The Study Of The Application Of Chemometric Methods To Capillary Electrophoresis, Fluorescence Spectroscopy And The Quality Control Of Traditional Chinese Medicine

There are nine chapters in this dissertation,which investigates the application of capillary electrophoresis(CE),high performance liquid chromatography and fluorescence spectroscopy combined with chemometric methods including wavelet transform,multivariate calibration and multivariate resolution to the quality control of Traditional Chinese Medicine(TCM)and the simultaneous determination of unseparated compounds.In chapter 1,the development of wavelet transform,multivariate calibration and multivariate resolution and the application of these chemometric methods in analytical chemistry in recent years were reviewed.In addition,there are a review of the application of chemometrics in the quality control of TCM and a brief introduction of chiral discrimination in the chapter.In chapter 2,the application of continuous wavelet transform(CWT)to resolving overlapping peaks from CE was described.The proposed approach was studied and examined on both simulated and experimental data.The quantification results are satisfactory.CE combined with CWT has been shown to be a practicable approach for quantitative determining coeluted compounds in overlapping peaks from CE.In addition,the quantification results from CWT were compared with those from numerical differentiation method.CWT has been shown to be superior to numerical differentiation method for processing experimental signals which contain noise.In chapter 3,CWT combined with experimental design(ED)was applied to determining the enantiomers of tryptophan(Trp).When the baseline separation of L-Trp and D-Trp was not reached by CE,by utilizing CWT to resolve the overlapping peaks of the Trp enantiomers and using ED to overcome the low reproducibility of CE, the composition of the L-Trp and D-Trp were determined precisely.Chapter 4 is concerned with the application of heuristic evolving latent projections (HELP)method to second-order CE data.The HELP method was applied to resolving the overlapping peaks of chloroquine phosphate and promethazine hydrochloride,the two main components in compound reserpine tablets.Satisfactory resolution results were obtained.In chapter 5,a novel strategy which combines ordinary fluorescence spectroscopy, guest-host chemistry and artificial neural networks(ANN),was proposed for determining Trp enantiomers.ANN was employed to establish a quantitative model between the fluorescence spectra of the prepared samples and the compositions of Trp enantiomers.The proposed approach was found to be a high-sensitivity method for determining Trp enantiomers.Chapter 6 is concerned with the application of genetic algorithm-multiple linear regression(GA-MLR)method to the fluorescence spectroscopy for determining the enantiomeric composition of Trp.The difference between the method described in this chapter and the method described in chapter5 is that the calibration model for calculating the concentration of Trp enantiomers was built with GA-MLR instead of ANN.The quantification results of this chapter is better than those of chapter 5,the GA-MLR method being used has been shown superior to the ANN method being used in chapter 5 for determining Trp enantiomers.It is important to correct the retention time shifts in chromatographic fingerprints of TCM for the quality control of TCM.In Chapter 7,it was investigated whether correlation optimized warping(COW)method is a practicable method for correcting the retention time shifts in chromatographic fingerprints of TCM for quality control purpose.The method was successful applied to 27 Radix Puerariae thomsonii(RPT) samples.Chapter 8 aims at developing a chemical pattern recognition method to discriminate processed and unprocessed Rhizoma Corydalis(RC)samples.By combining principal component analysis and ANN,the processed and unprocessed samples were successfully discriminated.In chapter 9,a data compression algorithm based on discrete wavelet transform was proposed to compress the chromatographic fingerprints of TCM.The proposed algorithm was successfully applied to the fingerprints described in chapter 7 and chapter 8.The storage space of the fingerprints of RPT and RC can be reduced by 92.03%and 93.91%respectively.In conclusion,CWT and HELP methods were applied to the resolution of overlapping peaks from CE and a novel strategy which combines fluorescence spectroscopy and multivariate calibration methods was proposed for determining Trp enantiomers.The proposed methods can be applied to quantitative determining the components in mixtures without separating them.In addition,the correction of the retention time shifts in chromatographic fingerprints of TCM,data compression of chromatographic fingerprints of TCM and the recognition of TCM samples were investigated.The proposed approaches can benefit the quality control of TCM.