Influence of Hydrothermal Pretreatment Temperature on the Hydration Properties and Direct Carbonation Efficiency of Al-Rich Ladle Furnace Refining Slag

Abstract

The influence of hydrothermal pretreatment temperature on the hydration products and carbonation efficiency of Al-rich LF slag was investigated. The results showed that the carbonation efficiency was strongly dependent on the morphology of hydration products and the hydration extent of the raw slag. Hydrothermal pretreatment at 20 °C or 80 °C favored the formation of flake-shaped products with a higher specific surface area and therefore resulted in a higher CO₂ uptake of 20 °C and 80 °C-pretreated slags (13.66 wt% and 10.82 wt%, respectively). However, hydrothermal pretreatment at 40 °C, 60 °C or 100 °C led to the rhombohedral-shaped calcite layer surrounding the unreacted core of the raw slag and the formation of fewer flake-shaped products, resulting in a lower CO₂ uptake of 40 °C, 60 °C and 100 °C-pretreated slags (9.21 wt%, 9.83 wt%, and 6.84 wt%, respectively).

1. Introduction

The growth of global greenhouse gas emissions was 2.0% in 2018 and there is no sign that any of these emissions are peaking yet. The six largest emitters of greenhouse gases, together accounting for 62% globally, are China (26%), the United States (13%), the European Union (more than 8%), India (7%), the Russian Federation (5%), and Japan (almost 3%) [1]. China’s carbon emission peak is a matter of international focus. Recently, China made a solemn promise to peak its carbon dioxide emission by 2030 and achieve carbon neutrality by 2060. Some of the main measures China will use to reduce CO₂ emissions over next 10 years are: changing energy and industrial structures, transforming the development mode, promoting clean energy, and appropriately increasing carbon sequestration ability. Among the current CO₂ sequestration routes, mineral carbonation is regarded as a potential technology because of its advantages; it is environmentally benign, it enables the permanent trapping of CO₂ in the form of carbonate, and it does not require post-storage surveillance for CO₂ leakage [2]. In general, mineral carbonation can be divided into two categories, namely direct carbonation and indirect carbonation. Direct mineral carbonation is accomplished through the reaction of a solid alkaline mineral with CO_2, either in gaseous or in aqueous phase [3].

Alkaline solid wastes such as red mud, steel slag, blast furnace slag, fly ash, etc., are used for direct mineral carbonation as efficient and economically available capturers of CO₂ [4,5,6,7,8,9]. For the direct mineral carbonation of steel slag, the formation of an increasingly thick and dense carbonate layer surrounding the unreacted core of the solid particle hinders further carbonation and results in the lower CO₂ capture capacity [10]. In our previous study [11], the improvement in the direct carbonation efficiency of Al-rich ladle furnace refining slag (LF slag) by hydrothermal pretreatment was investigated. The results showed that after hydrothermal pretreatment at 80 °C, the morphology of Ca₁₂Al₁₄O₃₃(C₁₂A₇) in the slag transformed from separated particles to the flake-shaped Ca₃Al₂O₆·xH₂O(C₃AH_x), resulting in an increased reaction surface area and carbonation efficiency. However, this study did not discuss the effect of hydrothermal temperature on the carbonation efficiency. In fact, the hydration product of C₁₂A₇ is dependent on the hydration temperature. Koplík et al. [12] reported that at 20 °C the major hydration products of C₁₂A₇ were Ca₂Al₂O₅·8H₂O(C₂AH₈) and CaAl₂O₄·10H₂O(CAH₁₀); at 30 °C CAH₁₀ disappeared and only C₂AH₈ remained; at 60 °C the only stable hydrates–Ca₃Al₂O₆·6H₂O(C₃AH₆) and Al(OH)₃(AH) were formed. Edmonds et al. [13] stated that both C₂AH₈ and CAH₁₀ can be produced during the hydration of C₁₂A₇ at 4 °C, while no trace of CAH₁₀ was spotted when C₁₂A₇ was hydrated at 20 or 40 °C. Given that the morphology of the hydration product of C12A7 has a significant effect on the carbonation efficiency of LF slag and the type of hydration product produced is related to temperature, the aim of this study was to investigate the influence of hydrothermal pre-treatment temperature on the hydration properties and the carbonation efficiency of Al-rich LF slag at ambient temperature and pressure. Moreover, the relation between the morphology of the hydration product and the carbonation efficiency was clarified in this work.

