Non-target screening of surface water samples to identify exposome-related pollutants: a case study from Luxembourg

Background Non-target screening of surface water samples collected over an extended period can reveal interesting temporal patterns in exposome-related pollutants. Additionally, geographical data on pollution sources close to the sampling sites, chemical classification data and the consideration of flow paths can provide valuable information on the origins and potential threat of tentatively identified chemical compounds. In this study, 271 surface water samples from 20 sampling sites across Luxembourg were analysed using high-resolution mass spectrometry, complementing routine target monitoring efforts in 2019–2022. Data analysis was performed using the open source R-package patRoon , which offers a customizable non-target workflow. By employing open source workflows featuring scoring terms, like spectral match and applying identification levels, tentative identifications can be prioritized, e.g. based on spectral similarity. Furthermore, by utilizing supplementary database information such as PubChemLite annotation categories and classification software such as classyFire , an overall assessment of the potential threats posed by the tentatively identified chemicals was conducted, enabling the prioritization of chemicals for future confirmation through targeted approaches. Results The study tentatively identified 378 compounds associated with the exposome including benzenoids, organoheterocyclic compounds, and organic phosphoric acids and derivatives (11 classyFire superclasses, 50 subclasses). The classification analysis not only revealed temporal variations in agrochemicals, with the majority of identifications occurring in May to July, but also highlighted the prevalence of pharmaceuticals such as venlafaxine in surface waters. Furthermore, potential sources of pollutants, like metallurgic industry or household products were explored by considering common uses and geographical information, as commercial uses of almost 100% of the identified chemicals are known. 41 chemicals were suggested for potential inclusion to governmental monitoring lists for further investigation. Conclusions The findings of this study complement existing knowledge on the pollution status of surface water in Luxembourg and highlight the usefulness of non-target screening for identifying temporal and spatial trends in pollutant levels. This approach, performed in a complementary manner to routine monitoring, can help to tentatively identify chemicals of concern for potential inclusion in target monitoring methods following additional confirmation and quantification efforts.