Kinetic Monte Carlo (KMC) simulations are used to study transport of guest molecules in a two-phase medium in which the minority phase forms closed regions. This type of model system resembles compositions of mixed-matrix membranes (MMMs) made of a matrix and imbedded filler particles with different permeabilities. Based on an ideal filler-matrix composite morphology as defined in [H. Vinh-Thang, S. Kaliaguine, Chem. Rev. 113, 4980, 2013], the effects of the filler-particle volume fraction, particle size, shape (aspect ratio), and the spatial particle distribution on gas transport through MMMs are addressed. The results obtained for nonoverlapping and randomly placed spherical filler particles are found to be in good agreement with the analytical models available in the literature and have proven their good accuracy also for thin membranes where finite-size effects are expected to play an appreciable role. Furthermore, the prominent influence of the aspect ratio for nonspherical particles on the effective permeability is shown and scenarios are discussed where the alignment of the asymmetric filler particles give rise to anisotropic transport properties potentially favorable for the performance of MMMs. The KMC approach developed guides the optimal spatial arrangement and orientation of the filler particles for the different strategies to increase the membrane permeability and separation selectivity by transport enhancement or inhibition.

Through IR microimaging the spatially and temporally resolved development of the CO₂ concentration in a ZIF-8 at 6FDA-DAM mixed matrix membrane (MMM) was visualized during transient adsorption. By recording the evolution of the CO₂ concentration, it is observed that the CO₂ molecules propagate from the ZIF-8 filler, which acts as a transport "highway", towards the surrounding polymer. A high-CO₂-concentration layer is formed at the MOF/polymer interface, which becomes more pronounced at higher CO₂ gas pressures. A microscopic explanation of the origins of this phenomenon is suggested by means of molecular modeling. By applying a computational methodology combining quantum and force-field based calculations, the formation of microvoids at the MOF/polymer interface is predicted. Grand canonical MonteCarlo simulations further demonstrate that CO₂ tends to preferentially reside in these microvoids, which is expected to facilitate CO₂ accumulation at the interface.