The research on the molecular spectroscopic recognition mechanism of microplastics in typical agricultural media

To study the identification mechanism of microplastics in agricultural environmental media by molecular spectroscopy, two typical media, soil and fishmeal, were selected for this study. Three common microplastics, PE, PP, and PS, were used as research objects. Near-infrared (NIR) and mid-infrared (MIR) spectroscopy combined with chemometric methods were used to explore microplastics' identification and analysis effects in different agricultural media and reveal their identification mechanisms. PCA analysis revealed that different soil types would affect the identification results of microplastics. PLS-DA discriminant analysis showed that the accuracy of NIR spectroscopy technology in identifying microplastics in different types of soil decreased in the order of sand > loam > clay. In contrast, the rule of MIR spectroscopy technology was the opposite. The sensitivity and specificity of the three microplastics in the NIR model of sand and the infrared spectroscopy model of fishmeal reached 1.000. NIR spectroscopy technology is suitable for identifying microplastics in soil, while infrared spectroscopy technology is more suitable for identifying microplastics in fish meal. Furthermore, based on the VIP values of each wavelength point in the spectrum, the characteristic bands that have essential contributions to identifying microplastics in soil were screened out. The NIR spectra of 4500–4300 ${\mathrm{cm}}^{-1}$, 4300–3900 ${\mathrm{cm}}^{-1}$ and 7100–5800 ${\mathrm{cm}}^{-1}$ are the most essential characteristic bands for identifying microplastics in clay, loam, and sand, respectively. The MIR of 3000–2900 ${\mathrm{cm}}^{-1}$ and 700–650 ${\mathrm{cm}}^{-1}$ were the most essential characteristic bands for identifying microplastics in soils, and the overlap of the characteristic spectra of the three soils reached 59.45%. The NIR spectra of 6050–5600 ${\mathrm{cm}}^{-1}$, 4700–4000 ${\mathrm{cm}}^{-1}$ and the MIR spectra of 2300–1900 ${\mathrm{cm}}^{-1}$ and 800–400 ${\mathrm{cm}}^{-1}$ are the most essential characteristic bands for the identifying microplastics in fishmeal. This study provides a more suitable technical "solution" for identifying microplastics in environmental media, which is of great significance for improving the accuracy of molecular spectroscopic identification in environmental media and clarifying the interaction between microplastics and environmental media.