Effect of Cooling Method on the Mineralogy and Stability of Steel Slag

Abstract

Steel slag, as a potentially active gelling material, has not been widely used in the field of building materials. The effective utilization of steel slag depends on its stability, which is related to the cooling process with phases changing. The relation between cooling method and the phases and stability of steel slag were carried out in the present study. The slags were first melted into liquid state at 1873 K and were then cooled using four cooling methods, namely, as-furnace cooling, air cooling, mist cooling and water quenching. The cooled slags were characterized by XRD, SEM, and chemical analysis. The results show that the slags under the four cooling methods mainly contained Ca₂SiO₄ (C₂S), Ca₂SiO₄–Ca₃P₂O₈ solid solution (C₂SP), Ca₃SiO₅ (C₃S), monoxide solid solution (RO), Ca₂Fe₂O₅ (C₂F), and f-CaO. The content of Ca₃SiO₅ in steel slag increased with the increase of cooling rate. Rapid cooling could reduce the content of RO phase in steel slag, and the content of RO phase for the furnace cooling, air cooling, mist cooling, and water quenching samples were 28.03%, 22.53%, 14.17%, and 13.30%, respectively. In addition, rapid cooling could effectively reduce the content of f-CaO in steel slag and improve the stability of steel slag.

1. Introduction

According to the report of National Bureau of Statistics of China (2020), China’s crude steel output is 1.06 billion tons, which means that the production of steel slag exceeded 100 million tons.¹⁾ However, the utilization rate of steel slag in China is as low as about 25%, and enormous steel slag was discarded and piled up into a slag hill.²⁾ Therefore, improving the comprehensive utilization of steel slag has become a major challenge for the steel industry to achieve the green, stable, and sustainable development.³⁾

As a by-product of steelmaking, steel slag is a mixture of different mineral phases: dicalcium silicate (C₂S), tricalcium silicate (C₃S), monoxide solid solution (RO), calcium ferrite (CF or C₂F), and free lime (f-CaO).⁴⁾ Because the chemical composition of steel slag is consistent with cement and concrete, and as well with the good strength and durability, steel slag can be used in road and building construction to effectively solve the problem of solid waste utilization.⁵⁾ However, in the field of construction materials, the difficulty in the use of steel slag is the low stability of the steel slag. The f-CaO in steel slag could undergo hydration reaction, resulting in volume expansion and cracking of slag products, which can greatly reduce the strength of construction materials.^6,7,8,9,10)

Slag cooling process is currently the main way to promote the utilization of slag resources. Generally, the molten slag is cooled by air quenching, water quenching, and other methods to obtain slag particles with uniform particle size, qualified stability, and good performance.^11,12,13,14) Many researchers have paid attention to the effect of different cooling methods on slag properties and crystallization behavior.^{15,16,17,18,19,20,21,22)} For blast furnace slag, the crystallization behavior are determined using time-temperature-transformation (TTT) and continuous cooling transformation (CCT) diagrams by the hot thermocouple technique. The relationship between the cooling method and the properties of blast furnace slag has been established successfully.²³⁾ High cooling rate can be used to transform liquid slag into a glassy state that can be used as raw material for cement manufacturing due to its high cementitious. Lin et al.²⁴⁾ found that different cooling methods can affect the crystal incubation time and the crystal onset temperature of blast furnace slag. For steel slag, due to the high basicity, high melting temperature zone and high viscosity, even under the condition of rapid cooling, steel slag can crystallize completely without the tendency to form glass phase.²⁵⁾ However, there are a few study on the cooling methods of steel slag. Choi et al.²⁶⁾ compared the crystallization of steel slags at different cooling rates and found that the cooling rate does not affect the crystallization procedure of the melted BOF slag, but affects the size and shape of crystals by changing the holding time in the range of some desired temperature for crystallization and driving force of the crystallization. In addition, Tossavainen et al.²⁵⁾ studied the influence of semi-rapid and rapid cooling treatment on the leaching of minor elements in steel slag crystallization.

Most of the studies on the cooling process of steel slag have focused on the results by the cooling process, the mechanism of the effect of cooling rate on mineralogy of steel slag has less studied. This has led to an unclear relationship between the cooling method and the volumetric stability of the steel slag. Therefore, the aim of this study is to clarify the influence of different cooling methods on the mineralogy and stability of steel slag.

