Reduction Mechanisms During the Recovery of Mo and Fe via Molten-Bath Smelting of Copper Slag and Waste MoSi₂

Abstract

Molybdenum (Mo) finds extensive applications in the steel industry, and the recycling of secondary molybdenum resources is crucial for the green development of the molybdenum sector. Meanwhile, the large-scale stockpiling of copper slag, a bulk industrial solid waste, poses severe environmental and resource-related challenges. Addressing the common issues of the refractory nature of waste molybdenum disilicide (MoSi₂) and the underutilization of iron resources in copper slag, this study proposes a synergistic smelting approach using copper slag and waste MoSi₂, aiming to realize the coordinated treatment of these two solid wastes and the simultaneous, efficient recovery of valuable metals (Mo and Fe). Under non-isothermal conditions, this work elucidates the phase evolution of copper slag and the decomposition–reduction behavior of MoSi₂; clarifies the dual role of coke as the primary reductant at the initial reaction stage and as a maintainer of a reducing atmosphere during smelting; and systematically investigates the effects of smelting temperature, slag basicity, and coke dosage on metal recovery. The results demonstrate that, under optimized process conditions, the recovery efficiencies of molybdenum and iron can reach 98.97% and 98.46%, respectively. This study provides a new strategy for the enrichment and extraction of metallic elements from waste MoSi₂ and copper slag.

1. Introduction

Molybdenum disilicide (MoSi₂), as an important high-temperature structural and functional material, is widely used as a heating element in high-temperature electric furnaces [1]. With its large-scale application, a substantial quantity of spent MoSi₂ heating elements (composed primarily of MoSi₂ with minor amounts of SiO₂, bentonite, and other additives) has been generated and requires effective treatment [2,3]. Owing to its high melting point, high strength, and excellent high-temperature stability, molybdenum is an indispensable strategic rare metal in the steel, alloy, and high-temperature material industries [4,5]. Consequently, the recovery of metallic molybdenum from molybdenum-bearing secondary resources, such as decommissioned MoSi₂ rods, is of great significance for the sustainable development of molybdenum metallurgy [6,7].

Because MoSi₂ is chemically stable and readily forms a protective SiO₂ layer, conventional hydrometallurgical or pyrometallurgical processes struggle to extract molybdenum from it economically and efficiently. Zhu et al. [8] employed waste MoSi₂ rods, ZrSi₂, and SiC as raw materials to fabricate MoSi₂–ZrSi₂–SiC coatings via spark plasma sintering (SPS); the coatings exhibited excellent oxidation resistance at 1700 °C and effectively protected the Nb substrate. Hou et al. [9] developed a novel oxidation-based approach at 1000 °C for spent MoSi₂ heating elements, producing single-crystalline MoO₃ belts ranging from micrometer to millimeter and even centimeter scales for applications in luminescence and high-temperature devices. At present, studies on the recycling of waste MoSi₂ remain limited and are mainly focused on its reuse for the preparation of oxidation-resistant coatings or the production of MoO₃. Copper smelting generates enormous quantities of copper slag (approximately 2.2–3.0 t of slag per tonne of copper) [10,11,12], with iron contents often as high as 30–40%, exceeding the grade of many low-grade iron ores [13,14]. The large-scale stockpiling of copper slag not only occupies land resources but also poses potential risks of heavy-metal contamination [15,16]. In recent years, increasing attention has been paid to strategies that exploit the low melting point of copper slag as a substitute for conventional fluxes, enabling the co-treatment of other metallurgical solid wastes and the simultaneous recovery of valuable metals. Yang et al. [17] treated stainless steel slag by smelting with coal tar and copper slag, achieving Fe and Cr recovery rates of 97.71% and 95.56%, respectively, under optimal conditions. Zulhan et al. [18] demonstrated the feasibility of a molten reduction process in which copper slag served as the flux and bituminous coal was used to treat manganese ore for ferromanganese production, realizing the recovery of Fe and Mn. Using copper slag as a molten bath to synergistically process other solid wastes thus offers further potential for the efficient recovery of valuable metallic elements.

