2023 Volume 63 Issue 5 Pages 799-809
It is expected that phosphorus-concentrated slag produced by using the pyro-reaction between high phosphorus hot metal, which was prepared by the reduction of usual steelmaking slag with a small amount of hot metal, and oxidizing slag can become an important phosphorus resource instead of phosphate rock. For separating phosphorus from the phosphorus-concentrated slag, the citric acid leaching method has been reported to be available. In this study, the effects of cooling condition and oxidation treatment of phosphorus-concentrated slag on the dissolution behaviors of phosphorus and other elements were discussed. Compared to slow-cooled slag, CO2 blowing-quenched slag did not affect dissolution of phosphorus but decreased dissolution of iron. Oxidation treatment increased the dissolution ratio of phosphorus by about twice and decreased that of iron by one-sixth. It was confirmed by the microscopic analysis of the mineral phases in these slags that the concentration of SiO2 and FeO solid-solved in 3CaO·P2O5 (C3P) phase increased by slow cooling, and they precipitated as fine Fe2O3–SiO2 compound in C3P phase by oxidation. Furthermore, the dissolution behavior of SiO2 and FeO solid-solved in C3P phase was investigated. Experimental results showed that the C3P–2CaO·SiO2 dissolved easily and the C3P–3FeO·P2O5 dissolved hardly. From those experimental findings, it was presumed that the change in dissolution behavior due to slag cooling conditions and oxidation treatment were attributed to the concentration of SiO2 and FeO in C3P phase.
Approximately 70% of the steelmaking slag generated in the steelmaking process of the blast furnace-converter method is currently used as construction materials and soil improvement materials in Japan.1) However, it is considered that the amount of steelmaking slag used in those applications will decrease due to competition with byproducts and wastes generated in other industries.2) Therefore, new applications of steelmaking slag for the promotion of phytoplankton growth and seaweed bed formation in coastal area3,4,5,6,7,8,9,10,11,12,13,14,15,16,17, 18,19,20,21,22,23,24,25,26,27,28,29,30,31,32,33,34) and the salt removal in farmland damaged by the tsunami35,36,37,38,39,40,41,42,43,44) have been explored. On the other hand, some proposals have been made to minimize the slag amount by reusing it in steelworks and to recover valuable components in the slag such as phosphorus and iron.45,46)
Phosphorus is one of the essential elements in contemporary society, but the uneven distribution of phosphate ores has been a cause of international supply concerns. Therefore, the phosphorus recovery from steelmaking slag has a great meaning in resource strategy. Since the phosphorus concentration in usual dephosphorization slags is around one percentage, the following phosphorus condensation process is required for the economically effective phosphorus recovery. At first, the phosphorus pentoxide in usual dephosphorization slag is reduced with carbon at high temperature to produce a hot metal with high phosphorus concentration. After then, the hot metal is subjected to oxidative dephosphorization using a small amount of flux. The obtained slag containing high concentration of phosphorous (called “P-concentrated slag”) can be utilized as a phosphorus resource. To demonstrate those advantages, studies on carbon reduction47,48,49,50,51,52,53) and oxidative dephosphorization54,55) have been actively conducted.
Although the P2O5 concentration in the P-concentrated slag is as high as 30 mass%,55) which is comparable to that of common phosphorus ore, the high concentration of iron oxide in the slag becomes disadvantageous for manufacturing phosphorus products. Consequently, if phosphorus in P-concentrated slag could be separated from iron oxide, it would be possible to use it as a substitute of phosphate ore. In addition, the residue after the separation can be recycled as flux in the steelmaking process, thus reducing the overall quantity of steelmaking slag output.
In the previous report,56) the differences in composition and mineral phases between ordinary steelmaking slag and P-concentrated slag were investigated, and it was found that phosphorus, which is included in the easily acid-soluble 2CaO∙SiO2−3CaO∙P2O5 (called C2S−C3P) phase in ordinary steelmaking slag containing 2−4 mass% P2O5, is present as the C3P phase with small amounts of C2S in quenched P-concentrated slag. Also, the effects of pH of the leaching solution, type of leaching acid, and leaching method on the dissolution of elements from P-concentrated slag were evaluated. The results indicated that the lower pH of the leachate, the higher the dissolution ratio of phosphorus. It was also found that citric acid added to the leachate increased the dissolution ratio of phosphorus even under mild acidity. In the case of ordinary steelmaking slags, the chemical composition, species, proportions, and size of the mineral phase57,58,59,60,61) and the dissolution behavior of elements, including phosphorus,58) were varied depending on the cooling method. Furthermore, the increment of the Fe3+/Total Fe concentration ratio in the slag enhanced phosphorus dissolution in acid solution.62,63) Therefore, in this study, the effects of cooling rate and oxidation treatment on the phosphorus dissolution behavior from P-concentrated slag are examined.
