Mechanisms of temperature-dependent thermal transport in amorphous silica from machine-learning molecular dynamics

Abstract

Amorphous silica (a-SiO${}_2$) is a foundational disordered material for which the thermal transport properties are important for various applications. To accurately model the interatomic interactions in classical molecular dynamics (MD) simulations of thermal transport in a-SiO${}_2$, we herein develop an accurate yet highly efficient machine-learned potential model that allowed us to generate a-SiO${}_2$ samples closely resembling experimentally produced ones. Using the homogeneous nonequilibrium MD method and a proper quantum-statistical correction to the classical MD results, quantitative agreement with experiments is achieved for the thermal conductivities of bulk and 190 nm-thick a-SiO${}_2$ films over a wide range of temperatures. To interrogate the thermal vibrations at different temperatures, we calculated the current correlation functions corresponding to the transverse acoustic (TA) and longitudinal acoustic (LA) collective vibrations. The results reveal that below the Ioffe-Regel crossover frequency, phonons as well-defined excitations, remain applicable in a-SiO${}_2$ and play a predominant role at low temperatures, resulting in a temperature-dependent increase in thermal conductivity. In the high-temperature region, more phonons are excited, accompanied by a more intense liquid-like diffusion event. We attribute the temperature-independent thermal conductivity in the high-temperature range of a-SiO${}_2$ to the collaborative involvement of excited phonon scattering and liquid-like diffusion in heat conduction. These findings provide physical insights into the thermal transport of a-SiO${}_2$ and are expected to be applied to a vast range of amorphous materials.

12 pages, 7 figures in main text; 15 pages, 12 figures in Supplemental Material

References in corpus (11)

• GPUMD: A package for constructing accurate machine-learned potentials and performing highly efficient atomistic simulations
• Improving the accuracy of the neuroevolution machine learning potential for multi-component systems
• The Ioffe-Regel criterion and diffusion of vibrations in random lattices
• Quantum-corrected thickness-dependent thermal conductivity in amorphous silicon predicted by machine-learning molecular dynamics simulations
• DYNASOR -- A tool for extracting dynamical structure factors and current correlation functions from molecular dynamics simulations
• Sub-micrometer phonon mean free paths in metal-organic frameworks revealed by machine-learning molecular dynamics simulations
• Accurate prediction of heat conductivity of water by a neuroevolution potential
• From Green-Kubo to the full Boltzmann kinetic approach to heat transport in crystals and glasses
• Hydrodynamic finite-size scaling of the thermal conductivity in glasses
• Variable thermal transport in black, blue, and violet phosphorene from extensive atomistic simulations with a neuroevolution potential
• Examining normal modes as fundamental heat carriers in amorphous solids: the case of amorphous silicon

Cited by in corpus (9)

• Molecular dynamics simulations of heat transport using machine-learned potentials: A mini review and tutorial on GPUMD with neuroevolution potentials
• Correcting force error-induced underestimation of lattice thermal conductivity in machine learning molecular dynamics
• Highly efficient path-integral molecular dynamics simulations with GPUMD using neuroevolution potentials: Case studies on thermal properties of materials
• PYSED: A tool for extracting kinetic-energy-weighted phonon dispersion and lifetime from molecular dynamics simulations
• Dissimilar thermal transport properties in $\kappa$-Ga${}_2$O${}_3$ and $\beta$-Ga${}_2$O${}_3$ revealed by machine-learning homogeneous nonequilibrium molecular dynamics simulations
• Probing the ideal limit of interfacial thermal conductance in two-dimensional van der Waals heterostructures
• Phonon coherence and minimum thermal conductivity in disordered superlattice
• Thermal Conductivity Calculation using Homogeneous Non-equilibrium Molecular Dynamics Simulation with Allegro
• Thermal conductivities of monolayer graphene oxide from machine learning molecular dynamics simulations