Separate treatment of waste MoSi₂ and copper slag is associated with considerable technical or economic challenges. The development of a novel process that employs copper slag simultaneously as a fluxing medium and an iron source, enabling the chemical reduction of molybdenum from MoSi₂ and the enrichment and recovery of iron from copper slag under a reductive smelting environment, would therefore offer substantial environmental and economic benefits. However, fundamental scientific issues—such as the high-temperature interaction mechanisms among multicomponent systems (Mo–Si–Fe–O–C) and the coke–MoSi₂ reduction pathway—remain insufficiently understood.

To address these challenges, this study proposes a coke-based molten-bath smelting process for the co-treatment of waste MoSi₂ and copper slag, aiming to extract molybdenum from spent MoSi₂ while recovering iron resources from copper slag. Non-isothermal heat-treatment experiments were conducted to elucidate gas evolution behavior and phase transformation pathways during the reaction process and to clarify the role of coke. Furthermore, thermodynamic calculations combined with systematic smelting experiments were performed to determine the effects of key process parameters on molybdenum and iron recovery efficiencies, thereby providing both experimental data and theoretical guidance for the extraction of metallic elements from these two solid wastes.

2. Materials and Methods

2.1. Materials

The copper slag used in the experiments was flotation-depleted copper slag obtained from a company in Yunnan, China. The major elemental contents were determined by XRF and chemical titration, showing a total iron content of 38.90 wt.%, SiO₂ content of 32.15 wt.%, CaO content of 5.60 wt.%, and Al₂O₃ content of 7.28 wt.%. Combined with the XRD results (Figure 1a), Fe in the copper slag is mainly present in the form of fayalite (ICDD-PDF:72-0297) and magnetite phases (ICDD-PDF:88-0315). The copper slag also contains 0.19 wt.% Cu, indicating a relatively low copper content. Revealing waste MoSi₂ material consisted of decommissioned grade-1800 MoSi₂ heating elements collected from the laboratory, which retained their original MoSi₂ composition after decommissioning [19]. XRD analysis (Figure 1b) confirmed that the material was predominantly composed of MoSi₂(ICDD-PDF:41-0612), with no detectable impurity diffraction peaks. The waste MoSi₂ was crushed into powder before use. The coke employed in this study had a carbon content of 85% and a particle size ranging from 0 to 5 mm.

2.2. Methods

Thermodynamic calculations: Thermodynamic calculations for the co-smelting system were performed using the FactSage 8.1 software package. The FactPS (for pure substances), FTstel (for molten steel), and FToxid (for oxides) databases within FactSage 8.1 were employed to construct predominance area diagrams and to predict and analyze phase transformation behaviors during heat treatment under different compositional ratios. Ternary phase diagrams of the slag system during smelting were established, and slag viscosity was calculated to support the optimization of the synergistic smelting process.

Non-isothermal heat-treatment experiments: Based on the optimal compositional ratio, 3 g of well-mixed raw material powders were pressed into cylindrical pellets with a diameter of 16 mm. Non-isothermal heat-treatment experiments were conducted in a horizontal tube furnace (Hefei Kejing, GSL-1200, Hefei, Anhui, China), with argon gas introduced at a flow rate of 500 mL/min as a protective atmosphere, to investigate phase transformations during the reaction between copper slag and MoSi₂ in the presence of coke. In addition, thermogravimetric–differential scanning calorimetry (TG–DSC; STA449F5, Netzsch, Selb, Germany) was used to examine mass changes in the system under the optimal composition during non-isothermal heating. The heating rate was set to 10 °C/min, with argon as the purge gas at a flow rate of 50 mL/min.

Molten smelting: Reduction smelting experiments were carried out to evaluate the effects of temperature, CaO/SiO₂ mass ratio, and coke addition on metal recovery during synergistic smelting. The charge consisted of 40 g of copper slag, 10 g of waste MoSi₂, and 3.7 g of CaF₂, with the CaO addition adjusted according to the desired CaO/SiO₂ mass ratio. Copper slag, coke, CaO, CaF₂, and waste MoSi₂ powders were thoroughly mixed and loaded into a corundum crucible (inner diameter: 50 mm; outer diameter: 55 mm; height: 60 mm), which was then placed in a vertical tube furnace (Hefei Kejing, GSL-1750X). The furnace was heated to the target temperature at a rate of 10 °C/min and held for 60 min, followed by furnace cooling. After cooling, the crucible was crushed to separate the secondary slag from the metal ingot. The recovery of valuable metallic elements in the smelting products was calculated using the equation R_i = 1 − m_i/M_i, where i denotes the target metal element (Mo or Fe), m_i, is the mass of the metal (Mo or Fe) remaining in the secondary slag after smelting, and M_i, is the mass of the metal (Mo or Fe) initially present in the raw materials.

