Ballistic transport from propagating vibrational modes in amorphous silicon dioxide: Thermal experiments and atomistic-machine learning modeling

Abstract

Understanding thermal transport in amorphous materials is critical for a wide range of applications, including buildings, vehicles, aerospace, and acoustic technologies. Despite its importance, the fundamental behavior of heat carriers in amorphous structures remains poorly understood and is often attributed to localized vibrational modes with mean free paths of about 1 nm, posing significant challenges for engineering their thermal functionalities. In this study, we present experimental measurements on mesoporous silica and atomistic analyses using Monte Carlo simulations and machine learning models to quantify the relationship between nanoarchitecture and effective thermal conductivity. Through rational chemical synthesis and ultrafast spectroscopy measurements, a strong size dependence within the sub-10 nm regime is observed, where the classical Fourier heat conduction theory fails to account for the effects of porosity and pore size. This deviation from diffusive transport is attributed to the significant contribution of propagating vibrational modes, in addition to non-propagating modes, revealing unexpectedly long mean free paths and ballistic thermal transport for heat carriers in amorphous silica. The fundamental vibrational modes in amorphous silica are further investigated using spectral-dependent Boltzmann transport equation simulations and molecular dynamics with machine learning potentials, showing good agreement with experimental results. This study provides valuable insights into nanoscale-modulated thermal transport properties in mesoporous silica and opens new opportunities for the rational design of thermally insulating materials.