Rapid Detection Of Pesticide Residues In Tea Based On Surface-enhanced Raman Spectroscopy

Tea is one of China's important agriculture products,but also traditional bulk export commodities.However,the reality of pesticide residues in tea exceeding the standard caused by all kinds of reasons not only seriously threatens consumers'health,but also severely impedes the development of China's tea export trade.Although the conventional detection methods for pesticide residues in tea posses the merits of high sensitivity,high accuracy and high reliability,the shortcomings of the conventional detection methods,including high cost,long detection time and the need of professional operator,make them not satisfying the requirement of rapid on-site detection of pesticide residues in the process of tea processing and circulation.In this study,three types of pesticides commonly used during tea cultivation,e.g.thiram,chlorpyrifos and acetamiprid,were taken as examples for the research on rapid detection method.Ultimately,the rapid and accurate detection of pesticide residues in tea was realized by using surface-enhanced Raman spectroscopy(SERS)coupled with chemometrics.The research contents are as follows:1.The construction of SERS detection system for pesticide residues in tea and the study of detection mechanism.In this section,a novel SERS detection system for the collection of SERS spectra of pesticide residues in tea was proposed.First,the SERS enhancement substrate,namely,Au@Ag nanoparticles(NPs)were synthesized by a seed growth strategy,and then the synthesized Au@Ag NPs were characterized by means of transmission electron microscope(TEM),scanning electron microscope(SEM)and UV-visible spectroscopy(UV-Vis);second,a low-cost and environmental paper-based microfluidics was fabricated;third,the novel SERS detection system was constructed by combining the fabricated paper-based microfluidics with the synthesized Au@Ag NPs,and the detection conditions of this detection system were optimized;finally,the novel SERS detection system was used for the collection of the SERS spectra of pesticide residues(thiram,chlorpyrifos and acetamiprid)in tea,the experimental results showed that the detection sensitivity of the novel SERS detection system for three pesticide residues in tea could be achieved at last 1.0×10^-9g/kg.The maximum residue limits(MRLs)for thiram,chlorpyrifos and acetamiprid in tea set by European Union(EU)is 2.0×10^-5,1.0×10^-5 and 5.0×10^-6 g/kg,respectively.In conclusion,the proposed novel SERS detection system can meet the sensitivity requirement of EU standard.2.The study of quantitative detection of a single pesticide residue in tea.In this section,a novel rapid detection approach for quantification of a single pesticide residue in tea was proposed.First,the SERS spectra of chlorpyrifos residue in tea samples were collected based on the SERS detection system;second,a novel characteristic wavelengths selection-modeling algorithm,namely,interval combination population analysis-minimal redundancy maximal relevance(ICPA-m RMR)algorithm,was developed;third,partial least squares(PLS),synergy internal partial least squares-genetic algorithm(si PLS-GA),competitive adaptive reweighted sampling-partial least squares(CARS-PLS)and ICPA-m RMR were all used for the selection of characteristic wavelengths of the SERS spectra of chlorpyrifos residue and constructing multivariate calibration models for quantification of chlorpyrifos residue in tea samples;finally,the prediction results of various models were contrastively analyzed and performed T test with the results obtained by standard method,respectively,the experimental results showed that ICPA-m RMR outperformed other models in terms of prediction ability,with RMSEC=0.1998,R_C²=0.992 in the calibration set,and RMSEP=0.2271,R_P²=0.989 in the prediction set,as well as there was no significant difference between the results predicted by ICPA-m RMR and the results detected by standard method.In conclusion,this proposed detection approach for quantification of a single pesticide residue in tea is feasible by combing the SERS detection system with ICPA-m RMR algorithm.3.The study of quantitative detection of mixed pesticide residues in tea.In this section,a novel rapid detection approach for quantification of mixed pesticide residues in tea was proposed.First,the SERS spectra of mixed pesticide residues(thiram+chlorpyrifos+acetamiprid)in tea samples were collected based on the SERS detection system;second,variable combination population analysis-SIMPLS(VCPA-SIMPLS)algorithm and variable iterative space shrinkage approach-SIMPLS(VISSA-SIMPLS)algorithm were improved,furthermore,a novel characteristic wavelength intervals selection-modeling algorithm,namely,interval combination iterative optimization approach-SIMPLS(ICIOA-SIMPLS),was developed;third,SIMPLS,VCPA-SIMPLS,VISSA-SIMPLS and ICIOA-SIMPLS were all used for the selection of characteristic wavelengths(intervals)of the SERS spectra of mixed pesticide residues and constructing multivariate calibration models for quantification of mixed pesticide residues in tea samples;finally,the prediction results of various models were contrastively analyzed and performed T test with the results obtained by standard method,respectively,the experimental results showed that ICIOA-SIMPLS possessed the best prediction ability when compared with other models,with RMSEC=0.0019,R_C²=0.990 in the calibration set,and RMSEP=0.0031,R_P²=0.980 in the prediction set,as well as there was no significant difference between the results predicted by ICIOA-SIMPLS and the results detected by standard method.In conclusion,this proposed detection approach by combing the SERS detection system with ICIOA-SIMPLS algorithm can be successfully applied for quantification of mixed pesticide residues in tea.4.The study of rapid identification of pesticide residues in tea.In this section,a novel qualitative analysis approach was proposed for identification of whether a pesticide residue in tea was excessive.First,by using the SERS detection system,the SERS spectra of chlorpyrifos and acetamiprid pesticide residues in tea samples were collected,respectively;second,one-dimensional convolution neural network(1D CNN)algorithm was introduced for constructing the rapid identification model;third,K-nearest neighbor(K-NN)model,back propagating-artificial neural network(BP-ANN)model,random forest(RF)model and 1D CNN model were used for identification of whether chlorpyrifos residue and acetamiprid residue in tea were excessive;finally,the identification results obtained from various models were contrastively analyzed,the experimental results showed that 1D CNN was superior to other identification models,with TPR=100%,FPR=0%,and ACC=100%for the identification results of chlorpyrifos residue in tea,and with TPR=100%,FPR=0%,and ACC=100%for the identification results of acetamiprid residue in tea,moreover there was no need for spectral preprocessing and reducing the dimensionality of the SERS spectra when 1D CNN handled SERS spectra.In conclusion,this proposed approach by combing the SERS detection system with 1D CNN can be successfully applied for rapid and highly accurate identification of whether pesticide residue in tea is excessive.5.The development of a portable Raman spectrometer detection system for pesticide residues in tea.In this section,on the basis of the previous four chapters,a portable Raman spectrometer detection system was developed to meet the needs of rapid on-site detection of pesticide residues in tea.First,the hardware part of the portable Raman spectrometer detection system,namely,a handheld Raman spectrometer,was developed;second,the software part of the portable Raman spectrometer detection system,namely,qualitative and quantitative analysis of SERS spectra software(V1.0)which included the qualitative and quantitative analysis models developed in chapter 3,chapter 4 and chapter 5,was developed by using Matlab GUI;finally,a portable Raman spectrometer detection system was developed by integrating the hardware and software parts,then the portable Raman spectrometer detection system was used for real samples test,the experimental results showed that there was no significant difference between the results detected by the portable Raman spectrometer detection system and the results detected by standard method.In conclusion,the portable Raman spectrometer detection system provides a new idea for rapid,on-site,and highly accurate quantitative and qualitative analysis of pesticide residues in tea.The purpose of this study is to realize the rapid and high-precision qualitative and quantitative analysis of pesticide residues in tea,which provides new methods and new ideas for the detection of pesticide residues in tea and the development of protable detection instruments.The results of this study are of great significance to the safety of tea consumers and the promotion of tea export trade.