2. Materials and Methods

2.1. Materials

The Al-rich LF slag used in this study was collected from the Xiangtan steel plant in Hunan province, China. The chemical composition of the slag determined by X-ray Fluorescence (XRF) is listed in

Table 1. Chemical composition of raw slag as determined by XRF analysis.

Component	CaO	Al2O3	SiO2	MgO	TiO2	SO3	Fe2O3	Others
wt%	52.0056	23.66	15.83	4.05	0.90	2.45	0.54	0.57

. Before pretreatment, the raw slag was crushed and ground into a powder <20 mesh particle size. Distilled water was used in this study for slag suspension preparation.

2.2. Hydrothermal Pretreatment of LF Slag at Different Temperatures

At first, the raw slag was fully mixed with water at a solid/water (S/W) ratio of 1:10 in a beaker. Then, the suspension was stirred for 30 min at 20, 40, 60, 80 and 100 °C (designated as 20H, 40H, 60H, 80H and 100H-slag, respectively). Next, the suspension was filtered and the obtained solid was sufficiently washed and dried to a constant weight at 105 °C for further characterization and for the following carbonation experiment.

2.3. Direct Aqueous Carbonation Process

The schematic diagram of the aqueous carbonation experimental system is shown in Figure 1. The pretreated slag suspension with a solid/water ratio of 1:10 was placed in a conical flask into an electric-heated thermostatic water bath equipped with a mechanical stirrer. The temperature of the water bath was kept at 40 °C. Then, 99.99% pure CO₂ from the CO₂ cylinder was injected into the suspension at a flow rate of 5 L/min controlled by a flowmeter, and was simultaneously stirred for carbonation. The suspension underwent the carbonation process for 60 min and was then filtered. The obtained solid was dried to a constant weight at 105 °C for further characterization.

2.4. Characterization of Slag

X-ray diffraction (XRD, Bruker AXS company D8 Advance, Karlsruhe, Germany) was conducted on the slags to identify their main mineral phases. The scanning range was from 5° to 70° 2θ at 2°/min. TG-DSC analysis was performed using a METTLER TOLEDO 1600 LF thermal gravimetric analyzer. A field-emission scanning electron microscope (FESEM, Hitachi company SU8010, Tokyo, Japan) was used to characterize the morphology of the slag. The specific surface area of the slags was measured by the N₂ gas adsorption Brunauer–Emmet–Teller (BET) method (ASAP 2020, Micromeritics, Norcross, GA, USA). The pH value of the slag suspension was determined by a PHS-3C pH meter.

2.5. Analysis of Carbonation Efficiency

In order to compare the carbonation efficiency of Al-rich LF slags under the hydrothermal pretreatment at different temperatures, the CO₂ uptake of the slags was measured based on the weight fraction of the TG curve (Δm_{600–800 °C}) and the dry weight (m) [14] expressed in terms of CO₂ (wt%), Equation (1):

$${\mathrm{CO}}_2(\mathrm{wt}\%)=\frac{{\mathsf{\Delta }\mathrm{m}}_{600\text{–}800{ }^{°}\mathrm{C}}}{\mathrm{m}}\times 100$$

(1)

