A correction method of mixed pesticide content prediction in apple by using Raman spectra

Authors: LI, YAN - China Agricultural University; PENG, YANKUN - China Agricultural University; Qin, Jianwei - Tony Qin; Chao, Kuanglin - Kevin Chao

Submitted to: Applied Sciences
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 5/31/2019
Publication Date: 6/7/2019
Citation: Li, Y., Peng, Y., Qin, J., Chao, K. 2019. A correction method of mixed pesticide content prediction in apple by using Raman spectra. Applied Sciences. 9(8):1699.

Interpretive Summary: Acetamiprid and deltamethrin are two pesticides commonly used on apples. Accurate measurement of mixed residue pesticides is required to for both food safety and regulatory concerns. In the measurement of the light intensity value of single pesticide vs. multiple pesticide mixture using optical instruments, the signal intensity (relative to noise) decreases when multiple pesticides are present. This unnecessarily increases the food safety risk of under measurement of pesticide contents on food products especially near its experimental and/or regulatory detection limit. In this study, the output of the regression equation for each pesticide was used in a quantitative model to accurately predict the concentration of individual pesticide in the mixture samples. The novel methodology presented in this study can be applied to correct spectral intensities in other pesticide mixture, and improve the accuracy of optical sensing measurement for rapid quantification of multiple pesticides in foods. The results of this research will benefit producers interested in monitoring or minimizing their pesticide applications as well as regulatory agencies interested in identifying pesticide residues on fresh produce products.

Technical Abstract: Two pesticides, acetamiprid and deltamethrin were used as test samples in apples to develop and evaluate an improved method to reduce Raman spectral measurement error in the quantitative analysis of mixed residue pesticides. Characteristic peak intensity values of gradient concentration pesticide from 100 ppm to 0.001 ppm were collected. A two-dimensional mathematical smooth surface model was built to describe the correction coefficients of the characteristic peak intensities. The correction method was applied to correct the spectral peak intensities of the mixed pesticides. The corrected intensities were used in the quantitative model to predict the concentration of each pesticide in the mixture sample. Correlation coefficient of model validation was 0.90. The root mean square error of model validation was significantly reduced, indicating the improvement from correction method in quantitative analysis. The improvement in precision is especially useful in increasing quantitative certainty at concentrations near detection limit. The spectral correction method developed in this study showed great potential for Raman sensing applications to detect pesticide mixture and other chemical contaminants in foods.