2.3. Characterization

The phases in the raw materials and heat-treated samples were analyzed by X-ray diffraction (XRD, Bruker D8, Bruker AXS GmbH, Karlsruhe, Germany). The chemical compositions of the raw copper slag, secondary slag phase, and alloy phase obtained after co-smelting were characterized using X-ray fluorescence spectroscopy (XRF, Bruker S8 Tiger, Bruker AXS GmbH, Karlsruhe, Germany). The microstructures, phase distributions, and elemental distributions of the heat-treated samples were examined via scanning electron microscopy (Hitachi SU 3500, Hitachi High-Technologies Corporation, Tokyo, Japan) in backscattered electron (BSE) mode coupled with energy-dispersive spectroscopy (EDS). The volume fractions of CO/CO₂ gases released during heat treatment of the optimal composition system were monitored using an infrared gas analyzer (Shanghai Xincheng, H-2100, Shanghai, China).

3. Results

3.1. Analysis of Phase Transformation and Microstructural Evolution During the Reaction of Waste MoSi₂ with Copper Slag

During the synergistic smelting process, the reactions that may occur include the reduction of copper slag, as well as the redox transformation and slag-forming reactions involving Mo and Si from MoSi₂. The specific reaction equations that may occur are shown in

Table 1. Possible reactions.

No.	Reactions
(R1)	2Fe2SiO4 + 4C = 4Fe + 2SiO2 + 4CO (g)
(R2)	CaMoO4 + 3.5C = 0.5Mo2C + CaO + 3CO (g)
(R3)	7Fe3O4 + MoSi2 = 21FeO + MoO3 + 2SiO2
(R4)	4Fe3O4 + MoSi2 = 12FeO + 2SiO2 + Mo
(R5)	2Fe2SiO4 + MoSi2 = 4Fe + 4SiO2 + Mo
(R6)	2Fe2SiO4 + MoSi2 + 0.5C = 4Fe + 4SiO2 + 0.5Mo2C
(R7)	1.333CaMoO4 + 1.167C + MoSi2 = 1.167Mo2C + 1.333CaO + 2SiO2
(R8)	2Fe2SiO4 + MoSi2 + 0.5C + 4CaO = 4Fe + 4CaSiO3 + 0.5Mo2C

, and the relationship between the Gibbs free energy of the possible reactions and temperature is illustrated in Figure 2a. Phase composition analysis of copper slag indicates that high-valence iron is mainly present in the fayalite and magnetite phases. Fe₃O₄ can be reduced by carbon at relatively low temperatures; when the temperature reaches approximately 800 °C, Fe²⁺ ions in Fe₂SiO₄ begin to be reduced by carbon to metallic Fe (reaction (R1)).

Figure 2b presents the equilibrium phase diagram at 1050 °C with different MoSi₂ consumptions, illustrating the reactions between copper slag and MoSi₂ and the associated phase transformations under the reducing atmosphere provided by carbon. Initially, Fe₃O₄ is reduced by MoSi₂ to form FeO (reaction (R4)), during which a small amount of metallic Fe is generated. Subsequently, FeO combines with SiO₂ to form the fayalite phase (Fe₂SiO₄). After the complete reduction of Fe₃O₄, Fe²⁺ ions in Fe₂SiO₄ are further reduced by MoSi₂ to produce Fe and SiO₂ (reactions (R5) and (R6)). CaO then reacts with SiO₂ to form the slag phase, ultimately yielding the more stable CaSiO₃ phase.

It is noteworthy that both Mo and Si in MoSi₂ possess reducing capability. Mo in MoSi₂ can be oxidized by Fe ions to form volatile MoO₃ (reaction (R3)). CaO reacts with MoO₃ to generate CaMoO₄, and the Mo in CaMoO₄ can subsequently be reduced to metallic Mo by carbon and MoSi₂ (reactions (R2) and (R7)). Adequate addition of reductant and CaO ensures effective recovery and directional enrichment of Mo, preventing Mo loss due to MoO₃ volatilization.

