Membrane separations provide a potentially attractive technology over conventional processes due to their advantages, such as low capital cost and energy consumption. The goal of this thesis is to design hybrid membranes that facilitate specific gas separations, especially olefin/paraffin separations. This thesis focuses on the designing dendrimer-based hybrid membranes on mesoporous alumina for reverse-selective separations, synthesizing Cu(I)-dendrimer hybrid membrane to facilitate olefin/paraffin separations, particularly ethylene/methane separation, and investigating the influence of solvent, stabilizing ligands on facilitated transport membrane. Reverse-selective gas separations have attracted considerable attention in removing the heavier/larger molecules from gas mixtures. In this study, dendrimer-based chemistry was proved to be an effective method by altering dendrimer structures and generations. G6-PIP, G4-AMP and G3-XDA are capable to fill the alumina mesopores and slight selectivity are observed. Facilitated transport membranes were made to increase the olefin/paraffin selectivity based on their chemical interaction with olefin molecules. Two approaches were explored, the first was to combine facilitator Cu(I) with dendrimer hybrid membrane to increase olefin permeance and olefin/paraffin selectivity simultaneously, and second was to facilitate transport membrane functionality by altering solvents and stabilizing ligands. Promising results were found by these two approaches, which were: 1) olefin/paraffin selectivity slightly increased by introducing facilitator Cu(I), 2) the interaction between Cu(I) and dendrimer functional groups are better known.

Propylene is one of the most important feedstocks of the petrochemical industry with an estimated 2015 worldwide demand of 100 million tons. Retrofitting conventional C3 splitters is highly desirable due to the huge amount of thermal energy required to separate propylene from propane. Membrane separation is among the alternatives that both academia and industry have actively studied during the past decades, however; many challenges remain to advance membrane separation as a scalable technology for energy-efficient propylene/propane separations. The overarching goal of this research is to provide a framework for development of scalable ZIF-based mixed-matrix membrane that is able to deliver attractive transport properties for advanced gas separations. Zeolitic imidazolate frameworks (ZIFs) were pursued instead of conventional molecular sieves (zeolites and carbon molecular sieves) to form mixed-matrix membrane due to their intrinsic compatibility with high Tg glassy polymers. A systematic study of adsorption and diffusion in zeolitic imidazolate framework-8 (ZIF-8) suggests that this material is remarkably kinetically selective for C3 and C4 hydrocarbons and therefore promising for membrane-based gas separation and adsorptive separation. As a result, ZIF-8 was used to form mixed-matrix dense film membranes with polyimide 6FDA-DAM at varied particle loadings and it was found that ZIF-8 significantly enhanced propylene/propane separation performance beyond the “permeability-selectivity” trade-off curve for polymeric materials. Eventually, this research advanced ZIF-based mixed-matrix membrane into a scalable technology by successfully forming high-loading dual-layer ZIF-8/6FDA-DAM asymmetric mixed-matrix hollow fiber membranes with attractive propylene/propane selectivity.