Influencing Factors and Mechanisms of Zinc Recovery from Electric Arc Furnace Dust via Microwave-Assisted Carbothermic Reduction

Abstract

Electric arc furnace dust (EAFD) is a zinc-containing solid waste generated during steelmaking, and advanced recycling strategies are needed to facilitate the recovery of valuable zinc. This study investigated the microwave-assisted carbothermic reduction in EAFD using coke as a reductant, with a focus on temperature (900–1100 °C), holding time (0–60 min), and the C/Zn molar ratio (3–5). The results demonstrated that the zinc removal rate exhibited positive correlations with both temperature and time. Under optimized conditions (1100 °C, 60 min), a zinc removal rate of 95.45% was achieved, accompanied by a complete decomposition of the ZnO phases. Furthermore, increasing the C/Zn molar ratio enhanced the zinc recovery efficiency and product purity. Isothermal kinetic analyses indicated that the reaction proceeds in two stages: during the initial stage (0–30 min), the process was governed by three-dimensional diffusion control with an activation energy of 146.50 kJ/mol, while the final stage (30–60 min) transitioned to chemical reaction control with an activation energy of 267.32 kJ/mol. Comparative assessments indicated that microwave processing significantly reduced the activation energy compared to conventional heating methods. These findings suggest that microwave-assisted reduction is capable of attaining a high-grade recovery of Zn from EAFD, thus opening up new avenues for the resource-oriented utilization of EAFD.

1. Introduction

As a dominant short-process steelmaking technology, electric arc furnace (EAF) production has progressively increased its global market share due to inherent advantages in energy efficiency and raw material flexibility [1]. Statistics from the World Steel Association indicate that global crude steel output reached 1892 million tons in 2023, with EAF processes accounting for 28.6% (541 million tons). Notably, China’s steel output totaled 1019.1 million tons, yet EAF only accounted for 9.9% [2]. A critical environmental challenge of EAF operations is the generation of 15–20 kg/t steel of electric arc furnace dust (EAFD), a hazardous byproduct containing zinc (10–35 wt%), iron (20–40 wt%), and trace heavy metals (Pb and Cd) alongside halogen contaminants (Cl and F) [3,4,5,6]. As a result, EAFD is recognized as hazardous waste under international regulations, including China’s National Hazardous Waste Inventory, the US Environmental Protection Agency, and the EU Waste Framework Directive [7,8,9]. If directly landfilled or stored openly without appropriate treatment, heavy metal ions can easily contaminate the soil and groundwater through rainwater leaching [10]. Moreover, the fine particles in the dust may disperse with the wind, leading to atmospheric pollution [11]. Therefore, the treatment of EAFD to be used as a resource is not only necessary for the sustainable development of the metallurgical industry but also an urgent issue for environmental protection.

Currently, the primary recycling and treatment processes for EAFD are pyrometallurgy and hydrometallurgy [12,13]. The hydrometallurgical process employs different leaching solutions to separate zinc from the mixture, purify the leachate, and obtain zinc metal through the electrowinning of a leach solution [14,15]. A series of hydrometallurgical methods have been developed, such as acid leaching methods, alkali leaching methods, and salt leaching methods [16]. In contrast, the pyrometallurgical process reduces metal oxides, such as zinc and lead, in the EAFD to metal vapors, thereby separating zinc from the residue [17,18,19]. For example, in the Waelz kiln and rotary hearth furnace (RHF) processes, due to reducing conditions in the furnace, Zn vapors are formed, which are re-oxidized above their previous charges and captured as dust [20]. At present, these pyrometallurgical processes remain predominant in industrial applications. Coleti et al. [21] found that the reduction in self-reducing briquettes made of EAFD at 700–1100 °C occurred in two parts: The first step is controlled by phase boundaries, where the apparent activation energy (Ea) was in the range of 137.6–200.3 kJ/mol. The second step is controlled by diffusion, where Ea is in the range of 360.5–378.8 kJ/mol. Kim et al. [22] investigated the reduction reaction of zinc oxide with C in the presence of different additives and found that the Ea of the ZnO-C reaction system was 224 kJ/mol. Nevertheless, conventional pyrometallurgical techniques, exemplified by the Waelz kiln and rotary hearth furnace (RHF) processes, are constrained by significant limitations including excessive energy expenditure and substantial environmental burdens associated with greenhouse gas emissions and solid waste generation [20,23].

