Converting Iron-Bearing Tailings from Recycling of Urban Steel Scrap to Direct Reduced Iron via Magnetic Separation Followed by Hydrogen Reduction Under Microwave Irradiation

Abstract

In this study, the feasibility of converting iron-bearing tailings from urban steel scrap recycling to value-added direct reduced iron (DRI) via magnetic separation followed by hydrogen reduction under microwave irradiation was investigated, with an emphasis on the effect of reduction temperature. The experimental results showed that by magnetic separation, the tailings sample with an iron content of 15.42 wt% could transit to a high-grade magnetic concentrate with an iron content of 60.04 wt% and good microwave absorption capability, as revealed by its short microwave penetration depth (D_p). After hydrogen reduction under microwave irradiation, the main iron-bearing phases, including magnetite, hematite, limonite, and martite, had stepwise deoxidation into metallic iron. As the reduction temperature increased from 750 °C to 1050 °C, the total iron content (TFe), reduction degree and iron metallization degree of the product increased rapidly and then became stable due to difficult reduction of FeO. As the reduction process proceeded, the dispersed iron particles gradually aggregated. At the optimum temperature of 950 °C, the reduction degree and iron metallization degree reached 90.10% and 88.71%, respectively. Meanwhile, the pore size, microporous volume, and specific surface area of the product were 1.943 nm, 1.767 × 10⁻⁵ cm³/g, and 0.3961 m²/g, respectively. The saturation magnetization (M_S) and coercivity (H_C) of the product remained 170.94 emu/g and 46.25 Oe, respectively. The product can act as a potential feedstock for electric arc furnace (EAF) steelmaking.

1. Introduction

As an essential metallic material, steel has become an indispensable infrastructure material in modern society due to its high strength, good plasticity, great toughness, etc. [1]. Currently, steel is predominantly produced through two steps [2], namely reduction of iron ore to metallic iron in a blast furnace (BF) or direct reduced iron (DRI) before further conversion to steel by decarbonization in a basic oxygen furnace (BOF) or electric arc furnace (EAF) [3]. In an EAF, steel scrap is typically utilized as a primary iron source with the addition of DRI for crude steel production [4,5]. This method is characterized by significantly lower energy consumption and reduced greenhouse gas emissions [6], primarily attributable to the elimination of energy-intensive sintering/pelletizing, coking, and BF operations [7]. With the rapid development of the steel industry, the depletion of high-grade iron ore resources, growing accumulation of steel scrap [8] and increasingly stringent low-carbon environmental protection requirements, it is essential to produce DRI from secondary iron-bearing resources.

Steel scrap is recognized as an important secondary resource for steelmaking [9]. Over the past few decades, the global steel scrap production has demonstrated substantial growth [10]. By 2050, the total amount of steel scrap available to the global steel industry is estimated to reach 760 million tons [11]. Furthermore, the global steel recycling percentage will keep increasing [12]. In China, the annual generation of steel scrap exceeds 120 million tons, constituting a great reservoir of recyclable iron resources [9].

Currently, steel scrap enrichment is primarily accomplished through established processing methods, including crushing, gravity separation, and other physical separation techniques [13]. The resulting tailings are generally utilized in cement production, construction work, or disposed of through landfilling [14], which not only results in significant waste of resources but also leads to severe environmental damage. Considering the tailings are mostly constituted by iron oxides, they may be reused as feedstock for producing DRI after reduction [15].

For converting iron-bearing tailings to DRI, it is critical to select a suitable reducing agent. From the perspective of eliminating carbon emissions [16,17], hydrogen gas is a good option [18,19,20] given its potential to enable near-zero CO₂ emissions. To intensify the reduction process, external fields, such as microwave, may also be used. Microwave refers to a category of electromagnetic radiation characterized by frequencies spanning from 300 MHz to 300 GHz, with corresponding wavelengths from 1 mm to 1 m [21,22,23]. It has a selective thermal effect, as revealed by the strong dependency of microwave absorptivity on the permittivity and permeability of the target material [24]. Owing to its relatively long wavelength, microwave can also dissipate throughout the entire volume of the material, producing a volumetric thermal effect that not only reduces processing time but also eliminates the temperature gradients commonly found in conventional heating [25]. Because of the above features, microwave heating has been applied for promoting endothermic reduction reactions of various metal ores and secondary metal-bearing resources which demand high thermal energy input [26,27].

