Detection Of Aflatoxin B₁ Contamination In Peanut Oil By Laser Induced Fluorescence

Peanut oil is susceptible to aflatoxin B₁(AFB₁)contamination during production and is difficult to control.In view of the problems such as cumbersome operation and poor timeliness of traditional detection methods for AFB₁ contamination in peanut oil.This study proposed the idea of rapid detection of AFB₁ by fluorescence technology based on the characteristics of AFB₁ fluorescence signal.An AFB₁ detection system based on 375 nm laser-induced fluorescence(LIF)was built autonomously to detect the AFB₁ in peanut oil contaminated with AFB₁,the light path of LIF was optimized,and the optimal conditions of LIF system for detecting AFB₁ in peanut oil were explored.A preliminary experiment was conducted on the contaminated AFB₁ peanut oil samples from different varieties,and a LIF recognition method for the contaminated AFB₁peanut oil was established by two-step modeling.The effect of interfering factors on the detection of AFB₁ by LIF technique was further discussed,the fluorescence spectra of AFB₁ and kojic acid derivative(KAD),a secondary metabolite of mold,were characterized to investigate the effect of KAD on the fluorescence spectra of AFB₁solution and contaminated AFB₁ peanut oils.The influence mechanism of temperature of interfering factors on the detection of contaminated AFB₁ peanut oil by LIF technique was explored,and a global model was established to accommodate sample temperature variation.In this study,a rapid and nondestructive detection method based on LIF technique for AFB₁ contamination of peanut oil was developed,which may provide a theoretical basis for advancing the application of fluorescence spectroscopy in the analysis of agricultural product quality and safety detection.The main research content and conclusions of this paper are as follows:(1)Laser induced fluorescence construction of a nondestructive detection system for AFB₁ of peanut oilA qualitative and quantitative analytical method of LIF technique for detecting AFB₁ contamination by peanut oil was established after verifying system stability and optimizing the light path.The results showed that the LIF technique was able to rapidly discriminate peanut oil contaminated AFB₁,and the model build based on the 400?600nm band spectra had a higher detection accuracy when the detection angle of the LIF system was 90 degrees,and the overall discrimination correct rates were all greater than90%,in which the support vector machine(SVM)discrimination model achieved 100%accuracy.Under these conditions,the partial least squares regression analysis(PLSR)of AFB₁ content in peanut oil model predicted the coefficient of determination R_P² to reach 0.95,the root mean square error of prediction(RMSEP)value of the model prediction set was 3.40 20?g kg^-1,the relative analytical deviation(RPD)was 4.37,the limit of detection(LOQ)of 11.06 20?g kg^-1,and the model accuracy was relatively good.(2)Identification of AFB₁ contaminated with peanut oil by laser induced fluorescence spectroscopyA rapid detection and analysis model of peanut oil AFB₁contamination was established based on a scaffolding LIF system,and the preliminary detection was performed by LIF for four kinds of peanut oil("Luhua","Duoli","Jinlong yu","Fulinmen")contaminated with different levels of AFB₁.Analysis of fluorescence spectra revealed that the spectrum of peanut oil at 400?600 nm was mainly related to the content of polyphenols and vitamin E in peanut oil,and at 600?800 nm was mainly related to chlorophyll and color in peanut oil.According to the limit standard of AFB₁in peanut oil in China(<20 20?g kg^-1 for not over spiked and?20 20?g kg^-1 for over spiked samples),a qualitative and quantitative analysis model of AFB₁ pollution level was established.Principal component analysis(PCA)results showed that there was an obvious trend for clustering of single varieties of peanut oil,while PCA showed a poor trend for classifying the four peanut oils together.A two-steps modeling approach was used to first discriminate peanut oil by variety based on spectral information in the range of 600 to 800 nm,and then a qualitative model of AFB₁ contamination levels was established for individual peanut oil.Its results provided at least a 5%improvement in the trueness rate compared to the one-step modeling results.The four peanut oils PLSR model results showed that building a quantitative model based on spectral data from400 to 600 nm could better realized the prediction of AFB₁ content in different varieties of peanut oil.(3)Effects of kojic acid derivatives on the fluorescence spectra of AFB₁ from peanut oilFluorescence spectra of different concentrations of KAD and AFB₁ solutions and peanut oil were collected by LIF system.The results showed that KAD did not change the apparent of AFB₁ solution,would cause a small red shift in the fluorescence spectrum of AFB₁ solution and made the AFB₁ solution fluorescence intensity enhanced.The effect on the fluorescence spectra of peanut oil from different cultivars was mainly reflected by the enhancement of the fluorescence intensity at about 470 nm.Using CARS to select the wavelength,and KAD content as a known variable to build the SVM model,the prediction set for the four peanut oils had a correct rate greater than92%,which was a 20%improvement over the model without KAD content added as an independent variable.(4)Mechanism of temperature effect on AFB₁ detection in peanut oilIn order to make the peanut oil samples more representative,LIF system was used to collect the fluorescence spectra of 10 kinds of contaminated AFB₁ peanut oil at five different temperatures(10,20,30,40,50?).The raw spectral intensity of peanut oil was decreased with increasing temperature.The temperature global model of LDA was strongly adaptable to five temperatures,and the modeling set had the lowest 86%correct rate when predicting samples with different temperatures.The temperature global model results of R²_Cand R_P² were 0.95 and 0.92,RMSEC and RMSEP were 2.1620?g kg^-1 and 3.03 20?g kg-¹,respectively,and LOD was 9.51 20?g kg^-1.