Effect of Fe₂O₃ Content and Acid on the Leaching Behavior of Phosphorus from Dephosphorization Slag

Abstract

Dephosphorization slag contains considerable quantities of valuable components, such as P₂O₅ and FeO_x. To recover P from dephosphorization slag, selective leaching has been adopted to separate the P-concentrating mineral phase. In this study, the effect of Fe₂O₃ content in slag and acid on the leaching behavior of P from dephosphorization slag was investigated. It was found that a higher Fe₂O₃ content in slag resulted in a higher P₂O₅ content in the C₂S–C₃P solid solution. Increasing the Fe₂O₃ content in slag promoted the dissolution of P and simultaneously suppressed the dissolution of other elements, facilitating the selective leaching of P. In the hydrochloric acid solution, more than 81% of P could be dissolved from dephosphorization slag at pH 4, and the dissolution ratio of Fe was nearly zero, achieving excellent selective leaching. Although better selective leaching was also realized in the citric acid solution at pH 5, hydrochloric acid was considered the appropriate leaching agent from the perspective of leaching cost. Through selective leaching, almost all the C₂S–C₃P solid solution was dissolved from dephosphorization slag, and the Fe-bearing matrix phase and magnesioferrite remained in the residue. The residue with low P₂O₅ content can be reutilized in ironmaking or steelmaking processes.

1. Introduction

Due to the shortage of natural resources and the deterioration of the ecological environment, the recovery of valuable components from metallurgical wastes and by-products is attracting increasing attention [1]. Dephosphorization slag, a major by-product of the steelmaking process, is generated from the dephosphorization of hot metal by adding metallurgical fluxes. The productivity of dephosphorization slag is approximately 0.10~0.15 tons per ton of molten steel [2]. Due to the enormous output of steel in China, the disposal and utilization of such large amounts of dephosphorization slag has become a crucial problem for the iron and steel industry [3]. Dephosphorization slag was principally composed of a CaO–SiO₂–FeO_x–P₂O₅ slag system and its slag basicity was lower than that of traditional converter slag [4]. At present, dephosphorization slag is normally applied as a concrete material or in road construction and landfills [5,6]. However, these applications are limited, and it is difficult to achieve the large-scale utilization of dephosphorization slag.

Dephosphorization slag contains considerable quantities of valuable components, such as FeO_x, MnO, and CaO, and its ideal utilization approach is via recycling in ironmaking or steelmaking processes as a flux [7]. However, the existence of P₂O₅ in slag hinders its recycling because phosphorus (P) is one of the most detrimental impurities in steels and the recycling of dephosphorization slag will lead to P enrichment in hot metal [8]. To achieve its resource utilization and exploit its potential value, it is necessary to separate P from dephosphorization slag. From another perspective, P is a nutrient element and a significant ingredient in fertilizers [9]. If the separation of P from dephosphorization slag becomes technically feasible, P could be developed as a new resource, and the P-removal slag could be effectively utilized in metallurgical processes, which would not only recover the valuable components but also drastically reduce the amounts of slag generated. Therefore, separation of P from dephosphorization slag is considerably significant to both the steel and phosphate industries.

In slag, the dephosphorization product of 3CaO·P₂O₅ reacts with 2CaO·SiO₂ to form a 2CaO·SiO₂–3CaO·P₂O₅ (C₂S–C₃P) solid solution [10]. The high distribution ratio of P₂O₅ between solid solution and other mineral phases indicates that the majority of P₂O₅ is concentrated in the C₂S–C₃P solid solution. The slag composition has a significant influence on the P₂O₅ enrichment [11,12]. Accordingly, the separation of P substantially depends on the separation of the P-concentrating solid solution from slag. Based on the difference in properties between C₂S–C₃P solid solution and other mineral phases, various methods have been proposed to treat dephosphorization slag [13]. Ono et al. [14] used the floatation of molten slag to recover the P-bearing mineral phase with low density. Miki et al. [15] studied the separation of C₂S–C₃P solid solution and the FeO-bearing liquid phase by capillary action in sintered CaO. Li et al. [16] adopted super gravity to make the P-concentrating solid solution gather in the upper molten slag.

