Abstract
The anthropogenic entry of organic micropollutants into the aquatic environment leads to a potential risk for drinking water resources and the drinking water itself. Therefore, sensitive screening analysis methods are needed to monitor the raw and drinking water quality continuously. Non-target screening analysis has been shown to allow for a more comprehensive investigation of drinking water processes compared to target analysis alone. However, non-target screening is challenging due to the many features that can be detected. Thus, data processing techniques to reduce the high number of features are necessary, and prioritization techniques are important to find the features of interest for identification, as identification of unknown substances is challenging as well. In this study, a drinking water production process, where drinking water is supplied by a water reservoir, was investigated. Since the water reservoir provides surface water, which is anthropogenically influenced by wastewater treatment plant (WWTP) effluents, substances originating from WWTP effluents and reaching the drinking water were investigated, because this indicates that they cannot be removed by the drinking water production process. For this purpose, ultra-performance liquid chromatography coupled with an ion-mobility high-resolution mass spectrometer (UPLC-IM-HRMS) was used in a combined approach including target, suspect and non-target screening analysis to identify known and unknown substances. Additionally, the role of ion-mobility-derived collision cross sections (CCS) in identification is discussed. To that end, six samples (two WWTP effluent samples, a surface water sample that received the effluents, a raw water sample from a downstream water reservoir, a process sample and the drinking water) were analyzed. Positive findings for a total of 60 substances in at least one sample were obtained through quantitative screening. Sixty-five percent (15 out of 23) of the identified substances in the drinking water sample were pharmaceuticals and transformation products of pharmaceuticals. Using suspect screening, further 33 substances were tentatively identified in one or more samples, where for 19 of these substances, CCS values could be compared with CCS values from the literature, which supported the tentative identification. Eight substances were identified by reference standards. In the non-target screening, a total of ten features detected in all six samples were prioritized, whereby metoprolol acid/atenolol acid (a transformation product of the two β-blockers metoprolol and atenolol) and 1,3-benzothiazol-2-sulfonic acid (a transformation product of the vulcanization accelerator 2-mercaptobenzothiazole) were identified with reference standards. Overall, this study demonstrates the added value of a comprehensive water monitoring approach based on UPLC-IM-HRMS analysis.Graphical abstract
Highlights
In drinking water treatment processes, the removal of particles by flocculation, filtration or slow-sand filtration is commonly the first step of treatment
Liquid chromatography coupled with high-resolution mass spectrometry (LC-HRMS) has the potential to detect a broad spectrum of organic substances and has been used previously for the analysis of drinking water by target, suspect or non-target screening analysis [8,9,10,11,12]
A combined approach including a target, suspect and non-target screening analysis was applied for the investigation of a drinking water production process
Summary
In drinking water treatment processes, the removal of particles by flocculation, filtration or slow-sand filtration is commonly the first step of treatment. For extended security of identification, retention times, fragment ions, collision cross section (CCS) values (if determined) and isotope ratios can be used This approach allows for the screening of a large number of compounds without the need for reference standards [14]. One of the challenges in non-target screening analysis is the data processing, including peak detection, grouping of peaks which may belong to one compound (adducts, isotopes and in-source fragments), annotation or subtraction of blank peaks, and alignment of samples and sample replicates [16,17,18] Such steps are important to reduce the complexity of the data. Models for the prediction of the retention time in LCMS are reported, which can be used for an improved identification in a suspect and/or non-target screening approach [28, 29] With these approaches, fully unknown chemicals which are far unreported cannot be identified. Bader et al [31] developed a classification strategy for feature signals based on observed fold changes during the drinking water treatment process
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
