Modification of Copper Slag Composite with Water-Quenched Silicon-Manganese Slag

Abstract

Metallurgical industries often discharge slag containing valuable elements that are poorly utilized when producing copper alloys and silicon-manganese alloys. To improve the utilization rate, in this study, a method to mix copper slag with water-quenched silicon-manganese slag and CaO for roasting and modification was proposed. In this work, FactSage 8.0, DSC-TG, and XRD were used to examine the phase change during the modification process and investigate the impacts of the CaO content, roasting temperature, and holding time on the modification effect. The results showed that the addition of water-quenched silicon-manganese slag and CaO could effectively promote the transformation of fayalite to (Mn, Mg, Fe)Fe₂O₄, with the highest conversion rate occurring at a 10% CaO content. An increase in the temperature and prolongation of the time facilitated fayalite transformation, but excessive temperature or time could result in iron loss. The optimal recovery rate and iron grade were achieved with roasting at 1400°C for 60 min. This method can provide a concentrate suitable for producing copper-containing antibacterial stainless steel and wear-resistant cast iron, and the tailings can be used to produce ceramic materials.

1. Introduction

Copper slag (CS) is a byproduct of copper alloy production, with nearly 2–3 tons of CS generated per ton of copper alloy produced.^1,2) In China, the annual CS emissions were 22.126 million tons for the entire year of 2022.³⁾ Domestic CS is currently mainly used in small amounts and can mostly only be heaped up and stored. Such open piles of CS not only waste the rich valuable metal in them but also severely pollute the environment.⁴⁾ CS contains valuable metals, including gold, silver, copper, iron, and zinc. The iron content in CS is approximately 35%, which is much greater than that in general iron ore.¹⁾ Fayalite (Fe₂SiO₄) is the name for the iron-rich phase of CS, and recovering iron from it is quite challenging because of its exceedingly small grain size, generally stable physicochemical characteristics, and interspersion with other mineral phases.⁵⁾ After fayalite in CS is reduced to metallic iron or oxidized to magnetite, it can be used as a raw material for ironmaking with high application value.^6,7) However, the reduction process produces a large amount of CO₂, and there are safety issues in the storage and use of the reduction gas.^8,9) Furthermore, oxidation of fayalite to magnetite requires strict conditions and may generate α-Fe₂O₃, which is difficult to recover through magnetic separation.^10,11) Cement mixed with CS has improved mechanical properties due to the excellent cohesion of CS,^12,13) but the free CaO and free MgO in CS can undermine the cement stability.^14,15,16,17)

Water-quenched silico-manganese slag (WSS) is an acidic alloy slag mainly composed of SiO₂, CaO, MnO, and MgO. In the first half of 2022, 7.2742 million tons of WSS emissions were recorded.^18,19) Currently, WSS is mostly utilized to make building materials such as mineral wool, aggregates, and permeable bricks. However, the building material application necessitates a significant number of additional chemical additives.^20,21,22) Considering the chemical composition of WSS, it can be blended into CS as an additive to promote the conversion of Fe₂SiO₄ in CS.

To address the above issues, we added WSS to CS for co-roasting modification. In this process, CaO, MnO, and MgO from WSS underwent chemical reactions with Fe₂SiO₄ in CS. (Mn, Mg, Fe)Fe₂O₄ was formed and could be separated by a magnetic force. The concentrates can be used to produce copper-containing antibacterial stainless steel and wear-resistant cast iron, and the tailings are suitable for ceramics or building materials. The modification process stabilizes free CaO, free MnO, and free MgO, thus facilitating iron precipitation after enrichment and effectively addressing the instability of tailings in the production of building materials. The liquid phase created by the melting of WSS, which has a low melting point, encourages particle migration and speeds up the modification reaction in the low-temperature stage. The CaO, MgO, and MnO in WSS can transform the Fe₂SiO₄ in CS into (Mn, Mg, Fe)Fe₂O₄ during roasting of CS and WSS in air. This not only enables the recovery of iron from CS but also lowers the amount of carbon needed during the reduction process to remove iron. In this study, the phase transition was analyzed through methods such as X-ray diffraction (XRD) and scanning electron microscopy (SEM)‒energy-dispersive X-ray spectroscopy (EDS), and the effects of the CaO content, roasting temperature, and holding time were investigated based on magnetic separation. The theoretical study of the modification process can be improved through thermodynamic calculations.

2. Experimental Section

2.1. Materials

The experimental materials used were CS (Shandong Hengbang Metallurgical Co., Ltd., 35.49 μm), WSS (Ningxia Baoliyuan Special Alloy Co., Ltd., 54.78 μm), and CaO (China National Pharmaceutical Group Chemical Reagent Co., Ltd., 74 μm). Table 1 and Fig. 1 present the chemical and phase compositions of the slag materials. Iron in CS mainly exists as fayalite and magnetite. Magnetite can be recovered via magnetic separation, while fayalite cannot be recovered due to its weak magnetism. The WSS contains a small quantity of Ca(OH)₂, which breaks down at approximately 400°C, and the released CaO can speed up the modification reaction.

Table 1. Chemical composition of the raw materials.

Slag	Components (wt%)
	SiO2	Al2O3	CaO	MnO	MgO	Fe2O3	P2O5	Pb	Cu	Others
CS	28.15	3.23	2.12	0.061	1.71	34.21	1.12	0.86	2.57	25.97
WSS	41.33	21.64	16.49	7.32	5.21	0.55	0.0097	0	0	7.3803

Fig. 1. XRD patterns: (a) CS; (b) WSS.