2. Experimental

2.1. Raw Materials

The steel slag can be divided into three categories according to basicity R = %CaO/(%SiO₂ + %P₂O₅), low basicity steel slag with basicity less than 1.8, medium basicity steel slag with basicity between 1.8 and 2.5, and high basicity steel slag with basicity greater than 2.5.²⁷⁾ The steel slag produced by Chinese steel factories is high basicity steel slag (Table 1), with a main chemical composition of CaO (43–48 mass%), SiO₂ (11.2–14.2 mass%), FeO (21.6–22.8 mass%), MgO (6.2–12.6 mass%), MnO (1.0–2.6 mass%), Al₂O₃ (0.8–2.9 mass%), and P₂O₅ (0.5–2.5 mass%). Depending on the difference in steel grade and production process, the steel slag also contains a certain amount of titanium and vanadium. According to the chemical composition of steel slag from different steel factories, synthetic 47 mass%CaO-23 mass%FeO-12 mass%SiO₂-11 mass%MgO-3 mass%MnO-2 mass%Al₂O₃-2 mass%P₂O₅ slags with basicity of 3.36, were prepared by using analytical pure reagents.

Table 1. Chemical composition of steel slag from different steel factories, wt%.

	CaO	SiO2	MgO	Al2O3	FeO	MnO	P2O5	TiO2	Basicity
Taisteel	47.59	11.23	6.29		22.50		0.524		4.05
Baosteel	45.45	11.38	10.71	1.50	21.60	2.59	2.50		3.27
Ansteel	43.31	13.23	12.6	2.9	22.72	1.13	0.81	0.41	3.08
Nansteel	43.55	14.22	7.32	0.83	21.83	2.52	2.32	1.05	2.63

2.2. Theoretical Calculation

The mineralogy of the slag with cooling was evaluated by comparing the measured observed phases with the expected mineralogy according to FactSage software (8.0 version).²⁸⁾ Considering the rapid cooling rate and the imbalance between slag phase and gas phase in the actual steel slag cooling process, gas phase has little effect on the precipitation of slag phase during cooling process. In the thermodynamic calculation, the influence of gas is ignored to simplify the calculation, and the FToxid database was selected. Two cases were considered: Equilibrium, and Scheil-Gulliver solidification. For the equilibrium solidification, it is assumed that diffusion is infinitely fast and that the entire system is in global equilibrium. These conditions are typical for very slow cooling process. In Scheil-Gulliver solidification, there is no diffusion or transformation in the solid phases. This means that all solid phases remain as solidified. This behavior is expected for very fast cooling. Real solidification problems are typically situated in between these two extreme cases.

2.3. Experimental Procedures

The synthetic slag with compositions comparable to the original steel slag which can be write as 47 mass%CaO-23 mass%FeO-12 mass%SiO₂-11 mass%MgO-3 mass%MnO-2 mass%Al₂O₃-2 mass%P₂O₅, was prepared by mixing the analytic grade oxides. All reagents were calcined in a muffle furnace to remove impurities like moisture, hydroxides, and carbonates before weighing and mixing. Then, the samples were divided into four groups and pressed into columnar blocks (30 mm × 30 mm × 10 mm) by a press machine. The sample blocks weighing 50±5 g were melted in an argon atmosphere of 5 L/min at a temperature of 1873 K for 2 hours to obtain homogeneous, followed by different experimental cooling methods (furnace cooling, air cooling, mist cooling, and water quenching) to room temperature. The Mo crucible were used for the pre-melting and a graphite crucible was used as a protection layer outside of the Mo crucible. The sample cooling process was shown in Fig. 1. The real-time temperature and time of steel slag cooling process were recorded by infrared temperature measuring device and stopwatch, and then the temperature curves were plotted to estimate the cooling rates of four cooling methods. With the increase of time, the cooling rate decreased gradually. The average cooling rates of the four cooling methods of a, b, c, d in high temperature stage from 1873 to 1173 K were 150 K/s, 80 K/s, 60 K/s and 9 K/s respectively, while the average cooling rates in the low temperature stage below 1173 K were reduced to 30 K/s, 6 K/s, 5 K/s and 3 K/s respectively.

Fig. 1.

Four cooling methods of the slag sample. (Online version in color.)

2.4. Characterization

The mineralogy of the slags was determined by X-ray diffraction analysis (XRD, Rigaku D/max 2500). Diffraction patterns were measured in 2θ range of 20 to 70 deg using Cu Kα radiation of 40 KV and 250 mA, with a 0.02 deg step size and scanning time per step of 200 seconds. Rietveld quantitative phase analysis was performed using HighScore^® software.