This study aimed to explore the feasibility of producing DRI from iron-bearing tailings generated during recycling of urban steel scrap via magnetic separation followed by hydrogen reduction under microwave irradiation. The results proved that the tailings could serve as a good raw material to produce DRI with a high iron metallization degree under optimal reduction conditions.

2. Experimental Section

2.1. Raw Materials

The sample of iron-bearing tailings used in this study was provided by a steel scrap recycling plant. Its main chemical composition is listed in

Table 1. Main chemical composition of iron-bearing tailings (wt%).

Element	Fe	Si	C	Ca	Al	Mg	Zn
Content	15.42	12.32	11.02	6.74	2.49	1.54	1.45
Element	Na	K	S	P	Cr	Ni	LOI
Content	1.44	0.56	0.31	0.16	0.050	0.031	25.42

. The iron content of the tailings was only 15.42 wt%, with many impure elements, including 12.32 wt% Si, 11.02 wt% C, 6.74 wt% Ca, 2.49 wt% Al, 1.54 wt% Mg, 1.45 wt% Zn, and 1.44 wt% Na. According to

Table 2. Phase distribution of iron in iron-bearing tailings (wt%).

Existence Form	Magnetite	Martite	Hematite and Limonite	Wüstite	Silicates	Sulfides	Total
Iron content (wt%)	5.2	0.65	7.5	0.64	1.41	0.024	15.42
Proportion (%)	33.7	4.2	48.6	4.2	9.1	0.2	100

, iron was mainly present in the form of magnetite (Fe₃O₄) with strong magnetism, along with hematite, limonite, martite, and wüstite, showing the feasibility for magnetic separation to enrich iron before reduction.

As shown in Figure 1a, the identified phases in the tailings included quartz (SiO₂), magnetite (Fe₃O₄), calcite (CaCO₃), goethite (FeO(OH), fayalite (Fe₂SiO₄)), hematite (Fe₂O₃), nepheline (KNa₃Al₄(SiO₄)₄), dolomite (CaMg(CO₃)₂), etc. The magnetic properties of the tailings, including saturation magnetization (M_S) and coercivity (H_c), were also measured. As shown in Figure 1b, the tailings had an M_s of 6.235 emu/g and H_c of 113.74 Oe.

2.2. Methods

2.2.1. Experimental Procedure

Before the experiments, the sample of iron-bearing tailings was ground until its particle size was smaller than 74 μm. It was then subject to magnetic separation using a magnetic separation tube with intensity of 0.245 T. The resulting magnetic concentrate was pressed into cylindrical briquettes of 10 mm in diameter and 15 mm in height with the addition of water as a binder, followed by drying in a vacuum oven at 105 °C for 4 h. These dried briquettes were reduced in a microwave tube furnace (2.45 GHz, maximum power of 1200 W, Hunan Huae Microwave Technology Co. Ltd., Changsha, China) at different temperatures (750–1050 °C) for 30 min with a H₂ flow rate of 0.5 L/min. Subsequently, 0.5 L/min N₂ was introduced as a production gas during the cooling process. The cooled reduction products were collected for determining reduction performance and other characterizations.

