High-Throughput Design of Non-oxide p-Type Transparent Conducting Materials: Data Mining, Search Strategy, and Identification of Boron Phosphide

High-performance p-type transparent conducting materials (TCMs) are needed in a wide range of applications ranging from solar cells to transparent electronics. p-type TCMs require a large band gap (for transparency), low hole effective mass (for high mobility), and hole dopability. It has been demonstrated that oxides have inherent limitations in terms of hole effective masses making them difficult to use as a high-performance p-type TCM. In this work, we use a high-throughput computational approach to identify novel, non-oxide, p-type TCMs. By data mining a large computational data set (more than 30,000 compounds), we demonstrate that non-oxide materials can lead to much lower hole effective masses but to the detriment of smaller gaps and lower transparencies. We propose a strategy to overcome this fundamental correlation between low effective mass and small band gap by exploiting the weak absorption for indirect optical transitions. We apply this strategy to phosphides and identify zinc blende boron phosphide (BP) as a very promising candidate. Follow-up computational studies on defects formation indicate that BP can also be doped p-type and potentially n-type as well. Our work demonstrates how high-throughput computational design can lead to identification of materials with exceptional properties, and we propose a strategy to open the design of TCMs to non-oxide materials.