, No. 4, pp. 161-171. An official publication of the Society for Mining, Metallurgy & Exploration Inc. http://dx.doi.org/10.19150/mmp.6837
Key words: Cyanide leaching, Gold processing, Refractory ores
Introduction
The recent decline in ore grades and metal prices have challenged mining companies to reduce capital and operating costs associated with developing new mineral resources. In some cases, metallurgical studies have been sacrificed to prevent cost overruns resulting in inadequate mineralogical, chemical and metallurgical characterization. Consequences have included longer construction schedules, higher initial and working capital, higher operating costs and less than desirable plant performance (Whincup, 2010; Torres, Chaves and Meech, 1999). Mineralogy, geology, mining type and sequence, and the metallurgy of an orebody will determine the process development metallurgical study requirements and eventual process plant design. Therefore, once the geologist has an early indication of a prospective orebody, geologists, metallurgists, mining engineers, and social and environmental responsibility scientists should convene and devise a plan on how to exploit the orebody economically (Lunt, Ritchie and Fleay, 1997). Geologists may find a prospective orebody solely based on grade. However, if the orebody cannot be mined and processed economically, then it is deemed not to be a deposit. An example of this could be a refractory gold deposit, either sulfidic, carbonaceous
or siliceous refractory or any combination of those, that may not be mined and processed economically due to reasons such as the physical location of the deposit, depth of the ore lode, or social and environmental concerns. Therefore, an early collaboration between all disciplines must be established. Each orebody may have its own unique mineralogical, geological and chemical characteristics (Guresin et al., 2012). These unique characteristics should be effectively sampled to reduce risks associated with process design and engineering efforts for the process development metallurgical studies. If the sample or samples are not representative of these unique characteristics, then the studies performed will be pointless. The project metallurgist in charge of conducting metallurgical studies should also possess general geology, mineralogy and mine engineering knowledge. Recommended readings are a book on understanding mineral deposits by Misra (2000) and articles on heap leach development by Scheffel (2014a and 2014b).
Process development metallurgical studies
Developing a metallurgical program depends on the type of ore under consideration. Gold ores may be divided into several categories (Marsden and House, 2006): 1. Oxide ores with or without coarse gold component. 2. Refractory ores with or without coarse gold component:
Paper number MMP-15-063. Original manuscript submitted August 2015. Revised manuscript accepted for publication May 2016. Discussion of this peer-reviewed and approved paper is invited and must be submitted to SME Publications Dept. prior to May 31, 2017. Copyright 2016, Society for Mining, Metallurgy & Exploration Inc. MINERALS & METALLURGICAL PROCESSING
161
Vol. 33 No. 4 • November 2016
Once the ore type is determined, obtaining the right samples — variability samples and master composites — for the process development metallurgical studies will easily be accomplished.
a. Sulfidic refractory ores, such as: • Pyrite, arsenopyrite and arsenian pyrite containing ores. • Telluride containing ores. • High cyanide and oxygen consuming ores (copper containing ores, ores with marcasite and/or pyrrhotite). • Enargite containing refractory ores. • High elemental sulfur containing ores. • Massive sulfide ores. • Antimony containing ores. b. Carbonaceous refractory (preg-robbing) ores. c. Siliceous refractory ores. d. Any combination of the above. 3. High silver containing gold ores.
Sampling the deposit for metallurgical studies
Sampling is conducted to assess geological and mineralogical variabilities within an orebody and the effects of these variabilities on metallurgical processes (Williams and Richardson, 2004). Metallurgical studies conducted on the samples are only attributable to the sample and what it represents. Therefore, sample quality is extremely important for the development of a metallurgical process for a particular deposit and troubleshooting during operations (Chieregati and Pitard, 2009; Dominy, 2009). The project metallurgist should always question, “Is
Table 1 — Ore types for a gold/copper/zinc ore deposit. Ore type
Au grade, opt
Cu grade (%)
Acidsoluble Cu:total Cu ratio (%)
Cyanidesoluble Cu:total Cu ratio (%)
Zinc grade (%)
Organic carbon content (%)
Lead grade (%)
Additional considerations
Gold ores Au-1
Oxide
0.008
Zn<0.1
a: CN/FA>40%
Au-2
Oxide
Au>0.1
Zn<0.1
b: CN/FA<40%
Au-3
Oxide
Au>0.008
Au-5
Sulfide
Au>0.03
Cu-1
Oxide
Cu>0.1
>40
Cu-2
Oxide
Cu>0.1
>40
Cu-3
Oxide
Cu>0.1
<40
<40
Cu-4
Oxide
Cu>0.1
<40
<40
Zn>0.1 oxide
Cu-5
Oxide
Cu>0.1
<40
>40
Zn>0.1 oxide
Cu-11
Sulfide
Cu>0.1
>40
Zn<0.1
d: Au<0.03 opt
Cu12
Sulfide
Cu>0.1
>40
Zn>0.1
e: 0.03
Cu13
Sulfide
Cu>0.1
>40
Zn>0.1
f: Au>0.1
Cu14
Sulfide
Cu>0.1
<40
Primary
Zn-0
Oxide
Zn>0.5
Zn-1
Sulfide
Zn>0.5
<0.2
Zn-2
Sulfide
Zn>0.5
<0.2
Zn-3
Sulfide
Zn>0.5
>0.2
Zn-4
Sulfide
Zn>0.5
>0.2
Zn>0.1 Cu<0.1
Zn<0.5 Copper ores a: Au<0.01 opt Zn>0.1 oxide
<40
b: 0.010.03
ore
Zinc ores
November 2016 • Vol. 33 No. 4
162
d: Au<0.03 opt >0.1
e: 0.030.1
>0.1
MINERALS & METALLURGICAL PROCESSING
