World’s reserves of PGM .
About 80% of the worlds’ reserves for platinum group metals (PGMs) are in South Africa’s Bushveld Igneous Complex. Processing of PGM involves comminution, flotation, smelting, converting, base metals refinery and precious metals refinery. Due to increasing chrome content in the feed and the challenges associated with operating high chrome feed, alternative routes to smelting of PGM are being investigated. Some hydrometallurgical routes have been proposed. However, none of the reported potential routes have yet been commercialised.
- platinum group metals
Platinum group metals (PGMs) are a group of six elements, namely iridium (Ir), osmium (Os), platinum (Pt), palladium (Pd), rhodium (Rh) and ruthenium (Ru) [1, 2]. PGMs together with gold and silver are classified as noble or precious metals because of their high corrosion and oxidation resistance [1, 2, 3]. The world’s reserves of platinum group metals/elements are estimated at 100 million kilogrammes, of which over 80% is contained in South Africa’s Bushveld Igneous Complex . Summary of the PGM reserves by country is shown in Table 1 . PGMs are used in a number of industrial processes and commercial applications including automotive, jewellery, electronics, dentistry, catalysts, needles, pivots, temperature measurements, special crucibles and investments amongst others [5, 6, 7].
|Country||PGM reserves (kg)|
|United States||900 000|
|Russia||1 100 000|
|South Africa||63 000 000|
|Other countries||800 000|
|World total (rounded)||66 000 000|
In South Africa, the PGM ore is mined in the western and eastern limb of the Bushveld Igneous Complex . Within the Bushveld Igneous Complex, the Merensky reef, the Platreef and Upper Group 2 (UG2) reef are exploited for platinum production [1, 4, 9]. Merensky and Platreef have similar chemical and mineral composition [6, 10]. These reefs typically have fairly low contents of sulphides. UG2 differs from the two reefs in that it has low nickel and copper content ; compared to UG2, Merensky reef has higher amounts of chalcopyrite, pentlandite and pyrrhotite . Chromite spinel can be up to 4% by mass in a UG2 concentrate, while a Merensky concentrate can have <1% by mass chromite spinel [8, 10].
A typical mineralogy of the flotation concentrate blend is shown in Table 2 . Eksteen (2011)  has given the fractions of typical minerals in the UG2 and Merensky concentrates. The abundant gangue minerals in a typical PGM concentrate are orthopyroxene, talc, clinopyroxene and plagioclase ; other gangue minerals are <5% by mass .
|Pentlandite||(Fe, Ni, Co)9S8|
|Pyrrhotite||Fe1 − xSx|
|Alteration silicates||Hydrated minerals|
|Pyroxenes||(Mg, Fe)SiO3 – Ca(Mg, Fe)Si2O6|
|Plagioclase/feldspar||NaAlSi3O8 – CaAl2Si2O8|
3. Sources of PGM
The principal sources of PGM are sulphide and arsenide minerals such as PtAs2, PtS, Pt(AsS)2, (Pt,Pd)S, (Pt,Pd,Ni)S, RuS2 and Pd3Sb and elemental ruthenium [3, 7]. Noble metals together with cobalt (Co), copper (Cu), iron (Fe) and nickel (Ni) belong to the class of transition metals in the periodic table . Geologically, PGM associates with base metal sulphides such as chalcopyrite (CuFeS2), millerite (NiS), pentlandite (Fe,Ni)9S8, pyrite (FeS2) and pyrrhotite (Fe1−xS) [3, 12, 13]. Troilite carries trace amounts of iridium, while chalcopyrite has trace amounts of Ru, Pd, Ir and Pt . Gangue minerals associated with PGM-containing minerals are feldspar, biotite, plagioclase and pyroxene .
4. PGM ore processing: an overview
The PGM ore is initially treated in primary and secondary crushers after which it is sent to rod and ball milling circuits. The milled PGM ore is treated using gravity separators and flotation cells; xanthate and dithiophosphate collectors are typical reagents used for flotation at a pH of 7.5–9 . A sulphide-rich PGM concentrate is produced in the flotation cells .
