Quantitative analysis of polycyclic aromatic hydrocarbons in soil by infrared spectroscopy combined with hybrid variable selection strategy and partial least squares.

Infrared spectroscopy (IR) combined with multivariate calibration technology can be used as a potential method to quantitative analysis of polycyclic aromatic hydrocarbons (PAHs) in soil, which provides a rapid data support for soil risk assessment. However, IR spectrum contains lots of useless information, its predictive performance is poor. Variable selection is an effective strategy to eliminate irrelevant wavelengths and enhance predictive performance. In this study, IR combined with partial least squares (PLS) was proposed to quantify anthracene and fluoranthene in soil. In order to improve the predictive performance of the PLS calibration model, the synergy interval PLS (siPLS) method was first used for "rough selection" to select feature bands; on this basis, "fine selection" was performed to extract the feature variables. In "fine selection", three different feature variables selection methods, such as successive projection algorithm (SPA), genetic algorithm (GA), and particle swarm optimization (PSO), were compared for their performance in extracting effective variables. The results show that the siPLS-GA calibration model receive a lowest root mean square error (RMSE) and a largest determination coefficient (R2). Results of external validation demonstrate an excellent predictive performance of siPLS-GA calibration model, with the R2 = 0.9830, RMSE = 0.5897 mg/g and R2 = 0.9849, RMSE = 0.4739 mg/g for anthracene and fluoranthene, respectively. In summary, siPLS combined with GA can accurately extract the effective information of the target substance and improve the predictive performance of the PLS calibration model based on IR spectroscopy.

Novel Method Based on Hollow Laser Trapping-LIBS-Machine Learning for Simultaneous Quantitative Analysis of Multiple Metal Elements in a Single Microsized Particle in Air.

Elemental identification of individual microsized aerosol particles is an important topic in air pollution studies. However, simultaneous and quantitative analysis of multiple constituents in a single aerosol particle with the noncontact in situ manner is still a challenging task. In this work, we explore the laser trapping-LIBS-machine learning to analyze four elements (Zn, Ni, Cu, and Cr) absorbed in a single micro-carbon black particle in air. By employing a hollow laser beam for trapping, the particle can be restricted in a range as small as ∼1.72 µm, which is much smaller than the focal diameter of the flat-topped LIBS exciting laser (∼20 µm). Therefore, the particle can be entirely and homogeneously radiated, and the LIBS spectrum with a high signal-to-noise ratio (SNR) is correspondingly achieved. Then, two types of calibration models, i.e., the univariate method (calibration curve) and the multivariate calibration method (random forests (RF) regression), are employed for data processing. The results indicate that the RF calibration model shows a better prediction performance. The mean relative error (MRE), relative standard deviation (RSD), and root-mean-squared error (RMSE) are reduced from 0.1854, 363.7, and 434.7 to 0.0866, 179.8, and 216.2 ppm, respectively. Finally, simultaneous and quantitative determination of the four metal contents with high accuracy is realized based on the RF model. The method proposed in this work has the potential for online single aerosol particle analysis and further provides a theoretical basis and technical support for the precise prevention and control of composite air pollution.