Because the solubility of C₂S–C₃P solid solution in water is higher than that of other mineral phases, it is possible to separate the C₂S–C₃P solid solution from slag by acid leaching [17]. Compared to the above pyrometallurgical methods, the hydrometallurgical method had an advantage in terms of effectiveness and cost. Numata et al. [18] investigated the dissolution behavior of the quenched slag of a CaO–SiO₂–Fe₂O₃–P₂O₅ system in a nitric acid solution at pH 3, proving that C₂S–C₃P solid solution was dissolved selectively, although the P dissolution ratio was not high. Qiao et al. [19] used a buffer solution of C₆H₈O₇–NaOH–HCl to leach P from dephosphorization slag and found that the majority of P was dissolved, whereas approximately 30% of Fe was also dissolved. Lv et al. [2] studied the dissolution behavior of P from quenched dephosphorization slag with different compositions. To promote the selective leaching of P, in previous studies, slow cooling, Na₂O modification, and oxidization were proposed to treat molten dephosphorization slag [20,21,22]. When a citric acid solution was used, the dissolution ratio of P from slag could exceed 80%, and the dissolution of Fe was not apparent [23].

To achieve excellent selective leaching of P in at low cost, the utilization of cheap inorganic acid as a leaching agent should be considered. In addition, due to the differences in the conversion process, the composition of dephosphorization slag varies in each plant, especially the Fe₂O₃ content. To promote the utilization of acid leaching in slag treatment, the leaching behavior of P from dephosphorization slag with different levels of Fe₂O₃ content was investigated in dilute citric acid and hydrochloric acid solutions, respectively.

2. Experimental

Reagent-grade MnO, MgO, Fe₂O₃, Ca₃(PO₄)₂, SiO₂, and CaO were utilized to fabricate the dephosphorization slag, which was similar to practical slag. In this study, Fe₂O₃ was selected as the iron oxide because the dissolution ratio of P was relatively high when Fe₂O₃ existed [22]; CaO was produced by calcining CaCO₃ at 1273 K for 10 h.

Table 1. Composition of dephosphorization slag with different Fe₂O₃ content (mass%).

	CaO	SiO2	Fe2O3	P2O5	MgO	MnO
Slag 1	40.7	25.3	24.0	3.0	4.0	3.0
Slag 2	36.7	22.8	30.5	3.0	4.0	3.0
Slag 3	33.3	20.7	36.0	3.0	4.0	3.0

lists the composition of three kinds of dephosphorization slag. These types of slag have the same P₂O₅ content and basicity (%CaO/%SiO₂). To investigate the effect of Fe₂O₃, the Fe₂O₃ content ranged from 24.0% to 36.0%. According to the slag composition, the mixture of reagents was fully mixed and placed in a Pt crucible. The sample was first heated to 1823 K to form a liquid slag in a MoSi₂ resistance furnace. After maintaining this temperature for 30 min, the molten slag was cooled to 1623 K at a speed of 3 K/min to precipitate the C₂S–C₃P solid solution. Then, the sample was cooled to 1273 K and taken out of the furnace. After cooling, the sample was crushed and sieved to a particle size of less than 53 μm to promote the dissolution of P in the following experiments. The composition and morphology of the mineral phases in the dephosphorization slag were analyzed by X-ray diffraction (XRD) (Rigaku Corporation, Tokyo, Japan) and electron probe microanalysis (EPMA) (JEOL, Tokyo, Japan).

Leaching experiments were conducted in a Teflon container with 300 mL of distilled water, which was immersed in a constant-temperature water bath at 298 K. The leaching parameters were determined based on previous studies [20,21,22,23]. First, 1.5 g of slag was added to the distilled water and the solution was agitated by a rotating stirrer. The dissolution of Ca from slag resulted in an increase in pH. To achieve the selective leaching of P, the pH of the solution was controlled at a constant value. Hence, dilute acid solution was automatically added to decrease the pH of aqueous solution by a PC control system. In this study, hydrochloric acid (HCl) solution (0.4 mol/L) was selected as the leaching agent and compared with citric acid solution (0.2 mol/L). The leaching experiments were conducted at pH 4 and 5, respectively. At appropriate intervals, approximately 4 mL of solution was withdrawn and filtered by a membrane filter (<0.45 μm). The concentration of each element in the filtered solution was analyzed using inductively coupled plasma atomic emission spectroscopy (ICP-AES) (SPECTRO, Kleve, Germany). After 120 min, the residue was separated from the solution by vacuum filtration. Following drying, the mass of residue was weighted, and its morphology and mineral composition were characterized by EPMA and XRD. After performing aqua-regia digestion, the chemical composition of residue was determined by ICP-AES analysis.

