Degree Name

Doctor of Philosophy


School of Civil, Mining and Environmental Engineering


The aim of this thesis was to investigate the impact of chemical cleaning on the rejection of trace contaminants (TrOC) by nanofiltration (NF) membranes. The impact of chemical cleaning on TrOC rejection was investigated by exposing three different virgin NF membrane samples - namely NF270, NF90 and TFC-SR100  to chemical cleaning solutions (i.e. alkalis, acids, surfactants, chelating reagents and a combination of these chemicals) of various strengths and temperatures. Filtration experiments of the chemically cleaned membrane samples were subsequently carried out in a laboratory-scale cross-flow membrane filtration system. The observed variations in rejection of TrOC were related to the variations of the membrane surface structure and properties, including charge, hydrophobicity, permeability and chemistry.

The results showed considerable impacts of chemical cleaning on the properties of the NF membranes as well as their capacity to reject TrOC. In this study, the observed changes in TrOC rejection were in good agreement with the changes in the membrane permeate flux. The predominant trend in rejection caused by chemical cleaning depends on the physicochemical properties of the TrOC, the nature of the cleaning solution (i.e. caustic or acidic) and the structure and composition of the membrane.

The rejection of negatively charged TrOC was the least affected by chemical cleaning in this study. This is because charge repulsion (which is the predominant rejection mechanism of negatively charged TrOC) was not affected by chemical cleaning. By contrast, the rejection of neutral TrOC by the NF270 membrane varied considerably due to chemical cleaning as a response to conformational rearrangements of polymeric chains in the membrane active skin layer. These conformational relocations of polymeric chains were mainly driven by variation in the membrane surface charge as a response to caustic and acidic cleaning. On the other hand, the impact of chemical cleaning on the solute rejection by the NF90 and TFC-SR100 membranes was insignificant, most likely because of their thicker and denser active skin layer. Caustic cleaning reagents increased the membrane permeability and the convective passage of neutral TrOC through the NF270 membrane most significantly among all cleaning reagents investigated in this study, particularly when surfactants and/or metal chelating reagents were incorporated in the caustic cleaning solution. In contrast, insignificant variations in the membrane performance were observed when the same metal chelating reagents and surfactants were used individually.

The observed variations in the membrane permeability and TrOC rejection due to caustic chemical cleaning are not permanent and can be recovered by subsequently rendering the membrane neutral (e.g. by subsequent acidic cleaning). Similarly, elevated cleaning temperature can compensate for the impact of caustic cleaning, since cyclic changes in the cleaning temperature tightened the membrane polymer significantly.

This study also explored the impact of repetitive fouling-cleaning cycles on the rejection of TrOC. Membrane fouling (i.e. by silica colloid, humic acid, alginate and secondary treated effluent) changed the solute rejection by pore blocking, modification of the membrane surface charge, and cake-enhanced concentration polarisation. Subsequent chemical cleaning using a commercial cleaning formulation was applied to recover the membrane flux. Although chemical cleaning effectively restored the membrane flux (over 100%), surface hydrophobicity and zeta potential was not recovered over multiple fouling-cleaning cycles. In agreement with this, microscopic investigations confirmed traces of remaining foulants on the membrane surface. The enlargement of the membrane pores due to caustic cleaning subsequently led to notable changes in the rejection of neutral TrOC, whereas repetitive fouling-cleaning cycles had negligible effects on the rejection of negatively charged TrOC.