Study On The Quality Analysis Of Xiangshayangwei Pills(Condensed) By Near-infrared Spectroscopy And Chromatographic Technique

This paper mainly study on the application of near-infrared spectroscopy and chromatography technology for the quality analysis of Xiangshayangwei pills(condensed).The main content is divided into the following three parts:The first part is the quality analysis of Amomum villosum, Citrus aurantium by near-infrared spectroscopy. Combine the near-infrared spectroscopy with partial least squares(PLS) is applied to establish the quantitative calibration model of bornyl acetate, synephrine and ethanol-extract. The correlation coefficients(R2) are 0.99259,0.98052 and 0.98489. The root mean square error of cross validation(RMSECV) are0.0714,0.02239 and 0.41789. The root mean square error of calibration(RMSEC) are 0.0145,0.00915 and 0.153. The root mean square error of predication(RMSEP) are 0.0167,0.0119 and 0.188. The performance index(PI) are 90.0,81.9 and 86.3. The method is fast, convenient and accurate and it’s feasible to determine the content of bornyl acetate in Amomum villosum, synephrine and ethanol-extract in Citrus aurantium by near-infrared spectroscopy.The second part is the quality analysis of Xiangshayangwei pills(condensed) by near-infrared spectroscopy. First, combine the near-infrared spectroscopy with partial least squares(PLS) is applied to establish the quantitative calibration model of magnolol and honokio and hesperidin in Xiangshayangwei pills(condensed)(made by A). The correlation coefficients(R2) are 0.99649 and 0.97771. The root mean square error of cross validation(RMSECV) are0.35900 and 0.12083. The root mean square error of calibration(RMSEC) are 0.0564 and 0.0994. The root mean square error of predication(RMSEP) are 0.0938 and 0.0958. The performance index(PI) are 86.0 and 81.8. Second, combine the near-infrared spectroscopy with partial least squares(PLS) is applied to establish the quantitative calibration model of magnolol and honokio and hesperidin in Xiangshayangwei pills(condensed)(produced by three different manufactures). The correlation coefficients(R2) are 0.98106 and 0.96258. The root mean square error of cross validation(RMSECV) are0.28900 and 0.18306. The root mean square error of calibration(RMSEC) are 0.112 and 0.124. The root mean square error of predication(RMSEP) are 0.151 and 0.128. The performance index(PI) are 78.3 and 86.9. Discriminant analysis and clustering analysisis are applied to identify Xiangshayangwei pills(condensed) of different manufactures.For the discriminant analysis model, it can accurately distinguish the sample of three different manufacturers, and the verify accuracy is 100%. For the clustering analysis model, it can clearly clustered the samples of three manufacturers into three groups, the boundaries is clear, and attribution is accurate, the result of clustering analysis is consistent the experimental result of discriminant analysis.The third part is the study of HPLC fingerprint of Xiangshayangwei pills(condensed). Through the study of the method of extraction, solvent extraction, chromatographic conditions, the methods for the analysis of HPLC fingerprint of Xiangshayangwei pills(condensed) is determined, and analyzes sample of three different manufacturers. The results show that the similarity of samples from three different manufacturers were above 0.88, and the samples from WanXi Henan is more stable. At the same time,through the clustering analysis and principal component analysis of HPLC fingerprint of the sample of three different manufacturers, the sample of three manufacturers is clearly divided into three categories. The method is accurate and specific, and it can provides reference for Xiangshayangwei pills(condensed) on quality evaluation.