An effective membrane separation process should have high flux (i.e., volume filtered per unit membrane surface area per unit time) and selectivity (i.e., passage of the desired species and rejection of undesired species). This dissertation examined two approaches, secondary membranes and lyotropic liquid crystal membranes, for improving flux and selectivity in aqueous liquid separations.

The first part of my work emphasizes the use of pre-deposited secondary membranes and backflushing for controlling membrane fouling in microfiltration and ultrafiltration of biological mixtures. Use of secondary membranes increased the permeate flux in microfiltration by several fold. Protein transmission is also enhanced due to the presence of the secondary membrane, and the amount of protein recovered is more than twice that obtained during filtration of protein-only solutions under otherwise identical conditions. In ultrafiltration, the flux enhancement due to secondary membranes is 50%, or less.

For the second part of my research, I developed and evaluated polymerized lyotropic liquid crystal (LLC) thin-film composite membranes. LLC assemblies provide an opportunity to make nanoporous polymer membranes with precise control over chemical and structural features on the nanometer scale, which is currently lacking in commercial reverse osmosis (RO) and nanofiltration (NF) membranes available today. These LLC composite membranes are prepared by photopolymerization of solution-cast films of LLC monomer on an ultrafiltration support membrane. These LLC membranes appeared to exhibit almost linearly increasing ionic rejection based on ionic diameter. LLC monomer was modified to achieve a 15% reduction in channel diameter, through the use of a larger multivalent Eu³⁺ cation as the carboxylate counterion. However, the monomers synthesized required use of solvents such as tetrahydrofuran, which resulted in the dissolution and damage of the support membranes used. Therefore, this direction for increasing ionic rejection was not pursued.

Solution-cast thin film bicontinuous cubic (Q_I) phase membranes were developed as proof-of-concept, in order to address alignment and low flux problems in H_II-phase LLC polymer membranes. Films with bicontinuous cubic LLC phases do not require alignment, have ca. 30° water contact angles, and have uniform 0.75 nm thick annular hydrophilic channels. Preliminary flux and ionic rejection properties of Qi-phase LLC membranes made by hot-pressing and solution-casting were studied. The hot-pressed Q_I membranes have a ca. 35 μm thick active layer while solution-cast Q_I-phase LLC membranes have a ca. 6-10 μm thick active layer. The permeability of the melt-processed Q_I-phase membranes was found to be ca. 4 x 10 ^-4 L m^-2 h^-1 μm bar^-1 and is similar to that of the commercial RO membranes. The Q_I-phase membranes exhibit greater than 95% rejection of small inorganic salts (such as NaCl, MgCl₂, Na₂SO₄) and organic solutes (such as glucose, sucrose, PEG-600).