Correlation between Thermodynamic Studies and Experimental Process for Roasting Cobalt-Bearing Pyrite

Abstract

Cobalt is a critical metal widely distributed in nature, but cobalt ore has hardly been found as an independent mineral. Cobalt-bearing pyrite tailings separated from iron ore is one of the resources for recovering cobalt. In the following study, roasting is carried out to oxidize cobalt-bearing pyrite tailings for preparing and recovering the cobalt by acid leaching. The further aim of the research is to determine and control the optimal technological regime for roasting by using thermodynamic modeling. The phase transition in Fe–S–O and Co–S–O systems and its mechanism are analyzed under the partial pressure of oxygen and sulfur dioxide at constant temperatures. Thermodynamic modeling proves that iron and cobalt sulfides can be intensively oxidized at a relatively high temperature (>900 °C) under an atmosphere of logp(O₂) > −5, leading to the formation of SO₂ (logp(SO₂) < 0). The results of the roasting experiment indicate 98% desulfurization degree upon holding for about 4–5 h and at > 1000 °C. Based on these thermodynamic modeling and experimental results, the roasting of cobalt containing pyrite can be optimized with substantial productivity with regard to the metal oxide and cobalt thereof. Oxidative roasting also allows the elimination of environmentally hazardous gases such as sulfur during the process.

1. Introduction

Cobalt is one of the critical metals classified by the European Union and used in several industrial, medical, and military applications due to its unique properties such as high density, hardness, resistance to corrosion, and good thermal conductivity [1,2,3]. Cobalt alloys are used to make super alloys that are durable against heat and corrosion and can sustain the high temperature needed for welding properties. In addition, cobalt compounds such as cobalt oxide, chloride, and sulfate are widely used in industrial applications. Recently, about 50% of cobalt produced worldwide is being utilized as cathode materials in lithium-ion or rechargeable batteries [4,5,6]. As the demand for cobalt has increased manifolds due to the growth of electric vehicles and other consumer electronics, to maintain a surplus, other sources of it, for example, urban mines and cobalt-bearing secondary resources, are being explored currently [7,8,9]. Secondary resources that contain cobalt include by-products that are separated from primary cobalt-bearing ores, as well as slags in which cobalt is dissolved during the smelting process and flotation tailings. Cobalt minerals are typically found in association with nickel, copper, zinc, and iron ores, which are known as their accessory elements [10,11]. Usually, cobalt is processed from a byproduct generated during the mining of base metal ores containing nickel and copper [12,13,14]. Among the approximately 30 cobalt-bearing minerals, the most common ones are classified as sulfides, arsenide, and sulphosalts existing in the deposits of magmatic Ni-Co sulfide, stratiform sediment-hosted Cu-Co, and laterite ore that are part of the ultramafic rocks with a high Ni and Co content [15,16].

Skarns, or tactites, are coarse-grained metamorphic rocks that contain Fe, Au, Cu, Zn, W, Mo, and Sn [17,18]. In Fe skarn, the magnetite with 40–50% Fe has around 0.007–0.028% Co combined with its minerals such as pyrite, pyrrhotite, and arsenide [19,20]. In the Tumur Tolgoi and Tumurtei regions of the Bayan Gol zone, Mongolia, magnetite iron ores were located, in which minerals such as martite, hematite, pyrrhotite, and pyrite were present. These high-grade iron deposits are believed to have 22–65% iron ores (a total of about 25–280 Mt), 1–4% sulfur, and 0.02–0.03% cobalt [21]. From these iron deposits, magnetite iron ore can be generated by employing a single magnetic separation process. Cobalt-bearing high-grade magnetite ores, as resources for 28% cobalt, have been discovered in China as well [20,22,23]. Two methods, namely flotation and hydrometallurgical, are followed for mineral processing as well as for recovering cobalt from tailings or industrial waste. Bethlehem Steel in Pennsylvania, US, applied the flotation method in order to produce iron ore and byproducts of cobalt, making it the largest domestic producer of cobalt [24]. Although high-grade iron concentrate with low sulfur and cobalt-bearing pyrite byproducts can be generated by the flotation-magnetic separation method. Many studies have also suggested the recovery of cobalt (95% recovery rate) from these byproducts or tailings by the leaching process after roasting [3,25,26]. Cobalt exhibits stability and does not undergo any reactions with the air at normal temperatures. However, if heated, it will undergo oxidation, first forming Co₃O₄ and then transitioning to CoO once the temperature reaches approximately 900 °C. Acid leaching is a widely adopted technique for extracting cobalt, which is typically performed after oxidative roasting. The presence of cobalt in two valence states, Co²⁺ and Co³⁺, leads to noticeable variations in its physical and chemical characteristics when it is in a solution. In the context of solubility in an acid solution, the divalent ion (Co²⁺) exhibits higher solubility compared to the trivalent ion (Co³⁺) [12,27,28].