Typical composition of UG2 and Merensky flotation concentrate is shown in Table 3 . A typical PGM content in the flotation concentrate is in Table 4 . The UG2 concentrate has higher chrome content when compared to the Merensky as seen in Tables 3 and 4. UG2 also has higher PGM content than the Merensky concentrate. The concentrate from the flotation cells is dried and smelted to separate the sulphides from the silicates .
|Merensky||1.6||2.2||0.3||2.1||22.3||18.2||3.2||9||41.4||100 - 250|
|UG2||5.0||2.4||2.9||0.8||12.6||21.0||1.7||3.6||44.0||300 - 600|
|Pt-% of 4E||63.5||56.7||45.1|
4.2. PGM ore smelting
To separate the PGM-rich sulphides from the gangue minerals, smelting is used. Rectangular six-in-line or circular three-electrode electric furnaces are typical in the PGM industry . Smelting is a high-temperature process step where the sulphides (valuable minerals) are separated from the silicates (gangue minerals). Energy required for melting the concentrate is provided by Joule heating when an electric current is passed through the resistive bath . The electrodes are used for electrical connections between the power supply and the bath. Graphite electrodes are inserted into the resistive bath such that when a current is applied through the electrodes, thermal energy is generated [6, 15].
The PGM concentrate is introduced into the smelter through the feed ports situated on the furnace roof. The concentrate forms a thick bed (~400 mm) on top of the molten bath. The heat generated in the resistive bath causes the concentrate bed to melt gradually .
The operating temperature of the smelter at the concentrate zone can range from 600 to 900°C . The liquidus temperature of base metal sulphides associated with PGMs is 850–875°C, whereas the liquidus temperature of the corresponding silicates is approximately 1350°C . The sulphides in a PGM concentrate start melting at the concentrate bed since temperature at the concentrate zone can exceed the liquidus temperature of the sulphides. The silicate minerals melt when they reach the molten bath (concentrate-slag interface) . The silicates and sulphides are immiscible; upon melting they form two separate layers.
The molten silicates and oxides form a fayalitic-forsteric slag layer, while the PGM-containing sulphides form a matte layer . The specific gravity of matte ranges from 4.8 to 5.6, and that of slag ranges from 2.8 to 3.8 [6, 15]. Owing to the difference in specific gravities of matte and slag, the slag forms a top layer. Matte being denser than slag falls through the slag layer and settles underneath the slag. Matte consists of base metal sulphides (cobalt, copper, iron and nickel). Matte serves as a collector for the PGMs [6, 7, 14].
Operating temperature of matte and slag varies with the composition of the concentrate . Typical smelter operating temperature for matte varies from 1380 to 1600°C, and that of slag varies from 1500 to 1680°C .
The molten slag and matte are tapped out of the smelter through the tap holes situated at the sidewall of the furnace. After tapping, the matte either can be fed directly to the converters as tapped or can be granulated before the conversion step; matte treatment varies with the producers.
4.2.1. Industrial PGM-furnace matte
Anglo American Platinum (Amplats), Impala Platinum (Implats) and Lonmin are three major producers of PGM in South Africa . Typical compositions of matte from selected smelters are shown in Table 5. Matte from different smelters seems to have comparable amounts of major components (Cu, Fe, Ni and S). The slight difference is that the amount of iron in Amplats matte is slightly higher than the iron in a Lonmin matte (Table 5).
4.2.2. Industrial PGM-furnace slag
Gangue minerals associated with PGM-containing minerals are feldspar, biotite, plagioclase and pyroxene ([Ca,Na]Al1–2Si3–2O8) . During smelting these gangue minerals form a slag that is a silicate rich phase. Typical composition of a PGM-furnace slag from different smelters is shown in Table 6 . The oxide compounds in Amplats and Lonmin-Merensky slag are comparable; Lonmin-UG2 slag has higher Cr2O3, CaO and MgO; and its FeO content is significantly lower when compared to that of the slag from Amplats and Lonmin-Merensky ore.