The chemical element distribution and microstructure characterization of the cooled slag were performed using a scanning electron microscopy (SEM, TESCAN VEGA2) equipped with energy dispersive spectrometry (EDS, ZRISS/ULTRA-55). The samples were coated with a conductive layer of gold before observation. The stability of steel slag was evealuated by the content of f-CaO. The more the content of f-CaO, the lower stability of steel slag.^29,30)

The content of f-CaO in steel slag was identified by chemical analysis according to Chinese standard YB/T 4328-2012. Since both Ca(OH)₂ and f-CaO in the slag can react with ethylene glycol, the total content of them (b₁) was obtained by chemical titration. The content of Ca(OH)₂ (b₂) was measured by thermogravimetric analysis. The content of f-CaO (b) in slag could be obtained by subtracting the b₂ from b₁ (b = b₁−b₂). The reaction equations were shown in Eqs. (1), (2), (3), (4).   

$$CaO+C_2{H}_6{O}_2\to {\left({CH}_{2}O\right)}_{2}Ca+H_{2}O\left(353\mathrm{\hspace{0.17em} }K-363\mathrm{\hspace{0.17em} }K\right)$$(1)

  

$$Ca{\left(OH\right)}_2+C_2{H}_6{O}_2\to {\left({CH}_{2}O\right)}_{2}Ca+2H_{2}O\left(353\mathrm{\hspace{0.17em} }K-363\mathrm{\hspace{0.17em} }K\right)$$(2)

  

$$\begin{array}{l}{C}_{10}{H}_{14}{N}_{2}N{\text{a}}_2{O}_8+{\left({CH}_{2}O\right)}_{2}Ca\to \\ C_{10}{H}_{14}{N}_{2}Ca{O}_8+{\left({CH}_{2}ONa\right)}_2\end{array}$$(3)

  

$$Ca{\left(OH\right)}_2\to CaO+H_{2}O$$(4)

3. Results and Discussion

3.1. Thermodynamics Calculation

The calculation results of Equilibrium and Scheil-Gulliver solidification of 47 mass%CaO-23 mass%FeO-12 mass%SiO₂-11 mass%MgO-3 mass%MnO-2 mass%Al₂O₃-2 mass%P₂O₅ mixture were shown in Figs. 2 and 3, respectively. The results showed that the steel slag was completely melted into liquid at 2273 K. Under the condition of equilibrium cooling, monoxide solid solution (RO) began to precipitate from the liquid at 2243.06 K, and then alpha-Ca₂SiO₄-Ca₃P₂O₈ solid solution (α-C₂SP), alpha-prime Ca₂SiO₄ (α′-C₂S), Ca₇P₂Si₂O₁₆, and Ca₃(Al,Fe)₂O₆ (C₃AF) were precipitated successively with the decrease of temperature. When the temperature was 1327.65 K, the liquid phase disappeared completely and the slag of 56.13 mass%RO-29.50 mass%C₂S-5.14 mass%C₃AF-9.22 mass%Ca₇P₂Si₂O₁₆ was obtained. However, compared with the equilibrium cooling, the phase transition based on Scheil-Gulliver solidification model was simple. The crystalline phases were precipitated from a single liquid phase, and there was no reaction from α-C₂SP to α′-C₂S. Therefore, the α-C₂SP could be obtained in Scheil-Gulliver solidification. The phase transition reactions of two cases were shown in Tables 2 and 3. Real solidification process was typically between these two cases. The theoretical calculations showed that the crystalline phases of steel slag mainly included silicate phase containing calcium, calcium ferrite, RO phase, and a small amount of solid solution containing P. The precipitation order was RO phase, silicate phase and calcium ferrite successively. The thermodynamic calculation results of this paper did not get Ca₃SiO₅ (C₃S) phase in the actual production process of rapid cooling steel slag, which was the result of equilibrium phase selection in the calculation process. Based on the minimum law of Gibbs energy under equilibrium conditions, when the corresponding data of α-C₂SP, α′-C₂S and RO phases were added to the calculation, the reaction Gibbs free energy of C₃S phase is larger than that of α-C₂SP and α′-C₂S and C₃S phase will not appear in the equilibrium calculation, as shown in Fig. 4. When the pure substance reacts, the formation reactions of α-C₂SP and α′-C₂S are selective with the change of temperature. The Gibbs free energy of the formation reaction of α-C₂SP at high temperature is smaller. When the temperature is reduced to 1715 K, the reaction is easier to form α′-C₂S phase, which is consistent with the calculation results of equilibrium and Scheil cooling conditions.

Fig. 2.

Thermodynamic calculation based on equilibrium cooling for steel slag. (Online version in color.)

Fig. 3.

Thermodynamic calculation based on a Scheil-Gulliver cooling model for steel slag. (Online version in color.)

Table 2. The phase transition reaction during equilibrium cooling.

Temperature range/K	Phase transition reaction
2243.06 to 1789.99	Liquid → RO
1789.99 to 1528.70	Liquid → RO + α-C2SP
1528.70 to 1343.30	Liquid + α-C2SP → RO + α′-C2S
1343.30	RO +α-C2SP → Liquid + RO + α′-C2S + Ca7P2Si2O16
1343.30 to 1330.93	Liquid → RO + α′-C2S + Ca7P2Si2O16
1330.93 to 1327.65	Liquid → RO + α′-C2S + C3AF + Ca7P2Si2O16
1327.65	56.13%RO-29.50%α′-C2S-5.14%C3AF-9.22%Ca7P2Si2O16

Table 3. The phase transition reaction based on Scheil-Gulliver solidification.

