The influence of non-idealities on the thermoelectric power factor of nanostructured superlattices

TL;DR

This study uses quantum simulations to evaluate how real-world imperfections affect the thermoelectric power factor improvements predicted for nanostructured superlattices, revealing their robustness or fragility under various non-idealities.

Contribution

It provides the first detailed analysis of the impact of non-idealities on superlattice thermoelectric performance, highlighting conditions that preserve or destroy power factor gains.

Findings

Power factor improvements are robust against thick barriers and random fluctuations in spacing and width.

Thin barriers and small fluctuations in barrier height can eliminate the power factor benefits.

Results explain why theoretical gains are often not observed in experimental superlattice thermoelectrics.

Abstract

Cross-plane superlattices composed of nanoscale layers of alternating potential wells and barriers have attracted great attention for their potential to provide thermoelectric power factor improvements and higher ZT figure of merit. Previous theoretical works have shown that the presence of optimized potential barriers could provide improvements to the Seebeck coefficient through carrier energy filtering, which improves the power factor by up to 40%. However, experimental corroboration of this prediction has been extremely scant. In this work, we employ quantum mechanical electronic transport simulations to outline the detrimental effects of random variation, imperfections and nonoptimal barrier shapes in a superlattice geometry on these predicted power factor improvements. Thus we aim to assess either the robustness or the fragility of these theoretical gains in the face of the types of variation one would find in real material systems. We show that these power factor improvements are relatively robust against: overly thick barriers, diffusion of barriers into the body of the wells, and random fluctuations in barrier spacing and width. However, notably, we discover that extremely thin barriers and random fluctuation in barrier heights by as little as 10% is sufficient to entirely destroy any power factor benefits of the optimized geometry. Our results could provide performance optimization routes for nanostructured thermoelectrics and elucidate the reasons why significant power factor improvements are not commonly realized in superlattices, despite theoretical predictions.