168c Methodological Assessment of Nanofiltration Membrane Performance Based on Surface Properties

Jermey N.A. Matthews1, Hisham Mohamed2, Don Szarowski2, Kimberly L. Jones1, and James Turner2. (1) Civil Engineering, Howard University, 2300 6th Street NW, LK Downing Hall, Washington, DC 20059, (2) Wadsworth Center, New York State Department of Health, Empire State Plaza (ESP/E115), Albany, NY 12201

Recent advances in instrumentation and experimental techniques have allowed for investigation of the chemistry of materials at the molecular scale. The molecular level is fundamental to an understanding of material morphology that propagates to macro-scale performance (such as steric rejection of particles by a membrane) and as well as surface force interactions (such as electrostatic hindrance of ions by a charged membrane). The nanofiltration (NF) membrane is an ideal investigative material as it exhibits retention of low molecular weight solutes due to nanopores (0.1 to 10 nm) as well as multivalent ions due to electrostatics. The challenge for NF membrane technology is to optimize solvent flux and solute rejection while minimizing fouling of the membrane. This mission, however, is rife with contradiction, since increasing membrane rejection usually requires reducing the pore size, which in turn normally leads to increased fouling and overall loss of performance. Hence our goal is to assess the effect of various surface properties on membrane performance. We assess the surface charge density as a function of pH by electrokinetic analysis, the surface roughness by AFM, the hydrophilicity by measuring the water contact angle by the sessile drop method, and the chemical structure of the active layer. The results of these characterization probes are weighed against pure water flux, organic solute and divalent salt rejection, and flux loss due to fouling by protein (hemoglobin and BSA). The membrane materials chosen for the active layer were polyamide and cellulose acetate, and for the support layer – Celgard (polypropylene), polysulfone and poly (ether sulfone). The fabrication methods employed were spin casting/wet inversion and interfacial polymerization. We measured the flux of pure water and protein and the rejection of mono- and di- valent ions and uncharged organic solutes in high pressure stirred cell (dead end) and cross-flow systems. Results of the cellulose acetate membrane shows a specific water permeability value on the order of 10-8 m s-1 kPa-1, characteristic of nanofiltration membranes. The limiting rejection of uncharged organic solutes increased with increasing percentage of acetate, and this value was used to calculate the effective pore radius. The zeta potential was negative over the operating pH range for all membranes and became increasingly negative with increasing pH.