Tracking Br-DBPs and bromine substitution factors by two-stage differential characterization of water matrix and NOM during chlorination

Abstract

Tracking disinfection byproducts (DBPs) upon chlorination becomes challenging whenever bromides occur because they alter DBP formation from chlorinated to brominated DBPs (Br-DBPs). Although epidemiological studies have emphasized the need to control Br-DBPs because of their notable carcinogen effects, work on Br-DBP monitoring is timely, since limited attention has to date been paid to this. This study proposes a two-stage differential characterization approach to a water matrix and natural organic matter (NOM) surrogates as a more robust way to trace Br-DBP precursors and bromine substitution factors (BSF). This two-stage approach monitors the differential variations (Δ) of both inorganic and organic precursors, e.g., dissolved organic carbon (DOC), bromide levels, Cl2 content, specific ultraviolet absorbance (SUVA), fluorescent spectra, before and after chlorination. To mimic practical conditions, chlorination experiments were conducted with two principle NOMs, including Suwannee River fulvic (SW) and M. aeruginosa organic matter (MA), with Br− levels from 0 to 2000 μg/L and chlorination times from 15 min-24 h. When Br− levels increased, all Br-DBP yields and BSF values correspondingly magnified, accounted for 70–90% of the total DBP yields from chlorinated SW or MA water. At each Br− level, prolonging chlorination time also increased the overall Br-DBPs and varied BSF values, but these differences were of no significance (p > 0.05). Correlation tests highlighted the significant insufficiency of the conventionally-used DBP surrogates (initial DOC, Cl2 dose, and SUVA) for Br-DBPs and BSF monitoring. By contrast, our two-stage characterization parameters are superior, with correlation coefficients ρ > 0.8 (p < 0.01), especially for those parameters including ΔBr as a numerator. Accordingly, regression models of Br-DBPs and BSFs built by our two-stage tertiary parameters exhibited high adjusted R2 values mostly >0.7 and >0.8 (n = 60), indicating a promising application for effective Br-DBPs and BSF monitoring.