Stepwise Utilization Process to Recover Valuable Components from Copper Slag

: Waste copper slag is a typical hazardous solid waste containing a variety of valuable elements and has not been effectively disposed of so far. In this paper, a stepwise extraction process was proposed to recover valuable elements (copper, iron, lead and zinc) from waste copper slag. The speciﬁc procedures are as follows: (1) A ﬂotation process was adopted to enrich copper, and when the copper grade in the ﬂotation concentrate was 21.50%, the copper recovery rate was 77.78%. (2) The ﬂotation tailings were pelletized with limestone, then the green pellets were reduced, and the magnetic separation process was carried out. When the iron and copper grades in the magnetic concentrate were 90.21% Fe and 0.4% Cu, 91.34% iron and 83.41% copper were recovered, respectively. (3) Non-magnetic tailings were mixed with clinker and standard sand to produce common Portland cement. Several products were obtained from the waste copper slag through the proposed process: ﬂotation concentrate, measured 21.50% Cu; magnetic concentrate, containing 90.21% TFe and 0.4% Cu; direct reduction dust, including 65.17% ZnO and 2.66% PbO; common Portland cement for building construction. The comprehensive utilization method for waste copper slag achieved zero tailing and has great potential for practical application.


Introduction
One of the greatest challenges in the mining and processing industries is waste management and waste copper slag is defined as one such waste [1]. Waste copper slag (WCS) is a by-product produced in the copper pyrometallurgical process. It is estimated that 30 million tons of WCS are generated and dumped globally every year [2,3]. In addition, the slag contains a lot of harmful elements, such as Cu, Pb and Zn, which cause pollution to the environment. These elements are not only non-biodegradable and toxic but also accumulate in human organisms, causing some physical disorders and health concerns [4][5][6]. Therefore, WCS is classified as hazardous waste [7][8][9]. Unfortunately, more than 80% of them are directly dumped without treatment, which may cause potential pollution [10,11].
In order to comply with strict environmental regulations, numerous studies have been conducted to upgrade WCS, including building material [12]. However, this method usually has new contamination, such as the leachability of zinc, arsenic and antimony, and can only accommodate a very small fraction [13,14].
Regarding the recovery of valuable metals from WCS, previous work has mainly focused on recycling elements from WCS using beneficiation, leaching and pyrometallurgical processes. These methods mainly include flotation [15,16], acid leaching [17,18], roasting followed by water leaching [19], bio-leaching [20] and direct reduction-magnetic separation or smelting reduction [21][22][23]. Roy et al. [15] adopted flotation to recover copper sulfide from copper smelter slag. The results show that under the optimal conditions, the recovery rate of copper sulfide reached 84.82%. There is no doubt that flotation can recover copper sulfide to a certain extent. However, the copper oxide within it cannot be recovered. Muhammad et al. [18] employed sulfuric acid to extract metals from copper-rich converter slag. A copper precipitation rate of 90% can be obtained. At the same time, the iron extraction rate is only slightly higher than 60%. Mikoda et al. [20] used acidithiobacillus thiooxidans to leach valuable metals from copper metallurgical slag. The extraction rate of rare earth element (REE) and Co can reach more than 90%. However, the process duration is too long. Physical beneficiation and hydrometallurgical process can efficiently extract Cu, Pb, REE, etc. However, these processes cannot effectively recover iron as the main element in copper slag.
The pyrometallurgical process is generally used to recover iron from copper slag [24]. Smelting reduction is a deep reduction process; during this process, FeO is reduced to pig iron by a reductant at a high temperature above 1450 • C [1]. It can obtain a high iron recovery rate, but shortens the use time of the kiln. It is effective to recover iron from copper slag (CS) by direct reduction-magnetic separation. By reducing CS at 1200-1300 • C, fayalite and copper sulfide can be transformed into metallic iron and copper, and then Fe-Cu alloy phase can be formed [25]. Subsequently, a magnetic separation process is used to recover the Fe-Cu alloy. However, magnetic separation tailings have not been effectively used.
In this paper, a comprehensive process for recovering valuable metals from WCS is developed. The main steps of the proposed process include flotation to enrich Cu and direct reducing-magnetic separation to produce crude Fe-Cu alloy, and subsequent non-magnetic tailings to replace clinker to produce common Portland cement. Meanwhile, lead and zinc were recovered as dust through this process. No tailings were produced during the entire process, indicating that this is a clean and green process.

