Extractive Metallurgy of Copper Part 8 doc

CHAPTER 12 Direct-To-Copper Flash Smelting Previous chapters show that coppermaking from sulfide concentrates entails two major steps: smelting and converting. They also show that smelting and converting are part of the same chemical process, Le.: oxidation of Fe and S from a Cu-Fe-S phase. It has long been the goal of metallurgical and chemical engineers to combine these two steps into one continuous direct-to-copper smelting process. The principal advantages of this combining would be: (a) isolation of SOz emission to a single, continuous gas stream (b) minimization of energy consumption (c) minimization of capital and operating costs. This chapter (i) describes direct-to-copper smelting in 2002 and (ii) examines the degree to which its potential advantages have been realized. The chapter indicates that the principal problems with the process are that: (a) about 25% of the Cu entering a direct-to-copper smelting furnace ends up dissolved in its slag (b) the cost of recovcring this Cu will probably restrict future expansion of direct-to-copper smelting to low-Fe concentrates (e.g. chalcocite (Cu2S) and bornite (Cu5FeS4) concentrates) rather than high-Fe chalcopyrite concentrates. 12.1 The Ideal Direct-to-Copper Process Fig. 12.1 is a sketch of the ideal direct-to-copper process. The principal inputs to the process are: 187 188 Extractive Metallurgy of Copper concentrate, oxygen, air, flux and recycles. The principal outputs are: molten copper, low-Cu slag, high-SO2 offgas. The process is autothennal. With highly oxygen-enriched blast, there is enough excess reaction heat to melt all the Cu-bearing recycle materials from the smelter and adjacent refinery, including scrap anodes. The process is continuous. The remainder of this chapter indicates how close we have come to this ideality. Concentrates Flux and reverts Scrap copper Oxygen and air SO2 - rich offgas Liquid copper ready for refining Slag low enough in Cu for direct discard Fig. 12.1. Ideal single-furnace coppermaking process. Ideally the copper is low in impurities, the slag is discardable without Cu-recovery treatment and the offgas is strong enough in SO1 for sulfuric acid manufacture. 12.2 Industrial Single Furnace Direct-to-Copper Smelting In 2002, single furnace direct-to-copper smelting is done by only one process - Outokumpu flash smelting, Fig. 1.4. It is done in two locations; Glogow, Poland (Czernecki et al., 1998, 1999a,b,c) and Olympic Dam Australia (Hunt et al., 1999a,b). Both furnaces treat chalcocite-bornite concentrates. For several years the Noranda submerged-tuyere process (Fig. 1.5) also produced copper directly (Mills et al., 1976). It now produces high-grade matte, 72-75% Cu. The change was made to increase smelting rate and improve impurity elimination. The products of direct-to-copper flash smelting (Table 12.1) are: Direct-To-Copper Flash Smelting 189 copper offgas 99% Cu, 0.04 to 0.9% S, 0.01% Fe, 0.4% 0, 1280°C 15 to 20 volume% SO2, 1350°C. slag 14 to 24% CU, -1300°C As with conventional matte flash smelting, the temperature of the furnace is controlled by adjusting; (a) the degree of oxygen enrichment of the blast, i.e. the amount of N2 'coolant' entering the furnace (b) the rate at which fossil fuel is burnt in the furnace. The O2 content of industrial direct-to-copper flash furnace blast is 50 to 90 volume% 02, depending on the furnace's solid feed mixture. Considerable fossil fuel is burnt in the reaction shaft and in settler burners, Table 12.1. 