Quantitative determination of heavy metal Pb content in soybean oil based on microwave detection technique combined with multivariate analysis

This study provides a reference for rapid determination of heavy metal Pb in soybean oil by microwave detection technique. The microwave transmission coefficients of soybean oil samples with different concentrations of the heavy metal Pb were collected using a portable microwave detector made by our team. Three feature optimization methods, which were bootstrapping soft shrinkage (BOSS), competitive adaptive reweighted sampling (CARS), and variable combination population analysis (VCPA), were employed to select the features of microwave transmission coefficients after least square filtering, and partial least squares (PLS) models constructed on features selected were used to rapid determine the heavy metal Pb content in soybean oil. The results showed that the three variable optimization algorithms can screen out highly targeted features. Especially, the CARS-PLS model obtained the optimal generalization performance, and its coefficient of determination ( R P 2 ${\mathrm{R}}_{\mathrm{P}}^2$ ) and residual predictive deviation (RPD) were 0.9857 and 8.3838, respectively. The results reveal that it is feasible to use the microwave detection technique combined with appropriate chemometrics methods to determine the heavy metal Pb in soybean oil with high precision. This study provides a practical technique for nondestructive determination of heavy metal residues in edible oil .