Exact heat flux formula and its spectral decomposition in molecular dynamics for arbitrary many-body potentials

Markos Poulos, Donatas Surblys, Konstantinos Termentzidis · 2025

Read Exact heat flux formula and its spectral decomposition in molecular dynamics for arbitrary many-body potentials by Markos Poulos, Donatas Surblys, Konstantinos Termentzidis (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2509-26044.txt.

Introduction

Markos Poulos, Donatas Surblys, and Konstantinos Termentzidis derive an exact formula for heat flow inside simulated materials, in this 2025 paper. Molecular dynamics tracks atoms one by one, and heat flux is computed from their energies and forces. Old formulas quietly assumed simple pair forces, so for the many-body force fields used in real materials they could misjudge heat conduction badly.

The new framework handles arbitrary many-body potentials and also splits the heat current by vibration frequency, showing which phonon modes carry the heat. Tests on two- and three-dimensional systems, including graphene and molybdenum disulfide sheets, matched the thermostat's own current under direct simulation, where older code strayed. The length dependence of each mode's contribution came into view as well. The method is built into the popular LAMMPS package for two common potentials.

Catalog record

CREATORMarkos Poulos, Donatas Surblys, Konstantinos Termentzidis
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE42 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2509-26044

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Exact heat flux formula and its spectral decomposition in molecular dynamics for arbitrary many-body potentials Markos Poulos, Donatas Surblys, Konstantinos Termentzidis Abstract In this study we have derived an exact framework for the calculation of the heat flux and its spectral decomposition in Molecular Dynamics (MD) for arbitrary many-body potentials. This work addresses several lacks and limitations of previous approaches and allows for the accurate computational study of thermal properties in a wide variety of many-body systems with MD. We have tested our modifications with Green-Kubo (GK) and Non-Equilibrium MD (NEMD) simulations for various 2D and 3D material systems using the Tersoff and Stillinger-Weber potentials as examples. The spectral decomposition of the heat current was also calculated for monolayer graphene (1LG) and MoS 2 , for different system lengths. Our results show that the heat current calculated by our method is consistently in agreement with the thermostat current in NEMD, while previous implementations can estimate quite poorly the thermal conductivity both under GK and NEMD simulations, and both for 2D and 3D materials. The decomposition of the heat current also sheds light on the contribution of different phonon modes to thermal conductivity and its dependence on length. Our methodology is implemented in the widely used LAMMPS code specifically for the Tersoff and SW potentials, and it is readily applicable to the vast majority of many-body MD potentials. I Introduction: Molecular Dynamics (MD) simulations are a powerful tool for studying the thermal properties of materials in the nanoscale de Sousa Oliveira and Neophytou 2019 ; Kinaci et al. 2012 ; Wei et al. 2019 ; Termentzidis et al. 2018 . Within MD there are two principal and widely used methods for calculating thermal properties, namely the Green-Kubo (GK) method and the Non-Equilibrium MD (NEMD) method. The GK method is a statistical technique that uses the autocorrelation function of the instantaneous heat flux of a system in equilibrium to calculate the thermal conductivity \kappa from the fluctuation-dissipation theorem Green 1954 ; Kubo 1957 , while NEMD is based on the application of a temperature gradient to a system by means of connecting thermal reservoirs at a temperature difference \Delta T . The resulting heat flux \bm{J} is calculated and, using Fourier’s law, thermal conductivity is determined as \kappa=-\bm{J}/\bm{\nabla}T (or sometimes seen with the gradient replaced by the average temperature slope \Delta T/L where L is the system length Hu et al. 2020 ). In both methods, the heat flux (HF) is a central quantity needed in order to calculate \kappa . For NEMD simulations, the HF is usually calculated by measuring the energy exchange rate between the thermostat regions and the rest of the system, while for GK the time-dependent HF is calculated as an average over the system