As the price and demand of cobalt rise, the pyrite tailings with 0.2–0.3% Co are considered suitable sources that can be subjected to the above methods at a low cost in order to recover the metal. Recently, Darkhan Metallurgical Plant, Mongolia, has investigated the extraction of high-grade iron concentrate with low sulfur and cobalt-bearing pyrite tailings from the iron magnetite ore by combining flotation and magnetic separation techniques. As per their studies, 1% Co from the iron ore can be recovered by using the above method. It has also been reported that the flotation of pyritic tailings containing about 1% Co from the iron ore can be achieved by using a flotation-magnetic separation. In our following studies, we propose a thermodynamic model for the oxidative roasting and desulfurization rates of cobalt-bearing tailings. Theoretical findings from thermodynamic models were also correlated with the experimental results in order to suggest an improved, effective, and low-cost process for recovering cobalt from the byproducts generated during the treatment of iron magnetite ore in an industry. The roasted tailing or product, via this optimized process, can be further subjected to acid leaching for easy meal (cobalt) recovery. Scheme 1 shows a production process for recovering cobalt from pyrite tailings. The tailing, or byproduct, is produced in the magnet-flotation separation step of an iron ore and is subjected to roasting under the atmosphere of O₂ in a successive step, releasing SO₂ and forming the oxides of metals (cobalt). The roasted tailing is almost free from sulfur content and, upon acid leaching, can be used for recovering cobalt at a low cost. Theoretical modeling and experimental results for the optimized roasting process can be employed in industry for the same purpose. This process is calibrated and optimized as per the thermodynamic models and their correlation with the experimental data to analyze the effect of heat and partial pressure of gases such as oxygen and sulfur dioxide on the phase stability of Fe–S–O and Co–S–O systems, which in turn can impact the roasting process and the release of cobalt thereof in a successive step. Additionally, the sulfur from the ore is eliminated in the form of SO₂, thereby reducing the release of sulfur gas into the environment and allowing safe roasting. The reaction-time dependent rate of desulfurization during the roasting of pyrite tailings was also monitored so that the proposed production process here can be implemented in industry as well for the recovery of cobalt from the byproducts or tailings.

2. Materials and Methods

2.1. Experimental Materials

The cobalt-bearing tailing used in the study is pyrite tailings extracted from the iron ore of the Tumurtolgoi deposit in Mongolia. The flotation experiment was performed at the Central Geological Laboratory in Mongolia. The chemical composition of the c cobalt-bearing tailing was analyzed by an X-ray fluorescence spectrometer (XRF), which was a Shimadzu XRF-1800 X-ray fluorescence spectrometer manufactured in Tokyo, Japan by Shimadzu Corporation. The sample was placed inside a platinum crucible and loaded into an XRF machine, where the beam was loaded into the sample holder, after which, the sample was ionized by x-rays under vacuum medium.

Table 1. Chemical composition of cobalt-bearing tailing.

Oxides	Fe2O3	MgO	SiO2	Al2O3	CaO	MnO	TiO2	SO3	Co	S
wt.%	67.12	13.67	6.66	1.32	0.69	0.26	0.14	10.12	1.28	8.73

shows the result of XRF analysis of the particles of the cobalt-bearing tailing, which indicated that the tailing consists of about 67 wt.% Fe₂O₃, 13 wt.% MgO, 6 wt.% SiO₂, and others. By utilizing the XRF analysis method, one can exclusively measure the concentration of elements and receive the results in the widely recognized oxide form. In addition, it is important to note that XRF analysis lacks the ability to distinguish between various oxidation states of elements. However, the search results only display information for the most dominant oxide form. Iron (Fe) can exist in various oxidation states, including FeO, Fe₃O₄, and Fe₂O₃. It should be noted, however, that the Fe K-alpha peak is specifically observed in the Fe₂O₃ state, as mentioned in reference [29]. The cobalt-bearing tailing contains 1.28 wt.% cobalt and 8.73 wt.% sulfur, as analyzed by ICP–OES.

