Defining valid criteria for the automatic identification of microplastics by micro-Raman using various spectrometers

The characteristics of plastic materials that motivate their commercial success, namely low cost and high chemical and physical resistance, and the inadequate disposal of these cheap materials, are responsible for the ubiquitous presence of lasting plastic particles in the environment. Plastic materials' type, size and shape affect their impact on the environment, where smaller particles can even travel inside living organisms with unknown impacts in these systems. Plastic particles can be classified according to their size as Macroplastics (> 25 mm), Mesoplastics (5 mm to 25 mm), Microplastics (1 µm to 5 mm) and Nanoplastics (< 1 µm) [1]. Parallel to determining the toxicological impact of these particulates, it is necessary to determine their abundance values and respective trends and the most relevant sources of these materials. This information should support setting up and monitoring the policies to reduce the impact of plastics on the environment and human health. The harmonisation of procedures for monitoring plastic contamination in food products and environmental matrices is still lacking. Additionally, development on how to assess the performance and evaluate the uncertainty of these monitoring is also needed [2,3]. Before counting plastic particles in a food or environmental sample, it is necessary to identify them. Micro-FTIR and micro-Raman are the most popular tools for plastic particle identification before characterising their size and shape. Reference and particle spectra are compared manually or automatically, where manual identifications are time-consuming and must be performed by qualified analysts, while automatic identifications are fast and do not require so much analyst expertise. Given the number of plastic particles observed in some samples, automatic identifications are thus advised. The automatic identification of particles requires quantifying the correlation or match between reference and particle spectra and defining a minimum match above which identification has an adequately high and low true and false positive results rate, respectively. The match threshold can be defined for reference and particle spectra collected using the same or different equipment, the diversity of equipment being an additional challenge for the analyst. This communication describes the development and validation of a procedure for the automatic identification of polyethylene terephthalate (PET) microplastics by micro-Raman fit for identifications supported by reference and particle spectra collected in different spectrometers. Identifications are considered valid if associated with a true positive rate not lower than 95% and false positive rates not larger than 5%. References: [1] Hartmann, N.B., Hüffer, T., Thompson, R.C., Hassellöv, M., Verschoor, A., Daugaard, A.E., Rist, S., Karlsson, T., Brennholt, N., Cole, M., Herrling, M.P., Hess, M.C., Ivleva, N.P., Lusher, A.L., Wagner, M., 2019. Are We Speaking the Same Language? Recommendations for a Definition and Categorisation Framework for Plastic Debris, Environ. Sci. Technol. 53(3), 1039–1047. [2] Morgado, V., Gomes, L., Bettencourt da Silva, R.J.N, Palma, C., Validated spreadsheet for the identification of PE, PET, PP and PS microplastics by micro-ATR-FTIR spectra with known uncertainty, Talanta 234 (2021) 122624. [3] Morgado, V., Palma, C., Bettencourt da Silva, R.J.N, Microplastics identification by Infrared spectroscopy – Evaluation of identification criteria and uncertainty by the Bootstrap method, Talanta 224 (2021) 121814.