A Computational Survey of Semiconductors for Power Electronics

Power electronics (PE) are used to control and convert electrical energy in a wide range of applications from consumer products to large-scale industrial equipment. While Si-based power devices account for the vast majority of the market, wide band gap semiconductors such as SiC, GaN, and Ga₂O₃ are starting to gain ground. However, these emerging materials face challenges due to either non-negligible defect densities, or high synthesis and processing costs, or poor thermal properties. We performed a broad computational search aimed at identifying promising materials for future power electronic devices beyond SiC, GaN, and Ga₂O₃. We consider 863 oxides, sulfides, nitrides, carbides, silicides, and borides reported in the crystallographic database and exhibit finite calculated band gaps. We utilize ab initio methods with models for intrinsic carrier mobility, and critical breakdown field to compute the Baliga figure of merit. We also compute the lattice thermal conductivity as a screening parameter. In addition to correctly identifying known power electronic materials, our survey has revealed a number of promising candidates exhibiting the desirable combination of high figure of merit and high lattice thermal conductivity, which we propose for further experimental investigations.