Computational identification of p-type transparent semiconductors

Abstract

Transparent oxide semiconductors have been used in various applications such as display and solar cell panels. The development of n-type oxides such as In2O3 and ITO realized the transparent electronics. In order to fabricate highly efficient transparent electronic devices, high performance p-type transparent oxide semiconductors comparable to the n-type oxides are required. However, the performance of present p-type oxides is much inferior to that of n-type oxides in spite of enormous efforts. Recently, several novel p-type candidates are suggested by using high-throughput approach with the density functional theory calculations. However, candidate materials have not been verified in experiment, which implies the necessity of a reliable theoretical descriptor. <br/> In this dissertation, we propose a reliable and computationally efficient descriptor for p-type dopability using the formation energy of hydrogen interstitial defect. The predictive power of hydrogen descriptor is demonstrated by showing that the hydrogen descriptor can distinguish well known p-type and n-type oxides. Using the hydrogen descriptor, we conduct high-throughput screening for most binary oxides and selected ternary oxides including Sn2+- and Cu1+-bearing oxides as well as oxychalcogenides. As a result, we introduce La2O2Te and CuLiO as promising p-type candidates, which are validated through the full intrinsic defect calculations. Additionally, we identify simple descriptors that correlate with hydrogen descriptor. By using the simple descriptors in the multi-step high-throughput screening, we screen most oxide compounds (~17,000 oxides) in Aflow database. As a result, we suggest new p-type candidates such as Na2PtO2, La2O2S2 and La2SiO4Se.