BP Neural Network Combined With FANWE And Three-dimensional Fluorescence Spectroscopy To Measure The Fluorescent Whitening Agent In Whitening Mask

Fluorescent whitening agent(FWAs)is a kind of fluorescent dye.Because of its brightening and brightening effect,FWAs is widely used in various industrial fields,including cosmetics.However,FWAs has no whitening effect and is harmful to human body.In recent years,people pay more and more attention to the ingredients in cosmetics(including whitening mask).How to detect FWAs in cosmetics has become a research hotspot.In this paper,the whitening principle,types and molecular structure of FWAs are first described.By analyzing the existing methods of measuring FWAs in cosmetics,the principle of fluorescence emission,the factors affecting fluorescence and the quantitative analysis of the use of fluorescence method are analyzed.This lays the groundwork for the feasibility of using three-dimensional fluorescence method to detect the feasibility of FWAs in whitening mask.The first detection and analysis method is trilinear decomposition combined with fluorescence spectrometry.In the experiment,the three most commonly used target analytes,FWAs VBL,CBS-X and BBU,were used to prepare the sample solution with methanol as the solvent,and the data were pretreated: the blank subtraction method was used to carry out Raman scattering,excitation emission correction for the initial three-dimensional fluorescence spectrum data,and the trained BP neural network method was used to compress the data.Because the spectrum of the three FWAs overlaps seriously,after using the core undia method to determine the component fraction,two algorithms,ATLD and anwe,are used to decompose the three-dimensional data,and effective and correct qualitative and quantitative analysis results are obtained.Then the detection and analysis method is: tetralinear decomposition combined with fluorescence spectrometry.In the experiment,the fourth dimension solvent was introduced to construct the four-dimensional fluorescence spectrum data.Three groups of experimental samples were obtained by using methanol(spectral level),ethanol(spectral level)and ultra pure water respectively,and the configuration method of the three groups of samples was the same as the amount of drugs added.Fs920 steady-state fluorescence spectrometer is used to detect the samples.After preprocessing the four-dimensional data,two kinds of fourdimensional spectral data are decomposed using four linear decomposition algorithms(aqld and fanwe).Among them,the fanwe algorithm is obtained by applying the three linear decomposition algorithm anwe to the four linear decomposition algorithm.Both the aqld and fanwe algorithms are efficient and correct,and are not sensitive to excessive group fraction.