Raman-infrared spectral fusion combined with partial least squares (PLS) for quantitative analysis of polycyclic aromatic hydrocarbons in soil
Spectroscopy is an effective method for the rapid detection of PAHs in soil, but the predictive performance is not satisfactory due to interference from other substances. Spectral fusion technology can further improve the accuracy and stability of quantitative analysis by including more comprehensive information. In this study, the feasibility of analyzing anthracene and fluoranthene in soil by Raman spectroscopy, infrared spectroscopy combined with spectral fusion technology and partial least squares (PLS) was investigated. Firstly, Raman spectra and infrared spectra of soil were collected. Then, in order to obtain a better PLS calibration model, the influence of different pretreatment methods on the predictive performance of the PLS calibration model was explored, with the coefficient of determination (R2) and root mean square error (RMSE) as evaluation indicators. Finally, low-level data fusion was achieved by concatenating Raman and infrared spectra into a matrix, and coupled with PLS to construct a quantitative analysis model of PAHs in soil. In comparison to the single applications of Raman and infrared spectroscopy, accurate prediction of PAHs was made by fusion spectroscopy, with R2 = 0.9514, RMSE = 0.9143 mg g−1, RPD = 4.3211 and R2 = 0.9609, RMSE = 0.8614 mg g−1 RPD = 5.0361 for anthracene and fluoranthene, respectively. In summary, the combination of Raman and infrared spectroscopy is a promising alternative to rapid quantitative analysis of PAHs in soil.