ISIJ International
Online ISSN : 1347-5460
Print ISSN : 0915-1559
ISSN-L : 0915-1559
Regular Article
Effect of TiO2 Addition on Crystallization Characteristics of CaO-Al2O3-based Mould Fluxes for High Al Steel Casting
Jiangling LiQifeng ShuXinmei HouKuochih Chou
Author information
JOURNALS OPEN ACCESS FULL-TEXT HTML

2015 Volume 55 Issue 4 Pages 830-836

Details
Abstract

Crystallization behaviors of new developed CaO–Al2O3 based mould fluxes with TiO2 addition for casting of high-Al steels were investigated by using DTA techniques combined with SEM-EDS and XRD analysis. XRD and SEM analyzed on the crystallized samples showed that the sequence of crystal precipitation for TiO2-free mould flux during cooling was MgO, and followed by Ca12Al14O33 during cooling. The sequence of crystal formation for TiO2-bearing mould fluxes during cooling is CaTiO3 to MgO, and then Ca12Al14O33. Continuous cooling transformation diagrams (CCT) were constructed for analysis of the crystallization behaviors. The Undercooling values for onset crystallization of various crystals were calculated by using liquidus temperature obtained by heating DTA and crystallization temperature of various crystals. The crystallization temperatures of CaO–Al2O3 based mould fluxes increased with increasing TiO2 content. The undercooling values for onset crystallization of CaTiO3 decreased with increasing TiO2 content, which indicating that the crystallization of CaTiO3 was enhanced with increasing TiO2 content. The undercooling values for onset crystallizations of Ca12Al14O33 and MgO only changed slightly with increasing TiO2. This indicated that crystallizations of Ca12Al14O33 and MgO crystals were only slightly influenced by TiO2 addition. The overall crystallization of mould fluxes was enhanced with increasing TiO2 content.

1. Introduction

High Alumium steel (e.g. high Al-TRIP and TWIP steel) has been received increasingly attention due to their excellent combination of high strength and superior formability.1) When casting high aluminum steel, due to the reaction between aluminum in steel and silica in the traditional lime-silica based mould flux as shown in Eq. (1), it is a great challenge to achieve good casting performance.   

4[Al]+ 3SiO 2 = 2Al 2 O 3 +3[Si] (1)
The dynamically change of chemistry for the mould flux due to reaction (1) leads to varied viscosity, solidification temperature and crystallization behaviours during the casting. As a result, it would deteriorate both the casting operation and the surface quality of cast slabs, and cause many problems such as increased crack frequency, non-uniform heat transfer across the mould flux, reduced lubrication and so on.2,3,4)

In order to solve the problem on chemical composition change of mould fluxes and develop optimal mould flux to meet the requirements of the high aluminum steel casting, previous researchers have proposed the nonreactive CaO–Al2O3 based mould flux to substitute for conventional lime-silica-based mould fluxes.2,5,6,7,8,9,10,11) Blazek et al.7) developed lime-alumina-based mould flux for casting high aluminum TRIP steel. The interaction between flux and steel was markedly reduced and the as-cast slab quality was improved compared with lime-silica-based mould flux. However, low consumption of mould fluxes and poor lubrication problems have not been solved effectively. Cho et al.5) found that lime-alumina based mould fluxes could suppress the occurrence of surface depressions and cracks by stabilizing the mould heat transfer at the initial stage of casting, but the slag film of lime-alumina mould fluxes were prone to crystallize easily during casting, which made the mould lubrication deteriorate rapidly. Jung and Sohn12) founded that the crystallization temperature trend to decrease with the increasing ratio of CaO/Al2O3 on CaO–Al2O3 based mould flux. Other researchers have suggested some additives such as B2O3, MnO13,14,15) to adjust the properties of mould fluxes to meet the demand for casting of high Al steel. However, since stability of B2O3 and MnO are weaker than SiO2,16) addition of B2O3 and MnO could enhance the possibility of reaction between mould fluxes and molten steel. It can be seen that there are still some problems existing, for example strong crystallization exists and causes poor lubrication between steel shell and the copper mould, which will in turn deteriorate the casting process as well as the surface quality of cast slabs. Accordingly, the study on physiochemical properties of the CaO–Al2O3 based system is still limited to meet the demand for development of optimal mould fluxes for casting high Al steel.

