Large and increasing numbers of chemicals are produced and used by modern society. Many known and unknown chemicals are present in the many materials and products we use, some of which can pose hazards to human health and to the environment. According to a report by the United Nations Environment Programme (UNEP), of the more than 13,000 chemicals associated with plastics and plastic production, only 7,000 have been screened for their hazardous properties (UNEP 2023). Of those, 3,200 chemicals were identified to have hazardous properties of concern with many found across a wide range of sectors and product value chains. The number of chemicals found just in plastics has been recently raised to over 16,000 with at least 4,200 chemicals (26%) presenting high human and environmental hazard (Wagner et al., 2024), Methods capable of mapping all hazardous substances present in materials and products are needed for proper regulation to minimize risks for human and environmental health (i.e. for identification and prioritization of chemicals/chemical groups for regulatory risk management measures under e.g. the REACH and CLP regulations). Most chemical analysis is performed using methods targeting specific compounds with already known structures. This usually provides quantitative data for a relatively small number of compounds; however, advances in technology and data evaluation software now allow broader analyses beyond the target analysis of only a few pre-defined chemicals. Non-target screening (NTS) and suspect screening (SS) methods can theoretically provide a broader screening of potentially hazardous chemicals in materials/products, but whether these methods can in reality and with sufficient reliability identify novel hazardous substances needing regulatory action is unclear and has not been extensively evaluated. This project aims to provide an overview of the extent to which these methods have been used to identify regulatory-relevant substances in articles, chemical products, and recycled materials. Regulatory-relevant substances were, in this context defined, as substances that could be subject for additional risk management measures based on their hazard properties and regulatory status.
NTS is a chemical analytical approach using high-resolution mass spectrometry (HRMS) coupled with liquid or gas chromatography (LC or GC) for identification of organic chemicals not limited to predetermined chemical compounds for which optimized analysis methods have been created. SS is generally considered a subcategory of NTS where a broad list of compounds of concern is screened. Traditional targeted analysis is limited to a much smaller number of pre-selected target compounds, while NTS using HRMS can identify thousands of different mass spectra or chemical fingerprints (USEPA, 2024). Using software these fingerprints can then be compared to catalogued chemical spectra for identification and prioritized by various criteria such as presence or absence in control samples. In SS, molecular features are compared against specific chemical suspect lists or databases to identify potential matches. As NTS aims at generating signals from as many compounds as possible from a given sample matrix, more general sample preparation and chromatographic separation methods are applied.
There are currently methodological issues hampering NTS and SS applicability and efficiency for detecting and identifying new and relevant substances for regulatory purposes. One of the main problems is the often time-consuming task of curating the data and validating the confidence of the identifications. Even though sophisticated digital workflows have been devised by many research groups, much time have to be spent in manually inspecting the output data. A single NTS experiment often gives rise to thousands of “features” in the data set, e.g., regions in the data where a signal increases sufficiently above the background noise and that can be described as being typical for a chromatographic peak. A final quality check by manual inspection is often required, even if most data processing is performed automatically, to ensure that adequate settings have been used in the automatic processing. A single mass spectrum can also have numerous matches or no chemical match, and not every chemical present in a sample will exhibit a spectrum since sample preparation, separation, ionization, and instrument choice will affect the chemical space detectable (USEPA, 2024). Data processing methods and software can also add additional variation. Erroneous assignment of the chemical formula, adduct formation, isotopic patterns or chromatographic peak shapes as well as matching with spectra in databases can all contribute to a false positive or negative identification of compounds in a NTS or SS experiment (Hollender et al., 2023).
As the relationship between the signal intensity and concentration generally is not established in NTS and SS experiments, these techniques are considered as more qualitative than quantitative. Quantification of concentrations is important for risk assessment and improvements in semi-quantification techniques have been made over the recent years (Malm et al., 2021, 2024) but may not provide the level of certainty required by regulators. Absolute confirmation of NTS-identified chemicals still requires comparison to reference standards, which necessitates they are already known and available. Since NTS is still relatively new, methods and reporting are not yet thoroughly standardized even though quality control guidelines and suggested confidence levels have been published with varying levels of uptake (BP4NTA et al., 2023; Schymanski et al., 2014). At the same time, the number of collaborative trials, e.g., studies where several laboratories have used their developed analysis workflows on the same samples, that have been performed is low. This is probably due to the significant effort involved in the evaluation of NTS and SS screening data as compared to the application of traditional target chemical analysis. These studies are discussed in relation to the findings in this literature review in the discussion part of this report.
In this work, Schymanski levels (either provided by the study or assigned within this work) are used to characterize confidence levels of chemical hits. Schymanski levels are a five-tiered classification system used to communicate the level of confidence in compound identification using HRMS in NTS and SS (Schymanski et al., 2014). The levels were first proposed in 2014 and have since been widely adopted by the environmental science community to enhance transparency and comparability of results across different studies (Hollender et al., 2023; Manz et al., 2023; Schulze et al., 2020; Sobus et al., 2018). NTS relies on accurate mass and MS2 fragmentation patterns, along with other evidence, to tentatively identify compounds in complex samples. However, achieving definitive structural confirmation often requires reference standards as mentioned above, which are not always readily available for all compounds, particularly emerging contaminants and transformation products. The Schymanski levels provide a method to communicate the degree of certainty associated with compound identifications given these challenges within NTS and are described in Table 1.