Basic Evaluation of Phase Relation in a Phosphorus-Containing System Saturated with CaSiO₃ at Elevated Temperatures for the Utilization of Steelmaking Slag and Sewage Sludge as Phosphorus Resources

Abstract

In view of obtaining fundamental information on phosphorus recovery from steelmaking slag and sewage sludge, a laboratory experiment using the model specimen of a slag/sludge mixture prepared at 1573 K was carried out to investigate phase relation in a [CaO-SiO₂-P₂O₅]-based system. The triangular compositional region, comprising of apices CaO·SiO₂ (CS), 3CaO·P₂O₅ (C₃P), and 2CaO·SiO₂-3CaO·P₂O₅ solid solution (C₂S-C₃Pss), was considered with particular interest. In this region, using SEM-EDX observation it was found that solid saturated CS and the solid C₂S-C₃Pss with a relatively high phosphorus content can coexist. With the addition of Al₂O₃ or Fe₂O₃ to the same specimens, the liquidus phase appeared as a third phase; however, CS and C₂S-C₃Pss phases were still observed for up to 5mass% addition. The further addition of Al₂O₃ or Fe₂O₃ to 10mass% resulted in dissolution of the solid CS phase, although C₂S-C₃Pss remained as the phosphorus concentrated phase. These results show that phase equilibria based on the ternary system would be stable and be beneficial for phosphorus recovery.

1. Introduction

Phosphorus is an essential element for life as it is found in bones, teeth, and even genes. Phosphorus is also important for industries such as plating, surface treatment of metal, and semiconductors. The global consumption of phosphorus increases year by year. One of the reasons for this is the growing needs of fertilizers for foods due to the growing world population. Another reason is the increasing demands of bioethanol. Bioethanol is produced from crops like corn, sugar cane, or soy bean, so more fertilizer is required to cultivate these plants. Therefore, phosphorus consumption is becoming larger.

Phosphorus is mainly produced from phosphate ore. The typical content of phosphorus in phosphate ore is about 30–40 mass%P₂O₅ [1]. Phosphate ore of a high quality is non-renewable and exhaustible. An alternative to natural phosphate is steelmaking slag, which contains phosphorus [2]. In an integrated steelmaking process, phosphorus originally comes from the iron ore and must be sufficiently removed as a typical impurity from molten pig iron into steelmaking slag. Even though not all steel mills use pig iron as a raw material, taking into account the worldwide contribution of the steel production of integrated mills, it is obvious that steelmaking slag contains a considerable amount of phosphorus and could be recognized as a promising phosphorus resource. An example of the composition of steelmaking slag containing phosphorus is given in

Table 1. Examples of composition of steelmaking slag and sewage sludge (in mass%).

By-Products	CaO	SiO2	P2O5	FeO	Fe2O3	Al2O3	MgO	MnO	C/S *
De-P Slag	26.3	22.2	3.0	24.7	7.6	5.1	6.1	4.9	1.2
Sewage Sludge	15.3	30.8	22.2	0.9	16.5	14.1		0.1	0.5

* (mass%CaO)/(mass%SiO₂).

[3]. Major components are CaO, SiO₂, FeOx, and P₂O₅. The P₂O₅ content is not sufficiently high, although the generation is huge. Some pioneering work on phosphorus recovery from steelmaking slag has been reported [4,5,6,7,8,9], but none were carried out industrially. Commonly, steelmaking slag is used for civil engineering, such as in road construction materials, which does not require a phosphorus resource.

In recent operations of hot metal dephosphorization (De-P) in steel refining, the basicity ((mass%CaO)/(mass%SiO₂)) of De-P slag has been found to be less than 1.5, as shown in Table 1. Previous mineralogical research on steelmaking slag has been mainly focused on a higher basicity [10,11,12,13,14,15,16]. If such De-P slag is considered as a phosphorus resource, the phase equilibria of the phosphorus-concentrated phase in the low basicity slag should be clarified [17].

Another alternative for a phosphorus resource is sewage sludge, which comes from human consumption and waste. An example of the composition of sewage sludge is given in Table 1 [18]. The major components of sewage sludge are CaO, SiO₂, FexO, P₂O₅, and Al₂O₃, which is similar to steelmaking slag. The basicity is less than unity and is lower compared to steelmaking slag. This property would make it difficult to apply pyro-treatment as a mass process to sewage sludge; therefore, most of the phosphorus recovery processes from sewage sludge are based on the wet technique using leaching [19].