After smelting, the furnace matte is treated in Pierce-Smith converters or Ausmelt process where iron sulphide is oxidised to ferrous oxide and sulphur is oxidised to sulphur dioxide . Sulphur dioxide is removed as an off-gas, and iron oxide is removed as a fayalitic slag. The slag phase of the converter contains significant amounts of entrained PGM and is recycled to the smelting furnace to recover the entrained PGM. The converter matte is cooled, milled and treated in base metals refinery (BMR) [1, 9].
Typical hydrometallurgical processes used for PGM purification are as follows: dissolution-precipitation (pressure oxidation leach), solvent extraction and ion exchange and molecular recognition technology . Pressure oxidation leach is a typical hydrometallurgical process used to separate base metals from the PGM residue. The base metals are leached, while the PGMs remain in the residue. The PGM residue is sent to a precious metal refinery where various solution extraction and precipitation methods are used to separate individual metals . Solvent extraction is another method by which PGMs can be separated from the base metals .
5. Operational challenges facing PGM industry
South Africa is faced with increasing cost of electricity which has direct impact on the energy-intensive processes such as smelting [1, 18]. South African PGM producers generally process a Merensky concentrate where available; otherwise, the Merensky and UG2 concentrates are blended together to achieve the required feed composition for the smelter. In recent years, the depletion of the Merensky reef has forced the PGM producers to mine the UG2 reef that has up to 60% chromite. As a consequence, the concentrate from the floatation cells has high chrome content than the amount that the smelters are designed to handle. In reducing conditions, the operating temperature limits the solubility of Cr2O3 in a typical slag. At 1450–1650°C the solubility of Cr2O3 in slag is limited to 1.8% by mass. As such the smelters have battled with numerous challenges owing to the treatment of a high chrome concentrate [9, 19].
5.1. Challenges associated with smelting high chromite concentrate
The challenges associated with high chrome feed to the smelter are the following :
Formation of chromite spinels which are insoluble in slag.
Chromite spinels have high melting points as such they increase the liquidus temperature of the slag and they lower the fluidity of the slag .
Chromite spinels increase the overall temperature of the constituents (slag and matte) .
When the matte temperature is above the liquidus temperature of the slag freeze-lining, the matte dissolves the freeze-lining; this leads to corrosion of the refractories by corrosive PGM melt .
There are FeO- and CrO-based spinels; the proportion of Fe and Cr in the spinel may vary, as such there are spinels that are heavier than the matte, while other spinels are slightly lighter than matte. Spinels which are denser than matte settle on the hearth; the accumulation of spinels on the hearth reduces the volume of the furnace .
Spinels with intermediate density form a ‘mushy’ layer at the slag-matte interface; this leads to the entrainment of matte in slag .
High chrome decreases electrical conductivity of the bath leading to electrical control problems in the furnace .
Higher chrome levels affect the matte temperatures during conversion step (matte temperatures above 1355°C have been observed); this causes damage to the refractory lining. Cold dope (revert) is normally used to lower the temperature of the matte during conversion [16, 19].
A number of approaches have been investigated to deal with the chrome problem. The following actions have eased chromite problem in the smelters:
Deep electrode immersion operating at high power densities causes sufficient mixing which keeps the solids in suspension [8, 21], but high power densities adversely affect the refractory life. To minimise the effect of high power input on the refractory life of the sidewall lining, the phase voltage is increased without increasing the current levels .
The control of furnace inputs and control of furnace parameters (power, furnace availability) are essential in controlling chrome content .
Selective reduction improves the solubility of chromium in the slag .
Tapping out the intermediate layer intentionally .
Decreasing the chromium input to the smelters .
Another innovation able to manage chrome-rich ores is the ConRoast process. This process involves removing and capturing sulphur from the concentrate prior to smelting in a DC arc furnace. Roasting a concentrate makes smelting more environmentally friendly; it also enables furnaces to accept any proportion of chromite, resulting in more efficient and cost-effective platinum production [9, 23].
