Modeling Organorhodium Catalysis

The first project considered in this dissertation was the improvement of an existing global optimization algorithm that uses extended dimensionality to find global minima of Lennard-Jones clusters. The speed of this algorithm was increased by three orders of magnitude, primarily by improving the algorithm for compressing the system from 4D to 3D at constant energy. The second project was modeling the adsorption of H2 molecules on the Si(100) surface using density functional theory (DFT) with the PW91 functional. Consistent with the experiments, the calculations predicted an energetic preference for clustering of occupied sites in a dimer row. However, our calculations did not verify the unbuckling induced by H2 adsorption reported by Buehler and Boland. The third project was modeling molybdenum and rhodium-catalyzed [2 + 2 + 1] cycloisomerization reaction using DFT with the B3LYP functional. We found that in the rhodium-catalyzed [2 + 2 + 1] cycloisomerization reactions of allenes the oxidative addition step determined both the rate and the product of the reaction. For the molybdenum-catalyzed reaction the rate was controlled not by oxidative addition, but by the next step, the attachment of a carbon monoxide molecule from the media to the molybdenum atom. The fourth project was modeling the transfer of hydrogen from one side of the heterocyclic ring to another in rhodium(I) catalyzed allenic Pauson-Khand type reactions. Our calculations showed that this process occurs after the cyclization step. We have also discovered a novel mechanism for this process - hydride transfer; however, we believe that in most cases the reaction proceeds by beta-hydride elimination