Temperature range/K	Phase transition reaction
2243.06 to 1780.86	Liquid → RO
1780.86 to 1498.70	Liquid → RO + α-C2SP
1498.70 to 1497.70	Liquid → RO + α-C2SP + α′-C2S
1497.70 to 1313.37	Liquid → RO + α′- C2S
1313.37 to 1292.16	Liquid → RO + α′-C2S + C3AF
1292.16 to 1292.03	Liquid → RO + α′-C2S + C3AF + Ca7P2Si2O16
1292.03	56.13%RO-26.98%α-C2SP-8.29%α′-C2S-5.11%C3AF-3.50%Ca7P2Si2O16

Fig. 4.

Reaction Gibbs free energy of α′-C₂S, α-C₂SP and C₃S. (Online version in color.)

3.2. Mineralogy

Figure 5 presents the XRD patterns of steel slag under different cooling methods. The peak position of slag samples did not change obviously under different cooling conditions. The main phase composition of four groups of steel slags included Ca₂SiO₄ (C₂S), Ca₃SiO₅ (C₃S), monoxide solid solution (RO), Ca₂Fe₂O₅ (C₂F), and free lime (f-CaO). Dicalcium silicate phase in steel slag is β-C₂S phase with monoclinic system. Different from α-C₂SP and α′-C₂S formed by thermodynamic calculation, β-C₂S was the main mineral of steel slag. The crystal form of dicalcium silicate changed from α′-C₂S to β-C₂S with the decrease of temperature during slag cooling.¹⁹⁾ Since the phase search matches were distinguished by structure, there were also C₂S solid solution with Ca₃P₂O₈ (C₂SP), C₂F solid solution with the Fe³⁺ substitution with Al³⁺ (C₂AF) in the steel slag. In addition, The XRD results of the rapid cooling in Figs. 5(c), 5(d) show no hump-shape in the 2θ range of 25 to 35° which represented the existence of the liquid phase. Compared with the cooling process of blast furnace slag, steel slag did not form glass phase under rapid cooling condition and all slag samples were completely crystallized. It might be due to the feature of high melting temperature and high basicity of steel slag.

Fig. 5.

XRD patterns of steel slag under different cooling methods. (Online version in color.)

However, the cooling rate had an important influence to the phase composition of steel slag. As the cooling rate increased, the peak intensity of the three strong peaks of the C₃S phase (2θ = 29.35°, 32.19°, 34.36°) on the fitted spectrum gradually increased. In order to further study the relationship between cooling rate and phase content, semi-quantitative analysis of XRD data was carried out. The weight percentage of mineral phase under different cooling methods was shown in Table 4. Under the condition of water quenching with fast cooling rate, the content of C₃S reached 25.3 mass%, the maximum among the four samples, while the content of lime reached the minimum. As a high-temperature stable phase, the theoretical crystallization temperature of C₃S is about 1723 K. When the temperature decreases to 1523 K, C₃S decomposes to C₂S and CaO.¹³⁾ The chemical reaction formula is shown in Eq. (5).   

$$C{a}_{3}Si{O}_5\to C{a}_{2}Si{O}_4+CaO\left(1\mathrm{\hspace{0.17em} }523\mathrm{\hspace{0.17em} }K\right)$$(5)

Table 4. XRD determined and mineralogical composition of four cooling methods, in weight percentage.

Mineral Phase	Chemical Formula	Cooling regime
		a. Furnace	b. Air	c. Mist	d. Water
C2S	Ca2SiO4	43.9	28.3	40	41.2
C3S	Ca3SiO5	19.2	20.4	23.1	25.3
RO	(Mg,Fe)O	36.4	31.6	31.9	28.3
C2AF	Ca2(Fe,Al)2O5	0.4	19.1	5.0	5.2
lime	CaO	0.1	0	0	0

Luxan et al.³¹⁾ found that rapid cooling inhibits the formation of C₂S. In our tests, the increase of cooling rate greatly could not only shorten cooling time and skip decomposition reaction of C₃S, but also reduce the content of free lime in steel slag. However, C₃S phase was not obtained by thermodynamic calculation. FactSage was calculated on the basis of Gibbs free energy minimization principle. Under the premise of reasonable selection of database, the Gibbs free energy of other solid solutions was smaller than that of C₃S in the equilibrium cooling process. The calcium-silicon phase in steel slag existed in the form of C₂S and C₂SP. The real cooling process did not reach equilibrium state, C₃S phase appeared in slag due to the different cooling condition.

