High-Throughput Screening for Thermoelectric Semiconductors with Desired Conduction Types by Energy Positions of Band Edges.

The conduction type of semiconductors is vitally important in many fields (e.g., photovoltaics, transistors, and thermoelectrics), but so far, there is no effective and simple indicator to quickly judge or predict the conduction type of various semiconductors. In this work, based on the relationship between the formation energy of charged defect and the Fermi level, we propose a simple and low-cost strategy for high-throughput screening the potential n-type or p-type semiconductors from the material database by using energy positions of band edges as indicators. As a case study, we validate this strategy in searching potential n-type thermoelectric materials from copper (Cu)-containing metal chalcogenides. A new promising thermoelectric material, CuIn 5 Se 8 , with potential intrinsic n-type conduction, is successfully screened from 407 Cu-containing metal chalcogenides and validated in the subsequent experiments. Upon doping iodine in CuIn 5 Se 8 , a peak thermoelectric figure of merit zT of 0.84 is obtained at 850 K. Beyond thermoelectrics, the strategy proposed in this study also sheds light on the new material development with desired conduction types in photovoltaics, transistors, and other fields.