Comparison of experimental and computationally predicted sulfoxide bond dissociation enthalpies

The accurate estimation of S-O bond dissociation enthalpies (BDE) of sulfoxides by computational chemistry methods has been a significant challenge. One of the primary causes for this challenge is the well-established requirement of including high-exponent d functions in the sulfur basis set for accurate energies. Unfortunately, even when high-exponent d functions were included in Pople-style basis sets, the relative strength of experimentally determined S-O BDE was incorrectly predicted. The aug-cc-pV(n+d)Z basis sets developed by Dunning include an additional high-exponent d function on sulfur. Thus, it was expected that the aug-cc-pV(n+d)Z basis sets would improve the prediction of sulfoxide S-O BDE. This study presents the S-O BDE predicted by B3LYP, CCSD, CCSD(T), M05-2X, M06-2X, and MP2 combined with aug-cc-pV(n+d)Z, aug-cc-pVnZ, and Pople-style basis sets. The accuracy of these predictions was determined by comparing the computationally predicted values to the experimentally determined S-O BDE. Values within experimental error were obtained for dialkyl sulfoxides when the S-O BDEs were estimated using an isodesmic oxygen transfer reaction at the M06-2X/aug-cc-pV(T+d)Z level of theory. However, the S-O BDE of divinyl sulfoxide was overestimated by this method.