Title:	Understanding Thermal Transport in Energy Materials
DNr:	NAISS 2024/22-1573
Project Type:	NAISS Small Compute
Principal Investigator:	Ashis Kundu <ashis.kundu@liu.se>
Affiliation:	Linköpings universitet
Duration:	2025-01-10 – 2026-02-01
Classification:	10304
Keywords:

Abstract

Engineering thermal transport is a critical factor in determining the efficiency of energy materials, especially thermoelectric materials, which convert waste heat into electric energy. In thermoelectric materials, low thermal conductivity with high electrical performance is essential for achieving a high figure of merit (ZT). Among these, direct band gap semiconductors have attracted considerable attention as promising candidates for thermoelectrics because they are efficient in carrier transport and have excellent optical properties. The accurate modeling of thermal-transport properties in these materials faces significant challenges due to the complex interplay of phonon-phonon interactions and phonon renormalization.  Thermal transport in direct bandgap semiconductors will be explored in thermoelectric materials, within the advanced computational methods made possible by machine learning force fields (MLFF). With the traditional methods using density functional theory (DFT) or ab initio molecular dynamics (AIMD), computational expense hinders the proper description of anharmonicity and phonon scattering processes. On the contrary, MLFFs are a computationally cheap and scalable alternative that learns the interatomic potentials from quantum-mechanical training data, hence allowing us to study thermal transport at a fraction of the computational cost but without sacrificing accuracy. Utilizing MLFFs, we aim to explore thermal conductivities for different direct band gap semiconductors. We will investigate the impact of higher-order phonon processes, including four-phonon scattering and the effect of phonon renormalization which presume to play a crucial role in accurately determining thermal conductivity. Our work will provide a more detailed insight into the thermal transport mechanisms in promising thermoelectric materials and give a complete workflow to use MLFFs in the computational study of energy materials. This work will help in designing the next-generation thermoelectric materials with tailored thermal transport properties toward more efficient energy conversion technologies.