12.3 Chemistry Direct-to-copper smelting takes place by the schematic (unbalanced) reaction: Cu2S,CugFeS4 + O2 + Si02 + Cu; + Fe0,Fe3O4,SiO2 + SO2 concentrate in oxygen flux molten slag in offgas enriched blast (12.1). Just enough O2 is supplied to produce metallic copper rather than Cu2S or Cu20. In practice, the flash furnace reaction shaft product is a mixture of overoxidized (oxide) and underoxidized (sulfide) materials. Individual particles may be overoxidized on the outside and underoxidized on the inside. The overoxidized and underoxidized portions react like: 2C~20 + CU~S -+ ~CU; + SO, (1 2.2) 2Fe304 + Cu2S + 2Cui + 6Fe0 + SO2 to produce molten copper, molten slag and SO2. Industrially, the overall extent of Reaction 12.1 is controlled by: (12.3). (a) monitoring the Cu content of the product slag and the S content of the product copper (b) adjusting the: 190 Extractive Metallurgy ofCopper 0, -in - blast inwt rate concentrate input rate ratio based on these measured Cu-in-slag and S-in-copper values. An increasing % Cu-in-slag is reversed by decreasing the Oz/concentrate ratio and vice versa. The % Cu-in-slag is kept between 14 and 24%. 12.4 Industrial Details Operating details of the two direct-to-copper flash furnaces are given in Table 12.1. The furnaces are similar to conventional flash furnaces. Differences are: (a) the hearths are deeply 'bowl' shaped to prevent molten copper from contacting the furnace sidewalls (b) the hearths are more radically arched and compressed to prevent their refractory from being floated by the dense (7.8 tonnes/m3) molten copper layer (Hunt, 1999) (c) the furnace walls are extensively water cooled and the hearth extensively air cooled to prevent metallic copper from seeping too far into the refractories (d) the refractories are monolithic to prevent molten copper from seeping under the bricks, solidifying and lifting them. Also, the copper tapholes are designed to prevent the out-flowing molten copper from enlarging the taphole to the point where molten copper contacts cooling water. Olympic Dam's molten copper passes through magnesite-chrome brick (inside), a silicon carbide insert and a graphite insert (outside) (Hunt et al., 1999b). The graphite insert is replaced after -1200 tonnes of tapped copper and the silicon carbide insert is replaced after -2400 tonnes. Excessive copper flow (i.e. an excessive taphole diameter) initiates earlier replacement. 12.5 Control The compositions of the industrial furnace products are controlled by adjusting the ratios: 0, -in -blast input rate concentrate input rate and Direct-To-Copper Flash Smelting 191 Table 12.1. Details of Olympic Dam and Glogow direct-to-copper Outokumpu flash furnaces. Note the high product temperatures as compared to matte smelting, Table 5.1. Smelter WMC Resources Olympic KGHM Polska Miedz Dam, Australia Glogow Poland Startup date 1999 Size, inside brick, rn hearth: w x 1 x h reaction shaft diameter height, above settler roof diameter height above settler roof gas uptake slag layer thickness, m copper layer thickness, m active slag tapholes active copper tapholes concentrate burners settler burners Feed details, tonnedday new concentrate (dry) oxygen silica flux recycle flash furnace dust other Blast details blast temperature, "C volume% O2 flowrate, thousand Nm3/hour Production details copper production, tonnesiday composition temperature, "C slag production, tonnesiday mass% Cu mass% Si02/mass%Fe temperature, "C Cu-from-slag recovery method offgas, thousand Nm'/hour volume% SO2, leaving furnace temperature, "C dust production, tonnedday burnt in reaction shaft Hydrocarbon fuel inputs, kg/hour 6.3 x 19.2 x 1.9 4.8 5.8 3.7 7.5 0-0.65 0.5-0.85 2 8 1 2 1200-1600: 41-56% CU 90-450 12-120 (95% Si02) 0-144 ambient SO-90 22 390-680 99% Cu, 0.7 to 0.85% S, 0.4% 0 1280 24 0.5 1320 electric furnace 25 19 1320-1400 boiler 65, ESP 55 oil, 0-200 620-883 1978 9.2 x 26.4 x 3.0 7.4 8.3 6.7 12.3 0.5 0.7 6 10 4 