2.2.2. Instrumental Characterizations

The chemical composition of magnetic concentrate was determined using an inductively coupled plasma optical emission spectrometer (ICP-OES, ICAP7400 Radial, ThermoFisher Scientific, Waltham, MA, USA). The phase distribution of iron in the concentrate was determined by chemical titration. The phase compositions of the concentrate and the non-magnetic residue after magnetic separation were determined using an X-ray diffraction spectrometer (XRD; D8 Advance, Bruker, Karlsruhe, Germany). The particle size distribution of the concentrate was determined using a laser particle size analyzer (Malvern Mastersizer 2000, Malvern Instruments Ltd., Worcestershire, UK). The microstructural features of the concentrate were determined using a scanning electron microscope (SEM; Sigma HD, Zeiss, Oberkochen, Germany) equipped with an energy-dispersive spectrometer (EDS; EDAX Inc., Mahwah, NJ, USA). The thermogravimetry–derivative thermogravimetry-differential scanning calorimetry (TG-DTG-DSC) analysis of the concentrate was carried out using a simultaneous thermal analyzer (TG-DSC; STA 449F3, Netzsch, Selb, Germany) with a ramp rate of 10 °C in air. The thermal stability of the concentrate under microwave irradiation was determined using a microwave thermogravimetric furnace operated at 2.45 GHz with a given power level (CY-TH1200C-M, Hunan Change Microwave Technology Co., Ltd., Changsha, China). The complex relative permittivity (ε_r) and permeability (µ_r) of the concentrate at room temperature were determined with a mass ratio of the sample powder to microwave transparent paraffin of 7/3 using a vector network analyzer (E5071C, Agilent Technology Co., Ltd., Palo Alto, CA, USA) by the transmission line method in the frequency range 2–4 GHz. The values of the microwave penetration depth (D_p) of the concentrate were calculated by Equation (1) using the measured values of permittivity and permeability [28].

$$D_p={\displaystyle \frac{{\lambda }_0}{2\sqrt{2}\pi }}{\left\{{\epsilon }_r^{″}{\mu }_r^{″}-{\epsilon }_r^{\prime }{\mu }_r^{\prime }+{\left[({\epsilon }_r^{\prime }{\mu }_r^{\prime }{)}^2+({\epsilon }_r^{″}{\mu }_r^{″}{)}^2+({\epsilon }_r^{\prime }{\mu }_r^{″}{)}^2+({\mu }_r^{\prime }{\epsilon }_r^{″}{)}^2\right]}^{{\displaystyle \frac{1}{2}}}\right\}}^{-{\displaystyle \frac{1}{2}}}$$

(1)

where λ₀ is the microwave wavelength in free space, m; ε_r′ and ε_r″ are the relative dielectric constant and relative dielectric loss factor, respectively; µ_r′ and µ_r″ are the relative magnetic constant and relative magnetic loss factor, respectively.

The thermodynamic analysis of hydrogen reduction of magnetic concentrate obtained by magnetic separation of iron-bearing tailings was performed using the software FactSage 8.1 (Thermfact/CRCT, Montreal, QC, Canada; GTT-Technologies, Herzogenrath, Germany).

For assessing the reduction performance, a few indexes, including total iron content (TFe), reduction degree (RD) and iron metallization degree (η), were used. Among them, the reduction degree was calculated using the following equation:

$$RD=\left({\displaystyle \frac{0.111w_1}{0.430w_2}}+{\displaystyle \frac{m_1-m_t}{m_0\times {0.430w}_2}}\right)\times 100\%$$

(2)

where w₁ is the FeO content of the concentrate before reduction, wt%; w₂ is the total iron content of the concentrate before reduction, wt%; m₀ is the weight of the concentrate before reduction, g; m₁ is the weight of the concentrate after reduction, g; and m_t is the weight of the concentrate at time t, g.

The iron metallization degree was calculated using the following equation:

$$\eta =\left({\displaystyle \frac{w_{MFe}}{w_{TFe}}}\right)\times 100\%$$

(3)

where w_MFe is the metallic iron content of metallized pellets, wt%, and w_TFe is the total iron content of metallized pellets, wt%. Both were determined by chemical titration.

The phase compositions of the reduction products were determined using the same apparatus as that for the concentrate. The semiquantitative analysis of the crystalline phase contents of the reduction products was conducted by Rietveld full-spectrum fitting of the obtained XRD patterns using the software Jade V8.8 (Materials Data Inc., Livermore, CA, USA). The microstructural features of the reduction products were determined using the same scanning electron microscope mentioned above. The BET specific surface area and micropore volume of the reduction products were analyzed using a specific surface area and porosity analyzer (BET; ASAP 2460, Micromeritics Instrument Co., Ltd., Atlanta, GA, USA). The magnetic properties of the reduction products, including M_S and H_c, were measured using a vibrating sample magnetometer (VSM; LakeShore 7404, LakeShore, Columbus, OH, USA).

3. Results and Discussion

3.1. Properties of Magnetic Concentrate

The chemical composition of magnetic concentrate obtained after magnetic separation of iron-bearing tailings from urban steel scrap recycling, with an iron yield of 8.98% and iron recovery of 40.24%, is shown in

Table 3. Main chemical composition of magnetic concentrate (wt%).

