Impact of groundwater quality and associated byproduct formation during UV/hydrogen peroxide treatment of 1,4-dioxane.

In this study, a semi-batch, bench-scale UV/hydrogen peroxide (UV/HO) advanced oxidation process system was used to investigate how typical groundwater quality parameters (pH, alkalinity, natural organic matter (NOM), nitrate, and iron) influence the treatment of 1,4-dioxane. Deionized (DI) water spiked with 1,4-dioxane (100 μg L), treated using HO (10 mg L) in a commercially available UV system (40 W low-pressure lamp) showed an UV fluence-based first-order rate constant (k') and electrical energy-per-order (EEO) of 4.32✕10 cm-mJ and 0.15 kWh-m-order, respectively. The most abundant byproduct generated in spiked-DI water was oxalic acid (up to 55 μg L), followed by formic and acetic acids. The k' showed no significant difference at pH ranging from 5 to 7 and at low alkalinity concentrations (<20 mg-CaCO L), typical of sandy aquifers. The k' declined by up to 85% with increasing NOM concentration. Elevated production (up to ∼400% increase) of aldehydes and organic acids was observed in NOM-spiked water, implying that NOM is a significant byproduct precursor during UV/HO treatment. High NO concentration (10 mg-N L) in source water reduced the k' by 25%, while no significant impact was observed at lower concentrations (<2 mg-N L). Addition of Fe(II) at 0.5 mg-L resulted in an instantaneous Fenton-reaction-assisted removal of ∼10% 1,4-dioxane in the presence of HO, but did not enhance the performance of UV/HO treatment over time. In contrast, both Fe(II) and Fe(III) addition lowered the k' by 15-27%. The decline of k' observed in these experiments was attributed to reduced UVT (Fe), OH radical scavenging (pH), or both (NO, NOM). Treatment of groundwater samples collected from three 1,4-dioxane-contaminated wells located in Long Island, NY, showed k' values of 13-40% lower than what was observed for DI water due to radical scavenging from a combination of high NO and NOM in the samples. A multiple linear-regression model, developed using water quality data as model input, showed good agreement with field observations (paired t-test: p > 0.05) in predicting k' for the removal of 1,4-dioxane from groundwater. This study provides the first systematic evaluation of the impacts of groundwater quality on UV/HO process to remove environmentally relevant levels of 1,4-dioxane and reports standardized performance-related parameters to aid in the design and evaluation of full-scale systems.