The Prediction of the Concentration of Artificially Added Microplastics in Different Foods and Feed Ingredients Using Near-Infrared Spectroscopy—Instrument Type, Limit of Detection and Quantification

Predicting microplastic (MP) concentrations in different foods and feed ingredients is a growing area of interest in near-infrared (NIR) spectroscopy. The objective of this paper is to evaluate the ability of NIR spectroscopy (benchtop and portable instruments) to predict the concentration (0% to 5% w/w) of the addition of polymers [polyethylene (PE), polypropylene (PP), and polyethylene terephthalate (PET)] into three different matrices commonly encountered as food and feed ingredients. This study used samples of banana, spinach, breadcrumbs, and black soldier fly larvae (BSFL) meal with different levels of polymer addition. Calibration models to predict the concentration of MP in the samples were developed using partial least squares (PLS) regression. The coefficient of determination in cross-validation (R2CV) and the standard error in cross-validation (SECV) were 0.76 and 0.72%, and 0.81 and 0.75% using cross-validation (CV) and the test set, respectively. Using CV and the test set, the R2CV and SECV were 0.92 (SECV: 0.38%) and 0.95 (SECV: 0.32%), respectively, for the portable instrument. The R2CV and SECV were 0.81 (SECV: 0.59%) and 0.81 (SECV: 0.60%) for the CV and test sets analysed using the benchtop instrument, respectively. The SEP was 0.41% and 0.65% using the portable and benchtop instruments, respectively. These results demonstrate the potential of both portable and benchtop NIR spectroscopy as rapid tools for predicting microplastic concentrations in a range of food and feed matrices.