2022 年 62 巻 6 号 p. 1116-1125
In this study, the rheological and crystallization behavior of CaO–BaO–Al2O3-based mold slags was investigated through the measurement of the viscosity-temperature relationship and the record of crystallization behavior during the continuous cooling process from 1300 to 600°C applying a modified confocal scanning laser microscopy. Variations of the viscosity, break temperature, initial crystallization temperature and crystallization phases of mold slags with the substitution of SiO2 by Al2O3 and Al2O3 by BaO at a gradient interval of 4 mass% were discussed, and crystallization parameters of average crystallization rate, Ozawa index and effective crystallization activation energy were calculated to explain the crystallization performance. The slag viscosity at 1300°C, the melting and break temperature increased with the substitution of SiO2 (16 to 0 mass%) by Al2O3 (20 to 36 mass%), while those decreased with the substitution of Al2O3 (32 to 16 mass%) by BaO (4 to 20 mass%). With the gradual substitution of SiO2 by Al2O3, the initial crystallization temperature increased from 820 to 1273°C at the cooling rate of 1°C/s, major precipitated phases gradually changed from CaF2 to CaF2, LiAlO2 and BaAl2O4, the average crystallization rate and the Ozawa index fluctuated but had the same tendency. With Al2O3 gradually replaced by BaO, the initial crystallization temperature decreased from 970 to 775°C at the cooling rate of 1°C/s, major precipitated phases changed from CaF2 to CaF2 and BaAl2O4, the crystallization rate of slags was affected by the difference of the nucleation and growth rate of different crystals.
Adequate crystallization in the slag film aids the control of horizontal heat transfer from solidifying shell to water-cooled copper plate, which is helpful to obtain mild cooling and to improve the surface quality of the steel products during the continuous casting process.1,2,3) However, a fully crystalline slag film is not an alternative, as it provides poor lubrication to the solidifying shell.4) This is a particular issue for the continuous casting of high-Al steel,5) due to the interfacial reaction between Al in liquid steel and SiO2 in conventional CaO–SiO2-based mold slag induces the transformation of initial silica-based mold slag into alumina-based mold slag, leading to a significant change of slag properties, such as increases in viscosity and crystallization of the slag, which disturb the lubrication and heat transfer by mold slag in the shell/mold gap.5,6,7,8,9,10,11) Hence, the low-reactivity of CaO–Al2O3-based mold slag was proposed by many researchers12,13,14,15,16,17) through reducing the content of reactive component of SiO2 and then replacing SiO2 with amphoteric oxide components, such as Al2O3, to stabilize the slag performance.
These years, a number of studies regarding the interaction between liquid steel and CaO–Al2O3-based mold slag and the fundamental performance of the slag have been investigated. Yang et al.11,18,19,20) reported that the increase of CaO/Al2O3 ratio from 1 to 3 accelerated the interfacial reaction between high-Al steel and CaO–Al2O3-based mold slag and increased the accumulation of Al2O3 in the slag,11) and the interfacial reaction was also responsible for the dynamic wetting between CaO–Al2O3-based mold slag and high-Al steel.11,18) Also, the content of F increased from 4.0 to 6.9 mass% could effectively reduce the viscosity of CaO–Al2O3-based mold slag,19) which also increased with increasing Al2O3/(B2O3+Na2O) ratio.20) Ou et al.21) proposed that CaO–Al2O3–10 mass%MnO–5 mass%Na2O–5 mass%MgO was promising for the continuous casting of high-Al steel through thermodynamic calculation and interfacial reaction experiment. Deng et al.22) studied the non-Newtonian behaviors of CaO–Al2O3-based mold slag to improve the slag lubrication. Many studies regarding the effect of slag compositions, such as F, Na2O, B2O3, Li2O, CaO/Al2O3, etc. on the crystallization of CaO–Al2O3-based mold slag were also investigated.23,24,25,26) Although casting trials for high-Al TRIP steel indicated that the surface quality of slab was improved, usually less than 3 heats of liquid steel could be continuously cast by applying CaO–Al2O3-based mold slag, and it still exhibited poor lubrication and inadequate consumption, due to the increase of viscosity and crystallization ability of the slag after the steel-slag reaction.15,27)
The reduction of SiO2 and the accumulation of Al2O3 that arose from the steel-slag reaction caused the drastic change of slag crystallization characteristic, which increased the risk of sticker breakout. Hence, it is essential to understand the effect of the substitution of SiO2 by Al2O3 on the crystallization of CaO–Al2O3-based mold slag, and the substitution of Al2O3 by BaO on slag properties was also investigated, with BaO helping to restrain slag crystallization and lower the slag viscosity.28,29,30) Moreover, the effect of the substitution of Al2O3 for SiO2 and BaO for Al2O3 on the rheological and crystallization behavior of the current CaO–BaO–Al2O3-based mold slag was not well understood.