To circumvent the limitations of traditional techniques, microwave-assisted carbothermic reduction has recently gained considerable scientific attention as an innovative thermal processing alternative [24,25,26]. Microwave heating, a non-contact thermal processing technology, generates heat energy through the interaction of an electromagnetic field with a material’s dielectric properties. This method is characterized by rapid volumetric heating, material selectivity, and precise thermal control that can be adjusted in real time [27]. Critically, EAFD can be rapidly heated under microwave irradiation due to its dielectric properties, particularly its high dielectric loss tangent and electrical conductivity [28,29,30]. Omran et al. [31] systematically evaluated the microwave-assisted carbothermic reduction of zinc from different steelmaking dust sources using blast furnace sludge as a reductant. Under 1100 W and a holding time of 20 min, their comparative analysis of three industrial dust sources revealed a strong correlation between zinc removal efficiency and oxide composition (FeO and CaO), achieving a peak zinc recovery of 90.43%. Al-harahsheh et al. [19] demonstrated a novel microwave-assisted pyrolysis approach by co-processing EAFD with polyvinyl chloride (PVC). Their findings revealed a complete decomposition of refractory franklinite (ZnFe₂O₄) phases during microwave irradiation (800 W, 15 min), which was corroborated by the disappearance of characteristic spinel XRD peaks.

However, there is still a significant gap in the existing literature on the kinetic study of EAFD reduction in microwave fields. In this study, the reduction mechanisms governing EAFD–coke composite systems were systematically examined to elucidate their reaction behavior. Key experimental parameters, including temperature, holding time, and the carbon–zinc molar ratio, were investigated to quantify their effects on zinc removal efficiency and phase evolution. The isothermal kinetic behavior of the zinc removal rate was evaluated based on different kinetic models, followed by the identification of limiting steps governing the reduction mechanism of electric arc furnace dust (EAFD)–coke composite systems. This study offers insights into the optimization of zinc recovery efficiency during microwave-assisted carbothermic reduction in EAFD.

2. Materials and Methods

2.1. Raw Materials

The electric arc furnace dust (EAFD) was sourced from a Chinese steel manufacturing facility. Prior to experimental procedures, EAFD was dried in a drying oven at 105 °C for 12 h to eliminate water. The particle size distribution of EAFD was analyzed using a laser particle size analyzer (Mastersizer 3000, Malvern Instruments Ltd., Malvern, UK), while its microscopic morphology was characterized by SEM, as shown in Figure 1. A granulometric analysis revealed a particle size distribution characterized by D90 = 9.77 μm, D50 = 4.07 μm, and D10 = 2.12 μm, indicating that 90% of EAFD particles measure less than 9.77 μm, 50% measure less than 4.07 μm, and 10% measure less than 2.12 μm. SEM images support the idea that the EAFD microstructure distribution was characterized by irregularly shaped, highly porous aggregates less than 10 μm in diameter. This morphological characteristic was attributed to the rapid condensation of high-temperature flue gas during steelmaking processes [4]. The chemical composition of EAFD, shown in

Table 1. Chemical composition of EAFD (wt%).

