Sulfoxides play important roles in organic and medicinal chemistry, as chiral handles, [1] ligands, [2] or bioisosteres.[3]The most common method for sulfoxide preparation is controlled sulfide oxidation, [4] but a more convergent approach would be to make them by formation of the C À S bond. Poli, Madec and co-workers recently disclosed a method for direct sulfoxide formation involving a Pd-catalyzed allylation or arylation of an intermediate sulfenate anion. [5] This strategy is illustrated by the reaction between tertbutyl 3-p-tolylsulfinylpropionate and cinnamyl acetate, which gave the cinnamyl p-tolyl sulfoxide. In this reaction, the sulfenate anion and h 3 -allyl complex are generated in situ and react with each other to afford the final product. Herein, we disclose our theoretical results for the formation of the allyl sulfoxides (Scheme 1).A simplified mechanistic scheme for the formation of phenyl allyl sulfoxide is shown in Scheme 2. Sulfenate anion and h 3 -allyl Pd complexes are formed in situ and react by nucleophilic displacement. However, it is not a priori clear whether initial attack occurs from the formally anionic oxygen atom or the formally neutral sulfur atom. The observed product is the allyl sulfoxide, but this would be expected to be formed also from a rapid Mislow-BravermanEvans (MBE) rearrangement [6] even if the allyl sulfenate would be the initially formed product. Since the two mechanisms are closely linked, we initiated our investigation by a detailed computational study of the MBE rearrangement. Our standard workhorse for computational investigations of reaction mechanisms is DFT at the B3LYP level [7] with a polarized double-z basis set (6-31G*), augmented by vibrational corrections, ECP basis sets for heavy elements, [8] continuum solvation, [9] and dispersion corrections.[10] However, Jorgensen and co-workers showed that even though trends in the MBE rearrangement kinetics could be well reproduced by similar methods, the equilibrium was erroneously calculated to favor the sulfenate product, not the sulfoxide.[11] We could verify that the same problem exists also when applying our standard methods to methyl allyl sulfoxide. The correct preference for the sulfoxide could only be reproduced by using large basis sets. The results seemed reasonably well converged at the cc-PVDZ level, but due to a slight shift in energy when going to the cc-PVTZ basis [12] we selected the latter basis set for further studies. Geometries were well converged already at the 6-31G* level, as evidenced by small differences in energy between cc-PVTZ and cc-PVTZ//6-31G* calculations. Single-point cc-PVTZ calculations at 6-31G*-derived IRC points also revealed that the transition state of the rearrangement shifted by less than 0.1 with the larger basis set. Details of these investigations can be found in the Supporting Information. In conclusion, to ensure proper energetics in our studies of the Pd-catalyzed process, we performed geometry optimizations at the B3LYP level with the lacvp* basis set, [8] val...