Classical trajectory calculations have been carried out to simulate the unimolecular decomposition of formaldehyde in the ground electronic state (Se). Global potential-energy surfaces were constructed using the empirical valence-bond (EVB) approach. Two sets of ab initio input were used to characterize two different EVB potential-energy surfaces, and trajectory calculations using one of these gives excellent agreement with experimental data for the product-state distributions of H2 and CO. The trajectory study of vector correlations with prompt dissociation of the parent molecule provides understanding of the dissociation dynamics in the molecular frame. From comparison with some of the experimental results and information from a few ab initio calculations, some improvements for the current potential surfaces are suggested.

A computational theory for determining electron transfer rate constants is formulated based on an instanton expression for the quantum rate and the self-consistent solution of the imaginary time nonadiabatic steepest descent approximation. The theory obtains the correct asymptotic behavior for the electron transfer rate constant in the nonadiabatic and adiabatic cases, and it smoothly bridges between those two limits for intermediate couplings. Furthermore, no assumptions regarding the form of the diabatic potentials are invoked (e.g., harmonic) and more than two diabatic states can be included in the calculations. The method thereby holds considerable promise for computing electron transfer rate constants in realistic condensed phase systems.

A general quantum-mechanical protocol for the study of nonrigid free radicals has been applied to the series CH 3 , CH 2 F, CHF 2 , and CF 3 . Electronic structures have been computed by highly correlated ab initio methods and vibrational modulation effects have been studied by a nonrigid invertor Hamiltonian. The effect of small amplitude vibrations perpendicular to the inversion motion has been taken into account by an adiabatic model. The results are in close agreement with experiment, and can be interpreted guite straightforwardly in terms of the interplay between the potential energy and the property hypersurfaces. This allows a more dynamically based analysis of hyperfine coupling constants extensible to large, nonrigid radicals.

The structure and EPR parameters of dihydronitrosyl radical H2NO were investigated by highly correlated ab-initio methods. The relative stabilities of planar and pyramidal structures were analyzed in detail, taking into account the effect of small-amplitude vibrations perpendicular to the inversion motion. Vibrational averaging of hyperfine coupling consts. was computed by quantum-mech. treatment based on the vibrational adiabatic zero curvature approxn. The general picture emerging from this study, substantiated by several checks, consists in a quasi-planar mol. with a nearly free inversion motion for out-of-plane angles as large as 30°. Due to compensation of different terms, vibrational averaging gives results very close to those obtained from a static treatment at an out-of-plane angle of about 20°. An equally important outcome of this work is the introduction of a general and reliable ab-initio strategy for the study of magnetic properties in nonrigid radicals

In this paper the calculation of the optical rotation (OR) of some rigid organic molecules, using the Hartree-Fock method with small (6-31G, DZP) basis sets, has been studied thoroughly to carefully evaluate the scope and limitations of this method, previously introduced by other authors. Calculations on test molecules (compounds 1-13) together with a careful analysis of their CD spectra allow a simple criterion for the reliability of this approach to be formulated: for unsaturated and/or aromatic (i.e., absorbing in the near-UV region) molecules, if the [alpha](D) is quantitatively determined by the lowest energy Cotton effect (at wavelengths >220 nm), then the HF/6-31G result is reliable. The usefulness of this method for the experimental organic chemist has been further demonstrated because the OR (sign and order of magnitude) of compounds 14-19 (i.e., large molecules having considerable interest in organic chemistry), which fulfill the above criterion and for which an extended basis set treatment is not feasible owing to their size, is correctly predicted.