Once the gold ore has been removed from underground and brought to surface, an entirely new challenge begins in separating the gold from the waste material. The separation process is particularly difficult due to the usually low quantity of gold relative to waste. Common gold grades are in the same magnitude as about 5 g/t, or 0.000005% gold content by weight. This means that upgrading the concentration by a factor of 3000 to 4000 is commonly required to achieve commercial ore (Marsden & House, 2009). In order to make extraction possible, significant comminution is required as well as the utilization of chemicals such as cyanide for leaching in the mineral processing circuit.
The efficiency and recovery of chemical processes involved in gold extraction is exclusive to the mineralogy of the ore. Gold may exist as free nuggets or particles in alluvial or elluvial deposits or as native metal in other minerals. The gold mineralogy of any given ore deposit is unique to all others due to variations in the following:
- Mineral mode of occurrence of gold
- Gold grain size distribution
- Host and gangue mineral type
- Host and gangue mineral grain size distribution
- Mineral associations
- Mineral alterations
- Variations of the above within a deposit or with time
It is therefore critically important, that when evaluating gold extraction techniques, the ore deposit and the mineralogical factors are well known prior to making any decision (Marsden & House, 2009). Although there is a large set of mineral processing technologies to select from, the list is distinct (Stewart, 2012). The ore classification techniques however, are not distinct; there is no universal characterization technique that is applicable to all gold-bearing rocks. La Brooy et al. (1994) provided a useful framework that is still used by many today outlined in Figure 1.
The selection of a flowsheet for free milling ores is relatively simple compared to others. The major factors include comminution circuit selection, heap leaching use, treatment of ores with high silver content, and flotation options for sulfides. Gold bearing ores that include base-metal mineralization, particularly copper, are classified as complex ores. These ores will often consume cyanide and create issues in CIP and elution. Preg-robbing carbon is defined as the phenomenon whereby the gold cyanide complex is removed from solution by constituents of the ore (Rees & van Deventer, 2000). The presence of preg-robbing carbon ultimately requires that certain flowsheet inclusions are made such that recovery does not drop to an unacceptable level and gold loss to CIL tailings is minimized. Refractory characteristics are particularly worth consideration if gold-bearing iron sulfides, such as pyrite, arsenopyrite, pyrrhotite, telluride and the stibnite family are present (Lunt & Weeks, 2005). These ore types require a pre-treatment before undergoing a typical processing circuit
Comminution is the process where ore particles are liberated from gangue material through progressive size reduction in the form of crushing, grinding, cutting or vibrating. The costs associated with power supply, grinding media, and liners used in comminution circuits almost always represent the single largest cost in gold extraction processes (Mosher, 2005). Thus, the efficiency of the entire processing operation is largely dependent on the efficiency of the comminution circuit. Sufficient comminution can be accomplished through a single stage or a multiple stage process. Each of the stages can be any combination of particle size-reduction techniques, including the following: jaw crusher, cone crusher, gyratory crusher, hammer crusher, ball mill, rod mill, pebble mill, semi-autogenous grinding (SAG) mill, semi-autogenous ball (SAB) mill, autogenous (AG) mill, or high-pressure-grinding-rod mill. The primary stage is usually a type of crusher that reduces feed size from a magnitude of about 1m down to about 100mm. Any subsequent stages have the function of reducing the 100 mm sized feed down to a magnitude of 1 mm or even 20 μm. Of the numerous major gold mining operations around the world, most utilize a primary crushing circuit which flows into a SAG mill circuit and then a ball mill circuit. The primary reason for the significant comminution circuits is because in gold mining the grades are notably low and thus excessive waste material is present, resulting in high mill feed rates. If the ore rock is highly fractured and/or brittle, it is likely that an AG mill be the most economical option in lieu of a SAG mill. AG milling can produce a finer grind than SAG milling for certain ores. Figure 2 summarizes the different comminution circuits and some relative factors compared to the other methods.
