Note
Reassessment of Oxide Stability Diagram in the Fe–Al–Ti–O System
2017 Volume 57 Issue 9 Pages 1665-1667
Details
2017 Volume 57 Issue 9 Pages 1665-1667
Phase equilibria of Fe–Al–Ti–O system is shortly reviewed, in particular for stability of oxide phases in equilibrium with liquid Fe alloy in dilute concentration. This system is relevant to Ultra Low Carbon steel with Ti alloying (Ti-ULC), often causing clogging of Submerged Entry Nozzle (SEN) during continuous casting. In order to predict stable oxides in the Ti-ULC steel, thermodynamic calculation using CALPHAD approach was employed. Previously available thermodynamic calculations were examined, and compared with available experimental data. A slight modification of thermodynamic modeling of previous investigation resulted in noticeable improvement for the prediction of stable phases by thermodynamic calculation. Significance of the present revision of the thermodynamic modeling is discussed in the view of nozzle clogging of Ti-ULC casting.
It has been a long-standing problem that clogging of Submerged Entry Nozzle (SEN) occurs during continuous casting of Ti added Ultra Low Carbon (Ti-ULC) steel. Increasing Ti content significantly increases extent of the nozzle clogging. Although cause and countermeasure of the nozzle clogging for Ti-free steel has been known, those for Ti-ULC are still unclear. Basu et al. reported interesting phenomena and possible mechanism during continuous casting of Ti-ULC.1) However, correct mechanism has not been elucidated yet. A number of researchers have paid attention to non-metallic inclusions in the Ti-ULC steel.2,3,4,5,6) It is easily thought that inclusions as a deoxidation product may consist of Al–Ti–O. Various kinds of oxides may be considered: Al2O3, Ti2O3, Ti3O5 as well as complex oxide such as Al2TiO5 or liquid oxide. It has been a debate whether the complex oxides exist in liquid steel as inclusions or not, and whether these complex oxide inclusions may be responsible for the serious nozzle clogging of Ti-ULC or not. So called “inclusion stability diagram” in the Fe–Al–Ti–O was then introduced, first by Ruby-Mayer et al.,7) subsequently by a number of researchers.2,8,9,10) Recently, the present authors have paid attention to the interfacial reaction between Ti-ULC steel and refractory of SEN.11) The refractory generates CO(g), which reacts with Ti-ULC steel at the interface between inner wall of the nozzle and the liquid steel. As a reaction product, it was found that liquid oxide formed along with solid alumina. Interpretation of this reaction also requires knowledge on thermodynamics and phase equilibria in the Fe–Al–Ti–O system. The oxide observed was not just an inclusion in liquid steel, therefore in the present note, phase diagram representing stability of various oxide phases in equilibrium with liquid steel is called “oxide stability diagram”.
The oxide stability diagram at 1520°C reported by Ruby-Mayer et al.7) is shown in Fig. 1. The diagram is governed by stable regions of Ti2O3, liquid oxide, Al2TiO5, and Al2O3. Liquid oxide and Al2TiO5 are seen to be stable between Ti2O3 and Al2O3. Jung et al.8) proposed a new diagram calculated at 1600°C using FactSage.12) They reported that Ti3O5 should also be stable, by replacing some part of the stable region of Ti2O3, and liquid oxide was not to be stable. Later, they revised their diagram that the liquid oxide appears between Ti3O5 and Al2O3.10) Matsuura et al.2) proposed a different type of oxide stability diagram at 1600°C where Al2TiO5 located between Ti3O5 and Al2O3. They did not consider liquid oxide phase due to limited thermodynamic information of liquid oxide.
Oxide stability diagram of the Fe–Al–Ti–O system at 1520°C reported by Ruby-Meyer et al.7)
Recently, Jo carried out a series of experiments in order to observe whether complex oxide (Al2TiO5, liquid oxide) is stable apart from Al2O3 and Ti3O5.13) He equilibrated liquid Fe–Al–Ti alloy in an Al2O3 crucible at 1600°C, and analyzed concentrations of Al and Ti in the liquid alloy using ICP, concentration of O using LECO, and interface between the liquid alloy and the crucible using SEM and EBSD. They confirmed that only Ti3O5 was newly formed between the liquid alloy and the Al2O3 crucible, when concentration of O in the liquid alloy was less than ~ 40 mass ppm. This means that only Ti3O5 and Al2O3 should appear in the oxide stability diagram in the range ([ppm O] < 40). This result does not conform to the previous thermodynamic calculations.2,7,8,9,10) Matsuura and Tsukihashi also reported that no intermediate phase was found between Al2O3 and TiOx solid phases in liquid Fe–Al–Ti alloys at 1600°C, when concentration of O was less than ~ 60 ppm.14) This also partly agrees with the result of Jo.13) It can be inferred from these investigations that there is no phase between Ti3O5 and Al2O3 in the Fe–Al–Ti–O system when concentration of O in the liquid alloy is low (less than ~ 60 ppm).
