Search

51V Magic-angle-spinning (MAS) NMR has been applied to V2O5 at two different magnetic field strengths (4.7 and 7.1 T). Both the magnitude and relative orientation of the quadrupole and chemical shift (CS) tensors have been determined by iterative fitting of the V MAS NMR lineshapes at the two magnetic field strengths. The reliability of the results is discussed. Moreover, it is shown that previous low-field single-crystal data are fully consistent with the high-field powder-sample MAS NMR results provided that a slight noncoincidence between the CS tensor and the crystal frame axes is considered. The electric field gradient tensor at the vanadium and lithium sites is subsequently used to test several electronic structure calculation at an ab initio Hartree-Fock level in V2O5 and γ-LiV2O5 crystals. It is shown that a wide distribution of oxygen charges must be considered to describe the particular environment of each type of oxygen atoms. Furthermore, this analysis supports the fact that the vanadyl bond is likely a short ionic bond. NMR is found to be a valuable experimental tool to get insight into the nature of chemical bonds in vanadium oxides.

Sterically hindered ligands generally form metal complexes with lower coordination numbers than less hindered ligands. In contrast to dogma, the solid state vanadium(V) complexes with ethylene glycol and pinacol contain four- and five-coordinate vanadium atoms, respectively. Ab initio and electrostatic potential distribution calculations were conducted on both experimental and optimized geometries of the four- and five-coordinate oxovanadium chloroalkoxides. Ab initio energy calculations favor the five-coordinate species for all ligands at all levels of theory examined. No significant differences in the electrostatic properties of the vanadium atoms in the two types of molecules were observed by using electrostatic potential distribution analysis. Methyl group substitutions on the ligand did not change the electronics of the vanadium atoms sufficiently to be a factor. Thus, we conclude that neither electronics nor sterics at the metal center explains the experimental geometries. In the absence of a significant observable electronic effect at the vanadium, we note that the experimental observations can be attributed to ligand geometric effects such as the Thorpe-Ingold and/or gem-dialkyl effects.

Electron correlation is the everlasting concern of people involved in ab initio Hartree-Fock calculations. If the investigation is restricted to the structure and properties of the ground state, the importance of correlation depends i) on the type of bonding in the molecule under scrutiny, and ii) on the level of accuracy requested for the calculation, that is, on the sensitivity of the investigated property. This sensitivity may be highly interrelated with the nature of the bonding: in the specific area of organometallic complexes, it has been recognized since more than a decade that no realistic description of multiple, direct metal-metal bonds can be obtained without an adequate treatment of the left-right correlation.1,2 Accidental near degeneracies connected for instance with the sd hybridization in complexes of Ni(0) should also be accounted for in a systematic way.3 Other cases where correlation is susceptible to qualitatively modify the description of the ground state can be detected from the occurrence of Hartree-Fock instability which makes the energy of the considered system symmetry-dependant.4,5 Introducing a systematic treatment of correlation for large systems in the frame of the Hartree-Fock methodology would lead to technical problems and to an intolerable computer cost. Our experience in the computation of the ground state electronic structure and properties of dimetallic complexes and of clusters with higher nuclearity argues for a flexible approach of the correlation problem.