The dissolution ratio of element M (R_M) from dephosphorization slag was calculated using the element concentration in the solution, as expressed in Equation (1):

$$R_{\mathrm{M}}=\frac{C_{\mathrm{M}}\cdot V}{m_{\mathrm{M}}}$$

(1)

where C_M is the M concentration in the solution (mg/L); V is the final volume of solution (L); $m_{\mathrm{M}}$ is the mass of M in slag (mg).

3. Results

3.1. Mineralogical Composition

Figure 1 presents the morphology of the mineral phases in dephosphorization slag with different amounts of Fe₂O₃ content. The average composition of each mineral phase is listed in

Table 2. Composition of mineral phases in different types of dephosphorization slag (mass%).

Mineral Phase		CaO	SiO2	Fe2O3	P2O5	MgO	MnO
Slag 1	A	1.4	0.2	76.5	0.0	12.8	9.1
	B	32.6	31.1	33.1	0.6	1.2	1.4
	C	59.9	29.7	1.8	7.1	0.8	0.7
Slag 2	A	2.3	0.1	78.2	0.0	10.4	9.0
	B	32.2	31.8	33.8	0.5	0.9	0.8
	C	59.7	28.4	1.9	9.2	0.4	0.4
Slag 3	A	2.1	0.1	79.1	0.0	9.9	8.8
	B	31.5	33.4	33.2	0.4	0.8	0.7
	C	59.5	27.4	1.6	10.8	0.4	0.3

. Three kinds of mineral phases were distinctly observed in each slag. It is worth noting that the area ratios of the mineral phases in each slag were different. The white phase had a net structure, and some dark gray phase was wrapped in it. The white phase rich in Fe₂O₃ was considered the magnesioferrite. The light gray phase consisting of the CaO–SiO₂–Fe₂O₃ slag system was regarded as the matrix phase, which had low P₂O₅ content. The dark gray phase primarily consisted of CaO and SiO₂, and it had a relatively high P₂O₅ content compared to other mineral phases. This phase was considered the C₂S–C₃P solid solution. As reported in a previous study [11], there was a high distribution ratio of P₂O₅ between the solid solution and matrix phase, demonstrating that P₂O₅ was enriched. Fe₂O₃ was mainly distributed in the matrix phase and magnesioferrite. Most of the MgO and MnO was distributed in the magnesioferrite. Because P₂O₅ and Fe₂O₃ were concentrated in different mineral phases, this provided an opportunity to separate P and Fe.

As the Fe₂O₃ content in slag increased, the P₂O₅ content in the C₂S–C₃P solid solution increased, indicating that P₂O₅ was further enriched. The P₂O₅ content in the solid solution could reach 10.8%, approximately 3.6 times higher than that in slag 3. The composition of the matrix phase had little change. The Fe₂O₃ content in the magnesioferrite increased with the Fe₂O₃ content in the slag. As shown in Figure 1, in the case of high Fe₂O₃ content, the area ratio of solid solution was lower than that in slag 1, illustrating that increasing Fe₂O₃ content was not beneficial for the formation of C₂S–C₃P solid solution. It was considered that a part of CaO and SiO₂ in the solid solution reacted with the excess Fe₂O₃. Due to the decrease in mass fraction, the P₂O₅ content in the C₂S–C₃P solid solution increased.

