High Throughput Density Functional Investigations of the Stabilities, Electronic Structures of XYZ2 Compounds

Abstract

The development of high-performance thermoelectric materials requires the reduction of the thermal conductivity as well as the enhancement of the power factor simultaneously, which may benefit from high-throughput materials simulations. Recently, we have revealed a new class of high-performance thermoelectric materials, namely XYZ2 (X,Y: rare earth or transition metals, Z: group VI element) based on high throughput simulations, among which TmAgTe2, YAgTe2, YCuTe2 have been successfully synthesized and proved to have extraordinary low thermal conductivity. But, similar to many other high-performance thermoelectric materials, e.g. PbTe, SnSe, XYZ2 thermoelectric materials could also have multiple phase transformations. Thus, it is important to systematically study the stability, electronic structure and transport property of different phases, and how these are related to the chemistry of the compound. The phase stabilities of XYZ2 compounds have been assessed by support vector machine methodology. We find that those compounds with larger X ionic radius have P3m1 space group with X ion in octahedral interstices, while smaller X ionic radius favors a tetrahedral site occupation with a space group of the compound of I-42d. Based on the gradient boosting algorithm, density of states effective mass mS* and fermi surface complexity factor NV*K* are identified as important descriptors to enhance zT values. Band structures and Crystal Orbital Hamilton Population analyses reveal that materials with larger mS* have a flat band to enhance Seebeck coefficient if the X-s (Y-d) orbitals have similar energy with Z-p orbital near the VBM. Cu and Ag based compounds with P3m1 space group are benefited from the Y-d orbital contribution to the high Fermi surface complexity factor. These intrinsic parameters including ionic size and molecular orbitals can alter stabilities and electronic properties, which provide us clues on how to design and improve the thermoelectric performance of XYZ2 compounds.