CS and WSS were combined based on a chemical composition of 1:1, and CaO was then added at a specific ratio. Samples C1-C4 were prepared and kept after thorough mixing. Table 2 provides the chemical compositions of the mixed samples and the related roasting schedules.

Table 2. Sample number and roasting conditions.

Num.	CS:WSS:CaO (%)	Components (wt%)						RTa) (°C)	HTb) (min)	CRc) (°C/min)	STd) (°C)
		SiO2	Fe2O3	Al2O3	MnO	MgO	CaO
C1	43:57	43	22	15	5	4	11	1400	60	1	1150
C2	43:57:5	41	21	14	4	4	16	1400	60	1	1150
C3	43:57:10	39	20	13	4	4	20	1400	60	1	1150
C4	43:57:15	37	19	13	4	4	23	1400	60	1	1150

Footnotes: a-Roasting temperature; b-Holding time; c-Cooling rate; d-Sampling temperature.

2.2. Methods

2.2.1. Process Flow

In this experiment, we mixed the raw materials and heated them to the molten state, simulating the actual conditions in industrial slagging. The specific process is shown in Fig. 2, and the numbering and roasting conditions of the samples are shown in Table 3. First, the raw material was mixed with agate grinding beads and anhydrous ethanol, followed by mixing on a ball mill for up to 12 hours. After the mixing was completed, a rotary evaporator was used to remove the anhydrous ethanol from the raw materials. Next, the samples were roasted at 1300°C, 1350°C, 1400°C, or 1450°C for 60 min to investigate the effect of temperature on the modification or at 1400°C for 30 min, 60 min, 90 min, or 120 min to investigate the effect of time on the modification. After completion of roasting, the samples were removed and allowed to cool to ambient temperature in air. Finally, the samples were poured into wet magnetic separation tubes for magnetic separation (particle size: 74 μm, magnetic field: 225 mT), which lasted for 10 min to achieve separation of the concentrate and tailings. The total iron content in the samples was determined according to the GB/T6730.73-2016 “EDTA Photometric Titration Method”, and the magnetic separation rate and recovery rate were calculated according to Eqs. (1) and (2):

  

$${\beta }_1=\frac{m_1}{m_0}\times 100\%\mathrm{\hspace{0.17em} };\mathrm{\hspace{0.17em} }{\beta }_2=\frac{m_2}{m_0}\times 100\text{\%}$$(1)

  

$$p=\frac{n_1\times {\beta }_1}{n_1\times {\beta }_1+n_2\times {\beta }_2}$$(2)

β1: the concentrate yield; β2: the tailing yield; m1: the concentrate mass; m2: the tailing mass; m0: the feed mass of modified slag subjected to magnetic separation; P: the final recovery ratio of iron; n1: the iron grade of the concentrate; n2: the iron grade of the tailing.

Fig. 2. Modification process. (Online version in color.)

Table 3. Sample number and roasting conditions.

Num.	RMa)	RTb) (°C)	HTc) (min)	CRd) (°C/min)	STe) (°C)	Num.	RM	RT (°C)	HT (min)	CR (°C/min)	ST (°C)
M1	C3	1300	60	1	1150	N1	C3	1400	30	1	1150
M2	C3	1350	60	1	1150	N2	C3	1400	60	1	1150
M3	C3	1400	60	1	1150	N3	C3	1400	90	1	1150
M4	C3	1450	60	1	1150	N4	C3	1400	120	1	1150

Footnotes: a-Raw material; b-Roasting temperature; c-Holding time; d-Cooling rate; e-Sampling temperature.

2.2.2. Sample Characterization

We used XRD to analyze the mineral phases in the sample. The parameters for the XRD scan were set to those for a Cu target, and a step scan method was used with a scan angle ranging from 10° to 70°. Each step was maintained for 2 seconds, with a step size of 0.02°. The voltage was set to 40 kV, and the current was set to 30 mA. To observe the microstructure and morphology of the sample, we used SEM with EDS and backscattered electron (BSE) mode at an electron beam acceleration voltage of 15 kV; the particle size distribution of each sample in the SEM images was measured using ImageJ software. In a comprehensive thermal analyzer (differential scanning calorimetry (DSC)-thermogravimetry (TG)), the sample was heated to a maximum temperature of 1400°C at a heating rate of 10°C/min in an air atmosphere and maintained for 60 min.

2.2.3. Thermodynamic Calculations

A thermodynamic analysis was performed using the composition provided in Table 2. In FactSage8.0, the temperature range was set to 1000–1500°C, the oxygen partial pressure was set to 0.21 atm, the total environmental pressure was set to 1 atm, and the databases used were FactPS and FToxid.

3. Results and Discussion

3.1. Thermodynamic Feasibility Analysis

(1) The modification reaction can be conducted in air. Figure 3(a) shows that when the temperature reaches 1400°C, (Mn, Mg, Fe)Fe₂O₄ starts to form in the system. This result indicates that the MnO and MgO enriched in WSS enables the conversion of the Fe₂O₃ produced by overoxidation in the system to (Mn, Mg, Fe)Fe₂O₄ with the purpose of enhancing the recovery. Figure 3(b) demonstrates that in the absence of CaO, the phase diagram contains a significant SiO₂ region. The SiO₂ region noticeably decreases and then completely disappears when CaO is introduced. This demonstrates that CaO consumes SiO₂ in the system to more efficiently release Fe from Fe₂SiO₄. However, as the CaO concentration increases, the monoxide region in the high-temperature phase area expands, which causes the spinel phase region to shrink. This suggests that altering the CaO ratio is essential for fine-tuning the spinel phase region.