Materials
The WCS used in this study was supplied by Anhui Province, China. The chemical composition of WCS is shown in Table 1. The WCS contains 39.54% Fe, 0.81% Cu and 32.11% SiO 2 . Meanwhile, the contents of Pb and Zn are 0.12% and 2.35%, respectively, which can be volatilized from the reduction process. As shown in Figure 1, the main mineral phases of copper slag are magnetite (Fe 3 O 4 ) and fayalite (Fe 2 SiO 4 ). The microstructure of the copper slag was observed by SEM-EDS, and the result is shown in Figure 2. It can be seen from Figure 2 that iron mainly exists in the form of fayalite (Fe 2 SiO 4 ) and magnetite (Fe 3 O 4 ), while copper mainly exists in the form of matte (Cu 2 S). Meanwhile, iron minerals and copper minerals exist as independent particles, which facilitates the recovery of copper through the flotation process.    Table 2).   Table 2).  Table 2).

Methods
The flow chart of the proposed WCS treatment process is presented in Figure 3, which was developed based on the basic understanding of the following steps: (1) the WCS is enriched with Cu by a flotation process; (2) direct reduction-magnetic separation of flotation tailing pellets to prepare Fe-Cu alloy and to remove Pb and Zn; (3) non-magnetic tailing to prepare Portland cement to achieve zero tailings. The experimental details of each step are described as follows.

Methods
The flow chart of the proposed WCS treatment process is presented in Figure 3, which was developed based on the basic understanding of the following steps: 1) the WCS is enriched with Cu by a flotation process; 2) direct reduction-magnetic separation of flotation tailing pellets to prepare Fe-Cu alloy and to remove Pb and Zn; 3) non-magnetic tailing to prepare Portland cement to achieve zero tailings. The experimental details of each step are described as follows.

Flotation Process to Recover Cu
The flotation tests were carried out in a 0.75 L XFD laboratory flotation cell (FGC5-35, Yunhao Mining and Metallurgical Equipment Co., Ltd., Wuhan, China). In all flotation tests, a 500 g sample was used, the impeller speed was set to 600 rpm and the aeration rate was 0.2 dm 3 /min. The specific operation of the flotation tests is outlined in Figure 4. In the flotation process, analytical grade Butyl xanthate (C 5 H 10 OS 2 ) (g/t) and terpene oil (C 10 H 18 O) (g/t) were used as collector and frother, respectively [7]. The recovery rate of copper is calculated as follows: where η Cu is the recovery rate of copper; w 1 and w 0 are the copper grades in the flotation concentrate and raw material, respectively; y is the yield of the flotation concentrate. The flotation tests were carried out in a 0.75 L XFD laboratory flotation cell (FGC5-35, Yunhao Mining and Metallurgical Equipment Co., Ltd., Wuhan, China). In all flotation tests, a 500 g sample was used, the impeller speed was set to 600 rpm and the aeration rate was 0.2 dm 3 /min. The specific operation of the flotation tests is outlined in Figure 4. In the flotation process, analytical grade Butyl xanthate (C5H10OS2) (g/t) and terpene oil (C10H18O) (g/t) were used as collector and frother, respectively [7]. The recovery rate of copper is calculated as follows: where is the recovery rate of copper; w1 and w0 are the copper grades in the flotation concentrate and raw material, respectively; y is the yield of the flotation concentrate.

Preparation of Crude Fe-Cu Alloy by Direct Reduction-Magnetic Separation
The flotation tailings and flux were fully mixed, and then the mixture was pressed into cylindrical briquettes (height: 10 mm and diameter: 10 mm). The briquettes were dried at 110 °C for 4 h.
The direct reduction tests were performed in a muffle furnace at 1250 °C for 80 min. For the reduction test, 60 g briquettes were mixed with 120 g soft coal (ash content: 4.49%; production location: Shenfu) (1-5 mm) and then encased into a corundum crucible with a depth of 80 mm and diameter of 80 mm. [21]. After reduction, the reduced briquettes were cooled to room temperature for the subsequent magnetic separation process. The specific operation process of magnetic separation is outlined in reference [26]. The metal recovery rate (iron or copper) is calculated as follows:

