Optimum Conditions for Phosphorus Recovery from Steelmaking Slag with High P₂O₅ Content by Selective Leaching

Abstract

Selective leaching of P-concentrated solid solution is considered an effective method for recovering P from slag with high P₂O₅ content. This plays a significant role in stimulating the utilization of high-P iron ores. To determine the optimum conditions for selective leaching of P, we investigated the effects of the cooling rate of molten slag, Na₂O content in slag, and pH on the dissolution behavior of the modified slag in the aqueous solution. Following leaching, a precipitation method was studied to recover P from the leachate. Compared to the quenched slag, the furnace-cooled slag exhibited a higher P dissolution ratio and lower Fe dissolution ratio, indicating that slow cooling was necessary to realize selective leaching. The addition of 2.5–4.0 mass% of Na₂O to the slag was sufficient to cause most of the solid solution to dissolve at pH 6, Fe being difficult to dissolve. The dissolution ratio of P from the modified slag increased significantly when the pH decreased from 7 to 5. A further decrease in the pH promoted Fe dissolution. Therefore, the pH of the aqueous solution should be controlled between 5 and 6. After leaching, with an increase in the pH of the leachate, the precipitation ratio of P from the leachate increased, while the P₂O₅ content in the obtained phosphate product decreased. In this process, approximately 70% of P in the slag was recovered in the form of the phosphate product, which can be used as a phosphate fertilizer.

1. Introduction

Utilization of high-P iron ores is gaining considerable attention because of large reserves and the depletion of high-grade iron ores. The use of low-grade iron ores results in an increase in the P content in hot metal. Therefore, P is removed from iron ores prior to its use in smelting.¹⁾ A variety of dephosphorization processes for high-P iron ores have been developed such as acid leaching, carbothermic reduction, and physical beneficiation.^2,3,4) Due to the large amount of iron ores, an enormous processing capacity is required and the treatment cost is also huge, which restricts the industrialization of these methods.

P existing in high-P iron ores is considered a potential source of phosphorus. We should therefore pay attention to the comprehensive utilization of high-P iron ores. Through efficient dephosphorization using multi-phase slag, high levels of P in hot metal can be removed, leading to its concentration in slag.⁵⁾ P is mainly found in the 2CaO∙SiO₂-3CaO∙P₂O₅ (C₂S–C₃P) solid solution in steelmaking slag, which has a high P₂O₅ content and high distribution ratio between C₂S–C₃P solid solution and other phases.⁶⁾ If the P-concentrated solid solution can be extracted from the slag, the obtained phosphate will be suitable for producing phosphate fertilizers. The residual slag with the low P₂O₅ content can be recycled during the steelmaking process. It is likely that the increased cost involved in utilizing high-P iron ores could be covered by the income generated by the sale of the recovered phosphate product and the savings resulting from the recycling of steelmaking slag after P separation.⁷⁾ Developing an effective and economic process to recover P from slag with high P₂O₅ content plays a significant role in stimulating the utilization of high-P iron ores, simultaneously ensuring the reliable supply of phosphate fertilizers as the global population increases.

Various methods have been explored for the separation and recovery of P from steelmaking slag based on differences in physicochemical properties between different compounds and phases in the slag. Li et al.⁸⁾ proposed the reduction of dephosphorization slag by C-saturated hot metal to remove P, Mn, and Fe. The obtained Fe–P–Mn alloy was further dephosphorized to produce slag with high P₂O₅ content as a fertilizer. Yokoyama et al.⁹⁾ studied the separation of the P-rich solid solution and Fe-rich phase with the aid of a strong magnetic field. Considering the density of each phase, Li et al.¹⁰⁾ tried to remove the P-concentrated phase from slag via super gravity.

The C₂S–C₃P solid solution has a higher solubility than other phases in water. This led to Teratoko and Kitamura et al.¹¹⁾ proposing a novel technique: selective leaching of the solid solution. It is achieved by dissolving and separating P-concentrated solid solution from slag, without dissolving other phases in aqueous solutions, at a constant pH condition; however, the dissolution ratio of P was not high.¹²⁾ In the case of high P₂O₅ content, the dissolution of the solid solution became more difficult. In addition, comparing with FeO and Fe₂O₃ as iron oxide in slag, the slag with FeO showed a lower dissolution ratio of P, which was not appropriate for selective leaching of P.¹¹⁾ To promote dissolution of the solid solution in the slag with high P₂O₅ content, we conducted a series of studies.^13,14,15) It was determined that using citric acid as the leaching agent, decreasing the cooling rate of molten slag, and modifying the slag by adding Na₂O were beneficial for achieving selective leaching of the solid solution. The dissolution of the solid solution was promoted under these conditions, while those of the other phases were suppressed.

On the basis of the above results, an innovative process for P recovery from the steelmaking slag with high P₂O₅ content via selective leaching is proposed. To determine the optimum conditions for the selective leaching of P, the effects of the cooling rate of molten slag, Na₂O content in the slag, and pH on the dissolution behavior of the modified slag in the aqueous solution were investigated. Following leaching, a precipitation method was applied to recover P from the leachate. The effect of pH on the composition of the precipitate and P recovery ratio was studied.