3.3. SEM-EDS Analysis

The SEM images and EDS analysis of the furnace-cooled steel slag were shown in Fig. 6 and Table 5. The chemical elements in the steel slag included Ca, Fe, Si, O, Mg, Al, Mg and P. From the element analysis in Fig. 6(b) and Table 5, it can be seen that the dark gray block or elliptical area was RO phase, the main component was MgO, the average particle size was 20 μm, and the grain boundary had obvious white edges. The light gray plate area was the silicate phase, the grain diameter spanned a wide range of 20–50 μm, the main components were C₂S and C₃S, and also contained a small amount of C₂SP. The light gray irregular area was calcium ferrite as brownmillerite, the main component was C₂AF, which was distributed between the RO phase and the silicate phase. The white area was free lime, and there was a large area of excess CaO that was not involved in the reaction during the cooling process, and there were also small particles of CaO precipitated from the C₃S in the crystal grain boundary due to the decomposition reaction. The bright white round area was iron. Figures 6(c)–6(j) shows the element distribution of the furnace-cooled steel slag. It can be seen that the distribution of P was similar to that of Ca and Si, where the element distribution area of Ca was the largest, Si was the second, and P was the smallest. P element in steel slag was mainly enriched in silicate phase. From Table 5, P was detected in the calcium silicate phase with a P atomic percentage of 1.99–2.92%. Combined with the morphological and thermodynamic analysis, the phase containing P was mainly the solid solution of Ca₂SiO₄ and Ca₃P₂O₈ (C₂SP).

Fig. 6.

SEM images and EDS analysis of furnace cooling steel slag: (a, b) BSE images; (c–j) elemental distribution maps of the areas in a. (Online version in color.)

Table 5. EDS analysis results of the furnace cooling steel slag.

Detecting position	Atomic percentage (%)								Mineral phase
	Ca	Fe	Si	Mg	Mn	Al	P	O
1	28.64		12.38				1.99	56.99	C2SP + C3S
2	28.54		11.6				2.32	57.54	C2SP + C3S
3	0.77	7.34		39.5	2.63			49.76	RO
4	0.79	8.49		42.25	4.57			43.90	RO
5	23.05		3.25			12.12		61.59	C2AF
6	27.93		9.29			3.04	2.92	56.83	C2SP +C2AF
7	24.28	1.97	1.52		2.90			69.33	f-CaO
8	15.92							84.08	f-CaO
9	2.77	97.23							Fe

Compared with the as furnace cooling slag, the light gray irregular area distribution area of the air cooling slag in Fig. 7 expanded, and the content of calcium ferrite increased with the increase of cooling rate. According to thermodynamic calculations, the precipitation temperature of calcium ferrite was lower than that of the silicate phase and RO phase. During the crystallization process, calcium ferrite finally crystallized and was used located along the grain boundary. From the SEM image in Fig. 7(b), it can be seen that the silicate phase grain size in the air-cooled slag became smaller, and C₂S and C₃S with an average diameter of less than 15 μm appeared. In mist-cooled and water-cooled slag, this trend became more and more obvious. In Fig. 8, the morphology of the silicate phase of the mist cooling slag was elongated, and in Fig. 9 calcium ferrite in the water quenching slag was network-shaped to wrap the grain size 2–15 μm granular silicate phase. Increasing the cooling rate could reduce the grain size of the silicate phase.

Fig. 7.

SEM images and EDS analysis of air cooling steel slag: (a, b) BSE images; (c–j) elemental distribution maps of the areas in a. (Online version in color.)

Table 6. EDS analysis results of the air cooling steel slag.

Detecting position	Atomic percentage (%)							Mineral phase
	Ca	Fe	Si	Mg	Mn	Al	O
1	30.41		11.23				58.35	C2S + C3S
2	29.98		8.77				61.26	C2S + C3S
3	0.87	8.85		41.28	4.21		44.79	RO
4	0.82	2.76		43.91	1.99		50.52	RO
5	21.87	6.56	1.15			12.26	58.18	C2AF
6	21.06	7.85			0.61	11.67	58.81	C2AF
7	25.47	1.16	9.45				63.92	C2S + f-CaO
8	18.97	6.51	1.53	1.84	4.03		67.11	RO + f-CaO + C2S

Fig. 8.

SEM images and EDS analysis of mist cooling steel slag: (a, b) BSE images; (c–j) elemental distribution maps of the areas in a. (Online version in color.)

Table 7. EDS analysis results of the mist cooling steel slag.