the sample representative of the orebody?” For process development metallurgical studies, samples should be selected from geological sections that spatially cover the orebody representing metallurgical, mineralogical and geological changes: that is, domains. Individual interval samples, which may be called variability samples, should be taken across a zone to obtain a response to changes in metallurgical type. The length of the variability sample should be, at minimum, a mineable length and restricted to a given domain. Variability samples should not cross over different domains and should include dilution. The mineable length will depend on the mining process, such as underground or openpit, and may vary from a 2.5-m adit opening to 10-m bench height. Master composites may be prepared to fulfill requirements for extensive process development studies by combining several variability samples representing a given domain while paying attention to spatial location. The following items should be carefully considered when sampling the deposit for process developmental testing (Guresin et al., 2012): 1. Changes in lithology — involving physical characteristics such as color, texture and grain size of a rock unit — and alteration, such as weathering, supergene alteration, hydrothermal alteration. 2. Information on structure and texture, such as fractures, faults, brecciation. 3. Oxidation state — relative intensity of oxidation and alteration, that is, oxide, transition and sulfide ores (Hoal, Woodhead and Smith, 2013). 4. Mineralization type, such as oxide versus refractory gold, oxide versus hypogene copper. 5. Ore and gangue mineralogy — such as mineral particle size and liberation characteristics, and valuable element deportment — and hardness. 6. Depth and spatial relations of different ore lodes and grade variations. 7. Chemical analysis, including valuable elements, elements that may constrain throughput and recovery, penalty elements and process indicators such as sequential copper assays, cyanide soluble gold assay, preg-robbing gold assay. 8. Mining method, plan and schedule. 9. Process options considered. A case study is presented here to demonstrate the variability sampling of a complicated gold deposit with copper and zinc credits. In this deposit, oxide gold ore zones are located closer to surface followed by the sulfidic refractory gold ore zone located at depth. A review of geological sections, drill logs and drill interval assays indicates that the deposit also contains a copper ore zone adjacent to the sulfidic refractory gold ore zone and a zinc ore zone at depth. Using available data such as lithology, alteration, and chemical and mineralogical characterizations as a guide, the following gold, copper and zinc metallurgical ore types may be identified as shown in Table 1, which demonstrates how complicated the variability sampling of this deposit would be. The gold ore zone may be divided into an oxide and sulfidic ore type zones. The oxide gold ore zone is further divided into several distinct ore types depending upon the process options, such as heap leaching or milling based on the gold grade and cyanide-soluble gold, Au(CN), to total gold, Au(FA), ratio. For example, Au2A ore type is defined as an oxide millable ore — indicated by a Au(CN) to Au(FA) ratio that is greater than 40 percent — with gold grades in excess of Au > 0.1 opt (3 g/t MINERALS & METALLURGICAL PROCESSING
163
Au). Special attention should also be paid to the copper and zinc contents for the sulfidic refractory gold ore type — indicated by a Au(CN) to Au(FA) ratio that is less than 40 percent — because they are known cyanide consumers, or cyanicides. Similarly, copper ores are separated into two distinct ore zones: oxide and sulfide. The oxide copper ores are further divided into several subtypes depending upon the acid- and cyanide-soluble copper contents — indicating different copper mineralogy — zinc mineralogy and gold grades. For example, Cu2B ore type is defined as oxide heap-leachable copper ore — having an acid-soluble copper to total copper ratio that is greater than 40 percent — with gold content of 0.01 > Au < 0.03 opt. For this ore type, the conducting of further laboratory studies is recommended to recover gold post-acid leaching for oxide copper minerals. Similarly, sulfidic copper ores may be further divided into four subtypes, each containing three different gold grade ranges. Zinc ores are also divided into two types: oxide and sulfide. The sulfidic zinc ores are further split into zones with varying organic carbon, lead and gold grades to determine process options. These ore types are schematically illustrated in Fig. 1, which shows a long section of the deposit. When preparing master Au, Cu and Zn composites for process development metallurgical studies, special attention should be paid to the spatial location, and mineralogical and chemical analyses of the individual variability samples. As a result, master composites would be prepared to represent a mining horizon with similar geometallurgical domains. The metallurgical laboratory conducting the process development study performed extensive studies to determine the process options for this deposit. The completed studies included: • Column and agitated bottle roll cyanide leach studies for oxide gold ores to determine heap leachability. • Agitated bottle roll cyanide leach studies for oxide gold ores to determine milling. • Whole ore roasting, pressure oxidation or biooxidation followed by cyanide leach for refractory gold ores. • Flotation followed by concentrate treatment for refractory gold ores and ores containing copper and/or zinc sulfides. • Column acid leach studies for copper and zinc oxide ores. • Column bioleach studies for copper and/or zinc sulfides. • Acid or bioleach followed by cyanide leach for copper and/or zinc ores with gold credits.