2. Experimental Method

2.1. Synthesis of Slags

A CaO–SiO₂–Fe₂O₃–P₂O₅–MgO slag system was prepared by mixing reagent-grade CaO (obtained by calcining CaCO₃ at 1273 K for at least 10 h), SiO₂, Fe₂O₃, Ca₃(PO₄)₂, and MgO. Na₂SiO₃ was added as the modifier. Fe₂O₃ was used as the iron oxide because it favored the selective leaching of P from slag. The compositions of each slag are listed in Table 1. Thoroughly mixed reagents (10 g) were placed in a Pt crucible and heated in an electrical resistance furnace to 1823 K under air. As shown in Fig. 1, after forming a liquid phase, the slag was cooled to 1623 K and kept at this temperature for 20 min to precipitate the solid solution. To quench, the slag was cooled quickly with water. During furnace cooling, the slag was cooled to 1323 K at a rate of 5 K/min, and then taken out of furnace. The composition of each phase in the slag was determined using an electron probe micro analyzer (EPMA).

Table 1. Slag composition (mass%) and leaching condition.

	CaO	SiO2	Fe2O3	P2O5	MgO	Na2O	Cooling	Leaching test
Slag A	35.7	22.2	30.0	8.0	3.1	1.0	Furnace cooling	pH=6
Slag B	35.1	21.9	29.5	8.0	3.1	2.5		pH=6
Slag C	34.5	21.5	29.0	8.0	3.0	4.0		pH=4, 4.5, 5, 6, 7
Slag D	32.1	19.9	29.0	8.0	3.0	8.0		pH=6
Slag E	34.5	21.5	29.0	8.0	3.0	4.0	Quenching	pH=5, 6

Fig. 1.

Heating pattern for synthesis of slags.

2.2. Leaching of Slags

The leaching apparatus is the same as that used in previous studies.¹³⁾ Ground slag (1 g, < 53 μm) was added to 400 mL deionized water in a container at 298 K. To increase the reaction rate, the aqueous solution was agitated using a stirrer at a speed of 200 r/min. During leaching, the dissolution of Ca from the slag would increase the pH of the aqueous solution. To maintain a constant pH, citric acid solution (0.1 mol/L) was automatically added as the leaching agent using a PC control system. Approximately 5 mL of the aqueous solution was sampled at adequate intervals and filtered using a syringe filter (< 0.45 μm). The concentration of each element was analyzed using inductively coupled plasma atomic emission spectroscopy (ICP-AES). After 2 h of leaching, the residue was filtered, dried, and weighed. X-ray diffraction (XRD) analysis and ICP-AES were used to determine the remaining phase and residue composition. Table 1 also lists the leaching tests conducted at various pH conditions.

2.3. P Precipitation from the Leachate

The leachate collected from the leaching experiment was further treated to recover P. Calcium phosphate compounds have little solubility at higher pH, and this made it possible to precipitate the dissolved P in the leachate via the addition of the Ca(OH)₂ solution. To investigate the effect of pH on P precipitation, the pH of the leachate were adjusted to 9, 10, 11, and 12, respectively, by adding saturated Ca(OH)₂ solution. The aqueous solution was settled for 24 h to separate the precipitate from the aqueous solution. The precipitate obtained was first dried at 373 K and then calcined at 873 K for 2 h to remove crystal water and form a crystalline substance. Finally, a phosphate product was obtained. The composition of this phosphate product was analyzed using XRD, and the content of each component was determined using ICP-AES.

3. Results and Discussion

3.1. Mineralogical Composition

Figure 2 shows the typical mineralogical compositions of slags with different Na₂O contents and cooling rates. The average composition of each phase is listed in Table 2. There were three phases in the furnace-cooled slags: magnesioferrite phase (rich in Fe₂O₃ and MgO), matrix phase (mainly consisting of CaO–SiO₂–Fe₂O₃ system), and solid solution (P₂O₅ concentrated). In the quenched slag, the magnesioferrite phase did not precipitate, and only the matrix phase and solid solution were observed. Some fine solid solution particles that precipitated during quenching were distributed in the matrix phase. In the case of furnace cooling, the effect of the Na₂O content on the composition of the magnesioferrite phase was not significant. P₂O₅ content in the matrix phase and in the solid solution decreased with increasing Na₂O content in slag. Further addition of Na₂O resulted in a higher Na₂O content in the solid solution. Compared to the furnace-cooled slag with the same Na₂O content, the composition of solid solution in the quenched slag was almost the same. P₂O₅ and Fe₂O₃ contents in the matrix phase in quenched slag were higher than those in furnace-cooled slag. This was due to the existence of fine solid solution particles, with no precipitation of the magnesioferrite phase.

Fig. 2.

Typical mineralogical composition of slags.

Table 2. Composition of each phase in different slags (mass%).

