High-throughput computational screening of thermal conductivity, Debye temperature, and Grüneisen parameter using a quasiharmonic Debye model

The quasiharmonic Debye approximation has been implemented within the aflow and Materials Project frameworks for high-throughput computational materials science (Automatic Gibbs Library, agl), in order to calculate thermal properties such as the Debye temperature and the thermal conductivity of materials. We demonstrate that the agl method, which is significantly cheaper computationally compared to the fully ab initio approach, can reliably predict the ordinal ranking of the thermal conductivity for several different classes of semiconductor materials. In particular, a high Pearson (i.e., linear) correlation is obtained between the experimental and agl computed values of the lattice thermal conductivity for a set of 75 compounds including materials with cubic, hexagonal, rhombohedral, and tetragonal symmetry.