On the Best Bandstructure for Thermoelectric Performance

TL;DR

This paper reevaluates the optimal bandstructure for thermoelectric performance, showing that the ideal transmission function is not a Dirac delta and depends on scattering physics, challenging previous assumptions.

Contribution

It introduces a simple tight binding model and analyzes different scattering mechanisms to identify the true optimal bandstructure for thermoelectric efficiency.

Findings

A Dirac delta transmission does not maximize ZT.

Optimal bandstructure varies with scattering physics.

High density-of-states near the band edge is not always beneficial.

Abstract

The conventional understanding that a bandstructure that produces a Dirac delta function transport distribution (or transmission in the Landauer framework) maximizes the thermoelectric figure of merit, ZT, is revisited. Thermoelectric (TE) performance is evaluated using a simple tight binding (TB) model for electron dispersion and three different scattering models: 1) a constant scattering time, 2) a constant mean-free-path, and 3) a scattering rate proportional to the density-of-states. We found that a Dirac delta-function transmission never produces the maximum ZT. The best bandstructure for maximizing ZT depends on the scattering physics. These results demonstrate that the selection of bandstructure to maximize TE performance is more complex than previously thought and that a high density-of-states near the band edge does not necessarily improve TE performance.