Study on quantitative methods of laser-induced two-dimensional fluorescence spectroscopy of multicomponent PAHs in soils

Laser-Induced Fluorescence(LIF) has become a powerful technology for quantitative analysis of Polycyclic Aromatic Hydrocarbons(PAHs) in soils due to its fast detection speed and low operational cost, without sample preparation. However, there are many types of PAHs in soils, and their similar structures lead to overlapped issues in their laser-induced fluorescence spectra. It is challenging to quantify a single PAH in complicated soil without the chemical separation. In this paper, fluorescence spectra of PAHs in agricultural soil are obtained by a 266 nm mobile LIF system, and quantification methods are investigated for PAHs, based on univariate linear regression, weighted non-negative least squares Multivariate Linear Regression(MRL) and Support Vector Regression(SVR). The results show that the correlation coefficients of anthracene and phenanthrene are both less than 0.90, and the average relative errors are both more than 20% by using univariate linear regression. Compared with univariate linear regression, MLR improves the prediction accuracy of anthracene and phenanthrene in the soil contaminated with bi-component PAHs. However, the average relative errors are still over 20% in the soil contaminated with multicomponent PAHs. Finally, a SVR model optimized by grey wolf optimization combined with differential evolution(GWO-DE) is applied for concentration measurement of anthracene and phenanthrene in agricultural soil contaminated with multicomponent PAHs. The average relative error of anthracene decreases from 23.1% (MLR) to 5.02%, while in the case of phenanthrene decreases from 20.8% (MLR) to 4.83%. This study provides an efficient method to improve the accuracy of LIF in quantifying multicomponent PAHs in soils.