Ability of DFT to evaluate the torsional barriers in neutral and CO* protonated aromatic carbonyl compounds

Before the recent development of new functionals, the Density Functional Theory (DFT) was considered as a failing quantum chemical method in accurately computing the rotational barrier height of the gaseous benzaldehyde. Since the 2004 polemical Speakman's paper about the accuracy of microwave value of this quantity L. D. Speakman, B. N. Papas, H. L. Woodcock, H. F. Schaefer, J. Chem. Phys. 2004, 120, 4247], the question is still relevant. This paper aims to display the ability of the DFT to evaluate the torsional barriers of a series of para‐substituted benzaldehydes in solution. The method is also tested in computing barriers of other solvated aromatic carbonyl compounds (in both neutral and carbonyl protonated forms) for which accurate experimental data are available. Computations have been carried out at two DFT methods using the popular hybrid generalized‐gradient‐approximation (GGA) density functional B3LYP and the global hybrid meta‐GGA Minnesota functional M06‐2X with a 6‐311++g (2d,2p) basis set. Solvent effects investigations were undertaken within the framework of the polarisable continuum model solvation approach. It has been concluded that the used computational methods are very satisfactory in predicting barriers of para‐substituted benzaldehydes in the liquid phase and less satisfactory for their CO*‐protonated forms. M06‐2X functional particularly leads to full agreement. For para‐substituted acetophenones, good agreement with experiment is observed only with M06‐2X applied to the CO*‐protonated forms. The present investigations have allowed to demonstrate that the Density Functional Theory does not constantly fail in accurately computing the rotational barrier heights of aromatic carbonyl compounds. Copyright © 2014 John Wiley & Sons, Ltd.