Viscosity and structure of new CaO–Al2O3 based mould fluxes with TiO2 addition was investigated in our previous work for casting of high-Al steel.6) As the stability of TiO2 is stronger than SiO2, the reaction between mould fluxes and steel could be reduced by using such a new mould flux. It was found that the addition of TiO2 would decrease the viscosity of mould fluxes. However, crystallization behaviors of CaO-Al2O3-based mould fluxes with TiO2 addition remains unclear. To evaluate crystallization behaviors of the of CaO–Al2O3 based mould fluxes with addition of TiO2, it is necessary to study the crystallization behaviors of CaO-Al2O3-based mould fluxes with varying TiO2 addition.

Crystallization of mould fluxes could be investigated by employing many techniques, such as differential thermal analysis (DTA)17) or differential scanning calorimeter (DSC),18) Single/Double Hot thermocouple technique (SHTT/DHTT)19,20) and laser confocal scanning microscopy (LCSM).21) In the present work, crystallization behaviors of CaO-Al2O3-based mould fluxes with varying TiO2 content were investigated by the differential thermal analysis (DTA). Continuous cooling transformation (CCT) diagrams of mould fluxes were constructed. The crystalline phases in the mould fluxes were identified by X-Ray diffraction (XRD) and the scanning electron microscope equipped with energy dispersive X-ray spectroscopy (SEM-EDS). Morphologies of crystalline phases were also investigated by using SEM-EDS.

2. Experimental

2.1. Sample Preparation

The raw materials of the mould flux were analytical grade TiO2, Al2O3, CaF2, MgO, Na2CO3, Li2CO3 and CaCO3 (Purity of all reagent>99.5%) were taken as raw materials, with Na2CO3, Li2CO3 and CaCO3 being substitutes for Na2O, Li2O and CaO due to their stability in air, respectively. The CaCO3 were calcined at 1323 K to obtain CaO in a muffle furnace for 10 hours. TiO2, CaF2 and MgO powders were also calcined at 773 K to remove moisture. High temperature melts which were melted at 1773 K were quenched into water to form the glasses. After quenching, the samples were subject to X-ray fluorescence (XRF) (XRF-1800 from Shimadzu) to determine the composition. Nominal Compositions of samples and composition of after the melting which is analyzed by XRF are listed in Table 1.

Table 1. Chemical composition of the studied slag system.
Sample No.Composition (mass %)
CaOAl2O3TiO2CaF2Na2OLi2OMgO
No. 1Nominal35.035.00.010.48.65.55.5
Analyzed39.229.40.013.56.35.53.0
No. 2Nominal33.233.25.010.28.25.15.1
Analyzed41.031.25.112.96.85.12.9
No. 3Nominal32.532.57.010.08.05.05.0
Analyzed35.227.67.513.36.55.02.9
No. 4Nominal31.531.510.09.77.74.84.8
Analyzed36.127.510.012.96.74.82.9

2.2. DTA Measurement

After carefully weighed, the powder mixture was ground in an agate mortar with ethanol as mixing media to ensure the homogenization. Then, the samples were put in a Pt crucible. Differential thermal analysis (DTA) were performed on the samples at the range of 373–1773 K using Netzsch STA449C TG-DTA calorimeter in Ar atmosphere at a flow rate of 45 ml/min. About 20 mg of sample powder was heated at a constant heating rate of 10 K/min from room temperature to 1773 K in a platinum crucible with a diameter of 5 mm and a height of 5 mm to homogenize its chemical composition. Subsequently, the melt sample was cooled at different cooling rate (10 K/min, 15 K/min, 20 K/min, 25 K/min). Figure 1 showed the temperature history in DTA measurement. α-Al2O3 was used as a reference material. Temperature was calibrated using the melting points of the high purity substance and an internal calibration file was established.

Fig. 1.

Thermal history in non-isothermal DTA measurement.

