Return to search

Mass transport studies in membrane filtration

First, a programme to model concentration polarisation was written using the finite difference approach developed by Ilias and Govind. It was validated with experiments using Centrisystem C-300 and C-400 cartridges and BSA solutions (1-5 g/L), and experimental data from Yeh and Cheng with an H1P30-20 Amicon cartridge and Dextran T-500. The next step was to incorporate fouling into this model. To calculate the configurational Derjaguin-Landau-Vervey-Overbeek (DLVO) forces and the resulting osmotic pressures, large use was made of the work of Bowen <I>et al</I>. Concentration dependent diffusivities were calculated from the generalised Stokes-Einstein equation, and used in the transport equation to describe the concentration polarisation profile. It was shown that, when the transport equation did not have a solution at the membrane (or membrane + cake) surface, and that concentration was greater than the highest-concentration local maximum for the diffusivity, coagulation would occur. In this case, a monolayer of globular protein was assumed to deposit, and concentration polarisation was recalculated with this additional resistance. Experiments with 1g/L BSA solutions and Amicon H1P30-20, for a range of transmembrane pressure, ionic strengths and pH, were compared with the model predictions. Both showed that fouling increased with ionic strength away from the Iso-Electric Point of BSA (IEP), and decreased with zeta potential. Simulation also showed that fouling could decrease with increased ionic strength around the IEP, in accordance with classical results. Total resistance to flux from experiments and simulation were in a similar range, although the lack of data relating zeta potentials and pH prevented further comparison. The model would also determine the critical pressure above which fouling occurred. However, observed values were significantly lower than predicted. Direct adsorption of the BSA onto the polysulfone membrane or the effect of high local pore velocities may both explain these discrepancies.

Identiferoai:union.ndltd.org:bl.uk/oai:ethos.bl.uk:657750
Date January 1998
CreatorsMignard, Dimitri
PublisherUniversity of Edinburgh
Source SetsEthos UK
Detected LanguageEnglish
TypeElectronic Thesis or Dissertation
Sourcehttp://hdl.handle.net/1842/12654

Page generated in 0.0096 seconds