Research Of The Rapid Detection For Aflatoxin B1 In Wheat Using Surface-enhanced Raman Spectroscopy

Wheat is one of the most important cereals in China,but mycotoxins,mainly aflatoxins seriously affects the quality and safety of wheat and its products as well as posing a significant health risk to consumers.Conventional technologies can meet the national standard for precision and sensitivity,but there are many shortcomings such as expensive equipment,complex operation,and elongated operation time to release data.Therefore,rapid,highly sensitive and specific mycotoxin detection strategies are proposed,which will be conducive to the healthy development of wheat industry.Surface-enhanced Raman spectroscopy(SERS)can efficiently detect organic pollutants,but to detect toxins,aptamers,antibodies,and other biological recognition agents are usually required which are prone to denaturation with high cost and technical threshold imperfections.Therefore,this study proposed non-biolabeled SERS strategies for aflatoxins detection in wheat,and an intelligent Web server was built as the carrier of chemometrics models with the main content as follows:1.The ionic media assisted SERS sensing for sensitive and rapid detection of AFB1 in wheat.To overcome the shortcoming such as high cost and denaturation in bio labeled SERS sensing,the non-biolabeled SERS strategy in this chapter was proposed on the basis of the specific binding between mercury ion(Hg(?))and adjacent oxygen of the carbonyl group in AFB1,spiral lactam in Rhodamine derivant respectively,and the‘nano gap'effect after particles aggregation.Firstly,the metal/metal nanostructure(Au@Ag CSNPs)was prepared and used as SERS substrate.After the modification of 5-aminotetramethylrhodamine(NH₂-Rh)on the surface of Au@Ag CSNPs,the sensing probe NH₂-Rh-Au@Ag CSNPs were obtained,and the quantification of AFB1 was performed.The strategy can be described as:Firstly,the wheat extract containing AFB1 was incubated with Hg(?)solution,the free Hg(?)would specifically bind with AFB1 and form a complex;then,the mixture was further incubated with NH₂-Rh-Au@Ag CSNPs solution,the SERS intensity could be significantly improved by the Hg(?)induced probe aggregation.A linear correlation between AFB1 and SERS intensity in the range of 0.1-100 ng m L^-1,(R²)=0.98 was constructed.The obtained limit of detection(LOD)and limit of quantification(LOQ)were 0.03 ng m L^-1=and 0.1 ng m L^-1,respectively.Further,it was shown that the correlation between the method and high performance liquid chromatography-fluorescence detection(HPLC-FLD)was 0.99;the relative standard deviation(RSD)of repeated experiments with different concentrations less than 5%,indicating the high reproducibility and stability.This study confirmed that the ion medium assisted SERS technology can realize the rapid and highly sensitive detection of AFB1 in wheat.Compared with the biosensor based SERS technology,it had the advantages of low cost and strong stress resistance.2.The linear affinity agent-assisted SERS sensing for rapid detection of AFB1in wheat.To obtain a more convenient non-biolabeled SERS technique for rapid detection of AFB1,linear polymers were prepared as affinity agents to capture AFB1molecules.Firstly,p NAGA polymer with a chain length of 15 was prepared by chain transfer polymerization to capture AFB1;then,p NAGA was modified on the surface of Au@Ag CSNPs surface ensured that the scattering signal of AFB1 could be amplified;then,combined with the 6-31G(d)basis set of density functional theory(DFT),the theoretical SERS spectrum of AFB1 was calculated,and the characteristic SERS signal peaks of AFB1 were selected by spectral comparison and vibration mode analysis;finally,the standard curve was drawn and quantified by using the intensity of AFB1 at1137cm^-1.The results showed that the quantification's linear range was 0.1-10 ng m L^-1with the R²=0.9522,LOQ and LOD were 0.1 ng m L^-1 and 0.05 ng m L^-1,respectively.The correlation between the established method and the standard method is 0.96,and the RSD of repeatability and reproducibility was less than 6.5%.The results showed that the p NAGA affinity agent could effectively capture AFB1 and achieve rapid SERS detection of AFB1 with the assist of DFT theoretical calculation,and the quantitative efficiency had improved.3.The chemometrics model-assisted SERS sensing for rapid detection of AFB1 in wheat.To further alleviate the complicated pretreatment and complicated substrate preparation,a more convenient non-biolabeled SERS detection was proposed in this chapter by using the three-dimensional nanocomposite SERS substrate combined with chemometrics models.Firstly,Zn O particles with different morphologies were synthesized by wet chemical method,and silver nanoparticles were grown in situ on the surface to obtain three-dimensional structure Ag@Zn O.Then,the SERS spectra of wheat extracts with different concentrations of AFB1 were collected,and various spectral pretreatment and variable screening methods were used to reduce the interference of spectral noise and redundant variables.The screened variables were introduced into the partial least squares(PLS)method to construct a multiple regression model.Finally,the prediction performance of different models was evaluated,and the best quantitative model was determined.The results