Influence of SiO 2 modification on phosphorus enrichment in P bearing steelmaking slag

Abstract

In order to effect enrichment of phosphorus in the converter slag, phosphorus enrichment was researched by slag modification with SiO₂. The SiO₂ modification process was evaluated thermodynamically, and phosphorus rich phase was separated from experimental slag by magnetic separation. The results show that n2CaO.SiO₂–3CaO.P₂O₅ (nC₂S–C₃P) solid solution is generated by the reaction of Ca₃(PO₄)₂ phase in the slag with precipitated Ca₂SiO₄ phase at 1623 K, and Ca₂SiO₄ in the solid solution is reduced with increasing SiO₂; hence, the phosphorus content in the solid solution is increased. If the addition of SiO₂ is excessive, the amount of Ca₂SiO₄ precipitation in the slag is decreased remarkably (and even disappeared), CaSiO₃ phase is generated and the generation of nC₂S–C₃P solid solution is reduced, which is not favourable to phosphorus enrichment. Through magnetic separation, the non-magnetic fraction was increased by 10–74·68% compared with the original slag. The distribution ratio of P₂O₅ was increased from 2·71 to 5·48. As 84·57% of the phosphorus in the slag is in the collected non-magnetic substances, the work demonstrates that most of the phosphorus is able to be effectively recycled.

Introduction

Increased dephosphorisation in steelmaking processes and the use of high phosphorus content iron ore in recent years have resulted in higher phosphorus content in the slag, which, if recycled to iron- and steelmaking processes, leads to phosphorus enrichment in hot metal; hence, the reuse of the slag is limited. Alternatively, if it is used directly as a raw material for phosphoric fertiliser, it is too low in P for effective use.1 Therefore, if P₂O₅ in the slag can be enriched and separated, the separated P₂O₅ phase may be used as phosphoric fertiliser additives, and the remainder may be recycled to iron- and steelmaking processes.

The form of phosphorus in steel slag has been researched2 ^– 4 and found to be nC₂S–C₃P solid solution. However, because the amount of precipitation of 2CaO.SiO₂ in high basicity slag is large, the phosphorus content is too low in the solid solution, and so, it is necessary to somehow enrich the phosphorus content of the slag.5 ^– 11 Some researchers12 ^– 14 used a melting–cooling method to study vanadium enrichment in V bearing steelmaking slag; however, the research on phosphorus enrichment behaviour in steel slag is limited.

This paper describes the work on the influence of melting modification on phosphorus enrichment in steel slag and the generation of a thermodynamic base of nC₂S–C₃P solid solution during modification, and phosphorus rich phase in the slag was separated by magnetic separation; resource utilisation of phosphorus steel slag provided the necessary basis of study.

Experimental

Reagent grade SiO₂ and P₂O₅ were added to the original slag to change the slag basicity, and the P₂O₅ contents were all adjusted to 10% (see Table 1).

Table 1.

Chemical compositions of original and modified slags/mass-%

Slag sample	CaO	SiO₂	Fe₂O₃	P₂O₅	Al₂O₃	MgO	TiO₂	CaF₂	MnO	Basicity CaO/SiO₂
Original slag	46·72	11·68	19·47	10	1·07	5·46	0·97	0·37	4·26	4
No. 1	44·78	14·93	18·66	10	1·03	5·23	0·93	0·35	4·08	3
No. 2	41·35	20·68	17·23	10	0·95	4·83	0·86	0·33	3·77	2
No. 3	33·63	33·63	14·01	10	0·77	3·93	0·70	0·27	3·07	1

The mixed slag (200 g) was placed in an alumina crucible 60 mm diameter×100 mm, placed inside a graphite crucible and heated in an MOSi₂ electric resistance furnace up to 1723 K; the temperature was maintained for 30 min to ensure the full melting of slag. The slag was then cooled to 1623 K at a cooling rate of 3 K min⁻¹ and this temperature was kept for 1 h in order to fully promote the precipitation of 3CaO.P₂O₅; it was then cooled to 1423 K at a cooling rate of 3 K min⁻¹. The furnace was then closed, and the slag sample was cooled within it (see Fig 1). Samples were taken from incubation at 1723 K for 15 min, incubation at 1723 K for 30 min, before incubation at 1623 K, incubation at 1623 K for 30 min, incubation at 1623 K for 60 min, and every 20 min of cooling from 1623 to 1423 K and quenched.