volume from the instantaneous atomic velocities and forces acting on the atoms (see Section II for details). The instantaneous heat current Q^{1\rightarrow 2} flowing across an interface separating the system into regions 1 and 2 is also an important quantity, as it is used to calculate the Interfacial Thermal Conductance (ITC) of interfaces and heterostructures López-Suárez et al. 2018 ; Hu et al. 2011 ; Merabia and Termentzidis 2014 . Q^{1\rightarrow 2} can also be decomposed to reveal the contribution of phonon modes of specific energy both to the the thermal conductivity and ITC, known as the Spectral Decomposition of the Heat Current (SDHC) Sääskilahti et al. 2014 ; Sääskilahti et al. 2015 ; Sääskilahti et al. 2016 ; Chalopin and Volz 2013 . The calculation of Q^{1\rightarrow 2} and its spectral decomposition requires a definition of the mechanical power exerted by the interatomic forces that is due only to atoms from across the interface, which is trivial for pair potentials Sääskilahti et al. 2014 , but not evident for many-body interactions. LAMMPS is a widely used open-source MD code, employing a vast variety of interatomic potentials to simulate various kinds of interactions (covalent, ionic, Van der Waals, etc.), and capable of performing both NEMD and GK simulations Plimpton 1995 . For NEMD, the calculation of the HF from the thermostat energy exchange rate is straightforward and a priori exact. However, for equilibrium methods like GK it is based on the calculation of the atomic stress tensor, for which LAMMPS originally used a formula that is valid only for pair potentials. This can lead to incorrect results for \kappa . The correct formulation has been known for a long time, since Torii et al. theoretically derived exact formulas for calculating the HF both for a control volume and a control surface, for arbitrary many-body potentials and an arbitrary partitioning of the many-body potential energy among the participating atoms Torii et al. 2008 . The correct calculation of \kappa has been demonstrated for some cases, such as molecular systems using simple bond, angle and torsion terms Surblys et al. 2019 ; Boone et al. 2019 , carbon systems using the Tersoff potential Fan et al. 2015 , and recently for some Machine Learning potentials Hamakawa et al. 2025 . In these works, the authors have proposed and published modifications to the LAMMPS HF formula; Surblys et al . Surblys et al. 2019 and Boone et al . Boone et al. 2019 have published corrections to the atomic stress tensor for bond, angle and torsion terms, while Surblys et al also extended this formalism to systems with constraint dynamics Surblys et al. 2021 . Hamakawa et al . corrected the HF formula for some Machine Learning potentials Hamakawa et al. 2025 , while Fan et al . Fan et al. 2015 have published an implementation of an exact formula for the calculation of the HF in an in-house, open-source GPU MD code, which implements the Tersoff potential Fan et al. 2017a . Therefore, implementations of exact formulas for the HF in a control volume are available but with limited applicability to specific systems. Calculation of Q^{1\rightarrow 2} through a control surface is also available, but only for pair-wise potentials, while it completely lacks many-body implementations for both Q^{1\rightarrow 2} and the SDHC. As a result, the SDHC has been calculated in the past only for systems with simple pairwise interactions Sääskilahti et al. 2016 , or by approximating the many-body forces as pair harmonic forces, under the so-called harmonic approximation Chalopin and Volz 2013 ; Sääskilahti et al. 2014 . Thus in this work we have extended and implemented in LAMMPS the Torii et al . formulas Torii et al. 2008 for calculating the heat flux for both a control volume and a control surface, and subsequently its spectral decomposition, for arbitrary many-body potentials. The article is organized as follows: In Section II we present the theoretical background based on the derivations by Torii et al . Torii et al. 2008 for the calculation of the heat flux in a control volume, the heat current through a control surface, and the Spectral Decomposition of the Heat Current (SDHC). In Section III we present some results of comparative tests of our implementation with the existing LAMMPS implementation, for GK and NEMD simulations and for various test systems (both 2D and 3D solids) using the Tersoff and the Stillinger-Weber potentials, as well as some results of SDHC

Read the complete text