Gas diffusion on the graphene surface is a crucial process for many applications, such as gas sensors and membrane separation; however, its mechanism differs from that of a typical solid surface. It resembles diffusion in the bulk phase governed by molecular collision, but the corresponding surface diffusion coefficient is less than that predicted based on bulk theory due to the adsorption effect. In this study, molecular dynamics simulations are used to examine gas diffusion on graphene surfaces under different pressures and gas-graphene interactions, and a theoretical model of the diffusion coefficient of gases on the graphene surface relying only on atomic potential energy parameters is proposed as an extension of Chapman-Enskog equation. The diffusion coefficient of gases on the graphene surface depends on the surface adsorption number, and increased gas phase pressure or gas-graphene interaction can reduce the surface diffusion coefficient by enhancing surface adsorption. Good agreement between the surface diffusion coefficient predicted using a preliminary modified Chapman-Enskog equation using the surface adsorption number from the simulations and that directly obtained from the simulation demonstrates that the higher density near the surface is the reason for the surface diffusion coefficient deviating from the original Chapman-Enskog equation. To accurately estimate the surface adsorption number, the potential field of gases on the graphene surface is deduced using the Lennard-Jones atomic potential model, and the spatial density distribution of gases is obtained according to the Boltzmann distribution. The pure theoretical model for the diffusion coefficient of gases on the graphene surface is finally established by substituting the surface adsorption number estimated using the atomic energy potential parameters into the original Chapman-Enskog equation. With low or moderate surface adsorption strength, the proposed model can efficiently predict the diffusion coefficient of gases on the graphene surface, with relative errors of less than 10% compared with the simulation results, which are comparable to that of the original Chapman-Enskog equation predicting diffusion coefficient in bulk phase (<8%). Further comparison with the literature regarding the surface diffusion coefficient of actual gases CH4 and CO2 demonstrates the validity and practicability of the proposed model. This research elucidates the underlying effect of surface adsorption on the diffusion coefficient of gases on the graphene surface and achieves precise theoretical prediction of the corresponding surface diffusion coefficient depending only on atomic potential parameters, which has significant implications for the optimal design of gas -related graphene devices.
