Adsorption and diffusion in nanoporous materials from stochastic and process-based reconstruction techniques

Digitized binary domains representing the porous structure of an alumina membrane are produced using a process-based technique following an algorithm of ballistic deposition of non-overlapping monosize spheres. The diffusivity in the Knudsen regime of inert, point-like tracers, is determined using a random walk simulation procedure. Computed and experimental data show very good agreement. The porous structure of Vycor glass is generated by a stochastic method respecting statistical properties of the original structure such as the porosity and the two point correlation function. The Knudsen diffusivity of inert, point-like tracers is again determined as above and the agreement with the experiment is satisfactory. In a further step, both adsorption and desorption of CH2Br2 in porous Vycor glass are simulated and the relative permeability curve of the system is computed and agrees well with measured data. It is shown that the adsorption process in Vycor can be simulated by the very same stochastic procedure used for the reconstruction of the dry porous medium. The same approach leads also to the simulation of the desorption branch, in the absence of thermodynamical hysteresis, by taking under consideration the accessibility of certain sites of the porous structure to the gaseous phase.