Path to predictive computations of thermoelectric effects for automotive waste heat recovery

Boris Kozinsky

Understanding the atomic-level origins of thermoelectricity is necessary in order to design and optimize higher-performing materials. Measurements of ZT at high temperature are prone to uncertainty and more fundamental investigations into heat and charge carrier dynamics are even more challenging. We demonstrate how ab-initio computation can be a valuable tool to complement experimental understanding. One way to increase performance is to reduce thermal conductivity through enhanced phonon scattering by alloying, rattler filling or nanostructuring. We show how density-functional perturbation theory, with the help of the Boltzmann transport theory and quasi-harmonic analysis, can deliver predictive accuracy in calculating electronic and thermal transport of bulk and nanostructured thermoelectrics. We analyze effects of alloying and nanostructuring on thermal transport in such materials as SiGe alloys and skutterudites, comparing to experimental measurements. We emphasize how these microscopic calculations provide design rules to gauge the importance of composition, disorder and nanostructure in enhancing thermoelectric performance.

Back to Workshop IV: Energy Conservation and Waste Heat Recovery