<i>Ab initio</i> high-throughput search methodology for next-generation power semiconductors candidates

Abstract

Since power semiconductor devices play crucial roles in electrical energy conversion and control, power semiconductor materials with more efficient and reliable power transmission and distribution are particularly worth exploring. The resource-intensive and time-consuming nature of traditional trial-and-error experimental approaches poses significant challenges in efficiently identifying high-performance ultra-wide-bandgap materials. In this work, we present a novel high-throughput search methodology based on the ab initio calculations to systematically evaluate and screen all 154 718 materials from the Materials Project database for critical properties, including low energy consumption, high melting point, wide bandgap, high electron mobility, and high thermal conductivity. We suggest 36 potential candidate materials. By comparing with third-generation power semiconductors (e.g., GaN and SiC), we further suggest that BeO and B2O3, which exhibit the highest Baliga figures of merit and Johnson figures of merit, might be the potential candidates for next-generation power semiconductors. The predictive capability of ab initio calculations in characterizing material properties prior to experimental validation significantly reduces the resources and time required for preliminary screening, offering a scalable pathway to accelerate the discovery of advanced materials for high-power electronics.