Elimination of isothiazolinone biocides in reverse osmosis concentrate by ozonation: A two-phase kinetics and a non-linear surrogate model

Elimination of commercial Kathon biocide (methyl-isothiazolinone (MIT) and chloro-methyl-isothiazolinone (CMIT) mixture) by ozonation was investigated in real RO influent and concentrate. MIT and CMIT had different reactivities (second-order-rate-constants) with molecular ozone and (OH)-O-center dot. Ozonation of biocides followed an instantaneous phase (16.6 %-36.9 % contributions) and then a gradual phase (33.6 %-78.8 % contributions). Newly developed kinetics including both phases demonstrated that O-3 oxidation contributed 25.6 %-39.8 % and < 10 % of MIT and CMIT eliminations, respectively, and (OH)-O-center dot oxidation contributed 60.2 %-74.4 % and > 90 % of MIT and CMIT eliminations, respectively. (OH)-O-center dot oxidation at the instantaneous phase accounted 15.7 %-37.9 % of total (OH)-O-center dot oxidation. Mass ratios of O-3/DOC of 0.24 and 0.32 were needed for similar to 80 % eliminations of MIT and CMIT in RO concentrate, respectively. The kinetics including both phases allowed a para-chlorobenzoic acid indicator model to predict MIT and CMIT elimination better than that including gradual ozonation only, with 58.9 %-96.0 % lower relative error. The attenuations of electron-donating-moiety indicated that O-3 may preferentially react with chromophores through aromatic cleavage and electrophilic extraction, while (OH)-O-center dot may non-selectively react with chromophores through predominant electrophilic addition. A surrogate model for biocide elimination by UVA(254) loss was proposed to be nonlinear rather than linear, which reduced 31.8 %-71.3 % surrogating error.