Microplastic Mass Estimation Using Two-Dimensional Chemical Images from Quantum-Cascade Laser-Based Infrared Spectrometers.

Modern spectroscopic chemical imaging techniques perform nondestructive, relatively fast sample analysis to count and chemically characterize individual particles. However, they do not offer information on particle mass, a relevant parameter for several disciplines, like toxicology. This work strives to estimate the mass of plastic particles using a tiered sequence of steps and models that employ bidimensional parameters (height, width, perimeter, area) reported by state-of-the-art tunable quantum-cascade laser-based infrared imaging spectroscopy. A hybrid model that does not need calibration steps for its routine application is proposed for the first time. The shape of each particle is assessed individually, rather than setting them initially. Fibers are modeled as cylinders using an equivalent cylinder concept, while fragments are approximated to either parallelepipeds or spheroids, based on their 2D circularity. Previously published models were tested, modified, and hybridized to assess the mass of known sets of plastic particles whose size ranged from 20 to 1500 μm and whose total weight ranged from 190 to 9400 μg. The hybrid approach estimated the mass of exemplary samples with relative errors lower than 20% (a satisfactory level for these estimations) and worked well in size and weight ranges barely tested before. An Excel-based spreadsheet (NOMME, Number Of Microplastics and their Mass Estimation) was developed to streamline the application of the hybrid model.