Multi-Objective Optimization of Thermoelectric (TE) Performance in Binary Chalcogenides through Computational Assessment of Extrinsic Dopants

We demonstrate a computational approach to optimize the dopant profile of a thermoelectric material, accounting for the effect of the dopant on multiple interrelated indicators of performance. While most computational studies of dopants in TE materials primarily focus on optimization of carrier concentrations with the assumption that bulk/intrinsic properties of the material remain unchanged, in cases of high concentrations doping can affect other aspects of performance through modifications to carrier mobilities, thermal conductivity, and density of states. These competing effects often make it challenging to identify optimal doping concentrations. In this work, we use the example of Ga and Sb co-doping as a case study for the multi-objective optimization of n-type achievable TE performance in the group IV-VI binary chalcogenides with AB stoichiometry. Although recent reports show improved performance in n-type PbS based on co-doping with Ga and Sb, a predictive model that describes doping effects on multiple properties would reveal ultimate performance achievable and the carrier concentrations at which it occurs. We present models to assess how co-doping with Ga and Sb affect carrier mobilities, thermal conductivity, electronic density of states, and carrier concentrations in PbS and PbSe. The density of states analysis on PbS and PbSe show increased dos curvature after Ga, Sb co-doping. The defect properties show the stability of the co-doped system over a single dopant, which shows the stability of co-doped Ga-Sb defect would be stable over single point defect Ga and Sb. Our analysis reveals opportunities for multi-objective optimization of TE properties through understanding the change in effective mass, mobility, thermal conductivity, and optimized carrier concentration.