Element	C	H	O	N	S	Cl	F	Fe	Zn	Si	Al	Ca	Pb	Mg	Mn	Na	K	Ni	Cr
Content	0.9	0.6	21.5	0.3	1.7	0.8	0.4	13.0	27.2	18.9	1.1	4.6	1.1	1.5	1.2	0.2	1.3	3.4	0.2

, was determined by XRF. The EAFD sample has a high concentration of zinc (Zn > 20 wt%) [32], while the concentrations of Fe, Si, and Ca are 13.0, 18.9, and 4.6 wt%, respectively.

The mineral phase composition of EAFD, shown in Figure 2, was determined by XRD. The analytical results demonstrated that zinc and iron constituted the predominant metallic elements, which primarily exist as franklinite (ZnFe₂O₄), zinc oxide (ZnO), and magnetite (Fe₃O₄).

The coke was sourced from sieved materials of an ironmaking plant, and an industrial proximate analysis yielded the following results: 86.89% fixed carbon (FCd), 1.70% volatile matter (Vd), and 11.41% ash content (Ad). To address potential reactivity limitations associated with the broad particle size distribution (2–5 mm) of raw coke, mechanical crushing followed by sieving was implemented to achieve a uniform particle size below 0.5 mm. Subsequently, coke was left in a drying oven at 80 °C for 24 h to effectively reduce the moisture content to <0.5 wt%.

2.2. Experimental Procedure

The EAFD and coke used in this experiment were both less than 0.5 mm, and the experiment was carried out in a self-designed microwave tube furnace (Figure 3) under a nitrogen atmosphere. The EAFD and coke were blended evenly, and the mixture was loaded into an alumina crucible and then into the furnace. The crucibles employed in the experiment were thermally treated in a muffle furnace at 200 °C for 3 h to eliminate moisture interference. A copper water-cooled condenser was positioned above the quartz reactor tube to condense the zinc vapor generated during the reduction process. During the experiment, a K-type thermocouple (temperature measurement accuracy: ±1.5 °C) was vertically inserted into the center of the mixture in order to measure temperature. The carbon proportion in the mixtures was determined based on the stoichiometric amount necessary for a complete reduction of zinc oxide (including zincite and zinc ferrite) in the system (ZnO + C). Subsequently, the carbon addition was adjusted to 3, 3.5, 4, 4.5, and 5 times the stoichiometric requirement. The experimental temperatures were set to 900, 950, 1000, 1050, and 1100 °C, where the samples were heated from ambient temperature, and the holding times were 0, 10, 20, 30, 40, 50, and 60 min, respectively. At the end of the experiment, the reduced sample was cooled to room temperature for analysis.

2.3. Methods

The thermodynamic feasibility of these reactions was evaluated using FactSage 7.2, with Gibbs free energy changes (ΔG) plotted as a function of temperature (Figure 4).

After cooling, the residues and products were collected and analyzed to determine the zinc removal rates under different conditions. X-ray diffractometry (XRD, 18 KW D/MAX 2500 V+/PC, Rigaku Corporation, Tokyo, Japan) was carried out to determine the main phase present in the sample. X-ray fluorescence (XRF, M4 Tornado, KAIYUE, Berlin, German) was used to determine the Zn content in the residues to calculate the zinc removal rate. The microscopic morphology of the products and the relative content of the elements were analyzed by scanning electron microscopy (SEM, Regulus 8230, HITACHI, Beijing, China).

The zinc removal rate of EAFD can be defined using Equation (1):

$$\alpha ={\displaystyle \frac{m_0{w}_0-m_t{w}_t}{m_0{w}_0}}\times 100\%$$

(1)

In this equation, α is the zinc removal rate (%), m₀ is the initial mass of the raw sample (g), w₀ is the zinc content in the raw sample (%), m_t is the mass of the post-experimental residue (g), and w_t is the zinc content in the treated residue (g).

In this study, the microwave heating technique was adopted, which resulted in rapid overall warming of the material being achieved within a few minutes; this allowed the system to rapidly enter and stabilize at the target temperature. This property allows for studying the carbothermal reduction mechanism of zinc in EAFD under near-ideal isothermal conditions. To further this investigation, the present study aimed to elucidate the kinetic control mechanism of the zinc reduction reaction in microwave fields by designing an isothermal kinetic experiment combining an unreacted nucleus model (UCM) with the fitting of the Arrhenius equation.