Seventy kg of hot metal with an initial phosphorus concentration of 1.04 mass% was melted at 1400°C in a high frequency induction furnace lined with magnesia. Oxygen gas was blown from the top nozzle onto the hot metal surface at 0.11 Nm3/min while stirring with argon gas bottom-blowing at 5 L/min to proceed the phosphorus enrichment reaction, that is the formation of C3P–C2S solid phases, although there was a possibility that the metal droplets became suspended in slag. Simultaneously, 1.05 kg of lime was supplied to the hot metal surface from a sub-nozzle with argon carrier gas at 20 L/min. After 20 min treatment, the slag was taken out from the furnace and put into a pail for rapid cooling by blowing CO2 gas to obtain “CO2-quenched slag” sample. The cooling rate of this CO2-quenching was 380°C/min by the measurement with a thermocouple directly contacted with the slag near the center of the sample. On the other hand, the slag removed from furnace was cooled in a pail covered with an Al2O3–SiO2–ZrO2 adiabatic sheet (50 mm thick) at a cooling rate of 45°C/min, which was also measured by thermocouple in direct contact with slag. This was regarded as “slow-cooled slag” sample. To evaluate only the effect of the cooling condition on the phosphorus dissolution, the procedure of synthesizing slags before cooling was kept almost same. The obtained slags were crushed and classified into 25–53 μm using sieves. When preparing “oxidized CO2-quenched slag”, the classified CO2-quenched slag was heated at 1000°C for 5 days under an air atmosphere.
2.1.2. 3CaO∙P2O5−2CaO∙SiO2 and 3CaO∙P2O5−3FeO∙P2O5 Solid Solutions3CaO∙P2O5−2CaO∙SiO2 (C3P−C2S) and 3CaO∙P2O5−3FeO∙P2O5 (C3P−F3P) solid solutions were synthesized by the reaction of pre-synthesized 2CaO∙SiO2 (C2S) and Fe3(PO4)2 with 3CaO∙P2O5 (C3P) reagent at high temperature.
C2S was synthesized by the following procedure: Reagent-grade CaCO3 and reagent-grade SiO2 were mixed in a 2:1 molar ratio and compacted into a cylindrical shape at a pressure of 2 ton/cm2. The pellet was inserted into a platinum crucible and held at 1000°C for 10 hours under an atmosphere to decarbonate CaCO3. At this time, CaO produced by decarbonation was temporarily sintered with SiO2.
Fe3(PO4)2 was synthesized by the following method, referring to Kinoshita et al.64) Reagent-grade FeSO4∙7H2O (139 g) and reagent-grade NH4H2PO4 (40 g) were dissolved by 700 mL distilled water in a sealed glass bottle. Five mol/L NaOH (185 mL) was added at a rate of 1.5 mL/min while stirring at 500 r.p.m. using a magnetic stirrer to obtain Fe3(PO4)2∙8H2O crystals. During the synthesis, N2 was bubbled into the solution at 200 mL/min to prevent Fe2+ oxidation. The crystals, which were separated from the solution by suction filtration, were washed by stirring in 500 mL of distilled water at 400 r.p.m. for 3 min, and separated again by suction filtration. The Fe3(PO4)2∙8H2O crystals were rinsed by repeating this treatment three times. The obtained Fe3(PO4)2∙8H2O crystals were dried at 40°C and then kept in a vacuum dryer under 0.01 atm at 240°C for 65 h to remove combined water.
The synthesized C2S or Fe3(PO4)2 (called F3P) was crushed and mixed with reagent-grade C3P powder at a respective mass ratio, and compacted into a cylindrical shape. The C3P–C2S pellet charged in a platinum crucible was held at 1500 or 1600°C under air atmosphere. The C3P–F3P pellet was heated at 1500°C in Ar flowing (200 mL/min). After 48 h, those pellets were taken out from the furnace and rapidly cooled by He gas blowing. The obtained samples were identified by X-ray diffraction analysis (called XRD). Each synthesized solid solution sample was crushed and classified into 25–53 μm before leaching experiment.
