The Identification And Adulteration Analysis Of Angelica Sinensis And Its Similar Products Based On Ultraviolet-visible Diffuse Reflectance Spectroscopy And Chemometrics

As one of the most important traditional Chinese medicines(TCMs),Angelicae Sinensis Radix(ASR,Danggui in Chinese)which derived from the root of Angelica sinensis(Oliv.)Diels,has been used for more than 2000 years in China.It has been widely used to treat gynecological disorders,enhance the immune system and relieve constipation because of its blood replenishing,pain reducing and intestine moistening effects.Due to high medical value and big market demand,the price of ASR keeps relative high.Intentional substitutes or adulteration are often made.Thus,precise identification and quantification of raw material is crucial for quality control of TCMs and safety of the consumers.This thesisinvestigates the feasibility of identifying and quantifying ASR and its similarities and adulterants by UVvisible diffuse reflectance spectroscopy(UV-Vis DRS)combined with chemometrics.The specific research contents are as follows:(1)A rapid and easy-to-use shortcut method was developed for nondestructive identification of ASR from its four similarities based on UV-Vis DRS and chemometrics.A total of 191 samples,including 40 ASR,39 Angelicae Pubescentis Radix(APR),38 Chuanxiong Rhizoma(CR),35 Atractylodis Macrocephalae Rhizoma(AMR)and 39 Angelicae Dahuricae Radix(ADR)were collected and divided into the training and prediction sets.Principal component analysis(PCA)was used for observing the sample cluster tendency of the training set.Different preprocessing methods were investigated and the optimal preprocessing combination was selected according to spectral signal characteristics and PCA clustering results.The final discriminant model was built by extreme learning machine(ELM).The results show that the classification of the five kinds of TCMs cannot be achieved by PCA on the raw spectra.Autoscaling,continuous wavelet transform(CWT)and SG smoothing can improve the clustering results in different degrees.Furthermore,their combination in the order of CWT + autoscaling + SG smoothing can enhance the spectral resolution and obtain the best clustering result.By using this combination,100% classification accuracy can be achieved by ELM modeling on both the training set and the prediction set.Therefore,the developed method could be used as a rapid,economic and effective method for discriminating the five kinds of TCMs(2)UV-Vis DSR combined with chemical pattern recognition to identify pure ASR and its adulterants.In this study,75 types of angelica and adulterated samples were designed,with mass fractions of 0-100%,and identified with 40 pure angelica samples.The effects of preprocessing methods such as autoscaling,CWT,SG smoothing and their combination were investigated.After proper pre-processing,a PLA-DA model is established.The results show that PCA of the original spectrum cannot achieve clustering of ASR and its adulterants.After CWT,the samples were grouped into three categories in the two-dimensional principal component space,which were respectively for pure ASR,forbidden APR adulterated products with adulteration below 60%,and for APR adulterated products adulterated above 60%.It shows that the spectrum can not only realize the classification of ASR and adulterated APR after CWT pretreatment,but also adulteration with different adulteration percentages can be further distinguished.After CWT + SG smoothing + autoscaling pre-processing,a partial least squares-discriminant analysis model is established.Using only one principal component can be used to identify 100% of ASR and its adulterants.(3)The feasibility of further quantitative analysis of adulterated ASR by UV-Vis DRS combined with multivariate calibration.First,81 binary adulterated samples of ASR and APR were prepared,and the UV-Vis DRS spectra of the samples were collected.Five multivariate calibration methods including principal component regression,partial least squares regression,artificial neural network,support vector regression,and extreme learning machine are compared.On the basis of the best modeling method,a variety of preprocessing methods such as SG smoothing,multiple scattering correction(MSC),standard normal variables(SNV),first derivative,second derivative,CWT,and their combined preprocessing effects were investigated.The results show that UV-Vis diffuse reflectance spectroscopy combined with chemometrics can accurately and quantitatively analyze the binary adulterated samples of ASR.