This research project, which treats three specific computational methods in molecular electronic structure theory, is supported by the NSF Theoretical and Computational Chemistry Program. Topic (1) extends Pulay's current Unrestricted Natural Orbital Complete Active Space method (UNO-CAS) to large molecules (e.g. molecules with 24 nonhydrogen atoms) and provides a more accurate description of excited electronic states. Topic (2) effectively includes higher-order corrections in low-order Generalized Moller Plesset perturbation theory by relaxing the reference function. Topic (3) employs gauge-invariant atomic orbitals for ab initio quantum mechanical calculation of NMR properties, and explores the use of similar techniques in density functional theory. Laboratory experiments and theoretical predictions, using computer models based on the quantum mechanical equations of motion of the electrons, today contribute as equal partners to the understanding of chemical behavior. The evolution of modern methods for molecular electronic structure computation has had a major impact on virtually every field of chemistry from the study of small molecules in the earth's atmosphere to large molecules of biochemical importance. The development of these mathematical methods and their implementation into large scale computer programs is the result of hundreds of man years of effort. This project continues this tradition of developing practical, widely applicable, theoretical tools. The goal of this project is to improve the speed and accuracy of molecular electronic structure computations, to extend these to larger molecules, to treat excited electronic states (important for optical properties and certain classes of chemical reactions), and to better understand fundamental relationships between molecular structure and observed nuclear magnetic resonance (NMR) spectra.