SAG Milling breaks down ore through impact breakage, attrition breakage and abrasion. The primary design variables in SAG mills include the aspect ratio and the rotation speed. The rotation should be at a rate at which the ore material falls a maximum distance from the head to the impact toe (refer to Figure 3). This rotational speed is referred to as the critical speed. Any faster and the material will ride the circumference of the mill and be exposed to less forceful or no impact at all. At lower speeds the material will not fall from a maximized height and the impact force will be reduced. SAG Mills are advantageous because they utilize the ore material as a grinding media, thus the actual grinding media costs are reduced. The balls in the mill serve to increase the amount of grinding done compared to if the mill was fully autogenous.
Ball mills can operate in either wet or dry conditions. They are advantageous because they can achieve a very fine particle product size (refer to Table 1). Generally ball mills are utilized after an initial material crushing or grinding has occurred. Thus, a disadvantage of the ball mill is that the feed size must already be small. Table 2 outlines some typical ball diameters found in ball mills dependent on the staging or purpose of the mill.
From the framework provided by La Brooy et al. (1994), there are three main classification categories for gold ores: free-milling, refractory, and complex. Free-milling gold ores are defined as an ore from which standard cyanidation (which takes about 20-30 hours) can extract more than 90 % of the gold, under the condition that the ore size is 80 % < 75 µm (Joe Zhou, 2004). Complex ores, meanwhile, require significantly higher cyanide or oxygen levels to attain an economic recovery (Brooy, 1994). Finally, refractory ores require additional chemical reagents or pre-treatment processes to attain sufficient gold recovery. This portion of the article will focus on the origins, characteristics, and required processing techniques for free-milling, complex and refractory ores.
At the most basic level, the response that an ore has to conventional cyanidation is the key factor in determining its gold ore classification. From La Brooy et al. (1994), there are three basic factors affecting response to conventional cyanidation:
- When the gold is entrapped in a mineral matrix leaching reagents are unable to access the gold component of the ore.
- The ore can contain reactive minerals that effectively consume leaching reagents, leaving insufficient cyanide or oxygen to leach the gold.
- Components of the ore can adsorb or precipitate the dissolved gold cyanide complex, and thus it will be removed from the leach liquor.
The degree to which each of these factors is relevant in the ore response to cyanidation determines its classification, and thus the processing strategy for gold extraction.
The complete definition of a free-milling ore is ore sized 80 % < 75 µm from which cyanidation can extract 90 % of the gold, and has not sustained high reagent consumption (John. O. Marsden, 2006). The gold is also commonly recovered through gravity concentration. The main classes of free-milling ores are placers, quartz veined gold ores, oxidized ores, and silver-rich ores. Epithermal deposits may also be free-milling in the oxidized portion, although frequently have higher concentrations of sulfide minerals wherein the ore is considered refractory.
Among free-milling gold ores placer gold is a special case, as there is no pre-treatment required for gold extraction. Gold can be extracted from placer ores using physical separation techniques, particularly through gravity treatment. In modern practice, concentrators use centrifugal force in order to liberate finer gold particles.
Major Processing Route
For most free-milling ores, a pretreatment process is necessary when chemical recovery is being employed. The gold is not sufficiently liberated as is the case with placer ores, and therefore crushing and grinding is required to allow the cyanide to access gold components. Permeability is the major indicator of how much communition is required prior to leaching. Although there are various leaching techniques for free-milling ores, the main technique in Australia, North America, and South Africa is agitation cyanide leaching. Gold recovery is then attained through carbon-in-leach (CIL) or carbon-in-pulp (CIP) methods (Brooy, 1994). The figure below from Lima (2007) shows a typical CIL/ CIP process.