Figure 1.

Experimental conditions for precipitation of C₂S–C₃P solid solution

The phase morphology of the prepared slag samples was observed by an S-360 scanning electron microscope, and each phase composition was analysed by a Tracor Northern spectrometer. After each experiment, the slag was ground [<300 mesh (48 μm)]; mineralogical phases were determined by X-ray diffraction analysis. Diffraction patterns were measured in a 2θ range of 10–90° using Cu K_α radiation of 40 kV and 30 mA, and the scan’s speed was 5° min⁻¹.

Magnetic separation experiments of the P bearing slag were performed with a DTCXG-ZN50 magnetic separator tube 50 mm in diameter. The swing frequency was 70 times per minute, the travelling stroke was 40 mm and the air gap between the magnetic poles was 52 mm. A 10 g slag sample was put into the solution mixed with alcohol and water. The magnetic field intensity of the separator tube was adjusted to a predetermined intensity of 350 mT. Sample solution was poured into the magnetic separator tube and magnetic substances attached to the tube wall, and non-magnetic substances flowed into the prepared container outside the tube with the aid of mechanical vibration. The magnetic and non-magnetic substances were dried at 105°C for 12 h, weighed and analysed.

Results and discussion

Influence of SiO₂ modification on phosphorus enrichment behaviour

Figure 2 is an SEM back scattered electron image corresponding to the Table 1 sample with furnace cooling. Table 2 shows the energy dispersive spectroscopy results corresponding to Fig. 2 sample compositions. Figure 3 are the corresponding slag X-ray diffraction results.

Figure 2.

Scanning electron microscopy back scattered electron image of experimental samples: 1 – P rich phase; 2 – RO phase; 3 – base phase

Figure 3.

X-ray diffraction patterns of four steel slags

Table 2.

Chemical compositions of each phase from sample in Fig. 2 by energy dispersive spectroscopy

Number	Main phase	CaF₂	Al₂O₃	SiO₂	P₂O₅	CaO	Fe₂O₃	MgO	MnO	TiO₂
Original slag	P rich phase	0·5	2·62	15·21	19·89	59·71	1·14	0·18	0·43	0·33
	RO phase	1·79	33·22	0·216	…	1·818	43·08	10·5	9·09	0·29
	Base phase	0·6	19·25	16·94	4·573	50·31	6·967	0·45	0·81	0·1
No. 1	P rich phase	0·43	0·76	13·08	31·28	53·39	0·48	0·15	0·07	0·37
	RO phase	…	0·95	0·18	0·23	2·01	69·84	10·61	16·12	0·05
	Base phase	0·05	32·89	21·61	0·31	42·17	2·39	…	0·39	0·19
No. 2	P rich phase	1·52	1·43	5·80	32·5	54·87	2·60	0·42	0·69	0·16
	RO phase	…	17·91	0·70	…	1·61	58·30	4·36	8·20	8·92
	Base phase	…	29·89	17·62	0·52	47·82	3·21	0·55	0·39	…
No. 3	P rich phase	…	…	2·56	34·39	63·06	…	…	…	…
	RO phase	…	23·22	18·82	3·04	18·11	27·67	9·14	…	…
	Base phase	…	19·38	25·57	6·72	28·05	12·39	3·12	4·77	…

From Fig. 1 and Table 2, the phosphorus rich phase in the original slag is mainly in the nC₃S–C₃P solid solution, including a small amount of fluorine apartite phase [Ca₅(PO₄)₂F]. The base phase is the liquid phase, mainly composed of calcium aluminium melilite (Ca₂Al₂SiO₇), some iron and phosphorus. The white RO phase is mainly iron oxides, Fe–Mn oxides or FeAl₂O₄. The iron in the slag mainly exists in this phase.

