CROSS-FLOW MICROFILTRATION BEHAVIOR OF A DOUBLE-CHAIN CATIONIC SURFACTANT DISPERSION IN WATER .1. THE EFFECT OF PROCESS AND MEMBRANE-CHARACTERISTICS ON PERMEATE FLUX AND SURFACTANT REJECTION

Microfiltration of a double-chain cationic surfactant, dioctadecyldime thylammonium chloride, in water has been studied using 1 or 0.2 mum tr ack-etched polycarbonate membranes. A computer-controlled filter rig w as used, in which the process conditions of crossflow velocity u, tran smembrane pressure drop DELTAp and temperature T were changed independ ently. In addition, permeate flux J and in some cases surfactant conce ntration in the feed cf and permeate c(p), were monitored during filtr ation. The effects of process and material variables on the permeate f lux decay, steady-state permeate flux J(infinity) and the permeate sur factant concentration were assessed. It was found that fouling of the membrane by surfactant is very rapid (within tens of seconds) although high crossflow velocity, large membrane pore size and low feed surfac tant concentration reduced the rate of surfactant deposition. Steady-s tate permeate flux increases significantly with crossflow velocity pro vided the surfactant is in a rigid chain (gel) lamellar dispersion sta te (T < 48-degrees-C). The steady-state permeate flux J(infinity) deca ys linearly with In (c(f)) provided the feed concentration is less tha n the so-called gel concentration (c(g)) which is obtained by extrapol ating the linear portion of the curves J(infinity) vs In (c(f)) so tha t c(f) = c(g) when J(infinity) = 0. Although the gel concentration, wh ich is independent of pore size and the process variables, is found to be 20 g l-1, experiments conducted at feed concentration well above c (g) still yield a reasonable permeate flux, albeit at a reduced level. Surfactant rejection studies indicate that 0.2 mum membranes yield lo wer rejection than 1 mum membranes while rejection is greater at 30-de grees-C than at 60-degrees-C. Other process variables do not have any significant influence on rejection. The effectiveness of a membrane an d the processing conditions can be quantified by considering the varia tion of J with c(p)/c(f). It is found that a 1 mum membrane is more ef fective than a 0.2 mum membrane, crossflow velocity increases the effe ctiveness of the process, and that temperature has no significant infl uence.