| Transition metal cluster research has attracted interest due to the importance of clusters and surfaces in catalysis. Basic research on small clusters is vital because interactions between metal atoms in clusters are not completely understood. Theory and experiment both contribute, with theory helping to interpret and understand experimental data. Theoretical methods that model electronic spectroscopy are crucial, as the data on small clusters is largely spectroscopic. Spectral interpretation is often difficult, so methods that predict state symmetries, transition energies and intensities, and spin-orbit effects play a key role. Unfortunately there are no methods for treating the spectroscopy of medium-sized clusters, as even diatomics pose difficulties for quantum chemistry. There is a need for methods applicable to larger clusters.;The primary difficulty in applying quantum chemistry to metal clusters is electron correlation, or the effect the motions electrons undergo to avoid each other has on energies and wavefunctions. Correlation has a major impact on the electronic structure of metal clusters, so diatomics are studied with highly-correlated ab initio methods, which are limited due to computational cost. Inexpensive semiempirical methods offer an alternative, but have historically assumed a low level of correlation. There are arguments supporting the application of semiempirical methods to clusters, however. The INDO/S model is a candidate because it effectively models the spectroscopy of metal complexes, including spin-orbit effects, and has shown promise with clusters of up to fifty atoms. The role of correlation in semiempirical methods is not well-understood, however, and must be elucidated if methods applicable to larger clusters are to be developed.;This dissertation describes correlated INDO/S calculations on V 2, VNb, and Nb2. Results obtained for the electronic spectroscopy of V2 and VNb serve as a proof of concept and illustrate the role of correlation in the calculations. An examination of ground state spin-orbit effects in the three molecules provides additional insight. The results obtained agree well with experiment. Basic questions left by the spectroscopy have been investigated as well, including identities of unassigned transitions and orbital symmetries of observed transitions. This work demonstrates that semiempirical methods can in principle be successfully applied to metal clusters. |