Sample types
A diversity of sample types may be used during various stages of metallurgical studies (Hanks and Barratt, 2002), including: • • • • •
Grab samples. Channel samples. Bulk samples. Reverse circulation drilling samples. Diamond core drilling samples, including large-diameter metallurgical core.
Depending upon availability, the preference should always be to use diamond drill core for variability samples and master composites during all stages of metallurgical studies. Exploration departments may use the diameter sizes of NQ (48 mm), HQ (64 mm) and PQ (85 mm) for core drilling. In most cases, one-half cut core is used for geochemical analysis and Vol. 33 No. 4 • November 2016
petrography studies, one-quarter cut core is stored as a catalog and the remaining one-quarter cut core is made available for metallurgical studies. Variability samples should be prepared for the metallurgical study using the quarter-cut drill core, and the amount of drill core available is usually a sore point between the geologist and the metallurgist. Most of the time, the project metallurgist will require several additional PQ and/ or 150 mm (6 in.)-diameter metallurgical core to be drilled to provide enough bulk sample and the top size required for some of the proposed studies. Table 2 summarizes the sample weights available for cores of NQ, HQ and PQ sizes.
diagnostic leach studies may be conducted for ores are refractory in nature. For all ore types, the following characterization studies are recommended: Chemical analysis. Chemical analysis must be conducted on variability samples and master composites to determine minerals of value and deleterious elements, including those that will have impacts on cost and environmental liabilities (Dietrich, 2014). The chemical analysis may include the following assays: Gold by fire assay. Gold by fire assay is the preferred method of analysis. If the estimated gold assay is Au < 5 ppm, fire assay with AA finish should be requested. If the gold content is exceeding 5 ppm, fire assay with a gravimetric finish will be required (McGuire, 1989; ALS Ltd., 2012; Hoffman, Clark and Yeager, 1999). Silver assay. Samples with substantial silver content of more than 300 ppm may also be measured for silver by fire assay, although inductively coupled plasma is the preferred method for Ag < 300 ppm. Cyanide-soluble gold assay. The cyanide-soluble gold (AuCN) procedure is an indicator of nonrefractory gold that
Developing a metallurgical program
Once the samples are obtained, all variability samples should be prepared according to crush or grind sizes for the perceived benchmarked process flowsheet. All variability samples and master composites must be subjected to chemical and mineralogical characterizations. Comminution (crushing and grinding) characterization studies should also be conducted on select master composites. Gravity-recoverable gold content studies may be conducted for ores containing free native gold, and
Figure 1 — A geological long section showing gold, copper and zinc variability samples as functions of depth. November 2016 • Vol. 33 No. 4
164
MINERALS & METALLURGICAL PROCESSING
Table 2 — Approximate sample weights for diamond drill cores with NQ, HQ and PQ sizes. Core size NQ
HQ
PQ
1
1
1
Core length
meter
Density
g/cm3
2.5
2.5
2.5
Core recovery
%
100
100
100
Void factor
%
0
0
0
inch
1.775
2.406
3.270
mm
45
61
83
Core diameter
Core size NQ
Weight
HQ
PQ
Full
Half
Quarter
Full
Half
Quarter
Full
Half
Quarter
kg
3.4
1.7
0.85
6.6
3.3
1.65
13.6
6.8
3.4
lb
7.5
3.75
7.875
14.5
7.25
3.625
30
15
7.5
is recoverable by cyanide leaching and is not a predictor of actual deposit gold recovery. The AuCN value is usually obtained by determining the gold solubility of a pulverized 5-g sample in 10 mL of 0.3 percent sodium cyanide (NaCN) solution after 60 minutes of contact time in a shaker flask. This AuCN value is then compared, by ratio, to the total gold content as determined through fire assay. The ratio of cyanidesoluble gold to total gold will give an indication of leaching difficulty due to several factors that may include the presence of sulfides, large particulate native Au, preg-robbing organic carbon and/or clay content. Preg-robbing gold assay. The preg-robbing gold (AuPR) assay is a diagnostic test used for estimating the preg-robbing gold capacity of the ore. It is identical to the AuCN protocol except that it also contains a gold spike, usually 0.1 opt Au (3.4 g/t Au). The AuPR number, which is calculated using the AuCN and AuPR assays, typically applies to any deposit containing organic carbon. The degree of preg-robbing may be determined by the preg-robbing number, PRN, calculated as follows: PRN = 3.4 g/t Au (Au in the spike solution) + AuCN (in g/t) (1) – AuPR (in g/t) where PRN = 0 indicates a nonpreg-robbing ore, while a PRN greater than 3 indicates an ore that is extremely preg-robbing. LECO series of assays. LECO is a trademark of LECO Corp. (St. Joseph, MI), and LECO instruments are used to determine the carbon and sulfur contents of ores (LECO