(528b) Application of Nonwoven Membrane Filtration for the Separation and Recycling of Colloidal Particles

This study applies a non-woven membrane filtration system similar to the submerged-type membrane bioreactor (sMBR) for separating submicron inorganic particles from water. Non-woven microporous membranes are isotropic with fibrous structure, and hence are generally less expansive and experience lower trans-membrane pressure as compared with thin-film composite membranes. A number of studies have been reported in applying non-woven membranes in the form of a sMBR to treat wastewater with high BODs or CODs. However, applications of non-woven membranes for separating micrometer- or sub-micrometer-scale inorganic and biological particles have not been extensively studied. Therefore, this study aims to investigate the performance as well as the possible fouling conditions of a non-woven membrane mimicking sMBR to separate inorganic submicron SiO2 and Al2O3 particles from water.

In this study, a tempered glass reactor having a full working volume of 22.4 L was divided into two parts by a baffle wall, namely a mixing zone (16 L) and a membrane separation zone (6.4 L). Non-woven membranes (KNH Enterprise Co., Ltd, Taiwan) with two different nominal pore sizes at 0.4 ìm and 1.0 ìm were used. The membranes were supported by rigid spacers and were fabricated into flat-plate type configuration in which permeate can be drawn through the membrane by suction with a peristaltic pump. The model solutions were prepared by diluting commercialized polishing slurries containing either fumed silica or alumina particles having particle size ranges between 50 and 500 nm. In addition, to avoid biofilm growth during the experiments, an UV-C lamp and sodium hypochloride were been added in the reactor as a disinfectant. This method effectively eliminated biofilm interference that causes biofouling of the membranes. During each test run, the trans-membrane pressure and the permeate flux were recorded, and the permeate quality was determined by turbidity that was well-correlated with the actual solid concentration. Fouled membranes were recovered by a sequence of backwashing, NaOCl soaking, and ultrasonication (10 s). Typical over than 90% of the original specific flux could be attained after each recovery process. Furthermore, membrane autopsy analysis was conducted using scanning electron microscope (SEM) to study the surface blocking and internal fouling of the membrane.

Preliminary study indicated that the optimal filtration parameters between membrane pore size and particle size were related to the specific flux and the permeate quality. The effect of concentration of filtrate on the trans-membrane pressure, and thus the filtration permeability, was inconclusive at this point, as experiments using a wider range of particle concentrations were ongoing. We also observed that biofouling, if not sufficiently controlled, represented a major obstacle in assessing the degree of fouling by the particles in the solutions. The type and the extent of fouling as a function of specific permeate flux will eventually be verified with SEM analyses.