Improved Modeling of Transition Metals, Applications to Catalysis and Technetium Chemistry

摘要

There is considerable impetus for identification of aqueous OM catalysts as water is the ultimate ''green'' solvent. In collaboration with researchers at Ames Lab, we investigated effective fragment and Monte Carlo techniques for aqueous-phase hydroformylation (HyF). The Rh of the HyF catalyst is weakly aquated, in contrast to the hydride of the Rh-H bond. As the insertion of the olefin C=C into Rh-H determines the linear-to-branched aldehyde ratio, it is reasonable to infer that solvent plays an important role in regiochemistry. Studies on aqueous-phase organometallic catalysis were complemented in studies of the gas-phase reaction. A Rh-carbonyl-phosphine catalyst was investigated. Two of the most important implications of this research include (a) pseudorotation among five-coordinate intermediates is significant in HyF, and (b) CO insertion is the rate-determining step. The latter is in contrast to experimental deductions, highlighting the need for more accurate modeling. To this end, we undertook studies of (a) experimentally relevant PR{sub 3} co-ligands (PMe{sub 3}, PPh{sub 3}, P(p-PhSO{sub 3{sup -}}){sub 3}, etc.), and (b) HyF of propene. For the propylene research, simulations indicated that the linear: branched aldehyde ratio (linear is more desirable) is determined by thermodynamic discrimination of two distinct pathways. Other projects include a theory-experiment study of C-H activation by early transition metal systems, which establishes that weakly-bound adducts play a key role in activity selectivity. By extension, more selective catalysts for functionalization of methane (major component of natural gas) will require better understanding of these adducts, which are greatly affected by steric interactions with the ligands. In the de novo design of Tc complexes, we constructed (and are now testing) a coupled quantum mechanics-molecular mechanics protocol. Initial research shows it to be capable of accurately predicting structure ''from scratch.'' Challenges include conformational, geometric, coordination, spin, and particularly linkage (e.g., Tc-SCN versus Tc-NCS) isomerism. In general, our protocol can rapidly (<1 day with desktop software/hardware) predict the structure of diverse Tc complexes with an accuracy commensurate to organics. Our de novo strategy is also being used to investigate tris-pyrazolyl borate (Tp) complexes. Data suggests a fundamental difference in methane activation between TpRe and related CpRe complexes. Furthermore, Tp is a more electronically ''flexible'' platform for catalysts modification than Cp.