Corp., 2016). Total carbon and sulfur contents should be measured for all samples using a LECO infrared combustion analyzer. Sulfide sulfur, S-Sulfide, measurement is required for samples for both metallurgical process selection and environmental acid generation potential estimation. Organic carbon, also called acid-insoluble carbon (AIC), is used to predict the preg-robbing potential of an ore. The organic carbon assay may also include nonpreg-robbing graphitic carbon. Organic carbon subtracted from total carbon represents inorganic carbon, or carbonate carbon, and is used to estimate the acid neutralization potential of the ore. Elemental analysis. Elements including much of the periodic table can efficiently be measured with a four-acid digestion and measured in combination with inductively coupled plasmaoptical emission spectrometry (ICP-OES), and inductively MINERALS & METALLURGICAL PROCESSING
165
coupled plasma-mass spectrometry (ICPMS). The ICP/ICPMS suite of assays generally requested are: Ag, aluminum (Al), arsenic (As), beryllium (Be), bismuth (Bi), calcium (Ca), cadmium (Cd), cobalt (Co), chromium (Cr), copper (Cu), iron (Fe), gallium (Ga), germanium (Ge), potassium (K), magnesium (Mg), manganese (Mn), molybdenum (Mo), sodium (Na), nickel (Ni), lead (Pb), rhenium (Re), antimony (Sb), selenium (Se), Tin (Sn), strontium (Sr), thorium (Th), titanium (Ti), thallium (Tl), uranium (U), vanadium (V), yttrium (Y), zinc (Zn) and zirconium (Zr). In sulfidic ores, a second aqua-regia digestion is necessary for volatile elements, primarily As. It is convenient to also measure Sb, Se and tellurium in this same digestion. Mercury analysis. Traditional mercury (Hg) analysis involved digestion followed by cold vapor measurement. However, in the presence of Au, Hg is suppressed by an amount equivalent to the Au concentration. Therefore the preferred method for Hg analysis is by using an Hg combustion analyzer, following the ASTM 6722-11 standard test method (ASTM International, 2015). Mineralogical characterization. Semiquantitative X-ray diffraction (XRD), X-ray fluorescence (XRF), petrography and optical microscopy, scanning electron microscopy/mineral liberation analysis (SEM/MLA), near infrared/cation exchange capacity (NIR/CEC) mineralogical requirements are determined as needed for each project in collaboration with the geology and metallurgical groups to quantify the primary ore minerals as well as gangue and clay minerals. Comminution characterization. A comminution characterization series of tests may include bond low energy impact, semiautogenous (SAG) mill comminution (SMC Testing, 2016), bond rod mill and ball mill indexes, bond abrasion index and high pressure grinding rolls tests. The sample weight and top particle sizes required for the comminution characterization tests will affect the metallurgical study sample requirements and preparation protocol (Hanks and Barratt, 2002; Morrell, 2008). Diagnostic leaching. Diagnostic leaching is a protocol to characterize gold association in the mineral matrix. Diagnostic leach is a series of acid leaching, each one more aggressive than the previous one. Diagnostic leaching is combined with interstage cyanide leaching to determine the amount of gold Vol. 33 No. 4 • November 2016
liberated from affected minerals. In general, acetic acid, sulfuric acid, hydrochloric acid and nitric acid are used either at room temperature or at elevated temperature depending upon the mineral matrix (Lorenzen, 1995; Armstrong and Malhotra, 1992; Tumilty and Schmidt, 1986).
to determine gold recovery, recovery rate, reagent requirements and sensitivity to feed size. The tests are usually conducted while maintaining 1 g/L NaCN and at pulp pH between 10.5 and 11 adjusted with lime. Minus 10 mesh bottle roll tests are conducted for at least 96 hours with recommended solution samples at 2, 4, 6, 8, 24, 48 and 72 hours to determine leach kinetics. Similarly, 80 percent −200 mesh bottle roll tests are conducted for 48 hours with recommended solution samples at 2, 4, 6, 8 and 24 hours. Cyanide concentration and pH should be determined for each kinetic solution sample and will be adjusted as necessary. Makeup water, equivalent to that withdrawn, should be added to the pulps. Lime will be added, when necessary, to maintain the leaching pH at between 10.5 and 11.0. After 48 or 96 hours, the pulps will be filtered and the final pH and cyanide concentrations will be determined. Leach residues will be washed, dried, weighed and assayed in triplicate to determine residual precious metal content. In addition to kinetic leach tests, carbon-in-leach (CIL) or carbon-in-pulp (CIP) tests may be conducted for some ores. Confirmation tests may be conducted at the NaCN concentration recommended for actual operation. For the milling process, once the cyanide concentration and grind particle size are optimized, then the variability samples should be subjected to an agitated bottle roll cyanidation study to determine deposit-wide gold recovery variations at the optimized condition.