To evaluate the degree of reduction as a function of time under isothermal conditions, numerous unreacted kernel models were developed and selected for data fitting. Common kinetic models include the following [33,34,35,36,37,38]:

Chemical reaction control (α is the zinc removal rate (%)):

$$\mathrm{n}=1:-\ln (1-\alpha )=\mathrm{k}\mathrm{t}$$

(2)

$$\mathrm{n}=2:{(1-\alpha )}^{-1}-1=\mathrm{k}\mathrm{t}$$

(3)

$$\mathrm{n}=3:[{(1-\alpha )}^{-1}-1]/2=\mathrm{k}\mathrm{t}$$

(4)

Diffusion control:

$$\mathrm{Two}-\mathrm{dimensional}\mathrm{diffusion}:\alpha +(1-\alpha )\ln (1-\alpha )=\mathrm{k}\mathrm{t}$$

(5)

$${\mathrm{Three}-\mathrm{dimensional}\mathrm{diffusion}\left(\mathrm{Jander}\mathrm{function}\right):[1-{(1-\alpha )}^{1/3}]}^2=\mathrm{k}\mathrm{t}$$

(6)

$$\mathrm{Three}-\mathrm{dimensional}\mathrm{diffusion}\left(\mathrm{Ginstling}\text{–}\mathrm{Brounshtein}\mathrm{function}\right):(1-2\alpha /3)-{(1-\alpha )}^{2/3}=\mathrm{k}\mathrm{t}$$

(7)

Phase-boundary control:

$$\mathrm{One}\mathrm{dimension}:\alpha =\mathrm{k}\mathrm{t}$$

(8)

$$\mathrm{Cylindrical}\mathrm{symmetry}:1-{(1-\alpha )}^{1/2}=\mathrm{k}\mathrm{t}$$

(9)

$$\mathrm{Spherical}\mathrm{symmetry}:1-{(1-\alpha )}^{1/3}=\mathrm{k}\mathrm{t}$$

(10)

According to the Arrhenius equation, the relationship between the reaction rate constant and temperature can be expressed as follows:

$$\mathrm{k}=\mathrm{A}{\mathrm{e}}^{(-\mathrm{E}\mathrm{a}/\mathrm{R}\mathrm{T})}$$

(11)

In this equation, α is the zinc removal rate (%), k is the reaction rate constant (min⁻¹), A is the exponential prefactor (min⁻¹), Ea is the apparent activation energy (kJ/mol), R is the gas constant, at 8.314 J/(mol K), and T is the thermodynamic temperature, K.

A logarithmic transformation of the Arrhenius equation yields the following:

$$\ln \mathrm{k}=\ln \mathrm{A}-\mathrm{E}\mathrm{a}/\mathrm{R}\mathrm{T}$$

(12)

Plotting lnk against 1/T yields a linear relationship (Arrhenius curve), whose slope (−Ea/R) is used to calculate the apparent activation energy (Ea). After substituting α into Equations (2)–(10), the calculated reaction rate constant k is substituted into Equation (12) to obtain a linear fit. When the coefficient of determination (R²) for a kinetic model exceeds 0.95, the mechanism function (e.g., diffusion-controlled or chemical reaction-controlled) corresponding to that model is statistically identified as the rate-limiting step.

3. Results and Discussion

3.1. Thermodynamic Analysis

The reduction process of EAFD aims to convert zinc oxide (ZnO) and zinc ferrite (ZnFe₂O₄) into metallic zinc (Zn). Under microwave heating, a series of coupled reactions is initiated within and on the surface of the particles. During the initial heating stage, solid–solid reduction dominates, as it is driven by direct interactions between zinc-/iron-bearing phases in the EAFD and coke [20]. As the reaction progresses, CO gas generated through the partial oxidation of carbon diffuses into the particle interior, facilitating the gas–solid reduction in residual zinc and iron oxides [39]. Meanwhile, the Boudouard reaction (C + CO₂→2 CO) regenerates CO, maintaining a reducing atmosphere.