To accurately test the crystallization behavior of CaO–Al2O3-based mold slag, a modified confocal scanning laser microscopy was employed in this study. The confocal scanning laser microscopy was replaced with a digital optical microscope in an infrared furnace (IR-MOP),31) which can observe the entire crystallization process of the mold slag. IR-MOP equipped with an in-house observation system has been proved to have a lower volatilization ratio of the slag than other methods and could accurately measure the crystallization of the volatile mold slag.31) The effect of Al2O3/SiO2 and BaO/Al2O3 ratios on the rheological and crystallization behavior of CaO–Al2O3-based mold slag was investigated at different cooling rates applying this method. The crystallization phases at different cooling rates were analyzed by X-ray diffraction (XRD). The crystallization kinetics parameters of average crystallization rate, Ozawa index, and effective crystallization activation energy were evaluated based on the measured data.
In the current study, two series of CaO–BaO–Al2O3-based mold slags with the substitution of SiO2 by Al2O3 (AS1–AS5) and Al2O3 by BaO (BA1–BA5) are concerned to understand the variation of the rheological and crystallization behavior, whose compositions are listed in Table 1. The substitution of SiO2 by Al2O3 and Al2O3 by BaO was set at a gradient interval of 4 mass%, respectively, to keep the relative content of other components constant. Before the measurement of slag properties, each slag sample of 250 g was prepared using the analytical grade reagents CaCO3, SiO2, Al2O3, BaCO3, Li2CO3, CaF2, MgO, B2O3 and Na2CO3 with the purity higher 99.0 mass%, and pre-melted in a MoSi2 furnace at 1300°C for 20 min to homogenize chemical compositions. Then, the mold slag was cooled, crushed and ground into powder for the further analysis of the melting temperature and the crystallization behavior.
| NO. | CaO | BaO | Al2O3 | SiO2 | F− | MgO | Li2O | Na2O | B2O3 |
|---|---|---|---|---|---|---|---|---|---|
| AS1 | 27 | 12 | 20 | 16 | 12 | 1 | 4 | 4 | 4 |
| AS2 | 27 | 12 | 24 | 12 | 12 | 1 | 4 | 4 | 4 |
| AS3 | 27 | 12 | 28 | 8 | 12 | 1 | 4 | 4 | 4 |
| AS4 | 27 | 12 | 32 | 4 | 12 | 1 | 4 | 4 | 4 |
| AS5 | 27 | 12 | 36 | 0 | 12 | 1 | 4 | 4 | 4 |
| BA1 | 27 | 4 | 32 | 12 | 12 | 1 | 4 | 4 | 4 |
| BA2 | 27 | 8 | 28 | 12 | 12 | 1 | 4 | 4 | 4 |
| BA3 | 27 | 12 | 24 | 12 | 12 | 1 | 4 | 4 | 4 |
| BA4 | 27 | 16 | 20 | 12 | 12 | 1 | 4 | 4 | 4 |
| BA5 | 27 | 20 | 16 | 12 | 12 | 1 | 4 | 4 | 4 |
The measurement of viscosity-temperature relationship was carried out using the rotating cylinder method according to Chinese industrial standards (YB/T 185-2001), and the break temperature (Tb), at which a marked change of viscosity along the viscosity-temperature curve occurred under the influence of the crystal precipitation in the liquid slag, was obtained.32) The melting temperature (Tm) of mold slags was measured through the hemisphere method according to Chinese industrial standards (YB/T 186-2001). The measurement of viscosity and melting temperature of the slag was conducted in the atmospheric condition without introduction of protected gases. To ensure the accuracy of the experiment results, repeat experiments were conducted until the difference values of two experiments were below 0.02 Pa·s for viscosity measurement and were below 5°C for the melting temperature measurement, and then the results of two experiments were averaged as final experiment values. Other details about the measurement of the viscosity and melting temperature could be found elsewhere.12) The melting and break temperature reflected the melting and solidification behavior of mold slags, which affected the depth of the liquid slag pool above the steel and the thickness of the liquid and solid slag film in the shell/mold gap, respectively.