When the basicity of no. 1 slag was lowered to 3·0, the C₃S phase and the phosphorus rich phase were mainly nC₂S–C₃P solid solution, including a small amount of fluorine apatite phase [Ca₅(PO₄)₂F]. The P₂O₅ content in the phosphorus rich phase is obviously increased. The liquid phase is mainly composed of calcium aluminium melilite (Ca₂Al₂SiO₇), and the white RO phase is mainly iron oxides, manganese oxides or FeAl₂O₄.

Compared with no. 1 slag, the main phase type in no. 2 slag is unchanged, but the SiO₂ content in the phosphorus rich phase is reduced, the P₂O₅ content is slightly increased and the particle size of phosphorus rich phase is increased.

The SiO₂ content in no. 3 slag is further increased, Ca₂SiO₄ content in phosphorus rich phase is decreased and a new phase (CaSiO₃) is generated. Meanwhile, calcium aluminium melilite (Ca₂Al₂SiO₇) has disappeared in liquid phase, and a new phase, anorthite (CaAl₂Si₂O₈), is generated. In addition to iron oxides in the RO phase, diopside [CaMg(SiO₃)₂] phase is also generated.

From the above analysis results, when the basicity of original slag (R = 4) is high, a large amount of 3CaO.SiO₂ is precipitated during cooling, as the solid solution that generated by 3CaO.SiO₂ phase and 3CaO.P₂O₅ phase cannot infinitely be dissolved in each other. The viscosity of the melted slag is high, which is not beneficial to phosphorus transfer. The P₂O₅ content in the phosphorus rich phase is 19–25% and is 4–5% in the base phase; the distribution of phosphorus rich phase is more dispersed, and particle size is small, usually 10–30 μm.

In slag no. 1 with a basicity of 3, the 3CaO.SiO₂ phase disappeared; the main precipitation is 2CaO.SiO₂, and the solid solution that is generated by 2CaO.SiO₂ and 3CaO.P₂O₅ can infinitely be dissolved in each other. Thus, the phosphorus rich phase content in the modified slag is higher than that in the original slag. The P₂O₅ content in the phosphorus rich phase is 31–32%, the phosphorus rich phase is claviform and the particle size is increased significantly to 20–100 μm.

The phosphorus rich phase content in no. 2 slag is 32–33%, slightly higher than that in the no. 1 slag, and the particle size is further increased to 30–150 μm. This is because 2CaO.SiO₂ precipitation is reduced in the low basicity slag, the phosphorus is enriched into the small precipitated 2CaO.SiO₂ phase, which improves the P₂O₅ content; the SiO₂ content of the phosphorus rich phase in the modified slag, which is lower than that in the original slag, also shows that. As the basicity of no. 1 slag is high, precipitation of 2CaO.SiO₂ phase needs a lower temperature; hence, the viscosity of the melted slag is increased, which is not favourable to phosphorus diffusion in slag and phosphorus rich phase growth. This is the reason that the particle size of the phosphorus rich phase in no. 1 slag is smaller than that in no. 2.

The phosphorus rich phase in no. 3 slag is claviform, but the size is finer than the phosphorus rich phase in no. 2 slag, and the P₂O₅ content in phosphorus rich phase is 34–37%. However, the basicity is too small, so slag composition deviated from the elementary 2CaO.SiO2 phase, the amount of precipitation of 2CaO.SiO₂ is reduced and precipitation of 2CaO.SiO₂ phase needs a lower temperature, resulting in higher viscosity slag, which is not favourable to phosphorus diffusion. In addition, the CaSiO₃ that is generated in the slag is not beneficial to phosphorus enrichment and growth of the phosphorus rich phase.

Therefore, when slag basicity is too low, the P₂O₅ content is higher in the other phase, and P₂O₅ content in phosphorus rich phase is also increased, but the particle size of phosphorus rich phase is smaller and the amount of precipitation is limited so it is difficult to achieve effective phosphorus enrichment in the slag. Thus, a proper amount of SiO2 modifying agent is added to the slag, which is advantageous to phosphorus enrichment of high concentration in the nC₂S–C₃P solid solution. The P₂O₅ content of phosphorus rich phase in the modified slag is >31%, and such a high phosphorus content can meet the requirement of P₂O₅ content in the phosphoric fertiliser industry. Phosphorus rich phase in slag can be separated to meet the requirement of phosphoric fertiliser production.