Both CIP and CIL are continuous processes, where activated carbon is added to the ore/ water pulp. The first step in gold extraction is leaching the gold which turns gold in solid form to liquid with cyanide in the presence of oxygen. Next, the gold adsorbs to the activated carbon. Finally, the carbon and gold complex is sent to elution where it is washed, and transferred to electrolysis where the gold is fully recovered. The main difference between CIP and CIL is that for CIP leaching and adsorption occur simultaneously, while in CIL dissolution of gold is complete when adsorption begins (Lima, 2007). The employed model is calibrated from experimental data to optimize recovery.
Complex ores require augmented cyanide or oxygen addition in order to be able to attain economic extraction. The main categories for complex ores are cyanide consuming, oxygen consuming, and preg-robbing. The various difficulties and extra considerations for extracting gold from each type of complex ore will be discussed below.
Cyanide consuming complex ores are associated with copper and reactive sulphides. Gold that is present as electrum causes a slower gold leach process in which cyanide consumption is increased. The cyanide consuming side reactions from oxides or sulphides drive excessive reagent consumption and thus higher production costs with a decrease in gold recovery. In order to combat this undesirable scenario, several processes have been introduced that aim to recover a portion of the cyanide. An example of this would be the Sceresini process in which copper cyanide is loaded onto the carbon prior to gold dissolution in order so that copper can be recovered as copper sulphate with cyanide regeneration (Sceresini, 1991). The most economical approach for sulphide ores is to utilize a gold-copper flotation circuit which uses smelting for gold recovery. The main consideration for this route is that the float tail can contain too much gold for being simply discarded.
Reactive sulphides containing gold, such as pyrrhottite, can require high amounts of oxygen as the iron (II) is oxidized to iron (III), and sulphide to sulphate (Brooy, 1994). Oxidant addition is employed in these cases to meet oxygen demands, and can be controlled by an oxygen electrode inserted into the leach tank. The main concerns with this process is that saline water can cause meter calibration issues (Komosa, 1991). The process options for oxygen consuming ores are thus to use alkaline pre-oxidation, or aeration and flotation, followed by a cyanide leach with oxygen addition.
Preg-Robbing Complex Ores
Preg-robbing occurs when a gold cyanide complex is removed from solution by some component of the ore. The preg-robbing component can be carbonaceous matter, such as organic carbon or other impurities which will adsorb the gold. Sulphides and clays can also cause preg-robbing. For carbonaceous preg-robbing components, the ore must be treated with chlorine, bacteria, oil or roasted to stop preg-robbing behavior prior to leaching. Preg-robbing sulphides can become an issue at low cyanide concentrations, although can be mitigated through pre-aeration. Clays will adsorb the gold cyanide complex as well, so heat or acid addition is used to lessen preg-robbing effects,
Refractory ores are classified as to their level of refractoriness, as shown in the table below.
The challenge of attaining economic extraction of gold from refractory ores is one of the main challenges in the mining industry today, as less refractory ores become more depleted. It is thus important to properly select an extraction method for refractory ores, based on mineralogy, precious metal grades, gold to sulphur ratios and hazardous impurities (Mark Aylemore, 2008). Refractory ores cannot attain sufficient extraction through conventional cyanidation techniques, and thus there are various alternative methods for extraction, depending on the refractoriness of the ore. Gold is locked into its host mineral due to any of the following factors from La Brooy et al. (1994).
- Physical locking (sulphides, silicates, oxides, etc…)
- Chemical locking as gold alloys or compounds (ie: electrum, gold tellurides)
- Gold substitution into a sulphide lattice
- Chemical layer formation causing, gold surface passivation
The main pretreatment options for “unlocking” the gold from refractory ores are chemical, biological, pressure oxidation, thermal and physical treatment. The figure below indicates the various pretreatment options before cyanidation can occur for refractory ores.
In order to choose a pretreatment option for a refractory ore, the main concern is attaining a solution which succeeds in preparing the ore for cyanidation, whilst remaining economically feasible.
