Rapid Analysis Of Viscous Or Darker Samples By Multiattenuated Total Reflection Or Raman Scattered Spectroscopy With Multi Variate Methods

Viscous samples can be found everywhere in life,such as edible oil,sesame paste and SBS modified asphalt.It has been found in markets that products of high prices are often adulterated with that of low prices or poor qualities.In food area,for examples,olive oil is adulterated with other edible oils of low prices,and sesame paste with peanut butter.Using repeated frying oil is harmful to human health.Those behaviors have seriously violated the interests of consumers.In high way construction area,the adulterated asphalt materials seriously reduce the service life of the high way and brought huge economic loss every year.Therefore,it is very important to supervise the quality of those products in the markets and to need rapid analytical methods for them.The traditional analytical methods for viscous samples are usually time consuming,high cost or using environmentally harmful solvents and not suitable to on site supervision of quality of those products.So,it is very urgent for the quality supervision in markets to develop rapid analytical methods for viscous samples.Molecular spectroscopy with multivariate analysis methods has been applied in many fields in virtue of fast,nondestructive and determining multi-properties simultaneously.However,the popular ways of measuring spectra are transmission and diffuse reflection,and not suitable to the most of viscous sample due to the difficulties in sample loading and cleaning.It is not easy for dark samples to get high quality infrared spectra using transmission and diffuse reflection.In this paper a new analytical method using attenuated total reflection Fourier transform infrared spectroscopy(FTIR-ATR)or Raman scattered spectroscopy with multi-variate methods was proposed to realize rapid analysis of viscous samples.Firstly,an identification method of class of edible oils is established using Raman spectroscopy with SIMCA.Ninety four samples of 6 classes of edible oils including canola oil,camellia oil,soybean oil,olive oil,peanut oil,corn oil were collected from different producing areas and different producers.PCA was applied on the raw Raman spectra of edible oils.It is indicated that all kinds of edible oils were not distinguished in PCA space due to the influence of fluorescence of sample matrix.air-PLS algorithm was applied on the raw Raman spectra to reduce the fluorescence signal.In PCA model,the clustering tendency of edible oils can be improved greatly,but it is still unable to distinguish canola oil from peanut oil.The air-PLS treated spectra were used using SIMCA to establish the identification model of 6 kinds of edible oils.All the samples can be identified by SIMCA on the basis of PCA with a recognition rate 100%.Next step,a fast quantitative analytical method of content of the total polar compound(TPC)in frying oils is established using FTIR-ATR with PLS.Forty-four samples of 4 classes of edible oils including canola oil,soybean oil,cottonseed oil and palm oil were prepared and the content of total polar compound were determined according to the standard GB/T 5009.202-2003.Using the correlation coefficient method and 2D-IR correlation infrared spectroscopy method to select the wavelength ranges for model building.The quantitative calibration model which the wavelength ranges were selected by 2D-IR correlation infrared spectroscopy has higher accuracy,with SEC 0.928 and SEP 0.995.Six frying oil samples were in random selected to test the repeatability of the model and the maximum standard deviation is 0.25.It provides a new method for the determination of total polar compound content in frying oils.Then,the sesame paste samples were prepared with a peanut butter content ranged from 2.0%to 96%.Quantitative calibration model was established by using FTIR-ATR with PLS,with SEC 3.264 and SEP 3.808.Sixteen samples and five samples in it were used to test the accuracy and repeatability of the model.It proves that PLS model has excellent accuracy and reproducibility.Finally,an identification method of class of qualified and unqualified SBS modified asphalt is proposed using similarity 0 and residual of FTIR-ATR spectra.The threshold value of similarity 0 and residual is determined by comparing the value of similarity 0 and residual between IR spectra of qualified SBS modified asphalt.Then,according to the threshold,the value of similarity 0 and residuals between IR spectra of qualified SBS modified asphalts and unknown samples are compared to determine the unknown sample’s quality.