Thermodynamic analysis of SiO₂ modification

The generation reaction of nC₂S–C₃P solid solution in P bearing slag is as follows (1) (2) (3) (4) where , a _CaO, , and are the activity of SiO₂, CaO, Ca₂SiO₄, Ca₃(PO₄)₂ and nCa₂SiO₄.Ca₃(PO₄)₂ respectively; ΔG ₁ and ΔG ₂ show the Gibbs free energy of reactions (1) and (3) respectively; nCa₂SiO₄.Ca₃(PO₄)₂, Ca₃(PO₄)₂ can be regarded as pure substances, and their activity is 1.

Figures 4 and 5 show phase diagrams of the Ca₂SiO₄–Ca₃(PO₄)₂ and CaO–SiO₂ systems calculated by FactSage 6·2. The Fact and FToxid database was used in the calculation process. When slag basicity is higher than 3, Ca, Si and O are mainly and first precipitated as Ca₃SiO₅ phase in the slag, which can be seen from the phase diagram. When the modified slag basicity is between 2 and 3, Ca, Si and O are mainly and first precipitated as Ca₂SiO₄ phase in the slag. When the slag basicity is <1·5, the Ca₂SiO₄ phase disappears, and Ca, Si and O are mainly precipitated as CaSiO₃ and Ca₃Si₂O₇.

Figure 4.

Phase diagram of Ca₂SiO₄–Ca₃(PO₄)₂ system

Figure 5.

Phase diagram of CaO–SiO₂ system

During slag cooling, Ca, Si and O are first precipitated as nCaO.SiO, Ca₂SiO₄ and Ca₃(PO₄)₂ phase can be dissolved with each other in any proportion. The enrichment effect depends on the generation of nC₂S–C₃P solid solution. From a thermodynamic point of view, when the temperature is reduced to ∼1623 K, the generation trend and amount of nC₂S–C₃P solid solution are decided by oxide activities in reactions (1) and (3). The effect of SiO₂ modified agent on generation of nC₂S–C₃P solid solution is mainly the effect on .

Figure 6 is the curve of the logarithm oxide activity change with basicity (CaO/SiO₂) on 1623 K as calculated by FactSage6·2. Figure 6 shows that the value of / is unchanged and the thermodynamic trend of 2CaO.SiO₂ phase generation has no influence with adding SiO₂ into the melted slag.

Figure 6.

Curve of logarithm of part oxides activity versus basicity (CaO/SiO₂) at 1623 K

With decreasing slag basicity, increases, and a _CaO decreases. is unchanged when basicity is >2·5, and is decreased slightly when basicity is <2·5, but decreases significantly when basicity is <1·5. This indicates that basicity >2·5 has little effect on the generation of elementary 2CaO.SiO₂. The amount of 2CaO.SiO₂ generated decreased significantly when basicity is <1·5, and a large amount of CaO.SiO₂ phase and 3CaO.2SiO₂ phase is generated.

Slag basicity is decreased with the addition of SiO₂; the value of is unchanged when R≧2·5. The value of is changed a little when 1·5≤R≤2·5, which has little effect on ΔG and thermodynamic trend of nC₂S–C₃P generation. The value of decreased when basicity is from 1·5 to 1·0, and the decrease is ∼66%, which has an evident effect on ΔG, and also makes the thermodynamic trend of nC₂S–C₃P generation weak.

According to the description above, when slag basicity is >3, slag contains 3CaO.SiO₂, which is not advantageous to P₂O₅ enrichment, and 3CaO.SiO₂ in modified slag disappears with the addition of SiO₂. 2CaO.SiO₂ is first precipitated during cooling, and the increasing SiO₂ has no influence on the thermodynamic trend of 2CaO.SiO₂ precipitation. However, when the SiO₂ content is increased, causing a slag basicity of 2·5, the amount of precipitation of 2CaO.SiO₂ in slag is reduced.

When the temperature is reduced to 1623 K, nC₂S–C₃P solid solution is formed by Ca₃(PO₄)₂ in the melted slag and precipitation of 2CaO.SiO₂. The increase in SiO₂ content (1·5<R<3·0) has little effect on the generation trend of nC₂S–C₃P solid solution, as the precipitation of 2CaO.SiO₂ phase is decreased slightly and P₂O₅ content in solid solution is increased relatively, which is favourable to phosphorus enrichment in solid solution.

