Rapid quantitative analysis of Hg²⁺ residue in dairy products using SERS coupled with ACO-BP-AdaBoost algorithm

In this study, surface-enhanced Raman spectroscopy (SERS) coupled with multivariate calibrations were employed to develop a rapid, simple and sensitive method for determination of mercury ions residues in dairy products. Initially, spherical Au@SiO₂ core shell nanoparticles with highly enhancement effect were synthesized to serve as the SERS substrate. Afterwards, an optical sensor system, namely micro-Raman spectroscopy system, was constructed for rapid acquisition of Au@SiO₂-mercury ions spectra. Then, ant colony optimization (ACO) and genetic algorithm (GA) were applied comparatively for selecting the characteristic variables from the Savitzky Golay-First derivative (SG-FD) processing data for subsequent quantitative analysis. Eventually, both linear (PLS and SW-MLR) and nonlinear (BPANN and BP-AdaBoost) methods were used for modeling. Experimental results showed that the variables selection methods significantly improved the model performance. Especially for the ACO algorithm, and the ACO-BP-AdaBoost model achieved the best results with the higher correlation coefficient of determination (R² = 0.997), and lower root-mean-square error of prediction (RMSEP = 0.092) than other quantification models. Paired sample t-test exhibited no statistically significant difference (sig > 0.05) between the reference concentrations determined by inductively coupled plasma mass spectrometry (ICP-MS) and the predicted concentrations by ACO-BP-AdaBoost model in adulterated foodstuffs.