Gravity (coarse) gold determination. Gravity gold recovery will be considered during project development for a number of reasons, such as the presence of coarse free gold content or the potential opportunity to improve cash flow by producing a more readily saleable product early in the flow sheet. There are many test procedures available to evaluate the amenability of a given sample to gravity gold recovery, including: • Screen fire assays. • Small-scale batch gravity tests. • Single- and three-stage GRG test (Laplante, Woodcock and Huang, 2000).
Process development metallurgical studies for oxide gold ores
Gold may be extracted from oxide ores through dump, heap or vat leaching, milling processes or any combination of the leaching and milling processes. The metallurgical program for oxide ores should therefore include dump/heap and vat leachability and milling studies (Fig. 2). In order to obtain the process design criteria, the following tests are recommended for oxide ores:
Column percolation cyanide leach studies. Column percolation leach tests should be conducted on master composite samples as a function of crush size, generally 80 percent passing 38 mm (1.5 in.), 19 mm (0.75 in.) and 3.35 mm (6 mesh Tyler) to determine gold recovery, recovery rate, reagent requirements and sensitivity to feed size under simulated heap leach conditions. Leaching is usually conducted by applying cyanide solutions, generally 0.5 g NaCN per liter, over the
Agitated bottle roll cyanidation leach studies. Kinetic cyanidation bottle roll leach tests should be conducted on master composites as a function of crush size, generally 80 percent passing 1.7 mm (10 mesh Tyler), and 75 µm (200 mesh Tyler)
Figure 2 — Oxide gold ores process development studies. November 2016 • Vol. 33 No. 4
166
MINERALS & METALLURGICAL PROCESSING
Figure 3 — Illustration of cyanide leach gold and silver extractions as functions of crush size.
charges at a rate of 10 L/hr/m2 of column cross-sectional area for a 10-m lift height for an industrial heap. Due to the difference in column size, the solution addition should be scaled down by factoring in ore mass. All columns should be leached until extinction before being washed and taken offline. In order to determine dump leachability, pilot-scale column leach tests using columns with large diameters, usually 2 m or 6 ft, should be conducted with 30-cm (12-in.) as-is or crushed bulk sample or samples. The gold and silver extractions for the agitated bottle roll and column leach tests should be plotted as functions of particle size (Fig. 3) to estimate gold and silver extractions from an industrial heap. The rule of thumb is to discount 3-5 percent from laboratory extractions when estimating industrial
heap gold extractions. For example, Fig. 3 indicates that gold extraction for a primary crushed ore with 80 percent passing 150 mm (6 in.) crush size would be about 53 percent. Rule of thumb would indicate using 48-50 percent gold extraction for an industrial heap at the same crush size. Similarly, rule of thumb for cyanide consumption for an industrial heap with oxide ores would be to use about one-third of the cyanide consumption indicated by laboratory testing. Column leach gold extractions are also plotted as functions of leach time (Fig. 4) and solution-to-ore ratio (Fig. 5). The data illustrated in Figs. 4 and 5 indicate that gold extraction is dependent on crush size and would be 55, 42 and 17 percent for 13 mm (0.5 in.), 25 mm (1 in.) and 76 mm (3 in.) crush sizes, respectively, considering the rule-of-thumb 5-percent deduction and solution-to-ore ratio of 1. The data presented in these
Figure 4 — Example of column leach gold extraction as a function of leach time for three crush sizes. MINERALS & METALLURGICAL PROCESSING
167
Vol. 33 No. 4 • November 2016
Hydraulic conductivity tests. Hydraulic conductivity tests, also called load-perm tests, should be conducted on select column leach residues to measure hydraulic conductivity rates for heap loading heights of 0, 20, 40, 60, 80 and 100 m. The measured hydraulic conductivity rates will indicate if solution flow/percolation problems are anticipated for simulated heap loading heights of up to 100 m. Solution flow problems are a result of fine particles, or slimes, migrating to the lower portions of the heap and blinding it, and also the presence of clay minerals, especially expansive clays, like montmorillonite and smectite. In order to improve solution percolation rates, the ore may require agglomeration with lime, cement or a combination of lime and cement (Garcia and Jorgensen, 1997).
figures indicate that this ore type may not be heap leachable as it will require considerable size reduction to obtain reasonable recoveries. It may be a good candidate for milling, assuming it contains high enough gold grades. Cyanide concentration and pulp oxygen levels are other variables that should be considered during process development studies. Some ores may contain cyanicides, such as copper minerals, or oxygen consumers, such as pyrrhotite. Lower gold extractions may be the outcome of testwork if the cyanide concentration or oxygen levels were not maintained at the levels that the ore required. Therefore, special attention must be paid to the pulp cyanide concentration and oxygen level while performing laboratory tests (Ellis and Senanayake, 2004).
Figure 5 — Example of column leach gold extraction as a function of solution-to-ore ratio for three crush sizes.