Considering the above comparison, the authors conceived to mix steelmaking slag with sewage sludge for recovering phosphorus through pyro-treatment, and started their preliminary research [17]. CaO, SiO₂, and P₂O₅ were chosen as common oxide components in the phosphorus-containing slag and sludge. This system is characterized by a solid solution that commonly exists between 2CaO·SiO₂ (C₂S) and 3CaO·P₂O₅ (C₃P) as end members [20,21,22,23,24,25,26,27,28,29,30,31,32,33,34]. For the CaO saturated region in this system, Hasegawa et al. reported a phase relationship at 1573 K [29]. The C₂S−C₃P solid solution (hereafter C₂S-C₃Pss) coexists with CaO in a wide compositional range, except for the composition near C₂S or C₃P. Matsugi et al. also showedthe phase relationship in the counterpart of the above region, that is, the system saturated with 3CaO·2SiO₂ (C₃S₂) or CaO·SiO₂ (CS), and C₂S-C₃Pss [31]. Following their work, the lower basicity and higher phosphorus content region, as shown by bold triangle in the figure, has remained uninvestigated.

Figure 1 shows the isothermal section of the [CaO-SiO₂-P₂O₅] ternary phase diagram at 1573 K of Saito et al. [33]. Some stoichiometric compositions such as 5CaO·P₂O₅·SiO₂ and 7CaO·P₂O₅·2SiO₂ are denoted among C₂S-C₃Pss, and these are expressed as C₅PS and C₇PS_2, respectively, throughout the paper. According to the coexistence of CS and C₃P in this pseudo-binary system [35], the connected line between these compounds is drawn.

The composition of De-P slag and sewage sludge are projected on the ternary diagram. The composition of the mixture moves along the tie line, depending on the mixing ratio, and forms a triangular region comprising of apices CS, C₃P, and C₂S-C₃Pss. Therefore, the phase relation in this region should be understood with particular interest on the separation of the P-concentrated solid phase. One can consider that CS and C₂S-C₃Pss would coexist by considering the aforementioned coexistence of CS and C₃P. However, the influence on phase relation by other components involved with the mixing of De-P slag and sewage sludge are still unknown.

In this study, aiming to utilize steelmaking slag and sewage sludge as a phosphorus resource, the phase equilibria of the common oxide system [CaO-SiO₂-P₂O₅] and additives under CaSiO₃ saturation were investigated at 1573 K in order to obtain a fundamental understanding of the phosphorus-concentrated phase in the slag.

2. Materials and Methods

2.1. Materials

The starting materials for the experiment were CaCO₃, SiO₂, Ca₂P₂O₇, Al₂O₃, and Fe₂O₃ of reagent grade (Fujifilm Wako Pure Chemical co., Osaka, Japan). These were weighed and mixed at a predetermined composition in advance to obtain the experimental samples, as listed in

Table 2. Composition of the samples employed (in mass%).

Sample No.	CaO	SiO2	P2O5	Al2O3	Fe2O3
M1-0	55.3	38.6	6.1
M2-0	53.6	33.6	12.8
M3-0	53.2	32.1	14.7
M3-A2	52.1	31.5	14.4	2.0
M3-A5	50.5	30.5	14.0	5.0
M3-A10	47.9	28.9	13.2	10.0
M3-F2	52.1	31.5	14.4		2.0
M3-F5	50.5	30.5	14.0		5.0
M3-F10	47.9	28.9	13.2		10.0

. The mixed material was pressed into a pellet shape.

2.2. Experimental Procedure

The experimental sample was heated to 1573 K in a resistance furnace. Figure 2 shows a schematic illustration of the self-produced resistance furnace. The furnace was equipped with 12 SiC heaters (A8-2, Siliconit Inc., Tokyo, Japan) and a mullite tube (HB, od50 × id42 × 1000 mm, Nikkato Inc., Tokyo, Japan). The sample was placed in a platinum plate, mounted on a refractory boat and set in a uniform temperature zone of the reaction tube, which was determined with preliminary heating. The sample was heated at 100 K/min to the experimental temperature in a dehydrated purified air stream at 100 mL/min, and maintained for 50 h. The temperature was measured with a Pt-PtRh13 thermocouple placed 10 mm from the samples, and was controlled within ±3 K.

