Ind. Eng. Chem. Res. 1997, 36, 5419-5424
5419
Studies in Reduction-Roast Leaching Ion Exchange of Copper Converter Slag from an Indian Copper Complex, Ghatshila M. G. Bodas* and S. B. Mathur Non-Ferrous Process Division, National Metallurgical Laboratory, Jamshedpur 831007, India
Utilization of industrial waste is of utmost important due to pollution constraints. Large quantities of converter and anode slags are generated in different unit operations for the production of copper in I.C.C. Ghatshila. The converter slag contains about 2.75% copper, 0.9% nickel, 52.0% iron, and 0.6% cobalt and cannot be rejected due to its economic importance. X-ray diffraction (XRD) studies revealed the presence of copper in elemental, oxide, and silicate phases and iron in silicate (feyalite) and oxide (Fe2O3) phases. From the earlier work on leaching by an acetic acid lixiviant up to 55% Copper could be extracted at 250 mL of acetic acid/100 g of slag, 125 °C temperature, and 35 atm of oxygen pressure. Low copper recovery was attributed to the presence of copper silicate and sulfide phases which could not be detected by XRD. Therefore, reduction roasting by noncoking coal was done to increase the copper recovery. A coal sample from Talchar which contained 30% fixed carbon, 40% volatile, and 30% ash was used for reduction roasting. Pellets were prepared manually by mixing powdered coal in different proportions with ground slag (-100 mesh). Roasting of the pellets was carried out in the temperature range of 650-800 °C. Maximum percent extraction of copper and nickel (97% Cu and 20% Ni) could be achieved at the temperature of 720 °C for 90 min of roasting with 20% reductant. The acetic acid requirement was 130 mL/100 g of reduced pellets, while leaching at the oxygen pressure of 35 atm for 120 min at 120 °C. The solid-liquid ratio was maintained at 1:5 in all the experiments. The copper concentration in the leach liquor was 4.5 g/L. Iron dissolution was minimum (max 1%) in all the experiments. The ion-exchange technique was employed to separate copper from acetate solution. Zeolite resin was found to be suitable for the separation of copper from leach liquor. About 92% copper was recovered by geolite. 2.5 N H2SO4 was found to be most effective for the stripping of the copper from resin. Introduction Present day, recycling of metallurgical waste has been a major area of interest in some of the non-ferrous industries. As the present reserves are dwindling out and the demand is continuously increasing, the only alternate source for the non-ferrous metals would be to utilize scrap, waste, slag, slime, etc. For example, aluminium, zinc, copper, etc., are recovered from scrap, dross, etc., by pyro- and hydrometallurgical processes. Recovery of copper and nickel from a converter slag, as well as from an anode slag, is a major problem in the copper plants. The copper plant at I.C.C. Ghatshila generates nearly 50 t/day of converter slag and 5 t/day of anode slag which so far are not fully utilized to recover the valuable metals such as Cu, Ni, and Co. The objective of the present investigation is to develop a suitable and economically viable technology by which Cu can be selectively recovered from the converter slag. From the available literature it is found that not much work has been done on the recovery of metal values from copper converter and anode slags. Anand et al. (1) recovered the metal values from the converter slag of I.C.C. Ghatshila with ferric chloride leaching. Various leaching parameters such as stirring rate, temperature, concentration of ferric chloride, S:L ratio, and the effect of particle size were studied for the extraction of Cu, Ni, and Co. Direct leaching of converter slag by ferric chloride (1) gave low recoveries of Ni and Co. Therefore reduction roast and the subsequent ferric chloride leaching route was tried for the recovery of copper, nickel, and cobalt by Anand et al. (2). The ground slag (-100 mesh B.S.S.) was mixed with various reductants such as charcoal, S0888-5885(97)00222-4 CCC: $14.00
bituminous coal from Talcher, lignite, and furnace oil. The mixed charge then reduced at different temperatures from 750 to 950 °C. The reduced slag was leached with different concentrations of ferric chloride. It was observed that the reduction above 750 °C, copper recovery decreases and nickel and cobalt recovery increases. This was attributed to the formation of copper ferrite under optimum conditions. Then, 80% copper, 95% nickel, and 80% cobalt could be recovered by reducing the slag at 750 °C with 10 wt % furnace oil followed by ferric chloride leaching. Sulfuric acid leaching of converter slag also gave quite good recoveries of Cu, Ni, and Co (3,4) along with a very high dissolution of iron in the leach liquor. The optimum conditions obtained for the recovery of these metals include a leaching time of 2 h, temperature of 100 °C, S:L ratio ) 1:10, particle size of -100 mesh B.S.S., 100 mL of 0.94 M sulfuric acid, and the recoveries of metals equal to 90.86% copper, 97.8% nickel, 98.7% cobalt, and 84.7% iron were achieved. The bulk of the copper is present as small droplets of Cu metal, coated with copper iron sulfide in copper slag. The chief metallic mineral in the sample is magnetite while the copper is present as chalcocite with minor occurrences of bornite, chalcopyrite, covelite, copper oxide, and metallic copper Chakraborthi (6). The main slag matrix consists of fayalite, pyroxene, mellilite, and glass. Nickel and cobalt are also present as a mixture of sulfide, oxide, and metal in varying proportions. Leaching only of copper values from copper smelter and converter slags with acidic ferric sulfate solutions and with ammonia-ammonium-carbonate solutions at atmospheric pressure has been reported by Shelley (7). © 1997 American Chemical Society
5420 Ind. Eng. Chem. Res., Vol. 36, No. 12, 1997 Table 1. Chemical Analysis of Converter Slag constituents
wt %
Cu Ni Co Fe Zn Mn SiO2 MgO CaO Al2O3 L.O.I.