Element	Fe	Si	C	Ca	Al	Mg	Zn	Na	K	S	P
Content	60.04	0.75	1.95	0.83	0.34	0.22	1.17	0.041	0.033	0.13	0.11

. The iron content of the concentrate reached 60.04 wt%, mainly existing in the forms of Fe₃O₄ (56.76%), Fe₂O₃ (34.05%) and FeCO₃ (8.63%), as shown in

Table 4. Phase distribution of iron in magnetic concentrate (wt%).

Existence Form	Magnetite	Martite	Hematite and Limonite	Carbonates	Silicates	Sulfides	Total
Iron content (wt%)	34.08	15.26	5.18	4.1	1.32	0.1	60.04
Proportion (%)	56.76	25.42	8.63	6.83	2.20	0.16	100

. It also contained 1.95 wt% C, 1.17 wt% Zn, 0.83 wt% Ca, 0.75 wt% Si, 0.34 wt% Al, 0.22 wt% Mg, 0.13 wt% S and 0.11 wt% P.

The main phase components of the concentrate were Fe₃O₄, Fe₂O₃, FeO(OH) and Fe₂SiO₄, as shown in Figure 2a, in agreement with the chemical titration results (Table 4). The main phase components of the non-magnetic residue obtained after magnetic separation of iron-bearing tailings were SiO₂, CaCO₃, and CaMg(CO₃)₂, as shown in Figure 3. According to the particle size distribution analysis (Figure 2b), the magnetic concentrate had a median particle diameter (D₅₀) of only 7.38 μm, with more than 90% of its particles (D₉₀) smaller than 82.57 μm, as shown in. Figure 4 shows the SEM-EDS mapping analysis of the concentrate. Evidently, Fe exhibited a spatial distribution similar to O, confirming the presence of iron oxides and silicates.

The thermal stability of the concentrate determined using the simultaneous thermal analyzer is shown in Figure 5. The weight loss process was divided into four stages. In the first stage, the weight loss was 0.58%, due to the loss of free water from the mineral, with a slight heat absorption peak. In the second stage, the weight loss reached 4.5%, attributed to the decomposition of iron carbonate, with a heat absorption peak formed at 200–300 °C (FeCO₃ → FeO + CO₂) [29,30]. In the third stage, a weight loss of 2.57% occurred, primarily ascribed to the removal of crystal water and the dehydroxylation of goethite (2FeO(OH) → Fe₂O₃ + H₂O). There were two heat absorption peaks that occurred at 416.8 °C and 515.8 °C, respectively. In the fourth stage, a weight loss of 6.01% was found, associated with the decomposition of calcium carbonate [28,31], as partially revealed by a broad heat absorption peak spanning from 600 °C to 900 °C (CaCO₃ → CaO + CO₂). The thermogravimetric data indicated that the concentrate exhibited good thermal stability until about 900 °C, above which it might be reduced for producing DRI.

As mentioned before, the determination of the microwave absorption capability of magnetic concentrate based on the permittivity and permeability measurements could reveal its microwave reduction feasibility. As shown in Figure 6, within the frequency range 2–4 GHz, the magnetic concentrate had minor changes in ε_r′, ε_r″, µ_r′, and µ_r″. Notably, ε_r″ and µ_r″ had close values, suggesting that both dielectric loss and magnetic loss would control initial microwave heating of the concentrate. Specifically, at the commonly used frequency, 2.45 GHz, the concentrate had ε_r′ of 2.91, ε_r″ of 0.29, µ_r′ of 0.95, and µ_r″ of 0.34. Its microwave penetration depth was calculated to be 27 mm at room temperature, indicating its good microwave absorption capability [28].