2.3. Crystallization Measurement and XRD AnalysisIn the current study, the crystallization behavior of mold slags in Table 1 was fully observed and recorded in the wide field of view by a digital optical microscope of the infrared furnace (IR-MOP),31) as shown in Fig. 1. During each run of crystallization measurement, mold slag of approximately 80 mg was placed into a platinum crucible with 6.2 mm-inner diameter, heated to 1300°C at a heating rate of 10°C/s, held for 60 s, and then continuous cooling at the cooling rate of 1, 2, 3 and 4°C/s, respectively. These chosen cooling rates of slags were related to the actual cooling rate in the shell/mold gap at some distance below the meniscus in the mold. With the gradual cooling of the slag, the crystallization process were recorded by a camera at 10 frames per second, and the relationship among the time, the temperature and the crystallization fraction X of the slag was obtained. In the current study, the initial crystallization temperature (Tc) of the slag was defined as the temperature at which the ratio of the crystalline area to the total area of the observed field was 5.0%, i.e. the crystalline fraction X was equal to 0.05,31) and the image processing software Photoshop was used to analyze the images and determine the crystallization fraction, as shown in Fig. 2. The crystallization fraction combined with the corresponding crystallization time and temperature was used to construct the continuous cooling transformation (CCT) curve for each mold slag. During the measurement of crystallization temperature, repeat experiments were conducted until the difference values of two experiments were below 5°C. The mass of the post-CCT experimental sample was too small to conduct XRD analysis, so continuous cooling experiments were repeated to obtain sufficient samples for XRD measurements. Then, the X-ray diffraction (XRD) was employed to identify crystalline phases of mold slags in the non-isothermal crystallization process at the cooling rates of 1 and 4°C/s during the crystallization measurement.
Schematic of in-house experimental apparatus for crystallization. (Online version in color.)
Crystallization process of molten slag: (a) molten slag, (b) initial crystallization (X=0.05), (c) crystal growth (X=0.5), and (d) complete crystallization (X=0.95). (Online version in color.)
During the crystallization measurement, the crystallization fraction of the slag along with the time and temperature was obtained, and then the crystallization kinetics parameters, including the average crystallization rate (va), Ozawa index (No), and effective crystallization activation energy (EX), were calculated to compare the crystallization rate, the growth dimension of crystals and the value of energy barrier during the crystallization process with different compositions of mold slags, which could reflect the crystallization behavior and then affect the heat transfer through the slag film in the shell/mold gap. These crystallization kinetics parameters were described below.
