Thermoelectric properties of semiconductor nanowire networks

TL;DR

This paper introduces a theoretical model for analyzing thermoelectric properties of semiconductor nanowire networks, demonstrating how junctions and disorder influence the figure of merit, $ZT$, with implications for large-scale thermoelectric device design.

Contribution

The paper develops a two-port network model for thermoelectric nanowire networks, enabling large-scale simulations and revealing the impact of junction scattering and disorder on $ZT$ enhancement.

Findings

Phonon scattering at junctions significantly increases $ZT$.

Disordered networks show an order of magnitude higher $ZT$ than ordered ones.

Preferential pathways in Cayley tree networks mimic branched nanowire behavior.

Abstract

To examine thermoelectric (TE) properties of a semiconductor nanowire (NW) network, we propose a theoretical approach mapping the TE network on a two-port network. In contrast to a conventional single-port (i.e., resistor) network model, our model allows for large scale calculations showing convergence of TE figure of merit, $ZT$, with an increasing number of junctions. Using this model, numerical simulations are performed for the Bi$_2$Te$_3$ branched nanowire (BNW) and Cayley tree NW (CTNW) network. We find that the phonon scattering at the network junctions plays a dominant role in enhancing the network $ZT$. Specifically, disordered BNW and CTNW demonstrate an order of magnitude higher ZT enhancement compared to their ordered counterparts. Formation of preferential TE pathways in CTNW makes the network effectively behave as its BNW counterpart. We provide formalism for simulating large scale nanowire networks hinged upon experimentally measurable TE parameters of a single T-junction.