The relationship between PC1 scores and the chemical compositions of the samples, as well as the influence on the precision and accuracy of PLSR from the difference of the spectrum regions set, were examined. In this paper, the shapes of PCA score plots in each spectral region were compared to determine suitable spectral regions that maximized the complete spectral information. ![]() Maximization of the spectral information for entire spectrum regions requires further examination. In that case, PCA and PLSR were performed on the spectrum regions of the carbonyl and quaternary backbone carbons. The second principal component (PC2) scores reflected the heterogeneity of comonomer sequences. 13 The first principal component (PC1) scores contained most of the information of the chemical compositions of polymers, and PLSR (without assigning resonance peaks) predicted the chemical compositions accurately and precisely. Recently, we have reported the PCA and PLSR of 13C NMR spectra of methyl methacrylate (MMA) and tert-butyl methacrylate (TBMA) copolymers, homopolymers and different blends of two methacrylates (poly MMA (PMMA) and poly TBMA (PTBMA)). PCA was also applied to analyze the Fourier transform infrared spectroscopy and differential scanning calorimetry data of various types of polyethylenes to classify chain-branching types, and chain-branching content and distribution. PCA and PLSR were utilized for the discrimination of ethylene-vinyl acetate copolymers with different compositions and the prediction of the content of vinyl acetate in the copolymers by infrared emission, 9 Raman 10 or near-infrared 11 spectroscopy. Multivariate analysis has also been applied to characterization of polymer materials. 6 An interesting application is the quality classification of Japanese green tea 7 or curative plants 8 through 1H nuclear magnetic resonance (NMR) spectra of their extracts. Principal component analysis (PCA) and partial least squares regression (PLSR), developed by Kowalski 1, 2 and Wold, 3 have been successfully applied to many applications such as the study of the impact of stress conditions on the plant metabolome, 4 the evaluation of neurological disease progression 5 and studies on toxicological mechanisms. Application of multivariate analysis to metabolite evaluations, so-called ‘metabolomics’, is well known. Multivariate analysis is a powerful tool that can transform complex information into more useful sets of information, and that can extract vital differences from data that may look similar by other methods. Multivariate analysis using properly prepared samples provided us with quantitative information of chemical compositions and comonomer sequence distributions, without assignment of the 13C NMR resonance peaks. ![]() Dyad sequence distributions of copolymers that were obtained at higher conversions were successfully determined by PLSR with those of copolymers obtained at low conversions as a training set. The second principal component was found to reflect the fraction of MMA-TBMA hetero dyad sequence. The chemical compositions of 16 copolymers were successfully predicted by partial least squares regression (PLSR). A better linear relationship was found between the first principal component score and the chemical composition in copolymers than was found between the results from spectra of the carbonyl and backbone quaternary carbons. To evaluate chemical compositions and heterogeneity of comonomer sequences in methyl methacrylate (MMA)- tert-butyl methacrylate (TBMA) copolymers, multivariate analysis was applied to the 13C nuclear magnetic resonance (NMR) spectra of the carbonyl, backbone quaternary and α-methyl carbons of the copolymers.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |