How well can new-generation density functionals describe protonated epoxides where older functionals fail?

In a recent article, Carlier et al. (J. Org. Chem. 2006, 71, 1592) examined the prediction of several DFT functionals and showed that the most popular density functional, B3LYP, and 15 others fail badly for the prediction of the structure of protonated 2-methyl-1,2-epoxypropane. In this note, we compare the performance of several recently developed density functionals for the calculation of structures and energetics of protonated cyclic ethers, including epoxides. We found that several of the newly developed DFT methods perform better than B3LYP or any of the other 17 functionals examined by Carlier. We conclude that a recently published functional, M05-2X, has greatly improved performance for an unsymmetrical protonated epoxide, and we recommend this functional for studies that involve protonated epoxides and protonated ethers.