Analytical strategies for the investigation of organic micropollutants in aqueous matrices by target, suspect and non-target screening analysis

Abstract

The occurrence of organic micropollutants in raw water used for drinking water production and in the drinking water itself causes a constant monitoring of the water quality to comply with the quality requirements. Commonly applied targeted approaches are based on gas chromatography coupled to tandem mass spectrometry (GC-MS/MS) or liquid chromatography coupled to tandem mass spectrometry (LC-MS/MS), which have been established as sensitive and robust methods. Nevertheless, with these methods the recognition of unexpectedly emerging contaminants usually stays off. With the development and commercialization of high-resolution mass spectrometers (HRMS), comprehensive screening approaches can be applied, meaning that high numbers of micropollutants can be detected and HRMS data can be useful to identify unknown substances. Therefore, this thesis deals with the development and application of analytical strategies, including target, suspect and non-target screening analysis for the investigation of known and unknown organic micropollutants in aqueous samples. An analytical method by means of ultra-performance liquid chromatography coupled to an ion mobility quadrupole time-of-flight mass spectrometer (UPLC-IM-Q-TOF-MS) was developed for a sensitive detection of contaminants over a large polarity range. Additionally, for non-target screening analysis a data evaluation workflow was created. In order to use HRMS data beside a suspect or non-target screening, a quantitative screening method was developed and validated for the detection of more than 140 micropollutants. Especially for water suppliers, the quantitative information is important in order to have the information on the relevance of a substance and additionally, this information can be useful for the selection of validated target methods. The usage of collision cross section (CCS) values, which can, e.g., contribute to the identification during suspect or non-target screening approaches, was investigated in relation to the question how far CCS databases with CCS values derived by different instruments (drift tube IM-MS and traveling wave IM-MS) can be used reciprocally. This comparative study was based on measurements of more than 120 substances with both systems and revealed mean values of absolute percentage errors of 1.0% for [M+H]+ and 1.1% for [M+Na]+ ions, meaning that a good correlation of CCS values derived with both instrument types resulted. However, the high deviations up to 6.2% obtained in isolated cases showed that a full comparability between both instruments does not exist. Different LC and MS settings had no significant impact on the resulting CCS values of exemplarily tested substances. Only for karbutilate changes in the drift spectrum by drift tube IM-MS measurements were observed by using different LC settings, which could be caused by different protonation sites in the molecule. A drinking water production process was investigated by a combined analytical approach by means of target, suspect and non-target screening analysis. Beside the known, also unknown substances could be identified or tentatively identified by database searches. Since for many unknown substances, comparisons with databases failed, it was investigated on example of an unknown feature with a mass-to-charge-ratio (m/z) of 326.1751 encountered by UPLC-IM-Q-TOF-MS measurements of a river Rhine sample, how extensive this feature can be characterized or even identified. In a comprehensive structural characterization by means of different analytical methods it was found that aromatic structures as well as a primary or secondary amino function is probably contained in the structure of the unknown feature. However, a possible molecular structure could not be proposed, which demonstrates the difficulty in identification attempts of unknown micropollutants. Overall, this thesis demonstrated, that the combined approach of target, suspect and non-target screening analysis for the prioritization of known and unknown contaminants and identification of unknowns by databases can be successfully applied and has therefore a great potential for the future water monitoring. With regard to the identification of unknowns, extensions of the freely accessible databases are required in order to identify further unknown organic micropollutants more easily and quickly in further studies. Additionally, in further studies, the workflow, in particular those of non-target screening, should be further developed to be more feasible in a shorter time to enable that changes in raw and drinking water can be quickly recognized because the workflows developed so far require a lot of time.