Materials Chemistry
Direct-to-Blister Smelting of Copper-Rich Concentrates Using SiO2-FeO-CaO Slag System
2024 Volume 65 Issue 10 Pages 1293-1300
Details
2024 Volume 65 Issue 10 Pages 1293-1300
The direct-to-blister process is considered as a short process, low energy consumption and environmentally friendly pyrometallurgical copper smelting process. In this study, direct-to-blister smelting experiments were conducted on a laboratory scale using high-grade chalcocite as the raw material and employing the SiO2-FeO-CaO slag system. The effects of smelting parameters including Fe/SiO2 ratio, oxygen blowing volume, and smelting temperature on copper recovery were investigated, and the optimal experimental scenario under the SiO2-FeO-CaO slag was ultimately obtained. The results showed that the maximum copper recovery of 96.23 wt.% was realized at the Fe/SiO2 ratio of 1.2 and with CaO addition of 2.8 wt.%. Moreover, the copper losses in the slag and the phases in the slag were analyzed in detail. The results of this paper may provide theoretical guidance for direct-to-blister of high-grade copper concentrates under SiO2-FeO-CaO slag system.
Traditional copper smelting technology consists of the two-step process of melting and converting, whose batch operation has resulted in highly fluctuating flue gas volume and serious SO2 emission. Moreover, potential safety risks and heat loss occur in the copper matte transfer process, causing serious pollution and high energy consumption in the copper smelting process [1, 2]. To resolve the shortcomings of the conventional smelting batch process, the above two steps are expected to be combined into a single direct-to-blister (DB) process for some particular concentrates, such as low-iron chalcocite and bornite concentrates [3]. This process has significant advantages such as short process, high productivity and low SO2 pollution [4]. Currently, several smelters, including the Olympic Dam smelter in Australia, two Glo smelters in Poland and the Chingola smelter in Zambia, have successfully commercialized the DB process [5–7].
The relevant researches [8–12] for DB process mainly focus on thermodynamic calculations and equilibrium systems under different slag types. For the current commercialized DB smelters, iron silicate slag and ferrous-calcium silicate slag as the two basic systems. Compared to iron silicate slag, ferrous-calcium silicate slag presents the advantages of lower viscosity and lower content of dissolved Cu in the slag [3]. Sun et al. [3, 13, 14] determined the slag liquid-phase line temperatures under the iron silicate slag and ferrous-calcium silicate slag systems in the DB process as well as quantified the solubility of Cu2O in the spinel phase. In addition, our group attempted to treat high-grade chalcocite and high-aluminum chalcocite with iron silicate slag and Al2O3-CaO-SiO2 slag systems, respectively, both of which were realized on a laboratory scale with the DB process, and the preliminary experiments obtained favorable results [15, 16]. Chen et al. [17, 18] quantified the benefits of adding appropriate amounts of CaO and SiO2 to the DB process to reduce the dissolved copper content in the slag via thermodynamic modeling and pilot-scale and test. The study by Taskinen et al. [19, 20] claimed that the pure iron silicate slag system was thermodynamically instable under DB smelting conditions, and that the addition of a small amount of CaO was favorable to increase the solubility of magnetite and olivine and to decrease the solubility of copper.
In the DB process, reducing copper losses in the slag was a prerequisite for improving copper recovery. Therefore, to ensure the smooth operation of the DB process, selection and design of a proper slag system based on copper concentrate fractions is of great significance. In this study, the DB process was attempted to be achieved on a laboratory scale by using a SiO2-FeO-CaO slag system with high-grade chalcocite as the raw material. The effects of smelting parameters such as oxygen blowing volume, smelting temperature Fe/SiO2 ratio and CaO addition on copper recovery were investigated with the aim of obtaining the optimal experimental scenario under SiO2-FeO-CaO slag system. Moreover, the results of this paper might provide theoretical guidance for the short process smelting of high-grade copper concentrates under SiO2-FeO-CaO slag system.
The chalcocite and pyrite used in the experiments were obtained from a local copper smelter. The concentrates were crushed and ground to 200 mesh, and the chemical analysis were listed in Table 1. The results showed that the Cu content of chalcocite was 65.03%, and the Fe/SiO2 ratio was only 0.43. The Fe and S contents in pyrite were 45.04% and 52.40%, respectively, and the Fe/SiO2 ratio was up to 51.77. The CaO used in the experiment was analytically pure which was purchased from Sinopharm Chemical Reagent Co., Ltd.