5.2. Premature failure of refractory lining
Due to high operating temperatures (1350 to >1600°C)  associated with PGM smelting, the smelter has to be lined with refractories at the hot face. To prolong the service life of the refractories, sufficient cooling of the refractories is required at the cold face of the refractory wall. Copper waffle coolers are typically used in the cold face of the refractories to extract heat away from the refractories . Due to high operating temperature and corrosiveness of the PGM melt, premature failure of the waffle copper coolers has been experienced in PGM smelters  in the upper sidewall region. Failure of waffle copper coolers causes explosions, loss of production and costs associated with furnace rebuild. The failure of waffle copper coolers was preceded by the consumption of conventional refractory bricks (MgOx-CrOx) which were used to form the furnace lining. The refractory brick (MgOx-CrOx) had low resistance to chemical attack by liquid PGM melt. To prevent the occurrences of copper cooler failures, conventional bricks have been replaced by the graphite blocks in recent designs of PGM smelter refractory walls . Graphite blocks are only applied at the hot face of the upper sidewall (against the concentrate and the slag zone).
With the graphite block lining, a frozen protective skull forms at the hot face of the refractory. The formation of the protective skull is attributed to the efficient cooling of the refractory wall by waffle coolers. The frozen skull is the melt that solidified due to the surface temperature of the graphite that was lower than the liquidus temperature of the melt .
Graphite blocks have increased the service life of the waffle copper coolers through the formation of a protective layer. However, high infiltration of melt is still observed at the matte-slag tidal zone. This is a challenge that still needs to be addressed; currently, conventional bricks or monolithics are used at lower sidewall of the PGM smelter refractory. It is desired to extend the graphite blocks to the lower sidewall of the PGM smelter refractory wall against the matte zone . It is envisaged that using carbon-based refractory at the hot face of the matte zone (lower sidewall) will improve the service life of the furnace lining in PGM smelters .
6. Opportunities in PGM processing
The PGM industry faces challenges with increasing chrome content in the feed and premature failure of refractory lining in the smelter. Alternative ways to process the PGM have become attractive due to the challenges associated with the conventional smelting process [1, 18]. A hydrometallurgy (Kell) process has been probed as an alternative to smelting PGM. The Kell process has three stages: stage 1 is the leaching of base metal sulphides in an acidic sulphate medium (pressure oxidation); stage 2 is roasting of the residue from stage 1; and stage 3 is atmospheric leaching of PGMs in a chloride media. The leached precious metals are further treated in refineries to recover metals . Other hydrometallurgical routes have been discussed in Ref. . These hydrometallurgical processes have advantages over the smelting process since they reduce the operating costs drastically . However, the alternative hydrometallurgical routes of processing PGMs have not yet been commercialised .
In this chapter, an overview of PGM processing has been presented. The conventional smelting process has challenges with high chromium feed, premature failure of refractory lining and increased operating cost associated with increasing cost of electricity in South Africa. The Kell process is an alternative way to process a PGM concentrate, and it has a number of advantages such as less energy consumption, less energy cost, less electricity consumption, less CO2 emissions and no restriction on chrome content of feed. Other hydrometallurgical routes have been investigated, but none has been commercialised yet.
Safarzadeh MS, Horton M, Van Rythoven AD. Review of recovery of platinum group metals from copper leach residues and other resources. Mineral Processing and Extractive Metallurgy Review. 2018; 39(1):1-17
Glaister BJ, Mudd GM. The environmental costs of platinum–PGM mining and sustainability: Is the glass half-full or half-empty? Minerals Engineering. 2010; 23(5):438-450
Xiao Z, Laplante AR. Characterizing and recovering the platinum group minerals – A review. Minerals Engineering. 2004; 17(9):961-979