When the addition of SiO₂ is very high (R<1·5), the amount of precipitation of 2CaO.SiO₂ is reduced and even disappears, and low calcium phase such as CaSiO₃ is generated. Although the phosphorus content in solid solution is increased, the thermodynamic condition of nC₂S–C₃P solid solution generation is difficult, and precipitation of CaSiO₃ cannot dissolve Ca₃(PO₄)₂ well. Thus, phosphorus transfer from base phase to phosphorus rich phase is not complete with high phosphorus content in the base phase. This is adverse for the generation of nC₂S–C₃P solid solution and phosphorus enrichment; hence, the suitable modification basicity of slag should be in the range of 1·5–2·0.

Magnetic separation of P bearing converter slag

The magnetic separation of the original slag and no. 2 slag was carried out. Figure 7 shows the proportion of magnetic substances and non-magnetic substances in both slags. The results show that non-magnetic proportion in no. 2 slag increased by 10·09% compared with original slag. Figure 8 shows the mass fraction of P₂O₅ in the magnetic and non-magnetic fractions in the two slags. The following expressions describe the phosphorus distribution in magnetic substances and non-magnetic substances (5) (6) where M ₁ and M ₂ are non-magnetic and magnetic substances mass (g) respectively, (%P₂O₅)₁ and (%P₂O₅)₂ are P₂O₅ mass fractions of non-magnetic and magnetic substances (%) respectively, L _P is the phosphorus distribution ratio of non-magnetic and magnetic substances and H _p is the P₂O₅ percentage from slag enters into non-magnetic or magnetic substances.

Figure 7.

Proportion of magnetic substance and non-magnetics in original slag and no. 2 slag

Figure 8.

Mass fraction of P₂O₅ enters into magnetic substance and non-magnetics in original slag and no. 2 slag

According to Fig. 8, 73·05% of P₂O₅ in the original slag enters into the non-magnetic substances, and 84·57% of P₂O₅ in modified slag enters into the non-magnetic substances. After slag modification, the phosphorus distribution ratio of non-magnetic and magnetic substances increases from 2·71 to 5·48. This is because the particle size of the phosphorus rich phase is larger and the P₂O₅ content is increased; the separation effect of phosphorus rich phase and other phase (especially RO phase) is better than original slag when the particle is ground to <300 mesh.

The composition of non-magnetic and magnetic substances for no. 2 modified slag is shown in Table 3. The phosphorus recovery rate and iron recovery rate in non-magnetic substances are 84·57 and 38·68% respectively, the phosphorus content is 13·90% and 746·8 kg of non-magnetic substances can be separated from 1 t of modified no. 2 slag and used for the phosphoric fertiliser industry, so most of the elemental phosphorus can be recycled. The phosphorus and iron recovery rates in magnetic substances are 14·43 and 61·32% respectively. T.Fe and the CaO contents in magnetic substances are 24·22 and 17·74% respectively. Therefore, magnetic substances can be recycled into iron- and steelmaking processes, such as sintering, hot metal desiliconisation and hot metal dephosphorisation. The process route of P bearing steelmaking slag utilisation is shown in Fig. 9.

Figure 9.

Schematic process route diagram of P bearing steelmaking slag utilisation

Table 3.

Composition of magnetic and non-magnetics for no. 2 modified slag after magnetic separation

	Mass ratio against initial slag/%	P recovery against initial slag/%	Fe recovery against initial slag/%	Composition/mass-%
	Mass ratio against initial slag/%	P recovery against initial slag/%	Fe recovery against initial slag/%	T.Fe	P₂O₅	CaO	SiO₂	MgO	MnO
Non-magnetics	74·68	84·57	38·68	5·18	13·90	38·79	23·21	2·79	2·28
Magnetic substances	25·32	15·43	61·32	24·22	7·48	17·74	9·74	10·02	7·53

Conclusions

The influence of SiO₂ modification on phosphorus enrichment behaviour in steel slag was studied. The key findings are as follows.