2.3. The Crystalline Phases Analysis

The powdered samples were melted in a platinum crucible of 16 mm inner diameter at 1773 K for 5 h to obtain a homogenous molten flux and subsequently cooled in the furnace to the specified temperature, and then held at specified temperature for 2 h. Afterwards, the samples were quenched by water. The crystalline phases in samples were identified by X-Ray diffraction (XRD) and the scanning electron microscope equipped with energy dispersive X-ray spectroscopy (SEM-EDS). Powder X-ray diffraction measurements were carried out on a M21X-SRA X-ray diffractometer (MACScience) equipped with graphite crystal monochromator in air. Powder X-ray diffraction measurements were carried out on a M21X-SRA X-ray diffractometer (MAC Science) equipped with graphite crystal monchromator in air. The XRD patterns were collected with Cu-Kα radiation. SEM examinations were carried out using ZEISS EVO MA18 equipped with an EDS. The working voltage was 25 kV.

3. Results and Discussion

3.1. Crystallization Behaviors of Mould Fluxes Investigated by DTA Measurement

The crystallization behaviors of the studied mould fluxes were investigated by employing non-isothermal DTA at various cooling rates. Figures 3(a)–3(d) showed the DTA curves of the mould flux at four different cooling rates with 10, 15, 20 and 25 K/min, respectively. It was observed that there were three exothermic peaks on DTA curves for samples with 5, 7 and 10% TiO2, and only two exothermic peaks for the TiO2-free sample at each cooling rate, indicating the presence of three successive crystallization events for the TiO2-containing samples and two successive crystallization events for the TiO2-free sample. The DTA results showed that the exothermic peaks on DTA curves shift toward lower temperature, and shape of exothermic peak becomes sharper with increase of cooling rate.

Fig. 3.

DTA curves of non-isothermal crystallization of mould fluxes at various cooling rates.

The crystallization temperature for crystalline phase corresponds to the temperature at which the crystallization just begins in the non-isothermal crystallization process and can be determined as onset temperature of exothermic peaks during cooling. Crystallization temperature for the first crystalline phase precipitated in mould flux is designated to be crystallization temperature of the mould fluxes.18) The onset temperature of exothermic peak for crystallization during cooling is dependent on the cooling rate, and decreases with increase of cooling rate. The crystallization temperature of mould fluxes with varying TiO2 content at fixed cooling rate of 10 K/min was determined to investigate the effect of TiO2 content on the crystallization behavior of mould fluxes. Figure 4 presented the crystallization temperature of mould fluxes with different TiO2 content at cooling rate of 10 K/min. It could be found that crystallization temperature increased with the increase of TiO2 content in the mould fluxes. The increasing crystallization temperature with increase of TiO2 content indicated that crystal growth starts at higher temperature and crystallization is enhanced. The effect of TiO2 on crystallization of fluoride free mould fluxes was investigated by Shu et al.19) by employing SHTT technique. They founded that incubation time for crystallization decreased and therefore crystallization was enhanced with increase of TiO2 content in mould fluxes. Our result is in consistence with the finding of Shu et al.19) regarding influence of TiO2 on crystallization of mould fluxes.

Fig. 4.

The change of crystallization temperature in the CaO–Al2O3 based mould fluxes at the cooling rate of 10 K/min with different content of TiO2.

CCT diagrams of mould flux for various cooling rates of 10, 15, 20 and 25 K/min were constructed by using crystallization temperatures and time-temperature profiles recorded in DTA. As shown in Fig. 5, crystallization temperatures of the crystalline phases in CaO–Al2O3 based mould fluxes with different TiO2 content shifted towards lower temperature with increasing of the cooling rate, which may be attributed by the fact that the nucleation and growth rate of crystals are functions of viscosity and undercooling. As viscosity increases quickly under the higher cooling rate, it requires a stronger driving force to initiate the mould flux nucleation. Therefore, the crystallization temperature decreased with the increase of continuous cooling rate.22) However, it could be found from Fig. 5 that the variation of crystallization temperatures for the second and third crystals formed at lower temperatures with cooling rate was less than that for the first crystal in all samples. Shi et al.18) also found similar behavior in crystallization investigation of calcium aluminate mould fluxes. It could be speculated that thermodynamic driving force (undercooling) has a dominant role for the nucleation and growth of the second and third crystals at lower temperature and kinetic factors (e.g. viscosity of liquid mould fluxes) have negligible effect on crystallization at lower temperature. Since precipitation of the second and third crystals was mainly determined by undercooling degree, crystallization temperatures do not vary with different cooling rate.