;%0A%09%09%09newWindow.document.open();%0A%09%09%09newWindow.document.write('<img src=%22./Graphics/MT-D2023014tab01.jpg%22>');%0A%09%09%09newWindow.document.close();%0A%09%09)
The results of X-ray diffraction (XRD) analysis of chalcocite and pyrite are shown in Fig. 1. The main phases of chalcocite are Cu2S, SiO2, CaFe(Si2O6), and Fe2O3, whereas pyrite has a more uniform XRD spectrum, with only the FeS2 phase detected.
XRD analysis of chalcocite and pyrite.
A series of high-temperature smelting experiments were carried out in a vertical tube electric-resistance furnace (Mita Electric Furnace Co., Ltd., Xiangtan, China), and the schematic diagram of the experimental device is shown in Fig. 2. Firstly, 80 g of chalcocite, pyrite and CaO were proportionally mixed into a corundum crucible (KPSX Chemical porcelain factory, Tangshan, China, 99% Al2O3 content, 32 * 120 mm size), and then put into the constant temperature zone of the furnace afterwards. Then, the lance (KPSX Chemical porcelain factory, Tangshan, China) was inserted into the constant temperature zone from the top of the furnace to keep the lance located above the crucible and not touched the melt before injecting the oxygen-enriched gas. The furnace was heated at a rate of 10°C/min. The temperature was held for 30 min after reaching the settings to ensure that the mixed material was completely melted. Subsequently, the corundum lance was inserted into the melt at a distance of 5 mm from the bottom of the crucible, and the gas mixture of 70 vol% oxygen (99.5% purity) and 30 vol% nitrogen (99.99% purity) was injected into the melt with a flow rate of 400 mL/min. The lance was removed from the furnace at the end of the oxygen injection, and the blister was settled naturally at the bottom of the crucible. The sample in the crucible was cooled to room temperature and collected for further analysis. The crucible with sample was weighed and subjected to separate the blister copper and slag. The slag amount and copper recovery were calculated using eqs. (1) and (2):
| \begin{equation} M_{4} = M_{1} - M_{2} - M_{3} \end{equation} | (1) |
| \begin{equation} R = \frac{M_{3} \times W_{3}}{M_{0} \times W_{0}} \times 100\% \end{equation} | (2) |
where M1, M2, M3 and M4 represent the mass of crucible with sample, the mass of empty crucible, the mass of blister copper and the amount of slag, respectively; R represents the recovery of copper; M0 and W0 are the weight of raw material and copper content in the ore, respectively; W3 is the copper content in the blister.
The schematic diagram of the experimental device.
This study mainly investigated the effects of smelting parameters including the oxygen blown, smelting temperature, the Fe/SiO2 ratio and CaO dosage on the copper recovery. It to be noted that the oxygen blowing volume ratio refers to the ratio of the actual oxygen blowing volume to the theoretical oxygen blowing volume. All the experimental parameter conditions involved are listed in Table 2.
;%0A%09%09%09newWindow.document.open();%0A%09%09%09newWindow.document.write('<img src=%22./Graphics/MT-D2023014tab02.jpg%22>');%0A%09%09%09newWindow.document.close();%0A%09%09)
X-ray diffraction (XRD) (Science D/max-R, Japan) with Cu Ka target (λ = 1.5406 A, 40 kV, 40 mA) was used to analyze the phases contained in the samples. The diffraction angle (2θ) was scanned from 10∼90° with a scanning speed of 10°/min. The magnetic content in the slag was determined with a SATMAGAN 135 magnetic analyzer with the accuracy of ±0.4%. The microscopic morphology, phase composition and elemental distribution of the samples were analyzed using Electro-Probe Microanalysis (EPMA, JEOL, JXA 8230, Tokyo, Japan). Based on the mineral composition, the viscosity module of Factsage software (7.2 version, ThermFact LTD., Montréal, QC, Canada) was used to calculate the pure liquid phase viscosity. While the viscosity of slag containing solid was determined based on the eq. (3) [21]:
| \begin{equation} \eta = \eta_{s}(1 - f)^{ - 2.5} \end{equation} | (3) |
where η is the viscosity of the slag solid containing, ηs is the viscosity of the pure liquid phase slag; and f is the volume fraction of solid.