Detecting position	Atomic percentage (%)							Mineral phase
	Ca	Fe	Si	Mg	Mn	Al	O
1	28.41	0.65	10.62				60.31	C3S + C2S
2	32.74	1.21	11.30		0.86		53.89	C3S
3	1.14	3.77		48.41	2.66		44.01	RO
4	0.85	4.27		48.02	2.82		44.04	RO
5	23.66	11.66	1.98		0.92	12.61	49.17	C2AF
6	28.73	7.40	1.75			18.17	43.94	C2AF
7	25.87	4.01	5.96			7.87	56.28	C2AF + C3S
8	16.22	4.37	4.73	2.02	2.98		69.69	C3S + C2S + RO
9		100						Fe

Fig. 9.

SEM images and EDS analysis of water quenching steel slag: (a, b) BSE images; (c–j) elemental distribution maps of the areas in a. (Online version in color.)

According to the statistics of Mg element area EDS analysis of four group slags in Figs. 6, 7, 8, 9, the proportions of RO phase as furnace cooling, air cooling, mist cooling, and water quenching were 28.03%, 22.53%, 14.17%, and 13.30%, respectively. With the increase of cooling rate, the mass percentage of RO phase in steel slag decreased, which was consistent with the semi analytical results of XRD. According to the thermodynamic analysis, it could be obtained that during the cooling process of steel slag, RO phase crystallizes first, followed by the silicate phase and calcium ferrite, respectively. It was speculated that the increase of cooling rate could inhibit the reaction of precipitation of RO phase from liquid slag. XRD results showed that silicate phase in steel slag include dicalcium silicate (C₂S) and tricalcium silicate (C₃S). C₃S was a metastable phase at room temperature and could be thermally decomposed to form C₂S and CaO at 1523 K. The increase of cooling rate could quickly pass through the unstable temperature stage, skip the decomposition reaction of components and solidify directly to obtain more C₃S. As shown in Fig. 9 and Table 8, f-CaO was not detected in the field of vision of water quenching steel slag, and the proportion of hexagonal plate-like area representing C₃S increased.

Table 8. EDS analysis results of the water quenching steel slag.

Detecting position	Atomic percentage (%)								Mineral phase
	Ca	Fe	Si	Mg	Mn	Al	O	P
1	27.06		11.10	1.20			60.63		C3S + C2S
2	27.45		10.65	2.09			59.81		C3S + C2S
3	29.37		11.50	1.74			57.39		C3S + C2S
4	28.58		10.48				58.25	2.69	C2SP + C3S
5	1.22	3.49		50.97	1.99		42.32		RO
6	0.84	3.70	0.29	51.04	2.62		41.5		RO
7	25.76	8.73	5.43		2.89	5.00	52.19		C2AF + C2S
8	18.51	16.41	2.42	5.48	8.35		48.82		C2AF
9	27.93	7.28	1.99	6.00		2.03	54.77		C2AF + C3S
10		100							Fe

Although the cooling rate does not change the crystallization tendency of steel slag,²⁶⁾ it could affect the content and morphology of crystalline phase. The grain size and quantity of steel slag crystalline phase could be controlled by different cooling methods. Therefore, this paper expected to further study the relationship between cooling rate and crystallization behavior of steel slag, and explain the effect of cooling rate on the stability of steel slag.

3.4. Stability Analysis

The volume stability of steel slag is an important index used to evaluate the quality of road construction materials mixed with steel slag. It is related to the composition of steel slag (f-CaO, f-MgO).^32,33,34) These components could react with water to produce phase transformation and produce volume expansion, which led to materials cracking and pulverization. According to Chinese standard GB/T 24175-2009, the autoclave chalked ratio is used to characterize the stability of steel slag, and the value for construction materials should not be more than 5.9%. Xu et al.³⁵⁾ measured the content of f-CaO w (%) and the autoclave chalked ratio r (%) in steel slags from 25 Steel factories in China, and the volume stability index can be calculated by Eq. (6):   

$$r=1.58w+1.67$$(6)

In present study, the content of f-CaO in steel slag with different cooling methods was obtained by chemical analysis. The results in Fig. 10 showed that the content of f-CaO in steel slag decreased with the increase of cooling rate. The content of f-CaO in samples was reduced from 14.75 mass% of furnace cooling slag to 2.64 mass% of water quenching slag. According to the empirical formula Eq. (6), the autoclaved chalked ratios of steel slag under furnace cooling, air cooling, mist cooling, and water quenching were 24.98%, 19.97%, 11.23%, and 5.84%, respectively. The stability index of water quenching slag was less than 5.9%, which met the requirements of steel slag for road construction materials. Therefore, increasing the cooling rate could effectively improve the volume stability of steel slag. With the increase of the cooling rate of steel slag, some crystallization reactions were avoid. Rapid cooling not only greatly reduced the occurrence of decomposition reaction of C₃S in steel slag and reduced the content of f-CaO in steel slag, but also the formed active minerals underwent hydration reaction, improving the volume stability of steel slag.

