Validated spreadsheet for the identification of PE, PET, PP and PS microplastics by micro-ATR-FTIR spectra with known uncertainty

The monitoring of microplastics in the environment is crucial to determine the relevance and trends of this contamination necessary to plan environmental protection policies. Monitoring data reliability is essential for this purpose. This work describes a methodology for the reliable identification of the most abundant polymer types in aquatic sediments (polyethylene, PE, polypropylene, PP, polyethylene terephthalate, PET, and polystyrene, PS) by micro-ATR-FTIR. Identifications with true and false result rates greater than 95% and lower than 5% are performed, respectively. The analysis is based on defining spectra requirements regarding characteristic and interfering bands intensity and selecting optimal assessed wavenumber range, signal processing, and algorithm to quantify the match/agreement between particle and reference spectra. It is also defined the minimum match value, P5 >> P, for reliable identifications. Examinations are performed in two stages where in the first stage PE and PP, PE&PP, are distinguished from other microplastics by taking the [4000-500] cm(-1) spectra and various Match Methods and P5 >> P depending on the polymer type. PE and PP are distinguished by quantifying weighted or unweighted Pearson correlation coefficients against a reference spectrum in the [3000-2800] cm(-1) range. The defined P5 >> P are above the 0.6 value considered in many references that do not quantify identification uncertainty. The MS-Excel files used in method development and validation are made available as Supplementary Material being applicable to other spectral techniques and analytical fields.

Automated summary

This study presents a methodology for reliably identifying the most common polymer types in aquatic sediments using micro-ATR-FTIR. Identifications with true and false result rates greater than 95% and lower than 5% are achieved. The analysis involves defining spectral requirements, selecting optimal assessed wavenumber range, and quantifying the match between particle and reference spectra.