Fig. 3. Thermodynamic diagram: (a) phase-stability diagram of Fe–Mn–Mg–O at 1400°C; (b) without CaO; (c) CaO/(Fe₂SiO₄+MnO+MgO)=0.1; (d) CaO/(Fe₂SiO₄+MnO+MgO)=0.2. (Online version in color.)

(2) According to the above thermodynamic analysis, the CaO content can change the spinel phase region, in addition to affecting the solid-phase line temperature. The main mineral phases of sample a include the spinel, CaMgSi₂O₆, CaAl₂Si₂O₈, and the liquid phase (Slag-liq). However, the solid-phase line temperature is high at this point (up to 1085°C), which can hinder the generation of the spinel because the liquid phase can effectively promote the reaction of Fe₂O₃ with MnO and MgO to generate the spinel.^23,24) In samples b-d, the solid-phase line temperature of sample c decreases by 56°C compared to that of sample a with the gradual increase in the CaO content, whereas the solid-phase line temperature of sample d increases by 26°C. This suggests that only a moderate amount of CaO can play a role in reducing the solidus line temperature. In addition, the generation temperature of the spinel increases with increasing amount of CaO. For example, in Fig. 3(d), which corresponds to the highest content of CaMgSi₂O₆, the presence of CaMgSi₂O₆ adversely affects the formation of the spinel because many phases are formed during the transition of the sample from the molten state to the semimolten state and then to the solid state, one of which is CaMgSi₂O₆. Since CaMgSi₂O₆ has a high melting point, its presence rapidly increases the viscosity of the system, which ultimately inhibits the formation of the spinel.^25,26)

3.2. Effect of CaO on the Modification

Figure 5 illustrates the phase composition of C1–C4. There is no Fe₂SiO₄ in C1, and (Mn, Mg, Fe)Fe₂O₄ is the main iron-containing phase, indicating that (Mn, Mg, Fe)Fe₂O₄ was formed by the transformation of Fe₂SiO₄ during the reaction. (Mn, Mg, Fe)Fe₂O₄ has a weak peak intensity in C1. CaO affects the (Mn, Mg, Fe)Fe₂O₄ composition, as evidenced by the peak intensity in C2–C4 increasing and then decreasing with the addition of CaO.²⁷⁾ The change in the strength of the (Mn, Mg, Fe)Fe₂O₄ peak is due to the increase in the number of reaction points in the system with the addition of CaO, which can convert Fe₂SiO₄ into Fe₃O₄ and CaSiO₃ according to Eq. (2), at which time the resulting CaSiO₃ can react with MgO and SiO₂ to form CaMgSi₂O₆, which has been fully verified thermodynamically, according to Eq. (8). At the same time, some of the Fe₃O₄ is overoxidized to Fe₂O₃, which then reacts with MgO and MnO₂ according to Eqs. (6) and (7) to produce MgFe₂O₄ and MnFe₂O₄. However, excess CaO leads to a significant increase in the CaMgSi₂O₆ content, as shown in Fig. 6(d). CaMgSi₂O₆ is a high-melting-point compound, and its excessive content leads to a decrease in the mobility of the slag, thus making the (Mn, Mg, Fe)Fe₂O₄ reaction difficult. This observation can be corroborated by the thermodynamic elevation of the solidus line, the increase in the CaMgSi₂O₆ content, and the difficulty of elemental diffusion due to excess CaO.^28,29)

Fig. 5. XRD patterns of C1–C4. (Online version in color.)

Fig. 6. SEM images of modified slags with different additions of CaO: (a) C1; (b) C2; (c) C3; (d) C4. (Online version in color.)

Figure 6 illustrates the microstructure morphology of the samples, and the atomic percentage of each mineral phase is given in Table 4. In Fig. 6(a), the main iron-containing phase is (Mn, Mg, Fe)Fe₂O₄, which has an irregular polygonal morphology and is embedded in the silicate phase with a sparse distribution, which is not favorable for Fe recovery. According to the analysis of the particle size statistics, the particle size of (Mn, Mg, Fe)Fe₂O₄ gradually increases with increasing amount of CaO, while the shape of Ca(Mg, Al, Fe)Si₂O₆ changes from the previous polygonal shape to an elongated shape. With increasing amount of CaO, the viscosity of the slag decreases, the fluidity is enhanced, and the elemental diffusion is enhanced.^30,31) For example, the higher contents of Mn and Fe in Table 4 c1 indicate that Mn and Fe can be better enriched in (Mn, Mg, Fe)Fe₂O₄ under the action of CaO, which promotes its growth. As the CaO content continues to increase, as shown in Fig. 6(c), (Mn, Mg, Fe)Fe₂O₄ grows into a bulk of approximately 27 μm, which is a size that is very favorable for better iron recovery.³²⁾ However, although CaO promotes the growth of (Mn, Mg, Fe)Fe₂O₄, excess CaO leads to an increase in the grain size of CaMgSi₂O₆. Figure 6(d) shows that the grain size of CaMgSi₂O₆ is larger, which obviously inhibits the growth of (Mn, Mg, Fe)Fe₂O₄. This implies that too much CaO affects the recovery of (Mn, Mg, Fe)Fe₂O₄, which is consistent with the thermodynamic calculations shown in Fig. 4(d).

Table 4. Atomic percentages at the spectral points in Fig. 6.

