Effect of Lime on Stability of Chromium in Stainless Steel Slag

Abstract

The stabilization of chromium is of great importance to the use of stainless steel slag (SSS), and the influence of lime on the stability of chromium is currently unclear. In this work, the variation of phase transformation and chromium distribution with slag basicity (CaO/SiO₂) were investigated experimentally, and the leaching ability of chromium was evaluated. Results showed that chromium-bearing phases were glass, dicalcium silicate (C₂S), spinel, and periclase, while the degree of enrichment of chromium in these phases was found to be closely related to the basicity. The optimal basicity obtained in this research was 1.5, with the chromium mainly present in the stable spinel and exhibiting the lowest leaching ability. The product layer structure of unmelted lime was studied as well, showing a periclase layer and a Ca₂SiO₄ layer. Some CaCr₂O₄ had formed in the periclase layer, which is potentially hazardous for the environment and living organisms.

1. Introduction

A large amount of chromium-bearing stainless steel slag (SSS) is generated during stainless steel production, and discharging this into the environment gives severe pollution issues [1]. Some chromium exists in unstable phases, like dicalcium silicate (Ca₂SiO₄, nominated as C₂S in this study), merwinite (Ca₃MgSi₂O₈), melilite (Ca₂Mg₂SiO₇), gehlenite (Ca₂Al₂SiO₇), and periclase (MgO). Cr²⁺ and Cr³⁺ can be released after phase decomposition, followed by oxidation to hexavalent chromium (Cr⁶⁺). The latter is known to be a carcinogen, and a dermatological and pulmonary sensitizer that is 500 times more toxic than Cr³⁺ [2,3,4]. In order to reduce the harmful effects of chromium, a lot of effort has been focused on controlling chromium distribution across the various phases. Our previous work has shown that chromium-bearing spinel does not notably corrode, even in an extremely acidic solution (80 wt % sulfuric acid, 160 °C) [5]. Therefore, the method of enriching the chromium into a stable spinel phase is attracting extensive attention [6,7,8], and a molten modification using spinel forming agents was reported to improve spinel growth and restrict chromium leaching [9].

For the purposes of dephosphorization and desulfurization, a certain amount of lime must be added to the molten slag [10]. The basicity of slag (the mass ratio of basic components to acidic components) increases with lime fusing, which causes a variation of elemental distributions in various phases [11,12,13]. However, there are discrepancies in the literature regarding the influence of the basicity of slag on the chromium migration behavior in the SSS. Mostafaee and Jönsson found that the basicity had almost no effect on the amount of spinel phase in slag [14], while some other researchers hold different opinions. Engström et al. reported basicity to have a strong impact on the chromium distribution in the SSS [15], and Albertsson et al. pointed out that a high basicity (>1.4 in their study) should be avoided, since it can lead to the formation of leachable chromium-bearing solid solutions [16]. The optimal value of basicity for enriching chromium in the spinel phase, as obtained from various research works, is closely related to the slag composition the researchers used [17,18]. Wang and Seetharaman investigated the oxidation state of chromium oxides in CaO–SiO₂–CrO_x and CaO–MgO–(FeO–)Al₂O₃–SiO₂–CrO_x using X-ray absorption near edge structure (XANES) analysis, and found that the mass ratio X_CrO/X_CrO1.5 increased with decreasing slag basicity, while, beyond a certain level, the ratio was unaffected by basicity [19].

Another noteworthy issue is that the lime readily reacts with SiO₂, forming a C₂S product layer at the interface, which could prevent the lime from dissolving into the molten slag [20,21]. The unmelted lime in the SSS would expand and blast when it contacts water [22,23]. As a consequence, some chromium-bearing powders erupt from the slag and after transfer by air, endangering a larger area. However, the details of the structure around the unmelted lime are poorly understood, and the possible pollution caused by this is, therefore, difficult to quantify.

In this paper, in order to clarify the effect of lime (melted, melting, or solid) on the stability of chromium in SSS, the variation of phase transformation and chromium distribution with basicity as well as the leaching ability of chromium were experimentally investigated. Moreover, the microstructure of the product layer around the unmelted lime was studied.