3.2. Leaching Results

The dissolution behavior of each dephosphorization slag in the hydrochloric acid solution at pH 4 is shown in Figure 2. The concentrations of the main elements in the solution increased continuously with leaching time. In the beginning, the dissolution of Ca, Si, and P from slag was very fast, resulting in higher concentrations in 20 min, indicating that the C₂S–C₃P solid solution was dissolved easily. The dissolution rate of slag gradually decreased, and the concentrations of the main elements showed a slight increase after 60 min. The Ca concentration was the highest among the dissolved elements, while the Fe concentration was very low in each case, illustrating that the dissolution of Fe-bearing mineral phases was difficult. A higher Fe₂O₃ content in slag led to lower Ca and Si concentrations in the solution. When the Fe₂O₃ content increased from 24.0% to 36.0%, the Ca concentration decreased from 1100.3 mg/L to 841.7 mg/L after 120 min. The Si concentrations from slag 2 and 3 were very similar, approximately 130 mg/L after 120 min. This is because the mass fraction of C₂S–C₃P solid solution was lower in the slag with high Fe₂O₃ content. Contrary to Ca and Si, the P concentration in the solution increased significantly with the Fe₂O₃ content in slag. For slag 3, the P concentration reached 49.7 mg/L. Due to the extremely low Fe concentration, the P-containing solution is suitable for phosphorus recovery. It has been reported that calcium phosphates could be extracted from this kind of solution by chemical precipitation [23].

Figure 3 shows the dissolution ratio of each element from dephosphorization slag in the hydrochloric acid solution at pH 4. The dissolution ratios of Ca, Si, and P were far higher than those of Fe, Mg, and Mn. In particular, Fe could hardly dissolve from each slag. For slag 1 with low Fe₂O₃ content, most of the Ca was dissolved, and the dissolution ratio of Si was 35.0%, higher than that of other slags. The dissolution of Mg and Mn indicated that a small part of the matrix phase or magnesioferrite was also dissolved. However, only 61.2% of P was dissolved from this slag, and the dissolution ratio of Fe was nearly zero. It was considered that the Fe³⁺ ions dissolved from slag reacted with phosphate ions to form ferric phosphate. As expressed in Equation (2), the equilibrium constant of this reaction (K) is high, indicating that this reaction can occur easily [24]. Increasing the Fe₂O₃ content in slag suppressed the dissolution of each element except for P. More than 81% of P was dissolved from slag 3, and the dissolution ratios of Mg and Mn were less than 5%, achieving excellent selective leaching. This result illustrated that the majority of the C₂S–C₃P solid solution was dissolved, without significant dissolution of other mineral phases.

$${\mathrm{Fe}}^{3+}+{\mathrm{HPO}}_4^{2-}{+2\mathrm{H}}_2{\mathrm{O}=\mathrm{FePO}}_4\cdot {2\mathrm{H}}_2{\mathrm{O}+\mathrm{H}}^{+}log\mathrm{K}=13.71$$

(2)

The dissolution behavior of P and Fe from dephosphorization slag in hydrochloric acid solution at pH 5 was compared with the results at pH 4, as shown in Figure 4. Increasing the pH significantly decreased the dissolution ratio of P, while it had no influence on the dissolution of Fe. This is because the dissolution of C₂S–C₃P solid solution became difficult when the concentration of H⁺ ions lowered. In addition, the phosphate ions dissolved from slag were easier to precipitate at a higher pH condition. At pH 5, the variation trend of P dissolution was the same as that at pH 4. The dissolution ratio of P increased with the Fe₂O₃ content in slag. However, only 68.2% of P was dissolved from slag 3. In summary, a higher Fe₂O₃ content and lower pH was beneficial for the dissolution of P from dephosphorization slag.

Citric acid was commonly used as a leaching agent in previous studies [21,22,23]. To understand the effect of acid and select an appropriate leaching agent, the dissolution behavior of dephosphorization slag with different amounts of Fe₂O₃ content in citric acid solution was also investigated. Figure 5 shows the dissolution ratio of each element from slag at pH 4. Compared to the leaching results in Figure 3, dephosphorization slag was easier to dissolve in the citric acid solution. Almost all the Ca, Si, and P was dissolved, and the dissolution ratios of Fe, Mn, and Mg were relatively high. In the citric acid solution, citrate ions (C₆H₅O₇³⁻) could combine with Ca²⁺ and Fe³⁺ ions to form complexes [25]. It has been reported that the formation of complexes on the surface of the mineral phase shifts the electron density toward the metal ion, which destabilizes the crystal lattice and facilitates the detachment of metal ions into aqueous solutions [26]. Therefore, citric acid was more effective to dissolve slag, resulting in a higher dissolution ratio of each element. As the Fe₂O₃ content in slag increased, the dissolution ratio of P increased slightly, while those of other elements decreased. The dissolution ratios of P and Fe from slag 3 were 96.7% and 21.7%, respectively, illustrating that a part of the Fe-containing mineral phases was also dissolved.

