B3LYP vs BP86 functionals for orgnometallics

["Dr. Peter Burger" <chburger' at \`aci.unizh.ch>]

Dear CCL netters,
 could perhaps someone point me towards a critical review/paper,
 in  which _experimentally_ derived relative energies for
 organometallic isomers have been compared to results from B3LYP and BP86
 DFT calculations?
 Somehow, the literatur seems to indicate that the B3LYP rel. energies are
 more accurate than those obtained with BP86 or BLYP functionals. But is
 this really justified? - in particular since B3LYP is "parametrized"
 for
 organic systems? What about results of groups that use exclusively pure
 DFT programs, e.g. DMOL or ADF? Are there studies less reliable
 (exaggerated) ???!
 Comments are also welcome.
 Regards,
 Peter
 --------------------------------
 Peter Burger
 Anorg.-chem Institut
 Universitaet Zuerich