The mechanism of macromolecular ultrafiltration with microporous membranes is discussed, focusing on factors that control membrane flux and solute retention. Flux is limited by mass transfer conditions on the feed solution side of the membrane (concentration polarization). Solute retention is determined by geometric properties of the membrane pores and the macromolecules in solution, as well as concentration polarization.Ultrafiltration data for solutions of Dextran fractions and a Carbowax fraction are presented and compared with theory. Agreement for turbulent flow, laminar flow, and stirred cell ultrafiltration systems i s good.uring the past decade ult'rafiltratiori has been advanced D from a laboratory curiosity to a n important industrial unit operation. Practical applications include a broad spectrum of solution concentrations and/or fractionations. I n an ultrafiltration process ( Figure 1) a feed solution is introduced into a membrane unit, where solvent and certain solutes pass through the membrane under a n applied hydrost'atic pressure. Solutes unable to pass the membrane are retained, concentrated, and removed in solution. The porosity of the membrane determines, primarily on the basis of size, which solutes pass through and which are retained by the membrane. The pore structure of this molecular sieve is such that i t is inherently "nonplugging," and stable fluxes for long-term operation are achievable. RIany ultrafiltration applications involve the retent,ion of relatively high molecular weight solutes, accompanied by the removal t'hrough the membrane of lower molecular weight impurities. Of current interest are the coiiceiitratioii and purification of enzyme solutions, and the fracbioiiation of cheese whey for protein recovery (deFilippi and Goldsmith, 1970).For this t'ype of operation, the use of a high-flux membrane leads to low-pressure operation, frequently below 50 psi. As 110 phase change occurs during ultrafiltration, it offers several attractions. Energy requirements for concentration by ult,rafiltration, compared with those for evaporation, are relatively loiv. In addition, sensitive macrosolutes such as funct,ional proteills may be treated without denaturatiori. This paper examines ultrafilt,ration parameters which determine ultrafiltration rate and select'ivity. Data for turbulent flow, laminar flow, and stirred-cell membrane systems are presented and compared with theoretical predictions.
ExperimentalUltrafiltration experiments were 1)erformed with membranes in three different ultrafiltration systems. A flow schematic, identical for each system, is shown in Figure 2.Solutions from a feed reservoir were pumped through the ultrafiltration system. Inlet and/or outlet pressures were measured with pressure gages to within 0.5 psi, and the retentate flow was measured with a rotameter. Ultrafiltration rates were measured volumetrically. The operating pressure was controlled with a back-pressure regulator. 130th retentate and permeate solutions were recycled to the feed reservoir. The feed sol...