Preparation of Crude Fe-Cu Alloy by Direct Reduction-Magnetic Separation
The flotation tailings and flux were fully mixed, and then the mixture was pressed into cylindrical briquettes (height: 10 mm and diameter: 10 mm). The briquettes were dried at 110 • C for 4 h.
The direct reduction tests were performed in a muffle furnace at 1250 • C for 80 min. For the reduction test, 60 g briquettes were mixed with 120 g soft coal (ash content: 4.49%; production location: Shenfu) (1-5 mm) and then encased into a corundum crucible with a depth of 80 mm and diameter of 80 mm. [21]. After reduction, the reduced briquettes were cooled to room temperature for the subsequent magnetic separation process. The specific operation process of magnetic separation is outlined in reference [26]. The metal recovery rate (iron or copper) is calculated as follows: where η is the metal recovery rate (iron or copper); w and w are the iron (or copper) grades in the magnetic separation concentrate and reduced pellets, respectively; and y 1 is the yield of the concentrate. The volatilization rate of zinc (or lead) is calculated as follows: where γ is the volatilization rate of zinc (or lead); w 2 and w 3 are the content of zinc (or lead) in the reduced pellets and flotation, respectively; and y 2 is the yield of reduced pellets.

Production of Common Portland Cement
A certain proportion of non-magnetic material (NMT), clinker and gypsum were thoroughly mixed. Ingredients of Portland cement are shown in Table 3. The mixtures were cast into steel molds (cuboid) in three layers and compacted using a vibrating table [27]. All samples were stored in the ambient environment under a plastic sheet, and then the samples were demolded and wet-cured until the test date. The samples were cured at room temperature (18-22°C) for 3 and 28 days separately. Concrete cuboids with a size of 40 × 40 × 160 mm were cast for the compressive strength test (Equipment: Universal testing machine, KY-D2305, Jilin Province Jilin Test Technology Co., Ltd., Jilin, China).

Characterization of Raw Materials and Products
The chemical compositions of WCS, flotation slag, Fe-Cu alloy and non-magnetical material tailings were measured by X-ray fluorescence spectroscopy (XRF, PANalytical Axios mAX, PANalytical B, V., Almelo, The Netherlands). The microstructure of reduced briquettes was observed and characterized by scanning electron microscopy (SEM) (TESCAN, MIRA3, LMH, FEI, Eindhoven, The Netherlands) and energy-dispersive spectrometry (EDS) (XMAX20, FEI, Eindhovern, The Netherlands). The mineral phases of WCS, flotation tailing, direct reduction dust and reduced briquettes were measured by X-ray powder diffraction (XRD, RIGAKU, D/Max-2500, Bruker corporation, Madison, WI, USA) with a 2θ scan range from 10 • to 80 • . The compressive strength of the Portland cement was measured according to the national standard (GB/T 175-2007). Table 4 lists the chemical composition of the flotation concentrates and tailings. When the overall copper recovery rate was 77.78%, the Cu content of the flotation concentrate was 21.50%. In addition, the contents of Fe total and SiO 2 in the concentrate were higher, wth values are 28.22% and 21.33%, respectively. Meanwhile, the contents of S, Pb and Zn were 10.15%, 0.77% and 0.72%, respectively. The contents of Fe, SiO 2 , Pb and Zn of the tailings were 39.88%, 32.44%, 0.1% and 2.4%, respectively. The size distribution of the Cu concentrate and flotation tailings is presented in Figure 5a,b. It can be seen that their particle size is fine, and 80% pass 100 µm. concentrate and flotation tailings is presented in Figure 5a,b. It can be seen that their particle size is fine, and 80% pass 100 μm.  The XRD patterns of the Cu concentrate and tailings are shown in Figure 6a,b. It indicates that iron in the flotation tailings mainly exists in the form of fayalite (Fe2SiO4) and magnetite (Fe3O4). The microstructure of copper concentrate was observed by SEM-EDS, and the results are shown in Figure 7a-d. It can be seen from Figure 7 that chalcopyrite (CuFeS2) and copper sulfide (Cu2S) are the main mineral phases of copper-bearing with a particle size of less than 50 μm.  concentrate and flotation tailings is presented in Figure 5a,b. It can be seen that their particle size is fine, and 80% pass 100 μm.  The XRD patterns of the Cu concentrate and tailings are shown in Figure 6a,b. It indicates that iron in the flotation tailings mainly exists in the form of fayalite (Fe2SiO4) and magnetite (Fe3O4). The microstructure of copper concentrate was observed by SEM-EDS, and the results are shown in Figure 7a-d. It can be seen from Figure 7 that chalcopyrite (CuFeS2) and copper sulfide (Cu2S) are the main mineral phases of copper-bearing with a particle size of less than 50 μm.

