Potential models for multicomponent oxides: Hexa-aluminates

Abstract Our recent experiences in the use of interatomic potential models for the simulation of complex multicomponent oxides are discussed, with particular reference to alkaline-earth hexa-aluminates. We have found that simply transferring parameters from models of binary oxides to these ternary oxides is not necessarily valid, and that the usual test of reproducing the experimental property data such as crystal structural parameters is not by itself sufficient. Additional checks on the viability of the potential model, such as considerations of thermodynamic stability, are also highly desirable. A re-examination of the potential model for strontium magnetoplumbite SrAl12O19, revealed the need to include a consideration of cation coordination, as had been found desirable in simulations of spinels. Furthermore, by taking account of cation coordination only within the spinel-structured blocks in hexa-aluminate structures, and otherwise transferring an appropriate set of binary oxide potentials, the thermodynamic stabilities of the related complex compounds are correctly predicted, thus allowing a discussion of the defect solid-state chemistry of these systems.