The
increased production of unconventional hydrocarbons emphasizes
the need to understand the transport of fluids through narrow pores.
Although it is well-known that confinement affects fluids structure
and transport, it is not yet possible to quantitatively predict properties
such as diffusivity as a function of pore width in the range of 1–50
nm. Such pores are commonly found not only in shale rocks but also
in a wide range of engineering materials, including catalysts. We
propose here a novel and computationally efficient methodology to
obtain accurate diffusion coefficient predictions as a function of
pore width for pores carved out of common materials, such as silica,
alumina, magnesium oxide, calcite, and muscovite. We implement atomistic
molecular dynamics (MD) simulations to quantify fluid structure and
transport within 5 nm-wide pores, with particular focus on the diffusion
coefficient within different pore regions. We then use these data
as input to a bespoke stochastic kinetic Monte Carlo (KMC) model,
developed to predict fluid transport in mesopores. The KMC model is
used to extrapolate the fluid diffusivity for pores of increasing
width. We validate the approach against atomistic MD simulation results
obtained for wider pores. When applied to supercritical methane in
slit-shaped pores, our methodology yields data within 10% of the atomistic
simulation results, with significant savings in computational time.
The proposed methodology, which combines the advantages of MD and
KMC simulations, is used to generate a digital library for the diffusivity
of gases as a function of pore chemistry and pore width and could
be relevant for a number of applications, from the prediction of hydrocarbon
transport in shale rocks to the optimization of catalysts, when surface-fluid
interactions impact transport.