Review of computational approaches to predict the thermodynamic stability of inorganic solids

Improvements in the efficiency and availability of quantum chemistry codes, supercomputing centers, and open materials databases have transformed the accessibility of computational materials design approaches. Thermodynamic stability predictions play a central role in the efficacy of these approaches and should be considered carefully. This review covers the fundamentals of calculating thermodynamic stability using first-principles methods. Stability is delineated into two main topics-stability with respect to decomposition into competing phases and stability with respect to phase transition into alternative structures at fixed composition. For each topic, a summary of the state-of-the-art is provided along with a tutorial overview of practical considerations. The application of machine learning to both kinds of stability predictions is also covered. Finally, the limitations of thermodynamic stability predictions are discussed within the context of predicting the synthesizability of materials.

Automated summary

This review discusses the fundamentals of calculating thermodynamic stability using first-principles methods. It covers stability with respect to decomposition into competing phases and stability with respect to phase transition into alternative structures at fixed composition. The state-of-the-art and practical considerations for each topic are summarized. The application of machine learning to stability predictions is also addressed. Finally, the limitations of thermodynamic stability predictions in predicting materials synthesizability are discussed.