< Back

Physical Chemistry Chemical Physics

Temperature Profiles and Heat Fluxes Observed in Molecular Dynamics Simulations of Force-Driven Liquid Flows

Authors: Jafar Ghorbanian; Ali Beskok

Publication Date: -0001-11-30  Article ASAP

This paper concentrates on the unconventional temperature profiles and heat fluxes observed in non-equilibrium molecular dynamics (MD) simulations of force-driven liquid flows in nano-channels. Using MD simulations of liquid argon flows in gold nano-channels, we investigate manifestation of the first law of thermodynamics for the MD system, and compare it with that of the continuum fluid mechanics. While the energy equation for the continuum system results in heat conduction determined by viscous heating, the first law of thermodynamics in the MD system includes an additional slip-heating term. Interaction strength between argon and gold molecules are varied in order to investigate the effects of slip-velocity on the slip-heating term and the resulting temperature profiles. Heat fluxes and temperature profiles from “continuum”, “continuum augmented with slip-heating”, and “heat conduction due to power input by the driving force” are modeled and compared with the MD results. The continuum model can neither predict the heat fluxes nor the temperature profiles from MD simulations. While the continuum model augmented with slip-heating matches the MD heat fluxes, the resulting temperature profiles do not agree with the MD predictions. Overall the analytical model based on “heat conduction due to power input by the driving force” matches the heat fluxes from MD simulation, while the temperature profiles match MD predictions using an effective thermal conductivity that is about 70% of the thermodynamic value. Using different liquid-wall pair affects the slip velocity, temperature jump, and the resulting thermal conductivity of the fluid, but results in similar physical observations. Inability of MD method in mimicking continuum fluid mechanics in energy transport for force-driven liquid flows is scale independent, and it is more likely a numerical artifact.  Read more