By combining with the SiO₂ generated during reactions to form the slag phase, CaO further decreases the reaction Gibbs free energy (reaction (R8)) and facilitates subsequent slag–metal separation. This process promotes the reduction of fayalite and the selective capture of Mo and Fe, which ultimately constitute the thermodynamically stable phases of the reaction system.

Figure 3 presents the TG–DSC results, which can be divided into three distinct stages. At approximately 400 °C, the sample mass decreases, and an endothermic peak appears, primarily due to the evaporation of crystalline water from CaO. Around 650 °C, the sample mass decreases further, and the rate of mass loss subsequently diminishes, mainly because carbon begins to react with Fe₃O₄ in the copper slag, generating CO gas. At about 1050 °C, a pronounced mass loss is observed, which is mainly attributed to the occurrence of the Boudouard reaction, as well as the reduction of Fe₂SiO₄ and iron oxides in the copper slag by carbon and CO, producing CO and CO₂.

In conjunction with the TG–DSC results, non-isothermal heat-treatment experiments were conducted on the mixed system of copper slag, waste MoSi₂, CaO, and C during synergistic smelting, in order to further elucidate the complex interactions among the components, as well as the phase transformations and microstructural evolution during heating. Figure 4a shows the XRD patterns of samples subjected to isothermal heat treatment for 0.5 h at different temperatures. When the temperature increases to 750 °C, the characteristic peaks of Fe₃O₄ (ICDD-PDF:88-0315) weaken, indicating its reduction to FeO, which partially reacts with SiO₂ to form the fayalite phase (ICDD-PDF:72-0297). With further temperature elevation, the DSC curve continuously decreases, reflecting sustained endothermic behavior of the system. Thermal expansion of the fayalite and MoSi₂ crystal structures causes their characteristic peaks to shift toward lower angles. Owing to thermal expansion, lattice energy decreases, interatomic distances increase, and chemical bond strength is weakened, thereby facilitating elemental diffusion and enhancing the reactivity between Fe₂SiO₄ and MoSi₂.

When the temperature reaches 1050 °C, Fe²⁺ in Fe₂SiO₄ is reduced, and Ca²⁺ ions enter the olivine crystal structure, substituting for part of the Fe²⁺ sites and inducing a transformation to a pyroxene structure. Consequently, the characteristic peaks of Fe₂SiO₄ nearly disappear, while weak diffraction peaks corresponding to the pyroxene phase CaSiO₃ (ICDD-PDF:42-0550) emerge. At 1150 °C, partial melting occurs in localized regions of the sample, leading to pronounced broadening of many diffraction peaks. Meanwhile, some MoSi₂ comes into contact with iron oxides in the copper slag and undergoes solid–solid redox reactions, during which Si is oxidized to Si⁴⁺ and Ca²⁺ ions continue to incorporate into the lattice. As a result, distinct diffraction peaks of the pyroxene phase CaSiO₃, the anorthite phase Ca₂Al(AlSiO₇) (ICDD-PDF:77-1146), and Ca₂SiO₄ (ICDD-PDF:83-0460) appear. Mo and the reduced Fe become enriched and combine with carbon, giving rise to characteristic peaks of the Fe₃Mo₃C phase (ICDD-PDF:47-1191).

Figure 4b presents the SEM–EDS results of the roasted samples. When the temperature reaches 1050 °C, Fe²⁺ in the outer layer of the fayalite phase is reduced, and the generated SiO₂ is transformed into columnar pyroxene under the action of CaO, while the reduced Fe particles gradually grow, enrich, and coalesce. Fine MoSi₂ particles react with iron oxides in the copper slag, and a large amount of light-gray Fe phase envelops the white Mo-rich Fe–Mo phase. A small fraction of Si does not fully diffuse and remains in the residual white phase. This core–shell structure indicates that the decomposition of MoSi₂ is primarily driven by the decoupled diffusion of Si and Mo: Si diffuses out of the MoSi₂ crystal lattice and acts as a reductant to reduce the surrounding iron oxides, generating Fe and SiO₂ in situ. Line-scan analyses further reveal the presence of a Si-rich oxide layer surrounding the Fe phase. EDS elemental mapping shows a large number of interlaced columnar pyroxene phases distributed within the slag, with some Fe²⁺ ions remaining incompletely reduced.

