The Development Of Predictive Modeling Of Nanofiltration Membrane Properties : A Review

Abstract

Nanofiltration (NF) membranes, a relatively recent type of membranes, possess properties in between those of ultrafiltration (UF) and reverse osmosis (RO). The ionic transport mechanisms are therefore governed by both of steric and charge effects. This combined effect offers a value added to the membrane separation abilities, which covers almost all range of liquid-liquid separation system. For all type of unit operation such as membrane separation system, having a good predictive tool is certainly vital in process performance prediction, hence, process design and optimization. In NF separation system, however, the understanding of ionic transport mechanisms have not been fully understood. The previously developed predictive models have been revised and modified by other researchers intermittently. Two main approaches have been used thus far to model the transport of ionic species through NF membranes. One approach is through the Spiegler-Kedem model (Schirg and Widmer, 1992; Leveinstein et al., 1996). This black box approach allows the membrane to be characterized in terms of salt permeability (Ps) and reflection coefficient (s). The second approach describes the transport of ions in terms of an effective membrane thickness/porosity ?x/Ak, effective pore radius, rp, and effective membrane charge density, Xd . This paper reviews these two approaches including the assumptions used in model derivation, descriptions of mathematical formulation, descriptions of transport mechanisms and membrane pore size and effective charge density quantification. The weakness and strength of the two approaches are also highlighted. Of these two approaches, researchers were found to be more interested in applying model that was based on the extended Nernst-Planck equation. This was due to its ability to describe and illustrate the ionic transport mechanisms comprehensively and in more detail.