When the slag basicity decreases to 3 with the addition of SiO₂, 3CaO·SiO₂ phase disappears, nC₃S–C₃P solid solution disappears, the phosphorus rich phase is mainly nC₂S–C₃P solid solution and the P₂O₅ content in solid solution increases, which is beneficial to phosphorus enrichment in solid solution. The increase (1·5<R<3·0) in SiO₂ has little effect on the generation trend of nC₂S–C₃P solid solution as the precipitation of 2CaO.SiO₂ phase decreases slightly, which makes the P₂O₅ content in solid solution increase relatively. When the addition of SiO₂ is high (R<1·5), the precipitation amount of 2CaO.SiO₂ phase decreases significantly and even disappears, and CaO.SiO₂ phase is generated, so the generation thermodynamic condition of nC₂S–C₃P solid solution is difficult, which is detrimental to phosphorus enrichment in slag. Therefore, the suitable modification basicity of slag should be 1·5–2·0.

The slag contains phosphorus rich phase, base phase and RO phase. The phosphorus in slag is mainly in the form of nC₂S–C₃P solid solution in the phosphorus rich phase. The phosphorus content in the phosphorus rich phase increases after suitable SiO₂ modification, the P₂O₅ content is >31% and the particle size of phosphorus rich phase increases to 30–60 μm, which can fully meet the requirement of phosphorus content of phosphatic fertiliser.

After SiO₂ melting modification and magnetic separation of P bearing converter slag, the phosphorus distribution ratio of non-magnetic and magnetic substances increased from 2·71 to 5·48, 84·57% of P₂O₅ in slag enters into recycling of non-magnetic substances and the extracted non-magnetic substances proportion is 74·68%. Therefore, recycling of most phosphorus element can be achieved.

It is feasible to make the phosphorus enriched effectively in slag by SiO₂ melting modification and separate phosphorus rich phase from modified slag by magnetic separation.

Footnotes

Acknowledgements

The authors would like to thank the Australian Hamersley Iron Ore Company for the financial support. The great help from Rio Trading (Shanghai) Company Limited is really appreciated.

References

Boom

Riaz

Mills

: Ironmaking Steelmaking, 2005, 32, 21.

Suito

Inoue

: ISIJ Int., 2006, 46, 180.

Kitamura

: Proceedings of the VIII International Conference on Molten Slags, Fluxes&Salts, proceedings of Molten 2009, Santiago, Chile, 2009, 858.

Wang

Chen

: Ironmaking Steelmaking, 2011, 38, 185.

Yang

Matsuura

Tsukihashi

: ISIJ Int., 2009, 49, 1298.

Kitamura

Saito

Utagawa

Shibata

Robertson

DGC

: ISIJ Int., 2009, 49, 1838.

Son

Kashiwaya

: ISIJ Int., 2008, 48, 1165.

Shimauchi

Kitamura

Shibata

: ISIJ Int., 2009, 49, 505.

Hironari

Kazuyo

Tetsuya

: ISIJ Int., 2010, 50, 59.

10.

Kazuyo

Hironari

Tetsuya

: ISIJ Int., 2010, 50, 65.

11.

Deo

Halder

Snoeijer

Overbosch

Boom

: Ironmaking Steelmaking, 2005, 32, 54.

12.

Dong

: Ironmaking Steelmaking, 2007, 34, 131.

13.

Dong

: Ironmaking Steelmaking, 2008, 35, 367.

14.

Dong

: Acta Metall. Sin., 2008, 44, 603.

Influence of SiO 2 modification on phosphorus enrichment in P bearing steelmaking slag

Abstract

Introduction

Experimental

Chemical compositions of original and modified slags/mass-%

Results and discussion

Influence of SiO2 modification on phosphorus enrichment behaviour

Chemical compositions of each phase from sample in Fig. 2 by energy dispersive spectroscopy

Thermodynamic analysis of SiO2 modification

Magnetic separation of P bearing converter slag

Composition of magnetic and non-magnetics for no. 2 modified slag after magnetic separation

Conclusions

Footnotes

Acknowledgements

References

Influence of SiO₂ modification on phosphorus enrichment behaviour

Thermodynamic analysis of SiO₂ modification