When the temperature increases to 1150 °C, the columnar pyroxene phase grows and coarsens, and Fe²⁺ within it is almost completely reduced. Based on the preceding experimental results and thermodynamic calculations, the stepwise reaction and decomposition process of MoSi₂ can be summarized into three stages. First, MoSi₂ comes into contact with iron oxides, during which Si reduces Fe²⁺ to Fe while forming SiO₂. Second, during the diffusion stage, Si diffuses toward iron oxides and is consumed by reaction, leading to the decomposition of MoSi₂ and the formation of the intermediate phase Mo-Si. Third, the intermediate Mo-Si phase further decomposes, with Si diffusing and reacting with Fe²⁺. A small amount of in situ-generated Fe combines with Mo and Si to form a Mo–Si–Fe alloy as an interfacial layer, while a minor fraction of Si enters Fe to form an Fe–Si alloy.

3.2. The Effect of Temperature

A charge consisting of 18.5 g CaO, 40 g copper slag, 10 g MoSi₂, and 3.7 g CaF₂ was prepared to achieve a CaO/SiO₂ mass ratio of 1, and the effect of temperature on the recoveries of Mo and Fe was investigated. FactSage calculations were employed to predict the temperature-dependent phase evolution during synergistic smelting (Figure 5a,b). The addition of CaF₂ lowers the melting point of the slag system and promotes the transformation of the high-melting Ca₃Si₂O₇ phase into the lower-melting cuspidine phase (Ca₄Si₂F₂O₇). When the temperature reaches approximately 1430 °C, the slag phase becomes fully molten, providing a favorable molten-bath environment for the reactions.

Based on the thermodynamic calculations, smelting experiments were conducted at 1450, 1500, 1550, and 1600 °C. XRD analysis of the resulting secondary slags (Figure 5c) shows that the slag obtained at 1450 °C is mainly composed of gehlenite (Ca₂(Al(AlSi)O₇)) and Ca₂SiO₄ phases. When the temperature increases to 1500 °C, the Ca₂SiO₄ phase disappears, and a small amount of cuspidine precipitates. With further temperature elevation, other high-melting phases in the secondary slag vanish, leaving gehlenite as the dominant phase. Figure 5e presents the SEM–EDS results of the secondary slag produced at 1450 °C, which consists primarily of three distinct phases. The dark-gray matrix, composed mainly of Ca, Al, Si, and O, is identified as gehlenite based on XRD and point-scan analyses. The light-gray short rod-like phase, enriched in Ca, Si, and O with minor amounts of Fe and F, corresponds to Ca₂SiO₄. This is attributed to the relatively low temperature, under which molecular thermal motion and mass transfer are limited, resulting in unfavorable kinetics for the reduction of Fe²⁺ by carbon; consequently, Fe²⁺ is not completely reduced and partially remains chemically dissolved in the slag. The bright white phase represents alloy particles that have not fully settled in the secondary slag, consisting mainly of Fe and C with only trace amounts of Mo.

Figure 5d illustrates the effects of temperature on slag melt viscosity and the recoveries of Mo and Fe. With increasing temperature, high-melting phases in the slag gradually melt, leading to a decrease in melt viscosity and thereby enhancing mass transfer and the settling of alloy droplets. As the temperature rises, the recovery of Mo initially increases and then decreases, reaching a maximum of 87.63% at 1550 °C. When the temperature is further increased to 1600 °C, the reduced melt viscosity promotes the reduction of Fe and the settling of iron droplets, resulting in a continuous increase in Fe recovery; however, partial oxidation and loss of Mo occur, causing a decline in Mo recovery. Figure 6 shows the alloy phase obtained at 1550 °C, which is mainly composed of Mo and Fe with a small amount of Si. This temperature is identified as optimal, yielding Fe and Mo recoveries of 97.06% and 87.63%, respectively.

3.3. The Effect of Basicity

CaO facilitates the release of Fe from the fayalite phase, while the SiO₂ liberated from fayalite decomposition combines with CaO to form a stable calcium silicate slag with favorable physicochemical properties [20], thereby promoting Fe reduction and Mo capture. An operating temperature of 1550 °C was selected to investigate the effect of the CaO/SiO₂ mass ratio on the recoveries of Fe and Mo.

The effect of CaO addition on the equilibrium phases of the melt was calculated using FactSage software (Figure 7a). With increasing CaO addition, the amount of liquid slag increases; however, when the CaO addition exceeds 28 g (corresponding to a CaO/SiO₂ mass ratio greater than 1.4), high-melting solid phases begin to precipitate from the slag melt, deteriorating melt fluidity and hindering slag–metal separation. Figure 7b presents the calculated slag phase composition and viscosity during synergistic smelting at different CaO/SiO₂ mass ratios. When the CaO/SiO₂ mass ratio exceeds 1.0, solid gehlenite begins to precipitate from the slag melt, and the amount of precipitated solid increases progressively with increasing basicity. As the basicity increases, slag viscosity initially decreases and then increases; the minimum viscosity of 0.14 Pa·s is achieved at a CaO/SiO₂ mass ratio of 1.1. This condition provides a favorable kinetic environment for reactions within the melt and facilitates the settling of alloy droplets, thereby promoting the selective enrichment and recovery of Mo and Fe.

XRD analysis was performed on the secondary slags obtained from synergistic smelting at different basicities (Figure 7c). At a CaO/SiO₂ mass ratio of 0.9, the secondary slag is predominantly amorphous, corresponding to a glassy phase. With increasing basicity, the characteristic diffraction peaks of gehlenite become progressively more pronounced. Figure 8 presents the SEM–EDS results of the secondary slags produced at CaO/SiO₂ mass ratios of 0.9 and 1.1. At a CaO/SiO₂ mass ratio of 0.9, the secondary slag mainly consists of residual alloy droplets entrapped in the slag and the slag matrix itself (Figure 8a). The slag phase shows a homogeneous distribution of Ca, Si, Al, and O; combined with the XRD results, it is identified as a CaO–Al₂O₃–SiO₂ glassy phase. The alloy droplets are mainly composed of Fe and C. At a CaO/SiO₂ mass ratio of 1.1, the secondary slag is composed of four distinct phases (Figure 8b). The bright spherical phase corresponds to incompletely settled alloy particles entrapped in the slag, consisting primarily of Fe and C and containing negligible Mo, indicating that most of the Mo has been effectively captured in the alloy ingot at the bottom of the melt. The light-gray, large spindle-shaped phase represents regions enriched in Ca and Al; it is composed of Ca, Al, Si, and O and is identified as gehlenite based on point-scan and XRD analyses. In the SEM images, the matrix phase appears similar in contrast to the gehlenite phase and is difficult to distinguish; however, elemental mapping reveals that the matrix is a Si-enriched region composed of Ca, Si, Al, O, and a small amount of F, corresponding to an amorphous glassy phase. In addition, a small amount of light-gray dendritic phase is present in the secondary slag, dispersed throughout the matrix; this phase is enriched in F and is likely attributable to cuspidine and CaF₂.

Figure 7d shows the recoveries of Mo and Fe at different CaO/SiO₂ mass ratios. As the basicity increases, the recoveries of both Mo and Fe initially increase and then decrease. When the CaO/SiO₂ mass ratio reaches 1.1, the recoveries of Mo and Fe attain their maximum values of 98.97% and 98.46%, respectively. Based on the previous experimental observations (Figure 4b and Figure 7c,d) and thermodynamic simulation calculations (Figure 2a and Figure 7b), the addition of CaO improves the thermodynamic conditions of the reactions, promotes the reduction of the fayalite phase, and facilitates the formation of Fe–Mo alloys. Moreover, CaO reacts with SiO₂ to form a stable Ca–Si–Al–O slag system, thereby enhancing melt fluidity, promoting mass transfer, facilitating alloy settling, and improving slag–metal separation.