Direct Reduction-Magnetic Process to Recover Fe, Cu, Pb and Zn
According to Figure 6, the main phases of flotation tailings are fayalite and magnetite. The reduction of magnetite, fayalite and pyrite has been investigated in previous research [10,21,23]. These studies have shown that it is difficult to reduce fayalite with CO and that it can only be reduced by carbon at high temperatures. In the presence of CaO, fayalite is more easily reduced by C and CO. Pyrite and copper sulfide are difficult to be reduced by solid carbon to form metallic iron and copper. Adding CaO can reduce the required Gibbs free energy of pyrite and copper sulfide reduction. Figure 8 presents the effect of additive dosage on the recovery of Fe and Cu and the volatilization of Zn and Pb. Without additives, the recovery rates of iron and copper are only 72.87% and 65.34%, respectively. Meanwhile, the iron grade of magnetic concentrate is 78.62%. Compared with no additives, in the presence of 20% additive, the recovery rates of iron and copper are 91.34% and 83.41%, respectively, and the grade of iron is 90.21%, and the content of copper is maintained at 0.34%. When the additives are 20% and 0%, the volatilization rates of zinc and lead are 98.89% and 95.33%, and 89.6% and 86.33%, respectively. Increasing the dosage to 25% has a slight effect on the indexes. From the above, it is recommended to add 20% as the best additive dosage.

Direct Reduction-Magnetic Process to Recover Fe, Cu, Pb and Zn
According to Figure 6, the main phases of flotation tailings are fayalite and magnetite. The reduction of magnetite, fayalite and pyrite has been investigated in previous research [10,21,23]. These studies have shown that it is difficult to reduce fayalite with CO and that it can only be reduced by carbon at high temperatures. In the presence of CaO, fayalite is more easily reduced by C and CO. Pyrite and copper sulfide are difficult to be reduced by solid carbon to form metallic iron and copper. Adding CaO can reduce the required Gibbs free energy of pyrite and copper sulfide reduction. Figure 8 presents the effect of additive dosage on the recovery of Fe and Cu and the volatilization of Zn and Pb. Without additives, the recovery rates of iron and copper are only 72.87% and 65.34%, respectively. Meanwhile, the iron grade of magnetic concentrate is 78.62%. Compared with no additives, in the presence of 20% additive, the recovery rates of iron and copper are 91.34% and 83.41%, respectively, and the grade of iron is 90.21%, and the content of copper is maintained at 0.34%. When the additives are 20% and 0%, the volatilization rates of zinc and lead are 98.89% and 95.33%, and 89.6% and 86.33%, respectively. Increasing the dosage to 25% has a slight effect on the indexes. From the above, it is recommended to add 20% as the best additive dosage. The XRD patterns of the roasted briquettes obtained under different additive dosages are presented in Figure 9. It can be seen from Figure 9 that during the reduction process, the peak intensity of iron is enhanced in the presence of additives, which is caused by the iron grain growing. Figure 10 shows the electron probe microanalysis (EPMA) of roasted briquettes with 20% additives. The roasted briquettes contain two parts,:one part is Fe-Cu alloy, and the other part is gangue. Most of the copper is combined with iron to form Fe-Cu alloy. Meanwhile, the boundary between the alloy and the gangue is obvious, which facilitates the recovery of iron and copper through the grinding process.  The XRD patterns of the roasted briquettes obtained under different additive dosages are presented in Figure 9. It can be seen from Figure 9 that during the reduction process, the peak intensity of iron is enhanced in the presence of additives, which is caused by the iron grain growing. Figure 10 shows the electron probe microanalysis (EPMA) of roasted briquettes with 20% additives. The roasted briquettes contain two parts,:one part is Fe-Cu alloy, and the other part is gangue. Most of the copper is combined with iron to form Fe-Cu alloy. Meanwhile, the boundary between the alloy and the gangue is obvious, which facilitates the recovery of iron and copper through the grinding process. The XRD patterns of the roasted briquettes obtained under different additive dosages are presented in Figure 9. It can be seen from Figure 9 that during the reduction process, the peak intensity of iron is enhanced in the presence of additives, which is caused by the iron grain growing. Figure 10 shows the electron probe microanalysis (EPMA) of roasted briquettes with 20% additives. The roasted briquettes contain two parts,:one part is Fe-Cu alloy, and the other part is gangue. Most of the copper is combined with iron to form Fe-Cu alloy. Meanwhile, the boundary between the alloy and the gangue is obvious, which facilitates the recovery of iron and copper through the grinding process.   Reduction was performed at 1250 °C for 80 min under the conditions of C/Fe mass ratio of 2, the basicity of 0.3 and additive dosages of 20%. The chemical compositions of the Fe-Cu alloy obtained at 20% additive are presented in Table 5. It can be seen that the Fe-Cu alloy contains 90.21% of TFe and 0.4% of Cu, and the contents of S and P are only 0.032% and 0.025%, respectively, which can be used as a burden for the weather resistance steelmaking. The results further confirm that the direct reduction-magnetic separation process is an effective means to upgrade the WCS. The chemical compositions of the direct reduction dust are shown in Table 6. The main compositions of dust are 65.17% ZnO and 2.66% PbO, which can be used as raw materials for extracting zinc. Many studies have been carried out on the comprehensive utilization of secondary zinc oxide [28,29]. Therefore, the treatment of reduction dust are not discussed in this paper.