2. Materials and Methods

2.1. Materials

Based on a typical composition of SSS from a stainless steel plant, experimental slags with different basicities (the mass ratio of CaO to SiO₂, B) of 1.0 (S1), 1.5 (S2), and 2.0 (S3) were prepared by chemical reagents calcium oxide (CaO ≥ 98.0 wt %), silicon oxide (SiO₂ ≥ 99.0 wt %), magnesium oxide (MgO ≥ 99.9 wt %), aluminum oxide (Al₂O₃ ≥ 99.0 wt %), ferrous oxalate (FeC₂O₄·2H₂O ≥ 99.9 wt %), chromium oxide (Cr₂O₃ ≥ 99.0 wt %), and calcium fluoride (CaF₂ ≥ 98.5 wt %), respectively. All reagents were purchased from Sinopharm Chemical Reagent Co., Ltd., Shenyang, China, and were dried at 110 °C for 10 h before weighing and mixing. One sample (S4) without Cr₂O₃ was also prepared for a control experiment with unmelted lime. The chemical compositions of the target SSS samples (S1–S4) are shown in

Table 1. Chemical composition of the target stainless steel slag, wt % and index of basicity, B.

No.	CaO	SiO2	MgO	Al2O3	FeO	Cr2O3	CaF2	B
S1	38.0	38.0	9.0	4.0	3.0	5.0	3.0	1.0
S2	45.6	30.4	9.0	4.0	3.0	5.0	3.0	1.5
S3	50.6	25.4	9.0	4.0	3.0	5.0	3.0	2.0
S4	48.6	32.4	9.0	4.0	3.0	0	3.0	1.5

. Some CaO lumps (6 mm × 6 mm × 8 mm) were prepared from CaO powder by briquetting and polishing treatments and used as unmelted lime.

2.2. Methods

The experimental furnace employed in this study is equipped with a MoSi₂ heater and is controlled by a PID controller equipped with a PtRh30%/PtRh6% thermocouple sensor, which is illustrated in Figure 1. The temperature deviation at the constant temperature zone is less than ±2 °C.

Slag (10 g) was poured into a molybdenum crucible inside a graphite crucible. The molybdenum crucible is used to protect the graphite crucible and for easy sampling. The graphite crucible was placed in the constant-temperature zone of the furnace. High purity argon was injected from the bottom of the furnace at a flow rate of 0.5 L·min⁻¹ to protect the samples from oxidation. The slags were heated to 1600 °C at a rate of 3 °C·min⁻¹, and held for 30 min at 1600 °C to achieve a fully molten state, after which the molybdenum crucible was taken out of the furnace followed by cooling. Two different cooling treatment routes were adopted: (1) direct quenching in water at 1600 °C after incubation for 30 min; (2) incubation for 30 min, cooled to 1300 °C at a cooling rate of 3 °C/min and then incubation for 30 min before quenching with water. Theoretical calculations using FactSage (developed jointly between Thermfact/CRCT and GTT-Technologies, Version 7.0, Montreal, Canada and Aachen, Germany) and pilot experiments both showed that the phase transformation and chromium migration hardly change below 1300 °C.

Moreover, 30 g of samples S2 and S4 were used for the unmelted lime study. When the slag completely molten, the as-prepared CaO lump was immersed into the melt and held in place for 20 min at 1600 °C, after which the crucible was taken out of the furnace and cooled under the air atmosphere. Polishing and gold spraying processes were conducted on the samples, to give them conductivity for scanning electric microscope-energy dispersive spectroscopy (SEM-EDS) analysis.

SEM-EDS analysis was conducted using the Hitachi S3400N scanning electron microscope (HITACHI, Tokyo, Japan) equipped with an INCA 3294 EDS system (HITACHI, Tokyo, Japan) to investigate the microstructure and chemical composition of the precipitated phases. The law of conservation of mass and linear least sum of squares methods were adopted to calculate the composition of the precipitated phase, and the distribution of chromium across the various phases was evaluated by calculating an enrichment degree of chromium (E_Cr,i). The E_Cr,i is defined as

$$E_{\mathrm{Cr},i}=\frac{{(\%\mathrm{Cr})}_i{X}_i^{*}}{{\displaystyle \sum _{i=1}^n{(\%\mathrm{Cr})}_i{X}_i^{*}}}\times 100\%,$$

(1)

where i is a specific phase; (%Cr)_i is the chromium content in a specific phase, wt %; and $X_i^{*}$ is the normalized mass fraction of a specific phase.