Mineralogical analysis of the cobalt-bearing tailing was determined by X-ray diffractometer, as the model of X’Pert-MPD PANalytical manufactured in Malvern, UK by PANalytical, has a 3 kw Cu-Kα X-ray tube. Figure 1 shows the pattern of the XRD analysis of the cobalt-bearing tailing. XRD data collection of the cobalt-bearing tailing was conducted with a scanning range of 10 to 80° and a time of 10 min with a step size of 0.02°. Although peaks peculiar to magnetite (Fe₃O₄) and quartz (SiO₂) were indicated, some peaks of pyrite (FeS₂) were observed. The mean particle size of the cobalt-bearing tailing was obtained using a laser particle size analyzer Microtrac S3500. This instrument was manufactured by MicrotracBel in Osaka, Japan. The result in Figure 2 demonstrates that the particles have an irregular shape, and the 50-pct passing size was about 78.65 μm. Measurements were realized in triplicate at a 90° angle at 25 °C under refractive index 1.81.

Utilizing the model of EM-30AX manufactured in Daejeon, Republic of Korea by COXEM, the Scanning Electron Microscope and Energy Disperse X-ray Spectrometer (SEM-EDS) were employed to conduct a detailed analysis of the morphology and distribution of elements in the cobalt-bearing tailing. Figure 3 provides a visual representation of the tailing particles, illustrating their shape, size, and distribution of elements. The particle size distribution of the tailing, as depicted in Figure 3a, exhibits irregularity and a range of different sizes. The surface of the particles, which consist of relatively large grains, has a rough texture resembling the aggregation of multiple small particles. The small particles in question appear as if they have been broken off from a larger entity, and their surfaces are characterized by smooth, sharp-edged crystals. Based on the distribution of elements shown in Figure 3b, it has been determined that iron is the predominant element across the entire area. Alongside iron, there are also elements such as magnesium, silicon, sulfur, and calcium that coexist.

The EDS mapping analysis results are presented in

Table 2. Chemical composition of cobalt-bearing tailing by EDS analysis.

Elements	Fe	O	Si	Mg	Ca	S
wt.%	60.34	24.21	4.09	6.43	0.90	4.03

. The distribution of iron was found to be present in all particles, with a content percentage of 60.34 wt.%. The oxygen content in this sample is measured at 24.21 wt.%, which suggests that the particles are predominantly present in the form of oxides. The impurities found in the ore include silicon, magnesium, and calcium. The results obtained from EDS mapping demonstrated identical findings to those obtained from XRF analysis, which determined the presence of the dominant impurity elements MgO and SiO₂. Additionally, the EDS mapping analysis provided insight into the composition of the cobalt-bearing tailing, showing a sulfur content of 4.03 wt.%, which indicates a significant presence of sulfur in this material.

2.2. Experimental Procedures

The roasting experiments were carried out in a horizontal tube heated by silicon carbide (SiC) elements at 900–1100 °C (Figure 4). The tube furnace is a horizontal tube consisting of a furnace controller, a heating box with a B-type thermocouple, and a horizontal alumina tube (working place) with a 75-mm diameter and 900-mm length. The tube furnace can heat the sample to a maximum temperature of 1600 °C and can be sealed to avoid infiltration by air. The sample was put in an alumina boat, which was placed in the center of the working place. A controlled flow of oxygen, argon, nitrogen, and carbon monoxide can be supplied to the furnace for the experiment. The operational temperature was monitored by a C type thermocouple (error of less than ±0.5% of °C) attached to the furnace center zone. The furnace is equipped with a temperature-controlling system and programmable segments for precise control of heating rate, cooling rate, holding, and time setting. An over-temperature controller and alarm are also mounted in the furnace for continuous operation without the need for an attendant. The experimental gas based on the oxygen partial pressure was controlled by a mass flow controller (MFC), injected from one side of the furnace through a working place, and supplied at a flow rate of 300 cc/min, ±2 cc.

2.3. Analytical Methods

2.3.1. Thermodynamic Modeling

HSC Chemistry empowers researchers to investigate the effects of a multitude of variables on chemical systems at equilibrium through its advanced computational capabilities. With a database containing information on over 17,000 chemical compounds, the program relies on thermodynamic data such as enthalpy, entropy, and heat capacity to carry out its calculations. The focus of this study was to analyze the stability areas of phases in the Fe–S–O and Co–S–O systems. The use of the HSC Chemistry database allowed for a comprehensive examination of how temperature, O₂ partial pressure, and SO₂ partial pressure influenced the calculation. Furthermore, calculations were conducted to determine the most suitable modeling for the transition from iron and cobalt sulfides to their respective oxides. By employing the Tpp diagram module, a temperature partial pressure diagram (T-p-diagram) was generated, enabling the determination of the isothermal predominance area in Fe–S–O and Co–S–O ternary systems by considering partial pressures on both axes. Additionally, the Lpp module was used to create diagrams that illustrate the specific partial pressures of O₂ and SO₂ on both axes. The Equilibrium module was utilized to perform calculations on the activity of multi-component equilibrium compositions found in cobalt-bearing tailing. The module is specifically designed to handle calculations of product amounts at equilibrium, taking into account both isothermal and constant pressure conditions.

2.3.2. Calculating the Desulfurization Rate

The evaluation of how efficiently sulfur was removed from cobalt-bearing tailings through roasting relied on the desulfurization rate. A comparison was conducted between the sulfur content in the raw material before roasting and the sulfur content in the roasted tailing after the roasting process. The desulfurization rate is represented by Equation (1).

$$\mathsf{\eta }=\frac{({\mathrm{S}}_0-{\mathrm{S}}_{\mathrm{i}})}{{\mathrm{S}}_0}\times 100\mathrm{\%}$$

(1)

where: η is the desulfurization rate, %; S₀ is the sulfur content in cobalt-bearing tailing before roasting, wt.%, S_i is the sulfur content in cobalt-bearing tailing after roasting, wt.%.

3. Results and Discussion

3.1. Thermodynamic Modeling for Optimal Roasting

The oxygen partial pressure, or concentration of oxygen, during roasting has a great impact on the phase transformations of pyrite. Phase transformation of metal sulfides (MeS₂) during the oxidation of pyrite by blowing oxygen at high-temperatures can lead to the formation of various phases such as sulfides (Me_xS/MeS), oxides (Me_mO_n/MeO), and sulfates (MeSO₄/Me_x(SO4)_x−1). As per thermodynamic principles, a phase transition determines how the physical process of transition between the metal sulfide state and oxide state occurs at a certain temperature and oxygen partial pressure. Figure 5 shows how the phase transition occurs in Fe–S–O and Co–S–O systems under the function of the partial pressure of oxygen depending on the temperature using the HSC Chemistry 6.2 software database. When logp(O₂) is below −30, the phase transitions of iron and cobalt sulfides take place in the order of FeS₂→Fe₂S₃→FeS→Fe and CoS₂→CoS_1.333→CoS_0.89→Co at a temperature up to 1100 °C (Figure 5a,b). However, when oxygen partial pressure increases above logp(O₂) ≥ −25, phase transitions of iron and cobalt sulfides occur as FeS₂→FeO→Fe₃O₄→Fe₂O₃ and CoS₂→CoS_1.33→CoO at 750 and 450 °C, respectively. Under an atmosphere of logp(O₂) ≥ 0, the phase transition of iron and cobalt follows as FeS₂→Fe₂O₃ at a temperature above 850 °C and CoS₂→CoO at 900 °C. It appears that intensive oxidative roasting of iron and cobalt sulfides in Fe–S–O and Co–S–O systems can be carried out in an oxygen atmosphere with logp(O₂) ≥ −5 at a temperature above 900 °C.

The phase stability of iron and cobalt sulfides was also analyzed in the form of the Fe–S–O and Co–S–O systems to determine how the phase transition occurs under the partial pressure of oxygen and sulfur dioxide at a constant temperature of 900, 1000, and 1100 °C (Figure 6a,b). The phase stability diagrams use logarithmic scales for partial pressures, which are expressed in bar units. As shown in Figure 6, a straight line at 900 °C, a dashed line at 1000 °C, and a dotted line at 1100 °C are shown, respectively. The transition of the iron sulfide phase moves to the right in the order of FeS₂→Fe_0.877S→FeS→Fe→Fe_0.975O→Fe₃O₄→Fe₂O₃ with the increase of oxygen partial pressure (logp(O₂)) at 900, 1000, and 1100 °C (Figure 6a). If the partial pressure of SO₂ (logp(SO₂) in the system is above zero (≥0), the phase transition of iron sulfide happens as FeS₂→FeSO₄→Fe₂(SO₄)₃. However, the partial pressure of SO₂ in the system is low as the gas is removed from the furnace during the roasting. When logp(O₂) is higher than −10 with logp(SO₂) ≤ 0, the phase transition of iron sulfide transfers towards FeS₂→Fe₃O₄, and FeS₂→Fe₂O₃ when logp(O₂) is above 2. For cobalt complexes, the phase transition follows the following order: CoS₂→CoS_1.33→CoS→CoS_0.89→Co→CoO→Co₃O₄ with the increase of logp(O₂) at 900, 1000, and 1100 °C (Figure 6b). The CoS₂ phase changes to CoO and Co₃O₄ phase when the logp(O₂) is above −10 and logp(O₂) ≥ 0, respectively. It must be noticed that as the temperature increases, the activity of the oxidation of sulfide rises, as can be seen from the phase stability diagram.

Sulfur in the cobalt-bearing tailing reacts with the oxygen in different possible pathways, as represented by Equations (2)–(7) and (8)–(12). In oxidative roasting, magnetite (Fe₃O₄) can be completely oxidized to hematite at high temperatures, which is illustrated by the phase stability diagram in the Fe–S–O system. MgO and CaO oxides existing in the cobalt-bearing tailing can form sulfates of MgSO₄ and CaSO₄ according to the following Equations (13) and (14).

FeS₂ + O₂(g) = FeS + SO₂(g)

(2)

2FeS₂ + 5.5O₂(g) = Fe₂O₃ + 4SO₂(g)

(3)

3FeS₂ + 8O₂(g) = Fe₃O₄ + 6SO₂(g)

(4)

Fe₂S₃ + 6.5O₂(g) = Fe₃O₄ + 4.5SO₂(g)

(5)

3FeS + 5O₂(g) = Fe₃O₄ + 3SO₂(g)

(6)

2FeS + 3.5O₂(g) = Fe₂O₃ + 2SO₂(g)

(7)

4Fe₃O₄ + O₂(g) = 6Fe₂O₃

(8)

CoS₂ + O₂(g) = CoS + SO₂(g)

(9)

3CoS + 5O₂ (g) = Co₃O₄ + 3SO₂(g)

(10)

CoS + 1.5O₂ (g) = CoO + SO₂(g)

(11)

2CoS + 4FeS + 10O₂(g) = 2CoO∙Fe₂O₃ + 6SO₂(g)

(12)

2CaO + O₂(g) + 2SO₂(g) = 2CaSO₄

(13)

2MgO + O₂(g) + 2SO₂(g) = 2MgSO₄

(14)

The thermodynamic feasibility of the reactions and their product stability were analyzed by Gibbs free energies (∆G) based on HSC Chemistry 6.2 data at temperatures up to 1100 °C. In thermodynamic analysis, ∆G > 0 indicates that a chemical reaction is not possible to occur, whereas if ∆G < 0 or is more negative, that chemical reaction can spontaneously take place [30]. As shown in Figure 7, the ∆G value for reactions 2 to 6, 9, and 11 is calculated as ∆G < −1000 $\mathrm{k}\mathrm{J}{\mathrm{m}\mathrm{o}\mathrm{l}}^{-1}$ at a given temperature, meaning that the reactions tend to take place rapidly. Also, reactions 12 and 13 can proceed (∆G < −500 $\mathrm{k}\mathrm{J}{\mathrm{m}\mathrm{o}\mathrm{l}}^{-1}$) at elevated temperatures (900 °C and below). However, when overheating is carried out above 900 °C, ∆G value increases indicating that the reaction of sulfates is thermodynamically difficult to proceed with. The Gibbs free energy values of cobalt oxides are $∆{\mathrm{G}}_{\left(\mathrm{C}{\mathrm{o}}_3{\mathrm{O}}_4\right)}<-1200 \mathrm{k}\mathrm{J}{\mathrm{m}\mathrm{o}\mathrm{l}}^{-1}$, and $∆{\mathrm{G}}_{\left(\mathrm{C}\mathrm{o}\mathrm{O}\right)}<-50 \mathrm{k}\mathrm{J}{\mathrm{m}\mathrm{o}\mathrm{l}}^{-1}$ which are slightly negative even at high temperatures, indicating that cobalt oxidation is likely to occur. At temperatures of 900, 1000, and 1100 °C (dotted line in Figure 7), ∆G values for the reactions listed as 1 to 13 were ∆G < 0 pointing towards the thermodynamically favorable pathways.

In order to accurately evaluate the compositional variation of the cobalt-bearing tailing and the effect of composition on the stability of a particular end component during reaction, the activity (a_i) of component i, one of the thermodynamic properties, should be considered [31]. In principle, a thermodynamic activity is a value that indicates how reactive the end component is in a raw material composition. The analysis of the activity of a specific component, which was formed during the roasting of the tailing, was conducted using the HSC Chemistry 6.2 database. This analysis was performed by considering the function of temperature at a 1 atm oxygen partial pressure, as shown in Figure 8. HSC Chemistry utilized the data from the XRF analysis of the tailings to calculate the activity of the products in roasting, taking into account the quantity of components. The main objective of conducting this analysis was to make predictions about the potential formation of products in the system during roasting. Specifically, the calculation focused on the target components, namely Fe, Co, and S, as well as the impurity components, namely Mg and Si. Additionally, it aimed to determine the possible compounds that could be formed between these components. With the increase in roasting temperature, there was a gradual decrease in the activity of the FeS₂ and CoS₂ phases, and it was observed that the activity sharply declined when temperatures exceeded 500 °C. However, it is important to note that there was a significant increase in the activity of SO₂ gas, reaching a value of 1. The activity of FeS and CoS sulfides showed an increase between the temperature range of 500 and 900 °C, followed by a decrease above 900 °C, and eventually reaching zero activity at approximately 1100 °C. According to the calculation data, it was found that the sulfur content in the Fe–S and Co–S systems was entirely transformed into the gas phase of SO₂ when the temperature was between 900 and 1100 °C. As the temperature rises above 600 °C, it becomes evident that the activity of the Fe₃O₄ phase declines, whereas, in contrast, the activity of the Fe₂O₃ phase rises. Through the process of roasting, iron magnetite oxide undergoes a transformation and is confirmed to be converted into iron hematite oxide. The process of roasting involved the interaction of FeO, MgO, and SiO₂ oxides, leading to the formation of new compounds such as Fe₂SiO₄ and MgFe₂O₄. When considering temperatures between 900 and 1100 °C, it can be concluded that the Fe₂SiO₄ compound exhibits a lower activity level in comparison to the MgFe₂O₄ compound. Even though the FeO oxide activity in the system continues to increase without decreasing, a significant transformation occurs when the partial pressure of oxygen is raised during roasting, leading to the complete conversion of iron monoxide into Fe₂O₃.

3.2. Efficient Desulfurization during Oxidative Roasting

Cobalt mostly remains associated with pyrite in the magnetite iron ore, which is confirmed by mineralogical analysis of the cobalt-bearing tailing and by detecting the pyrite phases. The pyrite contains cobalt substituted with iron as [Fe,Co]S₂ isomorphous from which cobalt can be extracted only through the oxidation of pyrite [32,33]. Previously, some studies have concluded that desulfurization in the roasting of pyrite concentrates is successful at 800–1200 °C [34,35,36]. Therefore, in our studies, the roasting experiment for removing sulfur from the cobalt-bearing tailing was performed at 900, 1000, and 1100 °C with oxygen blowing at a flow rate of 300 cc/min for 1 h of holding time. A summary of the results of the experiment is shown in Figure 9. The sulfur content in the initial tailing was found to be 3.9 wt.%, and a 55.3% degree of desulfurization was accomplished at 900 °C. However, with the increase in temperature, the degree of desulfurization increased to 81.6% and 89.3% at 1000 and 1100 °C, respectively. The roasting experiment conducted as a function of high temperature gave up to 90% of desulfurization degree with a holding time of 1 h. These experimental results confirmed that desulfurization can be successfully carried out at high temperatures, which is in accordance with thermodynamic modeling too.

To further improve the desulfurization rate, time-dependent roasting experiments were also performed. In Figure 10, the effect of holding time on the degree of desulfurization in a temperature range is shown. At 900 °C, the sulfur content in the cobalt-bearing tailing decreased with a longer holding time. For the experiment with a holding time of 5 h, the sulfur content in the tailing decreased to 1.09 wt.% with an 87.5% degree of desulfurization. However, for 1 h of holding time at 1000 °C, the sulfur content decreased to 1.62 wt.% within roasting for 1 h of holding time. Upon increasing the holding time further (3 h), the sulfur content was lowered to 0.81 wt.% with a 90.7% degree of desulfurization. After 5 h of holding time, the sulfur content was reduced to <0.18 with a 98.1%. degree of desulfurization. Interestingly, for the experiment performed at 1100 °C, the degree of desulfurization was observed to be 93%, 98.3%, and 98.5% for 3, 4, and 5 h of holding time, respectively. These observations suggest that the conditions for oxidative roasting and desulfurization are favorable and optimized when the experiments are carried out at 1000 and 1100 °C for a period of 4 and 5 h, permitting the elimination of sulfur up to 98.5%.

To analyze desulfurization and its products, the chemical composition of the roasted tailing was determined by XRF. The results are given in

Table 3. Chemical composition as determined by XRF of roasted tailings at different temperatures.

Temperature, °C	Element, wt.%
	Fe2O3	MgO	SiO2	Al2O3	CaO	MnO	TiO2	SO3	S
900	74.05	13.46	9.13	1.96	0.90	0.25	0.12	0.13	1.09
1000	78.22	10.37	8.64	1.76	0.92	0.26	0.18	0.08	0.16
1100	76.42	11.37	8.71	1.65	0.99	0.29	0.14	-	0.13

. The sulfur content in the roasted tailing was found to be 1.09 wt.%, 0.16 wt.%, and 0.13 wt.% at 900, 1000, and 1100 °C, respectively, as analyzed by ICP-OES at 900 °C. It must be noted that as the sulfur content in the sample began to decrease, the fraction of iron oxide varied from 67.12% (the initial fraction) to 74.05%, 78.22%, and 76.42% at 900, 1000, and 1100 °C consecutively. The XRF data showed an increment in iron oxide content, indicating the transformation of iron in pyrite to its oxide during the roasting process. However, the concentration of oxides such as MgO, SiO₂, and Al₂O₃ in the roasted tailing was found to be the same as in the initial sample.

To further examine the mineralogical structure of the roasted tailing, X-ray diffraction analysis (XRD) was conducted to gain insight into the crystalline domains in the cobalt-bearing tailing and their relationship with the composition. The results of XRD for the roasted tailings at 900, 1000, and 1100 °C with 5 h of holding time are shown in Figure 11. The cobalt-bearing tailing was composed predominantly of magnetite (Fe₃O₄) and several quartz and pyrite phases. The cobalt-bearing waste, which was a black powder, consisted primarily of magnetite (Fe₃O₄) as well as various phases of quartz and pyrite. It was evident that a profound alteration had taken place in the tailings as their color transitioned to a reddish-brown shade after undergoing roasting, thereby signifying the extensive transformation of iron magnetite oxide into hematite. The XRD analysis revealed that the roasted tailing, which was subjected to a temperature of 900 °C, predominantly consisted of hematite (Fe₂O₃) with a complete conversion of magnetite into hematite. Additionally, a small amount of quartz phase was also detected. The diffraction patterns for the pyrite phase were absent in XRD, which could be due to the minimal content of sulfur in the roasted tailing. In addition to the hematite and quartz phases, the crystalline peaks for a few magnesioferrite (MgFe₂O₄) phases were detected, from which FeO and MgO oxides have combined to form MgFe₂O₄ spinel during the roasting at a temperature above 1000 °C. It is worth noting that the XRD analysis result corresponds exactly to the calculated results derived from the HSC Chemistry database. Considering the relatively low activity level of the Fe₂SiO₄ compound, it is likely that it did not originate in the tailing during the roasting. However, it is important to note that when the temperature exceeds 1100 °C, there is a high likelihood of the formation of this compound. The XRD patterns show the presence of major phases, namely hematite and magnesioferrite, in oxidized and roasted tailing.

A thorough analysis was conducted on the particle size distribution of the tailings that underwent roasting for a duration of 5 h, specifically at temperatures of 900, 1000, and 1100 °C. The SEM image in Figure 3 confirmed that there was variation in the particle size of the initial tailing before roasting. When examining the tailing, it was found that the average diameter measured 45.15 μm, whereas the average width was 78.65 μm. However, it is important to highlight that the size of the tailing particles underwent a reduction subsequent to the roasting procedure, as evidenced by the graphical representation presented in Figure 12. As the temperature increased to 900, 1000, and 1100 °C, there was a corresponding decrease in the average diameter of the tailings to 25.17, 21.78, and 19.13 μm, and the average width decreased to 45.12, 38.19, and 31.77, respectively.

Following the roasting procedure, a detailed examination of the morphology and elemental distribution of the tailings was carried out using SEM-EDS analysis. The particles observed during the initial tailing phase were characterized by irregular shapes and exhibited a wide range of sizes. The smaller parts were characterized by their sharp edges and smooth surfaces, whereas the larger pieces were identified as agglomerations of multiple smaller components joined together (in Figure 3). Figure 13 shows the SEM-EDS analysis image, which provides evidence of a decrease in the particle size of the tailings and significant changes in both surface condition and elemental distribution after the roasting process. Figure 13a,b show the SEM and EDS mapping images of tailing that have been roasted at a temperature of 900 °C. Additionally, Figure 13c,e and Figure 13d,f show the SEM image and EDS mapping image of tailings roasted at temperatures of 1000 and 1100 °C, respectively. Upon examination, it was found that there were no significant disparities in the morphology of the roasted tailings. According to the EDS mapping results, iron emerged as the primary component, accompanied by magnesium and silicon acting as impurity elements. It is important to mention that there is currently no distribution of sulfur. Additionally, the particle size of the tailing differs from that of the initial tailing. However, the most distinct difference was the visible reduction in the size of particles. In addition, roughness and rough joints were observed on the surface of all the particles, indicating another optical change. The sharp-edged, smooth-surfaced pieces that had been so prominent in the initial tailing were no longer present. During the roasting process, it is believed that the size of the particles decreased due to the fragmentation of large agglomerates into smaller particles. This phenomenon could be one of the reasons for the observed decrease in particle size. The cracking of large particles can also happen as a result of thermal stresses and phase transitions that occur within the components of these particles. The smooth surface becomes porous as a result of sulfur release from the solid particles, and this leads to the creation of a rough surface through the formation of multi-layer stratification.

The distribution of elements in tailings roasted at temperatures of 900, 1000, and 1100 °C showed no significant differences. According to the results of the EDS analysis presented in

Table 4. Chemical composition of roasted tailing by EDS analysis.

Temperature, °C	Elements, wt.%
	Fe	O	Si	Mg	S
900	75.87	14.92	3.85	5.36	-
1000	78.85	16.75	2.19	2.21	-
1100	78.10	14.50	3.33	4.07	-

, it is evident that iron is the prevailing element, as indicated by the EDS mapping image displayed in Figure 13. The iron content in the tailings that were roasted at a temperature of 900 °C was approximately 76 wt.%, while the average iron content in the tailings roasted at temperatures of 1000 and 1100 °C was found to be 78 wt.%. The iron content detected through the EDS analysis demonstrated a close correlation with the iron content detected through the XRF analysis (Table 3). The consistent findings of both the EDS and XRF analyses indicate that magnesium and silicon are the main impurity elements. The EDS analysis did not identify elements such as S, Al₂O₃, CaO, MnO, and TiO₂ due to their low concentration levels.

4. Conclusions

Cobalt-bearing pyrite tailing extracted from iron ore by flotation contains 1.28 wt.% Co, 8.73 wt.% S, and major minerals such as magnetite, quartz, and some pyrites. In iron magnetite ore, cobalt is combined with iron (in pyrite) in a [Fe,Co]S₂ isomorphous state, from which cobalt can be recovered through the hydrometallurgical method followed by oxidative roasting and leaching successively. In our study directed towards the oxidation of pyrite, thermodynamic modeling using HSC Chemistry 6.2 software for potentially stable phases was adopted to investigate the effect of temperature and oxygen partial pressure on the phase transition of Fe–S–O and Co–S–O systems to optimize the degree of desulfurization from cobalt-bearing pyrite tailing and to ease the extraction of cobalt from the oxidized byproduct. The results of this work are presented in a carefully structured order.

• As a highlight, the poisonous sulfur gas is also eliminated as its oxide during the roasting process, thus allowing a rather environmentally friendly roasting of ore. The thermodynamic modeling in Fe–S–O and Co–S–O systems proved that the iron and cobalt sulfides can be oxidized moderately at relatively high temperatures (>900 °C) under the atmosphere, such as logp(O₂) > −5 supplied from outside and logp(SO₂) < 0 generated during roasting.
• The equilibrium module was employed in the multi-component system of the tailing to determine the products that are generated during the roasting process. The main target elements in these products are Fe, Co, and S, while impurity elements like Mg and Si are also present. By subjecting iron and cobalt sulfides to roasting within the temperature range of 900 to 1100 °C, it was determined that sulfur can be effectively and entirely eliminated, undergoing transformation into SO₂ gas. Furthermore, it should be mentioned that under these temperature conditions, there is a possibility for iron magnetite to undergo a conversion into hematite. Furthermore, it is intriguing to observe that the Fe₂O₃ that is present in the system can react with MgO, leading to the synthesis of a compound called MgFe₂O₄. The results of the XRD analysis strongly support the presence of the compound in the roasted tailing, as indicated by the close agreement.
• The roasting experiments performed at 1000 and 1100 °C for 4 and 5 h yielded a 98% degree of desulfurization. The magnetite phases (Fe₃O₄) in the cobalt-bearing tailing were completely converted to hematite (Fe₂O₃). The absence of the pyrite phase in the roasted tailing confirms that the oxidation of pyrite was successful, as evident by the XRF and XRD analyses. During the roasting process at 1000 and 1100 °C, the tailing particles underwent a division, resulting in their size being reduced by half. As a result, these particles had an average diameter ranging from 19 to 21 μm, accompanied by a width spanning from 32 to 38 μm. An observation was made that all of the grains trailing behind had a rough surface, suggesting a phase transition of iron oxide and the separation of sulfur from pyrite.

These optimized conditions dictated by the thermodynamic model can allow a high degree of desulfurization of the cobalt-bearing pyrite during the roasting process and thus permit the recovery of cobalt when applied in industry. Thermodynamic modeling predicted the occurrence of oxidation and sulfation reactions during roasting, which were in agreement with the experimental results. Our theoretical and experimentally correlated models presented here enable the metallurgists to adopt an optimized method for the specific roasting of cobalt-bearing pyrite at an industrial scale. Further focus of this research will be on the recovery of cobalt from the oxidized and roasted byproducts as obtained via the above thermodynamically optimized and improved method.