LC-MS/MS QTOF analysis of river water identifies contaminants of environmental concern by non-targeted profiling

Significance of the topic

The rising diversity of contaminants of emerging concern (CECs) in urban aquatic systems poses challenges for environmental monitoring and public-health related research. Non-targeted analysis (NTA) coupled with high-resolution LC/QTOF workflows enables broad, hypothesis-free detection of pharmaceuticals, personal-care ingredients, industrial additives and agricultural inputs that are not captured by conventional targeted lists. Such profiling supports assessment of wastewater treatment performance, identification of diffuse and point-source inputs, and early warning of novel or increasingly prevalent urban pollutants.

Objectives and study overview

The study evaluated the application of a non-targeted LC/QTOF workflow and Insight Profiler software to river water grab samples from two London river systems (River Thames and Hogsmill River). Objectives were to:

• Demonstrate that NTA can discriminate sample groups and spatial trends across tidal and freshwater locations;
• Identify CECs (pharmaceuticals, metabolites, industrial additives) with high confidence using large spectral/suspect libraries;
• Assess the influence of wastewater treatment facility (WWTF) discharges and urban transport-related inputs on contaminant distributions.

Methodology

Sampling and sample preparation:

• Grab samples were collected at 12 sites across two river systems, including upstream and downstream locations relative to WWTF discharge points; samples included tidal (Thames) and freshwater (Hogsmill) contexts.
• Aliquots from each site were pooled to create a quality control (pooled QC) for batch performance monitoring.
• Sample cleanup involved centrifugal filtration using Millipore Ultrafree centrifugal filters (PVDF, 0.22 µm) prior to analysis.

Chromatography and mass spectrometry:

• Direct injection of 40 µL per analysis using a reversed-phase forensic toxicology gradient.
• Column: Shim-pack Velox Biphenyl (100 x 2.1 mm, 2.7 µm); column temperature 40 °C; flow rate 0.3 mL/min; total run time 17 min; mobile phase: methanol/water with 2 mM ammonium formate and 0.002% formic acid.
• LC instrument: Nexera X2.
• Mass spectrometer: High-resolution QTOF (LCMS-9050) operated with data-independent acquisition (DIA) MS/MS. MS scan m/z 100–1000 (100 ms); DIA MS/MS comprised consecutive windows (e.g., m/z 100–500 with 20–35 Da isolation widths) with ca. 25 ms per MS/MS scan and ~0.925 s total cycle time.

Data processing and identification:

• Insight Profiler software was used as a single, configurable processing workflow covering feature detection, alignment, statistical filtering and compound identification.
• Feature detection used low threshold/default settings to capture broad chemical space, followed by statistical filters to remove high variance features.
• Compound annotation leveraged suspect lists and spectral libraries (including Shimadzu Forensic Toxicology database and MassBank .msp libraries) with identification criteria such as 5 ppm mass accuracy and retention time match within ~1 min for high-confidence IDs.

Used instrumentation

• LC system: Shimadzu Nexera X2.
• Column: Shim-pack Velox Biphenyl (100 x 2.1 mm, 2.7 µm).
• MS: Shimadzu LCMS-9050 high-resolution QTOF operating in DIA-MS/MS mode.
• Sample preparation: Millipore Ultrafree centrifugal filters (PVDF, 0.22 µm).
• Software and libraries: Insight Profiler (non-targeted workflow), Shimadzu Forensic Toxicology database, MassBank .msp spectral libraries.

Main results and discussion

Key findings:

• Non-targeted profiling resolved distinct contaminant signatures between the two river systems and between sites relative to WWTF discharges. Samples downstream of WWTFs, particularly on the small urban Hogsmill tributary, showed markedly elevated levels of prescription pharmaceuticals and metabolites.
• Common wastewater markers such as acetaminophen (paracetamol) and caffeine were detected and observed to decrease post-treatment, indicating partial removal by WWTFs.
• A broad array of pharmaceuticals and metabolites were identified with high confidence, including venlafaxine, desmethylvenlafaxine, tramadol and O-desmethyltramadol, gabapentin, ketamine, clarithromycin, metronidazole, carbamazepine, lidocaine, tapentadol and apixaban. In the Hogsmill downstream sites, prescription compounds and metabolites comprised approximately 70% of detected positive IDs.
• Industrial additives and transport-related contaminants were detected; notably, hexa-methoxymethyl-melamine (HMMM), a polymer additive linked to road runoff, was identified at highest levels near urban/transport proximities in the Thames system.

Interpretation:

• The results demonstrate the capability of LC/QTOF-DIA combined with NTA software to link chemical signatures to likely sources (WWTF effluent, urban runoff), reveal compounds not routinely targeted, and detect both ubiquitous wastewater markers and emerging urban contaminants like HMMM.
• High-confidence spectral matches (mass accuracy, isotope pattern and retention time concordance) supported robust putative identifications, while pooled QCs allowed monitoring of analytical consistency across the batch.

Practical benefits and applications

• Provides environmental managers and researchers with a broad screening approach to detect known and previously unreported contaminants without requiring compound-specific methods.
• Supports wastewater-based epidemiology and spatial source apportionment by tracking marker compounds and complex mixtures across catchments.
• Enables rapid suspect screening using large libraries (MassBank, proprietary forensic lists) to prioritize contaminants for follow-up targeted quantitation and risk assessment.

Future trends and potential uses

• Integration of NTA with quantitative targeted follow-up will improve exposure and risk assessments for identified CECs, prioritizing those with persistence or high biological activity.
• Expanded and curated spectral libraries, machine-learning assisted spectral matching and retention time prediction will increase identification confidence and reduce false positives.
• Combining NTA data with hydrological and metadata (tidal cycles, discharge volumes, rainfall events) will refine source attribution and temporal dynamics of contaminant loads.
• Application of automated batch workflows (such as Insight Profiler) will facilitate routine monitoring programs and inter-laboratory comparability for urban water quality surveillance.

Conclusions

This study demonstrates that high-resolution LC/QTOF with DIA-MS/MS paired with an integrated non-targeted processing workflow can effectively profile complex urban river water. The approach distinguished spatial contamination patterns, identified a wide range of pharmaceuticals and industrial additives (including emerging contaminants like HMMM), and provided actionable information on WWTF influence and urban runoff contributions. Such NTA workflows are valuable tools to expand environmental monitoring beyond conventional targets and to guide subsequent targeted analyses and management actions.

References

• Barnes A., Armitage E.G., Ault J., Loftus N.J., Egli M., Rapp Wright H.L., Barron L. LC-MS/MS QTOF analysis of river water identifies contaminants of environmental concern by non-targeted profiling. Shimadzu Corporation / Imperial College London. WP 290. (Authors and affiliations as provided in the source document).