2.75 0.92 0.56 52.10 0.17 0.035 18.80 4.50 0.16 7.70
Particle size of the charge was very much effective on the recovery. As the particle size decreased from +30 to -200 mesh, the recovery of metals showed an increasing trend. This was attributed to the liberation of copper, from the slag particles during grinding. Other investigators have also done considerable research work on the recovery of copper values from converter slag (57). Experimental Section Leaching of Converter Slag. The converter slag sample from I.C.C. Ghatshila, containing copper, nickel, and cobalt was crushed, then ground in a pot mill, and sieved through -100 mesh B.S.S. sieve. Only the -100 mesh fraction was used in all the experiments. The chemical analysis of the converter slag sample is given in Table 1. The leaching experiments were carried out on ground converter slag in a 2 L autoclave at different temperature and pressure to recover metallic value from it. The pressure leaching was done to suppress the iron which was also dissolved in the solution along with Cu, Ni, and Co at ambient temperature and pressure. The sketch diagram of the autoclave is given in Figure 1. The pressure leaching experiments were conducted to study the various parameters such as concentration of acetic acid, S:L ratio (wt/vol), leaching time, pressure, and temperature. Acetic acid concentration was varied from 5 to 25%. The pressure varied from 1 to 50 atm, leaching time varied from 30 to 180 min, and the temperature varied from 30 to 150 °C. The results are given in Figures 2-5. Roasting of Converter Slag. As mentioned earlier only 55% of copper could be extracted by pressure leaching of the converter slag. To enhance the copper extraction, roasting of converter slag prior to leaching was tried. The converter slag was roasted along with reductant coal having 30% fixed carbon, 40% volatile matter, and 30% ash at different temperatures from 600 to 800 °C for 90 min in a muffle furnace. The roasted slag was leached with acetic acid in an autoclave, keeping other parameters constant i.e. solid:liquid ratios at 1:5, pressure at 35 atm, temperature at 120 °C, and 90 min retention time. After roasting at 750 °C the copper was dissolved in a lixiviant (acetic acid), and about 97% Cu, 15%, Ni and 1% Fe were extracted. Separation of Copper by Ion Exchange Process. Synthetic cation-exchange resin was used for the separation of copper from the leach liquor. The resin was a strong cation exchanger. A known amount of resin (200 g) was taken in a glass tube with a diameter of 7.5 cm and length of 80 cm (Figure 11). The one end of the glass tube was fitted with a perforated disc. The resin was initially activated with dilute sulfuric acid and then
Figure 1. Schematic diagram of autoclave.