Figure 6 — Generic heap leach metallurgical study flow sheet. November 2016 • Vol. 33 No. 4
168
MINERALS & METALLURGICAL PROCESSING
Figure 7 — Generic refractory ore metallurgical study flow sheet.
Gravity-recoverable gold studies. If the ore has a coarse, gravity-recoverable gold component, it is recommended that this component be recovered by conducting gravity concentration tests using instruments such as centrifugal gravity concentrators, shaking tables and/or jigs. The gravity concentrate may be fire-refined or reground and cyanide leached. Gravity tails may be cyanide leached to complete the material balance. A generic heap leach metallurgical study flow sheet is illustrated in Fig. 6, which shows the major steps required to develop an oxide gold deposit using a heap leach process. In some cases, a pilot column leach study using large-diameter columns or demonstration heaps may be required to confirm the results and analysis of smaller-scale tests.
Things that could go wrong during process development studies
Process development metallurgical studies for refractory gold ores
Refractory gold ore may be defined as an ore in which gold is resistant to leaching with cyanide. Gold may be very finely disseminated in the ore matrix; in solid solution with sulfide minerals such as pyrite (FeS2), arsenopyrite (FeAsS) and arsenian pyrite or telluride minerals such as calaverite (AuTe2) and petzite (Ag3AuTe2); in the presence of oxygen consumers and cyanicides such as copper sulfides, marcasite (FeS2 with orthorhombic crystal structure), pyrrhotite [Fe(1 − x)S, where x = 0 to 0.2)] and elemental sulfur; or with naturally occurring carbonaceous preg-robbing material such as elemental carbon or humic acids. Refractory gold ores require a pretreatment step, either physical or chemical, in order to render the gold leachable with cyanide (Fernández, Collins and Marczak, 2007; Simmons, 1995; Thomas et al., 1998; Pyke, Johnston and Brooks, 1999; Brierley et al., 1995; Wan and Brierley, 1997). The choice of pretreatment option depends on the economics of the particular ore type treated. A physical pretreatment option is ultrafine grinding of the ore to expose surfaces of finely disseminated gold particles to cyanide leaching. Chemical pretreatment involves the oxidation of the gold-containing sulfide minerals using hydrometallurgical — such as pressure or atmospheric oxidation or biooxidation — or pyrometallurgical — that is, roasting — processes. During these processes, gold-containing refractory minerals are oxidized, resulting in a porous matrix MINERALS & METALLURGICAL PROCESSING
for the cyanide to saturate and leach gold. In addition to hydrometallurgical and pyrometallurgical pretreatment processes, blinding reagents such as kerosene or motor oil may be used for preg-robbing carbonaceous ores. These pretreatment options are applicable to whole ore or flotation concentrates. The metallurgical program for refractory ores should therefore include a pretreatment options study on the whole ore or gravity and/or flotation concentrates (Fig. 7). Sulfidic refractory ores process development metallurgical study options are summarized in Fig. 8. This paper will not go into the details of these individual processes and how development testwork should be conducted for them, if selected.
169
Several things might go wrong in the course of conducting process development metallurgical studies. These could be associated with the selection, preparation and characterization of variability samples and master composites, drafting of the appropriate test program for the contemplated process or processes, conducting of laboratory tests, and evaluating of data. Some examples of things that could go wrong and some precautions that should be taken are: 1. Variability sample does not represent a mineable zone: The mining method considered is openpit mining with 10-m bench height, and the selected variability sample is 25-30 m. As a result, data granularity is lost. 2. Variability sample does not represent a metallurgical domain: A variability sample should represent a “metallurgical response type.” If the variability sample is a combination of 2-m intercept samples with significantly different cyanide-soluble gold to total gold ratios, this would be the mixing of refractory intercepts with an oxide intercept. 3. Variability sample does not represent a geometallurgical domain: The variability sample is made out of several lithologies and alterations and does not provide useful data for the metallurgical process selected. 4. Master composite does not represent a metallurgical domain: The master composite is prepared by combinVol. 33 No. 4 • November 2016
combination of several variability samples irrespective of where they are located in the deposit. An example of this is the combination of variability samples from different depths, and as a result data granularity is lost. 7. Variability samples and master composites were prepared with incomplete characterization: Cyanide-soluble gold and preg-robbing gold assays along with sulfur and carbon speciation are extremely important when
