Simulation of lattice properties of SnSe based on Ab initio calculations and machine learning

A machine learning algorithm is used to propose three possible model phonon densities of states for a perspective thermoelectric material SnSe. Based on this and on the energy of the ground state calculated from the first principles, we develop a thermodynamic model of SnSe. The model is used to simulate temperature dependencies of the lattice properties of SnSe such as heat capacity, coefficient of thermal expansion, and bulk modulus. The anharmonicity of acoustic and optical phonons of SnSe is discussed.