Degree Name

Master of Engineering - Research


School of Civil, Mining, and Environmental Engineering


This dissertation aimed to examine and elucidate the effects of feed solution chemistry, including pH and ionic strength, and membrane fouling on the rejection of boron by different nanofiltration (NF) and reverse osmosis (RO) membranes. Membrane filtration experiments were conducted using a laboratory-scale cross-flow NF/RO membrane system and five different commercially available NF and RO membranes. The related rejection mechanisms were delineated on the basis of solution chemistry, membrane characteristics, and physiochemical properties of boric acid and model foulants. In every experiment, the rejection of conductivity and sodium were measured and compared with boron rejection data. The effects of membrane fouling on the performance of membrane system were evaluated by comparing the flux behaviours of the fouled membranes after 18 hour, and the quality of permeate produced by the virgin and fouled membranes.

Results reported here showed that solution pH appeared to be an important factor governing the rejection of boron by NF/RO membranes under both virgin and fouled conditions. Increasing solution pH led to an increase in boron rejection efficiency. The key role of solution pH in determining boron rejection efficiency was attributed to its influence on membrane charged properties and the speciation of boric acid, which could increase the rejection by both size exclusion and charged repulsion mechanisms. However, the influence of solution pH on membrane surface charged properties could be inhibited by the fouling layer, which resulted in a lower boron rejection by fouled membranes than that by virgin membranes under high pH conditions. Solution pH could also determine the impact of solution ionic strength on boron rejection efficiency. Increasing ionic strength was found to improve boron rejection in some certain conditions. However, this effect was observed only at pH values that higher than the intrinsic pKa of boric acid. The implication of this finding was that in double-pass membrane desalination systems where boron rejection was achieved at the second pass, the rejection can be improved by allowing a higher salt passage through the first pass.

Boron rejection by NF/RO membranes in fouled conditions revealed that different types of membrane fouling could induce different effects on the permeate flux behaviour and the separation of boron. Humic acid and sodium alginate fouling layer caused the most severe permeate flux decline, followed by silica colloid and CaSO4 scaling. Humic acid was the only model foulant investigated in this study could enhance boron rejection, especially at low solution pH. Possible factors that could influence boron rejection by fouled membranes included the extent of cake-enhanced concentration polarisation, charged property of foulant, and foulant‟s ability to clog the membrane surface.

Membrane pore size appeared to be a reliable parameter that determined the solute rejection efficiency of a membrane because membrane pore size was not affected by solution chemistry such as pH and ionic strength. Overall, solute rejection efficiency of a tighter membrane would be less affected by any changes in solution chemistry, and also less affected by membrane fouling conditions. This finding suggested a possibility of using NF membranes under high pH conditions to increase boron rejection for the second pass in double-pass membrane desalination plants. The utilisation of NF membrane could avoid over-demineralisation of the final product water, reduce energy consumption, and also decrease the potential of membrane scaling at high solution pH.