The average crystallization rate, va, can be defined as:
| (1) |
The Ozawa index (No), reflecting the growth dimension of crystals, can be obtained from Ozawa equation,33) which is usually applied to describe the non-isothermal crystallization of the melt at a constant cooling rate and is typically written as:
| (2) |
However, Ozawa equation is difficult to obtain a better linear relationship to describe non-isothermal crystallization.34) Mo35) derived a new kinetic equation of non-isothermal crystallization by combining the Ozawa and Avrami equations to better fit experimental results and determine an exact value of NO. It is written as:
| (3) |
| (4) |
The crystallization activation energy (EX) could be calculated through the Friedman equation written as:39)
| (5) |
In the continuous casting of steel, the melting and break temperature of the slag affect the depth of the liquid slag pool and the thickness of the slag film in the mold, while the viscosity of the slag determines the slag consumption and the lubrication of the slag on the shell. Hence, effects of the substitution of SiO2 by Al2O3 and Al2O3 by BaO on the viscosity at 1300°C (η1300°C), melting temperature (Tm), break temperature (Tb) and viscosity–temperature curves of slags are discussed, as shown in Fig. 3. From Fig. 3(a), with the gradual replacement of SiO2 by Al2O3 (Al2O3 increased from 20% to 36%), the η1300°C, Tb, and Tm significantly increased from 0.113 to 0.372 Pa·s, from 977 to 1250°C and from 1014 to 1220°C, respectively, while those of slags were gradually reduced with the gradual replacement of Al2O3 by BaO in Fig. 3(c) from 0.244 to 0.081 Pa·s, from 1140 to 1025°C and from 1143 to 979°C, respectively, which was related to the variation of the network structures of the slag.40,41) The molten slag constituent gradually migrated to the aluminate region with high melting temperature and polymerization degree with the increase of Al2O3 content, which resulted in the increase of slag viscosity.40,41) In addition, the free O2− released by BaO will depolymerize the network structures of the melt and thereby cause a decrease in the slag viscosity with the replacement of Al2O3 by BaO.41) The effect of Al2O3 and BaO on the crystallization of mold slags would be responsible for the variation of the break temperature, which would be introduced in the next text.
Viscosity at 1300°C (η1300°C), melting temperature (Tm), break temperature (Tb) and viscosity–temperature curves of mold slags with the substitution of (a–b) SiO2 by Al2O3, and (c–d) Al2O3 by BaO. (Online version in color.)
From Fig. 3(b), with the gradual replacement of SiO2 by Al2O3, viscosity–temperature curves of slags moved to the high-temperature region and the basic characteristics of slags increased, while viscosity–temperature curves of slags moved to the low-temperature region with the gradual replacement of Al2O3 by BaO in Fig. 3(d), this was related to the variation of crystallization behavior of mold slags. Viscosity–temperature curves of slag BA1 and BA2 with a higher level of Al2O3 contents fluctuated greatly in the high–temperature area, which might be related to the instability of Al–O network structure.42) With the substitution of Al2O3 by BaO, the viscosity–temperature relationship became smoother with the fact that BaO played a role in stabilizing the structure of Al–O network, and the amount of Al–O network structure was reduced within the mold slag.42) Hence, the substitution of SiO2 by Al2O3 and Al2O3 by BaO had an inverse effect on the melting and solidification performance of slags, and Al2O3 substituted by BaO in mold slags could effectively offset the change of slag performance caused by the steel–slag interfacial reactions and the floating inclusions, which was conducive to improve the lubrication behavior of mold slag.
The crystallization of the slag film in the shell/mold gap controls the heat transfer from the solidified shell to the copper plate, which was greatly related to the crystallization temperature, the crystallization phase and crystallization rate. Effects of the substitution of SiO2 by Al2O3 and Al2O3 by BaO on the initial crystallization temperature (Tc) of slags with different cooling rates are shown in Fig. 4. From Fig. 4(a), the initial crystallization temperature Tc increased with the substitution of SiO2 by Al2O3 in mold slags at the cooling rate of 1–4°C/s, indicating that Al2O3 as a replacement for SiO2 enhanced the crystallization ability of CaO–BaO–Al2O3-based mold slags, which caused the increase of break temperature and the viscosity–temperature curves of slags moved to the high-temperature region as shown in Figs. 3(a) and 3(b). The initial crystallization temperature of the same slag changed less at different cooling rates, and that increased from 820 to 1273°C with SiO2 gradually substituted by Al2O3 from AS1 to AS5 at 1°C/s, which has the same variation of the break temperature in Fig. 3(a).
Initial crystallization temperature (Tc) of slags in continuous cooling transformation (CCT) curves with (a) substitution of SiO2 by Al2O3 and (b) substitution of Al2O3 by BaO. (Online version in color.)
