1. Anhui Engineering Laboratory of Agro-Ecological Big Data, Anhui University, Hefei 230601, China
2. School of Electronic and Electrical Engineering, Anhui Sanlian University, Hefei 230601, China
3. Hefei Institute of Technology Innovation, Chinese Academy of Sciences, Hefei 230031, China
Abstract：Dynamic surface-enhanced Raman spectroscopy (SERS) is based on thestate transition of nanostructure from wetstate to dry state to realize spectra measurement, and it not only produces giant Raman enhancement but also provides reproducible and stable SERS signal. In the paper, we proposed a novel method for the rapid and quantitative analysis of edifenphos based on dynamicsurface-enhanced Raman spectroscopy with multivariate analysis method. In experiment, the spectra of 100, 50, 10, 5, 1, 0.5 and 0.1 mg·L-1edifenphos were measured using dynamic surface-enhanced Raman spectroscopy, and the baseline shift of spectra was removed by the polynomial correction method. Then, the spectra of full range (600~1 800 cm-1) and characteristic range (674~713，890~1 195，1 341~1 399 and 1 549~1 612 cm-1) were used to develop the regression model for the rapid and quantitative analysis ofedifenphos using support vector machine regression (SVR), respectively. Simultaneously, we also evaluated the effect of the principle component analysis (PCA) on the construction of the regression model. The experiments showed the model developed with the spectra of characteristic range had lower prediction error and PCA can improve the prediction accuracy of the corresponding model further. The best regression model (RMSECV=0.065 7 mg·L-1) was built with the spectra of characteristic range extracted by PCA, and the regression model can predict the concentration of edifenphos solutions accurately. Finally, to evaluate the effect of the method in practical application, the edifenphos residue on the apple peel were also detected using the proposed method, which was compared with the gas chromatography. The detection results showed the multiple detection value for each sample was in a small range, and the mean value was basically consistent with the detection value of gas chromatography, in which the maximal relative error was about 5.13 %. Additionally, the detection process of dynamic surface-enhanced Raman spectroscopy only needed 2 min and 2 μL samples volume, and the subsequent data analysis generally consumed several seconds. In a word, the dynamic surface-enhanced Raman spectroscopy can provide the high reproducible and precision detection of edifenphos, and the multivariate analysis method can realize the intelligent and rapid analysis of Raman spectra of edifenphos. Therefore, the dynamic surface-enhanced Raman spectroscopy with the multivariate analysis method is of great advantage for the rapid and accurate detection of edifenphos.
[1] Roberts E M, English P B, Grether J K, et al. Environ. Health Persp., 2007: 1482.
[2] Ferrer C, Martínez-Bueno M J, Lozano A, et al. Talanta, 2011, 83(5): 1552.
[3] Ozkan A. Fresenius Environ. Bull., 2015, 24(2a): 615.
[4] Cialla D, Mrz A, Bhme R, et al. Anal. Bioanal. Chem., 2012, 403(1): 27.
[5] Ock K, Jeon W I, Ganbold E O, et al. Anal. Chem., 2012, 84(5): 2172.
[6] Wen Z Q, Li G, Ren D. Appl. Spectrosc., 2011, 65(5): 514.
[7] Liu B, Zhou P, Liu X, et al. Food and Bioprocess Tech., 2013, 6(3): 710.
[8] WANG Xiao-bin, WU Rui-mei, LING Jing, et al(王晓彬, 吴瑞梅, 凌 晶, 等). Spectroscopy and Spectral Analysis(光谱学与光谱分析), 2016, 36(3): 736.
[9] Palonpon A, Ando J, Yamakoshi H, et al. Nat Protoc., 2013, 8(4): 677.
[10] Hu Y, Liao J, Wang D, et al. Anal. Chem., 2014, 86(8): 3955.
[11] Yang L, Liu H, Wang J, et al. Chem. Comm., 2011, 47(12): 3583.
[12] Yang L, Li P, Liu H, et al. Chem. Soc. Rev., 2015, 44(10): 2837.
[13] Liu H, Yang Z, Meng L, et al. J. Am. Chem. Soc., 2014, 136(14): 5332.
[14] Dong R, Weng S, Yang L, et al. Anal. Chem., 2015, 87(5): 2937.
[15] Gracie K, Correa E, Mabbott S, et al. Chem. Sci., 2014, 5(3): 1030.
[16] Weng S, Chen S, Li M, et al. Anal. Methods, 2014, 6(1): 242.
[17] Li P, Liu H, Yang L, et al. Talanta, 2013, 117: 39.
[18] Hung K-C, Lin K-P. Information Sciences, 2013, 224: 37.
[19] Weng S, Li M, Chen C, et al. Anal. Methods-UK, 2015, 7(6): 2563.