To further investigate the microwave absorptivity of the concentrate, microwave thermogravimetric tests were conducted at 1800 W with a hydrogen flow rate of 0.5 L/min. As shown in Figure 7, the temperature of the concentrate under microwave irradiation reached 1100 °C within 33 min and the weight loss could be divided into four stages based on the DTG curve. The first stage, between 100 °C and 200 °C, involved the removal of free water, resulting in a weight loss of 0.58%. The second stage, between 200 °C and 700 °C, showed a slower weight loss, attributed to the removal of crystal water, decomposition of iron carbonate, and dehydroxylation of goethite, with a weight loss of 11.82%. In the third stage, from 700 °C to 900 °C, the weight loss rate increased, which was a result of the reduction of iron oxides such as Fe₃O₄ to FeO, producing a weight loss of 21.2%. The fourth stage, from 900 °C to 1100 °C, exhibited a lower weight loss rate, likely due to the slower reduction of FeO, with a weight loss of 9.7%. These results confirmed that the concentrate exhibited a strong microwave response, consistent with the microwave absorption capability analysis.

3.2. Thermodynamic Analysis

According to the above analysis, the main iron-bearing phases of the concentrate were Fe₃O₄, FeO(OH), Fe₂SiO₄, and Fe₂O₃. From a thermodynamic perspective, Fe₃O₄ undergoes a reduction pathway, progressing from Fe₃O₄ to FeO, and ultimately to Fe, once the temperature exceeds 570 °C. Therefore, a higher temperature is necessary to effectively ensure the reduction of iron oxides. To further investigate the conversions of iron oxides in the concentrate during reduction, the Reaction module of FactSage 8.1 was used. As shown by the temperature (T) dependence of standard Gibbs free energy changes (△_rG_m^Θ) of relevant reactions in Figure 8a, with increasing temperature, iron oxides are gradually reduced to Fe. Concurrently, CaO from CaCO₃ decomposition reacts with SiO₂ to form CaSiO₃, which subsequently reacts with MgO to form CaMgSiO₄. According to Figure 8b, the partial pressure of hydrogen required for reduction of FeO to Fe decreases with increasing temperature. Therefore, to obtain high-quality DRI, appropriate reaction temperature and hydrogen partial pressure should be selected.

3.3. Microwave Reduction Analysis

The effect of temperature on the hydrogen reduction performance of magnetic concentrate under microwave irradiation is shown in Figure 9. When the reduction temperature increased from 750 °C to 1050 °C, the TFe and content of metallic iron increased from 80.97 wt% and 63.90 wt% to 89.32 wt% and 79.25 wt%, respectively. Correspondingly, the reduction degree and iron metallization degree increased from 81.05% and 78.91% to 90.37% and 88.73%, respectively. The content of metallic iron increased by only 1.2% when the temperature increased from 750 °C to 850 °C, consistent with the TG-DTG-DSC analysis that showed a small weight loss between 750 °C and 850 °C. The reduction degree and iron metallization degree reached 90.10% and 88.71% at 950 °C, increased by 9.22% and 8.10% compared to the counterparts at 850 °C. This agreed with the presence of a weight loss peak around 950 °C in the TG-DTG-DSC analysis. With further increase in temperature from 950 °C to 1050 °C, the increase in reduction degree and iron metallization degree became less obvious. It indicated that the reduction kept proceeding with a slower rate because FeO limited the reaction progress, agreeing with the TG-DTG-DSC analysis in which the weight loss rate gradually decreased until the temperature approached 1050 °C.

Along with changes in the reduction indexes, evident phase transformations might take place. To verify this, the XRD patterns of the products generated after hydrogen reduction of magnetic concentrate obtained by magnetic separation of iron-bearing tailings at different temperatures were obtained, as shown in Figure 10. When the temperature was 750 °C, the main phases were Fe and FeO, along with a small amount of CaMgSiO₄. Increasing reduction temperature facilitated the reaction process, which was proven by a decline in the intensity of the diffraction peaks of FeO and enhanced intensity of the diffraction peaks of Fe. When the temperature exceeded 950 °C, the diffraction peaks of FeO disappeared, indicating its complete conversion.

To further examine the reduction process, Figure 11 shows the SEM-EDS analysis of the products obtained by hydrogen reduction of magnetic concentrate under microwave irradiation at 750 °C, 850 °C, 950 °C, and 1050 °C. Note that for brevity, detailed EDS spectra of different spots are not provided. Obviously, there were multiple phases in the reduction products. With rising temperature, a greater number of Fe particles were observed in the absence of detectable FeO. The formation of CaMgSiO₄ originated from the thermal decomposition of CaCO₃, where the liberated CaO subsequently reacted with SiO₂ in the presence of trace magnesium impurities. As indicated by the dotted circles in Figure 12, these impurities no longer remained scattered when the temperature increased.

