Geometry Optimization of a Ru(IV) Allyl Dicationic Complex:  A DFT Failure?

Five pure and four hybrid DFT functionals associated with VDZP, VTZP, and VQZP basis sets are tested (Gaussian 03) for their performance on the geometry optimization of [Ru(η(5)-C5H5)(η(3)-CH2CHCHC6H5)(CH3CN)2](2+). When the calculated geometries were compared with the X-ray structure determination for the analogous complex with permethylated cyclopentadienyl, it was found that in all cases the coordination mode of the η(3)-allyl was very poorly described, despite the functional used. The Ru-C bond distance corresponding to the substituted allyl carbon was overestimated by 0.23-0.50 Å, depending on the functional and the basis set used. These results were reproduced by further testing carried out with the ADF program and larger basis sets. MP2 leads to an acceptable value for the same Ru-C distance, with an underestimation of 0.07 Å, suggesting that, at least in the case of the functionals tested, DFT does not provide an accurate description of a weak Ru-C interaction.