The thermodynamic calculations reveal that zinc ferrite (ZnFe₂O₄) is most easily reduced directly to zinc oxide (ZnO) and magnetite (Fe₃O₄) when the reaction initiates at 557 °C (Figure 4a), while metallic zinc (Zn) begins to form through ZnO reduction at 951 °C. The ΔG of all direct reduction processes decreases with increasing temperature, demonstrating that thermal elevation facilitates reduction kinetics. The initial reaction temperature of the carbon gasification reaction is 704 °C (Figure 4b). A comparison between direct and indirect reduction shows that ZnFe₂O₄, Fe₃O₄, and FeO react more readily with C. A progressive reduction of ZnFe₂O₄ leads to sequential phase transformations, generating intermediate iron oxides, such as magnetite (Fe₃O₄) and wustite (FeO), before ultimately producing metallic iron (Fe). At higher temperatures (>1000 °C), the equilibrium of the reaction shifts predominantly toward the formation of metallic Zn and Fe.

3.2. Effect of Holding Time on Zinc Removal Rate

As shown in Figure 5a, when the reduction temperature is fixed at 1000 °C and the C/Zn ratio is fixed at 4, the zinc removal rate demonstrates growth with a prolonged holding time, exhibiting a rapid increase during the initial stage (0–30 min) followed by slower growth in the subsequent stage (30–60 min). A zinc removal rate of 82.23% was achieved at 1000 °C after a 60 min reaction.

As demonstrated in Figure 5b, the initial phase composition (0 min) is dominated by ZnFe₂O₄ and ZnO. At 20 min, the ZnFe₂O₄ diffraction peak disappears with the emergence of the Fe₃O₄ phase and an attenuation of the ZnO peak, demonstrating a reduction of ZnFe₂O₄ and ZnO. At 40 min, the ZnO diffraction peaks continue to decrease. At 60 min, the ZnO peaks essentially disappear, indicating that ZnO is reduced completely.

3.3. Effect of Temperature on Zinc Removal Rate

As illustrated in Figure 6a, when the holding time was fixed at 60 min and the C/Zn ratio was fixed at 4, a temperature-dependent enhancement in zinc removal efficiency was systematically observed, progressing from 46.43% at 900 °C to 95.45% at 1100 °C. A notable increase in efficiency from 54.84% at 950 °C to 82.23% at 1000 °C was observed, resulting in substantial mass loss, which is attributed to the volatilization of zinc in the reduced sample under high-temperature conditions.

In addition, the phase evolution analysis in Figure 6b revealed a progressive structural transformation. At 900 °C, the initial phases were identified as ZnO and Fe₃O₄, with ZnO diffraction peaks dominating the pattern. Upon heating to 950 °C, a significant decrease in the peak intensity of ZnO was detected, coinciding with the initial efficiency improvement. At 1000 °C, the emergence of the FeO phase was observed alongside a complete reduction in ZnO. Further temperature increases to 1050 °C resulted in metallic Fe formation, while ZnO peaks were nearly undetectable. By 1100 °C, all ZnO signatures were eliminated, corresponding to the maximum zinc removal efficiency of 95.45%.

3.4. Effect of C/Zn Ratio on Zinc Removal Rate and Product

As shown in Figure 7a, when the holding time is fixed at 60 min and the temperature is fixed at 1100 °C, the zinc removal rate exhibits a staged enhancement: progressing from 67.95% at C/Zn = 3 to 88.66% at C/Zn = 3.5, ultimately reaching a stabilization plateau (95.45–96.66%) at C/Zn = 4 to C/Zn = 5. Notably, the efficiency exceeded 95% with little variation (less than 1%) when C/Zn ratios surpassed 4.0, indicating reaction saturation under excess-carbon conditions.