Samples	No.	Element (%)
		O	Si	Al	Ca	Mn	Mg	Fe	Cu
a	1	67.54	0	5.24	0.12	1.79	8.41	16.9	0
	2	69.45	11.66	3.97	6.71	0.41	5.08	2.72	0
	3	71.49	12.26	4.01	7.62	1.1	1.78	0.96	0.78
b	1	67.99	0	4.94	0.08	2.02	7.42	17.55	0
	2	70.01	11.85	4.43	6.79	0.54	3.95	2.43	0
	3	72.15	12.8	4.51	6.92	0.95	0.95	0.96	0.76
c	1	66.99	0	3.62	0.08	2.6	5.35	21.36	0
	2	68.36	13.38	4.26	5.62	0.78	5.17	2.43	0
	3	70.53	15.99	7.17	2.25	0.73	1.59	0.98	0.76
d	1	68.01	0	5.46	0.09	1.88	8.32	16.24	0
	2	69.3	11.43	4.1	6.93	0.33	4.69	3.22	0
	3	70.66	11.79	3.46	9.48	0.85	2.27	1.11	0.38

Fig. 4. Equilibrium diagram: (a) C1; (b) C2; (c) C3; (d) C4. (Online version in color.)

3.3. Effect of Temperature on the Modification

The phase compositions of the samples roasted at different temperatures are given in Fig. 7. The mineral phases of M1–M4 consist mainly of Fe₃O₄, MnFe₂O₄, MgFe₂O₄ and CaMgSi₂O₆. The intensities of the Fe₃O₄, MnFe₂O₄, MgFe₂O₄ and CaMgSi₂O₆ peak gradually increase with increasing temperature, which indicates that an increase in the temperature can increase the reaction rates of Eqs. (4), (6), (7) and (8). The reason for this is that an increase in the temperature helps increase the dissociation rate of Fe₂SiO₄ and the reactivity of FeO.^33,34,35) With increasing temperature, the peak intensity does not continue to increase indefinitely. A significant amount of Fe₃O₄ has been produced in the system by the time the temperature reaches 1400°C, which increases the viscosity of the fluid and prevents the particles from migrating.³²⁾ As illustrated in Fig. 7(b), the characteristic peaks of Fe₃O₄ and MgFe₂O₄ gradually shift toward higher angles as the temperature increases. Excessively high temperatures can lead to iron loss due to the replacement of Fe³⁺ by Al³⁺.³⁶⁾

Fig. 7. XRD patterns of M1–M4. (Online version in color.)

The microstructure of the samples is depicted in Fig. 8. M1 exhibits many small (Mn, Mg, Fe)Fe₂O₄ particles and dendritic crystals of CuO dispersed in the silicate phase. According to the thermodynamic analysis in Fig. 4, when the temperature is low, the content of the liquid phase in the system is low, and the growth of crystals is mainly affected by the oxidation reaction at the gas‒solid interface. Since the mass transfer of particles in the solid phase is low compared to that in the liquid phase, the number of small particles of (Mn, Mg, Fe)Fe₂O₄ is high, and some of the CuO particles show a dendritic distribution, which leads to unstable growth.³⁷⁾ As the temperature increases, the (Mn, Mg, Fe)Fe₂O₄ grain size increases, and the CuO dendrites disappear, indicating that liquid phase augmentation supports stable crystal growth. For the M3 sample, the system produces more liquid phase. With the increase in the amount of the liquid phase, the additional pressure at the bending liquid level promotes interparticle migration, which in turn leads to an increase in the densification of the (Mn, Mg, Fe)Fe₂O₄ grains. According to Table 5, the Fe/(Mn+Mg) ratio increases and then decreases with increasing temperature, while the Al content continues to increase. This result indicates that the amount of Al dissolved in (Mn, Mg, Fe)Fe₂O₄ gradually increases with increasing temperature, especially in the M4 sample, where the most Al is incorporated. This increase in the amount of dissolved Al not only darkens the color of the sample but also causes iron loss, and this conclusion is confirmed by XRD analysis.

Fig. 8. SEM images of slags modified at different temperatures: (a) M1; (b) M2; (c) M3; (d) M4. (Online version in color.)

Table 5. Atomic percentages at the spectral points in Fig. 8.

Samples	No.	Element (%)
		O	Si	Al	Ca	Mn	Mg	Fe	Cu
a	1	69.36	0	3.39	0.12	2.21	5.43	17.64	1.85
	2	70.49	10.75	4.48	6.64	0.48	3.63	3.53	0
	3	71.5	12.04	4.22	6.99	0.86	0.53	0.88	2.98
b	1	69.91	0	3.45	0	2.19	6.6	17.85	0
	2	71.1	12.02	4.61	6.22	0.67	3.26	2.12	0
	3	72.1	12.11	4.75	7.18	0.99	0.71	1.16	1
c	1	66.99	0	3.62	0.08	2.6	5.35	21.36	0
	2	68.36	13.38	4.26	5.62	0.78	5.17	2.43	0
	3	70.53	15.99	7.17	2.25	0.73	1.59	0.98	0.76
d	1	66.27	0	6	0	1.73	8.06	17.94	0
	2	69.63	14.52	5.82	4.82	0.77	2.73	1.44	0.27

3.4. Effect of the Holding Time on the Modification

Figure 9 displays the phase compositions of N1 to N4. When the holding time is extended to 60 min, the intensity of the (Mn, Mg, Fe)Fe₂O₄ peak significantly increases. The intensity of the (Mn, Mg, Fe)Fe₂O₄ peak still increases after 90 min of holding, although it increases more slowly. However, the intensity of the (Mn, Mg, Fe)Fe₂O₄ peak slightly decreases when the holding time is extended to 120 min, suggesting that the holding time affects the content of (Mn, Mg, Fe)Fe₂O₄. Longer holding times allow for complete melting of the sample and promote a homogeneous reaction in the slag, leading to sufficient reaction according to Eqs. (4), (6), (7), and (8).³⁸⁾ However, a longer time does not necessarily mean that the peak intensity will increase. In contrast, the peak intensity slightly decreases when the time is too long. This occurs because at 60 min, a large amount of (Mn, Mg, Fe)Fe₂O₄ has been generated in the system, which adheres to the surface of the unreacted modified slag, leading to a decrease in the contact area of Fe₂SiO₄ with the oxidizing atmosphere, which in turn prevents the oxidation reaction from proceeding. This reaction process is consistent with the thermodynamic feasibility of Eq. (9).