	CaO	SiO2	Fe2O3	P2O5	MgO	Na2O	Phase
Slag A	1.3	0.0	87.4	0.0	11.2	0.0	1. Magnesioferrite
	40.1	34.0	20.8	2.3	1.0	1.7	2. Matrix phase
	56.2	12.0	0.6	29.3	0.2	1.6	3. Solid solution
Slag B	1.5	0.2	86.5	0.0	11.5	0.2	1. Magnesioferrite
	38.1	35.1	20.7	1.9	1.7	2.6	2. Matrix phase
	55.1	13.9	0.6	26.6	0.3	3.4	3. Solid solution
Slag C	1.8	0.0	86.8	0.0	11.1	0.3	1. Magnesioferrite
	34.4	33.1	25.9	1.1	1.5	3.9	2. Matrix phase
	51.4	15.6	0.9	25.4	0.3	6.4	3. Solid solution
Slag D	2.0	0.0	84.1	0.0	12.7	1.2	1. Magnesioferrite
	24.2	29.4	34.8	0.4	1.2	9.9	2. Matrix phase
	48.6	16.6	0.7	24.5	0.3	9.4	3. Solid solution
Slag E	29.6	21.6	38.3	3.5	3.4	3.5	1. Matrix phase
	50.9	14.3	0.9	25.9	1.1	6.8	2. Solid solution

3.2. Leaching Results

The effect of the Na₂O content on the dissolution behavior of each element from the furnace-cooled slags at pH 6 is shown in Fig. 3. The concentrations of each element increased with leaching time. The dissolution rate gradually decreased owing to increased concentrations of each element in the aqueous solution. Ca concentration was the highest among the dissolved elements and it increased significantly with the increase in the Na₂O content in slag. Si and P concentrations were also increased by the addition of Na₂O. Fe concentration was the lowest in the aqueous solution: less than 10 mg/L in each case.

Fig. 3.

Change in the concentrations of each element with time at pH 6.

As discussed in previous studies,¹⁵⁾ the dissolution ratios of each element from slag were calculated using Eq. (1):   

$$R_{\mathrm{d}\mathrm{i}\mathrm{s}\mathrm{s}\mathrm{o}\mathrm{l}\mathrm{u}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}}^M=\frac{C_M\cdot V}{m_M}$$(1)

where $R_{\mathrm{d}\mathrm{i}\mathrm{s}\mathrm{s}\mathrm{o}\mathrm{l}\mathrm{u}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}}^{\mathrm{M}}$ is the dissolution ratio of element M from slag, C_M the concentration of M after 120 min (mg/L), V the final volume of the aqueous solution (L), and m_M the mass of M in the initial slag (mg). Figure 4 shows the calculated dissolution ratios at pH 6. There were significant differences between the dissolution ratios of the elements. The dissolution ratio of P was the highest. Fe was difficult to dissolve, and its dissolution ratio was negligible in each case. The dissolution ratio of Na exceeded 40% when the Na₂O content in slag was more than 2.5 mass%. With the increase in the Na₂O content, the dissolution ratios of Ca, Si, and P increased. These results indicated that Na₂O addition promoted dissolution of the P-condensed phase without dissolving phases containing Fe. When the Na₂O content in the slag exceeded 2.5 mass%, the dissolution ratio of P increased to 73%, approximately half of the Ca and one third of the Si were dissolved simultaneously. Therefore, the addition of 2.5–4.0 mass% of Na₂O to slag was sufficient to achieve good selective leaching of P.

Fig. 4.

Dissolution ratios of each element from slags with various Na₂O contents at pH 6.

Figure 5 shows the effect of pH on the dissolution behavior of P and Fe from the slag containing 4 mass% of Na₂O. Decreasing pH of the aqueous solution improved the dissolution rate of P from slag and also increased the P concentration. When the pH decreased from 7 to 5, there was a significant increase in the P concentration; however, further decrease in pH changed the P concentration slightly. Fe concentration was very low less than 20 mg/L when the pH exceeded 5. When the pH was less than 5, Fe concentration increased to approximately 140 mg/L, which was caused by significant dissolution of the Fe-containing phase.

Fig. 5.

Change in the concentrations of each element at various pH conditions.

Calculated dissolution ratios of each element at various pH conditions are shown in Fig. 6. Dissolution of slag increased when the pH value decreased, resulting in higher dissolution ratios of each element. The dissolution ratios of Fe and Mg were lower than those of other elements in each case, indicating that the phases rich in Fe and Mg were difficult to dissolve. The dissolution ratio of P increased from 49.7% to 76.9% when the pH varied from 7 to 6. A further decrease in pH had little effect in promoting P dissolution. The dissolution behavior of Na was similar with that of Ca and Si. At pH 5 or 6, about half of Na was dissolved from slag. There was a huge difference in the dissolution ratios of other elements between pH 5 and 4, especially Fe. Large amounts of Fe and Mg dissolution deteriorated selective leaching, which was not beneficial for P recovery in the following processes. The pH of the aqueous solution should be controlled between 5 and 6 to achieve better selective leaching of P from slag.

Fig. 6.

Dissolution ratios of each element from the slag with 4 mass% of Na₂O at various pH conditions.

