This research is supported by the NSF theoretical and computational chemistry program. Coupled-cluster and density functional methods in molecular electronic structure theory will be studied and compared to search for an ab-initio-derived-functional for density functional theory. An improved scheme will be explored for computing the triples perturbative correction in Brueckner doubles theory. These computational methods will be applied to the study of transition metal systems and combustion processes. In principle, the physical and chemical properties of molecular systems can be obtained from either the many-body wavefunction (traditional ab initio theory) or from the electronic charge density (density functional theory). There presently exist highly developed computer programs for applying either of these two theories to any of a wide variety of chemical problems. Ab initio equations, particularly those of coupled cluster theory, may require very heavy computations using supercomputers, but these equations are firmly based on well established mathematical and physical principles. The equations of density functional theory offer significant computational advantages, but serious questions remain concerning the precise form of these equations. The present comparative study of ab initio and density functional methods attempts to develop improved computational methods for molecular electronic structure computations which combine the strengths of these two different formalisms. The development of such an improved density functional theory is currently a hot topic in modern theoretical chemistry. A breakthrough in this area will greatly extend the size of molecular systems for which reliable theoretical predictions can be made using modern computers, and will almost certainly find wide and immediate application. The proposed test calculations on transition metal compounds constitute the initial steps toward the rational design of new industrial catalysts.