Fig. 9. XRD patterns of N1–N4. (Online version in color.)

The structures of the samples at the microscopic level are illustrated in Fig. 10. (Mn, Mg, Fe)Fe₂O₄ particles of 16 μm are dispersed within the silicate matrix after roasting at 1400°C for 30 min. However, the small size and sporadic distribution of these particles leads to poor recovery in the magnetic separation process. When the holding time is prolonged, the size of (Mn, Mg, Fe)Fe₂O₄ increases, and the particles gradually tend to agglomerate. However, when the holding time is too long, the size of (Mn, Mg, Fe)Fe₂O₄ significantly decreases, as in the case of (Mn, Mg, Fe)Fe₂O₄ in Fig. 10(d), which decreases to approximately 21 μm. Moreover, according to the data in Table 6, too long a holding time leads to a significant decrease in the iron content in the samples, which suggests that prolongation of the holding time promotes breaking of the silicate anionic bonds and enhances the fluidity of the slag, which leads to a larger size of (Mn, Mg, Fe)Fe₂O₄. However, too long of a holding time likewise promotes the growth of the silicate phase, leading to a lower yield of (Mn, Mg, Fe)Fe₂O₄.

Fig. 10. SEM images of slags modified for different times: (a) N1, (b) N2, (c) N3, and (d) N4. (Online version in color.)

Table 6. Atomic percentages at the spectral points in Fig. 10.

Samples	No.	Element (%)
		O	Si	Al	Ca	Mn	Mg	Fe	Cu
a	1	67.41	0	4.31	0.11	2.23	6.84	19.1	0
	2	68.44	11.76	4.89	7.27	0.46	3.46	3.72	0
	3	70.49	13.31	5.11	8.04	0.73	0.65	1.13	0.54
b	1	66.99	0	3.62	0.08	2.6	5.35	21.36	0
	2	68.36	13.38	4.26	5.62	0.78	5.17	2.43	0
	3	70.53	15.99	7.17	2.25	0.73	1.59	0.98	0.76
c	1	67.44	0	4.37	0	2.24	6.76	19.19	0
	2	70.06	12.92	4.83	7.35	0.78	2.1	1.66	0.3
	3	70.66	12.99	4.74	7.76	0.92	0.94	1.24	0.75
d	1	67.44	0	4.28	0	2.31	6.81	19.16	0
	2	69.05	12.03	4.63	7.06	0.49	3.75	2.99	0
	3	71.38	12.95	4.83	7.45	0.91	0.73	1	0.75

4. Magnetic Separation

Figure 11 shows the magnetic separation results. The addition of CaO increases the recovery rate from 58.32% to 72.56% and the iron grade from 33.12% to 38.49%. However, a further increase in the CaO addition decreases the recovery rate and iron grade to 52.41% and 31.38%, respectively. This suggests that CaO promotes the generation of (Mn, Mg, Fe)Fe₂O₄; therefore, the recovery and iron grade first increase with increasing CaO addition. However, when the CaO concentration is too high, the reaction kinetics deteriorate, which leads to a decrease in the recovery rate and iron grade.^39,40,41) In addition, the temperature positively affects the recovery rate and iron grade. The recovery rate increases from 62.22% to 72.56% and then slowly increases to 75.21%, while the iron grade increases from 31.19% to 38.49% and then decreases to 34.37%. This result suggests that high temperatures contribute to the conversion of Fe₂SiO₄ to (Mn, Mg, Fe)Fe₂O₄, and hence, the recovery and iron grade first increase with increasing temperature. However, when the temperature is too high, Al³⁺ will excessively replace Fe³⁺, as shown in Fig. 7(b), which leads to a decrease in the iron grade. Finally, an extension of 30–90 min can benefit the decomposition of Fe₂SiO₄, increasing the recovery rate from 60.34% to 76.03% and the iron grade from 30.18% to 40.01%. However, extending the time to 120 min decreases the recovery rate and iron grade to 71.01% and 35.32%, respectively. This means that a large amount of (Mn, Mg, Fe)Fe₂O₄ is generated within 30–90 min, and this (Mn, Mg, Fe)Fe₂O₄ adheres to the surface of the unreacted CS and interferes with the dissociation of Fe₂SiO₄ by CaO.

Fig. 11. Magnetic separation of modified slag. (Online version in color.)

Fig. 12. (a) Gibbs free energy diagram; (b) DSC-TG results. (Online version in color.)

Table 7 shows the chemical composition of the concentrate. Modified Si exists in the silicate phase, realizing selective separation of Fe and Si. Therefore, the concentrate can be used as a raw material for the production of copper-containing antimicrobial stainless steel and wear-resistant cast iron.

Table 7. Chemical composition of the concentrate.

Type	Brand	Components (%)
		C	Si	Mn	Ni	Cr	Cu	P	S	Mo
1a)	06Cr18NiCu2	≤0.07	≤1	≤2	8–11	17–19	1.5–2.5	–	–	–
	06Cr18NiCu3	≤0.07	≤1	≤2	8–11	17–19	2.5–4.0	–	–	–
2b)	HBW600Cr20Mo2Cu	2.6–2.9	≤1	≤1	≤1	18–21	1.4–2	≤0.06	≤0.06	1.4–2
3c)	Concentrate	–	1.3	1.8	–	0.6	1.5	–	–	–

Footnotes: a-Copper-containing antibacterial stainless steel; b-Wear-resistant cast iron; c-Concentrate.