Figure 7 shows the effect of the cooling rate on the dissolution ratios of Ca, P, and Fe from slag containing 4 mass% of Na₂O. At pH 6, the dissolution ratios of Ca and Fe from each slag were almost the same. A lower dissolution ratio of Fe indicated that the matrix phase was not dissolved easily, regardless of the cooling rate. The dissolution ratio of P from furnace-cooled slag was higher than that from quenched slag. When the pH decreased to 5, the dissolution of the quenched slag was promoted, resulting in higher dissolution ratios of Ca and Fe. In the case of furnace cooling, the dissolution ratios of each element increased slightly. The furnace-cooled slag exhibited a higher dissolution ratio of P and a lower dissolution ratio of Fe at pH 5 than quenched slag. A decrease in the cooling rate was necessary for the modified slag to achieve better selective leaching of P.

Fig. 7.

Dissolution ratios of Ca, P, and Fe from slags with different cooling rates.

3.3. Residue Composition

Table 3 lists the composition of residue obtained under various leaching conditions. P₂O₅ content in the residue decreased significantly relative to the original slag while Fe₂O₃ content increased. The further addition of Na₂O resulted in lower P₂O₅ content in the residue because of an increase in the dissolution ratio of P. Correspondingly, Fe₂O₃ and MgO contents increased. P₂O₅ content further decreased when the pH value decreased and a residue containing higher Fe₂O₃ content was obtained. For the slag with 4 mass% of Na₂O, P₂O₅ content was 0.9 mass% and the Fe₂O₃ content reached 48.5 mass% after leaching at pH 5. When this slag was leached at pH 4, the majority of the residue consisted of Fe₂O₃ and MgO.

Table 3. Composition of residue under various leaching conditions (mass%).

Residue of Sample	pH	CaO	SiO2	Fe2O3	P2O5	MgO	Na2O
Slag A (1.0% Na2O)	6	30.71	21.37	39.93	2.79	4.19	1.01
Slag B (2.5% Na2O)	6	27.76	20.04	43.53	2.02	4.53	2.12
Slag C (4.0% Na2O)	6	26.17	19.20	44.85	2.09	4.59	3.10
Slag D (8.0% Na2O)	6	18.32	16.30	50.70	1.49	5.14	8.06
Slag C (4.0% Na2O)	5	22.94	20.45	48.46	0.93	4.47	2.75
Slag C (4.0% Na2O)	4.5	9.12	2.19	77.92	0.27	9.25	1.25

Figure 8 shows XRD patterns of the slag with 4 mass% of Na₂O and its residues after leaching at pH 5 and 6. Peaks associated with solid solution and magnesioferrite were observed in the original slag. After leaching, intensities of the peaks associated with solid solution weakened; those peaks almost disappeared at pH 5. In contrast, the peaks associated with magnesioferrite intensified. The P-concentrated solid solution could be separated from the slag by leaching and the residue had the potential for recycling within the steelmaking process.

Fig. 8.

XRD patterns of the original slag and its residues after leaching at pH 5 and 6.

3.4. Discussion on Selective Leaching of P

Mass fractions of different phases in different slags were calculated on the basis of mass balance.¹⁵⁾ As described in Eqs. (2) and (3), the sum of the oxide MO content in each phase was equal to its content in slag.   

$$w{(\mathrm{M}\mathrm{O})}_{slag}=\sum w{(\mathrm{M}\mathrm{O})}_i\cdot X_i$$(2)

  

$$\sum X_i=1$$(3)

where w(MO)_slag is the MO content in slag, w(MO)_i is the MO content in phase i (shown in Table 2), and X_i is the mass fraction of phase i. For slags with different Na₂O contents, the phase fraction was compared with the mass ratios of dissolved portion and residue at pH 6 to estimate the dissolution behavior of solid solution. Figure 9 shows that with the increase in the Na₂O content, the mass fraction of the solid solution increased while that of the matrix phase decreased. In the case of unmodified slag, the mass ratio of the dissolved portion was lower than that of the solid solution, indicating that a part of the solid solution remained in the residue and did not dissolve. The mass ratio of the dissolved part increased due to Na₂O modification. It was almost similar to the mass fraction of the solid solution when the Na₂O content exceeded 2.5 mass%. Fe-rich matrix phase and magnesioferrite were difficult to dissolve because of the lower dissolution ratio of Fe. This result indicated that most of the solid solution was dissolved from slag. A further increase in the Na₂O content resulted in an increase in the dissolution of slag together with the mass fraction of the solid solution in the slag. The mass ratio of the dissolved part was a little higher than that of the solid solution. It illustrated that the solid solution could be dissolved and separated from slag by sufficient Na₂O addition, without dissolving large amounts of other phases.

Fig. 9.

Mass ratios of the residue and dissolved part at pH 6, compared with the phase fractions of each slag.

