Interlaboratory development and proposition for a new quality control sample for chemical forensics analysis of chemical warfare agents

• Published in: Talanta open 2023-12, Vol.8, p.100249, Article 100249
• Main Authors: Säde, Solja, Delaporte, Grégoire, Fraga, Carlos G., Hakulinen, Hanna, Höjer Holmgren, Karin, Spiandore, Marie, Åstot, Crister, Akmeemana, Anuradha, de Bruin-Hoegée, Mirjam, Doward, Justin, Jacques, Karen, Kesah, Saiful Anuar, See, Samantha, Thomson, Stuart, Flinck, Jens, Kalliovirta, Leena, Vanninen, Paula
• Format: Article
• Language: English
• Subjects: Chemical forensic signatures; Chemical forensics; Chemometrics; Gas chromatography–mass spectrometry; Interlaboratory study; Quality control
• ISSN: 2666-8319; 2666-8319
• DOI: 10.1016/j.talo.2023.100249

•A new quality control sample was designed and tested.•No statistically significant differences in GC–MS results from 11 laboratories.•Statistically similar results obtained when using the same data processing software.•Quality control acceptance criteria are proposed. A new quality control (QC) test sample for gas chromatography–mass spectrometry (GC–MS) was created and analysed to test the comparability and repeatability of chemical forensics results within the Organisation for the Prohibition of Chemical Weapons (OPCW)–designated laboratories. The QC test sample was designed in collaboration between four laboratories and consists of 27 compounds which evaluate the performance of GC–MS instruments. This solution was analysed with GC–MS(EI) in 11 laboratories, seven of which were OPCW designated. The participating laboratories analysed the sample multiple times on consecutive days, as well as after the analysis of a set of complex matrix samples. Retention times, retention indices, peak areas, peak tailing values, signal-to-noise ratios, and isotope ratios were extracted from the GC–MS data, and statistical multivariate analyses with principal component analysis and Hotelling's T2-tests were conducted. The results from these analyses indicate that differences between GC–MS analyses by multiple laboratories were not statistically significant at the 5% level, as the approximate p-value for the null hypothesis of “no differences between the runs” was 0.69. However, similar data processing methods and data normalisation are essential for enabling the reliable comparison of chemical fingerprints between laboratories. A composition for the QC sample and criteria for acceptable GC–MS performance for chemical forensics are proposed. The composition and criteria differ from the currently used chemical weapons verification analysis QC sample by e.g. broadening the range for retention index calculations by addition of new alkane compounds, including new chemicals with concentrations close to the limit of detection (10–100 ng/ml), and including compounds with higher polarity to emulate real-life forensic samples. The proposed criteria include monitoring of retention indices, isotope ratios, peak tailing, signal-to-noise ratios, peak height, mass spectra, and sensitivity of the instrument. The new compounds and criteria will be the subject of future confidence building exercises to validate their relevancy on a large scale. [Display omitted]