Diffusion and flow in porous domains constructed using process-based and stochastic techniques

Two model mesoporous materials, namely Vycor glass and alumina membrane made by compaction of spherical monosize alumina particles, are reconstructed and the transport properties (Knudsen diffusivity, molecular diffusivity and permeability) of the resulting 3-dimensional binary domains are investigated through computer simulations. For the alumina membrane, two reconstruction alternatives are used, a ballistic deposition of spherical particles (process-based approach) and a stochastic procedure based on the porosity and the two-point correlation function of the porous matrix. For Vycor glass, only the stochastic reconstruction technique is employed. The obtained 3-D samples are structurally characterized in terms of their two-point correlation and chord length distribution functions. This type of information along with the comparisons between computed and reported transport coefficients indicate that the random sphere pack obtained from the ballistic-deposition procedure represents quite well the porous structure of the alumina membrane while the stochastically reconstructed domain fails to do so. In contrast, for the Vycor glass it is shown that the stochastic reconstruction technique is sufficient for a faithful representation of the porous matrix. It is furthermore argued that in the case of Vycor the stochastic reconstruction technique constitutes a kind of process-based approach.