Computational Studies of Ziegler-Natta Catalysis and Concurrent Resonance Computations

Citation

Bierwagen, Erik Paul (1995) Computational Studies of Ziegler-Natta Catalysis and Concurrent Resonance Computations. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/3w1m-hk72. https://resolver.caltech.edu/CaltechETD:etd-09112007-104050

Abstract

NOTE: Text or symbols not renderable in plain ASCII are indicated by [...]. Abstract is included in .pdf document. This thesis discusses work on two different subjects. First, results from computational studies of Ziegler-Natta catalysts are presented. Quantum mechanical calculations of model Ziegler-Natta catalysts [...] are described, where X is either Cl or the cyclopentadienyl (Cp) ligand, M is a group three transition metal or group four transition metal cation, and R is a hydrogen or alkyl group. It is found that complexes based on group four cations have pyramidal structures (the R group is not in the [...] plane), whereas group three neutral complexes have planar structures. This difference in structure is considered in the context of syndiodirecting polymerization, leading to the conclusion that the group three metals are expected to show little, if any, syndiodirecting capabilities, while the group four cations, including thorium, are expected to show large syndiodirecting capabilities, in accord with experiments. Results from molecular mechanics simulations of zirconocene-based catalysts follow. These calculations are used to assess the steric demands of different ligand environments during propylene polymerization, to determine the relative importance of site and chain end control on the enantioface selectivity, both of which are found to be operative. The second part of the thesis describes the development of concurrent algorithms for the computation of resonance matrix elements. The algorithms are described for two different models of concurrent computing: parallel and distributed. A general program design and architecture that facilitate the program development are described. The scaling of both algorithms with the number of processors is found to be nearly ideal.

Thesis Files