The ore throughput rate of a mill is largely dependent on the mining production rate. Generally, it closely related to the mining rate of ore only. The goal is to have the mill operating with little to no downtime, without having to stockpile massive quantities of ore for a feed supply. The mill must be large enough to keep up with the mining operation and process the raw material at similar rates. Larger process plants have a lower operating cost per unit weight of material than a smaller plant using similar processes, but they also have a larger initial capital cost. For reference, a 1 Mt/a mill will have a capital cost within $25,000 /t/h and $50,000 /t/h. A 20Mt/a mill will have a capital cost within $10,000 /t/h and $20,000 /t/h (Lane, Fleay, Reynolds, & La Brooy, 2002). Typical operating costs for mills used in gold extraction are summarized in Table 3. The selection of a mill size is ultimately a function of the economics of the project and which mill size will maximize the project NPV. Thus, a case-by-case analysis must be performed with a proper costing model to determine a proper decision regarding mill size. Table 4 below is a sample cost model for a 1 Mt/a and 20 Mt/a plant.
Gold ore can be classified as free-milling, complex, or refractory and each type of ore has different challenges when it comes to processing methods. Cyanidation with CIL/ CIP continues to be the most common method, although there are environmental concerns due to toxicity. As mining companies continue to face pressure to reduce environmental impact, and as free-milling ores become less and less available, there will be increased focus on developing advanced gold processing techniques. Through various pre-treatment options, the goal is to achieve sufficient gold recovery in the most economical way possible. This article serves its purpose in providing an overall guide to gold processing, and the associated processing options for each gold ore classification.
Brooy, S. R. (1994). Review of gold extraction from ores. In Minerals Engineering Vol. 7, No. 10 (pp. 1213-1242). Wembley: Elsevier Science Ltd. Callow, M. I., & Moon, A. G. (2002). Types and Characteristics of Grinding Equipment and Circuit Flowsheets. In A. L. Mular, D. N. Halbe, & D. J. Barratt, Mineral Processing Plant Design, Practice, and Control Proceedings (pp. 698-709). Littleton, Colorado: Society for Mining, Metallurgy, and Exploration (SME).
Joe Zhou, B. J. (2004). Establishing The Process Mineralogy of Gold Ores. SGS Minerals - Technical Bulletin.
John. O. Marsden, C. I. (2006). The Chemistry of Gold Extraction. Littleton: Society for Mining, Metallurgy, and Exploration.
Komosa, D. (1991). Oxygen Requirements and Monitoring for Gold Ore Processing. Gold Forum on Technology and Practice, 165.
Lane, G. S., Fleay, J., Reynolds, K., & La Brooy, S. (2002). Selection of Comminution Circuits for Improved Efficiency. Kalgoorlie: GRD Minproc Limited.
Lima, L. R. (2007). Dynamic Simulation of the carbon-in-pulp and carbon-in-leach Processes. Department of Materials Science and Technology. Salvador: Federal University of Bahia.
Lunt, D., & Weeks, T. (2005). Process flowsheet selection. In M. D. Adams, Advances in Gold Processing (pp. 73-96). Guildford, Australia: Elsevier.
Mark Aylemore, A. J. (2008). Evaluating Process Options For Treating Some Refractory Ores. South Africa: Bateman Engineering Pty Ltd.
Marsden, J. O., & House, C. I. (2009). The Chemistry of Gold Extraction 2nd Ed. Littleton, Colorado: the Society of Mining, Metallurgy, and Exploration, Inc.
Mosher, J. B. (2005). Comminution circuits for gold ore processing. In M. Adams, Advances in Gold Ore Processing (pp. 253-276). Guildford, Australia: Elsevier.
Rees, K. L., & van Deventer, J. J. (2000). Preg-robbing phenomena in the cyanidation of sulphide gold ores. Melbourne, Australia: Elsevier.
Sceresini, B. (1991). Copper/ cyanide treatment of high-copper gold ores. Fifth Aus Extractive Metallurgy Conference (p. 123). Melbourne: IMM.
Stewart, P. (2012). Australian Gold Processing. Mineral Processing and Extractive Metallurgy, pp. 187-189.