Sulfide Perovskites for Thermoelectricity.

The search for new thermoelectric materials that directly convert (waste) heat into electricity is a high-cost and time-consuming experimental effort. To facilitate this process, we perform a systematic screening for synthesizable and stable ( and are metals; = S, Se) compounds using first-principles density functional theory calculations. A total of 40 compounds are predicted to be highly competent thermoelectric materials with nontoxic and earth-abundant advantages. The calculated power factors of some of them (e.g., -type SnHfS, -type SbGaS, -type PbHfS, and so forth) are comparable (even outperform) those of the well-known thermoelectric materials such as PbTe and BiTe. The detailed analysis of electronic band structure reveals that either one or a combination of "pudding-mold" type band structure, high valley degeneracy, and high orbital degeneracy is responsible for the high computed in this family of materials. Taking two representative cases, we validate a low lattice thermal conductivity in compounds by calculating the Boltzmann transport equation using the highly accurate anharmonic lattice dynamics methods. Third-order interatomic force constants reveal that the anharmonicity and soft phonon modes, rooted in the nature of unconventional chemical bonds between the -site metals and chalcogen atoms, lead to an ultralow lattice thermal conductivity in this family of materials. The combination of intrinsically low lattice thermal conductivity and high power factor has realized highly efficient -type and -type thermoelectric materials showing various anisotropic characteristics. Considering the thermal and moisture stability of chalcogenide perovskites, our results suggest that this unexplored family of materials is a host of highly efficient and practical thermoelectric materials awaiting further experimental validation.