Fig. 5.

CCT diagrams for CaO–Al2O3 based mould fluxes with different TiO2 content (a) 0% TiO2 (b)5% TiO2 (c) 7% TiO2 (d) 10% TiO2.

Figure 6 presented the CCT diagram for the first crystalline phases precipitated in the CaO–Al2O3 based mould fluxes. As shown in Fig. 6, there was a large difference in crystallization temperatures between mould fluxes free of TiO2 and with 5% TiO2. It could be found that the crystallization temperature continuously increased with increasing TiO2 addition in the mould flux. When TiO2 content was at a relatively high level, the increasing effect became weaker. This indicated that the TiO2 addition promoted the crystallization of CaO–Al2O3 based mould flux. The enhancement of mould flux crystallization ability with increasing the TiO2 content might be attributed to the decrease of viscosity and the degree of polymerization of the aluminate network structure with increasing of TiO2 content in mould fluxes. In our previous work,6) we measured viscosity of the same mould fluxes using rotating spindle method and investigate the structure of mould fluxes using Raman spectroscopy. It was found that viscosity of mould fluxes gradually decreased with increase of TiO2 content. Meanwhile, degree of polymerization for the AlO4 network was found to decrease with increase of TiO2 content. The mass transfer of mould flux components from liquid to interface was accelerated with the decrease of mould flux viscosity and the depolymerization of the aluminate network structure. Therefore, nucleation and growth of crystals from liquid mould fluxes was enhanced, and overall crystallization of mould fluxes was promoted.

Fig. 6.

CCT diagram for the first crystalline phases precipitated in the mould fluxes.

Since mould flux samples after DTA measurements are too small to determine crystalline phase by XRD and SEM, a series of heat treatment experiments were carried out in order to determine the sequence of crystalline phase precipitation for mould fluxes. The experimental apparatus was shown in Fig. 2. Samples were placed in a platinum crucible and melted at 1773 K for 5 h to ensure complete melting. Subsequently, it was cooled to the desired temperature at the cooling rate of 5 K/min and then held at the temperature for 2 hours. Finally, crucible was taken out rapidly and quenched by water. The cooling rate of heat treatment is less than that in the DTA experiments due to limited abilities of the furnace.

Fig. 2.

Experimental apparatus for the phase identified sample preparation.

Figures 7, 8 showed the XRD patterns and SEM results for the water-quenched mould flux samples at different temperatures. In heat treatment experiments, cooling rate is less than that in the DTA measurements, and 2 hours’ isothermal heat treatment was performed after cooling, so the temperature for crystal precipitation in heat treatment could be higher than that in DTA measurement. In view of different kinetic condition for crystallization in heat treatment, higher temperatures than peak temperature in DTA curves were adopted for quenching. The samples free of TiO2 and with 10% TiO2 were heat treated to investigate the sequence of crystallization. As shown in Figs. 7(a) and 8(a), 8(b), there was only MgO crystal precipitated in sample free of TiO2 heated at higher temperature (1623 K), whereas MgO crystal and Ca12Al14O33 crystal co-precipitated in sample free of TiO2 heated at lower temperature (1473 K). This indicated that MgO firstly crystallized from liquid and Ca12Al14O33 crystal precipitated at lower temperature. As shown in Figs. 7(b) and 8(c)–8(e), only CaTiO3 crystal could be found in sample with 10% TiO2 heated at high temperature (1613 K) and MgO and CaTiO3 could be found at intermediate temperature (1593 K). Crystals of MgO, CaTiO3 and Ca12Al14O33 could be investigated at the lowest temperature (1473 K). Table 3 summarized quenching temperatures and identification results of crystalline phase by XRD and SEM.

