Large-Scale Computational Identification of p-Type Oxide Semiconductors by Hierarchical Screening

With potentials to outperform traditional semiconductors, oxide semiconductors are penetrating into a wide range of energy, environmental, and electronic applications. However, an insufficient library of p-type oxides is a major obstacle against complete oxide electronics. To identify promising p-type oxides, we carry out a large-scale computational screening over 17,700 oxygen-containing compounds using the hydrogen descriptor that was shown to select the p-type oxide reliably. To enable high-throughput screening, we use in the initial stage coarse but simple descriptors that correlate with the hydrogen descriptor. Through a hierarchical screening procedure, we find 156 p-type oxides with the band gap larger than 1.1 eV, discovering promising candidates such as NaNbO2 and La2SiO4Se and also new types of p-type oxides. Furthermore, we classify the identified p-type oxides according to the valence-band character and reveal chemical principles underlying the p-type dopability.