TOWARD A SYSTEMATIC MOLECULAR-ORBITAL THEORY FOR EXCITED-STATES

This work reviews the methodological and computational considerations necessary for the determination of the ab initio energy, wave function, and gradient of a molecule in an electronically excited state using molecular orbital theory. In particular, this paper reexamines a fundamental level of theory which was employed several years ago for the interpretation of the electronic spectra of simple organic molecules: configuration interaction (CI) among all singly substituted determinants using a Hartree-Fock reference state. This investigation presents several new enhancements to this general theory. First, it is shown how the "CI-singles" wave function can be used to compute efficiently the analytic first derivative of the energy in order to obtain accurate properties and optimized geometries for a wide range of molecules in their excited states. Second, a computer program is described which allows these computations to be done in a "direct" fashion, with no disk storage required for the two-electron repulsion integrals. This allows investigations of systems with large numbers of atoms (or large numbers of basis functions). Third, it is shown how the CI-singles approximation can be corrected via second-order Moller-Plesset perturbation theory to produce a level of theory for excited states which further includes some effects of electronic correlation. The relative success of the model as a function of basis set indicates that a judicious choice of basis set is needed in order to evaluate its performance adequately. Application of the method to the excited states of formaldehyde, ethylene, pyridine, and porphin demonstrates the utility of CI-singles theory.