Innovative Methodology for Sulfur Release from Copper Tailings by the Oxidation Roasting Process

Key Laboratory of Ministry of Education for Solid Waste Treatment and Resource Recycle, Southwest University of Science and Technology, Mianyang, Sichuan 621010, China Institute of Mineral Materials and Applications, Southwest University of Science and Technology, Mianyang, Sichuan 621010, China City College, Southwest University of Science and Technology, Mianyang, Sichuan 621000, China Center of Forecasting and Analysis, Southwest University of Science and Technology, Mianyang, Sichuan 621010, China


Introduction
Copper tailings are solid wastes which are discharged in slurry after processing including crushing, grinding, sorting, and extracting target components from copper ores under specific economic and technical conditions [1]. In China, for every ton of copper mine, approximately 400 tons tailings were dumped. Subsequently, a large amount of copper tailings was stored up, not only encroached on a large number of land resources but also posed serious threats to the environment and also caused serious waste of resources [2]. Especially, for the copper tailings which contained metallic sulfide, it was easy to produce AMD which caused severe environmental pollution [3,4]. Meanwhile, sulfur in copper tailings also affected the grade of valuable metals extracted from copper tailings [5]. Additionally, it was possible for sulfur as a resource to be collected to make sulfuric acid which was the important application of sulfur [6,7]. erefore, it was necessary to release and recover sulfur from copper tailings. Innovative and cleaner methodology for recovery of S from copper tailings was increasingly needed to promote utilization of copper tailings more sustainable and respectful to ecological environment.
Generally, two methods to release sulfur from tailings rich in pyrite or pyrrhotite were involved, including flotation separation [8][9][10] and roasting [11]. e tailings rich in pyrrhotite had succeeded in desulfurization to less than 0.4 wt.% by optimizing the flotation process with respect to solution pH, activator, collector, frothier, and air-solid ratio [12]. Moreover, acid leaching and reverse flotation by using mixing barrel, flotation column, and flotation machine had also reduced the sulfur content of pyrite from 0.5% to 0.01% [13]. However, in the process of flotation, tedious step with some chemical reagents which were easy to bring secondary pollution was involved. Compared with flotation, roasting was more suitable for release and recovery of sulfur, including reductive roasting, sulphation roasting, and oxidation roasting. When the temperature exceeded 1430°C, hydrogen, as a reducing agent, could combine with sulfur from the surface of pyrite to form hydrogen sulfide [14]. Pyrite could also be translated to ferric sulfate by sulphation process. With temperature ranging from 1000°C to 1350°C, the sulfur component diffused and was released in gaseous SO x [15], where residual S in the roasted slag declined from 1.52% to 0.20% as a result of thermal decomposition of FeSO 4 . Moreover, pyrrhotite was formed by desulfurization during the thermal decomposition of pyrite. In an aerobic or carbon dioxide atmosphere, pyrrhotite could be oxidized to form iron oxides and gaseous SO x [16,17]. When the concentration of SO 2 in gaseous SO x reached or exceeded 3%, SO 2 could be collected together to produce sulfuric acid [18], and iron oxides could be recovered in subsequent processing. However, the change of the mineral phase, the release process, and mechanism of sulfur components during the roasting in the copper tailings had been rarely reported.
In this study, it involved resource utilization of sulfur from pyrrhotite in copper tailings. Based on the properties and AMD risks of copper tailings, oxidation roasting was proposed to release sulfur, and the sulfur release process, sulfur release rate, composition, and phase changes of the roasted slag were investigated. e objective of this research was to ascertain the sulfur release mechanism.

Materials.
e copper tailing samples (LW) used in this study were obtained from Liwu Copper, Sichuan, China. e collected samples were dried, crushed, and ground to 100 mesh. e samples were then mixed and stored for later use.