The dissolution behavior of above solid solutions was compared with that of C3P. In order to minimize the effect of thermal history on the dissolution behavior of solid solutions and C3P, the reagent grade C3P compacted into cylindrical shape was also held at 1500°C for 48 h under air atmosphere, and rapidly cooled by He gas blowing.
2.2. Leaching Procedure 2.2.1. P-concentrated SlagAn overview of the slag leaching apparatus using a nylon mill pot and nylon-coated steel balls was noted in a previous paper.56) In this study, the leaching medium is 0.01 mol/L citric acid solution, which had a high phosphorus dissolution ratio in the previous report.56) The citric acid solution (800 mL) and P-concentrated slag powder (1.00 g) were charged into the leaching apparatus, and the nylon mill pot was rotated at approximately 90 r.p.m., which corresponded to 80% of the critical rotational speed. At first, the leaching test was carried out for a specified time without pH adjustment, and then pH of the leachate was maintained at pH = 4 by dropping 1 mol/L NaOH solution using peristaltic pump connected with an automatic pH controller, while the pH of the leachate was continuously measured with a pH glass electrode immersed in the leachate. Subsequently, the leaching test was carried out with automatic titration of 1 mol/L HCl−0.01 mol/L citric acid solution to maintain pH=3, and then pH=2. The leaching test was carried out at room temperature (20–25°C) under atmosphere. During the experiment, 2 to 5 mL of the leachate was sampled using a plastic syringe attached to a membrane filter cartridge (open pore size: 0.2 μm) at specified time intervals. After the leaching test, the solid phase was separated by suction filtration with a membrane filter (open pore size: 0.45 μm), and the separated solid phase was dried at 20°C for at least 48 h in an incubator.
2.2.2. C3P-based Solid SolutionThe 0.01 mol/L citric acid solution with pH=4 was prepared by dissolving 0.38 g of reagent-grade citric acid and 0.091 g of reagent-grade NaOH in 200 mL of distilled water. Two grams of synthesized C3P-based solid solution powder was added to this solution and stirred at 400 r.p.m. using a magnetic stirrer. One mol/L HCl-0.01 mol/L citric acid solution was dripped using peristaltic pump connected with an automatic pH controller to maintain pH=4. The procedures for the sampling of the solution during leaching test and the collection of the solid phase after leaching test were the same as those mentioned in Section 2.2.1.
2.3. AnalysisThe slag powder (0.1 g) was mixed with reagent-grade Na2CO3 (2.0 g) and reagent-grade B2O3 (1.0 g) in a platinum crucible, and the mixture was melted at around 1000°C. After the treatment, the sample was digested with warm dilute hydrochloric acid and quantified using an inductively coupled plasma-atomic emission spectroscopy (ICP-AES). The synthesized C3P-based solid solution was dissolved by mixed acid of HCl and HNO3 and quantified using ICP-AES.
A XRD method was used to identify the mineral phases of the P-concentrated slags and the C3P-based solid solutions. Elemental mapping on the polished surface of the slag lump and quantitative analysis of each mineral phase were conducted using a field emission-electron probe microanalyzer (FE-EPMA). Furthermore, the observation and quantitative analysis of the solid phase after leaching was performed by FE-EPMA.
The chemical compositions of the synthesized CO2-quenched and slow-cooled slag are shown in Table 1. From the XRD analysis of those slags shown in Fig. 1, it is clear that the mineral phases of both CO2-quenched slag and slow-cooled slag are C3P, C3P–C2S (C5PS), and FeO, and no differences in the types of mineral phases due to cooling rate are observed. The oxidized CO2-quenched slag indicated in Table 1 and Fig. 1 will be explained later. Figures 2 and 3 are the results for elemental mapping given by EPMA on the cross sections of the slow-cooled slag and the CO2-quenched slag, respectively. Three phases are recognized in both slags: phase a is rich in calcium and phosphorus, phase b contains only iron or iron and magnesium, and phase c contains all elements. It should be noted that the phase b in the slow-cooled slag (Fig. 2) is the iron oxide-5 mass% MgO phase, which is thought to have grown during slow cooling and is considered as wüstite phase in the XRD result (Fig. 1), while it appears as a dendritic iron oxide without MgO in the CO2-quenched slag (Fig. 3). That is, although the constituent mineral phase “wüstite” is identified in both slags its composition and morphology are changed by the cooling rate. Phase d in Fig. 2, which is not iron oxide but metallic iron according to EPMA analysis, was also observed in the CO2-quenched slag in another field of view from Fig. 3. It is not clear from this study whether this metallic iron is a suspended granular iron or it is generated by Eq. (1) during cooling.