After being maintained for a predetermined duration, the sample was pulled outside the reaction tube with the aid of an alumina rod for quenching. Then, the sample was embedded in cured resin, and was finally polished to be served for the SEM-EDX analysis (JSM-IT300LV, JEOL Ltd., Tokyo, Japan). As water was not used when polishing the sample so as not to affect its composition, a scratch was appeared on the surface of the specimen, as shown in the SEM images below (backscattered electron image). The quantitative spot analysis of the composition of each phase was made for more than three points, avoiding such scratches, and the variations were within 1mass%. The area fraction of each phase in the SEM image was evaluated using ImageJ software for at least five images of each sample.

3. Results

3.1. CaO-SiO₂-P₂O₅ Ternary System

Figure 3, Figure 4 and Figure 5 show SEM images of M1-0, M2-0, and M3-0, respectively. All of the samples were not melted by appearance, and different solid phases in bright and dark contrast were observed. “SS” in the figures and

Table 3. Compositional analysis of the observed phases in the ternary samples (in mass%).

Sample No.	CaO	SiO2	P2O5	Note 1
M1-0	49.3	50.6	0.1	CS
	61.7	24.2	14.1	SS
M2-0	48.7	51.1	0.2	CS
	55.5	12.1	32.4	C5PS
	59.7	23.4	16.9	SS
M3-0	48.4	51.4	0.2	CS
	56.8	11.9	31.3	C5PS

¹ Estimated phase. SS: C₂S-C₃P solid solution except for C₅PS or C₇PS₂.

represents C₂S-C₃P solid solutions, except for stoichiometric compounds such as C₅PS or C₇PS₂. Throughout the results, the black regions in the SEM images are holes in the specimen filled with cured resin. The results of the compositional analysis of these phases with SEM-EDX are listed in Table 3, and plausible phases estimated therefrom are also given. All the employed data are the average of at least three measured points.

The composition of each phase is also plotted on an isothermal section of the [CaO-SiO₂-P₂O₅] ternary phase diagram at 1573 K in the figures. The bulk composition of the sample is also given in the figure with an X mark.

For samples M1-0 and M3-0, two phases were observed. One of them, in a bright color, had a relatively high phosphorus content and was identified as C₂S-C₃Pss for M1-0 and a stoichiometric component of C₅PS for M3-0. The other, in a dark color, had a high SiO₂ content and was almost free of P₂O₅, which was identified as CS. The composition of the observed C₂S-C₃Pss and C₅PS phases lay at opposite sides of the extension line from CS through the bulk composition, showing reasonable phase relation.

For sample M2-0, three phases with different contrast were observed. One of them, in the darkest color, was almost free of P₂O₅ and was identified as CS. Another two phases in a relatively brighter color, contained some P₂O₅ and had a composition of C₂S-C₃Pss and C₅PS, respectively. It was a different phase relation against M1-0 and M3-0 that the solid solution phase was further separated into two phases. A comparison on such separation is made in the Discussion, in comparison with the reported C₂S-C₃P pseudo-binary phase diagrams.

For sample M2-0, three phases with different contrasts were observed. One of them in the darkest color as almost free of P₂O₅ and was identified as CS. Another two phases in a relatively brighter color contained some P₂O₅ and had the composition of C₂S-C₃Pss and C₅PS, respectively. This was a different phase relation against M1-0 and M3-0, where the solid solution phase was further separated into two phases. A comparison of such a separation as made in the Discussion against the reported C₂S-C₃P pseudo-binary phase diagrams.

All the samples were separated into phases corresponding to CS and C₂S-C₃Pss. These results basically confirm that the compositional region focused on in this study was essentially saturated with CS as the almost P-free phase, and the solid solution as the P-concentrated phase. This finding would be advantageous for phosphorus recovery by phase separation.

3.2. CaO-SiO₂-P₂O₅–Al₂O₃/Fe₂O₃ System

Figure 6, Figure 7 and Figure 8 are SEM images of M3-A2, M3-A5, and M3-A10, respectively. All of the samples were partially melted in appearance. The results of the compositional analysis of these phases with SEM-EDX are listed in

Table 4. Compositional analysis of the observed phases in the quaternary samples containing Al₂O₃ (in mass%).

Sample No.	CaO	SiO2	P2O5	Al2O3	Note 1
M3-A2	48.6	51.2	0.2	0.0	CS
	56.0	11.7	32.3	0.0	C5PS
	43.8	37.3	4.7	14.2	Liq
M3-A5	49.0	50.6	0.3	0.0	CS
	56.1	11.8	32.0	0.0	C5PS
	42.3	38.4	5.0	14.3	Liq
M3-A10	55.0	11.6	31.6	1.8	C5PS
	40.4	36.8	6.0	16.8	Liq

¹ Estimated phase. Liq—liquid phase.