5. Reaction Mechanisms

The iron-containing phase that is difficult to recover from CS is Fe₂SiO₄. Fe₂O₃ and SiO₂ are created when Fe₂SiO₄ interacts with O₂. Fe₂O₃, which is a mineral phase with poor magnetic properties, affects the subsequent magnetic separation.⁴²⁾ Meanwhile, Fe₂SiO₄ can be regenerated through a high-temperature reaction between SiO₂ and free Fe₃O₄. To address these issues, CS was doped with CaO and WSS, which caused Fe₂SiO₄ to be converted into Fe₃O₄ and CaSiO₃ according to Eq. (4). Fe₃O₄ is beneficial for increasing iron recovery; however, its crystal structure is unstable, and it can be oxidized into Fe₂O₃ at high temperatures.¹⁰⁾ Notably, MgO and MnO₂ in WSS can react with Fe₂O₃ to create (Mn, Mg)Fe₂O₄, which is a highly magnetic mineral phase facilitating magnetic separation and increasing the recovery efficiency.⁴³⁾ We discovered via XRD analysis that the predominant iron-containing phase of the C3 sample is (Mn, Mg, Fe)Fe₂O₄, which supports the viability of the reaction mechanism. When the temperature exceeds 720°C, the Gibbs free energy of Eq. (6) is positive, suggesting the possibility of unreacted free MgO existing in the system. The Gibbs free energy of Eq. (8) is fortunately negative below 1400°C. The remaining MgO can therefore continue to interact with CaSiO₃ and SiO₂ to create CaMgSi₂O₆. By further depleting the free MgO, CaMgSi₂O₆ completely resolves the stability issue in tailings applications. The main chemical equations involved in the reaction are as follows:

  

$$2\mathrm{F}{\mathrm{e}}_2\mathrm{S}\mathrm{i}{\mathrm{O}}_4+{\mathrm{O}}_2=2\mathrm{F}{\mathrm{e}}_2{\mathrm{O}}_3+2\mathrm{S}\mathrm{i}{\mathrm{O}}_2$$(3)

  

$$3\mathrm{F}{\mathrm{e}}_2\mathrm{S}\mathrm{i}{\mathrm{O}}_4+3\mathrm{C}\mathrm{a}\mathrm{O}+{\mathrm{O}}_2=2\mathrm{F}{\mathrm{e}}_3{\mathrm{O}}_4+3\mathrm{C}\mathrm{a}\mathrm{S}\mathrm{i}{\mathrm{O}}_3$$(4)

  

$$4\gamma -\mathrm{F}{\mathrm{e}}_3{\mathrm{O}}_4+{\mathrm{O}}_2=6\alpha -\mathrm{F}{\mathrm{e}}_2{\mathrm{O}}_3$$(5)

  

$$\mathrm{F}{\mathrm{e}}_2{\mathrm{O}}_3+\mathrm{M}\mathrm{g}\mathrm{O}=\mathrm{M}\mathrm{g}\mathrm{F}{\mathrm{e}}_2{\mathrm{O}}_4$$(6)

  

$$2\mathrm{F}{\mathrm{e}}_2{\mathrm{O}}_3+2\mathrm{M}\mathrm{n}{\mathrm{O}}_2=2\mathrm{M}\mathrm{n}\mathrm{F}{\mathrm{e}}_2{\mathrm{O}}_4+{\mathrm{O}}_2$$(7)

  

$$\mathrm{C}\mathrm{a}\mathrm{S}\mathrm{i}{\mathrm{O}}_3+\mathrm{M}\mathrm{g}\mathrm{O}+\mathrm{S}\mathrm{i}{\mathrm{O}}_2=\mathrm{C}\mathrm{a}\mathrm{M}\mathrm{g}\mathrm{S}{\mathrm{i}}_2{\mathrm{O}}_6$$(8)

  

$$\mathrm{F}{\mathrm{e}}_2\mathrm{S}\mathrm{i}{\mathrm{O}}_4+\mathrm{C}\mathrm{a}\mathrm{O}+1/3{\mathrm{O}}_2=\mathrm{C}\mathrm{a}\mathrm{S}\mathrm{i}{\mathrm{O}}_3+2/3\mathrm{F}{\mathrm{e}}_3{\mathrm{O}}_4$$(9)

Experimental tests were conducted on the temperature ramp curve during the reaction process to confirm the accuracy of the XRD results. The DSC curve shows three endothermic peaks at 404°C, 1087°C, and 1208°C and one exothermic peak at 1397°C. Ca(OH)₂ in the WSS decomposes into H₂O and CaO when heated, and dehydration leads to weight loss in the TG curve at 404°C. During sample preparation, CaO reacts with CO₂ in the air to form CaCO₃, which decomposes at 1087°C, releasing CO₂, which leads to weight loss in the TG curve. Due to the action of CO₂ released from CaCO₃, Fe₂O₃ thermally decomposes into Fe₃O₄ at the same time, which accounts for the weight loss in the TG curve at 1208°C. The exothermic peak at 1397°C, in contrast to the endothermic peak, is induced by the interaction of MnO₂ with Fe₂O₃, during which the released O₂ causes weight loss in the TG curve. This finding is consistent with the XRD results and confirms the thermodynamic stability of Eq. (7).