The present authors’ recent investigation revealed that liquid oxide can form at relatively higher concentration of O, obtained by reaction with CO(g).11) Reaction between liquid Fe and CO(g) would increase the concentration of O up to ~ 500 ppm. Therefore, in such high O region, the liquid oxide should appear in the oxide stability diagram. This is important to interpret interfacial reaction between Ti-ULC steel and nozzle refractory. So far, the oxide stability diagrams reported in literature2,7,8,9,10) do not explain recent experimental data consistently. In the present study, a new oxide stability diagram in the Fe–Al–Ti–O system is presented, by reassessing available thermodynamic modeling of the system.
In the present study, CALPHAD type thermodynamic calculation was carried out using FactSage.12) The following phases were considered: liquid alloy (Fe–Al–Ti–O), liquid oxide (FeO–FeO1.5–TiO1.5–TiO2–AlO1.5), corundum (Al2O3 with limited solubility of Ti2O3), ilmenite (Ti2O3 with limited solubility of Al2O3), pseudobrookite (Ti3O5 with limited solubility of FeTi2O5), spinel (FeAl2O4 with limited solubility of Fe3O4), and Al2TiO5. Gibbs energies of these phases were initially taken from FactSage FTOxid and FTmisc databases. Formulation of Gibbs energies are described in previous reports.8,15,16,17,18,19) Using these thermodynamic information, initially calculated oxide stability diagram of Fe–Al–Ti–O system at 1600°C looks almost the same to that reported by Jung et al. in 2009.10) In order to make the liquid oxide phase less stable to conform with the data of Jo13) and that of Matsuura and Tsukihashi,14) it was decided to reassess Gibbs energy of liquid Al2O3–TiO2–Ti2O3 system.
Gibbs energy of the liquid oxide phase was modeled using the Modified Quasichemical Model in the pair approximation.20) In an AO-BO liquid oxide solution, the following pair exchange reactions are considered:
| (1) |
The previously estimated Gibbs energy of liquid Al2O3–TiO2–Ti2O3 system by Jung et al.10) was revised in the present study. The Gibbs energy of the ternary system was first estimated from that of sub-binary system using “Toop-like” interpolation method, with Al2O3 as an asymmetric component.21) Since TiO2–Ti2O3 binary liquid exhibits negative deviation16) while Al2O3–TiO2 and Al2O3–Ti2O3 exhibit positive deviation,10) it is thought that choosing Al2O3 as the asymmetric component is reasonable. A small ternary adjustable parameter, 41840 (XAlO1.5/(XAlO1.5+XTiO1.5+ (4/3)XTiO2)) (J/mol), was added to the ΔgTi3+–Ti4+,16) in order to reproduce the experimental data of Jo in the Fe–Al–Ti–O oxide stability diagram at 1600°C.13) This parameter represents an effect of Al3+ on the formation energy of (Ti3+-O-Ti4+), thereby making the (Ti3+-O-Ti4+) SNN pair less stable by Al3+ (TiO2–Ti2O3 binary liquid exhibiting less negative deviation). By changing the interpolation method as well as adding the positive ternary parameter, the Gibbs energy of Al2O3–TiO2–Ti2O3 liquid oxide becomes relatively positive over entire composition, and the ternary liquid oxide could be less stable, compared to the previous assessment.10) This resulted in narrowed liquid region in the ternary phase diagram.
Figure 2 shows a calculated liquidus surface of Al2O3–TiO2–Ti2O3 system, which is slightly different to that proposed by Jung et al.10) The minimum eutectic temperature was increased from 1496°C10) to 1553°C, for a eutectic reaction Liquid→Ilmenite+Corundum+Ti4O7. The decreased stability of ternary liquid then contributes to the narrowed liquid oxide region in Fe–Al–Ti–O system.