Fig. 10.

Relationship between cooling rate and the content of f-CaO in steel slag. (Online version in color.)

4. Conclusions

The main conclusions of the present study are summarized as follows:

(1) The increase of cooling rate has minor effect on the solidification tendency of steel slag, but affects the content of mineral phase. Mineral phases of steel slag contain C₂S, C₂SP, C₃S, monoxide solid solution (RO), C₂F, and f-CaO. With the increase of cooling rate, the content of C₃S increases, while the content of RO phase, calcium ferrite and free lime decrease gradually.

(2) The increase of cooling rate decreases the grain size, and the silicate phase with grain size less than 5 μm appears in the water quenching slag. The distribution of f-CaO decreases with the increase of cooling rate, and rapid cooling can skip the decomposition reaction of C₃S and reduce the precipitation of f-CaO. According to the chemical element mapping, the content of RO phase in steel slag decreases with the increase of cooling rate.

(3) The increase of cooling rate is beneficial to the volumetric stability of steel slag. The content of f-CaO in steel slag by as furnace cooling, air cooling, mist cooling, and water quenching are 14.75%, 11.58%, 6.05%, and 2.64%, respectively. With the increase of cooling rate, the autoclaved chalked ratios of steel slag calculated theoretically decreases from 24.98% to 5.84%, which can greatly improve the strength of the construction materials.

Acknowledgements

The authors wish to thank Baosteel and Green extraction metallurgy Research Group of polymetallic ore. This work was supported by the National Key R&D Program of China (2018YFC1900500) and the Graduate Scientific Research and Innovation Foundation of Chongqing (Grant No. CYB20002).

