Vanadyl bond dissociation energies are calculated by density functional theory (DFT). While the hybrid (B3LYP) functional results are close to the available reference data, gradient corrected functionals (BP86, PBE) yield large errors (about 50 to 100 kJ mol(-1)), but reproduce trends correctly. PBE calculations on a V(20)O(62)H(24) cluster model for the (001) surface of V(2)O(5) crystals virtually reproduce periodic slab calculations. The low bond dissociation energy (formation of oxygen surface defect) of 113 kJ mol(-1)(B3LYP) is due to substantial structure relaxations leading to formation of V-O-V bonds between the V(2)O(5) layers of the crystal. This relaxation cannot occur in polyhedral (V(2)O(5))(n) clusters and also not for V(2)O(5) species supported on silica or alumina (represented by cage-type models) for which bond dissociation energies of 250-300 kJ mol(-1) are calculated. The OV(OCH(3))(3) molecule and its dimer are also considered. Radical cations V(2)O(5)(+) and V(4)O(10)(+) have very low bond dissociation energies (22 and 14 kJ mol(-1), respectively), while the corresponding radical anions have higher dissociation energies (about 330 kJ mol(-1)) than the neutral clusters. The bond dissociation energies of the closed shell V(3)O(7)(+) cation (165 kJ mol(-1)) and the closed shell V(3)O(8)(-) anion (283 kJ mol(-1)) are closest to the values of the neutral clusters. This makes them suitable for gas phase studies which aim at comparisons with V(2)O(5) species on supporting oxides.
Read full abstract