The SiO2-FeO-CaO slag system were selected in present study based on the properties of the composition of raw materials used in the smelting process. Figure 3 shows the phase diagram of the SiO2-FeO-CaO slag system. It could be observed that the area of liquid-phase region increased as the temperature increasing, which mainly toward the FeO vertex and the SiO2-CaO edge. The chalcocite and pyrite were selected as the research materials in the present study, and the Fe/SiO2 ratio was set to 0.6. The composition of the slag was located in the SiO2 + slag liquid phase region (shown in Point 1), indicating that the SiO2 content in the slag was oversaturated in the SiO2-FeO-CaO system, meaning that the SiO2 content would be precipitated during the smelting process. As a result, the viscosity of the slag would be increased, which hindered the separation of slag and copper. In addition, the viscosity of the slag was calculated by Factsage at 0.71 Pa·s. Therefore, in order to promote the slag fluidity and improve the environment for the settlement of blister copper, it is suggested to further increase the Fe/SiO2 ratio (Scheme ①) or adding the appropriate amount of CaO (Scheme ②). With the above analysis, the shadowed area in Fig. 3 is located within the liquid-phase region, where the Fe/SiO2 ratio at the range of (0.6–1.4):1, and the CaO/SiO2 ratio in the range of (0.25–1.0):1 were selected as the region to investigate the slag compositions. This means that the viscosity of the slag in the selected scope is maintained at a relatively low value, whereby the DB process could be carried out normally.
Liquid phase diagram of SiO2-FeO-CaO.
The purpose of the DB process is to directly smelt copper concentrate into blister copper, which involves the conversion of sulfides into metals or oxides. Therefore, the oxygen blowing time (i.e., the amount of oxygen) has a crucial effect on the copper recovery. Figure 4 shows the effects of oxygen blowing volume ratio on the copper recovery and sulfur content in blister copper with Fe/SiO2 = 0.6, smelting temperature of 1300°C, and CaO addition of 4.8 wt.%. When the oxygen blowing volume ratio was 1, the blister copper recovery and sulfur content in blister copper obtained in the experiment were 88.15 wt.% and 2.3 wt.%, respectively. This might be attributed to the fact that when the theoretical oxygen volume was injected, the oxygen could not completely contact and react with the melt, thus a portion of Cu2S was not being blown into the blister copper. This resulted in a high copper content in the slag obtained and a low grade of blister copper. As the oxygen blowing volume ratio increased to 1.15, the content of Cu2S in the melt gradually decreased with the increase of oxygen blowing volume ratio. The corresponding blister copper recovery increased to 94.69 wt.%, and the sulfur content in the blister copper further dropped to 0.47 wt.%. However, when the blowing oxygen is too excessive, the generated blister copper continued to be oxidized to Cu2O, causing an increase in the loss of chemically dissolved copper in the slag, thereby reducing the copper straight recovery. It is crucial to appropriately control the volume of oxygen so that the copper sulfides in the melt could be converted to the metallic phase. Consequently, the oxygen blowing volume ratio should be controlled at 1.15 in the present study.
Effect of oxygen blowing volume ratio on copper recovery and sulfur content in copper.
Figure 5 shows the image of the slag-copper separation effect and the XRD pattern of the slag with the oxygen blowing volume ratio of 1.15. As observed from the image of slag-copper separation effect in Fig. 5(a), the copper particles could be settled successfully at the bottom of the crucible for 120 min settlement time, whereas a few particles were interspersed in the upper slag phase. The upper slag layer was collected and the viscosity was calculated to be 0.71 Pa·s based on its composition. The XRD pattern of the slag is shown in Fig. 5(b). The results showed that the main phases of the slag were fayalite, magnetite, and Cu. From the results of the magnetic analyzer and chemical analysis, the contents of Fe3O4 and Cu in the slag were 9.1 wt.% and 28.3 wt.%, respectively. Therefore, it is essential to further optimize the smelting parameters to reduce the Cu loss.
(a) Image of the vertical section of the sample after separation at Fe/SiO2 = 0.6; (b) XRD patterns of copper slag.
The microscopic morphology and EDS analysis of the slag are presented in Fig. 6. The bright white round corresponds to the blister copper phase, and the darkest are the fayalite phases. According to the results of EDS analysis, the copper in the slag mainly exists in metal form, and the remaining present in the slag as chemically dissolved phase (CuFeO2). The viscosity of the slag is the directly contributing factor to the reduction of the copper recovery. The SiO2 content in the slag reached 38.5 wt.% after the removal of copper and sulfur according to the material conservation calculation. It indicated that high melting point substances might be precipitated in the slag liquid, which would seriously hinder the settlement of copper particles by the dramatic increase of slag viscosity. This is consistent with the phenomenon that part of the blister copper was not settled to the bottom in Fig. 5(a). In addition, the temperature and slag composition have a significant effect on the viscosity. In the DB process, as the furnace temperature is relatively stable, the slag viscosity was mainly affected by its composition. The effects of temperature and slag composition on viscosity and copper recovery from DB process would be discussed in detail in the following sections.