From Fig. 4(b), the critical cooling rate of BA5 slag was approximately 1°C/s, beyond which only amorphous phase existed, so the cooling rates of 0.25, 0.5, and 1°C/s were selected to conduct continuous cooling experiments, indicating that the crystallization ability of the BA5 slag was weak. As depicted in Fig. 4(b), the higher the degree of BaO replacing Al2O3 in the mold slags, the lower the initial crystallization temperature Tc, indicating that the substitution of BaO for Al2O3 resulted in the weak crystallization ability of CaO–BaO–Al2O3-based mold slags, which caused the decrease of break temperature and the viscosity–temperature curves of slags moved to the low-temperature region as shown in Figs. 3(c) and 3(d). With BaO gradually replacing Al2O3 from BA1 to BA5 at the cooling rate of 1°C/s, the Tc decreased from 970 to 775°C, which has also the same variation as the break temperature in Fig. 3(c). Meanwhile, the Tc for individual mold slag decreased with the increase of cooling rate, which could be explained that the viscosity of mold slags increased with the decrease of the temperature in Fig. 3(d), followed by an enhanced resistance of molecular and ion migration, so a large degree of undercooling for crystal formation was needed.43)
In Fig. 5, crystalline phases of CaO–BaO–Al2O3-based mold slags are compared with the substitution of SiO2 by Al2O3 and Al2O3 by BaO at different cooling rates, and the precipitation of CaF2 was obtained in AS1–AS5 and BA1–BA5 mold slags at the cooling rates of 1 and 4°C/s. From Fig. 5(a), the major precipitated phase was CaF2 with the substitution of SiO2 by Al2O3 from AS1 to AS3 mold slags, while some other crystalline phases existed as BaAl2O4, LiF and AlF3, Al2SiO5, CaAlFSiO4 and LiAlO2 with the substitution of SiO2 by Al2O3 from AS3 to AS5 mold slags, and major precipitated phases changed from CaF2 to CaF2, LiAlO2 and BaAl2O4, due to the increase of aluminate network structures with increasing Al2O3 content.41) From Fig. 5(b), the major precipitated phase was CaF2 with the substitution of Al2O3 by BaO from BA1 to BA4 mold slags, while major precipitated phases changed from CaF2 to CaF2 and BaAl2O4 in slags from BA4 to BA5, indicating that BaO replaced Al2O3 inhibited the precipitation of CaF2 and prompted BaAl2O4 crystalline phases. The free O2− released by BaO in the molten slag resulted in the increase of the number of available network modifiers with the substitution of Al2O3 by BaO,41) which depolymerized the network structures of aluminate and promoted the combinations of Ba2+ and [Al2O4]2−. Hence, major crystalline phases in current mold slags existed as CaF2, LiAlO2 and BaAl2O4, and the initial crystallization temperature was affected by the crystalline phases of slags and the cooling rates.
Effect of (a) Al2O3 substituting SiO2 and (b) BaO substituting Al2O3 on crystalline phases of CaO–BaO–Al2O3-based mold slags at the cooling rate of 0.25, 1 and 4°C/s. (Online version in color.)
Based on the Mo equation35) (Eq. (3)), plots of lgβ vs lgt for various crystallization fractions in the first series are presented in Fig. 6. It was evident from the plots in Fig. 6 that Mo equation could well fit experimental data, and the values of the kinetics parameter m that estimated from the slopes of linear fits are summarized in Table 2. Based on the Avrami equation36,37,38) (Eq. (4)), plots of lg[–ln(1–X)] vs lgt were constructed using the non-isothermal crystallization data at various cooling rates, as shown in Fig. 7. The crystallization ratio of 0.3 was taken as the critical point for sectionalized fitting. Values for the kinetics parameter NA were estimated from the slopes of linear fits for both stages of crystallization, and different values of NA were found in the ranges of 0.1≤X≤0.3 and 0.4≤X≤0.9. Mean values of NA under four different cooling rates were taken, namely, the sectionalized NA was obtained, as listed in Table 3.
