Is the allylpalladium structure altered between solid and solutions?

Recent EXAFS measurements on [(Ph2PCH2CH2PPh2)Pd(H2CCHCMe2)]O3SCF3 (Tromp et al. J. Am. Chem. Soc. 2002, 124, 14814) were interpreted as evidence that, when the complex is dissolved in THF, the allyl ligand adopts an eta(2) structure with a dangling allyl CH2 substituent. DFT calculations of the Pd complex using H2P-CH2CH2-PH2 as a model for Ph2P-CH2CH2-PPh2 (dppe), in the absence or the presence of the triflate counteranion, and modeling the THF solvent by explicit Me2O molecules or by a continuum model give always a conventional eta(3)-H2CCHCMe2 structure with equal Pd-C bonds to the terminal carbon centers of the allyl. QM/MM calculations using the dppe ligand also fail to support an eta(2)-allyl structure as a global minimum. The EXAFS parameter space is shown to have multiple minima. These have very similar overall EXAFS, but have very different structural parameters. The minimum that was the basis for the previous structural conclusion gives a slightly better fit but has unrealistic Debye-Waller factors and threshold energies.