Chemical AMD Prediction Test.
To evaluate the AMD risks of copper tailings piled effectively, the net acid generation (NAG) test, which involved several parameters, including sulfur content of copper tailings, acid neutralization capacity (ANC), net acid-producing potential (NAPP), and net acid generation (NAG) [4], was performed firstly. e test of ANC was conducted as follows: 1 g ground sample was added to a 100 mL beaker containing 25 mL HCl (0.2 mol·L −1 ) at 90°C for 3 h. After being cooled, it was titrated with NaOH (0.2 mol·L −1 ) to pH 7, and the ANC was determined according to the following formula: where ANC is an index reflecting the sample's ability to neutralize acid by itself, kg H 2 SO 4 ·t −1 , and V is the volume of NaOH used in the titration process, mL. e NAPP was calculated as the following formula: where NAPP is a value which assessed the sample's net acid potential, kg H 2 SO 4 ·t −1 , and S is the sulfur content of the copper tailings, %. e test of NAG was conducted as follows: 2.5 g ground sample was added to a 500 ml conical flask containing 250 mL H 2 O 2 (15%). After being left in the fume hood for 24 h, the mixture was boiled for about 1 h to remove remaining H 2 O 2 . en, the mixture was cooled to room temperature, and it was titrated with NaOH (0.1 mol·L −1 ) to pH 7, and the NAG was calculated as the following formula: where NAG is a value of the sample's net acid generation, kg H 2 SO 4 ·t −1 , and V is the volume of NaOH used in the titration process, mL.
where APR is the acid potential ratio and S is the sulfur content of the copper tailings, %.

Oxidation
Roasting Test. Figure 1 shows the experiment equipment. e oxidation roasting test was carried out in a programmable energy-saving tubular furnace (LTKC-6-16A). From Figure 2, 20 ground samples (LW) were charged into a corundum crucible. Subsequently, the corundum crucible was put in a programmable energy-saving tubular furnace and heated at a temperature of 300°C-1200°C for a soaking time ranging from 20 min to 60 min. Meanwhile, the concentration of SO 2 was monitored at different temperatures in real time by a monitor of SO 2 . SO 2 could be collected to produce sulfuric acid. During the oxidation roasting, the heating rate was 10°C/min under 1000°C, and the heating rate was 5°C/min above 1000°C. In addition, the effect of the filling rate (0 L/min, 0.2 L/min, 2.4 L/min, 0.6 L/min, 0.8 L/min, and 1.0 L/min) was also investigated. After the oxidation roasting, the roasted slag was made natural cooling to room temperature, and then it was removed from the tubular furnace for further analysis. e content of residual S in the copper slag was determined by the combustion neutralization method [19], and the sulfur release rate (n) was calculated as the following formula: where M 1 and M 2 are the masses of samples before and after roasting, respectively, g and S 1 is the initial sulfur content of the copper tailings, while S 2 represents the residual sulfur content of the roasted slag, %.

Analytical Methods.
In this study, the mineralogical phase compositions of both raw tailings and roasted samples were tested by X-ray diffraction (XRD, X'Pert pro +3 KW, Netherlands), using Cu Kα radiation with a continuous scanning step of 0.02°in 2θ ranging from 3°to 80°. Subsequently, software X'Pert HighScore Plus was used for analyzing the diffraction patterns of XRD. e sulfur content of both raw tailings and roasted samples was determined using the combustion neutralization method. Moreover, the concentration of SO 2 in the roasting process was obtained continuously by a SO 2 monitor (SK-600) made in China.
Apart from sulfur, the contents of other chemical compositions in raw tailings were identified by the X-ray fluorescence spectrometer (XRF, Axios, Netherlands). ermogravimetry and differential scanning calorimetry analysis (TG-DSC) was observed by a thermal analyzer (NETZSCH STA 449F5) in the air atmosphere, heated from room temperature to 1500°C with a heating rate of 10°C min −1 .
In addition, HSC Chemistry Software 6.0 was used to work out basic thermodynamic calculations including the standard free energy (ΔG θ ).
ese calculations were employed to evaluate the possible reaction and explain the mechanism of sulfur release during oxidation roasting.