Relationships between lgβ and lgt for non-isothermal crystallization of mold slags (a) AS1, (b) AS2, (c) AS3, (d) AS4, and (e) AS5. (Online version in color.)
| X | 0.1 | 0.2 | 0.3 | 0.4 | 0.5 | 0.6 | 0.7 | 0.8 | 0.9 | |
|---|---|---|---|---|---|---|---|---|---|---|
| AS1 | m | 1.06 | 0.86 | 0.79 | 0.90 | 1.11 | 1.32 | 1.46 | 1.56 | 1.64 |
| R2 | 0.99 | 1.00 | 1.00 | 1.00 | 0.99 | 0.99 | 0.99 | 1.00 | 1.00 | |
| AS2 | m | 1.66 | 1.41 | 1.23 | 1.29 | 1.26 | 1.21 | 1.22 | 1.36 | 1.24 |
| R2 | 0.94 | 0.99 | 0.98 | 0.98 | 0.98 | 0.98 | 0.98 | 0.98 | 1.00 | |
| AS3 | m | 0.57 | 0.69 | 0.77 | 0.81 | 0.81 | 0.82 | 0.85 | 0.89 | 0.94 |
| R2 | 0.92 | 1.00 | 0.99 | 0.98 | 0.99 | 0.99 | 1.00 | 1.00 | 0.99 | |
| AS4 | m | 0.67 | 0.69 | 0.68 | 0.67 | 0.67 | 0.68 | 0.73 | 0.80 | 0.95 |
| R2 | 0.99 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 1.00 | 0.91 | |
| AS5 | m | 0.54 | 0.56 | 0.59 | 0.63 | 0.68 | 0.75 | 0.82 | 0.89 | 0.93 |
| R2 | 1.00 | 1.00 | 0.99 | 0.99 | 0.99 | 0.99 | 1.00 | 1.00 | 1.00 | |
Relationships between lg[–ln(1–X)] and lgt for non-isothermal crystallization of mold slags (a) AS1, (b) AS2, (c) AS3, (d) AS4, and (e) AS5 at different cooling rates. (Online version in color.)
| X | AS1 | AS2 | AS3 | AS4 | AS5 | |
|---|---|---|---|---|---|---|
| NA | 0.1–0.3 | 1.24 | 1.23 | 1.17 | 0.97 | 1.01 |
| 0.4–0.9 | 3.24 | 3.78 | 1.81 | 2.46 | 2.03 |
The values of m and NA in Tables 2 and 3 were used to calculate Ozawa index No from m=NA/NO, and the Ozawa index NO of each mold slag with nine different crystallization ratios were averaged as NO’, which was compared with the average crystallization rate va that calculated from Eq. (1), as shown in Fig. 8(a). The larger Ozawa index is, the larger the growth dimension and then the higher the growth rate should be. The effects of BaO replacing Al2O3 on NO’ and va are shown in Fig. 8(b). Crystal was not precipitated at a cooling rate of 4°C/s for BA5 in Fig. 4(b), so only four data of va at 4°C/s are shown in Fig. 8(b).
Relationship between NO’ and va of mold slags at the cooling rates of 1 and 4°C/s with the substitution of (a) SiO2 by Al2O3 and (b) Al2O3 by BaO. (Online version in color.)
From Fig. 8(a), it could be seen that the changes of NO’ and va were parallel with the substitution of Al2O3 for SiO2, and the larger the NO’ was, the greater the va would be, indicating that the large growth dimension of crystal resulted in large crystallization rate of AS1–AS5 slags. From Fig. 8(b), with the increase of BaO and decrease of Al2O3, the average crystallization rate va increased and then decreased, and reached the peak with the BaO content of 12 mass% and the Al2O3 content of 24 mass%. However, the changes in NO’ and va were opposite when BaO content increased from 4 to 8 mass% (from BA1 to BA2 slags). The va was affected by both the growth rate (related to the NO’) and nucleation rate of crystals, and the difference in changes of NO’ and va indicated that the nucleation rate of BA2 slag was faster than that of BA1 slag. Figure 8 indicates that the NO’ index could affect the va, but there was no obvious linear relationship between the two parameters due to the nucleation and growth rates discrepancy of the precipitated crystals in Fig. 5.44) The average crystallization rate va of AS1–AS5 slags was larger than that of BA1–BA5 slags, as a whole. The substitution of SiO2 by Al2O3 resulted in a slight fluctuation of NO’ in the range of 2.2±0.2, except for the 2.7 of NO’ for AS4, and the substitution of Al2O3 by BaO resulted in a drastic reduction from 2.7 of BA1 to 0.7 of BA5. Hence, the growth dimension of crystals was greatly related to the crystalline phases in Fig. 5 and the variation of crystallization temperature in Fig. 4, and then affected the crystallization rate.
