ION SEPARATIONS BASED ON ELECTRICAL POTENTIALS IN NANOPOROUS AND MICROPOROUS MEMBRANESByJason A. Armstrong This dissertation examines several types of ion separations in nanometer to micrometer pores in membranes. Membranes provide an attractive platform for ion separations, primarily because they operate continuously (i.e. not in a batch mode), and small pores offer the potential for ion separation based on charge and electrophoretic mobility differences. Initial studies employed charged, nanoporous membranes to separate monovalent and divalent ions. Adsorption of polyelectrolyte multilayers in nanoporous membranes afforded control over the surface charge and pore radii in track-etched membranes, and electrostatic ion-exclusion, particularly for divalent ions, occurred in these membranes because the electrical double layer filled the entire nanopore. Initial experiments employed adsorption of (PSS/PAH) multilayers in the 50-nm diameter pores of PCTE membranes to give a K+/Mg2+ selectivity of ~10 in pressure-driven dead-end filtration. Adsorption of (PSS/PAH)1 films in 30-nm pores gave a similar K+/Mg2+selectivity with a simpler modification procedure. Separations utilizing (PSS/PAH)1 films in 30-nm pores showed the lowest ion rejections with high ion concentrations, consistent with enhanced screening of the electrical double layer at high ionic strength. However, solutions with < 5 mM ionic strength exhibited essentially 100% Mg2+ rejections (the Mg2+ concentration in the permeate was below the method detection limit). Moreover, K+ rejections increased in the presence of Mg2+, which may stem from Mg2+-adsorption within the PEM and increased surface charge. Finally, separation of Br- and SO42- with a PSS1-modified, 30-nm PCTE membrane validated the exclusion mechanism for anions. The average Br-/SO42- selectivity was 3.4 ± 0.8 for a solution containing 0.5 mM NaBr and 0.5 mM Na2SO4. The low selectivity in this case likely stems from a relatively large pore. The membranes used for the separation of monovalent and divalent ions also facilitated separation of monovalent ions (e.g. Li+ and Cs+), via a streaming-potential mechanism. In these separations, flow through a negatively charged membrane yields a positive (permeate minus feed) streaming potential, which retards the transport of a more mobile cation to a greater extent than transport of a less mobile cation. Thus, (PSS)1-modified, 30-nm PCTE membranes enabled Li+ and Cs+ separation, whereas (PSS-PAH)1-modified membranes separated acetate- and Br-. Cation selectivities were ~3 for solutions containing 1.5 mM Li2SO4 and 1.5 mM Cs2SO4, whereas anion selectivities were ~6 for 0.5 mM Mg(Acetate)2, 0.5 mM MgBr2. The streaming potential method gave only modest selectivities, however, and required low ion concentrations. Electrical potentials applied across microporous glass membranes also facilitate separation of monovalent ions with different electrophoretic mobilities. This dissertation describes a filtration cell with porous electrodes to enable cross-flow filtration with an applied potential. With the appropriate potentials, the cell afforded some separation of K+ and Li+, but the average selectivities were ~3. Moreover, the rejection of both ions plateaued near 90% at sufficiently high current to flow rate ratios. Buffer depletion or nonuniform cross-flow and electric fields may lead to membrane areas with low rejection and prevent high selectivities. Fabrication of a dual cross-flow cell (cross-flow on feed and permeate sides) limits buffer depletion issues and may provide higher monovalent ion selectivities.
Read