Figure 6 shows the dissolution ratios of P and Fe from dephosphorization slag in the citric acid solution at pH 4 and 5. The dissolution ratio of P decreased slightly when the pH increased from 4 to 5, indicating that most of the P-concentrating solid solution was still dissolved under this condition. The dissolution ratio of P from slag 3 was 91.3% at pH 5. Increasing the pH significantly decreased the dissolution ratio of Fe, indicating that the dissolution of Fe-bearing mineral phases was suppressed. The dissolution ratio of Fe from slag 3 decreased approximately seven times, only 2.9% at pH 5. This illustrates that better selective leaching of P was also realized in the citric acid solution at pH 5.

In summary, increasing the Fe₂O₃ content in slag promoted the dissolution of P and simultaneously suppressed the dissolution of other elements. Better selective leaching of P from dephosphorization slag was achieved in the hydrochloric acid solution at 4. Although the dissolution ratio of P was slightly lower than that in the citric acid solution, the dissolution ratio of Fe was negligible in the hydrochloric acid solution. From the perspective of treatment cost, hydrochloric acid was considered an appropriate leaching agent.

4. Discussions

To understand the dissolution behavior of mineral phases, the mass fractions of mineral phases in dephosphorization slag were compared with those of the dissolved slag and residue. Using the composition of mineral phases, their mass fractions were calculated based on the mass balance of each oxide, as described in Equations (3) and (4).

$${(\%\mathrm{MO})}_{\mathrm{slag}}{=\alpha (\%\mathrm{MO})}_{\alpha }{+\beta (\%\mathrm{MO})}_{\beta }{+\gamma (\%\mathrm{MO})}_{\gamma }$$

(3)

$$\alpha +\beta +\gamma =1$$

(4)

where α, β, and γ are the mass fractions of solid solution, matrix phase, and magnesioferrite, respectively; (%MO)_slag is the MO content in slag; (%MO)_α is the MO content in the solid solution.

The mass fraction of dissolved slag was calculated using the dissolution ratio of each element. Figure 7 presents the mass fractions of mineral phases and dissolved slag in the hydrochloric acid solution at pH 4. In slag 1 with low Fe₂O₃ content, the mass fraction of solid solution was approximately 47%, and that of magnesioferrite was relatively low. With the increase in Fe₂O₃ content, the mass fraction of solid solution decreased, while that of magnesioferrite increased. This result was consistent with the EPMA analysis. There was no significant change in the mass fraction of the matrix phase. The mass fraction of dissolved slag decreased with the increase in Fe₂O₃ content. In each case, the mass fraction of dissolved slag was slightly lower than that of the solid solution. Because the C₂S–C₃P solid solution was readily dissolved, it was found that a large amount of the C₂S–C₃P solid solution was dissolved and separated from slag. However, the dissolution ratio of P from slag 1 was not high, and a part of Mg and Mn was also dissolved. As discussed above, it is considered that a part of the P dissolved from the solid solution was precipitated with Fe³⁺ ions dissolved from slag under this condition. For slag 3, the dissolution ratio of P was the highest, illustrating that almost all the C₂S–C₃P solid solution was dissolved and the phosphate precipitation was not significant.

Figure 8 shows the mass fractions of dissolved slag in the citric acid solution at pH 4. A large amount of slag was dissolved in the citric acid solution, and the dissolution of slag became difficult as the Fe₂O₃ content increased. Due to the large dissolution of slag, acid consumption was high during leaching. Approximately 60.9% of slag 3 was dissolved, and this value was far higher than the mass fraction of solid solution. This indicated that the C₂S–C₃P solid solution was sufficiently dissolved and a part of the matrix phase and magnesioferrite was also dissolved. Although the Fe³⁺ concentration in the solution was high, most of the Fe³⁺ ions formed complexes with citrate ions. The precipitation of ferrite phosphate was avoided in the citric acid solution. Therefore, almost all the P was dissolved from slag, even with a higher dissolution ratio of Fe.

