Degree Name

Doctor of Philosophy


School of Chemistry


Fresh water has become an increasingly scarce resource globally. However, the presence of various trace organic contaminants (TrOCs) in treated wastewater and other water sources remains a challenge to the operation of water reclamation and reuse systems for fresh water supply, especially TrOCs which are of increasing concerns with respect to their potential health effects. Membrane technology such as nanofiltration (NF) and reverse osmosis (RO) is likely to play a key role in removal of these organic contaminants.

In this thesis, a laboratory scale cross flow NF/RO cell membrane filtration system was used to investigate the effect of pH and temperature on the adsorption and rejection of trace organic contaminants from two membrane types: a NF270 nanofiltration membrane and an ESPA2 reverse osmosis membrane. Experiments were performed in both a background electrolyte and a buffered synthetic wastewater. The concentration of TrOCs in feed and permeate samples were analysed by gas chromatography - mass spectrometry (GC-MS) and liquid chromatography - mass spectrometry (LC-MS).

A study of the impact of factors and mechanisms on the adsorption and rejection of TrOCs under different pH conditions, employed sixteen TrOCs chosen from the major classes organic pollutants of endocrine disrupting chemicals (EDCs), pharmaceutically active compounds (PhACs) and disinfection by products (DBPs). A mass balance analysis showed that the correlation between adsorption and hydrophobicity (represented by the logarithm of the octanol-water distribution coefficient, log D) of the TrOCs observed in this study was rather weak after 24 hours of filtration at all experimental pH conditions of pH 4.7, pH 7 and pH 11, although many hydrophobic TrOCs (as reflected by log D ≥ 3) were observed to adsorb more significantly compared to many hydrophilic compounds (log D < 3). This was because the adsorption was not only dependent on hydrophobicity, but also on other physicochemical aspects of TrOCs and the membrane material such as molecular size, charge of the compounds, pore size, charge and surface roughness properties of the membranes. Therefore, it was suggested that these factors might also govern the adsorption (and subsequently rejection) of TrOCs to NF/RO membranes. The results also demonstrated that the rejection efficiency by the ESPA2 membrane of most of the TrOCs studied was considerably higher than for the NF270 membrane under all conditions, owing to its smaller pore size. Amongst these compounds, many negatively charged compounds were removed more efficiently than the neutral compounds for both the membranes, with electrostatic interaction being the main rejection mechanism. A good relationship between rejection of hydrophilic neutral and hydrophilic negatively charged compounds with their molecular weight (MW) for both the membranes was found at the experimental conditions of pH 4.7 and pH 7. However, this relationship was not apparent at pH 11, due to the contribution of rejection mechanisms such as size exclusion and electrostatic interaction at strong basic pH. For hydrophobic neutral and hydrophobic negatively charged compounds, there was no strong relationship between their rejection and molecular weight, suggesting that adsorption and electrostatic interaction at high pH were the overriding rejection mechanisms for these compounds.

In an investigation on the influence of the solution pHs (4, 7, and 11) on the rejection and adsorption behaviour of phytoestrogens, genistein and formononetin were used as representative compounds. Mass balance calculations indicated that significant adsorption of both phytoestrogens to the membranes occurred under all pH conditions, because of hydrogen bonding interactions between them and the membranes, the pore size of the NF270 membrane and surface roughness of the ESPA2 membrane. The rejection efficiency of the phytoestrogens by the ESPA2 membrane was considerably higher than for the NF270 membrane at all experimental pH conditions. For the NF270 membrane, at pH 4 and 7, the rejection of phytoestrogens decreased dramatically over the first 4 hours of operation and was relatively stable during the later stages of filtration, suggesting that size exclusion, adsorption and convection were the main rejection mechanisms for these compounds. By contrast at pH 11, there was only a slight reduction in the rejection of these compounds with time and that electrostatic repulsion became the overriding rejection mechanism. Conversely, the phytoestrogen rejection by the ESPA2 membrane was relatively stable at all pH conditions, which could be attributed to size exclusion being the dominant rejection mechanism.

A major study on the effects of feed temperature on membrane pore size and the rejection of twelve TrOCs (bezafibrate, sulfamethoxazole, trimethoprim, diclofenac, formononetin, genistein, pentachlorophenol, carbamazepine, primidone, caffeine, amitriptyline and linuron) by a NF270 membrane, filtration experiments were conducted at three different temperature conditions (20, 30 and 40 ºC). The membrane pore radius at each temperature was estimated using the pore-hindrance transport model and experimental data of the permeation of reference organic solutes (i.e. erythritol, xylose and glucose) through the membrane. The results suggested that the pore size of an NF membrane was dependent on the feed solution temperature. An increase in the feed temperature from 20 to 40 ºC led to an increase in the effective pore radius from 0.39 to 0.44 nm. Therefore, an increase in the feed temperature caused a considerable drop in the rejection of all TrOCs investigated in this study, and the rejection behaviour varied depending on the different physicochemical properties of the compounds. The decrease in rejection observed here could be attributed not only to the increase in the solute diffusivity but also the enlargement of the membrane pore size. As the feed temperature increased, the decrease in rejection of hydrophilic and hydrophobic neutral TrOCs was more severe than that of the hydrophilic negatively charged compounds with lower adsorption capabilities. This was because in addition to size exclusion (or steric hindrance) the rejection of these negatively charged TrOCs could also be driving by electrostatic interaction given that the membrane surface was also negatively charged. However, for hydrophilic negatively charged TrOCs with strong adsorption capacity onto the membrane, their rejection trends were very similar to that of hydrophilic and hydrophobic neutral TrOCs. The contribution of adsorption in addition to molecular size of the compounds (represented by their molecular widths or heights), were directly responsible for different rejection efficiencies between the hydrophilic and hydrophobic neutral TrOCs at different temperature conditions.