First principles methods for elpasolite halide crystal structure prediction at finite temperatures

Abstract

• We predict crystal structures at finite temperature for four elpasolite halides. • We compare three levels of theory: ground state, simple harmonic, and quasiharmonic. • Ground state energies alone are inadequate due to low temperature phase changes. • Quasiharmonic model outperforms simple harmonic level of theory. • Methods are useful for identifying materials with desirable crystal structures. Computational methods able to identify elpasolite halides that crystallize with cubic symmetry are desirable since the isotropic properties of these materials have attractive manufacturing and optical properties. We compare the performance of three levels of theory for four elpasolite halides, Cs 2 NaGdBr 6 , Cs 2 NaLaBr 6 , Cs 2 LiLaI 6 , and Cs 2 LiScI 6, in order to determine the minimum level of theory required to accurately predict the equilibrium crystal structures at finite temperatures. We evaluate ground state, simple harmonic, and quasiharmonic free energies for each material in the common cubic, tetragonal, and trigonal symmetries using Density Functional Theory (DFT) and phonon calculations. The highest level of theory based on a quasiharmonic model accounting for thermal expansion reproduces available experimental phase information for the studied materials and outperforms the simple harmonic model. As expected, ground state energies alone do not provide unambiguous information regarding expected finite temperature crystal structures and fail to identify interesting materials that crystallize with cubic symmetry. The methods we demonstrate will be useful for considering the large number of elpasolite halides that exist to identify desirable crystal structures.