Session: Research Posters

Paper Number: 173662

Pressure-Enhanced Lattice Thermal Transport in Graphite: Insights From Deep Neural Network Molecular Dynamics 

Layered van der Waals (vdW) materials such as graphite exhibit thermal transport behaviors that are crucial for a wide array of applications, including nanoelectronic devices, energy systems, and geophysical processes at elevated pressures. Owing to its highly anisotropic bonding—strong in-plane covalent bonds and weak cross-plane vdW interactions—graphite exhibits extreme directional disparity in thermal conductivity. While the in-plane thermal conductivity of graphite is exceptionally high under ambient conditions, the cross-plane conductivity remains significantly limited. However, recent studies suggest that this anisotropy can be substantially altered under high pressure, which reduces interlayer spacing, strengthens interlayer coupling, and modifies phonon dispersions—opening new regimes of thermal transport that challenge conventional models.

In this work, we employ molecular dynamics simulations based on deep neural network (DNN) interatomic potentials to investigate the evolution of lattice thermal transport in graphite under pressures from 0 to 70 GPa at 300 K. Our DNN potential, trained on extensive ab initio molecular dynamics data, enables accurate and computationally efficient modeling of pressure-dependent phonon properties far beyond the reach of traditional empirical potentials or first-principles simulations. Equilibrium molecular dynamics simulations are performed using the Green–Kubo formalism to compute the bulk-limit thermal conductivity, while spectral energy density and anharmonic lattice dynamics analyses reveal comprehensive phonon properties including phonon dispersion relations, group velocities, lifetimes, mode-resolved Grüneisen parameter, and three-phonon scattering space at finite pressures. These insights enable a deeper understanding of the underlying phonon transport mechanisms in graphite under compression.

Our results reveal that the cross-plane thermal conductivity of graphite increases by nearly two orders of magnitude with increasing pressure. This enhancement is primarily driven by the significant increase in enhanced cross-plane phonon group velocities and prolonged phonon lifetimes resulting from the suppression of anharmonic phonon scattering phase space. These pressure-induced changes are particularly pronounced along the Γ–A direction. In-plane thermal conductivity also rises, albeit more moderately (up to ~100%), due to increases in both phonon group velocities and lifetimes. The Grüneisen parameter analysis further confirms reduced anharmonicity in compressed graphite. In summary, our findings provide atomistic insights into how pressure fundamentally alters interatomic force constants and phonon dynamics in graphite.

This study advances the fundamental understanding of thermal transport in layered materials under external compression and suggests practical routes to tailor cross-plane thermal transport via pressure engineering. Our approach, combining DNN potentials with phonon-resolved analysis, establishes a robust framework for predicting thermal properties in a broad class of layered materials under high pressure.

Presenting Author: Haoran Cui University of Nevada

Presenting Author Biography: Haoran Cui is a 5th year PhD student at the University of Nevada, Reno

Authors:

Haoran Cui University of Nevada
John Crosby University of Nevada, Reno
Mehrab Lotfpour University of Nevada, Reno
Lei Cao University of Nevada, Reno
Yan Wang University of Nevada, Reno