Characteristics of low and high SUVA precursors: Relationships among molecular weight, fluorescence, and chemical composition with DBP formation

Abstract

Disinfection by-products (DBPs) formed upon water treatment is an emerging issue worldwide. While monitoring of DBP precursors can easily be achieved for high specific UV absorbance (SUVA) organic (>6 L/mg.m), low and extremely low SUVA precursors (<2 L/mg.m) are difficult to monitor or even to predict their DBP formation behaviour. This study investigated the relationships among NOM characteristics, such as molecular weight (MW), fluorescence, and chemical composition, with DBP formation resulting from the chlorination of relatively high and low SUVA precursors. High SUVA precursors were formed by C-rich substances (82-85% of total mass) corresponding with high C/N and C/O (>100 and >5, respectively). Such precursors exhibited the fluorescence of long-wavelength humic-like signal and occurred at a high MW range (>30 kDa). By contrast, low SUVA precursors were either N-rich and/or O-rich substances, associated with much lower carbon content (40-70%). Low SUVA, N-rich precursors particularly also occurred at a high MW region (>100 kDa) and produced a strong protein-like fluorescence signal. When SUVA values of O-rich precursors were extremely low (<1 L/mg.m) they were accompanyied by short-wavelength humic-like fluorescence. During DBP tests, high SUVA produced only high yields of carbonaceous DBPs (e.g trichloromethane, haloacetic acids, haloketones), whereas low SUVA Nrich precursors yielded high levels of both C and NDBPs (e.g. haloacetonenitrile, chloropicrin). By contrast, extremely low SUVA precursors produced significantly lowlevels of both C and NDBPs (total <30 mu g/mgC). Furthermore, 19 of 20 regression models of DBP formation using log-transformed MW gave R-2 = 0.50-0.97. The strong regressions and correlations of NOM characteristics with DBPs in this study provide a better understanding of the influence of precursors characteristics on DBP monitoring, especially for low SUVA NOM. (C) 2020 Elsevier B.V. All rights reserved.