The Analysis Of Complex System By Using High Performance Liquid Chromatography And Related Chemometric Methods

In this thesis, the developments and applications of multivariate calibration and resolution in chemometrics have been reviewed. Several three-way data resolution methods based on alternating least squares are mainly introduced, and these methods are applied to resolve the overlapping peaks of complex multi-component systems from high performance liquid chromatography-diode array detection (HPLC-DAD). The superiority is showed that chemometric multivariate resolution which can be regarded as "mathematical separation" can improve or replace "chemical separation".There are four parts in this thesis, the main works of author are summarized and presented as follows:1. Multivariate curve resolution with alternating least squares (MCR-ALS) and alternating trilinear decomposition (ATLD) are applied in the resolution of overlapping peaks of camptothecin (CPT) and 7-ethyl-10-hydroxy- camptothecin (SN-38) from HPLC-DAD. The results show that the resolved spectra coincide perfectly with real spectra and the quantitative results are satisfactory.2. MCR-ALS and Jack-knife PARAFAC are applied in the resolution of overlapping peaks of irinotecan (CPT-11), SN-38 and CPT (as impurity) from HPLC-DAD. The satisfactory resolution results are obtained, though high similarity present between the unit spectra of CPT and CPT-11.3. The spectral correlative chromatography method is used to identify the correlative component peak clusters of tetrahydropalmatine in 15 different chromatographic fingerprints of Rhizoma Corydalis. Other chemometrics methods are used to test the purity and determine number of components in the peak clusters of tetrahydropalmatine. The accurate quantitative results are also obtained.