Figure 9 shows the alloy phase obtained by smelting at a CaO/SiO₂ mass ratio of 1.1. The alloy mainly consists of two phases. The white rod-like phase represents the primary Mo-enriched region and contains Fe along with a small amount of Si, whereas the gray phase serves as the matrix and is enriched in Fe with minor solid-solution amounts of Mo and Si. The overall Mo content in the alloy is 32.68%.

3.4. The Effect of the Amount of Coke Addition

At 1550 °C and a CaO/SiO₂ mass ratio of 1.1, the effect of coke addition on the recoveries of Fe and Mo was investigated. Figure 10a shows the equilibrium phase diagrams of the synergistic smelting system at different carbon additions. With increasing carbon addition, Fe ions in the slag are progressively and completely reduced, leading to an increased Fe content in the alloy phase. Mo reacts with carbon preferentially to form Mo₂C, followed by the formation of Fe₃C through the reaction of Fe with carbon. Figure 10b presents the metal recoveries as a function of coke addition. As the carbon addition increases, the recovery of Mo rises gradually; the increased carbon provides a sufficient and stable reducing atmosphere, preventing the oxidation of Mo in MoSi₂ to volatile MoO₃ during synergistic smelting. In contrast, the recovery of Fe exhibits only minor variation with carbon addition.

Because the crystal structure of MoSi₂ is relatively stable and its melting point is high, it remains in the solid state during synergistic smelting, which restricts the diffusion of Si and limits its reducing ability at lower temperatures. Consequently, in the initial stage of the reaction, carbon acts as the primary reductant and reacts with iron oxides in the copper slag. The volume fractions of CO and CO₂ generated during the carbon-assisted synergistic smelting of waste MoSi₂ and copper slag were monitored using an infrared CO/CO₂ gas analyzer. As shown in Figure 11, carbon plays a reductive role throughout the synergistic smelting process and provides a reducing atmosphere.

Based on the gas evolution behavior, the reaction between MoSi₂ and copper slag can be divided into four stages.

• In the initial stage of heat treatment, the reaction proceeds slowly. When the temperature reaches approximately 700 °C, CO gas begins to evolve, followed by the release of CO₂. The volume fraction trends in CO and CO₂ are similar, both reaching a maximum at about 860 °C and then gradually decreasing. This indicates that within this temperature range, carbon primarily reacts directly with Fe₃O₄ in the copper slag to generate CO gas, while part of the CO also acts as a reductant, reacting with iron oxides in the slag to produce CO₂. These observations are in good agreement with the mass-change behavior revealed by the TG–DSC analysis discussed earlier.
• When the temperature reaches approximately 950 °C, the volume fraction curves of CO and CO₂ reach a minimum and then begin to increase. Combined with the XRD analysis discussed above, the reduction in CO and CO₂ evolution is mainly attributed to the complete reduction of the magnetite phase by C and CO. Owing to the complex crystal structure of fayalite, C or CO cannot readily access and reduce Fe²⁺ within fayalite, leading to decreased gas evolution after 860 °C. As the temperature increases to 950 °C, thermal expansion of the fayalite lattice and intensified molecular motion enhance the probability of direct contact between C/CO and Fe²⁺ in fayalite. Consequently, the kinetic conditions are improved, the reduction of fayalite by C and CO is significantly intensified, and the evolution of CO and CO₂ increases.
• Subsequently, the CO volume fraction curve reaches a maximum at around 1100 °C. Notably, at approximately 1100 °C, the CO₂ volume fraction curve exhibits a plateau. In conjunction with thermodynamic analysis, TG–DSC results, and the CO evolution profile, this behavior indicates that the Boudouard reaction becomes pronounced. The occurrence of the Boudouard reaction causes the carbon reduction process to gradually shift from direct reduction to indirect reduction mediated by CO [21].
• With a further temperature increase, the CO volume fraction decreases again, reaching a minimum at about 1230 °C, and then increases slowly. The CO₂ volume fraction follows a similar trend to that of CO; however, above 1230 °C, CO₂ is no longer detected, and only CO evolution is observed. This behavior is mainly attributed to partial melting of phases in the synergistic smelting system, leading to the formation of a liquid phase that promotes contact between MoSi₂ and iron oxides. Under these conditions, Si in MoSi₂ becomes the dominant reductant. Owing to the strong reducing capability of Si in MoSi₂, CO evolution is suppressed, while sufficient carbon ensures the continuous occurrence of the Boudouard reaction and the maintenance of a reducing atmosphere.