, and the plausible phases estimated therefrom are also given. All the employed data were the average of at least three measured points.

The composition of each phase was projected on an isothermal section of the [CaO-SiO₂-P₂O₅] ternary phase diagram at 1573 K in the figures. The bulk composition of the sample is also given in the figure with an X mark.

For M3-A2 and M3-A5, three phases with different contrast were observed, as shown in Figure 6 and Figure 7. Among them, the phases corresponding to CS and C₅PS still existed, despite the addition of Al₂O₃ to M3-0 as a ternary sample.

The third phase, in the darkest color, was characterized by a Al₂O₃ content with more than 10mass% and was considerably higher than that of bulk slag, and it was identified as liquid phase. The liquid compositions of M3-A2 and M3-A5 almost lay on the tie line between CS and C₅PS, and were close to CS compared to the bulk composition. Both liquid compositions were similar. As a comparison, the eutectic composition in CS-C₃P pseudo-binary (1693 K) [35] is shown in the figures. The projected liquid composition was apart from the eutectic composition, so the formation of the liquid phase in the samples could be critically influenced by Al₂O₃ addition. Further consideration on the liquid composition is given in the Discussion.

For sample M3-A10 containing 10mass% Al₂O₃, two different phases were observed, as shown in Figure 8. These phases correspond to the C₅PS and liquid phase, indicating that the CS phase was dissolved and formed a continuous matrix phase. It should be mentioned, at this point, that the phosphorus concentrated phase still existed under Al₂O₃ addition of up to 10mass%. The composition of the liquid phase was apart from the eutectic composition in the CS-C₃P pseudo-binary, and was rather close to the CS composition due to the dissolution of the CS phase in this sample.

Figure 9, Figure 10 and Figure 11 show the SEM images of M3-F2, M3-F5, and M3-F10, respectively. The composition of each phase obtained with SEM-EDX is given in the same manner as for the Al₂O₃-containing sample in

Table 5. Compositional analysis of the observed phases in the quaternary samples containing Fe₂O₃ (in mass%).

Sample No.	CaO	SiO2	P2O5	Fe2O3	Note 1
M3-F2	49.9	49.1	0.8	0.1	CS
	56.5	11.4	31.9	0.3	C5PS
	25.9	23.7	1.7	48.7	Liq
M3-F5	49.7	50.1	0.1	0.1	CS
	56.6	11.7	31.2	0.5	C5PS
	39.5	29.2	2.7	28.6	Liq
M3-F10	56.5	11.3	31.7	0.5	C5PS
	40.6	32.3	3.4	23.8	Liq

¹ Estimated phase. Liq—liquid phase.

and Figure 9, Figure 10 and Figure 11. For M3-F2 and M3-F5, three phases in different contrasts were observed. The phases corresponding to CS and C₅PS still existed with the addition of Fe₂O₃ to M3–0 as a ternary sample, as well as in the case of Al₂O₃ addition. The third whitish phase lay in the gap of the abovementioned two solid phases and was identified as the liquid phase. The liquidus composition of M3-F2 almost lay on the tie line between CS and C₅PS, and was close to CS compared to the bulk composition. The liquidus composition of M3-F5 was not on the tie line, but was still far from the eutectic composition in the system CS-C₃P.

For sample M3-F10 containing 10mass% Fe₂O₃, two different phases were observed, as shown in Figure 11. These phases correspond to the C₅PS and liquid phase, indicating that the CS phase was dissolved and formed a continuous matrix phase. Again, the phosphorus concentrated phase still existed under Fe₂O₃ addition of up to 10mass%. The composition of the liquid phase in M3-F10 was near that in M3-F5.

Throughout the investigation of the quaternary system, CS and C₅PS were observed in the samples containing 5mass% Al₂O₃ or Fe₂O₃, and the dissolution of the CS phase occurred with 10mass% addition of each. Such results reflect that the phase equilibria in the ternary system basically held in the quaternary system, at least under C₅PS saturation, within the composition range in this study.

In order to obtain a higher efficiency of phosphorus recovery, contamination of the gang component in solid C₅PS, which was Al₂O₃ or Fe₂O₃ in this system, should be avoided. From this point of view, the Al₂O₃ content in C₅PS was almost zero, except for M3-A10, and the Fe₂O₃ content was less than 1mass%. So, it is beneficial to note that the Al₂O₃ and Fe₂O₃ contents in the C₅PS phase were considerably low within the amount of addition of the fourth component in this study.

