On Using <i>ZT</i> to Optimize Thermoelectric Performance

Thermoelectric technology allows for the direct solid-state conversion of heat to electrical energy, enabling power generation in applications such as deep space exploration. Since the best thermoelectric materials are typically heavily doped semiconductors, an important step in the development of new materials with better energy conversion efficiencies is the optimization of thermoelectric properties by adjusting the doping density and/or composition of the material. In this work, we use simple semiconductor transport models to show that the metric most commonly used to identify the optimum doping density, the dimensionless thermoelectric figure of merit, ZT, can lead to incorrect conclusions about the optimum doping density and therefore a potential loss of device efficiency. Our analysis suggests it is safer to use modern efficiency models, in which the temperature dependence of the material properties are accounted for, rather than ZT to determine the optimal doping density for a thermoelectric material.