“…It is also seen from Tables 6 and 7 that different approaches have been used to model the non-ideality of solution phases in different software. For example, the molten oxide phase is described with the modified quasichemical model (MQM) 56,87,[136][137][138][139] in FactSage's (GTT) FToxid database, [129][130][131]134) with the associate model (AM) 56,91,120) in MTDATA's (NPL) NPLOX database, 54,132,134) with the cell model (CM) 56,140,141) in CEQCSI's (IRSID) 123,125,126,133,134) and MPE's (CSIRO) 127,128,134) oxide databases and with either cell model 56,140,141) or associate model 56,91,120) in ThermoCalc's SLAG2 and NOX3 databases, respectively. 134) Furthermore, the generalized central atom model (GCA) 124,126) developed based on the cell model has also been used to model both oxide and metal phases in CEQCSI, 125) whereas the reciprocal ionic liquid model (RILM), 56,142) which assumes two sublattices in liquid oxide phase has been used in ThermoCalc's own ION3 -oxide database.…”