Characteristics and AMD Risk Assessment of Copper
Tailings. Table 1 shows chemical compositions of copper tailings, and Figure 3 shows phase compositions of the copper tailings. From Figure 3, the main mineral in the copper tailings was quartz, biotite, clinochlore, pyrrhotite, lepidocrocite, and sulfur. erefore, the sulfur minerals in copper tailings, including pyrrhotite and natural sulfur, could be released by oxidative roasting. Table 2 shows the chemical AMD prediction test results for the copper tailings. pH of the sample (LW) was 6.32, so both of solid phase solubility and the salt content would gradually increase. en, it was qualified to produce AMD [20]. At the same time, the content of sulfur in the sample (LW) was relatively high. With oxygen and water, it was easy to form AMD due to the oxidation of sulfide minerals. From Table 2, ANC was smaller than NAPP and NAG, indicating that the copper tailings themselves could not effectively neutralize the acid generated by the oxidation of sulfide in some parts, which was exactly explained by the positive NAPP [21]. On the basis of classifications listed in Table 2, with NAPP > 20 and APR < 1, AMD was formed easily in the copper tailings by oxidation of sulfide minerals, and AMD was considered a delayed (chemical time bomb) and persistent geochemical disaster [22]. erefore, to remove or recover sulfur from the copper tailings was of great importance to not only control tailing pollution but also to realize recycling and utilization of the resource.

Phase Transformation, ermal Effect, and
Decomposition Kinetics 3.2.1. Phase Transformation of the Roasted Slag. Figure 4 shows the XRD patterns of the roasted slag at different roasting temperatures. When the temperature was 300°C,  the main phase of the roasting slag was quartz, biotite, pyrrhotite, and chlorite, indicating that no thermal decomposition reaction occurred. Compared with the raw material, the pattern of lepidocrocite and natural sulfur disappeared. With the increasing temperature, some pyrrhotite decomposed to form FeS and S. Subsequently, FeS was oxidized to form FeSO 4 and Fe 2 O 3 . When the temperature reached 600°C, quartz, biotite, clinochlore, pyrrhotite, and hematite existed in the roasted slag. However, the peak strength of clinochlore and pyrrhotite was reduced to a certain level, and a new phase (FeSO 4 ) was present. As the temperature continued to rise, FeSO 4 began to decompose to hematite and SO 2 . At 900°C, the main phases in the roasted slag were quartz and biotite, and hematite and anorthite formed. At 1100°C, the phase of the roasted slag did not change in kind, but the peak strength of biotite decreased gradually. Under the optimal roasting temperature of 1200°C, the main phases in the roasting slag were quartz, hematite, and anorthite.

3.2.2.
ermal Effect Change. In order to confirm the characteristics of the copper tailings during oxidation roasting and explain the path of the sulfur liberation reaction of pyrrhotite indirectly, the TG-DSC curves of copper tailings (LW) are performed in Figure 5. On the TG curve, the weight loss was about 6.48% from room temperature to 344°C, where the weight loss was about 0.72% as a result of loss of adsorbed water from room temperature to 110°C. From 110°C to 344°C, a small amount of natural sulfur sublimates emerged, and lepidocrocite lost crystallization water, with a weight loss of about 5.76%. Over the temperature range of 344°C-428°C, the mass of samples increased about 0.59%, which was related to the formation of FeSO 4 and Fe 2 O 3 from the oxidation decomposition of pyrrhotite [23,24]. e weight loss between 428°C and 600°C was 1.43%, which was in relation to the loss of interlayer structure water (528°C±) of clinochlore and sulfur release in pyrrhotite decomposition (550°C±) [25]. In the range of 600°C-900°C, FeSO 4 , which was formed by the decomposition and oxidation of pyrrhotite, was further decomposed to release SO 2 , with a weight loss of 1.75%. en, an obvious mass loss (13.81%) from 900°C to 1200°C was presented in relation to the removal of structural water from biotite and the loss of sulfur in the residual structure.
On the DSC curve, two endothermic peaks (251.8°C and 271.9°C) appeared, which corresponded to the endothermic effect, including desorption of crystallization water of lepidocrocite and phase change of pyrrhotite. e exothermic peaks (1003.8°C and 1155.3°C) were caused by the crystallization exothermic effects of calcium feldspar and hematite, respectively.

