Towards a predictive route for selection of doping elements for the thermoelectric compound PbTe from first-principles

Abstract

Striving for improvements of the thermoelectric (TE) properties of the technologically important lead telluride (PbTe) compound, we investigate the influence of different doping elements on the thermal conductivity, Seebeck coefficient, and electrical conductivity applying density functional theory calculations. Our approach combines total-energy calculations yielding lattice vibrational properties with the Boltzmann transport theory to obtain electronic transport properties. We find that doping with elements from the 1st and 3rd columns of the periodic table reduces the sound velocity and, consequently, the lattice thermal conductivity, while 2nd column dopants have no such influence. Furthermore, 1.6 at. % doping with 4th and 5th column elements provides the highest reduction of lattice thermal conductivity. Out of this group, Hf doping results in maximum reduction of the sound velocity from 2030 m s−1 for pure PbTe to 1370 m s−1, which is equivalent to ca. 32% reduction of lattice thermal conductivity. The highest power factor values calculated for 1.6 at. % doping range between 40 and 56 μW cm−1 K−2, and are obtained for substitution with dopants having the same valence as Pb or Te, such as those located at the 2nd, 14th, and 16th columns of the periodic table. We demonstrate how this method may be generalized for dopant-selection-oriented materials design aimed at improving TE performance of other compounds.