Visual characterization of microplastics in corn flour by near field molecular spectral imaging and data mining

As potential hazard to human health, microplastics have attracted increasing attention. Most current studies have addressed the characterization of microplastics from the environment. For microplastics in food, most detections focused on liquid systems such as alcohol, beverages, etc., while there has been quite rare research on microplastics in solid foods with complex matrices. Thus, this study attempted to use three molecular spectral imaging approaches, namely, Fourier transform infrared (FTIR), optical photothermal resonance infrared (O-PTIR), and confocal Raman spectral imaging, combined with chemometrics to characterize the presence of microplastics in corn flour. The results demonstrated that O-PTIR imaging can rapidly sense the presence of microplastics, but its data integrity and visualization were limited. By decomposing the image, FTIR and Raman acquired a more integral distribution. Wherein, microplastics were well depicted by Raman imaging coupled with independent component analysis. Moreover, O-PTIR imaging can quickly detect contaminants at low concentrations but with a low detection rate. Raman imaging underperformed in low-concentration samples but provided a better visualization in mid-concentration samples. Overall, the results confirmed that the visual detection of microplastics in powdered food can be realized by molecular spectral imaging coupled with data mining, which can provide a reference for the detection of microplastics in other foods.