An exploratory study on the use of near infrared spectroscopy to analyse the addition of different polymers and the concentration of PET to black soldier fly larvae

This study evaluated the effects of polymer type and concentration on both near-infrared (NIR) spectra and the performance of calibration models for identifying polymer type and concentration in black soldier fly larvae (BSFL) samples. Polyethylene (PE), polypropylene (PP), polyethylene terephthalate (PET), polystyrene (PS), and polyvinyl chloride (PVC) at one concentration (1% w/w) and PET at different concentrations (0.1-5% w/w) were added to BSFL. The polymers and BSFL samples were analysed using an FT-NIR instrument to develop partial least squares (PLS) regression models. The PLS model (all samples in the data set, NIR range 9000 to 4000 cm -1 ) achieved a coefficient of determination (R² CV ) of 0.84, a root-mean-square error in cross-validation (RMSECV) of 0.67% w/w, and an RPD of 2.6. The R 2 CV and RMSECV values obtained to predict the addition of PET ranged from 0.80 to 0.81 (RMSECV: 0.72% to 0.75%) using BSFL samples from different production sites, R 2 CV and RMSECV ranged from 0.79 to 0.81 (RMSECV: 0.72% to 0.75%) using different batches, and R 2 CV and RMSECV ranged from 0.75 to 0.83 (RMSECV: 0.69% to 0.83%) using different preprocessing (vortexing) methods. The RPD values for each PLS calibration model ranged from 2.1 to 2.6, depending on the dataset used. Overall, NIR spectroscopy has shown promise as a rapid tool for detecting MPs. These results highlight the potential of NIR spectroscopy as a rapid screening tool for MP contamination in insect-derived feed ingredients, supporting quality assurance in the animal feed industry.