Studying the viscosity of methane fluid for different resolution levels models using Poiseuille flow in a nano-channel

Abstract

In this work, the viscosity–temperature–density relationship of methane is investigated for three fully atomistic models (RISM, OPLS and MOPLS) and the corresponding coarse-grained models (CGRISM, CGOPLS and CGMOPLS) by studying the Poiseuille flow inside a silicon nano-channel using molecular dynamic simulation. In order to solve the interaction problem of nano-channel atom and methane molecule, the coarse-grained methane model and the classical Berthelot–Lorentz mixing rules are employed. The optimized coarse-grained methane models are determined using the relative entropy minimization method. The density distribution, stress force profile and velocity profile of coarse-grained models are compared with the fully atomistic models for different nano-channel widths. It is concluded that the results of coarse-grained model are in reasonable agreement with those of the corresponding fully atomistic model. Furthermore, the value of viscosity is calculated by fitting the velocity profile to the continuum solution from the Navier–Stokes equations and then compared to experimental value. The results show that the MOPLS and CGMOPLS models could well predict the viscosity of methane fluid, while the OPLS model performed worst. All simulation results indicate that the coarse-grained models can be used to predict the viscosity of methane fluid accurately.