To evaluate the dissolution behavior of each phase, the presumed dissolution ratio of element M ($R_M^{SS}$) from solid solution was calculated using Eq. (4):   

$$R_M^{SS}=\frac{C_M\cdot V}{m_M^{SS}}$$(4)

where $m_M^{SS}$ is the original mass of element M in solid solution (mg). The value of $R_P^{SS}$ indicated the dissolution ratio of solid solution since the dissolution of P occurred mainly from the latter. Figure 10 shows the relationship between dissolution ratio of solid solution and the molar ratio of Na₂O to P₂O₅ in solid solution. With increasing Na₂O content in slag, the molar ratio of Na₂O to P₂O₅ in solid solution increased. It has been reported that 2CaO·Na₂O·P₂O₅ had higher solubility than 3CaO·P₂O₅.¹⁶⁾ The addition of 0.1 mol Na₂O to 1 mol P₂O₅ of calcium phosphate was unfavorable but more than 0.2 mol of Na₂O favored the formation of soluble phosphate.¹⁷⁾ When the molar ratio of Na₂O to P₂O₅ increased from 0 to 0.29, the dissolution ratio of solid solution increased significantly. A further increase in the molar ratio improved the dissolution of the solid solution negligibly. This result indicated that a solid solution with good solubility was formed when the molar ratio of Na₂O to P₂O₅ exceeded 0.29, which agreed with previous studies.

Fig. 10.

Relationship between the dissolution ratio of solid solution and the molar ratio of Na₂O to P₂O₅ in solid solution.

Assuming that each phase was individually dissolved from slag, the dissolution ratios of each element from slag could be calculated using the mass fraction and phase composition of each phase (listed in Table 2). Figure 11 shows calculated dissolution ratios and experimental results of slag with 4 mass% of Na₂O at various pH conditions. The dissolution ratios of P and Ca from slag reached 92% and 45.5%, respectively, when the solid solution was totally dissolved. At pH 7, the dissolution ratio of P was lower than the calculated value, indicating that the dissolution of solid solution was poor. When the pH was between 6 and 5, the dissolution ratio of P increased and approached the calculated value. The dissolution ratios of Ca and Si were a little higher than the values calculated from solid solution. It demonstrated that most of the solid solution was dissolved and dissolution of other phases was insignificant. When the pH was lower than 4.5, the dissolution ratios of Ca and Si were higher than the values calculated from solid solution and lower when the solid solution and matrix phase were both dissolved. This result indicated that a portion of the dissolved Ca and Si was from dissolution of the matrix phase. The dissolution of Fe also went through similar conditions: when the pH varied from 5 to 4, the dissolution ratio of Fe gradually approached the value calculated from matrix phase. There was significant variation in the dissolution of matrix phase between pH 5 and 4.5.

Fig. 11.

Calculated dissolution ratios of some elements from slag when each phase was dissolved, and the experimental results.

Figure 12 shows the mass ratios of the dissolved slag and residue at various pH conditions, compared with each phase in the original slag. At pH 7, the mass ratio of the dissolved portion was lower than that of the solid solution. With the decrease in pH, the mass ratio of the dissolved portion increased and exceeded that of the solid solution at pH 6, indicating that a majority of the solid solution was dissolved. When the pH decreased to less than 4.5, approximately 70% of slag was dissolved, which was almost equal to the mass fractions of solid solution and a large proportion of the matrix phase. These results were consistent with the above discussion. To avoid dissolution of matrix phase and achieve selective leaching of solid solution, the pH should be controlled between 5 and 6.

Fig. 12.

Mass ratios of the residue and dissolved part at various pH conditions, compared with phase fraction.

The effect of the cooling rate on the dissolution behavior of each phase was investigated by calculating the presumed dissolution ratios of Ca and P from solid solution ($m_{\mathrm{C}\mathrm{a}}^{SS}$ and $m_{\mathrm{P}}^{SS}$) using Eq. (4). For this index, dissolution was assumed to occur from only the solid solution. If the calculated value was more than unity, it indicated that the dissolution of the element occurred from the solid solution and from the matrix phase. Figure 13 shows the presumed dissolution ratios of Ca and P from solid solution under different cooling rates. In the case of furnace cooling, the presumed dissolution ratios of Ca and P were close to unity at pH 5 and 6, illustrating that almost all of the solid solution was dissolved, and that the matrix phase was barely dissolved. In the case of quenching, selective leaching of solid solution occurred at pH 6. The dissolution ratio of P from quenched slag was lower than that from furnace-cooled slag. This was because the solid solution in furnace-cooled slag aggregated and fine solid solution particles were not distributed in the matrix phase, which facilitated leaching of the solid solution. At pH 5, the presumed dissolution ratio of Ca was far greater than unity, indicating that part of the matrix phase in quenched slag was dissolved. The possible reason was that the increasing cooling rate changed the structure of the silicate glass¹⁸⁾ and hindered precipitation of the crystalline phase,¹⁴⁾ promoting dissolution of the matrix phase. Slow cooling was thus beneficial for selective leaching of the solid solution.

Fig. 13.

Presumed dissolution ratios of Ca and P from solid solution at different cooling rates.