Preparation of Portland Cement from Non-Magnetic Tailings
The chemical composition of non-magnetic tailings is listed in Table 7. The non-magnetic tailings include 55.64% SiO2 and 16.83% CaO, and the contents of Cu, Pb and Zn are 0.01%, 0.06% and 0.02%, respectively. According to the USEPA standard, the toxicity characteristics of the non-magnetic tailings were determined by the leaching procedure tests, and the results are shown in Table 8. The concentrations of Cu 2+ , Pb 2+ and Zn 2+ in the leachate are 0.021 mg/L, 0.001 mg/L and 0.046 mg/L, respectively, which are lower than the critical value, indicating that the non-magnetic tailings harmless and can be used directly as the burden for preparing cement. Table 7. Chemical compositions of non-magnetic tailings/wt.%. Reduction was performed at 1250 • C for 80 min under the conditions of C/Fe mass ratio of 2, the basicity of 0.3 and additive dosages of 20%. The chemical compositions of the Fe-Cu alloy obtained at 20% additive are presented in Table 5. It can be seen that the Fe-Cu alloy contains 90.21% of TFe and 0.4% of Cu, and the contents of S and P are only 0.032% and 0.025%, respectively, which can be used as a burden for the weather resistance steelmaking. The results further confirm that the direct reduction-magnetic separation process is an effective means to upgrade the WCS. The chemical compositions of the direct reduction dust are shown in Table 6. The main compositions of dust are 65.17% ZnO and 2.66% PbO, which can be used as raw materials for extracting zinc. Many studies have been carried out on the comprehensive utilization of secondary zinc oxide [28,29]. Therefore, the treatment of reduction dust are not discussed in this paper.