The phase composition of each sample was analyzed using Philips X’pert X-ray diffraction (XRD) with a Cu Kα source (λ = 1.5418 Å) over the range 2θ = 10–80°, at a step of 0.008°, and specified by Crystallographica Search-Match (CSM) software (Oxford Cryosystems Ltd., Version 3.1.0.2, Oxford, UK) with the Powder Diffraction File (PDF) databases from the International Centre for Diffraction Data (ICDD). A standard leaching test was conducted according to the Standard of Environmental Protection Industry of the People’s Republic of China HJ/T 299-2007, to evaluate the leachability of chromium: for this, 10 g of each slag was crushed to below 74 μm and leached with 100 mL of acid solution (mass ratio of H₂SO₄ and HNO₃ was 2:1) at a pH value of 3.2 for 18 h at 23 ± 2 °C. Each experiment was repeated three times. The leached amount of chromium was determined using induced coupled plasma-optical emission spectroscopy (ICP-OES, Thermo Fisher Scientific, Waltham, MA, USA) with a detection limit of 0.01 mg·L⁻¹.

3. Results and Discussion

3.1. Phase Transformation

The effect of basicity (1~2) on the phase transformation behavior of SSS was studied in the temperature range from 1600 to 1300 °C at a cooling rate of 3 °C/min. XRD patterns of quenched samples are shown in Figure 2 and Figure 3. The phase composition showed different results with the variation of basicity.

It can be clearly seen in Figure 3a,d that sample S1 (B = 1.0) at 1600 °C was a in a glassy state, and two phases of isomorphous spinel (Mg(Al,Cr)₂O₄) and melilite (Ca₂(Al,Mg)[(Si,Al)SiO₇]) precipitated from the glass after cooling to 1300 °C (cf. Figure 2a). The particle size of spinel was measured to be of the order 10–30 micron, while the precipitation amount was small.

As for sample S2 (B = 1.5) (cf. Figure 3b,e), spinel and chromium-bearing C₂S were found at 1600 °C. The isothermal section diagram of CaO–SiO₂–MgO–4%Al₂O₃ slag system at 1600 °C was calculated using FactSage 7.0, and compositions of the S1, S2, and S3, respectively, were marked with “*”, as shown in Figure 4. It can be seen that, with the increase of basicity from 1.0 to 1.5, the equilibrium phases change, and C₂S could precipitate from the liquid. Moreover, increasing basicity was reported to have a positive effect on the activity coefficient of the MgO and CrOx, which leads to a decrease of spinel solubility in the liquid [24,25]. This is presumably the reason for the precipitation of spinel in S2 at 1600 °C. Furthermore, the C₂S disappeared, and a new phase of chromium-free merwinite precipitated when the temperature decreased to 1300 °C. This suggests that the growth of spinel crystal during the cooling process requires a Cr₂O₃ feed, which causes a driving force for chromium transport from other phases to the spinel. Moreover, the solid solubility of Cr₂O₃ in the precipitated silicate phases decreases with temperature, and eliminated Cr₂O₃ from the silicate phases.

As shown in Figure 2c and Figure 3c,f, besides the phases of the spinel, glass, and C₂S, a periclase phase formed in sample S3 (B = 2.0). Figure 4 shows the composition of S3 to be located in the phase equilibrium zone of liquid + MgO + C₂S. Therefore, the precipitation of periclase is explained by thermodynamics when the basicity is 2.0. Furthermore, chromium was only detected in the spinel and with periclase, while other phases of glass and C₂S were determined to be chromium-free. The explanation for this phenomenon can be found in the binary phase diagram of MgO–Cr₂O₃, showing that Cr₂O₃ can dissolve into MgO at high temperature [26]. No change in phase composition of S3 was observed after the cooling process from 1600 to 1300 °C.

3.2. Distribution and Stability of Chromium

The enrichment degree of chromium in each phase of the various samples was calculated and illustrated in the form of pie charts, as shown in Figure 5. Furthermore, a batch of standard leaching tests was conducted for these samples to evaluate the leaching ability of chromium, and the corresponding results are also provided in Figure 5.