Microstructure and energy spectrum analysis of copper slag.
Higher smelting temperature is conducive to decrease the viscosity of slag, thus improving the separation conditions of blister copper and slag in the settlement stage and raise the copper recovery. However, the energy consumption loss would have to be taken into account in actual smelting process, so the appropriate smelting temperature ought to be considered. Figure 7 shows the effect of smelting temperature on copper recovery and slag viscosity. As temperature was increased from 1260°C to 1280°C, copper recovery increased from 90.25 wt.% to 93.13 wt.%. Continuing to increase the temperature to 1340°C, the trend of the copper recovery remained generally constant. It can be noticed obviously that increasing the temperature has a positive effect on the experimental results. Moreover, the liquid phase region variation in Fig. 2 indicated that the increase of temperature expanded the liquid phase area and improved the mobility of the slag, which led to the copper particles settlement successfully. When the temperature reached 1260°C, the slag viscosity was 0.58 Pa·s, and it decreased to 0.28 Pa·s as the temperature increased to 1340°C, indicating that the temperature increased is conducive to improving the slag fluidity to promote the effective separation of the copper from the slag and to reduce the loss of mechanical entrainment of copper in the slag. In addition, raising temperature is favorable to reduce the activity coefficient of CuO0.5, the corresponding copper chemical dissolution loss would be reduced [22]. However, the higher temperature would cause more energy consumption and shorter furnace life [3, 23], which is not advantageous for the economy of the production process. Therefore, it is appropriate for the smelting temperature at 1280°C.
Effect of smelting temperature on copper recovery and sulfur content in copper.
The variation pattern of copper recovery under different Fe/SiO2 and CaO additions are shown in Fig. 8. It can be found that the addition of appropriate amount of CaO is favorable to improve the copper recovery at the lower Fe/SiO2 (Fe/SiO2 < 1.0). The copper recovery rate presented a trend of first increased and then decreased when Fe/SiO2 = 0.6, and the copper recovery was 94.67 wt.% with CaO addition of 5.8 wt.%. It is worth noted that the lower Fe/SiO2 ratio corresponds to the higher amount of proper CaO addition in the range of Fe/SiO2 ratio from 0.6 to 1.0. This phenomenon was mainly attributed to the ability of CaO to destroy the complex silicon-oxygen tetrahedron network structure, so that the slag viscosity was reduced, thus promoting the separation of slag and copper. CaO has a remarkable effect on reducing slag viscosity in melts with high SiO2 content. In contrast, when the Fe/SiO2 ratio exceeded 1.0, the addition of CaO reduced the copper recovery instead. This was due to the fact that the presence of CaO could consume part of SiO2 and increase the activity of FeO, leading to the conversion of FeO into Fe3O4, reducing the interfacial tension between slag and copper [24], which was unfavorable to the separation of slag and copper. Furthermore, increasing the CaO content in the slag also resulted in raising CuO0.5 activity coefficient, which exacerbated the chemical dissolution losses of copper in the slag. The copper recovery was maintained in the optimum range of 95.10–96.23 wt.% with the Fe/SiO2 ratio of 1.2, and the optimum addition of CaO was 2.8 wt.% in this case. Whereas, the copper recovery decreased sharply with an increase of Fe/SiO2 ratio of 1.4 and the higher activity coefficient of CuO0.5 in the slag, indicating that Fe/SiO2 was not suitable to be excessively high in the DB process.
Relationship between CaO addition and copper recovery at different Fe/SiO2.
Figure 9 shows the viscosity of the slag and the content of Fe3O4 in the slag with various Fe/SiO2 ratios and CaO additions. Figure 9(a) indicates that the slag viscosity remained within a relatively high amount range when the Fe/SiO2 ratio was 0.6, while the Fe3O4 content in the slag was within the range of 7.5 to 9.3 wt.% only (Fig. 9(b)). It is apparent that the slag viscosity depending on the amount of SiO2 crystals precipitated in the slag at this stage. The appropriate Fe/SiO2 ratio is conducive to reduce the slag viscosity. However, the slag viscosity showed a slightly increasing trend when Fe/SiO2 = 1.4, in which the Fe3O4 content in the slag was up to 17.3∼19.1 wt.%. The higher Fe/SiO2 ratio led to the increase of the oxygen potential and Fe3O4 activity in the slag, which caused the massive Fe3O4 generation [25]. Moreover, the above results also suggested that despite the fact that increasing Fe/SiO2 ratio could solve the problem of SiO2 precipitation resulting in the increase of viscosity, it would generate excessive foamy slag, which affected the separation of slag and copper.