ing several variability samples irrespective of metallurgical domains. 5. Master composite does not represent a geometallurgical domain: The master composite is a combination of several variability samples irrespective of geometallurgical domains. 6. Master composite does not represent a given minable horizon in the deposit: The master composite is a
Figure 8 — Sulfidic refractory gold ores process considerations. November 2016 • Vol. 33 No. 4
170
MINERALS & METALLURGICAL PROCESSING
differentiating oxide versus refractory gold ores. Attention also needs to be paid to deleterious elements such as Cu, As, Se and Hg. 8. When preparing variability samples and master composites, attention must be paid to the drill hole collar, depth and dip data so that the master composite is prepared with variability samples that represent a mining horizon and domain. 9. Several different tests on the master composites may be required for comminution characterization for comminution circuit process design. During sample preparation, samples may be erroneously crushed finer than the required particle size for a particular test, say, JK Drop Weight test on a −38 mm sample, resulting in noncompliance and requirement for a costly new series of metallurgical core drilling. 10. During sulfur speciation chemical characterization, the temperature at which sulfides are roasted must be agreed upon prior to analysis and documented, because 550 °C is used for the estimation of acid generation potential from pyrite while 650 °C is used for total sulfide estimation. XRD analysis must be complete before selecting an appropriate method as sulfide sulfur can be overstated or understated depending upon the presence of sulfur-bearing minerals. Common interfering minerals are alunite, barite, jarosite, orpiment, realgar, pyrrhotite and carbonates (Dietrich, 2014). 11. Similarly, it is recommended to perform a separate aquaregia digestion for volatile elements such as As and Sb. 12. The incomplete mineralogical characterization of head samples and also test residues or tailings for gold associations and clay characterization will result in lost revenue (Chryssoulis and McMullin, 2005). 13. Laboratory grinding should be conducted in mild steel mills instead of stainless steel mills. The flotation test responses for samples ground in mild steel and stainless steel mills may differ significantly, resulting in a wrongly designed flotation circuit. 14. Laboratory bench-scale investigations should always be extended to include pilot investigations. This is particularly important for pressure oxidation studies. Gathje, Oberg and Simmons (1995) reported that the cyanide leaching responses of oxidized solids from bench-scale batch tests can differ considerably from those achieved by the continuous pilot-plant autoclave.
References
ALS Ltd., 2012, Fire Assay Technical Note 2012. Armstrong, S.E., and Malhotra, D., 1992, “Can diagnostic leach procedures characterize refractory gold ores?” Proceedings of the Randol Gold Conference, Vancouver, British Columbia, Canada, 1992, Randol International, Golden, CO, pp. 155-157. ASTM International, 2016, “ASTM D6722 – 11, StandardTest Method forTotal Mercury in Coal and Coal Combustion Residues by Direct Combustion Analysis.” Brierley, J.A., Wan, R.Y., Hill, D.L., and Logan, T.C., 1995, “Biooxidation-heap biotreatment technology for processing lower grade refractory gold ores,” Proceedings of the International Biohydrometallurgy Symposium, Vina del Mar, Chile, Nov. 19-22, 1995. Chieregati, A.C., and Pitard, F.F., 2009, “The challenge of sampling gold,” Fourth World Conference on Sampling and Blending, SAIMM Symposium Series S59, pp 107-112, Oct. 21-23, 2009, Cape Town, South Africa. Chryssoulis, S.I. and McMullin, J., 2005, “Mineralogical investigation of gold ores,” Developments in Mineral Processing, Vol. 15 Advances in Gold Ore Processing, by Mike D. Adams, Elsevier, pp. 21-71, 2005, http://dx.doi. org/10.1016/s0167-4528(05)15002-9. Dietrich, M., 2014, senior consulting chemist, Denver, CO, personal communication. Dominy, S.C., 2009, “Grab sampling for underground gold mine grade control,” Fourth World Conference on Sampling and Blending, SAIMM Symposium
MINERALS & METALLURGICAL PROCESSING
171
