Indexed on: 13 Nov '10Published on: 13 Nov '10Published in: Journal of Physical Chemistry A
A systematic assessment of theoretical methods applicable to the accurate characterization of catalytic cycles of homogeneous catalysts for H(2) oxidation and evolution is reported. The key elementary steps involve heterolytic cleavage of the H-H bond and formation/cleavage of Ni-H and N-H bonds. In the context of density functional theory (DFT), we investigated the use of functionals in the generalized gradient approximation (GGA) as well as hybrid functionals. We compared the results with wave-function theories based on perturbation theory (MP2 and MP4) and on coupled-cluster expansions [CCD, CCSD, and CCSD(T)]. Our findings indicate that DFT results based on Perdew correlation functionals are in semiquantitative agreement with the CCSD(T) results, with deviations of only a few kilocalories/mole. On the other hand, the B3LYP functional is not even in qualitative agreement with CCSD(T). Surprisingly, the MP2 results are found to be extremely poor, in particular for the diproton Ni(0) and dihydride Ni(IV) species on the reaction potential energy surface. The Hartree-Fock reference wave function in MP2 theory gives a poor reference state description for these states that are electron rich on Ni, giving rise to erroneous MP2 energies. We present a detailed potential-energy diagram for the oxidation of H(2) by these catalysts after accounting for the effects of solvation, as modeled by a polarizable continuum, and of free energy estimated at the harmonic level of theory.