3.5. P Recovery from the Leachate

The composition of the leachate obtained by leaching of slag with 4 mass% of Na₂O is shown in Table 4. The major ions in the leachate were those of Ca, Si, and P. Following the addition of Ca(OH)₂ solution to adjust the pH, the leachate was separated into two layers via settling: the upper solution and the precipitate. Table 4 lists the composition of the upper solution at various pH conditions. The concentrations of each element in the upper solution decreased relative to the original leachate. The P concentration varied from 63 mg/L to ≤ 1 mg/L at pH 10. With increasing pH of the aqueous solution, the concentrations of Si, Fe, and P further decreased because of the precipitation of some substances. When the pH was adjusted to 12 the Si concentration reduced by approximately half.

Table 4. Concentration of the leachate and the upper solution after precipitation (mg/L).

Solution	Ca	Si	Fe	P	Mg	Na
Original Leachate	301.01	69.60	6.25	63.03	2.51	35.54
Upper Solution (pH=9)	230.61	64.33	2.38	4.14	2.00	30.09
Upper Solution (pH=10)	233.49	56.25	0.32	0.91	1.86	28.99
Upper Solution (pH=11)	247.13	51.44	0.04	0.24	1.50	28.38
Upper Solution (pH=12)	295.52	37.69	0.03	0.20	0.37	24.69

Ca²⁺ and phosphate ions in the leachate could form a variety of calcium phosphates such as DCPD (brushite, CaHPO₄∙2H₂O), OCP (octacalcium phosphate, Ca₈H₂(PO₄)₆∙5H₂O), TCP (tricalcium phosphate, Ca₃(PO₄)₂), and HAP (hydroxylapatite, Ca₁₀(PO₄)₆(OH)₂).¹⁹⁾ The precipitation reactions of these calcium phosphates are described below:^20,21)   

$$\mathrm{C}{\mathrm{a}}^{2+}+\mathrm{H}\mathrm{P}{\mathrm{O}}_4^{2-}+2{\mathrm{H}}_2\mathrm{O}=\mathrm{C}\mathrm{a}\mathrm{H}\mathrm{P}{\mathrm{O}}_4\cdot 2{\mathrm{H}}_2\mathrm{O}\hspace{1em}log\mathrm{\hspace{0.17em} }K=6.52$$(5)

  

$$\begin{array}{c}8\mathrm{C}{\mathrm{a}}^{2+}6\mathrm{P}{\mathrm{O}}_4^{3-}+2{\mathrm{H}}^{+}+5{\mathrm{H}}_2\mathrm{O}=\mathrm{C}{\mathrm{a}}_8{\mathrm{H}}_2{(\mathrm{P}{\mathrm{O}}_4)}_6\cdot 5{\mathrm{H}}_2\mathrm{O}\\ log\mathrm{\hspace{0.17em} }K=87.59\end{array}$$(6)

  

$$3\mathrm{C}{\mathrm{a}}^{2+}+2\mathrm{P}{\mathrm{O}}_4^{3-}=\mathrm{C}{\mathrm{a}}_3{(\mathrm{P}{\mathrm{O}}_4)}_2\hspace{1em}log\mathrm{\hspace{0.17em} }K=24.58$$(7)

  

$$\begin{array}{c}10\mathrm{C}{\mathrm{a}}^{2+}+6\mathrm{P}{\mathrm{O}}_4^{3-}+2{\mathrm{H}}_2\mathrm{O}=\mathrm{C}{\mathrm{a}}_{10}{(\mathrm{P}{\mathrm{O}}_4)}_6{(\mathrm{O}\mathrm{H})}_2+2{\mathrm{H}}^{+}\\ log\mathrm{\hspace{0.17em} }K=62.42\end{array}$$(8)

Figure 14 shows the solubility curves of calcium phosphates (the formation of OCP and HAP had relationships with the pH of solution) and the experimental results under various pH conditions. In the present study, the solubility of calcium phosphate in descending order was: DCPD, OCP, TCP, and HAP. Ca and P concentrations in the original leachate were lower than those required for HAP precipitation at pH 6, but were higher than the saturated concentration of calcium phosphates at higher pH conditions. It was thus possible to precipitate P in the leachate via adjusting the pH and increasing the Ca concentration. After precipitation, the observed points for the upper solution were located around the solubility curves of DCPD. At pH 11 and 12, they lay between the solubility curves of DCPD and OCP. DCPD was the major constituent of the precipitate and its solubility determined Ca and P concentrations in the upper solution.

Fig. 14.

Solubility curves of calcium phosphates and the experimental results.

