Multiscale study of anisotropic thermal expansion in β‑Ga2O3 and modulation of interface heat transfer conductance

Abstract

β-Ga₂O₃, with its ultra-wide bandgap and high breakdown electric field, holds great potential for power switching devices. However, its limited thermal conductivity under high-voltage and high-frequency conditions presents challenges to device performance. This study employs a machine learning-based deep potential (DP) potential to systematically investigate the anisotropic thermal properties of β-Ga₂O₃. By incorporating quantum effects via the quantum thermal bath method, we obtain specific heat values that align better with experimental results. The study further explores the anisotropy of the thermal expansion coefficient and its response to different strain conditions. Additionally, a β-Ga₂O₃(100)/SiC heterostructure is constructed to show that moderate tensile strain optimizes the interface structure and enhances interfacial thermal transport. This work provides a comprehensive theoretical framework for understanding the anisotropic thermodynamic behavior of β-Ga₂O₃ and offers valuable insights for optimizing thermal management in power electronics.