washed with distilled water several times for complete removal of sulfate ions. A known volume of leach liquor containing copper and other metal impurities was passed in the column. Copper content in the leach liquor was 4.8 g/L. The loaded resin was thoroughly washed with distilled water until it became a free copper acetate solution. Elution of the loaded resin was done with different concentrations of H2SO4 (1, 2, and 2.5 N). After elution the resin was again washed with distilled water until it was free from copper sulfate. All the solutions, i.e., leach liquor, raffinate, washings, eluted copper sulfate solution, were analyzed for copper, nickel, and iron content. The results are tabulated in Table 3. Results and Discussion Characterization of Converter Slag. The converter slag is generated during blasting and smelting of matte in the converter which contains nickel and copper along with iron. The chemical composition of slag is given in Table 1. XRD of the converter slag sample was done and the results are tabulated in Table 2 A-C). The observed dÅ values were compared with reported dÅ values for different phases. Copper was found to be present in elemental and oxide forms and iron as Fe2O3 and FeSiO3 phases. As the other elements like Ni and Co were in less concentration, these could not be identified. Leaching of Copper Converter Slag with Acetic Acid. Leaching tests were carried out using an appropriate quantity of -100 mesh converter slag with acetic acid. The leaching experiments were done in ambient conditions as well as under pressure using a 2 L capacity autoclave. Tests under ambient conditions gave copper recovery up to 15% along with 50% iron in leach liquor. Therefore to suppress the iron dissolution and to increase copper recovery, pressure-leaching experiments were carried out in an autoclave. The different parameters such as acetic acid concentration, leaching time, temperature, S:L ratio, pressure, etc., were studied. The schematic diagram of the 2 L autoclave is shown in Figure 1. The leaching results are given in Figures 2-5. Commercial acetic acid was used for carrying out the experiments. Acid concentration was varied from 25 to 150 mL for 50 g of slag. The solid:liquid ratio was kept at 1:5 (wt/ vol) for each set of experiments. The pressure and temperature were varied from 1 to 50 atm and from 30
Ind. Eng. Chem. Res., Vol. 36, No. 12, 1997 5421
Figure 2. Variation of percent extraction of metals with acetic acid.
Figure 4. Variation of percent extraction of metals with time. Figure 3. Effect of temperature on leaching of metal values.
to 125 °C, respectively, and the leaching time varied from 30 to 150 min. It was observed from Figure 2, as the concentration of acetic acid was increased, the copper recovery also increases but remained constant when 100 mL of acetic acid was used per 50 g of slag. The other variables were kept constant. Nickel and iron dissolution was negligible at this concentration but copper extraction was 46% in a 2 min leaching time. Under the same conditions nickel and iron dissolution was high but as the leaching time increased, the dissolution of these metals did not show any appreciable change. About 55% of copper could be extracted by pressure leaching only. It was also observed in ferric chloride leaching. Therefore roast reduction with pressureleaching route was followed in a view to increase the recovery of the copper metal. The converter slag was mixed with coal and then reduced at different temperatures in a muffle furnace. The reduced slag was
leached with acetic acid at a different temperature and pressure. The results were quite encouraging and plotted in Figures 8-10. After optimizing the roasting temperature, the experiments were conducted to find out the optimum roasting time at 750 °C. The results are plotted in Figure 9. It was observed as the roasting time increased from 30 to 120 min, the recovery of copper and nickel also increased but iron dissolution remained almost constant up to 2%. Copper and nickel recovery was increased up to a 90 min roasting time after this became constant. The amount of reductant was varied from 10 to 25%, keeping other parameters, i.e., roasting temperature and time, constant. The results are plotted in Figure 10. It was observed that as the reductant increased from 10 to 20%, the recovery of copper increased from 70 to 97.5%; a further increase of reductant did not show any effect on copper recovery. The nickel also showed the increasing trend with an increase of reductant and
5422 Ind. Eng. Chem. Res., Vol. 36, No. 12, 1997
Figure 6. Variation of extraction rate of copper with time.
Figure 5. Variation of percent extraction of metals with pressure.
became almost constant after 20%. Iron dissolution was also not much with 20% reducatant. Therefore 20% reductant was taken as the optimum. With the encouragement of laboratory scale results some bigger scale experiments were conducted in a 20 L autoclave using 500 g of roasted converter slag to verify the laboratory scale results. The optimum recovery of copper (97%) and nickel (15.5%) were obtained in this case. Effect of temperature on copper recovery was also studied and is shown in Figure 5. It was found that percent extraction of copper increased with a rise in temperature and no change was observed after 120 °C and copper recovery was 55%. Iron dissolution was found to be high at low temperature and decreased with a rise in temperature, indicating the precipitation of iron hydroxide. From the above data the activation energy was determined by using Arrhenius plots of log k vs 1/T. Kinetics of Leaching. Kinetics of the leaching seems to be a diffusion-controlled mechanism of the first order and obeys the following kinetic equation:
kt ) 1 - 2/3R - (1 - R)2/3
(1)
k ) 2FDC/r20 where k is a rate constant, F the molar density, D the diffusion coefficient, C the bulk concentration, r0 the intial particle radius, R the fraction dissolved, and where t the time of leaching. The variation of R is plotted against time and is shown in Figure 7. Activation energy calculated by plotting log k vs 1/T from the Arrhenius equation.
k ) e∆E/kt
(2)
k ) Aexp{∆E/RT} was found to be very low (∆E ) 1.7 kJ/mol) and is in agreement with the diffusion-controlled mechanism. Variation of percent extraction of copper with pressure showed a steady increase up to a certain oxygen pressure (35 atm) and afterward remained constant. In all the experiments the confidence level was +1%.