Fig. 7.

XRD patterns for the mould fluxes of contain 10% TiO2 and TiO2-free heat treated at different temperatures.

Fig. 8.

SEM results for the mould fluxes of contain 10% TiO2 and TiO2-free heat treated at different temperatures.

Table 3. Crystalline phases identified by XRD and SEM in the samples No. 1 and No. 4 quenched from annealing temperature.
Sample numberHeating temperature/KCrystalline phases identified by XRD and SEM
No. 11623MgO
1473MgO+Ca12Al14O33
No. 41613CaTiO3
1593MgO+CaTiO3
1473MgO + CaTiO3 + Ca12Al14O33

It could be found that there are large differences between crystallization temperatures in heat treatment and those in continuous cooling during DTA measurements. For example, for sample free of TiO2, the temperature for the precipitation of MgO crystal should be large than 1623 K and the temperature for the precipitation of Ca12Al14O33 crystal should be located between 1473 K and 1623 K during heat treatments (see Table 3). In comparison, it could be found from CCT curves in Fig. 5(a) that the temperature for precipitation of Ca12Al14O33 crystal at cooling rate of 10 K/min should be 1474 K and 1423 K respectively, which is smaller than precipitation temperature at heat treatment.

The difference between crystallization temperatures in continuous cooling DTA and those in heat treatment may be due to the follow reasons: 1) Cooling rates in heat treatment is lower than those in DTA measurement, which leads to the higher crystallization temperature 2) A larger volume of samples for the heat treatments may increase natural convection which is beneficial to the crystallization. Correspondences of crystallization temperature between heat treatments and DTA measurements are not good. However, it was assumed that kinetic factors have similar influences on crystallization of various phases from melts and increases of crystallization temperature due to more favorable kinetics factors for various phases are close to each other. Then it was postulated that the sequences of crystallization products in heat treatments could be similar to those in continuous cooling of DTA measurement despite of large differences of crystallization temperature. Therefore, it is proposed that MgO crystals precipitated at first, followed by Ca12Al14O33 in continuous cooling for the mould flux free of TiO2. For the sample of 10% TiO2 sample, CaTiO3 crystal firstly precipitated, the second and the third phases were MgO and Ca12Al14O33 crystals respectively.

The morphology of crystals precipitating in heat treatment could be also investigated by employing SEM. As shown in Fig. 8, non-faceted MgO crystals were precipitated from mould fluxes free of TiO2, while faceted and large Ca12Al14O33 crystal was found in samples heat-treated at low temperature. In mould fluxes with 10% TiO2, non-faceted CaTiO3 crystal was found to precipitate firstly, and some CaTiO3 aligned to form chains. With the decrease of heat treating temperature, non-faceted MgO further crystallized. Faceted and large Ca12Al14O33 crystal precipitated with further decrease of heat treating temperature. Sizes of CaTiO3 and MgO crystals also increased with decrease of heat treating temperature.

The liquidus temperature of mould fluxes investigated could be obtained from heating DTA curves for different samples.23,24) During heating of samples, several endothermic thermal effects could be investigated, and liquidus temperature could be obtained as the last endothermic peak temperature.24) The undercooling for onset crystallization could be defined as ΔT=TLTx, where TL represents liquidus temperature and Tx represents onset crystallization temperature for one kind crystal.22) The ΔT values of mould fluxes with different TiO2 content were calculated by employing liquidus temperature and crystallization temperature at the cooling rate of 25 K/min. As shown in Table 2, the ΔT values of CaTiO3 precipitation in mould flux decreased from 114 K to 79 K with TiO2 content increasing from 5% to 10%, which indicated that the crystallization ability of the CaTiO3 was enhanced with increasing content of TiO2 in the CaO–Al2O3 based mould fluxes. The ΔT values for crystallization of MgO increased slightly at first with increase of TiO2 content, and then decreased slightly with further increase of TiO2 content, indicating that TiO2 addition had the negligible effect on crystallization ability of MgO. The ΔT values for crystallization of Ca12Al14O33 also change slightly with increase of TiO2 content, indicating that TiO2 addition also has a slight effect on the crystallization. Based on the above analysis, it could be concluded that with increase of TiO2 content, the crystallization of CaTiO3 from CaO–Al2O3 based mould fluxes was promoted, whereas the crystallizations of Ca12Al14O33 and MgO were hardly affected.