4. Discussion

4.1. Phase Relation in Ternary Samples

C₂S-C₃Pss is an important component in slag, cement, and fertilizer industries. With respect to this pseudo-binary system, the phase diagrams reported by Nurse et al. [20] and Fix et al. [21] have been referred to for a long time, and are cited in Figure 12. The stoichiometric compounds of C₇PS₂ and C₅PS appear as “A” and “S”, respectively, in these diagrams. These diagrams are slightly different, especially at the C₃P content around 40–60mass% and temperature range between 1373–1773 K.

In Figure 12, the compositions of the C₂S-C₃Pss phase that appeared in samples M1-0, M2-0, and M3-0 are plotted. In the figure, the end member of the tie line passing through the CS and the bulk composition of each sample is shown with a dashed line. For samples M1-0 and M3-0, which are two phase coexistence of CS and C₂S-C₃Pss, the compositions of the C₂S-C₃Pss phase were almost similar to the end member of the tie line. Such a phase relation agreed well with both diagrams in Figure 12.

There seemed to be another relation for sample M2-0. Two types of C₂S-C₃Pss phases appeared, as shown in the figure. One of the compositions of those phases corresponded to C₅PS. The other was 35mass%C₃P, which is far from both C₅PS and the end member of the tie line passing through CS and the bulk composition. According to the phase diagram of Fix et al., the composition of the aforementioned end member of sample M2-0 lay in the single phase region of “R” at 1573 K. On the other hand, the same composition lay in the two-phase region of “S” (C₅PS) and “$\overline{\alpha }$” (C₂S-C₃Pss) in the phase diagram of Nurse et al. It is important to note, at this point, that the observed phase relation for sample M2-0 agreed with the diagram after Nurse. It is a matter of course that the deposition of the compound with a higher P content, such as C₅PS, would be advantageous for phosphorus recovery through the C₂S-C₃P compound.

With further attention on the phase relation for sample M2-0, it should be noted that the observed composition of C₂S-C₃Pss was precisely different from the phase boundary of the “$\overline{\alpha }$” phase in the diagram after Nurse. In this study, an SEM analysis was carried out for the samples after 50 h of heating and quenching. Nurse and Fix drew the phase diagram based on high temperature in-situ XRD measurements, however, the samples were maintainied at an elevated temperature for up to 20 h. To confirm the phase relation of this system in more detail, it was necessary to prepare the specimen under a sufficient heating time and temperature. Related work on nearby composition is in progress.

4.2. Phase Relation in Quaternary Samples (Liquid Composition)

As shown in Table 4, for quaternary samples M3-A2, M3-A5, and M3-A10, the major components of the liquid phases were CaO, SiO₂, and Al₂O₃, against a minor portion of P₂O₅. In Figure 13, the liquid compositions were projected on the isothermal section of the CaO-SiO₂-Al₂O₃ ternary phase diagram at 1573 K [36] (Plate 1). For M3-A2 and M3-A5, the liquid compositions were located nearly on the liquid line saturated with CS. For M3-A10, in which the CS phase dissolved into the liquid phase, the composition lay inside the homogeneous liquidus region. As most of the Al₂O₃ in the bulk sample went into the liquid phase, the solid CS phase would be dissolved to form liquid slag coexisting with CS, as in CaO-SiO₂-Al₂O₃ ternary phase diagram, through the addition of Al₂O₃. Such liquid compositions are reasonable in terms of the phase relationship. Furthermore, the CS phase would be relatively unstable due to the low melting point at 1813 K [35] compared to that of C₅PS at about 2273 K [21].

As shown in Table 5, for quaternary samples M3-F2, M3-F5, and M3-F10, the major components of the liquid phase were CaO, SiO₂, and Fe₂O₃. In Figure 14, the liquid compositions were projected on the isothermal section of the CaO-SiO₂-Fe₂O₃ ternary phase diagram at 1573 K [36] (Plate 10). Although the compositions were slightly scattered, they were located near the liquid region with the middle range of Fe₂O₃ content, in which the liquid boundary partly faced CS. So, one can consider that the solid CS phase would be dissolved to form liquid slag coexisting with CS, as in CaO-SiO₂-Fe₂O₃ ternary phase diagram, through the addition of Fe₂O₃, as in the case with Al₂O₃ addition. Such consideration would be reasonable, again based on the compositional change.