ermodynamic Analysis.
Under the condition of air, pyrrhotite was oxidized and decomposed to form SO 2 . e possible reactions involved during oxidation roasting of copper tailings are listed as follows: Figure 6 shows the standard free energy (ΔG θ ) changes for equations (6)-(9) over the temperature ranging from 0°C to 1250°C. ΔG θ for equations (6), (7), and (9) was negative over the temperature ranging from 0°C to 1200°C, suggesting these reactions were possible from the view of thermodynamics. Moreover, ΔG θ for equation (8) became smaller with the increasing temperature, even became negative after 720°C, indicating that FeSO 4 could be decomposed to metal oxides and SO 2 .
Take equation (6) for example: Meanwhile, ΔG θ could also be expressed as follows:  where K θ is the equilibrium constant. K θ is shown as rough equations (10)- (12), where a Fe 2 O 3 is the activity of Fe 3 O 4 in the roasted slag; a FeS is the activity of FeS in the roasted slag; P SO2 is the partial pressure of SO 2 generated during the oxidation roasting; and P O2 is the partial pressure of O 2 present during the oxidation roasting. When the thermodynamic temperature T, P O2 , P SO2 was determined, the relationship between a Fe 2 O 3 and a FeS could be expressed correctly by equations (10)-(13) and could be calculated by equation (13), where the parabolic relationship between a Fe 2 O 3 and a FeS was observed. In the case where the temperature was 1000°C and P SO2 was 1 atm simultaneously, the value of e − G θ and 1/P SO 2 was constant. en, equation (13) turned into equation (14), and the results are present in Figure 7.
where t is the constant, and t � e − ΔG θ × (1/P SO 2 ). As shown in Figure 7, it was indicated that, with an increase in P O2 , a Fe 2 O 3 increased rapidly, which meant FeS was oxidized to Fe 2 O 3 and SO 2 , respectively. As a consequence, oxygen added in roasting could promote sulfur release.

Sulfur Release Rate and Residual S Content during
Oxidation Roasting

Effect of the Roasting
Temperature. e effect of roasting temperature on sulfur release rate and residual S content was investigated with temperature ranging from 300°C to 1200°C over a roasting time of 30 min at an air flow � 1.0 L/min. e results are presented in Figure 8. As illustrated in Figure 6, with an increase in the roasting temperature which was beneficial to thermal decomposition of sulfide minerals, the sulfur release rate and the residual S content were observed from 13.60% to 99.45% and from 23.70% to 0.15%, respectively. Corresponding to the above thermal analysis, it released sulfur vapor slowly before 600°C on account of pyrrhotite considered as solid solution which caused the sulfur to leave the particle very slowly. Moreover, sulfur began to be released as SO 2 at 600°C [15]. When the temperature further increased, the sulfur vapor diffused through the Fe 1−x S layer quickly and could combine with oxygen to form SO 2 . However, the sulfur release rate and the residual S content were steady at 1200°C, indicating that a roasting temperature of 1200°C was recommended for the following tests.     Figure 9 shows the effect of residence time on sulfur release rate and residual S content. From Figure 9, by increasing the duration time from 20 min to 60 min, the sulfur release rate and the residual S content were observed from 96.32% to 99.78% and from 1.01% to 0.06%, respectively. When the duration time was over 60 min, the sample could be roasted excessively to form the liquid phase, which prevented SO 2 from escaping in the oxidation process [26,27]. Hence, the optimal duration was determined as 60 min to ensure proper sulfur fugacity.

Effect of Air
Flow. e influence of air flow on sulfur release rate and residual S content is illustrated in Figure 10, indicating that, with an increase in an air flow from Standard free energy ∆G θ (kJ·mol -1 ) Equation (7) Equation (6) Equation (9) Equation (    Journal of Chemistry 0 L/min to 1 L/min, the sulfur release rate and the residual S content were observed from 59.53% to 99.82% and from 11.10% to 0.05%, respectively. When the sample weight was fixed, oxygen as a reactant not only affected the concentration of SO 2 but also played an important part in the reaction. With an increase in the air flow, the sulfur release rate and the residual S content gradually became steady. erefore, an optimum air flow was selected as 0.8 L/min. Surprisingly, when air flow was 0 L/min, sulfur in the sample was released as sulfur vapor rather than SO 2. e results are observed in Figures 11(a) and 11(b). It was suggested that pyrrhotite lost sulfur firstly and the sulfur vapor diffused upward. en, sulfur vapor was cooled and was deposited on the top layer of the sample. In order to confirm the yellow material deposited on the surface of the sample, the XRD was analyzed as shown in Figure 11(c). From Figure 11(c), the yellow material on the layer of the sample was S (d 222 � 3.8540Å, d 206 � 3.2190Å, and d 026 � 3.4470Å). Moreover, the SEM image of S is illustrated in Figure 11(d) which showed that the yellow powder had a fluffy structure. e release pattern of SO 2 is presented in Figure 12. From Figure 12, the presence of SO 2 began when the temperature arrived 600°C, and it reached the maximum concentration when the temperature was 1168°C. Interestingly, the distribution of SO 2 concentration was parabolic, and the concentration of SO 2 in the flue gas was 0∼8.26%. Hence, SO 2 could be collected for sulfuric acid production through purification of SO 2 , conversion of SO 2 , and absorption of SO 3 when it met the requirements that the concentration of SO 2 in the gas reached 3% [28,29].

