*** 9707202 PULAY The Theoretical and Computational chemistry program is supporting Prof. Pulay of the University of Arkansas for work to improve the efficiency of SCF and density functional methods: the multi-Fock technique, and novel basis sets. In another project, further development of the local correlation method is proposed: gradients and a simplified treatment of distant pairs. The latter constitute the majority of electron pairs in large molecules. Density functional theory (DFT), in spite of its convincing success for ground states, is still not applicable to excited states. A project uill explore the use of empirical correlation terms, borrowed from DFT, together with exact exchange, to describe excited states and diradicals. Pulay proposes to build a powerful parallel computer system for quantum chemical calculations, using commodity PC parts. The price of the proposed system is estimated to be about a tenth of a comparable system using workstations. Several applications are proposed: (1) geometries, force fields and vibrational spectra of porphyrins, (2) systematic calculation and parameterization of NMR shifts in peptides as functions of the backbone torsional angles, (3) parameterization of the vibrational force fields of 200 organic molecules, generated under the preceding grant, (4) search for novel stable molecules, such as the cyclic polymers of phosgene and related molecules. The effort during this award is aimed at increasing the efficiency and accuracy of quantum mechanical simulation of molecular systems. Such simulations are likely to lead to fundamental innovations in chemistry, biochemistry and materials science in the future, by allowing the rapid computational determination of the properties of chemical systems and materials. Accurate solutions of the quantum mechanical equations for larger molecules requires very much computer power, and has become possible only recently due to the development of new algorithms and inexpensive fast computers. The proposal lists several projects that, if successful, will increase the efficiency of molecular quantum calculations, or extend the applicability of present methods, for example. to excited electronic states. The work includes the construction of a parallel computer, based on commodity personal computer parts. ***
