Stability and Equilibrium Structures of Unknown Ternary Metal Oxides Explored by Machine-Learned Potentials

Ternary metal oxides are crucialcomponents in a widerange ofapplications and have been extensively cataloged in experimental materialsdatabases. However, there still exist cation combinations with unknownstability and structures of their compounds in oxide forms. In thisstudy, we employ extensive crystal structure prediction methods, acceleratedby machine-learned potentials, to investigate these untapped chemicalspaces. We examine 181 ternary metal oxide systems, encompassing mostcations except for partially filled 3d or f shells, and determinetheir lowest-energy crystal structures with representative stoichiometryderived from prevalent oxidation states or recommender systems. Consequently,we discover 45 ternary oxide systems containing stable compounds againstdecomposition into binary or elemental phases, the majority of whichincorporate noble metals. Comparisons with other theoretical databaseshighlight the strengths and limitations of informatics-based materialsearches. With a relatively modest computational resource requirement,we contend that heuristic-based structure searches, as demonstratedin this study, offer a promising approach for future materials discoveryendeavors.

Automated summary

In this study, we utilize crystal structure prediction methods accelerated by machine-learned potentials to investigate ternary metal oxides. We discover 45 stable ternary oxide systems, which can contribute to future materials discovery endeavors.