To further investigate the hydrogen reduction behavior of the concentrate under microwave irradiation, Figure 13 shows the nitrogen adsorption–desorption isotherms of the reduction products. As revealed by the small loops [28], the reduction products had small and close average pore sizes at 750 °C, 850 °C, 950 °C, and 1050 °C, namely 1.943 nm, 1.942 nm, 1.943 nm, and 1.936 nm, respectively. In terms of pore volume, it decreased with increasing temperature, showing a more compact internal structure probably due to the growth of metallic iron particles.

For elucidating the reduction mechanism, the specific surface areas and pore volumes of the reduction products were determined using the multi-point BET method and t method, respectively [32,33], as shown in Figure 14. As the temperature increased from 750 °C to 1050 °C, the specific surface area of the reduction product decreased from 1.946 m²/g to 0.2764 m²/g. Meanwhile, the pore volume declined from 3.81 × 10⁻⁴ cm³/g to 9.421 × 10⁻⁶ cm³/g. Note that at the temperature with the optimal reduction indexes (Figure 9), 950 °C, the two parameters were 1.767 × 10⁻⁵ cm³/g and 0.3961 m²/g, respectively. These changes indicated a more compact structure of the product with increasing reduction temperature, probably attributed to the continuous accumulation of metallic iron particles produced from stepwise reduction of iron oxides [34].

The magnetic properties of the reduction products could be evaluated by measuring M_S and H_C, as shown in Figure 15. The magnetic concentrate had limited magnetism before reduction, as revealed by its low M_s values. After reduction at 750 °C, its reduction product exhibited a much higher M_s value. As the temperature increased continuously, there were minor increases in M_s, consistent with the variation in the iron metallization degree. The other parameter, H_c, decreased from 52.1 Oe to 29.3 Oe, probably due to the growth of iron particles. Note that after reduction at 950 °C, the M_s and H_c of the product remained at 170.94 emu/g and 46.25 Oe, respectively.

Overall, the above results proved that the iron-bearing tailings could be a valuable raw material for preparing DRI. After magnetic separation, the reduction product had the reduction degree of 90.10% and iron metallization degree of 88.71% under optimal reduction conditions, namely reduction temperature of 950 °C, dwell time of 30 min, and hydrogen flow rate of 0.5 L/min. The iron metallization degree of the product met the People’s Republic of China Ferrous Metallurgy Industry Standard for DRI used in steelmaking (YB/T 4170-2008) [35], showing its potential as an EAF feedstock. It is thus highly promising to use the method developed in the present study for processing iron-bearing tailings from urban steel scrap recycling, offering both environmental and economic benefits.

4. Conclusions

The feasibility of converting iron-bearing tailings from urban steel scrap recycling to DRI by magnetic separation followed by hydrogen reduction under microwave irradiation was investigated, with a focus on the effect of reduction temperature. The results showed that a magnetic concentrate with an iron content of 60.04 wt% could be obtained after magnetic separation. By hydrogen reduction under microwave irradiation, the main iron-bearing phases in the concentrate, such as magnetite, hematite, limonite, and martite, were reduced to metallic iron. When the reduction temperature increased from 750 °C to 1050 °C, the total iron content, reduction degree and iron metallization degree of the product increased considerably and then stabilized due to the difficulty of achieving full reduction of FeO. Dispersed iron particles gradually accumulated, producing a more compact structure which hindered H₂ diffusion to some extent, in agreement with the BET analysis. The reduction performance was also verified by the VSM test, which demonstrated the high saturation magnetization and coercivity of the product. At the optimum temperature of 950 °C, the reduction degree and iron metallization degree of the product reached 90.10% and 88.71%, respectively. The pore size, pore volume, and specific surface area were 1.943 nm, 1.767 × 10⁻⁵ cm³/g, and 0.3961 m²/g, respectively. The saturation magnetization and coercivity remained at 170.94 emu/g and 46.25 Oe, respectively. The results verified the feasibility of using iron-bearing tailings from urban steel scrap recycling for producing DRI to achieve evident environmental and economic benefits.