(541c) Coupled Effects of Salt Concentration Polarization and Colloidal Deposition on the Performance of Reverse Osmosis Membranes

The goal of this study was to investigate the coupled effects of salt concentration polarization and colloidal deposition on reverse osmosis (RO) membrane fouling by colloid-bearing feed waters. The fouling can be due to residual particulate matter that passed the pretreatment stage or due to large colloidal loadings in cases when pretreatment fails or simply does not exist. By recognizing possible interferences between rejected salts and colloids present in the feed and/or deposited on the membrane surface, the proposed work addressed the phenomenon of coupling between colloidal deposition and concentration polarization as main factors limiting the application of RO membranes. We hypothesized that the dynamics of the flux decline in RO systems can be predicted based on the knowledge of interrelationships between colloidal fouling and concentration polarization effect and that this coupling is indeed an essential feature of the overall fouling process.

Filtration of model colloidal particle suspensions of different ionic strengths was conducted using a bench-scale crossflow filtration apparatus that comprised two flat-sheet RO membrane modules connected in parallel. The collected real-time data included permeate flux, permeate conductivity, feed suspension temperature, retentate flow rate, transmembrane pressure, and the mass of particles deposited on the membrane surface. Based on measurements of osmotic pressure and salt flux in the absence of colloidal foulants, the concentration of rejected salt at the membrane wall and membrane salt permeability constant (B) were experimentally determined. The constant B was then used to calculate the concentration of rejected salt at the membrane wall and the resulting osmotic pressure when colloidal particles were present in the feed. This approach allowed for a clear identification of individual contributions of salt concentration polarization and colloidal deposit formation to the permeate flux decline. The results unequivocally pointed to the importance of the two-way coupling between these two phenomena: formation of the colloidal deposit resulted in enhanced concentration polarization of the salt while the structure (effective porosity) of the colloidal deposits exhibited a strong dependence on the ionic strength.

Baseline rejection was recorded in the experiments when no colloidal particles were present in the feed. A short-term increase in salt rejection upon introduction of colloidal particles into the feed water was observed. We attributed this initial increase to the mixing of the salt concentration polarization layer by colloidal particles. The following decrease in rejection below the baseline value was attributed to the effect of the enhancement of concentration polarization by the deposited layer of colloidal particles.

Importantly, the observed initial increase in salt rejection in the presence of colloidal particles in the feed could be sustained over the long term when spacers were introduced into the feed channel. The observed transient behavior of salt rejection and permeate flux can be explained by taking into the account the complex structure of colloidal deposits that consist of a stagnant layer and a flowing layer. At the earlier stages of colloidal fouling, the mass of colloidal particles in the flowing layer is significant in comparison with the mass of particles in the growing stagnant layer. Under these conditions, mixing of the salt concentration polarization layer by the flowing layer of colloidal particles is relatively important with respect to the enhancement of the concentration polarization due to the formation of the stagnant colloidal layer. As the stagnant part of the deposit grows, the relative importance of mixing becomes increasing smaller and the rejection decreases. By introducing a spacer, the formation of the stagnant deposit is inhibited and mixing by colloidal particles in the flowing layer remains the only effect of particles on the salt concentration polarization. These findings point to new, perhaps counter-intuitive, possibilities for controlling the salt rejection and permeate flux by using spacers in conjunction with introducing particles with low deposition propensity into the feed.