3. Results and Discussion

3.1. Influence of Hydrothermal Temperature on Hydration Properties of Al-Rich LF Slag

Figure 2 shows the XRD patterns of the raw slag and the pretreated slags. The main mineral phases of raw slag were C₁₂A₇ and Ca₂SiO₄(CS₂). For all the slags with pretreatment, the peaks of C₁₂A₇ were reduced, indicating its hydration. The 40H-slag and 100H-slag presented relatively more intense residual C₁₂A₇ peaks than the other slags, suggesting the lower hydration extent of these two slags. For the 20H-slag, 3CaO·Al₂O₃·CaCO₃·11H₂O(C₄ACH₁₁) was the dominant hydration product. A small amount of (C₄ACH₁₁) also appeared in the 40H-slag and C₃AH₆ was the other main product for this slag. Hydrocalumite (Ca₄Al₂(OH)₁₂CO₃·5H₂O) was only present in the 100H-slag, and this slag had the most intense C₃AH₆ peaks. With respect to the 60H-slag, only C₃AH₆ crystal was found. In general, C₃AHx, C₄ACH₁₁, and C₃AH₆ were the main hydration products, while the other products mentioned in the “Introduction” (such as CAH₁₀ and C₂AH₈) were not observed. This could be explained by the fact that CAH₁₀ and C₂AH₈ are the transition phases and can be converted to the ultimate stable products (e.g., C₃AH₆ and C₃AH_x), described by Equations (2)–(4), respectively [15].

2CAH₁₀ → C₂AH₈ + AH₃ + 9H

(2)

3C₂AH₈ → 2C₃AH₆ + AH₃ + 9H

(3)

3C₂AH₈ → 2C₃AH_x + AH₃ + (21 − 2x)H

(4)

Figure 3 displays the FESEM pictures of the raw slag and the pretreated slags. The raw slag appeared as irregular-shaped particles with dense and coarse surfaces (Figure 3a). After hydration at 20 °C, the slag surface became smooth due to the formation of flake-shaped C₄ACH₁₁ (Figure 3b). In addition, metastable hydrates in the form of hexagonal platelets [14] were observed (Figure 3b). The microstructure of the 40H-slag presented as a mixture of C₄ACH₁₁, metastable hydrated hexagonal-shaped platelets [16], and unhydrated slag particles (Figure 3c). The edge of the unhydrated particles in the 40H-slag was covered by rhombohedral-shaped CaCO₃ (calcite) particles and AH gel with a grain size of 0.5 μm, which may hinder the further hydration of C₁₂A₇ (Figure 3d). Once again, this verified that the thick and dense CaCO₃ layer surrounding the unreacted core of the solid particle was the main cause of the low carbonation efficiency of the slag without hydrothermal pretreatment, as illustrated in our previous studies [9]. CaCO₃ and AH gel should be generated by the indirect carbonation reaction between C₁₂A₇ and CO₂ in the air, which can be described by Equations (5) and (6) [17]. The occurrence of Equation (5) resulted in the alkalinity of the slag suspensions (the final pH values were 10.42, 10.13, 10.05, 10.56, and 10.02 for the slag suspensions treated at 20 °C, 40 °C, 60 °C, 80 °C, and 100 °C, respectively).

The FESEM image of the 60H-slag (Figure 3e) indicates the formation of flake-shaped and cubic hydrates that should be amorphous calcium carboaluminate (CCA) and C₃AH₆, respectively. In addition, a dense rhombohedral-shaped calcite layer covered part of the slag surface (Figure 3f). Therefore, it is assumed that CCA was formed through the reaction between C₃AH₆ and calcite [18,19]. This also provides an explanation for why the 20H-slag contained large amounts of C₄ACH₁₁ but small amounts of calcite formation (the reaction between C₃AH₆ and calcite can be described by Equation (7) [19]). After hydration at 80 °C, the morphology of the slag changed from separated particles to continuous gel (Figure 3g). The flake-shaped gel should be C₃AH_x, and calcite particles were scattered on the surface of C₃AH_X in the 80H-slag (Figure 3g). For the 100H-slag, flake-shaped hydrocalumite and cubic C₃AH₆ particles were embedded in the unhydrated slag particles (Figure 3h), and a rhombohedral-shaped calcite layer deposited on the edge of the slag particles (Figure 3i), similarly to the 40H and 60H slag.