| (1) |
| CaO | SiO2 | P2O5 | FeO | Fe2O3 | Al2O3 | MgO | |
|---|---|---|---|---|---|---|---|
| CO2-quenched | 39.7 | 6.1 | 28.0 | 21.9 | – | 0.7 | 2.3 |
| Slow-cooled | 42.5 | 8.3 | 23.4 | 20.7 | – | 1.6 | 2.2 |
| Oxidized CO2-quenched | 35.7 | 8.3 | 25.6 | – | 25.0 | 1.7 | 2.4 |
X-ray diffraction peaks obtained by XRD analysis of P-concentrated slag before leaching.
Elemental mapping on the cross section of slow-cooled slag obtained by EPMA.
Elemental mapping on the cross section of CO2-quenced slag obtained by EPMA.
Quantitative analysis by EPMA was performed on 40 to 60 phases containing high phosphorus concentration in each slag. The obtained values were converted to the CaO–SiO2–P2O5 pseudo-ternary system, and plotted in the CaO–SiO2–P2O5 ternary phase diagram65) in Fig. 4(a). The amount of C2S solid-solved in C3P–C2S phase in the CO2-quenched slag (marked with ●) is very small compared with that in the slow-cooled slag (marked with ▲). Figure 4(b) shows the relationship between FeO and SiO2 concentration in C3P–C2S phase. The concentration of solid-solved SiO2 in C3P–C2S phase of the slow-cooled slag (marked with ▲) is up to 23 mass%, which is higher than that of CO2-quenched slag (marked with ●). Phase c, which contains all slag components, is considered to be the final solidification section of the molten slag, since the XRD peaks corresponding this phase are not detected and the baseline shape suggests the presence of a glassy phase with an amorphous structure in Fig. 1. The average compositions of the 26–28 glassy phases determined by EPMA are plotted on the CaO–SiO2–FeOx ternary phase diagram66) in Fig. 5(a), and the relationships between basicity (CaO/SiO2 mass% ratio) and P2O5 concentration are shown in Fig. 5(b). It can be seen that the slow-cooled slag (marked with ▲) has higher basicity and P2O5 concentration than CO2-quenched slag (marked with ●) in Fig. 5(b), although the slow-cooled slag has a variety of low and high FeO concentration parts in Fig. 5(a).
Composition of P-enriched phases in CO2-quenched and slow-cooled slag projected on the CaO–SiO2–P2O5 ternary phase diagram65) (a) and plotted in the relations between SiO2 and FetO concentrations (b).
Composition of glassy phases in CO2-quenched and slow-cooled slags projected on the CaO–SiO2–FeO ternary phase diagram66) (a) and the relations between C/S ratio and P2O5 concentration (b).
Figure 6 shows the variation of pH and the concentration of each element in the leachate (citric acid solution) during the leaching test of CO2-quenched slag (marked with ○) and slow-cooled slag (marked with △). When pH is not controlled at the beginning of leaching test, the concentration of each element in the leachate increases immediately after the slag addition, and then increases gradually. In the case of CO2-quenched slag, the concentrations of iron and calcium are lower and those of phosphorus and silicon are higher compared with the case of slow-cooled slag. Subsequently, when controlled to pH=4 by dropping NaOH solution, the concentration of each element changes little. The final phosphorus concentrations at pH=4 in the leachate of the CO2-quenched and slow-cooled slag were 67 and 54 mg/L, respectively. Then, by adjusting pH at 3 by the addition of dilute HCl, the elution of phosphorus and calcium proceeds quickly in the first 60 min, but it almost stops after 60 min. In controlling pH=2, the dissolution of phosphorus and calcium proceeds rapidly in only the initial stage. By leaching at a lower pH, the higher phosphorus concentration in solution can be obtained whereas the other elements are also present at higher concentrations in the same solution. The reason for this is considered to be that the lower the pH of a solution, the higher the solubility of P-enriched phases (C3P and C3P–C2S) and P-containing secondary products (6-30 mass%P2O5‒0-25 mass%CaO‒0-40 mass%SiO2‒15-50 mass%FeO‒H2O), as mentioned in a previous report.56) In the present study, the pH control was continuously carried out from pH=4 to 2. Even if the pH is controlled to 3 or 2 initially, it can be considered that the phosphorus dissolution ratio will not change because this ratio is affected by the equilibrium solubility of phosphates. On the contrary, the phosphorus dissolution rate in the initial stage of leaching is expected to be greater in controlling the pH to 3 or 2 initially, when the rate is dependent on phosphorus concentration in solution.
