ISSN:
0749-1581
Keywords:
13C NMR substituent effects
;
Principal component analysis
;
Electronegativity correlation
;
Chemistry
;
Analytical Chemistry and Spectroscopy
Source:
Wiley InterScience Backfile Collection 1832-2000
Topics:
Chemistry and Pharmacology
Notes:
Principal component (PC) or factor analysis was applied to 13C substituent effects (literature SCS values) in α to δ positions of 47 aliphatic frameworks with 4-13 different substituents X. Two PC terms for each substituent X are sufficient to describe 99.3% of all C-α shifts within 〈 ± 0.6 ppm, even in the largest data matrix containing 13 × 47 = 611 shifts; shifts in β. γ and δ positions require a third PC term for a similar percentage SCS accuracy. Scaling problems with the different order of magnitudes between α to δ SCS variations are avoided by analysing the corresponding data sets separately; this approach, together with the large data base, leads to a dominating first PC term for C-α which unambiguously correlates with substituent electronegativities (EN), in contrast to an earlier factor analysis. The same result is obtained by direct comparison of α-SCS with different EN scales, including for the first time group electronegativities. PCs for β to γ SCS indicate only less consistent trends, which again agree partially with results from direct correlations, but not with EN values. The present PC analysis did not include stereochemical descriptors, but furnishes clear distinctions between clusters of (pseudo-) equatorially or axially substituted compounds. The PC loadings for C-α also lead to the unambiguous distinction of primary, secondary, tertiary and quaternary carbon atoms.
Additional Material:
8 Ill.
Type of Medium:
Electronic Resource
URL:
http://dx.doi.org/10.1002/mrc.1260300110
Permalink