Preparation of Portland Cement from Non-Magnetic Tailings
The chemical composition of non-magnetic tailings is listed in Table 7. The nonmagnetic tailings include 55.64% SiO 2 and 16.83% CaO, and the contents of Cu, Pb and Zn are 0.01%, 0.06% and 0.02%, respectively. According to the USEPA standard, the toxicity characteristics of the non-magnetic tailings were determined by the leaching procedure tests, and the results are shown in Table 8. The concentrations of Cu 2+ , Pb 2+ and Zn 2+ in the leachate are 0.021 mg/L, 0.001 mg/L and 0.046 mg/L, respectively, which are lower than the critical value, indicating that the non-magnetic tailings harmless and can be used directly as the burden for preparing cement.  Figure 11 and Table 9 present the particle size distributions of non-magnetic tailings obtained from different ball-milling times. It can be seen that after ball milling the original sample for 1 h, 90% of the powder is less than 30 µm. When ball milling was extended to 3 h, the particle size changed slightly. When the ball milling time was further extended to 4 h, the powder agglomerated, and the particle size D90 increased. The specific surface area (SSA) increased from 1.03 g/cm 2 to 2.73 g/cm 2 . Considering the cost, it is recommended to use a ball milling of 2 h. The samples obtained from the ball milling for 2 h were applied as raw materials for cement manufacturing.   Figure 11 and Table 9 present the particle size distributions of non-magnetic tailings obtained from different ball-milling times. It can be seen that after ball milling the original sample for 1 h, 90% of the powder is less than 30 μm. When ball milling was extended to 3 h, the particle size changed slightly. When the ball milling time was further extended to 4 h, the powder agglomerated, and the particle size D90 increased. The specific surface area (SSA) increased from 1.03 g/cm 2 to 2.73 g/cm 2 . Considering the cost, it is recommended to use a ball milling of 2 h. The samples obtained from the ball milling for 2 h were applied as raw materials for cement manufacturing.   T 175-2007). They explored the most suitable mixing ratio of copper slag tailings to replace clinker. They believed that the proportion of copper slag tailings ranges from 6% to 15%. Figure 12 shows the changes of the compressive strength and breaking strength ver- Figure 11. Particle size distributions of non-magnetic tailings obtained from different ball milling times. According to the Chinese Standard GB/T 175-2007 (Standard for Common Portland cement), the minimum compressive strength of Portland cement at 3 and 28 days should be greater than 17 and 42.5 MPa, respectively (GB/T 175-2007). They explored the most suitable mixing ratio of copper slag tailings to replace clinker. They believed that the proportion of copper slag tailings ranges from 6% to 15%. Figure 12 shows the changes of the compressive strength and breaking strength versus the clinker content replacement by non-magnetic tailings at 3 and 28 days. It can be seen that the compressive strength of the Portland cement at 3 and 28 days increased from 24.1 to 29.70 MPa and 47 to 50.89 MPa, respectively, and the addition of non-magnetic tailings increased from 0% to 9%. If the amount of non-magnetic tailings is further increased, the compressive strength is decreased.

Balance of Elements in the Proposed Flow Chart
From the above research, we proposed a stepwise extraction process for the recovery of Cu, Fe, Zn and Pb, and the production of common Portland cement from non-magnetic tailings, as shown in Figure 13. It can be seen that the reason for the increase in the yield of green pellets is the addition of limestone and binder. However, through direct reduction, the sum of dust yield and reduced pellet yield is less than the yield of green pellets. This is because of the loss on ignition of limestone and flotation tailings. Several products were obtained from WCS, Cu concentrate for copper smelting, magnetic concentrate for weather resistance steelmaking and non-magnetic tailings for building construction. The proposed process can recover more than 90% iron and copper. Meanwhile, the recovery rates of Zn and Pb exceed over 90%. Moreover, the proposed process achieves zero tailing emissions. Therefore, WCS can be effectively utilized through this process.

Balance of Elements in the Proposed Flow Chart
From the above research, we proposed a stepwise extraction process for the recovery of Cu, Fe, Zn and Pb, and the production of common Portland cement from non-magnetic tailings, as shown in Figure 13. It can be seen that the reason for the increase in the yield of green pellets is the addition of limestone and binder. However, through direct reduction, the sum of dust yield and reduced pellet yield is less than the yield of green pellets. This is because of the loss on ignition of limestone and flotation tailings. Several products were obtained from WCS, Cu concentrate for copper smelting, magnetic concentrate for weather resistance steelmaking and non-magnetic tailings for building construction. The proposed process can recover more than 90% iron and copper. Meanwhile, the recovery rates of Zn and Pb exceed over 90%. Moreover, the proposed process achieves zero tailing emissions. Therefore, WCS can be effectively utilized through this process.