Figure 7. Arrhenius plot for the extraction of copper in acetic acid.
Separation of Copper from Leach Liquor by Ion Exchange Technique Zeolite Molecular Sieves. Zeolites are crystalline, hydrated aluminosilicates of group I and group II elements, in particular sodium, potassium, magnesium, calcium, strontium, and barium. Structurally the zeolites are “framework” alumina silicates which are based on an infinitely extending three-dimensional network of AlO4 and SiO4 tetrahedra linked to each other by that sharing of all the oxygens (8). Zeolite may be represented by the empirical formula
M2/nO‚Al2O3‚xSiO2‚yH2O In this oxide formula x is generally equal to or greater than 2 since AlO4 tetrahedra are joined only to SiO4 tetrahedra and n is a certain valency. Commercial adsorbents which exhibit ultraporosity and which are generally used for the separation of gas and vapor mixtures include the activated carbons, activated clays, inorganic gels such as silica gel, and activated alumina and the crystalline aluminum silicate (zeolites).
Ind. Eng. Chem. Res., Vol. 36, No. 12, 1997 5423
Figure 8. Variation of percent extraction of metals with roasting temperature.
Figure 9. Variation of percent extraction of metals with roasting time.
Table 2. XRD of Free Copper in Converter Slag, of Fe2O3, and of FeSiO3 I/I0 100 45 20 17 5 3 9 8 I/I0
ASTM (dÅ)
I/I0
observed (dÅ)
(A) Free Copper in Converter Slag 2.0880 76.5 1.8080 90.8 1.2780 1.0900 90.8 1.0436 73.8 0.9038 0.8293 0.8083 ASTM-33-664 (dÅ)
2.0898 1.0835 1.0835 1.0447
I/I0
observed (dÅ)
25.3 46.8 10.0 26.5 60.0 73.8 64.8 100.0 92.6
1.6971 1.6018 1.4845 1.3110 1.1408 1.0447 0.9786 0.9611 0.9357
I/I0
observed (dÅ)
29.2
3.3506
27.2 27.9
2.9907 2.9624
(B) Fe2O3 45 5 30 10 7 7 4 5 2
1.6941 1.6033 1.4859 1.3115 1.1411 1.1035 0.9715 0.9606 0.9318
I/I0
ASTM 17-547 (dÅ) (C) FeSiO3
70 100 40 70 10 90 20 20 20 10 20 20 20
6.48 4.62 3.34 3.24 2.992 2.911 2.750 2.596 2.510 2.406 2.302 2.134 1.995
Figure 10. Variation of percent extraction of metals with coal. 100 36.3
2.5103 2.4123
Activated carbons, activated alumina, and silica gel do not possess an ordered crystal structure and consequently the pores are nonuniform. The distribution of the pore diameters within the adsorbent particles may
be narrow (20-50 Å) or it may range widely (twentyseveral thousand Å) as in the case for some activated carbons. Hence all molecular species, with the possible exceptions of high molecular weight polymeric materials, may enter the pores. Zeolite molecular sieves have pores of uniform size (3-10 Å) which are unequally determined by the unit structure of the crystal. The pores will completely exclude molecules which are larger than their diameter.
