Combining the D3 dispersion correction with the neuroevolution machine-learned potential

Abstract

Machine-learned potentials (MLPs) have become a popular approach of modeling interatomic interactions in atomistic simulations, but to keep the computational cost under control, a relatively short cutoff must be imposed, which put serious restrictions on the capability of the MLPs for modeling relatively long-ranged dispersion interactions. In this paper, we propose to combine the neuroevolution potential (NEP) with the popular D3 correction to achieve a unified NEP-D3 model that can simultaneously model relatively short-ranged bonded interactions and relatively long-ranged dispersion interactions. We show that improved descriptions of the binding and sliding energies in bilayer graphene can be obtained by the NEP-D3 approach compared to the pure NEP approach. We implement the D3 part into the gpumd package such that it can be used out of the box for many exchange-correlation functionals. As a realistic application, we show that dispersion interactions result in approximately a 10% reduction in thermal conductivity for three typical metal-organic frameworks.

Semianisotropic Interfacial Potential for Interfaces between Metal and 2D Carbon Allotrope

“…The commonly used two-body Lennard-Jones (LJ) potential ( V LJ ( r ) = 4 ϵ [( σ / r ) 12 –( σ / r ) 6 ]) is found to fall short in capturing the anisotropic nature of the vdW interaction at the metal/graphene interface. ,, Another potential avenue involves the development of machine learning potentials (MLPs), which facilitate materials simulations at extended lengths and time scales while maintaining near-ab initio accuracy. Although MLPs lack a direct physical interpretation, their practicality and broad applications render them valuable tools in material study and design. , However, existing methods for constructing and training MLPs encounter a significant challenge in describing long-range van der Waals interactions while ensuring both high accuracy and efficiency. − Therefore, the development of an empirical force field that accurately describes the vdW interactions at the metal/2D carbon allotrope interface is still of paramount significance.…”