Using fluorescence-parallel factor analysis for assessing disinfection by-product formation and natural organic matter removal efficiency in secondary treated synthetic drinking waters

Parallel factor (PARAFAC) analysis of fluorescence excitation-emission matrices (EEMs) was used to investigate the organic matter and DBP formation characteristics of untreated, primary treated (enhanced coagulation; EC) and secondary treated synthetic waters prepared using a Suwannee River natural organic matter (SR-NOM) isolate. The organic matter was characterised by four different fluorescence components; two humic acid-like (C1 and C2) and two protein-like (C3 and C4). Secondary treatment methods tested, following EC treatment, were; powdered activated carbon (PAC), granular activated carbon (GAC), 0.1% silver-impregnated activated carbon (SIAC), and MIEX® resin. Secondary treatments were more effective at removing natural organic matter (NOM) and fluorescent DBP-precursor components than EC alone. The formation of a suite of 17 DBPs including chlorinated, brominated and iodinated trihalomethanes (THMs), dihaloacetonitriles (DHANs), chloropropanones (CPs), chloral hydrate (CH) and trichloronitromethane (TCNM) was determined after chlorinating water sampled before and after each treatment step. Regression analysis was used to investigate the relationship between peak component fluorescence intensity (FMAX), DBP concentration and speciation, and more commonly used aggregate parameters such as DOC, UV254 and SUVA254. PARAFAC component 1 (C1) was in general a better predictor of DBP formation than other aggregate parameters, and was well correlated (R ≥ 0.80) with all detected DBPs except dibromochloromethane (DBCM) and dibromoacetonitrile (DBAN). These results indicate that the fluorescence-PARAFAC approach could provide a robust analytical tool for predicting DBP formation, and for evaluating the removal of NOM fractions relevant to DBP formation during water treatment.