Comparison of dissolution behavior of elements from CO2-quenched, slowly cooled and oxidized CO2-quenched slags.
From the experimental data shown in Fig. 6, the dissolution ratio of element M (RM) is calculated using Eq. (2).56)
| (2) |
Dissolution ratios of the elements contained in P-concentrated slags plotted against pH value.
The mineral phase and composition of C3P-based solid solutions, which were synthesized by sintering at 1500 or 1600°C, are listed in Table 2. It is confirmed by the characterization using XRD that 0 mass%C2S (0S in the table) is α-C3P (α phase), C3P-13 mass%C2S (10S) and C3P-52 mass%C2S (50S) are C3P–C2S solid solution (R phase), and C3P-32 mass%C2S (30S) is 5CaO∙P2O5∙SiO2 (S phase). According to the CaO–P2O5 phase diagram,67) the α-C3P is stable above 1125°C and β phase is stable below that temperature. Therefore, it is suggested that the quenching in this study was sufficiently fast for the retention of α-C3P. The compositions of the C3P–C2S solid solutions synthesized in this study are plotted in the C2S–C3P pseudo-binary phase diagram68) in Fig. 8. The 10S and 50S, which were synthesized at 1500 and 1600°C, respectively, are in a complete solid solution (R phase) area, corresponding to the fact that they are R phase in Table 2. In contrast, although 30S synthesized at 1600°C is also in R phase in Fig. 8, it was identified as S phase in Table 2. The reason for this discrepancy may be the following: Serena et al.69) reported that the R phase with C2S/C3P molar ratio=1/1 (64 mass% C3P) transformed inevitably to silicocarnotite (S phase) during rapid cooling from 1500°C. Even though the heating temperature in this study was 1600°C and the temperature difference to the transformation point of 30S (1450°C) was larger than in their case, the transformation from R phase to S phase proceeded because of insufficient quenching rate.
| Phase | CaO | SiO2 | P2O5 | FeO | |
|---|---|---|---|---|---|
| 0S | α-C3P | 56.4 | – | 43.6 | – |
| 10S | C3P–C2S (R) | 56.8 | 4.5 | 38.6 | – |
| 30S | 5CaO∙P2O5∙SiO2 (S) | 58.7 | 11.3 | 30.0 | – |
| 50S | C3P–C2S (R) | 59.8 | 18.0 | 22.1 | – |
| 5F | β-C3P | 53.4 | – | 43.5 | 3.1 |
| 10F | β-C3P | 51.0 | – | 43.0 | 6.0 |
| 10S10F | C3P–C2S (R) | 54.0 | 3.3 | 36.9 | 5.7 |
Composition and synthesized temperature of 2CaO∙SiO2–3CaO∙P2O5 compounds (0S to 50S) plotted in 2CaO∙SiO2–3CaO∙P2O5 phase diagram.68)
As listed in Table 2, C3P-5mass%F3P (5F in the table) and C3P-10mass%F3P (10F) are β-C3P phase (β phase) and C3P-10mass%C2S-10mass%F3P (10S10F) is R phase. Since there have been no reports for C3P–F3P pseudo-binary system, the analyzed compositions of 5F and 10F are plotted in the isothermal cross section of the CaO–P2O5–FeO ternary phase diagram at 1600°C70) in Fig. 9, although they were sintered at 1500°C. Only a single β phase was identified in 5F and 10F by XRD analysis, indicating that F3P solid-solves in C3P up to 10 mass%. Yoo et al.71) explained that iron ion, whose ionic radius is smaller than that of calcium ion, improves the thermal stability of β-C3P phase. However, it is not clear from this study whether the β phase is stable in the C3P–F3P solid solution at high temperature or was transformed from α phase during quenching.