4.3. P₂O₅ Content in Liquid Phase

With respect to effective recovery of the phosphorus through the phase separation of P-concentrated solid phase against the liquid phase in this system, the phosphorus content in the liquid phase should be low. The phosphorus contamination in the liquid phase was evaluated in terms of the chemical character of the slag. In Figure 15a, the content of P₂O₅ in the liquid phase is shown against the Al₂O₃ and Fe₂O₃ content in the quaternary bulk sample. The P₂O₅ content was lower in the Fe₂O₃-containing sample compared to the Al₂O₃-containing sample. In addition, the P₂O₅ content slightly increased with the Al₂O₃ and Fe₂O₃ content.

For a more detailed evaluation, optical basicity was introduced for an analysis of the present system. The theoretical optical basicity Λ_th is proposed as Equation (1) [37,38];

$${\Lambda }_{th}=\frac{{\displaystyle \sum }{n}_{\mathrm{i}}·X_{\mathrm{i}}·{\Lambda }_{\mathrm{i}}}{{\displaystyle \sum }{n}_{\mathrm{i}}·X_{\mathrm{i}}}$$

(1)

where n_i is the number of oxygen ions in oxide i, X_i is the molar fraction of oxide i, and Λ_i is the optical basicity of individual oxide i.

In Figure 15b, the P₂O₅ contents in the liquid phase are plotted against Λ_th for the liquid phases. Although Λ_th for the Fe₂O₃-containing samples are higher compared to the Al₂O₃-containing samples, the P₂O₅ contents are lower for the Fe₂O₃-containing samples. In the thermochemical point of view, P₂O₅ might tend to go into the matrix liquid phase with a higher Λ_th value, as P₂O₅ is an acidic oxide. Therefore, the lower P₂O₅ content in the Fe₂O₃-containing liquidus is difficult to explain in terms of Λ_th.

In Figure 16, the area fraction of each phase that appeared in the SEM image is shown for Al₂O₃-containing samples and Fe₂O₃-containing samples, respectively. The value of the area fraction of each phase is an average of the plural image analysis. With the addition of the fourth components, the area fraction of the liquid phase increased, and that of C₅PS decreased. For M3-A5 and M3-A10, the area fraction of CS drastically decreased and even disappeared. The area fraction of C₅PS in M3-A10 was slightly higher than that of M3-A5, due to the incorporation of Al₂O₃ in C₅PS at a higher Al₂O₃ content in the bulk composition, as shown in Table 4, over some variations of the analyzed data. With respect to the Fe₂O₃-containing samples, the liquid phase also increased with the addition of Fe₂O₃. Comparing the results of the same content of Al₂O₃ or Fe₂O₃, the area fraction of the liquid phase was relatively larger for Al₂O₃ addition.

If C₅PS as a P-concentrated phase dissolved into a liquid phase, the P₂O₅ content would become higher. It is suggested that a higher fluxing effect of Al₂O₃ in the present system would affect the dissolution of C₅PS and result in a higher P₂O₅ content in the liquid phase, rather than the aforementioned liquid property. In a multi-component slag system, such thermochemical properties should be understood based on the basic contribution of each component in any way.

With respect to phosphorus recovery from such a multicomponent and multiphase system, it is necessary to make a pair of phases with a high and low phosphorus content. Considering this, the sufficient addition of Al₂O₃ or Fe₂O₃ as in the present system may be effective, so that phosphorus is concentrated into solid C₅PS as a phosphorus-rich phase against the surrounding liquid phase with a low phosphorus content. Owing to such a simple structure, enlargement of phosphorus-rich phase would be advantageous to separate and recover them, based on controlling the crystal growth. Further research is in progress.

5. Conclusions

In view of obtaining fundamental information on phosphorus recovery from steelmaking slag and sewage sludge, phase relation in [CaO-SiO₂-P₂O₅]-based slag with low basicity at 1573 K was investigated using SEM-EDX, with particular interest in the triangular compositional region within the apices CS, C₃P, and C₂S-C₃Pss. [CaO-SiO₂-P₂O₅] ternary specimens based on such composition were essentially separated into CS and C₂S-C₃Pss as a phosphorus concentrated phase. The composition of the observed C₂S-C₃Pss phases was discussed in comparison with the related pseudo-binary phase diagram. With the addition of Al₂O₃ and Fe₂O₃ to the relevant specimens, although the liquidus phase appeared as the third phase, C₂S-C₃Pss were still observed. This finding shows that phase equilibria based on the ternary system would be stable, and would be beneficial for phosphorus recovery through the phase separation.