| Traditional Chinese Medicine (TCM) is one of the precious cultural heritages in our motherland. But its modernization and internationalization are blocked by its poor consistence and instability. As one of the main reasons, the absence of efficient process analytical methods greatly induces the difficulty in the process control of TCM production. Based upon the characteristics of near infrared spectroscopy (NIR), the process analytical theories and methodologies of the application of NIR in TCM production were studied and then the constructed methods were applied to some processes of TCM production successfully in this thesis.Studies on the preparation experimental methods of process analytical chemistry for TCM manufacture using NIRS were first carried out.Uniform design and regression analysis algorithms were first used to select the technical conditions of extracting Coptis chinensis. In the optimization of parameters, the contents of Jatrorrhizine, Palmatine and Berberine in the extraction of Coptis chinensis were determined by HPLC, three equations about the relationship between the process parameters and the contents of the three compounds in the extraction were obtained with stepwise multivariate regression, and the extracting technical conditions were then selected. The optimal conditions predicted by the regression equations were: 75% alcohol as extraction solvent, the ratio between herb and solvent 1/10, fluxing at 90 D, extraction 0.5 h for 2 times. The actual yields of Jatrorrhizine, Palmatine and Berberine obtained by the above optimal conditions were 10.52, 23.58, and 68.17 mg/g herbal powder, respectively, which were better than the results reported in literatures.As only three saponins contents in Panax notoginseng herb had been measured simultaneously by HPLC till now, a reverse phase HPLC method was developed to determine the contents of notoginsenoside RI, ginsenoside Rg1, Rb1 and Rd in the herb. The recovery rates of R1, Rg1, Rb1 and Rd were 89.54%, 90.08%, 82.82%, 84.46% respectively, linear ranges 0.244-6.110, 0.820-20.510, 0.396-9.890, 0.260-6.500 μg respectively, and the RSD of repeatability 5.94, 3.16, 4.17, 2.70% (n=5). The method had been successfully applied to the determination of the above 4 saponins in different specifications and in different parts of the herb from multiple origins, along with colorimetry for the determination of Panax notoginseng saponins (PNS).Furthermore, a new approach using colourimetric method for the determination of mannitol in fermented Cordyceps sinensis was proposed. The recovery was 102.6% and the RSD of repeatability was 1.34% (n=5 ).The present methods ensured the superiority of the extraction approach from natural herband the accuracy of determining the active components in TCM. These methods can be not only applied to the production and process quality control of TCM, but also effectively supported the further investigation about process analysis for TCM based on NIRS method.The methodologies of process analysis of TCM production using NIRS were then studied in detail. The reasonable approaches for sample preparation were investigated. The non-linear relationships between the contents of active components in natural herbs and the NIR spectra of the matrices can be usually found in the NIRS studies about natural herbs, The reasons of the phenomena were analyzed and artificial neural networks(ANN) were selected to model these non-linear relationships. It utilized the significant principal components of the NIR spectra data matrix as the input of the networks so as to overcome the co-linearity in the obtained spectra data. Leave-one-out cross validation procedure during the training of the networks was used to avoid the shortcomings of the canonical BP-ANNs, such as slow training and local optima. The architecture and parameters of networks were selected so as to reduce prediction errors.In order to solve the problem of the absence of process analysis methods in TCM production, fiber-optic NIR transmission spectroscopy (NIRTS) was applied...
|
PDF Full Text Request