Composition of 3CaO∙P2O5–3FeO∙P2O5 compounds plotted in CaO–FetO–P2O5 phase diagram.70)
Figure 10 shows the relationships between the dissolution ratio of each element in C3P–C2S and C3P–F3P solid solutions and the C2S or F3P concentration in those solid solutions as a function of leaching time (5, 10, and 20 min). It is found in Fig. 10(a) that when C2S solid-solved in C3P, the dissolution ratios of phosphorus and silicon became higher. Pietak et al.72) described that the substitution of Si to C3P generates lattice defects and leads to an increase of the C3P solubility, which is the same trend as in Fig. 10(a). On the other hand, Fig. 10(b) shows the suppression of phosphorus dissolution by the solid-solving of F3P to C3P. Du et al.63) suggested that the dissolution of elements becomes hard when another element with higher electron density substitutes. Namely, since Fe2+ has a higher electron density than Ca2+, the replacement of Ca2+ with Fe2+ reduces the phosphorus dissolution. The phosphorus dissolution ratio of C3P-10mass%C2S-10mass%F3P is slightly lower than that of C3P-10mass%C2S, but is greatly higher compared to C3P-10mass%F3P. Therefore, it is expected that the elution-enhancing effect of solid-solved C2S is greater than the elution-suppressing effect of solid-solved F3P.
Compositional dependence on dissolution ratio of phosphorus and silicon from C3P–C2S (a) and that of phosphorus and iron from C3P–F3P (b) as a function of leaching time.
The oxidized CO2-quenched slag was obtained by heating the CO2 quenched slag powder (25–53 μm) at 1000°C in an air atmosphere. The result of XRD measurement shown in Fig. 1(c) indicates that the characteristic peaks of FeO in the slag before oxidation (Fig. 1(b)) disappears and a new Fe2O3 phase is present after the oxidation treatment. Figure 11 is elemental mapping images on the polished cross section of oxidized CO2-quenched slag powder obtained by EPMA. By oxidation, small amounts of FeO and SiO2 solid-solved in the CaO–P2O5 phase (Fig. 4) changes to Fe2O3–SiO2 precipitates of less than 1 μm. In addition, this Fe2O3–SiO2 phase is also present densely outside the CaO–P2O5 phase. The other main phase observed in Fig. 11 is Fe2O3.
Elemental mapping on the cross section of oxidized CO2-quenched slag obtained by EPMA.
The dissolution behavior of elements from the oxidized CO2-quenched slag is given in Fig. 6 (marked with □). Compared to the CO2-quenched slag (marked with ○), the concentrations of phosphorus and calcium is promoted by oxidation treatment. Especially, the former becomes about 1.7 times larger under a weakly acidic condition of pH=4. On the other hand, more than 100 mg/L of iron is released from the slag before oxidation, but less than 20 mg/L after oxidation at any pH. This experimental finding suggests that the iron dissolution can be significantly suppressed by oxidation.
The relationships between the dissolution ratio of each element from the oxidized CO2-quenched slag and pH are shown by gray marks in Fig. 7. In comparison with that from the CO2-quenched slag (open marks), the oxidation treatment increases the dissolution ratio of phosphorus and calcium at pH=3 and 4, and decreases the dissolution ratio of iron and silicon at all pH. When the leaching is carried out at pH=3, the experimental result that the phosphorus dissolution ratio of the oxidized slag is 0.98 means almost complete elution of phosphorus from the slag. On the contrary, the lower dissolution ratios of iron and silicon from the oxidized slag suggest that the oxidation treatment of P-enriched slag is effective for selective phosphorus extraction.
Thermal history (smelting temperature, cooling rate, and quenching temperature) and slag composition are considered to be the factors that affect both the constitution and size of mineral phases and the dissolution behavior of elements in ordinary steelmaking slag. In this study, the effect of cooling rate is discussed. In addition, since the phosphorus dissolution was enhanced by increasing the Fe3+/Total Fe concentration ratio in the ordinary slag,62,63) the effect of oxidation treatment on P-concentrated slag is also discussed.
4.1. Cooling RateSince the C3P phase in the CO2-quenched slag contained FeO and a smaller amount of SiO2 in Fig. 4(b), C3P phase is presumed to be the primary crystal in this slag system. From the experimental finding that the SiO2 concentration in C3P phase of slow-cooled slag was higher than that of CO2-quenched slag, the primary C3P crystal is considered to be changed to the C3P–C2S phase containing FeO during cooling. On the other hand, it was found from Fig. 5(b) that the basicity and P2O5 concentration of the glass phase increased by slow cooling, while the FeO concentration range in Fig. 5(a) was not varied. It can be predicted from the plot of the glass phase of the CO2-quenched slag in Fig. 5(a) (marked with ●) that the Fe2SiO4 phase is formed with cooling, but the phase was not observed in the slow-cooled slag (marked with ▲). This may be due to the reason that the P2O5 concentration in P-enriched slag is significantly higher than that in ordinary steelmaking slag, resulting in complex changes in phase equilibrium relationships. Although there have been many studies73,74,75,76,77,78,79,80) on the phase transitions caused by the reactions between C2S–C3P phase and liquid slag, the reactions between C3P phase and liquid slag during cooling were not reported.
