NMR shieldings in benzoyl and 2-hydroxybenzoyl compounds. Experimental versus GIAO calculated data

GIAO-calculated NMR chemical shifts (H-1,C-13, and O-17) as Obtained at Various computational levels are reported for the three parent compounds phenol, benzaldehyde, and salicylaldehyde, and for 13 different benzoyl and the 13 corresponding 2-hydroxybenzoyl compounds. The data are compared with experimental solution data, focusing on the agreement with spectral patterns and spectral trends. The influence of different optimized geometries (HF, MP2, B3LYP, BLYP), basis sets (6-31G(d,p) up to 6-311++G(2df,2dp)), and levels of theory (HF, B3LYP, BLYP) was investigated systematically by exhaustive calculations on the three parent compounds. With regard to the results obtained from this foregoing study, the GLAO calculations for the compounds of the two series were performed at two levels of theory, HF/6-311++G(d,p) and BLYP/6-311++G(d,p) for both the B3LYP/6-31G(d,p) and the HF/6-31G(d,p) optimized geometries. It turned out that, with the exception of the nuclei of the hydrogen-bonded OH groups, B3LYP and HF optimized geometries yield rather similar results. For aromatic carbons and protons, because of systematic shortcomings, the GIAO-HF calculations are distinctly worse than the GIAO-BLYP calculations. In the latter case, interchanges with respect to the experimental spectral patterns are obtained only in few instances and concern nuclei with rather small chemical shift differences (within 4 ppm for carbons, within 0.5 ppm for hydrogens). For the nuclei of the C=O and O-H groups, the experimentally observed spectral trends are reproduced in similar quality at both the HF and the BLYP levels of theory.