The difference may arise from the database, as illustrated in the Database Documentation, in the presence of Ti, the database has been developed only for reducing conditions, and the liquid phase is generally well modeled only for binary systems, but for ternary and higher-order systems is only estimated from the model due to lack of any experimental. 6 that the calcu- lated 1 300?C isotherm has a good accordance with both the 1 300?C isotherm from present work and Sun et al., 25) while the calculated 1 400?C isotherm has a higher solubil- ity of SiO 2 when the TiO 2 content is smaller than 19 wt.%, even though experimental results and calculated data have the same variation trend with compositions. 6, the 1 300?C and 1 400?C isotherms calculated by factsage software 44) are also added for comparison. 25) established the phase diagram in dynamic conditions which is different from the static methods adopted in pres- ent work, otherwise, the normalization of present data can also cause minor shift. The deviation may be due to the different method employed in the experiments, Sun et al.

As for 1 400?C isotherm, the liquidus line changes with composi- tion below 15 wt.% TiO 2 from both have the same variation trend, however, the 1 400?C isotherm from present work shifts to a little higher solubility of SiO 2 when the TiO 2 content is higher than 15 wt.%. 25) have the same variation trend with compositions. As can be seen, both the 1 300?C liquidus line from present work and Sun et al. 25) are plotted as short dash lines for comparison.

Factsage Database