References

• 1)  National Bureau of Statistics of China, China Statistics Press Beijing: National Bureau of Statistics, http://www.stats.gov.cn/, (accessed 2022-01-21).
• 2)   J.  Guo,  Y.  Bao and  M.  Wang: Waste Manag., 78 (2018), 318. https://doi.org/10.1016/j.wasman.2018.04.045
• 3)   L.  Kang,  H.  Du,  H.  Zhang and  W.  Ma: Complexity, 2018 (2018), 6703908. https://doi.org/10.1155/2018/6703908
• 4)   L.  Li,  T.  Ling and  S.  Pan: Sci. Total Environ., 806 (2022), 150280. https://doi.org/10.1016/j.scitotenv.2021.150280
• 5)   Y.  Jiang,  T.  Ling,  C.  Shi and  S.  Pan: Resour. Conserv. Recycl., 136 (2018), 187. https://doi.org/10.1016/j.resconrec.2018.04.023
• 6)   E.  Belhadj,  C.  Diliberto and  A.  Lecomte: Cem. Concr. Compos., 45 (2014), 15. https://doi.org/10.1016/j.cemconcomp.2013.09.016
• 7)   Y.  Fang,  W.  Su,  Y.  Zhang,  M.  Zhang,  X.  Ding and  Q.  Wang: J. Therm. Anal. Calorim., 147 (2022), 6181. https://doi.org/10.1007/s10973-021-10914-z
• 8)   Z.  Li,  M.  Whitwood,  S.  Millman and  J.  van Boggelen: Ironmaking Steelmaking, 41 (2014), 112. https://doi.org/10.1179/1743281213Y.0000000108
• 9)   C.  Liu,  M.  Guo,  L.  Pandelaers,  B.  Blanpain and  S.  Huang: Metall. Mater. Trans. B, 47 (2016), 3237. https://doi.org/10.1007/s11663-016-0809-4
• 10)   W.  Kuo and  C.  Shu: Constr. Build. Mater., 50 (2014), 463. https://doi.org/10.1016/j.conbuildmat.2013.09.030
• 11)   J.  Shi,  J.  Tan,  B.  Liu,  J.  Chen,  J.  Dai and  Z.  He: J. Hazard. Mater., 403 (2021), 123983. https://doi.org/10.1016/j.jhazmat.2020.123983
• 12)   A.  Cramb: ISIJ Int., 54 (2014), 2665. https://doi.org/10.2355/isijinternational.54.2665
• 13)   F.  Engström,  D.  Adolfsson,  Q.  Yang,  C.  Samuelsson and  B.  Bjorkman: Steel Res. Int., 81 (2010), 362. https://doi.org/10.1002/srin.200900154
• 14)   Y.  Ta,  T.  Hoshino,  H.  Tobo,  K.  Watanabe and  K.  Takahashi: ISIJ Int., 61 (2021), 1716. https://doi.org/10.2355/isijinternational.ISIJINT-2020-632
• 15)   Y.  Ta,  T.  Hoshino,  H.  Tobo and  K.  Watanabe: ISIJ Int., 59 (2019), 1917. https://doi.org/10.2355/isijinternational.ISIJINT-2019-147
• 16)   L.  Kriskova,  Y.  Pontikes,  L.  Pandelaers,  O.  Cizer,  P.  Jones,  K.  Van Balen and  B.  Blanpain: Metall. Mater. Trans. B, 44 (2013), 1173. https://doi.org/10.1007/s11663-013-9894-9
• 17)   H.  Tobo,  K.  Watanabe,  M.  Kuwayama,  S.  Goto,  H.  Goto and  T.  Tanaka: ISIJ Int., 55 (2015), 2499. https://doi.org/10.2355/isijinternational.ISIJINT-2015-254
• 18)   P. K.  Son and  Y.  Kashiwaya: ISIJ Int., 48 (2008), 1165. https://doi.org/10.2355/isijinternational.48.1165
• 19)   L.  Jiang,  Y.  Bao,  Q.  Yang,  Y.  Chen,  G.  Liu,  F.  Han,  J.  Wei,  F.  Engström and  J.  Deng: Steel Res. Int., 88 (2017), 1700066. https://doi.org/10.1002/srin.201700066
• 20)   M.  Gautier,  J.  Poirier,  F.  Bodénan,  G.  Franceschini and  E.  Véron: Int. J. Miner. Process., 123 (2013), 94. https://doi.org/10.1016/j.minpro.2013.05.002
• 21)   C.  Myers and  T.  Nakagaki: ISIJ Int., 58 (2018), 211. https://doi.org/10.2355/isijinternational.ISIJINT-2017-478
• 22)   Z.  Liu,  L.  Chen,  B.  Blanpain and  M.  Guo: ISIJ Int., 58 (2018), 1972. https://doi.org/10.2355/isijinternational.ISIJINT-2018-176
• 23)   S.  Esfahani,  S.  Mostaghel and  M.  Barati: J. Non-Cryst. Solids, 436 (2016), 29. https://doi.org/10.1016/j.jnoncrysol.2015.12.013
• 24)   B.  Lin,  H.  Wang,  X.  Zhu,  Q.  Liao and  B.  Ding: Appl. Therm. Eng., 96 (2016), 432. https://doi.org/10.1016/j.applthermaleng.2015.11.075
• 25)   M.  Tossavainen,  F.  Engstrom,  Q.  Yang,  N.  Menad,  M.  Lidstrom Larsson and  B.  Bjorkman: Waste Manag., 27 (2007), 1335. https://doi.org/10.1016/j.wasman.2006.08.002
• 26)   M.  Choi and  S.  Jung: J. Non-Cryst. Solids, 468 (2017), 105. https://doi.org/10.1016/j.jnoncrysol.2017.04.042
• 27)   B.  Mason: J. Iron Steel Inst., 11 (1994), 69.
• 28)   C.  Bale,  E.  Bélisle,  P.  Chartrand,  S.  Decterov,  G.  Eriksson,  K.  Hack,  I.  Jung,  Y.  Kang,  J.  Melançon,  A.  Pelton,  C.  Robelin and  S.  Petersen: Calphad, 33 (2009), 295. https://doi.org/10.1016/j.calphad.2008.09.009
• 29)   R.  Jia,  J.  Liu and  R.  Jia: Cem. Concr. Res., 91 (2017), 123. https://doi.org/10.1016/j.cemconres.2016.11.011
• 30)   L.  Juckes: Miner. Process. Extr. Metall., 112 (2003), 177. https://doi.org/10.1179/037195503225003708
• 31)   M. P.  Luxán,  R.  Sotolongo,  F.  Dorrego and  E.  Herrero: Cem. Concr. Res., 30 (2000), 517. https://doi.org/10.1016/S0008-8846(99)00253-7
• 32)   H.  Yi,  G.  Xu,  H.  Cheng,  J.  Wang,  Y.  Wan and  H.  Chen: 7th Int. Conf. on Waste Management and Technology (ICWMT), (Beijing), Elsevier, Amsterdam, (2012), 791.
• 33)   R.  Inoue and  H.  Suito: ISIJ Int., 35 (1995), 272. https://doi.org/10.2355/isijinternational.35.272
• 34)   D.  Stewart,  A.  Bray,  G.  Udoma,  A.  Hobson,  W.  Mayes,  M.  Rogerson and  I.  Burke: Environ. Sci. Pollut. Res., 25 (2018), 9861. https://doi.org/10.1007/s11356-018-1260-7
• 35)   G.  Xu and  Y.  Huang: Ind. Saf. Environ. Prot., 41 (2015), 94 (in Chinese).