Research On Voltammetric Electronic Tongue For The Detection Of Aflatoxin In Rice

Food security is closely related to human's lives.Aflatoxins are the most common mycotoxins,can contaminating food crops and causing mildew.The content of aflatoxin in foodstuffs is strictly limited,especially the content of aflatoxin B1?AFB₁?,because it has the largest contaminant range,the greatest toxicity and the highest content.At present,the method of detecting AFB1 has the limitations of complex operation,long cycle and high cost,and cannot meet the increasing detection requirements.Electronic tongue technology is a new type of detection technology.Based on the sensor array,the whole signal of the unknown sample is measured,and then the pattern recognition and qualitative quantitative analysis are performed.The operation is simple and the repeatability is good.The applications in various fields are increasing.In this paper,a voltammetric electronic tongue system for detecting aflatoxin B1 was constructed.According to the mechanism of inhibition of acetylcholinesterase?AChE?by AFB1,the experimental conditions were optimized,and AChE was fixed to a glassy carbon electrode?GCE?by cross-linking.Preparation of AChE/GCE electrode.A volt-ampere type electronic tongue data acquisition unit with AChE/GCE electrode as working electrode,Ag/AgCl as reference electrode,and platinum electrode as auxiliary electrode was set up.The data acquisition application program of the upper computer was used to control the acquisition platform and receive data.The aflatoxin B1 standard solution and rice solution were tested under optimal conditions.The original data is preprocessed using the extraction special point method and wavelet compression to compress the data volume.Quantitative analysis was performed by partial least squares?PLS?and radial basis?RBF?neural networks.Two kinds of data pretreatment methods were used to extract the characteristics of aflatoxin B1 standard fluid data.PLS and RBF neural networks were used to establish two concentration prediction models for quantitative prediction of aflatoxin B1 in rice solution.The experimental results show that the average residual energy and the coefficient of zero-setting of the signal after wavelet compression are 98.75%and 98.81%,respectively;the root mean square error of the PLS and RBF neural network model established by using the wavelet compression coefficient as input is smaller and better than The model established by eigenvalues verifies the effectiveness of wavelet compression;PLS and RBF neural network models established by two data preprocessing methods have good quantitative analysis effect,and the root mean square errors of RBF neural network models are smaller than the mean square of PLS models.Root error,RBF neural network concentration prediction effect is better.In this paper,a voltammetric electronic tongue system based on enzyme sensor is constructed to realize the effective detection of aflatoxin B1 in rice.This study provides a rapid,economical and highly sensitive method for the detection of aflatoxin B₁.