Configurational diffusion of small gas molecules in nanostructured materials: A computational and experimental study.

摘要

Nanostructured materials, such as nanoporous carbons (NPC) and nanoconfined polymers, are of great interest owing to the superior gas separation properties exhibited by them. The motion of gas molecules through such materials is described as configurational diffusion where the pore size of the material is commensurate with the size of the diffusing species. A theoretical description of transport of gas molecules through nanostructured materials is a daunting task due to the complex, tortuous pathways and non-uniform force fields inside the material. The goal of this dissertation is to provide a fundamental understanding of the mechanism of gas diffusion through such materials primarily using molecular simulations.;NPCs are important materials in separation industry and are used extensively for gas separation using pressure swing adsorption. Membranes fabricated from NPCs have been observed to exhibit high selectivity for several gases such as O2/N2 (30:1), He/N2 (178:1) and H 2/N2 (331:1). However, the molecular mechanism responsible for such high selectivities is not fully understood. There is no general consensus on whether entropic (size) effects or enthalpic (energy) effects are responsible for separation of gases in NPCs. We intend to develop a hierarchical molecular model for gas transport in NPCs to resolve this controversy.;In order to study the diffusion of gases through NPCs using simulation a realistic representation of their structure is required. We have developed a novel Monte Carlo (MC) algorithm to generate molecular models of NPCs using Gaussian polymer chains as the starting structure. The NPC models are characterized using properties such as pair distribution function, fraction of non-hexagonal rings and bond anisotropy map. The structural properties of NPC models generated using the MC algorithm are in good agreement with those observed experimentally for real NPCs. The NPC models are composed of curved and randomly oriented graphitic sheets.;Grand Canonical Monte Carlo (GCMC) simulations are carried out using two potentials, namely the Steele potential (developed for graphite) and an ab-initio potential (developed for C168 schwarzite having curved carbon surfaces) to represent the gas-carbon interactions. N2 and O2 sorption in our model NPCs show good agreement with experimental data in terms of the adsorption isotherm and the isosteric heat of adsorption. The adsorptive selectivity of oxygen over nitrogen is higher for Steele potential and increases with pressure/loading.;Molecular dynamics (MD) simulations of small molecule (N2 and O2) diffusion through the atomistic NPC models show that the self-diffusion coefficients of N2 and O2 are similar (i.e., the kinetic selectivity is close to unity). The overall oxygen-to-nitrogen selectivity (product of adsorptive and kinetic selectivities) is much higher when using Steele potential (compared to the ab-initio potential). Thus, we conclude that the Steele potential is more promising for simulating N2 and O2 transport in NPCs. The current MD simulations have been carried out at infinite dilution. The next logical step would be to implement the simulations at finite concentrations and under a pressure gradient in order to more accurately mimic the actual separation conditions.;Configurational diffusion in glassy, amorphous polymers is dictated by the structure and dynamics of the polymer since the penetrant has to travel through a tortuous pathway afforded by the free volume within the polymer while being in close contact with the polymer matrix. Experimental studies reported in literature have shown that altering the local structure of the polymer by confining it to a pore or by adding nanoparticles to it results in improved separation (permeability and selectivity) properties. To understand this phenomenon, the transport of helium and methane in atactic polypropylene (aPP) confined to a model slit-shaped pore is studied using MD simulation. "Mirror" boundary conditions are used at the pore wall to make is impermeable to the polymer and penetrant molecules while regular periodic boundary conditions are used in the other directions. Confinement is observed to modify the polymer density and backbone structure in the region near the pore wall. A comparison with transport through bulk aPP and through aPP confined to a rectangular pore showed that helium permeability increases with increasing degree of confinement while methane permeability remains relatively unchanged. Helium perm-selectivity over methane showed considerable enhancement on increasing the degree of confinement. The results from the simulations provide a logical explanation of the experimentally observed enhancement in separation properties of polymers upon nano-confinement.