Ca₁₂Al₁₄O₃₃ + 33H₂O → 12Ca²⁺ + 14Al(OH)₄⁻ + 10OH⁻

(5)

Ca²⁺ + CO₂ + 2Al(OH)₄⁻ → CaCO₃↓ + 2Al(OH)₃↓ + H₂O

(6)

CaCO₃ + Ca₃Al₂O₆·6H₂O + 5H₂O → 3CaO·Al₂O₃·CaCO₃·11H₂O

(7)

The BET specific surface area (S_BET) of the slags were listed in

Table 2. The BET specific surface area of the pretreated slags.

	Raw Slag	20H-Slag	40H-Slag	60H-Slag	80H-Slag	100H-Slag
SBET (m2/g)	4.68	8.32	5.89	6.15	9.40	5.15

. The S_BET of the 20H-slag and the 80H-slag was more than two times that of the raw slag, while other pretreated slags demonstrated only a slight S_BET increase compared with the raw slag. This should be attributed to the larger amount of flake-shaped hydrates in the 20H-slag and the 80H-slag [20].

Figure 4 shows the TG-DSC analysis results of the pretreated slags. The endothermic peak of around 260–270 °C denoted the decomposition of Al(OH)₃ and appeared in all slags [19]. This peak was overlapped by the endothermic peak between 280 °C and 325 °C, which was attributed to the dehydration of C₃AH₆ in the 40H-slag, 60H-slag, and 80H-slag [18]. These results corresponded well with the XRD and FESEM analysis. The endothermic peak at 157 °C in the 20H-salg indicated the dehydration of C₄ACH₁₁ [16], while this peak became broad for the 40H-slag due to the low crystallinity of C₄ACH₁₁ [16], as was also reflected in the broad peak of XRD patterns (Figure 1). The absence of C₄ACH₁₁ in the other slags can be explained by its instability in temperatures above 40 °C [19]. A very broad endothermic peak between 80 and 200 °C was observed in the 60H-slag, generated by the dehydration of amorphous CCA [14]. The endothermic peak at 155 °C in the 80H-slag represented the dehydration of C₃AH_x which is close to the dehydration temperatures of CAH₁₀ and C₂AH₈ [15]. With respect to the 100H-slag, the dehydration of hydrocalumite was reflected in the endothermic peak around 146 °C. In general, the dehydration of hydrates mainly occurred over the temperature range of 105–325 °C, resulting in significant weight loss. The other significant weight loss region was between 600 and 800 °C, which was ascribed to the CaCO₃ decomposition. During the hydrothermal process, C₁₂A₇ was transformed into calcium aluminates hydrate (CAH), CAC, AH, and CaCO₃; therefore, the mass loss ratio of the slags (See

Table 3. The mass loss ratio of the pretreated slags between 105 and 800 °C.

Sample	Mass Loss Ratio between 105 and 800 °C (wt%)
20H-slag	20.58
40H-slag	13.80
60H-slag	15.95
80H-slag	20.70
100H-slag	15.07

) above the temperature range of 105–800 °C should be an indicator of C₁₂A₇ hydration extent. It may be concluded from the results in Table 2 that the 20H-slag and the 80H-slag had significantly higher hydration extent than other three slags, in good agreement with the XRD and FESEM analysis.

In conclusion, cubic C₃AH₆ was a main hydration product for the 40H-slag, 60H-slag, and 100H-slag. Part of C₃AH₆ could react with CaCO₃ to generate CCA while the unreacted rhombohedral-shaped CaCO₃ layer covered the slag surface, resulting in the hindrance of further hydration for these three slags. By contrast, flake-shaped C₄ACH₁₁ and C₃AH_x were the main hydration products for the 20H-slag and the 80H-slag, respectively, and their higher specific surface area may accelerate the carbonation reaction.

