This grant, in the Theoretical and Computational Chemistry program, provides support for development and application of an approximate quantum-mechanical method for predicting the structures and energies of molecules that are too large to be treated accurately. Professor Dewar's previous work has been of immense practical utility in providing theoretical guidance to experimental organic chemists, and the present project is designed to extend that utility to molecules containing the biologically important elements phosphorus, sulfur, and various metal atoms. Specifically, Professor Dewar's previous work has led to the development of parametric ("semiempirical") quantum mechanical models (MINDO/3, MNDO, AM1) which are unique in providing results of useful accuracy for large organic molecules at reasonable cost in computing time. The present work utilizes improvements in computers to develop a method for treating electron repulsion and core-electron attraction integrals analytically. This method is expected not only to be generally superior to AM1 but also extendable to higher valence states, P(V) being of major interest in biochemistry, and to transition metals.
