Study On Rapid Detection Method Of 9 Mineral Elements In Wheat And Its Origin Discrimination Based On EDXRF Method

The quality and safety of food and agricultural products are directly related to people's health and have been widely concerned by the public.Among them,the pollution of heavy metal elements in agricultural products has attracted everyone's attention.Once the heavy metal elements enter the human body,it is difficult to be excreted by the body,resulting in the occurrence of various diseases.For example,the accumulation of heavy metals such as Ni and As in the human body can lead to the occurrence of cancer.Therefore,the establishment of rapid detection methods for heavy metals in foods and agricultural products is particularly important.In addition,the establishment of traceability methods for the origin of food and agricultural products is very important to avoid counterfeit products and protect consumers' interests.Therefore,a rapid detection method for nine mineral elements(Mg,P,S,K,Ca,Mn,Fe,Cu,Zn)in wheat was studied and established based on energy dispersive X-ray fluorescence spectrometry(EDXRF),combined with certain chemometrics methods;The quick method of identifying the origin of wheat was explored based on the difference of characteristic mineral elements in wheat.The research content and main results are as follows:(1)The influence of tableting pressure and wheat grain size on the accuracy of the instrument during sample preparation was analyzed.Based on the X fluorescence intensity produced by the elements,the results of different grain sizes of unscreened,40 mesh,60 mesh,80 mesh,and 100 mesh sieves were analyzed,and measurement results of 25 t,30t,35 t,and 40 t different pressure on the sample preparations were also compared.The results showed that the wheat flour passing through 60-mesh sieve and sample preparation at a pressure of 35 t can improve the detection accuracy.(2)The instrument detection parameters of each element were optimized.Different elements have different X fluorescence yields,in order to excite the X fluorescence of the test element most effectively,two different instrument detection parameters are used for light and non-light weight elements.The results showed that a better element X fluorescence signal can be obtained under two different measurement conditions: for the light weight elements,the tube voltage was 7 kV,the tube current was 500 ?A,and the vacuum time was 35 s.No.1 filter and 8 mm collimator were used;for non-light elements,the tube voltage was 40 kV,tube current was 250 ?A,no vacuum,No.3 filter(0.3 mm Al)and 8 mm collimator were used.(3)The accuracy of the conventional peak area method modeling results was studied,and the elemental content correction model about peak value method based on the fluorescence intensity of each element and full spectrum method was established.The results showed that the accuracy of the peak area method model and calibration model of the peak value method based on the X fluorescence intensity of each element fails to achieve a satisfactory detection result.The main reason for the error is that peak area method modeling has large randomness when determining boundaries and the the peak value method based on the X fluorescence intensity of each element does not consider the mutual interference between elements.In contrast,the prediction accuracy of the element content of the full-spectrum correction model considerring the mutual interference between elements is significantly improved.During modeling of nine elemental,the spectral pre-processing method is preferred for each element,and the model accuracy prediction on the best correction model of the content of each element based on the full spectrum is performed.The results showed that prediction correlation coefficients of eight elements such as Mg,P,S,K,Ca,Mn,Fe,and Zn were all above 0.9,besides Cu(0.884).It could be seen that the accuracy of the PLSR model based on the full spectrum can meet the actual needs of production.(4)In the study of identification of wheat producing areas,variance analysis,principal component analysis and stepwise discriminant analysis was combined to explore the feasibility of determining the origin of wheat based on the content of mineral elements.The results of variance analysis showed that there were significant differences among the 9 different elements in different producing areas.After the principal component analysis,the pre-separation of sample data points from different producing areas was preliminarily achieved.Six characteristic elements such as P,S,Ca,Mn,Cu and Zn were selected by stepwise discrimination method.The discriminative model of origin based on the characteristic element spectrum was established.The accuracy rate of model back-generation discrimination is 87.6%,the correct rate of cross-validation is 84.1%,which can effectively distinguish wheat from five producing areas in Heilongjiang,Hebei,Shanxi,Shandong and Jiangsu.It could be seen that the discriminating method of wheat origin based on the characteristic element is feasible.