3.2. Carbonation Efficiency of Slags Pretreated at Different Temperatures

The XRD patterns of the slags after carbonation are shown in Figure 5. After carbonation, the peaks of C₃AH_x, C₄ACH₁₁, and hydrocalumite disappeared or showed a significant decrease in intensity while the calcite peaks increased in intensity. This suggests the carbonation of these hydrates. On the contrary, C₃AH₆ appeared less active in terms of its carbonation, which may be the main cause of the less intense calcite peaks in the carbonated100H-slag with C₃AH₆ as the main hydration product (see Figure 1). In addition, residual C₄ACH₁₁ peaks were observed, indicating the incomplete carbonation of C₄ACH₁₁ in the 20H-slag during carbonation. This resulted in the less intense calcite peaks in the carbonated 20H-slag compared with the carbonated 80H-slag.

Figure 6 exhibits the FESEM pictures of the slags after carbonation. For the carbonated 20H-slag, flake-shaped C₄ACH₁₁ was decomposed and cubic calcite crystals were observable (Figure 6a). In the carbonated 40H-slag and 60H-slag, some of the cubic calcite crystals were surrounded by unreacted hydrates (Figure 6b,c). Larger amounts of cubic calcite crystals appeared in the carbonated 80H-slag than in the other slags, leading to the breakdown of continuous C₃AH_x gel (Figure 6d). Moreover, the carbonation products were covered by a small amount of unreacted C₃AH_x debris (Figure 6d). The microstructure of the carbonated 100H-slag (Figure 6e) was similar to the carbonated 40H-slag and 60H-slag; it was composed of unreacted hydration products, unhydrated slag, and some cubic calcite crystals. In each of the slags, the cubic calcite was generated by the direct reaction of CO₂ with the hydrates and amorphous AH, as the other reaction product surrounded the cubic calcite crystals [11]. In each of the carbonated slags, calcite appeared as non-uniform aggregated crystal particles, which indicate direct carbonation [21]. Direct carbonation of alkaline slag involved two stages: CO₂ dissolution and carbonation reaction [21,22]. The simplified direct carbonation mechanism of slags with hydrothermal pretreatment in this study was summarized in

Table 4. Chemical reaction equations of the direct carbonation of the pretreated slags in this study.

Stage	Chemical Reaction Equation
CO2 dissolution	CO2(g) → CO2(aq) CO2(aq) + 2OH−(aq) → CO32−(aq) + H2O(l)
Carbonation	Ca3Al2O6·6H2O(s) + 3CO32−(aq) → 3CaCO3(s) + 2Al(OH)3(s) + 6OH−(aq) Ca3Al2O6·xH2O(s) + 3CO32−(aq) → 3CaCO3(s) + 2Al(OH)3(s) + 6OH−(aq) + (x − 6)H2O(l) 3CaO·Al2O3·CaCO3·11H2O(s) + 3CO32−(aq) → 4CaCO3(s) + 2Al(OH)3(s) + 6OH−(aq) + 5H2O(l)

.

Based on the TG curves of the carbonated slags (Figure 7), CO₂ uptake (wt%) was calculated with Equation (1) where m was the dry weight at 325 °C (at this temperature, free water and chemically bound water evaporated). The results of CO₂ uptake were listed in

Table 5. CO₂ uptake of pretreated slags.

