The Research On Rapid Quality Evaluation Of Sichuan Genuine Medicinal Materials Fritillariae Cirrhosae And Zanthoxyli Pericarpium Based On Portable Near-Infrared Spectrometer

The quality of Chinese medicinal materials（CMM）is the key factor affecting the quality of Chinese medicinal preparations（CMP）.The NIR technology can be used to evaluate the quality of CMM quickly,which is one of the effective means to control the quality source of CMP.Fritillariae cirrhosae and Zanthoxyli Pericarpium are both representative varieties of Sichuan genuine medicinal materials.Due to their high price and wide variety,there are many varieties confusion and adulterated sales phenomenon in the market,which seriously interferes with the effectiveness of the control of the quality source of traditional CMP.The traditional sensory evaluation methods for evaluating the quality of Fritillariae cirrhosae and Zanthoxyli Pericarpium have some shortcomings,such as strong subjectivity and poor inheritance.However,the chemical methods based on liquid chromatography also have disadvantages such as tedious operation,long detection time and environmental pollution.Therefore,this subject to a major science and technology projects in Sichuan province-"other high-end general scientific instruments and equipment development of（portable near infrared spectrometer）（No.2018 tzdzx0007）"as the backing,the quality of Fritillariae cirrhosae and Zanthoxyli Pericarpium from Sichuan Province were evaluated by a new portable near-infrared（NIR）spectrometer.A rapid analysis method for the variety,alcohol-soluble extract and total alkaloid content and adulteration of Fritillariae cirrhosae was established based on portable NIR spectrometer,and a real-time identification method for the origin,volatile oil content,crack rate and grade of Zanthoxyli Pericarpium was established.The main research and conclusions of this paper are as follows:（1）The NIR spectra of Fritillariae cirrhosae,non-Fritillariae cirrhosae,common forgeries Bulbus Tulipae and adulterated samples were collected.The contents of alcohol soluble extracts and total alkaloids in 72 batches of Fritillariae cirrhosae were determined according to pharmacopoeia method.The qualitative models of Fritillariae cirrhosae of different varieties,Fritillaria of different varieties,Fritillaria-adulterated Fritillariae cirrhosae,different adulterated Fritillariae cirrhosae and quantitative models of alcohol soluble extract,total alkaloid content and adulteration ratio of Fritillariae cirrhosae were modelling.The data showed that the accuracy of calibration set and validation set of Fritillariae cirrhosae of different varieties were 79.63%and 83.33%,respectively,using the powder spectrum pretreated with 1^st D,which indicated that the portable NIR spectrometer was more effective in identifying Fritillariae cirrhosae of different varieties than the original crude material,and could realize the rapid identification of different varieties of Fritillariae cirrhosae.In the quantitative model,the R_C²,R_p²,RMSEC and RMSEP of the alcohol soluble extract model after MSC pretreatment were 0.9363,0.8382,0.2975 and 0.4147,respectively,which indicated that the calibration and prediction error of the model were small.In the total alkaloid model,the R_C²,R_p²,RMSEC and RMSEP were 0.7144,0.6312,0.0168 and 0.0128,respectively,indicating that although the correlation effect of the model is not good,it still has the applicability in predicting the total alkaloid content of Fritillariae cirrhosae.Meanwhile,the best qualitative models for Fritillariae cirrhosae,other Fritillaria thun-bergli and Bulbus Tulipae were established by using unpretreated spectra.The accuracy of calibration set and validation set were 93.26%and 90.91%,respectively,indicating that Fritillariae cirrhosae could be clearly distinguished from other samples.In addition,in 72 Fritillariae cirrhosae samples and 570 batches of different types of adulterated Fritillariae cirrhosae,the recognition rates of Fritillariae cirrhosae and adulterated Fritillariae cirrhosae by the model were 99.49%（calibration set）and 100.00%（validation set）,respectively,which indicated that the NIR spectrometer could distinguish whether Fritillariae cirrhosae adulterated successfully.However,after SG-Smoothing pretreatment,the accuracy of calibration set and validation set of adulteration type are 70.47%and 73.68%,respectively,indicating that the effect is not very ideal.In the specific proportion of adulteration,the Fritillariae cirrhosae adulteration of Fritillaria thunbergia Miq.、Fritillaria ussuriensis Maxim、Fritillaria pallidiflora Schrenk、Fritillaria hupehensis Hsiao et K.C.Hsia、Bulbus Tulipae and Flour.These six quantitative model R_C² were 0.8811,0.9849,0.8288,0.9654,0.9768,0.9906,R_p² were 0.8402,0.9602,0.6792,0.9025,0.9574,0.9761.The results showed that the actual value of the adulteration ratio had a high correlation with the predicted value,and the portable NIR spectrometer could be used to determine the adulteration ratio quickly and accurately.（2）A new portable NIR spectrometer was used to collect the NIR spectrum of 274Zanthoxyli Pericarpium samples,and the volatile oil content and crake rate of the samples were determined.By comparing the modeling results of different spectral preprocessing methods,the NIR qualitative analysis model of Zanthoxyli Pericarpium origin and grade was established based on linear discriminant analysis（LDA）method,and the quantitative analysis model of volatile oil content and crack rate was established based on partial least squares（PLS）regression.The results showed that the origin identification rates of the model calibration set and the validation set were 91.26%and86.76%,respectively,after the 1^st D preprocessing spectrum,which indicated that the portable NIR spectrometer could be used to accurately identify the origin of Zanthoxyli Pericarpium.At the same time,the 2^nd D model is established after pretreatment with good performance,the accuracy of calibration set and validation set were 67.48%and58.82%respectively,using portable NIR spectrometer and LDA method not well forecast Zanthoxyli Pericarpium grade information,but the external rating information associated with inner NIR spectrum data with larger,indirect verify the scientific of theory“Quality Evaluation Through Morphological Identification”.In addition,the volatile oil content model had the best performance after 1^st D pretreatment,in which the calibration set R_C² was 0.8655,RMSEC was 0.3148,the validation set R_p² was0.8386,RMSEP was 0.3779,respectively.And the R_C²,R_p²,RMSEC and RMSEP of the crack rate model were 0.8201,0.8259,6.1484 and 6.6049 after 1st D+SNV pretreatment.