As demonstrated in Figure 7b, a sequential reduction mechanism was observed for varying carbon-to-zinc molar ratios (C/Zn). At a C/Zn ratio of 3, ZnFe₂O₄ diffraction peaks were eliminated, whereas ZnO signals were intensified, confirming the preferential reduction of ZnFe₂O₄ over ZnO under carbon-deficient conditions. When the ratio was increased to 3.5, an attenuation in the peak intensity of ZnO was observed, which occurred concurrently with the emergence of metallic iron (Fe), indicating progressive ZnO reduction. At C/Zn = 4, ZnO diffraction features were nearly eliminated, and Fe₃O₄ emerged as the main phase, demonstrating optimized reduction conditions for ZnO decomposition. Notably, at C/Zn = 4.5, distinct graphite diffraction peaks were detected with increasing intensities, indicating that carbon exceeded the stoichiometric requirements for metal oxide reduction.

At C/Zn = 3, the XRD pattern (Figure 8a) exclusively exhibited ZnO phases, with no detectable metallic zinc signatures. Corresponding SEM-EDS results (Figure 8b) revealed particles characterized by hexagonal prismatic morphology, smooth surfaces, and uniform oxygen distribution (O/Zn atomic ratio = 1.0), which is consistent with stoichiometric ZnO. This confirms that insufficient carbon (C/Zn = 3) fails to maintain a reducing atmosphere, which leads to the re-oxidation of Zn (i.e., Zn + CO₂ → ZnO + CO) [20].

When the ratio was increased to C/Zn = 4, metallic Zn diffraction peaks were first observed (Figure 8d), accompanied by an attenuation in the peak intensity of ZnO compared to the C/Zn = 3 sample. Concurrent SEM-EDS analyses (Figure 8e) identified a bimodal morphology comprising hexagonal prisms and rod-like structures (Zn/O ≈ 1.68), which are indicative of partial Zn re-oxidation. This confirmed that the increase in C/Zn could significantly suppress CO₂ generation and decrease Zn re-oxidation probability, facilitating the partial retention of metallic Zn.

At C/Zn = 5, the XRD pattern (Figure 8a) exclusively exhibited Zn phases, with no detectable ZnO signatures. SEM-EDS analyses (Figure 8h) demonstrated homogeneous rod structures with a highly concentrated zinc distribution. This confirms that excess carbon maintained a reducing atmosphere for the reaction and effectively suppressed CO₂ formation, reducing the likelihood of the re-oxidation of Zn to ZnO.

The impacts of reduction temperature and holding time on the zinc removal rate of EAFD are shown in Figure 9; at a high temperature of 1100 °C, even with a holding time of 0 min (i.e., a rapid temperature increase followed by immediate cooling), the zinc removal rate can still reach 24.14%, which is significantly higher than a removal rate of 2.27% at 900 °C with a holding time of 0 min. At a fixed holding time of 60 min, increasing the temperature from 900 °C to 1100 °C improved the zinc removal rate from 74% to 92% (an increase of 18%). When the temperature was increased to 1100 °C and the holding time was extended to 60 min, the zinc removal rate reached 95.45%, while that at 900 °C for 60 min was only 46.43%. When the reduction temperature was fixed at 1100 °C and the holding time was increased from 30 min to 60 min, the zinc removal rate increased by 26.66%, and when the holding time was fixed at 30 min and the reduction temperature was increased from 900 °C to 1100 °C, the zinc removal rate increased by 41.20%. This indicates that the reduction temperature exerts a more pronounced influence on the zinc removal rate than the holding time.

