Untitled

In this work, we report about cellulose acetate nanofiltration (NF) membrane preparation according to the phase inversion process. Pore size was monitored by using dope solutions of two polymer concentrations (20 and 22 wt-%) and annealing temperatures from 60–80°C. Membrane characterization included a morphologic analysis using scanning electron microscopy (SEM) and hydraulic permeability (Lp0) determined from pure water filtration. SEM cross-section micrographs exhibited the expected asymmetric structure with more opened sub-layer morphology for nanofilters prepared using the lower annealing temperatures. The Lp0 values indicative of the pore size increased by decreasing the polymer concentration in dope solutions and the temperature during the annealing operation. Desalting efficiency of the obtained nanofilters was first evaluated through the filtration of synthetic solutions containing 500–2000 ppm of NaCl using a dead end process. The variation of salt rejection vs. the permeation rate allowed us to determine the operating conditions corresponding to the optimal material performances as shown by the plateau related to limit rejection. Desalination experiments for two Tunisian brackish waters with a salt content between 3500 and 4000 ppm were carried out under a transmembrane pressure of 16 bar. Annealed membranes at 75 and 80°C showed the better performances.

This study dealt with the preparation and characterization of NF–UF cellulose acetate-based membranes and their application to desalination of two Tunisian brackish waters The characterization by pure water permeability coefficient determination showed that this parameter varies in an inversely proportional way with annealing temperatures of membranes and that it reflects a narrowing of the average pore diameter and enhancing of the polymer cristallinity. Morphological study by SEM of membrane cross sections evidenced the anisotropic structure of prepared films and showed differences in their bulk structure depending on the annealing temperature and polymer concentration of the dope solutions. The study of variation curves of rejection rates of NaCl vs. the permeation fluxes threw into relief a plateaus zone representing an optimum between these parameters for R-20-80 and R-22-75-80 membranes. Nanofiltration tests revealed that these samples are well adapted to desalination needs of Tunisian brackish waters. The other samples should be successfully used in UF.