Calculated liquidus surface of Al2O3–TiO2–Ti2O3 system.
Figure 3 shows a calculated oxide stability diagram in equilibrium with liquid alloy at 1600°C in the Fe–Al–Ti–O system, extended to lower concentrations of Al and Ti. Thin lines are calculated iso-[ppm O] lines, and thick lines are phase boundaries. The stability diagram is mostly governed by Al2O3-corundum, (Ti,Fe)3O5-pseudobrookite, and liquid oxide (FetO–TiOx–Al2O3). Stability regions of FeAl2O4-spinel and (Ti,Al)2O3-ilmenite are also seen in limited regions. Open circles are experimentally determined concentrations of Al and Ti in the liquid alloys, which was in equilibrium with Ti3O5 and Al2O3.13) The experimental method was described in Sec. 1. The experimental data were in good agreement with the present thermodynamic calculation, except for one datum at higher Al, Ti concentrations. Jo reported that there were no phases observed between Al2O3 and Ti3O5, where [ppm O] was less than ~ 40 ppm. In the calculated diagram, liquid oxide becomes stable when [ppm O] is higher than ~ 50 ppm. Indeed, the size of ternary model parameter for the liquid oxide phase mentioned earlier was optimized in order to reproduce the experimental data of Jo,13) but to be stable at high O content according to the present authors’ investigation.11) It should be noted that the “50 ppm” (lower limit of O concentration where the liquid oxide becomes stable) is only an approximated value. Therefore, it should be understood that the liquid oxide is not stable when [ppm O] is less than ~ 40 ppm,13,14) but it should be stable at relatively higher O concentration.11)
Calculated oxide stability diagram of the Fe–Al–Ti–O system at 1600°C in the present study. Thin lines are iso-[ppm O] lines. Open circles were taken from Jo.13)
When the Ti-ULC steel passes through SEN connecting tundish and casting mold, the liquid steel is in contact with the nozzle refractory, which emits CO(g) according to Sasai and Mizukami.22) Temperature and O concentration at the interface between the liquid steel and the nozzle refractory are different to that in bulk liquid steel in RH. The temperature is about 1520°C–1550°C. Equilibrium O concentration in liquid Fe was estimated in such a way that CO(g) dissolves into the liquid Fe by CO(g)=C+O, and saturated concentration of O at this temperature range is about 500 ppm. This high O would consume Al and Ti in the liquid steel. Hence, concentrations of Al and Ti near the interface should be lower than those in the bulk. Similarly, concentration of O should be higher than that in the bulk.
Figure 4 shows a calculated (predicted) oxide stability diagram in the Fe–Al–Ti–O system at 1540°C in order to simulate condition at the interface during continuous casting. When concentrations of Al and Ti are in the range of a few hundred ppm, concentration of O is seen to be a few ppm (marked as “Region A”). This corresponds to the bulk steel. Equilibrium oxide phase is Al2O3. This also tells that most stable inclusions in the bulk steel is alumina. On the other hand, when concentration of O in the liquid steel is high due to the dissolution of CO(g) from the refractory nozzle, it is reasonable to think that concentrations of Al and Ti are lower and concentration of O is higher than those in the Region A. This is marked as “Region B”. It is clearly seen that liquid oxide should form under this condition. This is also consistent with the present authors’ experimental observation.11)
Calculated oxide stability diagram of the Fe–Al–Ti–O system at 1540°C in the present study. “Region A” represents a condition in bulk liquid steel, and “Region B” represents another condition at interface between the liquid steel and the nozzle refractory.
Thermodynamic reassessment of the oxide stability diagram in the Fe–Al–Ti–O system was carried out. By decreasing thermodynamic stability of liquid Al2O3–TiO2–Ti2O3 system, stable region of “liquid oxide” in the stability diagram could be refined. This reassessment is consistent with recent experimental investigations, regarding the stability of liquid oxide.11,12,13) The thermodynamic reassessment and newly proposed oxide stability diagram can be used for figuring out what type of oxide phase including non-metallic inclusion and interfacial reaction product could form in Ti-ULC liquid steel.
The present authors would like to thank to Prof. J.-J. Pak, Hanyang University, Rep. of Korea for allowing us to use unpublished data of Dr. J.-O. Jo, and Prof. I.-H. Jung, McGill University, Canada for valuable discussion regarding phase equilibria and thermodynamic modeling.