According to Figure 9, the variation in the zinc removal rate from 0 to 30 min and from 30 to 60 min differed greatly, and the reduction process had different limiting steps at different stages. k is calculated as the reaction rate constant according to Equation (11); the plots of lnk versus T for different kinetic equations were plotted with a straight line with a slope of −Ea/R, and the apparent activation energy of each kinetic equation is obtained by substituting the slopes into Equation (11). The results of fitting the lnk versus T for each kinetic equation with the apparent activation energy (Ea) are summarized in

Table 2. Results of fitting different kinetic models.

Mechanism		0~30 min		30~60 min
		Ea/kJ·mol−1	R2	Ea/kJ·mol−1	R2
Chemical reaction	Equation (2)	73.57	97.00%	145.18	93.74%
	Equation (3)	114.62	97.09%	267.32	96.88%
	Equation (4)	114.50	97.08%	267.29	96.84%
Diffusion	Equation (5)	126.68	98.03%	133.27	87.25%
	Equation (6)	146.50	98.24%	190.34	93.22%
	Equation (7)	133.26	98.11%	152.87	90.13%
Phase-boundary reaction	Equation (8)	37.05	92.18%	36.67	71.22%
	Equation (9)	54.76	94.62%	89.26	80.19%
	Equation (10)	60.84	96.45%	107.54	89.40%

. The Arrhenius plots obtained by reducing the EAFD at 900–1100 °C are shown in Figure 10.

According to Table 2, the three-dimensional diffusion model (Jander equation, Equation (5)) achieves a fit of up to 98.24% at the early stage of the reaction (0–30 min), indicating that the rate of the reaction at this stage is controlled by the diffusion process of zinc vapor through the product layer (Fe₃O₄/FeO). The apparent activation energy (Ea) is 146.50 kJ/mol. When the reaction enters the later stage (30–60 min), the fit of the chemical reaction control model (Equation (2)) improves to 96.88%, which can be explained by the contact probabilities of the reactants (C and ZnO) decreasing with increasing zinc removal rates. The apparent activation energy (Ea) is 267.32 kJ/mol. Compared with conventional heating methods (

Table 3. Reduction kinetics of EAFD under conventional methods Adapted with permission from Ref. [28]. 2025 Springer Nature.

Reductant	Heating Form	Controlling Mechanism	Ea (kJ/mol)	Reference
Charcoal	Electric heating	Phase-boundary	137.6–200.3	[21]
		diffusion	360.5–378.8
Graphite	Electric heating	One-dimensional diffusion	305.01	[40]
		Chemical reaction	315.67
		Chemical reaction	288.22

), microwave-assisted carbothermic reduction can reduce the activation energy needed for the zinc removal of EAFD.

4. Conclusions

In this study, the zinc removal mechanism, physical phase evolution characteristics, and the kinetic control mechanism in electric arc furnace dust (EAFD) were systematically revealed through a microwave-assisted carbothermic reduction experiment. The major conclusions from the experiment are as follows:

In the process of zinc reduction by EAFD, the zinc removal rate increases with increases in the reduction temperature and holding time, where the reduction temperature has a more significant effect on the zinc removal rate. With the increasing C/Zn ratio, the zinc removal rate and the purity of product (Zn) both increase. Under 1100 °C and C/Zn = 5 conditions, a zinc removal rate of 96.66% was achieved. Phase evolution follows a sequential mechanism: ZnFe₂O₄ is first reduced to Fe₃O₄ and ZnO, and then ZnO is reduced to Zn.

The reduction process of the EAFD occurred in two stages: the initial stage (0~30 min) was governed by the three-dimensional diffusion model (Ea = 146.5 kJ/mol), and the later stage (30~60 min) was controlled by the chemical reaction model (Ea = 267.32 kJ/mol).

Compared with the conventional heating form, microwave-assisted carbothermic reduction optimizes the pore structure and reaction path through non-thermal effects, providing a more efficient solution for the resource-oriented utilization of EAFD.