Using First-Principles Calculations in CALPHAD Models to Determine Carrier Concentration of the Binary PbSe Semiconductor

Abstract

PbSe is a promising thermoelectric that can be further improved by nanostructuring, band engineering, and carrier concentration tuning; therefore, a firm understanding of the defects in PbSe is necessary. The formation energies of point defects in PbSe are computed via first-principles calculations under the dilute-limit approximation. We find that under Pb-rich conditions, PbSe is an n-type semiconductor dominated by doubly-charged Se vacancies. Conversely, under Se-rich conditions, PbSe is a p-type semiconductor dominated by doubly-charged Pb vacancies. Both of these results agree with previously performed experiments. Temperature- and chemical potential-dependent Fermi levels and carrier concentrations are found by enforcing the condition of charge neutrality across all charged atomic and electronic states in the system. The first-principles-predicted charge-carrier concentration is in qualitative agreement with experiment, but slightly varies in the magnitude of carriers. To better describe the experimental data, a CALPHAD assessment of PbSe is performed. Parameters determined via first-principles calculations are used as inputs to a five-sublattice CALPHAD model that was developed explicitly for binary semiconductors. This five sublattice model is in contrast to previous work which treated PbSe as a stoichiometric compound. The current treatment allows for experimental carrier concentrations to be accurately described within the CALPHAD formalism. In addition to the five-sublattice model, a two-sublattice model is also developed for use in multicomponent databases. Both models show excellent agreement with the experimental data and close agreement with first-principles calculations. These CALPHAD models can be used to determine processing parameters that will result in an optimized carrier concentration and peak zT value.