The transient crystallization rate (dX/dt) of each mold slag can be calculated at different crystallization fractions when keep the cooling rate constant. According to the Friedman equation39) (Eq. (5)), plots of ln(dX/dt) vs 1/T can be constructed in Fig. 9 based on the transient crystallization rate and crystallization temperature with different cooling rates, and the effective crystallization activation energy EX of each mold slag can be derived from the slope of the linear fits as a function of crystallization fraction, as shown in Fig. 10. The EX at different crystallization fractions were averaged as the average effective crystallization activation energy EX’, which can be considered as an energy barrier in the crystallization process. Zhou et al.45) reported that EX’ was negatively correlated with initial crystallization temperature Tc on the precipitation of cuspidine in CaO–SiO2-based mold slag. The relationship between the Tc at a cooling rate of 1°C/s and the EX’ of CaO–Al2O3–BaO-based mold slags with the substitution of SiO2 by Al2O3 and Al2O3 by BaO was shown in Fig. 11, and the negative correlation relationship between Tc and EX’ was also obtained. From Fig. 11(a), with the gradual replacement of SiO2 by Al2O3, the effective crystallization activation energy EX’ of mold slags significantly declined, while that of mold slags increased with the increased replacement of Al2O3 by BaO in Fig. 11(b). Hence, it also verified that the crystallization ability of the current CaO–BaO–Al2O3-based mold slags increased with the substitution of SiO2 by Al2O3 and decreased with the substitution of Al2O3 by BaO.
Relationships between ln(dX/dt) and 1/T for non-isothermal crystallization of mold slags (a) AS1, (b) AS2, (c) AS3, (d) AS4, and (e) AS5 at different crystallization fractions. (Online version in color.)
The effective crystallization activation energy EX at different crystallization fractions of mold slags (a) AS1–AS5, and (b) BA1–BA5. (Online version in color.)
Relationship between effective crystallization activation energy EX’ and initial crystallization temperature Tc of mold slags with the substitution of (a) SiO2 by Al2O3 and (b) Al2O3 by BaO. (Online version in color.)
In the current study, the rheological and crystallization behavior of CaO–BaO–Al2O3-based mold slags were investigated through the measurement of the viscosity, break temperature, crystallization temperature, crystallization phases and the calculation of crystallization parameters. Conclusions were drawn below.
(1) With the replacement of SiO2 by Al2O3 in current CaO–BaO–Al2O3-based mold slags, the viscosity at 1300°C, the melting and breaking temperature increased, while those of slags decreased with the replacement of Al2O3 by BaO, which was related to the variation of the network structures and crystallization of mold slags.
(2) The initial crystallization temperature of the same slag changed less at different cooling rates, and that increased from 820 to 1273°C with SiO2 substituted by Al2O3 at 1°C/s, indicating that Al2O3 as a replacement for SiO2 enhanced the crystallization ability of CaO–BaO–Al2O3-based mold slag. The substitution of BaO for Al2O3 resulted in the weak crystallization ability of slags, and the initial crystallization temperature decreased from 970 to 775°C with the substitution of Al2O3 by BaO at 1°C/s. The effective crystallization activation energy was negatively correlated with the initial crystallization temperature for current CaO–BaO–Al2O3-based mold slag.
(3) With the substitution of SiO2 by Al2O3, major precipitated phases gradually changed from CaF2 to CaF2, LiAlO2 and BaAl2O4, and major precipitated phases changed from CaF2 to CaF2 and BaAl2O4 with the substitution of Al2O3 by BaO. The crystallization rate was related to the Ozawa index, but there was no obvious linear relationship between the two parameters due to the discrepancy of the precipitated crystals.
The authors are grateful for support from the National Natural Science Foundation China (Grant No. U20A20270, No. 52004045 and No. 52074054), the Postdoctoral Foundation of Chongqing Natural Science Foundation (Grant No. cstc2020jcyj-bshX0003), Science and Technology Key Project of Panxi Experimental Area and College of Materials Science and Engineering and Chongqing Key Laboratory of Vanadium–Titanium Metallurgy and Advanced Materials at Chongqing University, China.
