Successive heterolytic cleavages of H2 achieve N2 splitting on silica-supported tantalum hydrides: a DFT proposed mechanism.

DFT(B3PW91) calculations have been carried out to propose a pathway for the N(2) cleavage by H(2) in the presence of silica-supported tantalum hydride complexes [(≡SiO)(2)TaH(x)] that forms [(≡SiO)(2)Ta(NH)(NH(2))] (Science 2007, 317, 1056). The calculations, performed on the cluster models {μ-O[(HO)(2)SiO](2)}TaH(1) and {μ-O[(HO)(2)SiO](2)}TaH(3), labelled as (≡SiO)(2)TaH(x) (x = 1, 3), show that the direct hydride transfers to coordinated N-based ligands in (≡SiO)(2)TaH(η(2)-N(2)) and (≡SiO)(2)TaH(η(2)-HNNH) have high energy barrier barriers. These high energy barriers are due in part to a lack of energetically accessible empty orbitals in the negatively charged N-based ligands. It is shown that a succession of proton transfers and reduction steps (hydride transfer or 2 electron reduction by way of dihydride reductive coupling) to the nitrogen-based ligands leads to more energetically accessible pathways. These proton transfers, which occur by way of heterolytic activation of H(2), increase the electrophilicity of the resulting ligand (diazenido, N(2)H(-), and hydrazido, NHNH(2)(-), respectively) that can thus accept a hydride with a moderate energy barrier. In the case of (≡SiO)(2)TaH(η(2)-HNNH), the H(2) molecule that is adding across the Ta-N bond is released after the hydride transfer step by heterolytic elimination from (≡SiO)(2)TaH(NH(2))(2), suggesting that dihydrogen has a key role in assisting the final steps of the reaction without itself being consumed in the process. This partly accounts for the experimental observation that the addition of H(2) is needed to convert an intermediate, identified as a diazenido complex [(≡SiO)(2)TaH(η(2)-HNNH)] from its ν(N-H) stretching frequency of 3400 cm(-1), to the final product. Throughout the proposed mechanism, the tantalum remains in its preferred high oxidation state and avoids redox-type reactions, which are more energetically demanding.