Amorphous silica membranes have molecular sieving properties for the separation of hydrogen from gas mixtures at high temperature. Consequently, they are considered to be applied in separation of a shifted syngas coming out of a water-gas-shift-reactor into the syngas and hydrogen. This separation is a key to an Integrated Gasification Combined Cycle (IGCC) plant, which would allow reducing the carbon footprint in power generation industry. The main objective of this thesis was to carry out a preliminary assessment of suitability of currently available amorphous silica membranes for this separation. However, the separation properties of amorphous silica membranes reported in the open literature vary by orders of magnitude. Therefore, in the first part of this thesis the separation properties of hypothetical silica membrane with different pore size distributions were predicted from first principles. Considering different possible gas transport mechanisms, it was concluded that gas transport in amorphous silica membranes is dominated by the activated and non-activated Knudsen diffusion. The activation energy for transport of different species was predicted using the concept of suction energy. Then, with arbitrary pore size distributions gas permeance of hypothetical silica membrane was predicted for different gas species. Since the pore size distribution of amorphous silica membrane cannot be known a priori, the developed model was used to determine the pore size distribution based on experimentally measured single gas permeances of three different species (kindly provided by Natural Resources Canada, CANMET Energy Technology Center (CETC) laboratory in Ottawa) by minimizing the error of the calculated permeance ratios with respect to the experimental values. The results indicate that, depending on how the objective function is defined, more than one pore size distribution can be found to satisfy the experimental permeance ratios. It is speculated that by increasing the number of experimentally determined permeances, a more unique pore size distribution for the tested silica membrane can be obtained. However, even at this early stage, the developed model provides a rational explanation for the effect of membrane densification on the properties of silica membranes. More specifically, a simultaneous decrease in membrane permeance and selectivity due to membrane densification, reported in the literature, is explained by shrinking the size of pores beyond a certain critical value, which depends on the kinetic diameter of gas molecules that are being separated. Comparing theoretically determined permeances, which match experimentally observed permeance ratios, revealed that the experimental permeances are considerably smaller than the theoretical values. The ratio of the two provided the basis for a scaling factor, a new concept that was introduced in this thesis. To simulate membrane module performance, a novel approach was introduced. More specifically, co- and counter-current flow configurations as well as cross-flow configuration were modeled by assuming no change in feed composition over an infinitesimally small element of membrane area. This led to a system of linear, rather than differential equations, which was readily solved numerically.

A literature survey was performed to evaluate the state-of-the-art membrane systems for CO₂/CH₄ separation which is critical in the natural gas industry. The systems that were reviewed included zeolite, carbon, polymeric, mixed matrix, amorphous silica, and supported ionic liquid membranes. Supported ionic liquid CO₂/CH₄ selective membranes were synthesized in our laboratory by applying room temperature ionic liquids (RTILs) to porous inorganic α-alumina supports. The supported ionic liquid membranes (SILMs) displayed CO₂ permeance of 1x10⁻⁹ to 3x10⁻⁸ mol m⁻² s⁻¹ Pa⁻¹ and CO₂/CH₄ selectivity of up to 50 which is comparable with the current polymeric separation systems. It is concluded that, although the RTIL membranes showed good CO₂/CH₄ selectivity, the CO₂ permeance was too low for industrial applications. A new type of SILM was prepared by dissolving 1-aminopyridinium iodide which contained amine functionality in other ionic liquids which improved the CO₂ permeance and selectivity of these membranes. The H₂ gas separation is an important process because it has many industrial applications in petroleum processing and chemical synthesis. Amorphous silica membranes for H₂ separation were prepared on hollow fiber (HF) inorganic supports using chemical vapor deposition (CVD) of tetraethyl orthosilicate (TEOS). These membranes exhibited good H₂ permeance on the order of 10⁻⁷ mol m⁻² s⁻¹ Pa⁻¹ together with H₂/CO₂ selectivity of over 100. The separation was achieved using a new hybrid intermediate layer that was developed by depositing a mesoporous silica layer on top of γ-alumina. / Master of Science