Table 2. The undercooling degree of the mould fluxes at the cooling rate of 25 K/min.
Sample
number
Liquidus temperature obtained by DTA (K)Crystallization
products
Crystallization
onset
Temperature (K)
Undercooling for
crystallization
(K)
No. 11633MgO1450183
Ca12Al14O331409224
No. 21628CaTiO31514114
MgO1436192
Ca12Al14O331392236
No. 31622CaTiO3152894
MgO1437185
Ca12Al14O331386236
No. 41618CaTiO3153979
MgO1435183
Ca12Al14O331386232

It could be found from Fig. 8 that CaTiO3 crystals were in non-faceted morphology, which suggested that the rate of crystallization of CaTiO3 was controlled by the diffusion of species from bulk melts to crystal-melt interface.25) In this case, viscosity of melts has a great effect on the crystallization of CaTiO3. The promotion of crystallization ability of CaTiO3 in mould fluxes could be attributed the improved kinetics by decreased viscosity of mould fluxes with addition of TiO2. It could be also found from Fig. 8 that MgO crystals have non-faceted morphology, which indicates that the crystallization of MgO was also controlled by diffusion of species in the melts. However, due to significant precipitation of CaTiO3, effect of the TiO2 addition on viscosity of melts could be weakened. Therefore, crystallization of MgO in mould fluxes was nearly unaffected by the addition of TiO2.

As seen from Fig. 8, Ca12Al14O33 crystals have larger faceted morphology, which indicates that the crystallization of Ca12Al14O33 was controlled by interfacial reaction. In such a case, viscosity could not have large influence on the crystallization of Ca12Al14O33. The crystallization of Ca12Al14O33 was mainly determined by thermodynamic driving force. Since liquidus temperature (see Table 2) is only slightly reduced by the TiO2 addition, the crystallization of Ca12Al14O33 was not influenced very much by the TiO2 addition. The ΔT values of mould flux for the crystal first precipitated (MgO for mould flux free of TiO2 and CaTiO3 for rest mould fluxes) decreased from 183 K to 79 K with increasing TiO2 from 0% to 10%. This indicated that the overall crystallization ability of mould fluxes is enhanced with increasing TiO2 from 0% to 10%.

4. Conclusion

The crystallization behaviors of the newly developed CaO–Al2O3 based mould fluxes with addition of TiO2 were investigated by DTA techniques. The crystalline phase precipitating from liquid mould fluxes was analyzed by XRD and SEM-EDS. The conclusions were summarized as follows:

(1) The MgO crystal precipitated first, and followed by Ca12Al14O33 crystal during continuous cooling of the mould flux free of TiO2. CaTiO3 crystals first precipitate from mould fluxes containing TiO2. The sequence of crystal precipitation during cooling of the mould fluxes containing TiO2 was determined to be CaTiO3 crystal to MgO crystal, and then Ca12Al14O33 crystal.

(2) The crystallization temperatures of CaO–Al2O3 based mould fluxes increased with increasing TiO2 content. The undercooling value for onset crystallization of CaTiO3 decreased with increasing content of TiO2, which indicating that the crystallization of CaTiO3 crystal product was enhanced with increasing content of TiO2. The undercooling values for onset crystallization of Ca12Al14O33 and MgO were hardly affected by TiO2 addition, indicating that the crystallization of Ca12Al14O33 and MgO remained unaffected by TiO2 addition. The overall crystallization of mould fluxes was enhanced with increasing TiO2 content from 0% to 10%.

Acknowledgement

Financial supports from Natural Science Foundation of China (NSFC contract nos. 51174018, 51174022) and Fundamental Research Funds for the Central Universities (FRF-TP-14-108A2) are gratefully acknowledged.

References
 
© 2015 by The Iron and Steel Institute of Japan
feedback
Top