Development of novel nanostructured polymer membranes based on polymerizable lyotropic liquid crystals for aqueous molecular-size selective separations

Recently, the development of polymer membranes with uniform pores in the nanometer-size regime has been of great interest because such membranes should be capable of extremely fine separations based on differences in the molecular size of the substrates. In addition, such membranes may also enable new ways of separating and detecting analytes for applications such as chiral separations and single-molecule sensing. Methods such as gold plating as well as a number of molecular self-assembly techniques (e.g., thermotropic liquid crystals, molecular squares, phase-separated block copolymers), have been employed to produce membranes with well-defined, uniform nanopores. A number of the nanoporous polymers made using these techniques have been shown to be capable of molecular-size selective gas or liquid phase separations.;This thesis describes investigations into the use and development of polymerizable ionic lyotropic liquid crystals (LLCs) as a novel means of generating nanoporous polymer membranes for molecular size-selective aqueous nanofiltration (NF). LLCs are typically amphiphilic molecules, with a long extended hydrophobic tail or tails, plus a highly hydrophilic headgroup (usually ionic). The shape and amphiphilic character of LLC molecules encourage them to self-assemble into highly ordered, phase-segregated assemblies with uniform, aqueous domains in the 1-10 nm range, with the polar hydrophilic headgroups localized exclusively at the hydrophilic/hydrophobic interface. The open aqueous domains vary in structure from lamellae to cylinders to three-dimensionally interconnected channels. These LLC assemblies (especially those with the inverted hexagonal (HII) phase containing ordered 1-D cylindrical nanopores, and the bicontinuous cubic (QI or QII) phases containing 3-D interconnected water nanodomains) would be ideal aqueous molecular separation media if their nanostructures could be (1) stabilized through covalent cross-linking, and (2) successfully fabricated into viable membrane film configurations. The successful formation of NF membranes based on polymerized ionic LLC assemblies is unprecedented.;This thesis describes work over the last 4 years to develop the first working examples of nanoporous LLC polymer membranes and investigate the factors that affect their flux and selectivity performance.;First, first-generation LLC membranes with the HII nanostructure using LLC monomer 1 were developed. In order to achieve higher water flux and better mechanical properties than the free-standing LLC bulk films, a solution-casting method was developed to prepare supported H II-phase membranes using commercial ultrafiltration polysulfone (PSf) membranes. Due to severe oxygen inhibition during radical photocross-linking, a study on controlling photopolymerization of HII-phase thin films to minimize oxygen inhibition and to achieve high degrees of polymerization was untaken. An optimal polymerization procedure using a specially designed photopolymerization cell for thin films (less than one micron thick) to achieve >90 % degrees of polymerization was developed. Furthermore, the potential of these supported HII-phase membranes in molecular size-selective, aqueous NF was evaluated by using various water-soluble probe molecules, such as rigid, charged molecules; monodisperse, flexible poly(ethylene glycol)s (PEGs); and cyclodextrins (CDs). It was found that these membranes were capable of completely rejecting water-soluble, rigid molecules with sizes greater than or equal to the nanopore diameter (ca. 1.2 nm). Also, they exhibited better molecular sieving properties than blank PSf supports as well as isotropic LLC membranes made from LLC monomer 1 when relatively flexible probe molecules (PEGs and CDs) were used. These results suggest that the HII nanostructure plays an important role in the molecular-size selective separation performance. However, one major disadvantage of these supported HII-phase membranes is extremely low water flux, which is mainly due to the non-uniform alignment of the uniaxial hexagonal domains in the membranes. This factor greatly limits their use in practical applications.;Second, another type of LLC membrane with a type I Q (QI) phase that does not require uniform nanopore alignment (and thus potentially has higher water flux compared with HII-phase membranes), was also developed. Unlike the HII membranes, this QI-phase LLC assembly contains two interpenetrating organic networks separated from one another by a continuous ultra-thin water layer surface, with overall cubic symmetry. A modified hot-infusion method was developed to prepare the first example of supported QI-phase membranes using gemini diene-tail LLC monomer 2, because the prior solution-casting method was not able to generate the QI-phase structure. Using this new processing method, a degree of diene polymerization of above 95% and a highly ordered QI-phase structure were achieved. In addition, the aqueous filtration performance and the effective pore size of these novel LLC membranes were also evaluated using various water-soluble probe molecules. As expected, these supported QI-phase membranes showed greatly improved water flux. In addition, the filtration results revealed that these supported QI -phase membranes have a ca. 0.75 nm gap spacing or effective molecular-size cut-off and unprecedented water desalination performance. This novel LLC membrane exhibited 95--99.9% rejection of dissolved atomic salts, neutral molecules, and molecular ions, as well as a thickness- and pressure-normalized water permeability comparable to the active layer materials in commercial RO membranes. To our knowledge, the design and demonstration of a water desalination material with a well-defined, nanostructured water transport and ion size rejection manifold based on Q-phase LLCs are completely unprecedented.