ABSTRACT
Coagulation is often applied as a pre-treatment for membrane processes to reduce dissolved organic matter and to prevent membrane fouling. Biopolymers (BPs) have repeatedly been reported as major organic foulants, and coagulation conditions such as pH or dose have been optimised to minimise the remaining BPs. Optimisation however remains problematic because of the complex and heterogenetic nature of BP. In this study, the behaviour of several BP fractions in a coagulation process was investigated by excitation-emission matrix-parallel factor analysis (EEM-PARAFAC) following liquid chromatography (LC)-fractionation. Using a series of jar tests, we found that BP removal depends on the type of source water, reflecting differences in charge neutralisation conditions in three samples of natural water despite nearly identical processes for removing humic substances. This result demonstrates the complexity of optimisation for BP coagulation. Fractionation of EEM-PARAFAC to BP by LC showed that at least three organic component groups (C1, C2 and C3) constitute BP. C1 is tryptophan-like organic matter that is often found in wastewater effluent, C2 is tyrosine-like organic matter that has a phenolic chemical structure, and C3 is a humic-like substance. C1 was removed thoroughly at acidic pH but not at neutral pH, while the removal of C2 was inefficient even with a significant change in pH or dose, indicating similar difficulties in a coagulation process. The difference in components C1 and C2 may partly explain the difference in efficiencies of removal of BP in water from different sources. Our investigation suggests that the optimisation or selection of appropriate pre-treatment processes for membrane systems should be substantially based on the composition of BPs (e.g., C1 and C2 components).
