A System-Level Approach to Thermoelectric Material Property Optimization

Abstract

The opportunities for creating the most effective thermoelectric (TE) material can be maximized by considering the full system for a given application when developing the material. If conversion efficiency is the only consideration in the design of a TE material, then maximizing average ZT over the largest temperature range may be the best choice. If the system or end application is unknown or not well defined, this design path may also be the best choice. However, there are more factors that should affect TE material design choice than just maximizing average ZT when considering the system-level attributes of an application. The following paper demonstrates how other aspects of the design affect how TE material properties (Seebeck coefficient, electrical resistivity, and thermal conductivity) can be tuned to get maximum performance for a given application. These other aspects include device and system level parasitic losses, constraints, and design objectives. Simulation results are provided that demonstrate how a TE material with a lower average ZT can be more effective depending on the design objective than a TE material with a higher average ZT.