normally none 2000 (28% Cu) self- fluxing 270 IO0 desulfurizing dust 140 75 32 392 0.007% Fe, 0.25% Pb 1280 1050 14 5.7* 1290 electric furnace 35 15 I320 260 0.04% S, 0.45?'00, oil, 300 burnt in settler burners oil, 900-1200 0 '32% SO2; 5.6% Fe; 10% A1203; 13.4% CaO; 6.9% MgO; 13.7% Cu; 3% Pb 192 Extractive Metallurgy of Copper flux input rate concentrate input rate The temperatures of the products are controlled by adjusting the oxygen- enrichment level of the blast (as represented by the N2/02 ratio) and the rate at which fossil fuel is burnt in the furnace. 12.5.1 Target: No Matte Layer to Avoid Foaming The Glogow and Olympic Dam furnaces are operated with 02/concentrate ratios which are high enough to avoid forming a Cu2S layer. This is done to avoid the possibility of foaming slag out the top of the furnace (Smieszek et al., 1985; Asteljoki and Muller, 1987; Day, 1989; Hunt et al., 1999a). A molten Cu2S layer, once built up between the molten copper and molten slag layers, has the potential to react with slag by reactions like: 2C~20 + CU~S + 6Cu; + SO, in slag matte 2cuo + cu2s -+ 4cu; + so, in slag matte (12.2) (12.4) 2Fe304 + Cu2S -+ 2Cui + 6Fe0 + SO2 inslag matte (12.3) all of which can produce SO2 beneath the slag layer. Foaming is particularly favored if the input 02/concentrate ratio is subsequently increased to shrink or remove an existing Cu2S layer. This results in a highly oxidized slag, fill of Fe304, CuO and Cu20, which has great potential for producing SO2 beneath the slag layer. The foaming problem is avoided by ensuring that the 02/concentrate ratio is always at or above its set point, never below. This may lead to high copper oxide-in-slag levels but it avoids the potentially serious operational problems caused by foaming (Hunt et al., 1999a). S-in-copper below -1% S guarantee that a Cu2S layer does not form (Fig. 9.2a)*. *Glogow copper contains 0.04% S, Le. much less than is necessary to prevent matte layer formation. This extra oxidation is done to oxidize Pb (from concentrate) to PbO, keeping Pb-in-copper below 0.3%. Direct-To-Copper Flash Smelting I93 12.5.2 High %Cu-in-slag from no-matte-layer strategy An unfortunate side effect of the above no-matte-layer strategy is high %Cu-in- slag, mainly as dissolved Cu20. It arises because there is no permanent layer of CuzS in the furnace to reduce Cu20 to metallic copper, Reaction (1 2.2). Simply stated, direct-to-copper smelting is operated in a slightly over-oxidizing mode to prevent the foaming described in Section 12.5.1. The downside of operating this way is 14 to 24% Cu in slag, Table 12.1. 12.6 Cu-in-Slag: Comparison with Conventional Matte Smelting/Converting A significant difference between direct-to-copper flash smelting and flash smelting/Peirce-Smith converting is the large amount of Cu in direct-to-copper slag. This extra Cu-in-slag arises because: (a) % Cu in direct-to-copper slags (14-24%, Table 12.1) is much greater than % Cu in conventional smelting slags (1-2% Cu) and converting slags (b) the amounts of slag produced by direct-to-copper smelting and (-6% CU) conventional smelting/converting are about the same. Also, direct-to-copper slags contain most of their Cu in oxidized form (Le. copper oxide dissolved in the molten slag) - so they must be reduced with carbon to recover their Cu. 12.6.1 Electric furnace Cu recovery Both direct-to-copper smelters reduce their flash furnace slag in an electric slag cleaning furnace. The slag flows from the flash furnace directly into an electric furnace where it is settled for about 10 hours under a 0.25 m blanket of metallurgical coke (Czernecki et al., 1999b). This coke reduces the oxidized Cu from the slag by reactions like: cu20 + c -+ 2cu; + co CUO + c -+ cu; + co Magnetite (molten and solid) is also rerluced: Fe304 + C + 3Fe0 + CO (12.5) (12.6). (12.7) and some FeO is inadvertently reduced to Fe by the reaction: 194 Extractive Metallurgy ofcopper FeO + C + Fe + CO (12.8). The Fe joins the newly reduced copper. Glogow results The Cu content of the Glogow direct-to-copper slag is lowered from -14% Cu to -0.6% Cu in an 18 000 kVA electric furnace. The metallic product is (Czernecki et u1, 1999b): 70-80% CU -5% Fe 15-25% Pb (from Pb in the concentrate). This product is too impure to be sent directly to anode-making. It is oxidized in a Hoboken converter (Section 9.6.1) to remove its Fe and Pb, then sent to anode- making. Olympic Dam results Olympic Dam lowers its direct-to-copper slag from 24% to -4% in its 15 000 kVA electric furnace (Hunt et al., 1999a). It could lower it more by using more coke and a longer residence time, but the copper product would contain excessive radioactive '"Pb and '"Po, from the original concentrate. Instead, the Cu-in-slag is lowered further by solidificationicommunitiodflotation in its mine flotation circuit, Section 11.5. 12.7 Cu-in-Slag Limitation of Direct-to-Copper Smelting The principal advantage of direct-to-copper smelting is isolation of SO2 evolution to one furnace. The principal disadvantage of the process is its large amount of Cu-in-slag. Balancing these factors, it appears that direct-to-copper smelting is best suited to Cu2S, Cu5FeS4 concentrates. These concentrates produce little slag so that Cu recovery from slag is not too costly. Direct-to-copper smelting will probably not, however, be suitable for most chalcopyrite concentrates, -30% Cu. These concentrates produce about 2 tonnes of slag per tonne of Cu so that the energy and cost of recovering Cu from their slag is considerable. Only about 60% of new Cu in concentrate would report directly to copper, 40% being recovered from slag. Direct-To-Copper Flash Smelting 195 Davenport et ai. (2001) confirm this view but Hanniala et al. (1999) suggest that direct-to-copper smelting may be economic even for chalcopyrite concentrates. 12.8 Direct-to-Copper Impurities The compositions of the anode copper produced by the direct-to-copper smelters are given in Table 12.2. The impurity levels of the copper are within the normal range of electrorefining anodes, Chapter 15. The impurity levels could be reduced further by avoiding recycle of the flash hrnace dust. Impurities do not seem therefore, to be a problem in the two existing direct-to- copper smelters. However, metallic copper is always present in the direct-to- copper furnace, ready to absorb impurities. For this reason, concentrates destined for direct-to-copper smelting should always be carefully tested in a pilot furnace before being accepted by the smelter. Table 12.2. Anode compositions from direct to copper smelters. Olympic Dam Glogow I1 ppm in copper Impurity pp m in copper Ag 200-300 1500-3500 AS 250-350 500-800 Au 10-20 Bi 100-150 10-30 Fe 20-200 200-400 Ni 20-40 500- 1000 Pb 10-50 2000-3000 S 20-30 Sb 5-15 50-200 Se 150-350 100 Te 30-50 12.9 Summary Direct-to-copper smelting is smelting of concentrate directly to molten copper in one furnace. In 1994, it is practiced in two smelters; Glogow I1 (Poland) and Olympic Dam (Australia). In both cases the smelting unit is an Outokumpu flash furnace. The main advantage of the process is its restriction of SOz evolution to a single continuous source of high S02-strength gas. In principal, the energy, operating and capital costs of producing metallic copper are also minimized by the single- furnace process. 