6. Conclusion

The concept of “waste for waste” is used in this study along with the principles of roasting and oxidizing Fe₂SiO₄, combining CS with WSS, and then adding a certain amount of CaO for modification treatment. The modification treatment can enhance the utilization of WSS, which is high in CaO, MnO, and MgO, in addition to the efficient recovery of the valuable elements in CS. The major conclusions are summarized as follows:

(1) In the context of the experimental study, adding 10% CaO when the ratio of CS and WSS is 1:1 will result in the best recovery and iron grade.

(2) When roasting is performed at a temperature of 1400°C, the particle size of (Mn, Mg, Fe)Fe₂O₄ can increase to approximately 27 μm, which is best for magnetic separation.

(3) Large amounts of (Mn, Mg, Fe)Fe₂O₄ can be produced in the system after being kept at 1400°C for 60 min. After magnetic separation, these materials can be classified into (Mn, Mg, Fe)Fe₂O₄ and tailings, which are the two main categories. The tailing can be used to make ceramics or building materials, while (Mn, Mg, Fe)Fe₂O₄ can be utilized as a raw material to make cast iron that is wear resistant and stainless steel that contains copper for its antibacterial properties.

Acknowledgments

The authors would like to thank the Ningxia Natural Science Foundation (Grant No. 2022AAC03254), the National Science Foundation (Grant No. 51964001), the Key Research and Development Project of Ningxia Hui Autonomous Region (Grant No. 2020BDE03001), and the Key Research and Development Plan of Ningxia Autonomous Region (Grant No. 2022BDE02002) for their financial support.

Conflict of Interest Statement

The authors declare no conflicts of interest.

