Detection and prioritization of unknown compounds in water extracts from drinking water production.

The goal of this study is to enhance the identification of unknown compounds in LC chromatograms of untreated raw water used for drinking water production by (1) preconcentration of samples and (2) prioritization of the unknown peaks based on their bioactivity. Preconcentration was done by solid phase extraction (SPE) of 3L large volume (LV) groundwater samples or by 6-week passive sampler (PS) deployment in the groundwater, using divinylbenzene as active phase. Bioactivity of the samples was tested as a decrease in bioluminescence in the Allivibrio fischeri bioassay, which was chosen as a broad-scale bioanalytical tool responding to many different types of pollutants. Samples were collected at three different drinking water production stations with low or high degree of anthropogenic influence. At two stations, different groundwater inlets were sampled. At one station, samples were taken at different stages in the drinking water production process. SPE and PS extracts were used for target analysis and non-target screening, and for testing in the bioassay before and after high-resolution fractionation. More target compounds were detected in the concentrated LV and PS extracts than in a 1 mL direct injection of the water. As expected least compounds and lowest bioassay responses were detected in the raw water from the station with least anthropogenic influence. PS extracts gave much higher bioassay responses than the LV extracts. Both chemical and bioassay analysis of samples collected during subsequent steps in the drinking water production process confirmed the efficient removal of (bioactive) contaminants. In addition to the 4 peaks detected in the bioassay chromatogram of the reference station, the more anthropogenically influenced stations showed additional peaks indicative for the presence of anthropogenic substances. Identification of suspect compounds responsible for these additional peaks will proceed according to a three-step procedure, i.e. based on (1) exact mass and isotopic pattern of compounds available in different suspect lists, (2) fragment ions assigned to a precursor available in spectral libraries, and (3) estimated elemental composition. Reference standards will be obtained for the suspect compounds to confirm their retention time and bioactivity. Results from the identification process are available by the end of May 2020, and will be presented at the SETAC Europe conference.