In a solution rich in Ca²⁺ and silicate ions, CaO–SiO₂–H₂O gel in the form of (CaO)₅(SiO₂)₆(H₂O)_5.5 generally precipitates in alkaline conditions.²²⁾ The reaction for the formation of calcium silicate hydrogel is given by Eq. (9):²²⁾   

$$\begin{array}{c}{(\mathrm{C}\mathrm{a}\mathrm{O})}_5{(\mathrm{S}\mathrm{i}{\mathrm{O}}_2)}_6{({\mathrm{H}}_2\mathrm{O})}_{5.5}+4{\mathrm{H}}^{+}=5\mathrm{C}{\mathrm{a}}^{2+}+6\mathrm{H}\mathrm{S}\mathrm{i}{\mathrm{O}}_3^{-}+4.5{\mathrm{H}}_2\mathrm{O}\\ log\mathrm{\hspace{0.17em} }K=8.7\end{array}$$(9)

Figure 15 shows the solubility curves of calcium silicate hydrogel and the experimental results. With increasing Ca concentration, the Si concentration in the aqueous solution decreased. High pH suppressed dissolution of this gel. The observed point for Ca and Si concentrations at pH 9 was located near the solubility line at pH 9, indicating that the solubility of calcium silicate hydrogel determined Si concentration. As the pH decreased, the observed points moved to the solubility line at pH 12. The experimental value however was far higher than the value determined by the solubility of the gel at pH 12. The reason is not clear.

Fig. 15.

Solubility curves of calcium silicate hydrogel and the experimental results.

A gray phosphate product was obtained through drying and calcination. As listed in Table 5, this product mainly consisted of CaO and P₂O₅. The CaO content was almost identical in each product, approximately 54 mass%. P₂O₅ content exceeded 26 mass% and it increased further with a decrease in pH. At pH 9, the P₂O₅ content was 34.5 mass%. With the increase in pH, the SiO₂ content in the product increased because of silicate precipitation. According to the mass balance calculation, approximately 9 mass% of the product was determined to be unknown. It is considered that this part consists of the hydroxyl (OH) which did not remove during heating and the carbide which came from the decomposition of the organic substance.

Table 5. Compositions of the obtained P products (mass%).

Phosphate product	CaO	SiO2	Fe2O3	P2O5	MgO	Na2O	Others
pH=9	54.28	0.51	1.65	34.48	0.21	0.26	8.61
pH=10	53.12	2.25	1.95	31.27	0.22	0.26	10.95
pH=11	53.71	3.91	2.00	30.18	0.35	0.24	9.61
pH=12	54.94	5.71	1.73	26.05	0.72	0.15	10.70

Figure 16 shows XRD patterns of the products obtained. Similar peaks associated with hydroxylapatite (HAP) and silicon substituted calcium hydroxylapatite were observed in each product, indicating that the same substance was formed after calcination of the precipitates. As shown in Fig. 14, HAP was the most stable calcium phosphate. The precipitation of phosphate from leachate depended on nucleation and growth.¹⁹⁾ The precipitation kinetics of DCPD was fast and DCPD precipitated instead of HAP, which determined the P concentration in the upper solution. Through calcination, DCPD contained in the precipitate was transformed to thermodynamically stable HAP.²³⁾ In the solubility tests,¹⁶⁾ about 97% of P₂O₅ in the obtained product dissolved in 2% citric acid solution, indicating most of the P was available in the soil. The phosphate product obtained had a high content of citrate-soluble P₂O₅ and could be used as a fertilizer.   

$$R_{\mathrm{p}\mathrm{r}\mathrm{e}\mathrm{c}\mathrm{i}\mathrm{p}\mathrm{i}\mathrm{t}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}}=\frac{m\cdot w_{{\mathrm{P}}_2{\mathrm{O}}_5}\cdot 2M_{\mathrm{P}}}{V\cdot C_{\mathrm{P}}\cdot M_{{\mathrm{P}}_2{\mathrm{O}}_5}}$$(10)

  

$$R_{\mathrm{t}\mathrm{o}\mathrm{t}\mathrm{a}\mathrm{l}}=R_{\mathrm{d}\mathrm{i}\mathrm{s}\mathrm{s}\mathrm{o}\mathrm{l}\mathrm{u}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}}^{\mathrm{P}}\cdot R_{\mathrm{p}\mathrm{r}\mathrm{e}\mathrm{c}\mathrm{i}\mathrm{p}\mathrm{i}\mathrm{t}\mathrm{a}\mathrm{t}\mathrm{i}\mathrm{o}\mathrm{n}}$$(11)

Fig. 16.

XRD patterns of the phosphate products obtained in this process.

The precipitation ratio of P in the leachate (R_{precipitation}), and the total recovery ratio of P from the modified steelmaking slag (R_total) via selective leaching and precipitation were calculated using Eqs. (10) and (11), where m is the mass of precipitate obtained, $w_{{\mathrm{P}}_2{\mathrm{O}}_5}$ the mass ratio of P₂O₅ in the precipitate, M the molar mass, V the volume of the original leachate, and C_P the P concentration in the original leachate. As shown in Table 6, most of the P in the leachate was concentrated in the precipitate after precipitation. The precipitation ratio of P in the leachate increased from 87.9% to 98.8% when the pH changed from 9 to 12. The dissolution ratio of P from the slag with 4 mass% of Na₂O reached 76.9% in this case. Most of the P in the steelmaking slag was thus recovered. The total recovery ratio of P from slag exceeded 73% when the pH exceeded 10. With the increase in the pH of the leachate, the total recovery ratio of P from slag increased, while the P₂O₅ content in the obtained product decreased.