Based on the mass fractions of mineral phases, the mass ratios of P and Fe distributed in the C₂S–C₃P solid solution were calculated. These results were compared with the dissolution ratios of P and Fe in the hydrochloric acid solution at pH 4, as shown in Figure 9. More than 93.7% of P was distributed in the C₂S–C₃P solid solution, while the mass ratio of Fe was less than 3.3%. If the C₂S–C₃P solid solution was totally dissolved and no precipitation occurred, the dissolution ratio should be equal to the mass ratio. However, there was a large difference between the mass ratio of P and the dissolution ratio of P. With the increase in Fe₂O₃ content, this difference became smaller. The variation trend of Fe was the same. Because most of the C₂S–C₃P solid solution was dissolved, it demonstrated that some of the phosphate ions precipitated with Fe³⁺ ions in the hydrochloric acid solution. In the case of high Fe₂O₃ content, due to the low dissolution ratio of Fe, phosphate precipitation could be suppressed.

Table 3. Composition of residues after leaching in hydrochloric acid solution at pH 4 (mass%).

	CaO	SiO2	P2O5	Fe2O3	MgO	MnO
Residue (slag 1)	10.0	32.3	2.1	45.5	5.0	5.1
Residue (slag 2)	11.6	27.4	0.9	50.4	5.1	4.6
Residue (slag 3)	11.4	22.7	0.7	56.5	4.5	4.2

lists the average composition of residues after leaching in hydrochloric acid solution at pH 4. Compared to the original slag composition, the P₂O₅ content in the residue decreased while the Fe₂O₃ content increased. A higher Fe₂O₃ content in slag resulted in a lower P₂O₅ content in the residue. The residue of slag 3 contained only 0.7% P₂O₅, demonstrating effective separation of P and Fe. This P-removal residue had a high Fe₂O₃ content, which can substitute for iron ores.

Because slag 3 showed better selective leaching at pH 4, the mineralogical composition and morphology of its residue were analyzed. Figure 10 shows the XRD patterns of slag 3 and its residue. The characteristic peaks of C₂S–C₃P solid solution and magnesioferrite were observed in the original slag. Because the matrix phase was an amorphous glass phase, it had no characteristic peaks. After leaching, the peaks of C₂S–C₃P solid solution almost disappeared, while the peaks of magnesioferrite intensified. This further demonstrated that the P-concentrating solid solution was separated from the slag and the Fe-bearing mineral phases remained in the residue.

Figure 11 shows the morphology of the residue after leaching in the hydrochloric acid solution. The composition of the identified domains on the residue surface is listed in

Table 4. Composition of mineral phases on the residue surface (mass%).

Point	Ca	Si	Fe	P	Mg	Mn
A	1.3	0	55.1	0	5.8	6.4
B	22.6	15.2	23.0	0.2	0.4	0.5

. Two mineral phases, similar to the matrix phase and magnesioferrite, were observed. It was difficult to detect the C₂S–C₃P solid solution. This shows that the residue principally comprised the matrix phase and magnesioferrite after the separation of the solid solution, which can be reutilized as a raw material in ironmaking or steelmaking processes. In this study, effective selective leaching of P was realized, which would promote the resource utilization of dephosphorization slag.

5. Conclusions

The effect of Fe₂O₃ content in slag and acid on the selective leaching of P from dephosphorization slag was investigated in this study. The P₂O₅ content in the C₂S–C₃P solid solution increased with the Fe₂O₃ content in slag. Increasing the Fe₂O₃ content in slag increased the dissolution ratio of P and simultaneously suppressed the dissolution of other elements, which was beneficial for the selective leaching of P. More than 81% of P was dissolved from dephosphorization slag in hydrochloric acid solution at pH 4, and the dissolution ratio of Fe was nearly zero, achieving excellent selective leaching. Although better selective leaching was also realized in citric acid solution at pH 5, hydrochloric acid was considered an appropriate leaching agent from the perspective of leaching cost. Almost all the C₂S–C₃P solid solution was dissolved from dephosphorization slag by acid leaching, and the Fe-bearing matrix phase and magnesioferrite remained in the residue. The residue contained 0.7% P₂O₅ and 56.5% Fe₂O₃ and can be reutilized as a flux in ironmaking or steelmaking processes.