Series S59, pp. 97-107, Oct. 21-23, 2009, Cape Town, South Africa. Ellis, S., and Senanayake, G., 2004, “The effects of dissolved oxygen and cyanide dosage on gold extraction from a pyrrhotite-rich ore,” Hydrometallurgy, Vol. 72, Issues 1-2, February 2004, pp. 39-50, http://dx.doi.org/10.1016/s0304386x(03)00131-2. Fernández, R., Collins, A., and Marczak, E., 2007, “Gold Recovery from HighArsenic Containing Ores at Newmont Roasters,” SME Annual Conference & Expo, Feb. 25-28, 2007, Denver, CO, Society for Mining, Metallurgy & Exploration, Englewood, CO, Preprint 07-037. Garcia, A.J., and Jorgensen, M.K., 1997, “Agglomeration and heap leach testing requirements for high clay ores,” Randol Gold Forum, 1997, Randol International, Golden, CO, pp. 143-146. Gathje, J., Oberg, K.., and Simmons, G., 1995, “Pressure-oxidation process development: beware of lab results,” Mining Engineering, Vol. 47, No. 6, pp. 520-523. Guresin, N., Lorenzen, L., Dominy, S.C., Muller, H., and Cooper, A., 2012, “Importance of effective sampling and test protocols for process plant design,” Sampling Conference, Perth, Western Australia, Aug., 21-22, 2012, pp. 95-107. Hanks, J., and Barratt, D., 2002, “Sampling a mineral deposit for metallurgical testing and the design of comminution and mineral separation,” Mineral Processing Plant Design, Practice and Control, Vancouver, British Columbia, Canada, Society for Mining, Metallurgy & Exploration, Oct. 20-24, 2002, pp. 99-116. Hoal, K.O., Woodhead, J., and Smith, K.S., 2013, “The importance of mineralogical input into geometallurgy programs,” The Second AUSIMM International Geometallurgy Conference, Brisbane, Queensland, Sept. 30-Oct. 2, 2013, pp. 17-35. Hoffman, E.L., Clark, J.R., and Yeager, J.R., 1999, “Gold analysis – fire assaying and alternative methods,” Canadian Institute of Mining, Metallurgy and Petroleum, Explor. Mining Geol., Vol. 7, Nos. 1 and 2, pp. 155-160. Laplante, A.R., Woodcock, F., and Huang, L., 2000, “Laboratory procedure to characterize gravity-recoverable gold,” Transactions of the Society for Mining, Metallurgy & Exploration, Vol. 308, 2000, pp. 53-59. LECO Corp., 2016, “844 Series: Carbon and Sulfur in Inorganic Materials by the Combustion Infrared Detection Technique,” http://www.leco.com/products/ analytical-sciences/carbon-sulfur-analyzers/844-series. Lorenzen, L., 1995, “Some guidelines to the design of a diagnostic leaching experiment,” Minerals Engineering, Vol. 8, No. 3, pp. 247-256, http://dx.doi. org/10.1016/0892-6875(94)00122-s. Lunt, D., Ritchie, I., and Fleay, J., 1997, “Metallurgical process development and plant design,” MINDEV 97: The International Conference on Mine Project Development Conference, Nov. 24-26, Sydney, Australia, The Australasian Institute of Mining and Metallurgy (AUSIMM) Publication Series, pp. 111-124. Marsden, J.O., and House, C.I., 2006, The Chemistry of Gold Extraction, Second Edition, Society for Mining, Metallurgy & Exploration Inc., Englewood, CO. McGuire, M.A., 1989, “Trial by Fire, A Fire Assay Short Course,” Newmont Metallurgical Services, June 2, 1989. Misra, K.C., 2000, Understanding Mineral Deposits, Kluwer Academic Publishers. Morrell, S., 2008, “A method for predicting the specific energy requirement of comminution circuits and assessing their energy utilization efficiency,” Minerals Engineering, 2008, Vol. 21, No. 3, http://dx.doi.org/10.1016/j. mineng.2007.10.001. Pyke, B.L., , P., 1999, “The characterization and behaviour of carbonaceous material in a refractory gold bearing ore,” Minerals Engineering, Vol. 12, No. 8, pp. 1999, http://dx.doi.org/10.1016/s0892-6875(99)00073-4. Scheffel, R.E., 2014a, “Heap leach development - Achieving the correct conceptual design - Part I,” Proceedings of Heap Leach Solutions, Lima, Peru, InfoMIne, 2014. Scheffel, R.E., 2014b, “Heap leach development - Achieving the correct conceptual design - Part II,” Proceedings of Heap Leach Solutions, Lima, Peru, InfoMIne, 2014. Simmons, G.L. 1995, “Pressure Oxidation Process Development for Treating Refractory Carbonaceous Ores at Twin Creeks,” SME Annual Conference & Expo, March 6-9, 1995, Denver, CO, Society for Mining, Metallurgy & Exploration, Englewood, CO, Preprint 95-65. SMC Testing, 2016, “Using the SMC Test to Predict Comminution Circuit Performance,” http://www.smctesting.com/documents/Using_the_SMC_Test.pdf. Thomas, K.G., Fleming, C., Marchbank, A.R., and Dreisinger, D., 1998, “Gold Recovery from Refractory Carbonaceous Ores by Pressure Oxidation,Thiosulfate Leaching and Resin-in-Pulp Adsorption,” U.S. Patent 5,785,736, July 28, 1998. Torres, V.M., Chaves A.P., and Meech J.A., 1999, “Process design for gold ores: A diagnostic approach,” Minerals Engineering, Vol. 12, No. 3, pp. 245-254, http://dx.doi.org/10.1016/s0892-6875(99)00002-3. Tumilty, J.A., and Schmidt, C.G., 1986, “Deportment of gold in the Witwatersrand system,” Gold 100 Proceedings of the International Conference on Gold, Vol. 2: Extractive Metallurgy of Gold, Johannesburg, South Africa, SAIMM, 1986, pp. 541-552. Wan, R.Y., and Brierley, J.A., 1997, “Thiosulfate leaching following biooxidation pretreatment for gold recovery from refractory carbonaceous-sulfidic ore,” Mining Engineering, Aug. 1997, pp 76-80. Whincup, P.R., 2010, “Guidelines for mineral process plant development studies,” Mineral Processing and Extractive Metallurgy (Trans. Inst. Min Metall. C), Vol. 119, No. 4, pp. 191-198. Williams, S.R., and Richardson, J.M., 2004, “Geometallurgical mapping: A new approach that reduces technical risk,” Proceedings of the 36th Annual Meeting of the Canadian Mineral Processors, Ottawa, Ontario, Canada, CMP, Jan. 20-22, 2004, pp. 241-268.
Vol. 33 No. 4 • November 2016
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