	20H-Slag	40H-Slag	60H-Slag	80H-Slag	100H-Slag
CO2 Uptake(wt%)	10.82	9.21	9.83	13.66	6.84

. The CO₂ uptake of slags followed this order: 80H-slag (13.66 wt%) > 20H-slag (10.82 wt%) > 60H-slag (9.83 wt%) > 40H-slag (9.21 wt%) > 100H-slag (6.84 wt%), which corresponded well with the XRD and FESEM results. This is attributed to the following reasons: (1) a dense CaCO₃ or AH gel layer covered the unhydrated slag surface in the 40H-slag and the 60H-slag, therefore resulting in the hindrance of further hydration and carbonation; (2) flake-shaped hydrates such as C₄ACH₁₁ in the 20H-slag or C₃AH_x in the 80H-slag provided a larger reaction surface aera than the cubic C₃AH₆ and raw slag particles, avoiding calcite and AH gel layer formation on the unreacted hydrates surface. In short, the carbonation efficiency was strongly dependent on the type and morphology of the hydrates of LF slag.

The maximum CO₂ uptake among these five pretreated slags was 13.66 wt% (namely 136.6 g of CO₂/1 kg of slag) which was attained by the 80H-slag. Compared with previous studies about wet direct carbonation of steelmaking slags conducted under ambient temperature and pressure (See

Table 6. Comparison of current with previous steelmaking slag carbonation studies conducted under ambient temperature and pressure.

Slag Type	CO2 Uptake (g of CO2/1 kg of Slag)	Conditions	Reference
Al-rich LF slag	136.6	Temperature: 40 °C	This study
		S/W ratio: 1:10
		Reaction time: 1 h
EAF steel slag	87	Temperature: 20 °C	[23]
		S/W ratio: 1:10
		Reaction time: 37 min
BOF steel slag	116.4	Temperature: 70 °C	[24]
		S/W ratio: 1:2
		Reaction time: 2 h
LF slag	67	Temperature: 20 °C	[25]
		S/W ratio: 1:20
		Reaction time: 1 h
BOF steel slag	168.32	Temperature: 60 °C	[26]
		S/W ratio: 1:30
		Reaction time: 10 h
LF slag	56	Temperature: 20 °C	[27]
		S/W ratio: 1:5
		Reaction time: 70 min
BOF steel slag	215	Temperature: 60 °C	[28]
		S/W ratio: 1:5
		Reaction time: 3 h

), the CO₂ uptake of the 80H-slag in this study was considerable and the process was attractive. With an annual (2019–2020) output of LF slag of about 5 MT in China, this waste could, under 80 °C hydrothermal pretreatment, capture about 0.67 MT CO₂ if the mineral carbonation process is applied in steel plants.

4. Conclusions

This study investigated the temperature of hydrothermal pretreatment on the hydrate formation and carbonation efficiency of Al-rich LF slag at ambient temperature and pressure. The main results are as follows:

During hydrothermal pretreatment, cubic C₃AH₆ was a main hydration product for 40 °C, 80 °C, and 100 °C-pretreated slags while C₄ACH₁₁ and C₃AH_x with flaked shapes were the main hydrates for 20 °C and 80 °C-pretreated slags, respectively. Rhombohedral-shaped CaCO₃ was generated by the reaction between C₁₂A₇ in the slag and CO₂ in the air; and then CaCO₃ reacted with C₃AH₆ to form flake-shaped CCA. Flake-shaped products presented higher BET specific surface area. In 40 °C, 60 °C, and 100 °C-pretreated slags, a dense CaCO₃ layer surrounded the unreacted core of the slag particle, resulting in the hindrance of further C₁₂A₇ hydration.

Flake-shaped products could provide a lager reaction surface area and avoid the calcite and AH gel layer formation on the surface of the unreacted hydrates. Therefore, 80 °C and 20 °C-pretreated slags containing a larger number of flake-shaped hydrates had larger CO₂ uptake (13.66 wt% and 10.82 wt%, respectively). Cubic C₃AH₆ crystal and unhydrated raw slag particles were less inactive for carbonation, resulting in the smaller CO₂ uptake for 40 °C, 60 °C, and 100 °C-pretreated slags (9.21 wt%, 9.83 wt% and 6.84 wt%, respectively). In short, the carbonation efficiency of the pretreated slag was strongly associated with the morphology of the hydration products and the hydration extent of LF slag.