The compositional changes of the slag melt with increasing SiO2 or FeO concentration in the C3P primary crystal during cooling are explained as follows.
When assuming a mechanism in which the C3P phase precipitates from the slag melt as a primary crystal and is transformed to C3P–C2S phase by the diffusion of Ca2+ and SiO44− into C3P phase during slow cooling, the basicity of the slag melt is reduced by the formation of C3P–C2S phase. Therefore, the difference in composition due to cooling rate shown in Fig. 5(a) cannot be explained. As another mechanism, when SiO44− in slag melt diffuses into C3P phase by replacing PO43−, the SiO2 concentration in slag melt is reduced, while the basicity and P2O5 concentration of the slag melt increase by the transportation of the excess CaO and P2O5 in C3P phase to slag melt. Regarding the solid solution of FeO in C3P–C2S phase shown in Fig. 4(b), Du et al.63) mentioned that Ca2+ in the C2S–C3P phase can be replaced by Fe2+ based on the Hume-Rothery rule because Ca2+ and Fe2+ have the same ionic valence and almost similar ionic radius. Since the FeO concentration in the C3P phase with low C2S concentration in the CO2-quenched slag (marked with ●) is comparable to that in the C3P–C2S phase in the slow-cooled slag (marked with ▲), it is considered that FeO solid-solves in the C3P phase by replacing Fe2+ with Ca2+.
The FeO concentration in the glassy phase of slow-cooled slag is almost similar to that of CO2-quenched slag, although the data points are widely dispersed in Fig. 5(a). The reason for the crystallization of dendritic wüstite in the glassy phase of CO2-quenched slag, as shown in Fig. 3, might be as follows. When the cooling rate is high and Fe2+ diffusion in the glassy phase is insufficient, the Fe2+ transfer could not correspond to the composition change of the glassy phase, and finally the excessive Fe2+ crystallized as dendritic wüstite phase.
The variation of dissolution behavior of elements with cooling rate of slag shown in Figs. 6 and 7, is explained as follows. From the results for the leaching experiments of the synthesized C3P mineral phases, the dissolution ratio of phosphorus from 10S10F (closed marks in Fig. 10), in which FeO and SiO2 were both solid-solved, is higher than that from 5F and 10F (open marks), in which only FeO was solid-solved. Since the C3P in the slow-cooled slag contains both FeO and SiO2, the phosphorus dissolution ratio of the slag must be high. However, there is no significant difference in the phosphorus dissolution ratio between the slow-cooled and the CO2-quenched slag in Fig. 7. From the changes of each concentration in the initial stage of leaching (Fig. 6), it is expected that when C3P–C2S with solid-solved FeO dissolved rapidly in the early leaching stage, CaO–SiO2–FeO–P2O5–H2O gel (C–S–F–P–H gel) is generated at the dissolution site on the slag surface and formed a film or suspended in eluate. The SEM image of CO2-quenched slag after leaching at pH=3.7–3.8 (uncontrolled) for 24 h and then at pH=4 for 24 h is shown in Fig. 12. The compositions of points 1 to 6 in this figure, which were quantitatively analyzed by EPMA, are listed in Table 3. Although the C3P phases (points 1–3) are not covered by C–S–F–P–H gel, the suspended C–S–F–P–H gel (point 6) is observed. Therefore, it is reasonable to say that the concentrations of ions in the eluate have reached the saturation value of the gel. Because the PO43− ion, which is contained in the C3P phase and its hydration products, is weak acid, the dissolution of those phases proceeds with decreasing pH thermodynamically. It was confirmed that the dissolution of oxides was accelerated at lower pH.81) As a result, the dissolution ratio of phosphorus and other elements becomes higher by decreasing pH.