Table 6. Precipitation ratio of P from leachate and total recovery ratio of P from slag in this process.

	Precipitation at various pH conditions
	pH=9	pH=10	pH=11	pH=12
Dissolution Ratio (pH=6)	0.769
Precipitation Ratio	0.879	0.950	0.972	0.988
Total Recovery Ratio	0.676	0.731	0.748	0.760

4. Conclusions

The optimum conditions for the selective leaching of P were determined by investigating the effects of the cooling rate of molten slag, Na₂O content in slag, and pH on the dissolution behavior of the modified slag in the aqueous solution. A precipitation method was applied to recover P from the leachate. The effect of pH on phosphate precipitation was studied. The results obtained are summarized below:

(1) With the increase in the Na₂O content in slag, the dissolution ratios of Ca, Si, and P from the modified slag increased at pH 6, while Fe was barely dissolved. The addition of 2.5–4.0 mass% of Na₂O to slag was sufficient to dissolve the majority of P-condensed solid solution in the aqueous solution.

(2) The dissolution ratio of P from the modified slag increased significantly when the pH decreased from 7 to 5. Further decrease in the pH caused little improvement in P dissolution, and resulted in dissolution of large amounts of Fe. To achieve better selective leaching of P, the pH should be controlled between 5 and 6.

(3) Compared to the quenched slag, the furnace-cooled slag exhibited a higher dissolution ratio of P and a lower dissolution ratio of Fe, indicating that slow cooling was necessary to realize selective leaching in the case of the modified slag.

(4) Most of the P in the leachate was precipitated via the addition of the Ca(OH)₂ solution. With the increase in pH of the leachate, the precipitation ratio of P in the leachate increased, while the P₂O₅ content in the obtained phosphate product decreased. In this process, approximately 70% of P in the slag was recovered in the form of the phosphate product, which can be used as a phosphate fertilizer.

References

• 1)   M. J.  Fisher-White,  R. R.  Lovel and  G. J.  Sparrow: ISIJ Int., 52 (2012), 1794.
• 2)   M.  Muhammed and  Y.  Zhang: Hydrometallurgy, 21 (1989), 277.
• 3)   E.  Matinde and  M.  Hino: ISIJ Int., 51 (2011), 220.
• 4)   G. W.  Qi,  A.  Parentich,  L. H.  Little and  L. J.  Warren: Int. J. Miner. Process., 34 (1992), 83.
• 5)   S.  Kitamura and  F.  Pahlevani: Tetsu-to-Hagané, 100 (2014), 500.
• 6)   K.  Shimauchi,  S.  Kitamura and  H.  Shibata: ISIJ Int., 49 (2009), 505.
• 7)   K.  Matsubae,  E.  Yamasue,  T.  Inazumi,  E.  Webeck,  T.  Miki and  T.  Nagasaka: Sci. Total Environ., 542 (2016), 1162.
• 8)   H.-J.  Li,  H.  Suito and  M.  Tokuda: ISIJ Int., 35 (1995), 1079.
• 9)   K.  Yokoyama, H. kubo, K. Mori, H. Okada, S. Takeuchi and T. Nagasaka: ISIJ Int., 47 (2007), 1541.
• 10)   C.  Li,  J.  Gao and  Z.  Guo: ISIJ Int., 56 (2016), 759.
• 11)   T.  Teratoko,  N.  Maruoka,  H.  Shibata and  S.  Kitamura: High Temp. Mater. Process., 31 (2012), 329.
• 12)   M.  Numata,  N.  Maruoka,  S. J.  Kim and  S.  Kitamura: ISIJ Int., 54 (2014), 1983.
• 13)   C.  Du,  X.  Gao,  S.  Kim,  S.  Ueda and  S.  Kitamura: ISIJ Int., 56 (2016), 1436.
• 14)   C.  Du,  X.  Gao,  S.  Ueda and  S.  Kitamura: ISIJ Int., 57 (2017), 487.
• 15)   C.  Du,  X.  Gao,  S.  Ueda and  S.  Kitamura: J. Sustain. Metall., 3 (2017), 671.
• 16)   R. P.  Gunawardane and  F. P.  Glasser: J. Mater. Sci., 14 (1979), 2797.
• 17)   J.  Ando: Bull. Chem. Soc. Jpn., 31 (1958), 196.
• 18)   J.  Tan,  S.  Zhao,  W.  Wang,  G.  Davies and  X.  Mo: Mater. Sci. Eng., B, 106 (2004), 295.
• 19)   E.  Valsami-Jones: Mineral. Mag., 65 (2001), 611.
• 20)   T.  Futatsuka,  K.  Shitogiden,  T.  Miki,  T.  Nagasaka and  M.  Hino: ISIJ Int., 44 (2004), 753.
• 21)   A. L.  Iglesia: Estud. Geol., 65 (2009), 109.
• 22)   D. A.  Kulik and  M.  Kersten: J. Am. Ceram. Soc., 84 (2001), 3017.
• 23)   W.  Kibalczyc: Cryst. Res. Technol., 24 (1989), 773.