5424 Ind. Eng. Chem. Res., Vol. 36, No. 12, 1997
Figure 11. Ion exchange column. Table 3. Separation of Copper from Acetic Acid Solution by Zeolite Resin Cu (g/L)
Ni (g/L)
Fe (g/L)
4.8
0.24
1.2
raffinate Cu% 8.8 4.2 1.4
Ni%
Ni%
Fe%
(A) Loading of Leach Liquor, 4.0 L 5.0 27.8 2.9 1.4 trace 2.1 trace 3.5 trace 1.5 trace 3.0
Cu%
52.8 95.7 99.3
eluted solution H2SO4 1N 2N 2.5 N
Acknowledgment
washing Fe%
Cu%
Ni%
loss Fe%
Cu%
Ni%
(B) Elution of Loaded Copper with H2SO4 87.9 93.5 18.5 0.6 0.1 95.3 96.3 nil 0.5 0.2 97.5 96.7 nil 0.3 0.3
copper at optimum conditions of 250 mL of acetic acid/ 100 g of slag, 120 °C temperature, 35 atm of oxygen pressure, solid:liquor ratio of 1:5 (wt/volume), and 20 min leaching period was obtained. 4. The dissolution of copper in acetic acid was found to be a diffusion-controlled mechanism. 5. Reduction roast with coal and subsequent acetic acid leaching enhanced the copper recovery to 97.5% at 720 °C temperature in a 90 min roasting time with 20% reductant. 6. Optimum conditions in the case of reduction roast acetic acid leaching were 260 mL of acetic acid/100 g of slag, 35 atm pressure, 125 °C temperature, and a 120 min leaching period. S:L was kept at 1:5 (weight/ volume). 7. Zeolite cation-exchange resin was found to be effective for the separation of copper from leach solution. The confidence level in each experiment was (1% observed.
Fe% 0.9 0.4 0.2
The ion exchange technique was employed for the separation of copper from copper acetate solution. Tulsion T-42, cation-exchange resin, was found to be very effective for the separation of copper from the leach liquor. The resin was kept in a perforated disc attached to the one end of the glass tube with 8 cm diameter and 60 cm length as shown in Figure 11. The flow rate of leach liquor was kept constant (6.5 L/h). The copper from the leach liquor was absorbed by the resin, and acetic acid was liberated in the raffinate which can be recycled for the next batch of leaching experiments. The loaded copper in the resin was thoroughly washed with distilled water a number of times. The loaded copper was eluted with 1, 2, and 2.5 N sulfuric acid solution. The results are given in Table 3. Conclusions 1. From X-ray powder diffraction analysis of converter slag copper was found to be present in its native state, oxide, and silicate forms and iron as Fe2O3 and FeSiO3. 2. Magnetic separation of converter slag was not possible due to the complex nature of the slag as the slag was highly magnetic. 3. Leaching of the converter slag gave about 55% copper recovery in solution. It indicates the presence of sulfide and silicate in the residue. The recovery of
The authors are very much indebted to the Director, National Metallurgical Laboratory, Jamshedpur for guidance and permission to publish this paper. The authors are also grateful to I.C.C. Ghatshila for sparing the slag samples. The authors thank ANC Division for chemical analysis of the samples. The authors also thank Mr. D. P. Rao, Pariti, Sr. Steno for typing the manuscript. Literature Cited (1) Anand, S.; Kanta Rao, P.; Jena, P. K. Recovery of copper, nickel and cobalt from copper converter and smelter slag through ferric chloride leaching. Hydrometallurgy 1980, 5, 355-365. (2) Anand, S.; Kanta Rao, P.; Jena, P. K. Reduction Roast Ferric Chloride Leaching of Converter Slag. Hydrometallurgy 1981, 7 (3), 252. (3) Anand, S.; Sarveswara Rao, K.; Jena P. K. Pressure leaching of copper converter slag using dilute sulphuric acid for the extraction of cobalt, nickel and copper values. Hydrometallurgy 1983, 10, 305-312. (4) Anand, S.; Kanta Rao, P.; Jena, P. K. Leaching behaviour of copper converter slag in sulphuric acid. Trans. Indian Inst. Met. 1980, 33 (1), 77-81. (5) Write, R. M.; Organs, J. R.; Harris, G. B.; Thomas, J. A. Development of a process for the recovery of electrolytic copper and cobalt from Rokhana converter slag. In Jones, M. J., Ed.; Advances in Extractive Metallurgy; The Institution of Mining and Metallurgy: London, 1977; pp 57-68. (6) Chakraborthy, D. M. Recovery of copper value from copper slag. Symposium on the Utilization of Minerals and Metallurgical Wastes and By-products, Regional Research Laboratory, Bhubaneswar, India, Dec. 3-4, 1977. (7) Shelley, T. R. Possible methods for recovering copper from waste copper smelting slags by leaching. Trans. Inst. Min. Metall., Sect. C. 1975, 84, 1-4. (8) Breck, Donald W. Zeolite Molecular Sieves: Structure, Chemistry, and Use; John Wiley and Sons, Inc.: New York, 1974.
Received for review March 19, 1997 Revised manuscript received September 17, 1997 Accepted September 22, 1997X IE970222E
X Abstract published in Advance ACS Abstracts, November 1, 1997.
