Extraction and identification of mixed pesticides' Raman signal and establishment of their prediction models

A nondestructive and sensitive method is developed to determine mixed pesticides of acetamiprid, chlorpyrifos and carbendazim in apple samples by surface‐enhanced Raman spectroscopy (SERS). Self‐modeling mixture analysis (SMA) was used to identify and extract the Raman signals of each pesticide from the spectra of apples contaminated with mixed pesticides. Results indicate that the obtained SERS signal intensities of each pesticide in their mixture have no obvious difference to the signal intensities of the corresponding pure pesticide at a low concentration. The lowest detectable level of acetamiprid, chlorpyrifos and carbendazim in apple are 0.0054 mg/kg, 0.064 mg/kg and 0.014 mg/kg, respectively, which are sensitive enough for identifying apple contaminated with pesticides above the maximum residue limit. The predicted values of each pesticide in their mixture are obtained using the prediction model based on the Raman signal of the single pesticide. The correlation coefficients of predicted values and actual values are 0.893 for acetamiprid, 0.926 for chlorpyrifos and 0.938 for carbendazim, respectively. The method presents the ultrasensitive SERS performance for quantifying residual pesticides in apple samples without sample pre‐treatment, showing great potential to serve as a useful means for monitoring pesticide residues used in mixed state. Copyright © 2016 John Wiley & Sons, Ltd. This method presents the ultrasensitive SERS performance for quantifying mixed pesticides in apple samples without sample pre‐treatment. The self‐modeling mixture analysis (SMA) method was applied to decompose the Raman signals of mixed pesticides and extract the Raman information of the pure pesticide. This work indicates that the obtained SERS signal intensities of each pesticide in their mixture have no obvious difference to the signal intensities of the corresponding pure pesticide at a low concentration.