Backscattered electron image of CO2-quenched slag after leaching at pH=3.7–3.8 (uncontrolled) for 24 h and then at pH=4 for 24 h.
| CaO | P2O5 | SiO2 | FeO | Mineral phase | |
|---|---|---|---|---|---|
| 1 | 43.4 | 33.0 | 0.6 | 2.0 | C3P |
| 2 | 44.7 | 38.9 | 0.4 | 2.9 | |
| 3 | 47.6 | 37.1 | 0.8 | 1.3 | |
| 4 | 0.9 | 1.0 | 0.2 | 83.1 | FeO |
| 5 | 10.5 | 0.5 | 26.2 | 29.0 | Matrix |
| 6 | 30.8 | 20.0 | 12.0 | 18.0 | Secondary product |
As described in Section 3.4, small amounts of FeO and SiO2 solid-solved in the CaO–P2O5 phase (Fig. 4) changed to Fe2O3–SiO2 precipitates and outer layer of the CaO–P2O5 phase by oxidation.
Compared to the concentration of FeO solid-solved in the C2S–C3P solid solution phase of CaO–SiO2–FeO–P2O5 slag, that of Fe2O3 in the C2S–C3P phase of CaO–SiO2–Fe2O3–P2O5 slag was lower above 1300°C.82,83) Similarly, when Fe2+ (FeO) in the C3P phase is oxidized to Fe3+ at 1000°C, the Fe3+ above the solubility limit is drained off from the C3P phase and precipitated. From the charge balance in the C3P–C2S phase, it can be said that SiO44− ejected from the C3P phase combines with the Fe3+ to form the Fe2O3–SiO2 crystalline product. Similarly in the glass phase, the formation of Fe2O3–SiO2 crystals is expected with the FeO oxidation. However, in the XRD result shown by Fig. 1(c), not Fe2O3–SiO2 phase but Fe2O3 phase was identified. Therefore, the concentration of SiO2 in the Fe2O3–SiO2 phase crystallized in the C3P–C2S phase is considered to be low.
The presence of Fe2+ in C3P and C3P–C2S phases suppresses the phosphorus dissolution as discussed in Section 3.3. The change to Fe3+ by oxidation decreased the concentration of iron oxides in the P-enriched phase, thereby facilitating the progress of the phosphorus dissolution. It is also shown in Fig. 6 that oxidation can significantly reduce the iron dissolution. This is due to the formation of mineral phases containing trivalent iron ions such as dicalcium ferrite (2CaO∙Fe2O3), magnesioferrite (MgO∙Fe2O3) and Fe2O3, which are highly resistant to dissolution into water,84,85) as seen in the elemental mapping results (Fig. 11). However, since the XRD peaks for the former two phases were not observed in Fig. 1, their amounts are estimated to be so small.
To prepare P-concentrated slags with different cooling rate, the slag melt of CaO–SiO2–FeO system containing 23 to 28 mass% P2O5 was cooled slowly under air atmosphere or quenched by blowing CO2. In addition, the CO2-quenched slag was held at 1000°C for 5 days under air atmosphere to be oxidized completely. Leaching experiments were conducted on these slags using citric acid as eluate to investigate the effect of cooling rate and oxidation on the phosphorus dissolution from the slags. Furthermore, the effects of dissolved components such as Si and Fe in the C3P phase on its dissolution behavior in citric acid solution were also examined by using synthesizing the C3P phase with solid-solved SiO2, FeO, or both. From those results, the following findings were revealed.
(1) The C3P phase in the slow-cooled slag had higher SiO2 concentration than that in the CO2-quenched slag. Dissolution ratio of phosphorus in citric acid solution did not differ between both slags, but those of iron and calcium for the CO2-quenched slag were lower and that of silicon was higher than those for the slow-cooled slag.
(2) After oxidation of the CO2-quenched slag, the concentration of iron in the C3P phase decreased due to a finely dispersed arrangement of Fe2O3-dominant compounds. In leaching test with citric acid solution, the phosphorus dissolution ratio from the oxidized CO2-quenched slag was about two times higher than that from the original slag even under the weak acidic condition of pH=4, while the iron dissolution ratio was suppressed to 1/6.
(3) The dissolution of the C3P phase in citric acid solution was enhanced by SiO2 and suppressed by FeO solid-solved in C3P phase.
From the above findings, it is concluded that rapid cooling and oxidizing treatment of the P-concentrated slag is desirable to promote phosphorus dissolution and to suppress iron dissolution in citric acid solution.
This paper is based on results obtained from a project, JPNP12004, subsidized by the New Energy and Industrial Technology Development Organization (NEDO).