4. Discussion

Compared with previously reported methods—such as Fe recovery from molten copper slag assisted by externally applied stirring vortices (Fe recovery rate: 91.76%) [20], coke reduction of molten copper slag (Fe recovery rate: 97.6%) [22], direct reduction with rubber seed oil (Fe recovery rate: 93.85%) [13], or biomass carbon roasting for metallic Fe recovery (Fe metallization: 88%) [21]—the synergistic smelting process used in this study achieves a higher Fe recovery rate (98.46%) while simultaneously processing two types of solid waste, thereby improving overall waste treatment efficiency.

Regarding Mo recovery from MoSi₂-based waste, most previous studies have focused on oxidative routes to generate MoO₃ or on reusing MoSi₂ as a functional material (e.g., coatings) [2,3,6,9], which offers limited applications. In contrast, this study demonstrates a distinct metallurgical approach, emphasizing the Si–Mo decoupled diffusion in MoSi₂ and the competitive Mo–Fe–Si–C oxidation in the high-temperature melt. In this process, MoSi₂ reacts within the melt bath provided by the copper slag, decomposes, and in situ generates a Mo–Fe alloy while SiO₂ dissolves into the slag, thereby directly enriching Mo and Fe into the iron–molybdenum alloy phase.

This study investigates the cooperative smelting process for Mo and Fe extraction from copper slag and waste MoSi₂. The effects of smelting temperature, CaO/SiO₂ ratio, and coke dosage on the recovery efficiency of Mo and Fe were systematically examined. Additionally, the reaction mechanisms and phase transformation behaviors during the interaction between copper slag and waste MoSi₂ were elucidated through non-isothermal heat treatment experiments. The primary findings are summarized below:

(1)

This study combines thermogravimetric–differential scanning calorimetry (TG–DSC), gas release behavior, XRD phase analysis, and SEM–EDS microstructural observations to analyze the transformation behavior of iron oxides in copper slag. During the initial reaction stage at lower temperatures, coke acts as the primary reductant, directly reducing copper slag. Fe₃O₄ is initially reduced, followed by the onset of Fe₂SiO₄ reduction at 950 °C. The phase transformation pathway proceeds as: Fe₂SiO₄ → iron-rich pyroxene phase → Fe + CaSiO₃. When the temperature reaches 1050 °C, the Boudouard reaction becomes prominent, marking a transition in copper slag reduction from direct carbon reduction to indirect CO reduction.

(2)

The Si element in MoSi₂ exhibits strong reducing capability, and the decomposition of MoSi₂ is mainly driven by the decoupled diffusion of Si and Mo. Si diffuses out of the MoSi₂ crystal lattice and acts as a reductant, reducing the surrounding iron oxides to generate Fe and SiO₂ in situ. When the temperature exceeds 1230 °C, partial phases in the copper slag melt to form a liquid phase, which enhances the contact and reaction between MoSi₂ and iron oxides in the slag, allowing MoSi₂ to function as the primary reductant. Meanwhile, coke provides a reducing atmosphere, preventing the oxidation of Mo in MoSi₂ to volatile MoO₃ and thus minimizing Mo loss.

(3)

CaO addition reduces ΔG of the reaction between copper slag and MoSi₂, thereby promoting MoSi₂ decomposition and the reduction of iron oxides in copper slag. Additionally, the addition of CaO can improve the melt viscosity (as verified by FactSage calculations), promote reaction mass transfer in the melt and the settling of alloy droplets, and optimize the recovery of Mo and Fe metallic elements.

(4)

The copper slag provides both a melt bath and a source of iron, supplying Fe²⁺ in the initial stage of smelting to promote the decomposition of MoSi₂ and the formation of Mo–Fe. Ultimately, the copper slag transforms into a stable CaO–SiO₂ slag system, facilitating the directional enrichment of Fe and Mo into the alloy phase and enabling the final separation of metal and slag. Under optimal experimental conditions, the recovery efficiencies of Mo and Fe reach 98.97% and 98.46%, respectively, with the metals consolidated into ingots. This method enables resource utilization of copper slag and waste MoSi₂.