e Characterization of the Roasted Slag.
e chemical composition of the roasted slag is shown in Table 3. Compared with raw materials, more Fe 2 O 3 and SiO 2 in the roasted slag existed. After sulfur was released in the form of SO 2 , the total mass of the sample decreased, and the percentage of Fe 2 O 3 increased obviously. Moreover, the residual S content of the tailings was 0.05%. According to the quality requirement of the ore suitable for steelmaking, the tailings with more than 55% TFe and with less than 1% S could be used in iron smelting [25].

Sulfur Release Mechanism and Phase Transition of
Pyrrhotite. Based on the previous discussion, the mechanism of sulfur release from pyrrhotite could be proposed in Figure 13. e possible process is shown in the following four steps: the oxidation decomposition of pyrrhotite, the formation of ferric sulfate, the decomposition of ferric sulfate, and the formation of hematite. During roasting, SO 2 was partly released after oxidation decomposition of pyrrhotite, and the sulfur release rate increased with the increase of temperature. Under the condition of sufficient air, pyrrhotite was directly oxidized to hematite when it was heated. With air supplied insufficiently, pyrrhotite was partially oxidized to hematite and partially oxidized to ferric sulfate. en, ferric sulfate was dissolved into hematite at higher temperatures.
In the oxidation process of pyrrhotite, sulfur from the middle layer to the surface was oxidized and was escaped as SO 2 , while hematite was formed in situ. e oxidation process of pyrrhotite was carried out from the surface layer to the inner layer with the increase of temperature and time. Finally, pyrrhotite was completely oxidized and decomposed, and hematite was generated in the roasted slag which could be used for iron recovery in the next step.

Conclusion
In this study, an innovative method was developed for release of S from copper tailings. e main findings from this work are concluded as follows: (1) e main metal sulfide mineral in the copper tailings was pyrrhotite. e contents of Fe-total and S-total reached 36.35% and 27.82%, respectively. erefore, S and Fe could be separated and recovered step by step successively where S release was necessary and was realized by oxidation roasting firstly. (2) e AMD risk assessment showed that the copper tailings had acid-generating potential, and the NAPP was tested as 634.43 kg H 2 SO 4 t −1 . e AMD would pose a severe threat to human health and ecological environment.

Journal of Chemistry
(3) e optimum conditions for sulfur release were obtained with a temperature of 1200°C, a residence time of 60 min, and an air flow of 0.8 L/min. Under these conditions, the sulfur release rate and the residual S content were 99.82% and 0.05%, respectively. (4) In the process of oxidizing roasting, pyrrhotite was oxidized and decomposed to form ferric sulfate, hematite, and sulfur dioxide. With the increasing temperature and time, ferric sulfate decomposed to hematite and SO 2 . (5) e concentration of SO 2 released exceeds 3% and could meet the requirements of sulfuric acid production. Moreover, hematite with magnetic properties and quartz mainly existed in the roasted slag, suggesting that hematite could be separated and recovered by magnetic separation in the next step.
Data Availability e data included the chemical composition of the tailings, the content of sulfur in the tailings, and the phases of the tailings under different conditions. Moreover, all data graphs in the manuscript were drawn with Origin software, and the original format of the graph was retained in the text.

Conflicts of Interest
e authors declare that they have no conflicts of interest.