References

• 1)   G.  Aleyna,  B.  Mikail and  K.  Elif Uzun: Resour. Policy, 80 (2023), 103232. https://doi.org/10.1016/j.resourpol.2022.103232
• 2)   G.  Wang,  H.  Zhang,  J.  Wang and  Q.  Xue: ISIJ Int., 62 (2022), 1. https://doi.org/10.2355/isijinternational.ISIJINT-2021-136
• 3)  Annual Data, National Bureau of Statistics of China, https://data.stats.gov.cn/easyquery.htm, (accessed 2023-03-01).
• 4)   L.  Xu,  Y.  Liu,  M.  Chen,  N.  Wang,  H.  Chen and  L.  Liu: Constr. Build. Mater., 375 (2023), 130982. https://doi.org/10.1016/j.conbuildmat.2023.130982
• 5)   P.  Jiang,  J.  Liu,  Y.  Xiao,  X.  Tan and  W.  Liu: J. Iron Steel Res. Int., 27 (2020), 796. https://doi.org/10.1007/s42243-020-00413-0
• 6)   J.  Du,  F.  Zhang,  J.  Hu,  S.  Yang,  H.  Liu and  H.  Wang: J. Cleaner Prod., 363 (2022), 132546. https://doi.org/10.1016/j.jclepro.2022.132546
• 7)   J.  Hao,  Z.  Dou,  T.  Zhang,  B.  Jiang,  K.  Wang and  X.  Wan: J. Cleaner Prod., 375 (2022), 134202. https://doi.org/10.1016/j.jclepro.2022.134202
• 8)   W.  Xiao,  S.  Yao,  S.  Zhou,  Y.  Wei,  B.  Li and  H.  Wang: J. Alloys Compd., 903 (2022), 163751. https://doi.org/10.1016/j.jallcom.2022.163751
• 9)   R.  Jiao,  P.  Xing,  C.  Wang,  B.  Ma and  Y.  Chen: Int. J. Miner., Metall. Mater., 24 (2017), 974. https://doi.org/10.1007/s12613-017-1485-3
• 10)   Q.  Li,  F.  Yang,  Z.  Wang and  S.  Liu: Trans. Indian Inst. Met., 72 (2019), 3223. https://doi.org/10.1007/s12666-019-01788-9
• 11)   S.  Anna: Metall. Mater. Trans. B., 43 (2012), 56. https://doi.org/10.1007/s11663-011-9576-4
• 12)   A.  Kumar and  M. L.  Tejaswini: Mater. Today: Proc., 60 (2022), 646. https://doi.org/10.1016/j.matpr.2022.02.264
• 13)   M.  Sridharan and  T. C.  Madhavi: Mater. Today: Proc., 46 (2021), 3225. https://doi.org/10.1016/j.matpr.2020.11.195
• 14)   J.  Zhao,  Q.  Liu and  K.  Fang: Mater. Lett., 280 (2020), 128528. https://doi.org/10.1016/j.matlet.2020.128528
• 15)   Z.  Chen,  J.  Liu and  R.  Li: Mater. Lett., 273 (2020), 127930. https://doi.org/10.1016/j.matlet.2020.127930
• 16)   L.  Huang,  Z.  Yang and  S.  Wang: Constr. Build. Mater., 262 (2020), 120776. https://doi.org/10.1016/j.conbuildmat.2020.120776
• 17)   S. L. T.  Jamilla Emi,  S.  Aecio Guilherme,  P.  Patricio Moreira,  F. C. B.  Veronica Teixeira and  M.  Henrique Barbosa: Transp. Res. Rec., 2673 (2019), 506. https://doi.org/10.1177/0361198119835513
• 18)   X.  Miao,  Z.  Bai,  G.  Qiu,  S.  Tang,  M.  Guo,  F.  Cheng and  M.  Zhang: J. Eur. Ceram. Soc., 40 (2020), 3249. https://doi.org/10.1016/j.jeurceramsoc.2020.02.029
• 19)  Mysteel: Silicomanganese 2022 H1 Market Review and H2 Outlook, Mysteel.com, https://thj.mysteel.com/22/0725/10/187A6E5A605EE96D.html, (accessed 2022-07-25).
• 20)   N.  Susanta Kumar,  R.  Navneet Singh and  K.  Sanjay: Resour., Conserv. Recycl., 176 (2021), 105946. https://doi.org/10.1016/j.resconrec.2021.105946
• 21)   P.  Tamayo,  G.  Angel,  J.  Setién,  A.  Soto and  C.  Thomas: Constr. Build. Mater., 364 (2022), 129938. https://doi.org/10.1016/j.conbuildmat.2022.129938
• 22)   P.  Lavanya and  N.  Venkatrao: Mater. Today: Proc., (2023). https://doi.org/10.1016/j.matpr.2023.03.438
• 23)   Y.  Gao,  J.  Wang,  X.  Tian and  X.  Hou: Materials, 16 (2023), 2086. https://doi.org/10.3390/ma16052086
• 24)   Y.  Wang,  R.  Zhu,  Q.  Chen,  G.  Wei,  S.  Hu and  Y.  Guo: ISIJ Int., 58 (2018), 2191. https://doi.org/10.2355/isijinternational.ISUINT-2018-533
• 25)   J. K.  Russell and  D.  Giordano: Geochim. Cosmochim. Acta, 69 (2005), 5333. https://doi.org/10.1016/j.gca.2005.06.019
• 26)   J.  Yang,  J.  Zhang,  Y.  Sasaki,  O.  Ostrovski, C, Zhang, D. Cai and Y. Kashiwaya: ISIJ Int., 56 (2016), 574. https://doi.org/10.2355/isijinternational.ISIJINT-2015-583
• 27)   L.  Kong,  J.  Bai,  X.  Wen,  X.  Li,  Z.  Bai,  Z.  Guo and  H.  Li: Fuel, 158 (2015), 968. https://doi.org/10.1016/j.fuel.2015.02.055
• 28)   A. K.  Verma and  B. B.  Karki: Am. Mineral., 97 (2012), 2049. https://doi.org/10.2138/am.2012.4123
• 29)   Z.  Zuo,  Q.  Yu,  J.  Liu,  Q.  Qin,  H.  Xie,  F.  Yang and  W.  Duan: ISIJ Int., 57 (2017), 220. https://doi.org/10.2355/isijinternational.ISIJINT-2016-402
• 30)   X.  Dai,  X.  Gan and  C.  Zhang: Trans. Nonferrous Met. Soc. China., 13 (2003), 1451. https://doi.org/JournalArticle/5af22558c095d718d8ef8b4d
• 31)   G.  Zhang,  N.  Wang,  M.  Chen and  Y.  Cheng: ISIJ Int., 60 (2020), 602. https://doi.org/10.2355/isijinternational.ISIJINT-2019-173
• 32)   H.  Zhang,  B.  Li,  Y.  Wei,  H.  Wang and  Y.  Yang: Metall Mater Trans B., 53 (2022), 1538. https://doi.org/10.1007/s11663-022-02464-y
• 33)   J.  Ku,  L.  Zhang,  W.  Fu,  S.  Wang,  W.  Yin and  H.  Chen: J. Hazard. Mater., 410 (2021), 124630. https://doi.org/10.1016/j.jhazmat.2020.124630
• 34)   Y.  Sun,  H.  Zhou,  K.  Yin,  M.  Zhao,  S.  Xu and  X.  Lu: J. Geophys. Res.: Solid Earth, 123 (2018), 3667. https://doi.org/10.1029/2018JB015452
• 35)   Z.  Wang,  B.  Peng,  L.  Zhang,  Z.  Zhao,  D.  Liu,  N.  Peng,  D.  Wang,  Y.  He,  Y.  Liang and  H.  Liu: JOM, 70 (2018), 539. https://doi.org/10.1007/s11837-017-2699-6
• 36)   Z.  He,  X.  Hu and  K.  Zhou: Process Saf. Environ. Prot., 165 (2022), 487. https://doi.org/10.1016/j.psep.2022.07.044
• 37)   B.  Li,  X.  Du,  Y.  Shen,  Z.  Zhang and  T.  Rong: Metall. Mater. Trans. B, 53 (2022), 1816. https://doi.org/10.1007/s11663-022-02491-9
• 38)   Z.  Guo,  D.  Zhu,  J.  Pan,  F.  Zhang and  C.  Yang: JOM, 70 (2018), 533. https://doi.org/10.1007/s11837-017-2671-5
• 39)   P.  Vadasz,  K.  Tomasek and  M.  Havlik: Arch. Metall., 46 (2001), 279. https://doi.org/10.1016/j.clae.2004.08.003
• 40)   R.  Zheng,  J.  Lu,  W.  Song,  M.  Liu,  H.  Li,  Y.  Liu,  X.  Lu and  Z.  Ma: Int. J. Miner., Metall. Mater., 30 (2023), 886. https://doi.org/10.1007/s12613-022-2569-2
• 41)   G.  Zhang,  N.  Wang,  M.  Chen and  Y.  Cheng: ISIJ Int., 60 (2020), 1863. https://doi.org/10.2355/isijinternational.ISIJINT-2019-670
• 42)   E.  Rivera,  R. A.  Munoz-Meneses,  L.  Marin,  M.  Mora,  J. A.  Tabares,  M.  Manotas-Albor,  L. A.  Rodriguez,  J. E.  Diosa and  E.  Mosquera-Vargas: Mater. Sci. Eng. B, 288 (2023), 116170. https://doi.org/10.1016/j.mseb.2022.116170
• 43)   Q. F.  Shu and  Y.  Liu: Ironmaking Steelmaking, 45 (2018), 363. https://doi.org/10.1080/03019233.2016.1274463