196 Extractive Metallurgy of Copper Metallic copper is obtained in a flash furnace by setting the ratio: 0, -in -blast input rate concentrate input rate at the point where all the Fe and S in the input concentrate are oxidized. The ratio must be controlled precisely, otherwise Cu2S or Cu20 will also be produced. Avoidance of forming a molten Cu2S layer in the furnace is particularly important. Reactions between Cu2S layers and oxidizing slag have caused rapid SOz evolution and slag foaming. Direct-to-copper flash smelting has proven effective for SO2 capture. However, 15-35% of the Cu-in-concentrate is oxidized, ending up as copper oxide dissolved in slag. This copper oxide must be reduced back to metallic copper, usually with coke. The expense of this Cu-from-slag recovery treatment will probably restrict future direct-to-copper smelting to concentrates which produce little slag. Chalcopyrite concentrates will probably continue to be treated by multi-furnace processes - either by conventional smeltingkonverting or by continuous multi-furnace processing, Chapter 13. Suggested Reading Czemecki, J., Smieszek, Z., Miczkowski, Z., Dobrzanski, J. and Warmuz, M. (1999) Copper metallurgy at the KGHM Polska Miedz S.A. - present state and perspectives. In Copper 99-Cobre 99 Proceedings of the Fourth International Conference, Vol. V Smelting Operations and Advances, ed. George, D.B., Chen, W.J., Mackey P.J. and Weddick, A.J., TMS, Warrendale, PA, 189 203. Davenport, W.G., Jones, D.M., King, M.J. and Partelpoeg, E.H. (2001) Flash Smelting, Analysis, Control and Optimization, TMS, Warrendale, PA (especially Chapters 19-2 1). Hunt, A.G., Day, S.K., Shaw, R.G. and West, R.C. (1999) Developments in direct-to- copper smelting at Olympic Dam. In Copper 99-Cobre 99 Proceedings ofthe Fourth International Conference, Vol. V Smelting Operations and Advances, ed. George, D.B., Chen, W.J., Mackey, P.J. and Weddick, A.J., TMS, Warrendale, PA, 239 253. References Asteljoki, J.A. and Muller, H.B. (1987) Direct smelting of blister copper - flash smelting tests of Olympic Dam concentrate. In Pyrometallurgy 87, The Institution of Mining and Metallurgy, London, England, 265 283. [...]... to 180 95 to 100 0 10 18 69 246 78 44 35 490 180 25 to 28 460 to 470 32 to 35 430 133 (99% 0 2 ) 900 to 1000 98. 4 0.3 0.7 600 14 0.4 480 31 85 0 to 900 98. 5 0.3 0.7 I60 to 180 14 0.4 450 25 82 0 98. 5 0.7 0.9 360 I5 0.34 410 24 206 Extractive Metallurgy of Copper It produces: (a) molten copper, -0.7% S (b) molten slag, 14% Cu (c) SO2 bearing offgas, 25-30 volume% S02 The molten copper continuously departs... Kidd Creek Copper smelter - an update on plant performance In Copper 91-Cobre 91 Proceedings of the Second International Conference Vol IV Pyrometallurgy of Copper, ed Diaz, C., Landolt, C., Luraschi, A and Newman, C J., Pergamon Press, New York, NY, 65 80 Newman, C.J., MacFarlane, G., and Molnar, K.E (1993) Increased productivity from Kidd Creek Copper operations In Extractive Metallurgy of Copper, Nickel... 400 35 I60 to 180 60 60 40 sludge from wastewater treatment plant 50 to 55 600 to 650 2109 (33.2% Cu) 340 42 240 67 61 5 compressed copper scrap 56 540 500 63 34 560 67 386 (82 % S O 2 ) 52 96 60 60 14 reverts 90 45 to 55 600 450 (99% 0 2 ) 140 30 600 to 650 60 31 570 60 matte and slag from smelting furnace 1400 68 1300 10 18 68. 8 1331 0 .8 0.9 500 ( . 28 460 to 470 85 0 to 900 98. 5 0.3 0.7 I60 to 180 14 0.4 450 25 10 18 68. 8 1331 0 .8 0.9 30 500 (<I50 ppm) 10 18 69 246 78 44 32 to 35 430 133 (99% 02) 82 0. Ideal Direct-to -Copper Process Fig. 12.1 is a sketch of the ideal direct-to -copper process. The principal inputs to the process are: 187 188 Extractive Metallurgy of Copper concentrate,. Muller, H.B. (1 987 ) Direct smelting of blister copper - flash smelting tests of Olympic Dam concentrate. In Pyrometallurgy 87 , The Institution of Mining and Metallurgy, London,