Results indicated that, at 1600 °C (cf. the left-hand side of Figure 5), when B = 1.0, all of the chromium existed in the glass phase, and when the basicity increased to 1.5, about 54.3% and 3.1% of chromium were in the spinel phase and C₂S, respectively. The migration of chromium from glass to spinel phase caused a decrease in the leached amount of chromium, from 2.82 to 2.26 mg·L⁻¹. When B = 2.0, about 61.7% of the chromium existed in the spinel phase and the remainder dissolved in the periclase phase. The leached amount of chromium reached 3.68 mg·L⁻¹ as a result of the low stability of the periclase phase in acid solution. When the samples were cooled to 1300 °C (cf. the right hand side of the Figure 5), more chromium migrated into the spinel phase: the E_Cr,S of S1, S2, and S3 were 51.1%, 91.2%, and 86.0%, respectively. Moreover, the lowest leached amount of chromium was 0.62 mg·L⁻¹, obtained when the basicity was 1.5. On the basis of the findings of this work, it can be proposed that the CaO content in SSS has a significant effect on the distribution and stability of chromium, and the optimal basicity was 1.5 from the viewpoint of chromium retention.

3.3. Product Layer Structure of Unmelted Lime

To study the effects of unmelted lime on the stability of chromium in SSS, CaO lumps were immersed into the molten samples S2 (5% Cr₂O₃, B = 1.5) and S4 (Cr₂O₃ free, B = 1.5), and held in place for 20 min at 1600 °C. XRD analysis was then carried out for investigating the phase composition of the S2 and S4 treated by the CaO lump, giving the results shown in Figure 6. The C₂S, merwinite, spinel, periclase phases were identified in both samples, and the diffraction peaks of CaCr₂O₄ was detected in the XRD pattern of Cr₂O₃-bearing S2.

Figure 7 shows the SEM images of the interface zone between the CaO lump and the slag. As shown in Figure 7a, four phases can be identified in for sample S4. The phase at lower left-hand corner is the unmelted CaO lump, while the top right-hand corner shows the slag containing precipitated spinel, merwinite, and C₂S phase. A 10–30 micron thick product layer was formed between the two phases, which consisted of a periclase layer (next to the CaO lump) and a C₂S layer (next to the slag). Based on the analysis of the Figure 4, it can be inferred that some CaO gradually dissolved into the molten slag with time, causing an increase in basicity around the CaO lump, which resulted in the precipitation of MgO and C₂S at the interface zone between the CaO lump and slag.

Besides the findings previously obtained, a clear crack in the periclase layer was observed in S2 (cf. Figure 7b,c). A needle-shaped phase was detected at the fracture surface which, using EDS analysis, was determined to be the CaCr₂O₄ [27]. The binary phase diagram of CaO–Cr₂O₃ is provided in Figure 8a, giving evidence for the possibility of CaCr₂O₄ precipitation around the CaO lump. An Eh–pH diagram of Ca–Cr–H₂O at 25 °C was calculated using FactSage 7.0, and given in Figure 8b, showing that CaCr₂O₄ is unstable in an acid solution and an oxidizing environment. The Cr³⁺ can be released from the slag and dispersed by water, and a toxic chromium chromate could be formed that causes serious pollution issues. Moreover, it was found that the periclase phase contained some chromium, while the C₂S was chromium-free. The sample S2 treated with a CaO lump was crushed and milled into powder, and a leaching test was conducted to determine the chromium stability. Results indicated that the chromium concentration in the leachate was about 3.73 mg·L⁻¹ chromium, which is much higher than that of the untreated S2 (cf. Figure 5).

It is noteworthy that the unmelted lime could react with water and produce Ca(OH)₂, which can result in volume expansion and pulverization of slag, which aggravates the possible chromium pollution. As stated above, the unmelted lime has an adverse impact on the security of SSS application.

4. Conclusions

The effect of lime on the stability of chromium in stainless steel slag (SSS) was investigated in this work. Experimental results showed that glass, dicalcium silicate (C₂S), spinel, and periclase are the chromium-bearing phases present in SSS, and that the enrichment degree of chromium in these phases is closely related to the basicity. The optimal basicity was found to be 1.5, at which the chromium mainly existed in the stable spinel phase, and exhibited the lowest leaching amount: 2.26 and 0.62 mg·L⁻¹ for 1600 and 1300 °C, respectively. Moreover, the details of the product layer structure of unmelted lime were studied. It was found that the product layer consisted of a periclase layer (next to the CaO lump) and a C₂S layer (next to the slag), while some CaCr₂O₄ was detected in the periclase layer, which poses a threat to the environment and living organisms. Therefore, the amount and particle size of lime for slagging need to be controlled in view of the chromium retention.