(a) Viscosity of slag and (b) Fe3O4 content in slag at different Fe/SiO2 and CaO additions.
When the Fe/SiO2 ratio was 1.2, the CaO addition increased from 2.8 wt.% to 6.8 wt.%, and the slag viscosity decreased from 0.31 Pa·s to 0.14 Pa·s, while the corresponding Fe3O4 content in the slag remained constant. This indicated that adding an appropriate amount of CaO could achieve the purpose of reducing the slag viscosity as the CaO content in the slag component was relatively low [26]. Nevertheless, the addition of massive amounts of CaO could increase the amount of slag, and as a result, the absolute amount of copper in the slag also increased [16], which was consistent with the results in Fig. 8. Meanwhile, excessive concentration of CaO in the slag would deteriorate the slag performance. For instance, the formation of anorthite leads to an increase in the liquidus temperature [3, 27]. Therefore, the addition amount of CaO in the slag should be appropriately controlled during the operational process. Under the present system with Fe/SiO2 ratio of 1.2, the addition of 2.8 wt.% CaO is considered reasonable.
When Fe/SiO2 = 1.2, the maximum copper recovery and the lowest slag viscosity were presented in the above experimental results. The slag phase analysis of the slag at different CaO additions was carried out, and the results are shown in Fig. 10. The diffraction peaks of CaFeSi2O6, CaAl2Si2O8, Ca2Al2SiO7, Fe3O4, CuFeO2, and low-intensity Cu phases were mainly present in the slag with low CaO content. It is interesting that the diffraction peaks of CuFeO2, Cu2FeO8 and Cu phases get more intense with the increase of CaO addition, indicating that the copper loss gradually increased, which is in line with the decrease of copper recovery in Fig. 8. When adding 6.8 wt.% CaO, the diffraction peaks mainly consisted of Ca2Al2SiO7, Fe3O4 and CuFeO2 phases, and it could be found that the Ca2AlSiO7 diffraction peaks were intensified whereas the CaFeSi2O6 diffraction peaks were weakened. The above phenomenon could be attributed to the fact that the ferrous oxides in the slag were not enough to consume the excessive CaO, and with the slag basicity progressively increasing, Al2O3 exhibited the characteristics of acidic oxides in the basic slag so as to neutralize the excessive CaO in the slag. In addition, the slag volume was increased with the addition of excessive CaO, which implied the further increase of the Cu loss. Therefore, excessive CaO content in the slag should be avoided at higher Fe/SiO2.
XRD analysis of slag with Fe/SiO2 = 1.2.
Figure 11 shows the blister images and micro-morphological analysis of the blisters obtained at Fe/SiO2 = 1.2 and with the addition of 2.8 wt.% CaO. EDS analysis results indicated that the blister copper consists only of Cu phase.
SEM-EDS analysis of copper at Fe/SiO2 = 1.2, 2.8 wt.% CaO addition: (a) blister images; (b) SEM analysis of blister (200).
For further investigation the phase distribution in copper slag of Fe/SiO2 = 1.2 and with CaO addition of 2.8 wt.% was performed by EPMA spectroscopy and the results are presented in Fig. 12. Copper was present as metal in the copper slag, and magnetite particles were gathered around the copper particles, and the presence of magnetite would hinder the settling of the copper particles. Whereas the other elements such as Ca, Si and O are existed in the glass phase.
EMPA analysis of slag at Fe/SiO2 = 1.2, 2.8 wt.%CaO addition.
In this study, a direct-to-blister smelting process for high-grade copper concentrates was realized on a laboratory scale using the SiO2-FeO-CaO slag system. The results indicated that the main factor contributing to the high slag viscosity was the precipitation of SiO2 crystals when the Fe/SiO2 ratio was 0.6. Increasing the CaO content in the slag at the present period could significantly improve the slag fluidity, and thus increasing the copper recovery. Under the condition of 1280°C, Fe/SiO2 ratio was 1.2, with added amount of CaO of 2.8 wt.%, the blister copper grade was higher than 98.5 wt.%, and the corresponding copper recovery was 96.23 wt.%, and the sulfur content was approximately 0.33 wt.%. The SEM-EDS detection of the slag revealed that the presence of high melting point Fe3O4 phase in the slag and the loss of copper mainly in the form of mechanical entrainment, indicating that the precipitation of high melting point Fe3O4 phase could be one of the dominant factors hindering the separation of slag and copper.
This work was supported by the National Natural Science Foundation of China (51974142).
