Download files
Abstract
Characteristics of irreversible protein residues on cleaned membranes are a crucial issue during membrane operations as the build-up of residues following repeated fouling and cleaning reduces membrane performance, purity and safety. In this study, the residual deposition of various proteins along a flat-sheet 30kDa Molecular Weight Cut-off (MWCO) polyethersulphone membrane was investigated. One Dimensional Sodium Dodecyl Sulphate ¬Polyacrylamide Gel Electrophoresis (10 SOS-PAGE) and Lowry method were utilized to determine cleaning effectiveness. Whey protein isolate, Bovine Serum Albumin (BSA) and Beta Lactoglobulin (BLg) solutions were used as foulants while Hydrochloric acid (HCI), Sodium Hydroxide (NaOH), and Protease M Amano enzyme were used as cleaners.
The deposition was generally arbitrary due to the random nature of attachment and aggregation of the protein molecules. However these fluctuations can be reduced when a spacer was inserted in the channel resting above the membrane. When recovering the membrane flux, sequential cleaning with HCI and NaOH was more effective than NaOH cleaning alone. The inclusion of Hel was beneficial for severely fouled membranes with high amount of residues. The highest flux recovered through sequential cleaning for repeated fouling and cleaning cycles was 88%, white protease cleaning alone recovered 77%. Chemical cleaning leaves behind residues under the 20kOa molecular weight however these proteins were undistinguished due to the limitations of 10 SDS-PAGE. On the other hand, protease cleaning leaves behind a distinct 3akOa protease residue on the surface.
The cross-flow ultrafiltration of BSA solutions at various conditions was studied using empirical mass transfer models and Computational Fluid Dynamics (CFO) simulations. The CFO simulates the steady state flux and wall concentrations based on varying physical properties. The mass transfer study estimates individual resistances, mass transfer coefficients. and diffusivities through various models.
The CFO shows good agreement between simulated and experimental fluxes over time however the simulation response to the varied conditions was only qualitatively similar. The effect of concentration polarization towards flux decline was high however it was underestimated in the CFO. While theoretical wall concentrations varied with pH, measured depositions were constant. The estimated diffusivities between CFD and mass transfer study concurred.