Junge M, Wirth R, Oberthür T, Melcher F, Schreiber A. Mineralogical siting of platinum-group elements in Pentlandite from the Bushveld complex, South Africa. Mineralium Deposita. 2015; 50(1):41-54
Jewell S, Kimball SM. Mineral Commodity Summaries 2015. U.S. Geological Survey. Reston, Virginia: CreateSpace Independent Publishing Platform; 2015; 196:120-121
Jones RT. An overview of Southern African PGM smelting, nickel and cobalt 2005: Challenges in extraction and production. In: 44th Annual Conference of Metallurgists, Calgary, Alberta, Canada: Canadian Institute of Mining, Metallurgy and Petroleum. 21-24 August 2005. pp. 147-178
Hayes PC. Process Principles in Minerals & Materials Production. 3rd ed. Australia: Hayes Publishing CO; 2003. pp. 278-279
Eksteen JJ. A mechanistic model to predict matte temperatures during the smelting of UG2-rich blends of platinum group metal concentrates. Minerals Engineering. 2011; 24(7):676-687
Cramer LA. What is your PGM concentrate worth? In: Third International Platinum Conference, ‘Platinum in Transformation’, Johannesburg, South Africa: South African Institute of Mining and Metallurgy; 2008. pp. 5-9
Nell J. Melting of platinum group metal concentrates in South Africa. Journal of the South African Institute of Mining and Metallurgy. 2004; 104(7):423-428
Andrew NJ, van Beek B, Lexmond A, Zietsman JH. Effect of Feed Composition Fluctuations on a Platinum Furnace Energy Balance and Slag Temperature. The Southern African Institute of Minng and Metallurgy, Pyrometallurgical Modelling-Principles and Practices, Kempton Park, South Africa, 4-5 August 2014. pp. 117-126
Jones RT. Platinum smelting in South Africa. South African Journal of Science. 1999; 95:525-534
Cabri LJ, Rudashevsky NS, Rudashevsky VN. Current approaches for the process mineralogy of platinum-group element ores and tailings. In: Ninth International Congress for Applied Mineralogy ICAM. Geological Society of India. 2009. pp. 9-17
Liddell KS, McRae LB, Dunne RC. Process routes for beneficiation of noble metals from Merensky and UG-2 ores. Mintek Review. 1986; 4:33-44
Habashi F. Textbook of Pyrometallurgy. Canada: Métallurgie Extractive Québec; 2002. pp. 237-242
Eksteen JJ, Van Beek B, Bezuidenhout GA. Cracking a hard nut: An overview of Lonmin’s operations directed at smelting of UG2-rich concentrate blends. Journal of the Southern African Institute of Mining and Metallurgy. 2011; 111(10):681-690
Bezuidenhout GA, Eksteen JJ, Akdogan G, Bradshaw SM, De Villiers JPR. Pyrometallurgical upgrading of PGM-rich leach residues from the western platinum base metals refinery through roasting. Minerals Engineering. 2013; 53:228-240
Liddell K, Newton T, Adams M, Muller B. Energy consumptions for Kell hydrometallurgical refining versus conventional pyrometallurgical smelting and refining of PGM concentrates. In: The 4th International Platinum Conference, Platinum in Transition ‘Boom or Bust’. Johannesburg, South Africa: The Southern African Institute of Mining and Metallurgy; 2010. pp. 181-186
Coetzee V. Common-sense improvements to electric smelting at impala platinum. Journal of the Southern African Institute of Mining and Metallurgy. 2006; 106(3):155-164
Hundermark RJ, Mncwango SB, de Villiers LPS, Nelson LR. The Smelting Operations of Anglo American’s Platinum Business: An Update, Southern African Pyrometallurgy 2011. Johannesburg: Southern African Institute of Mining and Metallurgy; 6-9 March 2011. pp. 295-307
Liddell KS, Adams MD. Kell hydrometallurgical process for extraction of platinum group metals and base metals from flotation concentrates. Journal of the Southern African Institute of Mining and Metallurgy. 2012; 112(1):31-36
Barnes AR, Newall AF. Spinel Removal from PGM Smelting Furnaces. Southern African Pyrometallurgy, 5-6 March 2006. Johannesburg, South Africa: Southern African Institute of Mining & Metallurgy; 2006. pp. 77-88
Jones RT, Geldenhuys IJ. The pros and cons of reductive matte smelting for PGMs. Minerals Engineering. 2011; 24(6):495-498
McDougall I, Eksteen JJ. Sidewall design to improve lining life in a platinum smelting furnace. In: International Smelting Technology Symposium (Incorporating the 6th Advances in Sulphide Smelting Symposium). Warrendale, PA: TMS (The Minerals, Metals & Materials Society); 2012. pp. 47-54
Thethwayo BM. Sulphidation of copper coolers in PGM smelters. [MSc thesis]. South Africa: University of Pretoria; 2010