This work presents a new approach to correcting for concentration polarization (CP) in pressure-driven membrane measurements. In the existing test cells (both cross-flow and stirred-batch) there are distributions of extent of CP over membrane surface. This complicates the interpretation of experimental data.A novel design of test cell with equally-accessible membrane surface has been developed based on the classical configuration of rotating disk combined with the possibility of applying trans-membrane hydrostatic pressure differences of up to 20. bar. Due to the equal accessibility, corrections for CP can easily be made even in multi-ionic systems, which would be much more difficult with other membrane test cells.Since the membrane has to be sealed at the edge the geometry somewhat deviates from the ideal case of infinite disk. The impact of these deviations has been quantified via CFD simulations. A major part of the membrane surface is shown to be equally accessible while there are some expectable deviations close to the sealed membrane edge. This zone could be “screened“ in the experiments. The approach could also be validated experimentally via studying the dependence of observed rejection on the rotation speed and demonstrating that intrinsic rejection was practically independent of it.Finally, to demonstrate the cell utility, we performed and interpreted a number of experiments using commercial NF270 membrane and various feed solutions (single salts and electrolyte mixtures). We conclude that this cell can be employed for systematic transport characterization of membranes and the obtained information can be used as input in the CFD modelling of membrane modules.
A number of CFD studies have demonstrated that there is a considerable inhomogeneity of extent of Concentration Polarization (CP) over the membrane surface especially in spacer-filled feed channels. However, the consequences of this inhomogeneity for the interpretation of measurements of solute rejection in pressure-driven membrane processes have received little attention.
This study uses a simple model of locally-1D CP combined with a postulated probability distribution of unstirred-layer thickness over the membrane thickness. In this way, we obtain transparent analytical results and can consider qualitative consequences of inhomogeneous distribution of CP over membrane surface. Our analysis shows that disregarding the CP distribution under-estimates the CP of strongly positively-rejected solutes and over-estimates the CP for the negatively-rejected ones. This observation is especially important for the interpretation of ion rejection from multi-ion solutions in nanofiltration where strong positive and pronounced negative rejections can occur simultaneously for solutes of different charges.
We conclude that for reliable interpretation of pressure-driven membrane measurements it is desirable to reduce the inhomogeneity of CP distribution to a minimum in membrane-testing devices
Wilson, D.; Florido, A.; Valderrama, C.; Fernandez de Labastida, M.; Alegret, S.; del Valle, M. International Symposium on Olfaction and Electronic Nose p. 98-99 DOI: 10.1063/1.3626320 Data de presentació: 2011-05-04 Presentació treball a congrés
Florido, A.; Vernet, J.; Fernandez de Labastida, M.; Valderrama, C.; Wilson, D.; del Valle, M. Ibero-American Congress on Sensors p. 671-675 Data de presentació: 2010-11-09 Presentació treball a congrés