showed that the enhancement factor of the Ag@Zn O substrate prepared at 60?could reach 8.46×10⁷;94 variables selected by standard normal variable transformation(SNV)combined with competitive adaptive reweighting algorithm(CARS)obtained the best prediction ability in PLS model,R_P=0.9555.Under this condition,the method can predict the AFB1concentration in the range of 0.01-500?g kg^-1;the correlation between the actual sample detection results and the standard method was 0.99,and the RSD of repeatability and reproducibility was less than 8.4%,indicating high stability and accuracy of the established methodstability and accuracy.The final results showed that the established procedure in this chapter could meet the accuracy requirements of AFB1 quantification in wheat with its improved detection efficiency without requiring any complex pre-processing steps.4.The pattern recognition assisted SERS sensing for rapid detection of AFB1 in wheat.To meet the needs of rapid detection of AFB1 pollution degree in practical production,spectral matching and semi quantitative prediction models are established by using spectral matching and pattern recognition algorithm respectively in this chapter.According to the safety standards of wheat in food and feed,the permissible contamination level is>20?g kg^-1,5-20?g kg^-1 and<5?g kg^-1,respectively.Three different contamination level corresponding to severe pollution,feed grade and food grade wheat,SERS spectra were collected in these ranges for spectral matching and semi quantitative model establishment;then,SERS spectra were used as the database,and different spectral distinguish algorithms were optimized based on the matching degree of prediction set;then,a variety of linear or nonlinear pattern recognition models were established,and the model parameters were optimized;finally,according to the prediction accuracy and chaotic matrix,the model performance was evaluated and the actual sample detection ability was verified.The results showed that the spectral angle matching(SAM)algorithm achieved better spectral discrimination under the condition of the threshold value of 0.9;the optimal spectral prediction result of the support vector machine(SVM)model for the three contamination level was achieved under the model parameters of C=128,PCs=9,?=0.758 and?=0.879 which obtained an accuracy of 94.44%;in the actual sample detection with HPLC-FLD as the standard method,obtained method were able to distinguish the spiked samples and the accuracy was100%.The results showed that the method could meet the needs of rapid identification of wheat pollution degree in the actual detection and further simplify the model,which was conducive to the rapid identification of different uses of wheat as raw materials.5.The cloud platform assisted SERS sensing for the establishment of an intelligent detection system of wheat AFB1.The detection methods established above can meet the accuracy requirements of AFB1 detection,but the preparation of SERS probe and the construction of the chemometrics model undoubtedly raise the technical threshold.Therefore,this chapter built an intelligent detection system and implants quantitative and quantitative wheat AFB1 to realize efficient data processing and highly shared quantification models and results.First,a web front-end management interface and a lightweight web server with the back-end programs of My SQL and Tomcat was developed;then,the quantitative and semi-quantitative prediction models were transcoded and implanted into the database under the Java environment to realize the efficient processing of spectral data;then,an APP in Android environment was developed as well to control the spectrometer and realize the data upload and download;and then,the integration of communication ports was achieved in the wireless local area network(WLAN),where server(?)smartphone(socket-TCP),smartphone(?)Raman spectrometers(Bluetooth 2.0)were bridged and the stable transmission,efficient sharing of data were realized as well;finally,two spectrometers with the same hardware were connected to the detection system respectively for SERS detection of actual samples and for model sharing performance.The results showed that the established detection system can successfully run the quantitative and semi quantitative models of wheat AFB1,and the functional modules run well;the actual detection results of the two spectrometers connected to the system had no significant difference(P>0.05)using the same model;the correlation analysis results confirmed that the correlation between the detection results of the system and those of the standard method was 0.99.According to the above results,the detection system was able to complete the rapid detection method in this chapter.It had the advantages of efficient data processing and strong ability of model sharing compared to conventional spectral equipment detections systems.In summary,the non biolabeled SERS detection method of AFB1 in wheat was explored from three perspectives of ion-mediated,polymer capture and data mining.The cloud platform technology was introduced to build an intelligent detection system to realize the interconnection and data sharing of user terminal,detection equipment and server,which provided a certain basis for the following research of agricultural product safety detection.